Embed Size (px)
Citation preview
EUROPEAN COMMISSION EUROSTAT
Eurostat ISP Task Force
Guide on developing an
Index of Services Production (ISP)
2
Members of the Index of Services Production (ISP) Task Force
Expert Country / organisation
Ms Michaela Marsalkova Czech Republic
Mr Søren Kristensen
Mr Søren Kühl Andersen
Denmark [email protected]
Ms Jutta Oertel
Ms Dorothee Blang
Mr Sebastian Metzl
Germany [email protected]
Mr Jean-François Loué
Ms Christelle Minodier
Mr Tony Vuillemin
France [email protected]
Mr Richard McMahon
Ms Marion McCann
Ireland [email protected]
Mr Roberto Iannaccone Italy [email protected]
Mr Tomas Rudys Lithuania [email protected]
Ms Ildikó Hamvaine Holocsy Hungary [email protected]
Mr Wilco de Jong Netherlands [email protected]
Ms Elisabeth Gruber
Ms Sabina Zach
Ms Michaela Lingler
Austria [email protected]
Ms Katarzyna Lipowska
Ms Agnieszka Kaźmierczak
Poland [email protected]
Mr Kari Rautio Finland [email protected]
Ms Pernilla Bengtsson Sweden [email protected]
Mr Steve Drew
Mr John Allcoat
Mr Pete Lee
United Kingdom
Mr Fitwi Wolday
Mr Jakob Kalko
Norway [email protected]
Mr Heinz-Christian Dieden ECB [email protected]
Mr Luis Biedma Eurostat.C.2 [email protected]
Mr Paolo Passerini Eurostat.C.5 [email protected]
Mr August Goetzfried
Ms Sophie Limpach
Mr Hakan Linden
Mr Stefan Schipper
Ms Carmen Lipp-Lingua
Ms Regina Hufnagel
Mr Ulrich Eidmann
Mr Thomas Jaegers
Eurostat.G.3 [email protected]
3
Contents
1. DEVELOPMENT OF THE EUROPEAN SERVICE SECTOR AND SHORT-TERM STATISTICS .................................................................................. 24
2. BACKGROUND AND OUTLINE ........................................................................... 27
2.1. Background ...................................................................................................... 27
2.2. Purpose and organisation of this guide ............................................................ 27
3. GENERAL OUTLINE OF A NEW ISP ................................................................... 28
3.1. Purpose of a new ISP ....................................................................................... 28
3.2. Scope of an ISP ............................................................................................... 29
3.3. Trade services: a special case .......................................................................... 31
3.4. Definition and measurement of outputs .......................................................... 35
3.4.1. Theoretical output concepts and approximations .............................. 35
3.4.2. Measurement of outputs — deflating value data ............................... 36
3.4.3. Measurement of outputs — direct volume measures ........................ 38
3.4.4. Measurement of outputs — (labour) input data................................. 40
3.5. Periodicity of an ISP ........................................................................................ 41
3.5.1. Benefits, cost and burden of monthly turnover data .......................... 42
3.5.2. Need for monthly turnover data — differences between monthly and quarterly results ............................................................ 44
3.5.3. Seasonal patterns in the turnover of services..................................... 45
3.5.4. Monthly turnover data and quarterly deflators .................................. 49
3.5.4.1. Temporal disaggregation methods .................................... 50
3.5.4.2. Practical test of the methods: Polish SPPIs ....................... 54
3.5.4.3. Forecasting of SPPIs.......................................................... 56
3.6. Seasonal adjustment ........................................................................................ 58
3.7. Statistical unit .................................................................................................. 58
3.8. Index construction ........................................................................................... 60
3.8.1. Choice of basic index type................................................................. 60
3.8.2. Level of deflation............................................................................... 60
3.8.3. Weights .............................................................................................. 62
3.9. Finding appropriate price indices for the deflation: B2B versus B2All .......... 62
4. REFERENCES FOR PART A .................................................................................. 64
5. LAND TRANSPORT (H49) ..................................................................................... 67
5.1. Relevant definitions ......................................................................................... 67
5.1.1. Index of services production (ISP) .................................................... 67
5.1.2. Turnover ............................................................................................ 67
5.1.3. Definition of a service ....................................................................... 67
5.2. Industrial classification .................................................................................... 68
5.3. Available data .................................................................................................. 71
5.3.1. STS Regulation .................................................................................. 71
5.3.2. Consumer price indices (CPIs) .......................................................... 72
4
5.3.3. Structural business statistics (SBS) ................................................... 72
5.3.4. Other volume indicators .................................................................... 72
5.4. Methodology .................................................................................................... 73
5.4.1. Statistical unit .................................................................................... 73
5.4.2. Price concept...................................................................................... 73
5.4.3. Passenger transport ............................................................................ 74
5.4.4. Recommended methods for passenger transport ............................... 75
5.4.5. Freight transport ................................................................................ 76
5.4.6. Transport via pipelines ...................................................................... 77
5.5. Compiling an ISP ............................................................................................ 78
5.5.1. Theoretical basis ................................................................................ 78
5.6. References ....................................................................................................... 79
5.7. Annex .............................................................................................................. 81
6. WATER AND AIR TRANSPORT (H50 & H51) .................................................... 83
6.1. Description of the water and air transport sector ............................................. 83
6.2. Available methodological information ............................................................ 83
6.2.1. OECD ISP for an index of service production .................................. 83
6.2.2. The Eurostat-OECD SPPI Guide ....................................................... 84
6.2.3. Voorburg Group ................................................................................ 84
6.3. Available data .................................................................................................. 84
6.3.1. Transport statistics ............................................................................. 84
6.3.2. ‘Foreign trade in services’ statistics .................................................. 84
6.3.3. The case of Denmark ......................................................................... 85
6.3.4. Services turnover and SPPIs in STS .................................................. 85
6.3.5. Structural business statistics (SBS) ................................................... 85
6.3.6. Conclusions ....................................................................................... 85
6.4. Further description of H50 and H51 ................................................................ 85
6.4.1. Air transport ....................................................................................... 89
6.5. Test calculation of an ISP for H50 and H51 .................................................... 89
7. WAREHOUSING AND SUPPORT ACTIVITIES FOR TRANSPORTATION (H52) ..................................................................................... 91
7.1. Definition of an ISP ......................................................................................... 91
7.2. Available methodological information ............................................................ 91
7.3. Activity classification: NACE Rev. 2 H52 ...................................................... 92
7.4. Structure of the industry .................................................................................. 94
7.5. Available data .................................................................................................. 99
7.5.1. Short-term business statistics (STS) .................................................. 99
7.5.2. Structural business statistics (SBS) ................................................. 100
7.5.3. Harmonised indices of consumer prices (HICPs) ........................... 100
7.6. Outline of methodology ................................................................................. 100
7.7. Experimental calculations of ISP .................................................................. 102
8. POSTAL SERVICES (H53) ................................................................................... 104
5
8.1. Introduction ................................................................................................... 104
8.2. Industrial classification .................................................................................. 104
8.3. Available data ................................................................................................ 105
8.3.1. Structural business statistics (SBS) ................................................. 106
8.3.2. National accounts (NAs) ................................................................. 108
8.3.3. Labour market.................................................................................. 109
8.3.4. Short-term statistics (STS) and prices ............................................. 110
8.4. An ISP for postal and courier activities — initial results .............................. 113
9. ACCOMMODATION AND FOOD SERVICES (I55 & I56) ................................ 117
9.1. Available methodological information .......................................................... 117
9.2. Classification issues ....................................................................................... 117
9.2.1. Accommodation services ................................................................. 117
9.2.2. Food and beverage service activities ............................................... 118
9.3. Structural profile of the sector in the EU ....................................................... 118
9.4. Data availability ............................................................................................. 121
9.4.1. Deflated output values ..................................................................... 121
9.4.2. Output volumes ............................................................................... 121
9.4.3. Input indicators ................................................................................ 121
9.4.4. Weights information ........................................................................ 121
9.5. Methodology for calculating an ISP for I55 and I56 ..................................... 122
9.5.1. Preferred (best practice) methodological approach ......................... 122
9.5.2. Alternative methodological approaches .......................................... 122
9.6. Calculation of experimental index for Ireland ............................................... 122
9.7. Key issues ...................................................................................................... 124
9.7.1. Data availability ............................................................................... 124
9.7.2. Secondary activities ......................................................................... 124
9.7.3. Use of CPI deflators ........................................................................ 125
9.7.4. Complementary data sources ........................................................... 125
9.8. Seasonality ..................................................................................................... 126
9.9. Conclusion ..................................................................................................... 126
10. MOTION PICTURE, VIDEO AND TV PROGRAMME PRODUCTION, SOUND RECORDING AND MUSIC PUBLISHING ACTIVITIES (J59) ........... 127
10.1. Available methodological information .......................................................... 127
10.2. Presentation of the industry ........................................................................... 127
10.2.1. Activities by NACE ......................................................................... 129
10.2.2. Secondary activities ......................................................................... 130
10.3. Available data ................................................................................................ 130
10.4. Experimental calculations of ISPs ................................................................. 131
10.5. Methodological issues ................................................................................... 134
10.5.1. Weights ......................................................................................... 134
10.5.2. Choice of deflator ............................................................................ 134
10.5.3. Methodological problems and advice .............................................. 135
6
11. PROGRAMMING AND BROADCASTING (J60) ............................................... 137
11.1. Available methodological information .......................................................... 137
11.2. Presentation of the industry ........................................................................... 137
11.3. Activities by NACE ....................................................................................... 139
11.4. Secondary activities ....................................................................................... 139
11.5. Available data ................................................................................................ 140
11.6. Experimental calculations of ISPs ................................................................. 140
11.7. Methodological issues ................................................................................... 141
11.7.1. Outer and inner weights ................................................................... 141
11.7.2. Choice of deflator ............................................................................ 141
11.7.3. Methodological problems and advice .............................................. 142
12. TELECOMMUNICATIONS (J61) ......................................................................... 144
12.1. Methodological information .......................................................................... 144
12.1.1. Reports from national and international organisations .................... 144
12.1.2. Statistical guides and manuals, statistical papers ............................ 146
12.2. Presentation of the industry ........................................................................... 148
12.2.1. Industry overview and trends .......................................................... 148
12.2.2. Classification aspects ....................................................................... 149
12.2.2.1. Classification by activity ................................................. 149
12.2.2.2. Classification by product ................................................. 152
12.3. Available data ................................................................................................ 154
12.3.1. Short-term statistics (STS) .............................................................. 154
12.3.2. Harmonised indices of consumer price (HICPs) ............................. 155
12.3.3. Structural business statistics (SBS) ................................................. 156
12.3.4. ICT statistics .................................................................................... 159
12.4. Experimental calculations of ISPs ................................................................. 159
12.4.1. Objectives of the calculation ........................................................... 159
12.4.2. Special characteristics of the telecommunications industry ............ 160
12.4.3. Input variables/deflators .................................................................. 161
12.4.4. Recommendations for calculation of an ISP ................................... 164
12.4.5. Test calculations .............................................................................. 164
12.5. Results of test calculations ............................................................................ 165
12.6. Conclusions and further recommendations ................................................... 167
13. COMPUTER PROGRAMMING, CONSULTANCY AND RELATED ACTIVITIES (J62) .................................................................................................. 168
13.1. Available methodological information .......................................................... 168
13.2. Presentation of the industry ........................................................................... 168
13.2.1. Turnover development for J62 ........................................................ 170
13.2.2. Service producer price indices (SPPIs) ........................................... 172
13.2.3. Number of persons employed .......................................................... 174
13.3. Available data ................................................................................................ 174
13.4. Experimental calculation of an ISP indicator ................................................ 175
7
13.4.1. Methodological considerations ........................................................ 175
13.4.2. Calculation of an experimental ISP indicator .................................. 176
13.4.2.1. Total output is basically homogeneous ........................... 176
13.4.2.2. Total output includes significant share of secondary activities ........................................................................... 178
13.5. References, classifications, and available data .............................................. 180
14. INFORMATION SERVICE ACTIVITIES (J63) ................................................... 184
14.1. Available methodological information .......................................................... 184
14.2. Presentation of the industry ........................................................................... 184
14.3. Turnover ........................................................................................................ 187
14.4. Service producer price indices (SPPIs) ......................................................... 188
14.5. Number of persons employed ........................................................................ 190
14.6. Available data ................................................................................................ 190
14.7. Experimental calculation of an ISP indicator ................................................ 190
14.7.1. Methodological considerations ........................................................ 190
14.7.2. Calculation of an experimental ISP indicator (for homogeneous and heterogeneous outputs) ...................................... 191
14.8. Summary ........................................................................................................ 196
14.9. References, classifications, available data ..................................................... 196
15. REAL ESTATE (L68) ............................................................................................. 200
15.1. Overview of activities in division L68 (real estate) ...................................... 200
15.1.1. NACE Rev. 2 ................................................................................... 200
15.1.2. The French classification (NAF rev. 2) ........................................... 202
15.2. The importance of real estate activities in the services sector ....................... 202
15.3. Method for calculating an ISP ....................................................................... 206
15.3.1. ISP 68.2: Renting and operating of own or leased real estate ......... 212
15.3.2. ISP 68.3: Real estate activities on a fee or contract basis ............... 219
15.3.3. ISP of the real estate division .......................................................... 223
15.4. Conclusion ..................................................................................................... 227
15.5. References ..................................................................................................... 227
16. LEGAL, ACCOUNTING AND MANAGEMENT SERVICES (M69_M702) ........................................................................................................... 228
16.1. Available methodological information .......................................................... 228
16.2. Description of the industry ............................................................................ 228
16.3. Activities by NACE ....................................................................................... 230
16.4. Secondary activities ....................................................................................... 231
16.5. Available data ................................................................................................ 231
16.6. Experimental calculations of an ISP .............................................................. 232
16.7. Methodological issues ................................................................................... 233
16.7.1. Outer and inner weights ................................................................... 233
16.7.2. Choice of deflator ............................................................................ 234
16.7.3. Methodological problems and advice .............................................. 234
8
17. ARCHITECTURE, ENGINEERING, TESTING (M71) ........................................ 236
17.1. Available methodological information .......................................................... 236
17.2. Presentation of the industry ........................................................................... 237
17.3. Economic classifications ............................................................................... 239
17.3.1. NACE Rev. 2 ................................................................................... 239
17.3.2. ISIC Rev. 4 ...................................................................................... 240
17.3.3. National economic activity classifications ...................................... 240
17.3.4. Typical products .............................................................................. 241
17.4. Industry description ....................................................................................... 242
17.5. Development .................................................................................................. 244
17.6. Available data ................................................................................................ 247
17.6.1. STS data ......................................................................................... 247
17.6.2. SBS data ......................................................................................... 249
17.6.3. Further data from other statistics ..................................................... 250
17.7. Typical business / pricing models ................................................................. 250
17.8. Experimental calculations of the ISP ............................................................ 254
17.8.1. Building elementary indices for the ISP .......................................... 254
17.8.2. Aggregation of elementary ISP data ................................................ 257
17.9. Issues in calculating the ISP .......................................................................... 259
17.10. Annexes ........................................................................................... 259
17.10.1. Sector papers and other sources....................................................... 259
18. OTHER PROFESSIONAL, SCIENTIFIC AND TECHNICAL ACTIVITIES (M74) ....................................................................................................................... 263
18.1. General description of the industry ............................................................... 263
18.2. Data availability ............................................................................................. 263
18.3. Industry profile .............................................................................................. 264
18.3.1. General overview ............................................................................. 264
18.3.2. Sectoral analysis .............................................................................. 265
18.3.3. Development .................................................................................... 266
18.3.4. Enterprise size.................................................................................. 267
18.3.5. Country analysis .............................................................................. 267
18.4. ISP calculation ............................................................................................... 268
19. RENTAL AND LEASING ACTIVITIES (N77) .................................................... 271
19.1. Available methodological information .......................................................... 271
19.2. Presentation of the rental and leasing service industry ................................. 272
19.2.1. Classification ................................................................................... 272
19.2.2. Description ...................................................................................... 274
19.3. Data availability at national and European level ........................................... 277
19.4. An experimental ISP for NACE 77 ............................................................... 278
20. EMPLOYMENT ACTIVITIES (N78) .................................................................... 280
20.1. Classification issues ....................................................................................... 280
20.2. Sectoral analysis ............................................................................................ 281
9
20.3. Available methodological information .......................................................... 283
20.3.1. Voorburg Group .............................................................................. 283
20.3.2. OECD ISP Manual .......................................................................... 284
20.3.3. National practices ............................................................................ 285
20.3.4. Further references and data sources................................................. 285
21. TRAVEL AGENCY, TOUR OPERATOR RESERVATION SERVICE AND RELATED ACTIVITIES (N79) .................................................................... 287
21.1. Classification issues and content of division N79 ......................................... 287
21.2. Available methodological information .......................................................... 287
21.2.1. OECD ISP Manual .......................................................................... 287
21.2.2. Voorburg Group .............................................................................. 288
21.3. Available data ................................................................................................ 288
21.3.1. Services turnover and SPPIs in STS ................................................ 288
21.3.2. Consumer price indices (CPIs) ........................................................ 288
21.3.3. Structural business statistics (SBS) ................................................. 288
21.4. Description of N79 ........................................................................................ 288
21.4.1. Difference between agencies and operators .................................... 289
21.5. Calculating an ISP for N79 ............................................................................ 289
21.5.1. Methodology .................................................................................... 289
21.5.2. Business models in N79 .................................................................. 290
21.5.3. Deflators for N79 ............................................................................. 290
21.5.4. Test calculations .............................................................................. 292
21.6. Conclusions ................................................................................................... 294
22. SECURITY AND INVESTIGATION ACTIVITIES (N80) .................................. 295
22.1. Summary ........................................................................................................ 295
22.2. Introduction ................................................................................................... 296
22.3. Methodological information .......................................................................... 296
22.4. Presentation of the industry ........................................................................... 296
22.4.1. Industry overview and trends .......................................................... 296
22.4.2. Classification issues — classification by activity............................ 297
22.4.3. Classification issues — classification by product ........................... 298
22.5. Available data ................................................................................................ 300
22.5.1. Short-term statistics (STS) .............................................................. 300
22.5.2. Harmonised consumer price indices (HICPs) ................................. 301
22.5.3. Structural business statistics (SBS) ................................................. 301
22.5.4. National accounts ............................................................................ 304
22.6. Experimental calculations of ISPs ................................................................. 304
22.6.1. Objectives of the calculation ........................................................... 304
22.6.2. Special characteristics of the security and investigation industry ......................................................................................... 305
22.6.3. Input variables/deflators .................................................................. 305
22.7. Detailed description of ISP calculation ......................................................... 309
22.7.1. Possible methods ............................................................................. 309
10
22.7.2. Description of test calculations ....................................................... 310
22.7.3. Results — ISPs for N80 .................................................................. 311
22.8. Conclusions and further recommendations ................................................... 313
23. SERVICES TO BUILDINGS AND LANDSCAPE ACTIVITIES (N81) ............. 315
23.1. Introduction ................................................................................................... 315
23.2. Services to buildings and landscape activities in Finland ............................. 316
23.3. Methodological considerations ...................................................................... 316
23.3.1. Statistical units ................................................................................. 317
23.3.2. Deflation method ............................................................................. 317
23.3.3. Weights used for the index construction ......................................... 318
23.4. References ..................................................................................................... 318
24. OFFICE ADMINISTRATION (N82) ..................................................................... 319
24.1. Available methodological information .......................................................... 319
24.2. Presentation of the industry ........................................................................... 320
24.2.1. Classification ................................................................................... 320
24.2.2. Description ...................................................................................... 321
24.2.3. Data and facts .................................................................................. 322
24.3. Available data ................................................................................................ 325
24.4. Experimental ISP ........................................................................................... 326
11
List of figures
Figure 1: Growth of European gross value added, 2000-2014 ...................................... 24
Figure 2: Growth rates (m/m) for nights in hotels and similar accommodation and turnover (index: 2010 = 100), Portugal (unadjusted data) ....................................... 40
Figure 3: Hours worked and production indices for the European food manufacturing industry ..................................................................................................................... 41
Figure 4: Turnover indices for services industries – average difference between months and their respective quarters (selected countries) ..................................................... 45
Figure 5: Seasonal pattern for turnover in accommodation services (I55) — summer peak ........................................................................................................................ 47
Figure 6: Seasonal pattern for turnover development in advertising and market research (M73) — ‘silly season’ ............................................................................................. 48
Figure 7: Seasonal pattern for turnover development in other professional, scientific and technical services (M74) — ‘hockey-stick’ effect ................................................... 49
Figure 8: Discrete interpolation — stylised example ..................................................... 51
Figure 9: Interpolated and re-scaled index — stylised example .................................... 52
Figure 10: Linear interpolation and cubic spline interpolation (simple and re-scaled) — stylised example ....................................................................................................... 53
Figure 11: Consumer prices and SPPIs for German air transport services (2010=100) ...... ........................................................................................................................ 54
Figure 12: Test of disaggregation methods — turnover in the Polish telecommunications industry ..................................................................................................................... 55
Figure 13: Test of disaggregation methods — turnover in the Polish postal services industry ..................................................................................................................... 55
Figure 14: Estimation of deflators for January, February and March .............................. 57
Figure 15: The market for air transport ............................................................................ 89
Figure 16: Experimental ISPs for water and air transport (EU-28) ................................. 90
Figure 17: Division H52 – value added at factor cost, number of persons employed and number of employees vis-à-vis total business economy, by country (%), 2010 .......... ........................................................................................................................ 96
Figure 18: Division H52 – turnover or gross premiums (EUR million and as % of total business economy), 2010 ......................................................................................... 97
Figure 19: Division H52 – number of enterprises, by country, 2010 .............................. 98
Figure 20: Turnover, SPPI and ISP for H52.1 ............................................................... 103
12
Figure 21: Turnover, SPPI and ISP for H52.24 ............................................................. 103
Figure 22: STS turnover data for postal and courier activities ...................................... 111
Figure 23: Postal and courier activities – STS SPPI data .............................................. 112
Figure 24: ISP for postal and courier activities – % change on same quarter of previous year (method A) ...................................................................................................... 114
Figure 25: ISP for postal and courier activities – % change on same quarter of previous year (method B) ...................................................................................................... 114
Figure 26: ISP for postal and courier activities — methods A and B ............................ 115
Figure 27: ISP for postal and courier activities — % change on same quarter of previous year (methods A and B) .......................................................................................... 115
Figure 28: NACE section I (accommodation and food service activities) –value added and employment as proportions of non-financial business economy total, 2009 (%).. ...................................................................................................................... 120
Figure 29: Accommodation and food and beverage sector, Ireland – monthly value and volume indices (Jan 2010-Feb 2013; base 2010=100) ........................................... 123
Figure 30: Distribution of total turnover in hotels in Ireland by service activity, 2007 ...... ...................................................................................................................... 124
Figure 31: NACE J50 (motion picture and sound recording activities) — value added and persons employed as proportions of non-financial service economy totals, 2012 (%) ...................................................................................................................... 128
Figure 32: NACE 59.1 (motion picture, video and TV programme activities) – ISP with different deflators, y-o-y change (2009-2013) ....................................................... 132
Figure 33: NACE 59.2 (sound recording and music publishing activities) – ISP with different deflators, y-o-y change (2009-2013) ....................................................... 132
Figure 34: NACE 59.14 (motion picture projection activities) – ISP (2009-2013) ............. ...................................................................................................................... 133
Figure 35: NACE 59 (motion picture, video and TV programme production, sound recording and music publishing activities) – ISP with different deflators, y-o-y change (2009-2013) ................................................................................................ 133
Figure 36: NACE 59 (motion picture, video and TV programme production, sound recording and music publishing activities) – indices of deflators (2005-2013) ........... ...................................................................................................................... 135
Figure 37: NACE J60 (programming and broadcasting activities) – value added and persons employed as proportions of non-financial service economy totals, 2012 (%) ...................................................................................................................... 138
Figure 38: NACE 60 (programming and broadcasting activities) – ISP with different deflators, y-o-y change (2009-2013) ...................................................................... 141
13
Figure 39: Telecommunications – turnover distribution by NACE Rev. 2 group, 2011–2013 (%, EU-28) .................................................................................................... 159
Figure 40: Structure of the telecommunications industries ........................................... 160
Figure 41: SA and WDA series, not deflated (TOVT) and deflated (TOVV), EU-28 – method 1 ................................................................................................................. 166
Figure 42: SA and WDA series, not deflated (TOVT) and deflated (TOVV), EU-28 – method 2 ................................................................................................................. 166
Figure 43: Comparison of methods 1 and 2 for EU-28 SA and WDA .......................... 167
Figure 44: Division J62 – turnover by Member State as proportions of the EU average, EU-27 (%) .............................................................................................................. 171
Figure 45: Computer programming, consultancy (J62) – annual SPPIs (2010 = 100) ........ ...................................................................................................................... 173
Figure 46: J62 – index of turnover and experimental ISP, EU-27 (2010=100) ............. 177
Figure 47: J62 – index of turnover and experimental ISP, euro area (EA-17) (2010=100) ...................................................................................................................... 177
Figure 48: J62 – index of turnover and experimental ISP(I), Hungary (2010=100) ........... ...................................................................................................................... 178
Figure 49: J62 – index of turnover at current prices and experimental ISP (II-1), Hungary (2010=100) ............................................................................................................. 179
Figure 50: J62 – index of turnover at current prices and experimental ISP (II-2), Hungary (2010=100) ............................................................................................................. 180
Figure 51: Division J63 — turnover by Member State as a proportion of EU average, EU-27 (%), 2012 .................................................................................................... 187
Figure 52: Information service activities (J63) – total output price indices (SPPIs), 2013 (2010 = 100) ........................................................................................................... 189
Figure 53: Index of turnover at current prices (TOVTJ63) and experimental ISP (TOVVJ63), EU-27 (2010 = 100) .......................................................................... 192
Figure 54: Index of turnover at current prices (TOVTJ63) and experimental ISP(I)J63 (TOVVJ63), Hungary (2010 = 100) ....................................................................... 192
Figure 55: Division J63 – total output price index and deflators, Hungary (2010 = 100) ... ...................................................................................................................... 193
Figure 56: Index of turnover (TOVTJ63) and experimental ISP(II)J63 (TOVV- Def1), Hungary (2010 = 100) ............................................................................................ 194
Figure 57: Index of turnover (TOVTJ63) and experimental ISP(II)J63 (TOVV- Def2), Hungary (2010 = 100) ............................................................................................ 195
14
Figure 58: Index of turnover at current prices (TOVTJ63) and experimental ISP(II)J63 (TOVV- Def3), Hungary (2010 = 100) .................................................................. 195
Figure 59: Proportion of enterprises in real estate activities (%) ................................... 203
Figure 60: Real estate services – proportion of value added in the total service sector, by country (%) ............................................................................................................. 204
Figure 63: Monthly old dwelling price .......................................................................... 210
Figure 64: Turnover, price and volume indices for class 68.10 ..................................... 211
Figure 65: SA-WDA indices obtained by the two methods ........................................... 212
Figure 66: Development of price indices for real estate activities ................................. 213
Figure 67: Turnover, price and volume indices for class 68.20 ..................................... 214
Figure 68: SA-WDA indices obtained by the two methods ........................................... 214
Figure 69: Turnover, price and volume indices, sub-class 68.20A ................................ 215
Figure 70: SA-WDA indices obtained by the two methods ........................................... 216
Figure 71: Turnover, price and volume indices, sub-class 68.20B ................................ 216
Figure 72: SA-WDA indices obtained by the two methods ........................................... 217
Figure 73: SA-WDA indices obtained by the two methods ........................................... 218
Figure 74: Comparison of SA-WDA volume indices (Method 1) obtained by the two approaches .............................................................................................................. 218
Figure 75: Development of price indices ....................................................................... 219
Figure 76: Turnover, price and volume indices, class 68.31 ......................................... 220
Figure 77: SA-WDA indices obtained by the two methods ........................................... 220
Figure 78: Turnover, price and volume indices, class 68.32 ......................................... 221
Figure 79: SA-WDA indices obtained by the two methods ........................................... 221
Figure 80: Gross volume index of group 68.3 ............................................................... 222
Figure 81: SA-WSA indices obtained by the two methods ........................................... 223
Figure 82: Gross volume index, division 68 .................................................................. 224
Figure 83: SA-WDA volume index obtained by the two methods ................................ 225
Figure 84: Volume indices, groups 68.1, 68.2 and 68.3 and division 68 ....................... 226
Figure 85: Real estate – Development of ISP and turnover index ................................. 226
15
Figure 86: NACE M69 and M70.2 (legal, accounting and management services) – value added and persons employed as proportions of non-financial service economy totals, 2012 (%) ................................................................................................................. 229
Figure 87: Legal activities (NACE 69.1) – ISP, SPPI and turnover, year-on-year change (2007-2014) ............................................................................................................ 232
Figure 88: Accounting, bookkeeping, auditing activities and tax consultancy (NACE 69.2) – ISP, SPPI and turnover, year-on-year change (2007-2014) ....................... 232
Figure 89: Management consultancy activities (NACE 70.2) – ISP, SPPI and turnover, year-on-year change (2007-2014) .......................................................................... 233
Figure 90: Legal, accounting and management services (NACE M69 and M70.2) – ISP, SPPI and turnover, year-on-year change (2007-2014) ........................................... 233
Figure 91: Division M71 – turnover index, EU-28 and Germany (calendar and SA) (2010=100) ............................................................................................................. 245
Figure 92: Division M71 – turnover index, Germany (original data) (2010=100) .............. ...................................................................................................................... 246
Figure 93: Division M71 – employment index, Germany (2010=100) ......................... 246
Figure 94: Producer price index for architectural and engineering services, and technical testing and analysis (2010=100) ............................................................................. 254
Figure 95: M71.11 (architectural activities) – price index, turnover ratios and experimental ISP (2010=100) ................................................................................ 255
Figure 96: M71.12 (engineering activities and related technical consultancy) – price index, turnover ratios and experimental ISP (2010=100) ...................................... 255
Figure 97: M71.2 (technical testing and analysis) – price index, turnover ratios and experimental ISP (2010=100) ................................................................................ 256
Figure 98: M71.11 – comparison of B2B and B2All volume and producer price indices (2010=100) ............................................................................................................. 256
Figure 99: M71.1 (architectural and engineering activities and related technical consultancy) – price index, turnover ratios and experimental ISP (2010=100) ..... 257
Figure 100: M71 (architectural and engineering activities; technical testing and analysis) – price index, turnover ratios and experimental ISP (2010=100) .......................... 258
Figure 101: M71 (architectural and engineering activities; technical testing and analysis – price index, calendar and SA turnover ratios and experimental ISP (2010=100) .. 258
Figure 102: Divisions in section M – shares in number of enterprises/turnover, EU-28 (2012) ..................................................................................................................... 264
Figure 103: Section M – shares of gross operating surplus, value added and personnel costs, by division, EU-28 (2012) ............................................................................ 265
16
Figure 104: Division M74 – sectoral analysis, EU-27 (2010) ...................................................................................................................... 265
Figure 105: Division M74 – seasonally adjusted turnover index EU-27 (%, 2010=100) ...................................................................................................................... 266
Figure 106: Division M74 – seasonally adjusted turnover index, selected countries (%, 2010=100) .............................................................................................................. 266
Figure 107: Division M74 – breakdown by enterprise size, EU-27 (2012) ...................................................................................................................... 267
Figure 108: Division M74 – value added, by country, EU-28 (2010) ...................................................................................................................... 268
Figure 109: Deflated indices (2008=100) for M74 (step 4) ...................................................................................................................... 269
Figure 110: Seasonally adjusted deflated indices (2008=100) for M74 (step 5) ...................................................................................................................... 270
Figure 111: Unadjusted and adjusted ISPs for M74, Poland (2008=100) ...................................................................................................................... 270
Figure 112: Division N77 – economic indicators by country ...................................................................................................................... 276
Figure 113: Division 77 – value added in relation to section N total (%) ...................................................................................................................... 277
Figure 114: NACE 77 – monthly turnover, Czech Republic ...................................................................................................................... 278
Figure 115: NACE 77 – experimental ISP (y-o-y indices) ...................................................................................................................... 278
Figure 116: N78 – development of turnover index (EU-28) ...................................................................................................................... 281
Figure 117: EU-28 turnover index, annual change ...................................................................................................................... 282
Figure 118: Quarterly ISP for division N79 (EU-28) ...................................................................................................................... 293
Figure 119: Quarterly ISP for division N79 (EU 28), SA ...................................................................................................................... 293
Figure 120: Division N80 – turnover distribution by group, 2011–2013, EU-28 (%) ...................................................................................................................... 304
Figure 121: Security and investigation industries – structure ...................................................................................................................... 305
17
Figure 122: Method 1 – SA and WDA series, not deflated (TOVT) and deflated (TOVV), EU-28 ..................................................................................................................... 312
Figure 123: Method 2 – SA and WDA series, not deflated (TOVT) and deflated (TOVV), EU-28 ..................................................................................................................... 312
Figure 124: Comparison between methods 1 and 2 for the EU-28 ...................................................................................................................... 313
Figure 125: N81 – turnover and volume index, Finland ...................................................................................................................... 317
Figure 126: Division 82 – value added (% of section N) ...................................................................................................................... 323
Figure 127: N82 – economic indicators by country ...................................................................................................................... 324
Figure 128: NACE 82 – quarterly turnover (EU-28) ...................................................................................................................... 325
Figure 129: SPPIs for N82, N74 and market services (2010 = 100) ...................................................................................................................... 326
Figure 130: N82 – experimental ISP, Czech Republic (2010 = 100) ...................................................................................................................... 327
Figure 131: N82 – experimental ISP, France (2010 = 100) ...................................................................................................................... 328
18
List of tables
Table 1: STS indicators — overview 26
Table 2: Potential inputs for deflators for G45 and G46 32
Table 3: Burden of turnover collection 43
Table 4: Cost of turnover collection 43
Table 5: Test for seasonality, assuming stability (grey cells — test found no seasonal pattern) 46
Table 6: Variance of quarterly turnover and quarterly SPPIs (arithmetic averages for countries for which data are available), 2010–2014 50
Table 7: Availability of turnover data and SPPIs 56
Table 8: Weighting B2B SPPIs and HICP — simplified example 64
Table 9: Land transport – correspondence between ISIC Rev. 4 and NACE Rev. 2 68
Table 10: Structure of H49 in NACE Rev. 2 69
Table 11: The road transport sector in the EU 81
Table 12: CPA 2008 — land and pipeline transport services 81
Table 13: NACE structure for H50 and H51 83
Table 14: Water and air transport — turnover by sector (mio. EUR), 2011-2013 (EU-28) 86
Table 15: Water and air transport — number of enterprises, 2011-2012 (EU-28) 86
Table 16: World merchant fleet by country of operator (April 2015) 88
Table 17: Methodological information on H52 from the Voorburg Group 92
Table 18: Warehousing and transport support activities in ISIC and CPA 93
Table 19: Size of group H52.2. in the EU 94
Table 20: Macroeconomic indicators of warehousing and support activities for transportation (EU-27), 2010 95
Table 21: Recommended variables and deflators and their sources 102
Table 22: Input variables for an ISP for industry H52 102
Table 23: NACE section H (transport and storage) 105
Table 24: Enterprises in the postal services sector (H53) as a percentage of total enterprises in the transport and storage division (H) 107
19
Table 25: Demographic and economic indicators for postal and courier activities (Italy) 108
Table 26: Postal and courier activities (ESA95) – households’ final consumption expenditure 109
Table 27: Index of wages under collective labour agreements per employee (base December 2010=100), Italy 109
Table 28: Postal and courier activities – turnover in services (annual WDA data; base 2010 = 100) 111
Table 29: Postal and courier activities – service producer prices (annual data; 2010 = 100) 112
Table 30: COICOP 081 postal services – HICP (annual data; 2005 = 100) 113
Table 31: Accommodation and food services – main economic indicators 119
Table 32: NACE Rev. 2 for J59 129
Table 33: Share of turnover by activity in the Netherlands, annual statistics, 2010 129
Table 34: Secondary activities by NACE in the Netherlands, annual statistics, 2010 (estimated %) 130
Table 35: Outer weights based on value added in the Netherlands, 2010 134
Table 36: Breakdown of turnover by activities in the Netherlands, 2010 139
Table 37: Percentage of secondary activities by NACE in the Netherlands, annual statistics, 2010 (estimates) 139
Table 38: IDI for EU Member States and other selected countries (2013 and 2012) 145
Table 39: Development of Austrian telecommunications turnover to final customers (2012–2014) 146
Table 40: Division 61, telecommunications (CPA 2008) 152
Table 41: Telecommunications – number of enterprises, turnover and value added at factor cost, EU-28 (2011–2013) 157
Table 42: Telecommunications – turnover distribution by NACE Rev. 2 group, 2013 (rounded % values, preliminary) 158
Table 43: Overview of main input variables for a telecommunications ISP 161
Table 44: Options for the calculation of missing input variables 163
Table 45: Methods for calculating an ISP for J61 164
Table 46: Division J62 – main economic indicators for as proportions of the total business economy (except financial and insurance activities), 2012 170
20
Table 47: Case 1 – estimated turnover structure and deflators, division J62 (main and secondary activity) 178
Table 48: Case 2 – estimated turnover structure and deflators, division J62 (main and secondary activity) 179
Table 49: CPA 2008 structure for J62 183
Table 50: Division J63 – main economic indicators for as proportions of business economy totals (except financial and insurance activities), 2012 186
Table 51: Division J63 – estimated turnover structure and deflators (main activity and secondary activities) 194
Table 52: NACE Rev. 2 – section J (Information and communication) 198
Table 53: CPA 2008 – section J (Information and communication services) 199
Table 54: Structure of real estate industry, France and EU as a whole 203
Table 55: Value added per person in the real estate sector, by country 205
Table 56: Distribution of value added in real estate in European countries (%) 206
Table 57: Availability of turnover and price indices for real estate activities 207
Table 59: Weights (value added) used for L68.20 215
Table 60: Solutions depending on availability of price indices 227
Table 61: Solutions depending on seasonality of price series 227
Table 62: NACE Rev. 2 — division M69 and group M702 230
Table 63: Share of turnover by activities in the Netherlands, 2010 (%) 230
Table 64: Secondary activities by NACE in the Netherlands, 2010 (%, estimates) 231
Table 65: Outer weights based on value added (EUR million), 2010 234
Table 66: NACE 69.2 – inner weights based on turnover (EUR million), 2010 234
Table 67: Division M71 – number of enterprises, EU-27 (2010) 243
Table 68: Division M71 – mean value added at factor cost per person employed in the EU, 2010 244
Table 69: Division M71 – turnover breakdown (SBS), % 247
Table 70: Correspondence of classifications (CPC Rev.2, ISIC Rev. 4, NACE Rev. 2 and CPA 2008) 262
Table 71: M74 – deflators used for the production of an ISP 268
Table 72: SBS weights used for compiling an index for M74 (Poland) 269
21
Table 73: NACE Rev. 2 N77 (rental and leasing activities) 272
Table 74: CPA 2008 for rental and leasing activities 273
Table 75: NACE 77 – economic indicators (EU-28) 275
Table 76: Deflators used for experimental ISP for N77 279
Table 77: N78 – main indicators 282
Table 78: Structure of N79 (travel agencies, tour operator and other reservation services and related activities), EU-28, 2011 289
Table 79: CPA 2008 for N79 291
Table 80: CPA 2008 for division N80 (security and investigation services) 299
Table 81: Division N80 – number of enterprises, turnover and value added at factor costs for security and investigation activities (2011-2013) 302
Table 82: Division N80 – turnover distribution by group, 2013 (rounded % values, preliminary) 303
Table 83: N80 – overview of main input variables for an ISP 306
Table 84: Problems and possible solutions 308
Table 85: ISP calculation methods 309
Table 86: Division 82 – NACE Rev. 2 classification and NACE Rev. 1 correspondence 320
Table 87: CPA 2008 for office administration services 321
Table 88: NACE N82 – economic indicators (EU-28) 323
22
List of abbreviations
API Agricultural price index
ARIMA Autoregressive integrated moving average model
B2B Business to business (transactions, sales)
B2C Business to consumers (transactions, sales)
CA Calendar adjusted
CAO Collectively agreed wages
COICOP Classification of individual consumption by purpose
CPA Classification of products by activity
CPC Central product classification
CPI Consumer price index
DNS Domain name system
ESA 1995 European System of National Accounts 1995
FRIBS Framework Regulation for the integration of business statistics
HICP Harmonized index of consumer prices
IMF International Monetary Fund
IPI Industrial production index
IPPI Industrial producer price index
ISIC International standard industrial classifcation
ISP Index of services production
ITU International telecommunications union
NA National accounts
NACE Statistical classification of economic activities in the European Community
n.e.c. Not elsewhere classified
NSI National statistical institute
OECD Organiation for economic cooperation and development
ONS Office for National Statistics (UK)
PEEI Principal European economic indicator
PPI Producer price index
PRBB Producer price, business to business (Code for data deliveries to Eurostat)
PRON Producer price, business to all (Code for data deliveries to Eurostat)
SA Seasonally adjusted
SBS Structural business statistics
SIC Standard industrial classification
SME Small and medium sized enterprises
SNA System of national accounts
23
SPPI Service producer price index
STESEG Short-term Economic Statistics Expert Group (OECD)
STISTF Short-term indicators for services task force (OECD)
STS Short-term business statistics
TOVT Turnover nominal (Code for data deliveries to Eurostat)
TOVV Turnover deflated (Code for data deliveries to Eurostat)
USO Universal service obligation (postal services)
USP Universal service provider (postal services)
WDA Working day adjusted
24
PART A GENERAL METHODOLOGICAL INFORMATION
1. DEVELOPMENT OF THE EUROPEAN SERVICE SECTOR AND SHORT-TERM
STATISTICS
European short-term business statistics (STS) provide a comprehensive set of production, turnover, producer (output) price and labour indicators for European industry, construction, trade and market service industries.
STS are used by the European Central Bank and national central banks, the European Commission, national governments, research institutes, businesses and business organisations. STS also provide important input for other statistical areas, such as national accounts (NAs). Economic trend analysis, forecasting and modelling are the most common uses. The data are also required by policy-makers and businesses, for scientific research and to check and validate data from other sources.
Users demand that STS give a picture of the business economy that is reliable in terms of coverage, timeliness and quality. In this context, it is important to note that the relative importance of the various areas of the business economy has changed in recent years. In particular, the economic importance of service industries has increased in absolute and relative terms over the past decade. Between 2000 and 2014, gross value added generated by European Union economies in trade and non-financial business services increased by 28 % (Figure 1). In the other service industries (financial, public), gross value added increased by 18 %. In the ‘traditional’ areas (agriculture, industry) growth was only 6 %.
Figure 1: Growth of European gross value added, 2000-2014
Source: Eurostat, NAs, data code: nama_10_a10
Since the introduction of the STS Regulation (Council Regulation (EC) No 1165/1998), the set of STS indicators in the area of trade and services has been augmented only twice: with the introduction of output prices for other services in
+6 %
+28%
+18 %
0
1,000,000
2,000,000
3,000,000
4,000,000
5,000,000
6,000,000
Agriculture, Industry,
Construction
Trade & Non-Financial
Business Services
Other Services
(Financial & Public)
Mio
un
its
of
na
tio
na
l cu
rre
ncy
(ch
ain
lin
ke
d v
olu
me
s 2
00
5)
EU 28 Gross value added
2000 2014
25
2005 (Regulation (EC) No 1158/2005) and hours worked and gross wages and salaries for trade and services in 2009 (Regulation (EC) No 329/2009).
Currently, the STS Regulation provides for 38 STS indicators (Table 1), of which 17 are for industrial activities (defined according to NACE Rev. 2), 11 for construction, five for retail trade1 and five for other services besides trade. Thus, the total service sector (including trade) is covered by only 10 indicators.
Apart from this numerical imbalance, the service sector is also comparatively badly covered as regards the periodicity of the indicators, their timeliness and the coverage of economic activities (neither industry nor construction are completely covered, but the gaps appear comparatively larger in the service sector).
A particular problem for STS is that there is no production measure for services similar to the production indicators for industry and construction. Such an indicator would be of particular value for NA forecasts, general business-cycle analysis and the regular monitoring of service industries.
1 The STS Regulation still refers to ‘retail trade and repair’. However, de facto it no longer covers repair
services for private households, since these dropped out of the service sections with the change from NACE Rev. 1.1 to Rev. 2 (now division S95). The only repair services covered by the current Regulation are car repair services.
Table 1: STS indicators — overview
Principal European economic indicators (PEEIs) in bold print , Q — quarterly indicator, M — monthly indicator
Industry Construction Retail trade and repair Other services
Production Industrial production (M) Production in construction (M/Q) • building construction • civil engineering
- -
Turnover Industrial turnover (M) • domestic industrial turnover • non-domestic industrial (euro
area and non-euro area)
- Turnover in retail trade and repair (M)
Turnover in other services (Q)
Costs - Construction costs (Q) Material costs Labour costs
- -
Building permits - Building permits (Q) • number of dwellings • m2 of useful floor area
- -
Number of persons employed
Number of persons employed in industry (Q)
Number of persons employed in construction (Q)
Number of persons employed in retail trade and repair (Q)
Number of persons employed in other services (Q)
Hours worked Hours worked in industry (Q) Hours worked in construction (Q) Hours worked in retail trade and repair (Q)
Hours worked in other services (Q)
Gross wages and salaries
Gross wages and salaries in industry (Q)
Gross wages and salaries in construction (Q)
Gross wages and salaries in retail trade and repair (Q)
Gross wages and salaries in other services (Q)
Output prices Industrial output prices (M) • domestic output prices • non-domestic output prices
(euro area and non-euro area)
Volume of retail trade (M) Output prices of other services (Q)
Import prices Industrial import prices (M) (euro area and non-euro area)
- - -
2. BACKGROUND AND OUTLINE
2.1. Background
The need to improve STS in the area of services has been on the European statistical agenda for some time. The STS Working Group discussed the introduction of an index of services production (ISP) as long ago as 2007.2 Also in 2007, Eurostat contributed financially to methodological work on a monthly ISP in the Netherlands and Sweden.3
Although the STS Regulation requires only a turnover index at current prices, Sweden and the United Kingdom already calculate a general ISP (‘index of services’ – IoS), some countries produce deflated service turnover indices for a wide range of services and many calculate such indicators for at least a selection of service activities.4 However, no ISP is yet available at European level.
In a workshop on STS on 8-9 December 2011, Eurostat (Unit G.3: short-term statistics, tourism) discussed the shortcomings of STS in the service sector with the national statistical institutes (NSIs) and users.5 There was wide agreement as to the importance of a new ISP and its potential usefulness for business-cycle analysis and NAs.
In line with the outcome of the workshop, an ISP Task Force was established in December 2011 to pave the way for introducing a European ISP in the range of STS indicators.
The Task Force held four meetings (on 14-15 June 2012, 6-7 June 2013, 18 June 2014 and 10-11 June 2015), which took place in parallel with preparatory discussions between Eurostat and NSIs regarding the future content of European STS and with legal preparations for a new Framework Regulation for the Integration of Business Statistics (FRIBS). This parallel approach stemmed from the need to provide statistical experts in the EU Member States with guidelines on the production of ISP data that would be available (at the latest) when the new requirements for European STS enter into force.
The Task Force took stock of available data that could be used to calculate an ISP and used them in test calculations. It investigated the gaps in STS and other data that hinder the calculation of a comprehensive ISP. It also made recommendations on various methods for calculating service-specific ISPs.
2.2. Purpose and organisation of this guide
This guide was prepared in order to give official statisticians practical advice on how to prepare an ISP in time when the relevant legal obligation enters into force.
2 Minutes of the STS Working Group meeting, 18–19 December 2007. 3 Grant agreements 44402-2007.004.343 (NL) and 44402-2007.004.356 (SE). 4 PEEIs in focus – a summary for the index of turnover in services (2009), p. 40. Available at:
http://ec.europa.eu/eurostat/documents/3888793/5851757/KS-RA-12-006-EN.PDF/d786c812-8bc7-495a-914c-5d9c2d53a55f?version=1.0.
5 Draft minutes of the STS workshop, Luxembourg, 8-9 December 2011.
28
The guide is to a large extent built on previous work by Eurostat and the OECD, notably the Handbook on price and volume measures in national accounts (Eurostat Handbook, 2001 edition)6 and the Organisation for Economic Cooperation and Development’s Compilation manual for an index of services production (OECD ISP Manual, 2007).7 It also takes into account the practical experiences of the UK’s Office for National Statistics (ONS) with the monthly IoS8 and Statistics Sweden with a similar monthly service volume indicator.9 The Producer price index manual (2004)10 and the Export and import price index manual (2009),11 both published by the International Monetary Fund (IMF), were also used as inputs.
The guide follows the example of the OECD-Eurostat Methodological guide for developing producer price indices for services (SPPI Guide) and is divided into two main parts: after introductory remarks in sections 1 and 2, section 3 outlines general methodological principles for developing ISPs. The remaining sections provide examples of ISPs for particular service industries in individual countries. We recognise that the findings in these sections reflect particular conditions in the countries concerned, but hope that they will at least provide some practical guidance for other countries.
Eurostat drafted sections 1, 2 and 3 on the basis of numerous papers and presentations, including from members of the Task Force. The service-specific sections were drafted by the members representing NSIs.
3. GENERAL OUTLINE OF A NEW ISP
3.1. Purpose of a new ISP
First and foremost, a new ISP should measure short-term changes in services production. Such an indicator would fill an important gap in STS and provide information on the most recent developments on services in deflated terms similar to the currently available production (volume) indicators for industry and construction, and the indicator for the volume of retail trade. Ultimately, an ISP could be combined with volume indicators for industry, construction and trade to generate an overall production volume indicator for the business economy as covered by STS.
STS volume indicators for industry, construction, trade and services have to be comparable, so ‘services production’ needs to be gauged in terms of value
6 http://ec.europa.eu/eurostat/ramon/statmanuals/files/KS-41-01-543-__-N-EN.pdf. 7 http://www.oecd.org/std/business-stats/37799074.pdf. 8 http://www.ons.gov.uk/ons/guide-method/method-quality/specific/economy/index-of-
services/index.html. 9 http://www.scb.se/en_/Finding-statistics/Statistics-by-subject-area/Trade-in-goods-and-
services/Domestic-trade/Index-of-Service-Production/. 10 http://www.imf.org/external/np/sta/tegppi/con0.pdf. 11 http://www.imf.org/external/np/sta/xipim/pdf/xipim.pdf.
29
added. Since ‘value added’ is a theoretical concept and cannot be directly measured, it has to be approximated (see below).12
The tracking of short-term movements in services production should improve business-cycle analysis in the area of services in general. As regards individual service industries, it should enable us to identify turning points in economic development and improve forecasts. The ISP should therefore focus on service industries that display a business cycle.
A new ISP should also provide important input for NA statistics, so the methodology should be broadly compatible with the NA approach.
3.2. Scope of an ISP
One frequently used definition of services is ‘intangible goods or commodities that cannot be stored’ (they are ‘produced’, ‘traded’ and ‘consumed’ in a single operation). For many types of service, each instance of service delivery is to a certain extent one-off. Also, it often makes a considerable difference by whom and under what circumstances the service is provided.13
Services can be classified in a multitude of ways: private versus public (usually depending on the status of the service provider), market versus non-market (depending on whether the service can be bought/sold or has to be allocated, e.g. by a public administration), business versus consumer services (depending on the main customers), financial versus non-financial, trade versus non-trade, etc.
When constructing an ISP, it is important to distinguish between measuring the production of services and measuring the production of service industries. The first includes services provided by businesses that are mainly manufacturers, (e.g. when the manufacturer of a machine also installs it and provides training for its use). However, it seems more pragmatic to measure the output of service industries (defining services as ‘what is supplied by businesses classified as belonging to the service sector’).
The statistical classification of economic activities in the European Community (Nomenclature statistique des activités économiques dans la Communauté européenne — NACE)14 does not contain a specific section on services. Implicitly, services are all economic activities not otherwise classified, i.e. all activities excluding agriculture, forestry and fishing (NACE section A), mining and quarrying (section B), manufacturing
12 For the definition of ‘production’ in STS, see Commission Regulation (EC) No 1503/2006 (OJ L 281,
12.10.2006, p. 15); http://eur-lex.europa.eu/legal-content/EN/ALL/?uri=CELEX:32006R1503. 13 The introductory guidelines to the CPA define services as ‘entities over which ownership rights cannot
be established. They cannot be traded separately from their production. Services are heterogeneous outputs produced to order and typically consist of changes in the conditions of the consuming units, due to the activities of producers at the demand of the consumers. By the time their production is completed these services must have been provided to the consumers’. The CPA is available at Eurostat’s Ramon (reference and management of nomenclatures) metadata server: http://ec.europa.eu/eurostat/ramon/index.cfm?TargetUrl=DSP_PUB_WELC
14 Regulation (EC) No 1893/2006; http://eur-lex.europa.eu/legal-content/EN/ALL/?uri=CELEX:32006R1893.
30
(section C), electricity, gas steam and air conditioning (section D), water supply, sewerages, waste management (section E), construction (section F) and certain activities of households (section T).
This maximum scope of ‘services’ therefore encompasses NACE sections G, H, I, J, K, L, M, N, O, P, Q, R and S. This is an extremely diverse range; for example, these sections cover public services, market services, trade services, financial services, business services and services for consumers.
For the purpose of this guide, the scope of ‘services’ has to be narrower than that in statistical classifications. There are several reasons for this, some of which are purely practical and stem from time and resource constraints, but (more importantly) our focus is also determined by the purpose of STS and political and economic considerations:
− the service areas analysed should be relevant for business-cycle analysis. Since one of the major objectives for the introduction of an ISP was its use in business-cycle analysis, we focus on business services;
− the coverage should avoid duplicating work in other areas. Where alternative indicators already provide a satisfactory representation of a service industry, an ISP appears to be less important;
− for some service industries, theoretical and empirical analysis is already available. It was necessary to strike a balance between providing information that was relatively new to a sufficiently large user group, on the one hand, and having some experience, on the other; and
− the analysis should focus on service areas for which the production of volume indicators appears feasible in terms of statistical burden and medium-term cost.
In the light of the above, this guide does not cover NACE section K (financial and insurance activities), as this area is already well covered by monetary statistics. Neither does it include service industries that are dominated by public-service providers, i.e. NACE sections O (public administration and defence, compulsory social security), P (education), Q (human health and social work activities). However, it is recognised that some formerly non-market activities provided by public or quasi-public institutions are increasingly becoming market activities. This guide should be considered a ‘living document’ and may have to be revisited in the future to include such services.
Many stakeholders involved in NACE section R (arts, entertainment and recreation) activities are also public or quasi-public. This section is therefore also excluded from a detailed analysis, but the general recommendations of this guide should apply to many of the activities in question.
NACE section S (other service activities) covers the activity of collective organisations (94.1) which are also quasi-public. However, it also contains the repair of computer and household goods (S95) and other personal services (S96), which were partly covered by STS before the introduction of NACE Rev. 2. Given the scale of these services (less than 0.5 % of the other
31
STS services) and their limited significance for business-cycle analysis, they are not covered here. The same applies to activities of households as employers and activities of extraterritorial organisations (NACE sections T and S).
In conclusion, this guide concentrates on NACE sections H, I, J, L, M and N, and for trade services section G (see below).
3.3. Trade services: a special case
Trade services are the services that wholesale businesses or retail traders provide for their customers, e.g. the selection and presentation of a consistent assortment of products, sales advice, packaging, transport, etc. With few exceptions, no special fee is charged; the price is included in the sales margin.
In principle, the volume of sales services could be determined following a method comparable to that used for other services, i.e. by deflating a value index with an appropriate price index (see below). However, there is currently only limited experience of producing special deflators for sales services.15 A possible way of approximating the volume of sales services is to develop the volume of sales, on the assumption that the former develop in parallel with the latter.16
For retail trade (G47), a trade volume indicator is already available in all EU countries. For the trade and repair of motor vehicles (G45) and wholesale trade (G46), turnover indicators are generally available, but only a few Member States have developed trade volume indicators because of the lack of appropriate deflators.17
For G45 and G46, a relatively high number of price indicators is available from other statistical areas, e.g. consumer price, industrial producer price and agricultural price indices, which could be re-used as proxies for wholesale price indicators (see Table 2). The basic methodological issues for the deflation of G45 and G46 would be similar to those for G47.
We can therefore conclude that for trade activities where data coverage is already relatively good, only minor efforts are required to calculate an ISP. For this reason, we do not address NACE sections G45 and G46 in detail in this guide.
15 For an example from Statistics Canada, see the chapter on wholesale and retail trade services in the
Eurostat-OECD SPPI Guide; http://ec.europa.eu/eurostat/web/products-manuals-and-guidelines/-/KS-04-14-661.
16 It should be noted that this approximation might not be adequate for individual trade businesses or in special situations. A trade business facing a declining turnover might improve and enlarge the quality of its sales services to attract customers.
17 In 2014, Eurostat launched a grant scheme for the development of trade deflators. The results will be published on CIRCABC.
Table 2: Potential inputs for deflators for G45 and G46
NACE HICP 1) IPPI 2) API 3)
COICOP
4) Input Output G45, G46 Wholesale exc. Motor vehicles, Trade and repair of motor vehicles and motorcycles - G45 Wholesale and retail trade and repair of motor vehicles and motorcycles - G45.1 Sale of motor vehicles 07.1.1 29.1 G45.11 Sale of cars and light motor vehicles 07.1.1 29.1 G45.19 Sale of other motor vehicles 07.1.1 G45.2 Maintenance and repair of motor vehicles 07.2.3 29.3 G45.3 Sale of motor vehicle parts and accessories 07.2.1 29.3 G45.31 Wholesale trade of motor vehicle parts and accessories 07.2.1 29.3 G45.32 Retail trade of motor vehicle parts and accessories 07.2.1
G45.4 Sale, maintenance and repair of motorcycles and related parts and accessories 07.1.2 07.2.1
30.91 29.3
G46 Wholesale trade, except of motor vehicles and motorcycles - - - G46.1 Wholesale on a fee or contract basis - -
G46.11 Agents involved in the sale of agricultural raw materials, live animals, textile raw materials and semi-finished goods
G46.12 Agents involved in the sale of fuels, ores, metals and industrial chemicals G46.13 Agents involved in the sale of timber and building materials G46.14 Agents involved in the sale of machinery, industrial equipment, ships and aircraft
G46.15 Agents involved in the sale of furniture, household goods, hardware and ironmongery
G46.16 Agents involved in the sale of textiles, clothing, fur, footwear and leather goods G46.17 Agents involved in the sale of food, beverages and tobacco G46.18 Agents specialised in the sale of other particular products G46.19 Agents involved in the sale of a variety of goods - -
G46.2 Wholesale of agricultural raw materials and live animals 201000 206000 92000
G46.21 Wholesale of grain, unmanufactured tobacco, seeds and animal feeds 09.3.4/5 42000 G46.22 Wholesale of flowers and plants 09.3.3 110000
33
G46.23 Wholesale of live animals 09.3.4/5 129000 G46.24 Wholesale of hides, skins and leather C15.1. - -
G46.3 Wholesale of food, beverages and tobacco - -
61000 / 62000/ 40000
G46.31 Wholesale of fruit and vegetables 01.1.6 / 01.1.7 C10.3
G46.32 Wholesale of meat and meat products 01.1.2 C10.1 120000 G46.33 Wholesale of dairy products, eggs and edible oils and fats 01.1.4 C10.5 70000 G46.34 Wholesale of beverages 01.2 / 02.1 C11 23000 G46.35 Wholesale of tobacco products 02.2.0 C12
G46.36 Wholesale of sugar and chocolate and sugar confectionery 01.1.8 C10.81 / C10.82
G46.37 Wholesale of coffee, tea, cocoa and spices 01.2.1 C10.83 G46.38 Wholesale of other food, including fish, crustaceans and molluscs 01.1.3 C10.2 G46.39 Non-specialised wholesale of food, beverages and tobacco G46.4 Wholesale of household goods G46.41 Wholesale of textiles 03.1. C13 G46.42 Wholesale of clothing and footwear 03.2. C14 / C15 G46.43 Wholesale of electrical household appliances 05.3-1/2 C27.51 G46.44 Wholesale of china and glassware and cleaning materials 05.4 C23.41 G46.45 Wholesale of perfume and cosmetics 12.1.2/3 C20.42 G46.46 Wholesale of pharmaceutical goods 06.1.1 C21
G46.47 Wholesale of furniture, carpets and lighting equipment 05.1.1 / 05.1.2
C31 / C13.93 / C27.4
G46.48 Wholesale of watches and jewellery 12.3.1. C32.1 G46.49 Wholesale of other household goods G46.5 Wholesale of information and communication equipment -
G46.51 Wholesale of computers, computer peripheral equipment and software 09.1.3 C26.1 / C26.2
G46.52 Wholesale of electronic and telecommunications equipment and parts 08.2/3.0 C26.3 G46.6 Wholesale of other machinery, equipment and supplies -
34
G46.61 Wholesale of agricultural machinery, equipment and supplies C28.3 G46.62 Wholesale of machine tools C28.4 G46.63 Wholesale of mining, construction and civil engineering machinery C28.92 G46.64 Wholesale of machinery for the textile industry and of sewing and knitting machines C28.94 G46.65 Wholesale of office furniture C31.01 G46.66 Wholesale of other office machinery and equipment G46.69 Wholesale of other machinery and equipment G46.7 Other specialised wholesale - G46.71 Wholesale of solid, liquid and gaseous fuels and related products 07.2.2 / G46.72 Wholesale of metals and metal ores C24
G46.73 Wholesale of wood, construction materials and sanitary equipment C16 / C23
G46.74 Wholesale of hardware, plumbing and heating equipment and supplies C23.42/ C25.21
G46.75 Wholesale of chemical products C20 G46.76 Wholesale of other intermediate products G46.77 Wholesale of waste and scrap G46.9 Non-specialised wholesale trade
Footnotes 1) Harmonised index of consumer prices according to classification of individual consumption by purpose (COICOP). 2) Industrial producer prices, available at two-digit level for medium-sized countries and three- and four-digit level for bigger countries. The coverage is the same for industrial
import prices. 3) Agricultural price indices (http://epp.eurostat.ec.europa.eu/portal/page/portal/agriculture/introduction). 4) Classification of individual consumption by purpose.
3.4. Definition and measurement of outputs
3.4.1. Theoretical output concepts and approximations
STS already produce two measures of production volume: an index of industrial production and an index of production in construction. Both are defined as changes in the gross value added of the industries in question. From the point of view of consistency within STS, it is preferable to define the ISP in a comparable manner. This would also be in line with NA requirements.
The concept of gross value added to be used in STS is defined by Commission Regulation (EC) No 1503/2006,18 which explicitly recognises that the definition is theoretical and must be approximated by practical measures (see below).
Gross value added is an output measure corrected for intermediate consumption by the measured sector:
Gross value added = output - intermediate consumption
It is not corrected for the use of fixed assets (depreciation). It includes subsidies on products, but excludes taxes on products.
Obviously, the above output concepts have to be measured at constant prices (i.e. corrected for mere price movements that do not affect volumes). Since the price development of gross output can differ from the price development of intermediate consumption output, the latter would ideally be deflated independently with appropriate price aggregates (i.e. double deflation).
However, the calculation of such an indicator requires data that are usually unavailable, especially in the short term. The OECD ISP Manual therefore recognises the possibility of substituting the preferred net output measure by a gross output measure (paragraphs 8 and 9). Implicitly, it is thus assumed that intermediate consumption and gross output are highly correlated and develop (more or less) in parallel.
Even leaving aside the question of double deflation, the calculation of net output values as defined above would require information on gross output and intermediate consumption, which generally has to be collected via special surveys. Thus, for practical reasons, alternative output measures, such as revenue, turnover or direct output volumes, have to be considered as approximations.
Revenue is a rather broad concept which encompasses all kinds of income.19 Revenue data might be comparatively easy to collect (e.g. from administrative tax data), but include revenues that have no direct relation to a company’s core economic activity (e.g. revenues from sales of assets). Because these revenues are special and exceptional, there is a risk that they
18 Commission Regulation (EC) No 1503/2006 (OJ L 281, 12.10.2006, pp. 15–29);
http://eur-lex.europa.eu/legal-content/EN/ALL/?uri=CELEX:32006R1503. 19 The OECD ISP Manual discusses a further breakdown of revenues into total revenues, operation
revenues and total receipts (paragraph 65).
36
will distort the information on the development of the main business activities.
Commission Regulation (EC) No 1503/2006 therefore excludes financial and extraordinary income from the measurement of production. It basically defines turnover (for the purposes of STS) as everything that is invoiced by a unit during a reference period. It includes rebates, discounts and charges for ancillary services, but excludes VAT and similar taxes which only pass through the seller’s account. The concept is much narrower than the concept of revenue, since it does not include other operational income.20
Currently, turnover definitions for STS and structural business statistics (SBS) vary slightly; in particular, STS exclude VAT and similar taxes and duties on services, while SBS only exclude VAT. STS include subsidies, but SBS do not.
Under the future FRIBS, the definition of ‘turnover’ for STS and SBS will be aligned, which will lead to some changes for STS: the new definition will exclude subsidies and be more oriented towards the accrual principle, i.e. turnover should be recorded for the period in which it occurs and not the period in which invoices are sent or payments made.
Generally, the collection of turnover data does not pose significant definition problems. However, there might be periodicity issues, since a service might be invoiced in one period and actually delivered in another. The problem should be smaller for services than for goods, since they typically cannot be sold from a company’s stocks. In some cases, however, turnover and service delivery/production may relate to different periods, for example under long-term contracts (e.g. for cleaning services) where services are delivered every month, but invoices are issued only once a quarter.
Using the development of turnover as a proxy for the development of gross value added/net output is admissible if it can be assumed that the ratio between turnover and input remains relatively stable. This depends on a number of factors, e.g. the economic cycle and notably the type of business. In trade, it is often not justified to assume a stable relation between the values of inputs of goods for resale and turnover, as selling prices often remain constant even if input purchase prices change.
3.4.2. Measurement of outputs — deflating value data
In principle, there are three ways of calculating a volume output measure:
− deflating a monetary gross output measure (i.e. turnover) with a suitable price index;
− using a direct volume output measure (e.g. passenger transport kilometres, established phone calls, number of nights of hotel accommodation); and
20 For details, see Commission Regulation (EC) No 1503/2006 (OJ L 281, 12.10.2006, p. 15);
http://eur-lex.europa.eu/legal-content/EN/ALL/?uri=CELEX:32006R1503.
37
− using input measures (e.g. hours worked or another labour indicator).
The first is the conceptually preferred method.21 This is also in line with the general preference of the System of national accounts 2008: ‘… in most cases it is preferable and more practicable to use price indices to deflate current value data ...’ (paragraph 15.103).
The Eurostat Handbook on price and volume measures in national accounts list gives some reasons for this preference (sections 2.3 and 2.4.3.):
• price observations in a competitive market environment are generally quite representative, since prices cannot vary greatly under such conditions; and
• price measurements usually account better for quality changes than is the case with quantity measurements.
The Handbook (section 3.1.1.3) also lists some characteristics of a price index that is suitable for the deflation of a value measure (e.g. turnover):
• It should generally be a producer price index (PPI). Such indices have a number of conceptual features that reflect their main purpose, i.e. the production of real volume measures (whereas consumer prices are used more for the deflation of income).
One major conceptual difference between PPIs and consumer price indices (CPIs) is that the former do not include VAT and similar taxes linked to sales, since they do not represent revenue from the point of view of the producer (but are simply collected and passed on to the tax administration). Thus, the price index should take into account basic prices and not (like a CPI) total prices.
A second important conceptual difference is that producer prices reflect prices for products from domestic producers. To a large extent, consumer prices include the prices of imported products.22
These conceptual differences also indicate in which cases and how consumer prices might be successfully used to approximate service producer prices for the production of service volume measures. Generally, basic price measures have to be used, so CPIs have to be corrected for possible changes in VAT and similar taxes. Ideally, such changes should not only encompass an arithmetical correction for the new tax rate, but should (to the extent possible) also take into account induced price changes set by the producer in the context of the new VAT arrangements (e.g. a producer might not simply pass on a higher VAT rate in a higher selling price, but reduce its own margin and pass on the higher rate to consumers only partially).23
21 Eurostat (2001), Handbook on price and volume measures in national accounts, chapter 2.3. and OECD
(2007) ISP Manual, paragraph 90. 22 Apart from the basic conceptual differences, there are a number of other, more technical, differences,
e.g. in the collection process, the weighting schemes, etc. 23 The opposite is also possible. In some European countries, VAT rates for hotel services were reduced
some time ago, but observed consumer prices remained more or less unchanged. This indicates an implicit increase in producer prices.
38
Moreover, an approximation of producer prices by consumer prices will be easier to justify if imports have a limited impact on consumer prices, as could be assumed for local transport services, for example;
• A second important qualification is that value and price measures should be based on the same structure of products;
• A third precondition for the use of a price index to deflate value measures is that it should take account of quality changes. These are to be considered as changes in ‘volume’ (if the quality increases, the customer receives ‘more’, in the sense of ‘better’). Quality adjustments may have a sizeable effect on service provider price indices (SPPIs) and thus also on ISPs.
A deflation approach also seems a sound option, as a range of gross output/turnover and service output prices already exists and it appears logical to seek better alignment of those indicators.
3.4.3. Measurement of outputs — direct volume measures
While the deflation approach is generally preferable, output volumes may also be measured directly in certain cases, e.g. where turnover and/or price data are not available for a given service industry or are not of sufficient quality.24 One might also opt for direct measurement where such data are already available, as official statistics or from business associations, etc., so that additional data collection by statistical offices is not strictly necessary.
In specific cases where the recording of value data and production does not coincide, data on output volumes might also help in avoiding timing mismatches. However, for services where production and consumption usually fall in the same period, this might be less relevant than for other industries (e.g. manufacturing).
Examples of direct volume measures which could in principle be considered as a basis for an ISP include:25
− passenger/kilometres and tonne/kilometres (H49, H50, H51);
− number of letters and parcels (H53);
− number of hotel nights (I55);
− number of customers (I56);
− number of sales and rentals/leases of real estate broken down by type and size (L68);
− number and size of rented billboards; and
− TV time bought (M73).
If direct volume data are used, they should be detailed enough to allow the identification of relatively homogeneous service products. Also, direct
24 The OECD ISP Manual gives the example of air transport, where price changes could be difficult to
measure (paragraph 97). 25 These and other examples can be found in the OECD ISP Manual (Annex 2).
39
output data will be more reliable for products that are not subject to frequent quality changes.26
It should be noted that direct volume data also have drawbacks. As far as official statistics are concerned, real output data are often available later than value (turnover) data and are thus not immediately useful for the purposes of STS (although they could be used for an ex post quality control of turnover data, since it is to be expected that turnover data and real output data developments are correlated).
There appear not to be many direct volume data in official statistics that might be used for the purposes of an ISP. The most prominent examples are tourism statistics and transport statistics. Tourism statistics provide monthly data on the number of nights guests stayed in hotels and similar accommodation. These can be compared with (unadjusted/gross) turnover in NACE section I55 (accommodation services). For the countries for which Eurostat publishes monthly service turnover data for I55 (the Czech Republic, Germany, Spain, France, Portugal, Romania, Finland and the UK), the correlation coefficient between the data series (2005-2014) ranges from 0.35 for Finland to 0.91 for Portugal. The average coefficient for the eight countries is 0.71.
Both tourism and STS turnover data show a very strong seasonality, which leads to a relatively high correlation of the time series. Figure 2 shows the example of Portugal, for which the correlation is particularly high.
Like data on accommodation, air transport data displays a strong seasonality, with turnover and passenger figures peaking during the summer months.
The correlation was tested for all countries for which monthly turnover data are published for this industry (Czech Republic, Spain, France, Luxembourg, Portugal, Romania and the UK). For passenger data, the series ‘on board’ and ‘carried’ were selected. The ‘on board’ figures refer to the passengers on board a plane at every stage of a flight. The ‘carried’ figures count passengers only once per flight (the differences in the growth rates between the two series are relatively small).
26 OECD ISP Manual, paragraph 98.
40
Figure 2: Growth rates (m/m) for nights in hotels and similar accommodation and turnover (index: 2010 = 100), Portugal (unadjusted data)
Source: Eurostat, STS, sts_setu_m, Tourism statistics, tour_occ_nim
Since passenger data are not adjusted for seasonal effects or working days, the correlation between them and turnover was tested for unadjusted (gross) turnover data for NACE division H51 (air transport). For the test period 2005-2014, the correlation coefficients range between 0.30 for Romania and 0.70 for the UK; the average coefficient was 0.59.
3.4.4. Measurement of outputs — (labour) input data
The third and generally least preferred option for the measurement of output is estimation by input data, in particular by labour input.
The Eurostat Handbook on price and volume measures in national accounts classifies the use of input data for the production of volume output data as a C method, i.e. generally not desirable (section 3.1.1.3). The OECD ISP Manual is more lenient, but sees a role for input methods only as proxy measures where no other variables are available (paragraph 68).
The main disadvantage of using input data to estimate output is that in cases of productivity changes there will be no clear link between the two data sets. If time changes in input and output are proportional to each other or input data are supplemented by other estimates (e.g. productivity adjustments), the use of input variables may produce meaningful estimates. However, if the assumption of input and output proportionality is not likely to hold or if adjustments cannot be made for productivity, it should be avoided.
Apart from the productivity problem, there is also a danger of a mismatch between labour input data and turnover. For example, employees in larger businesses could be registered in different NACE classes, so proper delineation is difficult.
41
STS encompass labour input variables in the form of employment, hours worked and gross wages and salaries. (The latter two only became an STS obligation in 2013). We know that employment figures react only with a certain delay to the development of production, since businesses will not immediately dismiss employees when production has to be reduced, but will first resort to other measures such as working-time reductions. Likewise, an increase in demand and production will not immediately result in the hiring of new employees. Neither is the development of gross wages and salaries directly linked to the development of production. The most obvious choice among the STS labour input indicators is therefore ‘hours worked’. In the OECD ISP Manual, the ‘use of hours worked is recommended when no other indicators are readily available’ (paragraph 70).
However, even the link between hours worked and production is questionable. As an example of the problems that might occur, Figure 3 illustrates the development of hours worked and production for food manufacturing (NACE C10). This is generally considered to be a ‘low tech’ sector where changes in productivity should affect the relation between input and production considerably less than in some service industries.
Figure 3: Hours worked and production indices for the European food manufacturing industry
Source: Eurostat, STS, data codes: sts_ind_prod and sts_ind_labo.
3.5. Periodicity of an ISP
In many respects, the methodological principles applying to the calculation of monthly and quarterly indices should be identical (index construction, level of deflation, etc.). However, monthly and quarterly indices will in some cases require different statistical surveys, which will have some consequences for the methodologies used.
42
3.5.1. Benefits, cost and burden of monthly turnover data
Regulation (EC) No 1165/98, which introduced STS at European level, provides for quarterly service turnover and producer price data and thus, at first glance, suggests a quarterly ISP. However, the production of a monthly indicator would have several tangible advantages:
� monthly data allow for a more precise identification of turning points and a more detailed analysis of the business cycle;
� the effect of special events (e.g. tax changes) or unusual calendar constellations can be better observed in monthly data;
� monthly data indicate economic developments earlier. If the publication deadline is two months after the end of the reference period, for example, a change in the business cycle in January will not be observed in quarterly data until June;
� STS turnover data provide important input for the forecast of the service sector’s contribution to GDP and monthly turnover data would be conducive to better and timelier GDP forecasts; and
� in general, the monthly frequency of STS is a key requirement of data used to monitor economic developments, in particular since quarterly data may be available from other statistical domains.
While monthly ISP data have several clear advantages, the introduction of monthly data collection in countries where it is not already the norm would also increase costs for statistical institutes (statistical costs) and businesses (response burden). These increased costs can be assessed roughly on the basis of 2012 data provided to Eurostat in the context of data collection for the 2014 STS quality report.27 Table 3 shows the total burden (expressed in hours per year) of collecting data for service turnover indices in various countries. This can be combined with the number of companies that have to provide data to calculate monthly averages for the majority of EU countries.
Some countries collect turnover data only quarterly, while some have monthly data collection. Some (DK, FR, LU and BE) rely on data from administrative sources (generally VAT data), so the STS data collection imposes no additional burden on businesses. Several countries combine administrative data collection with surveys. Overall, the various collection methods appear to result in a burden on the average business of around only six minutes per month.
27 The STS Regulation requires the Commission to submit a report to the European Parliament and the
Council every three years on the statistics compiled pursuant to the Regulation and in particular on their relevance and quality and the revision of indicators. The report must also address the cost of the statistical system and the burden on business arising from the Regulation in relation to its benefits. The most recent report (COM(2014) 381 final) was published in 2014, with cost and burden data for 2012; it is available at: http://ec.europa.eu/eurostat/web/short-term-business-statistics/methodology.
43
Table 3: Burden of turnover collection
Source: Eurostat, internal data for the preparation of the 2014 STS quality report to the European Parliament and the Council.
Table 4 provides similar information on the costs for NSIs. Here, the differences between the available data collection methods are more clearly marked. While it means a zero burden on businesses, relying on administrative data results in costs for NSIs (albeit comparatively low: 4.2 minutes per statistical unit).
It should be noted, however, that a precise calculation of such costs and burdens is notoriously difficult. The impact of a change of data collection frequency will very much depend on the circumstances in each country.
Table 4: Cost of turnover collection
Source: Eurostat, internal data for the preparation of the 2014 STS quality report to the European Parliament and the Council.
Burden
(h/y)
No. of
Units
Monthly
minutes
per unit
Burden
(h/y)
No. of
Units
Monthly
minutes
per unit
Burden
(h/y)
No. of
Units
Monthly
minutes
per unit
BG 5,120 8,520 3 CZ 12,000 12,247 5 DE 47,267 18,200 13
EE 3,360 4,381 4 IE 1,850 2,045 5 NL 14,755 23,051 3
EL 2,687 1,244 11 ES 10,100 10,100 5 SI 1,130 807 7
IT 38,800 19,400 10 PT na na na FI 2,400 600 20
CY 2,118 1,203 9 RO 16,214 8,107 10 SE 20,000 10,500 10
LV 4,280 3,933 5 SK 10,191 9,713 5 UK 14,051 24,500 3
LT na na na
HU 3,024 8,563 2
MT na na na
PL 10,159 10739 5
Total 69,548 57,983 6.0 Total 50,355 42,212 6.0 Total 99,603 77,658 6.4
Administrative data (burden = 0): DK, FR, LU, BE
AT: admin data plus voluntary survey, HR - no data collected for 2012
Quarterly survey Monthly survey Combination survey & admin data
Cost
(h/y)
No. of
Units
Monthly
minutes
per unit
Cost
(h/y)
No. of
Units
Monthly
minutes
per unit
Cost
(h/y)
No. of
Units
Monthly
minutes
per unit
BG 10,240 8,520 6 CZ 11,000 12,247 4 DE 44,240 18,200 12
EE 2,005 4,381 2 IE 9,464 2,045 23 NL 21,689 23,051 5
EL 4,520 1,244 18 ES 51,650 10,100 26 SI 6,317 807 39
IT 18,114 19,400 5 PT na na na FI na na na
CY 3,906 1,203 16 RO 21,712 8,107 13 SE 6,000 10,500 3
LV 16,498 3,933 21 SK 10,904 9713 6 UK 70,575 24,500 14
LT na na na
HU 1,767 8,563 1
MT na na na
PL 7,199 10739 3
Total 64,249 57,983 5.5 Total 104,730 42,212 12.4 Total 148,821 77,058 9.7
Administrative data: BE, DK, FR, LU, AT: Average Costs 4.2
HR - no data collected for 2012
Quarterly survey Monthly survey Combination survey & admin data
44
3.5.2. Need for monthly turnover data — differences between monthly and quarterly results
For several Member States, monthly turnover data are available for some service activities and we can gauge the advantages of monthly indicators, as compared with quarterly data, as follows:
i. for each country and industry, quarterly averages were produced on the basis of the available ‘monthly working days’ adjusted data;
ii. absolute differences between each monthly observation and the respective quarter were calculated and expressed as a proportion of the quarterly averages (e.g. ABS(January-Q1)/Q1, ABS(February-Q1)/Q1, …, ABS(April–Q2)/Q2, etc.); and
iii. averages were calculated for each month over the available years and finally for each industry over the available countries.
Figure 4 summarises the results. As a point of reference, the comparison between monthly and quarterly data was also made for manufacturing and retail trade (for the countries for which monthly service data were available). Figure 4 shows that, for example, the average monthly index for manufacturing will differ by 5.5 % from the quarterly average. With few exceptions (e.g. warehousing, postal services, food and beverage activities, rental and leasing, employment and cleaning activities), the differences between monthly and quarterly data are relatively large for service industries, as compared with manufacturing or retail trade, where there are well known deviations between monthly and quarterly data.
These results suggest that monthly data provide a much more accurate picture of the development of turnover industries; this could be approximated only insufficiently by quarterly data.
45
Figure 4: Turnover indices for services industries – average difference between months and their respective quarters (selected countries)
Source: Eurostat, STS, sts_setu_m, internal calculations.
3.5.3. Seasonal patterns in the turnover of services
If the variations between months and quarters were erratic, inexplicable and unpredictable, even great variations might not be sufficient to claim that monthly data are more useful than quarterly data. If it could be demonstrated that the variations are part of a relatively stable pattern and thus, in principle, explicable, the arguments in favour of the usefulness of monthly data would be strengthened.28
The available monthly service turnover data were therefore analysed for stable seasonality using the ‘F-test’, which compares the variance between months with the residual variance of the time series. It is assumed that the former is caused by seasonal factors and the latter is a result of irregular movements. A test coefficient of 7 or more indicates that the (regular) monthly variations are repetitive and regular enough to be reliably identified as seasonal movements.
The results are reported in Table 5. In general, the seasonal pattern is not as pronounced as for retail trade turnover, but it is often comparable with the seasonality of turnover in manufacturing.
28 One might argue that monthly data are not necessary when there are strong predictable patterns in a
time series, since they could be estimated on the basis of the pattern and the quarterly values. However, this is true only for extremely rigid patterns, e.g. we know that retail trade volume in December is higher than in October and November, but analysts still consider information on Christmas shopping to be useful. In any case, to make that argument, one first needs the monthly data.
Table 5: Test for seasonality, assuming stability (grey cells — test found no seasonal pattern)
Source: Eurostat, STS, sts_setu_m, internal calculations (software Demetra +, critical value: 7).
47
For manufacturing, the differences between individual months and their quarter is relatively strongly affected by the summer holidays. For retail trade, a strong influence of Christmas shopping on turnover development is apparent. In many cases, the seasonal influences are also relatively easy to explain for service industries. For example, there are clear seasonal peaks in the tourism industry, e.g. I55 (food and beverages), I56 (accommodation) and N79 (tour operators and travel agencies), especially in countries with a strong tourism sector, e.g. Spain and Portugal. These seasonal effects show country-specific patterns that are relatively easy to explain, e.g. in Spain, the peak month for restaurants and similar services is August; in Germany, the peaks occur just before and after the summer, in May and September.29 Nevertheless, the seasonal patterns have strong similarities, so the weighted and unweighted patterns for the countries for which data are available are fairly similar (Figure 5).
Figure 5: Seasonal pattern for turnover in accommodation services (I55) — summer peak
Source: Eurostat, STS, (sts_setu_m), internal calculations.
A counter-example is advertising and market research, in which turnover is typically low during the summer holidays when it is more difficult to reach an audience.
29 A comparison with data from tourism statistics shows that in certain countries tourism (as measured by nights
spent in hotels, for example) has a clear single peak in summer (mainly August), whereas in other countries tourism activities are more evenly spread between June and September.
48
Figure 6: Seasonal pattern for turnover development in advertising and market research (M73) — ‘silly season’
Source: Eurostat, STS, (sts_setu_m), internal calculations.
Many service industries display a ‘hockey-stick’ effect30 — a phenomenon well known in marketing and sales departments. One reason could be that (especially larger) companies have to meet quarterly business targets or sales goals and offer special incentives to potential customers towards the end of each quarter. (Since prospective buyers know about this, they might also tend to delay orders until the end of a quarter to benefit from discounts).
A typical example of the effect is illustrated in Figure 7, which represents the infra-annual pattern for NACE division M74 (other professional, scientific and technical services). Figure 7 shows unweighted monthly average index values (2010 = 100) for 2000-2012.31 Peaks occur in March, June and (after a lull in the summer) September, and finally in December.32
30 In general terms, a hockey-stick effect describes a certain pattern of time series in which the values of the
series change significantly upwards and downwards after a period of only minor changes. The level of the time series then often moves back to the original level.
31 M74 monthly turnover data from CZ, DK, ES, FR, LU, PT, RO, SI, FI and UK. 32 Ideally, the average of the individual months should be corrected for the development of the overall trend,
but the basic pattern would not be likely to change much.
49
Figure 7: Seasonal pattern for turnover development in other professional, scientific and technical services (M74) — ‘hockey-stick’ effect
Source: Eurostat, STS, (sts_setu_m), internal calculations.
A similar pattern — strong increase in March, slump in summer, increase in September and a particularly large increase in December — can be observed for some other business services, such as M68_70.2, M71. In particular, the big increase in December is quite a common phenomenon in business services (‘December Fever’).
3.5.4. Monthly turnover data and quarterly deflators
To the extent that monthly ISPs are produced by deflating turnover data, monthly deflators are needed that match the monthly frequency of the turnover data. This does not mean, however, that a monthly collection of price data is necessary; monthly deflators can also be produced on the basis of quarterly data collection.
Unfortunately, it is not possible to analyse the usefulness of monthly SPPI data in the same way as for turnover data (see above), because only minimal monthly SPPI data are available in the EU.33 However, analysis of available SPPIs for the EU shows that in many service industries changes between months are not very great. Table 6 shows the average variance in time series for (unadjusted) turnover indices and SPPIs. Other (anecdotal) evidence suggests that service prices do not vary much from one month to the next and a monthly collection of service producer prices would therefore not generate much additional information. It can even be argued that, in certain cases, a monthly collection of prices could reduce
33 One of the few exceptions in the EU is Poland, which produces monthly SPPIs for NACE section H
(transportation and storage), H49 (land transport and transport via pipelines), H50 (water transport), H52 (warehousing and support activities for transportation), H53 (postal and courier activities) and J61 (telecommunications). For other examples of monthly SPPIs, see Hong, Eun-pyo, McKenzie, Richard (2003).
50
the overall quality of the data, since respondents would have to provide (more or less) the same data for consecutive months. This could lead them to question the purpose of the data collection and detract from their willingness to respond accurately or at all.
Table 6: Variance of quarterly turnover and quarterly SPPIs (arithmetic averages for countries for which data are available), 2010–2014
Industry (NACE) Turnover SPPIs Air transport (H51) 2 364.4 57.9 Postal services (H53) 112.4 31.4 Telecommunication (J61) 47.0 39.7 Computer programming, consultancy (J62) 452.0 11.0 Information services (J63) 1 406.8 13.2 Legal, accounting, headquarters (M69_70.2) 253.4 11.5 Architecture, engineering, testing (M71) 341.5 11.5 Advertising services (M73) 387.8 36.0 Employment services (N78) 913.6 18.2 Security and investigation services (N80) 119.2 10.7 Cleaning services (N81.2) 192.2 9.5
Source: Eurostat, STS, sts_setu_q, sts_sepp_q
When quarterly SPPI data are used for the deflation, two problems need to be solved:
• monthly data have to be derived from quarterly data covering the month in question, and possibly from adjacent quarters (see sections 3.5.4.1 and 3.5.4.2); and
• at the end of the series, price data have to be forecast (see section 3.5.4.3).
3.5.4.1. Temporal disaggregation methods
There are several methods for converting time series with a low frequency (e.g. annual or quarterly) into series with a higher frequency (e.g. quarterly or monthly).34 These range from simple mathematical interpolation to sophisticated model-based techniques. An important distinction can be made between methods that disaggregate temporal data only on the basis of mathematical-statistical models and methods that input additional statistical data that are correlated with the estimated data (e.g. use CPIs to disaggregate SPPIs). Below we list a handful of tested methods that are also incorporated in available computer applications.
In the context of the Eurostat FAME (forecasting analysis and modelling environment) database:
− discrete interpolation;
− linear interpolation (with re-scaling); and
− cubic spline. 34 For some examples, see Di Fonzo (2003), Proietti (2004), Abad and Quilis (2005).
51
In the context of the temporal disaggregation software ‘ecotrim’:
− Denton adjustment method; and
− Feibes, Bott and Lisman smoothing method.
FAME is a database system for time series which Eurostat uses to produce and analyse STS. One feature of the database is that it temporally disaggregates quarterly time series when quarterly data are downloaded in a monthly format, for example. Different disaggregation methods can be selected for the download.
With discrete disaggregation, it is simply assumed that all months in a quarter have the same index value as the quarter. As a result, the disaggregation would yield a function with a characteristic ‘staircase’ appearance (see Figure 8).
Figure 8: Discrete interpolation — stylised example
This procedure implicitly assumes that the entire change in the index level between two quarters occurs between the last month of the first quarter and the first month of the second. There is, however, no evidence to suggest that price levels actually change in this way. A second problem with this method is that it provides no indication of a trend at the end of the time series and thus makes forecasts difficult. In consequence, this method should generally not be used.
With linear interpolation , monthly data are calculated as values from a linear function (yi = a + bxi) which joins the quarterly value of adjacent quarters. The change in the index level between two quarters (e.g. SPPIQ2 — SPPIQ1) is spread over the months.
Plotting the linear interpolated data into a graph basically amounts to joining the quarterly values by straight lines. If linear interpolation is used, it has to be decided where to put the quarter, i.e. on the first, second or third month. This will depend on the actual method of collecting prices. If prices are collected over all months, the quarterly SPPI reflects a monthly average and it seems appropriate to put the quarterly established SPPI at the middle of the quarter (see Figure 9, empirical index). However, the averages of the monthly data produced in this
94
95
95
96
96
97
Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec
52
way (light grey bar in Figure 9) will not correspond to the empirical quarterly data that were the starting point of the operation. The data therefore have to be re-scaled/corrected so that the estimated months and the empirical quarters correspond (Figure 9, re-scaled index).
Figure 9: Interpolated and re-scaled index — stylised example
It is also possible that the SPPI collection observation units report prices for the middle of the quarter, so the empirical index and index for the middle month of the quarter have to be identical. In this case, it is not necessary that the average of the monthly indices and the quarterly empirical index correspond and re-scaling is not required.35
Cubic splice interpolation is a form of polynomial interpolation,36 but it uses only cubic polynomials, i.e. polynomial functions with a maximum degree of 3 (yi = axi
3 + bxi2 + cxi + d). Each function (‘splice’) is used between two adjacent
data points only (i.e. two quarterly SPPI values). By assuming certain plausible constraints, a unique function can be found that connects all quarterly observations. One of these constraints, that the three-month averages of the
35 In both cases, the data collection method and disaggregation procedure should be explained and documented
in the metadata. 36 In general, polynomial interpolation establishes continuous polynomial function. If n+1 observations
(e.g. quarterly SPPIs) are available, a polynomial function of n degrees can be found which connects all given observations and provides estimates for all points between them. However, for polynomial functions of a higher degree (e.g. a function that uses all available quarterly SPPI observations), the overall result of the estimation becomes worse as the estimated function tends to oscillate heavily and will therefore produce some unrealistically high or low estimates for the monthly SPPIs. This problem is avoided by cubic spline interpolation.
53
interpolated monthly data should equal empirical quarterly data, is built into the FAME database.37
With interpolation using additional data sources, the quarterly data basis for SPPIs is enriched by monthly data from other areas. It is assumed that the monthly pattern underlying the quarterly collected SPPIs can be approximated by data from monthly series for other indicators. In some cases, it might be possible to approximate changes in SPPI data by changes in related consumer price data.
Figure 11 shows the SPPI for air transport (NACE H51) for Germany and the German harmonised index of consumer prices (HICP) for air transport. For the latter, consumer price data available on a monthly basis were re-scaled to base year 2010 and quarterly averages were calculated. The two series display rather similar patterns. In such cases, it might be justifiable to use the monthly seasonal pattern from the more detailed series to estimate monthly data for the SPPIs.
In order to improve the quality of the estimations further, the monthly data might first be corrected for effects that have an influence on them but not on the SPPI. For example, a change in the VAT rate would influence consumer prices, but since the definition of SPPIs excludes VAT, consumer prices would have to be corrected for VAT changes before using them in estimating SPPIs.
Figure 10: Linear interpolation and cubic spline interpolation (simple and re-scaled) — stylised example
37 A free version of a cubic spline interpolation can be found at:
http://www.srs1software.com/SRS1CubicSplineForExcel.aspx However, this version does not generate monthly data of which the averages tally with the empirical quarterly data.
70
80
90
100
110
Jan Feb Mar Apr May Jun Jul Aug Sept Oct Nov Dec
Empirical index Re-scaled index
Cubic spline Re-scaled cubic spline
54
Figure 11: Consumer prices and SPPIs for German air transport services (2010=100)
Source: Eurostat, consumer prices, prc_hicp_midx, STS, sts_sepp_q.
Prima facie, linear interpolation, cubic splice interpolation and an interpolation using additional information (e.g. from consumer price data) all appear admissible. Using additional data from other statistical areas probably produces the most realistic approximations. However, such data will not be available for all types of service. Where data are published, cubic splice interpolation might be preferable, as it avoids the problem of identical monthly growth rates within a quarter. However, if monthly data are used only as deflators for monthly turnover data, the greater simplicity might speak in favour of linear interpolation.
3.5.4.2. Practical test of the methods: Polish SPPIs
In order to test further the reliability of temporal disaggregation methods, Eurostat compared original monthly SPPI data from Poland with disaggregated quarterly data from the same source. First, quarterly averages were calculated from the monthly SPPIs. These averages were then disaggregated back to monthly data, which were compared with the original monthly data.
55
Figure 12: Test of disaggregation methods — turnover in the Polish telecommunications industry
Source: Central Statistical Office of Poland, calculations by Eurostat
Figure 12 shows the result for the telecommunications industry. All tested disaggregation techniques produce results that are almost identical with the empirical monthly data. Very similar results were found for other tested service industries.
Figure 13 shows the result for an industry with a relatively specific price development, i.e. occasional and relatively rare price changes. In such a case, all temporal disaggregation models over- and under-estimate the price changes. Where such specific price patterns exist, ad hoc estimation on the basis of additional results (e.g. announcements of price increases for postal services) will probably produce better results.
Figure 13: Test of disaggregation methods — turnover in the Polish postal services industry
Source: Central Statistical Office of Poland, calculations by Eurostat
56
3.5.4.3. Forecasting of SPPIs
It is assumed that an ISP will be produced by deflating a monthly turnover index by a monthly deflator. The deadline for the production of the ISP is t+60 days. The current STS Regulation deadlines for the availability of turnover data and SPPI data will remain unchanged in the future FRIBS (i.e. 60 days for turnover data and 90 for SPPIs), the turnover index will have a monthly periodicity, while the periodicity of the SPPIs will remain quarterly.
Table 7: Availability of turnover data and SPPIs
Month Turnover availability
(t+60)
Quarter SPPI availability
(t+90)
‘Gap’ (months)
January 31 March Q1 30 June 3 February 30 April Q1 30 June 2 March 31 May Q1 30 June 1 April 30 June Q2 30 Sept. 3 May 31 July Q2 30 Sept. 2 June 31 August Q2 30 Sept. 1 Table 7 illustrates the timing mismatch resulting from the different periodicities and deadlines. For example, turnover data for January will become available at the end of March, but SPPI data for the first quarter (and thus also for January) will not be available until the end of June, i.e. three months later. It should be noted that the gap between the availability of turnover and SPPI data is mainly due to the different periodicities and not the different deadlines (60 versus 90 days); even with identical deadlines, a gap would remain for the first two months in each quarter.
Therefore, for most recent periods, monthly SPPIs need to be estimated which could serve for deflation purposes. This could be done using an ARIMA model. Two basic approaches are possible:
− forecasting SPPIs on a quarterly basis and subsequent temporal disaggregation to derive monthly deflators (i.e. indirect calculation of monthly deflators); and
− directly forecasting monthly deflators on the basis of temporally disaggregated data from earlier periods.
For many SPPI series that display rather stable developments, the differences between the two methods will not be great. Where price series are relatively constant, a forecast based on an ARIMA model will generally be useable. One advantage of the first method is that the forecast would be based on real data that have not been derived using an interpolation process (which might have led to artificially smooth data).
Figure 14 illustrates some choices to be made in the case of an indirect approach. At the end of March, when turnover data for January are available and an ISP for January needs to be produced, there are no empirically established SPPI data for January. What should be available is the SPPI for the last quarter of the previous
57
year (Q4 t-1) — illustrated by the black bar over November. Here, an SPPI for the first quarter could be estimated (dark grey bar over February) and an index value for January established by interpolation (indicated by the solid black line).38
On 30 April, when an ISP for February needs to be produced, the estimate for the first quarter (maybe in a revised version, taking into account newer information) could be used as deflator.
On 31 May, a deflator for March is needed. Now two options are available:
− one could simply use the trend extrapolation from the estimation for the first quarter SPPI (see Figure 14, black solid line). It might, however, be known that in the industry in question the second quarter SPPI is usually lower than the first quarter SPPI. In this case, it would be better to estimate the second quarter SPPI (light grey bar over May) and only afterwards calculate the March deflator by linear interpolation (dashed line in Figure 14); or
− one could estimate the deflator for March on the basis of the seasonal distribution of the months within a quarter, which might be known from past data.
Figure 14: Estimation of deflators for January, February and March
A direct approach will be particularly useful if external monthly data are available that can be used to support the estimation. Also, this method could be considered where price series show greater volatility, as this renders ARIMA models for the estimation of quarterly SPPIs less reliable.
All the approaches discussed above are in principle possible and admissible. The choice of approach for a given industry will depend on available data sources, price developments in the industry and possibly the specific situation in the country in question.
38 For the sake of simplicity, the example involves no re-scaling.
58
3.6. Seasonal adjustment
Given the strong seasonal effects in the development of turnover in many service industries, data on services production will also display clear seasonal influences. Strong seasonal effects can also be observed in the real output data for some service industries (e.g. accommodation and transport). In order to allow a meaningful month-on-month comparison (which is essential for STS), data must to be adjusted for seasonal effects.
It can be assumed that, as volume data are produced by deflating turnover data, the largest seasonal effects on the volume data will result from turnover developments. However, in the producer prices of some service industries, e.g. H51 (air transport), I55 (accommodation), M73 (marketing and advertising), seasonal patterns can also be observed in many countries.
Where turnover data are deflated by SPPIs, therefore, three approaches are possible:
− indirect complete adjustment: TOsa / SPPIsa;
− direct adjustment: (TOgross / SPPIgross)sa; and
− indirect partial adjustment: TOsa / SPPIgross.
In the light of seasonal effects of some SPPIs, the third method (indirect partial adjustment) cannot be recommended for all service industries. However, it could be used in cases where SPPIs have no or only low seasonality.
According to the guidelines of the European Statistical System (ESS) on seasonal adjustment, both methods (direct and indirect) are admissible.39 Test calculations indicate that the differences between the two will be greater at the lower NACE levels (i.e. four-digit level or below); at higher levels of aggregation, the differences are generally smaller.
Several experts have expressed a preference for a direct seasonal adjustment approach (first deflate gross turnover by gross prices and then adjust the resulting time series), since it is held that volume data have their own specific seasonality. Moreover, the direct approach should be preferred for clarity, especially when component series show similar infra-annual patterns. The indirect approach might be chosen where the component series have rather different seasonal patterns.
3.7. Statistical unit40
Statistical units (observation units) are the entities for which information is collected. They must be defined in such a way that they cannot be confused with each other. In addition, the units in the statistical populations must be countable.
39 http://ec.europa.eu/eurostat/en/web/products-manuals-and-guidelines/-/KS-RA-09-006 40 The Task Force is aware that the definitions of statistical units are currently being revised. The analysis in
this report is based on the current legal situation, i.e. on the STS Regulation and Council Regulation (EEC) No 696/93 on the statistical units for the observation and analysis of the production system in the Community (the SU Regulation).
59
Currently, STS recognise two types of statistical unit: kind-of-activity units (KAUs) and enterprises. For service STS, only the latter is used. In methodological handbooks and similar publications, the local kind-of-activity unit (LKAU)41 is also mentioned as a possible statistical unit for service statistics.42 In the future (i.e. under the FRIBS), the KAU is envisaged as the only statistical unit for STS.
The main defining criterion of an enterprise is that it is the smallest combination of legal units that has a certain degree of autonomy in its economic decisions. An enterprise may consist of one or more several legal units and may carry out more than one economic activity in more than one local site of production.
A KAU is an enterprise or a part of an enterprise, in which case it groups together all offices, production facilities, etc. (which could be situated at different locations) of an enterprise that contribute to the performance of a specific economic activity defined at NACE Rev. 2 class (four-digit) level. In order to establish a KAU for statistical purposes, the enterprise’s information system must be capable of calculating some core variables for each KAU, e.g. value of production, intermediate consumption, cost of manpower, operating surplus, employment and gross fixed capital formation.
An LKAU or establishment is a part of an enterprise engaged in predominantly one kind of economic activity (like a KAU) in a specific location or geographical area.
Conceptually, the best observation unit for a service output indicator would be an establishment (LKAU).43 The second best would be a KAU. Enterprises often carry out more than one activity, but since they are classified according to their main activity44 a large part of their turnover might be generated otherwise than through the core activities responsible for their classification.
While the more specific KAU offers advantages in terms of homogeneity, there are also some disadvantages. Firstly, the choice for a statistical unit in STS services statistics cannot be made independent of other statistics, especially structural business statistics (SBS). A common statistical framework would facilitate the combination of the different statistics. Secondly, where statistical authorities use administrative source for the collection of data (esp. turnover) a collection at KAU-level might technically not be possible.
41 The LKAU is equivalent to the ‘establishment’ in ISIC and SNA. 42 The SU Regulation lists eight different statistical units. The institutional unit, the enterprise group, the unit of
homogeneous production, the local unit and the local unit of homogeneous production will not be discussed in the present report.
43 OECD ISP Manual, paragraph 32. 44 NACE Rev. 2, chapter 3, Classification rules for activities and units. The correct class (NACE four-digit
level) of a statistical unit is established in a top-down approach, i.e. a unit is first allocated to a section, then to a division within the section, then to a group within the division and to a class within the group. As a consequence, it is possible for a unit to belong to a certain class X even though the turnover generated by the unit in class Z is significantly higher than the turnover generated in X.
60
3.8. Index construction
3.8.1. Choice of basic index type
Various types of index or index formulae exist. A basic distinction can be drawn between unweighted (elementary) and weighted indices. The former result from combining elementary statistical observations (e.g. prices) for which no weighting schemes are available. Generally, these data are combined in the form of a geometrical average. Such indices should be used for relatively homogeneous types of observation (e.g. prices of a relatively narrowly defined product group); they form the basis for the higher-level, weighted indices.
Weighted indices are produced by adding weighted elementary indices (at the lowest level) or lower-level weighted indices. Depending on the weights used and the index formula, it is possible to produce various index types with various characteristics. Statistical index theory identifies a number of properties that an index should ideally have and that should guide the choice of index type. In practice, however, the index used most often is a Laspeyres or Laspeyres-type index (see below), which uses weights from a fixed base period (base year).
The basic formula for a Laspeyres index (IL) is given by: n ILt = ∑ (Iit wi0) i
The Laspeyres index for period t is the sum of the i = 1 … n sub-indices for period t, each multiplied with a weight from the base period 0.
Although the Laspeyres index does not offer all ideal properties, it represents a good compromise.45 It is relatively easy to interpret and less costly and data-intensive than other types of index. Its major disadvantage is that in dynamic economies the weights from the base period will after some time no longer be representative. The weights should therefore be updated at regular intervals, e.g. every five years, as provided for in the STS Regulation. In order to distinguish them from genuine Laspeyres indices, such indices (with regularly updated weights) are referred to as ‘Laspeyres-type’ indices.
The use of a Laspeyres-type index is recommended for the production of an ISP. Alternatively, a chained index could be used. No particular recommendations will be made in this guide as regards chain-linked indices, but attention is drawn to the recommendations in the final report of the ‘chain-linking in STS’ task force.46
3.8.2. Level of deflation
Where an ISP is produced by deflating a turnover aggregate, a decision has to be taken as to the level of aggregation at which the deflation should take place.
45 For a discussion of the various index types, see the OECD ISP Manual, chapter 5.1. 46 The final report of the Chain-Linking Task Force is available on CIRCABC (login necessary):
https://circabc.europa.eu/w/browse/5bacd0b8-7a6e-498c-b4fa-d68f86e3d212
61
According to the Eurostat Handbook on price and volume measures in national accounts (2011), which is based on the Commission Decision clarifying Annex A to Council Regulation (EC) No 2223/96,47 the elementary level of aggregation at which deflation should take place is at least as detailed as P60 in the European System of Accounts (ESA 1995),48 which corresponds to NACE division (two-digit) level. Under the STS Regulation, most service turnover aggregates and SPPIs are (or will be) defined at this level.
However, in many instances, this level of deflation will be too rough, since several divisions in the service sector (e.g. H52) contain quite varied service activities, each of which might still have considerable weight in the aggregation.
Another reason why the level of deflation should not be too high is the ‘Paasche problem’. When a nominal turnover index is deflated by a Laspeyres-type price index, the result will be a Paasche-type volume index, in which the prices of current turnover (pt) form the weights for the volume changes.
n n n n
∑ (pt qt) / ∑ (p0 q0) deflated by ∑ (pt q0) / ∑ (p0 q0) i=1 i=1 i=1 i=1
yields: n n
∑ (pt qt) / ∑ (pt q0) i=1 i=1
In order to arrive at a Laspeyres-type volume index, one would have to deflate with a Paasche-type price index. However, such indices are generally not available. The difference between the desired Laspeyres-type and the actually obtained/used Paasche-type volume index, i.e. the size of the deflation problem, is known to diminish with the homogeneity of the observations in the index, which is greater for more detailed indices.
In view of the above, a lower level of deflation is generally preferable. However, it also needs to be borne in mind that the share of secondary activities in the turnover aggregate will generally increase at a lower NACE level. If deflation takes place at a low NACE level, there is thus a risk that the deflator will actually correspond only to a fraction of the turnover aggregate that is deflated.
In general, deflation at NACE three-digit level will probably represent a workable compromise. Depending on the industry and the individual circumstances in the Member State in question, deflation could also take place at a lower level. Deflation at NACE two-digit (division) level should be considered only if a lack of more detailed data prevents the use of another solution.
47 Commission Decision clarifying Annex A to Council Regulation (EC) No 2223/96 on the European system
of national and regional accounts in the Community as concerns the principles for measuring prices and volumes (OJ L 340, 16.12.98, p. 33).
48 OJ L 310, 30.11.1996, p. 349.
62
3.8.3. Weights
The weights used for production indices should generally be based on the value added of the production sector (and not on turnover or a similar measure). This is in line with Commission Regulation (EC) No 588/2001, which provides definitions for the variables in the STS Regulation. The annex to this Regulation describes the desired contents of the STS production variables as follows:
‘It is the objective of the production index to measure changes in the volume of gross output at close and regular intervals. It provides a measure of the volume trend in value added at factor over a given reference period.’
Since the aim is to measure value added, this should also be reflected in the weights, as clearly spelled out in the OECD ISP Manual (paragraph 53):
‘As the primary aim of compiling an [ISP] is to measure short-term movements in the production activity of the services part of an economy, the ISP should be defined as a weighted average of the real output of these industries, where the weights are based on their shares in the value added of the services sector.’
Depending on the particular circumstances as regards timeliness, quality and degree of detail, value added weights can be derived from SBS or from NAs.
3.9. Finding appropriate price indices for the deflation: B2B versus B2All
As outlined above, the preferred method for producing service volume indices is to deflate turnover indices with an appropriate SPPI. The indices should represent total service production, including services produced for other businesses, consumers, the state and for exports. Moreover, turnover indices in STS reflect general turnover, i.e. turnover generated with all kinds of customer, so the chosen deflators must adequately represent total transactions in an industry. Currently, however, the STS Regulation provides only for SPPIs that reflect transactions with businesses or represent businesses, so current SPPIs are B2B.49
We now discuss some possible ways of finding suitable deflators for general turnover aggregates:
− using B2B SPPIs as approximations of B2All SPPIs;
− using CPIs as approximations of B2All SPPIs; and
− combining B2B SPPIs and CPIs.
In principle, the services covered by STS (NACE H, I, J, L, M and N, with some exclusions in division M) could be sold to all kinds of customer. However, in certain areas (e.g. advertising and market research, office administration, office support and other business support activities), the B2B proportion of the transactions will clearly dominate. In other services (e.g. land and water freight transport, transport via pipelines, and rental and leasing activities), there will be a proportion of transactions with consumers, but in comparison with the bulk of the
49 In future, under the Framework Regulation for the Integration of Business Statistics (FRIBS), SPPIs will be
B2All indicators.
63
business these will be negligible in most countries. (As a rough rule of thumb a turnover share of less than 10 % with private households could be considered negligible. Some examples are: land and water freight transport, transport via pipelines, and rental and leasing activities.)
In the above industries, a B2B SPPI will in most cases be a sufficient approximation for a B2All SPPI. Nevertheless, it cannot be excluded that a different picture may emerge in discussions with industry representatives which would call for an alternative solution.
In other areas, the B2B proportion might be relatively low and a CPI could serve as a proxy, if available. Possible candidates for this approach are passenger transport, accommodation and food services – industries in which a relatively large proportion of turnover will be generated with private customers. Even in these sectors, however, there will be B2B transactions, since only the final transaction in the sales/purchase chain will be a true B2C relationship. If a hotel sells an accommodation service to a tour operator, this will be considered a B2B transaction and only the subsequent sale of the service from the tour operator to a consumer will be B2C transaction.
Where a CPI is used as an approximation for an SPPI, it should ideally be calculated excluding VAT and similar taxes. In particular, the effects of changes in tax rates should be corrected.
In many instances, B2B indicators and (VAT-adjusted) CPIs will have to be combined, in which case they should be weighted according to relative (business and consumer) shares of turnover.
Table 8 provides a simple example of how deflators might be combined in this way. Turnover shares for the various sub-industries of land transport (H49) and air transport (H51) for all transactions (B2All) are taken from SBS. The shares of B2B and B2C trade are estimated at four-digit (class) level.
It is assumed that air passenger transport is dominated by B2C transactions, so a CPI is used for deflation. Air freight transport is exclusively B2B, so an SPPI should be used.
For land transport, the situation is slightly more complex. Some sub-industries (49.20, 49.41, 49.42 and 49.50) are dominated by B2B transactions and B2C transactions are negligible; in these cases, the SPPI is the only deflator. In other sub-industries, business and private transactions are split 20:80. Here, a combination of SPPIs and consumer prices (HICP) is used. Ultimately, the industry is split into 73 % B2B transactions and 27 % B2C transactions:
(B2B = 10.2 % * 20 % + 3.5.5 % * 100 % + … + 2.6 % * 100 %).
64
Table 8: Weighting B2B SPPIs and HICP — simplified example
Source: Eurostat, SBS, EU-28 2012 [sbs_na_1a_se_r2]
Various sources could be used to determine the appropriate split of total turnover into B2B and B2C transactions:
− direct (e.g. annual) collection in the SPPI questionnaires;
− collection in the context of SBS;
− NAs (input-output tables);
− information (estimates) from business authorities; and
− information from state authorities.
4. REFERENCES FOR PART A
Abad, A.; Quilis, E. M. (2005), Software to perform temporal disaggregation of economic time series, Eurostat (statistical working paper); http://www.uni-mannheim.de/edz/pdf/eurostat/05/KS-DT-05-004-EN.pdf
Chamberlin, G. (2010), ‘Temporal disaggregation’, Economic & Labour Market Review, Office for National Statistics, pp. 106-121.
Di Fonzo, T. (2003), Temporal disaggregation of economic time series: towards a dynamic extension, Eurostat (statistical working paper); http://ec.europa.eu/eurostat/documents/3888793/5816173/KS_AN-03-035-EN.PDF
Eurostat (2011), Handbook on price and volume measures in national accounts
Eurostat, Reference manual on air transport statistics, version 11; http://ec.europa.eu/eurostat/documents/29567/3217334/Aviation-Manual-V11.pdf/19f6a843-d7c1-4cae-b93e-fcd5e3f59187
McCormack, Kevin (2013), Central Statistics Office, Ireland, How does the producer price index differ from the consumer price index? http://www.cso.ie/en/media/csoie/surveysandmethodologies/surveys/prices/documents/PPI_differ_CPI1.pdf
Turnover shares (%) B2All B2B B2C49 Land transport & transport via pipelines 100.0 73.1 26.949.10 Passenger rail transport, interurban 10.2 20.0 10049.20 Freight rail transport 3.5 100.0 0.049.31 Urban & suburban passenger land transport 12.9 20.0 8049.32 Taxi operation 3.9 20.0 8049.39 Other passenger land transport n.e.c. 6.7 20.0 8049.41 Freight transport by road 59.3 100.0 -49.42 Removal services 1.0 100.0 -49.50 Transport via pipeline 2.6 100.0 -51 Air transport 100.0 10.1 89.951.10 Passenger air transport 89.9 - 100.051.21 Freight air transport 10.1 100.0 -51.22 Space transport 0.0 - -
65
OECD (2007), Compilation manual for an index of service production
Proietti, Tommaso (2004), Temporal disaggregation by state space methods: dynamic regression methods revisited, Eurostat (statistical working paper); http://ec.europa.eu/eurostat/web/products-statistical-working-papers/-/KS-DT-04-014
Eurostat:
http://epp.eurostat.ec.europa.eu/portal/page/portal/cpa_2008/documents/CPA2008introductoryguidelinesEN.pdf
http://ec.europa.eu/eurostat/documents/cpa_2008/documents/CPA2008introductoryguidelinesEN.pdf
66
PART B INDUSTRY-SPECIFIC INFORMATION
67
5. LAND TRANSPORT (H49)
Fitwi Wolday, Statistics Norway
This section starts by outlining the industrial classification in accordance with NACE Rev. 2 for H49 (land transport including pipelines), then takes stock of available data that can be used for ISP calculation. We then discuss the methodological basis for calculating the index and make recommendations on how to produce an ISP for the industry in question.
5.1. Relevant definitions
5.1.1. Index of services production (ISP)
As the primary aim of compiling an ISP is to measure short-term movements in the production activity of industries in the service sector, the ISP should be defined as a weighted average of the real output of these industries, where the weights are based on their shares in the value added of the services sector.
The OECD ISP Manual defines an ISP as a measure of changes over time in the output of the services sector. More precisely, it is the ratio of the volume of output produced by service industries in a given period to that produced by the same industries in a specified base period. The products covered are those that contribute to the gross output of the industries and may include products that are not primary to them.
An ISP for land transport and transport via pipelines is thus an aggregate volume index where all contributing industries are weighted together. The ISP Manual indicates that weights should be based on the value added of the service sector.
5.1.2. Turnover
Turnover comprises the total value invoiced by the observation unit during the reference period, which corresponds to gross sales of goods or services supplied to third parties. It includes all duties and taxes on the goods or services invoiced by the unit, with the exception of VAT invoiced by the unit vis-à-vis its customers and other similar deductible taxes directly linked to turnover.
5.1.3. Definition of a service
Transport is a highly regulated activity in many European countries in which quasi-market activities are prevalent. Passenger transport is a typical case in point. In France, for example, inland transport is provided by the public sector. As discussed at the first meeting of the ISP Task Force, the primary objective (at least initially) is to compile an ISP for market services only. But how do we draw the line between market and non-market services when the market is less free (regulated) and the public sector is a major service-provider? ESA 1995 distinguishes market and non-market services as follows:
‘Market producers are producers that sell their output at economically significant prices. Non-market producers are producers that provide most of their output to others free or at prices that are not economically significant.
68
Furthermore, an output is only sold at economically significant prices when more than 50 % of the production cost is covered by sales.’
ESA 1995 classifies services as market or non-market according to whether the service-provider is a market or non-market agent and by source of income. In areas where public involvement is prevalent and there is ambiguity as to whether the producer is a market or non-market agent, we follow the above definition, i.e. we include services that cover at least half of their costs from sales.
5.2. Industrial classification
The applicable standard industrial classification (SIC) for this guide is NACE Rev. 2, which corresponds closely to the International Standard Industrial Classification (ISIC) Rev. 4.
Division 49 (land transport and transport via pipelines) belongs to section H (transport and storage) in both NACE Rev. 2 and ISIC Rev. 4. In the latter, it further comprises three groups, while NACE Rev. 2 allows for more detailed disaggregation. A summary table of the correspondence between the two is presented below:
Table 9: Land transport – correspondence between ISIC Rev. 4 and NACE Rev. 2
ISIC Rev. 4 NACE Rev. 2 Description Section H Section H Transportation and storage Division 49 Division 49 Land transport & transport via pipelines 4911 49.10 Passenger rail transport, interurban 4912 49.20 Freight rail transport 4921 49.31 Urban and suburban passenger land transport 4922 49.32 Taxi operation 4922 49.39 Other passenger land transport n.e.c. 4923 49.41 Freight transport by road 4923 49.42 Removal services 4930 49.50 Transport via pipeline
Division 49 comprises all industries engaged in passenger and freight transport by rail and road, and pipe transport, i.e. the transportation of goods and people across land by rail, lorry, bus or taxi and the transport of gases, liquids, etc. via pipelines. It does not include self-drive car hire or the distribution of natural and manufactured gas, water and steam from the distributor to the final user.
In NACE Rev. 2, division 49 is further sub-divided into five groups: 49.1 (passenger rail transport), 49.2 (freight rail transport), 49.3 (other passenger land transport), 49.4 (freight transport by road and removal services) and 49.5 (transport via pipeline).
Depending on the availability of data for measuring gross value added as deflated turnover and the relative importance of an industry in a given country, it may be
69
desirable to sub-divide the groups further into classes and sub-classes or by activity. For example, other passenger land transport can be further sub-divided into classes or modes of transport.
A hierarchical classification of the division is presented below. The division is sub-divided into sub-classes where appropriate. Classifications of products by activity (CPA 2008) are presented in Appendix III.
The applicable SIC in Norway (SIC 2007) corresponds to NACE Rev. 2 up to four-digit level, but allows for a further five-digit disaggregation to accommodate national peculiarities.
Table 10: Structure of H49 in NACE Rev. 2
49 Land transport and transport via pipelines 49.1 Passenger rail transport, interurban 49.2 Freight rail transport 49.3 Other passenger land transport 49.31 Urban and suburban passenger land transport 49.311 Scheduled urban & suburban motor bus transport 49.312 Other urban and suburban transport 49.32 Taxi operation 49.39 Other passenger land transport n.e.c. 49.391 Scheduled long-distance bus transport 49.392 Charters and excursions bus services 49.393 Cableway transport and ski lifts 49.4 Freight transport by road and removal services 49.41 Freight transport by road 49.42 Removal services 49.5 Transport via pipeline
Passenger rail transport, interurban (H49.1)
This group covers long-haul passenger transport by interurban railways. It includes the integrated operation of sleeping cars (wagons) or dining cars by railway companies, but not by separate units.
It does not include passenger transport by urban and suburban transit systems, such as underground, metro, trams and similar systems, nor other activities such as passenger and freight terminal activities, cargo handling, storage and other auxiliary activities, maintenance and minor repairs of rolling stock.
Freight rail transport (H49.2)
This group includes freight transport on mainline rail networks and short-line freight railroads.
It does not include the operation of railroad infrastructure, related activities such as switching and shunting, storage and warehousing, freight terminal activities or cargo handling.
70
Other passenger land transport (H49.3)
This is a broad group encompassing all land-based passenger transport activities other than rail transport (except as part of urban or suburban transport systems). The diversity and scale of activities in this group may require further disaggregation if the service provided is to be measured accurately.
Urban and suburban passenger land transport (H49.31)
This class includes land transport of passengers by urban or suburban transport systems (e.g. bus, tram, trolley-bus, underground and elevated railways, etc.) on scheduled routes normally following a fixed time schedule, entailing the picking-up and setting-down of passengers at normally fixed stops.
It includes town-to-airport or town-to-station lines, the operation of funicular railways, aerial cableways, etc. as part of urban or suburban transit systems, but not passenger transport by interurban railways.
Taxi operation (H49.32)
This class also includes other renting of private cars with a driver.
Other passenger land transport n.e.c. (H49.39)
This class includes other passenger road transport such as scheduled long-distance bus services, charters, excursions and other occasional coach services, airport shuttles and the operation of cable cars, funiculars, ski and cable lifts, if not part of urban or suburban transit systems. It also includes the operation of school buses, buses for the transport of employees and passenger transport by man- or animal-drawn vehicle, but not ambulance transport.
Freight transport by road and removal services (H49.4)
This group includes all land-based freight transport activities other than rail transport.
Freight transport by road (H49.41)
This class includes all freight transport operations by road, such as logging haulage, stock haulage, refrigerated haulage, heavy haulage, bulk haulage (including haulage in tanker trucks, e.g. milk collection at farms), haulage of automobiles and the transport of waste and waste materials, without collection or disposal.
It also includes the renting of trucks with driver and freight transport by man- or animal-drawn vehicle.
It does not include log-hauling in forests as part of logging operations, the distribution of water by truck, the operation of terminal facilities for handling freight, crating and packing activities for transport, post and courier activities, and waste transport as an integrated part of waste-collection activities.
71
Removal services (H49.42)
This class includes removal (relocation) services to businesses and households by road.
Transport via pipeline (H49.5)
This class includes the transport of gases, liquids, water, slurry and other commodities via pipelines, and the operation of pump stations. Not included are the distribution of natural or manufactured gas, steam or water and the transport of liquids by truck.
5.3. Available data
5.3.1. STS Regulation
Annex D to the STS Regulation requires all European countries to compile and report the following data on services:
− turnover;
− number of persons employed; and
− output prices.
Turnover
Register-based data on turnover are readily available for all countries, mainly from VAT registers. The level of detail and the extent to which they match available price indicators may vary between countries, so NSIs have to assess whether the statistical unit for which turnover is collected also incorporates secondary activities.
Under the STS Regulation, the observation unit is the enterprise and turnover is to be disseminated at two-digit level, i.e. division 49 in this case. NSIs usually have turnover data at a more disaggregated level (e.g. LKAU level in the case of Norway), so matching SPPI and turnover at three- or four-digit level should not be a problem where greater disaggregation is considered in order to make more accurate measurements for some industries.
The fact that turnover data for many countries are mainly register-based poses a challenge as regards periodicity and frequency. In such cases, the shortest period for which turnover can be compiled is a quarter. Statistics Norway receives bi-monthly VAT data about 45 days after the end of the relevant period.
Number of persons employed
Detailed volume data on numbers of persons employed are available for almost all European countries. Annex D to the STS Regulation requires the use of KAU as the observation unit.
72
Price indicators
The appropriate deflator for the purpose of ISP compilation would be an output price reflecting all segments of the market, i.e. households, businesses and the external sector. However, the availability of SPPIs is largely determined by the requirements of the STS Regulation. Since this focuses on a few specific areas, the coverage is not broad and is mainly B2B. As a result, the area covered by price indices in NACE Rev. 2 division 49 is only one group: freight transport by road and removal services. Compiling an ISP for division 49 will require the use of data sources outside the current STS Regulation.
For countries that compile only the B2B variant, the market structure of the service should be assessed, i.e. is it predominantly B2B or B2C?
5.3.2. Consumer price indices (CPIs)
Although CPI data are readily available for all European countries, the level of detail and coverage may vary depending on how relevant a good or service is to the average consumer. Leaving aside country-specific differences, CPIs are best suited as deflators for groups 49.1 (interurban passenger rail transport) and 49.3 (other passenger land transport), which are mainly B2C.
Statistics Norway produces CPIs for the following areas in division 49:
− passenger transport by railway, i.e. inter-urban, urban and suburban passenger transport by rail (trams, subways, etc.); and
− passenger transport by road, i.e. urban and suburban passenger transport by road (e.g. bus), other passenger transport by road (e.g. scheduled long-distance bus services) and taxi operation.
CPI compilation follows the CPA as recommended by the System of Natinonal Accounts but the aggregates contain CPA units from different NACE three-digit groups, e.g. 49.1 and 49.3 for passenger transport by railways. However, reassigning these to their respective NACE groups or classes, as recommended by the ISP Task Force, should not be difficult.
5.3.3. Structural business statistics (SBS)
Regulation (EC) No 295/2008 on SBS requires Member States to compile and transmit data on turnover and a host of other variables. However, SBS data are annual and results are to be transmitted within 18 months of the end of the calendar year of the reference period. This time lag makes it difficult to use an ISP in a meaningful way, except as a control for quality assurance purposes.
5.3.4. Other volume indicators
Other widely available indicators are passenger/kilometres and tonne/kilometres. The level of detail for the former generally tends to be somewhat limited, whereas the latter is available in more detail (often broken down by mode of transport).
73
5.4. Methodology
The Eurostat Handbook on prices and volume measures for national accounts (2001) outlines methodology for compiling price and volume measures for the division and provides useful insights into the challenges involved. It also distinguishes between superior and inferior methods and classifies them as:
− A methods (the most appropriate);
− B methods (can be used where an A method cannot be applied); and
− C methods (considered inferior and not to be used).
5.4.1. Statistical unit
The transport sector (e.g. typically, ‘other passenger land transport’ and ‘freight transport by road’) is populated by many SMEs which provide services with a range of terms and specifications. This makes it difficult to collect service production information on a regular basis for an extended period, as the population landscape is constantly changing.
Similarly, a firm can engage in multiple activities across various sectors and a service activity may be the secondary activity of a firm whose predominant activity is another service activity or a non-service activity (e.g. manufacturing).
Therefore, the establishment is the preferred statistical unit. Where the type of service provided is very diverse, KAU could be an alternative primary or secondary information source. KAU is more appropriate for SPPI compilation than as a source of turnover information, because VAT is mostly organised by enterprise (see below). In the case of Norway, however, the establishment (LKAU) would be more appropriate, as turnover by establishment can be obtained from the VAT register.
The LKAU corresponds to the operational definition of an establishment, i.e. (part of) an enterprise or an enterprise that engages in (predominantly) one kind of economic activity at or from one location or within one geographical area, for which available data can meaningfully be compiled that allow the calculation of operating surplus.
5.4.2. Price concept
The Eurostat-OECD Methodological guide for developing producer price indices for services (SPPI guide) recommends that SPPIs should measure actual transaction prices reflecting the revenue received by the producer for products actually sold to customers. They should take into account any applicable discounts, rebates, surcharges, etc. that may apply to the customers. Consequently, the prices should exclude taxes on products, but include subsidies on products received by the producer. In other words, the recommendation is to use basic prices (the unit price received by the service producer) when measuring SPPIs.
74
5.4.3. Passenger transport
While using various modes of transport, passenger transport services all involve moving people from one place to another. However, quality differences should be accounted for within the different forms of travel (rail and road). Factors such as speed, convenience, comfort, reliability, timeliness should be considered, as shifts can occur that must be reflected correctly in price or volume. This calls for detailed measurement.
Urban and suburban passenger transport largely involve integrated modes of transport, but pricing mechanisms (ticket prices) are often time-based and do not differentiate between modes. For example, commuters can buy a ticket that enables them to use all modes of public transport in a specified geographical area (zone). In interurban passenger transport, on the other hand, a further disaggregation by mode (e.g. train versus bus) becomes important, as quality issues such as comfort and timeliness affect pricing mechanisms.
Some services, principally those by bus and train, can be purchased in different ways, e.g. a one-off ticket or a season ticket that gives the right to travel any number of times within a certain period. Each type of ticket could be seen as a separate product, giving different rights under varying conditions.
‘Output in current prices’ is essentially equal to the number of tickets sold multiplied by the price per ticket, for all different types of ticket, in the base period. This approach is also followed in the compilation of price indices, where the prices of different types of ticket are usually observed and weighted with total sales or purchases of each type in the base year.
The consequences of this approach should be carefully considered. It implies, for example, that a kilometre travelled by train on a one-off ticket is a different output from a kilometre travelled on a season ticket. It also implies that a train travelling without passengers still produces output, since season-ticket holders have already paid. In this case, output may be seen as being directly related not to the use of the transport service, but to the right of access to it.
In many instances, volume data on the number of passenger/kilometres produced are available and used as volume indicators in NAs. This typically assumes that only the actual transport of passengers yields output (i.e. an empty train does not produce output). The following formula can be used to compare the variation in passenger/kilometres and tickets sold (per type of ticket):
Number of tickets sold * average number of journeys per ticket * average distance per journey = number of passenger/kilometres for this type of ticket
We can easily see that, if (for one type of ticket) the travel habits of the population do not change over time, changes in the two indicators (tickets sold and passenger-kilometres) will show the same trend. For single (one-way) tickets, counting passenger-kilometres and the number of such tickets sold will give similar results (since the average number of journeys per ticket is by definition equal to one). However, where a ticket permits travel within a pre-defined area, rather than simply from one location to another (e.g. ‘city travel’ tickets), the
75
relation between number of tickets sold and passenger-kilometres depends on changes in average distance travelled per ticket.
For season tickets, the relation between passenger-kilometres and tickets sold depends both on the average number of journeys per ticket and the average distance travelled each time. It is possible to imagine a case where the ticket holder makes no actual journeys, i.e. the number of passenger-kilometres is zero, but output is still produced.
The opposite case is that of free travel, e.g. for students or elderly people. Such journeys will be included in the number of passenger-kilometres, but not in an index based on tickets sold. One could question, however, whether this travel is really free-of-charge, as the government will usually pay the transport company for it.
Tickets restricted to peak or off-peak, or first- or second-class travel, are characterised by different degrees of product quality (which we have already identified as an important issue for transport services). The prices of different types of ticket are readily available and conditions of sale can be monitored relatively easily. This should allow the correct estimation of price and volume, provided output at current prices can be identified separately for each type of ticket.
Discounts are often offered for passenger transport services, e.g. on presentation of a discount card. This also represents a different quality of product that needs to be reflected correctly in prices and volumes.
A significant complication in the deflation of output of passenger transport is frequent subsidisation. While transport companies have become market producers in most countries, they are often subsidised in order to keep prices affordable or to keep unprofitable routes open. The nature of the subsidies can vary considerably between countries. In some cases, they are paid per passenger-kilometre produced; in others, they take the form of a lump sum. ESA 1995 states (paragraph 4.35(c)) that such subsidies are to be seen as subsidies on products (D319) and are therefore part of the basic price.
5.4.4. Recommended methods for passenger transport
Given the very diverse nature of the products in the CPA for passenger transport, the Eurostat Handbook on price and volume measures on national accounts (2001) recommends that at least two areas be identified separately, provided that each makes a significant contribution to output:
− passenger transport via railways; and
− other passenger land transport.
Any method not involving this breakdown will be a C (inferior) method.
CPIs for transport services provided to households are generally available, but SPPIs for passenger transport are not. Volume indicators such as passenger-kilometres are widely available, but the level of detail may be limited.
76
The Handbook recommends the use of appropriate SPPIs as an A method. These would need to cover the various types of ticket available. The use of a CPI adjusted to basic prices for passenger transport is therefore regarded as a next-best (B) method, provided it takes sufficient account of quality changes. Subsidies (taxes and subsidies) are widespread in passenger transport and for a CPI to be used as a deflator in ISP compilation, it has to be made compatible with the concept of basic price, i.e. the effects of subsidies (taxes) have to be accounted for. The OECD ISP Manual adds that, where an industry’s output is consumed almost entirely by the household sector, the use of appropriately adjusted components of the CPI (i.e. adjusted to basic prices and combined using production value weights) is a preferred method.
Volume (passenger-kilometre) indicator methods are B methods, since it is to be assumed that the travel habits of the population do not change significantly. The more detailed these indicators, the better will be the result. For passenger transport, the indicators should at least distinguish between the different classes of travel, e.g. first and second class for rail transport.
Other methods, such as the use of volume indicator methods based only on the number of passengers transported, are C methods.
5.4.5. Freight transport
For freight transport, it is more difficult to compile price indices, because there are generally no fixed tariffs. The price for transporting goods from A to B can depend on a range of factors, such as distance, weight of the freight, domestic or international route, kind of product transported, etc. With the product specification constantly shifting, one can say that freight transport services have (to some extent, at least) the characteristics of one-off products.
SPPIs for freight transport by road are available in the STS Regulation, but freight transport by rail is still scarcely covered by SPPIs. On the other hand, volume (tonne-kilometre) data are widely available, generally broken down by mode of transport.
The Handbook on price and volume measures on national accounts recommends that at least rail transport, freight transport by road and transport via pipelines be identified separately, subject to each product making a significant contribution to output. Any method not involving this breakdown will be an inferior method.
By virtue of their inclusion in the STS Regulation, best practice in developing SPPIs for freight transport by road is well documented and widely available. The focus here will therefore be on freight transport by rail.
The summary of preferred methodologies below is drawn from the Voorburg Group’s sector paper based on the experience of Germany, Canada and Finland (Barzyk and Goldhammer, 2010).
In Germany, price-determining factors (wagon load, market segment, goods transported, wagon type and ownership, place of departure and arrival, intermodal transport/traction for block trains or wagon-load traffic/servicing of junctions,
77
stations/junctions serviced, etc.) are identified and a combination pricing method is used depending on the service:
− contract pricing for main haulage services;
− pricing of repeated services for construction, train and shunting services; and
− unit-value pricing for track construction services.
Canada, on the other hand, identifies a representative list of commodities transported and main routes based on the industry’s turnover. List/tariff prices for the commodities are posted on the internet.
Finland uses a unit-value method, with the respondent calculating the average price per tonne-kilometre. For this method to be used, however, detailed data are required so as to guarantee sufficient homogeneity of transactions. If the data are too general or broad, it is often difficult to determine what the source of ‘price change’ is, e.g. revenues, tonne-kilometres or both. In the absence of controls on shifts in the mix of items, changes within a group will be erroneously recorded as price changes.
Other preferred methods recommended by the Handbook on price and volume measures on national accounts are:
− model pricing, where a set of representative standard journeys could be formulated which would be priced by the producers of the services; and
− hedonic pricing, where the price of each journey is regressed on its characteristics.
Both approaches cater for the uniqueness of the services provided and monitor for quality aspects of the journeys, but the hedonic approach is very demanding to apply in practice, as it requires very detailed data.
5.4.6. Transport via pipelines
As the population for this group is small and the enterprises are often large, surveys can be an effective way of collecting price and turnover data. Volume indicators and tariffs are readily available. VAT data can also be used where necessary.
In2005, the EU-27 ‘transport via pipelines’ sector was made up of approximately 130 enterprises, which together generated EUR 11.0 billion of turnover (1.0 % of total transport service turnover).50
Preferred methods in the OECD ISP Manual are the use of turnover from surveys of appropriate units or administrative data, appropriate weighted price indices (e.g. PPIs for crude petroleum oil or road haulage) and certain volume indicators.
In the case of Norway, the group is entirely dominated by a few companies in the oil industry. In addition to VAT data, Statistics Norway collects monthly volume
50 Source: Eurostat, Statistics Explained, article on pipeline transport.
78
data (in standard m3 for gas and in barrels for oil) and turnover data at enterprise level. The volume data show actual volumes transported through the network of pipelines (allocated amount) and the length of the pipelines. The tariff paid, on the other hand, is based on volumes ordered. Since order contracts are signed some time (often years) ahead of delivery, volume delivered may differ from volume ordered; focusing on the former and on tariff paid accurately reflects the price actually paid.
5.5. Compiling an ISP
An ISP measures changes over time in the volume of output of the services sector. Eurostat favours gross constant price output indicators, i.e. deflated turnover, as the best proxy for short-term change in constant-price value -added. Turnover/sales deflated by an appropriate output price index is therefore the preferred approach (A method according to the Handbook). The price index used should be representative of the NACE activity being deflated. On the other hand, the Handbook generally refers to volume measures as B methods. However, if there is a detailed breakdown by type of commodity (to ensure reasonable homogeneity) and very little change in quality, a volume indicator (e.g. transport via pipelines) could be classified as an A method. ‘Input’ indicators (e.g. employment) are classified as C (least favoured) indicators, because they do not adequately detect changes in productivity.
Although the use of output price index as a deflator is the ideal choice in an ISP, in practice NSIs will have to combine price indices from different sources. In order to develop a representative output price index for use as a deflator, these have to be weighted together at a detailed level using production value data (ideally, from NA input-output or supply-and-use tables).
For ISPs for division 49 (at three-digit level), it may not be necessary to combine indices from different sources (SPPI, CPI, etc.), if one rules out the export sector, as the services in question are predominantly either B2C or B2B.
5.5.1. Theoretical basis
The ISP Manual recommends a chained Laspeyres index as a preferred method for compiling an ISP. Principle 2 in the Handbook states that volume measures available at the elementary level of aggregation are to be aggregated using the Laspeyres formula to obtain volume measures of all NA aggregates.
Weighting
Lower-level ISPs are aggregated into their respective higher NACE levels using gross value added as weights to represent their contribution to GDP. Information on gross value added can be obtained from input-output tables.
The appropriate frequency for updating weights depends on the rate at which innovation and other change alter production techniques in an industry. NACE Rev. 2 H49 is considered relatively stable and a three-to-five year interval would be appropriate.
79
Missing input variables (indices or weights)
When compiling an index of this magnitude, compilers will often be confronted with information gaps; a price or weight component might be missing from a given elementary group. The ISP Manual for an index of service production provides important methods for mitigating the effects of missing information.
Missing indices
If an index is not available for a lower-level activity, the weight of the activity should be distributed proportionately across the other activities that contribute to the same activity one level higher in the activity classification.
Missing weights
In the case of missing weights (e.g. disaggregated figures for value added do not cover the whole service sector), the preferred method is to apply the Jevons index (a ratio of geometric means of volume representatives), which gives individual volume components the same weight.
5.6. References
Barzyk, Fred and Goldhammer, Bernhard (2010), Sector paper on rail freight transportation services, 25th meeting of the Voorburg Group.
Eurostat (2001), Handbook on price and volume measures in national accounts, Office for official publications of the European Communities, Luxembourg.
Eurostat (2006), Methodology of short-term business statistics: interpretation and guidelines, European Commission and Eurostat.
Eurostat (2006), Methodology of short-term business statistics: associated documents, European Commission and Eurostat.
Eurostat (2008), CPA 2008: structure and explanatory notes, Eurostat.
Eurostat (2008), NACE Rev. 2: statistical classification of economic activities in the European Community, Eurostat methodologies and working papers, European Commission and Eurostat.
Eurostat (2009), PEEIs in focus: a summary for the index of turnover in services, Eurostat methodologies and working papers, European Commission and Eurostat.
IMF et al. (2004), Producer price index manual: theory and practice, IMF, Washington D. C.
Johannessen, R. et al. (2001), Consumer price index 1995-2000, Statistics Norway, Official Statistics of Norway.
Langer, T. and Murphy, J.(2009), Glossary of Voorburg Group terminology for mini-presentations, sector papers and revisited sector papers, Voorburg Group on Service Statistics.
OECD and Eurostat (2005), Methodological guide for developing producer price indices for services, OECD, Paris.
80
OECD (2007), Compilation manual for an index of service production, OECD, Paris.
United Nations (2008), International standard industrial classification of all economic activities, Rev. 4, New York
81
5.7. Annex
Table 11: The road transport sector in the EU
Table 12: CPA 2008 — land and pipeline transport services
49.1 Passenger rail transport services, interurban 49.10 Passenger rail transport services, interurban 49.10.1 Passenger rail transport services, interurban 49.10.11 Passenger rail transport services for sightseeing 49.10.19 Other passenger rail transport services, interurban 49.2 Freight rail transport services 49.20 Freight rail transport services
49.20.11 Railway transport services of freight by refrigerator cars 49.20.12 Railway transport services of freight by tanker cars, petroleum products 49.20.13 Railway transport services of freight by tanker cars, bulk liquids and gases
49.20.14 Railway transport services of intermodal containers 49.20.15 Railway transport services of letters and parcels 49.20.16 Railway transport services of dry bulk goods 49.20.19 Other railway transport services of freight 49.3 Other passenger land transport services 49.31 Urban and suburban passenger land transport services
49.31.10 Urban and suburban railway transport services of passengers 49.31.2 Other urban and suburban passenger land transport services
49.31.21 Urban and suburban scheduled road transport services of passengers 49.31.22 Mixed mode urban and suburban scheduled transport services of passengers
49.32 Taxi operation services 49.32.11 Taxi services 49.32.12 Rental services of passenger cars with driver 49.39 Other passenger land transport services n.e.c.
49.39.1 Interurban and special-purpose scheduled passenger
Number of enterprises
Turnover or gross premiums
written
Value added at factor cost
Number of persons
employedH49 Land transport and transport via pipelines 917,255 518,162.0 207,339.7 55,396
H49.1 Passenger rail transport, interurban 266 52,659.9 26,979.8 4,278
H49.2 Freight rail transport 538 18,332.1 6,051.6 1,492
H49.3 Other passenger land transport 343,094 121,291.9 65,720.5 19,885
H49.31 Urban and suburban passenger land transport 12,856 66,581.0 35,147.6 8,519
H49.32 Taxi operation 286,017 20,219.3 12,215.3 5,948
H49.39 Other passenger land transport n.e.c. 44,224 34,491.6 18,357.5 5,418
H49.4 Freight transport by road and removal services 573,148 312,121.2 99,887.5 29,457
H49.41 Freight transport by road 562,113 306,966.5 97,631.3 28,777
H49.42 Removal services 11,032 5,154.7 2,256.1 680
H49.5 Transport via pipeline 209 13,756.9 8,700.3 284
82
land transport services 49.39.11 Interurban scheduled road transport services of passengers 49.39.12 Interurban special-purpose scheduled road transport services of passengers 49.39.13 Other special-purpose scheduled road transport services of passengers
49.39.2 Passenger transport services by funiculars, teleferics and ski-lifts
49.39.20 Passenger transport services by funiculars, teleferics and ski-lifts
49.39.3 Non-scheduled passenger land transport services 49.39.31 Rental services of buses and coaches with driver
49.39.32 Road transport services of passengers for sightseeing 49.39.33 Non-scheduled local bus and coach charter services 49.39.34 Non-scheduled long distance bus and coach charter services 49.39.35 Road transport services of passengers by man- or animal-drawn vehicles
49.39.39 Passenger land transport services n.e.c. 49.4 Freight transport services by road and removal services 49.41 Freight transport services by road 49.41.1 Freight transport services by road
49.41.11 Road transport services of freight by refrigerator vehicles 49.41.12 Road transport services of freight by tank trucks or semi-trailers, petroleum products 49.41.13 Road transport services of freight by tank trucks or semi-trailers, other bulk liquids or gases
49.41.14 Road transport services of intermodal containers 49.41.15 Road transport services of dry bulk goods 49.41.16 Road transport services of live animals
49.41.17 Road transport services of freight by man- or animal-drawn vehicles
49.41.18 Road transport services of letters and parcels 49.41.19 Other road transport services of freight 49.41.2 Rental services of trucks with operator 49.41.20 Rental services of trucks with operator 49.42 Removal services 49.42.1 Removal services 49.42.11 Removal services for households 49.42.19 Other removal services 49.5 Transport services via pipeline 49.50 Transport services via pipeline
49.50.1 Transport services via pipeline
83
6. WATER AND AIR TRANSPORT (H50 & H51)
Søren Kühl Andersen, Statistics Denmark
6.1. Description of the water and air transport sector
H50 and H51 are fairly heterogeneous sectors characterised by large B2B and B2C markets, which is reflected in the NACE sub-divisions. Water transport is divided into four groups on the basis of two variables: passenger vs freight, and sea/coastal vs inland. Air transport is divided into three groups: passenger air transport, freight air transport and space transport (e.g. the launching of commercial satellites).
Table 13: NACE structure for H50 and H51
H50 Water transport H50.1 Sea and coastal passenger water transport H50.2 Sea and coastal freight water transport H50.3 Inland passenger water transport H50.4 Inland freight water transport
H51 Air transport H51.10 Passenger air transport H51.21 Freight air transport H51.22 Space transport
6.2. Available methodological information
6.2.1. OECD ISP for an index of service production
The material of most direct relevance to the ISP for water and air transport is the OECD ISP Manual. While the main text deals with general rather than sector-specific methodological issues, Annex 2 contains a very useful overview of various possibilities for calculating the ISP within each sector. This describes two alternative approaches for each section and sub-section: a deflated turnover approach (based on a survey or administrative data) and a volume approach. For each, it presents a preferred and an alternative method.
For the deflated turnover approach, the difference between the preferred and the alternative method lies in the price indices used as deflators. The preferred method is to use specific price indices, e.g. sea travel CPI or sea freight PPI. The use of less specific price indices, e.g. a more general CPI, is considered an alternative method. The choice of one over the other will probably depend on their availability.
For the volume approach, the difference between the preferred and the alternative method is generally that the former is weighted by distance, while the latter is not. An example is number of passenger-kilometres vs number of passengers.
84
6.2.2. The Eurostat-OECD SPPI Guide
The Eurostat-OECD Methodological guide for developing Producer Price Indices for services (SPPI guide) deals with one of the main elements of the ISP, the deflator, and contains quite a thorough description of the characteristics of each of the sectors covered here. It gives a short description of each sector and discusses sampling, pricing methods and various quality issues. The Guide draws on the work in the Voorburg Group.
For sea transport, the main focus of the Guide is freight transport prices, as it is argued that this is the most significant component of an SPPI for sea transport.
For air transport, on the other hand, the main focus is on passenger transport and one of the issues addressed is how to distinguish between leisure and business travel. The importance of this issue depends of course on how different these markets really are.
The Guide also describes various countries’ methods in each sector.
6.2.3. Voorburg Group
The Voorburg Group has also dealt with sea and air transport and many methodological papers are available on its website, including over 100 papers and presentations on the transport sector. Much of the focus is on prices (probably reflecting the work carried out at the time to draft the SPPI Handbook) and the measurement of output/turnover.
6.3. Available data
6.3.1. Transport statistics
There is quite a lot of data available on water transport and air transport. However, one should bear in mind that transport statistics and STS are compiled from different perspectives. While the main aim of the latter would be to measure the activity of businesses in the transport sector, the former focus on measuring the general level of activity, regardless of where the providers are located. For example, air transport statistics focus on the number of passengers arriving at and leaving from European airports; in STS, the focus would be on the number of passengers carried by airlines listed in European business registers.
These differences notwithstanding, there should be possibilities for comparison or benchmarking between transport statistics and the ISP for the transport sector.
6.3.2. ‘Foreign trade in services’ statistics
Exports of services rendered by businesses in sections H50 and H51 are covered in ‘foreign trade in services’ statistics. As the sectors are characterised by many large operators, the volumes will be quite substantial. The data could be used as an input or a quality check for the ISP, but will need to be supplemented with data on the domestic market.
85
One possibility could be to supplement the ‘foreign trade in services’ data with VAT data so as to cover both domestic and non-domestic trade.
6.3.3. The case of Denmark
For Denmark, by far the largest proportion of freight transport revenue is generated on the non-domestic market, as there are some very large shipping companies. These data are already collected for foreign trade in services. The remaining part of freight transport should be covered by Danish VAT data. However, VAT data cannot be used as a source for passenger transport, which is VAT-exempt.
6.3.4. Services turnover and SPPIs in STS
The service turnover variable covers sectors H50 and H51, but data are quarterly rather than monthly. They are presented as an index, which allows for comparisons of developments for different aggregates, but not for a description of the structure of the sector, e.g. in terms of share of turnover per Member State and NACE section. However, as Member States are required to collect turnover data from these sectors, quarterly turnover data should be available by country.
The STS output (producer) price variable in Annex D covers H51 and the sum of H50.1 and H50.2, i.e. the SPPI is aimed at two-digit level for these divisions, which will pose some challenges in terms of deflating turnover from areas as different as passenger and freight transport. Also, division H50 is only partially covered, as inland water transport is excluded. So, while there are SPPIs available in this area, the current coverage does not allow us to calculate series beyond the two-digit level.
6.3.5. Structural business statistics (SBS)
SBS contain more detailed and comprehensive data for divisions H50 and H51, e.g. value added, turnover, number of enterprises and number of employed persons.
6.3.6. Conclusions
Data are available, but there are issues of coverage and detail, so the options seem to be:
� to create an ISP within the current limitations; or
� to attempt to fill in the blanks.
Using foreign trade data, either as direct input or as a benchmark, could be an interesting option, but this will have to be investigated further.
6.4. Further description of H50 and H51
Table 14 shows turnover, or gross premiums written, for the EU-28, using data from Eurostat’s database. Sectors H50 and H51 are roughly the same size in terms of turnover, but the balance between passenger and freight transport is quite different.
86
Table 14: Water and air transport — turnover by sector (mio. EUR), 2011-2013 (EU-28)
NACE Description 2011 2012 2013 H50 Water transport 112 313 117 160 108 681 H50.1 Sea and coastal passenger water
transport 18 474 18 553 17 454
H50.2 Sea and coastal freight water transport 86 798 91 424 83 930 H50.3 Inland passenger water transport 1 639 1 568 1 728 H50.4 Inland freight water transport 5 378 5 595 5 520 H51 Air transport 126 808 133 878 132 795 H51.10 Passenger air transport 113 770 120 321 n.a H51.21 Freight air transport c c c H51.22 Space transport c c c Total Water and air transport 239 121 251 038 241 476 c - confidential In sector H50, freight transport accounts for most of the turnover; it generates approximately 80 % of turnover in water transport and most of this is from sea and coastal freight water transport.
In sector H51, the opposite applies: passenger transport accounts for most of the turnover by far. Around 90 % is generated by passenger air transport and the remaining 10 % by freight and space transport.
Table 15 shows the number of enterprises at EU-28 level in each sector. These probably do not equate to what are generally understood as ‘businesses’, as many operate internationally and have subsidiaries in each country. Airlines, for instance, typically have ticket sales and service offices in a number of countries.
Table 15: Water and air transport — number of enterprises, 2011-2012 (EU-28)
NACE Description 2011 2012 H50 Water transport 21 306 20 973 H50.1 Sea and costal passenger water transport 6 000 6 044 H50.2 Sea and costal freight transport 5 756 5 494 H50.3 Inland passenger water transport 3 482 3 489 H50.4 Inland freight water transport 6 000 5 941 H51 Air transport 4 000 4 130 H51.10 Passenger air transport c c H51.21 Freight air transport 593 635 H51.22 Space transport c c Total Water and air transport 25 306 25 103
c - confidential
Sector H50.2 (sea and coastal freight water transport) is dominated by large operators. In 2014, the 25 largest accounted for approximately 24 % of world gross tonnage. Of these, 10 are based in Europe: two in Denmark, two in
87
Switzerland, three in Germany, one in France, one in the UK and one in Greece (data from the Danish Ship Owners Association). A different picture emerges if one looks at national merchant fleets: Panama and Liberia account for a very large proportion of the world’s merchant fleet. Malta and Greece account for over 45 % of the EU merchant fleet.
For STS purposes, the population would be enterprises listed in national business register rather than ships listed in the national register. Sea transport businesses typically operate a mixture of own ships (some registered in a foreign country) and leased ships. For instance, the largest Danish ship owner, AP Møller-Mærsk, was operating 499 ships in April 2015, of which it owned 413. Businesses in sector H50.4 will tend to be smaller and operate on a more local or regional basis, so there would probably not be the same issue regarding the registration of ships in foreign countries.
Table 16 shows a breakdown of the world merchant fleet according to the country of the operator. Greece accounts for nearly 12 % of the total gross tonnage (GT), Germany for nearly 6 % and Denmark for 4 %. Quite a few other European countries are in the top 25. The table includes passenger ships, but they represent a relatively small proportion of total GT.
88
Table 16: World merchant fleet by country of operator (April 2015)
Country GT (1 000 t) GT (%) No. of ships Greece 138 667 11.8 3 973 Japan 133 152 11.4 7 154 China 92 028 7.9 5 104 Germany 66 332 5.7 3 035 USA 65 998 5.6 4 724 Singapore 64 875 5.5 4 350 South Korea 62 167 5.3 2 909 Denmark 48 228 4.1 1 807 Hong Kong 46 252 3.9 1 761 Switzerland 38 997 3.3 797 Great Britain 37 641 3.2 1 803 Norway 36 318 3.1 2 281 Taiwan 33 914 2.9 1 218 France 19 958 1.7 838 Bermuda 19 303 1.7 349 Italy 16 808 1.4 1 459 Brazil 15 200 1.3 695 Turkey 14 439 1.2 2 069 Russia 14 344 1.2 2 462 Netherlands 13 442 1.2 1 803 Belgium 13 231 1.1 476 Indonesia 13 090 1.1 8 211 India 11 972 1.0 1 456 UAE 11 348 1.0 1 870 Iran 10 444 0.9 806 Other countries 134 441 11.5 23 528 Total 1 172 589 100 86 938
Source: Danish Ship Owners Association. Data include passenger ships.
Sea transport freight enterprises typically operate on a world market, where the price structure can be quite different from area to area. Also, price structure and price development will depend on type of freight, e.g. container vs dry bulk. It might therefore be necessary to collect prices for different regions (and different types of cargo) from the shipping companies or to collect a weighted price that takes into account the difference in price structures.
Although there are differences between Member States in terms of how large a proportion of water transport is inland water transport , in most it tends to be negligible. The markets in which inland water transport enterprises operate are by definition more local than sea and coastal water transport markets.
Passenger transport also breaks down into sea and coastal transport on the one hand and inland transport on the other. As with freight transport, the former accounts for a much larger proportion than the latter.
89
Apart from the distinction between inland and sea, which is reflected in the NACE structure, there is also an important distinction between leisure travel (e.g. cruises) and regular transport from A to B. Clearly, the two markets (and their prices) will be quite different, so a single CPI is unlikely to be adequate.
Passenger ships (more specifically, ‘ro-ro’ vessels) typically carry freight (trucks) as well as passengers, so belong to a segment entirely different from passenger transport.
6.4.1. Air transport
As seen above, passenger transport is by far the largest division within air transport. The passenger air transport market is highly competitive and characterised by relatively few large enterprises. In more recent years, the structure has changed, as emerging low-cost airlines have challenged national airlines.
Figure 15 (data from Eurocontrol’s 2012 guide to European air traffic management) shows flights per day and various segments’ market share (i.e. share of flights) in 2011.
Low-cost airlines account for 22.3 % of flights, while traditional scheduled flights account for 57.9 %. Charter flights constitute 5.2 %, but there is probably a very strong seasonal element.
Figure 15: The market for air transport
Source: Eurocontrol
6.5. Test calculation of an ISP for H50 and H51
It is possible to calculate a simple quarterly ISP by simply deflating the turnover index by the SPPI. Figure 16 shows the result of this calculation at two-digit level. As mentioned above, there are some methodological issues with this approach and aspects to investigate further, including:
− the coverage of the SPPI vs the coverage of the turnover index; and
− the level of detail used for calculations.
This is simply the working-day adjusted turnover index divided by the corresponding SPPI for the EU-28 aggregate.
Seasonality in the air transport sector is obviously (and not surprisingly) very strong, whereas production in the water transport sector is less affected by
90
seasonality. This is in line with the sectors’ very different passenger transport ratios.
Figure 16: Experimental ISPs for water and air transport (EU-28)
70
80
90
100
110
120
130
140
Water transport Air transport
91
7. WAREHOUSING AND SUPPORT ACTIVITIES FOR TRANSPORTATION (H52)
Tomas Rudys, Statistics Lithuania
In this paper, we give a short description of warehousing and transportation support activities (NACE Rev. 2 H52). First, we describe available methodological information and then we give a more detailed presentation of the H52 industry. We analyse available data in individual countries and Europe as a whole and present an experimental calculation of ISP for H52 in Lithuania.
7.1. Definition of an ISP
As the primary aim of compiling an ISP is to measure short-term movements in the production activity of the services part of an economy, the ISP should be defined as a weighted average of the real output of these industries, where the weights are based on their shares in the value added of the services sector. The OECD ISP Manual defines an ISP a measure of change over time in the volume of output of the services sector. More precisely, it is the ratio of the volume of output produced by the services industries in a given time period to the volume produced by the same industries in a specified base period. The products covered are all goods or services that contribute to the gross output of the services industries, and may not be primary products for the industries in question.
7.2. Available methodological information
The ISP Manual is the most comprehensive document that deals directly with how to compile an ISP. It was prepared primarily to provide official statisticians with practical guidelines for compiling a short-term ISP, rather than just to discuss methodological aspects of measuring service activity. The recommendations are therefore just as relevant for compiling a quarterly ISP.
In the past, the Voorburg Group paid little attention to the warehousing and storage sector. Not until 2004 were two papers on SPPIs for warehousing and storage compiled in the United States and New Zealand. As regards turnover, there was no paper on this sector until 2010, when a ‘revisited sector paper’ on ISIC Rev. 4 52.10 (warehousing and storage) was presented at the Group’s 25th meeting. This identified the challenges associated with the classification of warehousing and storage, the collection of turnover data and developing PPIs. It put forward options, highlighted challenges and noted the implications of the choices to be made by those seeking to develop or revise turnover statistics and price deflators for warehousing and storage.
92
Table 17: Methodological information on H52 from the Voorburg Group
Year, location
Title Author Indicator Industry
2004, Ottawa
Cargo handling, storage and warehousing — New Zealand PPI
Statistics New Zealand
Prices Transportation and storage
2004, Ottawa
Cargo handling, storage and warehousing in the New Zealand PPIs (paper)
Parbhu, Charissa
Prices Miscellaneous business services
2004, Ottawa
Refrigerated warehousing and storage - NAICS 493120 (ppp)
Lucier, John L.
Prices Transportation and storage
2010, Vienna
Revisited sector paper on warehousing and storage services
Goldhammer, Bernhard
Revisited sector paper
Transportation and storage
2004, Ottawa
US PPI for refrigerated warehousing and storage (NAICS 493120)
Lucier, John L.
Prices Miscellaneous business services
7.3. Activity classification: NACE Rev. 2 H52
This division includes warehousing and support activities for transportation, such as the operation of transport infrastructure (e.g. airports, harbours, tunnels, bridges, etc.), the activities of transport agencies and cargo handling.
Warehousing and storage (H52.1/52.10)
This group/class includes the operation of storage and warehouse facilities for all kinds of goods (grain silos, general merchandise warehouses, refrigerated warehouses, storage tanks, etc.), the storage of goods in foreign trade zones, and blast-freezing.
Support activities for transportation (H52.2)
This group includes activities supporting the transport of passengers or freight, such as the operation of parts of the transport infrastructure or activities relating to the handling of freight immediately before or after transport or between transport segments. It includes the operation and maintenance of all transport facilities.
93
Table 18: Warehousing and transport support activities in ISIC and CPA
Division Group Class Description ISIC Rev. 4
CPA 2008
52 Warehousing and support activities for transportation
52 52
52.1 Warehousing and storage 521 52.1 52.10 Warehousing and storage 5210 52.10 52.2 Support activities for
transportation 52.2
52.21 Service activities incidental to land transportation
5221 52.21
52.22 Service activities incidental to water transportation
5222 52.22
52.23 Service activities incidental to air transportation
5223 52.23
52.24 Cargo handling 5224 52.24 52.29 Other transportation support
activities 5229 52.29
Service activities incidental to land transportation (H52.21)
This class includes activities relating to the transport of passengers, animals or freight by land, e.g. the operation of terminal facilities such as railway stations, bus stations, stations for handling goods, railroad infrastructure, roads, bridges, tunnels, car parks or garages, bicycle parking lots and the winter storage of caravans. It also includes switching and shunting, towing and roadside assistance, and the liquefaction of gas for transportation purposes.
Service activities incidental to water transportation (H52.22)
This class includes activities relating to the transport of passengers, animals or freight by water, e.g. the operation of terminal facilities such as harbours and piers, waterway locks, etc., navigation, pilotage and berthing activities, lighterage, salvage activities and the operation of lighthouses.
Service activities incidental to air transportation (H52.23)
This class includes activities relating to transport of passengers, animals or freight by air, e.g. the operation of terminal facilities such as airway terminals, and airport, air traffic control and ground-service activities. It includes firefighting and fire-prevention services at airports.
Cargo handling (H52.24)
This class includes the loading and unloading of goods or passengers’ luggage (irrespective of the mode of transport), stevedoring, and the loading and unloading of freight railway cars.
94
Other transportation support activities (H52.29)
This class includes a diverse range of activities such as the forwarding of freight, the arranging or organising of transport operations by rail, road, sea or air, the organisation of group and individual consignments (including pickup and delivery of goods and grouping of consignments), the issuing and procurement of transport documents and waybills, the activities of customs agents, sea-freight forwarders and air-cargo agents, brokerage for ship and aircraft space, goods-handling operations, e.g. temporary crating for the sole purpose of protecting the goods during transit, uncarting, sampling, weighing goods, etc.
7.4. Structure of the industry
In 2010, 127 600 enterprises with 24 900 employees were engaged in warehousing and support activities for transportation (H52) in the EU. These figures correspond to 0.6 % and 1.9% of the respective totals for the business economy as a whole. The sector accounted for 2.6 % of the total business aggregate for value added at factor cost and 1.8 % of total turnover.
Table 19: Size of group H52.2. in the EU
Source: Eurostat (online data code: sbs_na_sca_r2)
EU macroeconomic indicators and a breakdown by employment size classes are available for 2010. Over 80 % of enterprises in the sector had nine or fewer employees (see Table 20). Enterprises with 250 or more employees accounted for 50.2 % of employment in the sector, 42.5 % of its turnover and over 56 % of its value added.
95
Table 20: Macroeconomic indicators of warehousing and support activities for transportation (EU-27), 2010
Number of enterprises
Turnover or gross premium
written
Value added at factor costs
Number of persons
employed Size (no. of employees)
no. % EUR million
% EUR million
% no. %
Total 127 605 100.0 432 000 100.0 157 044 100.0 24 900 100.0 0 - 9 102 351 80.2 n.a. 13.7 n.a. 8.6 2 558 10.3 10 -19 10 691 8.4 32 000 7.4 8 997 5.7 1 500 6.0 20 - 49 8 200 6.4 55 651 12.9 15 111 9.6 2 600 10.4 50 - 249 5 400 4.2 101 365 23.5 31 360 20.0 5 751 23.1 > 250 1 200 0.9 183 387 42.5 88 092 56.1 12 491 50.2
Source: Eurostat (online data code: sbs_na_sca_r2)
The number of persons employed in the EU-27 in 2010 was 1 325 021. Warehousing and support activities for transportation employed 24 900. Value added at factor cost was EUR 5 933 billion in the total business economy and EUR 157 billion in NACE 52. Latvia and Estonia had the largest share of value added in warehousing and support activities for transportation: 7.5 % and 7.4 %, respectively, of the total business economy. Figure 17 shows shares of value added, persons employed and employees by country.
EU total turnover in warehousing and support activities for transportation amounted to EUR 432 billion in 2010. Again, Estonia and Latvia had the biggest shares.
Of the 21 761 617 enterprises in the EU-27, 127 605 engaged in warehousing and support activities for transportation.
96
Figure 17: Division H52 – value added at factor cost, number of persons employed and number of employees vis-à-vis total business economy, by country (%), 2010
Data for Greece not available. Source: Eurostat (online data code: sbs_na_sca_r2
97
Figure 18: Division H52 – turnover or gross premiums (EUR million and as % of total business economy), 2010
Data for Greece not available. Source: Eurostat (online data code: sbs_na_sca_r2)
98
Figure 19: Division H52 – number of enterprises, by country, 2010
Data for Greece not available. Source: Eurostat (online data code: sbs_na_sca_r2)
99
7.5. Available data
7.5.1. Short-term business statistics (STS)
STS provide information on a wide range of economic activities according to NACE Rev. 2. All data are index data. Percentage changes are also available for each indicator. The index data are generally presented in the following forms: unadjusted, calendar adjusted and SA. The STS Regulation determines whether they are monthly, quarterly or annual.
Turnover
The purpose of the turnover index is to show the development of the market for goods and services. Turnover comprises the total invoiced by the observation unit during the reference period, i.e. market sales of goods or services supplied to third parties. It includes all other charges (transport, packaging, etc.) passed on to the customer, even if these are listed separately in the invoice. It excludes VAT and other similar deductible taxes directly linked to turnover, and all duties and taxes on the goods or services invoiced by the unit.
Volume of sales
The volume of sales represents the value of turnover in constant prices and as such it is a quantity index. It is normally calculated as turnover at current prices, deflated on the basis of sales.
Producer prices (output prices)
The purpose of a PPI (or output price index) is to measure the monthly development of transaction prices of economic activities. The domestic output price index for an economic activity measures the average price development of all goods and related services resulting from that activity and sold on the domestic market. The non-domestic price index shows the average price development (expressed in the national currency) of all goods and related services resulting from that activity and sold outside the domestic market. When combined, these two indices show the average price development of all goods and related services resulting from an activity. It is essential that all price-determining characteristics of the products are taken into account, including quantity of units sold, transport provided, rebates, service conditions, guarantee conditions and destination.
Domestic and non-domestic price indices require separate output price indices to be compiled according to the destination of the product, which is determined by the residency of the third party that has ordered or purchased it. Output prices for the non-domestic market are further sub-divided into output prices for products despatched to euro-area countries and all other output prices.
Labour input indicators
The purpose of the ‘number of persons employed’ index is to show the development of employment. It reflects the total number of persons who work in the observation unit (including working proprietors, partners working regularly in
100
the unit and unpaid family workers) and persons who work outside the unit who belong to it and are paid by it (e.g. sales representatives, delivery personnel, repair and maintenance teams).
The ‘hours worked’ index shows the development of the volume of work.
The ‘gross wages and salaries’ index shows the development of remuneration in an industry.
7.5.2. Structural business statistics (SBS)
SBS cover NACE Rev. 2 sections B to N and division S95, which are organised in four annexes, covering Industry (sections B-E), Construction (F), Trade (G) and Services (H, I, J, L, M, N and S95). Financial services are covered in three specific annexes and separate metadata files have been compiled. SBS characteristics are laid down in Commission Regulation (EC) No 250/2009, which covers all business demographic, employment and monetary characteristics, and indicators in the form of ratios of monetary characteristics or per-head values. Key characteristics collected for SBS are business demographic variables (number of enterprises, number of local units), ‘input-related’ variables (number of persons employed, number of employees), cost of input (total purchases of goods and services, personnel costs), capital input (gross investment in tangible goods) and ‘output-related’ variables (turnover, production value, value added at factor costs).
7.5.3. Harmonised indices of consumer prices (HICPs)
HICPs give comparable measures of inflation for the countries and country groups for which they are produced. They are economic indicators that measure change over time in the prices of consumer goods and services acquired by households. In other words, an HICP is a set of CPIs calculated according to a harmonised approach and a single set of definitions. HICPs provide the official measure of consumer price inflation in the euro area for the purposes of monetary policy and the assessment of inflation convergence, as required under the Maastricht criteria. The data released monthly on Eurostat’s free dissemination database include price indices themselves and monthly, annual and 12-month moving average changes. In addition to the headline figure (the all-items HICP), around 100 sub-indices for various goods and services and over 30 special aggregates are made available. The relative weights for the indices and the special aggregates are published for individual countries and for country groups.
7.6. Outline of methodology
A major difficulty encountered by data compilers in the services sector is the non-availability of a single type of variable or source from which various service production activities can be measured. Only output measures in current prices may be available, without an appropriate deflator. Also, statistical information may be available at lower (e.g. annual or quarterly) frequencies, but not at higher (i.e. monthly) frequency.
101
The Eurostat Handbook on price and volume measures in national accounts classifies output indicators in three categories: A, B and C. Appropriately deflated turnover would be classified as an A method. Turnover deflated by a less appropriate deflator (e.g. with wider industry coverage) would be classified as a B method. Generally the Handbook classifies volume measures as B methods. However, if there is a detailed breakdown by type of commodity ensuring reasonable homogeneity and there is very little change in quality, a volume indicator could be classified as an A method. ‘Input’ indicators (e.g. employment) are classified as C indicators (to be avoided), because they do not adequately detect changes in productivity.
The Handbook seeks to identify a theoretical A (best) method for each industry category, regardless of whether this is practical. In contrast, in this guide we present preferred achievable measures.
Turnover data reflect receipts from sales, excluding VAT and other taxes on products, plus any subsidies on products. Preferred deflators to use are mainly PPIs, with some CPIs.
The Handbook considers turnover/sales deflated by an appropriate output price index as an A method, which will usually be the preferred data source. The price index used should be representative of the particular ISIC activity being deflated. In practice, many NSIs will sometimes need to combine price indices from different sources in order to develop a representative output price index for use as a deflator. If a combination of price indices is used, ideally these should be weighted together at a detailed level, using data relating to production values.
It is important to note that the aim of compiling such a hybrid index is to deflate service industry output, so it must be constructed using data on the same pricing basis (i.e. basic prices). For example, if components of a CPI adjusted to basic prices are used in conjunction with an SPPI, they will need to be weighted together using turnover data for the different components. Classification difficulties may arise when using a component of a CPI (usually classified according to purpose of product) to represent output from a particular industry.
Where turnover/sales is deflated by an appropriate output price index, this will be classified as a ‘preferred’ method. Deflating output (turnover) by a less appropriate, but satisfactory, price index would be an ‘alternative indicator’. The price index might be less appropriate because its scope and/or coverage does not relate directly to the output being deflated or because it is not adjusted for known changes in quality. Examples might include the use of an SPPI for another industry where price change is expected to be similar. Under other circumstances, an industrial PPI may be used where an SPPI is not available (e.g. to deflate wholesaling). If the deflator is less satisfactory, e.g. the total CPI or PPI, the approach would be classified as ‘other’.
102
Table 21: Recommended variables and deflators and their sources
NACE/ISIC level
Recommended process
Pre-processing Preparation of data (e.g. deflators, input variables, weights) suited for an automated compilation programme (e.g. estimation of missing values and treatment of series breaks)
four-digit Deflation, indexing, forecasting missing data, interpolation of higher- from lower-frequency series, seasonal adjustment
three-digit Consolidation to two-digit level two-digit Quality adjustment, benchmarking monthly or quarterly ISP
to GDP, consolidation to one-digit level one-digit Consolidation to general ISP ISP for total service sector
Dissemination
7.7. Experimental calculations of ISP
The primary aim of compiling an ISP is to measure short-term movements in the production activity of the services part of an economy. For an ISP calculation, the OECD ISP Manual recommends using gross turnover deflated by partially representative price indices. Available input variables are listed in Table 22.
Table 22: Input variables for an ISP for industry H52
Indicator Periodicity Breakdown Purpose STS turnover quarterly
(currently) Division 52 output indicator
to be deflated SBS turnover or value added
annual Division 52, groups 52.1, 52.2
weights
SPPI quarterly Group 52.1, class 52.24 (currently)
deflators
HICP monthly 07 (COICOP) deflators Currently, the SPPI breakdown for the H52 industry includes just group 52.1 and class 52.24, so it is impossible to derive the deflation for whole industry. As a result, compilers use alternative data sources or methods (e.g. volume indicators) or determine which CPI to use to deflate other classes in H52. For example, below we compile the separate ISP for group 52.1 and class 52.24 using turnover deflated by available SPPI.
103
Figure 20: Turnover, SPPI and ISP for H52.1
Figure 21: Turnover, SPPI and ISP for H52.24
0
20
40
60
80
100
120
140
160
180
200
Q1
07
Q2
07
Q3
07
Q4
07
Q1
08
Q2
08
Q3
08
Q4
08
Q1
09
Q2
09
Q3
09
Q4
09
Q1
10
Q2
10
Q3
10
Q4
10
Q1
11
Q2
11
Q3
11
Q4
11
Q1
12
Q2
12
Q3
12
Q4
12
Turnover 521 Sppi 521 ISP 521
0
20
40
60
80
100
120
140
160
Q1
07
Q2
07
Q3
07
Q4
07
Q1
08
Q2
08
Q3
08
Q4
08
Q1
09
Q2
09
Q3
09
Q4
09
Q1
10
Q2
10
Q3
10
Q4
10
Q1
11
Q2
11
Q3
11
Q4
11
Q1
12
Q2
12
Q3
12
Q4
12
Turnover 5224 Sppi 5224 ISP 5224
104
8. POSTAL SERVICES (H53)
Roberto Iannaccone, Valeria Quondamstefano (Istat)
8.1. Introduction
Division H53 (postal and courier activities) includes activities such as the picking-up, transport and delivery of letters and parcels under various arrangements. It also includes local delivery and messenger services, but not giro, postal savings and money order activities. Rapidly evolving postal services play a key role in the EU single market and are presenting national postal services with new challenges. In the new context of greater liberalisation (introduced with Directive 97/67/EC, as amended by Directives 2002/39/EC and 2008/6/EC) and (in some cases) privatisation, postal companies are adapting and developing their strategies and practices as key drivers of economic and social development.51
‘The postal sector has been opened to competition in a gradual and controlled way. The core aims of EU postal policy are to:
− complete the single market for postal services;
− ensure a high quality universal postal service for businesses and consumers at affordable prices;
− improve the quality of service, such as delivery times and convenient access; and
− ensure the correct implementation of EU legislation and propose any necessary changes to meet these objectives.’52
On this basis, Eurostat issued publications based on variables covering companies operating under a universal service obligation (USO), i.e. ‘universal service providers’ (USPs), until 2013 (latest data available were for 2011) and transferred its data collection and processing functions, together with the historical postal data, to the European Commission’s DG GROW.
8.2. Industrial classification
The International Standard Industrial Classification of all economic activities (ISIC Rev. 4) section H (transportation and storage) is made up of five divisions (Table 23). NACE Rev. 2 and the Italian classification of economic activities (Ateco 2007) are based on ISIC Rev. 4; they do not include any additional hierarchical level for this sector.
51 The process of liberalisation has differed from one Member State to the next. In Italy, for example, the
process is still ongoing and concerns some products only. 52 http://ec.europa.eu/growth/sectors/postal-services/index_en.htm
105
Table 23: NACE section H (transport and storage)
ISIC Rev. 4 Description H Transport and storage H49 Land transport H50 Water transport H51 Air transport H52 Support activities for transportation H53 Postal and courier activities
Division 53 is broken down into two classes:
− class 53.10 includes the activities of postal services operating under a universal service obligation, including the use of the universal service infrastructure, such as retail locations, sorting and processing facilities, and carrier routes to pick up and deliver mail. Items delivered can include letters, postcards, printed papers (newspapers, periodicals, advertising, etc.), small packets, goods or documents. Also included are other services necessary to support the universal service obligation; and
− class 53.20 includes courier activities operating outside the scope of a universal service obligation. One or more modes of transport may be involved and the activity may be carried out with self-owned (private) transport or via public transport.
In Italy, according to 2012 SBS, ‘postal and courier activities’ (division 53) accounted for 1.8 % of enterprises, 15.0 % of employees, 8.1 % of turnover and 14.5 % of value added at factor cost in ‘transport and storage’ (Section H).
8.3. Available data
In this section, we analyse the main available economic indicators for the postal services sector. The main domains are SBS, NAs, STS, prices and labour-market statistics. For these, we investigated what indicators were produced at the level of the postal and courier activities sector. In particular, the STS indicators for postal services are covered in Annex D to the STS Regulation. The Regulation requires Member States to produce quarterly data on:
− 120: turnover;
− 210: number of persons employed;
− 220: hours worked;
− 230: wages and salaries; and
− 310: producer prices.
In the following sections, we present information on the sector on the basis of the following variables:
− annual detailed enterprise statistics for services (SBS, Eurostat);
− final consumption expenditure (NAs, Eurostat);
− wages (labour market, Istat);
106
− turnover and services producer prices (STS, Eurostat); and
− HICP (prices, Eurostat).
8.3.1. Structural business statistics (SBS)
For SBS, European countries have to submit to Eurostat annual data at four-digit level on the main characteristics of enterprises with regard to:
− business demographic variables (e.g. number of enterprises);
− output-related variables (e.g. turnover, value added); and
− input-related variables:
o labour input (e.g. employment, hours worked);
o goods and services input (e.g. total of purchases); and
o capital input (e.g. material investments).
For reasons of confidentiality, some countries do not publish data at detailed level for postal services.
The ratio between the number of enterprises in the postal services sector (H53) and the number in section H (transport and storage) as a whole tends to vary (Table 24). Between 2005 and 2012, the ratio in the EU-27 was around 4 %. For some countries (e.g. the Netherlands, the UK and Germany), enterprises in the postal services sector represent over 10 % of the section H total, with a rising trend in the Netherlands and the UK. For other countries (e.g. Spain, France and Italy), the ratios are below the EU average.
107
Table 24: Enterprises in the postal services sector (H53) as a percentage of total enterprises in the transport and storage division (H)
Source: Eurostat, SBS
Table 25 presents results for Italy from a more in-depth analysis in an Istat report on the competitiveness of manufacturing and service sectors (2015 edition). For postal and courier activities, the number of enterprises rose in 2007 and 2008, fell by 4.0 % in 2009 and then rose (slightly) in 2010 and (more significantly) in 2011 (by 8.0 %) and 2012 (by 8.4 %). The number of employees increased slightly every year from 2006 to 2009, but then fell by 2.8 % in 2010. As a result, the average size of enterprises grew from 78.1 employees in 2008 to 81.5 in 2009 before dropping back to 66.6 in 2012. The enterprise ‘birth rate’ ( new enterprises in the reference year as a proportion of all active enterprises) spiked in 2010 (19.9 %) and 2012 (21.4 %). The ‘death rate’ peaked in 2010 (17.5 %). Consequently, the gross rate of turnover of enterprises (sum of birth and death rate) fluctuated, with high values in 2008 (30.6 %), 2010 (37.4 %) and 2012 (36.7 %) and low values in 2006 (27.0 %), 2007 (27.6 %) and 2011 (34.9 %). The net rate has been positive since 2010, with an 8.4 % increase between 2006 and 2012.
2005 2006 2007 2008 2009 2010 2011 2012EU27 3.8 - 4.1 - 4.5 4.4 4.6 -BE - - - 16.4 16.8 15.5 16 15.5BG 0.7 1 1.2 1.6 2.2 2.5 2.8 2.7CZ 1.2 1.1 1.1 1.2 1.2 0.9 0.8 0.8DK 6.9 7.6 8.5 9.2 9.1 9.6 11.2 11.9DE 11.6 12.1 12.9 9.8 10.1 9.8 11.4 10.5EE 0.8 0.9 1.2 1.3 1.3 1.7 1.8 1.7IE - - - 19.2 20.9 22 21.6 22EL 0.5 0.5 0.5 0.5 0.5 0.5 0.5 0.6ES 2.7 2.7 2.4 2 2.7 2.9 2.8 3.1FR 2 2.1 1.9 - - - 1.8 3.3IT 1.2 1.2 1.3 1.5 1.5 1.5 1.7 1.8LV 0.8 0.7 0.8 1.1 4.5 5.6 6.2 4.8LT 9 10.9 10.2 9.8 8.5 8.2 7.9 8.3LU 5.6 5.6 5.7 6 5.6 5.5 5.7 5.3HU 3.7 3.8 3.9 4 4.1 4.3 4.3 4.3NL 10.5 11.8 11.6 12.3 13.5 14.9 15.2 16.3AT 2.5 2.6 2.7 2.9 2.8 2.8 3.1 3.3PL 0.6 0.5 0.7 0.8 1.4 1.4 1.7 2PT 0.8 1 1.1 1.2 1.3 1.3 1.6 1.8SI 5.7 6.1 6.1 6.2 6.1 6.2 6.6 6.6SK 0.7 0.7 0.6 0.9 3.6 2.1 2.2 2.5FI 1.2 1.3 1.3 1.4 1.4 1.5 1.5 1.6SE 1.2 1.3 1.4 1.4 1.5 1.5 1.5 1.5UK 15.6 16.1 16.4 17 17.3 17.7 18.3 18.8NO 4.4 4.8 5.4 5.9 5.7 5.9 6.3 6.7
108
Table 25: Demographic and economic indicators for postal and courier activities (Italy)
Source: Istat (service sector competitiveness), Eurostat (SBS dataset)
Table 25 reports two variables as economic indicators: turnover and value added at factor cost. After a peak in 2008, turnover fell in 2009 (-3.7 %), recovered in 2010 (+1.8 %) and fell again in 2011 (-2.7 %) and 2012 (-0.2 %). Value added rose in the period to 2010 (except for 2008) and then began to fall (-0.8 %, -6.3 %).
8.3.2. National accounts (NAs)
NAs include data on households’ final consumption expenditure by consumption purpose at current prices (according to the COICOP classification). For postal services, growth rates since 2007 follow the same trends for the EU and the euro-area aggregates. Except in 2009, the trend for Italy differed from these. In particular, after growth in 2007 (4.1 %) and 2008 (1.6 %), annual percentage changes were negative from 2010 onwards.
2006 2007 2008 2009 2010 2011 2012Number of enterprises 2.025 2.113 2.147 2.061 2.067 2.234 2.422Number of employees 165.18 166.39 167.75 168 163.34 162.16 161.23Average size 81.6 78.7 78.1 81.5 79.0 72.6 66.6Enterprises Birth rate 12.3 13.3 14.7 14.3 19.9 18.9 21.4Enterprises Death rate 14.7 14.3 15.9 14.4 17.5 16.1 15.3Enterprises Surviving rate .... .... .... .... 32.4 28.5 36.1Gross rate of enterprises turnover 27.0 27.6 30.6 28.7 37.4 34.9 36.7Net rate of enterprises turnover -2.4 -1.0 -1.2 -0.1 2.4 2.8 6.0
Turnover (million) 11.67 12.038 12.314 11.858 12.067 11.743 11.723Value added (million) 7.746 8.113 8.077 8.126 8.204 8.14 7.625
Demographic indicators
Economic indicators
109
Table 26: Postal and courier activities (ESA95) – households’ final consumption expenditure
EUR million
2006 2007 2008 2009 2010 2011 2012 EU (27 countries) 9 407 9 712 9 500 8 850 9 183 9 383 .. EA (18 countries) 7 125 7 278 7 257 7 099 7 158 7 188 7 188 EA (12 countries) 7 044 7 182 7 148 6 932 7 118 7 283 .. Italy 850 885 899 805 787 751 714
Annual change (%)
2006 2007 2008 2009 2010 2011 2012 EU (27 countries) 3.2 -2.2 -6.8 3.8 2.2 .. EA (18 countries) 2.2 -0.3 -2.2 0.8 0.4 0.0 EA (12 countries) 2.0 -0.5 -3.0 2.7 2.3 .. Italy 4.1 1.6 -10.5 -2.2 -4.6 -4.9
Source: Eurostat (NAs)
8.3.3. Labour market
The Istat data warehouse (I.Stat) releases labour-market data (wages under collective labour agreements) for postal and courier activities at division and class level. The index showed positive and increasing growth rates until 2009, after which growth remained positive but was slower (except for a 3.2 % rise in 2013).
Table 27: Index of wages under collective labour agreements per employee (base December 2010=100), Italy
Source: Istat (I.stat)
2006 2007 2008 2009 2010 2011 2012EU27 9.407 9.712 9.5 8.85 9.183 9.383 ..EA18 7.125 7.278 7.257 7.099 7.158 7.188 7.188EA12 7.044 7.182 7.148 6.932 7.118 7.283 ..Italy 850 885 899 805 787 751 714
2006 2007 2008 2009 2010 2011 2012EU27 3.2 -2.2 -6.8 3.8 2.2 ..EA18 2.2 -0.3 -2.2 0.8 0.4 0.0EA12 2.0 -0.5 -3.0 2.7 2.3 ..Italy 4.1 1.6 -10.5 -2.2 -4.6 -4.9
Percentage Change on Previous Year
Million Euro
110
8.3.4. Short-term statistics (STS) and prices
STS turnover data and SPPIs are available for postal and courier activities.
Turnover
The STS Regulation requires turnover to be transmitted to Eurostat either as an index or in absolute figures. The turnover index is a simple value index (price multiplied by quantity) comparing the current period with the fixed period of the base year. In order to compile turnover indices at higher levels of NACE, the indices at the lowest level have to be aggregated using weights based on the turnover share of each activity in the base year, as provided by SBS.
The current definition of turnover comprises the totals invoiced by the observation unit during the reference period, which corresponds to market sales of goods or services supplied to third parties. Turnover also includes all other charges (transport, packaging, etc.) passed on to the customer, even if these are listed separately in the invoice. Excluded are VAT and other similar deductible taxes directly linked to turnover, and all duties and taxes on the goods or services invoiced by the unit. Price reductions, rebates and discounts are deducted, as is the value of returned packing. Price reductions, rebates and bonuses conceded later to clients, e.g. at the end of the year, are not taken into account. Income classified as other operating income, financial income and extraordinary income in company accounts is also excluded from turnover.
In Italy, STS for postal and courier activities have been produced since the first quarter of 2000. Turnover is measured as an index, currently for the base year 2010. STS aim to monitor economic trends and the indices are released 60 days after the end of the reporting quarter, unadjusted and in WDA and SA form. The indices are based on information gathered through a quarterly survey of a sample of companies that is representative of the economic sectors in question. The scale of structural differences among the sectors covered by the survey leads to different methods of sample selection. In the case of highly concentrated sectors, such as the postal and courier sector, where a few number of large companies determine the dynamics of the market, selection is based on a cut-off criterion, which includes larger companies to cover a sufficiently high proportion (usually over 80 %) of the sector’s turnover.
In Italy, average turnover for postal and courier activities fell every year from 2008 (Table 28). The EU (28 countries) and Italy saw the biggest drops in 2012 (by 4.6 % and 8.1 % respectively) and the euro area (19 countries) in 2009 (by 3.7 %).
111
Table 28: Postal and courier activities – turnover in services (annual WDA data; base 2010 = 100)
Source: Eurostat (STS)
Figure 22 shows the quarterly indices of turnover for postal and courier activities from the first quarter of 2006 to the first quarter of 2015 for the EU (28 countries), the euro area (19 countries) and Italy. The highest negative value was recorded in Italy for the third quarter of 2012 (-11.5 %), but Italy saw sharp falls in all quarters of 2012. The negative trend for Italy continued until the first quarter of 2015, while both European aggregates showed positive percentage change from the third quarter of 2013 onwards.
Figure 22: STS turnover data for postal and courier activities
Source: Eurostat (STS)
Price indices
The STS Regulation provides for the production of SPPIs as business-cycle indicators measuring gross change in the trading price of a number of services, including postal and courier services. SPPIs measure price changes from the perspective of the seller. The prices are transaction prices and take into account discounts, rebates, surcharges and similar price-determining elements, but not
2006 2007 2008 2009 2010 2011 2012 2013 2014EU28 97.3 100.8 102.5 98.3 100.0 104.2 99.4 100.3 102.2EA19 99.7 103.2 104.0 100.2 100.0 102.3 101.3 102.4 104.2Italy 105.5 108.2 107.4 100.8 100.0 98.1 90.2 88.3 84.6
EU28 3.6 1.7 -4.1 1.7 4.2 -4.6 0.9 2.0EA19 3.5 0.8 -3.7 -0.2 2.3 -1.0 1.1 1.7Italy 2.6 -0.8 -6.1 -0.8 -1.9 -8.1 -2.1 -4.2
Annual average index
Percentage change
75
85
95
105
115
125
I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
EU 28 EA 19 Italy
112
product taxes. SPPIs cover output produced by domestic producers, including output that is exported. In Italy, service price indices are calculated on a quarterly basis and data for postal and courier services are available from 2006 onwards.
HICPs for postal and courier activities measure change over time in the prices of consumer goods and services; they are calculated according to the COICOP classification.
The biggest annual growth rate for SPPIs was recorded for the EA-19 in 2013 (+1.9 %) and for the EU-28 in 2014 (+2.1 %). The situation in Italy was different: prices increased by only 0.1 % in 2013 and fell by 0.2 % in 2014.
Table 29: Postal and courier activities – service producer prices (annual data; 2010 = 100)
Source: Eurostat (prices database) and Istat (I.Stat)
Figure 23 highlights the different trends from 2012 onwards in Italy and Europe (EU and EA) on the basis of quarterly data starting in 2006.
Figure 23: Postal and courier activities – STS SPPI data
Source: Eurostat (prices database) and Istat (I.Stat)
The HICP increased by 3.5 % in 2014 for the euro area, with the biggest growth rate for the Netherlands (12.6 %). There were small increases for Germany, where prices increased by 6.1 % as compared with 2005. The situation is
2006 2007 2008 2009 2010 2011 2012 2013 2014EU28 97.7 99.4 100.8 99.7 100.0 101.2 101.9 103.6 105.8EA19 97.9 98.9 99.8 99.8 100.0 101.2 102.4 104.4 106.3Italy 100.6 100.8 102.1 100.5 100.0 99.6 100.2 100.3 100.1
EU28 1.8 1.4 -1.1 0.3 1.2 0.7 1.7 2.1EA19 1.0 1.0 0.0 0.2 1.2 1.2 1.9 1.8Italy 0.2 1.3 -1.6 -0.5 -0.4 0.6 0.1 -0.2
Annual average index
Annual average percentage change
95
100
105
110
I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
EU 28 EA 19 Italy
113
different in other countries, where prices show big increases compared with the annual average in 2005. For example, the HICP increased by 26.6 % in Italy and 33.8 % in Spain. The UK and the Netherlands saw the biggest HICP increases as compared with 2005 (121.7 % and 66.7 % respectively). These results are due to big rises in the most recent years: 2013 and 2014 for Netherlands and 2012 and 2013 for the UK.
Table 30: COICOP 081 postal services – HICP (annual data; 2005 = 100)
Source: Eurostat (prices database)
8.4. An ISP for postal and courier activities — initial results
An ISP should measure changes in the volume output of service sectors. Turnover deflated by an appropriate output/producer price index is the preferred means of compiling an ISP. The following indicators are available for the postal and courier activity sector (see above):
− turnover index calculated on a quarterly basis (2010=100);
− SPPI calculated on a quarterly basis (2010=100); and
− HICP calculated on a monthly basis (2005=100).
Depending on the industry concerned, either the SPPI or the HICP can be used to deflate turnover. For section H53 (postal and courier activities), the OECD ISP Manual recommends the deflation of turnover by appropriate weighted price indices of the SPPI and the HICP for measuring gross value added. As an alternative, the HICP or SPPI could be used in isolation. In this section, we show the results for an experimental ISP for postal and courier activities (base year 2010=100) using both approaches, i.e.:
− deflating turnover by the SPPI (A); and
− deflating turnover by a combination of SPPI and HICP (B).
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014EU28 100.0 102.9 105.5 107.8 111.5 114.3 117.5 121.8 127.4 132.3EA19 100.0 101.8 103.4 104.3 107.0 109.4 110.8 112.5 116.4 120.5Germany 100 100.5 99.7 99.7 100.9 102 101.4 101.9 103.9 106.1Spain 100.0 105.7 109.5 112.6 115.,8 120.9 125.2 129.2 133.3133.8Italy 100.0 103.1 104.8 104.8 110.7 116.8 117.9 119.2 124.2 126.6Netherlands 100.0 99.9 113.5 115.4 114.0 115.9 120.6 129.1 148.0 166.7United Kingdom 100.0 108.3 119.7 127.6 139.2 147.4 160.8 190.3 212.7 221.7
EU28 2.9 2.6 2.2 3.4 2.5 2.9 3.6 4.6 3.8EA19 1.8 1.5 0.9 2.6 2.2 1.3 1.5 3.5 3.5Germany 0.6 -0.9 0.0 1.2 1.0 -0.5 0.5 1.9 2.1Spain 5.7 3.6 2.8 2.8 4.5 3.6 3.2 3.2 0.3Italy 3.1 1.7 0.0 5.6 5.5 1.0 1.1 4.3 1.9Netherlands -0.1 13.5 1.7 -1.2 1.7 4.0 7.0 14.7 12.6United Kingdom 8.3 10.5 6.6 9.1 5.9 9.1 18.4 11.8 4.3
Annual average index
Annual average percentage change
114
For method B, using the supply-and-use tables at basic 2010 prices, we checked the breakdown of the output of division H53 between business use (intermediate consumption) and household consumption. For 2010, 86.8 % of the output was allocated to intermediate and 13.2 % to household consumption.
Figure 24 shows the results of the experimental ISP with method A, plotting the percentage change on the same quarter of the previous year for ISP (turnover deflated with SPPI), turnover and SPPI indices. The percentage changes for ISP are quite similar to those for turnover, given the flat trend for SPPI, except for the period from the third quarter of 2008 to the second quarter of 2011.
Figure 25 presents the results of the experimental ISP with method B, plotting the percentage change on the same quarter of the previous year for ISP, turnover, HICP and SPPI indices.
Figure 24: ISP for postal and courier activities – % change on same quarter of previous year (method A)
Figure 25: ISP for postal and courier activities – % change on same quarter of previous year (method B)
-14
-9
-4
1
6
I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I
2007 2008 2009 2010 2011 2012 2013 2014 2015
SPPI Turnover ISP
-15
-10
-5
0
5
10
15
I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I
2007 2008 2009 2010 2011 2012 2013 2014 2015
SPPI HICP Turnover ISP
115
The drops for the first and the second quarters of 2012 are bigger for the ISP than for turnover because of the rise in the SPPI. The opposite development is shown for the year.
Next, we compare the two ISPs calculated with method A and B (Figure 26). For the method B ISP, we consider a weighted average of the two price indices (SPPI and HICP) using the weights estimated with the supply-and-use table (0.868 and 0.131). The results in terms of percentage change on the same quarter (Figure 27) of the previous year are quite similar, except for the period from the second quarter of 2009 to the third quarter of 2010, when the HICP registered an increase, leading to a higher negative rate of change for the method B ISP.
Figure 26: ISP for postal and courier activities — methods A and B
Figure 27: ISP for postal and courier activities — % change on same quarter of previous year (methods A and B)
The results reported in this section are based on an experimental calculation of an ISP for postal and courier activities. As pointed out in the OECD Compilation manual for an index of service production ‘if the components of a CPI adjusted to basic prices are used in conjunction with an SPPI, they will need to be weighted
70
80
90
100
110
120
I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I
2006 2007 2008 2009 2010 2011 2012 2013 2014 2015
ISP A ISP B
-14
-9
-4
1
6
I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I II III IV I
2007 2008 2009 2010 2011 2012 2013 2014 2015
ISP A ISP B
116
together using turnover data for the different components’. The exclusive use of a B2B indicator (SPPI) would be reasonable if most of the turnover of a service industry was generated with business customers. This is not the case in NACE 53, where the use of a B2B SPPI deflator could be unsatisfactory. Given the availability of both SPPI and HICP for NACE 53, a B2All deflator, calculated as a combination of B2B and B2C price indices, is recommended.
117
9. ACCOMMODATION AND FOOD SERVICES (I55 & I56)
Richard McMahon, Central Statistical Office of Ireland (CSO)
This section presents practical recommendations for the compilation of an ISP for the accommodation and food services sector.
9.1. Available methodological information
The OECD Compilation manual for an index of service production provides guidelines and recommendations on the compilation of output indicators for the accommodation and food services sector. It aims to provide conceptual and practical recommendations (based on available input series) for the compilation of a production index to measure short-term economic activity in the services sector.
The Voorburg Group website presents comprehensive information on output and price measurement for the accommodation and food service sector.53 Mini-presentation papers are available outlining availability of output and price data, and country-specific issues for a number of countries. A sector paper presents a set of key methodological guidelines for the development of output and price statistics for this sector.
9.2. Classification issues
The accommodation and food services sector comprises establishments providing customers with short-stay accommodation and/or complete meals, snacks and beverages for immediate consumption. In many cases, the same establishment provides accommodation, food and beverage services and sometimes other leisure services. For example, hotels may provide any of the following services to guests and other members of the public: accommodation, restaurant, bar, laundry, gym, swimming, spa, casino, golf, etc.
9.2.1. Accommodation services
Accommodation services include the provision of short-stay accommodation for visitors and other travellers. The provision of longer-term accommodation for students, workers and similar individuals is also included. Some units may provide only accommodation, while others provide a combination of accommodation, meals and/or recreational facilities. This division excludes activities relating to the provision of long-term primary residences in facilities such as apartments typically leased on a monthly or annual basis and classified as real estate. Under NACE Rev. 2, accommodation is split into:
− hotels and similar accommodation;
− holiday and other short-stay accommodation;
− camping grounds, recreational vehicle parks and trailer parks; and
53 http://www.voorburggroup.org/index.asp?script=display_results&title=Industry/Industry Group&value=59
118
− other accommodation.
9.2.2. Food and beverage service activities
Food and beverage service activities include providing complete meals or drinks fit for immediate consumption, whether in traditional, self-service or take-away restaurants, or from permanent or temporary stands with or without seating. The key criterion is whether meals fit for immediate consumption are offered and not the kind of facility providing them. Leased food-service locations in facilities such as hotels, shopping malls, airports and department stores are included, but not food service activities within establishments such as hotels, civic and social associations, amusement and recreation parks, and theatres. In these cases, the provision of food services would be considered a secondary activity.
NACE Rev. 2 has the following breakdown of food and beverage service activities:
− restaurants and mobile food service activities;
− event catering and other food service activities;
− other food service activities (e.g. in sports facilities on a concession basis); and
− beverage service activities (e.g. bars, pubs, coffee shops, mobile beverage vendors, etc.).
A number of EU countries have their own derivations of the NACE Rev. 2 classification to meet their national requirements. For example, the Finnish standard industrial classification (TOL 2008) is identical to NACE Rev. 2 at four-digit level, but then has a more detailed breakdown.
9.3. Structural profile of the sector in the EU
The accommodation and food services sector is highly labour-intensive and is closely linked to the performance of the tourism economy. With its large number of enterprises and high levels of employment, it is of great importance to most economies. It is dominated by SMEs and characterised by seasonality and part-time and casual employment. People are a defining factor in the service delivery process, since the service is inseparable from the person providing it. A large proportion of the sector’s output is consumed by households.
The EU-27 sector recorded value added of EUR 186 800 million in 2009 and employed 9.9 million persons, many part-time; its workforce was also characterised by a large number of working proprietors and unpaid family workers. The sector’s contribution to employment in the non-financial business economy (sections B to J and L to N and division 95) was therefore much higher (7.4 %) than its contribution to value added (3.3 %), while it accounted for an even larger proportion (8.4 %) of the number of enterprises, reflecting the small average size of its 1.8 million enterprises.
119
Table 31: Accommodation and food services – main economic indicators
According to most indicators, division 56 (food and beverages) is larger than division 55 (accommodation), accounting for 85.0 % of all enterprises in the EU-27 accommodation and food services sector, 76.8 % of persons employed and 66.9 % of value added.
120
Figure 28: NACE section I (accommodation and food service activities) –value added and employment as proportions of non-financial business economy total, 2009 (%)
Activity in the sector is closely related to the performance of domestic and international tourism. As much of it is discretionary, expenditure on tourism-related activities is highly sensitive to the overall state of the economy. As a result, timely information for this sector is very important for STS purposes.
Countries in which tourism has significantly greater importance for the economy will have higher levels of activity in the accommodation and food services sector.
121
9.4. Data availability
It is possible to compile elementary service production indices from deflated output values (e.g. turnover, revenue, gross output deflated by consumer or producer prices) and/or output volumes (e.g. hotel bed-nights) or input indicators (e.g. employment). Information on gross value added is required for weighting the sectoral sub-indices. Also, data on the final destination of the sector’s output (i.e. B2B, B2C) is required in order to aggregate price deflators accurately.
9.4.1. Deflated output values
Turnover
There is good availability of quarterly aggregate turnover data for the sector, as it is covered under the STS Regulation. Turnover data are compiled from statistical surveys and/or administrative tax data, but monthly turnover data are not available for a number of EU countries. Some do not provide data disaggregated at NACE two-digit level.
Prices
Only a small number of countries compile SPPIs for these sectors, as these are not required under the STS Regulation. Such data are generally compiled on a quarterly basis and cover B2B transactions. Ideally, SPPIs would be compiled on a B2All basis. However, extensive monthly CPI data are available for the sectors.
Available product-level information is limited, as it is not required for STS and SBS purposes.
9.4.2. Output volumes
The performance of the sector is closely linked to that of the tourism economy. The system of statistics on tourism can provide useful indicators of output from the sector, e.g. the number of hotel bed-nights is a key tourism statistic generally collected monthly or quarterly.
9.4.3. Input indicators
There is very good availability of short-term industry-level employment data for the aggregate accommodation and food services sector, with all EU participants collecting and publishing these data under STS and social statistics regulations. Some countries do not provide data disaggregated at NACE two-digit level and/or compile the data on a quarterly basis only.
9.4.4. Weights information
Information on gross value added at factor cost for sectoral weights can be obtained from SBS or NAs.
As output from these sectors is subject to both household (B2C) and intermediate consumption (B2B), expenditure information broken down by B2B and B2C is required to weight price deflators accurately. Such information is not generally
122
directly available, as it is not required by EU regulations, but it can be derived from other official statistics (e.g. SBS, tourism, balance of payments, supply-and-use tables) or industry sources.
9.5. Methodology for calculating an ISP for I55 and I56
9.5.1. Preferred (best practice) methodological approach
The OECD Compilation manual for an index of service production indicates that the preferred approach for the accommodation sector is to compile the ISP using either gross turnover deflated by appropriate quality-adjusted price indices or volume indicators for representative categories (e.g. number of bed-nights by class of hotel).
The preferred approach for the food services sector is gross turnover deflated by appropriate quality-adjusted price indices.
The OECD Manual recommends the use of a range of CPIs as deflators for the sectors. It also specifies that SPPIs should be used for the deflation of accommodation for business travellers, conference rooms and catering for business functions.
The UK (ONS) is one of two Member States that compile an ISP. Until 2010, the deflators for service industries covered only the predominant industry and took no account of ancillary products. Since 2011, the ONS has used product deflators, so that deflators for each industry will reflect the products it supplies.
9.5.2. Alternative methodological approaches
The OECD Manual indicates that alternative approaches for the accommodation sector are to compile the ISP using gross turnover deflated by partially representative price indices or volume indicators for less detailed categories (e.g. number of bed- or pitch-nights).
The alternative approach for the food services sector is gross turnover deflated by partially representative price indices. For canteens and catering, volume indicators such as number of meals sold or customers are deemed acceptable.
The Manual indicates that the use of partially representative CPIs or general price indices as deflators for the sectors is acceptable. It classifies the use of input indicators, such as employment, as least preferred.
9.6. Calculation of experimental index for Ireland
We compiled an experimental ISP for the accommodation and food and beverage sector for Ireland (see Figure 29).
123
Figure 29: Accommodation and food and beverage sector, Ireland – monthly value and volume indices (Jan 2010-Feb 2013; base 2010=100)
The approach used is gross turnover deflated by appropriate quality-adjusted price indices.
A ‘monthly services inquiry’ collects turnover information from approximately 3 000 enterprises across the non-financial traded services sector. The data are used to compile a monthly services value index monitoring trends in output at current prices. This index is weighted by gross value added. Sub-indices for I55 (accommodation) and I56 (food services) are published nationally.
These value indices are then deflated by monthly price data from the CPI, which have been adjusted to exclude VAT. The following series are used:
Accommodation:
Personal: CPI – weekend hotels sub-index
: Guesthouses, B&Bs
: Hostels
Business: CPI — midweek rate per person for hotel room plus breakfast (typically on the second Tuesday of each month)
Secondary activities: food: CPI — restaurants, cafes, fast food and take-away food
Food services:
Restaurants and mobile food service activities
CPI — restaurants, cafes, fast food and take-away food
Event catering and other food service activities CPI — canteens
40
50
60
70
80
90
100
110
120
130
140
Accommodation &
food services -
volume
Accommodation &
food services -
value
124
The sectoral gross value added at factor cost weights were obtained from the SBS survey (Annual Services Inquiry, 2010). The weights for the B2C and B2B components of hotel accommodation are obtained from industry estimates.
9.7. Key issues
9.7.1. Data availability
Ideally, monthly turnover, CPI and SPPI information should be available to compile a monthly ISP. While all EU countries should have monthly CPIs, few have monthly turnover indices. Only a few compile SPPIs for the sector; while these are generally produced on a quarterly basis, it is possible to extrapolate monthly SPPIs.
Sectoral weights are readily available from SBS sources.
Derived estimates or industry sources of B2C and B2B expenditure can be obtained to aggregate price deflator sources.
9.7.2. Secondary activities
Turnover from secondary activities can contribute significantly to total turnover in the sector (Figure 30). Ideally, each industry should be deflated by multiple product deflators. In addition to product price information, good product level weights are required for each industry. These are not generally available, as they are not required under EU regulations.
In practice, it is a lot easier to use industry-level deflators for the predominant industry, although this may lead to issues of coherence with NA data. As indicated earlier, the ONS uses product deflators so that each industry uses deflators that reflect the products it supplies.
Figure 30: Distribution of total turnover in hotels in Ireland by service activity, 2007
Source: Ad hoc turnover by product inquiry in hotel sector (NACE Rev. 1.1 - 551), 2007.
51%
24%
18%
7%
Hotel services
Food serving
Beverage serving
Other
125
9.7.3. Use of CPI deflators
Private and business customers use the same hotel rooms and eat in the same restaurants, so the product consumed is unlikely to differ. It is possible, however, that the weighting structure will be rather different, with private persons consuming a greater proportion of cheaper products.
If trends in prices for business people and private consumers diverge or the composition of weights differs significantly, account should be taken of the type of consumer.
In the case of food and beverage sector activities, it could reasonably be assumed that price movements should not differ for business and non-business customers. However, some food services are specific to businesses (e.g. hospital catering) and would need to be covered in producing a SPPI for the sector.
Only a small number of Member States (e.g. the UK, Sweden, Finland) compile SPPIs for the accommodation and food and beverage sector. Where ‘representative’ SPPIs are available, deflation of output using these indices is considered an A method for NA purposes. However, national accountants use CPI data extensively for deflation of output for the accommodation and food and beverage sector, as:
− the lion’s share of output from the sector goes to final demand; and
− price data on most activities in the sector are collected for CPI purposes, so using these maximises their value and helps minimise the burden on respondents by avoiding the need to collect producer price data.
The CPI is considered an A method for deflation of output only if:
− it is adjusted to basic prices; and
− prices for business people and private consumers move in a similar way and the compositions of consumption are similar.
If it is not feasible to carry out dedicated SPPI surveys, it appears that the most cost-effective alternative approach to constructing an SPPI for the sector is to re-use CPI source data and adjust them to basic prices by deducting relevant taxes. Other factors to be taken into account include the potential for price trends for business customers to differ from those for private consumers. Adjusted CPI data could then be supplemented with other price data for areas not covered by CPIs, such as hotel conference charges, contract catering prices, etc. Such data would need to be collected by SPPI special survey.
9.7.4. Complementary data sources
As the performance of the accommodation and food and beverage sector is closely linked to that of the tourism economy, the system of statistics on tourism can provide useful indicators of output from the sector. For example, the number of hotel bed-nights is a key tourism statistic that is generally collected on a
126
monthly or quarterly basis. NA handbooks consider the use of bed-nights as a B method in the production of volume indicators for the accommodation sector.
9.8. Seasonality
The high degree of seasonality in these sectors can pose challenges for sub-annual estimation and combining different data sources. For example, an ‘implicit’ price for accommodation can be derived by dividing STS turnover data for accommodation by an index for hotel nights. If this ‘implicit’ price is then compared with CPI data for the accommodation sector, very clear, but opposite, cyclical effects are evident.
The likely cause of the cyclical divergences between ‘implicit’ prices and the HICP may be linked to the seasonal patterns of hotels’ secondary activities, such as food and bar sales. The revenue streams from these would not exhibit the same degree of seasonality as bed-nights, which would explain why the ‘implicit’ price of accommodation is lower during the high tourist season.
Also, the sector is characterised by high numbers of part-time and casual workers. This can lead to difficulty in measuring productivity in the sector.
9.9. Conclusion
It is feasible to compile a relatively good quarterly ISP for the accommodation and food and beverage sector using existing data sources. Quarterly gross turnover data can be deflated using VAT-exclusive CPI sub-indices as proxies for B2All SPPI deflators. Ideally, good product-level price and expenditure data would be available to cater for the deflation of the significant secondary activities.
Alternatively, monthly or quarterly ISPs for accommodation could be compiled using volume indicators such as hotel bed-nights from tourism statistics.
To produce a monthly ISP, many countries would have to introduce monthly turnover surveys.
127
10. MOTION PICTURE , VIDEO AND TV PROGRAMME PRODUCTION , SOUND RECORDING AND
MUSIC PUBLISHING ACTIVITIES (J59)
Herbert van Willenswaard, Wilco de Jong, Statistics Netherland
This section presents practical recommendations for the compilation of an ISP of motion picture and sound recording activities (J59).
10.1. Available methodological information
The following documents contain information about the methodology used to construct an ISP for this industry.
− Compilation manual for an index of service production, OECD; and
− NA manuals (information on the methodology for calculating production in NAs).
10.2. Presentation of the industry
Division J59 is very varied; it consists not only of services, but also of productions. It includes activities relating to theatrical and non-theatrical motion pictures whether on film, videotape or disc, for direct projection in cinemas or broadcasting on TV. These activities concern various stages from production, through post-production and distribution to projection. (Post-production activities include editing, film/tape transfers, titling, subtitling, credits, closed captioning, computer-produced graphics, animation and special effects, developing and processing motion picture film and the activities of motion picture film laboratories). The buying and selling of motion picture or other film production distribution rights is also included.
In addition to production on film, video tape or disc, division J59 also includes sound recording activities (i.e. the production, release, promotion and distribution of original sound master recordings), music publishing and sound recording service activities in a studio or elsewhere.
Motion picture and sound recording activities can be split into two parts, of which the larger by far in the EU is motion picture, video and TV programme activities (group 59.1), which accounted for just over 90 % of value added and employment in 2012. Value added was EUR 25 billion, which corresponds to 0.7 % of the total value added of the non-financial service economy in Europe. By far the most value added generated in the EU motion picture and sound recording activities sector was provided by SMEs (firms employing fewer than 250 persons).
128
Figure 31: NACE J50 (motion picture and sound recording activities) — value added and persons employed as proportions of non-financial service economy totals, 2012 (%)
Source: Eurostat
129
10.2.1. Activities by NACE
The activities included in NACE Rev. 2 division 59 are presented in Table 32.
Table 32: NACE Rev. 2 for J59
59 Motion picture, video and television programme production, sound recording and music publishing activities 59.1 Motion picture, video and television programme activities
59.11 Motion picture, video and television programme production activities 59.12 Motion picture, video and television programme post-production activities 59.13 Motion picture, video and television programme distribution activities 59.14 Motion picture projection activities
59.2 Sound recording and music publishing activities
Motion picture, video and TV programme activities, such as production and distribution, are mostly B2B. Sound recording and music publishing activities are also B2B. Motion picture projection activities tend to be B2C.
The turnover of division J59 in the Netherlands in 2010 was EUR 2.0 billion. Programme production activities are by far the largest activity. To give an idea of the scale of each activity, Table 33 shows the proportion of turnover by activity.
Table 33: Share of turnover by activity in the Netherlands, annual statistics, 2010
NACE Description Share of total industry (%)
59 Motion picture, video and TV programme production, sound recording and music publishing
100 (EUR 2.0 bn)
59.11 Motion picture, video and TV programme production activities
52
59.12 Motion picture, video and TV programme post-production activities
16
59.13 Motion picture, video and TV programme distribution activities
8
59.14 Motion picture projection activities 14 59.20 Sound recording and music publication activities 10 Annual statistics show that movie and TV show distribution is a service provided not only in NACE 59.13 (only an estimated 14 % of the total turnover of distribution is provided in NACE 59.13), but also in the other four-digit level classes of NACE 59. Therefore, it would seem inefficient to deflate with service prices below NACE two-digit level.
130
10.2.2. Secondary activities
Companies are classified on the basis of their primary activity, but they often have secondary activities too, the scale of which may vary by NACE section. In addition to service activities, production activities (e.g. of TV programmes) are also common in division J59. When calculating an ISP, we will not distinguish between services and production activities.
Table 34 gives estimates of secondary activities in share of turnover in the Netherlands. These activities (which include trade) accounted for only 11 % of turnover in division J59. They are more common in motion picture projection activities (59.14) and in sound recording and music publishing activities (59.20). Almost a third of turnover in cinemas comes from activities other than showing movies, such as selling drinks and snacks.
As it is difficult to measure the turnover of secondary activities for every period, it is not possible to take this into account when calculating an ISP.
Table 34: Secondary activities by NACE in the Netherlands, annual statistics, 2010 (estimated %)
NACE Description Share of other
activities (%)
59 Motion picture, video and TV programme production, sound recording and music publishing
11
59.11 Motion picture, video and TV programme production activities
7
59.12 Motion picture, video and TV programme post-production activities
8
59.13 Motion picture, video and TV programme distribution activities
1
59.14 Motion picture projection activities 29 59.20 Sound recording and music publication activities 2
10.3. Available data
The following sources are available for constructing an ISP for motion picture, video and TV programme production, sound recording and music publishing activities (J59):
− STS turnover — total turnover is available on a quarterly basis and at NACE two- or three-digit level. Deflating turnover is the preferred approach for compiling an ISP;
− STS SPPIs — prices are available on a quarterly basis and at NACE section level (J), but not at division level (J59). Turnover deflated with an SPPI of the same activity is the preferred method for compiling an ISP. The use of an SPPI with wider industry coverage is seen as less appropriate;
131
− CPI — prices (for use as a deflator) are available on a monthly basis for motion picture projection activities (59.14). At NACE section level, CPI can also be used to correct for inflation. The use of a CPI relating to films and videos was considered for NACE 59.11 and one relating to CDs, music DVDs and music downloads for NACE 59.20, but due to quality and coverage of only some activities, this was not pursued. As with SPPIs, deflation with a CPI is seen as an appropriate method, but using a CPI with a wider coverage is less appropriate;
− collectively negotiated wages and contractual wage costs – indices of the following are available on a monthly basis and can be used to deflate turnover:
o collective labour agreements for monthly and hourly wage rates, including and excluding special payments;
o contractual wage costs (per month and per hour); and
o contractual working hours.
The use of wages as a deflator is seen as a less appropriate method; and
− NA supply-and-use tables – value added available on a yearly basis at NACE three-digit level or lower. The tables are needed for weighting SPPIs, CPIs and other price information, and for constructing a weighting schedule.
10.4. Experimental calculations of ISPs
We performed test calculations for division J59, using all the above data. Below, we describe the resulting ISPs and then the weights and the choice of deflator.
Calculations were made at NACE three- and four-digit level and aggregated to two-digit level. Three different calculations were made on the basis of deflated turnover, each with a different deflator. For motion picture projection activities (NACE 59.14), a specific CPI is used in each calculation. The three different deflators used are:
− collectively agreed wages (CAO);
− CPI (inflation); and
− SPPI (sections J-N).
Figures 32–34 show the year-on-year change in the ISPs for NACE 59.1, 59.2 and 59.14. Figure 35 shows the aggregation to division J59.
132
Figure 32: NACE 59.1 (motion picture, video and TV programme activities) – ISP with different deflators, y-o-y change (2009-2013)
Figure 33: NACE 59.2 (sound recording and music publishing activities) – ISP with different deflators, y-o-y change (2009-2013)
-25%
-20%
-15%
-10%
-5%
0%
5%
10%
15%
20%
20
09
q1
20
09
q2
20
09
q3
20
09
q4
20
10
q1
20
10
q2
20
10
q3
20
10
q4
20
11
q1
20
11
q2
20
11
q3
20
11
q4
20
12
q1
20
12
q2
20
12
q3
20
12
q4
20
13
q1
20
13
q2
20
13
q3
20
13
q4
ISP - Collectively agreed wages ISP - SPPI
ISP - CPI Turnover STS
-30%
-20%
-10%
0%
10%
20%
30%
20
09
q1
20
09
q2
20
09
q3
20
09
q4
20
10
q1
20
10
q2
20
10
q3
20
10
q4
20
11
q1
20
11
q2
20
11
q3
20
11
q4
20
12
q1
20
12
q2
20
12
q3
20
12
q4
20
13
q1
20
13
q2
20
13
q3
20
13
q4
ISP - Collectively agreed wages ISP - SPPI
ISP - CPI Turnover STS
133
Figure 34: NACE 59.14 (motion picture projection activities) – ISP (2009-2013)
Figure 35: NACE 59 (motion picture, video and TV programme production, sound recording and music publishing activities) – ISP with different deflators, y-o-y change (2009-2013)
0
20
40
60
80
100
120
140
160
180
-20%
-10%
0%
10%
20%
30%
40%
50%
20
09
q1
20
09
q2
20
09
q3
20
09
q4
20
10
q1
20
10
q2
20
10
q3
20
10
q4
20
11
q1
20
11
q2
20
11
q3
20
11
q4
20
12
q1
20
12
q2
20
12
q3
20
12
q4
20
13
q1
20
13
q2
20
13
q3
20
13
q4
Ind
ice
s (2
01
0=
10
0)
yo
y %
yoy development ISP ISP
-20%
-15%
-10%
-5%
0%
5%
10%
15%
20%
20
09
q1
20
09
q2
20
09
q3
20
09
q4
20
10
q1
20
10
q2
20
10
q3
20
10
q4
20
11
q1
20
11
q2
20
11
q3
20
11
q4
20
12
q1
20
12
q2
20
12
q3
20
12
q4
20
13
q1
20
13
q2
20
13
q3
20
13
q4
ISP - Collectively agreed wages ISP - SPPI
ISP - CPI Turnover STS
134
10.5. Methodological issues
10.5.1. Weights
Two kinds of weight can be calculated:
• outer weights – these are calculated for the purpose of aggregating ISP indices from lower levels of NACE. They are based on value added in the NA supply-and-use tables; and
• inner weights – these are calculated for the purpose of combining different indices of turnover or prices. They are based on turnover from annual statistics.
Both outer and inner weights are based on the year 2010. For division J59, only outer weights are calculated; these are used to aggregate from NACE four-digit to three- or two-digit level.
Table 35: Outer weights based on value added in the Netherlands, 2010
NACE Description Value added (€ million)
% of total industry
59.11 – 59.13
Motion picture, video and TV programme activities
734 044 80
59.14 Motion picture projection activities 113 508 12 59.20 Sound recording and music
publication activities 73 200 8
10.5.2. Choice of deflator
In this industry, it is difficult to separate B2B and B2C activities. Also, as no SPPI is available for this part of the industry, a different deflator has to be used. Four different deflators may be eligible for compiling an ISP (see above):
− collectively agreed wages (CAO);
− CPI (inflation);
− CPI — for motion picture projection activities (590.14); and
− SPPI (sections J-N).
Figure 36 shows three different deflators. For confidentiality reasons, the CPI for motion picture projection activities is not shown.
135
Figure 36: NACE 59 (motion picture, video and TV programme production, sound recording and music publishing activities) – indices of deflators (2005-2013)
For NACE 59.14, most of the turnover is generated with business customers. A CPI is available in the form of the price of a cinema ticket. In the Netherlands, the turnover of NACE 59.14 is measured separately from that of the rest of NACE 59.1, so it is possible to use a different deflator.
Theoretically, a B2All SPPI for this industry is the best deflator. However, as this is not available, wages would seem to be the best choice (except as regards NACE 59.14). Wages are also used in the NAs in the Netherlands. In this industry, it is difficult to describe a particular product or service. For example, there is no standard format or budget when it comes to making a film. As labour seems to be the biggest cost item in this industry, a wage index seems to be the most suitable deflator. Because of the unique character of this service, the use of CAO wages is a B method (see Handbook on price and volume measures in NA). The second best option is the SPPI of sections J-N. Inflation can be used if other deflators are not available.
10.5.3. Methodological problems and advice
Level of detail for calculations
For the compilation/deflation of the European ISP, the deflation of turnover at NACE three-digit level is considered appropriate. Indicators for higher levels are then calculated by aggregating up from lower levels. For division J59, NACE three-digit level is very useful. To compile a deflator, it is sometimes necessary to use a more detailed level. For J59, we used price indices at four-digit level.
80
85
90
95
100
105
110
20
05
q1
20
05
q3
20
06
q1
20
06
q3
20
07
q1
20
07
q3
20
08
q1
20
08
q3
20
09
q1
20
09
q3
20
10
q1
20
10
q3
20
11
q1
20
11
q3
20
12
q1
20
12
q3
20
13
q1
20
13
q3
CPI Collectively agreed wages SPPI
136
Method for calculations
The preferred option for compiling an ISP is to use deflated turnover. The use of a B2B indicator (SPPI) as a deflator is justified if most of the turnover of a service industry is generated with business customers (e.g. market research) or if it can be assumed that there are no marked differences in price development between business and other customers. This is not the case with NACE 59: a B2B SPPI is insufficient and a B2All SPPI is needed. In addition, in the Netherlands the SPPI for this particular industry is not available. The SPPI of sections J-N as a deflator for NACE 59 is seen as a possible substitute for a B2B SPPI for this industry, but is not the first option.
The second choice is the use of B2C price indices or a combination of B2B and B2C price indices. If prices or consumed services are different for different groups of customers, a mixed approach (also requiring a breakdown of the weights with turnover) is recommended. This is the case with NACE 59, but in the absence of an SPPI for NACE 59 a mixed approach is not possible. Only NACE 59.14 can be seen as B2C, so a CPI can be used.
The third option, if labour accounts for a large proportion of the costs, is to use wages as a deflator. This is the case with NACE 59. In the Netherlands, this method is also used for the NAs.
If SPPI, CPI or wages are not available, the last choice is to use inflation as a deflator.
Seasonal adjustments
The test calculations show a clear seasonal pattern. Seasonal adjustments at NACE two-digit level seem necessary to calculate short-term developments of value added (quarter on quarter change) accurately. However, this depends on the purpose of the index. We have not made seasonal adjustments at this time.
Out-of-step development of value added/turnover and NAs
As a result of cyclical effects, it is possible that the development of value added and turnover are out of sync. Adjustments for such effects are not made when calculating an ISP.
In addition, NA calculations of value added are based on integrated figures. Integration is not part of an ISP, so this may cause small deviations between an ISP and NAs.
Recommendations
The following recommendations can be made on the basis of our experience of performing test calculations of an ISP for motion picture and sound recording activities (J59):
� make seasonal adjustments; and
� compile a B2All SPPI for NACE J59.
137
11. PROGRAMMING AND BROADCASTING (J60)
Herbert van Willenswaard, Wilco de Jong, Statistics Netherland
This section presents practical recommendations for the compilation of an ISP for programming and broadcasting activities (J60).
11.1. Available methodological information
The following documents contain information on the methodology for constructing an ISP:
− OECD Compilation manual for an index of service production;
− NA manuals (information on the methodology used to calculate production in the NAs).
11.2. Presentation of the industry
Division J60 is a fairly homogeneous industry. It consists of services and (for a small part) productions. It includes the activities of creating content (e.g. radio, TV and data programmes of entertainment, news, talk, etc.) or acquiring the right to distribute and subsequently broadcasting that content. The broadcasting can be via different technologies: over-the-air or via satellite, a cable network or the internet.
This division also includes the production of programmes that are typically ‘narrowcast’ (limited format, e.g. news, sports, education, and youth-oriented programming), on a subscription or fee basis, to a third party, for subsequent broadcast to the public.
In the EU, this industry generated value added of EUR 27 billion in 2012, or 0.7 % of the total value added of the non-financial service economy in Europe, and employed a quarter of a million people. By far the most value added generated in the EU motion picture and sound recording activities sector was from SMEs (firms employing fewer than 250 persons).
138
Figure 37: NACE J60 (programming and broadcasting activities) – value added and persons employed as proportions of non-financial service economy totals, 2012 (%)
Source: Eurostat
139
11.3. Activities by NACE
Division 60 includes the following activities:
60 Programming and broadcasting activities 60.1 Radio broadcasting 60.2 Television programming and broadcasting activities
Programming and broadcasting activities are mostly B2C. Sales of slots for commercials can be seen as B2B. In the Netherlands, about half of the turnover is generated from B2B.
The turnover of division J60 in the Netherlands in 2010 was EUR 1.7 billion. Table 36 shows a breakdown of turnover by activity.
Table 36: Breakdown of turnover by activities in the Netherlands, 2010
NACE Industry Share (%)
60 Programming and broadcasting activities 60.10 Radio broadcasting 57 60.20 TV programming and broadcasting activities 43
11.4. Secondary activities
Companies are classified on the basis of their primary activity, but they often have secondary activities too, the scale of which may vary by NACE section. In addition to service activities, production activities (e.g. the production of TV shows) are also common in division J60. In the Netherlands in 2010, production accounted for only 7 % of turnover. Therefore we will not distinguish between service and production activities when calculating an ISP.
Table 37 gives estimates of secondary activities by proportion of turnover in the Netherlands. Only 5 % of turnover in division J60 comes from secondary activities, including trade. Because it is difficult to measure the turnover of secondary activities for every period, it is not possible to take this into account when calculating an ISP.
Table 37: Percentage of secondary activities by NACE in the Netherlands, annual statistics, 2010 (estimates)
NACE Industry Share of secondary activities
(%) 60 Programming and broadcasting activities 5 60.10 Radio broadcasting 6 60.20 TV programming and broadcasting activities 3
140
11.5. Available data
The following sources are available for constructing an ISP for programming and broadcasting activities (J60):
− STS turnover — total turnover is available on a quarterly basis and at NACE two-digit level. Deflating turnover is the preferred method for compiling an ISP;
− STS SPPIs — prices are available on a quarterly basis and at NACE section level (J), but not at division level (J60). Turnover deflated with an SPPI of the same activity is the preferred method for compiling an ISP. The use of an SPPI with wider industry coverage is seen as a less appropriate method;
− CPI — prices are available on a monthly basis for TV subscriptions. At NACE section level or higher, CPI can also be used to correct for inflation. As with an SPPI, deflation with a CPI is seen as an appropriate method, but using a CPI with a wider coverage is less appropriate; and
− NA supply-and-use tables — value added available on a yearly basis at NACE two-digit level or lower. The tables are needed for weighting SPPI, CPI and other price information, and for constructing a weighting schedule.
11.6. Experimental calculations of ISPs
We performed test calculations for division J60 using all the above data. Below we describe the results of the ISP and then the weights and the choice of deflator. Three calculations are made at NACE two-digit level on the basis of deflated turnover, each with a different deflator. The three deflators used are:
− CPI cable (prices for TV subscriptions by cable or satellite);
− CPI (inflation); and
− SPPI (sections J-N).
141
Figure 38: NACE 60 (programming and broadcasting activities) – ISP with different deflators, y-o-y change (2009-2013)
11.7. Methodological issues
11.7.1. Outer and inner weights
Two kinds of weights can be calculated:
• outer weights – these are calculated for the purpose of aggregating ISP indices from lower levels of NACE. They are based on value added in the NA supply-and-use tables; and
• inner weights – these are calculated for the purpose of combining different indices of turnover or prices. They are based on turnover from annual statistics.
For division J60, only outer weights are needed to aggregate to section J. In the experimental calculations these weights were omitted.
11.7.2. Choice of deflator
In the Netherlands, about half of the turnover is generated from B2B. As no prices are available for this part of the industry, another deflator has to be used. Three different deflators may be eligible for compiling an ISP (see above):
− CPI cable (prices for TV and radio subscriptions);
− CPI (inflation); and
− SPPI (sections J-N).
Theoretically, CPI cable seems to be the best choice as a deflator for this division. It is also used for the NAs. Second best is the SPPI. Inflation can be used if other deflators are not available.
-20%
-15%
-10%
-5%
0%
5%
10%
15%
20%
25%
20
09
q1
20
09
q2
20
09
q3
20
09
q4
20
10
q1
20
10
q2
20
10
q3
20
10
q4
20
11
q1
20
11
q2
20
11
q3
20
11
q4
20
12
q1
20
12
q2
20
12
q3
20
12
q4
20
13
q1
20
13
q2
20
13
q3
20
13
q4
ISP - Cable CPI ISP - SPPI ISP - CPI Turnover STS
142
11.7.3. Methodological problems and advice
Level of detail for calculations
For the compilation/deflation of the European ISP, the deflation of turnover at NACE three-digit level is considered appropriate. Indicators for higher levels are then calculated by aggregating up from lower levels. For division J60, compilation at three-digit level is very difficult, so compilation and deflation is carried out at two-digit level.
Method for calculations
The preferred option for compiling an ISP is to use deflated turnover. The use of a B2B indicator (SPPI) as a deflator is justified if most of the turnover of a service industry is generated with business customers (e.g. market research) or if it can be assumed that there are no marked differences in price development between business and other customers. This is not the case in NACE 60: the use of SPPI is insufficient and a B2All SPPI is needed. In addition, the SPPI for this particular industry is not available in the Netherlands. The SPPI of sections J-N as a deflator for NACE 60 is seen as a possible substitute for a B2B SPPI for this industry, but is not the first option.
The second choice is the use of B2C price indices or a combination of B2B and B2C price indices. If prices or consumed services are different for different groups of customers, a mixed approach (also requiring a breakdown of the weights with turnover) is recommended. This could be the case in NACE 60, but because of the lack of an SPPI for NACE 60, a mixed approach is not possible.
For the consumer part of NACE 60, a CPI of TV and radio subscriptions is available. Cable companies are paid by consumers for providing TV and radio broadcasts by wire or over the air. This price should be a representation of the costs involved. The activities of cable companies, etc. are not included in NACE 60, but in NACE 61. In the Netherlands, the CPI for TV and radio subscriptions is the best option as a deflator for NACE 60 in total. This method is also used for the Netherlands’ NAs.
If no SPPI or CPI is available, a last choice would be to use inflation as a deflator.
Seasonal adjustments
The test calculations show a clear seasonal pattern. Seasonal adjustments at NACE two-digit level seem to be necessary to calculate accurately the short-term developments of value added (quarter on quarter change). However, this depends on the purpose of the index. We have not made seasonal adjustments at this time.
Out-of-step development of value added/turnover and NAs
As a result of cyclical effects, the development of value added and turnover may be out of sync. Adjustments for such effects are not made when calculating an ISP.
143
In addition, NA calculations of value added are based on integrated figures. Integration is not part of an ISP, so this may cause small deviations between an ISP and the NAs.
Recommendations
The following recommendations can be made on the basis of our experience of performing test calculations of an ISP for programming and broadcasting activities:
� make seasonal adjustments; and
� compile a B2All SPPI for NACE J60.
144
12. TELECOMMUNICATIONS (J61)
Michaela Lingler, Statistics Austria
12.1. Methodological information
We found methodological information for telecommunications on the internet, in statistical papers on telecommunications and available handbooks and statistics. We can distinguish between two types of paper:
− reports and studies from national and international organisations; and
− statistical papers (e.g. business statistics, NAs, price statistics).
While the latter provide information on harmonised data at European level, the former are useful in terms of learning more about telecommunications markets.
In the following sections, we give a brief overview of available documents with references to the production of an ISP. We also list links to the most relevant papers on the internet.54
12.1.1. Reports from national and international organisations
International Communications Market Report 2014, Ofcom (11 December 2014)55
Ofcom, the independent regulator and competition authority for the UK communications industries, has published comparative international data on the sector. The aim of this report was to benchmark the UK sector against a range of countries in order to assess its performance in an international context. The report compares the availability, take-up and use of services in the UK and 17 other countries.56 Some of the analysis focuses on a smaller sub-set of countries.
The report gives a broad overview of telecommunications markets and recent trends and more detailed information about key market developments, the telecoms industry structure and comparative international pricing.
Measuring the Information Society 2014, International Telecommunication Union (ITU 2014)57
The 2014 edition of Measuring the Information Society was prepared by the ICT Data and Statistics division in the ITU’s Telecommunications Development Bureau. It includes data from Eurostat, the OECD, the IMF, Informa, the UNESCO Institute for Statistics, the UN Population Division and the World Bank.
54 The last date of access to these links is 4 August 2015. The list of papers is not necessarily exhaustive. 55 http://stakeholders.ofcom.org.uk/market-data-research/market-data/communications-market-
reports/cmr14/international 56 France, Germany, Italy, the United States, Japan, Australia, Spain, the Netherlands, Sweden, Poland,
Singapore, South Korea, Brazil, Russia, India, China and Nigeria. 57 http://www.itu.int/en/ITU-D/Statistics/Pages/publications/mis2014.aspx
145
In 2008, the ITU developed an information and communication technology (ICT) development index (IDI), a benchmark indicator calculated from 11 weighted single indicators of ICT access, use and skills. The aggregated indicator monitors and compares ICT developments across countries.
Another of the ITU’s important analytical products was an ICT price basket (IPB) based on prices for fixed telephony, mobile-cellular telephony and fixed broadband internet services.
Table 38: IDI for EU Member States and other selected countries (2013 and 2012)
Country Rank 2013 IDI 2013 Rank 2012 IDI 2012 Denmark 1 8.86 2 8.78 Korea (Rep.) 2 8.85 1 8.81 Sweden 3 8.67 3 8.68 Iceland 4 8.64 4 8.58 United Kingdom 5 8.50 7 8.28 Norway 6 8.39 6 8.35 Netherlands 7 8.38 5 8.36 Finland 8 8.31 8 8.27 Hong Kong, China 9 8.28 11 8.08 Luxembourg 10 8.26 9 8.19 Germany 17 7.90 18 7.72 France 18 7.87 16 7.73 Estonia 21 7.68 21 7.54 Austria 24 7.62 23 7.46 Belgium 25 7.57 26 7.33 Ireland 26 7.57 22 7.48 Spain 28 7.38 29 7.14 Malta 30 7.25 30 7.08 Slovenia 31 7.13 31 6.96 Latvia 33 7.03 33 6.84 Italy 36 6.94 36 6.66 Croatia 37 6.90 34 6.70 Greece 39 6.85 35 6.70 Lithuania 40 6.74 40 6.50 Czech Republic 41 6.72 38 6.57 Portugal 43 6.67 39 6.57 Poland 44 6.60 37 6.63 Slovakia 45 6.58 45 6.30 Hungary 46 6.52 44 6.35 Bulgaria 49 6.31 47 6.12 Cyprus 51 6.11 48 6.09 Romania 58 5.83 58 5.52
Source: ITU
146
Table 38 shows that all Member States are in the top 60 for IDI; Denmark, Sweden and the UK are in the top five. We can conclude that telecommunications services are well developed in European countries.
Kommunikationsbericht 2014, Rundfunk & Telekom Regulierungs-GmbH58
The Rundfunk & Telekom Regulierungs-GmbH is one of the regulatory authorities for the telecommunications industry in Austria. The report covers its work and provides detailed information on the structure and development of Austrian communications markets. The information is very useful as regards the possible calculation of an ISP for the telecommunications industry.
Table 39: Development of Austrian telecommunications turnover to final customers (2012–2014)
2012
(€ million) 2013
(€ million) 2014
(€ million)
Difference 2012-2013
(%)
Difference 2013-2014
(%)
2014 share (%)
Mobile 2 361 2 240 2 263 -5.1 1.0 62.8
Broadband 731 769 782 5.2 1.7 21.7
Fixed lines 632 560 511 -11.4 -8.8 14.2 Dedicated circuit 55 49 46 -10.9 -6.1 1.3
Total 3 779 3 618 3 602 -4.3 -0.4 100.0
Source: RTR-GmbH
In general, turnover is declining for all telecommunications segments except broadband (broadband turnover includes turnover for combined broadband and fixed-line products). Call minutes for fixed lines and mobile have declined, while download volume for broadband has increased.
Reports from national and international organisations provide more or less useful information on the structure and recent development of telecommunications industries. However, a drawback of these sources is that information is not available in a standard format or with standardised classifications. Unfortunately, no comprehensive and comparable information is available as regards a breakdown of telecommunications turnover by clients, which would be very useful for producing an ISP.
12.1.2. Statistical guides and manuals, statistical papers
Methodology of short-term business statistics, Eurostat, 2006 edition, (Interpretation and guidelines and Associated documents)59
The STS methodology consists of two handbooks:
− Interpretation and guidelines for the compilation of STS; and
58 https://www.rtr.at/de/inf/KBericht2014 59 http://ec.europa.eu/eurostat/web/short-term-business-statistics/methodology
147
− Associated documents setting out the legal framework, recommendations and technical requirements.
The information on producing SPPIs is very limited. More information on this is available in the Eurostat-OECD SPPI Guide (see below).
Eurostat-OECD Methodological guide for developing producer price indices for services, 2nd edition (2014)60
The SPPI Guide, developed jointly by Eurostat and the OECD, focuses on service-specific aspects of PPI compilation. Price measurements for a wide number of service industries are described in detail. The first edition of the Guide was published in 2005 and a second updated and substantially enlarged version was released in 2014.
Updated sector paper on industry IS 61 (telecommunications), Aspasia Papa & Liam Murray, ONS, UK (27th Voorburg Group meeting on service statistics)61
This updated Voorburg Group sector paper sets out best practice for developing or revising turnover statistics and price deflators for telecommunication services. It describes the changes in this dynamic industry and will provide a basis for the telecommunications section in the updated SPPI Guide. The paper provides information on the structure of the telecommunications market, classification, turnover statistics and recommended approaches for SPPI development. This is key for the production of an ISP for telecommunications services.
Also available are mini-presentations from the 23rd Voorburg meeting (Mexico, September 2008), which looked into telecommunication services as a special topic. Papers with information on the structure of telecommunications industries and background information regarding the calculation of SPPIs were available from Austria, France, Germany, the Netherlands and Norway.
Inventory of methods for price and volume measures in the Austrian national accounts, Statistics Austria (October 2007) and Handbook on price and volume measures in national accounts (Eurostat)62
GDP and main aggregates of NAs have to be calculated at current and constant prices. Section 4 of the Eurostat Handbook describes recommended methods for the deflation of output for telecommunications services. The preferred (A) method is to use quality-adjusted SPPIs. Currently, an HICP at six-digit level (B method) is used in Austria. Sub-indices are weighted by their turnover. SPPIs will be used in the near future.
60 http://www.oecd.org/publications/eurostat-oecd-methodological-guide-for-developing-producer-price-
indices-for-services-9789264220676-en.htm 61 http://www.voorburggroup.org/Documents/2012%20Warsaw/Papers/1016%20-
%20Updated_Telecommunication_Paper.pdf 62 http://ec.europa.eu/eurostat/en/web/products-manuals-and-guidelines/-/KS-41-01-543
148
Compilation manual for an index of service production, OECD (2007)63
The main objective of the OECD Manual is to provide guidelines and recommendations on the compilation of output indicators for the service sector. It was prepared by members of the OECD’s short-term economic statistics working party task force on services. It contains terminologies, sources and methods for compiling an ISP and also covers the presentation and dissemination of an ISP. It therefore provides useful practical and technical information for the development and computation of an ISP in order to fill the current gap in service statistics.
12.2. Presentation of the industry
12.2.1. Industry overview and trends
Telecommunications is a very dynamic service industry. It is subject to rapid changes in regulation, technology and customer movement and has been one of the fastest developing industries as regards technological and structural change in recent years. At European level, a process of market deregulation started with the 1987 green paper.64 This was followed by EU directives and recommendations, which countries then implemented step by step. In Austria, the process of deregulation was initiated by the 1997 telecommunications law.
At the same time, the telecommunications sector has experienced considerable change in terms of user numbers and technologies. According to the ITU (2011), around 30 % of the world’s population (over 2 billion people) were internet users at the end of 2010, as compared with just 6 % at the end of 2000. Around 1.2 billion people accessed the internet via mobile networks, using technologies that were barely nascent 10 years earlier.
Increasing government deregulation, combined with further regulation of existing suppliers and lower entry requirements, has stimulated strong competition among business telecommunication service providers. This has led to rapid change characterised by:
− regulations that ensure effective competition and consumer protection as well as constituting the basis for a consistent environment in EU markets;
− new technologies;
− new services for consumers;
− an expansion of social networking;
− horizontal integration of IT, telecommunications and media markets;
− bundling of services (‘packages’); and
− strong price competition followed by many changes in market structure.
63 http://www.oecd.org/std/entrepreneurshipandbusinessstatistics/37799074.pdf
64 http://ec.europa.eu/green-papers/pdf/green_paper_telecom_services__common_market_com_87_290.pdf
149
Therefore, telecommunications services are an area with a wide range of fast-developing products. Currently, the market is dominated by fixed-line products, mobile phones, data network services, TV and radio transmission services and internet service provision. Telecommunications services are both B2B and B2C services. The range of products and clients and the pace of market developments have posed great challenges for NSIs, which are responsible for compiling statistical data (e.g. price and turnover statistics).
Fortunately, consumer and service producer prices for telecommunications services are regulated at European level, so structural, short-term and price information is available. Hence, the challenge of producing an ISP for telecommunication services relates more to using the available information than to data availability.
12.2.2. Classification aspects
12.2.2.1. Classification by activity
NACE Rev. 2 and ISIC Rev. 4 are identical for division 61 (telecommunications). At European level, consistency of classification is guaranteed to four-digit level. National versions of NACE may have a more detailed breakdown of economic activities.
NACE is revised in response to changes in economic structure and organisations, technological developments, new activities and products. NACE Rev. 2 contains a new section J (information and communication), which combines activities involving the production and distribution of information and cultural products, provision of the means to transmit or distribute these products, data or communications, IT activities, data processing and other information service activities.
The main components of section J are publishing activities, including software publishing (division 58), motion picture and sound recording activities (division 59), radio and TV broadcasting and programming activities (division 60), telecommunications activities (division 61), IT activities (division 62) and other information service activities (division 63). As these activities were previously included in NACE Rev. 1.1 sections D (Manufacturing), I (Transport, storage and communications), K (Real estate, renting and business activities) and O (Other community, social and personal service activities), there are significant issues as regards comparability between versions. However, the new version ensures a more consistent approach, based on the nature of the activities in question.
The following list shows the NACE Rev. 2 classification and explanatory notes for division 61:
61 Telecommunications
This division includes the activities of providing telecommunications and related service activities, i.e. transmitting voice, data, text, sound and video. The transmission facilities may be based on a single technology or a
150
combination of technologies. The common element of activities classified in this division is the transmission, rather than the creation, of content. The breakdown in this division is based on the type of infrastructure operated. With the transmission of TV signals, this may include the bundling of complete programming channels (produced in division 60) into programme packages for distribution.
61.1 Wired telecommunications activities
This class includes:
− operating, maintaining or providing access to facilities for the transmission of voice, data, text, sound and video using a wired telecommunications infrastructure;
− operating and maintaining switching and transmission facilities to provide point-to-point communications via landlines, microwave or a combination of landlines and satellite linkups;
− operating of cable distribution systems (e.g. for distribution of data and TV signals); and
− furnishing telegraph and other non-vocal communications using own facilities.
The transmission facilities may be based on a single technology or a combination of technologies.
This class also includes:
− purchasing access and network capacity from owners and operators of networks and providing businesses and households with telecommunications services using this capacity; and
− provision of internet access by the operator of the wired infrastructure.
This class excludes telecommunications resellers.
61.2 Wireless telecommunications activities
This class includes:
− operating, maintaining or providing access to facilities for the transmission of voice, data, text, sound and video using a wired telecommunications infrastructure; and
− maintaining and operating paging, cellular and other wireless telecommunications networks.
The transmission facilities provide omni-directional transmission via airwaves and may be based on a single technology or a combination of technologies.
This class also includes:
151
− purchasing access and network capacity from owners and operators of networks and providing business and households with wireless telecommunications services (except satellite) using this capacity; and
− provision of internet access by the operator of wireless infrastructure.
This class excludes telecommunications resellers.
61.3 Satellite telecommunications activities
This class includes:
− operating, maintaining or providing access to facilities for the transmission of voice data, text, and sound and video using a satellite telecommunications infrastructure; and
− delivery to consumers of visual or textual programming received from cable networks, local TV stations or radio networks via direct-to-home satellite systems. (The units in question do not generally originate programming material).
This class also includes provision of internet access by the operator of wireless infrastructure.
This class excludes telecommunications resellers.
61.9 Other telecommunications activities
This class includes:
− provision of specialised telecommunications applications, such as satellite tracking, communications telemetry and radar station operations;
− operation of satellite terminal stations and associated facilities operationally connected with one or more terrestrial communications systems and capable of transmitting telecommunications to or receiving telecommunications from satellite systems;
− provision of internet access over networks between the client and the internet service provider that are not owned or controlled by the latter, such as dial-up internet access etc.;
− provision of telephone and internet access in facilities open to the public;
− provision of telecommunications services over existing telecom connections;
− voice over internet protocol (VOIP) provision; and
− telecommunications resellers (i.e. purchasing and reselling network capacity without providing additional services).
This class excludes the provision of internet access by operators of telecommunications infrastructure.
152
In NACE Rev. 2, the most important distinction is between wired, wireless and satellite telecommunications activities. The provision of specialised telecommunications applications, reselling network capacity and internet service providers are the most important activities under NACE 61.9.
12.2.2.2. Classification by product
In contrast to activity classifications, the CPA categorises products (goods and services) according to common characteristics. It is part of an integrated system of statistical classifications, developed mainly under the auspices of the UN Statistical Division. The CPA is the European version of the UN’s central product classification (CPC).
The CPA is a product classification whose elements relate to activities as defined by NACE Rev. 2. Each product — be it a transportable or non-transportable good or a service — is assigned to a single NACE Rev. 2 activity. This means that the structure of the CPA follows that of NACE Rev. 2. However, the detailed linkage between products and activities could only be established to a certain degree: in some cases, products can be assigned to activities only at a level above class level or a class in CPA 2008 has no activity counterpart in NACE Rev. 2. Since the Regulation establishing CPA 2008 was adopted (April 2008), new concepts (such as ‘originals’ and intellectual property products) have been introduced and further details have been added to cater for new products and emerging services. National versions of the CPA exist, just as there are national versions of NACE Rev. 2.
As in NACE Rev. 2, a new section J (information and communication services) has been created in CPA 2008, combining services involving the production and distribution of information and cultural products, provision of the means to transmit or distribute these products, data or communications, IT services and the processing of data and other information services. The main components of this section are ‘publishing services’, including software publishing (division 58), ‘motion picture, video and television programme production services, sound recording and music publishing’ (division 59), ‘programming and broadcasting services’ (division 60),‘telecommunications services’ (division 61), ‘computer programming, consultancy and related services’ (division 62) and ‘other information services’ (division 63). As these services were previously included in CPA 2002 sections D (Manufactured products), I (Transport, storage and communication services), K (Real estate, renting and business services) and O (Other community, social and personal services), there are significant issues as regards comparability between versions. However, the new version ensures a more consistent approach than the previous version, which was based on the nature of the services in question.
Table 40: Division 61, telecommunications (CPA 2008)
61 Telecommunications 61.1 Wired telecommunications services 61.10 Wired telecommunications services 61.10.1 Data and message transmitting services 61.10.11 Fixed telephony services — access and use
153
61.10.12 Fixed telephony services — calling features 61.10.13 Private network services for wired
telecommunications systems 61.10.2 Carrier services for wired telecommunications
61.10.20 Carrier services for wired telecommunications systems
61.10.3 Data transmission services over wired telecommunications networks
61.10.30 Data transmission services over wired telecommunications networks
61.10.4 Wired internet telecommunications services 61.10.41 Internet backbone services
61.10.42 Narrow-band internet access services over wired networks
61.10.43 Broad-band internet access services over wired networks
61.10.49 Other wired internet telecommunications services
61.10.5 Home programme distribution services over wired infrastructure
61.10.51 Home programme distribution services over wired infrastructure, basic programming package
61.10.52 Home programme distribution services over wired infrastructure, discretionary programming package
61.10.53 Home programme distribution services over wired infrastructure, pay-per-view
61.2 Wireless telecommunications services 61.20 Wireless telecommunications services 61.20.1 Mobile telecommunications services and private
network services for wireless telecommunications systems 61.20.11 Mobile telecommunications services — access
and use 61.20.12 Mobile telecommunications services — calling
features 61.20.13 Private network services for wireless
telecommunications systems 61.20.2 Carrier services for wireless telecommunications
61.20.20 Carrier services for wireless telecommunications
61.20.3 Data transmission services over wireless telecommunications networks
61.20.30 Data transmission services over wireless telecommunications networks
61.20.4 Wireless internet telecommunications services 61.20.41 Narrow-band internet access services over
wireless networks 61.20.42 Broad-band internet access services over
154
wireless networks 61.20.49 Other wireless internet telecommunications
Services 61.20.5 Home programme distribution services over wireless
networks 61.20.50 Home programme distribution services over
wireless networks 61.3 Satellite telecommunications services 61.30 Satellite telecommunications services 61.30.1 Satellite telecommunications services, except home
programme distribution services via satellite 61.30.10 Satellite telecommunications services, except
home programme distribution services via satellite
61.30.2 Home programme distribution services via satellite 61.30.20 Home programme distribution services via
satellite 61.9 Other telecommunications services 61.90 Other telecommunications services 61.90.1 Other telecommunications services 61.90.10 Other telecommunications services
Harmonised classifications are a basic requirement if statistical data are to be comparable. The most important distinction is between wired, wireless, satellite and other telecommunications services. Enterprises are classified according to their main activity, so an enterprise engaged in wireless and wired telecommunication services is classified either in NACE 61.1 or 61.2, depending on the activity from which the higher proportion of value added (or turnover) is generated. Therefore, using data in the breakdown by activity means that turnover data of a certain industry may include turnover shares of other industries in NACE 61 — but also from other NACE categories, e.g. retail trade. Turnover by product (CPA) would provide more detailed information about shares of different telecommunications services. Unfortunately, there is no legal requirement at European level to provide a breakdown of turnover by CPA for telecommunications services. This is important as regards the interpretation of turnover data used for weights.
12.3. Available data
This section describes the data available at European level that can be used for the production of an ISP. It includes tables setting out the most recent data and recommendations for the use of the data source in question. The main data sources are STS, SBS, ICT and price statistics (SPPI, HCPI).
12.3.1. Short-term statistics (STS)
STS describe the short-term development of the economy. The STS Regulation entered into force in 1998, with a transition period of five years. The reference year (2003) is the first year for which the full set of basic STS indicators, as required in the Regulation, should be available. The availability of STS data
155
depends on the industry in question. Division 61 is regulated in Annex D (Other services) to the Regulation, which requires provision of:
− the variables ‘turnover’ and ‘number of persons employed’ for selected service industries on a quarterly basis;
− the additional variables ‘wages and salaries’ and ‘hours worked’ from 2013 onwards, starting with 2010;
− variables in the form of an index or in absolute figures; WDA data are required for ‘turnover’; SA data are produced by Eurostat;
− telecommunications data at NACE Rev. 2 two-digit level (division 61);
− SPPIs on a quarterly basis for selected activities covered by the ‘turnover’ variable (since 2006); and
− SPPIs for telecommunications (since 2006).
Transition periods in the STS Regulation, problems in the implementation of STS requirements and the 1 % rule for small Member States mean that there may be data gaps.
Time series with STS turnover indicators are the fundamental information required for the production of an ISP for telecommunications services. In the Eurostat database, these are available on a quarterly basis for almost all Member States, with the exception of Denmark and Ireland in August 2015. STS data are also missing for Switzerland and Norway. Time series are almost complete and available for gross (unadjusted), WDA and SA data. In the test calculations, we used turnover indicators with base year 2010 and up to the reference period (first quarter of 2015).65
As SPPIs will be one of the main data sources for producing an ISP for telecommunications, their complete availability for this industry is very important. Extraction from the Eurostat database shows that SPPIs have more data gaps than turnover indicators. Only six Member States have provided SPPIs for division 61 for 2006, 10 for 2007 and 11 for 2010 (a base year for STS). These data gaps will be a problem for test calculations for individual countries.
In Austria, SPPIs for telecommunications services are calculated using the unit-value method. Basic data are available from the national regulatory authority for telecommunications (RTR). In principle, the SPPIs are on a B2All basis.
12.3.2. Harmonised indices of consumer price (HICPs)
HICPs give comparable measures of inflation for individual countries. They measure the change of the prices of consumer goods and services over time. HICPs for all Member States are released on a monthly basis and classified according to COICOP four-digit categories and sub-categories.
65 Extracted from the database on 11 August 2015.
156
Telecommunications is included in COICOP division 08 (Communication), which is further divided into the following groups:
− 08.1 (postal services);
− 08.2 (telephone and telefax equipment); and
− 08.3 (telephone and telefax services).
For the telecommunications industry, groups 08.2 and 08.3 are more relevant for deflation than 08.1, as postal services are separated from telecommunications services in most countries.
HICP information in the Eurostat database is almost complete for all countries and available about 16 days after the end of the relevant reference month.
12.3.3. Structural business statistics (SBS)
SBS describe the structure, conduct and performance of businesses across the EU and cover industry, construction, trade and services. The main SBS indicators are enterprises, employment, turnover, value added and investment (for the EU-28 and by Member State) broken down by economic activity and size class. SBS are based mainly on data for enterprises classified according to NACE. NACE Rev. 2 has been applied since the reference year (2008), before which NACE Rev. 1 and Rev. 1.1 were applied.
SBS data are available on an annual basis, with a time lag of 18 months after the end of the reference year for individual Member States and about 21 months for EU aggregates. They include telecommunications as one of the service sectors covered by the SBS Regulation from the beginning. Nevertheless, changes in the NACE classification and the recasting of the SBS Regulation had an impact on the availability of time series.
SBS data are available for telecommunications from 1995. Reference years 1995-2007 are based on NACE Rev. 1 or Rev. 1.1, with a NACE three-digit breakdown (I 64.2 Telecommunications)
NACE Rev. 2 is applied from reference year 2008, with a more detailed NACE Rev. 2 four-digit breakdown of activities. So, the following activity breakdown is available for J61:
− 61.1 (wired telecommunications activities);
− 61.2 (wireless telecommunications activities);
− 61.3 (satellite telecommunications activities);
− 61.9 (other telecommunications activities).
NACE Rev. 1.1 and NACE Rev. 2 data are available for reference year 2008 due to double reporting. Backcasted data are not available for SBS in general, but some Member States have provided backcasted data back to 2005 on a voluntary basis. Double reporting for 2008 will allow backcasting on a macro scale.
157
Transition periods in the SBS Regulation, problems with the implementation of SBS requirements and the 1 % rule for small Member States may mean that there are data gaps. NACE Rev. 2 data prior to 2008 are available for some countries only, as there is no legal requirement for backcasting. SBS provide very detailed data for the activities in question and are an important information source for weighting in the production of SPPIs and an ISP.
The CPA is not a fundamental classification in the organisation of SBS, but some SBS will be compiled using a CPA product-based breakdown and disseminated for certain areas (trade, insurance, credit institutions and business services), as required by the SBS Regulation. Unfortunately, no CPA breakdown is required for telecommunications services.
The structural profile of telecommunications services in the EU shows that there were around 43 000 enterprises operating in this area in 2012. Estimated turnover was about EUR 400 million. Estimated value added at factor cost was EUR 169 million.
Table 41: Telecommunications – number of enterprises, turnover and value added at factor cost, EU-28 (2011–2013)
2011 2012 2013 Number of enterprises 41 138 43 137 . Turnover (€ million) 413 000.0(de) 400 000.0(de) 381 847.3(p) Value added at factor cost 174 752.4 169 000.0(de) . Number of persons employed 11 000.0(de) 10 000(deu) 10 246(p)
Source: Eurostat (d) definition differs, see metadata; (e) estimated; (p) provisional; (u) low reliability
The structure within telecommunications sectors is more interesting when it comes to compiling an ISP, so we extracted turnover distribution across sub-categories of telecommunications services (groups 61.1–61.9) from the Eurostat database. Table 42 shows the percentage turnover distribution by Member State and for the EU-28 for reference year 2010.
158
Table 42: Telecommunications – turnover distribution by NACE Rev. 2 group, 2013 (rounded % values, preliminary)
61.1 61.2 61.3 61.9 EU-28 40 32 2 26 Belgium 52 32 0 15 Czech Republic 16 82 1 1 Germany 52 33 0 14 Greece 39 53 1 7 Spain 56 35 1 8 France 49 39 6 7 Croatia 63 36 0 1 Italy 52 41 0 7 Cyprus 67 2 5 26 Latvia 31 56 0 13 Lithuania 40 59 0 1 Hungary 67 30 0 3 Poland 32 62 1 6 Portugal 55 37 1 7 Romania 36 57 1 6 Slovenia 64 23 1 12 Finland 42 57 0 0 Sweden 57 36 3 5 United Kingdom 2 3 2 93 Norway 29 58 11 2
Source: Eurostat, SBS
The table shows only countries for which the full matrix of data is available without confidential cells. In 2013, 40 % of turnover was generated by enterprises whose main activity was wired telecommunications services and 32 % through wireless telecommunications services. These two NACE groups were responsible for a considerable proportion of NACE 61 turnover in most of the countries. While the weight of sub-sector 61.3 (satellite telecommunications activities) is negligible (2 %), that of 61.9 (other telecommunications services) is considerably greater in some countries (particularly the UK: 93 %). This reflects structural differences between countries.
159
Figure 39: Telecommunications – turnover distribution by NACE Rev. 2 group, 2011–2013 (%, EU-28)
Source: Eurostat, SBS
Figure 39 shows that turnover distribution in the EU-28 is relatively constant in 2011–2013. Nevertheless, it can be affected by the restructuring and classification changes of large enterprises. In general, SBS data will provide useful information for weighting telecommunications services. Unfortunately, no turnover breakdown by product (CPA) is available from SBS. Enterprises (mainly large firms) that provide wired and wireless telecommunications services may also provide other (e.g. internet) services. Information society statistics (see below) are a possible source of information for a more detailed breakdown of telecommunication services.
12.3.4. ICT statistics
ICT statistics provide more detailed data about telecommunications services, e.g. a more detailed breakdown of turnover, employment data, international receipts and payments, international calls, traffic, SMS services, number of operators, access to networks, etc.
The data are available on an annual basis and are used mainly for weighting. They provide more detailed material than SBS, so will be a possible source of information for further breakdown of SBS data. ICT data were not used in the course of test calculations in this section.
12.4. Experimental calculations of ISPs
12.4.1. Objectives of the calculation
The main objective of the calculation is to produce an experimental ISP for telecommunications services. An ISP measures changes in the volume of output of the service sector over time. It is defined as a weighted average of the real output of the various telecommunications industries. Weights should be based on
.00 10.00 20.00 30.00 40.00 50.00
61.1
61.2
61.3
61.9
2013
2012
2011
160
their shares in the value added of the service sector. The change in price has to be removed from the nominal output to allow a comparison of change in real output. For telecommunications services, the volume of output of the economic activity in question has to be measured.
After describing the special characteristics of the telecommunications industry with regard to calculation of an ISP, we establish the requisite input variables and available data sources with regard to periodicity and completeness. Lastly, we elaborate and describe methods and recommendations for the calculation of an ISP for telecommunication services.
12.4.2. Special characteristics of the telecommunications industry
First, it is important to analyse the structure of the telecommunications industry and the levels of activity breakdown for which the variables relevant for calculating an ISP are available. Figure 40 shows the structure of the industry based on harmonised activity and product classifications. Detailed descriptions can be found in the section on classification aspects.
Figure 40: Structure of the telecommunications industries
Source: Classification database, Statistics Austria
The required publication level for an ISP for telecommunications is NACE Rev. 2 two-digit level. However, for reasons of quality, it will be necessary to base the calculation on a more detailed level. NACE 61 is split into four groups. As regards turnover or value added from SBS (see section 3.3.3), the most relevant groups for most of the countries are ‘wired telecommunications activities’ (61.1) and ‘wireless telecommunications activities’ (61.2).
In general, enterprises are classified according to their main activity, but normally they have secondary activities, e.g. enterprises classified in wired telecommunications activities may also have wireless telecommunication activities. Therefore, to produce a very detailed weighting scheme, one would ideally have information at product level. Unfortunately, there is no legal requirement at European level for a breakdown of telecommunications services
161
into CPA. Therefore, it is assumed that, for most countries, only NACE Rev. 2 data are available from SBS.
Another important factor is that a significant proportion of telecommunications services output is B2C. This is not covered in the existing scope of SPPIs for telecommunications for some countries and is not regarded as output for intermediate consumption (B2B). Therefore, it may be necessary to combine a set of price indices from SPPI and HICP, which have to be weighted together at a detailed level. In contrast, some countries produce a SPPI on a B2All basis.
12.4.3. Input variables/deflators
Table 43 lists possible input variables for production of an ISP for telecommunications services. It also describes the purpose of the variables in the calculations, their periodicity and timeliness.
Table 43: Overview of main input variables for a telecommunications ISP
A more detailed description of input variables with reference to calculation of an ISP for telecommunications services is provided below. Problems of availability are also discussed.
STS turnover
The relevant input variables to be deflated are turnover indicators from STS, which are available for division 61 on a quarterly basis as gross and working day adjusted values. Seasonal adjusted values produced by Eurostat were also used. The current legal basis requires calculation of the indicator on a quarterly basis 60 days after the end of the reference period. The draft FRIBS STS package proposes monthly turnover indicators — this proposal should also be considered. Test calculations will be carried out with the quarterly turnover indicators currently available, but the planned monthly periodicity should be kept in mind.
Price statistics
Telecommunications services are produced for households and business customers (including government) for intermediate consumption. This breakdown of clients has to be considered in the course of deflation. An ideal deflated index would be produced on a B2All basis, with separate price data for B2C and B2B.
Indicator Periodicity Deadline Breakdown Purpose STS turnover indicator
quarterly t + 60 days division 61 current price output deflator
SBS turnover indicator
annual t + 18 months
division 61 and groups 61.1-61.9
weighting
SPPI quarterly t + 90 days division 61 deflator CPI monthly t + 16 days 08.2 and 08.3 deflator (B2C)
162
If a SPPI for telecommunications services is produced on a B2All basis, this is the preferred price index for deflation. If not, additional data such as a CPI or HICP should be used as a proxy for the B2C component. Therefore, an SPPI and CPI for telecommunications services need to be weighted and added for production of a B2All index.
However, the compilation of SPPIs for telecommunications requires very detailed information on turnover shares for the most important products. In Austria, business statistics do not break turnover down by product (CPA), so enterprises in the sample have to be asked about their turnover breakdown for the calculation of SPPIs in order to have the required information for the weighting scheme. For telecommunications, the unit-value pricing method is applied in Austria.
The following should be borne in mind as regards SPPI activity/product coverage:
• ideally, all economic activities should be covered;
• satellite telecommunications and or other telecommunications are excluded in some countries (e.g. AT, CZ, DK, IT); and
• three countries (CZ, ES, NO) produce a separate B2B and B2C index; the remaining countries for which relevant information was available produce a B2B (BG, IT, CY, LU, SK, UK) or a B2All (PL, AT, SI, SE, HR) index.
Countries’ pricing methods vary. Nine countries use unit values exclusively and a further four use unit values in combination with other pricing methods. Three countries use component pricing exclusively. Three others use it with further methods. Contract pricing with other methods is used by three countries; one country uses model pricing with other methods. As regards product lists, the ideal would be a standardised list based on CPA, but this does not exist for telecommunications services. Specific products/transactions agreed with each statistical observation unit and product lists defined only for the SPPI with a regulatory authority are used in addition to the CPA.
Another issue is availability of price statistics. While CPIs are available in time for the ISP, SPPIs are available too late. Also, one should bear in mind the possible monthly periodicity of turnover indicators proposed in the STS package in combination with quarterly SPPIs.
SBS data
Annual information on value added at factor cost and turnover is available from SBS about 18 months after the end of the reference period in a NACE Rev. 2 four-digit breakdown. SBS provide basic information for weighting — EU legislation does not require a breakdown by CPA. Neither do SBS provide a breakdown by clients (B2B/B2C). So if relevant information is not available from the calculation of SPPIs, these data have to be estimated. Value added from NAs can also be used for weighting.
163
Recommendations for imputing missing input variables
An ISP for telecommunications services has to be calculated with predefined timeliness and periodicity. If the requisite input variables are not available, missing input variables have to be calculated using statistical estimation methods. Table 44, which is based on the OECD ISP Manual, presents methodological options for calculating missing input variables.
In general, there are various possible ways of imputing missing information. The most reliable method has to be determined using available information, according to the country-specific situation. Quality of estimates should be borne in mind.
In the event of missing information, it would also be possible to use volume indicators such as fixed line national, fixed line international or mobile voice data.
Table 44: Options for the calculation of missing input variables
Situation Problems Possible solutions
ISP deadline t + 60 days, SPPIs available at t + 90 days
Missing price information at end of time series
Preferred: forecast with information on other variables in the same NACE category or forecast with time series information only; Alternative: impute using information on other variables in the same NACE category with the same statistical behaviour; Other: forecast with simple method, e.g. simple moving average
ISP has to be calculated on a monthly basis with SPPI available on a quarterly basis
FRIBS STS package changes from quarterly to monthly and SPPI remains quarterly
Preferred: forecast quarterly data for three periods ahead and interpolate a monthly path from forecast data; Alternative: impute using information on other variables in the same NACE category and with the same statistical behaviour which are available at higher frequency; Other: interpolate using own information only
SPPI time series are shorter than required for calculation of ISP
Missing data at beginning of series
Preferred: backcast with information on other variables in the same NACE category or with series information only; Alternative: impute using information on other variables in the same NACE category with the same statistical behaviour; Other: backcast with simple method, e.g. simple moving average
Missing data for weighting
Missing breakdown of turnover by CPA and/or client
Preferred: use information from SPPI and NAs; Alternative: use turnover NACE three-digit breakdown from SBS and estimate breakdown by client; Other: use statistical estimation models to calculate required information
164
12.4.4. Recommendations for calculation of an ISP
An ISP can be produced using various methods. In general, the suitable method for a given country will depend on the availability of input data. In the course of test calculations for telecommunications services, two main methods were examined. Table 45 gives an overview of the methods. It should be noted that some assumptions have to be made because of data gaps. Test calculations were carried out for countries for which the full data matrix was available. In general, time series with base year 2010 were taken into account.
Table 45: Methods for calculating an ISP for J61
Method Recommendation Further information
Method 1: Deflation with B2All SPPI for J61 Deflation at NACE Rev. 2 two-digit level with SPPIs available at two-digit level for B2All Deflation for gross, working day and seasonal adjusted data (series first adjusted, then deflated)
Recommended method for all countries for which SPPI for telecommunications services is calculated at B2All level.
ISP calculated for countries for which full set of data was available.
Method 2: Deflation with a combined price index from a B2B SPPI NACE 61 and a B2C HICP communication Deflation at NACE Rev. 2 two-digit level: B2B part with SPPI for B2B and deflation of the B2C part with HICP Deflation for gross, working day and seasonal adjusted data (series first adjusted, then deflated)
Preferred method if SPPIs are available for B2B only; a precondition is the weight for the B2C part of telecommunications turnover and/or value added
Assumptions were made for the B2C part of turnover and applied for all countries for which test calculations were made.
Method 3: Deflation with B2All SPPI (NACE Rev. 2 three-digit level) Deflation at NACE Rev. 2 three-digit level with SPPIs available at three-digit level for B2All
More detailed than method 1; a breakdown of turnover by CPA (which is not the case for most countries) would be ideal — otherwise this would be the preferred method, if SPPIs are available on a B2All basis
No test calculations carried out
12.4.5. Test calculations
Test calculations were carried out for methods 1 and 2. Basic data for calculation of an ISP were extracted from the Eurostat database. STS gross data were used
165
for WDA and SA data (TOVT, TOVT_WDA, TOVT_SA). B2All (PRON) and B2B (PRBB) price indicators were extracted from the SPPI database. Base year 2010 was applied for test calculations in general, so HICP data for telecommunications (available for base year 2005 in the Eurostat database) were brought to base year 2010 mathematically.
For test calculation 1, STS gross series, WDA series and SA series were divided by the B2All SPPIs for NACE activity.
For test calculation 2, STS gross series, WDA series and SA series were divided by weighting the B2B SPPIs for telecommunications services and HICP for communication services for NACE activity 61.
For the test calculations, adjusted series were deflated; in general, gross series should be deflated and then seasonally and working-day adjusted.
In general, data sources used for test calculations (except HICP) are available on a quarterly basis. HICP can be brought from monthly to quarterly periodicity very easily by calculating the average of monthly data for each quarter under observation.
Finally, it should be noted that SPPIs are available with a time lag of t+90, so a forecast of price information will be necessary. The proposal for the FRIBS STS package involves a monthly periodicity for the ISP, while the periodicity for SPPIs is quarterly, with the option of producing monthly SPPIs. It is assumed that SPPIs will be produced on a quarterly basis for most of the Member States in the future, methods have to be developed for forecasting and interpolating monthly data. These methods were not tested in the course of the exercise.
Test calculations were carried out for countries and EU aggregates for which a full data matrix was available. Imputations were not made for missing data.
The possible method 3 (Table 45) was not tested.
12.5. Results of test calculations
This section provides a description of selected main results of the test calculations. In general, the methods tested are applicable for countries depending on data availability. If SPPIs are available at B2All level, this will be the preferred deflation method for telecommunications services. All customers of telecommunications industries are covered and the Austrian example showed that SPPIs have very detailed weightings according to the services produced by telecommunications enterprises. (e.g. fixed lines, mobile, broadband and leased lines). The following charts show selected results for the EU-28.
166
Figure 41: SA and WDA series, not deflated (TOVT) and deflated (TOVV), EU-28 – method 1
Source: test calculations, based on data from Eurostat
Figure 42: SA and WDA series, not deflated (TOVT) and deflated (TOVV), EU-28 – method 2
Source: test calculations, based on data from Eurostat
70.00
80.00
90.00
100.00
110.00
120.00
130.00
20
06
Q1
20
06
Q3
20
07
Q1
20
07
Q3
20
08
Q1
20
08
Q3
20
09
Q1
20
09
Q3
20
10
Q1
20
10
Q3
20
11
Q1
20
11
Q3
20
12
Q1
20
12
Q3
20
13
Q1
20
13
Q3
20
14
Q1
20
14
Q3
20
15
Q1
TOVT WDA
TOVT SA
PRON
TOVV WDA
TOVV SA
70.00
80.00
90.00
100.00
110.00
120.00
130.00
20
06
Q1
20
06
Q3
20
07
Q1
20
07
Q3
20
08
Q1
20
08
Q3
20
09
Q1
20
09
Q3
20
10
Q1
20
10
Q3
20
11
Q1
20
11
Q3
20
12
Q1
20
12
Q3
20
13
Q1
20
13
Q3
20
14
Q1
20
14
Q3
20
15
Q1
TOVT WDA
TOVT SA
TOVV WDA
TOVV SA
167
Figure 43: Comparison of methods 1 and 2 for EU-28 SA and WDA
Source: Test calculations, based on data from Eurostat
12.6. Conclusions and further recommendations
Compiling an ISP for telecommunications services is possible for countries for which SPPIs are available on a B2All basis. For countries for which a B2B SPPI is available, the appropriate CPI (HICP) should be taken as a proxy for the B2C part of turnover. Use of this method will require detailed information about the respective weights of the B2B and the B2C parts of turnover.
An ISP for telecommunications services can be calculated subject to the availability of:
• SPPIs for B2All;
• (if SPPIs are available on a B2B basis) adequate weights for the B2B and B2C part and HICP for communication services;
• statistical forecast and interpolation methods for producing monthly SPPIs;
• weighting information from SBS, NAs or other business statistics; and
• knowledge of the telecommunications markets and their developments.
70.00
80.00
90.00
100.00
110.00
120.00
130.00
20
06
Q1
20
06
Q4
20
07
Q3
20
08
Q2
20
09
Q1
20
09
Q4
20
10
Q3
20
11
Q2
20
12
Q1
20
12
Q4
20
13
Q3
20
14
Q2
20
15
Q1
Method 1 WDA
Method 1 SA
Method 2 WDA
Method 2 SA
168
13. COMPUTER PROGRAMMING , CONSULTANCY AND RELATED ACTIVITIES (J62)
Ildikó Holocsy, Central Statistical Office of Hungary
Within information and communication activities, the ‘computer programming, consultancy and related activities’ division (J62) plays an increasingly important role at the level of total production of services and the economy as a whole — both in terms of gross value added and net sales figures. The main objective of the calculation of an ISP is to determine the production of this industry not only at current prices, but also at constant prices.
According to the System of national accounts (2008 handbook), ‘the ideal way of producing volume estimates of macroeconomic aggregates is to work at a very detailed level, deflating each component by a strictly appropriate price index’. It follows that, to calculate an appropriate ISP for division J62, the ideal components are consistent output (turnover) and PPI (SPPIs) data at the same level and frequency according to the same classification system.
13.1. Available methodological information
Methodological guides and handbooks for the calculation of real volumes — developed in close alignment with the practice of international institutions (European Commission, IMF, OECD, United Nations, World Bank) — describe the main issues concerning the production of products and services generally and from the point of view of NAs. In contrast, the relevant section of the Eurostat-OECD SPPI Guide or Voorburg Group documents (sector papers, revisited sector papers, mini-presentations and issues papers) reflect the characteristics of classifications, market conditions, pricing methods and NA requirements for this division.
Initially, the main objective was to provide structural information about the industry, rather than price or turnover considerations. Work on describing price indices started in 2002 at the 17th meeting of the Voorburg Group in Nantes (INSEE, France), followed by mini-presentations, at the 23th meeting in Aguascalientes (Mexico) in 2008 by a sector paper, and in 2012 in Warsaw (Poland) by an updated sector paper. Other national studies and useful international publications describing the unique nature of this industry are available (see below).
13.2. Presentation of the industry
In NACE Rev. 2, division J62 is divided into four classes: computer programming activities (62.01), computer consultancy activities (62.02), computer facility management activities (62.03) and other information technology and computer service activities (63.09). It includes the following activities of providing expertise in the field of IT:
− writing, modifying, testing and supporting software;
− planning and designing computer systems that integrate computer hardware, software and communication technologies;
169
− on-site management and operation of clients’ computer systems and/or data processing facilities; and
− other professional and technical computer-related activities.
The CPA classification includes 11 sub-categories (product groups at six-digit level.
Most countries, including Hungary, apply harmonised national versions of NACE Rev. 2 and CPA 2008. Given the nature of the IT industry (unique, bundled services, sub-contracting, off-shoring, new market entrants face low barriers to entry), companies may rapidly change the structure and content of services provided to accommodate partners’ needs. For this reason, it is particularly difficult to divide activities into classes to determine primary activity. Secondary activities are common, as are bundled services. In addition, companies classified as belonging to other industries also offer programming, consultancy and related services.
In Hungary, the main secondary activities of companies classified in division J62 are ‘wholesale of information and communication equipment’ (46.5), ‘retail sale of information and communication equipment in specialised stores’ (47.4), ‘software publishing’ (58.2), ‘telecommunications’ (61), ‘information service activities’ (63), ‘accounting, bookkeeping and auditing activities; tax consultancy’ (69.20) and ‘business and other management consultancy activities’ (70.22). On the other hand, mainly companies whose primary activity is listed above (plus ‘manufacture of parts and accessories for motor vehicles’ (29.3)) also provide services classified in division J62.
Globalisation, in the form of the drive to maximise profit at world level, is also affecting companies operating in this industry. The issues of own-account software or whether to include transfer pricing of non-market transactions (where service providers sub-contract or outsource computer programming, consultancy and related activities, especially when the sub-contractor is an offshore establishment, affiliated or not) represent additional challenges for statisticians. However, if turnover of captive entities is included in the industry turnover or in the NAs, the SPPI as deflator should include transfer prices. In such cases, quality adjustment and quality control are necessary.
As regards STS indicators for the division (see Annex D to the STS Regulation: ‘Other services’), quarterly data are required for turnover, number of persons employed and SPPIs. With regard to the variable number of persons employed in countries with less than 4 % of the respective EU totals, the section J coverage is required.
According to NA statistics for computer programming, consultancy and information services (divisions 62 and 63), gross value added (at basic prices) increased by 38 % in the EU as a whole and by 89 % in Hungary in 2012, as compared with 2005. Total employment also increased, by 25 % in the EU as a whole and by 44 % in Hungary.
For division 62, based on annual detailed enterprise statistics for services in the EU, both the number of enterprises and turnover were 45 % higher.
170
Table 46 shows the main economic indicators (number of enterprises, turnover, value added and number of persons employed) as proportions of total values (based on 2012 annual business statistics for special aggregates of activities).
Table 46: Division J62 – main economic indicators for as proportions of the total business economy (except financial and insurance activities), 2012
EU-27 Hungary
Total business
economy…
Computer programming
%
Total business
economy…
Computer programming
%
Number of enterprises 22 346 729 : 528 519 3.5
Turnover (€ million) 26 074 073 1.6 254 233 1.3
Value added at factor cost (€ million)
6 184 837 3.3 46 497 2.8
Number of persons employed
(100 persons) 1 337 673 2.1 24 306 2.3
Source: Eurostat (online data code:sbs_na_sca_r2)
As regards the EU aggregates, the highest proportion (3.3 %) was calculated for value added.
In the EU, approximately 94 % of enterprises were in the lowest size (employment) category, while enterprises with 50 employees or more represented only 1.1 % of the total. However, for all other indicators, the proportions were highest in the category of enterprises with 250 or more employees (for turnover, over 46 %; for value added, over 43 %; the estimated proportion of persons employed was also over 34 %).
In Hungary, the industry is also dominated (well above 90 %) by micro-level companies (with fewer than five employees), but companies with more than 20 employees account for nearly 50 % of turnover.
13.2.1. Turnover development for J62
The main purpose of producing turnover statistics is to provide indicators of the economic cycle, from which several other indicators are derived. Annual data are important for economic benchmarks. Short-term data generally refer to flows during the quarter, and to stocks at the beginning and end of quarters.
In most countries, data collection is by statistical survey. In principle, it is compulsory to provide data. A combination of survey data and administrative data is also used. Missing data are typically estimated by using administrative (mainly tax) data to reduce response burden. Turnover data are collected annually, quarterly or monthly. In the majority of countries, an annual survey is carried out, typically in the framework of SBS, and quarterly or monthly in the
171
framework of STS. Eurostat uses the national data to compile EU and euro-area economic indicators.
Data collection is usually via an electronic questionnaire. In most countries, an integrated questionnaire is used to collect turnover and other data. Turnover is currently defined as total net sales value (excluding VAT). Countries usually collect data under standard sampling routines, first taking into account the industry structure (mainly available in a national business register). Reporting units are legal units considered as enterprises or kind of unit (KAUs).
In Hungary, enterprises with five or more persons employed are observed via an exhaustive survey. Enterprises with 50 or more persons employed are observed via a full scope survey. Enterprises with 5-49 persons employed are observed via a sample survey. The sample is selected from the sample frame through random stratified sampling. Data for enterprises with fewer than five persons employed are estimated from administrative (tax) data for the previous two years.
Figure 44 shows the turnover of European countries as proportions of the EU-27 total (based on 2013 annual business statistics for special aggregates of activities).
Figure 44: Division J62 – turnover by Member State as proportions of the EU average, EU-27 (%)
Source: Eurostat (online data code: sbs_na_1a_se_r2)
Germany (23.3 %) and the UK (20.5 %) accounted for over 20 % of EU turnover for the division. France was third with 12.3 %. Four countries accounted for a proportion of between 4 % and 8 %, and another six countries between 1 % and
1.3
1.6
1.6
2.0
2.1
2.9
4.4
5.4
5.5
7.5
12.3
20.5
23.3
0.0 5.0 10.0 15.0 20.0 25.0
Czech Republic
Finland
Austria
Poland
Denmark
Belgium
Sweden
Netherlands
Spain
Italy
France
United Kingdom
Germany
Turnover share, 2013
%
172
3 %. The remaining Member States accounted for less than 1 % of the total (Hungary 0.8 %).
As regards SBS statistics for this industry, in the majority of countries (based on data for companies with over 20 employees) the turnover split by products of the total industry turnover (primary activity, secondary activities) is also available at four-digit level. In addition, Annex VIII to the SBS Regulation requires Member States to collect data and estimate the turnover share of four-digit service groups by sub-category (CPA).
13.2.2. Service producer price indices (SPPIs)
The STS Regulation introduced SPPIs, including for computer programming, consultancy and related activities. In principle, SPPIs are available on the Eurostat and NSI websites as indices with a reference period of a quarter, currently 90 days after the end of the reference period.
To comply with European and national rules, most countries (including Hungary) started to develop SPPIs in 2006 and to collect data for IT activities in 2007 (with base year 2006).
Under the STS Regulation, NSIs follow mainly a B2B approach. As enterprises focus their activities on business clients, the B2B index could be a suitable approximation of the B2All index. The source of the SPPI is usually a statistical survey by NSIs, using the compulsory quarterly or monthly questionnaires. The survey frame is generally the business register and the reporting unit is in principle the enterprise.
Member States’ price indices are calculated on the basis of the prices of services in their national currency.
Figure 45 shows 2013 yearly total (B2All) SPPIs for European aggregates and individual countries (base year 2010).
The increase of both European average prices was minimal (EU-27: +0.7 %; euro area: only 0.1 %). The highest price increase was measured in Latvia (19.8 %), followed by Bulgaria (13.1 %). In seven countries, the increase was between 3 % and 10 %, and in many countries the average annual price change fell in a narrow (two percentage point) band. In contrast, prices moderated in four countries; the largest reduction was recorded for the Netherlands (-7.8 %).
The pricing method used depends on the service specifications and record-keeping practice. The observation of different services requires different methods. Some can be measured using a pricing method based on working time; for others, model or component pricing is proposed as a pragmatic compromise in the face of a difficult measurement challenge.
Hungary’s general SPPI observation system has been developed to compile both industry- and product-based indices, but currently the latter are calculated and published. The sample is selected at class level (62.01, 62.02, 62.03, 62.09). First, enterprises classified in these classes are selected, then companies classified
173
elsewhere, if J62 activities’ account for a relevant proportion of their turnover as a secondary activities.
Price data are collected via the internet. At the end of each quarter, ‘personified’, pre-printed questionnaires are available for the reporting units, developed in close cooperation with companies. Prices are an average for the reference period. Taxes are excluded, but discounts, rebates and surcharges are taken into account. In general, non-domestic sales are observed as well as prices for domestic sales. Basically, prices are collected in national currency (forint), but they are also received in foreign currency (and converted to forint according to the National Bank’s quarterly average exchange rate), so the price development is also influenced by the exchange rate.
Figure 45: Computer programming, consultancy (J62) – annual SPPIs (2010 = 100)
Source: Eurostat (online data code: sts_sepp_a).
As this sector is characterised by a range of unique, hardly comparable services (e.g. provided in packages), the pricing method used most is that based on working time (hourly rates or charge per day), taking into account staff education levels and experience. For regularly provided services, contract pricing (e.g. data processing or the operation of hardware systems for a given period, HUF/month) or direct use of prices of repeated services (e.g. fee for software licence) are used. In some other cases (e.g. data backup), the unit value method (or average prices based on contracts) is applied. The most important price-determining
-7.8
-1.2
-0.7
-0.2
0.7
1.0
1.1
1.3
1.5
1.8
2.0
2.0
3.6
4.2
4.2
5.9
7.6
9.0
9.2
13.1
19.8
0.1
0.7
-10.0 0.0 10.0 20.0
Netherlands
Germany
Slovenia
Lithuania
Hungary
France
Denmark
Poland
Sweden
United Kingdom
Czech Republic
Spain
Austria
Ireland
Luxembourg
Belgium
Finland
Cyprus
Romania
Bulgaria
Latvia
Euro area
EU-28
Total output price indices, 2013
%
174
characteristics are staff competences and skills, the characteristics and complexity of the services provided and the number of items. However, the price of the service depends heavily on the transaction partners (resident, non-resident, multinational).
In Hungary, it is common for multinational companies to sub-contract and provide services with the aim of maximising overall firm profits and reducing taxes. Thus one of the main challenges of the price collection appears to be the observation of sub-contractors and collecting transfer prices for transactions between members of multinational corporations.
For weighting purposes, Hungary uses two sources for the aggregation of SSPIs. For company-level indices and sub-indices of ‘big’ companies, product-based turnover data for the previous year (collected by SPPI survey) are used. The total four-digit and higher level weights are industry-based (SBS) turnover data from two years earlier. Weights are updated annually.
As mentioned above, one issue is how to collect data on sub-contractors or members of multinational companies. Many Hungarian enterprises operate as an associated member of a non-resident multinational. In the IT sector, such enterprises account for a significant proportion of the turnover at transfer prices. The relevant price index should reflect this fact. However, collecting intra-group price data appears to be very difficult. It might be possible to use an input type of price index to measure transfer prices, but the structure and number of services provided may vary and will not be comparable over time. For some cases (considered as a pilot), a list of relevant and recurring cost elements (such as wages and payroll taxes by categories, rents, lease payments, etc.) has been developed and priced by the respondents. According to the collected data, price development in such cases appears to be very unstable. Given the increasing role of multinational companies, the development of appropriate methods (perhaps models) would increase the quality of SPPIs for this industry.
13.2.3. Number of persons employed
The purpose of the ‘number of persons employed’ index is to show the development of employment. This is defined as the total number of persons working in the observation unit (including working proprietors, partners working regularly in the unit and unpaid family workers) and persons who work outside the unit but who belong to it and are paid by it (e.g. sales representatives, delivery personnel, repair and maintenance teams).
According to this indicator, division J62 accounts for about 2 % of the total business economy (except financial and insurance activities). The highest proportion of workers are employed by the largest companies (over 32 % in enterprises with 250 or more employees). The lowest proportion was found in enterprises with between 50 and 249 employees.
13.3. Available data
Member States are required to transmit SBS and STS time series on computer programming, consultancy and related activities to Eurostat, using a specified
175
format. They compile these data on the basis of elementary data and aggregate them. The primary observation unit is the enterprise or the KAU registered in the Member State’s business register. Generally, administrative data sources (mainly tax data) are also used to estimate related values for companies with fewer than five employees or to impute missing data.
Almost all of the indicators are published in the Eurostat database, which provides not only European but also national aggregates.
The time series relating to SBS and STS indicators, and other related statistics are available from the Eurostat online database under the heading ‘Industry, trade and services’.66
13.4. Experimental calculation of an ISP indicator
13.4.1. Methodological considerations
To summarise, the main elements required for estimating an experimental ISP indicator are output (turnover) data and deflators — B2All-type SPPIs are considered the most appropriate indicators.
Currently, the statistical unit for all STS variables is the enterprise, but some countries use the KAU for statistical observation.
The turnover variable is to be transmitted in SA form.
Ideally, the ISP index should be an industry-based indicator, so it should reflect the change of volume of all services provided by enterprises or KAUs classified in this division.
Depending on the type of observation unit or the homogeneity of total turnover by products (primary activity, secondary activities), the following alternatives for calculation of an ISP could be considered:
� where total output/turnover is basically homogeneous, an appropriate product-based SPPI for J62 could be used as deflator for total industry output (B2All could be estimated by B2B; industry-based SPPI could be estimated by product-based SPPI); and
� where total output includes a significant proportion of secondary activities (services other than computer programming, consultancy and related activities), an industry-based deflator could be used for total industry output (product-based SPPI for J62 combined with useable deflators for relevant secondary activities — as a weighted average applying appropriate turnover share of elements).
In Hungary, based on SBS data, the primary/main activity (MA, J62) accounted for about 85 % of total division turnover in 2010. Secondary activities with turnover share of over 1 % are classified in divisions G46, G47, J58, J63, M69
66 Data are updated regularly, so data in this publication may differ from data downloaded subsequently.
176
and N78, of which only two (G46 and G47) accounted for a proportion of 2 % to 5 %.
In this case, the total output is basically homogeneous and a product-based SPPI62 could be used as a deflator.
To improve the quality of such indicators and thereby the quality of the ISP, it is possible to calculate an industry-based deflator for J62:
• as a weighted average of available national-level deflators for the main activity (J62) and secondary activities (e.g. at two-digit level); and
• as a weighted average of the enterprise-level industry-based SPPIs (enterprises with main activity J62).
13.4.2. Calculation of an experimental ISP indicator
13.4.2.1. Total output is basically homogeneous
The calculation is based on the index of turnover at current prices, TOVT and SPPI, PRON (B2All, product-based), 2010=100 (Source: database, Eurostat).
Generally, an experimental ISP for J62 could be calculated as follows:
ISPJ62 = index of turnoverJ62 / deflatorJ62
For European countries, the formula could be expressed more accurately:
ISPJ62 = TOVTJ62 / PRONJ62
For the experimental calculation, SA indices of turnover and unadjusted SPPI data were used.
Since, for the EU, the J62 price indices remained essentially stable over the observation period, the difference between the indices of turnover at current and constant prices is negligible (less than one percentage point).
177
Figure 46: J62 – index of turnover and experimental ISP, EU-27 (2010=100)
Source: test calculations based on Eurostat data
Figure 47: J62 – index of turnover and experimental ISP, euro area (EA-17) (2010=100)
Source: test calculations based on Eurostat data
The trends of the series for the EU and the euro area are very similar, but in the latter the difference between the indices of turnover at current and at constant prices is slightly over 1 % in the first five quarters of the observation period.
90
95
100
105
110
115
EU27-TOVT 2010 SA EU27-TOVV-PRON-2010 EU27-PRON -2010
90
95
100
105
110
115
EU17-TOVT 2010 SA EU17-TOVV-PRON-2010 EU17-PRON -2010
178
Figure 48: J62 – index of turnover and experimental ISP(I), Hungary (2010=100)
Source: test calculations based on Eurostat data
In the Hungarian time series, the difference between related indicators is reversed: Hungarian experimental ISPs are slightly higher than the indices at current prices at the beginning of the observation period. In the fourth quarter of 2012, the EU-27 ISP is 14 percentage points lower than the Hungarian experimental ISP.
13.4.2.2. Total output includes significant share of secondary activities
The deflator is computed as a weighted average of available national-level deflators. For the Hungarian experimental ISP, two cases were examined:
− Case 1: using SPPI62 as a proxy for the missing deflators (Def1); and
− Case 2: using SPPISTS as a proxy for the missing deflators (Def2).
As mentioned above, all SPPIs required by the STS Regulation are available in Hungary.
Table 47: Case 1 – estimated turnover structure and deflators, division J62 (main and secondary activity)
Total G46 sec.
G47 sec.
J62 main
J63 sec.
M69_70 sec.
N78 sec.
Other sec.
Turnover share (%)
100 7 85 4 4
Deflator(s) Def1 J62 PPI C26 SPPI J62
SPPI J63
SPPI M69_70
SPPI N78
SPPI J62
Source: estimates based on Hungarian SBS data
60
70
80
90
100
110
120
130
HU-TOVT 2010 SA HU-TOVV-PRON-2010 HU-PRON -2010
179
Because the weight of the deflator (used for main activity J62, as well as for some secondary activities) covers almost the total turnover, the product-based SPPI62 is practically the same as the computed industry-based deflator SPPI62 (DEF1).
Figure 49: J62 – index of turnover at current prices and experimental ISP (II-1), Hungary (2010=100)
Source: test calculations based on data from Hungary (HCSO database)
Because the weight of the deflator used for some secondary activities reaches 15 %, some difference could be observed between the product-based SPPI62 and the industry-based deflator SPPI62 (Def2). As a result, turnover at current prices and turnover at constant prices (ISP) diverge slightly (by less than 1 %) in a few quarters.
Table 48: Case 2 – estimated turnover structure and deflators, division J62 (main and secondary activity)
Total G46 sec.
G47 sec.
J62 main
J63 sec.
M69_70 sec.
N78 sec.
Other sec.
Turnover share (%)
100 7 85 4 4
Deflator(s) Def2 J62 PPI C26 SPPI J62
SPPI J63
SPPI M69_70
SPPI N78
SPPI STS
Source: estimates based on Hungarian SBS data
70
80
90
100
110
120
130
HU-TOVV-DEf1-2010 HU-TOVV-SPPI-2010 HU-PRON -2010 HU-Def1 J62 -2010
180
Figure 50: J62 – index of turnover at current prices and experimental ISP (II-2), Hungary (2010=100)
Source: test calculations based on data from Hungary (HCSO database)
The industry-based deflators (computed from enterprise-level industry-based SPPIs) are still under development.
13.5. References, classifications, and available data
Handbooks
Essential SNA: building the basics, Eurostat, 2012
System of national accounts 2008 (European Commission, IMF, OECD, United Nations, World Bank)
Methodology of short-term business statistics interpretation and guidelines, European Communities, 2006
Handbook on price and volume measures in national accounts, European Communities, 2001
Methodological guide for developing producer price indices for services, OECD-Eurostat, 2005
Voorburg papers
Computer programming, consultancy and related activities, Wizner, Ruth; Warsaw, 2012
Effects of globalisation of IT services, Holdway, Mike; Oslo, 2009
Measuring the globalisation and the IT services industry in Ireland, McMahon, Richard; Oslo 2009
Towards the consistent measures of main economic aggregates in time of globalisation, Norman, Vera; Oslo 2009
70
80
90
100
110
120
130
HU-TOVV-Def2-2010 HU-TOVV-SPPI-2010HU-PRON -2010 HU-Def2 J62 -2010
181
Sector paper on computer programming, consultancy and related activities, Holdway, Mike; Aguascaliantes, 2008
The charge–out rates applied to Swedish computer consultancy services, Norman, Vera; Aguascaliantes, 2008;
Computer programming, information service and related activities – turnover measures and practices at the US Census Bureau, Murphy, John; Seoul, 2007
Turnover and output measurement for the computer service industry in the Netherlands, Berends-Ballast, Mieke; Seoul, 2007
Computer services and related activities in Sweden, Norman, Vera, Seoul; 2007
Session on IT industries: Turnover/output in France, Camus, Benjamin; Gac, Denis; Salvatori, Patric; Seoul, 2007
SPPI for IT industries in France, Gac, Denis; Seoul, 2007
PPI for software consultancy services in the Netherlands, Kirsten, Uwe, Seoul, 2007
National and international studies
Hungarian Central Statistical Office, 2012, A szolgáltatási kibocsátási árak alakulása 2011-ben. (The development of service producer prices, 2011), available at: http://www.ksh.hu/docs/hun/xftp/idoszaki/pdf/szolgkiboar11.pdf (Hungarian only)
‘Computer programming and consultancy statistics — NACE Rev. 2’, Eurostat, Statistics Explained, 2013
‘Information and communication service statistics — NACE Rev. 2’, Eurostat, Statistics Explained, 2013
‘Business services statistics — NACE Rev. 1.1’, Eurostat, Statistics Explained, 2012
‘Computer and information services statistics — NACE Rev. 1.1’, Eurostat, Statistics Explained, 2012
‘A comparison of methods used to compile PEEIs in short-term business statistics’, Statistics in Focus 52/2010, Eurostat
‘Service producer price indices (SPPIs): a new European economic indicator’, Statistics in Focus 103/2008, Eurostat
182
NACE Rev. 2
62.0 Computer programming, consultancy and related activities
62.01 Computer programming activities
This class includes the writing, modifying, testing and supporting of software, i.e. designing the structure and content of, and/or writing the computer code necessary to create and implement systems software (including updates and patches), applications (including updates and patches), databases, web pages, customising of software, i.e. modifying and configuring an existing application so that it is functional within the clients’ information system environment.
This class excludes publishing packaged software, translation or adaptation of non-customised software for a particular market on own account, planning and designing computer systems that integrate computer hardware, software and communication technologies, even though providing software might be an integral part.
62.02 Computer consultancy activities
This class includes the planning and designing of computer systems which integrate computer hardware, software and communication technologies. Services may include related users training.
This class excludes the sale of computer hardware or software, the installation of mainframe and similar computers, the installation (setting-up) of personal computers, and the installation of software, computer disaster recovery.
62.03 Computer facilities management activities
This class includes the provision of on-site management and operation of clients’ computer systems and/or data processing facilities, as well as related support services.
62.09 Other information technology and computer service activities
This class includes other information technology and computer related activities not elsewhere classified, such as: computer disaster recovery services, installation (setting-up) of personal computers, software installation services.
This class excludes the installation of mainframe and similar computers, computer programming, computer consultancy, computer facilities management, data processing and hosting.
183
Table 49: CPA 2008 structure for J62
62 Computer programming, consultancy and related services 62.0 Computer programming, consultancy and related services 62.01 Computer programming services 62.01.1 IT design and development services
62.01.11 IT design and development services for applications 62.01.12 IT design and development services for networks and systems 62.01.21 Computer games software originals 62.01.29 Other software originals
62.02 Computer consultancy services 62.02.1 Hardware consultancy services 62.02.10 Hardware consultancy services 62.02.2 Systems and software consultancy services 62.02.20 Systems and software consultancy services 62.02.3 IT technical support services 62.02.30 IT technical support services 62.03 Computer facilities management services 62.03.1 Computer facilities management services 62.03.11 Network management services 62.03.12 Computer systems management services 62.09 Other information technology and computer services
62.09.1 Installation services of computers and peripheral equipment
62.09.10 Installation services of computers and peripheral equipment
62.09.2 Other information technology and computer services n.e.c.
62.09.20 Other information technology and computer services n.e.c.
184
14. INFORMATION SERVICE ACTIVITIES (J63)
Ildikó Holocsy, Central Statistical Office of Hungary
Within information and communication, the ‘information service activities’ division (J63) plays an important role at the level of total production of services and the economy as a whole.
The main objective of an ISP for this division is to provide a volume indicator for information service activities which is comparable with the existing volume indicators for industry and construction.
According to the international methodological guidelines, the ideal way of producing volume estimates is deflating each component by a strictly appropriate price index. This means that, to calculate a suitable ISP for information service activities, the most appropriate indicators are consistent output (turnover) data and SPPIs at the same level and frequency, according to the same classification system.
14.1. Available methodological information
Several kinds of methodological information and proposals concerning the production of SPPIs and volume indicators for information service activities are available (guides, handbooks, international and national studies). The general methodology has been developed in close consultation with international institutions (European Commission, IMF, OECD, United Nations, World Bank).
In addition, Voorburg Group documents (sector papers, revisited sector papers, mini-presentations and issues papers) reflect the main characteristics of classification, market conditions of production and pricing methods, and NA needs as regards information service activities.
Work on describing information service and related activities, including turnover measures and practices, started in 2007 at the Voorburg Group meeting in Seoul (South Korea) and was followed by discussions on key issues, including the challenges of globalisation. As a result, papers on effects were presented at a meeting in Oslo (Norway) in 2009 and on measurement of globalisation at a meeting in Ireland.
14.2. Presentation of the industry
In NACE Rev. 2, ‘information service activities’ are classified in section J (information and communication) division 63, which is divided into four classes: data processing, hosting and related activities (63.11), web portals (63.12), news agency activities (63.91) and other information service activities n.e.c. (63.99). Division J63 includes the activities of web search portals, data processing and hosting activities, and other activities that primarily supply information.
The CPA classification includes 11 sub-categories (product groups at six-digit level).
Most countries, including Hungary, apply fully harmonised national versions of the NACE Rev. 2 (industry-based) and CPA 2008 (product-based) systems.
185
Because of the nature of the IT industry (unique and/or bundled services, continuous innovation, cross-border trade in services, sub-contracting, offshoring), companies operating in this field may rapidly change the structure and content of their services according to their partners’ needs. This can cause difficulties in the classification of activities, particularly in determining what is the primary activity at class level. Secondary activities are also common alongside the primary/main activity.
In order to serve the growing client base in a professional manner, companies continuously develop their network and server rooms. New services include operations such as IT security services, the monitoring, design and functionality testing of websites and newsletters, and domain centre services, as follows:
− domain registration: this has been designed so as to allow partners to buy and transfer web domains simply and quickly, ensuring management of the renewal system. Typically, the extensions that can be registered for domain names are com, net, org, info, biz and name;
− DNS management: through the management interface of a domain centre, it is possible to manage the DNS settings of domains purchased or transferred by the client. DNS management is integrated with the client’s cloud computing infrastructure;
− cloud computing: this service is based on the infrastructure as a service (IaaS) model. Clients are hosted on the public platform of a data centre in which they want to use their servers. This is the best way to benefit from the flexibility and strength of the cloud system — to activate and manage virtual machines (VM) running on VMware. Clients will pay only for the VM infrastructure that they actually use; and
− redundant internet connection: the available bandwidth may increase from the usual 1 gigabit to 2 gigabits without additional cost, and availability for internet access also increases significantly.
In Hungary, information service activities as a primary activity accounted for nearly half (45 %) of total J63 turnover in 2012. The main secondary activities of companies classified in division J63 are ‘office administrative, office support and other business support activities’ (N82), ‘scientific research and development’ (M72), ‘computer programming, consultancy and related activities’ (J62), ‘legal and accounting activities’ (M69) and ‘rental and leasing activities’ (N77). Companies whose primary activity is one of the above, or ‘publishing activities’ (J58), ‘telecommunications’ (J61), ‘activities of head offices; management consultancy activities’ (M70), also provide information service activities.
Globalisation has expanded opportunities to trade and maximise profit at world level, but the specific trade effects of globalisation are also affecting companies operating in this industry (like those in division J62). Unfortunately, details on specific issues are lacking, but if the total output/turnover of the industry includes transactions between group members, the SPPI as deflator should include transfer prices. In such cases, quality adjustment and quality control is necessary.
As regards STS indicators (STS Regulation Annex D – ‘Other services’), turnover, number of persons employed and SPPIs for division J63 are required
186
quarterly. As for variable SPPIs, for countries with less than 4 % of the respective EU totals, the division (J63) coverage is required. For other countries, indices at three-digit level (J63.1 and J63.9) are required. For number of persons employed, the compulsory coverage is section J.
Table 50 shows the main economic indicators (number of enterprises, turnover, value added and number of persons employed) as proportions of total values for special aggregates of activities in the European annual business statistics. The highest proportion was calculated for number of enterprises (EU-27: 0.5 %; Hungary: 1.1 %). The proportion of value added and number of persons employed was 0.5 % and 0.4 % in EU-27 and 0.5 % in Hungary, while the proportion of turnover was only 0.2 %.
Similarly to the situation in division J62, the proportion of enterprises in the lowest employment category was approximately 95 %, while only 1 % had 50 employees or more. Proportions for all other indicators were also highest among firms employing 0 to 9 persons.
In Hungary, this industry is dominated by micro-level companies (with fewer than five employees). The highest proportions for all indicators from the totals (as with the European aggregates) was calculated for firms employing 0 to 9 persons.
Table 50: Division J63 – main economic indicators for as proportions of business economy totals (except financial and insurance activities), 2012
EU-27 Hungary
Total business economy
J63 %
Total business economy
J63 %
Number of enterprises 22 346 729 0.5 528 519 1.1
Turnover or gross premiums written (€ million)
26 074 073 : 254 233 0.2
Value added at factor cost (€ million)
6 184 837 0.5 46 497 0.5
Number of persons employed (100 persons)
1 337 673 0.4 2 430 681 0.5
Source: Eurostat (online data code: sbs_na_sca_r2).
According to the breakdown by firm size, there is a significant difference between EU-27 and Hungarian data. For example, the proportion of turnover accounted for by the ‘over 20 persons employed’ category in the EU-27 was 67 %, while in Hungary it was only 36 %. As regards value added, the proportions were 70 % (EU-27) and 60 % (Hungary).
187
14.3. Turnover
The turnover index is one of the PEEIs, the main purpose of which is to provide indicators of the economic cycle. Several further statistics can be derived from turnover data.
The most important source of data in most countries is a statistical survey. In principle, it is compulsory to provide data. Combinations of survey and administrative data are also used. Reporting units are legal units considered as enterprises or KAUs.
Data collection is similar to that for division J62. Figure 51 shows individual countries’ turnover as a proportion of total EU turnover in 2012, according to the annual business statistics for special aggregates of activities.
Figure 51: Division J63 — turnover by Member State as a proportion of EU average, EU-27 (%), 2012
Source: Eurostat (online data code: sbs_na_1a_se_r2).
The UK accounted for about 23 % of total EU turnover in division J63, three other countries for between 14 % and 19 % (Germany 18.2 %, Italy 17.0 % and France 14.4 %), another 10 countries for between 1 % and 5 % (e.g. Hungary 1.0 %) and the remaining Member States for less than 1 %.
For the majority of countries, SBS for this industry include a turnover split by products of total industrial turnover (primary activity and secondary activities) at four-digit level (based on data for companies with over 20 employees). In addition, Annex VIII to the SBS Regulation requires Member States to collect data and estimate turnover shares within the four-digit level service groups by sub-categories (CPA).
In Hungary, the most important area of this division is ‘data processing, hosting and related services’, followed by ‘other information services n.e.c. activities’.
Turnover share, 2012
1,0
1,4
1,5
2,0
2,3
2,7
2,8
2,8
2,8
4,7
14,4
17,0
18,2
23,2
0,0 5,0 10,0 15,0 20,0 25,0
Hungary
Czech Republ i c
Denmark
Belgium
Luxembourg
Spai n
Sweden
Poland
Netherlands
Austri a
France
Ital y
Germany
Uni ted Kingdom
%
188
The service group 63.1 accounts for over 66 % of total sales. News agency services account for a limited proportion (around 5 %), while firms operating in other information services n.e.c. account for nearly 29 %.
14.4. Service producer price indices (SPPIs)
SPPIs including information service activities were introduced by the STS Regulation and are also considered as PEEIs. SPPIs are available on the Eurostat and NSI websites as indices with a reference period of one quarter, currently 90 days after the reference period.
The Hungarian Central Statistical Office (HCSO) started developing SPPIs covering IT activities in 2007 (with base year 2006) to comply with European and national rules.
In line with the STS Regulation, most NSIs follow a B2B approach. As enterprises focus their activities on business clients, the B2B index could be a suitable approximation for the B2All index. The SPPI is usually based on a compulsory statistical survey using quarterly or monthly questionnaires. The survey frame is generally the business register and the reporting unit is in principle the enterprise.
Member States’ price indices are calculated on the basis of the prices of services in their national currency.
Figure 52 shows the yearly total (B2All) SPPIs in 2013, as compared with base year 2010, for European aggregates and by country.
The prices of information service activities in the EU fell by 0.3 % between 2010 and 2013.
The highest price increase was measured in Lithuania (8.7 %), followed by Bulgaria (8.5 %) and Austria (6.2 %). Seven countries saw an increase of between 1 % and 4 %. In contrast, prices moderated in eight countries. The biggest reduction was recorded in Spain (5.6 %).
The European and Hungarian time series followed a similar declining trend until the end of 2011. Thereafter, prices fell faster in Hungary than in the EU as a whole. The biggest differences (one percentage point) were measured in 2012.
The pricing method used depends on several factors relating to specific features of the various service activities and record-keeping practice. Some activities can be measured by contract pricing or methods based on working time; for others, it is proposed that unit values or average transaction prices of homogeneous services be calculated.
The Hungarian general SPPI observation system was designed to compile both industry-based and product-based indices, but currently only the latter are calculated and published. The sample is selected at class level (63.11, 63.12). First, enterprises classified in these classes are selected, then companies classified elsewhere, if ‘information service activities’ as secondary activities account for a relevant share of their turnover.
189
Figure 52: Information service activities (J63) – total output price indices (SPPIs), 2013 (2010 = 100)
Source: Eurostat (online data code: sts_sepp_a).
The price collection methods and transfer pricing issue are identical to those described for division J62.
In Hungary, the pricing method used most for this industry is contract pricing for regularly provided services (e.g. for data recording or hardware system operations for the given period, HUF/month) and pricing methods based on working time (hourly rates or charge per day), taking into account staff education levels and experience, and unit values (or average prices based on contracts) are used as appropriate, or the prices of repeated services (e.g. HUF/access, HUF/download) are used directly.
The most important price-determining factors are the characteristics and complexity of the services provided, staff competences and skills, and the number of items. However, the price of the service provided depends above all on transaction partners (resident, non-resident, multinational), the special needs of contractors and the euro exchange rate.
Hungary uses two sources of weights for the aggregation of SPPIs. For company-level indices and sub-indices of ‘big’ companies, product-based turnover data from the previous year (collected by the SPPI survey) are used. For four-digit and higher level aggregates, industry-based (SBS) turnover data from two years previously are applied. Weights are updated annually.
Total output price indices, 2013
1,3
1,8
2,6
2,7
2,8
2,8
4,0
6,2
8,5
8,7
-5,6-3,9
-3,6
-3,4
-2,7
-1,8
-2,4
-1,0
-0,3
-6,0 -4,0 -2,0 0,0 2,0 4,0 6,0 8,0 10,0
Spain
Romania
Germany
Sweden
Slovenia
Cyprus
France
Hungary
Belgium
United Kingdom
Poland
Latvia
Czech Republ ic
Denmark
Finland
Austria
Bulgaria
Li thuania
EU-28
%
190
14.5. Number of persons employed
The purpose of the ‘number of persons employed’ index is to show the development of employment. The number of persons employed is defined as the total number of persons who work in the observation unit (including working proprietors, partners working regularly in the unit and unpaid family workers), and persons working outside the unit who belong to it and are paid by it (e.g. sales representatives, delivery personnel, repair and maintenance teams).
In 2012, persons employed in division J63 represented 0.4 % of the business total (except financial and insurance activities) in the EU-27 and 0.5 % in Hungary. The majority of workers are employed by enterprises in the smallest size category (in employment terms).
14.6. Available data
Member States are obliged to transmit to Eurostat time series on information service activities relating to SBS and STS indicators. They compile these on the basis of elementary data and then aggregate them. The primary observation unit is the enterprise or KAU registered in the Member State’s business register. Generally, administrative sources (mainly tax data) are also used to estimate values for companies with fewer than five employees or to impute missing data.
Most of the indicators are published in the Eurostat database, which contains national as well as European aggregates.
The time series relating to SBS and STS indicators and other related statistics are available in Eurostat’s online database under the heading ‘Industry, trade and services’.67
14.7. Experimental calculation of an ISP indicator
14.7.1. Methodological considerations
The methodology and steps in the calculation are the same as for division J62, so the main elements required for estimating an experimental ISP indicator are:
• output (turnover) data; and
• deflators (Def) — B2All type of SPPIs.
The STS Regulation requires the production and transmission to Eurostat of turnover and SPPI data for information service activities. Currently, the observation unit for all variables is the enterprise, but some Member States use the KAU for data collection and aggregation. The turnover variable is to be transmitted in SA form.
The ISP index should reflect the change of volume of all services provided by enterprises or KAUs classified in this division.
67 Data are regularly updated, so data in this publication may differ from data downloaded subsequently.
191
Depending on the type of the observation unit or homogeneity of total turnover by product (primary activity, secondary activities), several alternative solutions for calculation of an ISP are proposed:
• where total output/turnover is basically homogeneous, a product-based SPPIJ63 could be used as a deflator for total industry output (B2All could be estimated by B2B; industry-based DefJ63 could be estimated by product-based SPPIJ63); or
• where total output/turnover includes a significant proportion of secondary activities (SA, services other than information service activities), an industry-based Def63 deflator should be used (product-based SPPIJ63 combined with useable deflators for relevant secondary activities — as a weighted average applying appropriate turnover share of the elements).
SBS data for Hungary show that J63 main activity accounted for only 45 % of the total turnover of division J63 in 2010. Secondary activities with a turnover share of over 2 % were classified in divisions J62, M69, M72, N77 and N82; the proportion accounted for by ‘office administration, office support and other business support activities’ (N82) was close to 40 %.
As mentioned above, these SBS data cover enterprises with over 19 employees; for this case, an industry-based DefJ63 deflator should be used. To estimate such a deflator for the total output of division J63, three approaches were examined, based on the calculation of indicators Def1J63, Def2J63 and Def3J63.
14.7.2. Calculation of an experimental ISP indicator (for homogeneous and heterogeneous outputs)
Case 1: Total output/turnover is basically homogeneous
The calculation is based on the index of turnover at current prices, TOVT, 2010=100 (SA) and the SPPI PRON (B2All, product-based), 2010=100, from the Eurostat STS database. In general terms, an experimental ISP for J63 could be calculated as follows:
ISPJ63 = Index of TurnoverJ63 / DefJ63
Using the Eurostat data codes this can be expressed as:
TOVVJ63 = TOVTJ63 / PRONJ63
Because the price indices for J63 remained essentially unchanged in the EU over the observation period (as compared with base year 2010), the difference between the index of turnover and ISP is negligible (no more than one percentage point).
192
Figure 53: Index of turnover at current prices (TOVTJ63) and experimental ISP (TOVVJ63), EU-27 (2010 = 100)
Source: test calculations based on Eurostat data
In Hungary, the difference between the indicators is reversed; it does not exceed 2 pp. The Hungarian ISP for division J63 grew faster than the EU ISP between 2010 and 2013. In the fourth quarter of 2013, it was 57 pp higher.
Figure 54: Index of turnover at current prices (TOVTJ63) and experimental ISP(I)J63 (TOVVJ63), Hungary (2010 = 100)
Source: test calculations based on Eurostat data
90
95
100
105
110
115
EU2...-TOVT 2010 SA EU2...-TOVV-PRON-2010 EU2...-PRON -2010
60
70
80
90
100
110
120
130
140
150
160
170
HU-TOVT 2010 SA HU-TOVV-PRON-2010 HU-PRON -2010
193
Case 2: Total output includes significant share of secondary activities
The deflator is computed as a weighted average of available national-level price indices. For the Hungarian experimental ISP, three methods were examined:
Method 1 (Def1): using SPPIJ63 as proxy for the missing deflators;
Method 2 (Def2): using SPPISTS as proxy for the missing deflators;
Method 3 (Def3): using section-level SPPIM, SPPIN and SPPISTS as proxies.
As mentioned above, all SPPIs required by the STS Regulation are available in Hungary, as well as section-level indices (STSM, STSN) and the index aggregated at the level of the services groups, as required by the STS Regulation (SPPISTS).
Deflators Def1J63, Def2J63 and Def3J63 were calculated as weighted averages of available aggregates: SPPIJ63 for main activity J63 and other SPPIs for relevant secondary activities at two-digit level (Figure 55).
Figure 55: Division J63 – total output price index and deflators, Hungary (2010 = 100)
Source: Test calculations based on data from Hungary (HCSO database).
98
99
100
101
102
103
HU-PRON -2010 HU-Def1 J63 -2010 HU-Def2 J63 -2010 HU-Def3 J63 -2010
194
Table 51: Division J63 – estimated turnover structure and deflators (main activity and secondary activities)
Activities CPA 2008
Total J63 main
J62 sec.
M69 sec.
M72 sec.
N77 sec.
N82 sec.
Other sec.
Turnover share
100 % 45.1 % 6.2 % 48.7 %
Deflators Def1J63
SPPIJ63 SPPIJ62 SPPIJ69_70
SPPIJ63
Def2J63 SPPISTS
Def3J63 SPPIM SPPIN SPPIN SPPISTS
Source: estimates based on Hungarian SBS data
Method 1:
Because the weight for SPPIJ63, as used for the main activity and most of the secondary activities, covers almost 94 % of total turnover, the product-based SPPIJ63 is similar to the computed industry-based Def1J63 deflator. Therefore, ISP(I)J63 and ISP(II)J63 (TOVV-Def1) are also similar. The maximum difference between them over the observation period is only 0.2 pp.
Figure 56: Index of turnover (TOVTJ63) and experimental ISP(II)J63 (TOVV- Def1), Hungary (2010 = 100)
Source: test calculations based on data from Hungary (HCSO database)
Method 2:
The weight used for the secondary activities deflated by SPPISTS covers almost 49 % of total J63 industry turnover, but ISP(I)J63 and ISP(II)J63 (TOVV-Def2) are also similar. The maximum difference between them over the 2009-2013 period is only 0.5 pp.
60708090
100110120130140150160170180
HU-TOVT 2010 SA HU-TOVV-Def1-2010 HU-Def1 J63 -2010
195
Figure 57: Index of turnover (TOVTJ63) and experimental ISP(II)J63 (TOVV- Def2), Hungary (2010 = 100)
Source: test calculations based on data from Hungary (HCSO database)
Method 3:
The aggregated weight of indices used for missing SPPIs of secondary activities exceeds 48 % and some difference could be observed between the product-based SPPI63 and the industry-based Def3J63. As a result, the difference between ISP(I)J63 and ISP(II)J63 (TOVV-Def3) reaches 4 pp in a few quarters.
Figure 58: Index of turnover at current prices (TOVTJ63) and experimental ISP(II)J63 (TOVV- Def3), Hungary (2010 = 100)
Source: test calculations based on data from Hungary (HCSO database)
60
70
80
90
100
110
120
130
140
150
160
170
HU-TOVT 2010 SA HU-TOVV-Def2-2010 HU-Def2 J63 -2010
60
70
80
90
100
110
120
130
140
150
160
170
HU-TOVT 2010 SA HU-TOVV-Def3-2010 HU-Def3 J63 -2010
196
Industry-based deflators computed from the enterprise-level industry-based SPPIs are under development.
14.8. Summary
As the information service activities industry is dominated by business clients, the B2All index could be estimated using the B2B type of SPPIs. It is assumed that B2B and B2All are nearly identical.
On the other hand, it is possible that the proportion of J63 secondary activities is significant (in Hungary, it is close to 40 %). For this reason, applying different deflators — a product-based SPPIJ63 or an industry-based DEFJ63 deflator — may result in modest differences in volumes.
To summarise: a cost-efficient deflator for the total output (turnover) of division J63 with a homogeneous turnover structure can be estimated by the product-based SPPIJ63 (B2All or B2B). However, where the ratio of secondary activities is high or the variability of the price indices used is greater, applying an industry-based deflator may result in more accurate data.
14.9. References, classifications, available data
Handbooks
Essential SNA: building the basics, Eurostat, 2013
http://epp.eurostat.ec.europa.eu/cache/ITY_OFFPUB/KS-RA-13-003/EN/KS-RA-13-003-EN.PDF
System of national accounts 2008 (European Commission, IMF, OECD, United Nations, World Bank)
http://unstats.un.org/unsd/nationalaccount/docs/SNA2008.pdf
Methodology of short-term business statistics interpretation and guidelines, European Communities, 2006
http://epp.eurostat.ec.europa.eu/cache/ITY_OFFPUB/KS-BG-06-002/EN/KS-BG-06-002-EN.PDF
Handbook on price and volume measures in national accounts, European Communities, 2001
http://ec.europa.eu/eurostat/ramon/statmanuals/files/KS-41-01-543-__-N-EN.pdf
Methodological guide for developing producer price indices for services, OECD-Eurostat, 2005
http://epp.eurostat.ec.europa.eu/cache/ITY_OFFPUB/KS-BG-06-003/EN/KS-BG-06-003-EN.PDF
Voorburg papers
http://www4.statcan.ca/english/voorburg/
Thesaurus of producer price indices for services (SPPs), Kennesey A. et al., 22nd Voorburg Group meeting, Seoul 2007
197
http://www4.statcan.ca/english/voorburg/Documents/2007%20Seoul/papers/03.pdf
Cross-cutting topics linking services turnover/output prices to the national macroeconomic framework – Country issues: Hungary, Ildikó Hamvainé Holocsy; Tokyo 2013
http://www4.statcan.ca/english/voorburg/Documents/2013%20Tokyo/Papers/1019.pdf
Sample size and number of observations as a balance between quality and producibility, Blang, Dorothee; von Borstel, Johanna, Warsaw 2012
http://www4.statcan.ca/english/voorburg/Documents/2012%20Warsaw/Papers/4014%20-%20efficiency_germany.pdf
Effects of globalisation of IT services, Holdway, Mike; Oslo, 2009
http://www4.statcan.ca/english/voorburg/Documents/2009%20Oslo/Papers/2009%20-%2005.pdf
Measuring the globalisation and the IT services industry in Ireland, McMahon, Richard; Oslo 2009
http://www4.statcan.ca/english/voorburg/Documents/2009%20Oslo/Papers/2009%20-%2009.pdf
Towards the consistent measures of main economic aggregates in time of globalisation, Norman, Vera; Oslo 2009
http://www4.statcan.ca/english/voorburg/Documents/2009%20Oslo/Papers/2009%20-%2007.pdf
Computer programming, information service and related activities – turnover measures and practices at the US Census Bureau, Murphy, John; Seoul, 2007
http://www4.statcan.ca/english/voorburg/Documents/2007%20Seoul/papers/18.pdf
Session on IT industries – turnover/output in France, Camus, Benjamin; Gac, Denis; Salvatori, Patric; Seoul, 2007
http://www4.statcan.ca/english/voorburg/Documents/2007%20Seoul/papers/21.pdf
National and international studies
Hungarian Central Statistical Office, 2013, Szolgáltatások, 2008-2012 (Services, 2008-2012), available at (Hungarian only): http://www.ksh.hu/docs/hun/xftp/idoszaki/pdf/szolgaltatasok_2008_2012.pdf
Hungarian Central Statistical Office, 2013, A szolgáltatási kibocsátási árak alakulása 2012-ben (The development of service producer prices, 2012), available at (Hungarian only): http://www.ksh.hu/docs/hun/xftp/idoszaki/pdf/szolgkiboar12.pdf
‘Business economy by sector — NACE Rev. 2’, Eurostat, 2013
http://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Business_economy_by_sector_-_NACE_Rev._2
198
‘Information services statistics — NACE Rev. 2’, Eurostat, Statistics Explained, 2013
http://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Information_services_statistics_-_NACE_Rev._2
‘Information and communication service statistics — NACE Rev. 2’, Eurostat, Statistics Explained, 2013
http://epp.eurostat.ec.europa.eu/statistics_explained/index.php/Information_and_communication_service_statistics_-_NACE_Rev._2
Eurostat, 2010, ‘A comparison of methods used to compile PEEIs in short-term business statistics’, Statistics in Focus 52/2010
http://epp.eurostat.ec.europa.eu/cache/ITY_OFFPUB/KS-RA-12-005/EN/KS-RA-12-005-EN.PDF;
Eurostat, 2008, ‘Service producer price indices (SPPIs): a new European economic indicator’, Statistics in Focus 103/2008
http://epp.eurostat.ec.europa.eu/cache/ITY_OFFPUB/KS-SF-08-103/EN/KS-SF-08-103-EN.PDF
Table 52: NACE Rev. 2 – section J (Information and communication)
63.0 Information service activities 63.1 Data processing, hosting and related activities; web portals 63.11 Data processing, hosting and related activities
This class includes the provision of infrastructure for hosting, data processing services and related activities, specialised hosting activities such as web hosting, streaming services, application hosting, application service provisioning, general time-share provision of mainframe facilities to clients, data processing activities. This class excludes activities where the supplier uses the computers only as a tool are classified according to the nature of the services rendered.
63.12 Web portals This class includes the operation of web sites that use a search engine to generate and maintain extensive databases of internet addresses and content in an easily searchable format, the operation of other websites that act as portals to the internet, such as media sites providing periodically updated content. This class excludes the publishing of books, newspapers, journals etc. via internet, broadcasting via internet.
63.9 Other information service activities 63.91 News agency activities
This class includes news syndicate and news agency activities furnishing news, pictures and features to the media. This class excludes activities of independent photojournalists, activities of independent journalists. 63.99 Other information service activities n.e.c. This class includes other information service activities not elsewhere classified such as computer-based telephone information services,
199
information search services on a contract or fee basis, news clipping services, press clipping services, etc. This class excludes activities of call centres.
Table 53: CPA 2008 – section J (Information and communication services)
63.1 Data processing, hosting and related activities; web portals 63.11 Data processing, hosting and related services
63.11.1 Data processing, hosting, application services and other IT infrastructure provisioning services 63.11.11 Data processing services 63.11.12 Web hosting services 63.11.13 Application service provisioning
63.11.19 Other hosting and IT infrastructure provisioning services
63.11.2 Streamed video and audio content 63.11.21 Streamed video content 63.11.22 Streamed audio content
63.12 Web portal content 63.12.1 Web portal content 63.12.10 Web portal content 63.9 Other information service 63.91 News agency services 63.91.1 News agency services 63.91.11 News agency services to newspapers and periodicals 63.91.12 News agency services to audio-visual media 63.99 Other information services n.e.c. 63.99.1 Information services n.e.c. 63.99.10 Information services n.e.c. 63.99.2 Original compilations of facts/information 63.99.20 Original compilations of facts/information
200
15. REAL ESTATE (L68)
Tony Vuillemin, INSEE
This section presents NACE Rev. 2 division L68 (real estate) and an experimental volume index (index of production) for it.
15.1. Overview of activities in division L68 (real estate)
15.1.1. NACE Rev. 2
Division L68 includes lessors, agents and/or brokers in one or more of the following areas: selling or buying real estate, renting real estate and providing other real estate services, such as appraising real estate or acting as real estate escrow agents. Activities may be carried out on own or leased property and may be done on a fee or contract basis. Also included is the building of structures, combined with maintaining ownership or leasing of such structures. The division also includes real estate property managers.
The division comprises three groups (NACE Rev. 2):
− 68.1 (buying and selling of own real estate);
− 68.2 (renting and operating of own or leased real estate); and
− 68.3 (real estate activities on a fee or contract basis).
Group 68.1 (buying and selling of own real estate) is composed of only one class (68.10), which includes the buying and selling of self-owned real estate (apartment buildings and dwellings, non-residential buildings, including exhibition halls, self-storage facilities, malls and shopping centres and land), and sub-dividing real estate into lots, without land improvement. Class 68.10 excludes the development of building projects for sale and the sub-dividing and improving of land.
Associated services (CPA 2008) are:
− 68.10.11 (residential buildings and associated land sales or purchases services);
− 68.10.12 (time-share properties sales or purchases services);
− 68.10.13 (residential vacant land sales or purchases services);
− 68.10.14 (non-residential buildings and associated land sales or purchases services); and
− 68.10.15 (non-residential vacant land sales or purchases services).
Group 68.2 (renting and operating of own or leased real estate) is also composed of one class (68.20), which includes the renting and operating of self-owned or leased real estate (apartment buildings and dwellings, non-residential buildings, including exhibition halls, self-storage facilities, and land), the provision of homes and furnished or unfurnished flats or apartments for more permanent use, typically on a monthly or annual basis, the development of building projects for
201
own operation and the operation of residential mobile home sites. Class 68.2 excludes the operation of hotels, suite hotels, holiday homes, rooming houses, campgrounds, trailer parks and other non-residential or short-stay accommodation places.
Associated services are:
− 68.20.11 (rental and operating services of own or leased residential real estate); and
− 68.20.12 (rental and operating services of own or leased non-residential real estate).
Group 68.3 (real estate activities on a fee or contract basis) is divided into two classes: class 68.31 for real estate agencies and class 68.32 for the management of real estate on a fee or contract basis.
Real estate agencies’ activities (class 68.31) include intermediation in the buying, selling and renting of real estate on a fee or contract basis and advisory activities and appraisal services in connection with the buying, selling and renting of real estate, on a fee or contract basis, and real estate escrow agents’ activities. This class excludes legal activities.
Associated services are:
− 68.31.11 (residential buildings and associated land sale services on a fee or contract basis, except of time-share ownership properties);
− 68.31.12 (time-share properties sale services on a fee or contract basis);
− 68.31.13 (residential vacant land sale services on a fee or contract basis);
− 68.31.14 (non-residential buildings and associated land sale services on a fee or contract basis);
− 68.31.15 (non-residential vacant land sale services on a fee or contract basis); and
− 68.31.16 (real estate appraisal services on a fee or contract basis).
Class 68.32 (management of real estate on a fee or contract basis) includes the activities of rent-collecting agencies. It excludes legal activities, facilities support services, i.e. the combination of services such as general interior cleaning, maintenance and making minor repairs, waste disposal, guard and security, and the management of facilities, such as military bases, prisons and other facilities.
Associated services are:
− 68.32.11 (residential property management services on a fee or contract basis, except of timeshare ownership properties);
− 68.32.12 (time-share property management services on a fee or contract basis); and
− 68.32.13 (non-residential property management services on a fee or contract basis).
202
15.1.2. The French classification (NAF rev. 2)
The French classification has an additional fifth level (sub-class) and is more detailed than NACE Rev. 2. The sub-classes are coded with a letter after the class number.
In the real estate division, class 68.20 (renting and operating of own or leased real estate) is separated into two sub-classes:
− 68.20A for the renting of residential dwellings (houses, apartment buildings, etc.); and
− 68.20B for the renting of non-residential buildings (offices, exhibition halls, commercial spaces, room of conference, meeting, grounds for agricultural use, etc.).
Class 68.32 (management of real estate on a fee or contract basis) is also divided into two sub-classes:
− 68.32A for the management of residential and non-residential dwellings; and
− 68.32B is created to isolate transactions due to succession or joint possession.
For the other classes in the real estate division (68.10 and 68.31), there is no difference between NAF rev. 2 and NACE Rev. 2.
15.2. The importance of real estate activities in the services sector
In France, real estate activities are carried out by 150 000 enterprises, which employ 300 000 persons (2010 data) and account for 13 % of enterprises in the services sector (as defined by Eurostat) and 4.5 % of the services workforce. These proportions are close to the EU-27 mean values (11.6 % and 5 %). In Norway and Latvia, about a third of enterprises operate in real estate activities; Switzerland has the lowest percentage (3.6 %).
203
Figure 59: Proportion of enterprises in real estate activities (%)
Source: Eurostat, 2010 data
In France, over half of the enterprises engaging in real estate activities (78 000) deal with the rental and operating of own or leased real estate (group 68.2).
Table 54: Structure of real estate industry, France and EU as a whole
Group Distribution of enterprises
France (%) EU (%)
68.1 10.9 10.2 68.2 52.1 61.4 68.3 37.0 28.4
Source: Eurostat, 2010 data
In France and in the EU as a whole, ‘real estate’ is the division of the services sector with the highest value added: 9.4 % (EUR 36.4 billion) of value added in the services sector in France in 2010, 9.9 % in the EU as a whole and 18.5 % (the highest proportion) in Sweden. Cyprus is the European country where real estate activities account for the lowest proportion (2.4 %).
0.0
5.0
10.0
15.0
20.0
25.0
30.0
35.0
Sw
itzer
land
Slo
veni
a
Cyp
rus
Slo
vaki
a
Por
tuga
l
Pol
and
Net
herla
nds
Cro
atia
Uni
ted
Kin
gdom
Rom
ania
Aus
tria
Spa
in
Irel
and
Eur
opea
n…
Hun
gary
Cze
ch R
epub
lic
Fra
nce
Bel
gium Ita
ly
Sw
eden
Bul
garia
Fin
land
Luxe
mbo
urg
Ger
man
y
Lith
uani
a
Est
onia
Den
mar
k
Nor
way
Latv
ia
204
Figure 60: Real estate services – proportion of value added in the total service sector, by country (%)
Source: Eurostat, 2010 data
02468
101214161820
Cyp
rus
Irel
and
Sw
itzer
land
Cro
atia
Slo
veni
aP
ortu
gal
Spa
inB
elgi
umU
nite
d…Ita
lyR
oman
iaB
ulga
riaF
ranc
eN
ethe
rland
sE
urop
ean…
Lith
uani
aP
olan
dH
unga
ryE
ston
iaC
zech
…S
lova
kia
Fin
land
Ger
man
y…A
ustr
iaN
orw
ayD
enm
ark
Latv
iaS
wed
en
205
In France, ‘real estate’ is also the division in which value added per employee is highest (EUR 12 170/person). Productivity varies between Member States, ranging from EUR 1 430/person in Bulgaria to EUR 31 540/person in Norway.
Table 55: Value added per person in the real estate sector, by country
Country Value added per person (EUR)
Bulgaria 1 430 Latvia 1 600 Lithuania 1 910 Hungary 2 560 Romania 2 630 Estonia 3 060 Poland 3 090 Portugal 3 260 Croatia 3 430 Ireland 3 730 Slovakia 4 450 Czech Republic 4 850 Cyprus 5 110 Italy 5 340 Slovenia 5 350 Spain 6 130 United Kingdom 6 790 Belgium 8 930 France 12 170 Switzerland 13 390 Germany 13 550 Denmark 13 880 Netherlands 15 180 Austria 16 940 Sweden 17 640 Norway 31 540
Source: Eurostat, 2010 data
206
In France, the value added of division 68 is determined mainly (75.8 %) by group 68.2 (renting and operating of own or leased real estate). Again, the proportions vary across Member States. In Portugal, group 68.1 (buying and selling of own real estate) has the highest value added (49 %). In Poland and Switzerland, group 68.3 (real estate activities on a fee or contract basis) is the most important.
Table 56: Distribution of value added in real estate in European countries (%)
Group
Country 68.1 68.2 68.3 Austria 2.8 78.1 19.1 Belgium 5.0 66.4 28.7 Bulgaria 4.2 80.7 15.1 Croatia 11.9 63.2 25.0 Cyprus 9.3 27.2 63.4 Czech Republic 0.1 82.3 17.6 Denmark 1.8 84.8 13.4 Estonia -4.9 95.0 10.0 France 2.9 75.8 21.2 Germany 4.5 82.2 13.4 Hungary 32.4 57.3 10.3 Italy 16.7 71.0 12.3 Latvia 2.9 62.6 34.5 Lithuania -0.4 87.9 12.5 Netherlands 8.5 79.0 12.5 Norway 5.8 82.5 11.7 Poland 6.3 36.2 57.5 Portugal 48.9 28.9 22.2 Romania 15.2 68.7 16.1 Slovakia 2.1 63.6 34.2 Slovenia 23.6 45.2 31.2 Spain 0.1 83.0 16.9 Sweden 0.3 90.3 9.4 Switzerland 10.5 25.1 64.4 United Kingdom 1.2 61.9 36.9
Source: Eurostat, 2010 data.
15.3. Method for calculating an ISP
The aim is to calculate a volume index (index of production) for each class and group of division 68 and an overall volume index for the whole division. Each index will be calculated on a monthly basis and published 60 days after the end of the month.
207
In France, a monthly turnover index (indice de chiffres d’affaires – ICA) has been produced for each level of classification (sub-class, class, group) since 1995. Turnover indices are calculated from an administrative source: enterprises have to fill in a VAT tax return and send it to the French administration (DGFiP). These returns contain enterprise turnover, in value, without VAT (all activities included), which allows INSEE (the French statistical institute) to measure the monthly development of services activities in value terms.
At class level, the volume index is the ratio of the turnover index to the price index. SPPIs measure changes in transaction prices, excluding VAT, for services sold by French enterprises. They are calculated by markets and by products. Volume indices are used to measure changes in activity in volume terms (at constant prices).
Turnover indices are calculated each month and published 60 days after the end of the month (M + 60), while most price indices are produced on a quarterly basis and published 60 days after the end of the quarter (T + 60).
To produce a monthly volume index, it is necessary to interpolate a monthly path from quarterly price indices with statistical methods based on the hypotheses of continuous first and second derivatives.
We want to publish volume indices 60 days after the end of the month, when quarterly price indices are not available, so we have to forecast the price index of the next quarter on the basis of ARIMA models.
Table 57: Availability of turnover and price indices for real estate activities
Reference month April May June Publication of turnover index
June July August
Publication of Q2 price index
August August August
Volume index (ISP)
M + 60 M + 60 M + 60: ICA/price index
ICA/price index ICA/price index Q2 price index available
Forecast April price index
Forecast May price index
Conversion of quarterly price index into monthly index and definitive ISP for April and May
At the end of June, we want to calculate the volume index for April. The turnover index for April is available, but the price index will be available only 60 days after the end of the quarter, at the end of August. We forecast the price index for quarter 2 and interpolate a monthly path to have an estimate of the April price index and publish a volume index for April at the end of June. For the publication of the volume index for June (the last month of the quarter), turnover and price
208
indices are available and we can calculate a volume index without having to use forecasts. Then, by interpolating a monthly path from the quarterly price index, we can revise volume indices for April and May.
So, thanks to prediction and interpolation, we can calculate gross working-day and SA volume indices for each class, group and division of the services sector 60 days after the end of the month.
Class level
(Gross) volume indices are the ratio of turnover indices and price indices.
For each month, ISPgross = gross
gross
iceIndex
ICA
Pr
Two possible calculation methods are tested to produce an SA–WDA ISP:
� the first ISP is the ratio of SA-WDA turnover and a price index. We test the seasonality of the price index. If there is no seasonality in the index, we use gross prices for the calculation of SA-WDA ISP; otherwise, we use a SA-WDA price index.
ISPSA-WDA1= gross
WDASA
iceIndex
ICA
Pr− or ISPSA-WDA1 =
WDASA
WDASA
iceIndex
ICA
−
−
Pr
� the second method (ISPSA-WDA3) involves seasonally and working-day adjusting the gross ISP: ISPSA-WDA3 = (ISPgross)SA-WDA.
Group level
After the calculation of volume indices at class level, we can calculate ISPs at group level. For each group, we aggregate volume indices of classes in the group using weights reflecting relative value added from each class.
ISPgross = ∑∈
×Groupi
igroupVA
iVA)(ISP
)(
)(gross
For the two methods above, we measure SA-WDA ISP by the following formula:
ISPSA-WDA = ∑∈
−×Groupi
WDASA iISPgroupVA
iVA)(
)(
)(
where:
ISPgross(i) is the gross ISP of class i;
ISPSA-WDA(i) is the SA-WDA ISP of class i;
VA(i) is the value added of class i; and
VA(group) is the value added of the group.
209
Division level
For the division level, we calculate volume indices by aggregating (weighting: VA) volume indices of the various groups.
ISPgross = ∑∈
×Divisioni
gross iISPdivisionVA
iVA)(
)(
)(
For the two methods above, we calculate SA-WDA ISP by the formula:
ISPSA-WDA = ∑∈
−×Divisioni
WDASA iISPdivisionVA
iVA)(
)(
)(
where:
ISPgross(i) is the gross ISP of group i;
ISPSA-WDA(i) is the SA-WDA ISP of group i;
VA(i) is the value added of group i; and
VA(division) is the value added of division 68.
Source: INSEE
After forecasting the price index for the next quarter, we interpolate a monthly path from the quarterly price index to calculate a monthly volume index. This is done using statistical methods based on the hypotheses of continuous first and second derivatives. Figure 63 presents the development of the monthly price index from 2005 to 2015.
100
101
102
103
104
105
Available date Estimated data
210
Figure 61: Monthly old dwelling price
Source: INSEE
80
85
90
95
100
105
110
Monthly old dwelling price
Monthly price index Quarterly price index
211
We can calculate a gross ISP and a SA-WDA ISP for class 68.10 as follows:
For each month, ISPgross = icegOldDwellin
ICA
Pr10.68
Figure 64 presents volume, turnover and price indices for class 68.10 between 2005 and 2015.
Figure 62: Turnover, price and volume indices for class 68.10
Source: INSEE
The ISP and the turnover index have the same profile. The variations of the price index are very small compared with those of the turnover index and do not result in ISP variations.
Calculation of SA-WDA ISP
Two methods are tested for the calculation of the SA-WDA volume index:
� Method 1:
ISPSA-WDA1= gross
WDASA
iceIndex
ICA
Pr− or ISPSA-WDA1 =
WDASA
WDASA
iceIndex
ICA
−
−
Pr
� Method 2: ISPSA-WDA2 obtained by eliminating seasonality of a gross ISP
First, we test the seasonality of the old dwelling price index. The result is negative, so we use the gross price index to calculate the SA-WDA ISP with the first method. Figure 65 depicts the indices obtained as a result of the two methods.
0
50
100
150
200
250
300
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
Volume index Price index Turnover index
212
Figure 63: SA-WDA indices obtained by the two methods
Source: INSEE
The trend of the two series is nearly the same. Locally, there are small differences between the series, especially at the beginning.
15.3.1. ISP 68.2: Renting and operating of own or leased real estate
NACE group 68.2 is composed of one class. In NAF rev. 2, this class is divided into two sub-classes:
− 68.20A — renting of residential dwellings (houses, apartment buildings...); and
− 68.20B — renting of non-residential buildings (offices, exhibition halls, commercial spaces, room of conference, meeting, grounds for agricultural use, etc.).
France has monthly turnover indices for each sub-class and for the class as a whole.
Three different price indices can be used to calculate a volume index for group 68.2:
• a monthly CPI which represents renting prices paid by households;
• a quarterly SPPI (B2B) following prices for renting offices; and
• a quarterly SPPI (B2All), which includes office rental prices paid by enterprises and prices paid by households to rent residential dwellings.
60
70
80
90
100
110
120
130
140
150
Method1 Method2
213
Two prices are produced quarterly. To calculate a monthly volume index and publish it at M+60, we forecast the values of the two quarterly price indices and interpolate them by using the same method as used for the old dwelling index.
As the tests of seasonality are negative for the three prices, we use gross price indices to calculate an SA-WDA ISP with the first method.
Figure 66 shows the development of the three monthly price indices between 2005 and 2015.
Figure 64: Development of price indices for real estate activities
Source: INSEE
To calculate a volume index for class 68.20, we tested two approaches:
� the first does not distinguish between sub-classes 68.20A and 68.20B and uses SPPIs (B2All) to calculate the volume index for class 68.20; it involves deflating turnover index at class level:
ISP68.20 = ICA 68.20 / SPPI B2All68.20
� in the second, we use different price indices for the two sub-classes: CPI for sub-class 68.20A and SPPI B2B for sub-class 68.20B.
Figure 67 depicts volume, turnover and price indices for class 68.20 between 2005 and 2015. The ISP and the turnover index are very similar.
80
85
90
95
100
105
110
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
CPI SPPI (68.20) SPPI (68.20B)
214
Figure 65: Turnover, price and volume indices for class 68.20
Source: INSEE
In the next step we calculate an SA-WDA ISP using the following two methods:
� Method 1: ISPSA-WDA1=gross
WDASA
iceIndex
ICA
Pr−
� Method 2: ISPSA-WDA2 obtained by eliminating seasonality of gross ISP
Figure 66: SA-WDA indices obtained by the two methods
Source: INSEE
The second method gives very different results at the beginning of the observation period. Here, we use different deflators for sub-classes 68.20A and
40
60
80
100
120
140
160
180
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
Price index Turnover index Volume index
90
95
100
105
110
Method1 Method2
215
68.20B and calculate a volume index for each. Then, we obtain a volume index for class 68.20 by aggregating (weighting according to value added) volume indices of the two sub-classes.
Table 58: Weights (value added) used for L68.20
Group Weight (%) 68.20A (renting of residential dwellings) 59.7 68.20B (renting of non-residential buildings) 40.3
Source: INSEE, 2010 data
First, we calculate the volume index for sub-class 68.20A:
ISP 68.20A = ICA 68.20A / CPI
Figure 67: Turnover, price and volume indices, sub-class 68.20A
Source: INSEE
60
70
80
90
100
110
120
130
140
150
160
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
CPI Turnover index Volume index
216
Figure 70 depicts the development of the SA-WDA ISP obtained by the two methods.
Figure 68: SA-WDA indices obtained by the two methods
Source: INSEE
Secondly, we calculate the volume index for sub-class 68.20B:
ISP 68.20B = ICA 68.20B / price index B2B 68.20BCPI
Figure 69: Turnover, price and volume indices, sub-class 68.20B
Source: INSEE
90
92
94
96
98
100
102
104
106
Method1 Method2
40
60
80
100
120
140
160
180
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
Price index Turnover index Volume index
217
Figure 70: SA-WDA indices obtained by the two methods
Source: INSEE
Then we aggregate ISP 68.20A and ISP 68.20B to obtain gross ISP 68.20, using the following formula:
ISP 68.20 = 59.7 % * ISP68.20A + 40.3 % * ISP 68.20B
To calculate SA-WDA ISP, we use two different methods:
� Method 1:
68.20BISP 3% 40, 68.20AISP 59,7% ISP WDA1-SAWDA1-SAWDA1-SA ×+×=
� Method 2:
68.20BISP 3% 40, 68.20AISP 59,7% ISP WDA2-SAWDA2-SAWDA2-SA ×+×=
90
92
94
96
98
100
102
104
106
108
110
Method1 Method2
218
Figure 71: SA-WDA indices obtained by the two methods
Source: INSEE
Method 2 produces a less volatile series than Method 1.
In Figure 74, we compare the ISPsa-wda1 deflator at class level with the index deflated at the sub-class levels and subsequent aggregation.
Figure 72: Comparison of SA-WDA volume indices (Method 1) obtained by the two approaches
Source: INSEE
90
92
94
96
98
100
102
104
106
108
110
Method1 Method2
90
92
94
96
98
100
102
104
106
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
Approach1 Approach2
219
The two approaches yield very similar results. The values of the index obtained with the first method appear to be smaller between 2007 and 2009. Where possible, we will give priority to the second approach (using different deflators for each sub-class), which appears to be more rigorous. Unfortunately, we sometimes do not have price indices for the sub-class level; in such cases we will use the first approach.
15.3.2. ISP 68.3: Real estate activities on a fee or contract basis
We have monthly turnover indices for the two classes of group 68.3: the first for class 68.31 (real estate agencies) and class 68.32 (management of real estate on a fee or contract basis).
Quarterly SPPIs for classes 68.31 and 68.32 are used as deflators to calculate the volume index of group 68.3. For class 68.32, we do not distinguish between sub-classes (68.32A and 68.32B), but use the same deflator for the whole class.
SPPIs for classes 68.31 and 68.32 have been calculated from 2010. To interpret the development of volume indices over a longer period, we backcasted these using CPI of classes 68.31 and 68.32. Figure 75 presents the development of the two price indices between 2005 and 2015.
Figure 73: Development of price indices
Source: INSEE
We can calculate volume indices for class 68.31 from 2005 onwards using the following formula: ISP 68.31 = ICA68.31 / SPPI 68.31.
80
85
90
95
100
105
110
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
SPPI (68.31) SPPI (68.32)
220
Figure 74: Turnover, price and volume indices, class 68.31
Source: INSEE
Figure 77 compares two methods for calculating SA-WDA ISP:
� Method 1: ISPSA-WDA1= gross
WDASA
iceIndex
ICA
Pr−
� Method 2: ISPSA-WDA2 obtained by eliminating seasonality of gross ISP
Figure 75: SA-WDA indices obtained by the two methods
Source: INSEE
60
80
100
120
140
160
180
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
Price index Turnover index Volume index
80
85
90
95
100
105
110
115
120
125
130
Method1 Method2
221
The index obtained with the second method is more volatile than that obtained with the first method.
For class 68.32, we use the formula: ISP 68.32 = ICA 68.32 / SPPI 68.32.
Figure 76: Turnover, price and volume indices, class 68.32
Source: INSEE
Figure 77: SA-WDA indices obtained by the two methods
Source: INSEE
60
80
100
120
140
160
180
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
Price index Turnover index Volume index
80
85
90
95
100
105
110
115
120
Method1 Method2
222
Group 68.3 is divided into two classes: 68.31 — real estate agencies (weight in value added: 62.6 %) and 68.32 — management of real estate on a fee or contract basis (weight in value added: 37.4 %). We aggregate the two classes calculated previously (weighting: value added) to obtain an ISP for group 68.3:
ISP 68.3 = 62.6 % * ISP68.31 + 37.44 % * ISP 68.32
Figure 78: Gross volume index of group 68.3
Source: INSEE
� Method 1: 68.32ISP 37,4% 68.31ISP 62,6% ISP WDA1-SAWDA1-SAWDA1-SA ×+×=
� Method 2: 68.32ISP 37,4% 68.31ISP 62,6% ISP WDA2-SAWDA2-SAWDA2-SA ×+×=
60
70
80
90
100
110
120
130
140
150
160
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
Gross volume index
223
Figure 79: SA-WSA indices obtained by the two methods
Source: INSEE
15.3.3. ISP of the real estate division
With ISP 68.1, ISP 68.2 (second approach) and ISP 68.3 as calculated in this section, we can calculate a ISP for the whole division:
ISP 68 = 2.9 % * ISP68.1 + 75.8 % * ISP68.2 + 21.2 % * ISP68.3
The weightings represent the value added of groups 68.1, 68.2 and 68.3.
90
95
100
105
110
115
120
Method1 Method2
224
Figure 80: Gross volume index, division 68
Source: INSEE
As before, we calculate the SA-WDA ISP of division 68 using two different methods:
Method 1: 68.3ISP 21,2% 68.2ISP 75,8% 68.1ISP 2,9% ISP WDA1-SAWDA1-SAWDA1-SAWDA1-SA ×+×+×=
Method 2: 68.3ISP 21,2% 68.2ISP75,8% 68.1ISP %9,2ISP WDA2-SAWDA2-SAWDA2-SAWDA2-SA ×+×+×=
60
70
80
90
100
110
120
130
140
150
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
Gross volume index
225
Figure 81: SA-WDA volume index obtained by the two methods
Source: INSEE
As price indices for division 68 have no seasonality, we use the gross price index to calculate the SA-WDA ISP with the first method. In other divisions, e.g. division 55 (accommodation), price indices have a strong seasonality.
Regardless of the level, the results from methods 1 and 2 are quite similar in general, but can sometimes differ. The main drawback of the second method is that the development of the SA-WDA ISP can diverge from that of the turnover and price indices. This could be a problem, since users might not understand why. Nevertheless, some methodological experts prefer the direct approach, since they hold that the volume index has its own specific seasonal pattern.
Figure 84 presents SA-WDA ISP for the three groups in division 68 and the SA-WDA ISP for the whole division calculated with the first method. The graphs for ISP 68 and ISP 68.2 are very close, due to the predominance of group 68.2 (75.8 %).
90
92
94
96
98
100
102
104
106
108
Method1 Method2
226
Figure 82: Volume indices, groups 68.1, 68.2 and 68.3 and division 68
Source: INSEE
Role of deflation:
Figure 83: Real estate – Development of ISP and turnover index
Source: INSEE
The growth of the division 68 turnover index is due to inflation. In fact, the volume index has a decreasing trend.
60
70
80
90
100
110
120
130
140
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
Division 68 Group 68.1 Group 68.2 Group 68.3
90
92
94
96
98
100
102
104
106
108
Mar
-05
Sep
-05
Mar
-06
Sep
-06
Mar
-07
Sep
-07
Mar
-08
Sep
-08
Mar
-09
Sep
-09
Mar
-10
Sep
-10
Mar
-11
Sep
-11
Mar
-12
Sep
-12
Mar
-13
Sep
-13
Mar
-14
Sep
-14
Mar
-15
SA-WDA ISP SA-WDA turnover index
227
15.4. Conclusion
For the calculation of volume indices in the services sector, turnover indices should be deflated at the most detailed level, i.e. at five-digit (sub-class) level if price indices are available or at four-digit (class) level. The weightings used for aggregation should be gross value added at basic prices.
Table 59: Solutions depending on availability of price indices
Situation Method Price index for each sub-class Deflation at the most detailed level No price index for sub-classes Deflation at class level No price index for class Imputation (CPI, price for whole division) Most price indices are quarterly. To produce a monthly volume index, we have to interpolate a monthly path from quarterly price indices before deflation. It will also be necessary to forecast the value of the price index for the next quarter to publish a volume index 60 days after the end of the month. Volume indices will be revised when we have the real value of the price index.
To obtain SA-WDA volume indices, the preferred method is to deflate SA-WDA turnover index by the gross price index, since in the majority of cases price indices have no seasonality. Also, this method appears to result in less volatile volume indices.
Table 60: Solutions depending on seasonality of price series
Situation Method No seasonality in price index SA-WDA turnover index/gross price index Seasonality in price index SA-WDA turnover index/SA-WDA price
index The SA-WDA will be adjusted at the most detailed level and then the SA-WDA sub-indices will be aggregated.
15.5. References
Eurostat (2006), Methodology of short-term business statistics: interpretation and guidelines
Eurostat (2008), NACE Rev. 2: Statistical classification of economic activities in the European Community
OECD (2007), Compilation manual for an index of service production
UK Index of services (IoS) methodological manual
Statistics Sweden, Index of service production (ISP) methodological manual
228
16. LEGAL , ACCOUNTING AND MANAGEMENT SERVICES (M69_M702)
Wilco de Jong, Statistics Netherland
16.1. Available methodological information
The following documents contain information on the methodology for constructing an ISP:
− OECD ISP Manual;
− NA manuals (information on the methodology used to calculate production in the NAs); and
− Voorburg Group documents (sector papers, revisited sector papers, mini-presentations and issues papers); these reflect the main characteristics of classification, market conditions of production and pricing methods, and NA requirements as regards the activities in question.
16.2. Description of the industry
Division M69 (legal and accounting services) and group M70.2 (management services) are very varied.
Legal activities include the legal representation of one party’s interest against another party, whether or not before courts or other judicial bodies, the preparation of legal documents (e.g. articles of incorporation, partnership agreements or similar documents in connection with company formation, patents and copyrights) and other activities of notaries public, civil law notaries, bailiffs, arbitrators, examiners and referees.
Accounting activities include accounting and bookkeeping services, such as auditing of accounting records, preparing financial statements and bookkeeping.
Management services include public relations, communication activities, business and other management consultancy activities, the provision of advice, guidance and operational assistance to businesses and other organisations. They do not include educational consulting activities.
In the EU, the value added of this industry in 2012 was EUR 322 billion, or 9 % of the total value added of the non-financial service economy. The industry employs over five million people in the EU. Most of the value added in the sector is generated by SMEs (firms employing fewer than 250 persons).
229
Figure 84: NACE M69 and M70.2 (legal, accounting and management services) – value added and persons employed as proportions of non-financial service economy totals, 2012 (%)
230
16.3. Activities by NACE
Division 69 and group 70.2 include the following activities:
Table 61: NACE Rev. 2 — division M69 and group M702
691 Legal activities (Law court activities not included) 69101 Law firms 69102 Legal consultants 69103 Notary offices 69104 Bailiffs offices 69105 Patent offices
692 Accounting, bookkeeping and auditing activities; tax consultancy 69201 Chartered accountants 69202 Accounting and administration consultants 69203 Bookkeeping firms 69204 Tax consultants 69205 Other administrative offices
702 Management consultancy activities 7021 Public relations and communication activities 7022 Business and other management consultancy activities
Management and consultancy activities are mostly B2B, but tax consultancy in particular is also partly B2C.
Table 63 shows shares of turnover by activity in the Netherlands. The turnover of division M69 and M70.2 in 2010 was EUR 30 billion.
Table 62: Share of turnover by activities in the Netherlands, 2010 (%)
A service such as bookkeeping is not only provided in NACE 69.20.3 (an estimated 30 % of bookkeeping turnover), but also in other five-digit levels of NACE 69.2. Therefore, it is difficult to deflate with services prices below NACE three-digit level. The same is true for legal and management consultancy activities. Deflating therefore requires a SPPI with a good weighting schedule.
691 6 billion euro69101 Law firms 65%69102 Legal consultans 10%69103 Notary offices 15%69104 Bailiffs offices 7%69105 Patent offices 4%
692 10 billion euro69201 Chartered accountants 42%69202 Accounting and administration consultans 21%69203 Bookkeeping firms 18%69204 Tax consultans 11%69209 Other administrative offices 8%
702 14 billion euro7021 Public relations and communication activities 4%7022 Business and other management consultancy activities 96%
Accounting, bookkeeping and auditing activities; tax consultancy
Legal activities (Law court activities not included)
Management consultancy activities
231
16.4. Secondary activities
Companies are classified on the basis of primary activity, but they often have secondary activities too the size of which varies depending on the NACE category. Table 64 shows estimates of the secondary activities’ share of turnover. A distinction is made between types of service as described in NACE and activities other than services, such as trade.
Secondary activities account for only 5-6 % of Dutch turnover in NACE 69.1 and 69.2. They are more common in NACE 70.2, especially activities other than services. Because it is difficult to measure the kind of secondary activity, it is not possible to take this into account when deflating.
Table 63: Secondary activities by NACE in the Netherlands, 2010 (%, estimates)
16.5. Available data
The following sources are available and necessary for constructing an ISP for legal, accounting and management services (M69 and M70.2):
− STS turnover – total turnover is available on a quarterly basis and at NACE three-digit level. To compile an ISP, deflated turnover is the preferred method;
− STS SPPIs – prices are available on a quarterly basis and at NACE three-digit level. To compile an ISP, turnover deflated with an SPPI for the same activity is the preferred method. The use of an SPPI with wider industry coverage is seen as a less appropriate method;
− CPI – prices are available on a monthly basis for tax consultants (69.20.4). As with an SPPI, deflation with a CPI is seen as an appropriate method; using a CPI with wider coverage is less appropriate; and
− NA input/output tables – value added is available on a yearly basis at NACE three-digit level or lower. Input/output tables are needed for weighting SPPI and CPI for tax consultants (69.20.4) and also for constructing a weighting schedule.
% turnover of other services
% turnover of other activities then services
691 5 169101 Law firms 4 069102 Legal consultans 21 169103 Notary offices 1 069104 Bailiffs offices 5 169105 Patent offices 6 0
692 4 169201 Chartered accountants 3 169202 Accounting and administration consultans 3 169203 Bookkeeping firms 4 269204 Tax consultans 2 169209 Other administrative offices 12 5
702 3 67021 Public relations and communication activities 10 97022 Business and other management consultancy activities 3 6
Accounting, bookkeeping and auditing activities; tax consultancy
Management consultancy activities
Legal activities (Law court activities not included)
NACE
232
16.6. Experimental calculations of an ISP
We carried out test calculations for division M69 (legal and accounting activities) and group M70.2 (management consultancy activities) using all available data as mentioned above. First, we describe the results of the ISP and then the weights and choice of deflator.
The calculations are at NACE three-digit level. Figures 87–89 show the results for NACE 69.1, 69.2 and 70.2. Figure 90 shows the result for the combination of M69 and M70.2 (M69.1 and M69.2 are first aggregated to M69).
Figure 85: Legal activities (NACE 69.1) – ISP, SPPI and turnover, year-on-year change (2007-2014)
Figure 86: Accounting, bookkeeping, auditing activities and tax consultancy (NACE 69.2) – ISP, SPPI and turnover, year-on-year change (2007-2014)
-8.0
-6.0
-4.0
-2.0
0.0
2.0
4.0
6.0
8.0
10.0
1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
2007 2008 2009 2010 2011 2012 2013 2014
%
ISP SPPI STS turnover
-6.0
-4.0
-2.0
0.0
2.0
4.0
6.0
8.0
1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
2007 2008 2009 2010 2011 2012 2013 2014
%
ISP SPPI STS turnover
233
Figure 87: Management consultancy activities (NACE 70.2) – ISP, SPPI and turnover, year-on-year change (2007-2014)
Figure 88: Legal, accounting and management services (NACE M69 and M70.2) – ISP, SPPI and turnover, year-on-year change (2007-2014)
16.7. Methodological issues
16.7.1. Outer and inner weights
Two kinds of weight are calculated:
• outer weights are calculated for the purpose of aggregating ISP indices from lower levels of NACE. These are based on value added in the NA input/output tables; and
-20.0
-15.0
-10.0
-5.0
0.0
5.0
10.0
15.0
20.0
1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
2007 2008 2009 2010 2011 2012 2013 2014
%
ISP SPPI STS turnover
-10.0
-8.0
-6.0
-4.0
-2.0
0.0
2.0
4.0
6.0
8.0
10.0
12.0
1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4
2007 2008 2009 2010 2011 2012 2013 2014
%
ISP SPPI STS turnover
234
• inner weights are calculated for the purpose of combining different indices of turnover or prices. These weights are based on turnover from annual statistics. NA input/output tables are used to determine the breakdown of tax consultants’ turnover into B2B and B2C.
Table 64: Outer weights based on value added (EUR million), 2010
NACE Value added % 691 2 926 21 692 5 801 41 702 5 459 39
Table 65: NACE 69.2 – inner weights based on turnover (EUR million), 2010
NACE Description Turnover % 69.20.1 Chartered accountants 4 064 550 42 69.20.2 Accounting and administration consultants 1 994 998 21 69.20.3 Bookkeeping firms 1 711 058 18 69.20.4 B2B — tax consultants 1 081 872 11 69.20.4 B2C — tax consultants 10 000 0 69.20.9 Other administrative offices 721 483 8
16.7.2. Choice of deflator
Most of the turnover of M69 and M70.2 is generated with business customers. Only a small part is B2C (tax consultants, notaries and lawyers are partly B2C).
For deflating tax consultants (69.20.4), both the SPPI and CPI can be used. The inner weights based on turnover can be used to combine the SPPI and CPI in one price index. As shown in Table 66, the share of B2C in NACE 69.20.4 in the Netherlands is less than 1 %, so the SPPI is sufficient for deflating tax consultants. In other Member States where B2C is more important, it might be necessary to combine SPPI and CPI.
In contrast to most other industries, the SPPI of notary offices is constructed for both B2B and B2C, so its use is sufficient for deflating notary consultants (69.10.3). Because of the lack of data on lawyers’ B2C prices, it is assumed that there are no marked differences in price development between business and consumers for law firms (69.10.1) and legal consultants (69.10.2).
16.7.3. Methodological problems and advice
Method for calculations
The first choice is to use deflated turnover. The use of a B2B indicator (SPPI) is justified where most of a service industry’s turnover is generated with business customers (e.g. market research) or where it can be assumed that there are no
235
marked price differences between business and other customers. This is the case in NACE 69.1 and 70.2, where the use of SPPIs is sufficient.
The second option is a combination of B2B and B2C price indices. Where prices or consumed services are different for different groups of customers, a mixed approach (requiring also a breakdown of the weights with turnover) is recommended. This is the case in NACE 69.2. For deflating tax consultants (69.20.4), both the SPPI as CPI will be used.
Level of detail for calculations
For the compilation/deflation of the European ISP, the deflation of turnover at NACE three-digit level is considered appropriate. Indicators for higher levels are then calculated by aggregating up from lower levels. For sections M69 and M70.2, the NACE three-digit level is very useful.
Recommendations
� Perform seasonal adjustments;
� Use a combined price index (SPPI and CPI) for deflating tax consultants (69204) when B2C is more important. Data on B2C prices are preferred for law firms and legal consultants.
236
17. ARCHITECTURE , ENGINEERING , TESTING (M71)
Sebastian Metzl, Kerstin Wichmann, Dorothee Blang, Federal Statistical Office of Germany (Destatis)
17.1. Available methodological information
This section summarises the methodological information for calculating an ISP available from a number of handbooks, published in particular by Eurostat, the OECD and other international institutions. Many sector-specific papers on industry M71 were written by members of the Voorburg Group. These are not discussed in detail in this section, but are listed in the annex.
Compilation manual for an index of service production
The Compilation manual for an index of service production (OECD, 2007) sets out the purpose of, and possible methods for calculating, an ISP. It also provides recommendations for compiling an index of volume for services.
For NSI developers of the methodological basis of an ISP, the manual delineates the most important basics of business statistics, such as statistical units, classification systems and terminologies, and also explains the various formulas for calculating an index.
It also focuses on the relevant variables needed for an ISP and classifies the data into three groups (preferred, alternative and other). The annex is of particular interest for data collection, because it gives examples of elementary data for an ISP. Because the manual was published in 2007, the classification system by which the data are sorted is not up to date.
Handbook on price and volume measures in national accounts
The Eurostat Handbook gives an insight into indices, the uses of which are similar to those of an ISP, but which are calculated in NAs. It includes a detailed description of the aims of different indices and the consequences for the calculation method. In particular, it explains some problems relating to deflation.
Its conclusions are not for a particular industry, but apply to business statistics in general. The theoretical explanations are complemented by practical calculation examples. In addition, the elementary data needed for the calculation of acceptable indices are discussed in detail. Because the release date is 2001, the classification system is out of date.
Guidelines for compiling the monthly index of production in construction
The Guidelines for compiling the monthly index of production in construction (Eurostat, 2011) are a manual for calculating an index of volume for construction. This index is therefore similar to the ISP, but for another industry. The Guidelines include an overview of the most important issues regarding volume indices. Their structure follows that of the OECD manual, but they also give an insight into current practices in the Member States.
237
Methodology of short-term business statistics
Indices with similar purposes to those of the ISP are already calculated for other industries in STS. The methodology handbook (Eurostat, 2006) gives a concise overview of the various industries covered in STS, including the volume indices computed. In addition, the associated documents (Eurostat, 2006) give a detailed description of the calculation formulae used.
Producer price index manual
PPIs are important for calculating an ISP, because they are elementary data for the deflation of turnover. The Producer price index manual, (IMF, 2004) gives a detailed description of the method for calculating these important data. The purpose of PPIs is discussed, but different types of index and the aggregation of price information are also explained, together with the consequences for the result of the compilation. The manual describes price collection problems, methods and sources, and addresses important issues such as the quality change of goods.
Methodological guide for developing producer price indices for services
The Eurostat-OECD SPPI Guide (2nd edition, 2014) covers the most important issues as regards the construction of price indices. It complements the information in the Producer price index manual by concentrating on the service industry. It gives examples of how price indices are calculated in practice in various countries. It focuses on industry-specific issues, rather than index theory per se. In addition, it contains examples of calculations to help the reader understand the most important aspects of data collection and compiling methods.
Thesaurus of producer price indices for services
This short Voorburg Group paper (Kenessey et al., 2007) gives a description of several pricing methods in relationship to SPPIs. It discusses some examples and explains certain terms.
17.2. Presentation of the industry
This section presents an overview of the architectural, engineering and technical services sector, as covered by NACE Rev. 2 division M71; this comprises three economic activities:
• M71.11 — architectural activities;
• M71.12 — engineering activities and related technical consultancy,; and
• M71.20 — technical testing and analysis.
According to the NACE Rev. 2 notes, division M71 includes:
− the provision of architectural services (e.g. consulting, design and drafting services, town planning);
− engineering services (e.g. engineering design and implementation, geophysical, geological and seismic surveying and mapping services);
238
− the performance of physical, chemical and other analytical testing services of materials and products (e.g. testing activities in the field of food hygiene, reliability testing, testing and measuring of environmental indicators);
− the operation of police laboratories; and
− the certification of products such as transport equipment.
Architectural activities (M71.11)
Architects’ services range from small activities, such as compiling an environmental report, to large construction projects, such as a high-speed railway. All require a high level of professionalism, so certain qualifications are required of architects in most countries, possibly including an accredited university degree. In addition, it may be obligatory to undergo a training period before being able to use professional titles, including their derivatives (US Department of Labor’s Bureau of Labor Statistics, 2014). So the title ‘architect’ is restricted to individuals licensed by law.
Engineering activities and related technical consultancy (M71.12)
Engineers solve technical, economic and social problems. They have to be technical experts, because they are designers of materials, structures and systems to find solutions for production and maintenance. Engineers may analyse and supervise production processes. Additionally, they often assure the quality of the results. Some engineers have to be economic professionals, because they have to evaluate the time and cost required to complete projects or even supervise them generally. They have to respect the requirements of all stakeholders in an assignment and integrate them into the project plans, so engineers also need social expertise.
Engineering is grounded in applied sciences. Traditionally, engineers are developers, but unlike other scientists they do not work in basic research. Engineering involves applying basic natural scientific principles to practical problems, rather than establishing such principles. There are several engineering specialisations, from aerospace and agricultural to nuclear and surveying undertakings, which require a high degree of professionalism (US Department of Labor’s Bureau of Labor Statistics, 2014b).
Because of the high degree of professionalism required, in many countries, especially in continental Europe, engineers have to be licensed. In the UK, some engineering titles are protected, but not the term ‘engineer’ itself.
Technical testing and analysis (M71.20)
Testing and analysis services are mainly B2B. They vary according to the products and characteristics being tested, testing methods and standards. Most are one-off and complex, but some testing services are periodic and standardised, e.g. the inspection of motor vehicles.
239
Technical testing and analysis might often be to satisfy legal requirements or to confirm the safety and reliability of products and procedures.
17.3. Economic classifications
17.3.1. NACE Rev. 2
NACE Rev. 2 classes M71.11, M71.12 and M71.20 cover the following activities:
M71.11 — Architectural activities
This class includes architectural consulting activities, such as building design, drafting, town and city planning, and landscape architecture, but not interior decorating (M74).
M71.12 — Engineering activities and related technical consultancy
This class includes:
− engineering design (i.e. applying physical laws and principles of engineering in the design of machines, materials, instruments, structures, processes and systems) and related consulting activities;
− the elaboration of projects using air conditioning, refrigeration, sanitary and pollution control engineering, acoustical engineering, etc.; and
− geophysical, geologic and seismic surveying, geodetic surveying activities, such as land and boundary surveying activities, hydrologic surveying activities, sub-surface surveying activities, cartographic and spatial information activities.
It does not include test drilling in connection with mining operations (B09), activities of computer consultants and software development (J62), research activities relating to engineering (M72), industrial design (M74) and aerial photography (M74).
M71.20 — Technical testing and analysis
This class includes physical, chemical and other analytical testing of all types of material and product, such as:
− acoustics and vibration testing;
− testing the composition and purity of minerals, etc.;
− testing activities in the field of food hygiene, including veterinary testing and control in relation to food production;
− testing the physical characteristics and performance of materials, such as strength, thickness, durability, radioactivity, etc.;
− qualification and reliability testing;
240
− performance testing of complete machinery: motors, automobiles, electronic equipment, etc.;
− radiographic testing of welds and joints;
− failure analysis;
− testing and measuring environmental indicators: air and water pollution, etc.;
− the certification of products (including consumer goods, motor vehicles, aircraft, pressurised containers, nuclear plants, etc.), periodic road-safety testing of motor vehicles, testing with use of models or mock-ups (e.g. of aircraft, ships, dams, etc.); and
− the operation of police laboratories.
It does not include the testing of animal specimens (M75) or diagnostic imaging, testing and analysis of medical and dental specimens (Q86).
17.3.2. ISIC Rev. 4
In ISIC Rev. 4, M71 contains only two classes: 71.10 (architectural and engineering activities and related technical consultancy) and 71.20 (technical testing and analysis). NACE Rev. 2 is therefore more detailed at its lowest (class) level than ISIC Rev. 4.
17.3.3. National economic activity classifications
NSI are allowed to have a more detailed economic activity classification with an additional fifth or even a sixth level (sub-class). The 2008 edition of the German classification of economic activities, which is relevant for all German-speaking countries in Europe (Germany, Austria and Luxembourg), has an additional fifth level. Division M71 is sub-divided as follows:
71.11 — Architectural activities 71.11.1 – Consulting architectural activities in building construction 71.11.2 – Consulting architectural activities in interior design 71.11.3 – Consulting architectural activities in town, city and
regional planning 71.11.4 — Consulting architectural activities in landscape
architecture 71.12 — Engineering activities and related technical consultancy 71.12.1 – Engineering activities for projects involving civil,
hydraulic and traffic engineering 71.12.2 – Engineering activities for projects in specific technical
fields and engineering design 71.12.3 – Surveying activities 71.12.9 – Other engineering activities and related technical
consultancy
241
71.20 — Technical testing and analysis 71.20.0 – Technical testing and analysis
The German classification is therefore no more detailed than NACE Rev. 2 at the fifth level for class 71.20, but the Dutch economic activity classification (SBI 2008) is (Papa, 2012, p. 4):
71.20 — Technical testing and analysis 71.20.1 – Testing and analysis of agricultural products and food 71.20.2 – Technical testing and analysis of machinery, equipment
and materials 71.20.3 – Other technical testing and analysis
In Finland (Standard Industrial Classification TOL 2008), class 71.11 is not separated into sub-classes, but classes 71.12 and 71.20 are divided into more detailed activities:
71.12 — Engineering activities and related technical consultancy 71.12.1 – Town and city planning 71.12.2 – Civil engineering activities 71.12.3 – Structural engineering activities 71.12.4 – Heating, plumbing and air-conditioning design 71.12.5 – Electrical engineering design 71.12.6 – Other construction activities 71.12.7 – Mechanical and process engineering design 71.12.9 – Other architectural and engineering activities and related
technical consultancy 71.20 — Technical testing and analysis 71.20.1 – Motor vehicle inspection 71.20.2 – Other technical testing and analysis
17.3.4. Typical products
CPC Rev. 2 provides an international classification of the products of activities in M71. In Europe, CPA 2008 is applied for this purpose and many countries also have an extended version of the CPA. CPC Rev. 2, ISIC Rev. 4, CPA 2008 and NACE Rev. 2 correspondence tables for class M71 are set out in the annex. The CPA 2008 products are as follows:
71.11.10 – Plans and drawings for architectural purposes 71.11.24 – Architectural advisory services 71.11.21 – Architectural services for residential building projects 71.11.22 – Architectural services for non-residential building projects 71.11.23 – Historical restoration architectural services 71.11.31 – Urban planning services 71.11.32 – Rural land planning services 71.11.33 – Project site master planning services
242
71.11.42 – Landscape architectural advisory services 71.11.41 – Landscape architectural services 71.12.11 – Engineering advisory services 71.12.12 – Engineering services for building projects 71.12.17 – Engineering services for industrial and manufacturing projects 71.12.14 – Engineering services for transportation projects 71.12.13 – Engineering services for power projects 71.12.18 – Engineering services for telecommunications and broadcasting projects 71.12.15 – Engineering services for waste management projects (hazardous and non-hazardous) 71.12.16 – Engineering services for water, sewerage and drainage projects 71.12.19 – Engineering services for other projects 71.12.20 – Project management services for construction projects 71.12.31 – Geological and geophysical consulting services 71.12.32 – Geophysical services 71.12.33 – Mineral exploration and evaluation services 71.12.34 – Surface surveying services 71.12.35 – Map-making services 71.20.11 – Composition and purity testing and analysis services 71.20.12 – Testing and analysis services of physical properties 71.20.13 – Testing and analysis services of integrated mechanical and electrical systems 71.20.14 – Technical inspection services of road transport vehicles 71.20.19 – Other technical testing and analysis services
Germany has no product-based classification other than the CPA, but Norway’s office for statistics has created an additional, more detailed product, ‘certification services’, due to the importance of this specialised technical testing service in Norwegian industry (Papa, 2012, p. 4). Certification is a process involving a test (usually by a third party) and issuing of a certificate or attestation.
17.4. Industry description
Nearly 10 % of enterprises in the services sector (NACE Rev. 2 sections H, I, J, L, M and N, and division S95) were classified under the economic activity code M71 in 2010. Therefore, approximately every 10th firm generated its value added mainly from architectural or engineering activities or technical testing and analysis. These activities accounted for around 6 % of total value added at factor cost and net turnover in the services sector, thus underlining the importance of division M71 in the EU.
243
Table 66: Division M71 – number of enterprises, EU-27 (2010)
NACE Rev. 2 code Number of enterprises % M 71 970 924 100.0 M 71.1 910 416 93.8 M 71.11 316 233 32.6 M 71.12 594 182 61.2 M 71.2 60 508 6.2 M 71.20 60 508 6.2
Source: Eurostat. online code data code sbs_na_1a_se_r2 (update 17 August 2015)
In 2010, division M71 was structured at European level as follows:
• M71 was the main economic activity of approximately 1 million enterprises, the majority of which (nearly 94 %) were classified as architectural or engineering entities, so M71.2 is less important as regards the number of enterprises (see Table 67);
• enterprises covered by activity code M71 generated turnover of about EUR 293 billion, i.e. over EUR 300 000 per enterprise;
• over 50 % of the turnover in the EU was generated in France, the UK and Germany, although only a quarter of the enterprises were based in those Member States;
• the enterprises employed nearly 3 million persons, i.e. an average of three persons per enterprise;
• the persons employed were mainly employees (around 72 %); and
• personnel costs in the industry amounted to EUR 91 billion, or 31 % of turnover; this equates to around EUR 43 000 per employee, of which wages and salaries made up around EUR 34 000.
Value added at factor cost per person employed varies enormously across Member States. The EU-27 mean is EUR 49 511, but the lowest level according to available data (see Table 68) is in Lithuania (EUR 10 278) and the highest in Denmark (EUR 85 100).
244
Table 67: Division M71 – mean value added at factor cost per person employed in the EU, 2010
Member State EUR
European Union 49 511 Belgium 58 775 Bulgaria 13 294 Czech Republic ./. Denmark 85 100 Germany 56 808 Estonia 18 337 Ireland 47 000 Greece 20 558 Spain 39 599 France 64 567 Croatia 24 338 Italy 41 872 Cyprus 38 992 Latvia 13 650 Lithuania 10 278 Luxembourg 73 710 Hungary 14 330 Malta ./. Netherlands 63 771 Austria 55 562 Poland 18 535 Portugal 22 507 Romania 13 769 Slovenia 30 872 Slovakia 18 130 Finland 53 730 Sweden 58 930 United Kingdom 64 713
Source: Eurostat. online data code: sbs_na_1a_se_r2 (last update 17 August 2015)
17.5. Development
Since 2003, the turnover of industry M71 (architecture, engineering, testing) has grown by over 40 % in the EU (Figure 91). Between the end of 2008 and 2010, a negative development could be observed, but this is typical of most economic activities because of the global financial and economic crisis. Germany’s economy recovered earlier than the European economy as a whole, but it too saw an abrupt fall in the first quarter of 2010.
245
Figure 89: Division M71 – turnover index, EU-28 and Germany (calendar and SA) (2010=100)
Source: Eurostat (Luxembourg), 2015.
To gain a better understanding of industry M71, we need to split the turnover index into M71.1 and M71.2. However, this breakdown is not available for the EU-28, so the diagrams below present only the data for Germany.
As can be seen in Figure 92, division M71 in Germany is dominated by architectural and engineering services (M71.1). The relative weight of M71.2 is too low to have a real influence on the M71 turnover index. The situation in other Member States is similar (Table 69).
65
75
85
95
105
115
125
135
20
03
Q1
20
03
Q3
20
04
Q1
20
04
Q3
20
05
Q1
20
05
Q3
20
06
Q1
20
06
Q3
20
07
Q1
20
07
Q3
20
08
Q1
20
08
Q3
20
09
Q1
20
09
Q3
20
10
Q1
20
10
Q3
20
11
Q1
20
11
Q3
20
12
Q1
20
12
Q3
20
13
Q1
20
13
Q3
20
14
Q1
20
14
Q3
Turnover Index European Union Turnover Index Germany
246
Figure 90: Division M71 – turnover index, Germany (original data) (2010=100)
Source: Federal Statistical Office of Germany (Destatis), Wiesbaden, 2015.
Figure 91: Division M71 – employment index, Germany (2010=100)
Source: Federal Statistical Office of Germany (Destatis), Wiesbaden, 2015
60
70
80
90
100
110
120
130
140
150
20
03
Q1
20
03
Q3
20
04
Q1
20
04
Q3
20
05
Q1
20
05
Q3
20
06
Q1
20
06
Q3
20
07
Q1
20
07
Q3
20
08
Q1
20
08
Q3
20
09
Q1
20
09
Q3
20
10
Q1
20
10
Q3
20
11
Q1
20
11
Q3
20
12
Q1
20
12
Q3
20
13
Q1
20
13
Q3
20
14
Q1
20
14
Q3
Turnover Index M71 Turnover Index M71.1 Turnover Index M71.2
80
85
90
95
100
105
110
115
120
125
20
03
Q1
20
03
Q3
20
04
Q1
20
04
Q3
20
05
Q1
20
05
Q3
20
06
Q1
20
06
Q3
20
07
Q1
20
07
Q3
20
08
Q1
20
08
Q3
20
09
Q1
20
09
Q3
20
10
Q1
20
10
Q3
20
11
Q1
20
11
Q3
20
12
Q1
20
12
Q3
20
13
Q1
20
13
Q3
20
14
Q1
20
14
Q3
Employment Index M71 Employment Index M71.1 Employment Index M71.2
247
Table 68: Division M71 – turnover breakdown (SBS), %
Member State 71.1 71.2 European Union 88.4 11.6 Belgium 87.0 13.0 Bulgaria 90.9 9.1 Czech Republic ./. ./. Denmark 93.2 6.8 Germany 83.9 16.1 Estonia 80.5 19.5 Ireland 94.8 5.2 Greece 94.3 5.7 Spain 88.2 11.8 France 88.6 11.4 Croatia 85.4 14.6 Italy 86.4 13.6 Cyprus 86.8 13.2 Latvia 73.9 26.1 Lithuania 82.5 17.5 Luxembourg 90.8 9.2 Hungary 88.4 11.6 Malta ./. ./. Netherlands 87.2 12.8 Austria 91.5 8.5 Poland 88.4 11.6 Portugal 91.4 8.6 Romania 88.6 11.4 Slovenia 90.6 9.4 Slovakia 84.2 15.8 Finland 88.8 11.2 Sweden 92.6 7.4 United Kingdom 91.3 8.7
Source: Eurostat. online data code: sbs_na_1a_se_r2 (last update 17 August 2015)
17.6. Available data
17.6.1. STS data
NACE code M71 is covered by Annex D to the STS Regulation, which requires Member States to submit quarterly data on turnover (code 120), number of persons employed (code 210), hours worked (code 220), wages and salaries (code 230) and producer prices (code 310).
The turnover and producer price variables must be at two-digit level (M71) only. These are the focus of this paper, because they are needed to produce the price adjusted turnover index for services activities as a whole. Data on the other variables have to be submitted at NACE Rev. 2 section level.
248
Under Commission Regulation (EC) No 1503/2006, the purpose of the turnover index is to show the development of the market for goods and services. Turnover comprises the totals invoiced by the observation unit (for STS and SBS, the enterprise) during the reference period, which correspond to market sales of goods or services supplied to third parties. Turnover also includes all other charges (transport, packaging, etc.) passed on to the customer, even if these are listed separately in the invoice. Excluded are VAT and other similar deductible taxes directly linked to turnover and all duties and taxes on the goods or services invoiced by the unit. Price reductions, rebates, discounts and the value of returned packing must all be deducted. Price reductions, rebates and bonuses conceded later to clients, e.g. at the end of the year, are not taken into account. Income classified in company accounts as other operating income, financial and extraordinary income is also excluded.
Under the Regulation, the purpose of the output price index is to measure the monthly development of transaction prices for economic activities. The domestic output price index for an economic activity measures the average price development of all goods and related services resulting from that activity and sold on the domestic market. The non-domestic price index shows the average price development (converted to local currency) of all goods and related services resulting from that activity and sold outside the domestic market. When combined, these two indices show the average price development of all goods and related services resulting from an activity.
It is essential that all price-determining characteristics of the products are taken into account, including quantity of units sold, transport provided, rebates, service conditions, guarantee conditions and destination. The specification must be such that the observation unit is able to identify the product precisely in subsequent reference periods and to provide the appropriate price per unit.
The following rules apply for determining prices:
− the appropriate price is the basic price excluding VAT, similar deductible taxes directly linked to turnover, and all duties and taxes on goods and services invoiced by the unit, but including any subsidies on products received by the producer;
− if transport costs are included, this should be part of the product specification;
− in order to show the price developments accurately, actual transaction prices should be used, not list prices;
− the output price index should take into account quality changes in products;
− the price collected for period t should refer to orders booked in period t (moment of order), not the moment when the commodities leave the factory gates; and
− for output prices on the non-domestic market, the price should be calculated at national frontiers (free on board – fob).
249
In principle, the index should reflect the average price during the reference period. In practice, the information collected may relate to a particular day in the middle of the reference period and be selected as a representative figure for the period. For products with a significant impact on the national economy that are known to be subject (at least occasionally) to volatile price developments, it is important that the index does reflect average prices.
Some particularities must be borne in mind as regards output prices for services:
• they show price developments for services delivered to customers that are enterprises or persons representing enterprises;
• domestic and non-domestic services are monitored; and
• the appropriate price measure is the transaction price, reflecting the revenue received by the producer for products actually sold to customers. This should take into account any applicable discounts, rebates, surcharges, etc. that may apply to the customers. Because the price reflects revenue received by the producer, taxes on products should be excluded but any subsidies on products received by the producer should be added. The price should be recorded at the date of delivery of the service. If delivery is spread over a number of periods, appropriate adjustments must be made.
17.6.2. SBS data
Division M71 (architecture, engineering, testing) is covered by various EU regulations on statistics. For SBS, Regulation (EC) No 295/2008 determines which data have to be submitted to Eurostat. It consists of eight annexes (modules). Division M71 appears in Annex 1 (the common module for annual structural statistics) and Annex 8 (the detailed module for structural statistics on business services). Under Annex 1, Member States have to submit the following annual data for M71 down to four-digit level:
Code Variable
11 11 0 – Number of enterprises 12 11 0 – Turnover 12 12 0 – Production value 12 15 0 – Value added at factor cost 12 17 0 – Gross operating surplus 13 11 0 – Total purchases of goods and services 13 12 0 – Purchases of goods and services for resale in the same condition as received 13 13 1 – Payments for agency workers 13 31 0 – Personnel costs 13 32 0 – Wages and salaries 13 33 0 – Social security costs 15 11 0 – Gross investment in tangible goods 16 11 0 – Number of persons employed 16 13 0 – Number of employees
250
16 14 0 – Number of employees in full-time equivalents
The following variables have to be submitted down to three-digit level, broken down by number of persons employed:
Code Variable
11 11 0 – Number of enterprises 12 11 0 – Turnover 12 15 0 – Value added at factor cost 16 11 0 – Number of persons employed
The following variables have to be submitted down to two-digit level, broken down by region according to NUTS2:68
Code Variable
11 21 0 – Number of local units 13 32 0 – Wages and salaries 16 11 0 – Number of persons employed
In Germany, additional data surveyed include numbers of:
− female employees;
− apprentices;
− part-time workers; and
− workers on low pay.
17.6.3. Further data from other statistics
Division M71 is also covered by Regulation (EC) No 530/1999 on structural statistics on earnings and labour costs, which requires Member States to submit annual NACE Rev. 2 division-level data in a four-year cycle.
Regulation (EU) No 549/2013 on the European system of national and regional accounts requires the Member States to submit:
− annual NACE Rev. 2 two-digit-level data on most of activities covered by the Regulation (A64 breakdown); and
− quarterly data based on an A10 breakdown, which means that these data are available only on at aggregate M_N level (sum of sections M and N).
17.7. Typical business / pricing models
As regards the typical business/pricing models and weighting of NACE Rev. 2 division M71, we have to distinguish between sub-divisions M71.1 (architectural
68 Commission Regulation (EC) No 105/2007 of 1 February 2007 amending the annexes to Regulation (EC)
No 1059/2003 of the European Parliament and of the Council on the establishment of a common classification of territorial units for statistics (NUTS).
251
and engineering activities and related technical consultancy) and M71.2 (technical testing and analysis).
Architectural activities (M71.11)
The main pricing methods for architectural activities are:
• pricing based on working time; and
• model pricing
Realised hourly charge-out rates can be distinguished by staff type, type of service or type of customer. The downside of the pricing method based on working time is the productivity bias: when productivity increases, the price decrease is not always measured. Also, architects often disregard record-keeping and therefore have problems in calculating charge-out rates.
To overcome the productivity bias, model pricing can be used. It also suits architectural services, since the activities performed are mostly one-off. The statistical office develops a model in agreement with the company. The model(s) should reflect the most common activities performed by the company. For architecture, this can be quite a difficult task, since the activities are diverse. The downside of model pricing in this industry is that companies are very reluctant to price models. It is very time-consuming, they find it very complex and only a handful of people within a company can actually do this type of calculation (and those persons often have too little time on their hands as it is). Furthermore, a virtual model is difficult to price, since the ‘stress factor’ is very low (much lower than with real market conditions).
The following steps have to be taken to collect model prices:
i. a description of a representative assignment is drawn up, in which the determining factors are fixed (type of assignment, characteristics of the building, complexity and hours of work);
ii. the persons and the corresponding hours of work are filled in; from this, respondents can calculate a price for the model (total price without negotiation);
iii. to capture the final price for the assignment, companies then adjust the calculated price to market conditions.
Alternative pricing methods include:
− percentage fee (Germany, France and Spain);
− contract pricing (Hungary); and
− component pricing based on repeated services for one client (USA).
Many architects in the Netherlands use percentage fees (% of total construction cost) as a check to see if the price of their service reflects market conditions. Germany is one of the few countries in Europe that have a scale of fees for architects. The Honorarordnung für Architekten und Ingenieure (HOAI) covers services in the construction sector. It describes in detail the services of architects
252
in several fields of activity and applies a key for the complexity of the building to determine a range for the architect’s charges.
Weighting
A very common approach to aggregation is to use turnover data as weights. These can be used for the company level, but also for the aggregation levels. Price changes are averaged per activity group using the geometric mean. At company level, turnover statistics are used, or turnover data from the companies themselves.
With help of building activity surveys, it is possible to estimate the proportions of buildings designed for residential and for business purposes. One can assume that owners of non-residential buildings are businesses. In the residential sector, the proportion of private and business house-builders can be estimated using the results of building activity surveys. Thus, for architectural activities we can derive a weighting scheme distinguishing between B2B and B2C services. Using construction costs, we can produce a B2B SPPI and combine it with the weighting scheme to produce a B2All estimator.
Engineering activities and related technical consultancy (M71.12)
Engineering services are mostly one-off and they are subject to changes in productivity, which complicates price measurement.
The model pricing method is the most appropriate for industries where service outputs are predominantly one-off. It can also cope (at least in theory) with changes in productivity. However, the following issues arose when we applied this method:
� models used are not representative;
� engineering companies are reluctant to estimate model prices; and
� the survey is subjective.
Although theoretically not the ‘best’ method, time-based pricing methods are commonly used to measure prices for engineering services. A method based on percentage fee of construction costs can be used for construction engineering services.
It is assumed that only these construction-related activities account for a significant proportion of private consumption. The proportion of construction engineering services represented by B2C services might be estimated in the same way as for architectural services. As with those services, the B2B SPPI may be supplemented by indices of construction prices to produce a B2All estimator.
Technical testing and analysis (M71.20)
There is a wide range of services in the technical testing and analysis industry because of the variety of products and characters being tested, testing methods and standards. Therefore, the majority of services are non-recurring and rather complex. As a result, the most appropriate pricing method(s) depends on the
253
service provided, the availability of alternative data sources and the cooperation of data suppliers.
The periodic road-safety testing of motor vehicles is an exception. As this is usually a repeated standard service, the direct use of prices, contract pricing and model pricing (or a combination of these methods) seems to be the most suitable approach. It is important, however, to ensure the necessary coverage of products sold as well as turnover accounted. A key distinction is that between administratively regulated and freely negotiated prices. Existing tariff systems on the most important testing services can easily be used as price quotations; this method is quite simple and effective.
As mentioned above, CPI as proxy for SPPI could be useable for services to final consumers (e.g. periodic road-safety testing services). CPIs may need additional conversion before being combined with an SPPI, because of differences in periodicity (monthly for CPI, quarterly for SPPI), methodology (taxes, e.g. CPIs include VAT, but SPPIs do not) or weighting structure.
The price of other testing and analysis services depends on the product and characteristic being tested, testing method and standard used, number of samples, promptness of results, etc. Pricing based on working time is a poor price indicator, because an extremely varied range of tests can take an hour. Although these services are often complex and one-off, it is possible to find services that reoccur; these include food production control, purity tests, air and water pollution tests, etc., where the same tests and analysis are performed regularly with the same testing methods and standards. Therefore, pricing based on actual contracts and simple model pricing are the most appropriate methods. Model pricing is very close to ‘contract prices’ or ‘prices for repeated services’; the main difference is that the model is still priced (by expert estimate) for periods in which the service was not provided.
Weighting
As regards the weighting structure, several countries get turnover data directly from companies, statistical sources or the regulatory authority with a breakdown for sub-markets (business and consumers). Other countries use data from NAs. Some countries aggregate CPI and SPPI using estimated sub-market shares, in particular for periodic road-safety testing services. In the field of technical, chemical and physical tests, it can be assumed that the share of final consumption is negligible. Nevertheless, it should be noted that the SPPI weighting pattern is highly sensitive to the choice of a B2B or B2All approach, because of the significant number of periodic road-safety tests on private vehicles.
In Germany, the information needed to construct B2B and B2C weighting schemes can be derived from federal office for motor traffic data on the number of safety tests broken down by kind of vehicle, technical inspection agency and type of owner (business or private). Price developments for road-safety tests on private vehicles are collected and published as part of the HICP (see above).
254
Figure 92: Producer price index for architectural and engineering services, and technical testing and analysis (2010=100)
Source: Federal Statistical Office of Germany (Destatis), Wiesbaden, 2015
17.8. Experimental calculations of the ISP
The ISPs in this section are calculated by dividing German turnover data by a price index at elementary NACE level (four-digit level for M71.1 and three-digit level for M71.2). The elementary indices are aggregated to the higher levels with SBS value added information for 2010.
17.8.1. Building elementary indices for the ISP
The time series for indices of turnover, prices and volumes in M71.11 is shown in Figure 95. A seasonal peak in turnover and volume data can be identified in the fourth quarter. It can be seen that the peak in turnover data is based on price increases in 2009, as the volume of production did not change at all. The effect of prices is even bigger in 2013. These examples show the interaction between prices, volumes and turnover.
In M71.12 (engineering activities and related technical consultancy), there is no such effect in the fourth quarter of 2009. There is a constant seasonal figure. An exception is the beginning of 2009 (Figure 96), where the financial crisis leads to a different pattern.
90
95
100
105
110
115
120
20
06
Q1
20
06
Q3
20
07
Q1
20
07
Q3
20
08
Q1
20
08
Q3
20
09
Q1
20
09
Q3
20
10
Q1
20
10
Q3
20
11
Q1
20
11
Q3
20
12
Q1
20
12
Q3
20
13
Q1
20
13
Q3
20
14
Q1
20
14
Q3
20
15
Q1
NACE 71.1 NACE 71.2
255
Figure 93: M71.11 (architectural activities) – price index, turnover ratios and experimental ISP (2010=100)
Source: Federal Statistical Office of Germany (Destatis), Wiesbaden, 2015
Figure 94: M71.12 (engineering activities and related technical consultancy) – price index, turnover ratios and experimental ISP (2010=100)
Source: Federal Statistical Office of Germany (Destatis), Wiesbaden, 2015
For M71.2 (technical testing and analysis), there is a different figure. The seasonal peak in the fourth quarter is not as high as for M71.1 and there is a trough in the first quarter. In 2009, the volume data development differs from that in other years: the spike in the second quarter is missing (Figure 97).
70
80
90
100
110
120
130
140
150
20
06
Q1
20
06
Q2
20
06
Q3
20
06
Q4
20
07
Q1
20
07
Q2
20
07
Q3
20
07
Q4
20
08
Q1
20
08
Q2
20
08
Q3
20
08
Q4
20
09
Q1
20
09
Q2
20
09
Q3
20
09
Q4
20
10
Q1
20
10
Q2
20
10
Q3
20
10
Q4
20
11
Q1
20
11
Q2
20
11
Q3
20
11
Q4
20
12
Q1
20
12
Q2
20
12
Q3
20
12
Q4
20
13
Q1
20
13
Q2
20
13
Q3
20
13
Q4
20
14
Q1
20
14
Q2
20
14
Q3
Turnover Index Price Index ISP
70
80
90
100
110
120
130
140
150
20
06
Q1
20
06
Q2
20
06
Q3
20
06
Q4
20
07
Q1
20
07
Q2
20
07
Q3
20
07
Q4
20
08
Q1
20
08
Q2
20
08
Q3
20
08
Q4
20
09
Q1
20
09
Q2
20
09
Q3
20
09
Q4
20
10
Q1
20
10
Q2
20
10
Q3
20
10
Q4
20
11
Q1
20
11
Q2
20
11
Q3
20
11
Q4
20
12
Q1
20
12
Q2
20
12
Q3
20
12
Q4
20
13
Q1
20
13
Q2
20
13
Q3
20
13
Q4
20
14
Q1
20
14
Q2
20
14
Q3
Turnover Index Price Index ISP
256
Figure 95: M71.2 (technical testing and analysis) – price index, turnover ratios and experimental ISP (2010=100)
Source: Federal Statistical Office of Germany (Destatis), Wiesbaden, 2015
In general, German PPIs are B2All, but B2B data were available for M71. Figure 98 compares B2All and B2B price and volume indices for M71.11.
The price index for business demand increases more than the B2All index in the third quarter of 2013. This leads to different volume indices for the various customer groups, so that the B2All ISP is higher than the B2B volume index.
Figure 96: M71.11 – comparison of B2B and B2All volume and producer price indices (2010=100)
Source: Federal Statistical Office of Germany (Destatis), Wiesbaden, 2015
70
80
90
100
110
120
130
140
150
20
06
Q1
20
06
Q2
20
06
Q3
20
06
Q4
20
07
Q1
20
07
Q2
20
07
Q3
20
07
Q4
20
08
Q1
20
08
Q2
20
08
Q3
20
08
Q4
20
09
Q1
20
09
Q2
20
09
Q3
20
09
Q4
20
10
Q1
20
10
Q2
20
10
Q3
20
10
Q4
20
11
Q1
20
11
Q2
20
11
Q3
20
11
Q4
20
12
Q1
20
12
Q2
20
12
Q3
20
12
Q4
20
13
Q1
20
13
Q2
20
13
Q3
20
13
Q4
20
14
Q1
20
14
Q2
20
14
Q3
Turnover Index Price Index ISP
70
80
90
100
110
120
130
140
150
20
06
Q1
20
06
Q2
20
06
Q3
20
06
Q4
20
07
Q1
20
07
Q2
20
07
Q3
20
07
Q4
20
08
Q1
20
08
Q2
20
08
Q3
20
08
Q4
20
09
Q1
20
09
Q2
20
09
Q3
20
09
Q4
20
10
Q1
20
10
Q2
20
10
Q3
20
10
Q4
20
11
Q1
20
11
Q2
20
11
Q3
20
11
Q4
20
12
Q1
20
12
Q2
20
12
Q3
20
12
Q4
20
13
Q1
20
13
Q2
20
13
Q3
20
13
Q4
20
14
Q1
20
14
Q2
20
14
Q3
Ind
ex
Price Index B2B Price Index B2All ISP B2B ISP B2All
257
17.8.2. Aggregation of elementary ISP data
The elementary ISP volume information calculated in the last section has to be aggregated to the upper levels of NACE. Because the aim is to get information on value added at constant prices, this is done with value added of an elementary level as a proportion of total value added from the industry in the base year. This information is taken from the German SBS.
First, the M71.11 and M71.12 ISPs are aggregated to produce a volume index at three-digit level (M71.1). Figure 99 shows the result, including the corresponding turnover ratio and price index. Because the value added of industry M71.12 dominates at three-digit level, the ISP seasonal effect of M71.1 is very similar to that for M71.12. The resulting graphs for M71 (architectural and engineering activities and related technical consultancy) constructed by aggregating the three-digit level indices are displayed in Figure 100.
As the price trend is upward, the turnover ratio and the volume index diverge with time. There is a constant seasonal figure in turnover and volume data, except in 2009.
Figure 97: M71.1 (architectural and engineering activities and related technical consultancy) – price index, turnover ratios and experimental ISP (2010=100)
Source: Federal Statistical Office of Germany (Destatis), Wiesbaden, 2015
70
80
90
100
110
120
130
140
150
20
06
Q1
20
06
Q2
20
06
Q3
20
06
Q4
20
07
Q1
20
07
Q2
20
07
Q3
20
07
Q4
20
08
Q1
20
08
Q2
20
08
Q3
20
08
Q4
20
09
Q1
20
09
Q2
20
09
Q3
20
09
Q4
20
10
Q1
20
10
Q2
20
10
Q3
20
10
Q4
20
11
Q1
20
11
Q2
20
11
Q3
20
11
Q4
20
12
Q1
20
12
Q2
20
12
Q3
20
12
Q4
20
13
Q1
20
13
Q2
20
13
Q3
20
13
Q4
20
14
Q1
20
14
Q2
20
14
Q3
Turnover Index Price Index ISP
258
Figure 98: M71 (architectural and engineering activities; technical testing and analysis) – price index, turnover ratios and experimental ISP (2010=100)
Source: Federal Statistical Office of Germany (Destatis), Wiesbaden, 2015
For M71, we produced calendar and SA turnover ratios. Figure 101 is the result of dividing these data by the price index.69
Figure 99: M71 (architectural and engineering activities; technical testing and analysis – price index, calendar and SA turnover ratios and experimental ISP (2010=100)
Source: Federal Statistical Office of Germany (Destatis), Wiesbaden, 2015
The volume index shows the effect of the economic crisis of 2008 and 2009.
69 Note that the seasonal adjustment method was not in line with the general approach taken in this guide.
70
80
90
100
110
120
130
140
150
20
06
Q1
20
06
Q2
20
06
Q3
20
06
Q4
20
07
Q1
20
07
Q2
20
07
Q3
20
07
Q4
20
08
Q1
20
08
Q2
20
08
Q3
20
08
Q4
20
09
Q1
20
09
Q2
20
09
Q3
20
09
Q4
20
10
Q1
20
10
Q2
20
10
Q3
20
10
Q4
20
11
Q1
20
11
Q2
20
11
Q3
20
11
Q4
20
12
Q1
20
12
Q2
20
12
Q3
20
12
Q4
20
13
Q1
20
13
Q2
20
13
Q3
20
13
Q4
20
14
Q1
20
14
Q2
20
14
Q3
Turnover Index Price Index ISP
70
80
90
100
110
120
130
140
150
20
06
Q1
20
06
Q2
20
06
Q3
20
06
Q4
20
07
Q1
20
07
Q2
20
07
Q3
20
07
Q4
20
08
Q1
20
08
Q2
20
08
Q3
20
08
Q4
20
09
Q1
20
09
Q2
20
09
Q3
20
09
Q4
20
10
Q1
20
10
Q2
20
10
Q3
20
10
Q4
20
11
Q1
20
11
Q2
20
11
Q3
20
11
Q4
20
12
Q1
20
12
Q2
20
12
Q3
20
12
Q4
20
13
Q1
20
13
Q2
20
13
Q3
20
13
Q4
20
14
Q1
20
14
Q2
20
14
Q3
Turnover Index Price Index ISP
259
17.9. Issues in calculating the ISP
For the final calculation of GDP, more detailed structural data are available, as for the STS. German NAs deflate output values of architectural activities, engineering activities and technical testing and analysis separately using sub-indices of the SPPI for industry M71. The output values of engineering activities and engineering activities as intermediate consumption are broken down even further into several service products. Also, we can use the results of the ‘turnover by product type and residence of client’ survey, which has been conducted since 2008 (in addition to the structural statistics on business services) to fulfil the data requirements of Annex VIII to the SBS Regulation.
Significant shares of sales from services for the final consumer are generated by architects and motor vehicle inspection services. German SPPIs are based on the B2All approach, so the ‘services of architects’ sub-index incorporates the (possibly divergent but suitably weighted) development of prices for architectural services for consumers. The same applies to the sub-index for the CPA product ‘technical inspection services of road transport vehicles’ and its weight within the SPPI for technical testing.
Where the SPPI is based on the B2B approach, it is extremely difficult to estimate revenues generated in retail business and subject them to an alternative deflator. None of the basic statistics include the information needed to sub-divide revenues by destination of outputs. At best, this is obtained at five-year intervals from producer price surveys to collect data for SPPI weighting structures.
17.10. Annexes
17.10.1. Sector papers and other sources
Professional, scientific and technical activity statistics - NACE Rev. 2 (website), European Union, 2015
United States Department of Labor’s Bureau of Labor Statistics, Occupational Outlook Handbook – Architects (website), BLS, 2014
United States Department of Labor’s Bureau of Labor Statistics, Occupational Outlook Handbook - Architecture and Engineering Occupations (website), BLS, 2014b
Papa, Aspasia, Updated sector paper on Industry ISIC 712 (technical testing and analysis), Voorburg Group, 2012
Wikipedia contributors, Architect (website), Wikipedia, The Free Encyclopedia, 2012
Wikipedia contributors, Engineer (website), Wikipedia, The Free Encyclopedia, 2012
von Borstel, SPPI for NACE 71.2 Technical testing and analysis - Germany (ppp), Voorburg Group, 2011
Holilokk, Kalko, Turnover and output measurement for technical testing and analysis in Norway (paper), Voorburg Group, 2011
260
Jenkins, Introduction : Technical testing and analysis (ppp), Voorburg Group, 2011
von Borstel, SPPI for technical testing and analysis in Germany, Voorburg Group, 2011
Kalko, Mini-presentation: NACE 71.2 Technical testing and analysis in Norway (SPPI) (ppp), Voorburg Group, 2011
Guidelines for compiling the monthly index of production in construction, Publication Office of the European Union, 2011
Matulska-Bachura, Mini-presentation on turnover / output for technical testing and analysis in Poland (ppp), Voorburg Group, 2012
Oertel, Turnover and output measurement for technical testing and analysis in Germany (ppp), Voorburg Group, 2011
Oertel, Röstel, Mini-presentation on turnover / output: turnover and output measurement for technical testing and analysis in Germany (paper), Voorburg Group, 2011
Spanjaard, SPPI for technical testing and analysis in the Netherlands, Voorburg Group, 2011
Eidman, A comparison of methods used to compile PEEIs in short-term business statistics, Eurostat, 2010
Goldhammer, ISIC Rev. 3.1 7421 / Rev. 4 7110 Architectural and engineering services and related technical consultancy (ppp), Voorburg Group, 2008
Goldhammer, ISIC Rev. 3.1 7421/Rev. 4 7110 Architectural and engineering activities and related technical consultancy (paper), Voorburg Group, 2008
Rémond-Tiedrez, Service producer price indices (SPPIs): a new European economic indicator, Eurostat, 2008
Sura, Architecture, engineering and technical testing, Eurostat, 2008
Alajääskö, Exports of business services, Eurostat, 2007
Kennesey et al.: Thesaurus of producer price indices for services (SPPs), Voorburg Group, 2007
Compilation manual for an index of service production, OECD, 2007
Assaf, Service price index for architectural services (paper), Voorburg Group, 2006
Bernard, High-tech industries and knowledge-based services, Eurostat, 2006
Methodology of short-term business statistics, European Communities Publications Office, 2006
Methodology of short-term business statistics - associated documents, European Communities Publications Office, 2006
OECD, Eurostat, Methodological guide for developing producer price indices for services, OECD, 2005
Alajääskö, Blackburn, Architectural and engineering activities and related technical consultancy, Eurostat, 2004
261
ILO et al.: Producer price index manual, ILO, 2004
Bylin, Ribe, Challenges in the development of a price index for engineering services in Sweden (paper), Voorburg Group, 2002
Rosenbaum, Price index for engineering services (paper), Voorburg Group, 2002
Handbook on price and volume measures, European Communities Publications Office, 2001
Enge, Developing PPIs on architectural and civil engineering services - the Norwegian experience (paper), Voorburg Group, 2000
Meguerditchian, Developing price indexes for consulting engineers: the Canadian experience (paper), Voorburg Group, 1994
Lowe, Pricing services of consulting engineers - update October 1992 (paper), Voorburg Group, 1992
Olsson, Engineering and other technical activities and the CPC (paper), Voorburg Group, 1992
April, Nijhowne, Engineering services and the CPC (paper), Voorburg Group, 1991
Lowe, Pricing of services of consulting engineers (paper), Voorburg Group, 1991
Ansmits, Quantity and price measures for business services: the case of engineering and architectural services (paper), Voorburg Group, 1989
Statistics Canada, Proposal for a classification of the primary products of ISIC 892 - technical activities (paper), Voorburg Group, 1987
262
Table 69: Correspondence of classifications (CPC Rev.2, ISIC Rev. 4, NACE Rev. 2 and CPA 2008)
CPC Rev 2
Description CPC Rev 2 ISIC Rev 4
NACE Rev 2
CPA 2008
Description CPA 2008
32550 Plans and drawings for architectural, engineering, industrial, commercial, topographical or similar purposes, being originals drawn by hand; handwritten texts; photographic reproductions and carbon copies of the foregoing
7110 7111 71 11 10 Plans and drawings for architectural purposes
83211 Architectural advisory services 7110 7111 71 11 24 Architectural advisory services 83212 Architectural services for residential building projects 7110 7111 71 11 21 Architectural services for residential building projects 83213 Architectural services for non-residential building projects 7110 7111 71 11 22 Architectural services for non-residential building projects 83214 Historical restoration architectural services 7110 7111 71 11 23 Historical restoration architectural services 83221 Urban planning services 7110 7111 71 11 31 Urban planning services 83222 Rural land planning services 7110 7111 71 11 32 Rural land planning services 83223 Project site master planning services 7110 7111 71 11 33 Project site master planning services 83231 Landscape architectural advisory services 7110 7111 71 11 42 Landscape architectural advisory services 83232 Landscape architectural services 7110 7111 71 11 41 Landscape architectural services 83310 Engineering advisory services 7110 7112 71 12 11 Engineering advisory services 83321 Engineering services for building projects 7110 7112 71 12 12 Engineering services for building projects 83322 Engineering services for industrial and manufacturing projects 7110 7112 71 12 17 Engineering services for industrial and manufacturing
projects 83323 Engineering services for transportation projects 7110 7112 71 12 14 Engineering services for transportation projects 83324 Engineering services for power projects 7110 7112 71 12 13 Engineering services for power projects 83325 Engineering services for telecommunications and broadcasting
projects 7110 7112 71 12 18 Engineering services for telecommunications and
broadcasting projects 83326 Engineering services for waste management projects (hazardous
and non-hazardous) 7110 7112 71 12 15 Engineering services for waste management projects
(hazardous and non-hazardous) 83327 Engineering services for water, sewerage and drainage projects 7110 7112 71 12 16 Engineering services for water, sewerage and drainage
projects 83329 Engineering services for other projects 7110 7112 71 12 19 Engineering services for other projects 83330 Project management services for construction projects 7110 7112 71 12 20 Project management services for construction projects 83411 Geological and geophysical consulting services 7110 7112 71 12 31 Geological and geophysical consulting services 83412 Geophysical services 7110 7112 71 12 32 Geophysical services 83413 Mineral exploration and evaluation 7110 7112 71 12 33 Mineral exploration and evaluation services 83421 Surface surveying services 7110 7112 71 12 34 Surface surveying services 83422 Map-making services 7110 7112 71 12 35 Map-making services
263
18. OTHER PROFESSIONAL, SCIENTIFIC AND TECHNICAL ACTIVITIES (M74)
Katarzyna Lipowska, Statistics Poland
The aim of this paper is to present an experimental ISP for NACE Rev. 2 division 74 (other professional, scientific and technical activities). We describe the industry, review the data availability. present an industry profile and finally propose a method for calculating an ISP.
18.1. General description of the industry
NACE division M74 covers the provision of professional scientific and technical services, with the exception of legal and accounting activities (M69), architecture and engineering activities, technical testing and analysis (M71), management and management consultancy activities (M70), research and development (M72) and advertising activities (M73). It is further split into:
− specialised design activities (M74.1);
− photographic activities (M74.2);
− translation and interpretation activities (M74.3); and
− other professional, scientific and technical activities n.e.c. (M74.9).
18.2. Data availability
Division M74 is covered by:
− Council Regulation (EC) No 1165/98 (STS);
− Regulation (EC) No 295/2008 (SBS); and
− Regulation (EC) No 530/1999 (structural statistics on earnings and labour costs).
The data required by the STS Regulation include turnover, number of persons employed, number of hours worked and gross wages and salaries. These are gathered on a quarterly basis and transmitted to Eurostat in unadjusted form. Only turnover variable is submitted at division level; the other variables are aggregated. Also, turnover has to be working-day adjusted.
The SBS Regulation lists statistics to be transmitted annually to Eurostat. These mainly concern structure, activity, competitiveness and performance of businesses. Examples of variables include turnover, number of enterprises, value added, personnel costs and number of persons employed. The data are broken down to NACE class level. Regional statistics include number of local units, number of persons employed and wages and salaries to be provided at division level and NUTS level 2.
The Regulation on structural statistics on earnings and labour costs requires Member States to deliver annual data every four years. These are provided at division level, including information on region (NUTS 1) and size class. In the case of labour costs statistics, total annual labour costs, average annual number of employees, annual number of hours worked and annual number of hours paid are
264
provided. Statistics on the structure and distribution of earnings include information on the form of economic and financial control and the type of collective pay agreement in force, as well as economic activity, region and size class. Employee characteristics are also supplied.
18.3. Industry profile
18.3.1. General overview
NACE division 74 includes almost 622 000 enterprises, which constitute nearly 16 % of enterprises in section M and nearly 5.9 % of non-financial services enterprises (NACE Rev. 2 sections H-J, L-N and division S95). These enterprises employ over 10 000 people, or roughly 9 % of persons employed in section M and almost 2 % of non-financial services workers. Figure 102 compares the divisions making up section M, in terms of number of employees and turnover.
Enterprises belonging to M74 generate turnover of EUR 90 billion (approximately 7 % of the section M and 1.5 % of the non-financial services total), value added of EUR 39 billion (almost 6.5 % and 1.5 %) and gross operating surplus of EUR 20 billion (roughly 9 % and 2 %). Figure 103 shows the shares of each division in section M as a whole.
Figure 100: Divisions in section M – shares in number of enterprises/turnover, EU-28 (2012)
Source: Eurostat
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
M 69 M 70 M 71 M 72 M 73 M 74 M 75
Number of enterprises Turnover or gross premiums written
265
Figure 101: Section M – shares of gross operating surplus, value added and personnel costs, by division, EU-28 (2012)
Source: Eurostat
18.3.2. Sectoral analysis
Regardless of how one assesses the structure of division M74 (be it on the basis of shares in value added or employment), ‘other scientific and technical activities n.e.c.’ (group M74.9) constitutes by far the most prominent sub-category (over 40 % of M 74), followed by M74.1 ‘specialised design activities’ (about 20 %) and M74.2 ‘photographic activities’ (over 10 %) and M74.3 (translation and interpretation activities’. For details, see Figure 104 (some data are unavailable for M74.3).
Figure 102: Division M74 – sectoral analysis, EU-27 (2010)
Source: Eurostat (online data code: sbs_na_1a_se_r2)
0.0%
5.0%
10.0%
15.0%
20.0%
25.0%
30.0%
35.0%
40.0%
45.0%
M 69 M 70 M 71 M 72 M 73 M 74 M 75
Value added at factor cost Gross operating surplus Personnel costs
0.0%
10.0%
20.0%
30.0%
40.0%
50.0%
60.0%
70.0%
M741 M742 M743 M749
Number of enterprises Value added at factor cost
Number of persons employed
266
18.3.3. Development
The turnover quarterly index for the EU-27 rose from about 90 % in 2003 (base year 2010) to 115 % in the second quarter of 2008. It then fell to 93 % in the first quarter of 2010 and rose to 110 % in the first quarter of 2013 (see Figure 105).
Index values varied greatly between countries: from about 20 % for Malta (2003Q1, 2003Q2, 2005Q4 and 2007Q3) and Bulgaria (2003Q3) to 315-520 % for Ireland (2012Q1 – 2013Q1). The biggest changes were in Ireland (2012Q2 - Q1), Hungary (2004Q1 – 2003Q4) and Romania (2005Q1 - 2004Q4); of these, only Ireland’s involved a rise (see Figure 106).
Figure 103: Division M74 – seasonally adjusted turnover index EU-27 (%, 2010=100)
Figure 104: Division M74 – seasonally adjusted turnover index, selected countries (%, 2010=100)
80
85
90
95
100
105
110
115
120
125
20
03
Q1
20
03
Q3
20
04
Q1
20
04
Q3
20
05
Q1
20
05
Q3
20
06
Q1
20
06
Q3
20
07
Q1
20
07
Q3
20
08
Q1
20
08
Q3
20
09
Q1
20
09
Q3
20
10
Q1
20
10
Q3
20
11
Q1
20
11
Q3
20
12
Q1
20
12
Q3
20
13
Q1
0
100
200
300
400
500
600
EU
Ireland
Hungary
Romania
267
18.3.4. Enterprise size
Micro enterprises (0-9 workers) employed the most people in M74 (69.2 %) and their share of value added was also the largest (58.6 %). 16.6 % of employees worked for small enterprises (10-49 people employed), producing 20.4 % of value added. Medium-sized enterprises (50-249 employees) employed 8.0 % of the workers in the sector, generating 11.2 % of value added. Large enterprises (over 250 workers) accounted for the lowest proportions both of number of persons employed (6.1 %) and value added (9.5 %).
Figure 105: Division M74 – breakdown by enterprise size, EU-27 (2012)
Source: Eurostat (online data code: sbs_sc_1b_se_r2)
18.3.5. Country analysis
The largest number of enterprises was registered in Italy (128 000; 188 000 persons employed), followed by France (81 000; 142 000), the UK (64 000; 158 000), Germany (54 000; 133 000) and the Czech Republic (45 000; 37 000).
UK enterprises generated most turnover and value added (EUR 18.9 billion and EUR 10.2 billion respectively), ahead of Germany (EUR 14.5 bn; EUR 7.1 bn), Italy (EUR 14.3 bn; EUR 6.6 bn), France (EUR 8.7 bn; EUR 3.6 bn) and the Netherlands (EUR 7.8 bn; EUR 2.2 bn). Spanish enterprises produced EUR 2.7 bn value added (see Figure 108).
0.1
0.1
0.20.6
Value added at factor cost
large
medium
small
micro
268
Figure 106: Division M74 – value added, by country, EU-28 (2010)
Source: Eurostat (online data code: sbs_sc_1b_se_r2)
18.4. ISP calculation
The experimental calculation of the ISP for division M74 for Poland was based on 2008-2012 data, with 2008 as a reference year. As class level corresponds to group level in the case of M74, the calculation was carried out at NACE three-digit level and later consolidated to two-digit (division) level. It involved the following steps:
1) pre-processing — data for STS are collected from the DG-1 monthly survey of economic activity, which provides up-to-date information and basic indicators of the economic activity of non-financial enterprises classified to industry, construction, retail trade and repair, and other services. This provides various kinds of information on a monthly and an accrued basis, including turnover in current prices.
Since 2008, SBS monthly turnover in current prices and quarterly weights have been available at three-digit level.
As no specific SPPI is available for M74, we used other deflators (see Table 71).
Table 70: M74 – deflators used for the production of an ISP
NACE Deflator Deflator details M74.1 (specialised design activities)
SPPI (quarterly) architectural and engineering activities; technical testing and analysis (M71)
M74.2 (photographic activities)
HICP-CT (quarterly) recreational and cultural services
M74.3 (translation and interpretation activities)
HICP-CT (quarterly) miscellaneous goods and services
0.30.2 0.2
0.1 0.1 0.1 0.00.1
1.0
0.0%
20.0%
40.0%
60.0%
80.0%
100.0%
120.0%
Value added at factor cost
269
M74.9 (other professional, scientific and technical activities n.e.c.)
HICP-CT (quarterly) miscellaneous goods and services
2) aggregation of monthly current price turnover series into quarterly data;
3) deflation by dividing quarterly current price turnover data by quarterly deflators for all series and each quarter of the analysis;
4) conversion of deflated series into indices with the average base value for 2008=100 (data not SA);
5) seasonal adjustment of the deflated indices using Demetra 2.2; and
6) consolidation of three-digit level deflated turnover indices (both raw and SA series) to two-digit level with SBS weights of 2008 for turnover using the chain-linked method. The indices series for the division level was calculated using a Laspeyres-type index according to the following formula:
Lt = ∑ (wi, t * Ci,t)
where: wi, t is the weight of group i at time t and Ci,t is the index for group i at time t.
Table 71: SBS weights used for compiling an index for M74 (Poland)
74.10 74.20 74.30 74.90 7.94 9.19 11.96 70.91
Figure 107: Deflated indices (2008=100) for M74 (step 4)
0
50
100
150
200
250
300
350
400
Q1
08
Q2
08
Q3
08
Q4
08
Q1
09
Q2
09
Q3
09
Q4
09
Q1
10
Q2
10
Q3
10
Q4
10
Q1
11
Q2
11
Q3
11
Q4
11
Q1
12
Q2
12
Q3
12
Q4
12
74.1 74.2 74.3 74.9
270
Figure 108: Seasonally adjusted deflated indices (2008=100) for M74 (step 5)
Figure 109: Unadjusted and adjusted ISPs for M74, Poland (2008=100)
0
20
40
60
80
100
120
140
160
180
200
Q1
08
Q2
08
Q3
08
Q4
08
Q1
09
Q2
09
Q3
09
Q4
09
Q1
10
Q2
10
Q3
10
Q4
10
Q1
11
Q2
11
Q3
11
Q4
11
Q1
12
Q2
12
Q3
12
Q4
12
74.1 74.2 74.3 74.9
0.0
20.0
40.0
60.0
80.0
100.0
120.0
140.0
160.0
Q1
20
08
Q2
20
08
Q3
20
08
Q4
20
08
Q1
20
09
Q2
20
09
Q3
20
09
Q4
20
09
Q1
20
10
Q2
20
10
Q3
20
10
Q4
20
10
Q1
20
11
Q2
20
11
Q3
20
11
Q4
20
11
Q1
20
12
Q2
20
12
Q3
20
12
Q4
20
12
M74 - raw M74 - adjusted
271
19. RENTAL AND LEASING ACTIVITIES (N77)
Michaela Maršálková, Czech Statistical Office
19.1. Available methodological information
The measurement of outputs and prices of rental services is discussed both in the Eurostat Handbook on price and volume measures in national accounts (Eurostat, 2001) and the Compilation manual for an index of services production (OECD ISP Manual (2007). Both were compiled with reference to earlier classifications (ISIC Rev 3.1/CPA 2002), but remain largely relevant, since the classifications correspond reasonably closely to NACE and revisions have not been extensive in this case (mainly involving the split of renting of transport equipment from the renting of cars). One important exception is the addition of the leasing of intellectual property and similar products (NACE 77.4), which was not included in these groups and therefore is not addressed specifically.
The Handbook recommends collecting gross turnover and actual rental prices, with corrections for changes in quality (both in the service and the rented good), or model prices for unique contracts. For services supplied to consumers, it recommends appropriate CPIs, although in cases where the services are supplied to both markets, use of CPI only is regarded as a B method (since it does not cover the whole industry). It allows for the possibility of using the actual price of the rented item, but only as a B method.
The ISP Manual also recommends deflating turnover by PPIs for appropriate rental activities, with CPIs used where applicable. Proposed alternatives are deflating turnover by PPIs or CPIs for the good in question, or volume indicators such as number of items leased or rented out.
The development of price indices for rental services is addressed in detail in the Voorburg Group’s revisited sector paper on rental and leasing services,70 which focuses mainly on developing a price index for ISIC categories 7710 (renting and leasing of motor vehicles) and 7730 (renting and leasing of other machinery, equipment and tangible goods), where it distinguishes several pricing methods (direct use of repeated services, average unit price and model pricing). According to the report, a major problem lies in variation of contract terms and possible quality changes of the leased item. These must be accounted for, but with care taken to avoid double-counting when the quality change was already included in the producing industry. The possible difficulties with properly excluding financial leases where a unit engages in both activities are also mentioned.
NACE 77 is included in the UK’s Index of services (IoS) and its treatment can be found in the methodology (GDP: output approach methodological guide, revised, Sharp P., 2003). The renting industry was subject to an industry review, which is also available on the ONS website. The review report followed the recommendations of the Handbook and deals with specific challenges of the industry sub-groups. The resulting approach was to cover the renting industry by
70 Fred Barzyk, Yana Mazin, Statistics Canada (24th Voorburg Group Meeting, 2009).
272
deflating gross turnover at NACE four-digit level by available deflators, i.e. CPIs for rental of personal and household goods, and renting of automobiles, SPPI for rental of construction and civil engineering machinery and equipment, and the underlying goods’ PPIs for the other groups.
19.2. Presentation of the rental and leasing service industry
NACE division 77 (rental and leasing activities) covers the renting and leasing of goods and assets in return for a one-time or periodic payment. It excludes the renting of real estate and the renting of cars or equipment with operator. In terms of business models, it covers only operating leases and specifically excludes financial leasing. The difference lies in the economic ownership of the leased asset: while in financial leasing the customer assumes all the risks and benefits from ownership of the asset, in operating leasing the risks are assumed by the legal owner. The owner is usually responsible for maintenance and repair of the rented good. The asset is rented for periods of time shorter than its expected lifetime. (This issue is further discussed in SNA 2008).
The division consists of four sub-divisions, which differs as regards business models and usual customers. The construction of the index should reflect this, preferably by deflating the lower NACE levels separately using appropriate deflators or, if that is not possible, by properly weighing together relevant price indices.
19.2.1. Classification
In NACE Rev. 2, group N77 is divided as follows:
Table 72: NACE Rev. 2 N77 (rental and leasing activities)
77 Rental and leasing activities 77.1 Renting and leasing of motor vehicles 77.11 Renting and leasing of cars and light motor vehicles 77.12 Renting and leading of trucks 77.2 Renting and leasing of personal and household goods 77.21 Renting and leasing of recreational and sports goods 77.22 Renting and leasing of video tapes and disks 77.29 Renting and leasing of other personal and household goods 77.3 Renting and leasing of other machinery, equipment and tangible goods 77.31 Renting and leasing of agricultural machinery and equipment
77.32 Renting and leasing of construction and civil engineering machinery and equipment 77.33 Renting and leasing of office machinery and equipment (including computers) 77.34 Renting and leasing of water transport equipment 77.35 Renting and leasing of air transport equipment 77.39 Renting and leasing of other machinery, equipment and tangible goods n.e.c.
77.4 Renting and leasing of intellectual property and similar products, except copyrighted works
273
77.40 Renting and leasing of intellectual property and similar products, except copyrighted works
The current UN classification, ISIC Rev. 4, is similarly constructed, but it does not further divide groups 77.1 and 77.3.
Product-wise, the CPA 2008 product classification has the same structure, but offers more detailed separation (Table 74).
Table 73: CPA 2008 for rental and leasing activities
77 Rental and leasing services 77.1 Rental and leasing services of motor vehicles 77.11 Rental and leasing services of cars and light motor vehicles 77.12 Rental and leasing services of trucks
77.12.11 Rental and leasing services of goods transport vehicles without driver 77.12.19 Rental and leasing services of other land transport equipment without driver
77.2 Rental and leasing services of personal and household goods 77.21 Rental and leasing services of recreational and sports goods 77.22 Rental services of video tapes and disks
77.29 Rental and leasing services of other personal and household goods
77.29.11 Rental and leasing services of televisions, radios, video cassette recorders and related equipment and accessories 77.29.12 Rental and leasing services of furniture and other household appliances 77.29.13 Rental and leasing services of musical instruments 77.29.14 Rental and leasing services of household linen 77.29.15 Rental and leasing services of textiles, clothing and footwear 77.29.16 Rental and leasing services of do-it-yourself machinery and equipment 77.29.19 Rental and leasing services of other personal and household goods n.e.c.
77.3 Rental and leasing services of other machinery, equipment and tangible goods 77.31 Rental and leasing services of agricultural machinery and equipment 77.32 Rental and leasing services of construction and civil engineering machinery and equipment
77.33 Rental and leasing services of office machinery and equipment (including computers)
77.33.11 Rental and leasing services of office machinery and equipment (excluding computers) 77.33.12 Rental and leasing services of computers
77.34 Rental and leasing services of water transport equipment 77.35 Rental and leasing services of air transport equipment
77.39 Rental and leasing services of other machinery, equipment and tangible goods n.e.c. 77.39.11 Rental and leasing services of railroad vehicles
274
77.39.12 Rental and leasing services of containers 77.39.13 Rental and leasing services of motorcycles, caravans and campers
77.39.14 Rental and leasing services of telecommunications equipment 77.39.19 Rental and leasing services of other machinery and equipment without operator and tangible goods n.e.c.
77.4 Licensing services for the right to use intellectual property and similar products, except copyrighted works
77.40 Licensing services for the right to use intellectual property and similar products, except copyrighted works
77.40.11 Licensing services for the right to use research and development products 77.40.12 Licensing services for the right to use trademarks and franchises 77.40.13 Licensing services for the right to use mineral exploration and evaluation 77.40.19 Licensing services for the right to use other intellectual property and similar products, except copyrighted works
19.2.2. Description
Sub-division 77.1 (renting and leasing of motor vehicles) represented 42 % of division 77 by value added in 2012 (EU-28), 89 % of which was from category 77.11 (renting and leasing of cars and light motor vehicles). The industry serves both consumers and businesses and provides short- and long-term leases. Business customers more often make use of long-time leases as a way of outsourcing care of their vehicle fleet, while private customers generally rent on a short-term basis for vacations and other one-off uses.
Prices for car rental are determined by the characteristics of the vehicle in question (size, type of drive, equipment, etc.), often according to a standard categorisation (e.g. ACRISS car classification code). There may be fees for additional equipment, insurance, mileage or services. Business leases may be individually tailored on the basis of the size of the customer’s fleet, services needed and other considerations.
Sub-division 77.2 (the renting of recreational and sports equipment and personal and household equipment) is almost exclusively oriented to consumers and the rentals are largely short-term. Typical activities are the renting of costumes, videos or sports equipment (skis, etc.). There is room for long-term leases to industries, e.g. rental of textiles. Pricing and conditions are generally standard and easy to define.
In Europe, 23 % of enterprises in NACE 77 belong to group 77.2, but the value added of this group in 2012 represented only 8 % of value added of the whole division.
Sub-division 77.3 (renting and leasing of other machinery, equipment and tangible goods) concerns almost exclusively B2B services. Lease contracts in this area are mostly long-term and often one-off. The goods rented are diverse,
275
ranging from water and air transport (without crew), to construction machinery and equipment (including scaffolds and platforms), communication equipment containers, computers and even animals.
The sub-division accounted for 44 % of EU value added in 2012. The largest sub-group is 77.39 (renting and leasing of other machinery, equipment and tangible goods n.e.c.) and second is 77.32 (renting and leasing of construction and civil engineering machinery and equipment) or, in some countries, 77.33 (renting and leasing of office machinery and equipment).
Sub-division 77.4 (leasing of intellectual property and similar products, except copyrighted works) was a new addition in NACE Rev 2. It forms only 5 % of European value added, although in some countries its share is markedly higher (e.g. in Luxembourg, where it makes up over 40 % of enterprises). The products are leased to business entities and the licensing fees or royalty payments are negotiated on a one-off basis. It includes franchising and payments for the use of trademarks and brand names, and rights for mineral exploration.
According to SBS data, EU-28 value added for division N77 was EUR 81 200 million in 2012 or 19.2 % of the section N total and 3.3 % of value added for services (NACE H+I+J+L+M+N). There are 169 000 firms engaged in the industry, 53 % of which are in sub-division 77.3, 23 % in 77.2 and 21 % in 77.1. The 646 000 employees (4.8 % of section N employment) are spread similarly across the sub-divisions.
There are differences between countries as to the composition of the industry; while groups 77.1 and 77.3 generally produce more value added than groups 77.2 and 77.4, their proportions vary greatly. On the one hand, there is Spain, with 73 % of value from 77.1 and 14 % from 77.3; on the other, there is Estonia, with 87 % from 77.3 and 8 % from 77.1. While the share of group 77.4 is usually small, in four countries (Luxembourg, Hungary, Norway and Denmark) it reaches over 10 % and in Luxembourg it accounts for over 50 %.
Table 74: NACE 77 – economic indicators (EU-28)
2011 2012 Number of enterprises 165 415 169 579 Turnover (EUR million) 156 869 162 000 Value added (EUR million) 78 168 81 200 Persons employed (thousands) 622 646
77.1 77.2 77.3 77.4 Number of enterprises 38 887 38 713 89 974 5 006 Turnover (EUR million) 66 746 14 292 73 593 7 400 Value added (EUR million) 34 601 6 459 36 019 4 100 Persons employed (thousands) 160 143 326 16
Source: Eurostat, data codes sbs_na_sca_r2, sbs_na_1a_se_r2
276
Figure 110: Division N77 – economic indicators by country
Source: Eurostat, data code sbs_na_1a_se_r2
NACE 77 generally accounts for around 15-20 % of value added in section N, but in some countries the proportion exceeded 30 % (Austria, Estonia, Ireland) or only just 13 % (Slovenia). The outlier is Luxembourg (almost 50 %). The majority of European value added for the division, as well as turnover and employees, comes from three countries with the highest values: the UK, Germany and France. A significant number of enterprises are located in Spain.
EU
28
BE
BG CZ
DK
DE
EE IE EL
ES
FR IT CY
LV LT LU HU NL
AT
PL
PT
RO SK FI
SE
UK
NO
Number of enterprises
EU
28
BE
BG CZ
DK
DE
EE IE EL
ES
FR IT CY
LV LT LU HU NL
AT
PL
PT
RO SK FI
SE
UK
NO
Value added
N771 N772 N773 N774
277
Figure 111: Division 77 – value added in relation to section N total (%)
Source: Eurostat, data code sbs_na_sca_r2
19.3. Data availability at national and European level
The STS Regulation does not currently cover NACE 77 either for turnover or for price indices. There is detailed PPI coverage of industries manufacturing the goods rented, the most relevant being NACE 28 (manufacture of machinery and equipment n.e.c.), 29 (manufacture of motor vehicles, trailers and semi-trailers) and 30 (manufacture of other transport equipment).
While there is no COICOP grouping to ensure availability of CPIs for rental services, countries may collect prices for renting some goods under groupings such as 03.1.4 (cleaning, repair and hire of clothing), 05.5.1 (major tools and equipment), 07.2.4 (other services in respect of personal transport equipment), 09.1.4 (recording media ) and 09.4 (recreational and cultural services).
The SBS Regulation requires the submission of annual turnover and value added at factor cost broken down to NACE Rev. 2 four-digit (class) level, i.e. including all sub-classes of NACE 77.
Some European countries go beyond regulatory requirements and publish quarterly or even monthly turnover for NACE 77; some publish SPPI indices for the whole division or selected sub-divisions (most often 77.11 or 77.32).
Figure 114 shows the monthly turnover index for the Czech Republic over four years. A pronounced seasonal patt ern can be observed, albeit obscured by irregular influences.
0%
20%
40%
EU
28
BE
BG CZ
DK
DE
EE IE EL
ES
FR
HR IT CY
LV LT LU HU
MT
NL
AT
PL
PT
RO SK FI
SE
UK
NO
278
Figure 112: NACE 77 – monthly turnover, Czech Republic
19.4. An experimental ISP for NACE 77
Pictured below are the results of computation of an ISP for NACE 77 for the Czech Republic, presented as monthly year-on-year indices for 2010-2012. The ISP is compared with turnover indices in current prices.
Figure 113: NACE 77 – experimental ISP (y-o-y indices)
The experimental ISP was calculated as deflated turnover data. Monthly turnover indices were available for NACE groups 77.11 (renting and leasing of cars and light motor vehicles), 77.2 (renting of recreational and sports equipment and personal and household equipment) and 77.12 (renting and leasing of trucks) together with 77.3 (renting and leasing of other machinery, equipment and tangible goods).
75
80
85
90
95
100
105
110
115
20
10
01
20
10
03
20
10
05
20
10
07
20
10
09
20
10
11
20
11
01
20
11
03
20
11
05
20
11
07
20
11
09
20
11
11
20
12
01
20
12
03
20
12
05
20
12
07
20
12
09
20
12
11
20
13
01
20
13
03
20
13
05
20
13
07
20
13
09
20
13
11
ye
ar
20
10
= 1
00
77 – gross 77 – sa
80
85
90
95
100
105
110
115
120
20
10
01
20
10
03
20
10
05
20
10
07
20
10
09
20
10
11
20
11
01
20
11
03
20
11
05
20
11
07
20
11
09
20
11
11
20
12
01
20
12
03
20
12
05
20
12
07
20
12
09
20
12
11sa
me
mo
nth
of
pre
vio
us
ye
ar
= 1
00
Turnover Price Volume
279
Table 75: Deflators used for experimental ISP for N77
NACE group Deflator
77.11 PPI for vehicle rental (NACE 77.1)
77.2 CPI — weighted mean of low-level indices
77.3 & 77.12 PPI for rental of other machinery (NACE 77.3)
The PPI for vehicle rental in fact covers only car rental (77.11), since truck rental (77.12) was considered too small for a dedicated index. The survey follows prices for the short-term and long-term rental of the most commonly leased cars (as reported by the unit) in selected categories. The PPI should represent only B2B prices; we presume that B2C prices move similarly and that the B2C market share is significantly smaller. If these assumptions did not hold, it would be better to weigh the PPI together with an appropriate CPI, if available.
We lack a specific deflator for truck rental but, since it is less significant than both 77.11 and 77.3, it is probably sufficient (if not ideal) to use deflators for similar groups. We decided that the development of machinery rental should be reasonably close to the development of truck rental, so we used the same deflator. If truck rental behaves more similarly to car rental, the relevant PPI could be used instead.
The PPI for machinery rental is composed of PPIs for renting of other machinery (77.39) and renting of construction engineering machinery (77.32), which together form a majority of NACE 77.3. For 77.32, the survey tracks several products, including crane lorries and scaffolds; group 77.39 is represented by the rental of waste containers.
No higher-level CPI was available for NACE 77.2, so two low-level indices had to be used. An index of DVD rental represents NACE 77.22 (renting and leasing of other personal and household goods), and an index of wedding dress rental represents NACE 77.29 (renting of video tapes and disks). The weights are computed from weights in the consumer basket. More representative CPIs would of course be desirable, but even the combining of two indices offers some protection against excessive fluctuations. The weight of the group is not very great (under 10 %), so the price index does not have much influence.
Each group was deflated with an appropriate deflator and the resultant indices combined with value-added weights from SBS. The resulting volume index shows difference in 2011, when turnover was depressed by lower prices, the actual volume of services staying roughly the same and declining only a year later.
280
20. EMPLOYMENT ACTIVITIES (N78)
Agnieszka Kaźmierczak, Statistics Poland
The purpose of this analysis is to present an overview of the employment activities sector (division 78) in the EU and at national level. We first present the statistics for the division and then describe national practices as regards producing an ISP.
20.1. Classification issues
The employment activities sector is covered by NACE Rev. 2 division 78 (ISIC Rev. 4 78) and consists of all services relating to the search for and selection of candidates, and all ancillary services such as skills assessments. It can be split into the following groups:
− activities of employment placement agencies (78.1)
− temporary employment agency activities (78.2)
− other human resources provision (78.3).
The activities of employment placement agencies include:
− listing employment vacancies and referring or placing applicants for employment, where the individuals referred or placed are not employees of the agencies;
− personnel search, selection referral and placement activities, including executive placement and search activities;
− activities of casting agencies and bureaux, e.g. theatrical casting agencies; and
− activities of online employment placement agencies.
In terms of product, the provision of personnel consists of all forms of personnel supply, including temporary employment. A distinction is then made between the types of personnel supplied: office staff, domestic staff, labourer or technician, paramedical staff and other types of personnel. A temporary employment company supplies on a temporary basis personnel, whom it recruits and remunerates via temporary employment agencies, under conditions stipulated by the employment code. The climate in temporary employment is closely linked to the general economic situation.
In the case of the provision of personnel through ‘help services’ (in particular for temporary work), the client pays the service company, which passes on part of the amount to the person supplied. In temporary employment, the temp agency pays the temporary worker his salary. It is important to recognise that the price to be collected is that charged to the client rather than that paid by the agency to the employee. The price may depend on factors such as geographical location, type of work, the relationship between the temporary agency and the client, the length of the contract and the skill level. In some countries, invoicing involves applying a multiplier coefficient to the temporary worker’s hourly gross salary in order to obtain a price per hour, which is then multiplied by the number of hours charged. The hourly salary is negotiable between the temporary employment company and
281
the client company, according to the temporary worker’s qualifications and the economic circumstances.
20.2. Sectoral analysis
In 2005, the EU employment activities sector accounted for 1.2 % of all services covered by Annex D to the STS Regulation (i.e. services except retail trade and repair).
As regards the recent market situation, the sector has recorded growth of 33.2 % since 2010. In 2014, the most significant rises were observed in countries such as Poland (225.3 %), Slovakia (224.5 %) and Lithuania (221.4 %)
Figure 114: N78 – development of turnover index (EU-28)
Source: Eurostat database (12.8.2015); employment activities turnover index, annual data (2010=100) (NACE Rev.2); data adjusted by working days in current prices (TOVT)
Employment activities saw a significant (18.8 %) y-o-y decrease in 2009 (Ireland: 36.6 %; Czech Republic: 34.2 %; Slovakia: 31.6 %). This did not affect all Member States, however: Cyprus, Malta, Hungary, Poland and Latvia reported growth, by 24.3 %, 22.1 %, 8.4 %, 5.7 %, and 2.6 % respectively.
81.75
96.60
108.96 111.67
90.64100.00
113.87 114.73120.92
133.23
0
20
40
60
80
100
120
140
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
European Union (28 countries)
282
Figure 115: EU-28 turnover index, annual change
Source: Eurostat database (12.11.2012); employment activities turnover index, annual data, % change compared with corresponding period of the previous year (NACE Rev. 2); data adjusted by working days in current prices (TOVT)
Nearly 70 500 enterprises were operating in the EU-28 employment activities (division 78) sector in 2012. Almost three quarters (72.0 %) of these employed no more than nine persons, while the biggest enterprises (over 250 persons employed) accounted for 2.7 %. Together, they employed 4.4 million persons, i.e. 0.1 % of the non-financial business economy71 workforce, and generated EUR 116 billion of value added, or 4.6 % of its total.
Of the three NACE groups that make up the sector in the EU, temporary employment agency activities (division 78.2) generated the most significant share (82.1 %) of value added. This sub-sector accounted for a very high (83.9 %) proportion of employment in the sector and almost half (45.3 %) of the enterprises (31 920 units, of which 33.2 % were established in the UK).
Table 76: N78 – main indicators
Main indicators (EU-28) Value Number of enterprises (1 000) 70.5 Number of persons employed (1 000) 44 402 Turnover (EUR million) 167 000 Personnel costs (EUR million) 106 000 Value added (EUR million) 116 000 Gross operating surplus (EUR million) 10 100
Shares in non-financial business economy % Number of enterprises 0.7 Number of person employed 8.1 Value added 4.6
71 Sections H-J_L_M_N_S95.
7.2
18.2
12.8
2.5
-18.8
10.313.9
0.8
5.4
10.2
-25
-20
-15
-10
-5
0
5
10
15
20
25
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014
European Union (28 countries)
283
In terms of value added, the shares of the other sub-sectors connected with employment activities were less significant. ‘Employment placement agencies’ (78.1) accounted for an estimated 8.9 % and ‘other human resources provision’ (78.3) for 7.8 %. Despite its low contribution to value added, group 78.1 accounted for over four in 10 (40.7 %) of the enterprises and 8.1 % of employment in the sector. The figures for ‘other human resources provision’ were 14.1 % and 8.1 % respectively.
Three countries generated over two thirds (67.8 %) of the EU’s total value added from the sector in 2012: the UK (27.3 %), Germany (20.4 %) and France (20.2 %).
Most of the enterprises were established in the UK (27.4 %). The second highest proportion (15.8 %) was recorded for Netherlands, which contributed 7.1 % of the EU sector’s value added.
The largest workforce in employment activities was registered in France (828 900 persons), ahead of Germany (774 000) and the UK (663 600).
20.3. Available methodological information
20.3.1. Voorburg Group
ISP issues have been presented at Voorburg Group meetings in the following papers:
• OECD activity on the short-term indicators for services (Eun-Pyo Hong, OECD, 2003 — Tokyo).
This paper summarised the OECD’s activity in the field of short-term indicators for the services sector via the activities of the short-term indicators for services task force (STISTF). Until June 2003, the task force’s work was focused on monthly and quarterly ISPs for the entire services sector, on the basis of the common practice of OECD member countries. The following similarities were observed from information collected from Canada, the UK and Korea:
− publication of monthly results within two months after the reference month;
− surveys complemented by administrative information were the main source of data; and
− the scope of activities covered was almost the same.
The analysis also identified differences as regards:
− the compilation methodologies applied;
− the inclusion of sub-sectors or activities within an activity; and
− the type of variables and deflators used to convert value data to volume.
• Prototype compilation manual for a monthly ISP (STESEG Task Force on Services, 2004 — Ottawa); and
284
• Presentation on short-term indicators in the service sector (Jonas Färnstrand, Statistics Sweden, 2008 — Aguascalientes).
The author presents a short abstract of the methodology for compiling an ISP in Sweden (theoretical form based on the Laspeyres-type volume index, practical index calculation based on deflated turnover for each activity weighted by SBS data updated yearly).
All the above address ISPs in general, rather than focusing on individual kinds of activity.
Employment activities were touched on repeatedly at the Voorburg Group meeting on SPPIs:
− Mini-presentation on producer price indices – development of a price index for employment services in the UK (Nick Palmer and Steve Lewis, ONS — 2002, Nantes);
− US producer price index for help supply services (Ryan Broderick, US Bureau of Labor Statistics — 2002, Nantes);
− Mini-presentations on producer price indices – development of a UK price index for recruitment services (Anthony Luke and Pam Davies, ONS — 2004, Ottawa);
− The French PPI for provision of temporary workers and labour recruitment (Benoît Buisson, INSEE, France – 2004, Ottawa);
− Producer price indexes for recruitment and the provision of temporary workers within Australia (Jo-Marie Moore and Daryl Williams, Australian Bureau of Statistics – 2004, Ottawa);
− Labor recruitment in Japan’s corporate service price index (CSPI) (Kiyohito Utsunomiya and Nao Sudo, Research and Statistics Department, Bank of Japan — 2005, Helsinki);
− Recruitment in the UK’s corporate services price index (CSPI) (Tim Clode — 2005, Helsinki); and
− Principal paper: The French SPPI on recruitment and provision of temporary workers (Benoît Buisson, INSEE, France) – 2005, Helsinki).
20.3.2. OECD ISP Manual
In 2002, the OECD’s Short-term Economic Statistics Expert Group (STESEG) and the Short-term Indicators for Services Task Force (STISTF) were set up to cover supply and demand indicators for short-term service activities, with the monthly ISP as a priority. The STISTF’s end-product was the ISP Manual, which provides theoretical and practical recommendations for compiling the ISP. However, it should be noted that the Manual is based on ISIC Rev. 3.1, which classifies employment activities under code 74.91 (labour recruitment and provision of personnel).
According to the ISP Manual, the best variable for measuring the output of employment activities is gross turnover deflated by an appropriate
285
quality-adjusted price index. Turnover data should be obtained ‘from a survey of enterprises supplying these types of services (employment services) or from administrative data taken from tax returns which could be split by type of other business service supplied’.
In the case of employment activities, the PPI for employment agencies is specified as the best deflator. Turnover data are defined as receipts from sales excluding VAT and other taxes on products, plus any subsidies on products. Where long-term projects are involved, turnover should be adjusted for number of hours worked but not billed for. Partially representative72 price indices or volume indicators, e.g. number of persons employed, are presented as alternative deflators.
20.3.3. National practices
The UK ’s index of production activity for the service sector (IoS) covers almost all the categories of service featuring employment activities. The main source of information for employment services is the monthly turnover survey, which provides information on two types of variable: gross turnover (value) and number of employees. The UK uses the three- to five-digit levels of variables – four digits in the case of employment services, for which it uses an SPPI as a deflator. A final aggregate ‘administrative and support service activities’ index is published two months after the reference period.
Sweden releases its service production index for employment activities 35 days after the reference month as an aggregate for ‘administrative and support service activities’. Data are derived from a monthly turnover survey and cover NACE Rev. 2 divisions 45-96 (excluding 64-66). As a deflator for employment activities, Swedish statistics use a quarterly SPPI. Value added weights are based on the SBS survey and updated annually.
Canadian estimates of monthly value added are obtained from deflated SA data ‘weighted by benchmarking them to the output values of the deflated annual input output tables using, in general, the proportional Denton method as modified by Cholette’. The ‘real value of service industries’ series in Canadian GDP covers NAICS 41 to 91, which include sub-sector 56.13 (employment services).
20.3.4. Further references and data sources
Statistical Classification of Economic Activities in the European Community, Rev. 2 (2008) (NACE Rev. 2)
Methodological guide for developing producer price indices for services, 2005
Revisited sector paper on ISIC Rev. 4 78 (employment activities), Agnieszka Matulska-Bachura; 24th meeting of the Voorburg Group (Oslo, Norway; 14-18 September 2009)
For UK methodological issues, see: http://www.ons.gov.uk/ons/guide-method/user-guidance/index-of-services/introduction/index.html
72 i.e. partially representative for employment services (PPI, CPI or general price index).
286
The development of the index of services in the UK (presentation), Steve Drew, (ONS), 1st meeting of the ISP task force (14-15 June 2012)
Overview of activity of the Short-term Indicators for Services Task Force, Eun-Pyo Hong, OECD (May 2003)
Review of indices of service production for OECD member countries, OECD STESEG, Benoit Arnaud, Anne Hery, Eun-Pyo Hong, Frederic Parrot (Paris, June 2003)
Index of service production (presentation), Pernilla Bengtsson, Statistics Sweden, 1st meeting of ISP task force (14-15 June 2012)
SBS — services: annual detailed enterprise statistics for services (NACE Rev. 2 H-N and S95) (sbs_na_1a_se_r2), 17.12.2012
STS — services: other services turnover index — annual data (2005=100) (NACE Rev. 2), 12.11.2012
STS — services: other services turnover index — annual data (% change compared with corresponding period of previous year) (NACE Rev. 2), 12.11.2012
287
21. TRAVEL AGENCY , TOUR OPERATOR RESERVATION SERVICE AND RELATED ACTIVITIES
(N79)
Søren Kühl Andersen, Statistics Denmark
This section describes NACE Rev. 2 division N79 (travel agency, tour operators reservation service and related activities). After a brief description of the content of the division, we present available methodological information and data. The main part of the section is devoted to discussing how an index of service volume could be produced for this division.
21.1. Classification issues and content of division N79
Division N79 consist of only three classes divided into two groups:
79 Travel agency, tour operator and other reservation service and related activities 79.1 Travel agency and tour operator activities 79.1.1 Travel agency activities 79.1.2 Tour operator activities 79.9 Other reservation service and related activities 79.9.0 Other reservation service and related activities
The first group is by far the dominant one. Its two classes are not easily distinguishable, despite the fact that there are, by definition, big differences: between travel agencies (primarily engaged in selling travel, tour, transportation and accommodation services on a wholesale or retail basis) and tour operators (which arrange and assemble tours that are sold through travel agencies or directly). The tours may include transportation, accommodation, food and visits to a variety of cultural sites or events.
The big difference is obviously that a travel agency functions mainly as an agent, linking various travel businesses to consumers and other businesses, whereas a tour operator also arranges combinations of various travel services.
The third class in division N79, other reservation service and related activities, which is by far the smallest, covers various other related activities, of which the most important are probably ticket sales activities for theatrical, sports and other amusement and entertainment events, and visitor assistance services.
21.2. Available methodological information
21.2.1. OECD ISP Manual
The material of most direct relevance to the ISP for N79 is the OECD Manual. While the main body of it deals with general, rather than sector-specific, methodological issues, Annex 2 does contain a very useful overview of various possibilities for calculating the ISP within each sector.
In the case of N79, the preferred method mentioned is, as in most other cases, to deflate gross turnover with appropriate price indices. Mentioned here are the CPI of package holidays and the SPPI of travel agents and tour operators.
288
21.2.2. Voorburg Group
The Voorburg Group has also dealt with the activities of travel agents and tour operators and there are several methodological papers available on its website. Many of these deal primarily with price measurements, but some also cover issues relating to output/turnover. Several papers give valuable insights into the anatomy and history of division N79, which are very useful for constructing an ISP.
21.3. Available data
21.3.1. Services turnover and SPPIs in STS
N79 is currently included in services turnover in STS. The published figures are indices and most countries deliver only quarterly data to Eurostat, although some publish and transmit monthly data. Data are only available at two-digit level, so there is only one index series for the entire division.
N79 is currently not included in the producer prices in STS and no SPPIs for N79 are available in the Eurostat database.73
21.3.2. Consumer price indices (CPIs)
The CPI is suggested as an alternative deflator for package holidays. This is available for as a monthly index for all EU countries.
21.3.3. Structural business statistics (SBS)
The SBS obviously contain more detailed and comprehensive data for division N79, such as value added, turnover, number of enterprises and number of employed persons.
21.4. Description of N79
Table 78 draws on Eurostat’s database to show number of enterprises, turnover and value added for division N79 in 2011 (the latest year for which most of the data are not estimated, missing or confidential).
73 Neither does the OECD SPPI guide provide methodological information on division N79.
289
Table 77: Structure of N79 (travel agencies, tour operator and other reservation services and related activities), EU-28, 2011
Number of enterprises
Turnover (€ million)
Value added at
factor cost (€ million)
N79 (travel agency, tour operator and other reservation service and related activities)
91 898 151 114 24 430
N79.11 (travel agency activities) 47 647 78 035 14 591 N79.12 (tour operator activities) 20 000 65 717 7 669 N79.90 (other reservation service and related activities)
confidential 7 332 2 141
The division is dominated by travel agencies and tour operators. Measured by turnover, the two are close in size, but unsurprisingly the former have on average much smaller turnover. It is somewhat surprising, on the other hand, that they generate higher value added per turnover. In general, one would expect value added in a pure agent activity to be relatively small as compared with the business of arranging and assembling travel services.
21.4.1. Difference between agencies and operators
Most people would find it difficult to distinguish between tour agencies and tour operators. Although the definitions are quite clear and the activities of the two are actually very different, it is probably wise not to rely too much on the split between the two classes.
A look at the Danish business register reveals that, in Denmark at least, there are classification issues within group N79.1. Many of the larger enterprises classified as N97.11 are obviously not mainly agents, but arrange tours and package deals themselves and should therefore be classified in N97.12.
21.5. Calculating an ISP for N79
Calculating an ISP for N79 is not very different from calculating an ISP for most other service industries, but it is worth mentioning some general and more sector-specific issues.
21.5.1. Methodology
In theory, an ISP should show short-term movements of gross value added, which can be calculated as gross output minus intermediate consumption. In most cases, it is close to impossible to do this calculation on a monthly (or even quarterly) basis, so other methods have to be used.
The standard approach is using the development of (deflated) turnover as a proxy for the development of gross value added. Though other methods (using a volume output measure or even an input measure) are possible, the turnover measure is generally considered most suitable for an ISP.
290
The method assumes that the ratio between turnover and input is relatively stable. It also assumes that an appropriate price index is available for deflation.
21.5.2. Business models in N79
The activities of travel agencies have changed a great deal over the past 15 years. The industry used to be dominated by enterprises whose primary source of revenue was commissions on the sale of airline tickets. The service they provided was finding and booking airline tickets that satisfied the customers’ needs. The commission was often paid by the airlines, but in some cases charged directly to the customer.
More recently, most airlines have reduced or discontinued commission payments, as they are now able to sell tickets directly over the internet. Some travel agencies now charge customers, mainly business travellers, service fees for arranging and booking flights. Some also charge commissions or booking fees on selling package holidays or cruises. Travel agencies also include internet pages offering price comparison and booking services to many airlines.
The activities of tour operators have not undergone changes of the same magnitude. Their main activity is to assemble and arrange package holidays and custom-designed trips, but obviously the ability to sell them over the internet is relatively new in this industry as well. The need for travel agents to sell the packages assembled by tour operators has probably diminished as a result.
The activities of the remainder group in N79, ‘other reservation service and related activities’, are quite diverse. In many countries, the dominant enterprises in this class will probably be engaged in selling tickets for theatrical, sports and other amusement and entertainment events. Of course, over the past 15-20 years this has gone from primarily taking place at physical ticket shops or by phone to being mainly an internet activity.
An important point when calculating an ISP for N79 is that the turnover relating to the service being provided is much smaller than the total price being paid by the customers. A customer will usually pay the full price for an airline ticket, a package holiday or a concert ticket when completing the transaction, but the main part of the price usually has nothing to do with the service provided by the travel agent/tour operator/ticket agent. It is turnover belonging to airlines, hotels, bus companies, concert organisers, etc. Whether the full price is actually registered as turnover at the N79 enterprise depends on the business model. In most cases, it will probably never count as turnover at a travel agency, whereas tour operators might actually ‘own’ all the components in a package holiday and then the full price will count as turnover.
21.5.3. Deflators for N79
The question of business models is very relevant when compiling an ISP for N79 by the deflated turnover method, since it is important to know the composition of turnover in order to be able to deflate it appropriately.
291
In time, SPPIs will be available, including for N79. The price codes relating to N79, according to CPA 2008 (five-digit level) are as follows:
Table 78: CPA 2008 for N79
79 Travel agency, tour operator and other reservation services and related services 79.11.1 Travel agency services for transport reservation
79.11.2 Travel agency services for reservation accommodation, cruises and package tours 79.12.1 Tour operator services 79.90.1 Tourism promotion and visitor information services 79.90.2 Tourist guide services 79.90.3 Other reservation services n.e.c.
At six-digit level, there are many more prices, especially for travel agency services.
An SPPI based on the above prices will of course be very suitable for deflating the turnover relating to the actual services performed by travel agencies, tour operators, etc., but the turnover might in some cases include much more (see above). There are certainly tour operators whose turnover includes the entire price on the packages sold and the result of deflating this with an SPPI could be misleading.
In a sense, the situation is similar to that of the retail and wholesale trade sectors. A retail trade enterprise usually owns the goods it is selling, so the turnover will be much higher than the value of the service provided. Deflating such a turnover with an SPPI for the retail trade sector (if such an SPPI was available) would not make much sense, since there is actually no reason to think that the price on retail trade services (e.g. the selection and presentation of a certain consistent assortment of products, sales advice, etc.) is closely aligned with the price on the goods sold. In such cases, the turnover is normally deflated with an appropriate CPI, resulting in a volume of sales index. Calculating an actual ISP for trade services would be close to impossible, as the services are almost always included in the price of the goods.
The situation is not exactly the same in N79, but for certain tour operators it is certainly similar. Some of the larger tour operators might actually purchase the components in a package holiday before they sell it. They ‘own’ the accommodation and the seat on the plane and when they sell the package holiday to a customer, the price includes these services as well as the service provided by the tour operator. Only the latter is relevant for the ISP of N79. In such a case, to deflate the entire turnover with an SPPI for N79 would not necessarily be very accurate.
Ideally, the turnover relating to the service actually provided by a tour operator should be identified. However, this would either impose a very large response burden on the enterprise or depend on some knowledge of intermediate consumption in such an enterprise. Either way it seems a hard task and in general it would probably be advisable to accept a less-than-ideal method.
292
If it is accepted that what is available for deflation is total turnover, consideration should be given (at least for tour operators) to deflating with the CPI for package holidays. Of course, this will not result in an accurate ISP, but as the service provided by the tour operators might constitute a smaller part of the total turnover than the net worth of the other services included in the price (accommodation, transport, etc.), it will at least give a better result than deflating the entire turnover with the SPPI.
Other options are using other output measures or even input measures. The output might be measured by accommodation or transport data, but it will have the same disadvantages as the turnover approach and to an even larger degree: it will not reflect the development of the services performed by tour operators. Input measures are generally to be avoided and in this case the deflated turnover method, though not perfect, is probably better.
In addition to the above issues, it is important to consider the level of deflation. If misclassifications seem to be a concern, it might be best to compile one deflator at least for N791. If, on the other hand, one can be confident that travel agencies and tour operators are classified correctly, it is worth considering deflating N79.11 and N79.12 separately, possibly using the CPI for package holidays in the deflator for N79.12, as described above.
21.5.4. Test calculations
There are currently no SPPIs available for N79, so the best option for calculating an ISP for N79 is probably to deflate turnover with the CPI for package holidays, even though this probably does not cover much of the turnover in N79.11 and N79.9.
The STS turnover for N79 is currently available for all EU countries on a quarterly basis only. Some countries also submit monthly turnover.
Figure 118 shows the turnover index for N79 (adjusted for working days) for the EU-28 aggregate, the HICP for package holidays (calculated from monthly values) and the result of using the latter as a deflator.
293
Figure 116: Quarterly ISP for division N79 (EU-28)
The turnover has a very distinct and stable seasonal pattern. The activity is rather low in the fourth and first quarters, increases by 20-30 % in the second and again by 20-30 % more in the third. The price index has a similar pattern, but with much smaller fluctuations, and the result is that the ISP has a seasonal pattern much like that for turnover. This is not surprising, as it is well known that tour operators at least have a high peak in the summer season.
A seasonal adjustment of the ISP (default settings with X12-Arima) naturally results in a much smoother index curve.
Figure 117: Quarterly ISP for division N79 (EU 28), SA
40
60
80
100
120
140
160
180
20
00
Q1
20
00
Q3
20
01
Q1
20
01
Q3
20
02
Q1
20
02
Q3
20
03
Q1
20
03
Q3
20
04
Q1
20
04
Q3
20
05
Q1
20
05
Q3
20
06
Q1
20
06
Q3
20
07
Q1
20
07
Q3
20
08
Q1
20
08
Q3
20
09
Q1
20
09
Q3
20
10
Q1
20
10
Q3
20
11
Q1
20
11
Q3
20
12
Q1
20
12
Q3
20
13
Q1
20
13
Q3
20
14
Q1
20
14
Q3
20
15
Q1
Turnover index HICP, package holidays ISP
60
65
70
75
80
85
90
95
100
105
110
20
00
Q1
20
00
Q3
20
01
Q1
20
01
Q3
20
02
Q1
20
02
Q3
20
03
Q1
20
03
Q3
20
04
Q1
20
04
Q3
20
05
Q1
20
05
Q3
20
06
Q1
20
06
Q3
20
07
Q1
20
07
Q3
20
08
Q1
20
08
Q3
20
09
Q1
20
09
Q3
20
10
Q1
20
10
Q3
20
11
Q1
20
11
Q3
20
12
Q1
20
12
Q3
20
13
Q1
20
13
Q3
20
14
Q1
20
14
Q3
20
15
Q1
ISP for N79, EU-28, sa
294
21.6. Conclusions
Division N79 is structurally simple, in the sense that it consists of only three classes. There could be classification issues, in particular as tour operators are sometimes misclassified as travel agents. This is definitely the case in Denmark.
The sector has seen big changes in recent years, where a lot of the sales activity has moved from physical offices to online sales and the activity of airline bookings by travel agents has decreased rapidly.
When compiling an ISP for N79, it is important to be aware of the business models in use. In particular, some tour operators will have a turnover that includes the full price of package holidays, including travel expenses and accommodation. These parts of the turnover do not actually have anything to do with the service production that an ISP should cover. This is particularly important when calculating an ISP by deflating turnover with an appropriate deflator, which after all seems to be the best option.
295
22. SECURITY AND INVESTIGATION ACTIVITIES (N80)
Michaela Lingler, Statistics Austria
22.1. Summary
More comprehensive and detailed information about service sectors is a basic requirement in the future development of European business statistics. The variable ISP was added to the list of indicators in the FRIBS package for STS. Eurostat established a task force to take stock of available methodological information, investigate current data gaps and run test calculations for an ISP.
In general, the following statistical data sets are currently (August 2015) available (in accordance with the STS-Regulation and other Regulations):
− quarterly turnover indicators at current prices from STS (t + 2 months);
− turnover, employment, etc. from SBS at NACE four-digit level (t + 18 months);
− intermediate and final consumption from NA — for the aggregate NACE N80 to N82 only (t + 36 months);
− quarterly SPPIs for NACE division N80 (t + 3 months)
o PRON = total output price index
o PRBB = service producer prices for businesses; and
− monthly HICP data with base year 2005 (available 45 days later)
o no detailed information for security and investigation activities
o general HICP has to be taken, if necessary.
For this section, we first collected data from European databases. We then contacted entities in Statistics Austria responsible for data input (NAs, SPPIs and CPIs) to find out more about the structure and availability of the data. For the industry section, background information from the Voorburg Group was also available.
Two possible methods were tested (the choice of method will depend on data availability in the country in question):
• in general, SPPIs will be the preferred deflation method for security and investigation services if they are available at B2All level, because they cover all customers and should be weighted according to the specific services;
• countries for which a B2B SPPI is available have to take the appropriate CPI, HICP or the general HICP as proxy for the B2C part of turnover. This approach will require detailed information on the relative weighting of the B2B and the B2C parts. It is recommended that this be done on the basis of NA use tables, if available. If only data for N80-N82 are available, it may be possible to use the B2B SPPI as a B2All SPPI.
296
An approximate ISP can be calculated with information currently available. Nevertheless, continuous calculations will require a monthly ISP at t+60 calculated with quarterly SPPIs available at t+90. The development of harmonised methods for the calculation of missing information will be a big challenge for future methodological work.
22.2. Introduction
Eurostat launched the FRIBS project in 2011 in order to establish a common legal framework for the collection, compilation and dissemination of European business statistics. The main issues to be addressed are the harmonisation, consistency and streamlining of the statistics. The project is made up of various modules, one of which concerns STS (currently based on the STS Regulation). Detailed requirements for STS have to be set out, in particular to meet user needs for more comprehensive and detailed information on the service sectors. As the variable ISP was added to the list of indicators in the FRIBS STS package, Eurostat set up a task force to take stock of available methodological information, investigate current data gaps and run test calculations for an ISP.
22.3. Methodological information
Methodological information for security and investigation activities was derived from the Voorburg Group, information from methodological handbooks and statistics.
In Austria, these activities are partly regulated. Those wishing to establish a business in the industry have to meet certain requirements, but there are no supervisory authorities. Enterprises are members of the Austrian Economic Chambers (WKÖ).
According to the Voorburg Group revisited sector paper on NACE/ISIC 80,74 countries use various systems to calculate turnover, partly involving administrative data and partly surveys. A further point to consider is that security and investigation companies sometimes offer other services, e.g. cleaning services. In this paper, we make a number of suggestions for improving data quality by checking administrative data.
22.4. Presentation of the industry
22.4.1. Industry overview and trends
The EU-28 security and investigation activities industry is a quite small in terms of SBS coverage, with approximately 54 000 enterprises generating turnover of around EUR 46 billion in 2012. To put this in context, there are about 1.26 million enterprises in NACE section N (administrative and support service activities), with turnover of EUR 880 billion. In NAs, division N80 is combined with N81 (services to buildings and landscape activities) and N82 (office administrative, office support and other business support activities). In this group,
74 Revisited sector paper on NACE/ISIC 80 (security and investigation activities), 24th Voorburg Group
Meeting (Oslo, Norway; September 2009).
297
over 80 % of consumption is B2B in most countries. SPPIs are available for division N80, but due to the small turnover this service is generally not provided by HICPs. The test calculations faced two challenges:
− NA data are only available for N80 to N82; and
− the HICP contains no positions for security services, as they account for such a small proportion of the total.
22.4.2. Classification issues — classification by activity
NACE Rev. 2 has been used from reference period 2008 or 2009 for business statistics. Statistics produced on the basis of NACE are comparable at European and, in general, at worldwide level (where ISIC Rev. 4 is currently used). The two systems are identical for division N80. At European level, consistency of classification is guaranteed down to four-digit level. National versions of NACE may have a more detailed breakdown of economic activities.
NACE is revised in response to changes in economic structures and organisations, technological developments, new activities and products. NACE Rev. 2 has a new section N (administrative and support service activities) which distinguishes different activities such as real estate activities and computer and related activities.
The main components of section N are rental and leasing activities (division 77), employment activities (N78), travel agency, tour operator and other reservation service and related activities (N79), security and investigation activities (N80), services to buildings and landscape activities (N81) and office administrative, office support and other business support activities (N82).
The NACE Rev. 2 classification and explanatory notes for division 80 are as follows:
N80 Security and investigation activities
This division includes security-related services, such as investigation and detective services, guard and patrol services, picking up and delivering money, receipts or other valuable items with personnel and equipment to protect such property while in transit, the operation of electronic security alarm systems, such as burglar and fire alarms, where the activity focuses on remote monitoring of such systems, but often involves also sale, installation and repair services. If such components are provided separately, they are excluded from this division and classified in retail sale, construction, etc.
N80.1 Private security activities
This class includes the provision of one or more of the following: guard and patrol services, picking up and delivering money, receipts, or other valuable items with personnel and equipment to protect such property while in transit. It includes armoured car services, bodyguard services, polygraph services, fingerprinting services, security guard services and security
298
shredding of information on any media. It excludes public order and safety activities (see 84.24)
N80.2 Security systems service activities
This class includes the monitoring or remote monitoring of electronic security alarm systems, such as burglar and fire alarms, including their installation and maintenance. It also includes installing, repairing, rebuilding and adjusting mechanical or electronic locking devices, safes and security vaults in connection with later monitoring and remote monitoring.
The units carrying out these activities may also engage in selling such security systems, mechanical or electronic locking devices, safes and security vaults.
This class excludes the installation of security systems, such as burglar and fire alarms, without later monitoring (see 43.21), the retail sale of electrical security alarm systems, mechanical or electronic locking devices, safes and security vaults in specialised stores, without monitoring, installation or maintenance services (see 47.59), security consultants (see 74.90), public order and safety activities (see 84.24) and providing key duplication services (see 95.29).
N80.3 Investigation activities
This class includes investigation and detective service activities and the activities of all private investigators, independent of the type of client or purpose of investigation.
22.4.3. Classification issues — classification by product
As a product classification, the CPA is designed to categorise products (goods and services) that have common characteristics. It is part of an integrated system of statistical classifications, developed mainly under the auspices of the UN Statistical Division. CPA is the European version of the CPC, the UN’s central product classification.
Every product in the CPA (be it a transportable or a non-transportable good or a service) is assigned to a single NACE Rev. 2 activity, so its structure reflects that of NACE Rev. 2. However, a detailed linkage between products and activities could only be established to a certain degree and in some cases products can be assigned to activities only at a level above class level or a class in CPA 2008 has no activity counterpart in NACE Rev. 2.
The Regulation establishing CPA 2008 was adopted in April 2008. New concepts, such as originals and intellectual property products, were introduced and further detail added to reflect new products and emerging services. National versions of the CPA exist, just as there are national versions of NACE Rev. 2.
Like NACE Rev. 2, CPA 2008 has a new section N (administrative and support service activities) combining services such as rental and leasing activities (N77),
299
employment activities (N78), travel agency, tour operator and other reservation service and related activities (N79), security and investigation activities (N80), services to buildings and landscape activities (N81) and office administrative, office support and other business support activities (N82).
Table 79: CPA 2008 for division N80 (security and investigation services)
80 Security and investigation services 80.1 Private security services 80.10 Private security services 80.10.1 Private security services
80.10.11 Armoured car services (This sub-category includes services consisting of providing an armoured car to pick up and deliver money, receipts or other valuable items with hired personnel to directly protect property while in transit, i.e. bank collection and deposit services and securities transfers services.) 80.10.12 Guard services (This sub-category includes services consisting of providing protective services through hired personnel to ensure the safety of people or private, industrial and commercial properties by guarding them against fire, theft, vandalism or illegal entry, i.e. security patrol services, security guard services, bodyguard services, watchdog services, parking control services, access control services. This sub-category excludes: public order and safety services.) 80.10.19 Other security services (This sub-category includes training of guard dogs, polygraph services, fingerprinting services, security shredding of information on any media.)
80.2 Security systems services 80.20 Security systems services 80.20.1 Security systems services
80.20.10 Security systems services (This sub-category includes, services consisting of monitoring and maintaining security systems devices, such as burglar and fire alarms, by receiving alarm signals, confirming or checking that all systems are properly functioning, and dispatching police officers, fire department or other designated parties This sub-category also includes remote monitoring services. This sub-category excludes installation services of security systems, such as burglar and fire alarms, without later monitoring, selling security systems, mechanical or electronic locking devices, safes and security vaults, without monitoring, installation or maintenance services, security consulting as a stand-alone service, providing key duplication services.)
80.3 Investigation services 80.30 Investigation services
300
80.30.1 Investigation services 80.30.10 Investigation services (This sub-category includes investigation and detective services such as services consisting of investigating cases submitted by the client, relating to crimes, theft, fraud, shoplifting, dishonesty, missing persons, domestic relations and other lawful or unlawful practices, internal and undercover investigation, shoplifting protection services. This sub-category excludes credit investigation services.)
Harmonised classifications are a basic requirement for comparable statistical data. The most important distinction is between private security activities and other activities. As enterprises are classified according to their main activity, using data according to a breakdown by activity means that turnover data for a certain industry may include turnover shares from other industries in NACE 80 — but also from other NACE categories, e.g. retail trade. Turnover by products (CPA) would provide more detailed information on shares of specific security and investigation services, but there is no legal requirement at European level to provide a breakdown of turnover by CPA for security and investigation services. This is important with regard to the interpretation of the turnover data used for weights.
22.5. Available data
22.5.1. Short-term statistics (STS)
STS describe the short-term development of the economy. The STS Regulation entered into force in 1998, with a transition period of five years. The reference year (2003) is the first year for which the full set of basic STS indicators required in the Regulation must be available. The availability of STS data depends on the industry in question. Security and investigation activities (NACE Rev. 2 division 80) are regulated in Annex D (Other Services) to the Regulation, under which:
− the ‘turnover’ and ‘number of persons employed’ variables have to be provided on a quarterly basis for selected service industries (t+2 months);
− since 2013, the list of variables has included ‘wages and salaries’ and ‘hours worked’ for NACE Rev. 2 section N, starting with 2010 (t+3 months);
− variables have to be transmitted either as an index or in absolute figures; working day adjustment is required for ‘turnover’; seasonal adjustment is carried out by Eurostat;
− data for division 80 have to be provided at NACE Rev. 2 two-digit level;
− since 2006, SPPIs have had to be produced on a quarterly basis for selected activities of the coverage of the turnover variable; and
− for security and investigation activities, SPPIs have been obligatory since 2006.
301
Transition periods in the STS Regulation, problems with implementing STS requirements and the 1 % rule for small Member States mean that there may be data gaps.
Time series with STS turnover indicators are fundamental for producing an ISP for security and investigation services. The Eurostat database contains quarterly gross turnover indicators for all Member States except Denmark, Italy and Slovakia. Time series are also almost complete and available for WDA and SA turnover. In the test calculations, we used turnover indicators up to the reference period 1st quarter 2015 (extracted from the database on 11 August 2015) with base year 2010.
SPPIs will be one of the main data sources for producing an ISP for security and investigation activities and the complete availability of SPPIs for this industry is very important. An extraction from the Eurostat database shows that SPPIs have more data gaps than turnover indicators. SPPIs for NACE Rev. 2 division 80 are available for only 13 Member States for 2006 and for 19 countries (PRON) for 2007. These data gaps were a problem for test calculations for individual countries. A B2All SPPI (PRON) is available for 21 Member States and a B2B SPPI (PRBB) for 14 Member States for base year 2010.
In Austria, SPPIs for security and investigation activities are calculated on the basis of a combination of product transaction price and model pricing. In principle, these are B2B SPPIs.
22.5.2. Harmonised consumer price indices (HICPs)
HICPs give comparable measures of inflation in individual countries. Preliminary HICPs for all Member States are available about 16 days, and final data about 45 days, after the end of the reference month. Data are released on a monthly basis. The HICPs are classified according to COICOP four-digit categories and sub-categories. For the test calculations, we used the general HICP after consulting the team for CPIs. There is no more detailed HICP available, because the share of security and investigation activities is so small that there is no differentiation for the HICP.75
22.5.3. Structural business statistics (SBS)
SBS describe the structure, conduct and performance of businesses across the EU and cover industry, construction, trade and services. The main SBS indicators are enterprises, employment, turnover, value added and investment (for the EU-28 and for Member States, broken down by economic activity and size class). SBS are based mainly on data for enterprises classified according to NACE. NACE Rev. 2 applies from reference year 2008; NACE Rev. 1 and Rev. 1.1 apply for previous years.
SBS data are available on an annual basis, with a time lag of 18 months after the end of the reference year for individual Member States and about 21 months for EU aggregates. As part of the service sectors covered by the SBS Regulation,
75 Data are available from the Eurostat database: http://ec.europa.eu/eurostat/web/hicp/data.
302
security and investigation activities are included in SBS data. Nevertheless, changes in the NACE classification and recasting of the Regulation had an impact on the availability of time series. SBS data for security and investigation activities are available from 1995 (except for 1996, in the case of Austria).
The breakdown for reference years 1995–2007 is at three-digit level of NACE Rev. 1 or Rev. 1.1 (K74.6 Investigation and security activities). From reference year 2008, there is a more detailed breakdown of activities at NACE Rev. 2 four-digit level. An activity breakdown is therefore available for N80.1 (private security activities), N80.2 (security systems service activities) and N80.3 (investigation activities). For reference year 2008, data for both NACE Rev. 1.1 and Rev. 2 are available, due to double reporting. Backcasted data are not generally available for SBS, but some Member States have provided them on a voluntary basis back to 2005. Double reporting for 2008 will allow a macro backcasting approach.
Transition periods in the SBS Regulation, problems with implementing SBS requirements and the 1 % rule for small Member States mean that there may be data gaps. Pre-2008 data according to NACE Rev. 2 are available for selected countries only. SBS provides very detailed data in this area and will be important for weighting in the course of producing SPPIs and the ISP.
The CPA is not a fundamental classification for SBS, but for some areas (trade, insurance, credit institutions and business services) results based on a CPA product breakdown are compiled and disseminated in line with the SBS Regulation. Unfortunately, no CPA breakdown is required for security and investigation activities.
In 2012, there were around 54 000 enterprises operating in the field of security and investigation services in the EU-28, generating turnover of around EUR 46 billion.
Table 80: Division N80 – number of enterprises, turnover and value added at factor costs for security and investigation activities (2011-2013)
EU-28 2011 2012 2013 Number of enterprises 53 827 53 844 . Turnover (EUR million) 45 000.0(d)(e) 45 871.3 44 364.6(p) Value added at factor costs (EUR million)
30 673.5 31 669.3 .
Numbers of persons employed 14 026 14 105 14 212(p)
Source: Eurostat, SBS; (d) definition differs, see metadata; (e) estimated; (p) provisional
For compiling an ISP, it is worthwhile taking account of the structure within security and investigation activities sectors, so we extracted data on turnover distribution across sub-categories (groups 80.1–80.3) from the Eurostat database. Table 82 shows the percentage distribution for individual countries and the EU-28 for reference year 2013.
303
Table 81: Division N80 – turnover distribution by group, 2013 (rounded % values, preliminary)76
Country / Group 80.1 80.2 80.3 EU-28 83 16 2 Belgium 78 22 0 Bulgaria 82 18 1 Czech Republic 87 11 2 Denmark 47 50 3 Germany 90 9 1 Greece 92 4 3 Spain 91 9 1 France 79 20 1 Croatia 88 12 0 Italy 92 2 6 Cyprus 97 0 3 Latvia 44 56 0 Lithuania 64 35 1 Hungary 78 20 3 Austria 93 2 5 Poland 59 38 2 Portugal 87 13 0 Romania 94 5 1 Slovenia 92 7 1 Slovakia 85 14 1 Finland 82 18 0 Sweden 59 41 0 United Kingdom 88 11 1 Norway 64 36 0 Switzerland 92 8 0
Source: Eurostat, SBS
In 2013, enterprises whose main activity was in private security activities generated approximately 83 % of the turnover. In the EU-27, security systems service activities accounted for around 16 %. These two groups were responsible for a considerable proportion of NACE 80 turnover in most of the countries. While the weight of sub-sector 80.3 (investigation activities) is rather negligible (2 %), it varies across countries and is much greater in some (e.g. Italy: 6 %; Austria: 5 %), indicating considerable structural differences between countries.
76 The table shows only countries for which the full matrix of data is available without confidential cells.
304
Figure 118: Division N80 – turnover distribution by group, 2011–2013, EU-28 (%)
Source: Eurostat, SBS
Figure 120 shows that turnover distribution in the EU-28 is relatively constant over the 2011–2013 period. In general, SBS data will provide useful information for weighting security and investigation services, but unfortunately no turnover breakdown by products (CPA) is available from SBS.
22.5.4. National accounts
NA supply and use tables were assessed for usability for weighting. Information is available only for NACE N80-N82, i.e. ‘security and investigation services’, ‘services to buildings and landscape’ and ‘office administrative, office support and other business support services’. In 2013, N80 accounted for approximately 11 % of N80-N82 turnover in the EU-28, ranging between 6 % in Italy and 49 % in Bulgaria. Therefore, we do not recommend using NA data for weights; we carried out test calculations just in order to compare the methods. Generally, in the case of a B2B SPPI, we would only recommend using method 2 if more detailed NA and HICP data are available; otherwise, we would propose using the B2B SPPI as a proxy for the B2All SPPI.
22.6. Experimental calculations of ISPs
22.6.1. Objectives of the calculation
The main objective of the computation is to produce an experimental ISP for security and investigation activities. The ISP will measure changes over time in the volume of sector output and price changes will be removed from nominal values to allow us to compare these. We need to describe the relevant characteristics of the security and investigation industry and determine the requisite input variables and available data sources on the basis of periodicity and completeness. Finally, we elaborate and describe methods and recommendations for the calculation of an ISP.
.00 20.00 40.00 60.00 80.00 100.00
80.1
80.2
80.3
2013
2012
2011
305
22.6.2. Special characteristics of the security and investigation industry
First, it is important to analyse the structure of the industry and the levels of activity breakdown for which the variables relevant for the calculation of an ISP are available. Figure 121 shows the structure of the industry on the basis of harmonised activity and product classifications.
The ISP has to be published at NACE Rev. 2 two-digit level, but for reasons of quality it may be necessary to base the calculation on a more detailed level. NACE 80 breaks down into three groups, of which 80.1 (private security activities) and 80.2 (security systems service activities) are the most relevant in terms of turnover or value added from SBS (see section 3.3.3). In general, enterprises are classified according to their main activity, but normally they have secondary activities (e.g. enterprises classified in private security activities may also have security systems service activities). Therefore, one would ideally have information at product level in order to produce a very detailed weighting scheme. Unfortunately, there is no legal requirement at European level for a breakdown of security and investigation activities into CPA, so it is assumed that the most detailed information for most of the countries is available from SBS (NACE Rev. 2 four-digit level).
Figure 119: Security and investigation industries – structure
Source: Classification database, Statistics Austria
22.6.3. Input variables/deflators
Table 83 lists possible input variables for production of an ISP for security and investigation services and indicates their purpose for the calculations and their periodicity and timeliness.
NACE Rev. 2.0.
group
NACE Rev. 2.0.
division
80 Security and
investigation
activities
80.1 Private
security activities
80.2 Security
systems service
activities
80.3 Investigation
activities
306
Table 82: N80 – overview of main input variables for an ISP
Below, we give a more detailed description of input variables with reference to the calculation of the ISP for security and investigation services and highlight availability issues.
STS turnover
STS turnover indicators (available for division 80 on a quarterly basis as gross and WDA values) are the relevant input variables to be deflated. We also used SA values produced from Eurostat. The current legal basis requires calculation of the indicator on a quarterly basis 60 days after the end of the reference period. The draft FRIBS STS package proposes a monthly periodicity of turnover indicators, but for the test calculations we used the quarterly turnover indicators that are currently available.
Price statistics (SPPI, HICP)
Security and investigation activities are produced for households and for business customers (including government) for intermediate consumption. In general, this breakdown of clients should be considered in the course of deflation. An ideal deflated index would be produced on a B2All basis using B2C and B2B price data.
If an SPPI for security and investigation activities is produced on a B2All basis, this is the preferred price index for deflation. If not, additional data such as CPI or HICP should be used as a proxy for the B2C component, but there are two problems for this industry:
� no HICP is available for security and investigation activities, so the general HICP has to be used; and
� NA data are available only for N80-N82, in which N80 produces only a small proportion of turnover.
As a result, analysis has shown that an SPPI and a CPI for these services should not be weighted together to produce a B2All index, because the NA and HICP data are not suitable for use in N80 weighted calculation.
Indicator Periodicity Deadline Breakdown Purpose STS turnover indicator
quarterly t + 60 days division 80 current price output deflator
SBS turnover annual t + 18 months division 80 and groups 80.1. to 80.3
weighting
NA (value added, distribution of final consumption)
annual t + 10 months, t + 4 years
division 80-82
weighting (only if detailed national data are available
SPPI quarterly t + 90 days division 80 deflator CPI monthly t + 16 days general HICP deflator (B2C)
307
No B2C indices are produced for security and investigation activities, only two countries (PL, SE) provide a B2All index and 12 (CZ, EE, IE, EL, ES, CY, LV, LU, AT, SI, SK, UK) provide a B2B index.
Countries use different pricing methods: some (CZ, DE, EL, CY, PL, FI, SE) use contract pricing, two (FR, SK) use the transaction price of a product, one (NL) uses model pricing and others (DK, EE, IE, ES, LV, LT, LU, HU, AT, SI, UK) use combinations of methods. Austria uses the transaction price of a product combined with model pricing.
Another issue is the availability of price statistics. While CPIs are available in time, SPPIs are available one month too late. In addition, we have to bear in mind the possible monthly periodicity of turnover indicators, as proposed in the STS package, in combination with quarterly SPPIs.
SBS data
Annual information on value added at factor costs and turnover is available from SBS about 18 months after the end of the reference period at NACE Rev. 2 four-digit level. SBS provide basic information for weighting — unfortunately, EU legislation does not require a breakdown by CPA. Neither do SBS provide a breakdown by clients (business customers or households).
Recommendations for imputing missing input variables
The ISP has to be calculated with predefined timeliness and periodicity, and the requisite input variables have to be available. If this is not the case, missing variables have to be calculated by applying statistical estimation methods. Table 84 gives an overview of the options.
Various methods can be used to impute missing information. Quality of estimates should be kept in mind and the most reliable method chosen in the light of what information is available for the country in question.
It should be noted that forecasting price indices is difficult and probably unreliable, so it will result in a greater need for revisions. Also, suitable methods need to be developed as regards the use of quarterly price indices to deflate monthly turnover indices. Member States are engaged in further analysis with a view to preparing a methodological section.
308
Table 83: Problems and possible solutions
Situation Problem Possible solutions
ISP deadline is t+60 days, but SPPIs are available at t+90 days
Missing price information at the end of time series
Preferred: forecast with information on other variables in the same NACE category or with time series information only; Alternative: impute using information on other variables in the same NACE category with the same statistical behaviour; Other: forecast with simple method, e.g. simple moving average
ISP has to be calculated on a monthly basis, with SPPI available on a quarterly basis
FRIBS STS package changes periodicity from quarterly to monthly and SPPI remains quarterly
Preferred: forecast quarterly data for several periods (number of periods depends on the interpolation) ahead and interpolate a monthly path from forecast data; Alternative: impute using information on other variables in the same NACE category and with the same statistical behaviour that are available at higher frequency; Other: interpolate using own information only
SPPI time series are shorter than required for calculation of ISP
Missing data at the beginning of the series
Preferred: backcast with information on other variables in the same NACE category or with series information only; Alternative: impute using information on other variables in the same NACE category with the same statistical behaviour; Other: backcast with simple method, e.g. simple moving average
Missing data for weighting
No breakdown of turnover by CPA and/or clients
Preferred: use information from SPPI and NAs; Alternative: use NACE three-digit turnover breakdown from SBS and estimate breakdown by clients; Other: use statistical estimation models to calculate required information
Recommendations for imputing missing input variables
Calculation of an ISP for security and investigation activities has to be carried out with predefined timeliness and periodicity. At that time required input variables have to be available. If this is not the case missing input variables have to be calculated by applying statistical estimation methods. Table 84 can be used to gain an overview over the options for calculation of missing input variables.
309
In general, for imputation of missing information different methods are possible. Depending on the special country-specific situation the most reliable method has to be elaborated using available information. Quality of estimates should be kept in mind.
It should be noted, that the forecasting of price indices is difficult and most likely unreliable and therefore will result in a higher need for revisions. Also, suitable methods need to be developed regarding the use of quarterly price indices for deflation of monthly turnover indices. Further analysis is done by MS in preparation of a methodological chapter.
22.7. Detailed description of ISP calculation
22.7.1. Possible methods
An ISP can be produced using various methods. In general, the method that is suitable for a given country will depend on the availability of input data. Data gaps require some assumptions to be made. We tested two methods in relation to countries for which the full data matrix was available, generally using time series with base year 2010.
Table 84: ISP calculation methods
Method Recommendation Further information
Test calculation 1: Deflation with B2All SPPI for N80 Deflation at NACE two-digit level with B2All SPPIs at two-digit level
Deflation for gross, WDA and SA data (first adjustment, then deflation)
Recommended method for all countries for which a B2All SPPI for these services is calculated
ISP calculated for countries for which full set of data was available — this was possible for selected countries from Q1/2007 onwards.
Test calculation 2: Deflation with a combined price index from a NACE 80 B2B SPPI and a B2C general HICP Deflation at NACE two-digit level: B2B part with B2B SPPI and B2C part with HICP
Deflation for gross, WDA and SA data (series first adjusted, then deflated).
Preferred method if SPPIs are available for B2B only; a precondition is the weight for the B2C part of turnover and/or value added
Assumptions were made for the B2C part of turnover and applied for all countries for which test calculations were made
Possible method 3: Deflation with B2All SPPI at NACE three-digit level Deflation at NACE three-digit level with
More detailed than method 1; breakdown of
Test calculations were not
310
B2All SPPIs available at three-digit level
turnover by CPA would be ideal, but not available for most countries (otherwise, this would be the preferred method, if B2All SPPIs are available)
carried out.
22.7.2. Description of test calculations
Test calculations were carried out for two different methods. Basic data were extracted from the Eurostat database. We extracted WDA and SA data (TOVT, TOVT_WDA and TOVT_SA) from STS gross data and B2All (PRON) and B2B (PRBB) price indicators from the SPPI database. In general, base year 2010 was applied, so we brought HICP data (available for base year 2005 in the Eurostat database) to base year 2010 mathematically. We extracted the most recent available use tables from NA.
� for the first calculation, we divided STS gross, WDA and SA series by the B2All SPPIs for NACE activity 80;
� for the second calculation, we divided STS gross, WDA and SA series by weighting the B2B SPPIs for security and investigation services and the general HICP for NACE activity 80. Weights were based on NA use tables extracted from the Eurostat database. Exports in the tables were distributed to intermediate and final consumption in proportion to their share. Weights from NA caused problems with this calculation, because they are based on the N80-N82 aggregate and the general HICP.
If no B2All SPPI is available, but more detailed data are available from NA (for N80) and HICP (sub-indices for security) nationally, we would recommend the second method. However, if no detailed data are available, we suggest using the B2B SPPI as a proxy for the B2All SPPI.
For the test calculation, we deflated adjusted series, but this is not the proposed methodology in practice: in general, gross series should be deflated and then seasonally and working day adjusted.
The data sources we used (except HICP) are mostly available on a quarterly basis. HICP can be brought from monthly to quarterly periodicity very easily, by calculating the average of monthly data for each quarter under observation.
Finally, it should be noted that SPPIs are available with a time lag of t+90, so it will be necessary to forecast price information. The proposal for the FRIBS STS package envisages a monthly ISP, while the periodicity for SPPIs is quarterly, with an option of producing monthly SPPIs. It is assumed that SPPIs will be produced on a quarterly basis for most Member States in the future, so methods will have to be developed for forecasting and interpolating monthly data. These methods were not tested in the course of this exercise.
311
We performed test calculations for countries and EU aggregates for which a full data matrix was available, so did not impute missing data.
22.7.3. Results — ISPs for N80
The following section provides a description of selected main results of the test calculations. In general, the national applicability of the methods tested depends on data availability. If B2All SPPIs are available, this will be the preferred deflation method in most cases. The following figures show selected results for the EU-28.
312
Figure 120: Method 1 – SA and WDA series, not deflated (TOVT) and deflated (TOVV), EU-28
Source: Test calculations, based on data from Eurostat
Figure 121: Method 2 – SA and WDA series, not deflated (TOVT) and deflated (TOVV), EU-28
Source: Test calculations, based on data from Eurostat
70.00
75.00
80.00
85.00
90.00
95.00
100.00
105.00
110.00
115.00
120.00
20
06
Q1
20
06
Q3
20
07
Q1
20
07
Q3
20
08
Q1
20
08
Q3
20
09
Q1
20
09
Q3
20
10
Q1
20
10
Q3
20
11
Q1
20
11
Q3
20
12
Q1
20
12
Q3
20
13
Q1
20
13
Q3
20
14
Q1
20
14
Q3
20
15
Q1
TOVT WDA
TOVT SA
PRON
TOVV WDA
TOVV SA
70.00
75.00
80.00
85.00
90.00
95.00
100.00
105.00
110.00
115.00
120.00
20
06
Q1
20
06
Q3
20
07
Q1
20
07
Q3
20
08
Q1
20
08
Q3
20
09
Q1
20
09
Q3
20
10
Q1
20
10
Q3
20
11
Q1
20
11
Q3
20
12
Q1
20
12
Q3
20
13
Q1
20
13
Q3
20
14
Q1
20
14
Q3
20
15
Q1
TOVT WDA
TOVT SA
TOVV WDA
TOVV SA
313
Figure 122: Comparison between methods 1 and 2 for the EU-28
Source: Test calculations, based on data from Eurostat
The results of the two methods for the EU-28 are almost identical.
Since there are no suitable weights available from NA and no detailed price information from HICP (best suited is the general HICP), the calculations for the second method show nearly the same results as for the first. Therefore, it seems to be preferable to use method 1 or to use the PRBB indices as proxies for PRON indices.
22.8. Conclusions and further recommendations
It is possible to compile an ISP for NACE Rev. 2 division 80 (security and investigation activities) for countries for which B2All SPPIs are available. For other countries, there may be the following problems:
� there is no appropriate CPI (HICP) available which could be taken as proxy for the B2C part of turnover (the general HICP is the best suited);
� weighting could be based on NA, but only data for N80 to N82 are available; since the N80 share is rather small, these are not suitable; and
� application of the second method will require detailed information on the weight of the B2B and B2C parts of enterprises’ turnover and price development in the B2C part, which cannot be derived from available data.
It can be concluded that an ISP can be calculated if the following are available:
� B2All SPPIs;
� statistical forecast and interpolation methods to produce a monthly SPPI;
� if data are available on a B2B basis, suitable weights for the B2B and B2C parts and a suitable HICP;
80.00
85.00
90.00
95.00
100.00
105.00
110.00
115.00
120.00
20
06
Q1
20
06
Q3
20
07
Q1
20
07
Q3
20
08
Q1
20
08
Q3
20
09
Q1
20
09
Q3
20
10
Q1
20
10
Q3
20
11
Q1
20
11
Q3
20
12
Q1
20
12
Q3
20
13
Q1
20
13
Q3
20
14
Q1
20
14
Q3
20
15
Q1
Method 1 WDA
Method 1 SA
Method 2 WDA
Method 2 SA
314
� information from SBS, NAs or other business statistics for weighting; and
� knowledge of the security and investigation markets and market developments.
In addition, it should be noted that statistical methods will need to be developed for estimating missing monthly SPPIs at t+60, for almost all economic activities to be covered by an ISP. Apart from this development work, additional analysis and test calculations will be required in almost all countries.
315
23. SERVICES TO BUILDINGS AND LANDSCAPE ACTIVITIES (N81)
Kari Rautio, Statistics Finland
23.1. Introduction
The primary aim of compiling an ISP is to measure short-term movements in the production activity of the service part of an economy. Generally, this is the most important sector in the economy. In Finland, for example, its share of GDP is now over 70 % and that of the industry and construction sector is less than 30 %. Despite this, only a few Member States produce a volume ISP. However, some countries have many decades of experience of producing monthly indices for industrial production.
ISIC Rev. 4 division 81 (services to buildings and landscape activities) comprises activities that are important in modern economics, not least because the demand for such services is growing. The private and public sectors have outsourced administrative and support service activities in the past and today enterprises in this sector provide clients with a broad range of services, sometimes complete facility management.
The broad range of services in division 81 can cause difficulties from a statistical point of view. Also, statistical turnover and service producer price surveys do not cover all three-digit codes in the division. Most NSIs have turnover data and SPPI data for N81.2 (cleaning activities), but not for N81.1 (combined facilities support activities) and N81.3 (landscape care and maintenance service activities).
For quality reasons, the best means of collecting data for this service industry is probably a combination of survey instruments and administrative data. In general, NSIs can impose a higher response burden on large enterprises, so survey instruments are more suitable for enterprises with a significant role in the sector. Administrative data should be used for small enterprises, due to the need to limit response burden. Additional estimation procedures and data processing facilities have to be developed and put in place together with the administrative data. Obviously, specific needs (e.g. revenues may include service categories not under examination, etc.) dictate that administrative data are prepared for turnover statistics.
Cleaning, laundry services and gardening services are provided not only by enterprises, but also by private individuals. The difficulty here is to determine this turnover or to classify all household-related services in a single branch on the basis of a classification by economic activities. Nonetheless, such services are important, because demand for them is rising on the back of demographic trends in many countries (ageing and increasing employment for women).
Outsourcing has changed the N81 business pattern considerably in recent decades. In the past, for example, hospitals employed their own cleaners and buildings had caretakers. Nowadays, large national or international companies provide cleaning services and there are also many new small enterprises in the N81 service sector.
316
Another issue, mentioned at the Voorburg Group meeting in Vienna (2010), concerns illegal employment by cleaning enterprises. In Norway, a trade organisation has estimated that the black economy affecting N81.2 (cleaning services) accounts for 10-20 % of total turnover in the sector. If this is true and the Norwegian figures are representative for other countries, the level of turnover is probably systematically and considerably higher than reported by NSIs. There could be several reasons for an extensive black economy:
− the cleaning service sector is highly competitive and labour-intensive; and
− the pressure on costs for enterprises in the (non-specialised) general cleaning service sector may be higher than for some other branches, resulting in higher illegal employment and/or earning.
In Finland, tax credits are granted for domestic cleaning costs in order to address this problem. Households cannot claim reductions if the enterprise they hire does not pay business taxes.
23.2. Services to buildings and landscape activities in Finland
Figure 125 shows the unadjusted turnover index for N81 in Finland, a deflated unadjusted and an SA index. There appears to be a strong ‘hockey-stick’ effect in this sector. Invoice-based recording means that turnover is concentrated at the end of the year. Activity is also concentrated, with the four biggest enterprises responsible for 60 % of the industry’s turnover, which is mainly generated inside Finland.
23.3. Methodological considerations
Since the primary aim of compiling an ISP is to measure short-term movements in the production activity of the services part of an economy, the ISP should be defined as a weighted average of the real output of these industries, where the weights are based on their shares in the value added of the services sector.
Where no value information is available, certain volume indicators can also be useful (e.g. in Finland, how much snow is moved). Possible input indicators include employment and hours worked. These are often easily available, but the results are often difficult to interpret due to uncertainties regarding the development of labour productivity.
317
Figure 123: N81 – turnover and volume index, Finland
23.3.1. Statistical units
Various statistical units can be used: enterprise, enterprise group, KAU or establishment.
The administrative data in Finland are based on enterprise turnover. It is important to distinguish between large multi-national enterprises and smaller groups, as enterprises are classified according to the main source of their revenue. If cleaning is an important source of revenue, but not the main source, the enterprise is not classified as belonging to the cleaning industry, but another industry. To measure the turnover from cleaning activity, we have to separate the enterprise into smaller parts.
23.3.2. Deflation method
The deflation of a turnover/sales index by an appropriate output price index is the most suitable method for producing an ISP. In Finland, monthly administrative data for turnover are available, as are direct survey data for large enterprises.
From a methodological point of view, the best deflators are SPPIs, but these are often not available on a monthly basis.
In Finland, industry N81 is highly concentrated. Four big enterprises account for over 60 % of the total industry turnover. Labour shortages prevent enterprises from growing fast. Hundreds of small enterprises have been established due to tax reductions for cleaning in households (this mainly affects CPI prices). Thresholds to starting up or going out of business are low.
SPPIs have been developed in Finland since 2002 and published since 2004. Enterprises provide us with price information on cleaning contracts that they
318
have selected themselves. Statistics Finland collects prices from enterprises on a quarterly basis. Price data are collected via email questionnaire.
It is also possible to use a PPI or CPI as a deflator for the service sector.
23.3.3. Weights used for the index construction
The recommended weights for the index compilation are gross value shares. In the case of a Laspeyres index, weights for production indices are normally updated every five years. Many countries use moving weights for the IPI and this is also recommended for the ISP.
23.4. References
Compilation manual for an index of service production (OECD)
Several Voorburg Group presentations at the meeting in Newport, UK (19-23 September 2011)
Elina Pääkkö, Producer price index for services in Finland, Luxembourg (24 April 2008)
319
24. OFFICE ADMINISTRATION (N82)
Michaela Maršálková, Czech Statistical Office
After a brief introduction of available methodological information on the compilation of an ISP, this section analyses key aspects of the office administration industry (NACE Rev. 2 N82). We explain what available statistical information can be used to compile an ISP and present experimental ISPs for the Czech Republic and for France.
24.1. Available methodological information
Division 82 (office administrative, office support and other business support activities) was a new group in NACE Rev. 2. It was formed largely from parts of NACE Rev. 1 division 74 (sub-groups 74.82, 74.85, 74.86 and 74.87).77
Any sources compiled under the previous classification are thus of limited use when considering the new division 82, since they would have to go down to (and preferably even below) four-digit level. Also, the old NACE Rev. 1 division 74 was large and very diverse, so detailed exploration is quite difficult. As it is, the OECD ISP Manual recommends deflating turnover with appropriate PPIs and mentions some possible ones, including three that are relevant here: contract cleaning, contract packaging and stenographic services.
The Eurostat Handbook does not cover the relevant ISIC groups (74.82, 74.83 and 74.84) specifically, but states only that ‘most of the photographic, packaging and secretarial services will be standard type services, for which prices in principle can be collected. There will however also always be some part of the services that is of a unique nature. Other business services are an amalgam of special types of services that will usually be of a unique nature’. The general recommendation is to deflate output with appropriate and representative price indices, with corrections for changes in quality (A method). Use of partially representative price indices or lack of correction for quality changes would constitute a B method.
As for national methodologies, NACE Rev. 1.1 is covered in methodological documents for the UK Index of Services (GDP: Output approach methodological guide — Revised, Sharp P., 2003). Originally, the turnover for the three relevant four-digit ISIC codes was deflated, in the absence of appropriate deflators, by a mixture of average earnings for the whole section and the retail price index. This approach was criticised in a later business activities industry review, so the method was changed to include SPPIs for packaging activities and secretarial activities. The SPPIs for these classes under the new classification (82.10 and 82.92) are computed by direct use of prices of repeated services from several categories described in Services producer price indices – methods and guidance (ONS, 2014).
From the general methodologies available, it follows that the best approach is to deflate turnover by appropriate SPPIs. The price indices should be representative
77 For the correspondence between NACE Rev. 1 and Rev. 2, see Table 86.
320
and controlled for quality changes. Ideally, the deflation should be at a lower (NACE three-digit or even four-digit level) and the resulting series aggregated.
24.2. Presentation of the industry
NACE Rev. 2 division 82 includes units performing service activities, usually in support of other businesses. They do not provide staff to carry out the complete operations of a business. More specifically, the group comprises office administrative services and ongoing routine support functions, but also other support activities not classified elsewhere.
By definition, businesses in division 82 cater mainly to business customers. Only limited services may also be available to consumers (photocopying, gift-wrapping, public stenography or credit bureaux) and the proportion of value from consumer services is negligible.
24.2.1. Classification
The division belongs to section N (administrative and support service activities). Table 86 shows how it is made up of four groups and how its components correspond to the previous NACE Rev. 1 classes.
Table 85: Division 82 – NACE Rev. 2 classification and NACE Rev. 1 correspondence
NACE Rev. 2 Description NACE Rev. 1 82.1 82.11 Combined office administration activities 74.85
82.19 Photocopying, document preparation and other specialised office support activities
74.85 (64.11)
82.2 82.20 Activities of call centres 74.86 (whole) 82.3 82.30 Organisation of conventions and trade shows 74.87 82.9 82.91 Activities of collection agencies and credit
bureaus 74.87
82.92 Packaging activities 74.82 (whole) 82.99 Other business support activities n.e.c. 74.85
The ISIC v. 4 classification is the same, but under ISIC v. 3.1 the relevant activities comprised classes 74.95 (packaging activities) and (parts of) 74.99 (other business activities n.e.c.).
The CPA 2008 product classification has a more detailed breakdown based on the same structure.
321
Table 86: CPA 2008 for office administration services
82 Office administrative, office support and other business support services 82.1 Office administrative and support services 82.11 Combined office administrative services 82.11.1 Combined office administrative services 82.11.10 Combined office administrative services
82.19 Photocopying, document preparation and other specialised office support services
82.19.1 Photocopying, document preparation and other specialised office support services
82.19.11 Duplicating services 82.19.12 Mailing list compilation and mailing services
82.19.13 Document preparation and other specialised office support services
82.2 Call centre services 82.20 Call centre services 82.20.1 Call centre services 82.20.10 Call centre services 82.3 Convention and trade show organisation services 82.30 Convention and trade show organisation services 82.30.1 Convention and trade show organisation services 82.30.11 Convention organisation services 82.30.12 Trade show organisation services 82.9 Business support services n.e.c. 82.91 Collection agency and credit bureau services 82.91.1 Collection agency and credit bureau services 82.91.11 Credit reporting services 82.91.12 Collection agency services 82.92 Packaging services 82.92.1 Packaging services 82.92.10 Packaging services 82.99 Other business support services n.e.c. 82.99.1 Other business support services n.e.c.
82.99.11 Verbatim reporting and stenotype recording services
82.99.12 Telephone based support services 82.99.19 Other miscellaneous business support services n.e.c.
24.2.2. Description
Group 82.1 (office administrative and support activities) comprises enterprises providing administrative services and other ongoing routine support activities for others.
Class 82.11 (combined office administrative service activities) covers day-to-day office administrative services, such as record-keeping, billing, logistics, mail or reception, on a contract or fee basis. However, it does not include firms that perform only one of these tasks; similarly, the sole provision of operational staff without supervision should be classified in NACE 78 instead.
322
Class 82.19 (photocopying, document preparation and other specialised office support activities) includes services such as editing, proofreading, word processing, secretarial services and copying, without full-blown printing services. Copy centres belong to this class and are the exception in that they serve consumers as well as business customers.
Group 82.2 (activities of call centres) include call centres, both outbound and inbound, whether operated by human operators or automated systems. Typical functions are customer assistance, product sales and market research. A call centre may be dedicated to one enterprise or serve various business customers. While there are relatively few enterprises in this group (2.1 % of division 82), they are larger operations and employ a lot of people (22 %).
Group 82.3 (organisation of conventions and trade shows) has the fewest employees (6.7 % of division 82). These businesses organise, manage and promote various events. Units are not classified under this heading if they just lease the facilities in which the event takes place. Turnover in this group is often seasonal and irregular, with few events providing high earnings.
Group 82.9 (business support services n.e.c.) is made up of three parts. Enterprises in class 82.91 (activities of collection agencies and credit bureaux) collect payments owed to their clients or compile information on businesses and individuals financial histories, including employment histories.
Class 82.92 (packaging activities) provides bottling and packaging to other units on a fee or contract basis. This includes labelling, stamping, imprinting and also parcel-packing and gift-wrapping.
24.2.3. Data and facts
According to SBS data, there were almost 480 000 enterprises in division 82 in the EU-28 in 2012, most (75.6 %) in sub-division 82.9. The number of employees reached 2.1 million and while sub-division 82.9 still accounts for a majority (62.1 %), almost a quarter (22.2 %) were in sub-division 82.2. Value added for the division was EUR 82 000 million, or 19.4 % of value added in section N and 3.3 % of value added for services (NACE H+I+J+L+M+N).
The importance of NACE 82 varies across countries; in terms of value added, it generally accounts for 10-20 % of section N, but the figure ranges from 35 % in Portugal and 34 % in Slovakia to 7 % in Luxembourg and 4 % in Croatia. In absolute terms, most of the European value added for the division is generated in the UK, Germany and France.
323
Table 87: NACE N82 – economic indicators (EU-28)
2011 2012 Number of enterprises 477 749 479 666 Turnover (EUR million) 190 000 207 000 Value added (EUR million) 76 800 82 000 Persons employed (thousands) 2 049 2 100 82.1 82.2 82.3 82.9 Number of enterprises 81 269 10 083 36 932 351 384 Turnover (EUR million) 13 600 20 500 20 818 152 000 Value added (EUR million) 6 300 12 180 6 574 57 060 Persons employed (thousands) 196 466 140 1 034
Source: Eurostat, data codes sbs_na_sca_r2, sbs_na_1a_se_r2
Figure 124: Division 82 – value added (% of section N)
Source: Eurostat, data code sbs_na_sca_r2
The composition of the industry also varies across countries; while group 82.9 generally produces the most value added, in Ireland 52 % of value added comes from group 82.2 and in Luxembourg most comes from group 82.1.
A significant proportion (or even a majority) of enterprises can be categorised in group 82.1 (69 % in Spain), but they generally do not contribute to the same extent to value added (only 17 % in Spain). This difference is less pronounced in other countries, e.g. Denmark (48 % of enterprises in group 82.1, producing 40 % of the value added).
On the other hand, group 82.2 comprises only relatively few enterprises (under 1 % in Slovakia, Portugal and Estonia; at most 16.4 % in Finland), but accounts for far more employees and value.
0%
10%
20%
30%
40%
EU
28
BE
BG CZ
DK
DE
EE IE EL
ES
FR
HR IT CY
LV LT LU HU NL
AT
PL
PT
RO SI
SK FI
SE
UK
NO
324
Figure 125: N82 – economic indicators by country
Source: Eurostat, data code sbs_na_1a_se_r2
Quarterly STS data for the EU-28 show an upward trend in recent years, with a slight decline in 2008 and 2009 due to the economic crisis. Figure 128 shows a strong seasonal pattern, with a major peak in the fourth quarter and a minor peak in the second quarter, which grew more pronounced around 2008. Analysis of available monthly series suggests the biggest peak is mainly in December, but other behaviour over the year differs across countries.
EU
28
BE
BG CZ
DK
DE
EE IE EL
ES
FR
HR IT LV LT LU HU NL
AT
PL
PT
RO SI
SK FI
SE
UK
NO
Number of enterprises
EU
28
BE
BG CZ
DK
DE
EE IE EL
ES
FR
HR IT LV LT LU HU NL
AT
PL
PT
RO SI
SK FI
SE
UK
NO
Value added
EU
28
BE
BG CZ
DK
DE
EE IE EL
ES
FR
HR IT LV LT LU HU NL
AT
PL
PT
RO SI
SK FI
SE
UK
NO
Persons employed
N821 N822
N823 value unknown
325
Figure 126: NACE 82 – quarterly turnover (EU-28)
Source: Eurostat, data code sts_setu_q
24.3. Available data
The STS Regulation requires Member States to submit quarterly turnover data for division 82 at two-digit level. Non-deflated turnover (gross and WDA) is a mandatory indicator. SA data may be supplied optionally. Currently, 11 Member States also supply these data voluntarily on a monthly basis (the Czech Republic, Spain, France, Luxembourg, Portugal, Romania, Slovenia, Finland, Sweden and the UK, plus Denmark, which sends the data confidentially).
The Regulation does not require SPPIs for NACE 82, but the UK and France already supply quarterly SPPIs for the division. Some Member States publish SPPIs for selected sub-divisions: France (quarterly, 82.19, 82.20, 82.30), the UK (quarterly, 82.10, 82.92), the Czech Republic (monthly, 82.30).
The Regulation requires labour data, but only at section level and on a quarterly basis. Mandatory indicators are number of persons employed, hours worked and gross wages and salaries. At this time, four countries — Spain, Portugal, Turkey and the UK (only hours worked) — provide quarterly labour data for division 82; Spain and Portugal also provide some of the data on a monthly basis.
Under the SBS Regulation, annual turnover and value added at factor cost broken down to NACE Rev. 2 class (four-digit level) should be provided; this includes all sub-classes of NACE 82.
Since the majority of production in division 82 is business-oriented, CPIs are not applicable. Some low-level indices of copy centre services may be available for deflating class 82.19, if practicable.
It seems clear that the successful introduction of an ISP for division 82 will depend on the production of additional SPPIs.
40
60
80
100
120
140
20
00
Q1
20
00
Q4
20
01
Q3
20
02
Q2
20
03
Q1
20
03
Q4
20
04
Q3
20
05
Q2
20
06
Q1
20
06
Q4
20
07
Q3
20
08
Q2
20
09
Q1
20
09
Q4
20
10
Q3
20
11
Q2
20
12
Q1
20
12
Q4
20
13
Q3
20
14
Q2
20
15
Q1
ye
ar
20
10
= 1
00
326
24.4. Experimental ISP
The experimental ISP will be constructed as SPPI-deflated turnover. Since NACE 82 turnover is readily available, the main problem will be obtaining appropriate SPPIs. First, we tried to produce an ISP for the Czech Republic.
While the Czech Republic publishes an index for NACE 82, it includes only group 82.3 (organisation of conventions and trade shows), which is not necessarily representative of the whole division.
Other possibilities for deflation are the price index for market services as a whole or, as a last option, NACE 74, which shares some similarities with NACE 82. The behaviour of these indices varies, but not radically:
Figure 127: SPPIs for N82, N74 and market services (2010 = 100)
An experimental ISP for the Czech Republic was constructed by deflating the turnover for NACE 82 with the available (non-representative) SPPI for NACE 82. An alternative ISP (ISP 2 in Figure 130) was derived by deflating the same turnover with an SPPI for market services.
95
96
97
98
99
100
101
102
103
104
20
10
M0
1
20
10
M0
3
20
10
M0
5
20
10
M0
7
20
10
M0
9
20
10
M1
1
20
11
M0
1
20
11
M0
3
20
11
M0
5
20
11
M0
7
20
11
M0
9
20
11
M1
1
20
12
M0
1
20
12
M0
3
20
12
M0
5
20
12
M0
7
20
12
M0
9
20
12
M1
1
20
13
M0
1
20
13
M0
3
20
13
M0
5
20
13
M0
7
20
13
M0
9
20
13
M1
1
20
14
M0
1
20
14
M0
3
82 Market services 74
327
Figure 128: N82 – experimental ISP, Czech Republic (2010 = 100)
The alternative series presents an opposite price movement, but the difference is quite small (4 percentage points at the most).
As a comparison, we tried to construct an experimental ISP for France, since the French statistical office produces more detailed turnover and SPPIs. Series were taken from the INSEE website.
Monthly turnover indices are available at NACE three-digit level and quarterly SPPIs are available for NACE 82.19, 82.2 and 82.3.
Each three-digit turnover series was deflated by an SPPI — 82.1 by SPPI 82.19, 82.2 by SPPI 82.2, 82.3 by SPPI 82.3 and 82.9 by an average of the three. It seems that the resultant ISP is also quite close to the original turnover. However, finding a suitable deflator for NACE 82.9 is a problem; the group is very disparate, and in the absence of a representative price index, we have to approximate to some extent. In Figure 131, ISP 2 represents an alternative ISP, where NACE 82.9 was deflated by an SPPI for section N. Since this group is quite large and the price very seasonal, this produces a very different ISP.
80
90
100
110
120
130
140
20
10
M0
1
20
10
M0
3
20
10
M0
5
20
10
M0
7
20
10
M0
9
20
10
M1
1
20
11
M0
1
20
11
M0
3
20
11
M0
5
20
11
M0
7
20
11
M0
9
20
11
M1
1
20
12
M0
1
20
12
M0
3
20
12
M0
5
20
12
M0
7
20
12
M0
9
20
12
M1
1
20
13
M0
1
20
13
M0
3
20
13
M0
5
20
13
M0
7
20
13
M0
9
20
13
M1
1
20
14
M0
1
20
14
M0
3
Turnover ISP ISP 2
328
Figure 129: N82 – experimental ISP, France (2010 = 100)
Source: INSEE — indices des prix à la production (PPI) division
80
85
90
95
100
105
110
115
120
125
20
10
Q1
20
10
Q2
20
10
Q3
20
10
Q4
20
11
Q1
20
11
Q2
20
11
Q3
20
11
Q4
20
12
Q1
20
12
Q2
20
12
Q3
20
12
Q4
20
13
Q1
20
13
Q2
20
13
Q3
20
13
Q4
Turnover ISP ISP 2
