THE GTAP DATABASE, GTAPinGAMS STATIC MODEL AND DATABASE
5.4 AUGMENTATION OF THE GTAP DATABASE
The GTAP database is augmented here by the inclusion of capital and labour data estimated in Chapter 4. The three data sources are referenced using different industry classifications, which necessetaties a method for achieving consistency.
Labour estimates are referenced with respect to the 1992 Standard Industrial Classification (SIC(92)), capital estimates by major industry groups, and GTAP sectors by the United Nations’ Central Product Classification (CPC) or the International Standard Industry Classification, Rev. 3 (ISIC), depending on the type of production.
Concordance between the three data sources is most heavily influenced by the capital stock estimates. Only 22 sectors are identified in this series compared to 50 in the GTAP database and there are several GTAP sectors that do not map directly onto a major industry group. Consequently, aggregations of GTAP sectors and major industry groups into consistent industry groups are required to achieve uniformity. Aggregations and the composition of each consistent industry group in terms of SIC(92) two-digit classifications, major industry groups, and GTAP sectors are presented in Table 5.3. The starting point for aggregations not involved in agriculture or food processing, GTAP sectors 13-18 and 27-50, is the classification of GTAP sectors according to ISIC codes in McDougall (1998, Table 8.4, p. 16). Combined with the listing of SIC(92) classifications for each 1995 heading used to create major industry groups in
Table C of the Input-Output Supply and Use Balances: 1992-96, allows a concordance to be generated.7
On several occasions there is an exact one-to-one mapping between major industry groups and GTAP sectors. For example, major industry group 4, water, exactly matches GTAP sector 45. On other occasions GTAP sectors are aggregated to obtain a direct correspondence. For example, major industry group 6 (basic metals) is equivalent to GTAP sectors 35 (ferrous metals), 36 (metals nec) and 37 (metal products). In the case of the trade and transport consistent industry group, 14, major industry groups and not GTAP sectors need to be aggregated to achieve a concordance. Finally, in a minority of cases major industry groups are comprised of ISIC classification from several GTAP sectors and vice versa. Aggregations of major industry groups and GTAP sectors are required to generate consistent industry groups in such circumstances, as is the case for the machinery and electrical equipment major industry group, 7.
Although sufficiently high aggregation will yield an exact concordance (an extreme example is one sector under each classification), an exact correspondence between major industry groups and GTAP sectors is not achieved here in all consistent industry groups. For example, the SIC(92) three-digit industry 52.7, repair of personal and household goods, is included in the GTAP industry assigned to consistent industry group 15, financial and public services, but is recorded in a major industry group assigned to the consistent industry group 14,
7 At the two-digit level, SIC(92) and ISIC Rev. 3 are identical in content and coding.
trade and transport.8 Small sacrifices are made regarding the level of precision in order to maintain a reasonable level of industry disaggregation.
Sectors in agriculture and food processing are referenced by CPC codes in the GTAP database as it affords finer aggregations of industries than that allowed by ISIC headings. However, as a detailed analysis of the agriculture and food processing is not required here, GTAP sectors referenced by CPC are simply aggregated to agriculture or food and beverages and assigned to appropriate consistent industry groups.
Correspondence between GTAP sectors and industries for which labour types are identified is more precise than that between GTAP sectors and major industry groups. The New Earnings Survey database recognises SIC(92) industries at the five-digit level, which allows an exact mapping from SIC(92) industries to GTAP sectors identified by ISIC.9 Industry concordance for agricultural and food and beverage sectors are calculated in a similar fashion to those for capital, by aggregating these sectors and assigning data from appropriate SIC(92) industries.
Data for labour types for consistent industry groups is then generated by simple aggregation.
8 A complete list of imperfect concordances is given in the notes to Table 5.3.
9 Below two-digit groups, although some code numbers differ and aggregation of SIC(92) classes are occasionally required, SIC(92) codes can be directly converted to ISIC headings.
Table 5.3
Correspondence between major industry groups and GTAP sectors
Consistent Industry Group Major Industry Group GTAP Sector
# Description SIC(92) # Description # Code Description
1 Agriculture 01-05 1 Agriculture 1 pdr Paddy rice
2 wht Wheat
3 gro Cereal grains nec
4 v_f Vegetables & fruit
5 osd Oil seeds
6 c_b Sugar
7 pfb Plant-based fibers
8 ocr Crops nec
9 ctl Bovine cattle etc
10 oap Animal products nec
11 rmk Raw milk
12 wol Wool, silk-worm cocoons
13 frs Forestry
14 fsh Fishing
2 Oil and gas 11-14 2 Oil and gas 15 col Coal
16 oil Oil
17 gas Gas
18 omn Minerals nec
3 Fuels and chemicals 23, 24 & 40 3 Fuel 32 p_c Petroleum & coal
7 Chemicals 33 crp Chemical, rubber, plastic products
43 ely Electricity
44 gdt Gas distribution
4 Water 41 4 Water 45 wtr Water
Continued
Table 5.3
Correspondence between major industry groups and GTAP sectors (continued)
Consistent Industry Group Major Industry Group GTAP Sector
# Description SIC(92) # Description # Code Description
5 Other minerals 26 5 Other minerals 34 nmm Minerals nec
6 Basic metals 27-28 6 Basic metals 35 i_s Ferrous metals
36 nfm Metals nec
37 fmp Metal products
7 Machinery & electrical 29-33 8 Machinery equipment 40 ele Electronic equipment
equipment 9 Electrical equipment 41 ome Machinery and equipment nec
8 Transport equipment 34-35 10 Transport equipment 38 mvh Motor vehicles
39 otn Transportation equipment nec
9 Food and beverages 15-16 11 Food and beverages 19 cmt Bovine cattle, sheep meat etc
20 omt Meat products nec
21 vol Vegetable oils
22 mil Dairy products
23 pcr Processed rice
24 sgr Sugar
25 ofd Food products nec
26 b_t Beverage and tobacco products
10 Textiles and leather 17-19 12 Textiles and leather 27 tex Textiles
28 wap Wearing apparel
29 lea Leather products
11 Paper & printing 21-22 13 Paper & printing 31 ppp Paper & publishing
Continued
Table 5.3
Correspondence between major industry groups and GTAP sectors (continued)
Consistent Industry Group Major Industry Group GTAP Sector
# Description SIC(92) # Description # Code Description 12 Other manufacturing 20, 25 & 36-37 14 Other manufacturing 30 lum Wood products
42 omf Manufactures nec
13 Construction 45 15 Construction 46 cns Construction
14 Trade & transport 50-52, 55 & 16 Trade 47 t_t Trade, transport
60-64 17 Hotels and catering
18 Transport
19 Post and telecom.
15 Financial and public services 65-67, 70-74 20 Financial services 48 osp Financial and business services etc
80-85, 90-95 21 Public admin 49 osg Public admin, education and health
& 99
16 Dwellings n.a. 22 Dwellings 50 dwe Dwellings
Note: The mapping between GTAP sectors and major industry groups is not exact for all consistent industry groups. Specifically, the SIC(92) three-digit industry 24.7, manufacture of man-made fibres, is assigned to consistent industry group 10 when assignments are made according to GTAP sectors but is included in consistent industry group 3 when distinctions are made using major industry groups. Similar disparities arise for SIC industries 25 (assigned to consistent industry groups 3 and 12 using GTAP and major industry group classifications respectively), 22.14 and 22.3 (7 and 11), 37 (14 and 12), 71.1 and 71.2 (14 and 15), and 52.7 (15 and 14).
Concordance between industries identified in the three data sources allows GTAP factor data to be augmented. Labour data for the UK are incorporated into the GTAP database by dividing total labour payments identified by the GTAP database among different types of labour according to proportions identified in the New Earnings Survey.10 Payments to different capital assets are estimated by assuming that the return to each efficiency unit is equal across assets. This allows total capital payments in each consistent industry group to be divided amongst capital assets in a manner consistent with the capital stock estimates in Table 4.7.
Specifically, payment shares are equal to efficiency-unit stock shares.
5.5 CONCLUSIONS
The GTAP database contains information on production, including inter-sectoral linkages, bilateral trade, taxes and tariffs for 50 sectors across 45 regions in 1995.
In this study the data are operationalised using Rutherford’s (1998) GTAPinGAMS core static model.
Modifications to the GTAP database are made for the UK. Cost shares by labour and asset types are included by appropriately aggregating sectors and dividing total labour and total capital cost shares identified in the GTAP database between labour types and capital assets in a manner consistent with the estimates of Chapter 4.
This finest level of aggregation for which this is possible generates 16 industry groups, as detailed in Tables 5.3.
10 The New Earnings Survey in 1996 is used for this assignment, as this is this first year that the data source employed the SIC(92) coding system.