6. Compilation of the database for the GEM-E3 model: the example of Hungary
6.7. Bilateral trade matrix
As we have seen in section 6.1 for Hungary import matrices in 21 sectors break-down were officially published for 1995 as a supplement for the I-O table.
Bilateral (country by country) trade matrices had to be compiled for the export and import turnover to and from the European Union (EU 15), the other ECE countries and the ‘rest of the world’ (inclusive tourist expenditures).
In Hungary, the main problem with the available mentioned published data (Foreign Trade Statistical Yearbook) was that in the case of bilateral trade it contained only the largest shipments, so in many cases the structure of trade with smaller EU- and accession countries could only be estimated. In addition, the published data by country was available only for the 20 main categories (sections) of the HS nomenclature (for custom free zones the country break-down was not available at all). Therefore, we had to split Section 2, 5, 15 and 16 to get the break-down by our 18 industries. A minor problem is that within Section 1-3 there are intermediate goods too, while we assign them to “Consumer goods” exclusively. Similar problems can be mentioned in connection with other sections too.
Trade matrices for merchandise exports and imports have been compiled by converting the foreign trade data given in HS classification into NACE classification, and aggregated into the sector format of the GEM-E3 model.
Table 6.2: The aggregate balanced Hungarian SAM for 1995 (in million Forints, 1 Euro=125,69 HUF)
USERS--> Branches
Consumpt
ion Firms Exports Tourist Total
Investmen
ts Change in Total
Branches TOTAL Labour Capital Total
Househol
ds Govern. Banks export Export
Househ.+
NPIS Private Govern. Total Stocks
TOTAL of branches 5748257 0 0 0 3723955 617700 219248 0 1849552 223495 2073047 281848 678717 164824 1125389 218346 13725942
SSC 729886 0 0 729886
Wages (withoutSSC) 1905116 0 0 0 1905116
Capital 2297832 0 0 2297832
Total Value Added 4932834 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 4932834
Actual Output-Subs 10681091 0 0 10681091
HHS 0 2635002 999682 3634684 1673645 0 543290 0 0 0 5851619
FIRMS 0 1049968 1049968 361326 47998 0 0 1459292
Indirect Taxes (net) 134653 0 0 134653
Direct Taxes 0 0 411965 0 125419 0 0 537384
Social Security 0 0 894523 0 0 894523
Subsidies&mine rent -63883 0 0 0 -63883
VAT taxes 580257 0 0 0 0 0 580257
Duties 249430 0 0 249430
Gov. Foreign 0 0 0 0
Gov. firms 0 248182 248182 0 248182
Total Taxes 900457 0 248182 248182 1306488 0 0 125419 0 0 0 0 0 0 0 0 2580546
Distr. Output 0 0 0 0
EC 1249273 0 0 1249273
NON-EC 762810 0 0 762810
Total Imports 2012083 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2012083
Expend.abroad 132312 0 1634 0 23418 0 157364
Total Imports 2144395 0 0 0 1634 0 23418 0 0 0 0 0 0 0 0 2169447
Savings 0 0 459850 239568 0 547917 96400 96400 1343735
Total Resources 13725943 2635002 2297832 4932834 5851619 2580545 219248 1240044 1945952 223495 2169447 281848 678717 164824 1125389 218346 32063416
Table 6.3: The aggregate balanced Hungarian SAM for 1995 (in Millions of Euro)
Total Consumption Total
Exports
Investments Total Change in
Stocks Total Total Labour Capital Total Househ. Govern. FIRMS Exports Househ. Private Govern. Investm.
Total 36 701 0 0 0 22 896 3 798 0 12 745 12 745 1 733 4 173 1 013 6 919 1 342 84 402
Wages 11 713 0 0 0 0 0 0 0 0 0 0 0 0 11 713
SSC 4 487 0 0 0 0 0 0 0 0 0 0 0 0 4 487
Capital 12 779 0 0 0 0 0 0 0 0 0 0 0 0 12 779
Total Value
Added 28 980 28 980
Actual Output 65 681 65 681
HHS 0 16 200 6 146 22 347 0 10 290 3 340 0 0 0 0 0 0 35 977
FIRMS 0 0 6 455 6 455 2 222 295 0 0 0 0 0 0 0 8 972
Indirect Taxes 828 0 0 0 0 0 0 0 0 0 0 0 0 828
Direct Taxes 0 0 0 0 2 533 0 771 0 0 0 0 0 0 3 304
Social Security 0 0 0 0 5 510 0 0 0 0 0 0 0 0 5 510
Subsidies -393 0 0 0 0 0 0 0 0 0 0 0 0 -393
VAT taxes 3 568 0 0 0 0 0 0 0 0 0 0 0 0 3 568
Duties 1 535 0 0 0 0 0 0 0 0 0 0 0 0 1 535
Gov. Foreign 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Gov. firms 0 0 1 526 1 526 0 0 0 0 0 0 0 0 0 1 526
Total Taxes 5 537 0 1 526 1 526 8 043 0 771 0 0 0 0 0 0 0 15 877
Distr. Output 65 681 0 65 681
Imports 13 184 0 0 0 0 10 144 0 13 338
Total Imports 13 184 0 0 0 0 10 144 0 0 13 338
SAVINGS 0 0 0 0 2 827 1 473 3 369 593 0 0 0 0 0 7 669
Total Resources 84 403 16 200 14 127 30 328 35 987 15 866 7 624 13 338 12 745 1 733 4 173 1 013 6 919 1 342
It is worth mentioning, that in spite of these problems, the resulting data (see in the file Trade95.xls) are fairly consistent with what we could have got, provided we had access to the more detailed (electronic) trade statistics (like the Polish partners had) in time. Later obtained data (in 2 and 4 digit code breakdown of the HS nomenclature and in branch and sub-sector breakdown of the origin of the products) converted to the GEM-E3 model’s 18 sectors break-down (see the TradAgAc.Xls, TradSzag.Xls files for the exports to the accession countries, the TrSzagEU.Xls file for Hungary’s trade with the EU15, and the EU15’s exports to Hungary as registered in the Comext database and shown in the Tradenis.xls file) show a satisfactory coincidence with what we estimated from incomplete (in the case of small countries fragmentary) data. The exceptions are Hungary’s export to Austria, Greece, Great Britain and Spain, and Hungary’s import from Austria, Denmark, Finland, and Sweden, where in at least 3 branches the ratio of the two corresponding data (coming from the different estimations) is close to or above 2. On the aggregate level (which means that in quite a few countries) we observe that trade data in branch of origin break-down resulted in significantly lower estimates for the GEM-E3 sectors of Agriculture, Coal, Metal products, and Electrical goods than estimates based on the HS-code commodity break-down dataset. On the other hand, Chemical industry, Energy intensive industries and Equipments the former method’s estimates were lower than those of the latter.
In any case, in the following step we had to adjust the trade data (already in the 18 sector break-down) to the I-O table’s (also already aggregated to the 18 sector break-down) corresponding (export or import) trade figures. However, since the value of exports in the national accounts and the input-output tables include export subsidies as well (they are at basic prices), while the trade statistics data are essentially at users prices (contract prices, usually on fob parity), they had to be separated from the I/O table figures and displayed among the subsidies. Export is measured at f.o.b. prices, whereas import at c.i.f. prices21.
The adjustment eliminated not only the above mentioned (mainly aggregation) errors, but also eliminated those discrepancies which were due to the different methodology of the I-O table and the trade statistics (as discussed earlier, in relation with the processed materials and the like). However, in the SAM tourist exports are also added to the total exports, although we do not have any information of its country break-down. In this case we can assume that the country-structure of the tourist spending is the same as that of the company conducted foreign trade. If this again causes assymmetricity in the estimated trade flows between countries, it can be eliminated either pair-wise or only at the aggregate level of the EU or the world market.
Data for foreign trade of services (as found in the I-O table) is usually not available in country break-down either, and even its industry break-down information is not very reliable (usually they are rather different from the Foreign Balance of Payments statistics, which, by the way, has a different break-down of the service trade). In any case, the GEM-E3 model has a subroutine which deals with the balancing of this service trade on the aggregate EU- and world market level.
21 Note the calibration module of the GEM-E3 estimates the difference between the export and import prices as linked export of transportation and other services.
To estimate the custom duty data, one could use various sources on foreign trade data and official tariff rates. Based on the above information duty matrices have been estimated, making use of proportionality assumptions (within the blocks of the EU and ECE countries, as well as commodity groups). For Slovenia there was more detailed information available so that it produced more differentiated duty rates than the other countries. It is interesting to note that the Hungarian custom policy treated Poland and Slovenia almost the same way as the EU countries.