1.
Belgian inventory of sources and methods for measuring prices and volumes in the national accounts
January 2003
1. General approach......
1.1. Introduction......
1.2. Structure of the supply and use table......
1.3. Provisional method for calculations in constant prices......
1.4. General information on the available indices and the method applied......
2. General information on the main sources......
2.1. Indices of industrial producer prices......
2.2. Indices of import and export prices......
2.3. Indices of consumer prices......
3. Method applicable to each product......
3.1. General observations......
3.2. CPA A - Products of agriculture, hunting and forestry......
3.3. CPA B - Fish......
3.4. CPA C - Products from mining and quarrying......
3.5. CPA D - Manufactured products......
3.6. CPA E - Electrical energy, gas, steam and hot water......
3.7. CPA F - Construction......
3.8. CPA G - Wholesale & retail trade services, repair services of motor vehicles, motor-cycles & personal & household goods
3.9. CPA H - Hotel & restaurant services......
3.10. CPA I - transport, storage and communication services......
3.11. CPA J - Financial intermediation services......
3.12. CPA K - Real estate, renting and business services......
3.13. CPA L - Public administration and defence; compulsory social security......
3.14. CPA M - Education services......
3.15. CPA N - Halth and social services......
3.16. CPA O - Other community, social and personal services......
3.17. CPA - P Private households with employed persons......
4. Method for other parts of the system......
4.1. Added value......
4.2. Taxes and subsidies linked to products......
4.3. Components of value added......
5. Annexes......
5.1. Annex 1 Indices of industrial producer prices......
5.2. Annex 2 SUT products......
5.3. Annex 3 Methods per SUT product for the calculation of P1 in volume terms......
5.4. Annex 4 Deflators for consumption expenditure......
1.general approach
1.1.Introduction
This inventory describes how price and volume indices are obtained in the national accounts, and how the results in constant prices are arrived at.
Essentially, the description is based on the available information for the calculation for the year 2000. Where better information has become available which will be used in the calculations for a particular year in the near future, this will be indicated.
To date, supply and use tables (SUT) in constant prices have not been compiled in Belgium. It is therefore not yet possible to describe a general method using a particular year as an example, and the procedures described are those which have been followed for the years for which SUTs are not yet available.
An explanation of the structure of the SUT is therefore in order.
1.2.structure of the SUPPLY AND USE TABLE
The SUT contains more industries and products than the reports which are submitted to Eurostat. This permits detailed analysis and adjustment of statistical discrepancies between supply and use. In practice, however, the quantity and quality of the source data impose constraints on the reporting format.
The reporting format has 60 industries (A60), which correspond to the NACE 2-digit level. The SUT format identifies 120 industries. Industries were defined with reference to Eurostat division A60, the importance (as measured by turnover) of the various NACE-BEL 4-digit classes, the homogeneity of SUT industries and the difference between market and non-market activities.
The reporting format identifies 60 products (P60), which correspond to the CPA 2-digit level. The number of products in the SUT format is 320. Goods in the SUT are usually defined at CPA 4-digit level, and services at CPA 3-digit level. This rule is occasionally deviated from in order to reflect the importance of the product for the Belgian economy or for technical reasons.
The structure of the industries and products reflects the growing importance of services in the economy. Nearly 30% of the products relate to services. Service industries account for 43% of the total number of industries.
Broadly speaking, the SUT variables are broken down as prescribed in the reporting format. Further breakdown is sometimes indicated, on practical grounds.
ESA95 makes a distinction between consumer expenditure and actual consumption. The former shows who effects the expenditure, while the latter shows who acquires the consumption goods or services. The SUT applies the concept of consumer expenditure, whereas the sector accounts use both concepts. As consumer expenditure in the sector accounts is derived from the SUT, the use table makes the following distinction:
P3S14 household consumer expenditure
P31S13 individual consumer expenditure by general government
P32S13 collective consumer expenditure by general government
P3S15 consumer expenditure by NPISHs.
In order to ascertain which items of fixed capital formation (P51) should be given priority for balancing, and in order to maintain a link to the sector accounts, fixed capital formation (P51) in the SUT format is distributed among the sectors. The amounts per sector and per product are obtained by means of a separate investment module which also differentiates between industries.
1.3.provisional method for calculations in constant prices
An integrated calculation of gross domestic product (GDP) in current prices from the production, expenditure and income perspectives is made within the framework of the SUT.
When the definitive method is applied, the aggregates in constant prices, which are published in the autumn of reporting year t, will be based on a definitive SUT for year t-3, while those for t-2 will be based on a provisional SUT. It will not be possible to compile an SUT for year t-1, and a more rudimentary method will have to be applied.
As from the SUT for 2001, the calculation will be done in constant prices in the framework of a supply and use table, using the format described above. In this case, the income perspective is irrelevant for the market branches.
As things stand, the production method is the most important for the calculation in constant prices. For each SUT product, the best available index figure is used to deflate output and intermediate consumption at a detailed level. If no suitable price index is available, an implicit price index figure is calculated by dividing the value index by the volume index.
Indices for output deflation
The aggregate to be deflated is output per SUT branch (P1). It comprises output from the principal activity and from secondary activities (e.g. trade and transport margins for certain branches).
The output structure per SUT branch of the most recent definitive SUT in current prices, is considered sufficiently reliable for calculating a weighted price index for total output per SUT branch in subsequent years.
Because there is a degree of uncertainty about the price indices for trade margins per product, it is assumed for branches other than trade that:
-the distribution of trade margins by products is proportional to the distribution of non-trade output;
-changes in prices of the trade margins per product is the same as change sin the prices of the products to which those margins relate.
These assumptions imply that the trade margins in each SUT branch are irrelevant to the calculation of the output price index.
For trade branches SUT 50B1 (motor fuels), SUT 51A1 (wholesale trade) and SUT 52A1 (retail trade), the deflator comprises two components: the deflator for non-marginal output, and one for marginal output. The first component is weighted by non-marginal output, and the second with marginal output. The deflator for non-marginal output is calculated using the general method. Trade margins per product are used as the weighting in the calculation of the deflator of marginal output[1].
The output of the transport margins of the transport sectors is reallocated for each SUT branch to the transport product which is the main product of the branch in question[2]. The output of significant transport margins by branches other than the haulage sectors is reallocated to the transport of goods by road.
Reallocating the transport margin to a transport product therefore increases the weight of the transport output in the weighted deflator for the total output of this branch.
Indices for the deflation of intermediate consumption
The aggregate which is to be deflated is intermediate consumption (P2) per SUT branch. It comprises products from domestic output and imports. The deflator must therefore take account of both components. If no deflator for imports is available, the P1 deflator is sued to deflate imports.
For each product (the most detailed level of breakdown), a weighted index is calculated from the indices for domestic output and imports. The weighting coefficients used for this purpose are Belgian output and imports respectively of the product in question. It is assumed that, for each product, the share of imports in the intermediate consumption of every branch is the same, which means that the weighted index can be used for deflation purposes in every branch in which the product is consumed.
The index used to deflate total intermediate consumption is the weighted average of the indices of the various input products. The weighting is the input structure of the branch concerned. This input structure is based on the most recent available definitive SUT in current prices.
Once P1 and P2 have been deflated, the P2/P1 ratios are compared with those for previous years. If there is a significant discrepancy, further arbitrage may be indicated.
Estimate of taxes and subsidies linked to products
Taxes and subsidies linked to products are deflated using the deflator for the product to which they relate.
From the expenditure perspective, aggregates are required for imports and exports (P7 and P6), consumptive expenditure and investments (P5).
Imports and exports are deflated per SUT product. In the case of goods, unit values are available from the foreign trade statistics. In the case of imports and exports of services, the P1 deflator is used in the absence of better information.
Household consumer expenditure is estimated at a detailed level in current prices, and deflated using the consumer price indices.
The deflator used for the consumer expenditure of general government and of non-profit institutions serving households, the deflator is based on the deflator of output in the branches concerned and social payments in kind.
Fixed capital formation (P51) is deflated using the weighted average of the deflator for investments from domestic production and imports. The indices are calculated as described above for intermediate consumption.
Changes in stocks in constant prices are currently calculated using the commodity flow method.
1.4.general information on the available indices and the method applied
The available price indices for producer prices and consumer prices are of the Laspeyres type, with a fixed base year. Detailed information on values and volumes is available for imports and exports of goods, which basically makes it possible to calculate all types of indices at the level of the SUT product breakdown.
The series in the national accounts are compiled using a fixed base year (1995). In 2004 (accounts up to 2003), the transition to a variable base year t-1 will be made (chain indices).
Volume data are published in absolute terms. Changes in volume compared with the previous year, expressed in percentages, are also published in summary tables.
2.general information on the main sources
This chapter looks at the main sources for deflators. Specific sources for particular products are discussed in Chapter 3.
2.1.Indices of industrial producer prices
Council Regulation (EC) No. 1165/98 concerning short-term statistics provides, inter alia, for the compilation of industrial producer prices (referred to in the Directive as "output prices"), a distinction being made between output prices for the domestic market and those for the "non-domestic" market[3]. The output prices for NACE sections C to E have to be provided monthly, at NACE 2-digit level[4]. The term "output price" is exclusive of VAT and other deductible taxes.
A new base year has to be adopted every five years (with the most recent year ending in a 0 or 5). All index figures must be renewed within three years of the end of the new base year. At the end of 2003, the base year is 2000.
Up until 2001, the indices of industrial producer prices (PPII)[5] were compiled by the Indices Service of the Ministry of Economic Affairs. Since 2001, the PPII's have been compiled on a monthly basis, as required by the Regulation on Short Term Statistics, by the National Institute for Statistics (NIS) (cf. Annex 1). The data are collected via a computer-assisted telephone survey in which the data are input immediately and subjected to automatic checks (outbound call centre, with Blaise software).
The unit of observation in the survey is the local unit of economic activity. Participation in the survey is voluntary. The information providers are selected from a representative sample of the Prodcom population[6]. When the survey was first started, 784 relevant 8-digit Prodcom products were selected (26% of all Prodcom products in the database), which together make up 75% of the total value of output. The main producers of each product were approached with a view to asking them to take part in the survey.
After grouping on grounds of confidentiality, 285 variables eventually became available at CPA 4-digit level for 2000 (and 2001) . These indices have been in use since 2001.
A major problem with output price indices is quality change: if a product (say, a television set) undergoes changes which significantly affect its quality, the product is no longer the same, and the prices can no longer be compared. This problem is resolved by selecting an overlapping period (a month): if, for whatever reason, a highly specific reference product produced by an information provider for a given Prodcom heading can no longer be monitored, it is replaced by a different reference product in the same heading, in consultation with the information provider; one month with overlapping price quotations is therefore sufficient to maintain the price series.
The PPII (Indices Service) was compiled for CPA sections C, D and E, and contained 84 headings. These indices were used for calculations in constant prices in the national accounts up to 2000.
In view of the thorough revision of the method, it is not possible to create a link at a detailed level between the PPIIs produced by the Indices Service and the NIS respectively.
2.2.Indices of import and export prices
Foreign trade statistics are compiled by the National Bank of Belgium (NBB) for account of the National Accounts Institute (INR) at the most detailed level of the Combined Nomenclature (CN8).
The index figures are not genuine index figures, but rather, unit values, which are calculated using the Fisher formula. The calculations are always made with respect to the previous year. For each CN8 product code, prices are observed for each unit of mass or for each supplementary unit. A filter eliminates excessively small initial values and extreme variations.
The available data on quantities and values make it possible to calculate indices using various formulae. These calculations are performed at SUT product level.
Product groups which are heterogeneous, and for which the unit values do not satisfy European rules on constant prices, are replaced by genuine index figures.
2.3.Indices of consumer prices
The indices used for deflating final consumption are derived from the national index for consumer prices (CPI_N). In addition to the CPI_N, pursuant to the Treaty of Maastricht, a Harmonised Index of Consumer Prices (HICP) has also been compiled in recent years.
The Indices Service of the Ministry of Economic Affairs is responsible for recording prices and calculating consumer price indices. Every month, on about the dame date, 20 officials collect the prices of a large number of the various products in the basket. The CPI_N comprises a total of 481 products, while the HICP comprises more than 500 products[7].
The data are collected at nearly 10000 sales points (always the same ones) in 65 different places, giving more than 105000 price observations[8] per month. Two-thirds (in weighted terms) of price observations are made locally, the remaining one-third being made centrally (electrical energy, natural gas and water, passenger vehicles, postal and telephone services, leasing, medical services, foreign package holidays, newspapers, rail transport, etc.).
The monthly calculation of the CPI_N is done in three stages. Firstly, an index is calculated for each product at every location (481 x 65 = 31265 product indices). Secondly, a weighted average of these indices is calculated for each location. The weightings are based on the results of the household budget surveys. Lastly, the national index is calculated as the weighted average of the indices for all 65 locations. Geographical weightings are based on population figures for the first day of the index reference year[9].
CPI_N is a Laspeyres index with a fixed base year (currently 1996).
Price changes of products which are included in the HICP but not (or not yet) in the CPI_N, are used to calculate the national accounts in constant prices.
For the purpose of application to the SUT, a weighted price index for household consumption expenditure is calculated for each SUT product from the elementary indices of consumer prices.