1. Practices of Norway
28.29. Current practices. The Department of Economic Statistics of Statistics Norway uses unit-value data from customs records and combines them with survey-based price data to compute price and volume indices of external trade in commodities. Survey price data are used for some commodity codes for which customs data are not considered acceptable as a source of price statistics. When this is the case, priority is given to Norway’s data sources, and only if they, too, are not adequate, are data collected from international sources (such as the United States Bureau of Labor Statistics).
28.30. Use of foreign trade indices in national accounts compilation. Detailed data on exports and imports at the HS eight-digit level are aggregated to the national accounts product level (going from about 6,500 commodities to approximately 700 product categories). Values of exports and imports are used in balancing the national accounts at current prices. Price information is applied to national accounts at the detailed CPA level to derive exports and imports at constant prices, and to estimate price changes of components where no price observations exist (namely, for intermediate consumption, gross fixed capital formation, and final consumption).
28.31. Future developments. It is expected that, in the future, large importers and exporters will be allowed to lodge consolidated customs declarations on a monthly basis. Although this will have a positive effect in terms of trade facilitation, statisticians will need to rely on fewer data records (owing to the consolidation of single transactions into monthly totals) ; hence, error detection procedures may be further complicated. Also, the general trend towards removal of tariffs and duties, and the inherent interest of customs in simplifying and facilitating trade procedures, means that additional efforts will have to be made in order to preserve the quality and coverage of data from administrative customs records. Currently, the goal of Statistics Norway is to replace unit-value indices with survey-based price indices for both exports and imports. For more detailed information on the practices of Norway, see annex XXVIII.A.
2. Practices of Canada
28.32. Current practices. Statistics Canada compiles an International Merchandise Trade Price Index (IMTPI), which is a composite price index designed to express, in a single index, price changes that involve a range of commodities. In order that realities of price movement may be accurately reflected, a fixed basket of goods is chosen which are representative of, and correlated to, the rest of the commodities in the trade universe. The index is based on a non-random sample of import and domestic export commodity classes. Data are extracted from administrative files and derived from other Statistics Canada surveys and/or other sources. International trade price and volume indices are constructed on the basis of unit values derived from detailed customs-based data and survey price indices taken from Canadian and foreign sources. As a general rule, unit values are retained for relatively homogeneous commodities such as primary and semi-manufactured goods; and proxies are used for heterogeneous commodities, particularly manufactured goods ready for final use. Several organizations provide the International Trade Division with proxies that are used as price relatives in the calculation of the Laspeyres and Paasche price indices. As Canada’s economy is very much interconnected with that of theUnited States, the United States Bureau of Labour Statistics producer price index is used as a proxy for the prices of some Canadian imports from theUnited States. For more detailed information onCanada’s practices, see annex XXVIII.B.
3. Practices of the Czech Republic
28.33. Current practices. The export and import price index has been calculated for theCzechRepublic since 1998. Prices are measured through a national monthly statistical questionnaire, with price relatives collected from both production enterprises and enterprises engaged in foreign trade only, about 580 of them engaged in exports and about 590 in imports. At present, the weight pattern includes approximately 2,050 exported and 2,100 imported products, raw materials and supplies, which take up a significant share in the value of external trade (both exports and imports). Foreign trade indices are calculated on the basis of invoiced prices (without duties, VAT and consumer tax) which are converted into national currency using the average monthly exchange rates declared by the National Bank. The price indices thus reflect changes in foreign-exchange rates. The breakdown of the export and import price index adjusted by exchange-rate influences has been published monthly since January 2011. The index is compiled according to the Harmonized System and is transformed to conform to the SITC, Rev.4, breakdown into its main groups and the Czech Republic Classification of Product by Activities (CZ-CPA 2008) , so as to meet the needs of national accounts and Eurostat. The first estimates are definitive, which means that ordinary revisions (revisions for the purpose of obtaining more precise previous estimates without methodological changes or modification of computation concepts) are not carried out.
4. Practices of Germany
28.34. Current practices: calculation of unit-value indices. The external trade price indices constitute an important module withinGermany’s overall system of price statistics, which is intended to measure the development of prices across all of the main stages of the economy in a methodologically consistent way. InGermany, (Laspeyres) volume indices and (Paasche) unit-value indices (UVIs) are calculated and published on a monthly basis in a breakdown by detailed commodities and country groups. The compilation is based on the results of external trade statistics. Even though price indices based on survey data are generally preferred for methodological reasons, UVIs are calculated as well, since they can be derived easily from already existing foreign trade figures and allow a breakdown by detailed commodities in combination with partner countries. In this way, a methodologically consistent set of nominal and real export/import figures is available which is based exclusively on foreign trade statistics. By means of volume indices and UVIs the nominal figures of foreign trade statistics can be split up into a quantity and a value component. Another advantage of UVIs is that they are calculated on the basis of up-to-date weighting factors (while calculation of price indices requires unchanging weighting factors of the relevant base year).
28.35. Compilation of price indices. In respect to the analysis of price changes in foreign trade, besides UVIs, “real” price indices are also calculated monthly as Laspeyres indices for a large number of commodity groups and by groups of countries. To a large extent, the data are obtained through an enterprise survey. Only in exceptional cases are other sources (e.g., stock market valuation, market reports) appropriate as providers of actual trade prices. The generation of the survey sample follows a multi-stage procedure. First, a basket of representative commodities is defined, which can be derived from foreign trade statistics. Second, the reporting enterprises are selected with regard to their share in the total value of the relevant commodities. Finally, the enterprises decide on specific “price representatives” (single cases of sale or purchase) for each of the selected commodities. As a result, the survey covers 6,000 enterprises reporting nearly 10,000 single prices per month. The advantages and disadvantages of a price survey have been described above. The main methodological issue is to keep the parameters underlying the calculation (goods and enterprises selected, sample of price representatives, weighting scheme) as constant as possible (according to the Laspeyres-concept).
5. Practices of Italy
28.36. Current practices. External trade UVIs inItaly are chained Fisher-type indices, with each monthly link calculated as the squared root of a Laspeyres- and Paasche-type index, both based on the previous year. Each Laspeyres and Paasche link is a weighted average of elementary unit value indices belonging to “product-country of origin/destination-flow” strata, the products being classified according to the Combined Nomenclature at the eight-digit level. The total number of elementary strata is about 220,000. Outlier detection procedures are applied to the log-distribution of the elementary unit values (levels) in each stratum, to control for the variability inside the strata, due mainly to errors on reported quantities. No imputation method is applied to replace the deleted observation so as not to introduce imputation bias in the resulting distributions. Deleted unit values are taken into account through maintaining their original weights in the calculation of the aggregated indices.
 The Statistics Canada Industrial Product Price Index (IPPI) and Automotive Price Index by Model, as well as data on exports of crude petroleum and natural gas (by pipeline), are derived from its Manufacturing, Construction, and Energy Division. Price indicators derived from other organizations include the Bank of Japan Export Price Index, price data on electricity from Canada’s National Energy Board, and the Computer Price Index by Component from the United States Bureau of Economic Analysis (BEA).
 The survey is not based on a random sample: the choice of enterprises is “targeted”. Hence, it is not possible to estimate sampling errors. Nevertheless, the reporting companies can be considered representative, since they are the leaders in the relevant market segment.