Annex 9.D. Conduct of reconciliation studies: the experience of Brazil
9.D.1. Reasons for divergences in bilateral trade. In the experience of Brazil, divergences in bilateral trade statistics can be caused by differences in coverage, different methods for the treatment of certain goods (e.g., military goods, ship’s stores and confidential data), value increases in intermediary countries, differences in the classification of goods, time lags in reporting, differences in valuation, including CIF/FOB differences, currency conversion, methods of partner-country attribution, and of trade through third country intermediaries. Although such divergences may be substantially reduced by the adoption of the concepts and definitions recommended in IMTS 2010, there are also divergences that may originate from variations in data sources, reporting errors, errors in data collection or in the processing and forwarding of results, the use of fraudulent documents or the inability of traders to furnish accurate information. Therefore, a certain amount of non-comparability will remain.
9.D.2. Establishment of a working group on statistics harmonization.Brazil has participated in several statistical harmonization groups within the framework of bilateral trade and investment negotiations, with the aim of improving the quality of statistical data for national purposes and to provide negotiators with trade figures that are accepted by both countries. In particular, statistical harmonization aims to discover the causes of the numerous discrepancies and to recommend corrective actions through the application of homogeneous and more compatible methodologies. When trade negotiations start,Brazil proposes the establishment of a working group on statistics harmonization, composed of specialists responsible for statistical production in both countries, in order to enable the exchange of data and the analysis of existing divergences. In some cases,Brazil also proposes the exchange of information on methods of data quality control and dissemination
9.D.3. Meetings and tasks of the working group on statistics harmonization. Normally, more than two meetings are necessary. In the first meeting, the experts from both countries establish the work methodology and define the data to be exchanged for further bilateral analysis. This meeting is also useful for the purpose of establishing agreement on the layout of the files,a the agenda and the schedule for exchanging data, so that statistical harmonization between the two countries may be achieved or so that the causes of differences may be identified and solutions proposed to reduce them. In the second meeting, the focus is on the analysis of the flows that present the biggest divergences. In the third meeting, the study concludes with the analysis of the other flows and the preparation of a document by both countries containing their conclusions.
9.D.4. Data exchange. Initially, data are exchanged for the previous three years and the available months of the current year. Later, the analysis is complemented by the remaining monthly data at the greatest possible level of detail. To facilitate the exchange of data, the Secretariat the Ministry of Development, Industry and Foreign Trade (MDIC/SECEX) creates mechanisms for the electronic transfer of data in American Standard Code for Information Interchange (ASCII) format according to an agreed layout through the File Transfer Protocol (FTP), granting access to the representatives of the partner country to upload and download files. After the initial exchange of files, each month’s data are exchanged by the twentieth day of the following month.
9.D.5. Characteristics of the work of the working groups on statistic harmonization. The work of the working group on statistic harmonization may be characterized as follows:
(a) Confidential data are not exchanged. The data exchanged are used only for the purposes of the working group on statistic harmonization and are not transferred to other entities or government agencies. Further publication of these data requires the prior written consent of the other party and an appropriate reference in the text of the publication;
(b) The parties exchange information on the methodology for processing adopted by each country. The statistical reconciliation demands the adoption of the same methodologies for the generation of data, with the exclusion of those aspects that, for operational reasons, cannot be adopted by the parties;
(c) If both parties adopt IMTS 2010, the concepts and definitions that may have more than one interpretation are prioritized for the bilateral analysis of data divergence. Indeed, some of the discrepancies result from different uses of these recommendations, based on different interpretations.
9.D.6. Steps to be taken during reconciliation. To identify data differences and their causes, the working group on statistic harmonization usually follows the steps below:
Step 1 - An initial comparison is made at the two-digit (chapter) level of HS, considering always more than one year.
Step 2 - After identifying the differences at the 2-digit level, the analysis is expanded to a greater level of detail (the four-, six-, or, if possible, eight-digit level), even if there are differences in the tables of commodity codes.
Step 3 - The working group then checks to determine whether the various differences within the same positions at the four-digit or six-digit levels compensate each other, in which case the problem would lie with the different classification of goods for the same group of products. Further, the working group checks to determine whether differences in tariffs may induce companies to classify products in codes that represent the payment of lower duties.
Step 4 - The working group also checks to determine whether differences compensate each other across years and months. In that case, the problem is the time of registration, mainly on imports under the special trade system. Many factors may contribute to differences in time of recording, including the characteristics of the products traded, time spent at port of entry, distance between countries and type of transport, as well as different administrative and customs procedures (see chap. XII for further details). The working group evaluates the distance between the two countries, the predominant type of product and the more frequently used type of transport. Averages of differences of about 4 per cent are considered normal.
Step 5 - The working group examines the persistent divergences to determine whether they are in terms of values or in terms of quantities. If the quantities are similar, there may be a problem of undervaluation or overvaluation of goods which then becomes a fiscal problem. In these specific cases, the working group checks the original records of each country relating to the same operation. It is highly important that the technicians who are undertaking the harmonization have unrestricted access to all the data of the operations of export and import in sufficient detail to enable them to make a clear identification of the cause of divergence. If the values are similar, but not the quantities, there may be differences related to the unit of statistical quantity adopted by each country. The use of different units for the same product generates significant distortions and is in most cases, an important reason for differences of quantities. This arises frequently in the case of errors in the classification of the same product.
Step 6 - Other factors to be considered by the working group are the effects of indirect trade and the criteria adopted for the registration of the partner country. In general, the values and quantities of imports where the partner country is not the country of origin must be excluded from the analysis. For instance, when the partner country is a member of the European Union, the “Rotterdam effect” may cause significant divergences.
Step 7 - When the data that are compared do not have the same valuation, the working group checks the characteristics of the goods sold and the average cost of freight and insurance. Those factors mean, in some cases, differences of over 100 per cent in the case of primary products (soybeans, iron ore, etc.). In average, the costs of freight and insurance, representing the difference between FOB and CIF of industrialized products, is between 8 and 10 per cent of the cost of goods. To conciliate them, it is necessary that both countries make available import and export data on an FOB basis, at the finest level of detail possible. InBrazil, the totality of foreign trade statistic is valued on an FOB basis.
9.D.7. There are other sources of divergences, such as coverage, trade system and method of currency conversion, which must be examined according to IMTS 2010. In this matter, special attention is given to transactions in products like energy, on-board consumption, airplanes, ships, etc., or to the case where countries have special customs regimes in their territory.
9.D.8. The most advanced working group study resulted in significant benefits in terms of the quality of statistics produced in both countries. Some suggestions for correction are implemented immediately by the authorities entrusted with statistical production, e.g., regarding issues related to coverage and valuation. Other issues are of a fiscal nature—for example, differences in goods classification and under-invoicing—and are not caused by methodological differences. In those cases, reports are sent to customs agencies identifying all problems found for each item in the customs nomenclature of the two countries, so that they can take appropriate action.
9.D.9. After the implementation of corrective measures, the data exchange and the statistical monitoring of bilateral trade are continued.
a The files always include the following variables: Reporting country, trade flow (exports or imports), reference year, reference month, partner country (last known country of destination for exports, country of origin for imports), country of consignment (for imports only), commodity code (HS at the six- or eight-digit level), supplementary quantity, supplementary quantity unit measurement code, net weight (in kilograms), FOB value (in United States dollars), insurance (in United States dollars, only for imports) and freight (in United States dollars, only for imports).