United States Department of COMMERCE NEWS Washington, D.C. 20230 ECONOMICS AND STATISTICS ADMINISTRATION ------------ Bureau of the Census This release contains sensitive economic data not to be released before 8:30 a.m. Monday, October 1, 2001 CB-01-161 Press Copy For information contact: Bureau of the Census: Kathy Puzzilla (301) 457-3251 MERCHANDISE TRADE RECONCILIATION UNITED STATES - MEXICO - CANADA 1998 - 1999 The Bureau of the Census announced today the results of a study into the differences between the official trade statistics released by the United States and Mexico, and Canada and Mexico during 1998 and 1999. This is the second report on the ongoing study being conducted jointly by the relevant agencies in the three countries. The first report, covering 1996 and 1997, was released in August of 2000. The statistics produced by countries on their merchandise trade with the rest of the world frequently differ from the statistics published by their trading partners. These differences reflect both legitimate conceptual differences between import and export statistics as well possible errors. The reconciliation study was undertaken to identify and quantify the causes of the differences. It is important to note that the results of this study do not constitute revisions to any country's official statistics; however, they do help improve the understanding of the trade statistics of the three countries. The significance of commercial relations among the United States, Mexico, and Canada can be gauged in part by the size of their trade flows. Roughly thirty percent of the total U.S. merchandise trade occurs with Canada and Mexico. North American trade represents about 80 percent of Mexico's total trade and a similar percent of Canada's total trade. Therefore, it is important to understand the reasons behind the differences between the partners' statistics. The participants in the project are: the U.S. Customs Service and the Foreign Trade Division of the U.S. Bureau of the Census; the International Trade Division of Statistics Canada; and the Mexican Working Group on Foreign Trade Statistics, which is comprised of the Secretariat of Treasury and Public Credit (SHCP), the Bank of Mexico, the Secretariat of the Economy (SE), and the National Institute of Statistics, Geography and Informatics (INEGI). The U.S. Bureau of Census, Statistics Canada and INEGI have co-operated, not only in developing the reconciliation estimates, but also in jointly preparing this report. Canada and the United States participate in a 'data exchange', in which the export statistics of each country are derived from the counterpart import data; therefore, there are no unexplained differences in their trade statistics. However, differences in the official trade statistics of the United States and Mexico, and of Canada and Mexico are sizeable. According to the trade statistics of the partner countries, Mexico's import statistics exceeded exceeded the United States' exports to Mexico by $14.5 billion in 1998 and $18.4 billion in 1999. On the other side of the coin, Mexico's exports to the United States exceeded the United States import from Mexico by $8.5 billion in 1998 and $10.7 billion in 1999. In the case of trade between Canada and Mexico, Mexico's import from Canada exceed Canada's export to Mexico by $1.3 billion in 1998 and $1.9 billion in 1999. On the other side of the coin, Mexico's exports to Canada fell below Canada's reported imports from Mexico by $3.6 billion in 1998 and $4.1 billion in 1999. I. United States - Mexico trade reconciliation As in 1996 and 1997, Mexico's trade statistics for 1998 and 1999 exceeded those of the United States for both directions of trade, with the southbound flow showing the larger discrepancy. Southbound Northbound U.S. exports, Mexican imports U.S. imports, Mexican exports In millions of U.S. dollars 1998 1999 1998 1999 Mexican statistics 93,252 105,267 103,109 120,393 U.S. statistics 78,772 86,909 94,629 109,720 Difference 14,480 18,358 8,480 10,673 The reconciliation study initially focused on southbound trade because it showed the greater discrepancy. The reconciliation study on southbound trade identified the following main sources of discrepancy: a) Geographic Coverage. The United States includes trade with Puerto Rico and the U.S. Virgin Islands in its merchandise trade statistics. In contrast, Mexico treats them as separate trading partners. This coverage difference resulted in a relatively minor difference in the trade statistics of the two countries, accounting for $157 and $186 million dollars in 1998 and 1999, respectively. b) Partner Country Attribution. The Mexican import entry form or pedimento only allows the reporting of one country of origin. When more than one country of origin is involved, as is common with summary declarations filed by the maquiladoras, the country accounting for the largest value is credited with the total value. Consequently, some imports are misattributed to the United States. Mexico's survey of maquiladoras estimated the value of imports misattributed to the United States at $6.0 billion in 1998 and $6.5 billion in 1999, accounting for 42 and 36 percent of the original differences, respectively. Note: This situation changed on May 21, 2001 with the introduction of the new entry form or pedimento that allows for the reporting of more than one country of origin. c) Nonfiling of U.S. exports. U.S. export regulations require the reporting of all export transactions valued over $2,500, some companies however, do not submit all the required declarations, especially companies exporting out of foreign trade zones. The reconciliation was not able to quantify the total effect of nonfiling on U.S. exports to Mexico; however, it could account for a substantial part of the remaining unreconciled difference between the southbound trade statistics of the two countries. d) Low value transactions. The United States does not include transactions valued below $ 2,501 in its export trade statistics instead an estimate of the total value of these transactions by country is included in the statistics. In contrast, Mexico records all trade transactions, regardless of value. The U.S.- Mexico reconciliation decreased the discrepancy in southbound trade by almost 49 percent in 1998 and 42 percent in 1999. Although no specific analysis of northbound trade was conducted, a portion of the discrepancy was explained by geographical coverage, classification and methodological criteria in the examination of the information. Reduction in the northbound discrepancy was more modest, roughly 16 percent in 1998 and 17 percent in 1999. II. Mexico - Canada trade reconciliation The 1998 and 1999 published trade statistics of Mexico and Canada also showed a significant difference, with the import statistics of each country exceeding counterpart exports. Southbound Northbound Canadian exports, Mexican imports Canadian imports, Mexican exports In millions of U.S. dollars 1998 1999 1998 1999 Mexican statistics 2,292 2,949 1,521 2,391 Canadian statistics 989 1,099 5,175 6,480 Difference 1,303 1,850 (3,654) (4,089) Source: Statistics Canada/INEGI Totals may not add due to rounding Negative numbers shown in brackets The reconciliation study identified indirect trade as the main source of discrepancy between Canada and Mexico in both directions of trade. The research has shown that record filers frequently treat the shipment as destined for the intermediate country rather than the country of final destination. For example, Canadian goods shipped to Mexico via the United States may be recorded as exports to the United States in the Canadian trade statistics, while in Mexican statistics these imports are recorded as originating in Canada. The estimation of this source of discrepancy was derived from Mexican import data. In 1999, this estimation explained 54.4 percent of the differences in the southbound trade statistics and 87.0 percent of the differences in the northbound trade statistics. III.. Steps to Improve the Quality of Trade Reporting United States: The United States has conducted extensive outreach and education efforts to improve the quality and thoroughness of trade reporting, with particular emphasis on trade with Mexico. In addition, during 1999 many companies began reporting through the new Automated Export System (AES), which eliminates the logistical problems of collecting paper documents, and includes on-line edits to ensure complete and accurate data. Educational efforts as well as increases in AES participation have continued into 2001. Mexico: In the case of Mexico, the results of the study have been taken into account in the design of a new Mexican Customs System . The new system will improve recording of the country of origin and destination for merchandise trade. This new system tackles one of the main sources of statistical discrepancy and remedies one of the main sources of statistical discrepancy between Mexican statistics and those of the United States and Canada. Canada: This study reinforces the findings of the previous reconciliation work undertaken for the 1996- 1997 data years. In both sets of studies, the major contributing factors to the discrepancies between the Canadian and Mexican published data are non-filing of export documents and differences in the way that indirect trade was reported. On Canada's part, considerable work has been undertaken to improve the reporting of exports to non-U.S. destinations. These activities involve not only Statistics Canada but also the Canada Customs and Revenue Agency (CCRA) and the U.S. Bureau of the Census. For example, electronic filing of export documents is being actively promoted to ease the burden of reporting such transactions, CCRA is undertaking various administrative measures to improve document filing and Statistics Canada and the USBC are jointly reviewing methodologies to provide better estimates of indirect trade. Finally, the utility of this study in identifying reporting errors has resulted in the agreement between all three countries to explore the possibility of undertaking further reconciliation work, with the expectation that more detailed work may be possible. IV. Reconciliation Results The following sections provide a more detailed presentation of the results of the study, as well as explanatory notes and a description of the methodology used in the reconciliation of the foreign trade statistics. TABLE 1 Reconciliation of 1998-1999 Merchandise Trade Statistics United States - Mexico Southbound Trade In millions of $U.S. Current Customs Basis 1998 1999 Mexican Published Imports 93,252 105,267 Differences attributed to: Geographic Coverage 157 186 Partner Country Attribution (6,029) (6,563) Low Value Transactions (1,113) (1,265) Electricity (47) (29) Residual differences (7,448) (10,687) U.S. Published Exports 78,772 86,909 Negative numbers are shown in brackets Totals may not add due to rounding Source: U.S. Bureau of the Census./INEGI UNITED STATES -MEXICO SOUTHBOUND EXPLANATORY NOTES Geographic Coverage: The United States includes the foreign trade of Puerto Rico and the U.S. Virgin Islands in its merchandise trade statistics. In contrast, Mexico treats them as separate trading partners. For comparison purposes, add Mexico's reported imports from Puerto Rico and the Virgin Islands to the Mexican imports from United States. Partner Country Attribution: The Mexican pedimento (customs document) only allows the reporting of one country of origin. Summarised pedimentos filed by the maquiladoras often combine merchandise with more than one country of origin. In these cases, the total value is attributed to the country accounting for the largest value. Consequently, some imports are misattributed to the United States. The working group on foreign trade statistics in Mexico surveyed import firms to determine the proportion of goods attributed to the Unites States that were in fact of other countries' origin. These proportions were used to obtain estimates of the annual value of misattributed trade and part of the results were added to Mexico's imports of Canadian goods. Low Value Transactions: The United States does not include transactions valued below U.S. $2,501 in its export trade statistics. Instead, an estimate of the total value of these transactions by country is included in the statistics. In contrast, Mexico tabulates all transactions, regardless of value. The value of the difference attributed to low value transactions was calculated from the difference between the sum of Mexican transactions valued less than $2,501 and the U.S. low-value transactions total estimate. This difference likely reflects both U.S. underestimation of low value transactions and differences in classification codes and reporting practices. Electricity: The United States does not include electricity in its U.S. and Mexico trade statistics. The value shown reflects the value of electricity imported from the United States obtained in Mexico's southbound statistics. Other: This is the remaining unexplained difference between U.S. and Mexican statistics. The sources of this discrepancy may include, among others, treatment of goods re-exported from U.S. foreign trade zones, additional differences in the attribution of partner country, over or under- estimations of the reconciliation estimates, revisions, and non-filling of U.S. export declarations. TABLE 2 Reconciliation of 1998-1999 Merchandise Trade Statistics United States - Mexico Northbound Trade In millions of $U.S. Current Customs Basis 1998 1999 U.S. Published Imports 94,629 109,721 Differences attributed to: Geographic Coverage (193) (195) Re-exports 1,955 2,516 Indirect Trade (421) (509) Electricity 3 2 Residual Difference 7,136 8,858 Mexican Published Exports 103,109 120,393 Negative numbers are shown in brackets Totals may not add due to rounding Source: U.S. Bureau of the Census/INEGI UNITED STATES - MEXICO NORTHBOUND EXPLANATORY NOTES Geographic Coverage: The United States includes the foreign trade of Puerto Rico and the U.S. Virgin Islands in its merchandise trade statistics but Mexico treats them as separate trading partners. They reduced Mexico's reported exports to Puerto Rico and the Virgin Islands from U.S. import totals for comparison purposes. Re-Exports: The U.S. import data are based on the principle of country of origin, whereas Mexico's export data include goods of national and foreign origin sold to the United States without any significant transformation. The U.S. customs documents show two countries, the country of origin and the country from which the goods were directly shipped to the United States. The value of these transactions showing Mexico as the country of shipment, but not of origin, were used for the estimate of re-export value. Indirect Trade: These are goods of Mexican origin shipped to the United States from other countries. This adjustment is based upon country of origin and country of shipment information reported by the U.S. importer. Electricity: Mexico includes the movement of electricity in its trade statistics and the United States does not. The value shown reflects the value of electricity imported from the United States obtained from Mexican northbound statistics. Other: This amount represents the remaining unexplained difference between U.S. and Mexican statistics. The sources of this discrepancy may include, among others: additional differences in valuation, attribution of partner country, over or under-estimations of the reconciliation estimates, or revisions. TABLE 3 Reconciliation of 1998-1999 Merchandise Trade Statistics Canada - Mexico Southbound Trade In millions of $U.S. Current Customs basis 1998 1999 Mexican Published Imports 2,292 2,949 Differences attributed to: Re-Exports 44 71 Low Value Transactions (43) (46) Indirect Trade: Via United States (716) (940) Via Other Countries (38) (66) Partner Country Attribution 271 189 Residual difference (821) (1,058) Canadian Published Exports 989 1,099 Source: Statistics Canada, INEGI Total may not add due to rounding Negative numbers are shown in brackets CANADA-MEXICO SOUTHBOUND EXPLANATORY NOTES Re-Exports: Mexico's import data are based on the country of origin principle, in which only goods grown, extracted or produced in Canada are considered as merchandise trade with Canada. However, the published Canadian export information includes both goods of Canadian origin and goods of foreign origin that have entered Canadian consumption and have been subsequently sold to Mexico without any substantial transformation occurring in Canada. This type of transaction is recorded separately in Canadian statistics; consequently, the amount attributed to this difference has been obtained from Canadian export data. Low Value Transactions: Canadian exports to Mexico are not required to be reported if the total value of the shipment is less than $2,000 (Canadian funds). However, the Mexican import statistics include all transactions regardless of value. This estimate was obtained from Mexico's import data, which allows identification of the total value of small transactions. Indirect Trade: Frequently, Canadian exporters at the time of shipment are not aware of the final country of destination or consumption of goods shipped from Canada. Research has shown that there is a tendency to record the intermediate country(ies) involved in the transportation of goods to the final destination as the country of final destination. For example, in the case of Canadian goods moving through the United States en route to Mexico, the goods are frequently not recorded as Canadian exports to Mexico because of the perception that the United States is the final destination. Therefore, the shipments continuing to Mexico cause a discrepancy between Canadian and Mexican statistics because Mexico records them as imports originating in Canada while Canada does not record them as exports to Mexico. This estimated adjustment was derived from Mexican import data. Mexico records both the country of origin and the country of seller (i.e. the country in which the goods were invoiced). The comparison of these two data sets permitted the derivation of the amounts that, in all probability, have been recorded as Canada/other country trade rather than Canada/Mexico transactions. Partner Country Attribution: As previously noted, during the study period, Mexico's published import data were based on the principal country of origin concept. The Mexican Pedimento (Customs document) did not provide for reporting imports from more than one country. This caused some goods from other countries to be included in Mexico's reported imports from the United States. The Working Group for Foreign Trade Statistics of Mexico surveyed import firms to determine the proportion of goods attributed to the United States that were in fact of Canadian origin. These proportions were used to obtain estimates of the annual value of misattributed trade and the results were added to Mexico's imports of Canadian goods. Other: This is the remaining unexplained difference between Canadian and Mexican statistics. The sources of this discrepancy may include, among others: additional differences in valuation, attribution of partner country, over or under-estimations of the reconciliation estimates, revisions or the non-filing of Canadian export declarations. TABLE 4 Reconciliation of 1998-1999 Merchandise Trade Statistics Canada - Mexico Northbound Trade In millions of $U.S. Current Customs Basis 1998 1999 Canadian Published Imports 5,175 6,480 Differences attributed to: Mexican Re-Exports: Canada Origin 15 15 United States Origin 8 15 Other Country Origin 54 27 Indirect Trade: Via United States (2,848) (3,491) Via Other Countries (5) (47) Timing Residual differences (878) (608) Mexican Published Exports 1,521 2,391 Source: Statistics Canada/INEGI Totals may not add due to rounding Negative numbers are shown in brackets CANADA-MEXICO NORTHBOUND EXPLANATORY NOTES Re-Exports: Canadian import data are based on the country of origin principle. Mexican published export trade data includes both goods of domestic origin and goods of foreign origin sold to Canada without any substantial transformation occurring in Mexico. Canadian import documentation records two countries, the country of origin and the country from which the goods were shipped directly to Canada. The values for transactions that showed Mexico as the country of shipment, but not of origin, were totaled and used as the basis for the re-export estimate. Indirect Trade: This estimate is based on Canadian import data using the country of origin/shipment information. It reflects the value of the goods of Mexican origin that were shipped directly to Canada from a country other than Mexico. That value is subtracted from the Canadian import data for comparison purposes. Timing: This estimate represents shipments of crude oil that were recorded in Mexican export statistics in one period but recorded in Canadian import data in another period. Other: This is the remaining unexplained difference between Canadian and Mexican statistics. The sources of this discrepancy may include, among others: additional differences in valuation, the attribution of partner country, over or under-estimations of the reconciliation estimates, and any revisions made during the reconciliation study period.