From the July 2000 SURVEY OF CURRENT BUSINESS: U.S. International Transactions Revised Estimates for 1982-99

By Christopher L. Bach

AS IS customary each June, the estimates of U.S. international transactions have been revised to incorporate statistical and methodological revisions. This year, like last year, a number of improvements have been implemented as part of continuing efforts by the Bureau of Economic Analysis (BEA) to address gaps in coverage of transactions. In large part, the gaps have arisen because of the dynamic nature of international markets. The major improvements this year respond to rapid changes in both the capital markets and services markets.

Net U.S. purchases of foreign securities are revised for 1995-99 as a result of a more complete accounting for large-scale foreign acquisitions of U.S. companies. Net U.S. purchases of foreign securities are also revised for 1995-99 to account for other transactions that are not completely captured by the statistical reporting system.
"Other" private income receipts are revised for 1998-99 to incorporate the final results of the U.S. Treasury Department's Benchmark Survey of U.S. Portfolio Investment Abroad as of December 31, 1997. Previously, only preliminary results were available. "Other" private income receipts are also revised to reflect the previously mentioned changes made to net U.S. purchases of foreign securities.
"Other" private services receipts and payments are revised for 1997-99 to reflect revisions to financial services receipts and payments. "Other" private services receipts are also revised as follows: For 1986-99, to incorporate improved estimates of expenditures of international organizations in the United States; for 1996-99, to incorporate improved estimates of expenditures of foreign embassies in the United States; and for 1986-99, to incorporate newly developed estimates of expenditures of temporary nonagricultural workers in the United States.
Direct investment income and capital flows are revised for 1982-99 to reflect revised estimates of the current-cost adjustment. Revised estimates of prices for equipment and structures are now incorporated into the current-cost adjustment.

The newly available benchmark data, improved methodologies, and improved coverage of the accounts are discussed in the remaining sections of this article. In addition to these major changes, revisions also result from the incorporation of regularly available data from BEA's annual and quarterly surveys, from the U.S. Treasury Department's and Federal Reserve System's quarterly and monthly surveys, and supplemental data from other U.S. Government agencies and private sources. For 1999, as a result of all the changes, the current-account deficit is reduced $7.4 billion, to $331.5 billion (table 1). By account, $1.3 billion is added to goods exports and $0.2 billion is removed from goods imports, resulting in a deficit that is $1.6 billion lower than previously estimated. For services, $5.2 billion is removed from services exports and $6.2 billion is removed from services imports, resulting in a surplus that is $1.0 billion higher than previously estimated. For income, $2.3 billion is added to income receipts and $4.0 billion is removed from income payments, resulting in a deficit that is $6.3 billion lower than previously estimated. For net current unilateral transfers, $1.4 billion in outflows is added, resulting in an increase in net transfers of the same amount. Net financial account inflows were revised down $54.8 billion, to $323.4 billion. Details on revisions to individual series are shown in table 2 at the end of the article.

Foreign securities

Additional offsets to direct investment.--Estimates of U.S. transactions in foreign securities are adjusted to account more completely for large-scale acquisitions that have occurred over the past 5 years.

The treatment in the international accounts of foreign acquisitions of U.S. companies can involve entries in the direct investment, foreign securities (portfolio investment), and banking accounts, depending on the type of financing.

For acquisitions financed by an exchange of stock, the amount of the acquisition is entered as a financial inflow in the foreign direct investment in the United States account. This amount is probably captured completely and valued correctly in the direct investment statistical reporting system. However, net U.S. purchases of foreign securities in the Treasury International Capital (TIC) portfolio investment reporting system, which records the contra- or offsetting entry, often does not effectively capture the receipt by U.S. investors of stock in a foreign company in exchange for stock in a domestic company, because this exchange of securities does not normally go through the TIC reporting system. Consequently, when BEA can confirm that an exchange of stock has occurred and that net U.S. purchases of foreign securities are underreported, it has adjusted its estimates to assure more complete coverage of securities transactions. BEA makes these adjustments for some medium-size transactions and for large-size transactions. These additional securities offsets have been entered in the appropriate periods for 1995- 99--a period in which foreign acquisitions were prevalent and in which their number and size reached new levels: $7.4 billion was added for 1995, $10.8 billion for 1996, $3.0 billion for 1997, $10.5 billion for 1998, and $8.9 billion for 1999.

For acquisitions financed either partly or entirely with cash, the cash portion of the acquisition is included in the U.S. bank-reported accounts, typically as a reduction in foreign-held dollar deposits, and it is believed to be completely captured and correctly valued.

Other adjustments for undercoverage.--Estimates of U.S. transactions in foreign securities are also adjusted to account for other sources of incompleteness. Coverage problems were partly confirmed by information from the U.S. Treasury Department's Benchmark Survey of U.S. Portfolio Investment Abroad at yearend 1997, which indicated a 20-percent discrepancy between BEA's position estimates, based on accumulations and revaluations of transactions in TIC reports, and the benchmark survey results. The discrepancy had arisen in the 3 years since the previous benchmark survey for yearend 1994.

When BEA adjusted its international investment position estimates last year using preliminary benchmark results, it attributed all of the discrepancy to valuation changes and none to the less than complete coverage of transactions, based on its experience in presenting the accounts up until that time. BEA is now changing that practice and attributing a large part of the discrepancy to transactions. This change permits the international transactions accounts to more accurately reflect several major changes that were taking place in financial markets: Strong growth in direct transactions between U.S. and foreign residents that was not captured by a U.S. reporting system based primarily on recording transactions between financial intermediaries; a new emphasis on global investing by mutual and pension funds that may not have conducted transactions through financial intermediaries; and greatly improved electronic communications networks that permitted direct transactions with foreign institutions at lowered transactions costs. The adjustment for undercoverage of transactions is assumed to rise in each of the 3 years between the benchmark surveys. Net transactions in foreign stocks and in foreign bonds are each adjusted proportionately. The amount of adjustment for additional net U.S. purchases in foreign securities is $15.0 billion for 1995, $23.2 billion for 1996, and $26.8 billion for 1997. Adjustments for subsequent years will be held constant at the 1997 level until the next outbound benchmark portfolio survey reveals the extent of undercoverage. The next benchmark survey is planned for yearend 2001.

BEA is working with the Federal Reserve System and the U.S. Treasury Department to improve the coverage of the quarterly TIC reports.

Related income receipts.--A result of the changes for offsets to direct investment and for other undercoverage is to boost related dividend and interest income receipts. Dividend receipts are raised $0.8 billion for 1998 and $2.4 billion for 1999, and interest income receipts are raised $0.8 billion for 1998 and $1.6 billion for 1999. No revisions are made for 1995-97, because the position estimates and related income estimates had already been adjusted by the preliminary benchmark survey results last year.

Other private income receipts

The final results of the U.S. Treasury Department's benchmark survey of U.S. holdings of foreign securities at yearend 1997 showed a slight increase in U.S. holdings from the preliminary results introduced into the accounts last year. In addition, the survey's dividend and interest data available this year showed lower average yields than had been used by BEA to estimate income.

Dividend rates on foreign stocks obtained from the benchmark survey at yearend 1997 were lower than those included in BEA's accounts. The difference is mostly attributable to the geographic composition of U.S. holdings. The benchmark data showed higher holdings in emerging market countries; many dividend rates in these countries were especially low and lower than those used in BEA's accounts. When BEA's market-based average dividend rate was adjusted downward on the basis of the new, lower Treasury benchmark dividend rate, the result was to lower dividend income receipts $5.0 billion for 1998 and $3.6 billion for 1999.

Interest yields on foreign bonds on average were close to those included in BEA's accounts at yearend 1997. A minor adjustment to average interest yields in BEA's accounts was necessary to account for a slightly lower proportion of dollar-denominated bonds and for a slightly higher proportion of foreign currency-denominated bonds. When BEA's market-based average interest yield was adjusted on the basis of the new Treasury benchmark interest yield, the result was to raise interest receipts $1.0 billion for 1998 and to lower receipts $1.5 billion for 1999.

Revisions are not made for 1995-97, because the position estimates and related income estimates had already been adjusted by the preliminary benchmark survey results last year.

Financial services

This year, BEA completed its quinquennial Benchmark Survey of Financial Services Transactions Between U.S. Financial Services Providers and Unaffiliated Foreign Persons for 1999. The benchmark survey is more comprehensive than BEA's annual surveys of financial services transactions and included a refinement in the categories of cross-border services reported by U.S. financial services providers. The current major categories are brokerage commissions, underwriting and private placement fees, financial management fees, credit-related fees, credit card services, financial advisory and custody services, securities lending fees, electronic funds transfer charges, and all other financial services.

The benchmark survey results led to significant changes for brokerage commissions and for underwriting and for private placement fees. The revisions to brokerage commissions were larger than revisions to underwriting fees, and for both, the revisions were larger for receipts than for payments. Both receipts and payments were revised down.

Based on the benchmark survey results, BEA is now assuming a more rapid decline than previously in the explicit fees and commissions charged on brokerage transactions and, to a smaller extent, on underwriting transactions. These declines are directly attributable to increased competition among financial institutions, mergers among large-scale financial services providers, improved telecommunications networks, greater ease in conducting transactions on foreign exchanges, and technological advances that have dramatically lowered unit transactions costs in recent years. Conversations with industry participants confirm the new survey results. In addition, an increasing portion of securities transactions may be occurring through affiliated companies, which would result in a lower level of financial services transactions with unaffiliated foreign persons. Therefore, the estimates of commissions and underwriting fees are reduced for 1997-99 to bring them in line with current developments in the financial services industry. Receipts were reduced $1.3 billion for 1997, $2.4 billion for 1998, and $1.8 billion for 1999. Payments were reduced $0.2 billion for 1997, $0.2 billion for 1998, and $0.3 billion for 1999.

Other services

Several changes are introduced for "other" private services receipts.

Foreign embassies and international organizations.--Improved estimates for noncompensation expenditures of foreign embassies and consulates and for international organizations in the United States are introduced. Previously, these estimates were included only implicitly as part of total receipts; now, the estimation techniques have been improved. As extraterritorial entities located in the United States, both foreign embassies and consulates and international organizations incur noncompensation expenses in the U.S. economy.

For foreign embassies and consulates in Washington, DC, source data on noncompensation expenditures in the United States (such as expend-itures for office supplies, contractual services, equipment, rent, utilities, and food) were used to calculate an average noncompensation expenditure per employee. This average expenditure was multiplied by the total number of personnel employed by all embassies and consulates in the United States, as supplied by the Department of State, to derive total noncompensation expenditures of foreign embassies and consulates. Revisions are for 1996-99. For international or-ganizations, a similar approach was used, based on data for noncompensation expenditures and the number of personnel provided by the organizations themselves. Revisions are for 1986-99.

An additional aspect of the improved estimates is that they now include a measure of the spending of foreign employees of the foreign embassies and consulates and of foreign employees of international organizations in the United States. These expenditures are estimated as a share of foreign employees' total earnings. Estimates are for 1986-99.

Expenditures of temporary nonagricultural workers.--In last year's annual revision, a new measure of the earnings of temporary undocumented nonagricultural workers in the United States was introduced.⁽¹⁾

The estimate of total compensation was calculated as the number of such workers multiplied by annual hours worked and an average hourly wage, based on data obtained from several sources. This year, based on the same data sources, estimates of these workers' expenditures in the United States are introduced and entered as receipts in the "other" private services account. Estimates are for 1986-99. For 1999, receipts are raised $1.6 billion.

Current-cost adjustment

The current-cost adjustment to direct investment income and capital has been revised to reflect revised estimates of economic depreciation and updated source data for historical-cost depreciation, depletion, and expensed exploration and development expenditures reported by direct investment affiliates. (The current-cost adjustment consists of (1) the difference between historical-cost economic depreciation, which is computed using consistent service lives and prices of the current period, and depreciation reported by direct investment affiliates using financial accounting principles, and (2) adjustments to reported earnings for charges taken by direct investment affiliates for depletion and for expensed exploration and development expenditures.)

The revised estimates of economic depreciation reflect revised prices for equipment and structures investment in the United States, based on estimates incorporated in the 1999 comprehensive revision of BEA's national income and product accounts (NIPA's), and revised prices for equipment and structures investment in foreign countries. U.S. prices are revised for all years, but the largest revisions are for recent years. Foreign prices are revised only for recent years.

Additional revisions to economic depreciation reflect revised investment data reported by direct investment affiliates and revised assumptions about the relationship between equipment and structures that are used to compute separate estimates of equipment and structures. For 1999, revisions to the current-cost adjustment for U.S. direct investment abroad raised income receipts $1.0 billion, and revisions for the current-cost adjustment for foreign direct investment in the United States reduced income payments $0.6 billion. Offsetting entries were made in the direct investment capital accounts.

This work extends the significant improvements to the current-cost adjustment estimates that were introduced last year. This year's revisions are for 1982, the first year for which current-cost adjustments are included in the accounts, through 1999.⁽²⁾

Taxes

Taxes received from nonresidents by the U.S. Government are revised for 1997-99 to incorporate updated source data from the Internal Revenue Service (IRS), and for 1982-99 to include an additional component that was introduced in the 1999 comprehensive revision of the NIPA's.

Previously, tax receipts from nonresidents were estimated on the basis of IRS-reported receipts from nonresident aliens only; these receipts were entered in the unilateral current transfers account as offsets to corresponding entries in the services and income accounts. Now, the methodology is adjusted to account for taxes received from nonresident U.S. citizens. This adjustment is made to maintain consistency with the NIPA's. For 1999, the revision raised U.S. tax receipts $2.2 billion.

Panama Canal

The U.S. Government's assets in the Panama Canal Commission have been revalued to reflect prices of the current period. The revaluation affects the transaction value of the transfer of the U.S. assets to the Republic of Panama in the fourth quarter of 1999 in the U.S. international transaction accounts. The revaluation also affects the value of the assets in the U.S. international investment position from October 1, 1979, when the Panama Canal Commission was created, to December 31, 1999 (at noon), when the United States last owned the assets.

The net stock of fixed assets on the Panama Canal Commission's balance sheet is revalued from historical cost to current cost. The current-cost net stock is constructed using a perpetual inventory method. This method is consistent with the method that BEA uses to estimate the current-cost value of the net stock of fixed assets and consumer durable goods in its domestic wealth estimates and the current-cost direct investment positions in its international accounts.⁽³⁾ In the perpetual inventory method, each year's capital investment is first deflated from historical cost to constant cost using capital goods investment price indexes. The constant-cost net capital stock for a given year is then calculated as the cumulative value of past investment less the cumulative value of past depreciation and discards. The constant-cost capital stock is then reflated to current cost using capital goods investment price indexes.

The data required to construct the current-cost value of the net stock of fixed assets on the Panama Canal Commission's balance sheet were assembled from various sources and, in cases in which sources were less than fully adequate, were derived using assumptions based on BEA's experience in its other capital stock work. Investment data were derived from accounting statements of the Panama Canal Company, the Panama Canal Commission, and the Budget of the United States. Service lives were based largely on rates of depreciation implied by accounting statements. A price index for capital investment in canals does not exist, so BEA chose a NIPA index that is used to deflate investment in certain heavy construction.

"Residual" seasonality

BEA and the Bureau of the Census seasonally adjust estimates of goods exports and goods imports at the five-digit end-use commodity category level, which is the most detailed level of end-use classification available. Aggregate goods series--total exports, total imports, and all major end-use categories--are derived as the sum of detailed seasonally adjusted series. Differences between directly adjusted aggregate series and corresponding series that are derived indirectly as the sum of individually seasonally adjusted series are sometimes called "residual" seasonality.

This year, building on the gains in reducing "residual" seasonality in recent years, BEA and the Bureau of the Census applied adjustments for trading-day variation at the five-digit level, which is the same level at which seasonal adjustments are applied. The change allows a consistent application of trading-day and seasonal factors to the most detailed level of unadjusted data available. As a result, "residual" seasonality was reduced significantly for exports and changed little for imports.

It was possible to make this change this year because of the adoption last year of a regression methodology to calculate trading-day factors. The regression method is better able to distinguish irregular movements from trading-day variation than the old multiplicative method. Therefore, it is no longer necessary to apply trading-day factors at the three-digit level. The old procedure was also found to contribute to more, rather than less, "residual" variation (1) because it sometimes forced a combination of series that were unrelated by nature of product, (2) because it sometimes forced the combination of series that had offsetting trading-day patterns, and (3) because trading-day factors for large series were sometimes applied to many smaller series that when adjusted individually, showed no trading-day variation. Development and testing for trading-day variation at the five-digit level should remove the problems sometimes encountered by grouping series to a three-digit level.

The process that BEA and the Bureau of the Census use to develop seasonally adjusted estimates is complex and detailed. Nearly 300 series are tested for seasonality and trading-day variation each year. Tests are conducted using the X-12 ARIMA program, which provides the diagnostic measures used in making both seasonal adjustment and trading-day adjustment decisions. Currently, based on diagnostics developed for each individual series, 95 percent of the value of total exports and 97 percent of the value of total imports receive at least one type of adjustment. About 115 of 144 export series and 127 of the 149 import series receive at least one type of adjustment. Revisions are made for 1997-99.

Tables 2and 2b follow.

Acknowledgments

The revised estimates were prepared under the general direction of Anthony DiLullo, with the assistance of Paul Farello and other staff of the Balance of Payments Division. Harlan King prepared the new estimates of foreign securities and related changes to "other" private income; Russell Scholl and Jane Newstedt, the new estimates of financial services; Michael Mann and Vicki Schepker, the new estimates of other services; Steve Baldwin, the new estimates of expenditures of temporary nonagricultural workers; Douglas Weinberg, the new estimates of the current-cost adjustment; D. Timothy Dobbs, William McCormick, and Doug-las Weinberg, the new estimates for the revalued assets of the Panama Canal Commission; and Kwok Lee and the staff of the Goods Branch, the revised estimates of goods.

The revisions to the estimates of U.S. direct investment abroad were prepared under the supervision of Patricia Walker of the International Investment Division (IID), the revisions to the estimates of foreign direct investment in the United States were prepared under the supervision of Gregory Fouch of IID, and the revisions of several of the estimates of unaffiliated private services were prepared under the supervision of Christopher Emond of IID.

Special assistance was provided by William L. Griever of the Federal Reserve Board, who conducted the benchmark survey of U.S. portfolio investment abroad, and by Diane Oberg and the staff of the Bureau of the Census' Foreign Trade Division, who conducted the study of "residual" seasonality of goods.>

See Christopher L. Bach, "U.S. International Transactions, Revised Estimates for 1982-98," Survey of Current Business 79 (July 1999): 70-71.
See Bach, "Revised Estimates for 1982-98," 65-67.
For a discussion of the methods used to derive net stocks, see U.S. Department of Commerce, Bureau of Economic Analysis, Fixed Reproducible Tangible Wealth of the United States, 1925-94 (Washington, DC: U.S. Government Printing Office, August 1999): M-1--M-36.