between residents of that country and residents of foreign countries and since, also by definition, international payments include changes in the short-term foreign assets and liabilities of the country, it follows that aggregate receipts from foreigners and aggregate payments to foreigners are equal. For any cash claims upon foreign countries not taken up by those with payments to make in foreign countries (an apparent excess of credits 14) would appear automatically as a net addition to short-term banking assets in foreign countries (or as a net reduction in short-term banking liabilities to foreigners) that is, as an exactly offsetting debit. Conversely, cash claims in favor of foreigners not matched by cash claims against foreigners (an apparent excess of debits 15) would appear as a net addition to short-term liabilities to foreigners (or as a net deduction from short-term assets abroad)—that is, as a counterbalancing credit. In spite of this evident fact, the terms "favorable" and "unfavorable" or "adverse," which are permissible if used in an objective sense to identify an excess of receipts or of payments on current account, are still applied incorrectly to a country's balance of payments as a whole. Unless conventional definitions and concepts are discarded, these adjectives should be dropped permanently from balance-of-payments terminology.16

The net difference between the sum of recorded and estimated credit transactions and the sum of recorded and estimated debit transactions is labelled "Residual item," and the sign affixed is that of the smaller aggregate. A more complete designation would be "Net discrepancy as a result of errors, omissions, duplications, unestimated items, and unreported transactions." Aside from indicating that the limitations of available statistical data make impossible the identification and accurate estimation of all balance-of-payments transactions, the residual item is without significance.

For the past 4 years, the residual item in the balance of international payments of the United States has been a large but variable credit. That is, the sum of the identified debit transactions has exceeded in each year the sum of the identified credit transactions, including the reported net inflow of capital funds. During the same period, the balance of payments of the country has been influenced almost continuously by heavy inward movements of gold. Throughout the 4 years, however, these extraordinarily large gold importations were almost entirely unrelated to the position of the country on trade and service account, but were rather the counterpart of a steady transfer of capital from foreign countries to the United States. This inflow consisted of inward movements both of foreign-owned and of American-owned capital both in short-term banking funds and in security transactions and was reflected in annual net credits on capital account in the balance of payments. International operations of this character are exceedingly elusive, and their accurate measurement involves

14 Or, in other words, an excess of receipts from the sale to foreigners of merchandise, services, gold, silver, currency, and evidences of indebtedness. 15 That is, an excess of payments to foreigners for merchandise, services, gold, silver, currency, and evidences of indebtedness. 16 This is not to deny that shifts in the relative magnitudes of the various items among the international transactions of a country or a change in the international short-term investment position of a country in the direction of a smaller excess of assets or of a larger excess of liabilities may be considered undesirable on economic grounds or from the point of view of public policy, while other shifts or a change in the opposite sense may be deemed desirable. Nevertheless, the terms "favorable" and "unfavorable" carry inevitably confusing mercantilistic connotations and, even in the case of trade and service items, should be abandoned in favor of the purely objective terms, "net credits (or receipts)" and "net debits (or payments)."

considerable difficulty. The countervailing movements of gold, on the other hand, which appeared as annual debits in the balance-ofpayments statement, were recorded with a high degree of accuracy (except for earmarking operations for the account of exchange stabilization funds).

At the same time and especially during the early part of this recent period, the system of exchange controls, blocked currencies (the rates for which have departed widely from stated or nominal parities), and other factors affecting the foreign exchanges accentuated the difficulties of measuring or estimating trade and service transactions in the international balance.

In 1934, the unmeasured or incorrectly estimated elements in the balance of payments of the United States aggregated, on the basis of revised figures, approximately $482,000,000. It was assumed in the contemporary analysis of that discrepancy that it comprised four principal types of transactions, as follows:

1. The aggregate of such factors as repatriated balances of American corporations, unrecorded currency exports, and short-term capital items representing the counterpart of the possible importation of gold which was held under earmark abroad by governmental agencies at the end of 1933 and which had therefore been carried as a gold "import" in the 1933 balance.

2. Repatriation of foreign dollar bonds and sales in foreign markets of American securites by methods which precluded the recording of their value in the questionnaire returns.

3. Inflow of private funds (balances in foreign banks and/or hoarded abroad) expatriated roughly between April 1933 and the end of January 1934.

4. Unrecorded foreign-owned funds sent to the United States after devaluation, as a result of political uncertainty and economic instability abroad.

At the end of 1934, the Treasury Department began the collection and compilation of statistics of capital movements between the United States and foreign countries upon the basis of reports from practically all banks and brokers involved in international capital transactions. At the same time, a number of special studies, designed to extend and improve the statistical information relating to several of the more important balance-of-payments items, were inaugurated, and a method was devised for estimating unrecorded exports of paper currency. The residual items in the following 2 years, representing in each case the net credit required to balance an excess of recorded or estimated debits, were, on the basis of revised data, $354,000,000 in 1935 and $170,000,000 in 1936.

In 1937, however, the excess of identified payments to foreigners over identified receipts from foreigners, including unreported stabilization-fund operations, rose to $676,000,000. For the year as a whole, and particularly in the last quarter, the estimates of service items and the reported merchandise, gold, silver, and capital transactions were offset by this comparatively large balance of "Other transactions and residual," which included, in addition to possible errors and omissions in the estimated items, unreported stabilization-fund operations 17 and other transactions not exactly reflected for balance-of-payments

17 Since the active portion of the stabilization fund is limited to $200,000,000, fund operations would not affect the size of the residual either way by more than this amount.

purposes in the reported figures. Data covering the fourth quarter of the year indicate specifically that there was a substantial loss of gold through unreported gold earmarking operations during the final months of 1937. These operations, if ascertainable, would have the effect of reducing the stated discrepancy in some degree, but the residual of unidentified credit transactions would still be relatively large.

It is recognized that the possible errors, from a balance-of-payments point of view, in the trade and service items may be large individually and perhaps substantial on balance. However, the fact that the basic data and methods used in the preparation of estimates of transactions on current account are identical in successive years suggests that significant changes in the residual item from year to year cannot be accounted for on this basis. Even if the degree and direction of the error were constant, an increase in the total turnover on commodity and service account of approximately 25 percent from 1936 to 1937 would account for no more than a proportionate rise in the net discrepancy. But this amount would be comparatively small as compared with the actual increase; and the decrease in the residual item in 1935 and 1936, during which turnover on current account also increased, could not be explained on the same grounds.

The progress of studies designed to reduce the margin of error in current items is set forth below.

MERCHANDISE ADJUSTMENTS

In its annual bulletin on the balance of international payments the Finance Division has from time to time called attention to various factors which tend to produce substantial differences between the aggregate recorded foreign-trade values and the respective dollar receipts or payments which enter into the balance of international payments. Certain important adjustments involve primarily the addition to the official foreign-trade figures of values involved in transactions which are essentially of a "commodity" nature but not reported in the merchandise export and import statistics. Examples of such transactions are (1) the sale of bunker fuel in the United States to foreign. vessels and corresponding purchases abroad by United States vessels, (2) the international sale and purchase of private yachts and other vessels, (3) unrecorded values of parcel-post exports and imports, and (4) transactions conducted with a view to avoiding payment of duties, penalties, and other charges or to evade customs requirements or other legal obligations. A second class of "corrections" relates chiefly to such factors as (1) terms of sale, (2) possible overstatement or understatement of invoice values, (3) nominal price bases in invoice values, (4) arbitrary import entries (largely in case of imports subject to specific duties or on nondutiable list), (5) inclusion of nondutiable charges in entries (again chiefly in case of imports either duty-free or subject to specific rates), and (6) additions to invoice values by importer to avoid penalty duties or to place dutiable value on basis required by special provisions of existing tariff legislation.

With a view to compiling the basic data essential to a satisfactory evaluation of the second class of adjustments the Finance Division of the Bureau of Foreign and Domestic Commerce has sponsored a W. P. A. project which was inaugurated early in 1938 through the

Works Progress Administration of the City of New York. It is believed that two definite purposes, each related to the balance of international payments, will be effectively served by the results of this project. In the first place, the basic data, compiled according to a schedule which requires the examination of invoices and entries covering more than 70 of the principal import classes, will make possible a determination of actual statistical values on a balance-of-payments basis. Secondly, it is believed that the collected statistics will, in a large number of cases, provide valuable supplementary data relating to the inter-company and intra-company accounts of corporations whose direct investments abroad, as well as the returns therefrom, influence the balance-of-payments entries, not only in the merchandise trade items but also in the interest and dividend categories.

AMERICAN DIRECT INVESTMENTS IN FOREIGN COUNTRIES

A survey of American direct investments in foreign countries, representing the first exhaustive reappraisal of this type of foreign holdings since 1930, was issued in 1938 as Economic Series No. 1, American Direct Investments in Foreign Countries-1936. This study, prepared in the Finance Division under the direction of Dr. Paul D. Dickens, was the third of a series of special researches designed to improve the available data relating to the international financial position of the United States.

FOREIGN INVESTMENTS IN THE UNITED STATES

The detailed results of a comprehensive survey of foreign investments in the United States, with the basic estimates carried forward to the end of 1936, were published by the Finance Division in 1937 in a separate bulletin-Foreign Investments in the United States. This survey was undertaken primarily as a means of establishing a more satisfactory basis for estimating interest and dividend payments by the United States to foreign holders of domestic securities. It constitutes, nevertheless, an important contribution to the general history of foreign investments in this country and bears directly upon contemporary international financial problems.

INTERNATIONAL INSURANCE TRANSACTIONS

A special study of the balance-of-payments aspects of international insurance transactions was issued in 1936 as Trade Information Bulletin No. 834 under the title "Insurance Transactions in the Balance of International Payments of the United States, 1919-35." This study, prepared by Dr. August Maffry of the Finance Division staff, provided the basis also for the discussion of the investment aspects of the operations of foreign insurance companies in this country which appeared in Foreign Investments in the United States (pp. 38-40 and 43). Current data relating to international insurance transactions are reported in this bulletin (see section, "Miscellaneous Service Transactions").

STUDY OF CURRENCY MOVEMENTS AND RELATED ITEMS

In 1936 the Finance Division reexamined the significance of paper currency movements for the balance of international payments of the

United States and attempted to measure the unrecorded export of dollar notes. The method tentatively adopted for estimating exports of United States paper currency through nonbanking and nonreporting channels was described briefly in The Balance of International Payments of the United States in 1935 (pp. 44-48), and current estimates were reached by the use of the same technique.

Since unrecorded currency exports consist primarily of notes carried out of the country by travelers and sent abroad as gratuities by alien and other residents of the United States, it was essential in this connection to undertake also a study of tourist expenditures and immigrant remittances and of their trends during the post-war years. These related studies promise to provide the bases for more accurate and more detailed current estimates of major balance-of-payments items. They assume, therefore, an importance independent of the investigation of currency movements.