Section 6(4) of P.L. 93-479 directs the Secretary of the Treasury "to the extent he deems feasible" to "analyze the effects of foreign portfolio investment on the United States balance of payments and the United States international investment position." Section 6(3) calls on the Secretary to "identify...the effects of foreign portfolio investment on American financial markets." This chapter represents the Department's response to these sections.

External Aspects: A Theoretical Overview

The requisite analyses cannot be based solely on the benchmark data collected in the survey, but must also utilize data on past and potential capital flows, and on payments of dividends and interest to foreign investors. Partly because foreign portfolio investments after netting out purchases and sales are small in relation to other balance of payments transactions, partly because of the effects of the major change from a system of par values to a system of more flexible exchange rates, and partly because of the complex economic relationships involved, this analysis will use theoretical and deductive reasoning as well as the actual statistical data. We turn first to the effect on the U.S. balance of payments.

The term "balance of payments" can be used to describe a statistical statement setting forth the international transactions of a country during a given period of time. Since all international transactions involve the exchange of economic values-goods, services and financial assets each transaction has a "debit" and "credit" phase; the record of international transactions is, therefore, a double entry accounting system in which the sum of all debits will equal the sum of all credits during a given period, if the data are correctly reported or estimated.

Popularly, however, the term "balance" has been, and still is, used to measure the net excess of debits or credits resulting from a given set of transactions (e.g., all those measuring the transfer of goods and services), which, by definition, has to be equal, with an opposite sign, to the

balance on all other transactions. Thus the Department of Commerce has in the past divided the balance of payments statistics in various ways to show the goods and services balance, the balance on current account, the basic balance (current and long-term capital transactions combined), the liquidity balance, and the official settlements balance.1 In this chapter, however, we shall discuss principally the relationships between portfolio investment and other international transactions, and not just the effect of portfolio investment on any particular measure of the "balance" in the sense just described.

The Potential Relationships Among International
Transactions

If the international current account transactions of the United States during a particular period of time are taken as determined by general economic conditions at home and abroad (an assumption which may be broadly valid, but only in the very short run), foreign net purchases of U.S. securities must result in an offsetting reduction in some other foreign claim on the United States, or in an increase in a U.S. claim on some foreign country. Given the dominant position of the U.S. dollar as a world currency, both for private transactions and as official international reserves, the former is the more likely alternative.

Typically, a foreign purchaser of securities will enter the exchange markets to buy dollars, say from his local bank, in exchange for local currency. The bank may provide the dollars by drawing on deposits held with its U.S. correspondent; the deposits will then be transferred to the former U.S. owner of the securities. If the sequence stops there, the only effect is that foreigners now have more long-term, and less short-term, claims on the United States. Another possibility is that some U.S. bank

'The balance of payments statistical presentation has been substantially modified in the light of the changed international monetary system and will no longer include any overall measure of the "balance."

will supply the dollars through loans denominated in dollars or-less likely-in exchange for a local currency deposit in a foreign bank. In these cases foreigners have acquired a long-term claim on the United States, and the latter has acquired a claim (usually short-term) on foreigners. In neither case will any other items in the balance of payments be affected.

But, of course, the sequence does not stop there. In the first, and more general, case, the foreign bank will want to rebuild its depleted working balances with its U.S. correspondent, and will, in turn, seek to acquire more dollars. What happens next depends on conditions in capital markets and other effects of monetary policies in the United States and abroad, and, perhaps even more importantly, on whether exchange rates are relatively free to move as they are under the system prevailing since 1973.

Before proceeding to the next stage, however, it should be noted that-as in all economic analysis-it is difficult to determine whether an observed relationship between two series of economic events results from the fact that one is affecting the other, or from the fact that both are being affected by other developments. For instance, changes in transactions in U.S. securities may be associated with a broader pattern of changes in economic transactions, such as those caused by cyclical developments at home and abroad. An expansion in domestic business activity would tend to stimulate merchandise imports and, at the same time, to raise profits of domestic corporations. The latter, in turn, would tend to raise share prices and attract foreign portfolio investment into U.S. equities. These two developments would tend to have offsetting effects on the net international demand for U.S. dollars.

But the possibility also exists that a rise in the demand for dollar exchange by foreign residents to enable them to increase their purchases of U.S. securities will induce changes in other transactions that would reduce the foreign demand for dollar exchange for other expenditures in the United States, and thus stimulate transactions which would increase the supply of dollars offered for sale in the exchange markets, by either foreign or U.S. residents.

It cannot be assumed, however, that the changes in the demand for dollars and in the offers of dollars at any given set of exchange rates will always balance. Changes in foreign purchases of U.S. securities may fail to lead to, or be accompanied by, changes in other transactions that would automatically have an exactly offsetting effect on the exchange markets. Two major reasons for such possible failure are (a) that most major countries follow general monetary and fiscal policies designed to achieve a desired level of output and employment, and to avoid undesirable price movements, and (b) that international transactions are frequently influenced by expectations as well as by current conditions. If the transfer of purchasing power from residents of foreign countries to those of the United States become so large as to threaten major deviations from policy goals, offsetting measures are likely to be taken.

Fixed Exchange Rates

Under the Bretton Woods system of par values which prevailed until 1973, and in the special case of the United States, changes in private capital flows that did not result from economic forces which tended to create offsetting changes in other balance of payments transactions were usually associated with changes in official holdings of U.S. securities-mainly U.S. Treasury bills. The demand for dollars created by foreign portfolio investors tended to cause the exchange rate of the dollar vis-a-vis foreign currencies to rise, approaching support limits set by international agreement. To prevent that limit from being breached, and often well before the limit was reached, some foreign central banks would sell dollars in the market, liquidating their holdings of U.S. Treasury bills or other dollar reserves in the process.

Although under that exchange rate system, foreign purchases and sales of dollar funds associated with foreign demand for, or sales of, U.S. securities did not significantly affect the exchange rate of the dollar because they were offset by official transactions, they affected the capability of the foreign authorities to maintain this system. When foreign central banks sold dollars (used reserves), their ability to resist future downward pressures on their currencies was weakened. In view of the limited amounts of international reserve assets held by central monetary authorities, and their limited ability to borrow funds, prolonged downward pressures on the exchange rate of any country sooner or later would force that country to change its domestic economic policies, or to restrict payments to foreigners, or to change its exchange rate.

Even the opposite situation caused problems on occasion. Initially a reserve increase would strengthen the ability of a foreign country to resist downward pressures on its exchange rate at a subsequent time when the situation in the exchange markets was reversed. If reserve additions continued over a longer period and, as was likely, involved a major expansion in the domestic money supply, the authorities were apt to become concerned that this expansion would create upward pressures on prices, distort the structure of the economy, intensify social and political tensions, and thus ultimately impair or even reverse their country's economic relationships with other countries which had contributed to the initial strengthening of its international reserves. Thus, even prolonged and intensive upward pressures on exchange rates would induce countries to adopt offsetting economic policies, and, if that were not desirable or successful, to change the agreed-upon exchange rates.

Floating Exchange Rates

Under today's exchange rate system, exchange rates respond more quickly than formerly to changes in the demand for, or supply of, a country's currency in the exchange markets. Thus, if associated or induced transactions at the existing exchange rates are not large enough to offset an increase in the foreign demand for dollars to purchase United States securities, the exchange rate of

the dollar will rise until changes in other transactions are induced which would either raise the supply of, or lower the demand for, dollars in the exchange market. Which of the various categories of international transactions will adjust as a consequence of changes in the exchange rate of the dollar will depend not only on the res onse of buyers and sellers to changes in relative international prices, but also on their judgment as to whether, and to what extent, these changes will continue or be reversed. The adjustments will also depend on the time needed to change the available supplies of goods or financial assets, which, in turn, may be a function of general business conditions both in the United States and abroad.

The earliest responses to changes in exchange rates are likely to occur in the capital accounts of the balance of payments because trade adjustments are relatively slow. The latter may require changes in trading relationships, in the placement of orders, and in production and delivery schedules; in many instances they may also require new investment plans and actual construction of productive facilities. Since changes in the exchange rates have to continue until the demand for a currency by potential buyers equals the offer of that currency by potential sellers, and trade can respond to such changes only after some time lag, the short-run gaps between demand and supply in the exchange markets are likely to be bridged by capital transactions, including those of speculators. Such capital transactions may involve shifts in the timing of acquisitions or sales of foreign currency assets as well as increases or decreases in such acquisitions in anticipation of changes in trading activities. It is also true, however, that longer run trends of either net purchases or net sales of U.S. securities by foreign residents are likely to induce changes in merchandise trade and services transactions, and that this effect is more likely to occur sooner under the new monetary system than under fixed exchange rates.

In the next part, the actual course of events over the last decade will be analyzed, in an effort to discover what effect, if any, actual transactions in U.S. securities have had on the U.S. balance of payments.

Foreign Purchases and Sales of U.S. Securities, and Transactions Related Thereto, in the Balance of Payments of the United States

General Magnitudes

During the 11 years 1965-75, net expenditures by private foreign residents on U.S. stocks amounted to about $14.3 billion; net expenditures on bonds were about $11.5 billion, on long-term loans to U.S. corporations about $4.2 billion, and on long-term deposits in U.S. banks (including certificates of deposits with an original maturity of one year or more) about $0.3 billion. These balance of payments figures differ in several respects from those reflected in the value of private foreign portfolio assets in the U.S. as of the end of 1974, and from those used elsewhere in this report to measure capital flows.

First, the figures represent the total of capital flows during 1965-75. Thus, they exclude capital flows prior to 1965, but include 1975, for which data are now available.

Second, they include funds obtained by U.S. corporations from financing subsidiaries incorporated in the Netherlands Antilles, which the latter had obtained through bond issues and long-term bank loans in the Euro-capital markets. Through the end of 1975 the net capital inflows resulting from such operations amounted to about $4 billion, including about $3.1 billion obtained through bond issues and $0.9 billion through bank loans. In the benchmark survey, foreign claims on these NA companies were not considered as portfolio investments in the United States, nor were the loans by the NA subsidiaries to their parent companies.

Third, the balance of payments figures do not reflect changes in market value which, of course, affect not only the assets acquired by foreigners during the period covered here, but also those acquired earlier.

Fourth, the figures on securities represent net receipts. after taking account of commissions on purchases and sales. Thus, the actual values of securities purchased by foreigners are smaller, and the values of securities sold by foreigners larger, than the recorded data, so that the net payments by foreigners for the securities themselves would be less than the amounts shown in the balance of payments figures.

The aggregate amount of the commissions depends on the rate per unit of value of purchases and sales, and the total turnover-not on the amounts of net purchases and sales. On the basis of data reported by members of the New York Stock Exchange,2 the commissions on stock transactions in 1974 were about 1.21 per cent of the value of purchases and sales, and 0.84 per cent in 1975. Applying the 1974 rate to that and prior years, the commissions on stock transactions paid by foreigners during the years 1965 through 1975 may have amounted to about $2.4 billion. Commissions on bonds issued abroad, which account for $9.9 billion of the total net foreign. purchases, and on corporate loans were presumably paid by the U.S. borrowers, and commissions paid by foreigners on other transactions in U.S. bonds were relatively small.

Based on these estimates, net foreign purchases of U.S. stocks during that period were about $12 billion, instead of the $14.3 billion mentioned above. In principle, at least, these adjusted figures are conceptually compatible with the figures on the total value of U.S. stocks owned by foreigners as reported in the survey.

Finally, for the purpose of this analysis, transactions in U.S. stocks by foreign official agencies are omitted. Liquidations of the U.K. Treasury's portfolio of U.S. stocks acquired in World War II have been eliminated (on an estimated basis) from the data in earlier years; also eliminated are OPEC countries' net purchases in 1974 and 1975 and (relatively small) net purchases by international agencies throughout the decade.

2Reported by the New York Stock Exchange, Research Department-International Finance.

Stocks

Foreign expenditures on U.S. stocks (including commissions) were relatively small prior to 1967. The largest amount in one year during the preceding 20 years had been in 1959 when net purchases were about $360 million. In 1967, however, net U.S. receipts from that source rose to nearly $800 million, and reached $2.1 billion in the following year. That first wave of foreign. investment in U.S. equity securities presumably reflected, in part, the growing incomes and wealth in advanced foreign countries and the rapid development of foreign investment funds.

This first post-war wave of net foreign investment in U.S. stocks subsided after the first quarter of 1969, in part, at least, because a major participant in the expansion of foreign investment funds had overextended its activity and had to curtail its operations. In the first half of 1970 foreigners sold more U.S. stocks than they purchased, but net purchases were resumed in the second half of the year. The second wave of net foreign expenditures reached a peak annual rate of about $5 billion in the first quarter of 1973, then declined to reach another period of small disinvestment in the second half of 1974, to be followed by another upsurge during 1975.

Cyclical movements. The waves in net foreign expenditures on U.S. stocks appear to have become steeper with the succeeding cycles. During the upward phases of these cycles, the rate of increase in net receipts over the cycle (measured by the growth in quarterly net foreign purchases from the trough to the peak quarter divided by the number of quarters in the upward phase) grew with each succeeding cycle. The growth in net purchases so measured averaged $100 million per quarter in the period 1966 IV to 1968 IV, about $180 million from 1971 II to 1973 I, and about $380 million from 1974 IV to 1975 III. Similarly, in the two distinguishable downward swings, the average rate of decline was greater in the second. (See Table 15.)

These data may, of course, merely reflect a growing foreign interest in United States stocks over the decade. On the other hand, they seem to indicate increasing variability over time, a characteristic that is not inconsistent with other measures of volatility of foreign stock holdings, discussed later in this chapter.

there were only few periods when these transactions resulted in net outflows of funds from the United States. Even during these periods, net payments by the United States hardly exceeded $200 million in any quarter, and continuous periods with net payments did not last more than two quarters. Thus the fluctuations in the transactions in U.S. stocks with private foreign residents did not-with minor exceptions-involve shifts between net foreign purchases and net foreign liquidations, but rather mainly between larger and smaller net purchases.

Impact on exchange markets. In view of the relatively large fluctations in the balance of transactions in U.S. stocks of private foreign residents, the question arises as to whether and to what extent these fluctuations increased the instability in total external transactions of the United States, or whether they counteracted instabilities in other transactions. In this connection, it is illuminating to compare the timing in these fluctuations with those in the balance on merchandise trade.

A comparison of high and low points in the two series shows that until the end of 1974 the highs and lows in both series3 were no more than one quarter away from the opposite turning points in the other. The peak in the net private foreign purchases in United States stocks in the second quarter of 1963 followed by one quarter a low in the merchandise trade balance in the first quarter of 1963; the next peak in the second quarter of 1966 preceded a low quarter in the trade balance by one quarter; the same applies to the stock purchases peak in the fourth quarter 1968; and the peak in the first quarter of 1973 followed a four-quarter period of relatively low trade balances that ended in the last quarter of 1972.

Similar relationships may be found between lows in the balance on transactions in stocks and highs in the trade balance. The low in the fourth quarter 1964 coincided with the end of a four-quarter period of relatively high trade balances, and the low in the second quarter 1970 was associated with a high in the trade balance that occurred in the same quarter and continued in the next.

Starting in 1973, the trade balance was strongly affected by rising prices and other factors bearing on exports of agricultural products, and, starting in 1974, by the rise in price of imports of petroleum and petroleum products. This was also the period in which the floating of exchange rates became widespread. Comparisons with the balances on transactions in United States stocks and U.S. trade in the years 1973 to 1975 are perhaps more valid if the commodities just mentioned are omitted from the trade balance. If this is done, the high in the balance on transactions in United States stocks in the first quarter of 1973 coincides with the low in the adjusted trade balance; from that point on, the trade balance showed a steady rise for nine quarters to the next peak in the second quarter of 1975, while the balance on transactions in United States stocks declined for seven quarters and reached a low in the fourth quarter of 1974. These opposite movements in the two balances follow the pattern exhibited in the earlier years.

3A "high" trade balance is defined in terms of the surplus of exports over imports.