In the early to middle 1960s the U. S. authorities adopted three main measures in an attempt to reduce capital outflows and to strengthen the U. S. balance of payments: the Interest Equalization Tax (IET), which applied to purchases by U. S. residents of foreign stocks, bonds, and other equity and debt securities; the Voluntary Foreign Credit Restraint (VFCR) program, which established ceilings on loans to foreigners by U. S. financial institutions, banks, and nonbanks; and the Foreign Direct Investment (FDI) program, which regulated the supply of funds from U. S. nonfinancial corporations to their overseas affiliates. These controls were terminated on January 29, 1974.1
It is the purpose of this paper to attempt to quantify the effects of these capital control measures on the capital account of the U. S. balance of payments—with particular emphasis on their impact on short-term capital transactions. For this purpose, a portfolio model of the relevant items of the capital account of the balance of payments is presented and estimated using quarterly data for the period covering the first quarter of 1964 through the first quarter of 1973.
Section I of the paper briefly describes the U. S. control measures, Section II presents the essence of our model, and Section III sets forth the empirical results and the main themes and conclusions of the paper.
I. U.S. Measures to Control Capital Outflows
Against a background of large balance of payments deficits and increasing private capital outflows, the IET was announced and became effective in July 1963, although it was not formally enacted until September 1964.2 As a tax on purchases of foreign securities by U. S. residents, the IET was designed to reduce capital outflows arising from portfolio investment abroad by offsetting the advantages of higher foreign interest returns. Several major categories of transaction were exempt from the IET, including direct investments, investments in developing countries, investments in Canadian obligations, and bank loans made by foreign branches of U. S. banks to foreign residents.
In its initial form the IET did not apply to long-term bank lending, which, as Governor Brimmer has noted,3 encouraged foreign borrowers to resort to bank financing rather than to the sale of securities in the United States. Simultaneously, other types of capital outflow (particularly short-term bank financing and direct investment) that also were not subject to the IET increased noticeably. The result was heavy capital outflows throughout 1964. By the last quarter of that year private capital outflows were at an annual rate in excess of $8 billion, and the overall balance of payments (measured on a liquidity basis) was in deficit at an annual rate of $5 billion.
In response to these developments, the U. S. authorities introduced in March 1965 a voluntary balance of payments program that consisted of the VFCR program and the FDI program. At the same time, the IET was extended to cover long-term bank loans to foreigners. Although these controls were generally assumed to be of a temporary nature—to be used “until the basic strength of our trade surplus, rising earnings on direct investment, and other favorable elements could coalesce to restore a lasting balance” 4—the continued deterioration in the balance of payments, particularly in 1967, led U. S. authorities to make significant changes in these programs in January 1968. In particular, the scope of the IET was widened, the FDI program established mandatory guidelines, and the Federal Reserve Board was given the power to make the VFCR guidelines mandatory if the situation warranted.
As established in 1965, the VFCR program requested financial institutions to observe voluntary ceilings on their lending to foreigners. Although there were separate guidelines for banks and nonbanks,5 both were requested “to give absolute priority to credits needed to finance exports of U. S. goods and services, and in the nonexport category to give highest priority to credits to, and investments in, less developed countries.” 6 Perhaps the only significant difference in the treatment of banks and nonbanks under the VFCR program was that the proportion of banks’ foreign assets subject to the program was higher than that for nonbank financial institutions.7
Although the ceilings were modified several times, the general structure of the VFCR program changed little during the period that it was in effect.8 The Federal Reserve chose not to make the program mandatory in 1968 “in view of the degree of cooperation exhibited by financial institutions since the program was begun.”9 However, several modifications were made to the program in 1968, and among them was the establishment of a separate ceiling on nonexport credits to the developed countries of continental Western Europe.10 The other notable change in the VFCR program occurred in 1970 when a separate ceiling was established for term credits that banks used to finance exports of U. S. goods and services.11 The abolition of this ceiling on November 11, 1971, together with a revision of the formula for calculating VFCR ceilings, represented a substantial loosening of those limitations on foreign lending by U. S. banks—export credits were no longer subject to guideline limitations and the new ceilings implied additional leeway of somewhat more than $1 billion from the situation prevailing before November 11.
The voluntary FDI program, administered by the Department of Commerce, requested U. S. nonfinancial corporations to restrain their direct foreign investment. Although this program remained voluntary through December 1967, as Herring and Willett have pointed out, “the restrictions on direct investment flows became increasingly stringent” during the period.12 The mandatory program announced on January 1, 1968 called for a reduction in the balance of payments impact of foreign direct investment by $1 billion in 1968. The 1968 program required the direct investor to choose one of three authorization limitations for investment in all countries except Canada, where investment was unrestricted:
(a) a world-wide minimum investment allowable of $1 million;
(b) an “earnings” allowable for each of three predetermined geographical areas, where the allowable for each area was an amount of direct investment equal to 30 per cent of the annual earnings of the direct investor’s foreign affiliate;
(c) an “historical” allowable for each of these three predetermined geographical areas, based on investment experience during the period 1964–66. Investments made with the proceeds of “long-term foreign borrowing” were not counted against the allowables, but repayment of such borrowing counted as direct investment subject to controls; there were also provisions relating to the repatriation to the United States of “unused” foreign borrowing. Finally, direct investors were prohibited from holding month-end “liquid foreign balances” that exceeded the average month-end outstandings in 1965–66.13
Any attempt to quantify the impact of these control programs on the U. S. balance of payments must take into account that the three programs were intended to be interrelated and mutually reinforcing.14 It would, therefore, seem desirable to consider these controls in “packages,” and to quantify the impact of these packages on the balance of payments as a whole. Hence, for the purposes of this paper three “packages” are identified: the 1964 package (enactment of the IET), the 1965 package (introduction of voluntary programs together with the amplification of the IET), and the 1968 package (imposition of mandatory controls under the FDI program together with refinements and extensions of the IET and VFCR programs).
There have been some published attempts to gauge the impact of these U. S. control measures on the balance of payments, and most such studies have been restricted to examining the direct impact of the control measures on the balance of payments items that they were designed to influence—for example, the impact of the FDI program on an “adjusted” balance of payments definition of direct investment, or the impact of the VFCR program on bank or short-term credit.15 Such a “micro” approach encounters empirical difficulties because it attempts to consider each control measure separately from its package—the observed effect of a capital control measure on a particular item of the balance of payments may be the sum of the direct impact of the control measure in question plus some indirect effects of other controls that are in force. Moreover, by concentrating on just one item of the balance of payments, this approach tends to overstate the impact of the control measure on the balance of payments as a whole because it neglects the indirect (in particular, “offsetting”) effects on other payments items.
Although several studies have discussed the possibility of “offsets” to the control measures,16 few, if any, have attempted to quantify these “offsets.” A typical approach is that taken in one study of the FDI program, where it was asserted:
The apparent tradeoff between U. S. capital outflows and overseas borrowing which was prompted by the 1965 voluntary restraint program may have been partly offset through round-about dollar drains. But the practical importance of such offsets seems doubtful. For expositional simplicity, the following discussion assumes that offsets do not exist.17
Although quantification may be difficult, offsets are potentially important. Consider the imposition of, say, a ceiling on U. S. bank lending to foreigners. For given asset demands and interest rates, the imposition of the ceiling would tend to increase the demand for loans from sources outside the United States. U. S. banks may try to meet this increased demand simply by transferring their accounts from the books of their head offices in the United States to those of their foreign branches. Hence, the ultimate effect of such controls on the U. S. balance of payments need not be unambiguously positive.18 Of course, the U. S. authorities could try to limit such offsets by increasing their control over the transactions of U. S. residents and banks, but offsets through the activities of nonresidents still could be large. If these offsets through the activities of nonresidents turned out to be significant, this would bring to the fore questions as to whether the U. S. authorities could, or should, impose controls on foreign-held capital in the United States. Important offsets have occurred through flows of foreign-held capital—for example, it has been suggested that following the introduction of the VFCR program, private foreigners repatriated their short-term deposits from the United States to Europe.19
To avoid the aforementioned problems, this paper adopts a “macro” approach to the question of the effect of capital controls. For U. S. controls, this necessitates concentrating on the impact of packages of controls on the whole balance of payments, even though this approach may not result in the best possible specification of the control variables. Specifically, it was necessary to rely on the use of dummy variables to gauge the impact of these packages of controls. Although we do not consider this the most desirable approach, the considerable difficulty encountered in attempting to specify variables for each of the control programs convinced us that it was the only viable alternative.20 Furthermore, in this paper our efforts are concentrated on the impact of these packages on financial capital flows. A complete analysis of these measures would also consider, of course, their impact on the items concerning trade and transactions in invisibles in the balance of payments, but this is more appropriately the subject of a separate investigation. Finally, the analysis concentrates on what may be called the short-term impact of these control measures in that we neglect (although we do not consider unimportant) the longer-run effects of controls on the allocation of resources because of changes in the costs of financial intermediation.
II. The Model
The framework for analyzing the U. S. capital control programs is a two-country (United States and “Europe”) general equilibrium portfolio model that was developed in relation to our previous work on the Euro-dollar market. As originally specified, this model21 assumed that the U. S., “European,” and Euro-dollar interest rates were endogenous. However, since attention in this paper is confined to the external account of the United States, for simplicity it is assumed that all U. S. and “European” interest rates are exogenous.
It was decided to specify equations for the following short-term items of the capital account of the U. S. balance of payments: short-term bank claims on private foreigners; total (bank and nonbank) short-term claims on private foreigners; liquid liabilities (of banks and nonbanks) to private foreigners; and other short-term liabilities (of nonbanks) to private foreigners. In addition, since one of the most interesting questions with respect to the effectiveness of controls concerns the “potential” for “leakages” through the errors and omissions item, a separate equation was estimated for errors and omissions.22 Each of these equations (designated as a demand or supply function) was specified, in accordance with standard portfolio theory, as a function of the relevant interest rates, constraint variables, and other exogenous policy control variables.
For the specification of the bank-claims equations, it was assumed that during the period considered in our empirical investigation there was always an excess demand for U. S. claims on private foreigners—that is, that the actual credits granted were determined largely by supply conditions. Since the capital control measures (specifically, VFCR) were in effect for most of the period considered, a supply-determined specification of the bank-claims equation seemed justified.
In these circumstances, the loanable funds of U. S. commercial banks was considered the appropriate constraint variable’23 in the bank-claims equation. When loanable funds (defined as total deposits minus required reserves) increased, banks would lower their loan rate to expand loans. In addition, it was assumed that banks always attempted to meet the loan demands of their prime domestic customers, so that when the demand for prime domestic loans was high (for example, in times of credit restraint in the United States), banks were induced to restrict their external credits.
Given these two constraints, U. S. banks were assumed to determine their demand (supply) for various assets (liabilities)—external credits, external liabilities, domestic securities, free reserves, etc.—on the basis of relative rates of return and risk. Assuming that relative risk remained unchanged except during periods when speculative forces were important, the outcome would reflect mainly interest rate configurations in the United States and abroad. The higher the foreign interest rate, represented in this paper by the Euro-dollar rate, the more U. S. banks would have been induced to lend abroad; the higher the U. S. rate, the less they would have been prepared to lend abroad.
Since a substantial portion of bank credit is related to foreign trade, the U. S. export variable was included in the equation to represent the pressure of demand for trade-related credits. The inclusion of a “demand” variable in essentially a supply-oriented equation is not necessarily inconsistent with the other assumptions of the paper, inasmuch as banks regard trade flows as a particularly desirable collateral for loans and, therefore, attempt to meet trade loans even when their financial conditions are tight. This export variable may be seen as representative of the prime loan demand of foreigners.
It was decided to use the same specification for the total-claims equation as for the bank-claims equation, even though the former includes items other than banking claims. The decision was in part a result of difficulties encountered in specifying an equation for the difference between total claims and bank claims, and of the difficulties of constructing appropriate nonbank-wealth variables. But it also seemed to provide a reasonable approximation, since credits given by nonbank institutions and banks are determined by similar variables.
To capture the impact of the capital control programs in the claims equations, three separate dummies (to be defined in the next section of the paper) were included in each equation. Since both bank credit and total credit were largely subject to the VFCR program, it was reasonable to expect these control variables to enter with a negative sign in these equations, on the assumption that the controls were effective in terms of their direct impact.
In addition, in view of the recurring importance of speculative activity during the period considered, several speculative variables were included in each equation. Since speculation was generally against the U. S. dollar, the signs of the speculative variables in the claims equations were expected to be positive.
The equations for bank claims and total claims, therefore, may be expressed as follows:
BC = total short-term bank claims on private foreigners
TC = total short-term claims on private foreigners reported by banking and nonbanking concerns
RED = Euro-dollar interest rate
RUS = U. S. interest rate
W = bank-wealth variable, defined as deposits minus required reserves of large U.S. commercial banks
EX = U. S. exports
CL= “prime” domestic loans
CAPC = dummy variables to represent the capital control programs
SP = speculative variables
CRED, CW, CEX, CSP > 0
CRUS, CCL < 0
On the liabilities side, liquid liabilities to private foreigners consisted mainly of U. S. banks’ borrowing from their foreign branches, borrowings by the U. S. branches and agencies of foreign banks from their head offices, demand and time deposits of private foreigners (including commercial banks) in U. S. banks, and private foreign holdings of U. S. Treasury marketable bonds and notes.24 The first-mentioned component, U. S. banks’ borrowing from their foreign branches, showed the widest fluctuation during the period considered, reaching a peak of $14.6 billion in August 1969, at which time it represented roughly half of the total liquid liabilities to private foreigners. For this component of liquid liabilities25 the appropriate wealth or constraint variables were the loanable funds variable and a variable to represent the prime loan demand of residents (as defined earlier in the discussion of the claims equations). An increase in the availability of loanable funds from domestic sources would have reduced the need to borrow from foreign branches, and, therefore, the loanable funds variable was expected to have a negative sign in the liabilities equation. The prime loan variable was expected to take on a positive sign, since foreign borrowing was resorted to by U. S. banks when needed to meet their prime loan demands; moreover, this variable would also seem to represent the demand by U. S. residents for loans from the U. S. branches of foreign banks.
However, for other components of liquid liabilities to private foreigners, the choice of the appropriate constraint variable was not obvious. In view of the difficulties encountered in attempting to develop wealth variables for foreign banks and other private foreigners, and because the inclusion of several wealth and income variables in the one equation could be expected to produce severe multicollinearity problems, it was decided to use “European” income as a proxy for all other wealth and income variables apart from those described earlier for borrowings by U. S. banks from their foreign branches. Given these wealth and income variables, three interest rates—European, Euro-dollar, and U. S.—were included in the equation for this liability item. These rates determined the allocation of assets and liabilities by the designated portfolio owners, assuming that relative risks remained unchanged except during periods when speculative factors took on particular importance.
To complete the discussion of the specification of the equation for liquid liabilities to private foreigners, it is necessary to indicate the relevant capital control variables. As demonstrated in a previous paper,26 Regulation Q and the reserve requirement on U. S. banks’ borrowing from their branches exerted a significant impact on the U. S. banks’ borrowing from these sources. When the Regulation Q ceiling was effective in 1966 and 1969, U. S. banks were unable to compete with Euro-banks, as well as with other segments of the U. S. money and capital market, in attracting funds. Consequently, the head offices of U. S. banks resorted to their foreign (largely Euro-dollar) branches for funds. The 10 per cent reserve requirement imposed by the Federal Reserve in September 1969 (and its subsequent increase to 20 per cent in January 1971) raised the effective cost of borrowing and induced banks to limit, and eventually to reduce, their indebtedness to foreign branches.
Although the capital control programs were designed to avoid having a direct impact on liquid liabilities to private foreigners, it seems plausible to believe that these controls produced indirect offsetting outflows in this item in the form of reductions in private foreign holdings of U. S. demand and time deposits. It can be assumed, for example, that if the VFCR program was effective in curtailing U. S. credits to private foreigners, they were induced to increase their borrowings from Euro-dollar banks. To the extent that these borrowers maintained working balances that bore some relationship to their loans, or wished to hold deposit and loan accounts with the same bank, it is conceivable that this shift in borrowing from the U. S. to the Euro-dollar market also may have led to a shift of deposit accounts to the Euro-dollar market. Furthermore, such a shift in liquid liabilities might also have been initiated by the movements of funds between head offices and foreign branches of U. S. banks as U. S. head offices needed less funds because of the ceiling on their foreign lendings.
The equation for U. S. liquid liabilities to private foreigners was expressed as follows:
LL = total U. S. liquid liabilities to private foreigners
RE= “European” interest rate
YE =“European” income
QNL = Regulation Q variable
RQ= reserve requirement on U. S. banks’ borrowing from their foreign branches
1RED, 1RE 1W 1RQ, 1SP < 0
1RUS, 1YE, 1CL, 1QNL > 0
Other short-term liabilities to private foreigners are nonliquid liabilities to private foreigners reported by nonfinancial corporations—in particular, nonbank financial institutions and multinational corporations. Again, the constraint variable most appropriate for this purpose was not obvious, but it was decided to represent constraints on the activities of U. S. institutions and corporations by two variables—U. S. income and a corporate liquidity variable (accumulated retained earnings). The choice of these wealth and income proxies was somewhat arbitrary, but in view of the lack of aggregate financial data on major U. S. multinational corporations, it seemed a reasonable second-best solution.
The equation for other liabilities to private foreigners was expressed as follows:
OL = other U. S. short-term liabilities to private foreigners
YUS = U. S. income
LIQ = corporate liquidity variable
hRED, hRE, hSP, hLIQ < 0
hRUS, hYUS > 0
Since it is difficult to determine the nature of the errors and omissions item, it was decided to hypothesize, along with Branson and others, that at least in part this item represented unrecorded financial claims of U. S. corporations and nonbank financial institutions transacted largely through the Euro-dollar as well as other financial markets.27 It seemed useful to test whether these corporations and institutions (especially the so-called multinational corporations) attempted to offset the impact of the U. S. capital control programs through transactions that would not normally be captured in the recording process by increasing their claims on their branches and subsidiaries. However, errors and omissions includes the effects of leads and lags in trade, probably some unrecorded banking transactions, and errors of statistical measurement. Hence, it was not readily apparent which constraint variables should be included in this equation. Rather than neglect these altogether (as Branson has done), it was decided to experiment with the wealth and income variables used in other equations of the model. If one of the banking-constraint variables turned out to be significant, it would perhaps give an indication of the importance of some unrecorded banking transactions, and, similarly, if one of the scale variables for non-U. S. residents (for example, the “European” income variable) turned out to be significant, it would perhaps indicate that some transactions between the foreign subsidiaries of U. S. multinational corporations and their head offices had gone unrecorded. The errors and omissions equation was expressed as follows:
E and O = accumulated errors and omissions
W1 = some “appropriate” wealth variable
eRUS < 0
eRED, eRE, eSP > 0
In addition to these equations for short-term claims and liabilities, it was decided to estimate an equation, on the basis of the portfolio model, for aggregate short-term capital and for aggregate capital, although it was recognized that the conceptual justification for the last-mentioned equation was somewhat less satisfactory. However, there was some merit in estimating these equations, since they enabled us to arrive at an estimate, albeit crude, of the impact of the capital control programs on the aggregate capital account without estimating individual equations for direct investment and portfolio investment, to which our theoretical structure was not directly applicable.
Finally, as a supplementary check on the effects of the U. S. capital control measures, a reduced form Euro-dollar rate equation was estimated. If a particular capital control measure was effective, it should have had an influence on interest rates as well as on capital flows. The basic assumption with respect to the Euro-dollar market was that Euro-banks acted as “pure” financial intermediaries in the sense that they had no influence on the equilibrium interest rate for Euro-dollars. It was also assumed that there was a constant margin between the Euro-dollar deposit and loan rates; thus, Euro-banks were assumed to adjust the deposit rate so as to simultaneously equate deposits to loan demands. Given the specification of the various deposit and loan functions, the general equilibrium model was solved for a reduced form Euro-dollar rate equation of the following form.28
X = a vector of all other exogenous variables in the demand functions for Euro-dollar deposits and loans
fRUS, fRE > 0
III. The Empirical Results and Conclusions
The empirical investigation was conducted using quarterly data for the period covering the first quarter of 1964 through the first quarter of 1973. Although the equations estimated did not represent the entire system of capital flows, with all the interdependency that is involved, they were estimated by two-stage least-squares to reduce the simultaneous bias relating to the endogenous nature of the Euro-dollar interest rate.
BC = total short-term U. S. claims on private foreigners reported by banks in the United States—U. S. Department of Commerce, Survey of Current Business (SCB)
TC = total short-term U. S. claims on private foreigners reported by banking and nonbanking concerns in the United States (SCB)
LL = total U. S. liquid liabilities to private foreigners (SCB)
OL = other U. S. short-term liabilities to private foreigners (SCB)
E and O = accumulated errors and omissions (SCB)
STK = TC − (LL + OL) + E and O
TK = total net capital (SCB)
REDC = Euro-dollar call rate—Morgan Guaranty Trust Company of New York, World Financial Markets (WFM)
RED3 = 90-day Euro-dollar deposit rate (WFM)
RUSC = U. S. federal funds rate—Board of Governors of the Federal Reserve System, Federal Reserve Bulletin (FRB)
RUS1 = U. S. corporate AAA bond rate (FRB)
RE = German call rate (covered)—International Monetary Fund, International Financial Statistics (IFS)
W = total loanable funds of U.S. commercial banks defined as total deposit liabilities minus required reserves (FRB)
YUS = U. S. gross national product (SCB)
YE = “European” income defined as wholesale price index multiplied by index of industrial production in six major countries (excluding the United States)—Canada, Japan, the United Kingdom, France, Italy, and Germany. (Production indices and weights from Organization for Economic Cooperation and Development, Main Economic Indicators; price indices from IFS.)
CL = total commercial and industrial loans by U.S. commercial banks (as an approximation to the prime loan demand of domestic residents) (FRB)
EX = U. S. exports (as an approximation to foreign prime loan demand) (IFS)
LIQ = corporate liquidity variable defined as accumulated retained earnings—Data Resources Inc. (Data Bank)
SP = a dummy variable to represent the general speculative environment after the suspension of convertibility in August of 1971; it was defined as:
0 - 64:1 to 71:2
0.5 - 71:3
1.0 - 71:4 to 73:1
SPUK = forward discount on U. K. pound sterling (IFS)
SPG = forward premium on deutsche mark (IFS)
SP712 = a dummy variable to represent the speculative activity from May 1971 to the time of the announcement of the suspension of convertibility of the U. S. dollar in August of 1971; it was defined as:
0 - 64:1 to 71:1
1 - 71:2 to 71:3
0 - 71:4 to 73:1
a variant of this dummy was designated SP713 and was defined as:
0 - 64:1 to 71:2
1 - 71:3
0 - 71:4 to 73:1
SP73 = a dummy variable to represent speculative activity during the first quarter of 1973; it was defined as:
0 - 64:1 to 72:4
1 - 73:1
QNL = Regulation Q variable defined as the ratio of the U. S. federal funds rate to the three-month certificate of deposit issue rate (FRB and WFM)
RQ= reserve requirement on borrowing by U. S. banks from their foreign (largely Euro-dollar) branches; it was defined as:
0 - 64:1 to 69:2
10 - 69:3 to 70:4
20 - 71:1 to 73:1
CAPC1 = a dummy variable to represent the “1965 package” of U. S. capital controls; it was defined as:
0 - 64:1 to 65:1
1 - 65:2 to 73:1
CAPC2 = a dummy variable to represent the “1968 package” of U. S. capital controls; it was defined as:
0 - 64:1 to 67:4
1 - 68:1 to 73:1
IET = a dummy variable to represent the period following the enactment of the IET in September 1964 and prior to the imposition of the voluntary program in March 1965; it was defined as:
0 - 64:1 to 64:3
1 - 64:4 to 73:1
All variables involving value were in billions of U. S. dollars, and all interest rates were in per cent per annum.
total short-term capital(TC − LL − OL + E and O)
In interpreting the impact of the capital control measures from equations (6)–(12), one should bear in mind the meaning to be attached to the signs of the coefficients on the capital control variables (IET, CAPC1, and CAPC2). In the equations for claims, errors and omissions, total short-term capital, and total capital, a positive sign is to be interpreted as a “perverse” effect, or an “offset,” while a negative sign would indicate that the control measure had a favorable effect on the balance of payments item in question. In the liabilities equations, a positive sign would indicate that the capital control measure had a favorable effect, whereas a negative sign would indicate that it had an offsetting effect.
Consider each of the control packages in turn. With respect to the impact of the IET, perhaps it should be noted that its failure to improve the U. S. balance of payments relates only to the impact of the enactment of the IET legislation in September 1964—deficiencies in the data ruled out consideration of the impact on the period immediately after the program was announced in June 1963. In the two quarters following the enactment of the IET, the capital account of the U. S. balance of payments deteriorated substantially. As revealed by the coefficient on IET in equations (11) and (12), outstanding short-term capital appears to have increased by some $2.7 billion (of which total claims—equation (7)—accounted for some $2 billion) while total capital may have increased by as much as $4.8 billion. These outflows, which tended to offset the IET program, may have occurred in anticipation of the imposition of capital controls on other U. S. capital transactions; initially, the IET program did not cover long-term bank credits.31 As was noted earlier, in the period between the announcement of the IET in July 1963 and its enactment in September 1964 some uncertainty existed as to the precise form of the tax, which probably encouraged some foreign borrowers to delay their borrowing decisions. It seems reasonable, therefore, to expect that following the enactment of the legislation and the removal of this uncertainty, foreign borrowers who had previously delayed their decision to borrow from the United States would look for “loopholes” in the legislation and for alternative sources of funds—particularly, for sources such as U. S. commercial bank credits, which were exempt under the IET.
The 1965 package of controls appears to have reduced total claims (see CAPC1 in equation (7)) on private foreigners (through the VFCR program) by as much as $1.5 billion, of which roughly $0.9 billion took the form of banking claims. However, this direct effect was offset by a reduction in liquid liabilities to private foreigners (equation (8)) of some $1.9 billion.32 Private foreigners, including the subsidiaries and branches of U. S. multinational corporations, moved deposits from the United States to Europe, particularly to the Euro-currency market. This shift in deposits was mostly the direct consequence of the shift in the locus of borrowing activity, and U. S. banks simply transferred the accounts from the books of their head offices to those of foreign branches to undercut the impact of the controls. The rapid growth in the size of the Euro-currency markets after 1965 can be attributed to an important extent to the shift in both deposit and loan business from the United States to Europe that occurred after the VFCR program was instituted.
A central feature of this expansion of Euro-currency banking was the surge of interest on the part of U. S. banks in foreign branches. Since assets of foreign branches were not subject to VFCR guidelines, U. S. banks were encouraged to shift their foreign credits from the books of the U. S. head office to the books of the foreign branch. Speaking in 1970, Governor Brimmer noted that there had been approximately a sixfold increase in the number of banks with foreign branches abroad, including both “full” branches in Europe and elsewhere and the so-called shell branches in Nassau, etc. Governor Brimmer considered this rapid growth in the number and scope of foreign branches of U. S. banks as “one by-product of the VFCR program [which] may not be as helpful to the long-run international position of the United States as it may appear at first glance.” 33 Specifically, shifting the emphasis of the bank’s foreign operations toward foreign branches could lessen the incentive for U. S. banks to engage in activities that might promote exports.34
Examination of the equation for net short-term capital (11) seems to confirm the conclusion drawn earlier that the favorable impact of the 1965 package on total claims was (at least) completely offset by movements in other short-term capital transactions. Although the coefficient on CAPC1 in equation (11) entered with a positive sign (indicating a net offset), it was not significantly different from zero. Furthermore, on the basis of the equation for total capital transactions—the conceptual basis for this equation is quite weak—it would appear that the 1965 package of controls also brought into play an offset, of perhaps as much as $1 billion net, in long-term capital.
As was noted earlier, the most important aspect of the 1968 package was the adoption of the mandatory FDI program. Although the estimated equations do not allow us to quantify the direct impact of the 1968 package on direct investment (which has been attempted by several other authors),35 they do allow us to investigate whether this package had any effect on short-term capital transactions (largely as a result of the increased stringency in the VFCR program) and also roughly to gauge the impact of this program on the total capital account.
From the equation for total net short-term capital (11) it would appear that, if anything, the 1968 program produced a slight offset, although the coefficient on CAPC2 of $1.5 billion in this equation is not significantly different from zero. While the 1968 package may have had a marginally increased impact over the 1965 package in the bank-claims equation (6), nonbank claims may have decreased.
On the basis of the total capital equation—which is conceptually less satisfactory—it would appear that the 1968 package may have had a slightly favorable impact on the total capital account, reversing the net unfavorable impact of the 1965 package, although the coefficient on CAPC2 is not significantly different from zero. Furthermore, to the extent that there was a net offset to the 1968 package in short-term transactions, the conclusion that might be drawn is that the 1968 package had some favorable impact on direct investment outflows, which is not inconsistent with the results of other studies that have confined their analysis to the direct impact of the 1968 FDI program on direct investment.36
The foregoing conclusions with respect to the general ineffectiveness of the U. S. capital control programs, particularly on short-term capital transactions, are borne out by the reduced form Euro-dollar rate equation:
If these capital control programs had been effective in reducing net outflows from the United States, the variables CAPC1 and CAPC2 would have entered the Euro-dollar rate equation with positive signs. However, none of the control packages had any significant effect on the Euro-dollar rate, confirming our contention that the offsets to the 1965 and 1968 packages were statistically complete.37 It may be noted that both Regulation Q (QNL) and the reserve requirement on U. S. banks’ borrowing from their foreign branches (RQ) had a significant impact on the Euro-dollar rate.
Apart from their conclusions in respect of controls, the foregoing equations have other features that merit comment. The partials with respect to both the U. S. and Euro-dollar rates in the claims equation are lower than those of the equation relating to liquid liabilities to private foreigners. This result reflects the impact of the VFCR program on the equations for claims; that program led to the development of an excess demand for the credits from U. S. sources at the prevailing loan rates, and thus impelled borrowers to resort to more expensive Eurocurrency and other foreign borrowing as substitutes.
Regarding equations (6)–(10), it would appear that during the period considered in this paper—covering a timespan in which capital transactions were generally subject to some form of control—short-term capital transactions in the U. S. balance of payments were particularly sensitive to relative financial conditions in the United States and abroad. An increase of one percentage point in “the” Euro-dollar rate induced, ceteris paribus, a deterioration of as much as $6.4 billion in the short-term component of the balance of payments (including errors and omissions), whereas a comparable increase in “the” U.S. interest rate improved the same items by some $7.1 billion.38 However, the U. S. interest rate is a major determinant of the Euro-dollar rate, as is revealed from the estimation of the reduced form Euro-dollar rate equation—see equation (13). When account is taken of the dependence of the Euro-dollar rate on the U. S. rate, it can be seen that the impact of a change of one percentage point in the U. S. rate on the short-term items of the balance of payments is reduced from $7.1 billion to roughly $1.9 billion.
It is of some interest that in the errors and omissions equation the significant constraint variables turned out to be the “European” income variable and the U. S. domestic prime loan variable (U. S. commercial and industrial loans)—all other constraint variables were insignificant. This may indicate that errors and omissions include some transactions between the head offices of U. S. multinational corporations and their “European” subsidiaries and some unrecorded banking transactions perhaps relating to the operations of foreign banks in the United States.
Finally, as was to be expected, the speculative activity that followed the suspension of convertibility of the U. S. dollar in August 1971 had a significant impact on the U. S. balance of payments. As evidenced by the coefficient on (SP) in equations (6)–(10), such speculative activity produced a deterioration in the “short end” of the U. S. balance of payments of as much as $20.25 billion, a little less than half of which was through the errors and omissions item.
The general conclusion to be drawn from these results with respect to the impact of the capital control measures on the U. S. balance of payments is that while the direct impact of these measures was favorable on the particular items that they were designed to influence, the indirect or perverse effects on other items of the balance of payments tended to balance the direct effect, so as to render these controls virtually ineffective as instruments to improve the U. S. balance of payments.
BehrmanJack N.“Assessing the Foreign Investment Controls” Law and Contemporary Problems Vol. 34 (Winter1969) pp. 84–94.
BransonWilliam H.Financial Capital Flows in the U. S. Balance of Payments (Amsterdam1968).
BransonWilliam H. “Monetary Policy and the New View of International Capital Movements,” Brookings Papers on Economic Activity: 2 (1970) pp. 235–62.
BransonWilliam H. and RaymondD.HillJr.“Capital Movements in the OECD Area: An Econometric Analysis”OECD Economic Outlook: Occasional Studies (ParisDecember1971).
BransonWilliam H. and Thomas D.Willett “Policy Toward Short-Term Capital Movements: Some Implications of the Portfolio Approach,” in International Mobility and Movement of Capitaled. byFritzMachlupWalter S.Salant and LorieTarshisNational Bureau of Economic Research (Columbia University Press1972) pp. 287–310.
BrimmerAndrew F. “Capital Outflows and the U. S. Balance of Payments: Review and Outlook,” paper delivered at Federal Reserve Bank of Dallas Dallas TexasFebruary111970.
BrimmerAndrew F. “Commercial Bank Lending Abroad and the U. S. Balance of Payments,” in The International Monetary System in Transition a symposium sponsored by the Federal Reserve Bank of ChicagoMarch16–171972 pp. 76–91.
BrimmerAndrew F.“American International Banking: Trends and Prospects”paper delivered at the 51st Annual Meeting of the Bankers Association for Foreign Trade, Boca Raton, FloridaApril21973.
BryantRalph C. and Patric H.HendershottFinancial Capital Flows in the Balance of Payments of the United States: an Exploratory Empirical StudyPrinceton Studies in International Finance No. 25 (Princeton University Press1970).
BryantRalph C. and Patric H.Hendershott “Empirical Analysis of Capital Flows: Some Consequences of Alternative Specifications,” in International Mobility and Movement of Capitaled. byFritzMachlupWalter S.Salant and LorieTarshisNational Bureau of Economic Research (Columbia University Press1972) pp. 207–40.
EllicottJohn “United States Controls on Foreign Direct Investment: The 1969 Program,” Law and Contemporary Problems Vol. 34 (Winter1969) pp. 47–63.
Federal Reserve System Board of GovernorsFederal Reserve Bulletin various issues and Annual Report (19651967and1968).
HerringRichard and Thomas D.Willett “The Capital Control Program and U. S. Investment Activity Abroad,” Southern Economic Journal Vol. 39 (July1972) pp. 58–71.
HewsonJohn and EisukeSakakibara (1973 a) “A General Equilibrium Approach to the Euro-Dollar Market” (unpublished1973).
HewsonJohn and EisukeSakakibara (1973 b) “The Effect of U.S. Controls on U.S. Commercial Bank Borrowing in the Euro-Dollar Market and the Euro-Dollar Interest Rate” (unpublishedInternational Monetary FundDecember281973).
HewsonJohn and EisukeSakakibaraThe Euro-Currency Markets and Their Implications: A “New” View of International Monetary Problems and Monetary ReformChapter 2 (Lexington, Massachusetts1975).
HolbikKarel “United States Experience with Direct Investment Controls,” Weltwirtschaftliches Archiv Band 108 Heft 3 (1972) pp. 491–513.
HufbauerG. C. and F. M.AdlerOverseas Manufacturing Investment and the Balance of PaymentsU. S. Treasury, Tax Policy Research Study No. 1 (Washington1968).
International Monetary FundInternational Financial Statisticsvarious issues.
KwackSung Y. “A Model of U.S. Direct Investment Abroad: A Neoclassical Approach,” Western Economic Journal Vol. 10 (December1972) pp. 376–83.
LafferArthur B. “The U. S. Balance of Payments—A Financial Center View,” Law and Contemporary Problems Vol. 34 (Winter1969) pp. 33–46.
MillerNorman C. and Marina v.N.Whitman “A Mean-Variance Analysis of United States Long-Term Portfolio Foreign Investment,” Quarterly Journal of Economics Vol. 84 (May1970) pp. 175–96.
Morgan Guaranty Trust Company of New YorkWorld Financial Marketsvarious issues.
Organization for Economic Cooperation and DevelopmentMain Economic Indicatorsvarious issues.
PrachownyMartin Frederick JacobA Structural Model of the U. S. Balance of Payments (Amsterdam1969).
ScaperlandaAnthony E. and Laurence JayMauer “The Impact of Controls on United States Direct Foreign Investment in the European Economic Community,” Southern Economic Journal Vol. 39 (January1973) pp. 419–23.
SpitällerErich “A Survey of Recent Quantitative Studies of Long-Term Capital Movements,” Staff Papers Vol. 18 (March1971) pp. 189–220.
StevensGuy V. G. “Capital Mobility and the International Firm,” in International Mobility and Movements of Capitaled. byFritzMachlupWalter S.Salant and LorieTarshisNational Bureau of Economic Research (Columbia University Press1972) pp. 323–53.
Mr. Hewson was an economist in the North American Division of the Fund’s Western Hemisphere Department when this paper was prepared. He is a graduate of the University of Sydney and the University of Saskatchewan, and received his doctorate from the Johns Hopkins University. He is presently Visiting Economist, Reserve Bank of Australia.
Mr. Sakakibara, economist in the Exchange Rate Practices Division of the Exchange and Trade Relations Department, is a graduate of the University of Tokyo and received his doctorate from the University of Michigan. He is presently on leave of absence from the Japanese Ministry of Finance.
In addition to colleagues in the Fund, the authors are grateful to L. Girton, D. Henderson, and P. Clark of the Federal Reserve Board for their helpful comments and suggestions.
In announcing the devaluation of the dollar in February 1973, Secretary of the Treasury George Shultz indicated that these controls would be phased out by the end of 1974, adding, “The phasing out of these restraints is appropriate in view of the improvement which will be brought to our underlying payments position by the cumulative effect of the exchange rate changes, by continued success in curbing inflationary tendencies, and by the attractiveness of the U. S. economy for investors from abroad. The termination of the restraints on capital flows is appropriate in the light of our broad objective of reducing governmental controls on private transactions.” (See “Statement on Foreign Economic Policy,” February 12, 1973.) As the U. S. payments situation improved more rapidly than expected during 1973, these controls were partially relaxed in November and December. In the Treasury Department press release (S-355) announcing the termination of controls on January 29, 1974, the basis for the decision was expressed in the following terms: “The actions are appropriate in light of the recent improvements in the U. S. balance of payments position, the strong position of the dollar in the exchange markets, and the desirability of avoiding official restrictions on the flow of capital to points of need at a time when the balance of payments positions of many countries have been sharply changed by the repercussions of the higher oil prices.”
As the availability of data (particularly, consistent series of Euro-dollar interest rates) restricted the empirical analysis of this paper to beginning in 1964, it was impossible to examine the direct impact of the IET on capital flows. However, available empirical evidence suggests that the IET had a significant impact on capital flows during the period July 1963–September 1964 (see Branson (1968) and Branson and Hill (1971)). There seemed little doubt during this period that the IET would be enacted, although there was some uncertainty regarding the final form and pervasiveness of the tax. The empirical investigation presented in this paper attempts to quantify the offsets that have been suggesed to have followed the enactment of the IET. (See reference to Federal Reserve Board Governor Andrew F. Brimmer in the text.)
See Brimmer (1970), p. 6.
Brimmer (1970), p. 2.
For example, as announced on March 3, 1965, the guidelines requested that banks hold loans and other foreign assets subject to the program to 105 per cent of the amount of credits outstanding on the base date of December 31, 1964. Similarly, for nonbanks the guidelines suggested that liquid funds held abroad, apart from minimum working balances, should be limited to the amount held at the end of 1964, and that they be reduced in an orderly manner to the December 1963 level. Further, it was requested that loans and investments with maturities of five years or less be held to 105 per cent of the amount of such loans and investments outstanding on December 31, 1964. See Board of Governors of the Federal Reserve System, Annual Report, 1965 (Washington, 1966), p. 23.
See Brimmer (1970), p. 9.
For a detailed discussion of the various changes in these ceilings, see Board of Governors of the Federal Reserve System, Annual Reports, 1964–72, and Fedral Reserve press releases for the more recent changes.
Federal Reserve System, Annual Report, 1968, p. 57.
With effect from January 1, 1968, banks were requested “to reduce their ceilings monthly by the amount of scheduled repayments of term loans outstanding on December 31, 1967, to developed countries of continental Western Europe, and further, during the course of the year, by 40 per cent of the amount of short-term credits to those countries outstanding on December 31, 1967.” In addition, banks were requested “not to make any new loans with maturities in excess of 1 year during 1968” to this area. See Board of Governors of the Federal Reserve System, Annual Report, 1967, pp. 20–21.
This ceiling was announced on December 17, 1969, at ½ of 1 per cent of total assets on December 31, 1968; it related to export loans of original maturity of one year or more that were for $250,000 or more.
See Herring and Willett (1972), p. 59—stringent “to the extent that the proposed voluntary program for 1968 called for the limitation of direct investment from 1967 through 1968 to 100 per cent of the 1962 through 1964 average.”
For the summaries of the FDI regulations, see Behrman (1969), Ellicott (1969), Herring and Willett (1972), and Holbik (1972). Several changes made after 1968 introduced more flexibility into the program without affecting the three basic features.
As Governor Brimmer noted, “Each part of this set of restraints reinforces the others, none of them would be effective without the others. Moreover, the administration of the parts (and this is especially true of the VFCR and the direct investment regulations) is coordinated as much as possible.” See Brimmer (1970), p. 8.
On the impact of the IET, see Branson (1968), Prachowny (1969), and Branson and Hill (1971); on the impact of the 1965 voluntary control programs, see Branson (1970), Miller and Whitman (1970), Branson and Willett (1972), Kwack (1972), and Stevens (1972); and on the 1968 package, see Herring and Willett (1972) and Scaperlanda and Mauer (1973). A useful review of some of the earlier studies of the impact of the capital control programs on long-term flows is found in Spitaller (1971).
See Hufbauer and Adler (1968), p. 13.
Behrman appears to argue implicitly along similar lines when, in discussing the FDI program, he notes that “the foreign financing encouraged by the controls appears to benefit payments by substituting for U. S. dollars. But if funds which would have come to the U. S. market are diverted to Europe or if U. S. capital is pulled into Europe to supply demands abroad, the benefit is reduced.” See Behrman (1969), pp. 90–91.
Furthermore, even if it were possible to create appropriate control variables for each of the control programs, it would be almost impossible econometrically to isolate the separate effects of each program in some of the capital equations estimated. Most investigations of the impact of particular capital control programs on the item of the balance of payments that they were designed to influence also have relied on dummy variables, even though the “package effect” is less important in these studies—see, for example, Kwack (1972), Stevens (1972), and Scaperlanda and Mauer (1973). However, Bryant and Hendershott (1970 and 1972), in a study of Japanese private borrowing from U. S. banks, attempted to construct a variable to capture the “restrictiveness” of the VFCR program.
As stock figures are not available for the errors and omissions item, it was decided to use an accumulated errors and omissions variable. A constant difference would not matter in a linear regression.
The justification for the various constraint variables in the bank-claims equation was discussed more fully in the paper on U. S. commercial bank borrowing in the Euro-dollar market; see Hewson and Sakakibara (1973 b).
Those with an original maturity of one year or less.
For a detailed analysis of U. S. banks’ borrowing from their foreign branches, see Hewson and Sakakibara (1973 b).
With respect to the Department of Commerce classification of balance of payments items (for references, see Table 2, Survey of Current Business, September 1973, p. 42), these dependent variables (in flow terms) are defined as follows:
BC = line 42 + line 43
TC = BC + line 45 + line 46
LL = line 54
OL = line 52
E and O = line 64
TK = line 38 − (line 47 − line 48 − line 55 − line 56 − line 57)
In calculating stock data, the relevant flows were subtracted from the corresponding outstandings at the end of March 1973 (outstandings were taken from Survey of Current Business, June 1973).
In the equation for total capital, the CAPC2 variable was assumed to start in the second quarter of 1968 to reflect the probable lag in adjustment of long-term capital flows to the imposition of the 1968 package.
Governor Brimmer has emphasized the offset to the IET through increases in bank borrowing; see Brimmer (1970), p. 6.
As indicated by the sign of CAPC1 in equation (9), there also may have been some offsetting capital movements through other short-term liabilities. However, this coefficient was not significantly different from zero.
See Brimmer (1970), p. 16.
Ibid., p. 18. As Governor Brimmer further noted, “… one of the considerations underlying the decision to provide more leeway under the VFCR for export financing” was the undesirable implication of the program, as initially instituted, for future export growth.
The inclusion of the control variables in equation (13) produced the following result:
Although the Euro-dollar call rate was used in the equations for total short-term claims, liquid liabilities, and errors and omissions, the three-month Eurodollar rate was used in the equation for other short-term liabilities. Similarly, although the corporate AAA bond rate was used as “the” U. S. rate in the equation for other short-term liabilities, the federal funds rate was used in other equations for short-term capital flows.