INTERNATIONAL RESERVES are held chiefly for the purpose of enabling the authorities of a country to maintain a fixed foreign exchange rate during periods of balance of payments deficit. Thus, the level of reserves which the authorities desire to hold is likely to be related to the expected magnitude of potential future deficits. Other things being equal, expected potential deficits will tend to be larger (1) the greater are the external or internal disturbances affecting the balance of payments with which the country’s authorities expect to be faced as a result of events beyond their control, (2) the larger are the gross transactions affected by these disturbances, and (3) the less inclined are the authorities to modify domestic objectives or accustomed economic policy designs in order to adjust the balance of payments.

Abstract

INTERNATIONAL RESERVES are held chiefly for the purpose of enabling the authorities of a country to maintain a fixed foreign exchange rate during periods of balance of payments deficit. Thus, the level of reserves which the authorities desire to hold is likely to be related to the expected magnitude of potential future deficits. Other things being equal, expected potential deficits will tend to be larger (1) the greater are the external or internal disturbances affecting the balance of payments with which the country’s authorities expect to be faced as a result of events beyond their control, (2) the larger are the gross transactions affected by these disturbances, and (3) the less inclined are the authorities to modify domestic objectives or accustomed economic policy designs in order to adjust the balance of payments.

INTERNATIONAL RESERVES are held chiefly for the purpose of enabling the authorities of a country to maintain a fixed foreign exchange rate during periods of balance of payments deficit. Thus, the level of reserves which the authorities desire to hold is likely to be related to the expected magnitude of potential future deficits. Other things being equal, expected potential deficits will tend to be larger (1) the greater are the external or internal disturbances affecting the balance of payments with which the country’s authorities expect to be faced as a result of events beyond their control, (2) the larger are the gross transactions affected by these disturbances, and (3) the less inclined are the authorities to modify domestic objectives or accustomed economic policy designs in order to adjust the balance of payments.

Additions to a country’s international reserves widen its scope for the pursuit of domestic policy objectives in times of external deficit. However, the holding of reserves also entails a cost in terms of the foregone alternative foreign or domestic assets which the residents of a country could have acquired. The level of desired reserves, i.e., the reserves which appear appropriate to the authorities and which they aim to achieve, will tend to reflect—however imperfectly—the balance of these considerations relating to costs and benefits of an incremental amount of reserves.

When desired reserves deviate appreciably from actual holdings, the authorities will sooner or later readjust their economic policies so as to reduce the magnitude of the deviation. If an increase in reserves appears desirable, they will pursue less expansionary or more contractionary financial policies, impose restrictions on external trade and payments or increase the severity of existing restrictions, or depreciate the exchange rate; if actual reserves exceed the desired level, they will follow more expansionary or less contractionary financial policies, reduce the severity of restrictions on trade and payments, or appreciate the exchange rate. Choice of these policies or combination of policies is likely to depend not only on general attitudes toward, and obligations with respect to, these alternative policy instruments but also on the requirements of the domestic economic situation.

In any discussion of the prospective adequacy of international liquidity, some consideration has to be given to probable future trends in countries’ desired reserves. By providing estimates of the trend in payments imbalances over a recent past period, this paper presents some background material relevant in this context.1

Payments Imbalances of Individual Countries and of All Countries Combined

On the assumption that the priorities given in individual countries to domestic and external objectives of economic policy and the attitudes toward the use of various policy instruments remain unchanged, desired reserves would tend to rise chiefly as a result of the increase in the size of expected payments fluctuations. In this context, the argument has been advanced that payments imbalances, and thus the prospective external deficits for which the national authorities must be prepared, should be expected to increase with the growth of international transactions.2 In a recent contribution, Mr. Weir M. Brown questioned this hypothesis on empirical grounds.3 On the basis of an examination of the over-all surpluses and deficits of eleven major industrial countries for the period 1953-62,4 he concluded that “two or possibly three of the eleven countries under review showed some increase in the absolute size of their net foreign balances within the ten-year period, though some of these fell again; and the majority of countries in the Group showed a stable or declining trend.” 5 Brown’s sample of countries is very narrow—it includes neither developing countries nor developed nonindustrial countries—and his method (inspection of the data on imbalances for each individual country in turn) also leaves room for improvement.

Over a relatively short period, any trend which, other things being equal, may exist in the size of payments fluctuations of an individual country is likely to be obscured by the effects of a multitude of other factors on the country’s balance of payments. For an individual country, certain events (such as a sudden spurt in domestic demand, fluctuations in price or output of one of the country’s major export products, emergence and subsequent reversal of inflationary expectations, or the sharp impact of policies designed to promote domestic stability) may have given rise to peak deficits or surpluses in one year or another at the beginning, at the end, or in the middle of the period under study. As a result, these events may produce the appearance of a falling, rising, or broken trend in the country’s payments imbalances, or—more likely—they may be reflected in a time series of such erratic appearance as to suggest the absence of any statistically significant trend. When payments imbalances for a large number of countries are added together, unusually large surpluses or deficits of individual countries in particular years tend to be submerged in the total so that any upward or downward trend in the sum of imbalances will be more prominently revealed. For this reason, an aggregative approach to the assessment of any tendency of the size of payments imbalances to change over time would seem more promising than computation of separate trends for individual countries.

Two methods, both reflecting this consideration, are used in this study. The first determines the statistical trend in aggregate payments imbalances, that is, in the sum of payments imbalances for all countries or a subgroup of countries. In the second approach, the data for payments imbalances of individual countries are scaled (so as to make them comparable) and pooled in order to arrive at a measure of the trend of payments imbalances of the typical, or average, country. These two approaches will be described in turn in the two following sections.

Unless otherwise indicated, the term “payments imbalances” is used in this paper to indicate over-all balance of payments surpluses or deficits without regard to sign, i.e., both surpluses and deficits are considered with a positive sign. These imbalances are measured by the sum of changes in official gold and foreign exchange assets, in net IMF positions,6 in liabilities resulting from swap transactions with the United States, and, for the United States and the United Kingdom, in other liabilities to foreign monetary authorities; the balances are further adjusted for advance debt repayment and certain other items of extraordinary financing.7

The statistical analysis of this paper covers 65 countries for which annual data on payments imbalances were available for the period 1952-64. These 65 countries account for about 90 per cent of world trade, measured by imports, and for a slightly higher share of total world reserves. Some of the calculations are also carried out with data for a subgroup of 25 “developed countries.” 8

Owing to the consideration of a somewhat longer time period and a much larger number of countries than were taken by Brown, and perhaps also as a result of the different method of investigation, the conclusion of this paper differs from the one reached by Brown. In particular, it is found that payments imbalances of the 65 countries and of the subgroup of 25 developed countries have indeed exhibited a statistically significant rising trend over the period 1952-64.

First Approach: Aggregate Payments Imbalances

For the 65 countries of the study and for the subgroup of 25 developed countries, three sets of time series are computed for the years 1952-64: (1) the sum of deficits of all countries having deficits in the year in question; (2) the sum of surpluses of all countries having surpluses; and (3) the sum of “payments imbalances,” i.e., of deficits and surpluses without regard to sign. These time series are shown in Chart 1. (The data are given in Appendix II, Table 5, p. 394.) The dollar amount of the sum of payments imbalances—involving, as it does, “double counting” of corresponding surpluses and deficits—cannot be given a direct economic interpretation, but it can serve as an indicator of the degree of payments imbalance in the group of countries under consideration.9

Chart 1.
Chart 1.

Sums of Actual Annual Payments Imbalances, 1952–64 1 (Including United States 2)

(In billions of U.S. dollars)

Citation: IMF Staff Papers 1966, 003; 10.5089/9781451969108.024.A001

1 Source: Appendix II, Table 5.2 U.S. data: Over-all balance.

Two types of trend have been computed from these time series by least-squares regression, that is to say, by the method determining the trend line of best statistical fit. One is a linear trend which indicates the average annual increase in terms of the units of measurement, i.e., millions of U.S. dollars. The other trend computation indicates the average percentage increase, or growth rate, of the series in question. The results of these computations are shown under 1(a) in Table 1. During 1952-64, aggregate deficits, surpluses, and total imbalances of the 65 countries of the study taken together grew at the rate of about 6½ per cent a year, and those of the 25 developed countries at rates of 7 to somewhat more than 9 per cent. In dollar terms, surpluses and deficits of the 25 developed countries, as well as all 65 countries, each increased by somewhat more than US$200 million a year, and the corresponding sums of all imbalances, of course, increased by about twice that figure.

Table 1.

Growth of Aggregate Payments Imbalances, 1952–64 1

article image
article image

This table shows the results of regression equations of imbalances, or their logarithms, on a trend term. An asterisk (*) indicates that the coefficient to which it is attached is not statistically significant at the 5 per cent level, i.e., the probability that the true coefficient is zero is greater than 1 out of 20.

The sum of imbalances is the sum of all deficits and surpluses added without regard to sign.

The derivation of the hypothetical imbalances is described in the text.

The 3-year imbalances are the algebraic sums of annual imbalances in 3 consecutive years divided by 3, i.e., the 3-year moving averages of annual imbalances taken with regard to sign. These computations cover the years 1953-63.

Over the same period, aggregate imports of the 65 countries grew by 6.2 per cent a year, and those of the 25 developed countries by 6.7 per cent.10 From these computations, therefore, it may be concluded that payments imbalances rose roughly in step with the total value of trade.

In considering these results, the question occurs as to what extent they are dominated by the rise over the period studied in the magnitude of the payments imbalance of the United States.11 For this reason, a number of computations have been made with the aim of reducing the influence of the U.S. balance of payments on the sum of imbalances of all countries, though a complete elimination of this influence is hardly possible. U.S. surpluses and deficits affect the data in two ways. First, the amounts of these imbalances are added to the sums of surpluses or deficits of all countries. This effect can easily be removed from the time series by omitting the imbalances of the United States. Second, U.S. imbalances tend to be reflected in the surpluses and deficits of the remaining countries: e.g., the larger the U.S. deficit, the larger will tend to be the surpluses, and the smaller will tend to be the deficits, of other countries included in the study. Correction of the data for this second influence is more difficult and requires some hypothesis concerning the effect of a change in the U.S. balance of payments on the balances of payments of countries other than the United States.

The first influence of the U.S. balance of payments on aggregate surpluses and deficits is modified or removed in the computations shown under 1(b) and (c) in Table 1. In one computation, the U.S. surplus or deficit is measured by the movement in its gross reserves (gold, foreign exchange holdings, and net IMF position), adjusted for advance debt repayment but excluding any changes in U.S. liabilities to foreign official holders. The economic justification for this modification of the data is based on the view that, to the extent of any increase in these liabilities, the United States may be considered part of the reserve-creating mechanism of the international monetary system; in this view, a rise in U.S. liabilities is no more regarded as part of the U.S. deficit than the sale of newly produced gold is counted as part of the deficit of a gold-producing country. In an alternative computation, payments imbalances of the United States are left out of account entirely. In this computation, the data on payments imbalances are the sums of 64 countries and 24 developed countries, respectively; they are shown in Chart 2. (Both sets of data are given in Appendix II, Table 5.)

Chart 2.
Chart 2.

Sums of Actual Annual Payments Imbalances, 1952–64 1

(Excluding United States)

(In billions of U.S. dollars)

Citation: IMF Staff Papers 1966, 003; 10.5089/9781451969108.024.A001

1 Source: Appendix II, Table 5.

In both these alternative computations, the growth of total aggregate imbalances for all countries is about 5 per cent a year, and that of surpluses 7 per cent when U.S. imbalances are included on the basis of gross reserve changes and about 8 per cent when the data for the United States are excluded. The corresponding imbalances of the developed countries grew at a somewhat higher rate than those of all countries. The growth rates of the sums of deficits, however, are much smaller (in one instance even negative) and, furthermore, are not significantly different from zero in the statistical sense.

Neither of these two procedures intended to minimize the influence of the U.S. data on the sum of payments imbalances is very satisfactory. Modification of the U.S. data by defining surpluses and deficits as movements in gross reserves, though not entirely implausible, introduces an arbitrary asymmetry with respect to the concept of payments imbalance. One could with equal justification alter the data for any country by omitting that part of an external deficit which the country was able to finance without making any net use of its international reserves, say, by drawing from the Fund in the credit tranches or using swap facilities. Moreover, when increases in officially held dollar balances are not counted as part of any U.S. deficit, the discrepancy between the sum of deficits and the sum of surpluses widens, and the sum of deficits, taken by itself, ceases to be an adequate measure of the degree of world payments imbalance. For countries other than the United States, any tendency toward an increase in the sum of deficits resulting from the growth in international transactions might well fail to manifest itself in a particular historical period, if it were offset by the effect of a rise in U.S. net expenditures abroad financed by an increase in foreign official holdings of U.S. dollars.12

An even stronger objection can be made against the use of data series for the sums of deficits, surpluses, and total imbalances from which the U.S. data have been omitted entirely. The resulting time series describe a notional historical sequence which could not have occurred in reality. As already indicated, if the assumption is made that the U.S. balance of payments had been in balance in every year of the period under consideration, it must also be assumed that the imbalances of some or all other countries would have been different from those actually observed. There is no way in which this alternative balance of payments history of countries other than the United States could be deduced in any even remotely realistic manner, even if one were to construct a very detailed and comprehensive model of the world economy.

It is possible, however, to construct time series for hypothetical payments imbalances that are at least intuitively plausible and logically consistent with the assumption that the U.S. balance of payments had been in balance. Two such sets of data have been generated by allocating the actually observed over-all U.S. deficits and surpluses among the other 24 developed countries in proportion to (1) their average reserves and (2) their average imports. This procedure amounts to assuming that the United States had a more (or less) favorable overall balance of payments with each of these 24 developed countries in such a way that its over-all deficit (or surplus) was eliminated, and that the 24 countries shared in the worsening (or improvement) of their consolidated balance of payments with the United States in proportion to their average reserves or average imports.13 The payments imbalances of the 40 less developed countries are assumed to be unaffected.

It turned out that the sums of hypothetical imbalances computed on these two alternative bases differed only slightly. The data for the sums of imbalances resulting if the U.S. balance of payments is allocated on the basis of average reserves of the other developed countries are given in Appendix II, Table 5, and are shown in Chart 3.

Chart 3.
Chart 3.

Sums of Hypothetical Annual Payments Imbalances, 1952-641 (U.S. Over-All Balance Allocated to 24 Developed Countries by Average Reserves)

(In billions of U.S. dollars)

Citation: IMF Staff Papers 1966, 003; 10.5089/9781451969108.024.A001

1 Source: Appendix II, Table 5.

The results of the trend computations from these hypothetical imbalance data are listed under 2(a) and (b) in Table 1. The computed growth rates are identical, at the degree of rounding shown in the table, for the two sets of hypothetical data generated under the alternative allocation principles. The sum of all imbalances and the sum of surpluses increased by roughly 4 per cent a year for both the entire group of 64 countries and the subgroup of 24 developed countries. The sum of deficits for the 24 developed countries grew by 6 per cent a year and for all 64 countries by somewhat less than 4 per cent.14

In view of the relatively crude method of deriving these hypothetical data on payments imbalances, it would not be appropriate to place much reliance on these computed trends. They do indicate, however, that total imbalances, i.e., the sums of deficits and surpluses, still show a statistically significant upward trend after the dominant influence on the data series of the U.S. balance of payments has been removed.

The trend computations discussed so far refer to any tendency of the magnitude of annual imbalances to grow over time. Payments imbalances, however, may change in magnitude not only through changes in the size of annual surpluses or deficits but also as a result of a lengthening or shortening of the period during which countries are either in deficit or in surplus. The period under study, encompassing only 13 years, is not sufficiently long for a direct assessment of any tendency of the duration of periods of surplus or deficit to have, on average, increased or decreased. An indirect approach was, therefore, chosen by comparing trends in annual payments imbalances with those in imbalances during overlapping periods of somewhat greater length. For this purpose, surpluses and deficits for each set of 3 successive years were summed with proper regard to sign, i.e., in such a manner that surpluses and deficits offset one another. These 3-year sums were divided by 3 to obtain annual averages for the years 1953-63. The resulting average overlapping 3-year imbalances were then aggregated without regard to sign as in the case of the original annual imbalances.15 The resulting data are shown in Chart 4 and listed in Appendix II, Table 6.

Chart 4.
Chart 4.

Sums of Actual Three-Year Payments Imbalances, 1953-631

(Including United States 2)

(In billions of U.S. dollars)

Citation: IMF Staff Papers 1966, 003; 10.5089/9781451969108.024.A001

1 Source: Appendix II, Table 6. Data are expressed at annual rates.2 U.S. data: Over-all balance.

A comparison of the values plotted in Charts 1 and 4, and of corresponding columns in Tables 5 and 6, indicates that the 3-year imbalances (expressed as annual amounts) are smaller than the annual imbalances, reflecting the offsetting of surpluses and deficits for individual countries within each 3-year period. On average over all years considered in this study, the 3-year imbalances are about two thirds of the annual imbalances. The sum of imbalances taken over a full 3-year period centered on a particular year is 3 times the amount actually shown for that year in Chart 4, Chart 5 (referred to below), and Table 6 and is thus, on average over the entire period, about twice the amount of the corresponding sum of annual imbalances given in Chart 1 and Table 5.

Chart 5.
Chart 5.

Sums of Hypothetical Three-Year Payments Imbalances, 1953–63 1

(U.S. Over-All Balance Allocated to 24 Developed Countries by Average Reserves)

(In billions of U.S. dollars)

Citation: IMF Staff Papers 1966, 003; 10.5089/9781451969108.024.A001

1 Source: Appendix II, Table 6. Data are expressed at annual rates.

The trend computations for these 3-year imbalances are given under 3(a) and (b) in Table 1. The trend growth of all imbalances and of surpluses alone are considerably higher than the corresponding growth rates computed for annual imbalances; they range between 9 and 12 per cent a year, and these figures are not much affected by the omission of the U.S. data from the computations. Leaving aside the results for the sums of deficits excluding U.S. deficits,16 a comparison of line 3(a) with the corresponding line 1(a), and line 3(b) with 1(c), shows clearly that the 3-year imbalances grew much faster than the annual imbalances. For all 65 countries, the 3-year imbalances rose at rates of about 11–12 per cent, and for the 25 developed countries at rates of 12 per cent for surpluses and 15 per cent for deficits. The growth of 3-year surpluses for countries excluding the United States was somewhat higher and, again, exceeded the corresponding growth rate computed from annual surplus data. This faster growth rate suggests that imbalances did indeed increase during the period under study not only as a result of a rise in the annual magnitude of surpluses or deficits but also because of an increasing tendency for surpluses and deficits to persist for a larger number of consecutive years.

Hypothetical data series were also derived for 3-year imbalances by allocating the U.S. imbalances to the other 24 developed countries alternatively on the basis of average reserves and average imports, the former being shown in Chart 5 and in Appendix II, Table 6. The trend computations in Table 1 under 4(a) and (b), in comparison with those for hypothetical annual imbalances under 2(a) and (b), on the whole confirm the impression obtained from the calculations with actual annual and 3-year imbalances. Hypothetical 3-year imbalances of all 64 countries grew by about 5 per cent a year compared with 4 per cent for hypothetical annual imbalances. The sum of hypothetical 3-year surpluses of the 24 developed countries also grew more rapidly than that of hypothetical annual surpluses, whereas the trend growth of the corresponding 3-year deficits is statistically more uncertain and no larger than that of the annual deficits.

These results are evaluated in the concluding section of this paper.

Second Approach: Average Trend in Individual Imbalances

In the second approach to the determination of the trend in payments imbalances, the data on individual countries are pooled. Data for surpluses and deficits are not separated. In order to remove the influence of the size of countries and their external transactions, the imbalances—surpluses or deficits, both taken with a positive sign—are scaled by expressing them as percentages of average reserves over the 13-year period under study. For the group of 65 countries, 845 observations are obtained (65x13). The regression of these observations on a linear trend variable, which runs from 1 to 13 for each of the 65 countries, shows the trend increase, for the average (or typical) country, in the ratio of its annual imbalance to its average reserves.

These computations have again been carried out for the sample of all countries with actual annual and 3-year imbalances, and with hypothetical annual and 3-year imbalances. The results are given in Table 2. Over the period 1952-64, the trend value of the ratio of actual annual imbalances to average reserves for the 65 countries rose by about three fourths of 1 per cent per annum, i.e., imbalances showed an annual upward trend of three fourths of 1 per cent of average reserves. The same result is obtained when the U.S. data are excluded from the computation. Since average reserves for all countries are about 7 times as high as the 1952–64 average of the sums of all countries’ imbalances, the implication is that the latter grew at a rate of somewhat more than 5 per cent (7x3A per cent). This result is thus roughly consistent with that obtained with the aggregate approach.

Table 2.

Average Annual Trend Increase in Payments Imbalances, 1952–64 1

(In per cent of average reserves at end of year, 1951-63)

article image

These computations, in which the data for all countries are pooled, show the annual trend increases, on average or for the “typical” country, in the ratio of payments imbalances to average reserves. For instance, the first regression coefficient shown can be interpreted to mean that, on average, actual annual payments imbalances of the 65 countries included in the study have increased by somewhat less than three fourths of 1 per cent of each country’s average reserves (end of year, 1951-63) per annum. All regression coefficients shown in this table are statistically significant at the 1 per cent level.

The same result is obtained if the United States is omitted from the computation.

The U.S. over-all balance is allocated among the other 24 developed countries of the study on two alternative bases: in accordance with average reserves (end of year, 1951-63) or with average imports (1952-64).

The 3-year imbalances are the algebraic sums of annual imbalances in 3 consecutive years divided by 3, i.e., the 3-year moving averages of annual imbalances taken with regard to sign. These computations cover the years 1953-63.

The trend increase of hypothetical annual imbalances, with the U.S. over-all surpluses or deficits allocated to the other 24 developed countries on the basis either of average reserves or of average imports, was somewhat less than two thirds of 1 per cent. The corresponding computation with actual 3-year imbalances shows an annual trend increase of almost 1 per cent of average reserves, and those with hypothetical 3-year imbalances result in trend increases of three fourths to four fifths of 1 per cent of average reserves. All these trend coefficients are statistically highly significant.

Summary and Conclusion

Though payments imbalances were clearly larger toward the end of the period examined than they were at the beginning, Charts 1-5 show that their growth was not nearly uniform. Actual imbalances of all countries including the United States declined over the last 4 years for which data are available, and hypothetical imbalances of countries other than the United States remained roughly unchanged during the second half of the period studied. Only the future can tell whether the trend itself has changed in the middle of the period considered in this paper or whether recent observations are merely large negative deviations from an essentially unbroken upward trend. In this connection, it must be emphasized that the conclusions as to trends in payments imbalances presented here relate only to the historical period from which the data of the study have been taken; they do not purport to be implicit forecasts of future developments of payments imbalances.

The two alternative approaches to the determination of the rate of growth of payments imbalances chosen in this paper yield compatible results and show that payments imbalances (i.e., surpluses and deficits taken without regard to sign) exhibited a significant upward trend over the period 1952-64. Essentially the same conclusion emerges when the growth of surpluses alone is examined. Deficits, taken separately, grew at rates similar to those of all imbalances when U.S. over-all deficits are included in the computations. When the influence of the U.S. balance of payments on aggregate deficits is partly removed, by omission either of the over-all U.S. deficits or of that part of any deficit which was financed through an increase in U.S. liabilities to foreign official holders, the aggregate value of deficits no longer shows a clear and unambiguous rise during the period studied. As has been pointed out above, outright deletion of the U.S. deficits or any part thereof is arbitrary and logically inconsistent. For this reason, not much weight should be given to the results obtained from the series from which the U.S. imbalances have been, entirely or partly, omitted without any further adjustment of the remaining data. The summary of results given in Table 3 is, therefore, based on the computations from data which, alternatively, include the United States or allocate U.S. imbalances among the other developed countries. Where the rates of growth for surpluses, deficits, and all imbalances are roughly the same, only one rate is shown in Table 3; where they are substantially different, the range in which these rates of growth fall is indicated.

Table 3.

Annual Growth Rates of Payments Imbalances, Imports, and Reserves, 1952-64

(In per cent)

article image

U.S. imbalances allocated among the other 24 developed countries on the basis of average reserves.

Source, Table 1. Rates or range of rates are computed separately for aggregate deficits, surpluses, and the sum of deficits and surpluses combined without regard to sign.

Source, Table 2. The trend increase in payments imbalances for the “average” country expressed in Table 2 as percentage of its reserves has been multiplied by the ratio of total reserves to total imbalances (averaged for 1952-64) for all countries, or all countries excluding the United States, to yield this alternative estimate of the growth rate of imbalances.

Growth rates for 1953-63.

For the sample of 65 countries, the growth rate of annual imbalances was between 3½ and 6½ per cent, the higher figure applying to the actual data including those of the United States, and the lower to computations made with hypothetical data constructed to remove the dominant effect which the sharp upward trend in the U.S. deficit had on the sum of payments imbalances of all countries or all developed countries. For the developed countries as a group, the proportionate increase in imbalances tended to be somewhat larger (5 to 8 per cent).17

When balance of payments surpluses or deficits for overlapping 3-year periods are considered, imbalances are found to have grown at annual rates ranging from 5 to almost 12 per cent for all countries and from 5 to 13 per cent for the developed countries,18 with the higher values again applying to the data including the United States and the lower ones to hypothetical data constructed by allocating U.S. imbalances to the other developed countries. The faster growth of 3-year imbalances than of annual imbalances indicates that surpluses or deficits not only increased in magnitude but also tended to be, on average, of somewhat longer duration toward the end of the period studied than they were at the beginning.

Over the period 1952–64 the annual growth of international trade, as measured by aggregate imports, was approximately 6 to 7 per cent (see Table 3). The estimated rates of increase in payments imbalances cluster around this range for the rate of growth of imports, though some of the estimates lie somewhat below it and others above. While there is, thus, no reasonable doubt that payments imbalances showed a rising trend over the period 1952-64, some uncertainty remains concerning the relation of this trend to the growth of external transactions. Nevertheless, the weight of the evidence supports the contention that, for the world as a whole, payments imbalances measured over 3-year periods have grown at least as rapidly as imports. Moreover, even the one estimate—that of the growth of 3-year imbalances of all countries excluding the United States—which lies below the growth rate of imports does not fall short of it by a statistically significant amount, considering the probable error attaching to the estimate. Under the assumption that the growth of trade can be taken as an indicator of the increase in all external transactions, the statistical results do not, therefore, contradict the hypothesis that, in the absence of major changes in countries’ balance of payments adjustment policies, payments imbalances tend to increase in proportion to gross transactions.

International reserves of all 65 countries of the study rose over the period studied by 2.5 per cent a year. This low rate of increase reflects, however, the large reduction in U.S. reserves. For all countries of the study excluding the United States, the reserves grew by 6.0 per cent a year. Leaving aside the loss of reserves by the United States, reserves of all countries appear, therefore, to have grown roughly in proportion to the value of trade and to the size of payments imbalances.

APPENDICES

I. Acceleration of Reserve Increases and Payments Imbalances

Some of the observed payments imbalances, i.e., surpluses and deficits, may be directly related to the increase in reserve assets generated by the international monetary system. For instance, the sum of surpluses must always exceed the sum of deficits by the amount of increase in the stock of monetary gold available to all countries. If the stock of other reserve assets were to remain unchanged, the annual addition to the gold stock would therefore equal the minimum sum of payments imbalances that could occur in the year in question. (For example, if there are no deficits at all, the sum of imbalances equals the sum of surpluses, which in turn equals the increase in the gold stock.) The sum of payments imbalances, of course, may be larger than this minimum whenever some countries are in deficit and transfer existing reserve assets to surplus countries. The minimum sum of annual payments imbalances will rise if the annual increase in the monetary gold stock becomes larger.

These considerations can be extended to apply to changes in all reserve assets. Assume a situation in which (1) there is some increase in the monetary gold stock,19 (2) reserve centers are in over-all deficit and finance all or part of these deficits by increases in their liabilities to foreign monetary authorities, and (3) some Fund drawings in the credit tranches—necessarily by deficit countries—take place. The minimum sum of deficits consistent with a given increase in reserves in the form of (2) and (3) is the sum of deficits of reserve centers financed through increases in their liabilities to foreign monetary authorities (i.e., essentially the increase in the world’s foreign exchange reserves20) plus the sum of deficits financed through drawings in the credit tranches. At the same time, the minimum sum of surpluses consistent with the observed reserve increases under (1), (2), and (3) equals the minimum sum of deficits defined above plus the necessary excess of surpluses over deficits resulting from the increase in the monetary gold stock. The minimum sum of imbalances (surpluses and deficits added together without regard to sign) is obtained by adding minimum deficits and minimum surpluses; it equals the increase in the monetary gold stock plus twice the amount of the increase in foreign exchange holdings and in reserve positions in the Fund arising out of credit tranche drawings. The actual sum of imbalances will exceed this minimum to the extent that deficits other than those financed by additional liabilities of reserve centers and net credit tranche drawings occur and are financed by the transfer of existing reserve assets.

The relevant data are shown in Table 4. Annual averages for three periods of three years each are compared: the years 1953-55, close to the beginning of the period considered in this paper, were characterized by relatively small payments imbalances of the 65 countries of the study, as well as by small U.S. deficits; the period 1958–60 was one of very large payments imbalances and large U.S. deficits; the period 1962–64 is chosen because it includes the last three years of the period studied.

Table 4.

Changes in Reserves and “Minimum” and Actual Sum of Payments Imbalances

(In billions of U.S. dollars)

article image

The data refer to “all countries” as given in International Financial Statistics.

Equals changes in foreign exchange and reserve positions in the Fund excluding those arising from gold subscriptions.

Equals changes in total reserves, i.e., minimum deficits plus the change in gold plus IMF gold subscriptions.

Equals sum of minimum deficits and minimum surpluses.

Includes 65 countries; computed from data in the first three columns of Table 5.

Equals sum of actual deficits and actual surpluses.

Somewhat surprisingly, it is found that reserve increases were smaller in the period of largest imbalances and largest U.S. deficits (1958-60) than in the base period 1953-55. In particular, the increase in foreign exchange balances (as well as the U.S. dollar component of these balances) fell from the earlier to the later period. The rise in surpluses and deficits between these two periods, therefore, cannot be explained in terms of the necessary association between minimum payments imbalances and reserve increases. Annual additions to reserves were somewhat higher during the final three years covered by the study than they had been in the base period, but the maximum increase in payments imbalances that can be directly ascribed to this acceleration in the rate of reserve increase was about $1 billion,21 whereas imbalances rose by more than $4 billion. It may be concluded that the observed rising trend in the sums of payments surpluses and deficits is only to a minor extent, if at all, attributable to an acceleration in the rate of increase in reserve assets.

Table 5.

Sums of Annual Payments Imbalances, 65 Countries and 25 Developed Countries, 1952–64 1

(In billions of U.S. dollars)

article image

The second and third columns under each heading show, respectively, the sums of deficits (taken with a positive sign) of countries having deficits in the particular year and the sums of surpluses (also taken with a positive sign) of countries having surpluses. The first column gives in each case the sums of the figures in the second and third columns, i.e., the sums of all payments imbalances added without regard to sign.

’The hypothetical annual imbalances shown in this table are arrived at by allocating the amount of the U.S. deficit or surplus among the 24 other developed countries in proportion to their average reserves for the period under consideration (year-ends 1951-63). The imbalances of the 40 less developed countries of the study are left unchanged. The resulting hypothetical imbalances of the countries other than the United States (whose imbalances are by assumption zero) are then summed as indicated in footnote 1.

Table 6.

Sums of 3-Year Payments Imbalances, 65 Countries and 25 Developed Countries, 1953–63 1

(In billions of U.S. dollars)

article image

See footnote 1, Table 5. The 3-year imbalance figures are computed by adding, for each individual country, the surpluses or deficits for 3 consecutive years with proper regard to sign (so that deficits offset surpluses). These imbalances for overlapping 3-year periods are divided by 3 (to form annual averages) and are entered under the center year. These data for 1953–63 are then summed as indicated in footnote 1, Table 5.

The hypothetical 3-year imbalances are computed in correspondence with the method used for hypothetical annual imbalances; see footnote 2, Table 5.

Tendances des déséquilibres de paiements, 1952-64

Résumé

Cet article examine la question de savoir si les excédents et déficits globaux des paiements, mesurés par les variations des réserves, ont eu tendance à s’accroître avec le temps et, dans l’affirmative, s’ils ont augmenté à peu près au même rythme que le commerce extérieur. L’article se fonde sur des renseignements concernant 65 pays pour la période 1952-64. L’auteur utilise deux méthodes statistiques : dans la première, il additionne les excédents et les déficits pour tous les pays, ainsi que pour un sous-échantillon de 25 pays développés, et il détermine les taux de croissance des sommes des déséquilibres de paiements en construisant les droites des moindres carrés représentatives des tendances enregistrées. Dans certains de ces calculs, on ajuste les données afin d’éliminer l’influence dominante des importants déséquilibres américains de paiements sur les sommes des déficits et excédents de tous les pays, ou de tous les pays développés, tout en évitant de constituer les séries chronologiques non cohérentes qui en résulteraient si l’on omettait simplement les déséquilibres américains. Dans la seconde méthode, l’auteur rend commensurables les excédents et les déficits de chaque pays en les divisant par les réserves internationales des pays, et il les rassemble pour calculer l’aug-mentation du trend des déséquilibres de paiements pour un pays type.

Par l’une ou l’autre de ces deux méthodes, on aboutit à la conclusion que les déséquilibres de paiements ont accusé, du point de vue statistique, une importante tendance à la hausse en 1952-64. Les excédents et les déficits annuels effectifs augmentèrent à peu prés au même rythme que les importations (6 à 7 pour cent). Les déséquilibres de paiements mesurés sur des périodes triennales se recouvrant partiellement se sont accrus un peu plus vite que les déséquilibres annuels, ce qui suggère que la période au cours de laquelle les balances des pays sont excédentaires ou déficitaires tend à s’allonger progressivement. Les trends déterminés à l’aide de séries chronologiques hypothétiques ajustées pour éviter l’influence de l’accroissement des déficits américains au cours de la période étudiée ont fait ressortir des taux de croissance quelque peu inférieurs à ceux des trends correspondant aux données effectives.

Tendencias en los desequilibrios de pagos, 1952-64

Resumen

Este trabajo examina la cuestión de si con el transcurso del tiempo los superávit y los déficit globales de pagos, midiéndolos de acuerdo con los movimientos de reservas, han mostrado una tendencia a aumentar y, caso de que así sea, si han tendido a aumentar aproximadamente al mismo ritmo que el comercio exterior. El estudio se basa en datos de 65 países para el período 1952-64. En la primera de dos formas de estudiarlos estadísticamente, se suman los superávit y déficit de todos los paises, así como los de una submuestra de 25 países desarrollados, y los ritmos de aumento de esas sumas de desequilibrios de pagos se determinan mediante un ajuste, por mínimos cuadrados, de las líneas indicativas de tendencia. En algunos de estos cómputes se han ajustado los datos para eliminar la influencia predominante de los grandes desequilibrios de pagos estadounidenses sobre las sumas de los déficit y superávit de todos los países, o de todos los países desarrollados, evitando al mismo tiempo la compilacioó de series cronológicas lógicamente incompatibles, que resultarían de omitir sin más los desequilibrios de Estados Unidos. En la segunda forma se escalonan los déficit y superávit de cada país, clasificándolos según sus reservas internacionales, y se los reune en un cómpute que determina el aumento en la tendencia de los equilibrios de pagos en el caso del país “típico”.

Ambas formas llevan a la conclusión de que los desequilibrios de pagos acusaron una tendencia creciente, estadísticamente significativa, en 1952-64. Los superávit y déficit reales de cada año aumentaron aproximadamente al mismo ritmo que las importaciones (entre el 6 y el 7 por ciento). Los desequilibrios de pagos, medidos por períodos trienales superpuestos, aumentaron algo más de prisa que los desequilibrios anuales, lo que sugiere la existencia de una tendencia hacia un alargamiento gradual del tiempo durante el cual los paises se encuentran en superávit o, bien, en déficit. Los cómputos de tendencias realizados usando series cronológicas hipotéticas, ajustadas de modo de eliminar la influencia del aumento de los déficit de Estados Unidos, durante el período estudiado, dieron tasas de crecimiento algo menores que las derivadas de los datos reales.

*

Mr. Rhomberg, Chief of the Special Studies Division of the International Monetary Fund, is a graduate of the University of Vienna and of Yale University and has been a member of the faculty of the University of Connecticut. He has contributed chapters to several books on economic subjects and articles to economic journals.

1

See also H. R. Heller, “A New Measure of International Liquidity,” National Banking Review, U.S. Treasury Department, Vol. 2 (December 1964), pp. 249-55. Heller proposes an index expressing the adequacy of a country’s reserves on the basis of the relation between its current reserves and past fluctuations in its reserves.

2

E.g., International Monetary Fund, Annual Report of the Executive Directors for the Fiscal Year Ended April 30, 1964 (Washington, 1964), p. 29.

3

Weir M. Brown, The External Liquidity of an Advanced Country, Princeton Studies in International Finance, No. 14 (Princeton, New Jersey, 1964).

4

The countries are those belonging to the so-called Group of Ten and Switzerland. For some countries, data were available only through 1961.

5

Op. tit., p. 20.

6

In contrast to reserve positions in the Fund, net IMF positions can be negative as well as positive; they will be negative when the Fund’s holdings of a country’s currency exceed 100 per cent of its quota, so that its use of the Fund’s resources exceeds its gold subscription. The net position vis-à-vis the Fund of the United States is further adjusted by subtracting the Fund’s gold investments in U.S. Treasury bills, which amounted to $200 million in 1956, $300 million in 1959, and $300 million in 1960.

7

Before 1958, certain U.S. foreign aid payments made as a matter of balance of payments support were treated as extraordinary financing in the Fund’s balance of payments statistics of both the United States and the recipient countries; from 1958 on they are so treated only in the accounts of recipient countries. Following this practice, such transfers for the years 1952–57 are in this paper taken not to have reduced the surpluses or increased the deficits of the United States. The reason for the change in the Fund’s accounting practice from 1958 on was the emergence at that time of an increased concern on the part of the U.S. authorities about the large and persistent U.S. balance of payments deficit; in these circumstances, it seemed more appropriate to consider such foreign aid payments by the United States as autonomous transactions (much like private capital outflows) rather than as accommodating transactions (such as changes in monetary gold).

8

The “developed countries” are the 14 listed as Industrial Countries and the 11 listed as Other Developed Areas in International Financial Statistics, issued by the International Monetary Fund. The Industrial Countries are Austria, Belgium-Luxembourg, Canada, Denmark, France, Germany, Italy, Japan, the Netherlands, Norway, Sweden, Switzerland, the United Kingdom, and the United States. The countries listed as Other Developed Areas are Australia, Finland, Greece, Iceland, Ireland, New Zealand, Portugal, South Africa, Spain, Turkey, and Yugoslavia.

9

Neither the sum of surpluses nor the sum of deficits, each taken by itself, can be regarded as a fully satisfactory indicator of the degree of imbalance, especially since the group of countries considered does not comprise the entire world economy. For instance, surpluses could decrease, though deficits remained unchanged, if (1) the additions to the monetary gold stock declined or (2) the consolidated surplus of the group of countries covered in the analysis with the rest of the world declined.

10

The growth rates of imports and reserves cited in this paper are computed in the same manner as the rates of increase in payments imbalances, i.e., by a trend line of best statistical fit showing a constant proportionate rate of annual increase.

11

A somewhat related problem has to do with the fact that an acceleration in the rate at which certain reserve assets are increased tends to be associated with a rise in payments imbalances. The extent to which the observed increases in payments imbalances may reflect an acceleration, during the period studied, in the growth of certain reserve assets is examined in Appendix I, page 392. It is concluded that this factor cannot explain more than a small part of the observed rise in payments imbalances.

12

Assuming unchanged economic policies in all countries but one, the sum of deficits of these countries will tend to be smaller, the larger is the deficit of the excluded country. If the deficit of the excluded country is large relative to the deficits of the remaining countries and if it shows a rising trend, any tendency toward a rising trend in the deficits of the remaining countries will be to that extent offset. See also Appendix I.

13

This allocation reduces the sum of surpluses and the sum of deficits by the same amount. If the United States has a deficit, part of it goes to reduce the surpluses of countries in surplus (possibly changing some small surpluses into deficits) and part to increase the deficits of countries initially in deficit. The sum of surpluses is thus reduced by the portion of the U.S. deficit that is allocated so as to reduce surpluses, and the sum of deficits, which initially included the entire U.S. deficit, but thereafter only that portion of it which is allocated so as to increase the deficits of other countries, is reduced by the same amount. Similarly, if the United States had a surplus, the sum of surpluses and the sum of deficits would both be reduced by the portion of the U.S. surplus allocated so as to reduce the deficits of other countries.

14

Two of these growth coefficients were not statistically significant at the 5 per cent level, namely, those for the deficits of all countries and for the surpluses of the developed countries.

15

In other words, the computations carried out on the annual imbalance data were also performed on the 3-year moving averages of the original data.

16

When the U.S. data are excluded, no significant upward trend is discernible in the sums of deficits. In fact, the trend computation for deficits of the 24 developed countries shows a negative coefficient, though it is not significantly different from zero in the statistical sense.

17

For the developed countries, surpluses alone grew slightly less rapidly (4 to 7 per cent) and deficits alone somewhat more rapidly (6 to 9 per cent), with the higher figures again applying to data including the United States and the lower to hypothetical data series constructed to exclude the influence of U.S. imbalances.

18

Actual 3-year deficits of developed countries grew at an even higher rate, namely by 15 per cent.

19

For purposes of this analysis, reserve positions in the Fund resulting from gold subscriptions of new members, or of existing members on the occasion of quota increases, should be included in countries’ gold holdings, since increases in Fund positions resulting from gold subscriptions are not associated with the financing of deficits.

20

Foreign exchange reserves could increase even if the reserve centers were in over-all surplus, namely, to the extent that other countries altered the composition of their reserves by selling gold and acquiring reserve currencies. In the text it is, however, assumed that each reserve center is in over-all deficit by an amount which equals or exceeds the increase in foreign official holdings of its currency.

21

I.e., the rise between 1953–55 and 1962–64 in the minimum sum of payments imbalances consistent with the observed rates of increase in the various reserve assets was $1 billion. These computations may well overstate the effect on the minimum sum of payments imbalances of the rise in the rate of reserve increases between 1953–55 and 1962-64, since part of the average annual increase in recorded foreign exchange holdings in the latter period, in contrast to the earlier period, may have been in currencies of countries that were not in over-all deficit (e.g., in French francs).

IMF Staff papers: Volume 13 No. 3
Author: International Monetary Fund. Research Dept.
  • View in gallery

    Sums of Actual Annual Payments Imbalances, 1952–64 1 (Including United States 2)

    (In billions of U.S. dollars)

  • View in gallery

    Sums of Actual Annual Payments Imbalances, 1952–64 1

    (Excluding United States)

    (In billions of U.S. dollars)

  • View in gallery

    Sums of Hypothetical Annual Payments Imbalances, 1952-641 (U.S. Over-All Balance Allocated to 24 Developed Countries by Average Reserves)

    (In billions of U.S. dollars)

  • View in gallery

    Sums of Actual Three-Year Payments Imbalances, 1953-631

    (Including United States 2)

    (In billions of U.S. dollars)

  • View in gallery

    Sums of Hypothetical Three-Year Payments Imbalances, 1953–63 1

    (U.S. Over-All Balance Allocated to 24 Developed Countries by Average Reserves)

    (In billions of U.S. dollars)