Article

Projection of U.S. Current Account Balance for 1964 from a World Trade Model

Author(s):
International Monetary Fund. Research Dept.
Published Date:
January 1964
Share
  • ShareShare
Show Summary Details

PROJECTIONS of a country’s current account balance are some-times made with the help of statistically estimated relations between imports or exports, on the one hand, and various “explanatory” magnitudes, such as incomes and prices, on the other. It is ordinarily easier to “explain” imports on the basis of such statistical relations—that is, to account for variations in imports in terms of their presumed causes—than it is to explain exports. Imports are fairly closely related to domestic economic activity and to import prices in relation to domestic prices. Exports depend on similar magnitudes abroad, but it is much more difficult to design appropriate indices of world economic activity and world prices to which a particular country’s exports should be related.

It is obviously not practicable to estimate exports to each foreign country on the basis of developments in that country’s economic activity and prices, if for no other reason than that information about these variables is for many countries not available. But even if this difficulty were absent, it would be incorrect in the case of a country as dominant in world trade as is the United States to project that country’s exports without taking account of the influence which its own imports and its other foreign transactions have on its exports. In brief, U.S. exports should be projected from a model of the world economy in which changes in U.S. economic activity, and thus in its imports, as well as changes in its capital exports and foreign aid payments, are allowed to exercise their full “feedback” effects on U.S. exports through incomes, foreign exchange earnings, and prices abroad. Such a model was described in a recent issue of Staff Papers.1

In that paper, tentative projections were made of the U.S. current account balance for 1964 and 1968. The projections for the latter year were to serve the purpose of comparison with similar projections made from an earlier version of this model by a Brookings study group.2 When they were first made by the Brookings group in 1962, these projections to 1968 spanned a considerable period of time. In some of the subsequent discussions of the Brookings Report,3 its authors were criticized for the attempt to make such long-run projections of the balance of payments.4 As was pointed out in these discussions, the two difficulties which may arise are (1) that the values of the explanatory variables may be incorrectly projected, and (2) that the relation between the explanatory and the explained variables, estimated on the basis of statistical data for a past period, may change over the years. If such long-run projections are ventured at all it is, therefore, appropriate that the performance of the model should be checked as frequently as possible at intermediate points in time.

The equations of the model were generally estimated from data ended in 1962 or, in some instances, in 1961. It is now possible to evaluate the performance of the model for one year outside the “sample period” (that is, the period used in the statistical estimation), namely, for 1963. Also, revised forecasts of gross national product (GNP) and prices are now available for the United States and many of the European countries, which permit an improved projection of the U.S. current account and its components for the calendar year 1964. Before these new projections are presented (in Section III), it is interesting to ascertain (1) the extent to which the projections to 1968 of economic activity and prices in the United States and Western Europe used by the Brookings group still appear plausible in view of recent trends (Section I), and (2) the extent to which the statistical relations of the model seem to hold true beyond the sample period (Section II).

I. Recent Income and Price Trends

The model5 divides the world into three regions, the United States, Western Europe,6 and the Rest of the World. Merchandise imports into each of the first two regions are estimated as related to GNP and inventory changes in the importing region and to a price variable.7 Payments for various categories of services are related to GNP, imports, or, for interest and dividend payments, to the outstanding indebtedness of the paying region to the receiving region. In the model, the imports of goods and services of the Rest of the World are determined by the foreign exchange receipts of this region from exports of goods and services to the two industrial regions and from the net receipts of capital and foreign aid from these regions. The division of the Rest of the World’s purchases of goods and services from the two industrial regions is affected by price developments in these two supplying regions, particularly by the ratio of U.S. export prices to the export prices of Western Europe. U.S. exports of goods and services depend, therefore, indirectly on U.S. imports from, and capital exports and aid to, the Rest of the World.

In accordance with this structure, the U.S. current account balance depends chiefly (1) on economic activity in the United States and Western Europe as measured by GNP in constant prices and by inventory investment, (2) on domestic prices and export prices in these two regions, as well as on the export prices of the Rest of the World, (3) on the foreign investment position of the United States vis-à-vis the two other regions, and (4) on the Rest of the World’s net foreign exchange receipts on account of capital and foreign aid less the accumulation of foreign exchange reserves by the Rest of the World.8 The actual values of these variables for 1960–63 and the assumed values for 1964 are shown in Appendix I, Table 7 (p. 429).

In regard to the U.S. current account balance, considerable interest attaches to trends in production and prices in the United States and Western Europe. The Brookings group made two sets of assumptions about production and price developments in these two regions between 1960 and 1968. According to their “initial” assumptions, GNP was projected to rise by 4.8 per cent per annum in the United States and by 4.2 per cent per annum in Western Europe; domestic prices, as measured by the implicit GNP price deflator, were assumed to rise at an annual rate of 1.5 per cent in the United States and of 2.75 per cent in Western Europe; and export prices were projected to rise at a considerably slower rate than GNP prices, namely, at 0.5 per cent in the United States and at 1.5 per cent in Western Europe. Under the “alternative assumptions,” the growth rate in GNP was somewhat lower both for the United States and for Western Europe, 4.2 per cent and 3.8 per cent, respectively ; while the price assumptions for the United States were the same as in the initial assumptions, those for Western Europe were considerably reduced—to annual growth rates of 1.75 per cent for GNP prices and 1.0 per cent for export prices.

Table 1 shows a comparison of these rates of change assumed by the Brookings group in making their projection of the U.S. balance of payments to 1968 and the actual rates of change recorded between 1960 and 1963; the table also shows the annual rates of change between 1960 and 1964 on the assumption that present forecasts of the respective magnitudes for 1964 will be realized.9 The year 1964 is the half-way mark in the projection period 1960-68 used by the Brookings group. The actual rates of change for real GNP, GNP prices, and export prices in the United States during the first half of the projection period conform quite closely to the rates used in the Brookings Report.10

Table 1.Annual Rates of Change of Incomes and Prices: Brookings Assumptions and Actual Changes(In per cent)
Brookings Assumptions, 1960–68Actual,

1960–63
Actual and

Projected,

1960–343
“Initial”1“Alternative”2
GNP in constant prices
United States4.84.23.94.2
Western Europe4.23.84.44.5
GNP prices
United States1.51.51.31.3
Western Europe2.751.753.63.5
Exoort prices
United States0.50.50.50.6
Western Europe1.51.00.60.7
Rest of the World−0.40.4

From Brookings Report (cited above, p. 415, fn. 2), Appendix Table 3, p. 283, and text p. 81.

Ibid., Appendix Table 5, p. 285, and text p. 88.

Actual 1960–63 and projected 1963–64.

From Brookings Report (cited above, p. 415, fn. 2), Appendix Table 3, p. 283, and text p. 81.

Ibid., Appendix Table 5, p. 285, and text p. 88.

Actual 1960–63 and projected 1963–64.

For Western Europe, the assumptions of the Brookings group are not borne out so well by the actual rates of change. The rate of growth of real GNP for 1960–64 has been slightly higher than the rate that forms part of the initial assumptions, and considerably higher than that used for the computations under the alternative assumptions. The largest discrepancy, however, arises in respect of Western Europe’s GNP prices. The rate of change for 1960–64 of 3.5 per cent per annum exceeds by a substantial margin the rate of 2.75 per cent used in connection with the initial assumptions. When the Brookings Report first appeared, there was considerable criticism of this particular assumption about the rise in Western Europe’s prices. Critics observed that, even if there were a tendency for European prices to rise at such a fast rate, it would be naïve to assume that the authorities would permit such a high rate of price inflation to continue for eight years. Nevertheless, price increases in Western Europe have to date been much larger than those projected by the Brookings group even under the initial assumptions. While the rate of domestic price increases may not persist, experience has shown that it is very difficult—even impossible—to achieve a sudden deceleration of upward price trends that have been going on for several years. It would, therefore, appear that the Brookings group’s assumptions about the trend in Western Europe’s domestic prices were quite conservative.

In this connection, it is of interest that the trend of prices in the United States, with a rate of increase of 1.3 per cent per annum, has to date been somewhat below that assumed by the Brookings group. In this case, to be sure, one may be inclined to say that the test is still to come, when the acceleration of growth in real output that has been experienced recently and that is expected to continue in the near future will put greater pressures on the available supplies of resources than has been the case in recent years.

Although the Brookings group, as far as the evidence to date indicates, underestimated the rising trend in Western Europe’s domestic prices, it failed to foresee the extent to which export prices would be kept stable in the face of rising GNP prices. Since 1960, Western Europe’s export prices have risen little more than U.S. export prices. This pronounced discrepancy in the trends in GNP prices and export prices undoubtedly reflects, in part, considerable improvements in productivity in Europe’s export industries relative to the productivity of the European economy in general, and, in part, the effects of lower tariffs in intra-European trade. In the Brookings Report it was argued that rising domestic wages and prices would eventually be reflected in higher export prices, though perhaps with a considerable lag. The rise in European domestic prices and wages during the early 1960’s may still find reflection in a faster rise in European export prices in the second half of the Brookings projection period.

In the preliminary projection to 1964, which was made in the earlier paper,11 the alternative assumptions of the Brookings group were used for GNP projections, and the initial assumptions for projections of prices. It is now seen that, while this was on the whole correct, it still resulted in an underestimate of the increase in Western Europe’s domestic prices, as well as a slight underestimate in Western Europe’s GNP, but in an overestimate of the increase in that region’s export prices. Projected values for 1964 of other independent variables occurring in the model, chiefly the Rest of the World’s net receipts of foreign exchange on capital and foreign aid account, also had to be revised in the present recomputation. Countries included in the Rest of the World have in the last two years (1962–63) accumulated reserves at an appreciable rate, perhaps on average close to $1 billion per annum, whereas the assumption in the Brookings Report was that there would be no such accumulation. Also, the increase in U.S. foreign aid between 1960 and 1963 has been much smaller than a proportionate share in the increase assumed by the Brookings group for the seven-year period 1961–68. While the capital flow from the United States to the countries of the Rest of the World has continued high in recent years, any slight upward trend which it may have shown appears to have been more than offset by an increasing capital outflow from the countries of the Rest of the World to Western Europe. As a result, the Rest of the World’s net capital and aid receipts during 1962 and 1963 were somewhat lower than they had been on average during the preceding five-year period. The present projections are computed on the assumption that the Rest of the World’s foreign exchange receipts from these sources will continue at the lower level of the last two years rather than at the higher level of an earlier period. (See Appendix I, Table 7.)

In brief, some of the revisions necessary in the new projection to 1964 tend to improve the projected U.S. current account balance, while others tend to worsen it. As will be seen in Section III, the effects of these two influences are in close balance, and the U.S. current account balance projected from the revised data is not very different from that projected in the earlier paper.

II. Test of the Model Against Data for 1963

It is now possible to check the performance of the model for the year 1963, which lies outside the period used for the statistical computation of the structural relationships of the model. Comparison of actual and computed values of the explained variables outside the sample period is ultimately the only reliable test of a model. Over the years, such comparisons will show whether the structural coefficients of the model, that is to say, the various estimated marginal propensities to import, price elasticities, etc., continue to be a sufficiently accurate description of the relevant relations in the real world.

Tables 2, 3, and 4 serve the dual purpose of showing (1) the degree of accuracy attained in the computation of the 1963 values of the variables explained in the model and (2) the computed values for the calendar year 1964 on the basis of projected values of the independent variables available at the time these computations were made.12 In this section, only the first of these matters, the degree of correspondence between actual and computed values in 1963, will be discussed.

Table 2.United States: Actual and Computed Values of Current Account Balance,1 1958–64(In billions of U.S. dollars)
ActualComputedError
19585.486.40−0.92
19593.103.56−0.46
19606.728.55−1.83
19618.379.02−0.65
19627.647.510.13
19637.987.360.62
19648.47
Source: Actual current account balance from U.S. Department of Commerce, Survey of Current Business; figures for 1960–63 are taken from the June 1964 issue, figures for 1958–59 from earlier issues.

Excluding military transactions and expenditures and miscellaneous government transactions.

Source: Actual current account balance from U.S. Department of Commerce, Survey of Current Business; figures for 1960–63 are taken from the June 1964 issue, figures for 1958–59 from earlier issues.

Excluding military transactions and expenditures and miscellaneous government transactions.

Table 3.United States: Actual and Computed Current Account Components for 1963 and Projections for 1964(In billions of U.S. dollars)
1963Computed

Increase,

1963 to

1964

(col. 4 – 2)

(5)
Actual

(1)
Computed

(2)
Residual

(col. 1 – 2)

(3)
1964

Projected

(4)
Balance of goods and services (excluding transfers and military grants)5.51(6.2)
Less: Balance of items not computed1−2.47(−2.3)
Equals: Balance of goods and services computed in the model7.987.360.628.471.11
Merchandise trade balance5.004.200.805.070.87
Exports, total21.9421.090.8523.162.07
To Western Europe7.567.270.298.160.89
To Rest of the World14.3813.820.5615.001.18
Imports, total16.9416.890.0518.091.20
From Western Europe4.724.81−0.105.270.46
From Rest of the World12.2212.080.1412.820.74
Service balances2.973.15−0.183.410.25
Service receipts8.878.97−0.109.710.73
From Western Europe2.502.55−0.052.780.22
From Rest of the World6.376.42−0.056.930.50
Service payments5.905.820.086.300.48
To Western Europe2.892.750.142.950.21
To Rest of the World3.013.07−0.063.350.27
Source: Actual values for 1963 are from U.S. Department of Commerce, Survey of Current Business, June 1964.

Military transactions and expenditures and miscellaneous government services.

Source: Actual values for 1963 are from U.S. Department of Commerce, Survey of Current Business, June 1964.

Military transactions and expenditures and miscellaneous government services.

Table 4.Western Europe: Actual and Computed Merchandise Trade Components for 1963 and Projections for 1964(In billions of U.S. dollars)
1963Computed

Increase,

1963 to

1964

(col. 4 – 2)

(5)
Actual

(1)
Computed

(2)
Residual

(col. 1 – 2)

(3)
1964

Projected

(4)
Merchandise trade balance−8.13−7.36−0.77−8.06−0.70
Exports, total23.5423.490.0525.502.01
To United States4.724.81−0.095.270.46
To Rest of the World18.8218.680.1420.231.55
Imports, total31.6730.850.8233.562.71
From United States7.567.270.298.160.89
From Rest of the World24.1123.580.5325.401.82
Sources: Data for actual merchandise trade with United States are from U.S. Department of Commerce, Survey of Current Business, June 1964; those for actual merchandise trade with the Rest of the World are from Organization for Economic Cooperation and Development, Statistical Bulletins: Foreign Trade, April 1964.
Sources: Data for actual merchandise trade with United States are from U.S. Department of Commerce, Survey of Current Business, June 1964; those for actual merchandise trade with the Rest of the World are from Organization for Economic Cooperation and Development, Statistical Bulletins: Foreign Trade, April 1964.

Actual and computed values of the U.S. current account balance, excluding military transactions and expenditures and miscellaneous government transactions, are shown in Table 2. This table, which corresponds to Table 2 in the earlier paper,13 gives the values starting with 1958 (as did the earlier table) in order to show three kinds of revisions that have been made. First, the U.S. balance of payments data published by the U.S. Department of Commerce have been revised, especially the data for service payments and receipts; these revisions affect the actual values of the U.S. current account balance. Second, the values of the independent variables used in these computations have been revised for prior years, particularly the data on Western Europe’s GNP and inventory investment; these revisions affect the computed values of the current account balance. Third, a minor revision of the computed values is due to an error in the table of multipliers used for these calculations in the earlier paper.14 Although there were substantial errors in the computed current account balance for early years of the period, notably in 1960, the computed balance for 1963 is fairly close to the actual. The error of $620 million amounts to only 8 per cent of the actual value.

Close correspondence between actual and computed figures for a net balance of many different items could, of course, be a coincidence or the result of cancellation of large errors with opposite signs in individual components. It is, therefore, interesting to note the correspondence of the computed and actual values of individual current account components for 1963, shown in Table 3. The residual error of the computation for the current account balance as a whole is, to some extent, the result of offsetting errors in the merchandise trade balance and the service balance. But these errors are themselves not very large; the error of $800 million in the merchandise trade balance is 16 per cent of the actual balance, while the error in the service balance is only 6 per cent of the actual balance. The largest absolute error (of $850 million) which occurs in total merchandise exports amounts to 3.9 per cent of the actual value. This error arises from a roughly proportionate underestimation of U.S. exports to Western Europe and to the Rest of the World. The performance of the model with respect to U.S. imports is very satisfactory. Computed values for 1963 of U.S. merchandise imports from the two other regions are very close to the actual values, as is also true for the service transactions.

The principal failing of the model in the extrapolation to 1963 seems thus to lie in the underestimate of U.S. exports, which the model yields. From 1962 to 1963, U.S. merchandise exports rose by $1.4 billion, or almost 7 per cent. The computed increase resulting from changes in income, inventory investment, and prices was close to $1 billion, but part of this increase was offset, in these computations, by the effect of a lower amount of the Rest of the World’s foreign exchange receipts from capital imports and aid funds (minus the amount of reserve accumulation). As a result, the computed increase was only $0.4 billion (see Table 6). The better performance of actual exports, compared with that computed from the model, may in part reflect increased tying of aid expenditures, but this factor would not account for the better than expected performance of U.S. exports to Western Europe. Exports of wheat to the U.S.S.R. did not affect the 1963 data, although they lend strength to U.S. exports in 1964. There was, however, a sharp rise in agricultural exports in the last quarter of 1963, which was due partly to temporary circum-stances. It may also be that the effect of the change in relative prices, which was favorable to the United States in 1963, is underestimated in the model; only further testing of the model in future years will show whether this is true.

Table 5.United States, Western Europe, and Rest of the World: Net Exports of Goods and Services, Actual 1961–63 and Projected 1964(In billions of U.S. dollars)
1960196119621963Projected

1964
United States13.04.43.84.55.3
Endogenous items6.78.47.68.08.5
Exogenous items−3.7−3.9−3.9−3.4−3.22
Western Europe12.62.3−0.1−0.8−1.0
Endogenous items−4.1−4.6−6.5−7.7−7.9
Exogenous items6.86.96.46.926.92
Rest of the World−5.3−6.2−3.3−2.8−3.3
Increase in world monetary
gold stock0.30.60.30.91.0
Sources: U.S. data are from U.S. Department of Commerce, Survey of Current Business; Western European data, from Organization for Economic Cooperation and Development, Statistical Bulletins: General Statistics; monetary gold stock, from International Monetary Fund, International Financial Statistics.

Net exports of goods and services as given in the national income accounts, including items not estimated in the model (exogenous items). Endogenous items are those estimated in the model. Data for 1963 are partly estimated.

Fund staff estimate.

Sources: U.S. data are from U.S. Department of Commerce, Survey of Current Business; Western European data, from Organization for Economic Cooperation and Development, Statistical Bulletins: General Statistics; monetary gold stock, from International Monetary Fund, International Financial Statistics.

Net exports of goods and services as given in the national income accounts, including items not estimated in the model (exogenous items). Endogenous items are those estimated in the model. Data for 1963 are partly estimated.

Fund staff estimate.

Table 6.United States: Analysis of Computed Changes in Current Account Balance, 1962 to 1963 and 1963 to 1964(In billions of U.S. dollars)
1962 to 19631963 to 1964
Total

Current

Account

Balance
Merchan-

dise

Exports
Merchan-

dise

Imports (-)
Services

(net)
Total

Current

Account

Balance
Merchan-

dise

Exports
Merchan-

dise

imports (-)
Services

(net)
Computed change in U.S. current account balance1−0.150.44−0.760.171.112.07−1.200.25
Effect of change in2
U.S. real GNP−0.610.18−0.62−0.17−0.880.26−0.90−0.24
W.E. real GNP0.650.510.140.800.630.17
U.S. inventory investment0.06−0.020.080.01−0.050.02−0.07
W.E. inventory investment−0.36−0.32−0.040.600.530.07
Income and inventory effects−0.260.35−0.54−0.060.471.44−0.97
U.S. GNP prices−0.160.05−0.16−0.05−0.150.05−0.15−0.05
W.E. GNP prices0.440.360.080.390.320.07
U.S. export prices−0.07−0.07
W.E. export prices0.080.060.010.010.080.060.010.01
R.O.W. export prices0.050.11−0.070.010.070.16−0.100.01
Price effects0.410.58−0.220.050.320.52−0.240.04
U.S. and foreign investment positions0.15−0.070.220.15−0.060.21
R.O.W. “autonomous foreign exchange receipts”3−0.44−0.42−0.020.180.170.01

Excluding military transactions and expenditures and miscellaneous government services.

Actual or assumed changes can be derived from Table 7.

U.S. and Western European current account items not estimated in the model plus Rest of the World’s net capital inflow minus Rest of the World’s reserve accumulation plus increase in world’s monetary gold stock.

Excluding military transactions and expenditures and miscellaneous government services.

Actual or assumed changes can be derived from Table 7.

U.S. and Western European current account items not estimated in the model plus Rest of the World’s net capital inflow minus Rest of the World’s reserve accumulation plus increase in world’s monetary gold stock.

Table 4 shows the accuracy of the model in the extrapolation to 1963 of Western Europe’s merchandise trade balance. The total balance is extrapolated with an error of somewhat less than 10 per cent, mainly accounted for by the error in computed imports, both from the United States and from the Rest of the World. This error of $820 million amounts to 2.6 per cent of total Western European imports. Here again it may be that the model underestimates the price elasticity of demand for imports in Europe and, therefore, gives insufficient weight to the effect of Europe’s rising domestic prices on imports in 1963.

III. Projection to 1964

The U.S. current account balance projected to 1964 from the three-region world trade model amounts to about $8.5 billion15 (Table 2). After allowance for the autonomous items that are not computed in the model (military transactions and expenditures and miscellaneous government transactions), the balance of goods and services (excluding transfers under military grants), as shown in the official U.S. balance of payments statistics, is projected at $6.2 billion (Table 3). This would be an improvement of $0.7 billion over 1963. The computed increase for 1963–64 amounts, however, to $1.1 billion. It may well be that factors not fully allowed for in the model, such as the effects of tying aid or of the current wheat sales to the U.S.S.R., will tend to make the U.S. current account balance somewhat stronger than the projected balance. In that event, the balance may, in the projection to 1964, tend to be underestimated just as it was in the computation for 1963, and the proper procédure for a projection to 1964 would be to add to the actual balance for 1963 the full amount of the computed improvement. The balance of current account components projected in the model would then be slightly over $9 billion. When account is taken of the estimated improvement by $0.2 billion in current account items not computed in the model, the balance of goods and services (excluding transfers under military grants) as published by the U.S. Department of Commerce would be $6.8 billion.

As Table 3 shows, the expected improvement is mainly the result of increased U.S. exports, the computed increase of which from 1963 to 1964 amounts to $2 billion. The computed increase of $1.2 billion in imports over the same period would only partially offset the improvement in exports. The service balance would also show an improvement in 1964 over 1963.

Western Europe’s merchandise trade balance is projected for 1964 at a deficit of more than $8 billion, a deterioration of approximately $0.7 billion from 1963. Although Western Europe’s exports are projected to increase by about $2 billion, its imports, both from the United States and from the Rest of the World, are estimated to rise even faster, by 9 per cent, or $2.7 billion.

Table 5 shows the projection for 1964 of net exports of goods and services16 of the three regions of the model, in addition to actual values for 1960–63. The recent improvement in the U.S. balance is expected to continue. For Western Europe, a further small deterioration in net exports of goods and services is indicated. As a result, the projected current account balance of the Rest of the World shows a slight deterioration from the relatively small deficit in 1963, but the expected 1964 deficit of $3.3 billion, no larger than that for 1962, is well below the high deficits of this region in 1960 and 1961. This reduction in the Rest of the World’s current account deficit in 1962–63 from that of the preceding two years reflects, and is indeed implicit in, the reduction in the net capital flow from the industrial countries to the Rest of the World from its particularly high level during 1960–61 and the substantial accumulation of reserves of the Rest of the World during the last two years, compared with a decumulation during 1960–61. The recent reduction in the net capital inflow of the Rest of the World appears not to be due to a decline in the gross capital outflow from the United States to the countries of the Rest of the World, but rather to the emergence of a net flow of capital from the Rest of the World to the countries of Western Europe.

In regard to the U.S. current account balance, it is of particular interest to ascertain the extent to which the improvement in the computed balance is ascribed in the model to the influence of changes in incomes, prices, and other factors. Table 6 shows the extent to which computed changes in the total current account balance, in merchandise exports and imports, and in the net service balance may be ascribed to the various actual or projected changes in the variables measuring U.S. and Western European economic activity and prices, and in certain other variables. The small improvement in the balance of the U.S. current account components from 1962 to 1963 explained in the model is not correctly projected by the model, which computes instead a small deterioration. This computed reduction was the result of offsetting influences. Income and inventory effects worked toward a reduction in the balance, whereas price effects made for a considerable improvement. This improvement was, however, almost entirely offset by a computed deterioration resulting from reduction in the Rest of the World’s “autonomous foreign exchange receipts.”

In the equivalent computation for the change from 1963 to 1964, income effects and price effects reinforce each other in improving the U.S. current account balance. This is chiefly the result of an expected increase in inventory investment in Western Europe, as well as of the expected rise in Western Europe’s GNP prices. The other factors listed also work toward an improvement of the U.S. current account balance from 1963 to 1964.

IV. Conclusion

The results of the extrapolations to 1963 from the three-region world trade model are, on the whole, encouraging. However, extrapolations for one year are not sufficient to form an opinion on the adequacy of the model. It is necessary to test the model over several years beyond the sample period and to revise it in the light of deficiencies noted. As indicated in the earlier paper, it may be desirable to expand the model by increasing the number of regions treated, and perhaps also by disaggregation of merchandise trade flows by commodity groups.

The projections to 1964 of the U.S. current account balance are, on the whole, plausible and in line with what would be expected on the strength of arguments from recent trends. No great claims of accuracy are made for the performance of the model at this stage. A probable error of about $0.5 billion, or perhaps slightly more, must be expected in these projections, and occasionally the error might be larger. Moreover, the model needs always to be supplemented by information not contained in it, such as data on the recent wheat sales to the U.S.S.R., or the increased tying of foreign aid. Supplemented in this fashion, the model can be a useful tool of analysis, particularly as an aid in sorting out the effects on the balance of payments and its components that should be ascribed to changes in incomes, prices, and other factors.

APPENDICES

I. Assumptions for 1964 Projection

Table 7 lists the exogenous variables used in the model and shows their actual values for the years 1960–63 and the assumed values for 1964. Data for incomes, prices, and other variables relating to the United States were readily available for 1963 at the time these calculations were made (June–July 1964). Moreover, forecasts for 1964 had been published by the Council of Economic Advisers.17

Table 7.United States, Western Europe, and Rest of the World: Values of Exogenous Variables, Actual for 1960–63 and Assumed for 1964
19601961196219631964
Billion 1954 U.S. dollars
U.S. GNP440.1447.7474.8492.9519.1
W.E. GNP270.3284.4296.0307.8322.3
U.S. inventory investment13.11.64.74.04.6
W.E. inventory investment16.44.23.63.04.0
Billion current U.S. dollars
W.E. investment in U.S.24.0427.4426.6930.0331.29
R.O.W. investment in U.S.14.7415.8716.8918.4819.74
U.S. investment in W.E.20.4421.2822.3424.3425.98
U.S. investment in R.O.W.43.5248.6152.92.58.3362.88
Autonomous items
U.S. current account−3.72−3.87−3.94−3.38−3.22
W.E. current account6.776.906.356.926.92
R.O.W. net capital receipts35.646.823.593.724.32
1954 = 1
U.S. GNP prices1.1421.1581.1691.1871.204
W.E. GNP prices1.2421.2771.3281.3801.426
U.S. export prices1.0861.1081.1011.1011.112
W.E. export prices1.0301.0401.0401.0501.060
R.O.W. export prices0.9500.9260.9220.9400.966
W.E. implicit exchange rate
deflator1.0991.0871.0851.0851.085
Sources: U.S. data are from U.S. Department of Commerce, Survey of Current Business; Western European data, from Organization for economic Cooperation and Development, Statistical Bulletins: General Statistics; Rest of the World net capital receipts, from International Monetary Fund, International Financial Statistics and unpublished Fund data. For sources and method of computation of Rest of World export prices, see Rudolf R. Rhomberg and Lorette Boissonneault, “Effects of Income and Price Changes on the U.S. Balance of Payments,” Staff Papers, Vol. XI (1964), Table 19, p. 118.

Inventory investment in current dollars deflated by the index of GNP prices; for description of GNP prices, see text, footnote 6.

Fund staff estimates.

Less excess of reserve accumulation over increase in the world’s monetary gold stock.

Sources: U.S. data are from U.S. Department of Commerce, Survey of Current Business; Western European data, from Organization for economic Cooperation and Development, Statistical Bulletins: General Statistics; Rest of the World net capital receipts, from International Monetary Fund, International Financial Statistics and unpublished Fund data. For sources and method of computation of Rest of World export prices, see Rudolf R. Rhomberg and Lorette Boissonneault, “Effects of Income and Price Changes on the U.S. Balance of Payments,” Staff Papers, Vol. XI (1964), Table 19, p. 118.

Inventory investment in current dollars deflated by the index of GNP prices; for description of GNP prices, see text, footnote 6.

Fund staff estimates.

Less excess of reserve accumulation over increase in the world’s monetary gold stock.

No comparable data were available for Western Europe as a whole, either for the actual values of the variables in 1963 or for the 1964 forecast. However, information on actual changes from 1962 to 1963 in real GNP, GNP prices, and inventory investment could be obtained from various publications in all of the larger countries of Western Europe. The weighted average percentage changes in GNP and prices for these countries were then applied to all countries of Western Europe, as this region is defined in this paper, to arrive at the estimate of GNP and GNP prices for 1963 shown in Table 7. A similar procedure was followed with respect to inventory investment.

For the 1964 projections of these variables it was possible to utilize forecasts of changes in GNP and in inventory investment for 1964, made in selected European countries by official or semiofficial bodies. These data were used to arrive at the projections of the corresponding variables for Western Europe as a whole. Projections of changes in GNP prices for 1963–64 are published for only a few countries. The assumed rise by 3.3 per cent of Western Europe’s GNP price index (shown in Table 7) was arrived at with the help of unpublished information and Fund staff estimates. It appears to be a conservative estimate, which is more likely to be too low than too high.

The extrapolation of export prices to 1964 merely follows the recent trend, although in some instances it was possible to make use of expectations concerning export prices informally expressed in various publications in the respective countries. The index of export prices of the Rest of the World was projected on the basis of commodity price trends apparent in the first quarter of 1964. Western Europe’s exchange rates were assumed to remain unchanged.

Projections of U.S. investment positions and of the Rest of the World’s “autonomous foreign exchange receipts” (i.e., the autonomous items in the U.S. and Western European current accounts, and the Rest of the World’s net foreign exchange receipts on capital and foreign aid account plus the excess of that region’s reserve accumulation over the increase in the world’s monetary gold stock) were based on Fund staff estimates. These estimates make allowance for a continued net outflow of U.S. capital to the Rest of the World (though at a somewhat lower rate than in 1963), an increase in U.S. foreign aid of the same size as in the previous year, an accumulation of $0.5 billion of gold and foreign exchange reserves by the Rest of the World, and a continued large increase in the world’s monetary gold stock. As a result of these assumptions, it is estimated that the “autonomous foreign exchange receipts” will be $0.6 billion greater in 1964 than they were in 1963.

II. Correction of Table of Multipliers

As indicated in the text, the table of multipliers in the earlier paper18 contains errors in two of the columns. Table 8 is a correction of the earlier table. The changes are in the columns under Px1 and Px2. Table 4 of the earlier paper,19 which gives selected values from Table 12 of that paper, is also subject to these corrections.

Table 8.Multipliers of Model 3: Multiplier Effects of Unit Changes in the Variables in the Column Headings on the Variables Appearing in the Stubs1
Y1/P1Y2/P2H1/P1H2/P2P1P2Px1Px2Px3rx2I12I13I21I31B3213
N1−0.03360.0551−0.08090.602−9.108.48−6.747.712.68−8.25−0.0184−0.01850.06160.02170.4488.37
M120.00.01850.00.4080.04.60−0.2900.00.0−4.750.00.00.00.00.0−0.520
S120.00.003690.00.03150.00.835−0.02240.00.0−0.3660.00.00.06160.00.0−0.907
M130.009900.02470.03530.1222.832.28−6.106.106.19−2.350.00.01420.0−0.01670.425−7.99
S130.002740.008650.01100.04260.7580.798−0.4190.4190.995−0.8240.00.0009750.00.03820.0292−2.04
M210.01490.00.02920.03.670.00.0−1.070.00.00.00.00.00.00.0−5.88
S210.005820.00.005960.01.760.00.0−0.2180.00.00.01840.00.00.00.0−2.56
M310.01930.00.07970.05.270.00.00.03.840.00.00.00.00.00.0−7.62
S310.006200.0004020.01230.001981.980.0372−0.09950.09950.665−0.03830.00.03360.0−0.0002730.00693−3.75
M310.00.06500.00.3200.06.000.00.010.7−6.190.00.00.00.00.0−9.02
M230.01290.03210.04580.1583.672.966.42−6.428.03−3.050.00.01850.0−0.02170.552−1.04

For list of variables, see Rudolf R. Rhomberg and Lorette Boissonneault, “Effects of Income and Price Changes on the U.S. Balance of Payments,” Staff Papers, Vol. XI (1964), p. 86.

B3 = K3B1B2.

Constant term.

For list of variables, see Rudolf R. Rhomberg and Lorette Boissonneault, “Effects of Income and Price Changes on the U.S. Balance of Payments,” Staff Papers, Vol. XI (1964), p. 86.

B3 = K3B1B2.

Constant term.

Projection de la balance des comptes courants des Etats-Unis pour 1964, d’après un modèle du commerce mondial

Résumé

Cette étude met à l’épreuve le modèle triregional du commerce mondial décrit dans le numéro de mars 1964 des Staff Papers, en comparant pour 1963 les résultats prévus avec les données connues, et présente des projections révisées d’après ce modèle de la balance des comptes courants des Etats-Unis et de ses éléments pour l’année civile 1964.

Les relations fonctionnelles de ce modèle ont été estimées à partir de données allant jusqu’à l’année 1962 inclusivement. L’extrapolation de ces relations pour l’année 1963 donne une sous-évaluation de la balance des comptes courants des Etats-Unis (à l’exclusion des transactions militaires et de certains postes afférents aux services de l’Etat) qui s’élève à environ 0,6 milliard de dollars, soit 8 pour cent de la balance réelle. Cette erreur s’explique principalement par une sous-estimation des exportations de marchandises des Etats-Unis, qui ont augmenté dans des proportions inhabituelles de 1962 à 1963.

La projection pour l’année civile 1964 indique une amélioration, par rapport à 1963, de la balance des comptes courants des Etats-Unis (en tenant compte des exclusions susmentionnées) ; cette amélioration se chiffre à un peu plus de 1 milliard de dollars, et résulte surtout d’une nouvelle et très forte augmentation des exportations américaines qui n’est contrebalancée qu’en partie par l’augmentation projetée des importations des Etats-Unis. Parallèlement, on s’attend à ce que les balances des comptes courants de l’Europe occidentale et du reste du monde se détériorent quelque peu dans ces projections. Bien que des projections de ce genre soient sujettes à des erreurs considérables, on peut dire avec un certain degré de probabilité que la balance des comptes courants des Etats-Unis aura tendance à s’améliorer plutôt qu’à se détériorer en 1964, et que cette amélioration pourra se chiffrer par une somme comprise entre 0,5 et 1,5 milliard de dollars.

Proyección para 1964 de la balanza de las cuentas corrientes estadounidenses obtenida de un modelo del comercio mundial

Resumen

Este estudio explora la eficacia del modelo trirregional del comercio mundial, descrito en un artículo publicado en la edición de marzo de 1964 de Staff Papers, para pronosticar las cifras para 1963, y presenta, además, revisiones de las proyecciones de este modelo de la balanza de las cuentas corrientes estadounidenses y sus componentes para el año civil de 1964.

Las relaciones de las variables del modelo se calcularon haciendo uso de datos que abarcaban hasta 1962, inclusive. La extrapolación de esas relaciones para 1963 da por resultado una subestimación de la balanza de las cuentas corrientes de Estados Unidos (exceptuando las transacciones militares y ciertas partidas relativas a servicios gubernamentales) de US$600 millones, aproximadamente, o el 8 por ciento de su verdadero monto. El error se puede atribuir principalmente a una estimaciôn insuficiente de sus exportaciones las cuales experimentaron un aumento inusitado de 1962 a 1963.

La proyección para el año civil de 1964 indica un mejoramiento en comparación con 1963, de un poco más de US$1.000 millones en la balanza de la cuenta corriente estadounidense (hechas las excepciones antes indicadas), debido sobre todo a otro agudo aumento de las exportaciones de ese país que está neutralizado sólo en parte por el aumento proyectado en sus importaciones. Análogamente, según estas proyecciones, cabe esperar cierto deterioro en las balanzas en cuenta corriente tanto de la Europa Occidental como del Resto del Mundo. Aunque dichas proyecciones pueden variar considerablemente, cabe afirmar, con cierto grado de probabilidad, que en 1964 la balanza de las cuentas corrientes de Estados Unidos mejorará en vez de deteriorarse, y que la cuantiá de ese mejoramiento será entre US$500 millones y US$1.500 millones.

Mr. Rhomberg, Chief of the Special Studies Division, 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.

Miss Fortucci, research assistant in the special Studies Division, is a graduate of the George Washington University, and of the Johns Hopkins Bologna Center for Advanced International Studies.

Rudolf R. Rhomberg and Lorette Boissonneault, “Effects of Income and Price Changes on the U.S. Balance of Payments,” Staff Papers, Vol. XI (1964), pp. 59–124.

Walter S. Salant et al., The United States Balance of Payments in 1968 (The Brookings Institution, Washington, 1963); also published under the same title by the U.S. Congress (88th Congress, 1st Session), Joint Economic Committee (Washington, 1963) ; hereinafter referred to as the Brookings Report.

See, in particular, The United States Balance of Payments: Statements by Economists, Bankers, and Others on the Brookings Institution Study, “The United States Balance of Payments in 1968” (materials submitted to the Joint economic Committee, 88th Congress, 1st Session, Washington, 1963) ; and Harry G. Johnson, “The International Competitive Position of the United States and the Balance of Payments Prospect for 1968,” Review of Economics and Statistics, Vol. XLVI (1964), pp. 14–32.

Professor Machlup introduces his critical remarks in the volume cited in the preceding footnote by saying, “It is notoriously difficult to do the impossible” (p. 301).

For a detailed description of the structure of the model, see Rhomberg and Boissonneault, op. cit., pp. 61–65.

In this paper, Western Europe is defined as European countries, excluding Spain, in the Organization for Economic Cooperation and Development (OECD), that is, Austria, Belgium, Denmark, France, Federal Republic of Germany, Greece, Iceland, Ireland, Italy, Luxembourg, Netherlands, Norway, Portugal, Sweden, Switzerland, Turkey, and United Kingdom.

The price variable used is the ratio of the implicit GNP price deflator in the importing region to the export price index of the region supplying the imports in question. The implicit GNP price deflator is obtained by dividing GNP in current prices by GNP in prices of the base period. In the following discussion, this will be referred to simply as GNP prices.

Since the sum of the current account balances of the three regions is equal to the increase in the world’s monetary gold stock, the explanatory variable actually used in the model is the item described under (4) in the text, plus the increase in the world’s monetary gold stock. Unfortunately, this was not made sufficiently clear in the discussion in the earlier paper referred to above. The Rest of the World’s current account deficit falls short of the sum of the current account surpluses of the two industrial regions by the increase in the world’s monetary gold stock; and it is only this reduced current account deficit which is, in fact, financed by the Rest of the World’s net receipts of capital and aid and by a decumulation of its gold and foreign exchange reserves.

Projections of output and domestic prices for the United States and Western Europe are shown in Appendix I, Table 7. The sources on which these projections are based are given in Appendix I.

The actual growth rate in 1960–64 of U.S. real GNP is equal to that in the alternative assumptions.

Rhomberg and Boissonneault, op. cit., especially Appendix III, pp. 110–12.

The computations were made in the spring and early summer of 1964.

Rhomberg and Boissonneault, op. cit., p. 70.

The error is in Table 12, ibid., p. 109. The corrected table is shown below in Appendix II (Table 8). It should be noted that the same error, is found in Table 4 of the earlier paper (p. 73), which shows selected values from Table 12 in that paper. See also Appendix II below.

This compares with a projected value of $8.3 billion in the March 1964 paper, ibid., p. 79. As mentioned above, the effects of various revisions canceled one another to some extent, leaving the projected balance almost unchanged.

Net exports of goods and services as defined in the national income accounts (rather than the corresponding current account balances from balance of payments statistics) are used in the model; this is done in part, because there is no consolidated balance of payments statement for Western European countries as a group, whereas the Consolidated national income statistics of European countries of the OECD show the net exports of goods and services of this region.

Economic Report of the President (together with the Annual Report of the Council of Economic Advisers, Washington, January 1964), pp. 51–52.

Rhomberg and Boissonneault, op. cit., Table 12, p. 109.

Ibid., p. 73.

Other Resources Citing This Publication