In view of the narrow focus of indicators based on measures of goods market competitiveness, the analysis normally has to be extended to take into account the many factors other than relative price movements that may exert major persistent effects on a country’s overall balance of payments, including capital transactions and current transactions. These factors encompass such developments as long-run changes in comparative advantage and in propensities to save and invest, as well as policy changes that affect the rates of domestic absorption for a number of years. Thus, the aim of this more comprehensive analysis becomes the assessment of the “underlying” external position of the country, after adjustment for the main temporary factors that may influence the currently observed level of its balance of payments, and taking account of any factors that are expected to have a persistent effect over the next few years. Once this assessment has been made on the basis of the present level of the exchange rate, it can then be used to evaluate whether the currently observed rate is likely to be sustainable in the medium term.
The major departure from the simple analysis of competitiveness is that instead of assuming that there is a tendency to return to some base period level of the real exchange rate, this analysis is specifically concerned with the determination of the sustainable real effective exchange rate for the period ahead, whether or not that rate has actually been observed in the past. Furthermore, attention is focused not only on foreign trade and the current account balance, but also on the longer-term evolution of capital flows. Finally, an assessment of the underlying balance of payments explicitly takes into account the impact of the stance of economic policies on the sustainable level of a country’s exchange rate.
The analysis of countries’ external positions and its use for the assessment of exchange rates have of course been at the center of Fund activities since the creation of the institution. However, it was mainly as a result of the systemic crises that occurred in the early 1970s that attempts were made to develop the Fund’s work on the measurement of underlying payments positions in the industrial countries, and to apply it on a comprehensive and multilateral basis. In October 1971, a Fund staff report suggesting ways of estimating desirable changes in payments balances on current account was presented as a contribution toward attaining an agreement regarding the extent to which payments balances on current account should be made to change (through a change in the structure of parities) from those that could be expected to prevail with exchange rates at existing parities. The move toward greater exchange rate flexibility in 1973, and the major exchange rate swings that occurred subsequently, led to an intensification of this type of work. Following discussion of some initial results during 1975 and January 1976, the Fund decided that officials in the industrial countries should make a technical evaluation of the work, and a conference was held for this purpose at the Fund’s Paris office in February 1976. Subsequent to that conference, the Fund staff’s work in this field began to be used systematically within the Fund. This led to a series of bilateral meetings with officials of all the industrial countries during the period June-October 1978 and a second conference at the Fund’s Paris office in October 1978.
This section presents a brief review of the Fund’s approach to the estimation of underlying payments balances and illustrates it by assessing the underlying payments positions of the three largest industrial countries on the basis of the average exchange rate levels that prevailed in November 1983. The numerical estimates in this section are presented exclusively for the purpose of illustration.
Procedures for Assessing a Country’s Underlying External Payments Position
From an analytical standpoint, the assessment of a country’s underlying payments position and its corresponding sustainable exchange rate can be divided into three separate steps.
(1) Estimation of the current account position adjusted for temporary factors and anticipated developments. Temporary factors include disturbances such as crop failures, strikes, etc. Anticipated developments may be divided into three broad categories: (i) expected effects of past exchange rate and relative price changes, (ii) effects resulting from the unwinding of “abnormal” relative cyclical positions, and (iii) other types of developments that are expected to occur in the medium term.
First, if large changes in prices and exchange rates have occurred in recent years, it is necessary to take into account those price effects that are still “in the pipeline,” in order to avoid reaching an incorrect judgment of the underlying payments position. Such judgments therefore require an assessment of the lagged response of merchandise trade and services flows to relative price movements that have already taken place.
Second, account must be taken of relative cyclical positions and their effects on trade flows among industrial countries. In the 1960s and early 1970s economic fluctuations tended to be moderate in amplitude, so that it was meaningful to speak of a cyclically neutral level of economic activity corresponding to a reasonable norm of full employment, and to expect this neutral level of economic activity to prevail, on average, over the medium term. For all practical purposes, what was involved was the calculation of a full-employment current account balance. Cyclical movements in trade flows could be viewed as transitory phenomena of little importance for assessing exchange rates. Owing to the spread of stagflation and the rise in average levels of unemployment, however, it became clear by the mid-1970s that such an approach had ceased to be realistic. Consequently, the staff began to define “abnormality” of cyclical positions in relation to what could be expected to prevail after a medium-term period of two or three years, rather than in relation to the notional full-employment level of economic activity. Because it is extremely difficult to project cyclical developments, the staff also began to consider a number of scenarios embodying alternative assumptions as to future policy stances and other economic developments.
Third, each country’s current account position must be adjusted for other types of expected future developments, including both longer-term trends that are apparent in the components of each country’s current and capital accounts and significant anticipated shifts. Examples in this category include allowance for the effects of expected shifts in items such as private and official international transfers; movements in trade balances on account of energy products arising from conservation measures or other factors; and changes in the size and direction of private capital flows owing to such major factors as tax and regulatory reforms that affect saving and investment decisions. It is in attempting to estimate the effects of these other developments that the most serious difficulties are encountered. An unconditional medium-term forecast is not feasible, since what happens will depend on policies and conditions that will be in effect a few years hence. Even projections that are contingent on certain policy assumptions are difficult to make, because the implications of various policy choices are uncertain. The problem is already present in the estimation of cyclical effects, but it is even more severe in the case of the many longer-run structural changes that affect foreign trade flows.19
(2) Calculation of the current account position that corresponds to a sustainable structure of the overall balance of payments, including some estimate of the level of capital flows that is likely to prevail in the medium term and a reasonable target for growth in the country’s official holdings of (net) international reserves. The estimate of the sustainable balance of payments structure should reflect factors that influence the longer-run evolution of aggregate flows of domestic saving and investment, including any longer-term tendency for net saving or dissaving by the public sector. Major structural or other noncyclical developments—such as the exploitation of the North Sea oil fields by Norway and the United Kingdom, or the recent moves toward import liberalization by Japan—must also be taken into account, as they may warrant a sustained change in the balance of payments structure.
(3) Estimation of the sustainable exchange rate. The comparison of the present underlying current account position resulting from step (1) with the normal current account position resulting from step (2) provides an estimate of the underlying current account disequilibrium. The deviation of the actual exchange rate from the sustainable exchange rate can then be estimated by judging the change in the exchange rate that would bring about the needed current account adjustment.
Illustrative Calculation of the Underlying Current Account Position
To illustrate the approach described above, Table 3 presents estimates of underlying payments positions on current account for the three largest industrial countries and for the other industrial countries as a group. These estimates were obtained by adjusting the most recent estimate of current account positions for 1983 in the following ways.20
Underlying Payments Balances on Current Account, 1983
(In billions of U.S. dollars at constant 1983 prices)
Balance on goods, services, and private transfers.
In Japan, recent impon liberalization measures increased import volumes: adjustments are required to give appropriate balances for the whole of 1983, both for Japan and for its trading partners.
Estimates of the effects of charges in exchange rales and domestic price levels that occurred from 1980 through November of 1983, but which are not reflected in 1983 trade flows.
The effect of assumed changes in relative cyclical positions to 1986 on trade and services balances in terms of 1983 dollars.
Effects over three years of estimated trend factors, including changes in fuels balances.
The 1983 balances adjusted by the factors in columns (2) through (5).
A further adjustment is made in column (7) for changes in relative cyclical positions which arise if prospective U.S. demand growth is assumed to be ½ of 1 percent a year higher, and prospective demand growth in the Federal Republic of Germany and Japan is assumed to be ½ of 1 percent lower than in the central scenario.
Includes effects of estimated medium-term trends in official transfers.
The sum of columns (6) and (8).
The sum of columns (7) and (8).
Reflects errors, omissions, and asymmetries in reported statistics, plus balance with other countries including U.S.S.R. and other nonmember countries of Eastern Europe. In the table, this asymmetry is assumed to remain constant at the projected 1983 level.
Underlying Payments Balances on Current Account, 1983
(In billions of U.S. dollars at constant 1983 prices)
Adjustment for: | Scenario I | Scenario II | Scenario I | Scenario II | ||||||
---|---|---|---|---|---|---|---|---|---|---|
Estimated Current Balance 19831 (1) | Temporary disturbailees2 (2) | Recent relative price chanees3 (3) | Changes in relative cyclical positions4 (4) | Medium-term tendencies5 (5) | Underlying current balance6 (6) | Underlying current balance7 (7) | Official Transfers8 (8) | Underlying current balance including official transfers9 (9) | Underlying current balance including official transfers 10 (10) | |
United States | – 34.3 | 0.2 | –28.2 | – 10.2 | – 7.8 | –80.3 | –85.5 | –5.6 | –85.9 | –91.1 |
Germany. Federal Republic of | 9.0 | — | 10.9 | 0.8 | –3.1 | 17.6 | 19.7 | –6.5 | 11.1 | 13.2 |
Japan | 22.3 | –0.6 | 0.8 | – 1.6 | 6.4 | 27.3 | 30.0 | –2.0 | 25.3 | 28.0 |
Other industrial countries | 1.5 | 0.1 | 16.5 | –0.2 | 3.1 | 21.0 | 20.7 | –8.0 | 13.0 | 12.7 |
Total industrial countries | – 1.5 | –0.3 | — | – 11.2 | – 1.4 | – 14.4 | – 15.1 | –22.1 | –36.5 | –37.2 |
Other countries | –72.6 | 0.3 | — | 11.2 | 1.4 | –59.7 | –59.0 | 22.1 | –37.6 | –36.9 |
Total11 | –74.1 | — | — | — | — | –74. 1 | –74.1 | — | –74.1 | –74.1 |
Balance on goods, services, and private transfers.
In Japan, recent impon liberalization measures increased import volumes: adjustments are required to give appropriate balances for the whole of 1983, both for Japan and for its trading partners.
Estimates of the effects of charges in exchange rales and domestic price levels that occurred from 1980 through November of 1983, but which are not reflected in 1983 trade flows.
The effect of assumed changes in relative cyclical positions to 1986 on trade and services balances in terms of 1983 dollars.
Effects over three years of estimated trend factors, including changes in fuels balances.
The 1983 balances adjusted by the factors in columns (2) through (5).
A further adjustment is made in column (7) for changes in relative cyclical positions which arise if prospective U.S. demand growth is assumed to be ½ of 1 percent a year higher, and prospective demand growth in the Federal Republic of Germany and Japan is assumed to be ½ of 1 percent lower than in the central scenario.
Includes effects of estimated medium-term trends in official transfers.
The sum of columns (6) and (8).
The sum of columns (7) and (8).
Reflects errors, omissions, and asymmetries in reported statistics, plus balance with other countries including U.S.S.R. and other nonmember countries of Eastern Europe. In the table, this asymmetry is assumed to remain constant at the projected 1983 level.
Underlying Payments Balances on Current Account, 1983
(In billions of U.S. dollars at constant 1983 prices)
Adjustment for: | Scenario I | Scenario II | Scenario I | Scenario II | ||||||
---|---|---|---|---|---|---|---|---|---|---|
Estimated Current Balance 19831 (1) | Temporary disturbailees2 (2) | Recent relative price chanees3 (3) | Changes in relative cyclical positions4 (4) | Medium-term tendencies5 (5) | Underlying current balance6 (6) | Underlying current balance7 (7) | Official Transfers8 (8) | Underlying current balance including official transfers9 (9) | Underlying current balance including official transfers 10 (10) | |
United States | – 34.3 | 0.2 | –28.2 | – 10.2 | – 7.8 | –80.3 | –85.5 | –5.6 | –85.9 | –91.1 |
Germany. Federal Republic of | 9.0 | — | 10.9 | 0.8 | –3.1 | 17.6 | 19.7 | –6.5 | 11.1 | 13.2 |
Japan | 22.3 | –0.6 | 0.8 | – 1.6 | 6.4 | 27.3 | 30.0 | –2.0 | 25.3 | 28.0 |
Other industrial countries | 1.5 | 0.1 | 16.5 | –0.2 | 3.1 | 21.0 | 20.7 | –8.0 | 13.0 | 12.7 |
Total industrial countries | – 1.5 | –0.3 | — | – 11.2 | – 1.4 | – 14.4 | – 15.1 | –22.1 | –36.5 | –37.2 |
Other countries | –72.6 | 0.3 | — | 11.2 | 1.4 | –59.7 | –59.0 | 22.1 | –37.6 | –36.9 |
Total11 | –74.1 | — | — | — | — | –74. 1 | –74.1 | — | –74.1 | –74.1 |
Balance on goods, services, and private transfers.
In Japan, recent impon liberalization measures increased import volumes: adjustments are required to give appropriate balances for the whole of 1983, both for Japan and for its trading partners.
Estimates of the effects of charges in exchange rales and domestic price levels that occurred from 1980 through November of 1983, but which are not reflected in 1983 trade flows.
The effect of assumed changes in relative cyclical positions to 1986 on trade and services balances in terms of 1983 dollars.
Effects over three years of estimated trend factors, including changes in fuels balances.
The 1983 balances adjusted by the factors in columns (2) through (5).
A further adjustment is made in column (7) for changes in relative cyclical positions which arise if prospective U.S. demand growth is assumed to be ½ of 1 percent a year higher, and prospective demand growth in the Federal Republic of Germany and Japan is assumed to be ½ of 1 percent lower than in the central scenario.
Includes effects of estimated medium-term trends in official transfers.
The sum of columns (6) and (8).
The sum of columns (7) and (8).
Reflects errors, omissions, and asymmetries in reported statistics, plus balance with other countries including U.S.S.R. and other nonmember countries of Eastern Europe. In the table, this asymmetry is assumed to remain constant at the projected 1983 level.
(i) Allowance was made for temporary (noncyclical) disturbances affecting the estimated current account position for 1983. In the present illustrative exercise, the main factor taken into account was the import liberalization policies of Japan, which began to take effect during the second half of 1983 and which are therefore only partially reflected in the estimated actual external payments figures for that year.
(ii) Real exchange rates were assumed to remain at their estimated November 1983 levels. Adjustment was then made for the effects of past changes in exchange rates and relative prices on foreign trade flows that were still in the pipeline by the end of November 1983—that is, for effects that were not fully reflected in the trade flows that occurred during calendar 1983 but which may be expected to exert an influence within a two to three year period. These estimates were obtained using the Fund’s world trade model.
(iii) The effects of projected changes in relative cyclical positions to 1986 were calculated on the basis of two alternative scenarios with respect to growth rates. Scenario I, corresponding to the medium-term “base scenario” in the World Economic Outlook, 1984,21 assumes that the average annual rate of growth of real gross national product (GNP) for the medium term is 3½ percent for the United States, 4 percent for Japan, 2½ percent for the Federal Republic of Germany, and 2 percent for the group of other industrial countries. Scenario II assumes average rates of growth which, relative to Scenario I, are ½ of 1 percentage point higher for the United States, ½ of 1 percentage point lower for Japan and the Federal Republic of Germany, and the same for the other industrial countries as a group. Estimates of the effects of projected changes in relative cyclical positions were also obtained from the world trade model.
(iv) Adjustment was made for medium-term tendencies that seem to be independent of considerations of price competitiveness. The adjustment was estimated by reference to the trend components of world trade model projections and on the basis of an ad hoc evaluation of tendencies on account of invisible flows.
In Table 3, column 1 presents the estimated current account balances for 1983, while columns 2 to 5 give the adjustments that were made to arrive at the illustrative calculation of the underlying current account positions. Columns 6 and 7 are the estimates of the underlying current account balances (under Scenarios I and II, respectively), excluding official transfers, and columns 9 and 10 are the same figures inclusive of official transfers.
Assessing the Sustainability of the Underlying Current Account Position
To determine whether an underlying current account balance estimated by steps (i)–(iv) is sustainable in the medium term at the given exchange rate, it is necessary to assess whether this flow is consistent with a “normal” level of net private capital flows and with a reasonable accumulation of (net) official foreign exchange reserves.
It is evident that projecting the sustainable and appropriate levels of these balance of payments flows is a very difficult task. A major role of capital movements is to contribute to a better allocation of world savings by promoting a tendency for the equalization of real rates of return on investment among countries. Many of the factors that lead to these private capital transfers are relatively stable or fluctuate with the economic cycle around stable longer-term trends. For example, propensities to save are usually stable, after adjustment for cyclical movements. Similarly, the geographic distribution of investment opportunities tends to be stable, particularly because it partly reflects basic factors related to income distribution and to the overall growth potential of various economies. To the extent that these factors remained stable over time, there would be an associated stable component in private capital movements. However, structural developments can certainly lead to a change in this component, and the change may be difficult to detect at an early stage because, in the short run, capital movements tend to be dominated by the effects of temporary variations in the relative expected yields on financial assets denominated in different currencies. Furthermore, major changes in fiscal positions may also have implications for the level and structure of private capital flows in the medium term.
In its work on the estimation of normal levels of private capital flows, the Fund staff has tended to rely in some instances on information or targets provided by the national authorities. Whenever this is not possible, the only alternative is to extrapolate past tendencies and to adjust these extrapolations for known factors that may involve a change from past trends. This eclectic and judgmental method is obviously subject to a significant margin of uncertainty.
Chart 6 presents annual data on private capital flows (including errors and omissions) expressed as a percentage of GNP for 12 industrial countries for the years 1963 to 1982. For four of the major countries—France, the Federal Republic of Germany, Japan, and Canada—the capital account balance expressed as a percentage of GNP fluctuated around a relatively stable longer-term level during the period under consideration, with major deviations related to special events, such as the breakdown of the Bretton Woods system in 1971–73 and the two waves of oil price increases in 1973–74 and 1979–80. For the first three of these countries the data suggest that, when averaged over a number of years, net private capital flows have been relatively small, both as a proportion of GNP and in absolute terms. In the case of Canada, net capital inflows tended to be significant throughout most of the period under consideration, possibly in part because of the close integration of its economy with that of the United States. For Japan, however, a break with the established trend may now be in prospect because the removal of controls on capital flows and the increased propensity of Japanese firms to invest abroad is leading to large capital outflows.
In the case of Norway and the United Kingdom in the 1970s (and the Netherlands in the late 1960s) large increases in capital flows appear to have been associated with the financing of major oil and gas projects. Capital flows for these countries were also influenced by temporary factors such as the U.K. authorities’ attempt to resist the appreciation of the pound sterling in 1977, as well as structural factors, such as the removal of controls on capital outflows in 1979. For other industrial countries, such as Belgium, Denmark, Italy, the Netherlands, and Sweden, the data suggest a rising trend of capital inflows since the mid-1970s, associated in particular with a substantial increase in official or quasi-official borrowing.
Although Chart 6 shows that the net private capital outflow from the United States including errors and omissions has generally been small both as a proportion of GNP and in absolute terms, the net flow is the outcome of large flows in opposite directions. Identified private capital movements (exclusive of errors and omissions) have recorded a relatively stable net outflow of somewhat more than 1 percent of GNP over a long period. By contrast, the errors and omissions item itself rose sharply from a negligible amount in the mid-1970s to an average of 1 percent of GNP in 1981–82, before falling off sharply again in 1983.
This brief review of capital account developments suggests that, even though net capital flows have often tended to cumulate to relatively small amounts over a number of years, it is rather difficult to arrive at a view of what is the “normal pattern” of capital account balances for the industrial countries. To continue the illustration of the underlying payments balance approach, estimates of the normal levels of private capital flows were obtained by averaging the ratio of capital flows (including errors and omissions) to GNP for the period 1975–82, excluding the years 1979 and 1980 (which are affected by the second oil price change) and scaling the results to 1982.22 For Japan, a special allowance was made for the effects of recent changes in capital controls and for increased direct investment outflows. The results appear to suggest an outflow on the order of $5-10 billion for the United States and Japan, a roughly balanced position on capital account for the Federal Republic of Germany, and an inflow on the order of $10-20 billion for the other industrial countries as a group.
The inherently large uncertainty in any attempt to project normal underlying positions on both the current and private capital accounts is of course greatly increased by the recent tendency toward a huge global asymmetry in the recording of balance of payments data.23 In the present illustrative exercise, errors and omissions have been included with capital flows, and the growth of industrial country receipts under this item in recent years has been interpreted as a decline in the magnitude of normal capital outflows from this group of countries. It is also possible, however, that errors and omissions reflect a persistent flow (possibly related to current transactions) in some countries, and a temporary flow (perhaps associated with volatile financial flows) in others. In this case, the estimates cited above could easily give a biased impression of prospective underlying payments imbalances among major industrial countries. Moreover, the uncertainty is also greatly increased as a result of the medium-term evolution of fiscal positions, particularly in the United States. For example, if the U.S. fiscal deficit were to be about 6 percent of GNP over the next few years (rather than about 1 percent as it was during the 1960s and 1970s) and if only 20 percent of the additional deficit were to be financed through an absorption of foreign savings, either directly or indirectly, this could change the “normal” capital flows of the United States by some $30 billion, with a roughly opposite effect on the normal private capital flows of other industrial countries,24 In this case, a large part of the previously projected imbalances between underlying payments positions on current account and the positions necessary to “offset” normal capital flows would disappear.
Reliability of Estimates of Underlying Payments Balances and Sustainable Exchange Rates
In contrast to judgments of the sustainable exchange rate based on the competitiveness indices discussed in Section III, assessments based on underlying payments balance considerations are explicitly concerned with developments in both the current and capital accounts, and they allow for the impact of the stance of economic policies on the sustainable level of a country’s exchange rate. Nevertheless, as already suggested above, the effort to take account of these considerations increases the complexity of the analysis on which exchange rate assessments are based. The remainder of this section considers the difficulties inherent in projecting underlying payments balances, the experience of past underlying payments exercises undertaken by the Fund staff and, finally, the problems that arise in moving from these estimates to an assessment of the sustainable exchange rate.
The first technical issue that arises in estimating underlying payments balances is that of making allowance for the effects of past relative price changes on the underlying value of merchandise trade. For these purposes, the staff has made use of the world trade model. The most recent econometric estimates of the price elasticities of merchandise import and export volume in this model are presented in Table 4. In contrast to Table 3. which was intended to indicate the responsiveness of merchandise trade flows to the standard indicators of international competitiveness, the elasticities in Table 4 measure the direct responsiveness of merchandise trade to export and import prices. To give an idea of the extent to which merchandise trade flows respond to price changes with a lag, Table 4 presents both an “impact” elasticity for the response during the first six months and a “long-run” elasticity, which represents the total estimated response over a two-year to three-year interval. The estimates presented in the table are plausible in that all price effects operate in the expected direction. Nevertheless, the size of the standard error attached to each of these estimates makes it clear that it is very difficult to obtain accurate estimates of the effects of price developments on merchandise trade flows and of the time profile of these effects.
World Trade Model: Price Elasticities of Demand for Imports and Exports of Manufactures1
The table gives, for each country, the price elasticity of domestic demand for imports and of the foreign demand for its exports. The relative prices are defined such that these elasticities are expected to be negative. The impact elasticity gives the response in the current semester; the shortrun elasticity gives the response after 2 semesters (1 year); and the long-run elasticity gives the total response. Standard errors are given in parentheses.
World Trade Model: Price Elasticities of Demand for Imports and Exports of Manufactures1
Imports | Exports | |||||
Countrt | Impact elasticity | Short-run elasticity | Long-run elasticity | Impact elasticity | Short-run elasticity | Long-run elasticity |
Austria | – 0.03 | –0.36 | –0.80 | –0.39 | –0.71 | – 1.37 |
(0.16) | (0.65) | (0.16) | (0.22) | |||
Belgium | — | — | –0.70 | –0.18 | –0.59 | –1.55 |
(0.22) | (0.24) | (0.50) | ||||
Canada | –0.72 | –0.72 | –0.72 | –0.08 | –0.40 | –0.71 |
(0.17) | (0.17) | (0.22) | (0.26) | |||
Denmark | –0.55 | –0.93 | – 1.14 | –0.82 | –1.13 | – 1.13 |
(0.24) | (0.55) | (0.27) | (0.15) | |||
France | — | –0.49 | –0.60 | –0.20 | –0.48 | –1.25 |
(0.18) | (0.18) | (0.33) | ||||
Germany, Federal | –0.57 | –0.77 | –0.77 | — | — | – 1.41 |
Republic of | (0.20) | (0.11) | (0.44) | |||
Italy | –0.94 | –0.94 | –0.94 | — | –0.56 | –0.64 |
(0.23) | (0.23) | (0.79) | ||||
Japan | –0.16 | –0.72 | –0.97 | –0.59 | –1.01 | – 1.61 |
(0.22) | (0.27) | (0.16) | (0.25) | |||
Netherlands | –0.71 | – 1.22 | – 1.22 | –0.24 | –0.49 | –0.89 |
(0.31) | (0.43) | (0.26) | (0.50) | |||
Norway | — | –0.01 | –0.71 | –0.40 | –0.74 | – 1.49 |
(0.61) | (0.08) | (0.29) | ||||
Sweden | — | — | –0.94 | –0.27 | –0.73 | – 1.59 |
(0.53) | (0.19) | (0.32) | ||||
Switzerland | –0.25 | –0.25 | –0.25 | –0.28 | –0.42 | –0.73 |
(0.24) | (0.24) | (0.17) | (0.58) | |||
United Kingdom | –0.60 | –0.75 | –0.75 | — | — | –0.31 |
(0.27) | (0.21) | (0.20) | ||||
United States | — | – 1.06 | – 1.06 | –0.18 | –0.48 | – 1.67 |
(0.36) | (0.18) | (0.21) |
The table gives, for each country, the price elasticity of domestic demand for imports and of the foreign demand for its exports. The relative prices are defined such that these elasticities are expected to be negative. The impact elasticity gives the response in the current semester; the shortrun elasticity gives the response after 2 semesters (1 year); and the long-run elasticity gives the total response. Standard errors are given in parentheses.
World Trade Model: Price Elasticities of Demand for Imports and Exports of Manufactures1
Imports | Exports | |||||
Countrt | Impact elasticity | Short-run elasticity | Long-run elasticity | Impact elasticity | Short-run elasticity | Long-run elasticity |
Austria | – 0.03 | –0.36 | –0.80 | –0.39 | –0.71 | – 1.37 |
(0.16) | (0.65) | (0.16) | (0.22) | |||
Belgium | — | — | –0.70 | –0.18 | –0.59 | –1.55 |
(0.22) | (0.24) | (0.50) | ||||
Canada | –0.72 | –0.72 | –0.72 | –0.08 | –0.40 | –0.71 |
(0.17) | (0.17) | (0.22) | (0.26) | |||
Denmark | –0.55 | –0.93 | – 1.14 | –0.82 | –1.13 | – 1.13 |
(0.24) | (0.55) | (0.27) | (0.15) | |||
France | — | –0.49 | –0.60 | –0.20 | –0.48 | –1.25 |
(0.18) | (0.18) | (0.33) | ||||
Germany, Federal | –0.57 | –0.77 | –0.77 | — | — | – 1.41 |
Republic of | (0.20) | (0.11) | (0.44) | |||
Italy | –0.94 | –0.94 | –0.94 | — | –0.56 | –0.64 |
(0.23) | (0.23) | (0.79) | ||||
Japan | –0.16 | –0.72 | –0.97 | –0.59 | –1.01 | – 1.61 |
(0.22) | (0.27) | (0.16) | (0.25) | |||
Netherlands | –0.71 | – 1.22 | – 1.22 | –0.24 | –0.49 | –0.89 |
(0.31) | (0.43) | (0.26) | (0.50) | |||
Norway | — | –0.01 | –0.71 | –0.40 | –0.74 | – 1.49 |
(0.61) | (0.08) | (0.29) | ||||
Sweden | — | — | –0.94 | –0.27 | –0.73 | – 1.59 |
(0.53) | (0.19) | (0.32) | ||||
Switzerland | –0.25 | –0.25 | –0.25 | –0.28 | –0.42 | –0.73 |
(0.24) | (0.24) | (0.17) | (0.58) | |||
United Kingdom | –0.60 | –0.75 | –0.75 | — | — | –0.31 |
(0.27) | (0.21) | (0.20) | ||||
United States | — | – 1.06 | – 1.06 | –0.18 | –0.48 | – 1.67 |
(0.36) | (0.18) | (0.21) |
The table gives, for each country, the price elasticity of domestic demand for imports and of the foreign demand for its exports. The relative prices are defined such that these elasticities are expected to be negative. The impact elasticity gives the response in the current semester; the shortrun elasticity gives the response after 2 semesters (1 year); and the long-run elasticity gives the total response. Standard errors are given in parentheses.
A second technical issue arises in attempting to take account of the effects of factors other than relative price developments on merchandise trade. In recent years of low growth rates, the large gaps that have appeared between actual and potential output—together with uncertainties about such developments as the likely course of commodity prices and the influence of rigidities in goods and factor markets on economic activity—have made it much more difficult to project cyclical developments in the industrial countries. Since changes in relative cyclical positions are major determinants of the pattern of observed current account balances among industrial economies, the problem of determining each country’s underlying balance of payments position becomes especially complicated. The estimation of medium-term tendencies is even more unreliable, since the assumption that recent historical trends will continue for the next two or three years is often questionable.
A third problem that tends to increase the degree of uncertainty associated with the projected current account position is that little quantitative information is available on the responsiveness of trade in invisibles to price developments or changes in relative cyclical positions among the industrial countries.25 As a result, the balance on invisibles must be projected on an ad hoc basis.
Finally, and most important, a significant margin of uncertainty arises in implementing the underlying payments balance approach because, from an international standpoint, there is no unique pattern of current account balances among the industrial countries that is clearly superior to all others. Even when it is possible to identify the “normal” levels of net capital flows for the various countries, one can reasonably assume that deviations from these normal levels would have to be quite sizable before they became a source of concern. Thus, in addition to the margin of error arising from imperfect knowledge of important parameters such as the price elasticities of supply and demand in foreign trade, an inherent margin of uncertainty exists in the calculations.
Once the imbalance in the pattern of current account positions of industrial countries has been estimated on the basis of current exchange rates by comparing the present underlying current account positions of the various countries with their normal capital account balances, the last step is to calculate the deviations of actual exchange rates from their sustainable levels by estimating the change in exchange rates that would bring about the needed current account adjustments. The two formal economic models available to the staff for making such calculations are the world trade model and the multilateral exchange rate model (MERM). As already noted, the world trade model is an econometric model that provides estimated values for trade price elasticities. In contrast, the MERM is a mathematical simulation model in which direct and cross-price elasticities in the goods markets for each commodity category in each country are generated from a few “basic parameters” chosen from various econometric studies.26
In practice, a major difficulty in using these models for calculating the sustainable exchange rate, as already noted earlier, is that estimates of the price elasticities of exports and imports may not be sufficiently reliable to undertake this task with a high degree of confidence. Referring again to the world trade model’s estimated elasticities in Table 4, it is clear that even the cumulative long-run elasticities are subject to a substantial margin of error, as reflected in the relatively large standard errors.
Broadly speaking, similar conclusions apply to the empirical estimates of price elasticities of demand for imports and exports that are available in the academic literature. The most recent survey of this work undertaken by members of the Fund staff supports the widely held view that the long-run (i.e., two-year to three-year) price elasticity of import demand for a typical industrial country lies in or above the range –0.5 to – 1.0.27 For the price elasticity of demand for exports, this survey gives a “consensus estimate” of – 1.25 to –2.5. As Table 5 indicates, however, that the ten recent empirical studies that were surveyed in the staff’s study yield quite a wide range of estimated medium-term trade-price elasticities for each industrial country. The reasons for this lack of unanimity in the estimates probably lie in the poor quality of statistical information on foreign trade prices, the difficulty of distinguishing between price effects and nonprice factors such as quality, delivery dates, “after-sales service,” development of new products, and aggregation bias.
Long-Run Price Elasticities of Demand for Total Exports and Imports: Representative Estimates from Recent Empirical Studies
Long-Run Price Elasticities of Demand for Total Exports and Imports: Representative Estimates from Recent Empirical Studies
Long-Run Elasticity of Demand for Total Exports | Long-Run Elasticity of Demand for Total Imports | |||||
Lowest estimate | Highest estimate | Range of estimates | Lowest estimate | Highest estimate | Range of estimates | |
Austria | – 0.93 | – 1.21 | 0.28 | – 0.82 | – 1.42 | 0.60 |
Belgium | – 0.84 | – 1.57 | 0.73 | – 0.48 | – 2.90 | 2.42 |
Canada | – 0.23 | – 1.10 | 0.87 | – 0.20 | – 2.50 | 2.30 |
Denmark | – 0.56 | – 1.28 | 0.72 | – 0.23 | – 1.66 | 1.43 |
France | – 0.34 | – 2.27 | 1.93 | – 0.33 | – 1.80 | 1.47 |
Germany, Federal | ||||||
Republic of | – 0.29 | – 1.90 | 1.61 | – 0.24 | – 1.48 | 1.24 |
Italy | – 0.25 | – 3.29 | 3.04 | – 0.13 | – 1.42 | 1.29 |
Japan | – 0.50 | – 2.38 | 1.88 | – 0.72 | – 1.47 | 0.75 |
Netherlands | – 0.88 | – 2.72 | 1.84 | – 0.02 | – 1.65 | 1.63 |
Norway | – 0.80 | – 1.16 | 0.36 | – 1.19 | – 1.20 | 0.01 |
Sweden | – 0.47 | – 1.99 | 1.52 | – 0.76 | – 1.30 | 0.54 |
Switzerland | – 0.58 | – 1.51 | 0.93 | – 0.84 | – 1.35 | 0.51 |
United Kingdom | – 0.08 | – 1.47 | 1.39 | – 0.21 | – 1.38 | 1.17 |
United States | – 0.32 | – 2.32 | 2.00 | – 1.03 | – 1.73 | 0.70 |
Mean | – 0.51 | – 1.87 | 1.36 | – 0.51 | – 1.66 | 1.15 |
Median | – 0.49 | – 1.74 | 1.46 | – 0.41 | – 1.48 | 1.21 |
Long-Run Price Elasticities of Demand for Total Exports and Imports: Representative Estimates from Recent Empirical Studies
Long-Run Elasticity of Demand for Total Exports | Long-Run Elasticity of Demand for Total Imports | |||||
Lowest estimate | Highest estimate | Range of estimates | Lowest estimate | Highest estimate | Range of estimates | |
Austria | – 0.93 | – 1.21 | 0.28 | – 0.82 | – 1.42 | 0.60 |
Belgium | – 0.84 | – 1.57 | 0.73 | – 0.48 | – 2.90 | 2.42 |
Canada | – 0.23 | – 1.10 | 0.87 | – 0.20 | – 2.50 | 2.30 |
Denmark | – 0.56 | – 1.28 | 0.72 | – 0.23 | – 1.66 | 1.43 |
France | – 0.34 | – 2.27 | 1.93 | – 0.33 | – 1.80 | 1.47 |
Germany, Federal | ||||||
Republic of | – 0.29 | – 1.90 | 1.61 | – 0.24 | – 1.48 | 1.24 |
Italy | – 0.25 | – 3.29 | 3.04 | – 0.13 | – 1.42 | 1.29 |
Japan | – 0.50 | – 2.38 | 1.88 | – 0.72 | – 1.47 | 0.75 |
Netherlands | – 0.88 | – 2.72 | 1.84 | – 0.02 | – 1.65 | 1.63 |
Norway | – 0.80 | – 1.16 | 0.36 | – 1.19 | – 1.20 | 0.01 |
Sweden | – 0.47 | – 1.99 | 1.52 | – 0.76 | – 1.30 | 0.54 |
Switzerland | – 0.58 | – 1.51 | 0.93 | – 0.84 | – 1.35 | 0.51 |
United Kingdom | – 0.08 | – 1.47 | 1.39 | – 0.21 | – 1.38 | 1.17 |
United States | – 0.32 | – 2.32 | 2.00 | – 1.03 | – 1.73 | 0.70 |
Mean | – 0.51 | – 1.87 | 1.36 | – 0.51 | – 1.66 | 1.15 |
Median | – 0.49 | – 1.74 | 1.46 | – 0.41 | – 1.48 | 1.21 |
Summary
Assessments of a country’s sustainable exchange rate based on its projected medium-term overall payments position have the advantage that they attempt to allow for a comprehensive set of factors that influence the current and capital accounts of the balance of payments. In particular, they make explicit allowance for the effects of policy stances and cyclical positions. Although such an approach encounters substantial difficulties the effort to assess underlying payments balances in a multilateral framework helps to ensure some degree of consistency across countries in the judgments that are made about the set of sustainable exchange rates for industrial countries.
A solution that is attractive at first sight is to do the balance of payments analysis exclusively on the basis of current conditions, after adjustment for transitory disturbances and lagged effects of past relative price changes. The problem with this solution is that there is little doubt that private market participants’ expectations of future events do affect their current behavior. If there are major developments on the horizon that private market participants can anticipate, ignoring these developments will lead to the estimation of a sustainable rate that could in fact be unsustainable not only in the longer run, but also even in the short run.
The 1983 estimates are obtained from the World Economic Outlook: A Survey by the Staff of the International Monetary Fund, IMF Occasional Paper, No. 27 (Washington, April 1984), p. 187. (Hereinafter referred to as World Economic Outlook, 19–, No. –.)
Ibid., pp. 21–22 and 67–73.
For the United States, where the errors and omissions item has generally been very large, this calculation was made in two steps. First, the average ratio of capital flows excluding errors and omissions for the period 1975–78 and 1981–82 was applied to the figure for nominal GNP in 1982, to give an estimated outflow of $38 billion. Then the average of the errors and omissions item in 1981 and 1982 (equivalent to an inflow of $32 billion) was added to this figure to give a net outflow, including errors and omissions, of approximately $6 billion. For other countries, the average ratio was simply applied to total capital flows including errors and omissions.
For discussion, see IMF, World Economic Outlook, 1983, No. 21, pp. 161–67.
Whether this opposite effect would be of the same magnitude is uncertain because part of the increase in the absorption of foreign savings by the United States could have as a counterpart a decrease in the absorption of foreign savings by the developing countries.
See Marian E. Bond, “The World Trade Model: Invisibles,” Staff Papers, International Monetary Fund (Washington), Vol. 26 (June 1979), pp. 257–333.
See Jacques R. Artus and Rudolf R. Rhomberg, “A Multilateral Exchange Rate Model,” Staff Papers, International Monetary Fund (Washington), Vol. 20 (November 1973), pp. 591–611. Also Jacques R. Artus and Anne Kenny McGuirk, “A Revised Version of the Multilateral Exchange Rate Model,” Staff Papers, International Monetary Fund (Washington), Vol. 28 (June 1981), pp. 275–309.
See M. Goldstein and M.S. Khan, “Income and Price Effects in Foreign Trade,” Chapter 20 in Handbook of International Economics, Vol. 20, ed. by Peter B. Kenen and Ronald W. Jones (Amsterdam: North-Holland, 1984).