Background and Caveats
The quantification of equilibrium exchange rates is a somewhat murky area of economics. Conceptual frameworks typically rely on strong simplifying assumptions, and the empirical support for different approaches is mixed. From a policymaking perspective, however, such exercises are important: the IMF cannot effectively exercise its surveillance responsibilities without forming views about possible misalignments of exchange rates and considering the risks of sudden and sharp corrections of currency misalignments.
Success in explaining the behavior of exchange rates—and, hence, in providing meaningful empirical support for methodologies of assessing equilibrium exchange rates in terms of economic fundamentals—depends importantly on time horizons. Based largely on research that flourished during the 1970s and early 1980s, economists have concluded that models of the links between currency values and economic fundamentals perform little if any better than random walks in predicting the behavior of exchange rates in the short run.3 Furthermore, from surveys of commercial banks and other private financial institutions that actively participate in exchange markets, policymakers have learned that market participants condition their short-run trading strategies to a large extent on “technical analysis” of very recent trends or other patterns in the observed behavior of exchange rates, although they regard fundamentals as much more important than technical analysis in their long-term decision making.4 These perspectives discredit the notion that exchange rates should be expected to adjust, in the short run, to fundamentals-based equilibrium levels.5 Partly for that reason, most methodologies for defining equilibrium exchange rates have relied on models of behavior over a medium-or long-run horizon.6
One traditional methodology for defining equilibrium exchange rates is the purchasing power parity (PPP) approach, which hypothesizes that the nominal exchange rate between any two currencies should reflect closely the relative purchasing powers of the two monetary units, as indicated by national price levels.7 It is commonly expressed as a hypothesis that, over a given period of time, the percentage change in the nominal exchange rate between any two currencies will equal the difference between the percentage changes in the price levels of the two corresponding countries—or, equivalently, that the real exchange rate between the two currencies remains constant.
Although PPP has been discredited as a hypothesis about the relationship between exchange rates and national price levels in the short run,8 the econometric evidence is much more favorable to PPP as a hypothesis about medium-run or long-run behavior.9 This is illustrated by Figure 2. Changes in nominal exchange rates are measured along the horizontal axis, inflation differentials are measured along the vertical axis, and the diagonal 45 degree line represents the set of points that are consistent with the PPP hypothesis. The figure shows that PPP becomes a much more respectable hypothesis as the time horizon is lengthened, and that—at least over the past quarter century—the long-run PPP hypothesis fits the data for industrial countries remarkably well.10
Exchange Rate Changes Versus Inflation Differentials Over Different Time Intervals 1
1 Based on Flood and Taylor (1996). The plots are constructed from annual average data on the nominal exchange rates of 20 industrial country currencies versus the U.S. dollar, along with corresponding consumer price indices, for the period 1976–2000. Changes in exchange rates are measured along the horizontal axes; changes in CPIs along the vertical axes. The first panel plots 480 one-year changes (24 for each country); the second plots 80 nonoverlapping six-year changes (at annual rates) corresponding to the periods 1976–82, 1982–88, 1988–94, and 1994–2000; and so forth.Policymakers have relied on the PPP formula on various occasions to help them decide whether, and by how much, to adjust nominal exchange rates. Such use of the PPP formula to calculate an appropriate or “equilibrium” level for the nominal exchange rate requires a decision about the type of price (or cost) index on which the calculations are to be based, as well as an assumption about the equilibrium level of the real exchange rate. The latter is generally taken to be the observed level of the real exchange rate during some selected base year, or the average level of the real exchange rate over an extended period of time.
Sensitivity to the price (or cost) measure and base period that are chosen limits the usefulness of PPP as a normative approach. Figure 3 plots five measures of the real exchange rate between the pound sterling and the deutsche mark from 1970 through 2000. Each of the measures suggests that the pound was stronger than its equilibrium rate (proxied by long-term averages) against the deutsche mark (hence, the euro) in 2000; but the magnitude of the estimated deviation from equilibrium ranges from 10 percent to 40 percent, depending on the type of price or cost index used and the averaging period chosen.
Real Exchange Rates Between the United Kingdom and Germany, 1970–2000
(1970=100)
Sources: IMF, International Financial Statistics; and IMF staff calculations.Given the limitations of the PPP hypothesis as a normative approach, economists have been attracted to other methodologies for assessing exchange rates. CGER relies heavily on the macroeconomic balance approach, which makes quantitative assessments of exchange rates that are consistent with “appropriate” current account positions (external balance) when economies are operating at potential output (internal balance). Other approaches include an extended PPP methodology that takes account of Balassa-Samuelson effects, estimated reduced-form models of exchange rate behavior, and the approach of relying on general equilibrium frameworks. Appendix I outlines these alternative approaches and why CGER has chosen not to rely on them.
Key features of the macroeconomic balance approach have been prominent in calculations by the IMF staff since at least as far back as 1967, when views were developed about the appropriate size of the prospective devaluation of sterling.11 As refined by the IMF staff during the 1970s and early 1980s,12 and also used by Williamson (1985) and others in their early work on “fundamental equilibrium exchange rates” (FEERs), the macroeconomic balance approach was rooted in the balance of payments identity. In particular, these applications tended to define external balance in terms of “balanced” or “normal” or “target” or “underlying” capital flows, and then estimated the levels of real exchange rates that would equate current account balances—at positions of internal balance—to these notions of equilibrium capital flows. By contrast, CGER applications, as elaborated below, are rooted in the national income accounting identity that links the current account to the saving-investment balance. While the two identities are closely related, the shift in emphasis has been in the direction of relying less on ad hoc judgments about equilibrium capital flows, which are difficult to make in an environment of high and rising capital mobility, and more on models of the behavior of the saving-investment balance over the medium run.13
It is important to recognize at the outset that CGER’s analysis is subject to considerable limitations in generating definitive estimates and to various caveats in interpreting the assessments. A key limitation is that the quantitative assessments are inherently imprecise. In general, economists do not have the conceptual basis or empirical methodology for generating precise estimates of the exchange rate levels that are consistent with medium-run macroeconomic fundamentals. As will become evident below, imprecision enters CGER’s quantitative analysis in several different places. Such imprecision provides the rationale for generating assessments with both the macroeconomic balance framework and a PPP approach, and for characterizing CGER’s summary assessments as approximations or ranges that involve a significant element of judgment.
A further important caveat is that large deviations of exchange rates from their medium-run equilibrium levels (even if these could be estimated precisely) do not necessarily represent currency misalignments. In some cases, such deviations may reflect short-term cyclical factors rather than medium-run disequilibria. Moreover, regardless of their cause, deviations of exchange rates from medium-run equilibrium can sometimes be helpful from a short-run perspective; in particular, an appreciated currency can help cool an overheated economy, just as a depreciated currency can help reinvigorate a weak one. Furthermore, quite apart from cyclical considerations, deviations of exchange rates from medium-run equilibrium sometimes reflect a need for policy adjustment rather than an indication that markets are wrong.
Two related points should be kept in mind when interpreting CGER’s assessments. First, the assessment that a currency is substantially stronger or weaker than its medium-run equilibrium level does not necessarily imply a high probability of a sudden and sharp correction. Consequently, it is difficult to evaluate the assessments on the basis of their track record in anticipating exchange rate movements. The failure of the U.S. dollar to weaken significantly since early 1997, when CGER began to assess it as substantially stronger than its medium-run equilibrium level, can be interpreted as either a persistent error in CGER’s analysis or an implication of the possibility that deviations from medium-run equilibrium can be persistent. Second, there is no general answer to the question of whether policy actions should be taken when exchange rates appear to deviate substantially from their medium-run equilibrium values. The latter issue needs to be addressed on a case-by-case basis in the context of a broader assessment of macroeconomic circumstances.
Macroeconomic Balance Assessments
As noted earlier, the macroeconomic balance framework is based on the accounting identity that links a country’s current account balance (CUR) to the excess of domestic saving (S) over domestic investment (I).14
The current account balance is explicitly recognized to depend on the real exchange rate, which influences the volumes and prices of exports and imports. A complicating feature is that the effects of changes in the exchange rate on CUR usually take some time to materialize fully. CUR also depends on the levels of domestic and foreign aggregate demand (or other measures of economic activity) and various other factors.
The broad outline of CGER’s approach remains unchanged from the overview presented in Isard and Faruqee (1998) and is recapitulated here for convenience. The first step in applying the framework is to estimate each country’s underlying current account position, which is the relevant measure of CUR for assessing exchange rates from a medium-run perspective. The underlying current account is defined as the value of CUR that would emerge at prevailing exchange rates if all countries were producing at their potential output levels and the lagged effects of past exchange rate changes had been fully realized. Under normal assumptions, a country’s underlying current account position will be inversely related to the prevailing level of its real exchange rate, as depicted by the negatively sloped line labeled UCUR in Figure 4; that is, a decline (or depreciation) in the real exchange rate will normally improve the underlying current account. If the real exchange rate was R1, the first step in the macroeconomic balance approach would identify the underlying current account position as UCUR1.
The second step is to derive an estimate of each country’s equilibrium saving-investment position, which is interpreted as a balance that can be regarded as “normal” from a medium-run perspective. Such estimates focus on the right-hand side of equation (1) and also assume that countries are operating at potential output. In Figure 4, the equilibrium saving-investment balance is assumed to be independent of the real exchange rate,15 as depicted by the vertical S-I line.
The third step is to calculate how much real exchange rates would have to change, other things being equal, to be consistent with medium-run fundamentals—that is, to equilibrate, for each country simultaneously, the underlying current account position with the medium-run equilibrium saving-investment balance. Although this calculation is made in a multilateral framework, it is broadly similar to estimating the difference between R1 and R*, where R* corresponds to the medium-run equilibrium exchange rate at which the UCUR and S-I lines intersect.
The methodology also involves applications of judgment, both in implementing each of the steps in the macroeconomic balance approach and in considering whether and how much to adjust the resulting macroeconomic balance assessments from PPP perspectives and other information that may be relevant. Additional judgment is required in deciding whether the final quantitative assessments imply a need for policy adjustment in the near term or point to significant risks for the macroeconomic outlook.
Underlying Current Account Positions
CGER has focused on two approaches to generating estimates of underlying current account positions. One approach, developed by the IMF’s Research Department (RES), is based on a standard trade model that has a relatively simple structure and employs common equation specifications and parameter values across countries (see Appendix II). The model implicitly treats all current account outflows (inflows) as merchandise exports (imports); treats goods produced in different countries as imperfect substitutes;16 assumes that export (import) volumes depend on both the level of foreign (domestic) activity and the current and lagged values of the real effective exchange rate; and assumes that exchange rate changes are fully passed through into import prices, with no effect on (domestic-currency-denominated) export prices. The model-based estimates of underlying current accounts have the positive attributes of global consistency and transparency, but the aggregation of non-oil trade, oil trade, and other current account items (investment income flows and transfers) and the lack of country-specific detail are disadvantages.
A second approach is based on the current account projections generated by the IMF’s country experts in connection with the World Economic Outlook exercise. For industrial countries (and most other countries), the WEO projections are conditional on unchanged real exchange rates. Moreover, the projections for the final year in the five-year WEO horizon assume that economies are operating at potential output and, in that sense, can be interpreted as estimates of underlying current account positions. These projections have the advantage of incorporating the country-specific knowledge (including that from models and projections of national authorities) and judgments of the IMF’s area department staff. A disadvantage is that the separate projections for individual countries may lack global consistency;17 indeed, at the time of CGER’s semiannual assessment in March 2001, the staff’s WEO projections summed to a global current account deficit of 1 percent of projected 2006 world GDP, compared with a global deficit in 2000 of about 0.6 percent of GDP. This global inconsistency, which the WEO forecasts appear to share with most other forecasts, presents a problem for the CGER exercise, as discussed in more detail below.
CGER has chosen to rely primarily on the WEO projections as estimates of underlying current account positions. To avoid relying uncritically on the WEO projections, however, it has disciplined its assessments by generating alternative estimates of underlying current account positions from the globally consistent RES model and taking those estimates into account when shaping its summary judgments on exchange rates. It has also continued to use the RES model in the third step of the CGER process (see below) for calculating the changes in real exchange rates that would be needed to make the underlying current account balances consistent with medium-run equilibrium levels of saving-investment balances.
To strengthen the assessment process, the staff has recently begun to develop benchmarks that can be used as comparators for the individual components of the WEO projections (see Appendix II). The development of these benchmarks amounts, in effect, to an effort to construct an improved version of the RES model that disaggregates the current account into different components and employs a globally consistent and common (across countries) framework for modeling each component.
The benchmarks are not intended as an alternative set of projections. The simple analytic models on which they are based, while having the important property of global consistency, do not take account of country-specific factors. Moreover, history reveals considerable variation over time in observed outcomes relative to benchmark calculations for individual countries. Nevertheless, the benchmarks provide CGER with a consistent and disciplined approach for considering the possible directions of bias in estimates of underlying current account balances based on the WEO projections and, accordingly, in the corresponding assessments of exchange rates.
Saving-Investment Norms
The basis for CGER’s estimates of equilibrium or normal saving-investment balances, here referred to as S-I norms, is an estimated equation generated in the Research Department. This aspect of the present CGER framework remains considerably oversimplified but can be viewed as a significant advance in terms of global consistency.18
The assumptions and derivation of the S-I model are described in Appendix III.19 Updated estimates for the period 1971–99 yield the following long-run relationship:
where CUR is the current account (expressed as a ratio to GDP), SUR is the fiscal surplus (as a ratio to GDP) relative to the industrial country average, YPCAP is income per capita relative to that of the United States, and the ci are country-specific constant terms.
Several points may be noted about the general structure and interpretation of this equation. First, current account positions would not change if all countries were to reduce their fiscal positions eqiproportionately, or to experience the same rates of per capita income growth. Second, the use of the equation to estimate the medium-run equilibrium levels of saving-investment balances relates the S-I norms to structural fiscal positions and cyclically adjusted levels of relative per capita incomes, since the calculations are based on projections at the end of the WEO horizon, where output gaps are zero. Third, the structural fiscal positions that enter the S-I norms do not necessarily correspond to “desirable” fiscal balances; nor should connotations of desirability be attached to the equilibrium or “historically normal” S-I balances.
As elaborated in Appendix III, the long-run relationship described by equation (2) is derived from a model that takes account of cyclical variables, but country-specific interest rates play no explicit role. Accordingly, the present version of the S-I model abstracts from various factors that are relevant to capital flows and portfolio allocation decisions in a world in which claims on different countries are perceived to offer different expected returns and to be subject to different risks.20 Future efforts to strengthen the S-I model should attempt to capture the influence of such factors, but hopes for major improvements in this area are dimmed by the fact that economists have not yet had much success in empirical efforts to explain how such factors systematically influence the behavior of country-specific interest rate premia.
The parameter estimates in equation (2) imply that a 1 percentage point of GDP increase in a country’s relative structural fiscal surplus improves its current account by 0.23 percentage points of GDP. Changes in fiscal positions are thus found to have “non-Ricardian” effects; that is, an increase in public saving is not fully offset by a reduction in private saving.21 A 1 percent improvement in a country’s relative per capita income raises its saving-investment balance by 0.11 percentage points of GDP, consistent with the view that higher-income countries tend to be larger (net) exporters of capital, other things being equal. The framework implicitly captures the countercyclical effects of monetary policy on the output gap, which is included in the estimated equation (see Appendix III), but abstracts from any explicit influence of monetary policy on the medium-run behavior of S-I balances or real exchange rates.22
The saving-investment model has been estimated with country-specific constant terms to capture the effects of omitted variables that may influence the relative saving and investment rates of different countries. These constant terms capture, for example, the fact that a relatively low S-I balance has been historically normal for the United States, other things being equal, consistent with a relatively low observed saving rate for the United States, and perhaps also indicating a relatively strong desire of nonresidents to accumulate claims on the United States.
Equation (2) is a reduced-form relationship and should be interpreted carefully. The relative structural fiscal positions that appear on the right-hand side of the equation should not be regarded as exogenous; in addition to reflecting tax rates and expenditure levels that might reasonably be treated as exogenous, structural fiscal positions can be influenced substantially by factors that affect productivity growth and the level of potential output. Accordingly, changes over time in the relative importance of such factors can make the reduced-form parameters sensitive to the sample period over which they are estimated. It is also widely recognized that the effects of fiscal policy changes on the overall S-I balance can depend (both in magnitude and in sign) on how the policy changes affect perceptions about debt sustainability and the outlook for macroeconomic growth. Further work to better capture the effects of fiscal policy changes on overall saving-investment positions is an important direction for enriching the S-I model.
Although equation (2) is highly simplified and leaves considerable scope for improvement through future research efforts, it provides a consistent way of capturing the sensitivity of overall saving-investment balances to relative fiscal positions, stages of development (per capita incomes), and other unspecified factors operating through the country-specific constant terms. It can be regarded as a formula for generating estimates of equilibrium medium-run current account positions by adjusting historical averages—as captured by the country-specific constant terms—to account for changes over time in variables that one might reasonably expect to affect saving-investment balances. The estimated model has served the important purpose of providing a basis for pinning down a set of judgments on equilibrium saving-investment balances and maintaining a high degree of consistency in those judgments over time. Extensions of the model that succeeded in capturing the influence of additional explanatory variables would also result in different estimates of the country-specific constant terms. Such extensions would affect the time profiles of the S-I norms as functions of the explanatory variables, but they would not affect the historical average levels of the S-I norms, which are determined by observed data. Moreover, to the extent that moving to a more sophisticated model significantly changed the calculated values of the norms for the particular year on which the assessments were based (i.e., the final year of the WEO horizon), discontinuities in the assessments could be avoided—where warranted in CGER’s judgment—by making adjustments to the country-specific constant terms.
Figure 5 shows the S-I norms that are implied by equation (2) for the period from 1990 through 2006. The estimates for 2001–06 reflect the WEO projections of explanatory variables as of March 2001. The United States, the United Kingdom, Canada, Australia, and New Zealand are countries in which current accounts have on average been in deficit over the past three decades, which is reflected in the deficit positions of their S-I norms. The other industrial counties are all estimated to have equilibrium S-I surpluses. In the cases of Australia, New Zealand, and Canada, the historically normal deficits in their S-I balances presumably reflect the combined influence of relatively abundant natural-resource-based investment opportunities and relatively sparse populations. The largest surplus norms are associated with Norway and Switzerland, consistent with the effects on S-I balances of substantial oil wealth in the case of Norway and a relatively large net foreign asset position in the case of Switzerland.23
Saving-Investment Norms, 1990–2006
(In percent of GDP)
Source: IMF staff estimates.The changes in the norms from one year to the next, as derived from equation (2), stem from changes in the observed or projected levels of structural fiscal balances and potential output (cyclically adjusted income) per capita. In the United States the gradual upward trend (decline in the norm deficit) during the 1990s mainly reflects an improvement in the public sector’s structural budget position. In Japan the time profile of the norm reflects the widening of the relative structural fiscal deficit24 and relatively slow growth of potential output during the 1990s, along with the projection for a gradual reduction in the structural fiscal deficit during the five years ahead. For the euro area, the norm has been constructed by aggregating estimates for the individual member countries, with an adjustment to account for temporary effects of German unification in the early 1990s.25 For most of the other countries, except Switzerland, the variation in the norms primarily arises from changes over time in relative structural fiscal balances. For Switzerland, the downward trend in the norm mainly comes from the relatively slow growth of potential output per capita.
The simple structure of the equation used to calculate the S-I norms is one of the reasons that CGER’s assessments should be viewed as imprecise.26 This poses a challenge for IMF staff in continuing their efforts to generate improved econometric estimates of S-I behavior for the industrial countries in general and to take account of specific factors or structural changes that may have major influences on S-I behavior in individual countries.27 It bears repeating, however, that the general levels of the S-I norms reflect the average historical values of the S-I balances and would probably not be affected substantially by more sophisticated explanations of the observed data. Thus, the main message that emerges from recent CGER assessments—that the U.S. dollar is considerably stronger than its medium-run equilibrium level—would probably not be altered substantially by a more sophisticated explanation of the historical behavior of the S-I balance for the United States.28
Multilateral Exchange Rate Calculations
Step three of the process calculates the direction and magnitude of the implied exchange rate changes that, assuming no changes in policies or other variables, would bring currency values into line with medium-term fundamentals. This section highlights aspects of the analytical framework that are relevant for the interpretation of these calculations and also describes the treatment of the global discrepancy.
As noted above, the UCUR line in Figure 4 is negatively sloped to reflect the presumption that a lower real effective exchange rate is associated with an improvement in the current account over the medium term. In the logic of CGER’s WEO-based assessments, the position of the UCUR line is assumed to reflect the projected values of economic variables at the end of the five-year WEO horizon, when output is at potential and the lagged effects of past exchange rate changes have been realized. Thus, projected changes in economic fundamentals, including effects of announced policy changes, are already reflected in the position of the UCUR line. By contrast, any unanticipated changes in relevant economic variables over the WEO horizon, including changes that arise from unanticipated policy actions, would shift the position of the UCUR line.
The vertical S-I line in Figure 4 shows the normal level of the saving-investment balance determined in step two. The line is vertical because the normal level of the S-I balance (at potential output) is assumed not to depend on the exchange rate.29 Its position, like that of the UCUR line, reflects the projections for relevant economic variables (in this case, per capita income levels and structural fiscal balances) at the end of the medium-run WEO horizon.
Given the initial real exchange rate (R1) and underlying current account position (UCUR1), the amount of exchange rate adjustment that is needed to equilibrate the underlying current account with the equilibrium S-I balance depends on the slope of the UCUR line. The assumptions on which the calculations are based (see below) imply that the slope of the UCUR line depends on the openness of the economy. Countries with relatively high ratios of exports and imports to GDP have relatively flat UCUR lines and require smaller percentage changes in their real exchange rates, other things being equal, to achieve given changes in their trade volumes or underlying current account positions (as shares of GDP).
Unexpected changes in economic fundamentals—that is, deviations from the changes projected in the WEO—can shift the position of either the vertical S-I line or the negatively sloped UCUR line or both, and can thereby alter the real effective exchange rate that is consistent with medium-run fundamentals. For example, a greater-than-projected increase in (relative) per capita income or the (relative) structural fiscal surplus will shift the S-I line in Figure 4 to the right. The size of the implied change in the equilibrium exchange rate will depend on the extent of the shift in the vertical S-I line, on the slope of the UCUR line, and on the extent of any shift in the UCUR line. By themselves, shifts in the position of the UCUR line (due to unexpected changes in medium-run fundamentals that affect the current account through channels other than the real exchange rate) change the equilibrium level of the real effective exchange rate that is consistent with medium-run fundamentals, but they do not change the medium-run levels of either the saving-investment balance (assuming a fixed and vertical S-I line) or, correspondingly, the current account balance.
For a more complete picture of how the exchange rate assessments are affected by policy changes or other exogenous shocks, one needs to look behind the UCUR line shown in Figure 4. An understanding of how the UCUR line shifts requires an understanding of how the shocks affect the WEO projections that CGER relies upon as measures of underlying current account positions. Moreover, in judging the plausibility of the exchange rate assessments, it is important to reflect on the adequacy of the reduced-form S-I equation for analyzing the shock under consideration.
This can be illustrated by considering how a permanent reduction in tax rates would affect the equilibrium exchange rate depicted in Figure 4. Equation (2) describes the normal effect on the S-I balance; the reduction in the public sector surplus is largely but not completely offset by an increase in the private sector saving-investment balance, with the leftward shift in the S-I line (decline in the overall saving-investment balance) tending to raise the equilibrium level of the real exchange rate, other things being equal. But that is only part of the story. In addition, a permanent tax cut would normally lead to upward adjustments in the projections for potential output, absorption, and imports in the medium run, and the associated leftward shift in the UCUR line (decline in the projected current account surplus) would tend to lower the equilibrium level of the real exchange rate. Thus, in the normal case the shifts in the S-I line and UCUR line act in opposite directions on the equilibrium exchange rate. To the extent that the UCUR line shifted more than the S-I line, the equilibrium exchange rate would decline. In some cases, moreover, the effects on the S-I line may differ from the normal response. Just as fiscal consolidation can be expansionary if it eases concerns about debt sustainability and results in reductions in the interest rate premia faced by the country’s residents, tax cuts that created strong concerns about debt sustainability could increase precautionary saving, reduce investment, and shift the S-I line to the right. In such cases the effect on the equilibrium exchange rate would be unambiguously negative.
As a second example, an increase in a country’s per capita income that resulted from a permanent shock to productivity in its tradable goods industries would shift the S-I curve to the right. It would also improve the country’s international competitiveness (provided that factor costs in the tradables sector rose less than proportionately) and be reflected in a rightward shift of the UCUR line. The outcome could well be a higher real effective exchange rate, depending on the balance of the two effects.
In moving from the analysis of the equilibrium level of the real effective exchange rate of a single country, as depicted in Figure 4, to conclusions about the equilibrium configuration of exchange rates in a multicountry world, it needs to be recognized that an n-country world has only n-1 independent exchange rates. It is not feasible to apply Figure 4 independently to all countries (or regions) without imposing a mathematical requirement for global consistency. The general procedure for producing such multilaterally consistent calculations is described elsewhere.30, 31
The calculations reflect two sets of specific assumptions. For purposes of characterizing the responsiveness of current account flows (and hence the underlying current account) to changes in real exchange rates, CGER relies on the aggregated RES model described earlier. This model treats each country in the same globally consistent manner, and its parameters have been calibrated to reflect representative estimates from the literature,32 with the elasticities (percentage responsiveness) of export and import volumes to a given percentage change in the real exchange rate assumed to be identical across countries. Accordingly, in relatively open economies—that is, in countries with relatively high ratios of exports and imports to GDP—a given percentage change in the real exchange rate generates relatively large changes in trade volumes and current account balances as shares of GDP.
The second set of specific assumptions relates to the treatment of the global current account discrepancy. Although the multilateral calculation procedure itself is globally consistent (to a very close approximation), the WEO current account projections for individual countries add up to a global total current account balance that has tended to differ substantially from the sum of the estimated saving-investment norms.
Table 1 illustrates how openness and the global discrepancy influence the multilateral calculations. The table focuses on three countries, with country B representing a much more open economy than countries A and C. For countries A and B, the underlying current account positions are 2½ percentage points of GDP more in deficit (less in surplus) than the equilibrium S-I balances, while for country C the underlying position coincides with the S-I norm (column 1). The calculated estimates of how much prevailing exchange rates deviate from their medium-run equilibrium levels in multilateral or real effective terms (column 2) have two noteworthy features. First, the assessments of multilateral exchange rates for countries A and B are quite different, despite the identical gaps between their underlying current accounts and S-I norms. Second, the exchange rate of country C is assessed as needing to appreciate somewhat to reach its medium-run equilibrium level, despite the fact that the underlying current account of country C is exactly equal to the S-I norm.
Illustrative Assessments
↓Indicates that currency would need to depreciate to reach its medium-run equilibrium level; ↑ Indicates need to appreciate.
Illustrative Assessments
Multilateral Real Exchange Rate | |||
---|---|---|---|
Underlying Current Account | (As percentage deviation from estimated | ||
Minus S-I Norm | medium-run equilibrium level)1 | ||
(In percent of GDP) | Calculated | Summary judgment | |
(1) | (2) | (3) | |
Country A | −2.5 | ↓ 26 | ↓ >20 |
Country B | −2.5 | ↓7 | ↓ about 10 |
Country C | 0 | ↑ 4 | ↑ about 10 |
↓Indicates that currency would need to depreciate to reach its medium-run equilibrium level; ↑ Indicates need to appreciate.
Illustrative Assessments
Multilateral Real Exchange Rate | |||
---|---|---|---|
Underlying Current Account | (As percentage deviation from estimated | ||
Minus S-I Norm | medium-run equilibrium level)1 | ||
(In percent of GDP) | Calculated | Summary judgment | |
(1) | (2) | (3) | |
Country A | −2.5 | ↓ 26 | ↓ >20 |
Country B | −2.5 | ↓7 | ↓ about 10 |
Country C | 0 | ↑ 4 | ↑ about 10 |
↓Indicates that currency would need to depreciate to reach its medium-run equilibrium level; ↑ Indicates need to appreciate.
Part of the difference between the assessments for countries A and B is explained by the fact that the calculations reflect not only the discrepancies between underlying current accounts and saving-investment norms, but also the different degrees of openness of the two countries. Because country A is a relatively closed economy in comparison with country B, each percentage point of GDP adjustment in the current account requires a larger multilateral exchange rate adjustment in country A. Another part of the explanation, which also pertains to the assessment for country C, is associated with the global discrepancy.
As of early 2001, the WEO projections for individual countries summed to a global current account deficit of $422 billion in 2006, or 1 percent of the projected level of world GDP in 2006, compared with an estimated global discrepancy of -$182 billion in 2000.33 The average global discrepancy over the past two decades was a deficit of 0.4 percent of GDP, which CGER uses as a norm for the global S-I balance. This historical discrepancy ratio and the WEO projection for global GDP in 2006 implied a global discrepancy of -$168 billion in the S-I norms for 2006.
The large difference between the global discrepancy in the underlying current account positions and the global discrepancy in the S-I norms has a significant effect on the calculations. If the world as a whole was treated as a single country, a depreciation of 4.4 percent (relative to a hypothetical mirror-image country) would be required, other things being equal, to adjust its current account from the projected aggregate current account deficit ($422 billion) to the deficit level consistent with the historical average global discrepancy ratio ($168 billion). Analogously, independent comparisons of the underlying current account positions and S-I norms for individual countries would tend to suggest estimates of equilibrium exchange rates that were biased by about 4.4 percentage points on average. The mechanical calculation procedure is appropriately constrained to be multilaterally consistent, and it essentially removes the bias in a manner that reduces (increases) by 4.4 percentage points the amount that each currency needs to depreciate (appreciate) in multilateral terms to reach its medium-run equilibrium level, other things being equal. This is why column 2 suggests that the currency of country C needs to appreciate by about 4 percent in multilateral terms. It also explains, together with openness considerations, why country B’s estimated deviation from equilibrium is so small in proportion to that of country A.34
Given the significant adjustment implemented to remove the global discrepancy bias, as well as the limitations of the models used to generate the estimates of underlying current account positions and saving-investment norms, the exchange rate assessments that emerge from the application of the macroeconomic balance framework should not be regarded as precise. Retrospective applications to episodes that are widely regarded as extreme misalignments of the major currencies during the past two decades (in particular, exchange rates among the U.S. dollar, the Japanese yen, and the deutsche mark in February 1985 and April 1995, and among major European currencies in June-July 1992) would generally have delivered correct signals at the time.35 Nevertheless, the various sources of imprecision point to both the desirability of continuing efforts to improve the inputs to the calculations, including the global consistency of the WEO projections, and the need to also take other considerations into account in making judgments about equilibrium exchange rates.
Other Relevant Considerations and Summary Judgments
In moving from the macroeconomic balance calculations (Table 1, column 2) to summary judgments about exchange rates,36 which are expressed as approximations or ranges (column 3), CGER takes account of PPP perspectives. For countries in which deviations of underlying current account balances from S-I norms suggest substantial over- or undervaluation relative to medium-run fundamentals, it is important to ask whether PPP-based perspectives indicate that the country’s price or cost competitiveness is substantially weaker or stronger than its historical trend. Given the imprecision of each approach, CGER’s confidence in its judgments about currencies is significantly enhanced when both approaches support the same assessment.
The benchmark comparators for the WEO projections (as described earlier and in Appendix II)—which provide perspectives on possible directions of bias in the macroeconomic balance assessments—can also be useful in forming summary judgments. Similarly, it can be useful to consider different sets of PPP-based perspectives; CGER routinely looks at perspectives based on GDP deflators, consumer price indices, and unit labor costs. Other information that CGER considers includes private forecasts of exchange rates and market forward exchange rates; but this information has not played a significant role in CGER’s assessments to date.
For the illustrative assessments shown in Table 1, the PPP-based perspectives supported the assessment that the currency of country A was substantially stronger than its medium-run equilibrium level. The PPP-based perspectives (as well as the benchmark WEO comparators) also supported judgments that the currency of country B was somewhat further above its equilibrium level, and that of country C somewhat further below its equilibrium, than the amounts suggested by the macroeconomic balance calculations.
Applications in IMF Surveillance
The main motivation for CGER’s assessments of industrial country exchange rates is to identify cases in which currency values appear to be substantially out of line with medium-run macroeconomic fundamentals. The identification of such cases and the characterization of how much exchange rates deviate from medium-run equilibrium levels reflect a combination of mechanical calculations and judgment. Starting with its assessments in early 1997, CGER has judged the U.S. dollar to be substantially stronger than its medium-run equilibrium level, based on both macroeconomic balance calculations and PPP-based perspectives. It is clear, however, that CGER would have felt much less confident in drawing that conclusion from the macroeconomic balance calculations in the absence of evidence (based on several different PPP-based perspectives) of a substantial erosion of international price and cost competitiveness. In that sense, it is difficult to separate the influence of the quantitative macroeconomic balance methodology from the role played by price- and cost-competitiveness indicators. Both factors contribute importantly to CGER’s assessments, and judgments are involved in weighing them, particularly when they suggest significantly different conclusions.
At times, countries need to make decisions about the precise values at which to set nominal exchange rates. This was the case when the Italian lira rejoined the European Exchange Rate Mechanism in 1996, and when decisions were made about the rates at which to lock the euro candidate currencies to the euro. While mechanical applications of CGER’s macroeconomic balance framework generated point estimates of appropriate exchange rates on those occasions, the IMF staff was cognizant of the imprecision of its estimates and would not have argued with any confidence that those estimates were more appropriate than the rates that the European authorities eventually agreed upon. The IMF’s assessments in those cases also reflected the combined influence of mechanical macroeconomic balance calculations, inspection of various price- and cost-competitiveness indicators, and judgment. Had IMF management concluded that the exchange rates under consideration by the European authorities were substantially out of line with medium-run fundamentals, it would have found an appropriate channel for communicating its analysis and conclusions.
As emphasized earlier, the assessment that a currency is substantially out of line with medium-run fundamentals may have different explanations or interpretations. Accordingly, the implications for the IMF’s policy advice in the exercise of its multilateral and bilateral surveillance responsibilities depend on the circumstances. One issue is whether prevailing exchange rates are appropriate or helpful from a short-run perspective. This perspective received prominence during much of the 1997–99 period in considering the implications of the relatively strong U.S. dollar and pound sterling when the U.S. and U.K. economies appeared to be at risk of overheating, while Japan and many European economies were relatively weak.
A second issue is whether deviations of exchange rates from their estimated medium-run equilibrium levels signal a need for policy adjustment. In considering this possible interpretation, it is relevant to note that even after policy changes have been proposed by national authorities, market participants may not view implementation of the proposed measures as fully credible. This was the case in Italy during 1995, when a large underlying current account surplus suggested that the lira was undervalued from a medium-run perspective, while large market interest rate premia on lira-denominated assets suggested that the weak lira reflected market concerns that the political process might not deliver the fiscal adjustment that national authorities were seeking. After the fiscal adjustment was legislated and the political uncertainty was resolved, the lira appreciated to a level that appeared to be broadly consistent with medium-run fundamentals, and in late 1996 Italy rejoined the European Exchange Rate Mechanism.
In some circumstances, adjustment of exchange rates toward estimated medium-run equilibrium levels would appear to be helpful from a cyclical perspective. This was judged to be the case in the spring of 1995, when the U.S. dollar was weak relative to the Japanese yen and the deutsche mark and the U.S. economy seemed at risk of overheating, while the Japanese and European economies were weakening. In those circumstances, the staff and management of the IMF pushed for coordinated interest rate actions by United States, Japan, and Germany, reinforced by medium-term fiscal adjustment in the United States and Europe and market opening measures in Japan.37 Such proposals for policy adjustments were aimed primarily at moving the economies toward internal balance and, secondarily, at bringing exchange rates into better international alignment. In the event, while monetary policy was eased in Germany and Japan in response to weakening cyclical conditions, the case for higher interest rates in the United States was subsequently eroded by signs of a greater-than-expected slowdown in activity (partly because of spillovers from the economic crisis in Mexico), setting in motion a lowering of U.S. interest rates from early July.
In still other circumstances, deviations of exchange rates from their estimated medium-run equilibrium levels may create concerns at times when prevailing interest rate levels are appropriate for the needs of the domestic economy. Such circumstances raise the question whether policy authorities should attempt to influence market perceptions through foreign exchange intervention or public pronouncements.
These examples indicate that the quantitative estimates that emerge from CGER’s analysis can only be regarded as the starting point for judgmental assessments of how to interpret cases in which exchange rates appear to deviate substantially from their medium-run equilibrium levels. Such deviations can have a range of different interpretations and policy implications and need to be considered on a case-by-case basis.
In addition to considering whether any near-term policy actions might be called for when exchange rates appear to deviate substantially from their medium-run equilibrium levels, it can be important to consider the risks to national economic performances and the global macroeconomic outlook. Although substantial deviations from equilibrium do not necessarily imply a high probability of sudden and sharp corrections, an analysis of what such corrections might imply is an integral part of the exercise of the IMF’s multilateral surveillance responsibilities. In that connection, the World Economic Outlook has paid considerable attention in recent years to addressing the possible implications of a sudden and sharp depreciation of the U.S. dollar.