Their interrelationship and the yen/dollar experience, 1979–82
The question of the role of interest rates in sustaining the current configuration of exchange rates has evoked considerable attention in recent years but no simple explanation of the relationship has emerged. Even close observation of the behavior of foreign exchange markets offers conflicting evidence. Over a period of several weeks or even months the movements of exchange rates between major currencies can appear to follow movements in interest rate differentials quite closely; then suddenly the relationship seems to dissolve, and changes in other factors appear to be the major determinants of exchange rate movements. One problem with much of the discussion on the relationship between interest rates and exchange rates has been a tendency to base conclusions on simple two-variable correlations rather than on analyses that incorporate the effects of a number of influences. Data for the dollar/yen exchange rate show that once the effects of other variables on exchange rates are accounted for, the relationship between interest rates and exchange rates seems quite robust.
Determinants of variability
Evidence suggests that over long periods of time, exchange rates tend to move to maintain purchasing power parity between currencies. While PPP between two currencies is not always easy to define, in general terms it can be viewed as the structure of relative prices that will give rise to sustainable current account balances between countries. In the absence of major disturbances, PPP is assumed to imply that exchange rates should move to offset fluctuations in relative prices between countries.
In fact, PPP seems to hold only over relatively long periods of time. In the short to medium term, the far more volatile asset markets tend to dominate exchange rate movements. Asset markets are central for two reasons. First, because an exchange rate can be viewed as the relative price of two assets, it is one of the main variables that adjusts to balance the supply of and demand for the stock of assets denominated in a currency. Second, capital markets are capable of responding much more rapidly than goods markets to perceived changes in conditions. The primary factors in asset markets that affect actual or potential capital flows are changes in actual and expected relative rates of return—arising from, among other factors, changes in actual or expected interest rate differentials, in expected movements of the exchange rate, and in the relative riskiness of holding various assets.
Interest rate differentials. To the extent that exchange rates are expected to move in line with inflation rate differentials, relative real rather than nominal interest rates provide the best measure of relative rates of return. It should be emphasized, however, that real interest rates are not directly observable and measures for them are only as accurate as the proxies used for inflationary expectations. Nevertheless, nominal interest rate differentials may give a misleading indication of relative rates of return since, like expected exchange rate changes, they are often related to expected rates of inflation.
For example, an increase in the nominal interest rate differential between, say, the United States and Japan in favor of the United States may reflect an increase in expected U.S. inflation and a corresponding expected depreciation of the dollar. This higher nominal U.S. interest rate would not represent an increased attractiveness of dollar relative to yen assets. Alternatively, an increase in the nominal interest differential in favor of the United States could occur with no increase in inflationary expectations. There would then be no change expected in the equilibrium exchange rate and, indeed, with a higher expected return on dollar assets, the immediate impact would be to strengthen the dollar. If the response of the real exchange rate to the real interest rate differential is examined, the difference between these two cases is clear. In the first case, there would be no change in either the real interest rate differential or the real exchange rate. In the second case, the real interest rate differential in favor of the United States would increase, and the real value of the dollar would appreciate.
Since interest rates differ according to the period to maturity of each asset, it is difficult to choose the asset maturity for which interest rates should be compared to exchange rate developments. Strictly speaking, interest rates on the whole spectrum of maturities should affect exchange rate behavior. Relating exchange rate movements to changes in interest differentials based on two different maturities may give quite different results when the term structure of interest rates in either country is not stable. Generally, however, a given change in long-term interest rate differentials should evoke a larger change in exchange rates than the same change in the short-term differential. Since interest rates provide an indication of expectations about interest rate developments, an equal change in long-term and short-term interest rate differentials will indicate expectations that the new differential will persist beyond the short term.
For given expectations about the future level of the exchange rate, the longer a given change in the interest rate differential between two countries is expected to last, the larger will be the immediate change in the exchange rate. This occurs because the immediate exchange rate change must be just large enough to make room for an expected change in the exchange rate during the holding period that will offset the interest rate differential and bring into balance relative rates of return over the entire period. Thus, to the extent that a change in the long-term interest rate differential reflects expectations that the differential will persist, it should have a larger effect on the current exchange rate. Nevertheless, most empirical analyses relate short-term real interest rate differentials to exchange rate behavior. This is because comparable short-term interest rates for a variety of countries are relatively easy to identify and it is generally more plausible to use proxies for expected inflation rates for short periods in the future.
Exchange rate expectations and risk. While neither exchange rate expectations nor the market’s views on prospective developments in variables affecting such expectations are observable, it is commonly believed that they are continuously subject to revision on the basis of “news” concerning virtually every facet of the economy. This continuous revision makes exchange rate expectations and, in turn, actual exchange rates, more volatile than the actual behavior of many of the variables that affect them. Although the movements of these are often obscure, they should not be misinterpreted as negating the more easily observable relationship between interest rate differentials and the exchange rate.
Generally, among the wide variety of factors that affect exchange rate expectations, many are political and not amenable to quantification; others are simply elements in a broad analysis of the prospects for an economy’s growth, inflation rate, and external balance. Among the more obvious factors for which proxies, at least, can be observed are expected inflation differentials, as captured in calculations of real interest rate differentials, and current account developments, which indicate the need for exchange rate adjustments to alter competitiveness to return to an equilibrium external position. In this context, the link between the current account and exchange rate expectations is based on the premise that unexpected changes in the current account alter the market’s view of the level of the real exchange rate that would be compatible with a sustainable current account. On a highly simplified level, it is assumed that expectations about the current account are based on its present level, so that actual changes in the current account affect changes in the exchange rate or, alternatively, the level of the current account affects the level of the exchange rate.
As with any portfolio allocation decision, investors’ international portfolio choices depend not only on expected rates of return but also on the expected variability of the return on the total portfolio. In selecting assets denominated in different currencies, investors are therefore influenced by the effect on the variability of the total return on their portfolios (in domestic currency terms) of additional holdings of assets denominated in any particular foreign currency. This influence is reflected in the patterns of variability between various exchange rates and in changes in the amounts of net outstanding foreign assets of each country (which is approximately equivalent to a country’s cumulated current account balance).
The yen/dollar experience
By the early part of the period 1977–82, the integration of Japanese and international financial markets was already underway as a result both of the gradual lifting of Japan’s restrictions on capital inflows and outflows and of the emergence of wider and deeper markets for Japanese financial assets that were internationally competitive. Consequently, in examining developments in the dollar/yen exchange rate during this period, a generally strong link is evident between movements in the exchange rate and in interest rate differentials between the United States and Japan. In addition, the influence of the other variables discussed above on the dollar/yen rate is examined.
One of the key determinants of long-term movements in the exchange rate was relative costs or prices between Japan and its major trading partners or competitors. Chart 1, which shows bilateral and effective exchange rates adjusted for relative prices and costs between Japan and its trading partners, suggests that over long periods (of about a decade), exchange rates of the yen have tended to adjust to maintain PPP. Nevertheless, deviations from average relative prices have been as large as 40 percent. These short- to medium-term deviations from PPP appear to reflect changes in factors affecting trade in assets rather than in factors affecting trade in goods.
Chart 1Measures of the real exchange rate, 1970–831
Source: IMF, Data Fund.
1 Rising indices imply falling competitiveness.
2Japanese values vis-à-vis the weighted average of the other currencies.
Chart 2 shows real short-term interest rates for the United States and Japan, the real short-term interest rate differential, and the real dollar/yen exchange rate. The relationship between the real exchange rate and the real interest rate differential must be examined with two caveats in mind. First, the chart shows only one dimension of the relationship—that between current interest rates and the exchange rate. Since expectations of future interest rate developments are probably even more volatile than actual interest rates, an important element explaining exchange rate variability is missing from the chart. Second, as mentioned above, numerous other factors can overwhelm and obscure the influence of interest rates on the exchange rate. Nevertheless, for periods when other factors obviously dominated exchange rate behavior, the chart shows a reasonably clear, positive relationship between the two.
Chart 2Real interest rates and the real exchange rate, 1977–83
Sources: IMF, Data Fund and International Financial Statistics.
Note: Vertical lines indicate the six subperiods discussed in the article.
1The real interest rate at time t is calculated as the nominal interest rate less the annualized percentage change between t and t-6 in a six-month centered moving average of the consumer price index. The nominal interest rate used for Japan is the three-month Gensaki rate and that for the United States is the three-month Treasury bill rate.
2The real exchange rate is an index of the nominal exchange rate (dollar yen) multiplied by the ratio of the consumer price index in Japan to that in the United States. The average value in 1980 equals 100.
• From early 1977 through mid-1978 (subperiod I), the real exchange rate and the real interest differential were positively related—the interest differential moved in favor of Japan by 3.8 percentage points and the yen appreciated by 27 percent. Clearly, however, not all of this appreciation reflects the influence of interest rate differentials. A highly simplified attempt to quantify the relationship between the real short-term interest rate differential and the real exchange rate between the yen and the dollar, which incorporates the influence of other factors, indicates that a 1 percentage point increase in the interest differential (Japan-United States) gives rise, on average, to a 1.3 percent appreciation of the yen.
• This relationship dissolved almost completely during the latter half of 1978 and the first half of 1979 (subperiod II) when the real value of the yen appreciated to a peak in late 1978 before depreciating sharply, while the interest differential moved slightly more in favor of Japan. The large current account imbalances of both Japan and the United States were reversed quite suddenly during this period—a factor which probably altered exchange rate expectations and therefore dominated the interest rate/exchange rate relationship (Chart 3).
Chart 3Current accounts of Japan and the United States and the real exchange rate, 1974–83
Source: IMF data.
• Between the third quarter of 1979 and mid-1980 (subperiod III) a positive relationship reemerged. Between August 1979 and March 1980 the real interest differential edged downward by 1.3 percentage points, while the yen depreciated by 15 percent in real terms. From March until June 1980, a 4.5 percentage point increase in the real interest differential appears to have been a major factor behind the dramatic 13 percent rebound in the real value of the yen.
• During the latter half of 1980 (sub-period IV), real interest rates in the United States increased sharply in relation to those in Japan, yet the yen appreciated strongly. This negative relationship appears to have been related at least to some extent to a sharp improvement in economic fundamentals in Japan—the relatively favorable wage settlement for major corporations in April 1980, the sharp turnaround in Japan’s current account deficit in the second quarter of the year, and the emerging evidence that Japan had more successfully dealt with the inflationary impact of the oil price increase than had other major industrial countries.
• During 1981 and the first two months of 1982 (subperiod V), a positive relationship again became apparent. Considerable variability of the real interest rate differential between January and August 1981 coincided with a 17 percent depreciation of the real value of the yen. Then, between August and November, a 2.7 percentage point increase in the interest differential was accompanied by a 4 percent appreciation of the yen. The reversal of the interest differential in December, associated with a further sharp rise in interest rates in the United States, closely coincided with the renewed weakness of the yen.
• Developments after February 1982 (subperiod VI) are often cited as evidence that the relationship between the exchange rate and interest rate differentials is tenuous at best. Between February and September, the differential moved in favor of Japan by 6.5 percentage points, while the yen, interrupted only by a brief show of strength in May, depreciated by more than 12 percent in real terms. During the last two months of the year, the interest differential again moved in favor of the United States by 1.7 percentage points, while the yen staged a dramatic recovery against the dollar, appreciating by almost 12 percent.
The apparent suspension of the conventional link between the interest differential and the exchange rate during the period after February 1982, however, may, for a number of reasons, more accurately be interpreted as an example of the ability of other factors to overpower that link. First, during the first half of the year, it became evident that the widely expected strengthening of Japan’s current account during 1982 would not materialize owing to an unanticipated weakening of export growth. Second, during the middle of the year, there appears to have been a marked deterioration in public perceptions of the strength of the emerging recovery in Japan and of the ability of the authorities to reverse the growing fiscal imbalance. A substantial shortfall in the outcome of budgetary revenues for fiscal year 1981 (ending March 31, 1982) was announced in May 1982, while the previous acceleration of private consumption expenditures was reversed during the third quarter of the year. These factors may have created some uncertainty among market participants as to the future course of economic policies in general, and interest rates and financial market conditions in particular.
It also seems likely that this uncertainty was aggravated both by the highly publicized official statements made in mid-September 1982 that the country was suffering from a “financial crisis” in connection with the fiscal situation and by events surrounding the change in leadership of the country during October and November. The fact that, throughout the year until November, the yen weakened not only against the dollar but also against virtually all major currencies suggests that the uncertainty aroused by these developments depressed expectations of the yen’s future value, which dominated the effect of the movement of interest differentials in favor of Japan.
Finally, during much of 1982, other major currencies were also considerably weaker against the dollar than developments in interest rate differentials would have indicated. This suggests that some change in the attractiveness of dollar assets may have occurred during this period—as a result, for example, of the decline in inflationary expectations in the United States and of a change in the perception of the relative safety of U.S. assets in the face of uncertainties in international financial markets.
As noted earlier, the relative attractiveness of assets denominated in any particular currency is affected not only by interest rate differentials but also by the expected future value of the exchange rate. It would be impossible to list, let alone review, the enormous variety of factors that affect expectations, but two easily observed factors appear to play an especially important role in exchange rate movements—relative expected rates of inflation (as embodied in real interest differentials) and current account balances. Chart 3 shows that between 1974 and early 1981, a strong positive relationship existed between the Japanese current account and the real value of the yen vis-à-vis the dollar. A factor contributing to the strength of this relationship was the generally negative correlation between the Japanese and U.S. current accounts, which helped provide unambiguous signals on needed changes in competitiveness. After early 1981, however, the relationship between the Japanese current account and the value of the yen weakened considerably. During the second and third quarters of 1981, the sharp increase in the Japanese current account surplus and some weakening of the U.S. current account coincided with a depreciation of the yen. Subsequently, though the Japanese current account generally remained in surplus and showed no trend, the real value of the yen generally declined. Nevertheless, the sharp deterioration in the U.S. current account in the third quarter of 1982 preceded by only several months the rebound in the value of the yen.
Three factors may be responsible for some apparent weakening of the current account/exchange rate link. First, after late 1980, U.S. and Japanese current account developments showed far more similarity than during the 1970s, and, therefore, jointly indicated less of a need for changes in competitiveness. Second, as pointed out above, the actual current account surplus for Japan in 1982 was substantially lower than original forecasts. Despite its relative strength, the current account outcome probably, therefore, had a depressing effect on the exchange rate. Finally, the threat that current account imbalances may be corrected through the intensification of protectionist measures against Japanese goods rather than through exchange rate adjustments, as in the past, may have eroded the influence of current account developments on expectations about exchange rates.
* * *
This discussion has identified a positive and empirically significant relationship between movements in the real short-term interest rate differential between Japan and the United States and the dollar/yen exchange rate. However, because of the important influence of highly volatile expectations on exchange rate movements, there were lengthy periods between 1977 and 1982 during which the link between the exchange rate and interest rate differentials has been obscured. In particular, current account developments, which indicated the need for substantial changes in competitiveness, at times led to large changes in exchange rate expectations, the effects of which dominated the relationship between interest differentials and exchange rates. Although the relationship of the exchange rate to both the real interest rate differential and current account developments is robust, the behavior of these two explanatory variables alone cannot fully explain many of the more dramatic movements in the dollar/yen exchange rate. For example, during the period of very rapid depreciation of the yen, between the first quarter of 1981 and the second quarter of 1982, movements in the real interest rate differential and in current account developments explained only about one half of the real depreciation of the yen. Clearly, therefore, other factors affecting public perceptions about such things as the future course of economic policies and financial conditions have had a large impact on the pattern of exchange rate movements.
Do you wish to receive
FINANCE & Development regularly?
New readers must either type or print in capital letters their names and addresses and identify their professional category from the list given at right when applying for inclusion in our mailing lists.
Regular readers are requested to send us their latest mailing label or previous address when requesting a change of address.
Professional category (Mark one only)
Commercial bank Other financial institution
International or regional organization
Faculty—college or university
Other government agency
Finance & Development,
Washington, D.C. 20431 USA