Abstract

Economists have shown little ability to explain exchange rate movements from day to day, month to month, or even year to year. As revealed by several extensive studies undertaken in the early 1980s, which focused on the currencies of major industrial countries, the most widely used structural models of exchange rates could not outperform a naive random-walk model over horizons of twelve months or less, even in predicting the realized behavior of exchange rates using the realized values of the “exogenous” explanatory variables (Meese and Rogoff,1983a and 1983b). Moreover, subsequent efforts to improve the structural models have achieved very limited success in outperforming the random-walk model over short-term horizons.1

Economists have shown little ability to explain exchange rate movements from day to day, month to month, or even year to year. As revealed by several extensive studies undertaken in the early 1980s, which focused on the currencies of major industrial countries, the most widely used structural models of exchange rates could not outperform a naive random-walk model over horizons of twelve months or less, even in predicting the realized behavior of exchange rates using the realized values of the “exogenous” explanatory variables (Meese and Rogoff,1983a and 1983b). Moreover, subsequent efforts to improve the structural models have achieved very limited success in outperforming the random-walk model over short-term horizons.1

Much greater success has been achieved, however, in explaining the behavior of exchange rates over relatively long time intervals. More specifically, nominal exchange rates and ratios of national price or cost levels have been observed to change by similar orders of magnitude over long periods. This can be seen in Figure 2-1, for example, which plots for each APEC member the average annual percentage change in its nominal exchange rate against the U.S. dollar over the period 1973–94 (vertical axis) versus the average annual percentage change in the ratio of its GDP deflator to the U.S. GDP deflator (horizontal axis). The chart shows a strong correlation between changes in nominal exchange rates and changes in ratios of national price levels over the past two decades. Although the average annual changes have been approximately the same in many cases,2 in some instances they have differed considerably. Consistently, a growing literature has challenged the appropriateness of models that assume that nominal exchange rates and ratios of national price levels have identical trends over the long run.3

Figure 2-1.
Figure 2-1.

Nominal Exchange Rates and Ratios of GDP Deflators1

(Average annual percent change, 1973–94)

1Nominal exchange rates are measured in U.S. dollars per unit of local currency. The ratio of price levels is the U.S. GDP deflator divided by the domestic GDP deflator. Annual averages are based on the data available within the 1973-94 period. For Chile and Mexico, not shown, the horizontal and vertical coordinates are (–21.3, –21.7) and (–24.6, –23.8), respectively.

Models of the long-run behavior of exchange rates are particularly relevant for assessing the extent to which observed changes in currency values are likely to persist. Accordingly, economists are continuing to address a number of unresolved issues about the nature of the long-run relationships between exchange rates and national price or cost levels. This section reviews these issues and summarizes existing attempts to shed light on them empirically. It also examines empirical evidence from the APEC region,4 which presents a wide range of economic experiences, including a number of economies in which real output has expanded rapidly, thus providing an attractive focus for a study of the long-run behavior of exchange rates.

Historical Perspectives for the APEC Region

The theories addressed in this section focus on the behavior of nominal exchange rates relative to various ratios of national price or cost levels. The ratio of the domestic national price level (P) to the foreign price level (P*), after multiplying by the nominal exchange rate (S) to convert domestic prices into foreign currency units, is commonly referred to as the real exchange rate (Q):

Q=SP/P*.(2-1)

Widely used measures of real exchange rates include those based on ratios of GDP deflators, consumer price indices, wholesale price indices, export price indices, and unit labor costs.5

Figure 2-2 shows time series of annual data on four different real exchange rate measures for each of the APEC members vis-á-vis the United States.6 Three of the real exchange rate measures—in particular, those based on GDP deflators, consumer price indices, and export price indices—are constructed as the nominal exchange rate (in U.S. dollars per unit of domestic currency) multiplied by the relevant domestic price level divided by the corresponding U.S. price level.7 By contrast, the real exchange rates based on Penn price relatives are constructed from estimates of the relative domestic and U.S. prices for individual product categories, without using data on nominal exchange rates (a detailed description is provided in the discussion of empirical evidence, below).8 The graphs in Figure 2-2 start in 1960 but in some cases are limited by data availability. Figure 2-3 shows corresponding time series of multilateral real exchange rates based on national-source GDP deflators, along with the associated bilateral rates vis-á-vis both Japan and the United States.

Figure 2-2.
Figure 2-2.
Figure 2-2.

Real Exchange Rates vis-á-vis the United States

(1973 = 100)

Figure 2-3.
Figure 2-3.
Figure 2-3.
Figure 2-3.

Multilateral and Bilateral Real Exchange Rates1

(1973 = 100)

1 Real exchange rates are based on GDP deflators. Multilateral rates are constructed using IMF Information Notice System (INS) total competitiveness weights, in which the weight of countryc in the multilateral exchange rate for countryjreflects the degree to which goods produced by the two countries compete directly in the home markets of countriesc and j and indirectly in third markets (see McGuirk, 1987).

Figure 2-2 reveals that, for most APEC members, the different measures of real exchange rates have diverged significantly over certain periods while also exhibiting fairly similar cyclical behavior.9 The discussion of the empirical evidence and the conclusions suggested by the analysis (below) focus on the relevance of both the cyclical similarities and the differences in long-run trends. Figure 2-3 shows that in virtually all cases the overall changes in multilateral rates have been somewhere in between the overall changes in bilateral rates against the dollar and against the yen; this reflects the importance to APEC members of trade with both Japan and the United States, together with the overall appreciation of the yen against the dollar. Figures 2-2 and 2-3 also reveal striking contrasts in exchange rate trends among different APEC members. With regard to the real exchange rates based on GDP deflators, consumer price indices, and Penn price relatives, Japan, Korea, and Taiwan Province of China have experienced substantial real appreciations vis-á-vis the United States since 1973; several other APEC members have experienced small-to-moderate depreciations; and China and Indonesia have had substantial depreciations.

In analyses of long-run trends in real exchange rates among the industrial countries, the real appreciation of the yen over the past two decades is generally associated with the relatively rapid expansion of output in Japan. As the analysis in this section shows, however, data for the APEC region suggest that the long-run relationship between output growth and real exchange rate appreciation is not robust. In particular, there are several members of APEC in which output growth since 1973 has been significantly more rapid than in Japan, while real exchange rates (vis-á-vis the United States) have not appreciated. Some reasons why rapid output growth has not always been accompanied by real exchange rate appreciation are addressed in the discussion of the empirical evidence, below.

Table 2-1 provides perspectives on the history of payments restrictions among APEC members, which may have some bearing on the range of exchange rate experiences. In particular, the table lists the time periods since the end of 1973 during which the members of APEC maintained restrictions on payments for current and capital transactions. For example, China restricted payments for both current and capital transactions during the entire period from the end of 1973 through the end of 1993. Thus, nominal exchange rates for the yuan during that period did not reflect market forces to the same extent as nominal exchange rates for the yen, which may partly explain why China is a significant outlier in Figure 2-1. Accordingly, theories about the behavior of real exchange rates in Japan (or other countries without extensive payments restrictions) cannot be applied to, or rejected by, the behavior of real exchange rates in China (or other countries with extensive payments restrictions). More generally, in analyzing the behavior of real exchange rates for APEC economies during the period since the end of 1973, it may be important to bear in mind that six APEC members (Chile, China, Korea, Mexico, Philippines, and Taiwan Province of China) maintained restrictions on payments for both current and capital transactions during substantial parts of the period.

Table 2-1.

Payments Restrictions Maintained by APEC Members, 1974–93

article image
Source: Summary tables in the IMF’s Annual Report on Exchange Arrangements and Exchange Restrictions (Washington, various issues).

Refers to period beginning on October 9, 1975, when Papua New Guinea became a member of the IMF.

Not included in summary tables for the period since 1979.

Theories of the Long-Run Behavior of Exchange Rates

Theories of the relationship between nominal exchange rates and ratios of national price or cost levels have been traced back to the sixteenth century and have been linked to the development of the quantity theory of money.10 Spanish economists were particularly influential in the development of these theories, which presumably were catalyzed by the large increases in money supplies and prices that Spain experienced in association with substantial inflows of gold and silver from newly discovered America. Today, the theory that nominal exchange rates move in parallel with ratios of national price or cost levels is commonly referred to as the hypothesis of “purchasing power parity,” a term coined by Gustav Cassel (1918, 1922). Cassel’s prolific writings on the purchasing-power-parity (PPP) hypothesis were stimulated by the general inflation that accompanied World War I.

The PPP hypothesis has received considerable attention in the economic literature during the past two decades.11 PPP has been thoroughly discredited as a hypothesis about the behavior of nominal exchange rates in the short run. However, extensive econometric work has verified that real exchange rates for industrial countries tend to revert toward their historical averages over long periods of time, which suggests that PPP has some degree of validity in the long run.12

An important modification or refinement of the long-run PPP hypothesis has come from the observation that prices of nontradable goods and services, relative to prices of tradables, tend to be higher in high-income countries than in low-income countries. This observation emerged from attempts to make quantitative comparisons of living standards in different countries in a series of projects sponsored by the United Nations and other international organizations, and spearheaded to a large extent by economists from the University of Pennsylvania.13 These studies have established that the methodology of comparing international standards of living by converting national accounts data at market exchange rates into a common currency unit generally understates the living standards of low-income countries relative to those of high-income countries.14 Samuelson (1994) referred to this empirical regularity as “the Penn effect.”

Balassa (1964) and Samuelson (1964) attempted to explain the Penn effect, along with the associated tendency for exchange rates to deviate systematically from PPP over the long run,15 by conjecturing that the tendency for the relative price of nontradables to be higher in high-income countries reflected a tendency for productivity in the tradable goods sector to rise relative to productivity in the nontradables sector as real incomes expanded. Given competitive pressures within each country for workers with similar skills to receive similar wages in the two sectors, relatively rapid productivity growth in the tradables sector would tend, other things being equal, to push up the relative cost of production in the nontradables sector and, hence, the relative price of nontradables. Under conditions in which the relative price of tradable goods across countries remained constant, such an increase in the relative price of nontradables would in turn give rise (as clarified in the equations below) to an appreciation of the real exchange rate.

Although changes in the relative price of nontradables have received emphasis in models of deviations from long-run PPP, other factors can also affect the behavior of real exchange rates over the long run. To provide a simple framework for analysis, it is useful to write the aggregate price levels for two countries (P and P*) as arithmetic weighted averages of the prices of nontradables (PN and P*N) and the prices of tradables (PT and P*T):

P=nPN+(1n)PT(2-2)
P*=n*PN*+(1n*)PT*(2-3)

The weights (n and n*) can be regarded as the shares of nontradables in production when P is interpreted as a GDP deflator, or as shares of nontradables in consumption when P represents a consumer price index.

It is useful to denote the relative common-currency price of tradables in the two countries as

b=SPT/PT*(24)

where, under the convention that the variables marked with an asterisk (*) refer to the foreign country, S is the nominal exchange rate measured in foreign currency per unit of domestic currency. (Hence, an increase in S corresponds to a nominal appreciation of the domestic currency.) It is often hypothesized that the relative price of tradables across countries is time invariant and equal to unity, reflecting the assumption that tradable goods have identical common-currency prices in all countries.16 Here, however, it is instructive to consider the more general case. It is convenient to define

F=P/PT=(1-n)+n(PN/PT)(2-5)
F*=P*/PT*=(1-n*)+n*(PN*/PT*)(2-6)

And it is then straightforward to show that

Q=SP/P*=bF/F*=b[(1-n)+n(PN/PT)]/[(1-n*)+n*(PN*/PT)*)](2-7)

Condition (2-7) describes a general relationship between the real exchange rate (Q), the relative prices of nontradables within each of the two countries (PN/PT and P*N/P*T), the weights of nontradables in the domestic and foreign price indices (n and n*), and the relative price of tradables across countries (b). The equation confirms that, other things being equal, a rise in the domestic relative price of nontradables (PN/PT) leads to a real appreciation. More generally, however, it indicates that changes in real exchange rates over the long run can logically be related to different combinations of changes in the relative price of tradables across countries, changes in both domestic and foreign price-index weights, and changes in both the domestic and foreign relative price of nontradables. In turn, changes in relative prices within countries, price-index weights, and perhaps the relative price of tradables across countries may be reflections of changes in deeper fundamentals, such as changes in productivity or shifts in the composition of aggregate demand.17

As is readily apparent, condition (2-7) provides a framework for describing the behavior of real exchange rates when only two types of goods are distinguished. In some contexts it may be preferable to distinguish among three types of goods—nontradables and two classes of tradables—in order to consider shifts in the terms of trade between exportables and importables, or to analyze supply shocks that are beneficial to an emerging tradable goods sector but harmful to an older tradables sector.

Empirical Methodologies and Findings from Studies of OECD Economies

A number of economists have conducted empirical tests of structural models of deviations from PPP.18 A major stimulus to the literature came from Balassa (1964, 1973 1973), who regressed a measure of the deviation from PPP on GNP per capita and found a significant positive correlation using data for 1960 from a cross section of 12 industrial countries. As mentioned earlier, Balassa hypothesized that the correlation reflected the effects of output growth on relative productivity levels and the relative price of nontradables.

Subsequent tests have looked for more direct evidence that real exchange rates are correlated with productivity levels or other determinants of the relative price of nontradables, such as the composition of aggregate demand. For the most part, these tests have looked at real exchange rates based on GDP deflators or consumer price indices and have focused on time-series data for member countries of the Organization for Economic Cooperation and Development (OECD).19 Although some of these tests have sought to explain changes over time in such measures of the real exchange rate, others have sought to explain changes over time in the relative price of nontradables (PN/PT).20

The Balassa-Samuelson hypothesis focuses on the effects of supply-side factors (that is, productivity levels) on relative prices and real exchange rates. It ignores the effects of demand-side factors on relative prices, which may be reasonable for purposes of long-run analysis.21 In particular, demand-side factors will not have long-run effects on relative prices if factors of production are perfectly mobile between the tradables and nontradables sectors over the long run and if there are constant returns to scale in each sector. In the short-and-medium run, however, demand-side factors, such as the composition of government spending, may have significant effects on the relative price of nontradables. Accordingly, several studies based on annual time-series data have investigated the relevance of demand-side factors.22

The empirical evidence from studies of OECD economies suggests two general conclusions about the Balassa-Samuelson hypothesis. One finding is that changes in relative productivity levels for the tradables and nontradables sectors are fairly significant in explaining changes in the relative price of nontradables (PN/PT) (De Gregorio, Giovannini, and Wolf, 1994; Asea and Mendoza, 1994). A second finding is that, for the specification forms tested by existing studies, neither productivity differentials nor relative prices have been highly significant in explaining the behavior of the broader concept of the real exchange rate (Q) (Froot and Rogoff, 1991; Asea and Mendoza, 1994). As will be shown below, these findings can be reconciled with each other by recognizing that the latter studies have been based on specifications embodying an assumption that is not supported by the data—that the relative price of tradable goods across countries is time invariant.

Empirical Evidence for the APEC Region

As a first step in analyzing real exchange rates for the APEC region, it is interesting to focus on the extent to which the Penn effect is evident in cross-section data. For this purpose, Figure 2-4 shows data on real exchange rates constructed from Penn price relatives, paired with per capita GDP levels, for the cross section of APEC members during 1973, 1983, and 1992. As defined and estimated in international comparisons of living standards, the real exchange rates shown in the chart are constructed solely from price and quantity data, without using data on market exchange rates.23 Specifically, a country’s real exchange rate based on Penn price relatives represents the ratio of the cost of its bill of goods at national prices to the cost of the same bill of goods at international prices.24 The per capita GDP measures (throughout this section) represent national accounts data converted at 1990 exchange rates into constant (1990) U.S. dollars.25

Figure 2-4.
Figure 2-4.

Real Exchange Rates and Levels of Per Capita GDP1

1 Real exchange rates are based on Penn price relatives.2 1975 for Chile.3 1990 for Taiwan Province of China; 1991 for Korea.

For each of the three selected years, Figure 2-4 shows a reasonably high positive correlation between real exchange rates and per capita GDP levels. To the extent that it was valid to assume that tradable goods have fairly similar common-currency prices across countries, the correlations shown in the chart would imply that the ratio of nontradables prices to tradables prices was positively correlated with per capita GDP levels.

By contrast with the high cross-section correlations between the levels of real exchange rates and per capita GDPs, the correlation between percentage changes in real exchange rates and growth rates of GDP per capita have been relatively weak. From Table 2-2, for example, it may be noted that among the eight APEC economies that grew faster than Japan during the 1973–92 period, all but Korea and Taiwan Province of China experienced little if any appreciation of their real exchange rates.26

Table 2-2.

Real Exchange Rates and Per Capita GDP1

article image

Real exchange rates are based on Penn price relatives.

1990 for Taiwan Province of China; 1991 for Korea.

Most empirical attempts to explain the long-run behavior of real exchange rates have looked at conventional measures (for example, real exchange rates based on GDP deflators or consumer price indices) rather than real exchange rates based on the Penn price relatives. Moreover, such studies have focused primarily on time-series data.27 As noted earlier, these studies have been relatively successful in explaining the behavior of the ratio of nontradables prices to tradables prices, but not in explaining the behavior of the broader concept of the real exchange rate (Q). In this connection, Figure 2-5 provides scatter plots for APEC members of average annual percentage changes in real exchange rates over the period 1973–92 versus average annual growth rates of GDP per capita. The two different scatter plots reflect two different measures of real exchange rates. The left panel uses bilateral real exchange rates vis-à-vis the United States, based on GDP deflators from national sources. The right panel reflects real exchange rates constructed from the Penn price relatives by using GDP weights. Both plots show weak correlations between the two variables, suggesting that per capita output growth by itself provides a poor explanation of long-run trends in real exchange rates. As in Table 2-2, moreover, Figure 2-5 indicates that among the eight APEC members that exhibited more rapid output growth than Japan during the period 1973–92, only Korea and Taiwan Province of China experienced significant real appreciations.28 In contrast, of the eight APEC members with average growth rates below that of Japan, none experienced a significant real appreciation.

Figure 2-5.
Figure 2-5.

Changes in Real Exchange Rates and Per Capital GDP

(Average annual percent change, 1973–92)1

11973 to 1990 for Taiwan Province of China; 1973 to 1991 for Korea; 1975 to 1992 for Chile.2Vis-a-vis the United States; based on GDP deflators.3Based on Penn price relatives.

Why is it that most of the rapidly growing APEC members did not experience real appreciations? One place to look for an explanation is in the behavior of productivity growth in the tradables sector relative to productivity growth in nontradables—or, if that is precluded by data limitations, in the behavior of the price of nontradables relative to the price of tradables. Recall that the Balassa-Samuelson conjecture maintains that the ratio of tradables productivity to nontradables productivity is positively correlated with output per capita, and that increases in the productivity ratio push up the price of nontradables relative to the price of tradables, which is positively correlated with the real exchange rate under the assumption that the relative price of tradable goods across countries remains constant.

Young (1995) has recently published relevant estimates of productivity growth for three of the fastest-growing APEC members: Korea, Singapore, and Taiwan Province of China. Young’s estimates suggest that, in apparent contradiction of the Balassa-Samuelson hypothesis, for Taiwan Province of China total factor productivity grew significantly faster in services than in manufacturing during 1966–90; while for Singapore, total factor productivity grew significantly faster economy-wide (hence, presumably, in the nontradables sector) than in manufacturing.29

For most members of APEC, available data are not adequate for constructing measures of labor productivity or total factor productivity for the tradables and nontradables sectors. Annual data exist, however, on value added broken down into several different categories of output, and on the implicit price deflators associated with each category of output. It is thus possible to test whether productivity growth in tradables has been relatively rapid by examining whether the change in the price of nontradables (services) relative to the price of tradables (manufactures) has been positively correlated with the growth rate of per capita output.30 Although it is certainly incorrect to assume that manufacturing output is entirely tradable and services entirely nontradable, the ratio of services prices to manufacturing prices may be a reasonable proxy for the relative price of nontradables.31

Figure 2-6 provides a scatter plot of average annual percentage changes in the relative price of nontradables and average rates of growth of output per capita during 1973–92. Korea and China provide examples in which rapid per capita output growth has been associated with a substantial increase in the relative price of nontradables;32 in contrast, all of the eight APEC members with average growth rates below that of Japan have experienced declines (or very small increases) in the relative prices of their nontradables. Yet it also appears from Figure 2-6 that the Balassa-Samuelson hypothesis is not valid in all cases. In particular, Singapore, Malaysia, Thailand, and Indonesia provide examples of rapidly growing economies that did not experience increases in the relative prices of nontradables.33

Figure 2-6.
Figure 2-6.

Changes in Relative Prices of Nontradables and Per Capita GDP1

(Average annual percent change in normalized levels, 1973–92)

1The vertical axis measures the average annual percent change in the ratio of the price of nontradable goods to the price of tradable goods divided by the corresponding ratio for the United States. The horizontal axis measures the average annual percent change in the ratio of domestic per capita GDP to per capita GDP for the United States. Annual averages are based on the data available within the 1973-92 period.

A caveat is that such evidence may be distorted to some extent by measurement error in the relative price data, including the distortions that may be introduced in using the ratio of services prices to manufacturing prices as a proxy for the ratio of nontradables prices to tradables prices.34 In addition, the analysis abstracts from the possibility that the relative price of nontradables may have been affected by changes in government policies—including changes in the composition of demand and changes in the stances of monetary and fiscal policies, as well as external and internal liberalization measures.

Regardless of the role of output (or relative productivity) growth in driving the relative price of nontradables, the real exchange rate is linked to the relative price of nontradables through the system of identities developed above (recall condition (2-7)). In particular, both the long-run trend and the variability over time in the real exchange rate (Q) are reflections of trends or variability in other variables—the relative (common-currency) price of domestic and foreign tradable goods (b), the domestic and foreign relative prices of nontradable goods (PN/PT and PN*/PT*), and the shares of nontradables in the domestic and foreign price indices (n and n*).

Figure 2-7 shows the relative contributions of these “proximate determinants” in explaining the behavior of real exchange rates over time for individual APEC members. For this purpose the real exchange rate measures have been constructed using condition (2-7), which ensures that the relative contributions of the proximate determinants add up to the total.35Table 2-3 provides additional information on the magnitudes of the changes over twenty years (1973 through 1992) in the proximate determinants. The most striking feature of the chart is the behavior of the relative price of tradables across countries. In general, the relative price of tradables across countries has varied widely over time, “explaining” most of the year-to-year variation in the real exchange rate. Moreover, in half the cases the trend change in the real exchange rate over the twenty-year period can be largely explained by the relative price of tradables across countries. As is evident, however, there are a number of cases—Canada, Korea, Malaysia, Mexico (since 1982), Papua New Guinea, Philippines (since 1982), Singapore, and Thailand (since 1985)—in which changes in real exchange rates have been related significantly to other factors, primarily to changes in the price of nontradables relative to the price of tradables.

Figure 2-7.
Figure 2-7.
Figure 2-7.

Real Exchange Rates and Contributions of Proximate Determinants1

1 The heavy line shows the time path of the real exchange rate relative to the United States (Q) as constructed using condition (2-7) in the text. The line labeled “due to b” (alternatively: PN/PTand P*N/P*T; or n and n*) shows the path that Q would have taken if only b (alternatively: PN/PT and P*N/P*T; or n and n*) varied while all other proximate determinants remained constant.
Table 2-3.

Real Exchange Rates and Proximate Determinants, 1973 and 19921

article image

Real exchange rates are those shown in Figure 2-7. The proximate determinants n and PN/PT are measured as ratios to U.S. levels.

1971 for Canada; 1975 for Chile; 1977 for New Zealand; and 1980 for Papua New Guinea.

1990 for Canada and New Zealand; 1988 for Chile; and 1983 for Malaysia.

Figure 2-8 shows the behavior of real exchange rates based on GDP deflators and two alternative measures of the relative price of tradables across countries—the relative price of manufactures across countries (essentially equivalent to b in Figure 2-7) and the relative price of exports (identical to the real exchange rate based on export prices in Figure 2-2). The latter measure of the relative price of tradables across countries has also exhibited wide year-to-year variability and, in some cases, substantial cumulative change over time. In a number of cases the two alternative measures of the relative price of tradable goods across countries do not move closely in parallel. In most of these cases, the divergence between the relative export price and the relative price of manufactures across countries may reflect a composition of exports that includes large proportions of agricultural products and other nonmanufactured goods. In some cases, however, it may also have been associated with significant divergence between the export prices and domestic prices of manufactured goods, reflecting either a widening or a narrowing trend in profit margins on sales of domestic manufactures accompanied by the opposite trend in profit margins on export sales.

Figure 2-8.
Figure 2-8.
Figure 2-8.

Real Exchange Rates and Relative Prices Across Countries1

(1973 = 100)

1Real exchange rates are based on GDP deflators. Relative prices across countries are constructed as products of the nominal exchange rate and the ratio of the domestic export price index (or price deflator for manufacturing output) to the U.S. export price index (or price deflator for manufacturing output).

The year-to-year variability in the relative prices of tradables across countries and the high correlation between these changes and the year-to-year changes in real exchange rates undoubtedly reflect fluctuations in nominal exchange rates (recall the definitions in conditions (2-1) and (2-4)). Moreover, the year-to-year changes in relative prices across countries, as exhibited by aggregate price indices for tradable goods, show deviations from the law of one price for individual tradable goods. One theory of why there can be deviations from the law of one price—in the sense that the same good or similar goods sell for different prices in different markets—is that imperfect competition in the short run enables oligopolistic suppliers to “price to market” and charge different prices for similar products in different countries.36 A second reason that changes in exchange rates may not be immediately reflected in the relative prices of a good in domestic and foreign markets is that firms may incur various costs when they adjust their prices. A number of economists have attempted to assess the adjustment-cost and pricing-to-market theories of movements in the relative prices of tradable goods across countries.37 The two possibilities are not mutually exclusive, and this has made it difficult to determine their relative importance.

In contrast to the year-to-year changes, the substantial cumulative changes over the twenty-year period in the relative prices of tradable goods across countries may be reflections of three different phenomena. One likely part of the explanation is that the composition of tradable goods across countries tends to change over time. As economies develop, their production activities tend to shift toward more sophisticated technologies and higher-quality products.38 Thus, to the extent that price indices for tradable goods are constructed from unit-value data or are not fully adjusted for quality changes, the influence of economic development on the composition of tradable goods is likely to be reflected in gradual trends in the relative prices of tradables across countries.39 Another likely part of the explanation is that the observed trends in the relative prices of tradables across countries may reflect trends in the terms of trade among different categories of tradable goods interacting with cross-country differences in price-index weights.

As a third possibility, the explanation may be partly related to changes over time in the costs of “goods arbitrage,” reflecting the liberalization of trade restrictions, reductions in transportation costs, or changes in other components of the costs of market penetration. As noted earlier, the liberalization of trade and payments restrictions was an important phenomenon for several APEC members during the 1973–92 period (see Table 2-1).

Concluding Perspectives

Although economists remain mystified by movements in exchange rates from day to day, month to month, and even year to year, there is strong evidence that nominal exchange rates move broadly in parallel with ratios of national price levels over long periods of time. The data also indicate, however, that the parallel is not exact. For some countries—for example, Japan—the trend appreciation in the nominal exchange rate has differed significantly from the trend in the ratio of national price levels.

Whether rapid output growth tends to be associated with real exchange rate appreciation has been a central issue of investigation in conceptual and empirical research on the long-run behavior of exchange rates. The issue has received stimulus from the Balassa-Samuelson conjecture, which states that relatively rapid output growth tends to be associated with more rapid productivity growth in the tradables sector than in the nontradables sector, putting upward pressure on the price of nontradables relative to the price of tradables. To the extent that little change occurs in the relative prices of tradable goods across countries, a rise in the domestic relative price of nontradables is in turn associated with real exchange rate appreciation. This story appears to be widely accepted as a large part of the explanation of the Japanese experience.

Although several fast-growing APEC economies besides Japan have also experienced real exchange rate appreciations over the past two decades, a number of others have not.40 To help to understand these different experiences, this section developed a definitional identity in which the real exchange rate (as conventionally based on GDP deflators or consumer price indices) is expressed as the product of the ratio of the relative price of domestic and foreign tradable goods, and an expression that increases monotonically with increases in the ratio of the domestic price of nontradables to the domestic price of tradables, other things being equal. The identity suggests that the fact that some rapidly growing APEC economies did not experience real exchange rate appreciations during recent decades may have been related either to changes over time in relative prices of tradables across countries or to decreases in relative domestic prices of nontradables.

This section has examined the empirical evidence on both of these possibilities. To the extent that the data are reliable (and that services and manufactured goods can reasonably be regarded as nontradables and tradables, respectively), they show that Indonesia (since 1985), Malaysia, Singapore, and Thailand all provide counterexamples to the Balassa-Samuelson hypothesis (see Figures 2-6 and 2-7). In these cases, rapid output growth did not lead to increases in relative domestic prices of nontradables during the 1973–92 period and, therefore, did not lead to real exchange rate appreciation through the Balassa-Samuelson channel.41 Thus, the data suggest that, in comparisons of these economies with the United States, the “Penn effect”—or the degree to which simple exchange rate conversions distort international comparisons of living standards—would have tended to widen over the two decades had other things been equal.42 Consistently, despite strong evidence of the Penn effect in cross-section data, timeseries data for the 1973–92 period indicate that these fast-growing economies experienced little if any real exchange rate appreciations during the two decades (see Figures 2-4 and 2-5).

The data also reveal that the relative prices of tradable goods across countries have varied considerably from year to year and, for most APEC members, have exhibited substantial trends over time. Thus, even for rapidly growing economies in which productivity growth is higher in the tradable goods sector than in the nontradables sector, the associated pressures on real exchange rates may conceivably be offset by changes in the relative prices of tradable goods across countries. Persistent changes in the latter relative prices may reflect several different phenomena, including changes over time in the composition of tradable goods across countries, changes in the terms of trade among different categories of tradable goods interacting with cross-country differences in price-index weights, or changes over time in the costs of “goods arbitrage” as a result of changes in trade restrictions, transportation costs, or other market penetration costs.

In sum, this section has verified that there is strong evidence of the Penn effect in cross-section data, while also challenging the presumption that fast-growing economies can generally be expected to experience real exchange rate appreciations. In challenging the latter presumption, it was noted, first, that time-series data for some rapidly growing APEC economies appear to provide counterexamples to the Balassa-Samuelson hypothesis and, second, that for aggregate measures of tradable goods prices, there is widespread evidence of trends over several decades in the relative prices of tradable goods across countries. An interesting issue that the analysis in this section has not addressed is the extent to which the rejections of the Balassa-Samuelson hypothesis can be attributed to the influence of government policies.

References

  • Asea, Patrick K., and Enrique G. Mendoza 1994The Balassa-Samuelson Model: A General-Equilibrium Appraisal,Review of International Economics, Vol. 2 (October), pp. 24467.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Balassa, Bela, 1964The Purchasing-Power-Parity Doctrine: A Reappraisal,Journal of Political Economy, Vol. 72 (December), pp. 58496.

  • Balassa, Bela, 1973Just How Misleading Are Official Exchange Rate Conversions? A Comment,Economic Journal, Vol. 83 (December), pp. 125867.

  • Bergstrand, Jeffrey H, 1991Structural Determinants of Real Exchange Rates and National Price Levels: Some Empirical Evidence,American Economic Review, Vol. 81 (March) pp. 32534.

    • Search Google Scholar
    • Export Citation
  • Breuer Boucher, Janice, 1994An Assessment of the Evidence on Purchasing Power Parity,in Estimating Equilibrium Exchange Rates, ed. by John Williamson (Washington: Institute for International Economics), pp. 24577.

    • Search Google Scholar
    • Export Citation
  • Cassel, Gustav, 1918Abnormal Deviations in International Exchanges,Economic Journal, Vol. 28 (December), pp. 41315.

  • Cassel, Gustav, 1922 Money and Foreign Exchange After 1914 (New York: Constable).

  • Clark, Peter B., Leonardo Bartolini, Tamim Bayoumi, Steven Symansky 1994Exchange Rates and Economic Fundamentals: A Framework for Analysis,Occasional Paper 115 (Washington: International Monetary Fund).

    • Search Google Scholar
    • Export Citation
  • De Gregorio, José Alberto, Giovannini Thomas H. Krueger, 1994The Behavior of Nontradable-Goods Prices in Europe: Evidence and Interpretation,Review of International Economics, Vol. 2 (October), pp. 284305.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • De Gregorio, José Alberto Giovannini Holger Wolf 1994International Evidence on Tradables Nontradables Inflation,European Economic Review, Vol. 38 (June), pp. 122544.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Dornbusch, Rudiger, 1992Purchasing Power Parity,in The New Palgrave Dictionary of Economics, Vol. 3, ed. by Peter Newman, Murray Milgate, John Eatwell (New York: Stockton) pp. 23644.

    • Search Google Scholar
    • Export Citation
  • Edison, Hali J., Jan Tore and Klovland, 1987A Quantitative Reassessment of the Purchasing-Power-Parity Hypothesis: Evidence from Norway and the United Kingdom,Journal of Applied Econometrics, Vol.2 (October), pp. 30933.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Einzig, Paul, 1962 The History of Foreign Exchange (London: Macmillan).

  • Faruqee, Hamid, 1995Pricing to Market and the Real Exchange Rate,Staff Papers, International Monetary Fund, Vol. 42 (December), pp. 85581.

  • Frankel, Jeffrey A., Andrew K. Rose, forthcomingA Survey of Empirical Research on Nominal Exchange Rates,in Handbook of International Economics, Vol. 3, ed. by Gene Grossman and Kenneth Rogoff (Amsterdam: North-Holland).

    • Search Google Scholar
    • Export Citation
  • Froot, Kenneth A., and Paul D. Klemperer 1989Exchange-Rate Pass Through When Market Share Matters,American Economic Review, Vol. 79 (September), pp. 63754.

    • Search Google Scholar
    • Export Citation
  • Froot, Kenneth A., Kenneth Rogoff, 1991The EMS, the EMU, and the Transition to a Common Currency,in 1991 NBER Macroeconomics Annual, ed. by Olivier Blanchard Stanley Fischer (Cambridge, Mass.: MIT Press), pp. 269317.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Froot, Kenneth A., Kenneth Rogoff, forthcoming, “Perspectives on PPP and Long-Run Real Exchange Rates,in Handbook of International Economics, Vol. 3, ed. by Gene Grossman Kenneth Rogoff (Amsterdam: North-Holland).

    • Search Google Scholar
    • Export Citation
  • Ghosh, Atish R., and Holger C. Wolf, 1994Pricing in International Markets: Lessons from the Economist,NBER Research Paper 4806 (Cambridge, Mass.: National Bureau of Economic Research, July).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Gilbert, Milton, and Irving B. Kravis, 1954 An International Comparison of National Products and the Purchasing Power of Currencies: A Study of the United States, the United Kingdom, France, Germany, and Italy (Paris: Organization for European Economic Cooperation).

    • Search Google Scholar
    • Export Citation
  • Gilbert, Milton, and associates, 1958 Comparative National Products and Price Levels: A Study of Western Europe and the United States (Paris: Organization for European Economic Cooperation).

    • Search Google Scholar
    • Export Citation
  • Grice-Hutchinson, Marjorie, 1952 The School of Salamanca (Oxford: Clarendon Press).

  • Harrod, Roy F., 1939 International Economics, 5th ed. (1973) (London: Nisbet).

  • Hsieh, David A., 1982The Determination for the Real Exchange Rate: The Productivity Approach,Journal of International Economics, Vol. 12 (May), pp. 35562.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Isard, Peter, 1983 review of Kravis, Heston, Summers, (1982) Journal of International Economics, Vol. 15 (August), pp. 17781.

  • Isard, Peter, 1995 Exchange Rate Economics (Cambridge and New York: Cambridge University Press).

  • Knetter, Michael, 1989Price Discrimination by U.S. and German Exporters,American Economic Review, Vol. 79 (March), pp. 198210.

    • Search Google Scholar
    • Export Citation
  • Knetter, Michael, 1993International Comparisons of Pricing to Market Behavior,American Economic Review, Vol. 83 (June), pp. 47386.

    • Search Google Scholar
    • Export Citation
  • Kravis, Irving B., Alan Heston, Robert Summers, 1982 World Product and Income: International Comparisons of Real Gross Product (Baltimore: Johns Hopkins University Press).

    • Search Google Scholar
    • Export Citation
  • McGuirk, Anne K., 1987Measuring Price Competitiveness for Industrial Country Trade in Manufactures,Working Paper 87/34 (Washington: International Monetary Fund, April).

    • Search Google Scholar
    • Export Citation
  • Marsh, Ian W., Stephen P. Tokarick, 1994Competitiveness Indicators: A Theoretical and Empirical Assessment,IMF Working Paper 94/29 (Washington: International Monetary Fund, March).

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Marston, Richard C., 1987Real Exchange Rates and Productivity Growth in the United States and Japan,in Real-Financial Linkages Among Open Economies, ed. by Sven Arndt J. David Richardson (Cambridge, Mass.: MIT Press).

    • Search Google Scholar
    • Export Citation
  • Meese, Richard A., Kenneth Rogoff 1983aEmpirical Exchange Rate Models of the Seventies: Do They Fit Out of Sample?Journal of International Economics, Vol. 14 (February), pp. 3 24.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Meese, Richard A., Kenneth Rogoff 1983bThe Out-of-Sample Failure of Empirical Exchange Rate Models: Sampling Error or Misspecification?in Exchange Rates and International Macroeconomics, ed. by Jacob A. Frenkel (Chicago: University of Chicago Press), pp. 67112.

    • Search Google Scholar
    • Export Citation
  • Officer, Lawrence H., 1982 Purchasing Power Parity and Exchange Rates: Theory, Evidence and Relevance (Greenwich, Conn.: JAI).

  • Rangan, Subramanian, Robert Z. Lawrence, 1993The Responses of U.S. Firms to Exchange Rate Fluctuations: Piercing the Corporate Veil,Brookings Papers on Economic Activity: 2, pp. 34169.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Ricardo, David, 1951 On the Principles of Political Economy and Taxation, in The Works and Correspondence of David Ricardo, Vol. 1, ed. by Sraffa Piero (Cambridge and New York: Cambridge University Press).

    • Search Google Scholar
    • Export Citation
  • Rogoff, Kenneth, 1992Traded Goods Consumption Smoothing and the Random Walk Behavior of the Real Exchange Rate,Bank of Japan Monetary and Economic Studies, Vol. 10 (November), pp. 129.

    • Search Google Scholar
    • Export Citation
  • Rogoff, Kenneth, 1996The Purchasing Power Parity Puzzle,Journal of Economic Literature, Vol. 34 (June), pp. 64768.

  • Samuelson, Paul A., 1964Theoretical Notes on Trade Problems,Review of Economics and Statistics, Vol. 46 (May), pp. 14554.

  • Samuelson, Paul A., 1994Facets of Balassa-Samuelson Thirty Years Later,Review of International Economics, Vol. 2 (October), pp. 20126.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Summers, Robert, Alan Heston, 1991The Penn World Table (Mark 5): An Expanded Set of International Comparisons, 1950–1988.Quarterly Journal of Economics, Vol. 106 (May), pp. 32768.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • Taylor, Mark P., 1995aThe Economics of Exchange Rates,Journal of Economic Literature, Vol. 33 (March), pp. 1347.

  • Taylor, Mark P., 1995bExchange Rate Behavior Under Alternative Exchange Rate Regimes,in Understanding Interdependence: The Macroeconomics of the Open Economy, ed. by Peter B Kenen (Princeton, N.J.: Princeton University Press).

    • Search Google Scholar
    • Export Citation
  • Turner, Philip, and Jozef Van’t dack, 1993Measuring International Price and Cost Competitiveness,Economic Papers, No. 39 (Basel: Bank for International Settlements).

    • Search Google Scholar
    • Export Citation
  • Young, Alwyn, 1995The Tyranny of Numbers: Con-fronting the Statistical Realities of the East Asian Growth Experience,Quarterly Journal of Economics, Vol. 110 (August), pp. 64180.

    • Crossref
    • Search Google Scholar
    • Export Citation
1

See the recent surveys by Frankel and Rose (forthcoming), 1Isard (995), and Taylor (1995a, 1995b).

2

Figure 2-1 includes a 45-degree line for comparison.

3

See Boucher Breuer (1994), Froot and Rogoff (forthcoming), and Rogoff (1996) for reviews of the relevant literature.

4

The extent to which the relevant theories can be tested empirically depends on whether the available data on employment and capital stocks are adequate for constructing quantitative measures of labor productivity or total factor productivity in the tradables and nontradables sectors. For most APEC members such data are inadequate. Nevertheless, other data available for APEC members are sufficient to illustrate a number of relevant points about the long-run behavior of exchange rates.

5

See Clark and others (1994), Marsh and Tokarick (1994), and Turner and Van’t dack (1993) for recent theoretical and empirical assessments of how well the different measures of real exchange rates capture international price and cost competitiveness.

6

There is no chart for the United States because all variables are measured relative to the U.S. dollar, or for Brunei Darussalam, whose currency notes are exchangeable at par with those of Singapore.

7

The time series on these three measures, based on data from national sources, are from the IMF’s International Financial Statistics and World Economic Outlook databases.

8

Data on the Penn measures, in which price relatives (domestic versus United States) for individual products are aggregated with weights that reflect shares in domestic real GDP, are from the Penn World Table; see Summers and Heston (1991).

9

Some of the divergence between the different real exchange rate measures might be attributable to problems of data quality. Accordingly, as a crude way of identifying data series that might be suspect, we looked for contrasts among four real exchange rate measures: the real exchange rate based on GDP deflators, the Penn measure constructed with GDP weights, the real exchange rate based on consumer price indices, and an alternative Penn measure constructed with weights that reflect consumption shares. Except in the cases of Indonesia and Korea, the four measures appear to be reasonably consistent. For Indonesia, the real exchange rate based on national-source GDP deflators is a clear outlier, while for Korea the measure based on national-source consumer price indices is the outlier.

11

See Dornbusch (1992), Isard (1995), and Rogoff (1996) for summary discussions.

12

See Boucher Breuer (1994) and Froot and Rogoff (forthcoming) for recent surveys of the extensive econometric literature.

13

See, for example, Gilbert and Kravis (1954); Kravis, Heston, and Summers (1982); and Summers and Heston (1991).

14

The nature of these studies is described in more detail in the discussion of the empirical evidence, below.

15

Harrod (1939) provided an earlier discussion of some of the key arguments made by Balassa and Samuelson, the seeds of which have been traced back to Ricardo in 1821 (see Ricardo, 1951).

16

If PT and P*T are index numbers, the assumption of identical common-currency prices would imply a time-invariant value of b, but not necessarily that b = 1.

17

The direction of causation in condition (2-7) is not simply from the right-hand-side variables (or their deeper determinants) to Q. For example, changes in nominal exchange rates, which affect both Q and b directly, may induce adjustments in tradable goods prices and, hence, in the relative price of nontradables. Thus, the year-to-year behavior of the relative price of nontradables may reflect nominal exchange rate variability, as well as shifts in the composition of demand or longer-run trends in productivity and other supply-side factors. To this extent, economies that have maintained relatively tight capital controls might be expected to have experienced smaller year-to-year fluctuations in the relative prices of their nontradables, other things being equal.

18

Froot and Rogoff (forthcoming) provide a recent review of the literature.

19

The first applications of time-series data to test the Balassa-Samuelson hypothesis included Hsieh (1982), xMarston (1987), and Edison and Klovland (1987). These tests employed measures of relative labor productivity levels in the tradables and nontradables sectors, constructed from data on value added and employment. The classification of sectors was fairly arbitrary, with differences among the studies. Hsieh (1982) equated the tradable goods sector with manufacturing and treated GDP other than manufacturing as nontradables. Marston (1987) defined agriculture and manufacturing as tradables and defined construction and all other services except electricity-gas-water as nontradables; mining and electricity-gas-water were excluded from his analysis because of a desire to abstract from products with a high energy content. Edison and Klovland (1987) defined agriculture, mining, manufacturing, construction, and electricity-gas-water as “commodity sectors” (tradables) and all services except construction and electricity-gas-water as “service sectors” (nontradables). In some subsequent studies, measures of labor productivity have been replaced with estimates of total factor productivity; see De Gregorio, Giovannini, and Wolf (1994) and Asea and Mendoza (1994). As another innovation, in some studies the methodology for classifying individual production sectors as tradables or nontradables has been based on whether the share of exports in sectoral value added exceeds some predecided threshold level; this approach was introduced by De Gregorio, Giovannini, and Wolf (1994).

20

The latter ratio is sometimes also referred to as a real ex-change rate in the context of models in which tradables have identical (common-currency) prices across countries.

21

This does not imply that it is reasonable to ignore the longrun effects of demand-side factors on the other terms that enter condition (2-7), such as the shares of nontradables in production or consumption. This issue is discussed in Froot and Rogoff (1991).

22

For example, Froot and Rogoff (1991); Rogoff (1992); De Gregorio, Giovannini, and H. Krueger (1994); and De Gregorio, Giovannini, and Wolf (1994). See also Bergstrand (1991), who found evidence of demand-side effects in cross-section data.

23

The data were obtained from the Penn World Table; see Summers and Heston (1991).

24

See Summers and Heston (1991); Kravis, Heston, and Summers (1982); or Isard (1983). The United Nations International Comparison Program (ICP) has collected price and expenditure data during several benchmark years for selected countries. Each country in the sample provides data for several hundred items, which are grouped into about 150 “detailed” product categories. The ICP has taken considerable care to define products in ways that are reasonably comparable across countries. For each country in the sample, the ICP records the domestic price of each product category relative to the corresponding price in the United States (the numeraire country), and also divides this relative price into the level of domestic expenditure on the product category to calculate a quantity measure valued at the U.S. price. The international price of the product is then calculated as a quantity-weighted average of national prices. Each country not included in the sample is assumed to have relative price and quantity structures corresponding to the average structures for some particular subset of the countries included in the sample, with the correspondences based on considerations of region and per capita GDP. For years between the sampling benchmarks, extrapolation procedures are used to obtain estimated price and quantity data.

25

Figure 2-4 would not be altered much if Penn (PPP-adjusted) measures of real GDP were plotted instead of national accounts measures. Only China and Papua New Guinea show significant differences between the two measures of real GDP.

26

Similarly, of the nine APEC economies that grew faster than Japan during the 1983-92 period, all but Hong Kong, Korea, and Taiwan Province of China experienced little if any real appreciation.

27

Cross-section regressions of conventional real exchange rate measures on GDP per capita or other variables make little sense. In particular, since price indices can be normalized to unity in any year, such specifications would be equivalent to regressing the nominal exchange rate on GDP per capita or other variables.

28

These findings apply as well to a scatter plot based on multi-lateral real exchange rates.

29

Young’s research focused primarily on explaining the rapid output growth experienced by these three economies and Hong Kong. His findings attribute much of the output growth to the accumulation of factor inputs (physical and human capital, including increases in labor participation rates) rather than to rapid productivity growth.

30

The price data are taken from the national accounts files of the World Bank’s Economic and Social Database. These files contain annual data on nominal value added and real value added (that is, value added at base-year prices) for major sectors of the economy; however, they do not contain the employment data necessary to construct measures of labor productivity by sector. For most APEC members, the available data span the 1973-92 period, at least for the manufacturing and services sectors.

31

In this section, the agricultural sector is not included in either the tradables sector or the nontradables sector. This may bias the analysis when agriculture has a large share in total output and its price moves in a different way from prices for the other sectors.

32

It may be noted from Figure 2-5, however, that China experienced a substantial depreciation of its real exchange rates.

33

For Hong Kong and Taiwan Province of China, data on the relative prices of nontradables were not available.

34

Such distortions may be particularly significant for countries in which much of the services sector consists of business and financial services, which tend to be traded internationally.

35

The real exchange rates (Q) in Figure 2-7 are analogous to the measures based on GDP deflators in Figure 2-2 and elsewhere in this section; however, the measures in Figure 2-7 are constructed as if GDP consisted only of the output of the manufacturing and services sectors.

36

See Faruqee (1995) and references cited therein. This could explain the upward trend in relative tradable goods prices in Japan in Figure 2-7 compared with the relatively stable behavior of Japanese export prices in Figure 2-2.

37

Recent contributions include Froot and Klemperer (1989), Knetter (1989, 1993), Rangan and Lawrence (1993), and Ghosh and Wolf (1994). Froot and Rogoff (forthcoming) provide a brief review of the literature.

38

Although in this section tradable goods are restricted to manufactured products, the phenomenon of shifting commodity composition is more widespread. Many APEC members, for example, have undergone a significant shift from being primary commodity producers at the beginning of the sample period to being economies with sizable manufacturing sectors in more recent years.

39

Likewise, to the extent that price indices reflect unit-value data, a significant part of the year-to-year fluctuations in the relative price of tradables across countries may reflect year-to-year changes in the commodity composition of traded goods.

40

Regardless of the correlation between output growth and real appreciation over the long run, any rapidly growing economy may sometimes encounter circumstances in which policies of allowing greater upward flexibility in nominal and real exchange rates are conducive to achieving the macroeconomic objectives of internal and external balance over the medium run.

41

Both policy design and inherent comparative advantages may contribute to cases of fast-growing economies in which productivity advances are concentrated in the services sector.

42

Specifically, in the absence of any changes in the relative price of tradable goods across countries, the quantity of nontradables that could have been obtained in these fast-growing economies in exchange for a given amount of tradables would have increased over time relative to the quantity of nontradables that could have been obtained in exchange for the same amount of tradables in the more slowly growing United States.

Cited By