The Japan-US Exchange Rate, Productivity, and the Competitiveness of Japanese Industries
The Japan-US Exchange Rate, Productivity, and the Competitiveness of Japanese Industries
Abstract and Keywords
This chapter explores the relationship between the nominal exchange rate and the real exchange rate in Japan, along with the process by which industries adjust to higher real exchange rates. It looks at the yen–dollar exchange rates as well as the productivity and competitiveness of Japanese Industries, separating the Japanese economy into three sectors: High-productivity-growth manufacturing, low-productivity-growth manufacturing, and low-productivity-growth services. The chapter shows that labor sometimes moves from more productive sectors to less productive sectors and that long-run movements in the real exchange rates are influenced by the relative productivity performance of producers of traded goods. It also examines the effects of the yen’s movements on the sectoral allocation of Japanese employment and how the high yen increased the production costs of Japanese industries relative to those in the United States.
In September 1985, representatives of the United States, Japan, Germany, the United Kingdom, and France met at the Plaza Hotel in New York to engineer a depreciation of the dollar to help eliminate the continuing trade deficits of the US. The yen-dollar exchange rate, about 250 before the Plaza Accord, approached 180 by February 1986 (figure 4.1).
The yen continued to appreciate, reaching 120 in early 1988. There was another spurt of appreciation in early 1995, with the yen briefly below 90. Subsequently, the yen weakened to 130, although it was mostly in the 110–120 range during the decade ending with the 2008 financial crises. (The effects of the crisis on the yen-dollar exchange rate and on the competitiveness of Japanese industries are too recent to be analyzed here.) These trends in the Japanese currency apply also in terms of the Nominal Effective Exchange Rate (NEER). In terms of the Real Effective Exchange Rate (REER), yen appreciation was more muted: the REER was at pre-Plaza levels by early 2007.
The focus here is on the effects of the yen’s movements on Japanese industries, and on the sectoral allocation of Japanese employment. The appreciations of 1985 and 1995 significantly hurt the ability of Japan to compete with the US by raising the relative production costs of Japanese industries. This relative cost gap with US industries narrowed from 1995 owing to faster wage growth in the US, and especially to higher productivity growth in some Japanese industries. In fact, in these high-productivity Japanese manufacturing industries, such as chemicals and transport equipment, relative production costs were essentially back to pre-1985, pre-Plaza Accord levels by 2004. In contrast, the relative production costs of Japanese low-productivity manufacturing industries, such as textiles and wood products, have remained high. Clearly, in aggregate, the appreciation of the yen was not matched by an increase in Japanese productivity.
We use our model to examine the relationship between these Japan-US differences in average production costs and the allocation of employment across the three sectors. In both Japan and the US, deindustrialization is an important policy concern. While our model simulations follow the decline in employment shares in high-productivity manufacturing quite well, it fails to capture the large flow of employment from low-productivity manufacturing to services in both countries. This failure is not simply because our long-run model cannot (p.107) capture the temporary employment effects of the overvaluation of the yen. The dollar was undervalued against the yen, but the US also experienced a decline in employment in low-productivity manufacturing. We suggest that some other long-term structural factor, such as the rise of China and India, may be responsible for the surge in service-sector employment in Japan and the US.
Hamada and Okada (2007) and McKinnon and Ono (1997) note that the Plaza Accord shifted the exchange rate expectations of market participants upward, toward a sharp yen appreciation. After the Accord, market participants became convinced that the US and Japanese governments would continue to undertake policies to appreciate the yen as long as Japanese trade surpluses—especially against the US— remained large. The yen continued to appreciate, reaching a peak of slightly under 90 yen in 1995. Given the sluggishness in the adjustment of prices, the real exchange rate appreciated in tandem.
This appreciation of the real exchange rate made Japanese industries uncompetitive against comparable US industries. The competitive discrepancy was adjusted over time through changes in relative Japan-US wages, capital costs, and differential rates of technical progress. In the long run, these adjustments, collectively, should restore the real exchange rate to its equilibrium value. However, empirically, the speed at which the real exchange rate reaches equilibrium parity is very slow. According to Froot and Rogoff (1995), standard estimates of the real exchange half-life are in the range of three to five years. In our macroeconomic-trade model, we show that it was only in 2003 when the real exchange rate was restored to its equilibrium value—almost a full 20 years after the Plaza Accord.
In our model the long- and very long-run equilibrium real exchange rates are determined largely by the productivity differentials between Japan and the US—that is, the Balassa-Samuelson effect. Japan’s real exchange rate appreciates over time because of the very rapid productivity growth in its high-productivity manufacturing sector relative to its services sector. In contrast, in the US, the gap in productivity growth between high-productivity manufacturing and services is not as large.
Compared to our long- to very long-run analysis of Japan-US real exchange rates, Lane (chapter 5) is concerned with more medium-run movements. In the medium run, large creditors such as Japan experience an appreciation of their real exchange rate, as net investment inflows lead to an excess of domestic absorption relative to domestic production (the transfer effect). In our very long-run model, trade is balanced and there is no transfer effect.
(p.108) Obstfeld (chapter 3) characterizes Japanese real exchange rates from the short- to the long-runs, and depicts how business cycles and monetary policies can change real interest rates, and cause the short-run real exchange rate to deviate from its long-run value. He shows that the Balassa-Samuelson effect appears to hold in the long- to the very long-runs, although empirically it may take more than 20 years—the same time it takes in our model—for the real exchange rate to revert to equilibrium value. Neither Lane nor Obstfeld examine the effects of exchange rates on Japanese industries, or on Japanese sectoral employment, as is done here.
0.1 Outline of Chapter
In the next section, we analyze how the high yen raised the production costs of Japanese industries relative to those in US industries. (See Dekle (2005) for an analysis using firm-level data on how the high yen hurt the profitability of Japanese firms.) We compare—in dollar terms—the average production costs of 14 Japanese and US manufacturing industries. Japanese average costs at the industry level rose substantially between 1985 and 1995. In related work, Jorgenson and Nomura (2005) calculate, using detailed commodity-level prices, the purchasing-power parity (PPP) exchange rates between Japan and the US, and found that for many of the 42 industries examined, the actual yen-dollar exchange rate was much higher than the PPP exchange rates.
Since 1995, Japan-US gaps in average costs have narrowed somewhat, largely owing to more-rapid wage increases in the US. In some key Japanese industries, productivity growth was rapid, helping close the average cost gaps. These Japanese high-productivity industries are optics, transport equipment, and chemical products. In these, and in some other industries, Japanese average costs in dollar terms were almost the same as US average costs by 2004. In industries in which Japanese productivity growth was relatively low—such as rubber and plastics, and textiles and apparel—the gaps in average costs remained substantial even in 2004.
In section 2 we build a simple long-run Ricardian Japan-US macroeconomic-trade model that endogenizes the real exchange rate and wage responses. The model has three sectors: low- and high- productivity manufacturing, and services. We simulate the Japan-US real exchange rates, average costs, and sectoral employment shares. Our simulated Japan-US real exchange rates can be interpreted as an equilibrium benchmark that depends mostly on relative productivity differences between Japan and the US (and model parameters). This benchmark can be used to assess whether actual Japan-US real exchange (p.109) rates are misaligned. We focus on understanding the changing sectoral employment shares in Japan and the US, given the widespread policy interest in de-industrialization and the rise of the service sectors in both countries.
Our baseline model simulations follow the Japan-US real exchange rates, and the Japan-US gaps in average costs in both low- and high-productivity manufacturing quite well over the long run. As expected in our static, long-run framework without money, we miss most of the short-run fluctuations. With regards to sectoral employment shares, our baseline model simulations follow the employment shares in high-productivity manufacturing quite well, and in services somewhat well, but fail to follow the employment shares in low-productivity manufacturing.
1 Changing Relative Average Production Costs Since 1985
This section analyzes how competitiveness (measured as US-Japan relative average costs) have changed over time. We measure changes in the competitiveness of Japanese industries by estimating the changes in their average costs of production in comparison to the changes in the average cost of production of US industries. The change can be decomposed into the changes in capital costs, wages, cost of input materials, and total factor productivity.
Suppose there exists a well-behaved constant returns to scale production function for a representative firm in industry i of country J of the following form:
where ϒiJ(t) denotes the real gross output of this firm at time t, LiJ(t) is the labor input, KiJ(t) the capital service input, XiJ(t) the input of intermediate goods, and TiJ(t) the technology level.
The average cost of production of this firm, CiJz(t), is expressed by
where wiJ(t) denotes the nominal wage rate (measured in country J’s currency) for workers in industry i of country J at time t, riJ(t) represents the capital service price, and qiJ(t) the intermediate input price. We assume that each firm is a price taker in all factor markets. We also (p.110) assume that factor prices and the technology level, TiJ(t), are continuous functions of time.
By differentiating equation (4.2) over time and using cost minimization conditions, we obtain
where the circumflex denote the growth rate of variables. sL,i,J(t), sK,i,J(t), and sX,i,J(t) denote the cost share of each production factor:
Ai,J(t) denotes the total factor productivity (TFP) level of industry i in country J at time t. The TFP growth rate can be defined by
Using growth accounting, we can estimate the TFP growth rate.
Equation (4.2) shows that we can explain the relative competitiveness (which we measure by the gap in the average costs of production) of the two countries in each industry by the gap in factor prices between the two countries, and the gap between their TFP levels.
In order to apply equation (4.2) to discrete time-series data, we use a Tornqvist-type approximation of this equation:
We use year 1980 as the benchmark year and set Ci,J(1980) = 100.
(p.111) We can measure the inter-temporal changes in the competitiveness of industry i in country J in comparison to that of industry i in country U by the changes in the ratio of the average cost of production in the two countries measured in the same currency: e(t)(Ci,J(t)/Ci,U*(t)), where e(t) denotes the nominal exchange rate index between the currencies of the two countries (the value of country J’s currency in terms of country U's currency). We use year 1980 as the benchmark year and set e(1980) = 100.
We have data on 19 industries in both countries: 13 in manufacturing, 6 in services. Data on the following factor inputs and on the TFP growth rate are from the EU KLEMS Database (March 2008 version).
wiJ(t)LiJ(t): nominal labor compensation in industry i in country J;
wiJ(t) labor compensation divided by the quality-adjusted labor input index (1995 = 100) of industry i in country J;
qiJ(t)XiJ(t): nominal intermediate input cost of industry i in country J;
qiJ(t): intermediate input cost divided by the real intermediate input index (1995 = 100) of industry i in country J; and
ln(AiJ(t) − ln(AiJ(t − 1)): TFP growth rate of industry i in country J.
We estimated the service price of capital as follows:
riJ(t)=gross fixed capital formation price index (all assets, 1995 = 100) × (interest rate + depreciation rate − capital gains).
The gross fixed capital formation price indices are from the EU KLEMS Database. We calculated the capital gain terms from these indices. For interest rates, we use the yield of newly issued 10-year government bonds for Japan and the market yield on 10-year Treasury securities for the US. Depreciation rates were obtained from the Japan Industrial Productivity ( JIP) Database for Japan and from the BEA web site for the US. To measure capital service inputs, we used the real capital stock in 1995 prices from the EU KLEMS Database. For the yen-dollar exchange rates, we used the annual average interbank market exchange rate from Nikkei NEEDS. Further details on data construction are in annex 4.1.
1.3 Japan-US Productivity and Average Cost Comparisons
If the appreciation of Japan’s real exchange rate (figure 4.1) had been accompanied by superior productivity growth in comparison with other countries, then the appreciation of the yen until the mid-1990s would (p.112) not have reduced the international competitiveness of Japan’s manufacturing sector. However, in many manufacturing industries, the TFP growth achieved was not sufficient to cancel out the effects of yen appreciation. In particular, from the 1990s, Japan’s TFP stagnated and its manufacturing industries lost competitiveness.
Figure 4.2 compares TFP growth of manufacturing in Japan and in the US. For 1980–90, TFP growth rates in most manufacturing industries in Japan were higher than those in the US. However, during 19902004, TFP growth in almost all industries became very low in Japan, with the exception of electrical and optical equipment, and US TFP growth in most manufacturing industries exceeded Japan’s.
Figure 4.3 Panel A shows how average production costs have changed over time in Japan and the US in the electrical and optical equipment industry. Production costs fall when there is a decline in wages, capital costs, or materials costs, or a rise in TFP. In dollar terms, Japanese production costs will rise when the yen appreciates. Holding the yen constant, Japanese production costs in the electrical and optical equipment industry have trended along with US production costs. However, in terms of dollars, Japanese production costs in this industry surged after the yen appreciation in 1985, reaching a peak in about 1995. Thereafter, Japanese production costs in terms of dollars started to decline, gradually approaching US levels by 2005.
To see the role of wage rate and productivity changes on Japan-US relative average production costs, figure 4.3 Panel B shows that, relative to Japanese wages, US wages have surged, especially since 1995, contributing to the convergence of Japan-US average costs. Japanese TFP growth rates have generally been lower than US TFP growth rates, except for the period after 2002 in the electrical and optical equipment industry.
Figure 4.4 shows how average production costs, wage rate, and productivity have changed over time in Japan and the US in the rubber and plastics industry.
Generally, the patterns observed in all manufacturing industries is similar to the pattern observed in the electrical and optical equipment industry (figure 4.3) and rubber and plastics industry (figure 4.4). (Figures for the other manufacturing industries are in annex 4.2.)
First, the yen appreciation sharply deteriorated the competitiveness of Japanese industries, starting in 1985. However, the rapid increase in US wage rates has partially offset the drop in Japanese competitiveness. The decline in capital cost and intermediate input prices in Japan also contributed to offsetting the effect of the yen appreciation. (Figures for (p.113)
Second, Japan’s competitiveness relative to the US reached bottom in 1995. This result is consistent with the fact that Japan’s real effective exchange rate (REER) was most-appreciated in 1995 (along with the yen-dollar rate and the nominal effective exchange rate) (figure 4.1). After that, the yen gradually depreciated in REER terms, and by 2006 the REER approached its 1985 level.
Third, in industries where Japan’s TFP growth was on par with, or higher than, that of the US before 1990—electrical and optical equipment; transport equipment; chemicals and chemical products; general machinery; and pulp, paper, printing and publishing—the Japan-US average cost gap in dollars did not become very large even in 1995, and was negligible by 2004. In contrast, in industries where Japan’s TFP growth was much smaller than that of the US before 1990—rubber and plastics; textiles, apparel, leather, and footwear; basic metals and fabricated metal; and food, beverages, and tobacco—the Japan-US gap in the average cost of production measured in dollars became very large in 1995, and even now remains sizable.
2 The Model
As the yen appreciated, Japanese average production costs rose relative to US average production costs on an industry-by-industry basis. The relative rise in Japanese average costs was larger in industries where increases in TFP growth were lower. For example, after the yen appreciation, Japanese average costs rose substantially relative to US average costs in textiles—a low-productivity industry—while in transport equipment—a high-productivity industry—Japanese average costs relative to US average costs were relatively constant over time.
The analysis in the previous section is of the short- to medium-runs when nominal exchange rates can deviate from Purchasing Power (PPP) values. In the short- to medium-run, tight monetary policies in Japan can appreciate the yen and raise Japan’s relative average costs, as happened between 1990 and 1995 (Hamada and Okada 2007). However, there is a tendency for PPP to hold in the long run, in the sense that nominal exchange rate changes approach the changes in relative prices (Taylor and Taylor 2004). Thus, there is likely to be movement of nominal exchange rates toward PPP within our sample period of 20+ years. The real exchange rate simulated by our long- to very long-run model (p.117) below should be viewed as an equilibrium benchmark by which the overvaluation or the undervaluation of the actual real yen-dollar exchange rate can be assessed.
We focus on the impact of productivity differences between Japan and the US on Japan-US real exchange rates and Japanese sectoral employment patterns. To adequately analyze the impact of productivity differences on real exchange rates and the allocation of employment across sectors, we need a general equilibrium model. A negative shock to Japanese productivity will directly raise average costs by lowering efficiency, and also will lower average costs by lowering wages and depreciating the yen. Thus, a productivity shock has offsetting effects on average costs, and it is unclear which will dominate.
Moreover, industry-by-industry average cost comparisons cannot give a complete picture of a country’s international competitiveness. Differences in Japan-US average costs in a third sector can alter the Japanese manufacturing sector’s competitiveness through the Balassa-Samuelson effect.
We aggregate our 19 industries from the EU-KLEMS data described in section 1 and annex 4.1 into three sectors: 1) low labor-productivity manufacturing; 2) high labor-productivity manufacturing; and 3) services. There are two countries, Japan and the United States, and each produces and consumes goods in the three sectors. The manufacturing sectors in the two countries are differentiated. In fact, we find, after calibrating from the data, that Japanese low- (and high-) productivity manufacturing and US low- (and high-) productivity manufacturing goods have very low substitutability in the consumption decisions of Japanese and US households.
2.1 Aggregating into Sectors and Data on Sectoral Labor Productivities
The model is a simple Ricardian type with no capital and labor as the only factor of production, Yij = θijLij, where θij is labor productivity in industry i and country j. (For more details on the model, see annex 4.3.) We can relate θijLij to TFPij in the previous section if the production function is Cobb-Douglas, as then θ in each industry (omitting the subscripts) is equal to , where 1 − α is the labor share of income, r is the real rate of interest, assumed constant, and A is the level of TFP. Growth in labor productivity is then simply . Using (p.118) growth in TFP by industry and labor shares by industry calculated in the previous section, we calculate growth in labor productivity for each sector (13 in tradeable manufacturing, 6 in non-tradeable services).
Each of the 13 manufacturing industries is ranked according to the average growth of Japanese labor productivities in the industries between 1978 and 2003. As a result, 6 industries are classed low-productivity manufacturing: food; textiles; wood; pulp and paper; refined petroleum; and basic and fabricated metals. The 7 high-productivity industries are: chemicals; rubber; non-metallic metals; basic machinery; electronics and optical machinery; transport equipment; and manufacturing not elsewhere classified. Services are hotels; transport and storage; posts and telecommunications; financial intermediation; real estate; and other business services.
To simulate our model, we need data on labor productivity, labor supply, output (in dollars), and prices (in dollars) for each sector, for both Japan and the US. Labor supply is defined as the total number of hours worked in that particular sector. Labor productivity is thus in terms of output per hour. For each sector, we form weighted averages of the data for the industries that comprise it, where the industry’s output share in that sector is used as the weight for that industry.
Figure 4.5 depicts the data for labor productivities so constructed from 1980 to 2003. These are the data fed into the model to simulate our
In low-productivity industries, labor productivity in Japan was declining slightly during this period, while in the US, labor productivity was rising. In high-productivity industries, labor productivity growth was high in both countries, especially in the US, until about 1997. Since then, Japanese labor productivity growth has outstripped US labor productivity growth in this sector, especially since 2000. In services, labor productivity growth in both countries was flat. The level of labor productivity in services in Japan compared to that in the US is especially low.
2.2 The Balassa-Samuelson Effect and Real Exchange Rates
Figure 4.6 compares the real exchange rate in the data with the real exchange rate simulated from the model. For the data, the real exchange rate is defined as the ratio of the weighted GDP deflators in the two countries in terms of dollars. To convert the Japanese GDP deflator into dollars, we use 1) the nominal yen-dollar exchange rate, and 2) the yen-dollar exchange rate adjusted for PPP (from Heston, Summers, and Bettina 2006). The model-simulated real exchange rate, the market real exchange rate, and the PPP real exchange rate are all assumed to be equal around 1980. That is, all three exchange rates are assumed to be in equilibrium in 1980.
In the model, the real exchange rate is defined as the ratio of the weighted averages of the three sectors in Japan and in the US, in terms of a common numeraire, in our case, low-productivity manufacturing in Japan. In the data, the gap in productivity growth between services and in manufacturing is wider in Japan than in the US (figure 4.6), so we should find aggregate prices in Japan rising faster than aggregate prices in the US (the Balassa-Samuelson effect).
In the long to very long runs, the real exchange rate as simulated from our model matches the data well, especially when the nominal yen-dollar exchange rate is used to convert the Japanese GDP deflator to US dollars. Specifically, we observe the Balassa-Samuleson effect—an appreciation of the Japanese real exchange rate over the long-run. If we take our simulated real exchange rate series as a long-run benchmark, (p.121)
Obstfeld (chapter 3) shows that over the short- to medium-runs, deviations from the Balassa-Samuelson benchmark are substantial. Lane (chapter 5) finds that the yen-dollar exchange rate is cointegrated with Japan-US productivity differentials, suggesting that Balassa-Samuelson holds in the long to very long runs. We find more evidence than does Obstfeld that the yen-dollar real exchange rate reverts to its benchmark equilibrium over the long run. This is probably because our model is somewhat more general than the usual Balassa-Samuelson model, where the real exchange rate depends only on the two countries’ “difference in differences” between their tradeable-sector and non-tradeable-sector productivity growth rates.
In our model, in addition to the usual Balassa-Samuelson productivity differentials, we have three (rather than two) sectors, with each sector producing highly differentiated goods in Japan and in the US. We also have home production in each country, which, given our parameter values (annex 4.3 table4.A1), tends to raise the relative price of services in Japan, resulting in an appreciation of the model-simulated (p.122) real exchange rate that better-matches the long-run appreciation of the actual yen-dollar real exchange rate.
2.3 Sectoral Average Costs and Employment Shares
Average cost in our framework is where, in both countries, k is a constant, and w is wages in dollars. Average costs in Japan will rise relative to average costs in the US when 1) Japanese productivity growth is lower than in the US or 2) when Japanese wages grow faster than US wages, because of either faster wage growth or an appreciation of the yen (which raises Japanese wages in dollar terms). Note, however, that both w and the real dollar-yen exchange rate depend on economy-wide labor productivities, including productivity in services, with lower Japanese relative productivity in services lowering Japanese relative wages, but appreciating the yen. Thus, the net effect of productivity growth on Japan-US average costs is ambiguous.
Figure 4.7 depicts the ratios of Japanese average costs to US average costs in both the data and the model. Japanese average costs rise sharply relative to US average costs starting from about 1985 and peaking around 1995. From 1995, the closing of Japan’s average-cost gap with the US is much slower in the low-productivity sector because Japan’s productivity growth is lower than in the US in this sector. Thus, the closing of the average-cost gap is due entirely to the rise in relative US
The model tracks the long-run trends in relative costs quite well, although it misses the yen-appreciation episodes between 1985 and 1995. These are clearly episodes when the nominal yen appreciation departed from economic fundamentals, and our model, driven entirely by productivity growth and other fundamentals, cannot capture that.
We have shown that slower relative productivity growth in services tends to appreciate the real exchange rate in the long run. Many observers also tie the rise in services to the faster growth of productivity in tradable goods, invoking a de-industrialization hypothesis. The economic reasoning is simple: with a higher volume of manufacturers being produced by fewer workers, some workers have to take service jobs.
For our parameter values (table 4.A1), the model predicts that employment shares of high-productivity industries in both countries should be decreasing. The employment shares in low-productivity industries and in services should be increasing, as seen in figure 4.8.
Our model tracks the actual employment shares in high-productivity manufacturing extremely well in both countries. However, employment shares in services are increasing more sharply in the data than in the model. Also, in the data, the employment shares in low-productivity manufacturing are declining dramatically, while in the model, these shares are increasing.
Thus, while our model captures the trends in high-productivity industries reasonably well, it does less-well in low-productivity industries and services. In particular, our model fails to capture, in both countries, the dramatic actual movements of labor from low-productivity manufacturing to services. Our simulation results are robust to raising the Armington elasticities in low-productivity industries to 1.5; and to raising the consumption utility weights of low-productivity goods from 0.1 to 0.2 or 0.3 in both countries.
Why does our model fail in this important dimension? In both the US and Japan, actual productivity growth in low-productivity manufacturing is just too low to drive out labor to the service sector. In our model, labor productivity growth in low-productivity manufacturing needs to be over 4% per year to induce sufficient labor migration, a rate much higher than what is observed in the data of both countries.
The failure is not simply because our long-run model cannot capture the temporary employment effects of an overvaluation or misalignment of the yen. The dollar was undervalued against the yen, but the US also experienced a decline in employment in low-productivity manufacturing. The deviation of the yen-dollar exchange rates from economic fundamentals cannot explain the decline in low productivity manufacturing employment in both countries. There must be other factors.
The rise of China and India, countries specialized in industries classed as low-productivity manufacturing in Japan and the US, may be the key (p.126) factor. As Broda and Weinstein (2008) note, in 1992 the US exported three times as much to Japan as to China; by 2005, China was exporting twice as much to Japan as to the US. Thus, China may be a more important trading partner to Japan than the US. As Coleman (2007) points out, the rise of these newly industrialized countries can theoretically lead to de-industrialization in industrial countries. An interesting empirical exercise using our basic framework would be to combine China, India, and the US as the “rest-of-the-world” and examine how the changing sectoral labor productivities in the rest-of-the-word affects the evolution of low-productivity manufacturing in Japan.
The main objective of this chapter is to better understand the relationships among Japanese productivity growth, the Japan-US real exchange rate, and long-run economic outcomes such as Japanese employment. As the yen appreciated starting in 1985, Japanese competitiveness—defined as the ratio of US average costs to Japanese average costs—sharply declined. However, the decline in Japanese competitiveness differed among industries. In industries where Japanese productivity growth was relatively high, the decline was much smaller.
A comparison of the actual real exchange rate with this chapter’s long-run equilibrium benchmark shows that the Japanese real exchange rate was near equilibrium in 2003, when the nominal yen-dollar exchange rate was around 120 yen/dollar. The appreciation of the yen to below 90 yen/dollar, owing to the global financial crisis, is thus clearly excessive, and has resulted in large profit and employment losses for Japanese manufacturers. Our analysis suggests that a return of the yen-dollar rate to above 120 should help restore profitability to Japanese manufacturers.
The following annexes are available on the web. Please visit http://mitpress.mit.edu/japansbubble
Annex 4.1 Definition of Variables Used in the Empirical Analysis and Data Sources
Annex 4.2 Graphs of average production costs, wage rates, productivity, capital costs, and intermediate input prices of manufacturing industries discussed in section 1.3.
Annex 4.3 The Two-County, Three-Sector Ricardian Model
We thank Gianluca Benigno, James Harrigan, Ann Harrison, Sam Kor-tum, and the editors of the ESRI Project, especially Koichi Hamada, for very helpful comments on earlier drafts. Robert Dekle thanks the Center for International Research on the Japanese Economy at Tokyo University for hosting him in 2008 and Yanyu Wu for research assistance.
Broda, C. and D. Weinstein, 2008, “Exporting Deflation? Chinese Exports and Japanese Prices,” in R. Feenstra and S. Wei (editors), China’s Growing Role in World Trade, University of Chicago Press.
Coleman, J., 2007, “Accommodating Emerging Giants,” mimeo, Duke Univ.
Dekle, R., 2005, “Exchange Rate Exposure and Foreign Market Competition: Evidence from Japanese Firms,” Journal of Business, 78 (1): 281–99.
Froot, K. and K. Rogoff, 1995, “Perspectives on PPP and Long-Run Real Exchange Rates,” in G. Grossman and K. Rogoff (editors), Handbook of International Economics, Vol. III. Amsterdam: North-Holland.
Fukao, Kyoji, Sumio Hamagata, Tomohiko Inui, Keiko Ito, Hyeog Ug Kwon, Tatsuji Makino, Tsutomu Miyagawa, Yasuo Nakanishi, and Joji Tokui, 2007, “Estimation Procedures and TFP Analysis of the JIP Database 2006,” RIETI Discussion Paper 07-E-003.
Hamada, K. and Y. Okada, 2007, “Monetary and International Factors Behind Japan’s Lost Decade: From the Plaza Accord to the Great Intervention,” Journal of the Japanese and International Economies 23 (2): 200–19.
Heston, A., R. Summers, and A. Bettina, 2006, Penn World Table Version 6.2, Center for International Comparisons of Production, Income, and Prices at the University of Pennsylvania. http://pwt.econ.upenn.edu/.
Jorgenson, D. and K. Nomura, 2005, “The Industry Origins of Japanese Economic Growth,” Journal of the Japanese and International Economies, 19 (4): 460–81.
Lane, P., 2011, “International Financial Integration and Japanese Economic Performance,” Chapter 5 in this volume.
Matsuyama, Kiminori, 2009, “Structural Change in an Interdependent World: A Global View of Manufacturing Decline,” Journal of the European Economic Association, 7 (2–3): 478–86.
Messina, J., 2006, “The Role of Product Market Regulations in the Process of Structural Change,” European Economic Review, 53 (3): 1863–90.
McKinnon, R., and K. Ohno, 1997, Dollar and the Yen, Cambridge: The MIT Press.
Obstfeld, M. and K. Rogoff, 1996, Foundations of International Macroeconomics, Cambridge: The MIT Press.
(p.128) Obstfeld, M., 2011, “Time of Troubles: The Yen and Japan’s Economy, 1985–2008,” Chapter 3 in this volume.
Taylor, A. and M. P. Taylor, 2004, “The Purchasing Power Parity Debate,” Journal of Economic Perspectives, 18 (2): 135–58.
Timmer, Marcel, Mary O‘Mahony, and Bart van Ark, 2007a, “The EU KLEMS Growth and Productivity Accounts: An Overview,” Univ of Groningen and Univ of Birmingham.
Timmer, Marcel, Mary O‘Mahony, and Bart van Ark, 2007b, “EU KLEMS Growth and Productivity Accounts, Version 1.0, Part I, Methodology,” Univ of Groningen and Univ of Birmingham.
Ungor, M., 2008, “De-industrialization of the Riches and the Rise of China,” mimeo, Univ of Southern California.