Journal Issue
Share
Article

Japan: 2011 Spillover Report—Selected Issues

Author(s):
International Monetary Fund
Published Date:
July 2011
Share
  • ShareShare
Show Summary Details

Chapter I. Japan’s Role in Regional Trade1

A. Changing Patterns in Japan’s Trade

Japan’s role in regional trade. Intra-regional trade has expanded rapidly since 1990, largely owing to dynamic economies such as China (Figure 1). Nonetheless, Japan’s intra-regional exports as a share of global GDP have remained remarkably stable—even during the crisis—and account for more than two-thirds of industrial countries’ intra-regional trade. Japan’s deepening regional integration has largely been driven by the outsourcing of production by Japanese firms to neighboring countries, especially China, Hong Kong SAR, and Singapore. This integration has implications for the interpretation of changes in Japan’s export structure.

Figure 1.Intraregional Trade

(exports in percent of world GDP)

Source: DOTS and WEO.

Shifting export structure. The share of high-technology goods in Japan’s overall exports has fallen from 34 percent in 1995 to 23 percent in 2005. This partly reflects a shift in Japan’s trade structure—from the export of high-tech final products toward a focus on sophisticated intermediate inputs, combined with an outsourcing of the low-tech stages of production to emerging Asian countries. Japan has thus established itself as an important supplier of sophisticated manufacturing inputs at the global and regional levels, especially in the transport and electrical-machinery sectors (Table 1). Even though they may not constitute a large share in Japan’s overall exports, these items account for a significant share of global exports in the semiconductor and auto subsectors, and are an important input not only for Asian countries but also for the United States and European Union.

Table 1Japan’s Share in Global Markets, 2010
8486840885418703
JPN exports (% world exports; reporting)34.310.015.017.1
Imports from Japan (% 4-digit imports):
China35.737.119.623.5
Hong Kong23.27.615.230.6
India12.210.07.815.0
Indonesia48.631.135.120.1
South Korea40.634.625.318.7
United States53.220.217.827.7
EU27 (Ext Trade)27.225.26.433.3
Notes: 8486: Boilers and reactors: mach & appl for mnf semiconductors, 8408: Boilers and reactors: compression-Ignition for combustion engines, 8541: electrical machinery: semiconductor devices, 8703: vehicles excl railways: autosSource: Global Trade Atlas.
Notes: 8486: Boilers and reactors: mach & appl for mnf semiconductors, 8408: Boilers and reactors: compression-Ignition for combustion engines, 8541: electrical machinery: semiconductor devices, 8703: vehicles excl railways: autosSource: Global Trade Atlas.

Rising similarity with export structures of emerging Asia. Although it continues to compete with other advanced countries—based on the export similarity index (ESI)2—the export structures of countries such as China and Thailand are also converging with that of Japan (Figure 2). Further, competitive pressure from Korea appears to have increased recently, in part due to its ability to brand and market products in China and other Asian countries. Japan’s increased outsourcing and upstream position has facilitated the shift in technology content to other Asian countries, adding to the apparent convergence in export structures. Rising similarity could thus reflect increased complementarity, as well as competition.

Figure 2.Japan--Export Similarity Index

(exports in percent of world GDP)

Source: COMTRADE and Fund staff calculations.

Position in the supply chain. Japan is clearly upstream in the Asian supply chain and is an important source of foreign value added (FVA) in the gross exports of other Asian countries (Table 2). Japan’s contribution to FVA is especially high in countries engaged in assembly or processing activities, such as Singapore, Taiwan and China—particularly for high-tech exports, such as electronic equipment and motor vehicles (Table 3). Japan has thus become more integrated in Asian regional trade, implying that a disruption in production of key intermediate inputs could spill over to other countries in the supply chain.

Table 2Measures of Vertical Specialization across Borders: 2004
Country(1) Imported contents

embodied in gross exports
(2) Indirect exports sent

to third countries1
(3) Upstream or downstream

position, (2)/(1)
Advanced economies
EU-1511.420.91.8
Japan12.230.82.5
United States12.926.92.1
Asian Newly Industrialized Countries
Korea33.923.10.7
Hong Kong27.519.50.7
Taiwan41.127.20.7
Emerging
China35.712.50.4
EU accession countries30.811.30.4
Mexico48.010.00.2
Source: Koopman and others (2010).

Includes indirect exports that return to home country.

Source: Koopman and others (2010).

Includes indirect exports that return to home country.

Table 3Sources of Value-Added for Asian Countries Gross Exports, 2004
Electronic equipmentMotor vehicles and parts
CountryGross

exports
DVAFVAGross

exports
DVAFVA
o/w JPNo/w

JPN
CHN20,40560.539.510.165,00263.436.68.8
HKG1,15638.661.416.95,73479.720.34.2
IDN11,17444.155.912.352474.625.46.4
IND6,25663.736.33.11,90577.822.22.8
KOR1,43058.241.811.51,06674.625.45.6
MYS29950.249.89.420859.041.017.7
PHL6,04552.947.114.53,33552.647.417.0
SGP1,78018.981.114.54,49945.454.68.7
THA22,57342.157.915.02,62156.044.018.3
TWN1,14753.846.212.21,65659.340.718.6
VNM5,92144.655.410.04,86657.442.67.0
Source: Koopman, and others (2010).Analysis is based on GTAP data for 2004.Note: Gross exports are in millions of U.S. dollars; DVA = domestic value added share in gross exports;FVA = foreign value added share in gross exports; and o/w JPN = Japanese value added share in total FVA.
Source: Koopman, and others (2010).Analysis is based on GTAP data for 2004.Note: Gross exports are in millions of U.S. dollars; DVA = domestic value added share in gross exports;FVA = foreign value added share in gross exports; and o/w JPN = Japanese value added share in total FVA.

Role of Japanese FDI. Japan’s role in the Asian supply chain is strongly linked to its vertical FDI and its role in the spread of technology. Vertical FDI by Japanese multinationals has traditionally been motivated by factor-price differentials; in contrast, U.S. FDI has generally been motivated by market-access considerations. Broadly speaking, the labor-intensive stages of Japanese production, such as final assembly, have been moved to countries with lower unskilled labor costs, while activities that are relatively intensive in skilled labor—such as marketing, patenting and innovation—have been retained in headquarters. As such, even though the share of Japan’s high technology exports may have declined, it has retained those aspects of production with the highest value added. Analysis of the iPod suggests that the second largest share of value added is indeed captured by Japanese firms, which produce the high-value components of the product. This also explains why Japan has consistently outperformed other G-7 countries in terms of increasing the income level of its exports (EXPY) (Figure 3).3

Figure 3.Income Level of Exports--G7

(USD billions)

Source: COMTRADE and Fund staff calculations.

B. Estimating Spillovers: A Sectoral Trade Elasticities Approach

Analytical framework. A simple model that combines a partial equilibrium approach with input-output analysis is used to analyze the response of sectoral trade flows to changes in relative prices.4 The simulation assumes a relative price decline that could arise from a 10 percent real depreciation; and then outlines changes in the structure of Japanese trade across sectors owing to differences in import demand and substitution elasticities, as well as in the amount of imported intermediate goods. The analysis does not model the drivers of these exchange rate changes, and does not account for indirect supply-side effects such as inter-industry reallocation of production factors.

Aggregate effects. The simulation suggests that a change in relative prices results in important long-run responses on trade flows. Assuming full pass-through of exchange rate changes to import prices, the Japanese trade balance improves by more than 3 percentage points of GDP, largely driven by a strong export response. The relatively large trade response reflects Japan’s upstream position in the Asian supply chain and the limited share of imported content in its exports. This likely represents an upper bound; however, as imperfect exchange rate pass-through and pricing-to-market are likely to mitigate the adjustment in trade flows to exchange rate changes. Also, adjustment is likely to be gradual, given high fixed costs in production and trade relationships.

Figure 4.Japan--Response to Relative Price Decline

(percent change)

Source: COMTRADE and Fund staff calculations.

Sectoral effects. A depreciation results in an increase in the share of medium-high technology exports, largely driven by the auto sector. In fact, a depreciation would reinforce Japan’s comparative advantage in medium-high technology exports and allow it to recover its relative specialization in the auto sector, which has been increasingly lost to countries such as Germany and France. Medium-high technology exports are generally more responsive to relative price changes, reflecting both higher domestic value added, and the discretionary consumer character of this sector, which is captured in income elasticities.

Supply chain effects. Trade with Asian partners is less sensitive to relative price changes, so that trade-balance adjustment takes place mainly outside the supply chain (Figure 4). In response to a depreciation, exports to (imports from) supply-chain partners would increase (decline) by a smaller amount, compared to the rest of the world. The net result is a greater outward reorientation of Japanese trade beyond the region, and a rebalancing with the rest of the world. Note that this effect is symmetric i.e., in response to an appreciation, trade with Asian partners would react less, resulting in greater regional reorientation of trade flows.

Chapter II. Analysis of International Spillovers Through the Asian Input-Output Table5

The Asian International Input-Output Tables (Asian IO tables) provide a systematic description of intermediate- and final-goods trade flows, and allow a quantitative assessment of regional interdependence among ten economies, including the NIEs3, ASEAN4, China, United States, and Japan.6 The analysis in this chapter covers the period from 1995-2008: actual published tables are available until 2000, whereas the 2005 and 2008 tables are extrapolated from past data (see April 2010 Asia-Pacific Regional Economic Outlook, Chapter III).

First, the analysis considers production-inducement coefficients (PICs), which capture the amount of production in a country that is induced by an additional unit of final demand in another country. These are based on information in the IO Table’s inverse-Leontief matrix, and reflect the strength of economic linkages between economies through multiple rounds of intermediate trade. PICs attributed to foreign economies have risen in most cases, suggesting deepening interdependence.

Figure 1.Production Inducement Coefficents to Home and Foreign Countries

Sources: Asian International Input-Output Table; and IMF staff calculations.

1 Coefficients provide quantitative information of production by countries induced by an additional unit of final demand in one country.

2 For example, one unit increase in final demand in Indonesia in 2008 induce total 1.3 unit of production, in which 1.0 unit to home while 0.3 unit to foreign countries.

Note: The four bars represent data for respectively, 1995, 2000, 2005, and 2008.

Japan plays key role as supplier of basic and capital-intensive parts/products at the upstream stage of regional production chain. Although the size and share of production induced by Japan has fallen over the past 15 years, it still remains sizeable, at around 20 percent of total foreign inducement in 2008. While the degree of production inducement to NIEs3/ASEAN4 economies has remained constant, the amount induced by China has increased dramatically through the 2000s, offering a clear contrast with Japan. This is consistent with: (i) the recent increase in overseas operations of Japan’s multinational enterprises, especially in Asia; and (ii) the technological progress of other Asian economies.

Figure 2.Breakdown of Production Inducement Coefficients to Foreign Countries

1 The coefficients to NIEs3 and ASEAN4 are calculated by excluding those to their home countries.

A breakdown of the value-added content of exports confirms Japan’s upstream position in the production chain. The general increase in the share of foreign value-added for most countries’ exports suggests a process of deepening trade integration. Countries with a higher foreign value-added share (such as China) focus more on assembly and processing and are considered -downstream”. In contrast, Japan is considered -upstream” with the high domestic value-added share in its own exports, and the high contribution to the foreign value-added shares of other Asian countries’ exports.

Table 1Source of Value-added in Exports
1995IDMYPHSGTHCHTWKRJP
Domestic90.9%65.8%82.7%46.8%75.1%86.0%71.6%80.5%95.6%
Asia3.6%19.5%8.0%29.0%11.1%4.4%9.7%6.0%1.1%
Japan1.9%11.1%3.4%13.9%6.3%2.3%6.2%3.7%
China0.3%1.0%0.6%1.5%0.8%0.8%0.9%0.3%
US1.1%5.4%3.0%7.7%2.9%1.3%4.1%3.1%0.8%
ROW4.5%9.3%6.3%16.5%10.9%8.4%14.6%10.3%2.6%
Total100%100%100%100%100%100%100%100%100%
2005IDMYPHSGTHCHTWKRJP
Domestic92.1%58.4%78.1%38.1%72.5%79.5%71.8%80.6%94.2%
Asia4.2%27.3%13.2%23.7%16.8%8.5%16.7%7.7%2.2%
Japan1.3%7.6%5.3%6.7%7.1%2.9%7.7%3.3%
China1.0%5.6%2.1%5.0%3.1%2.8%2.1%0.8%
US0.8%7.0%4.3%6.4%2.5%1.6%3.5%2.2%0.8%
ROW2.9%7.3%4.4%31.9%8.2%10.3%8.1%9.5%2.8%
Total100%100%100%100%100%100%100%100%100%

Moreover, the income dependence analysis illustrates Japan’s key role as one the regions’ largest consumer markets, following the United States. The income dependence analysis captures the extent to which an Asian country’s income depends on Japan’s final demand. Japan remains important, although its share has fallen recently—in contrast, China is gaining importance as final demand destination, and has become the largest consumer market for NIEs3/ASEAN4, exceeding even the United States in 2008.

Figure 3.Countries’ Dependence on Japan final Demand

Sources: Asian International Input-Output Table; and IMF staff calculations.

1 “ROW” denotes the rest of world.

2 “NIEs/ASEAN” of NIEs3 and ASEAN4 countries are calculated by including their income dependences on their home countries.

Chapter III. Japan’s Outward Foreign Direct Investment7

Japan’s stock of outward FDl is concentrated mainly in the United States, followed closely by Asia. Since 2000, Japan’s FDI has moved progressively away from the United States towards Asia, reflecting the increased presence of Japan’s corporations in the region. The Euro Area has continued to attract around one quarter of Japan’s outward FDI.

Table 1JPN’s Outward FDI by Region (billions of USD)
200120062009
DollarsPercentDollarsPercentDollarsPercent
Total300100387100741100
Asia5318882317624
P.R.China103256557
North America145481564024032
U.S.140471503923131
Central and South America2173399913
Cayman Isl.93185659
Middle East102041
Africa101061
EU6923922417524
U.K.3311246314
Others124164405
Oceania83133365

Japan’s outward FDI flows can be modelled by employing a gravity-model framework.

The specification is as follows:8

The dependent variable is Japan’s outward FDI flows to host country i as of year t. Y is the host country’s GDP, which captures its market size. |DpcY| is absolute difference in per capita GDP between Japan and host country, which proxies for differences in labor costs or factor endowments. Trade is bilateral trade between Japan and the host country, lagged by one period to account for potential endogeneity. Other control variables include one-period lagged variables of host-country characteristics (GDP growth as a proxy for productivity and the ratio of private credit to GDP as a proxy for financial depth). The variable ui is a fixed-effect (capturing host-country invariant factors, such geographical distance from Japan and domestic institutions); vt is a time-effect (which captures factors affecting all host countries in a similar fashion); and εtt is error-term.

The model is estimated using panel OLS (two-way fixed effects), based on a panel of 129 economies over 1989-2008. To check for robustness, other estimation methods were tested (e.g., Tobit estimation), with broadly similar results.

Table 2Determinants of Japanese Outward FDI
(Dependent variable: log of Japanese outward FDI)
1989-20081999-2008
All countriesEMs/LICsEMs/LICs
(1)(2)(3)(4)(5)(6)(7)(8)(9)
GDP (host country)-0.564-0.678-0.4400.2180.3160.7818.176**9.538***7.832**
(1.343)(1.398)(1.398)(1.482)(1.541)(1.551)(3.561)(3.772)(3.580)
Difference in GDP per capita2.310**2.323**2.003**13.580*13.311*16.430**38.631***37.511***39.934***
(0.993)(0.993)(1.033)(7.479)(7.520)(7.842)(14.463)(14.665)(14.511)
Bilateral trade (t-1)0.764**0.710**0.806**0.723*0.659*0.5410.2690.2500.277
(0.361)(0.363)(0.379)(0.388)(0.391)(0.420)(0.689)(0.695)(0.703)
GDP growth (host country) (t-1)0.0240.009-0.082
(0.030)(0.031)(0.071)
Private credit (host country) (t-1)-0.0050.028**0.012
(0.009)(0.013)(0.029)
Specificationtwo-way fixed effecttwo-way fixed effecttwo-way fixed effecttwo-way fixed effecttwo-way fixed effecttwo-way fixed effecttwo-way fixed effecttwo-way fixed effecttwo-way fixed effect
Number of observations230122912239172417141672937933932
Number of countries129129129979797979797
R20.5680.5690.5730.5040.5050.5110.4520.4520.454
Note: Standard errors are in parentheses.A *, **, and *** represent statistical significance at 10, 5, and 1 percent respectively.
Note: Standard errors are in parentheses.A *, **, and *** represent statistical significance at 10, 5, and 1 percent respectively.

Key results include:

  • The coefficient on income differences is positive and statistically significant in all samples (equations (1) to (9)). This suggests that labor cost differentials—vertical integration—has been main driver of Japan’s FDI.

  • The coefficient on the host country’s GDP is significant only for developing economies after 2000 (equations (7) to (9)), suggesting that the host country’s market size has recently become another important driving factor for Japan’s outward FDI. This is also consistent with the expansion of Japan’s multinational operations, especially in Asia, to develop local demand.

  • There is a clear positive relationship between Japan’s outward FDI and bilateral trade over the full sample (equations (1) to (6)), although this relationship weakens in the latter part of the sample for developing economies (equations (7) to (9)). This is consistent with the idea that originally, Japan’s outward FDI was complementary to its trade pattern (e.g., exporting parts/capital goods to factories/subsidiaries financed by Japan’s FDI); while since 2000 it has become increasingly aimed at servicing local markets, substituting for its exports.

Chapter IV. The Global Role of Japan’s Capital Markets and Investors: Stylized Facts9

Figure 1.Japan: Foreign Investors in Japanese Capital Markets

Sources: Bank of Japan, Ministry of Finance, IMF staff calculations.

Figure 2.Japan: Foreign Borrowers in Japan’s Capital Markets

Sources: Bank of Japan, Ministry of Finance, IMF staff calculations.

Figure 3.Japanese Investors in Foreign Markets

Source: Bank of Japan, CPIS, IMF staff calculations.

Chapter V. Extracting Aggregate Spillover Indices From the Gvar Model10

A. The GVAR Model

The GVAR (Global Vector Autoregressive) model is designed to examine the role of unobserved global factors on country-specific variables. The model is estimated in two steps.

First, individual country-level models are constructed, with country variables such as real GDP as a function of a constant, a trend, their own lags and, finally, contemporaneous and lagged trade weighted averages of corresponding foreign variables:

where x is a vector of endogenous variables for country i at time t, and x* is a vector of trade weighted averages of endogenous variables for the trading partners of country i, α is the intercept and t denotes the time trend.

Second, these individual models are combined in a consistent manner to create a global model, which can generate impulse responses or forecasts for all of the variables simultaneously. The country specific models (1) are subsequently aggregated by recognizing that the explanatory variables in equation (1) can be written as:

where W matrix contains trade weights for every country. The variables include: log of real GDP, inflation, real equity price, nominal exchange rate to the U.S. dollar, short and long-term interest rates and the oil price. The data set covers 25 countries and the Euro Area.11 The sample period is from 1979Q4 to 2009Q4.

B. Aggregate Spillover Indices

It is possible to use the GVAR to assess the importance of a particular country as a source of spillovers.12 One of the outputs from the GVAR model is the generalized variance decomposition matrix, which shows how much a shock to a particular variable contributes to the forecast error variance of another variable, accounting for the correlation structure. This information can be used to define an -aggregate spillover index”.

The aggregate spillover index is calculated for every variable according the following three-step procedure: (i) subtract the total contribution of the variance of domestic variables from the total forecast error variance; (ii) add up the variance shares due to the variables of every particular country and divide them by the total contribution due to variance of foreign variables; and (iii) normalize the resulting scores by subtracting the mean and dividing by standard deviation.

The resulting index shows the importance of all Japanese variables to the forecast error variance of a particular variable in all other countries. When the aggregate spillover index is less (greater) than 0, Japanese variables contribute less (more) to forecast error variance than the average contribution of other trading partners. Since the index is in terms of standard deviations, a number greater than 2 can be interpreted as high.

The results can be summarized as follows:

  • When the intensity of trade linkages is averaged over the period from 1980 to 2009, Japan’s importance to its trading partners is higher than the importance of the average trading partner (Chart 1). In some variables (output, inflation and real equity prices) the average aggregate spillover index is higher only relative to the region.13 For exchange rate, short-term and long-term interest rate the situation is reversed presumably because these variables are more linked with other advanced economies.

  • The importance of Japan to its trading partners has declined over time (Charts 2 and 3). When focusing on the second half of the sample, Japan contributes less to the forecast error variance of its trading partners than the average country and its aggregate spillover indices are less than zero.

  • Among the S4 countries, Japan contributes mostly to the forecast error variances of inflation and exchange rate (Chart 4). The Euro Area and the United States are by far the most important economies when compared to Japan or the United Kingdom. Japan is generally ranked third after the United States and the Euro Area.

Figure 1.Aggregate Spillover Indices (1980-2009)

Figure 2.Average agggregate spillover Indices (1980-94)

Figure 3.Average Aggregate spillover Indices (1995-2009)

Figure 4.Average Spillover Indices for the S4 countries

Chapter VI. Global and Regional Bank Linkages14

This analysis traces the network spillovers resulting from hypothetical credit events to specific banking systems.15 It is based on the methodology in Espinosa-Vega and Sole (2010) and relies on a matrix of BIS country-level bank exposure- and capital data for 26 countries.16 Two simulations are performed: a simulation of a banking system under distress that is unable to repay interbank loans to others (a credit shock); and a simulation of a distressed banking system that is not only unable to repay its loans, but is also unable to rollover its funding to others (credit shock plus funding shock) (Figures 1a and 1b).17

Figure 1a.Simulating a credit shock

Source: April 2009 GPSR http://www.imf.org/External/pubs/FT/GPSR/2009/01/index.html#ch1box

Note: The assets of the banking system, i, comprise interbank loans, xjt, to all j countries, plus other assets, The liabilities comprise interbank deposits from the rest of the j banking systems, deposits aid bond issuances. Following a credit event from country h (Figure 1a), country i incurs a credit loss of λ times exposure to h, which is absorbed by i”s capital, k. If capital is unable to absorb the loss, the banking system fails. Similarly, if i is unable to replace a traction ρ of its funding from (defaulting) h (Figure 1b), it needs to fie sell its assets, which trade at a discount, and hence amounts to (1+δ)ρ times the loss in book-value terms.

The loss is absorbed by the capital; if not absorbed, the banking system fails due to the finding-shortfall. The analysis assumes λ=1, δ=1, ρ=0.35.

Figure 1b.Simulating a funding and credit shock

Bilateral exposure data suggest that Japanese banks are mostly exposed to the U.S. and U.K. banking systems. Among Asian countries, Japan’s largest exposures are to Australia, South Korea and China, while the remaining countries account for less than 1 percent of Japan’s bank global exposures. Among Asian countries, China, Taiwan and Australia are the main funding countries, but these linkages remain small.

Network analysis results confirm that Japan is most at risk from exposures in the United States and the United Kingdom. Under the extreme scenario (λ=1, ρ=0.5, δ=2), there are eight instances in which Japan is expected to fail owing to distress in various U.S. and European banking systems. In particular, Japan’s banking system is expected to lose 57.6 percent (35.8 percent) of its pre-shock capital if the U.K. (U.S.) banking system suffers a credit event (Table 1). If the event is combined with a funding shock, then the United States could trigger distress in Japan, but only in the third contagion round.

Table 1Capital Impairment (in percent of pre-shock capital)
Impact on Japan if trigger country

defaults
Impact on others if Japan defaults
Trigger countryCredit shock 1/Credit & Funding

shock 2/
Affected

countries
Credit shock 1/Credit & Funding

shock 2/
Australia-4.4-4.5Australia-2.2-8.8
France-6.2-72.3France-10.8-13.8
Germany-7.2-72.3Germany-2.6-7.2
Ireland-3.1-72.3Ireland-10.5-11.7
Italy-0.5-72.3Italy-0.2-0.5
Portugal0.00.0Portugal0.0-0.3
Spain-0.6-0.7Spain-0.8-1.5
UK-57.6-72.3UK-25.3-39.7
US-35.8FullUS-9.6-14.6
China-1.6-1.9China-1.3-2.2
Taiwan-0.2-0.4Taiwan-5.9-6.8
India-0.4-0.5India-0.2-1.3
Indonesia-0.4-0.4Indonesia-1.6-5.7
Malaysia-0.2-0.2Malaysia-1.4-3.0
Philippines0.0-0.1Philippines-5.4-6.0
South Korea-3.1-3.2South Korea-4.1-14.7
Thailand-0.6-0.6Thailand-2.9-7.2
Vietnam-0.1-0.1Vietnam-0.9-2.6

Assumes loss-given-default is 1. The figures represent the direct and indirect effects of failures.

The results of this shock are highly sensitive to the choice of parameters. The benchmark assumes δ=1, ρ=0.35.

Assumes loss-given-default is 1. The figures represent the direct and indirect effects of failures.

The results of this shock are highly sensitive to the choice of parameters. The benchmark assumes δ=1, ρ=0.35.

Among Asian countries, Japan is most at risk from Australia and South Korea, but the potential impairment is below 5 percent of the pre-shock capital of the Japanese banking system; the combined shock does not substantially increase the distress from these countries (Table 1).

If Japanese banks become distressed, Australia and South Korea are most at risk. Even under extreme circumstances, this is not expected to have systemic consequences in Asia, with the exception for South Korea, which could suffer a credit event owing to second-round contagion effects from European and the U.S. banks. The most affected are the European and U.S. banks. (Table 1).

Network analysis confirms the limited spillovers that Japanese banks could have on the region. Given the modest role of Japanese banks in the region (compared to the rest of the world), very few banking systems would face systemic difficulties if the Japanese banking system were to become distressed.

Chapter VII. The Transmission of Japanese Financial-Sector Stress18

The global recession demonstrated forcefully how strains in the financial sector of one country can rapidly affect financial stability in another. Although Japan has not been at the core of the recent financial crisis, the size of its financial sector suggests that it might nonetheless be a potential source for spillovers.

Figure 1.Financial Stress Indices

(standardized index)

This chapter explores the extent to which financial strains in Japan transmit to other financial markets, using the IMF’s stress index (FSl).19 The FSI is a market-based measure that combines information on country’s securities, exchange markets, and the banking sector into a composite index. It tracks market-price movements relative to past levels or trends and defines stress as the deviation from historical norms. The index captures the most important stress episodes identified in the literature,20 and has been compiled for 17 advanced countries and 26 emerging economies using monthly data. A detailed description of the FSI is provided in the 2009 Spring World Economic Outlook.

To explore the spillovers from stress in Japan to other emerging regions, staff have employed a two-step approach. In the first step, staff extract a common stress component for Asian and other emerging economies (FSI-EM) derived the from common time effects of an unbalanced panel of emerging-economy stress indices from 1997-2010. In the second step, the common time-component of emerging-economy stress is related to the FSI for Japan (FSI-Japan), as well as the FSI for other developed regions (FSI-G6), and a range of global control variables (Globalfactors).

Asian economies are found to share a large common stress time-component. For Asian economies, the common time-component explains about 80 percent of the time variation of the FSI across economies. This compares to about 55 percent for other non-Asian emerging economies (e.g., Turkey, Brazil, and Russia). The FSI subindices with the strongest time co-movement are security and exchange markets, while comovement in banking-sector strains are less common among emerging economies.

Table 1Financial stress in Asian emerging Economies
Whole

sample
Whole

sample
1997-Pre

Lehman
Whole

sample
Whole

sample
1997-Pre

Lehman
AsiaAsiaAsiaOther EmsOther EMsOther EMs
FSI_Japan0.44***0.15*0.21**0.40***0.060.03
-0.048-0.079-0.096-0.031-0.043-0.045
FSI_G60.34***0.49***0.40***0.52***
-0.074-0.089-0.04-0.042
Libor0.22**0.17**0.120-0.05-0.14***
-0.088-0.083-0.099-0.057-0.046-0.046
Commodity
price index1.681.670.9-2.57***-2.60***-2.70***
(log)
-1.226-1.155-1.239-0.801-0.63-0.582
Industrial
production-16.80***-16.92***-15.94***0.910.770.49
(log)
-4.098-3.86-3.98-2.68-2.107-1.869
Constant70.72***71.66***70.77***6.837.9410.16
-14.367-13.535-13.815-9.394-7.386-6.486
Observations161161139161161139
R-squared0.5150.5720.5910.5590.7290.747
Source: IMF staff estimates, Dependent variable: common time component of financial stress index for emerging economies. EM Asia: China, India, Indonesia, Korea, Malaysia, Philippines, and Thailand EM other: all other EMs including Brazil, Russia, Turkey, South Africa, and others.
Source: IMF staff estimates, Dependent variable: common time component of financial stress index for emerging economies. EM Asia: China, India, Indonesia, Korea, Malaysia, Philippines, and Thailand EM other: all other EMs including Brazil, Russia, Turkey, South Africa, and others.

Financial stress in Japan has measureable spillovers to financial markets in Asia. A one standard-deviation increase in Japan’s FSI raises financial stress in the region by around 0.2 standard deviations. This effect is small relative to the spillovers from other G7 economies (with a coefficient of 0.3-0.5) but is robust to the exclusion of the global crisis from the sample (3rd column). Stress spillovers from Japan to other emerging economies are small and not statistically significant (last three columns).

Rapid transmission of financial stress illustrates the strength of international linkages across securities markets. Lag-structure tests (not shown) indicate that transmission of financial stress occurs within one month. And robustness tests confirm that the results are not sensitive to changes in the method for extracting the common stress component, or to the potential endogeneity of advanced-economy stress in the second-stage regression.

Chapter VIII. GIMF Simulations of Fiscal Consolidation and Growth Strategy21

Several scenarios are run to show spillovers from Japan’s policies to other regions using IMF’s Global Integrated Monetary and Fiscal model (GIMF). 22 The first set of simulations show the implications of financing earthquake-related expenditures and subsequent fiscal consolidation based on staff’s debt sustainability analysis; the second set examines the spillovers from the authorities’ growth strategy; and the final set presents the spillovers from their joint implementation.

Simulations show that fiscal reforms would benefit the rest of the world in the long term once the adjustment is fully completed, but may involve short-term costs. In the medium-term and during the transition, the net effects would depend on relative importance of trade linkages, permanent-income effects, exchange-rate flexibility, and monetary accommodation. Once the adjustment is fully completed, however, lower long-term real interest rates supported by higher public savings would benefit all regions. Some of the short-term negative spillovers to the other regions could be eliminated by implementation of the growth strategy and monetary accommodation in Japan. As the model focuses mainly on macro-level trade channels, spillovers could be larger if supply-chain linkages and other financial contagion channels, such as carry trades, are considered.

A. Fiscal Consolidation

The government has announced broad outlines of a medium-term fiscal strategy, but key details are not yet clear. The government’s Fiscal Management Strategy (released in June 2010) aims to halve the primary deficit by FY2015 and put the debt-to-GDP ratio on a downward path from FY2021 onwards. On June 30, 2011, the authorities outlined their social security reform plans to support their medium-term fiscal strategy. The plan proposes to double the consumption tax to 10 percent in stages by the mid-2010s, and to use the proceeds to fund social security. The tax increase would allow the government to meet its deficit target for FY2015. The plan also proposed to raise the pension retirement age and adjust nominal pension benefits for deflation, but did not stipulate steps beyond FY2015 for meeting the final target of reducing the debt ratio starting in FY2021 at the latest.

According to staff’s analysis, stabilizing the net debt ratio by 2016 and reducing it to around 135 percent of GDP by 2020 would require a 10 percent of GDP adjustment in the structural primary balance starting in 2012. While there are various possible options to achieve such adjustment, given the limited scope for cutting expenditure, fiscal adjustment would need to rely mainly on new revenue sources and constraints on spending growth.

The consolidation scenario assumes a gradual 10 percentage point increase in the consumption tax (Chart 1). A moderate increase in the consumption tax could start in 2012, when a cyclical recovery is underway, to limit bond issuance and strengthen the commitment to fiscal reforms. A gradual but sustained fiscal consolidation starts in 2013. Adjustment of about 2½ percent of GDP would come from the expiry of the fiscal stimulus package and modest expenditure adjustment, which are already incorporated in staff’s central WEO projections. The scenario further assumes a phased increase of the consumption tax (with some frontloading) raising revenues by 5 percent of GDP, and an increase in personal income tax by ½ percent of GDP. At the same time, corporate income tax is lowered, reducing revenues by ½ percent of GDP. In addition, the scenario builds in a decline in government consumption by 1¾ percent of GDP and in public investment by ¼ percent of GDP. The rest of the adjustment comes from transfers. The simulations assume that the package is fully credible in that the entire adjustment is anticipated, so that private agents adjust their behavior starting from the initial period.

Consolidation without supporting structural reforms would lower Japanese growth in the short-run but is likely to bring long-run benefits (Chart 2).

  • In the short-run, fiscal adjustment would depress GDP growth for several years by about ½ percentage points relative to a non-adjustment scenario. The increase in the consumption tax, personal income tax, lower government consumption and investment reduce domestic demand.

  • Some of the negative effects in the short-run can be offset by accommodative monetary policy in Japan (Chart 3).23 Keeping nominal interest rates low while inflation returns to its steady-state growth would lower the short-run real interest rates.24

  • In the medium-run, real GDP could rise above the baseline, but would depend on various factors, including the impact from lower long-run real interest rates, a fall in precautionary savings, a switch to less distortionary corporate taxes, and confidence effects.

  • Fiscal consolidation increases the trade balance. While a decline in private savings could offset some of the increase in government savings, world real interest rates would also change. As overall savings increase, Japan’s trade balance improves, requiring a real depreciation of the yen.

Relative to the impact in Japan, the spillovers into other economies are muted, with the largest impact on emerging Asia. Exports decline in all regions as a result of lower demand from Japan. The impact on imports and overall GDP, however, depends on various factors, including a) exchange rate flexibility and b) the increase in domestic demand—particularly in the United States and Euro Area—arising from the impact of lower long-run interest rates on investment, supply, and permanent income.

  • Over the next 5-10 years, fiscal consolidation reduces demand for imports by Japan, but the net effect on GDP in the region depends on the monetary-policy response, permanent income effects in other regions, and the flexibility of exchange rates.

  • On the one extreme, if the exchange rate is fixed, tradable inflation declines for Japan’s regional partners, pulling up real interest rates and reducing domestic demand and imports. As a result, real GDP declines. This negative impact on domestic demand would be lower for countries with restricted capital mobility. In such a situation, the interest rate is not forced to increase as much, limiting the negative spillovers.

  • At the other extreme, if the exchange rate adjusts fully, real interest rates do not increase as much, dampening the impact on GDP (Chart 4).

  • In other regions where the trade linkages are more modest, permanent income effects dominate—so that lower real interest rates result in higher investment and consumption.

Once fiscal consolidation is complete, all regions benefit from lower long-run interest rates. As a result of fiscal adjustment in Japan, world real interest rates are lower, pulling up investment and consumption in all regions.

B. Growth Strategy

The simulations also capture the impact of increasing productivity and enhanced competition in labor markets, in line with the authorities’ growth strategy. The growth-strategy scenario assumes that trend growth will gradually increase by about 1 percentage point over a 10 year period, owing to productivity increases in both tradable and nontradable sectors, as well as reductions in labor market mark-ups by 2 percentage points (Chart 5). The government’s growth strategy sets a target of 2 percent real growth for the coming decade, focusing on key sectors, such as environment, health, Asian integration, and tourism

A broad-based productivity increase will reduce the trade balance in Japan. With higher productivity in both tradable and nontradable sector, consumption and investment start to improve even in the short-run, and this together with the associated appreciation of the currency leads to higher imports than otherwise.

In the absence of fiscal consolidation, spillovers to other economies from the growth strategy are small relative to the impact on the Japan, with the net gains depending on the relative importance of trade and real interest rate effects.

  • In the short run, all regions benefit through trade linkages with higher demand from Japan. However, the short-term impact on real GDP depends on the monetary policy reaction in the United States and the Euro Area. If monetary policy does not react to higher inflation in other regions, real GDP increases along with the higher trade balance.

  • Over the medium- to long-run, all regions benefit from higher productivity and lower real interest rates, but again the benefits are relatively small, particularly in the absence of fiscal consolidation. While the benefits to Emerging Asia accrue over the shorter horizons through trade linkages, the benefits to other regions accrue over longer horizon.

C. Combined Policy Package of Fiscal Consolidation and Growth Strategy

The short-term negative effects from fiscal consolidation would be mitigated by structural reforms. Overall, structural reforms would help Japan’s GDP increase gradually, limiting the decline in imports and rise in current account surplus under the fiscal consolidation scenario, thereby reducing the negative trade spillovers to the rest of the world. The short-run growth spillovers from structural reforms in Japan would depend on exchange rate and monetary policy responses, but negative spillovers to emerging Asia could be significantly dampened through the trade channel. Other regions would benefit from lower real interest rates caused by a credible fiscal consolidation plan and higher productivity, particularly over the medium-run.25

Figure 1.Fiscal Consolidation and Growth Strategy

Chapter IX. Debt Sustainability, Borrowing Costs, and the Impact of a Fiscal Crisis26

A. Context

The earthquake has interrupted Japan’s nascent recovery, and has placed greater attention on the dynamics of Japan’s public debt. Large fiscal deficits and sluggish activity have pushed public debt to unprecedented levels, leaving the government’s financing requirements at around 50 percent of GDP, almost twice that of the United States. The fact that JGB yields remain at historic lows suggests that creditors are confident about the authorities’ ability to come up with a sound stabilization plan. To stabilize debt and place it firmly on a downward path, staff have recommended a 10 percent of GDP improvement of the structural primary balance over the next 10 years.

Japan’s debt-sustainability projections are particularly sensitive to assumptions on the future path of interest rates. For example, with a 200 bps increase in borrowing costs, even the staff’s recommended adjustment effort would fail to stabilize the public debt ratio. This raises a concern that, even in the event of a modest hike in borrowing costs, the implied fiscal burden of adjustment may be increasingly perceived as infeasible. And as events in Europe over the past year have demonstrated, once confidence in fiscal sustainability erodes, the authorities may rapidly face an adverse feedback loop between rising yields, a deteriorating fiscal situation, falling market confidence, a more vulnerable financial system, and a contracting real economy.

Figure 1.Japan: Net Public Debt

(percent of GDP)

B. Fiscal Crisis Scenarios

Staff have simulated a range of fiscal crisis scenarios originating in Japan, featuring different assumptions regarding the impact of the crisis on worldwide market confidence. These results are derived from a refined version of the structural macroeconometric model of the world economy documented in Vitek (2010), which features extensive linkages between the real and financial sectors, both within and across G20 economies.27

  • The first scenario features a fiscal crisis that is contained within Japan. A sudden loss of confidence in fiscal sustainability is represented by a positive term-premium shock, which raises the long term nominal interest rate by 450 basis points on impact (bringing Japanese yields in line with other similarly rated sovereigns). Heightened risk aversion also hits the stock market, represented by an equity risk-premium shock which reduces equity prices by 60 percent on impact (similar to stock-market declines in many other financial crises). In addition, households and firms postpone their consumption and investment, owing to reduced confidence, decreasing domestic demand by 1 percent, while a fiscal consolidation reduces it by a further 2 percent. Finally, there is a run on the yen, represented by an exchange-rate risk premium shock which results in a 30 percent nominal effective depreciation on impact. Overall, this fiscal crisis is estimated to generate a weighted-average peak output loss of 4.4 percent in Japan, 0.1 percent in peripheral European countries, 0.2 percent in other advanced economies, and 0.4 percent in emerging economies.

  • Under the second and third scenarios, heightened risk aversion in Japan spreads progressively to bond and stock markets in the European periphery and emerging markets. Although Japanese financial markets are relatively isolated, during periods of uncertainty global financial markets face elevated risks of falling market confidence and herd behavior. In this context, the scenarios augment the first with additional shocks to foreign long term nominal interest rates, equities, exchange rates, and demand (Table 1). Depending on the extent of spillovers to market confidence, the costs to other countries can reach as high as 3 percent of GDP.

Table 1Output Losses arising from a Fiscal Crisis
JapanPeripheral EuropeEmerging MarketsOther AM
Scenario 1LT Interest Rates450 bps------
Equity Prices60 percent drop------
Exchange Rate30 percent------
Private Demand1 percent of GDP------
Fiscal Consol.2 percent of GDP------
Peak GDP Loss4.40.10.40.2
Scenario 2LT Interest Rates450 bps300 bps----
Equity Prices60 percent drop40 percent drop----
Exchange Rate30 percent15 percent----
Private Demand1 percent of GDP1 percent of GDP----
Fiscal Consol.2 percent of GDP1 percent of GDP----
Peak GDP Loss4.42.90.50.3
Scenario 3LT Interest Rates450 bps300 bps450 bps
Equity Prices60 percent drop40 percent drop60 percent drop
Exchange Rate30 percent15 percent30 percent
Private Demand1 percent of GDP1 percent of GDP1 percent of GDP--
Fiscal Consol.2 percent of GDP1 percent of GDP----
Peak GDP Loss4.83.23.10.6

Chapter X. The Impact of a Fiscal Crisis on the Region: Financial-Sector Spillovers28

Most JGBs are held by Japanese financial institutions. This suggests that a shock to JGB yields might have a direct spillover to other markets, by impacting Japan’s financial-sector balance sheets and prompting a withdrawal by Japanese financial firms from foreign markets. Banks and insurance companies combined account for almost 90 percent of the financial sector’s JGB holdings. They also account for the majority of the financial sector’s foreign loans and investments. Therefore, this chapter looks at the impact of a JGB shock on local banks and insurance companies, focusing in particular their financial soundness and potential spillovers to Japan’s regional neighbors.

Table 1Share of Financial Institutions in JGB holdings and Foreign Securities and Loans
JGBForeign Securities and Foreign loans (%)
Banks60.844.5
Insurance25.323.4
Pension Funds4.710.8
Investment Trusts1.919.6
Others7.21.7
Total100.0100.0
Source: Bank of Japan, “Flow of Funds”.Note: Figures are the ratio of each financial institution’s JGB holdings, foreign securities and loans to total financial institutions excluding the central bank at the end of 2010. JGB is the sum of treasury discount bills, government securities and Fiscal Investment and Loan Program (FILP) bonds. All foreign loans are assumed to be conducted by banks.
Source: Bank of Japan, “Flow of Funds”.Note: Figures are the ratio of each financial institution’s JGB holdings, foreign securities and loans to total financial institutions excluding the central bank at the end of 2010. JGB is the sum of treasury discount bills, government securities and Fiscal Investment and Loan Program (FILP) bonds. All foreign loans are assumed to be conducted by banks.

The effect of a JGB shock on the Japanese banking sector.

Japanese banks have strengthened their capital recently, but still hold a large amount of JGBs and Japanese equities. They have also recently started to increase loans to foreign countries, especially Asian economies. Megabanks in particular—MUFG, SMFG, and Mizuho FG—account for most of the banking sector’s JGB and equity holdings, and are also responsible for most foreign loans. We therefore stress test the balance sheets of the three megabanks to gauge the effect of a JGB shock on bank stability and foreign lending.

We consider five stress scenarios: a 100bps, 200bps, 300bps, and 400bps parallel shift in the yield curve, as well as the fiscal-crisis scenario considered in Chapter IX. The Q-JEM (Quarterly Japanese Economic Model)29 is used to estimate the follow-on impact of an interest-rate shock on stock prices and GDP growth in the first four scenarios. For the fiscal-crisis scenario, the assumed shock entails a 450bps increase in interest rates, a 60 percent decline in equity prices, and 4.4 percent decline in growth.

The stress test estimates the impact on profits, tier l capital, and foreign loans. The first step provides an estimate of the immediate losses on banks’ JGB and equity holdings, as well as expected losses in the loan portfolio.30 The second step calculates the resulting Tier I capital ratio, assuming that risk-weighted assets remain unchanged. Finally, the drop in foreign lending is estimated by assuming that megabanks maintain the targeted Tier I ratio by reducing their foreign loans.31

Table 2TOPIX and credit cost rate in each scenario
Scenario iScenario iiScenario iiiScenario ivScenario v
-100bps--200bps--300bps--400bps--Crisis-
TOPIX828722629549392
Credit cost rate50 bps60 bps72 bps85 bps101 bps
Source: Fund staff estimates.
Source: Fund staff estimates.

The results show that the megabanks are resilient to JGB shocks. Only in the extreme scenario do they reduce their foreign portfolios significantly. In the first three scenarios their Tier I ratios remain above 8 percent. In the fourth 400bps scenario, the average Tier I ratio drops to 7.8%, prompting a slight scaling back of foreign loans to bring the ratio back up to 8 percent. In the crisis scenario, the Tier I ratio drops to 6.7 percent. Although this meets the minimum requirement level of 6 percent, in order to bring capital back to 8 percent, the banks would have to reduce foreign loans by 45 percent.

Table 3Estimation Results
Scenario iScenario iiScenario iiiScenario ivScenario v
-100bps--200bps--300bps--400bps--Crisis-
Tier 1 (%)12.210.69.27.86.7
Rate of reduction in foreign loans----7.2-45.0
Source: Fund staff estimates.Note: Tier 1 ratio when Tier 1 capital decreases and risk-weighted assets remain unchanged.
Source: Fund staff estimates.Note: Tier 1 ratio when Tier 1 capital decreases and risk-weighted assets remain unchanged.
Table 4Impact due to withdrawal of 45 percent of inter-bank funding from Japan
Affected countriesEffect on capital (in percent of pre-shock capital)
Australia-8.6
China-1.3
Taiwan-1.1
India-1.4
Indonesia-5.2
Malaysia-2.0
Philippines-0.8
South Korea-13.6
Thailand-5.5
Vietnam-2.2

Even under the most severe scenario, the regional impact of a reduction in foreign loans is limited. Assuming that the banks reduce their foreign loans in proportion to their share of loans to each jurisdiction, the impact on local banking systems is relatively minor, ranging from 0-2 percent expressed as a fraction of total domestic credit. The key exceptions are the offshore financial centers, Hong Kong SAR and Singapore, where the impact ranges from 3-6 percent. As these centers are effectively cross-border intermediaries, the effect on the local economy will likely be limited.

Interbank network analysis confirms that a withdrawal of Japanese funding-would not be severe enough to trigger systemic distress in other countries.32 In the event that regional banking systems experience a 45 percent withdrawal their funding from Japan,33 the most vulnerable country is South Korea, which could experience a 14 percent decline in pre-shock bank capital. Other exposed countries are Australia, and to a limited extent, Indonesia and Thailand. In no case would a JGB-initiated shock push any of the regional banking systems to failure.

The presence of Japanese investors in regional capital markets is limited, so a withdrawal of Japanese banks will have a relatively small impact. Japan’s equity holdings, as a fraction of local market capitalization, are significant in both the United States and Europe—ranging from 3-6 percent—but are much smaller in Asia (0-2 percent). A similar pattern applies to Japan’s debt holdings.

The effect of JGB shock on the insurance sector

Japanese insurance companies have substantial financial buffers. The average solvency margin of the major firms is 932 percent, well above the minimum requirement of 200 percent, and representing a sizable cushion against JGB shocks. Applying same stress scenarios, the result shows that even in the most severe scenario, the insurance companies’ solvency margin ratio remains above 300 percent. Therefore, JGB shocks would not force them to reallocate their financial assets or liquidate their foreign investment positions.

Table 5Results of stress test on major insurance companies
Scenario iScenario iiScenario iiiScenario ivScenario v (%)
-100bps--200bps--300bps--400bps--Crisis-
Solvency margin ratio871.7696.9532.8395/6341.9
Source: Fund staff estimates.Note: Major insurance companies are Dai-ichi Life Insurance, Meiji Yasuda Insurance, MS&AD Insurance, Nippon Life Insurance, NKSJ, and Tokyo Marine. In estimating the effect of interest rates on the market value of JGBs, it is assumed that the average maturity of JGBs is 10 years and the average coupon rate is 1.2 per cent.
Source: Fund staff estimates.Note: Major insurance companies are Dai-ichi Life Insurance, Meiji Yasuda Insurance, MS&AD Insurance, Nippon Life Insurance, NKSJ, and Tokyo Marine. In estimating the effect of interest rates on the market value of JGBs, it is assumed that the average maturity of JGBs is 10 years and the average coupon rate is 1.2 per cent.

Chapter XI. Capital Market Contagion and Extreme Tail Dependence34

The size of the public debt and the potential cost of post-earthquake reconstruction raise the issue of whether investors will continue to enthusiastically buy Japanese government bonds (JGBs). While it is difficult to estimate the risk of a sudden withdrawal, past episodes of large changes in JGB yields could shed light on the potential spillover effects to G7 financial markets.

This chapter assesses the cross-border spillovers of large changes in the 10-year JGB yields and equity returns. Key stylized facts about the correlations between excessive equity returns and bond-yields include:35

  • Large negative equity returns and large increases in bond yields are not necessarily correlated within major countries, such as the U.S., Germany, and Japan. This means that changes in these two market types are driven by separate events. However, greater correlations are found within some smaller countries, such as Greece and Ireland.

  • Large negative shocks in equity returns are significantly correlated across countries, but vary in degree. For instance, correlations are <0.30 with Japan, >0.6 within Europe, and around 0.5 between U.S.-Canada and U.S.-U.K.

  • Large positive changes in 10-year bond yields are not very correlated across countries. In fact, there are no coincident yield-spikes in Japan and the major countries at the 99th percentile tail. If the threshold is lowered to 95th percentile (where 10-year yield change>18bps counts as exceedance), then there are small but significant correlations with European countries. The correlations lie between 0.2-0.3 between U.S.-France and U.S.-Germany but are higher within Europe.

To control for common shocks, an extreme value theory (EVT) approach is used to examine the conditional-correlation between extreme returns in Japanese financial markets and those elsewhere. This is done by means of logit-regressions, in which exceedances in (say) the United States are regressed on exceedances elsewhere (including in Japan), controlling for common factors such as global equity returns, changes in global risk aversion (denoted by changes in the VIX), and extreme events in the euro-area periphery. The regressions also include lagged dependent variables to absorb other country-specific effects.

Four sets of results are presented in Tables 1-4. Spillovers from extreme increases in 10-year bond yields to similar bonds in other countries (-bond-to-bond”) are presented in Table 1. The matrix shows the association (+/-, significance) between extreme increases in bond yields in -trigger” countries (rows) to the -affected” countries (columns). The effects are in terms of the probability of experiencing extreme movements in the affected countries, given an extreme movement in the trigger. Table 2 shows the association between the affected countries’ probability of experiencing extremely negative equity returns conditional on the event that one of the triggers is also experiencing extremely negative returns (-equity-to-equity”). Tables 3 and 4 present results for cross-asset correlations across countries.

Table 1.Bond-to-bond: Effect of large increases in ten-year bond yields in the trigger countries on the probability of the same in recipient
TriggerFRAGERJPAUKUS
FRA---
GER++++
JPA++++
UK-+-+
US++++
Other
VIX 1/-+---
MSCI wrld 1/-++-+
Exc GIPS 2/+--+ 3/+
Table 2.Equity-to-equity: Effect of large negative equity returns in the trigger countries on the probability of the same in recipient
TriggerFRAGERJPAUKUS
FRA+-+-
GER+-+-
JPA+--+
UK+--+
US+--+
Other
VIX 1/---++
MSCI wrld 1/-----
Exc GIPS 2/+-+--
Table 3.Bond-to-equity: Effect of large changes in ten-year bond yields in the trigger countries on the probability of large changes in equity returns in recipient
TriggerFRAGERJPAUKUS
FRA-+++
GER+---
JPA++++
UK+--+
US--+-
Other
Own yield -
exceedance++-+-
VIX 1/+-+++
MSCI wrld 1/-----
Exc GIPS 2/-----
Table 4.Equity-to-bond: Effect of large negative equity returns in the trigger countries on the probability of large incrases in yields in recipient
TriggerFRAGERJPAUKUS
FRA-++-
GER++--
JPA+++-
UK-+-+
US--+-
Other
Own equity return
exceedance-+-++
VIX 1/++--+
MSCI wrld 1/+++++
Exc GIPS 2/-+-++
Note: The shadings denote statistical significance at different levels:1 percent5 percent10 percent

5-day percent change

Sum of exceedances in Greece, Portugal, Spain, Ireland. For instance, if all 4 are in exceedance at the same time, it takes a value of 4.

Exceedance in Irish bond yields

Note: The shadings denote statistical significance at different levels:1 percent5 percent10 percent

5-day percent change

Sum of exceedances in Greece, Portugal, Spain, Ireland. For instance, if all 4 are in exceedance at the same time, it takes a value of 4.

Exceedance in Irish bond yields

The results can be summarized as follows:

  • Extremely large increases in 10-year JGB bond yields are associated with a higher likelihood of a hike in France’s 10-year bond yields and, to a much smaller extent, the United States (Table 1). The equity returns of European stock indices are more likely to be adversely affected than their 10-year bond yields, especially in France and Germany (Table 2).

  • Large negative equity returns in the Japanese stock index tend to be correlated with large negative returns in the United States, and to a much smaller extent, France (Table 3). There is almost no impact of large negative equity returns on large increases in bond yields in advanced countries (Table 4).

  • When global equity markets (MSCI world) are doing better, chances of large negative equity returns fall for all countries, including Japan. There are some cases in which the chances of a large change in yields increase when the world equity markets are doing better, but the evidence on that is thin (mainly on Germany and the United States).

Annex 1. Methodology used in the extreme-value analysis

Step 1: We use daily (5-day week) equity returns and week-on-week changes in 10-year bond yields for France, Germany, Italy, Japan, U.K., and the United States. Other variables, used as controls, include the VIX, MSCI-World equity returns, and the equity returns and bond yields for Spain, Portugal, Greece and Ireland. The extreme value threshold for the data is calculated by the 5th percentile tail.

Step 2: All observations exceeding the threshold are assigned a value of 1; others are 0. These series are called exceedances.

Step 3: The spillover analysis is done by estimating the probability that the equity return of a country is in exceedance (takes the value of 1), conditional on Japan and other countries being in exceedance. This coexceedance is estimated from a logistic regression with an exceedance as the dependent variable (say country X’s exceedance) and other exceedances (countries Y1, Y2…) and a lagged-dependent variable as explanatory variables. The weekly changes in MSCI-World and VIX, and the sum of exceedances in Greece, Ireland, Portugal and Spain (0-4 variable), are added as controls for common factors. A significant positive coefficient on any right-hand side variable is interpreted as: the probability of X being in exceedance is higher if Y1 is in exceedance (positive coefficient), but not so if Y2 (non-significant coefficient) is in exceedance (for instance). Then, country X and Y1 are said to coexceed with each other. Four sets of regressions are estimated for each of the large advanced countries. The first one looks at coexceedances in bond-yields; the second, coexceedances in equity-returns; the third, coexceedances from bond-yields to equity returns; and, the fourth, from equity returns to bond-yields.

Figure 1.Histogram of Stacked Equity Returns

(week-on-week change in equity price index, %) Threshold (5th percentile left tail) = -5.3%

Figure 2.Histogram of Stacked 10-year Bond-yield changes

(week-on-week basis points) Upper (95th percentile) =18bps

Chapter XII. Assessing Distress Dependence Among Sovereigns36

The behavior of CDS spreads over the past few years suggests that financial-market spillovers between sovereigns tend to be elevated during times of market distress. This distress dependence might be due to fundamental factors, such as trade or capital-flow linkages. It may reflect the cross-border activities of globally-active financial institutions. Or, it may instead result from psychological factors, such as herding behavior, or a global shift in risk aversion.

As a measure of these potential spillovers, it is possible to compute the probability of sovereign distress in one country given default in another.

The probability of sovereign distress in country A given a default by country B—P(A|B)—is obtained in three steps:

  • The marginal probabilities of distress (PoD) for countries A and B, P(A) and P(B) respectively, are extracted from the individual-country CDS spreads, using data from Bloomberg.

  • The joint probability of distress (JPoD) of A and B, P(A∩B), is obtained using the methodology developed by Segoviano.37 This is a non-parametric approach that estimates the JPoD without imposing a (pre-determined) distributional form; subject only to the constraint that the implied PoD for each country is the same as that extracted from market data. This differs from traditional approaches, in which parametric copulas have to be chosen and calibrated explicitly—usually a difficult task.

  • Finally, the conditional probability of distress (CoPod) P(A|B) is obtained from:

The joint distribution is estimated for each date, providing a series of time-varying probability estimates for each country pair. Such pair-wise estimates provide insights into market views concerning the potential for confidence spillovers from one country to another. In the case of Japan, overall CoPods tend to be relatively low, but are elevated during times of global turmoil. Moreover, they appear to be currently elevated for some European countries with perceived fiscal vulnerabilities, as well as for some European banks.

Figure 1.Japan: Conditional Probability of Distress, Given Distress in Japan

Sources: Bloomberg, Datastream, IMF Staff calculations.

Chapter XIII. Monetary Policy Spillovers38

The Local and International Impact of Large-Scale Asset Purchases

The Bank of Japan has pursued powerful monetary easing since Dec 2009, introducing a new fixed-rate provision of funds, an asset purchase program take involves government and corporate bonds, and equity exchange-traded funds and real estate trusts. It has also introduced new growth-focused lending.

Event-study analysis suggests that the easing has had a modest impact on local sovereign yields and equity returns (Table 1), but the impact on corporate credit and economic activity is not yet clear.39 The impact on the exchange rate, or on foreign markets, has not been significant (Tables 2 and 3).

Table 1Impact of Bank of Japan’s Large-Scale Asset Purchases on Japanese Financial markets
(in basis points, unless stated otherwise)
DateEventsGovernment BondsShort-term interest ratesTerm premium (yield curve)Inflationary expectationsExchange rate JPY/USDCorporate yieldsEquity MarketRisk Premium
1yr JGB2-year JGB10-year JGB1-year futures3-month futuresshort endlong end5-year break-evenSpot Rate3-month Forward ratesAA-ratedBBB-ratedIndex FuturesNikkei FuturesImplied VolatilityJ-REITsCorporate Spreads
19-Dec-08Liquidity and Financial Stability Measures-7.3**-7.7**-7.3**-2.6-7.0**0.90.40.000.930.84-0.06**-0.010.650.92-7.05**3.91*5.9*
Powerful Monetary Easing (PME)
1-Dec-09Enhancement of Easy Monetary Conditions-4.3**-4.5**-5.8**-3.80-6.5**-0.9-1.3-0.031.121.12-0.0602.823.22-2.535.74**5.8*
17-Mar-10Expansion of measures to encourage-0.100.502.703.401.50-0.12.20.030.080.070.0200.210.19-1.451.19-2.7
decline of long-term rate
30-Aug-10Enhancement of Easy Monetary Conditions-0.30-1.50-3.90-3.50-0.50-1.1-2.4-0.02-1.20-1.2-0.04-0.04-1.86-1.93.21.17-0.6
5-Oct-10Comprehensive Monetary Easing-1.10-0.90-10**-5.4*0.00-0.4-9.1**-0.01-0.50-0.5-0.09**-0.09**3.313.74-1.172.251.2
Powerful Monetary Easing (PME)
Cumulative Sum-13.1**-14.1**-24.3**-11.9*-12.5**-1.6-10.2*-0.030.430.33-0.23**-0.14*5.136.17-9.0014.26**9.6
Average-2.62**-2.82-4.86*-2.38-2.50-0.3-2.0-0.010.090.07-0.05-0.0281.0261.234-1.82.8521.92
VariousIntroduction of new measures/facilities-13.1**-15.9**-21.8**-12.6*-11.5**-2.2*-5.9**-0.031.291.2-0.2**-0.16.017.11-14.19**18.5**11.9*
(See Table 2)
VariousExpansion of selected easing measures-1.7-7.5-6.8-6.27.0*-0.30.70.04-2.41-2.40-0.06-0.09-1-1.53-4.411.74-2.2
(See Table 2)
VariousExits of selected measures/facilities0.2-2.7-2.1-2.81.00.10.60.03-0.17-0.17-0.01000.35-1.94.062.1
(See Table 2)
Control Groups
Jul 08 - Dec 10Typical Trading Day
Average-0.2-0.2-0.1-0.1-0.10.00.10.01-0.07-0.070.000.00-0.05-0.04-0.010.00-0.09
s.d.1.32.03.53.62.11.33.30.071.161.170.030.042.953.123.353.023.72
MPC releasesMPC meeting release (excl. monetary
Jul 08 - Dec 10easing announcements)
Average-0.5-0.9-1.1-0.7-0.3-0.6-0.20.01-0.12-0.12-0.010.00-0.89-0.901.41-0.580.74
s.d.2.03.23.44.72.42.44.70.081.281.310.030.054.214.684.043.974.37
Source: Bank of Japan, Bloomberg, IMF staff calculations
Source: Bank of Japan, Bloomberg, IMF staff calculations
Table 2Impact of Bank of Japan’s Large-Scale Asset Purchases on US Financial markets
(In basis points, unless stated oteherwise)
DateEventsGovernment BondsShort-term interest ratesTerm premium (yield curve)Inflationary expectationsNominal Trade-weighted Exchange RateCorporate yieldsEquity MarketRisk Premium
1yr JGB2-year JGB10-year JGB1-year futures3-month futuresshort endlong end5-year break-evenSpot Rate3-month Forward ratesAA-ratedBBB-ratedStock IndexIndex FuturesImplied VolatilityCorporate Spreads
19-Dec-08Liquidity and Financial Stability Measures-3.015.3**9.431.4**1.58.00*-5.90-0.14*1.25**1.20**0.03-0.02-1.54-2.69-5.87-9.40
Powerful Monetary Easing (PME)
1-Dec-09Enhancement of Easy Monetary Conditions1.07.213.68.2-2.0-0.106.400.01-0.45-0.440.070.071.241.20-13.83*-13.60
17-Mar-10Expansion of measures to encourage0.04.42.92.93.00.40-1.50-0.010.170.16-0.01-0.010.550.57-6.05-2.90
decline of long-term rate
30-Aug-10Enhancement of Easy Monetary Conditions-2.0-6.8-17.9*-13.4*-0.8-2.20-11.1*-0.090.340.34-0.14*-0.14-1.43-1.456.5417.9*
5-Oct-10Comprehensive Monetary Easing-2.0-2.2-8.2-6.3-3.0-1.80-6.000.09-0.9*-0.88*-0.06-0.062.021.84-8.678.20
Powerful Monetary Easing (PME)
Cumulative Sum-6.017.9-0.222.8-1.34.30-18.1-0.140.410.38-0.11-0.160.84-0.53-27.88*0.2
Average-1.23.60.04.6-0.30.86-3.62-0.030.080.076-0.022-0.030.168-0.106-5.5760.04
VariousIntroduction of new measures/facilities0.018.713.141.4**4.55.20-5.600.100.260.220.180.090.36-1.80-17.05-13.10
(See Table 2)
VariousExpansion of selected easing measures-1.03.91.04.1-6.32.80-2.900.27-1.21-1.190.070.021.671.90-3.02-1.00
(See Table 2)
VariousExits of selected measures/facilities0.02.45.44.4-3.0-0.703.000.11-1.12-1.09-0.01-0.050.670.6730.49**-5.40
(See Table 2)
Control Groups
Jul 08 - Dec 10Typical Trading Day
Average-0.6-0.6-0.1-0.2-0.80.100.500.000.020.020.000.000.030.030.330.11
s.d.6.58.611.99.88.55.718.450.100.650.640.100.112.612.669.8211.93
MPC releasesMPC meeting release (excl. monetary
Jul 08 - Dec 10easing announcements)
Average-0.40.91.11.0-0.70.990.180.02-0.07-0.070.010.020.16-0.03-1.10-1.11
s.d.13.211.411.711.311.810.528.580.110.680.660.100.123.773.849.9511.74
Source: Bank of Japan, Bloomberg, IMF staff calculations
Source: Bank of Japan, Bloomberg, IMF staff calculations
Table 3Impact of Bank of Japan’s Large-Scale Asset Purchases on Euro-Area Financial markets
(in basis points, unless stated otherwise)
DateEventsGovernment BondsShort-term interest ratesTerm premium (yield curve)Inflationary expectationsExchange rate EUR/USDCorporate yieldsEquity MarketRisk Premium
1yr JGB2-year JGB10-year JGB1-year futures3-month futuresshort endlong end5-year break-evenSpot Rate3-month Forward ratesAA-ratedBBB-ratedStock IndexIndex FuturesImplied VolatilityCorporate Spreads
19-Dec-08Liquidity and Financial Stability Measures-5.3-3.7-1.2-6.1-8.58.02.50.08-2.14**-2.23**-3.60-2.90-2.47-1.70-10.96*-1.7
Powerful Monetary Easing (PME)
1-Dec-09Enhancement of Easy Monetary Conditions-1.8-1.51.2-1.3-1.5-0.12.7-0.060.270.28-7.101.202.772.49-10.96*0.0
17-Mar-10Expansion of measures to encourage0.4-1.4-1.1-1.5-0.50.40.30.05-1.13-1.12-1.30-3.000.690.69-1.82-1.9
decline of long-term rate
30-Aug-10Enhancement of Easy Monetary Conditions-2.5-4.5-3.8-6.7-0.5-2.20.7-0.14*-0.66-0.62-2.80-2.00-0.44-0.762.461.8
5-Oct-10Comprehensive Monetary Easing-1.9-2.8-1.6-1.7-2.0-1.81.20.011.78*1.66*-4.20-5.302.232.03-9.45-3.7
Powerful Monetary Easing (PME)
Cumulative Sum-11.1-13.9-6.5-17.3-13.04.37.4-0.06-1.88-2.03-19.00-12.002.782.75-30.73**-5.5
Average-2.2-2.8-1.3-3.5-2.60.91.5-0.01-0.38-0.41-3.80-2.400.560.55-6.15-1.1
VariousIntroduction of new measures/facilities-9.8-9.925.9*6.8-13.55.235.8**0.25*-0.79-0.68-12.1-33.9**0.961.47-37.34**-59.8**
(See Table 2)
VariousExpansion of selected easing measures-12.6-5.128.9*12.3-10.52.834*0.163.243.42-0.50-2.60-0.310.38-14.36-31.5**
(See Table 2)
VariousExits of selected measures/facilities-4.2-6.0-8.3-1.6-4.5-0.7-2.3-0.100.950.98-2.30-7.600.870.966.380.7
(See Table 2)
Control Groups
Jul 08 - Dec 10Typical Trading Day
Average-0.6-0.6-0.1-0.2-0.80.10.50.000.020.020.000.000.030.030.330.1
s.d.6.58.611.99.88.55.78.50.100.650.640.100.112.612.669.8211.9
MPC releasesMPC meeting release (excl. monetary
Jul 08 - Dec 10easing announcements)
Average-0.40.91.11.0-0.71.00.20.02-0.07-0.070.010.020.16-0.03-1.10-1.1
s.d.13.211.411.711.311.810.58.60.110.680.660.100.123.773.849.9511.7
Source: Bank of Japan, Bloomberg, IMF staff calculations
Source: Bank of Japan, Bloomberg, IMF staff calculations

The Impact of Further Easing: Model Simulations

Spillovers from additional easing will likely be limited. Simulations are based on a refined version of the structural G-20 model.40 The scenario investigates a larger easing effort, and is calibrated by scaling up the above event-study results, assuming that the authorities increase their asset purchases up to the current allowable limit. The result is a sequence of term premium shocks that reduce the long term nominal interest rate by around 50 basis points in Japan, and by 1-15 basis points in the rest of the world. Equity markets in Japan also rise by over 10 percent, matched by increases of 2-6 percent elsewhere.41 Monetary policy is constrained by the zero lower bound on the short term nominal interest rate in the Euro Area, Japan, the United Kingdom, and the United States.

Figure 1.Peak Impulse Response to Japanese LSAP (percent of output)

Rising Interest Differentials and the Return of the Carry Trade

Prior to the global financial crisis, persistently low interest rates and historically low volatility made the yen a favored funding currency for carry trades.42 Moreover, the strong appetite for risk that characterized 2003-07 led to a steady build up in these positions, and rendered the carry trade a significant driver of cross-currency positioning.

Quantifying the size and destination of these positions is challenging. The range of instruments associated with the carry trade has grown over the years, including complex off-balance sheet transactions that are less easily detected in BoP and capital-flow statistics. The trade has also come to encompass a range of different investor classes, from -Mrs. Watanabe,” to more sophisticated global brokerage houses and hedge funds. In 2007, near the peak, estimates of the yen-funded carry trade ranged from $100 billion to $2 trillion.

Currently, the prospects for a return of the carry trade do not seem strong. Compared to the precrisis period, forward-looking measures of risk-adjusted yields are relatively low for the Australian, New Zealand, and U.S. currencies; reflecting narrow interest-rate differentials against the yen and a higher level of implied volatility. Notably, only the Brazilian real appears to be climbing towards the elevated levels of 2008, and market contacts indicate that Brazilian assets have become a favorite destination for Japanese retail investment trust accounts.

Position and leverage indicators suggests less capacity to hold yen-funded carry trades. On the Chicago Mercantile Exchange, noncommercial traders are currently holding net long positions in the yen, whereas net short positions were the norm prior to the crisis. Leverage indicators, such as the call-money liabilities of foreign banks in Japan, also suggest a significant decline in the trade’s attractiveness. This is consistent with anecdotal reports that hedge funds and other speculative investors now find it more difficult building up leverage in the post-Lehman shock environment.

Forecasting the Carry Trade

Looking forward, monetary normalization in other advanced markets suggests that interest differentials will widen once more. Following the methodology of Shin (2009), we can take the net interoffice assets of foreign banks operating in Japan as an indicator of the scale of the yen-funded carry trade. Arguably this is a better guide than using foreign-bank liabilities in the Japanese cash market, as it excludes funds used to purchase Japanese securities.

These net interoffice assets can then be modeled as a function of international policy-rate differentials (JPN vs. average of AUS/US/EUR) and the VIX.

Figure 2.Following the trail of leverage bets

Table 4Determinants of the Japanese Carry Trade
Dependent Variable: Net

Interoffice Assets
OLS: ShinOLS: Full SampleOLS: Post-Lehman

controls
Dynamic

Specification 1/
Interest Rate Differential-37.281***-22.828***-37.499***-36.695***
(-10.27)(-8.09)(-10.74)(-3.56)
Post-Lehman interaction16.877
(-1.45)
VIX-3.599***-0.871*-3.439***-5.425***
(-7.44)(-2.36)(-7.44)(-4.79)
Post-Lehman interaction3.916***2.891**
(-5.63)(-3.00)
Post-Lehman Dummy-5.421
-0.21)
Constant-134.971***-125.937***-136.115***-15.15
(-7.79)(-8.66)(-8.18)(-1.94)
Lagged dependent Variable0.826***
(-19.86)
R-sq0.5880.3620.568
No. Obs.110146146146
Parentheses contain t-statistics. * p<0.05, ** p<0.01, *** p<0.001

Long-run coefficients reported.

Parentheses contain t-statistics. * p<0.05, ** p<0.01, *** p<0.001

Long-run coefficients reported.

Key results:

  • Extending Shin’s original regression to an updated dataset, the relationship with the VIX breaks down—again, this may reflect that fact that, after the Lehman shock, banks and hedge funds are more constrained in using their balance sheets for speculative purposes.

  • By including a post-Lehman interaction term on the VIX, however, we recover Shin’s original relationship. A similar interaction term on the interest differential, and a post-Lehman dummy, are both insignificant.

  • The residuals of the OLS specification suggest substantial autocorrelation. The preferred model, therefore, includes a lagged dependent variable. (The choice of specification, however, makes little material difference to the ultimate projection of net interoffice assets).

Projections

Assuming Japanese rates remain unchanged over the next two years, we can then use the expected increase in policy rates abroad (median forecast from Bloomberg) to project the likely increase in the scale of the carry trade out of Japan. To summarize, the average interest differential is expected to widen by about 220 bps by end 2012, prompting an increase in the carry trade of around ¥4.3 Trillion ($51 billion).

The next step is to map the change in the carry trade into actual exchange rates. A good price indicator is the JPY/AUD pair. Drawing from the recent relationship between this rate and net interoffice assets, staff project the (marginal) impact on the JPY/AUD of the anticipated increase in the carry trade. Projections are based on a simple VARX framework.

Overall, the carry-trade increase is expected to prompt a 5 percent depreciation of the yen against the AUD. The impact on other rates, such as the JPY/USD is uncertain. But as a general guide, 5 percent might be considered as an upper bound against other rates.

Figure 3.Japanese Carry Trade: Net Interoffice Assets of Foreign Banks in Japan, 1999-2012

Figure 4.Japan: Carry Trade Projections

Chapter XIV. The Transpacific Partnership Agreement—Impact on Japan and Other Members43

Background. The origin of the Trans-Pacific Partnership (TPP) is the 2006 Economic Partnership Agreement (EPA) between Brunei, Chile, New Zealand and Singapore. In 2010, an additional alliance between the United States, Australia, Malaysia, Peru, and Vietnam was announced and talks began about extending the TPP to these five countries. Japan, Korea, Thailand, Canada and Mexico have expressed interest in joining. Negotiations are progressing under U.S. leadership. The TPP is envisioned as a high-standard, 21st century trade agreement that includes commitments covering all aspects of trade and investment. The TPP is also seen as a starting point for a broader Free Trade Area of the Asia-Pacific (FTAAP). In June 2010, as part of its new growth strategy, Japan announced its intention to join TPP as a means of opening up the country and revitalizing agriculture.

Key proposals under the TPP. Trade liberalization would extend to all chapters of the Harmonized System; coverage is therefore expected to include agreements on agriculture, a potentially controversial sector for both Japan and the United States. The TPP also features a strong focus on services liberalization—an area of particular interest to U.S. service suppliers. At the Sixth Round of negotiations in April 2011 in Singapore, the United States tabled proposals related to labor rights, environmental protection, and intellectual property protection—potentially contentious issues in TPP talks. Ambitious demands by the United States in these three areas may induce TPP countries to demand more in market access to the United States or to give less. At the same time, the United States also tabled a legal text on regulatory coherence; the first time this issue has featured in a trade agreement.44 The plan is for TPP to be concluded during the APEC-leaders meeting hosted by the United States in November 2011.45

Potential impact of Japan’s membership. The Global Trade Analysis Project (GTAP) is used to assess the impact and benefits of TPP on Japan as well as other Asian countries.46 The analysis is static’ in the sense that it only captures the economic efficiency impact of a tariff removal; no allowance is made for more dynamic adjustments such as incorporating the impact of capital accumulation and productivity improvements, as in Kawasaki (2010).47 Following Wignaraja (2011), the analysis covers 10 countries.48 The analysis considers different scenarios regarding coverage (i.e., with and without agriculture liberalization) and membership (i.e., with and without Japan). The impact should be read as a one-time effect on GDP, exports, and utility (in the form of higher purchasing power as a result of tariff removal). The results are presented in Figure 1 (detailed data in Table 1).

Figure 1.Marginal Contribution of Japan’s Membership in TPP on:

Source: GTAP and author estimates.

Table 1Marginal Impact of Japan’s Membership in TPP
TPP with JPN - % change
All goods except Agriculture and processed

Food
All goods including Agriculture and

processed Food
Real GDPExport

volume
Utility per capita

from expenditures
Real GDPExport

volume
Utility per capita

from expenditures
AUS0.031.830.010.072.220.57
NZL0.0210.020.071.190.43
CHN-0.04-0.26-0.16-0.04-0.28-0.22
HKG0-0.05-0.110-0.09-0.13
JPN0.020.930.140.053.250.11
KOR0.083.10.290.363.840.39
TWN-0.02-0.1-0.22-0.03-0.11-0.27
KHM-0.11-0.2-0.85-0.11-0.19-0.86
IDN-0.02-0.1-0.11-0.02-0.18-0.19
LAO-0.020.15-0.12-0.030.75-0.06
MMR-0.01-0.06-0.050-0.15-0.14
MYS0.41.651.430.51.991.34
PHL-0.030.09-0.18-0.050.04-0.32
SGP-0.010.120.410.020.040.88
THA-0.090.91-0.61-0.111.18-0.91
VNM1.19.952.991.0211.685.72
IND-0.020.05-0.06-0.020.17-0.11
ROASOC-0.03-0.11-0.15-0.03-0.13-0.19
CAN-0.01-0.1-0.07-0.01-0.11-0.15
USA01.21001.630.04
MEX-0.050.04-0.11-0.050.15-0.17
CHL0.020.290.110.020.370.18
ROAmerica-0.020.03-0.05-0.020.09-0.09
PER-0.023.42-0.11-0.024.26-0.16
EU_25-0.010.05-0.03-0.010.1-0.05
RUS0-0.01-0.010.01-0.10
RestofWorld-0.010-0.03-0.01-0.01-0.04
Sources: GTAP and author estimates.
Sources: GTAP and author estimates.

Key results include:

  • TPP membership generates welfare gains for Japan and most other members, especially if agriculture is included. Potential losses for nonmembers are also higher when Japan is included due to higher trade diversion.

  • For Japan, the marginal contribution of TPP membership in terms of a one-time real GDP increase is somewhat modest—about 0.05 percent (including agriculture)—but is much higher for export volume (more than 3 percent).

  • For other members, projected welfare gains are highest for Vietnam and Malaysia (poorer members gain more from FTAs). The marginal contribution of Japan’s membership in terms of a one-time increase in real GDP ranges from 0.5 percent for Vietnam to 0.03 percent for Korea; for export volume the gains range from 2 percent for Vietnam to about 0.07 percent for Chile.

  • Our results at the aggregate level are qualitatively similar to Kawasaki (2010) but much smaller in magnitude. Again, the latter allows for the impact of dynamic aspects of capital formation and productivity improvements on economic outcomes.

In general, our results are consistent with the balance of existing literature on the impact of FTAs suggesting that: (i) FTAs in Asia generate welfare gains for members and modest losses for nonmembers (mainly through trade diversion); (ii) broadening the membership of the FTA generates more gains; (iii) results vary depending on the assumptions underlying liberalization and membership but the sign—gain or loss—is quite robust; and (iv) production of sectors with comparative advantage will increase under the FTA.

Preliminary considerations. Our empirical analysis suggests a positive welfare impact from membership, both for Japan and other Asian countries. Second, Korea (Japan’s key competitor in Asia) has been quite active on the FTA front and could have FTAs in place with the United States by the end of 2011 and with the European Union and China by 2012. Japanese export sectors may therefore find themselves at a competitive disadvantage if they find themselves facing a widening tariff gap. Finally, financial liberalization is a key feature of TPP. Japanese financial institutions could therefore benefit from improved access to rapidly growing emerging nations’ financial markets.49

Table 2TPP without JPN - % change
All goods except Agriculture and

processed Food
All goods including Agriculture and

processed Food
Real GDPExport

volume
Utility per capita

from expenditures
Real

GDP
Export

volume
Utility per capita

from expenditures
AUS0.011.190.040.021.30.17
NZL00.630.010.040.950.35
CHN-0.03-0.15-0.11-0.03-0.16-0.12
HKG0-0.05-0.060-0.05-0.08
JPN00.06-0.0300.12-0.03
KOR0.071.490.460.332.130.56
TWN-0.01-0.04-0.09-0.01-0.04-0.11
KHM-0.09-0.18-0.74-0.1-0.17-0.81
IDN-0.01-0.12-0.06-0.01-0.14-0.09
LAO-0.020.11-0.11-0.030.58-0.08
MMR0-0.04-0.040-0.03-0.07
MYS0.041.030.830.131.380.77
PHL-0.020.04-0.11-0.040.05-0.19
SGP-0.010.140.560.020.10.98
THA-0.040.36-0.27-0.040.48-0.37
VNM0.637.872.440.59.65.12
IND-0.010.03-0.05-0.010.08-0.07
ROASOC-0.02-0.14-0.11-0.01-0.12-0.12
CAN0-0.04-0.04-0.01-0.02-0.06
USA00.80.0200.980.02
MEX-0.010.06-0.05-0.010.06-0.06
CHL0.020.250.10.020.30.13
ROAmerica-0.01-0.01-0.03-0.010.01-0.05
PER-0.023.16-0.1-0.014.04-0.13
EU_25-0.010.02-0.02-0.010.04-0.02
RUS0-0.0100.01-0.040.01
RestofWorld0-0.02-0.020-0.02-0.01
Sources: GTAP and author estimates.
Sources: GTAP and author estimates.
Table 3Gains from Japan joining on members (% change) [TPP with JPN - TPP without JPN]
All goods except Agriculture and processed

Food
All goods including Agriculture and processed

Food
Real GDPExport

volume
Utility per capita

from expenditures
Real GDPExport

volume
Utility per capita from

expenditures
AUS0.020.64-0.030.050.920.40
NZL0.020.370.010.030.240.08
JPN0.020.870.170.053.130.14
KOR0.011.61-0.170.031.71-0.17
MYS0.360.620.60.370.610.57
SGP0.00-0.02-0.150.00-0.06-0.10
VNM0.472.080.550.522.080.60
USA0.000.41-0.020.000.650.02
CHL0.000.040.010.000.070.05
PER0.000.26-0.01-0.010.22-0.03
Sources: GTAP and author estimates.
Sources: GTAP and author estimates.

Prepared by Nagwa Riad (SPR) based on analysis in a Board paper on Changing Patterns in Global Trade (forthcoming).

The ESI measures the similarity of export patterns across pairs of countries, and takes a higher value for pairs with similar shares of each product category.

For each product category, the EXPY index notes the average income level of those countries producing the same product—capturing the fact that goods produced by industrialized countries will likely embody higher quality/value added. See Hausmann, and others (2007) -What You Export Matters” J. of Ec. Growth, Vol.12.

See Changing Patterns of Global Trade (forthcoming), for a detailed description of the methodology and additional results. Data on imports at the 6-digit level is used for the full set of 162 countries for data available in COMTRADE.

Prepared by Hitoshi Sasaki (SPR).

NIEs3 includes Korea, Taiwan, and Singapore, while ASEAN4 includes Indonesia, Malaysia, the Philippines, and Thailand.

Prepared by Hitoshi Sasaki (SPR).

Variables are transformed using a semi-log procedure of the form, x=sign(x)log(1+|x|), in order to include entries with recorded values of zero.

Prepared by Akira Otani (MCM) and Andrew Tiffin (SPR).

Prepared by Sergejs Saksonovs (SPR).

Not all variables are available for all countries and some missing data is interpolated from annual levels. Euro Area variables are defined as a GDP weighted aggregate of eight countries: Austria, Belgium, Finland, France, Germany, Italy, Netherlands and Spain.

See Pesaran, and others (2004), -Modeling Regional Interdependencies Using a Global Error-Correcting Macroeconometric Model”, J. of Bus. & Ec. Statistics 22(2) and Dees, and others, (2007) -Exploring the International Linkages of the Euro Area: A Global VAR Analysis”, J. of Applied Econometrics, 22.

Regional countries are Australia, China, India, Indonesia, Korea, Malaysia, New Zealand, Singapore, Philippines and Thailand.

Prepared by Srobona Mitra (MCM).

The analysis is based on BIS locational statistics as of September 2009. This allows for a broad sectoral breakdown and rich set of Asian countries.

Countries included in the analysis are Australia, Austria, Belgium, Canada, France, Germany, Ireland, Italy, Japan, Netherlands, Portugal, Spain, Sweden, Switzerland, United Kingdom, United States, China, Taiwan, India, Indonesia, Malaysia, Philippines, South Korea, Thailand, and Vietnam. Note that global exposures of banks in China, Indonesia, Philippines, Thailand and Vietnam are extracted from data on liabilities of the countries’ counterparties.

Marco Espinosa-Vega and Juan Solé, 2010, -Cross-border Financial Surveillance: A Network Perspective,” IMF Working Paper 10/105.

Prepared by Stephan Danninger (APD).

Balakrishnan, and others, 2009, World Economic Outlook, Spring 2009, Chapter 4; Fall 2008, Chapter 3.

Following the literature, an episode of financial stress is identified as a period when the FSI exceeds 1.5 standard deviations above its mean.

Prepared by Pelin Berkmen (APD).

The model covers five regions: Japan, the United States, Euro area, emerging Asia, and remaining countries. The calibration is slightly different from the version used for Japan 2010 Article IV, with updated monetary policy parameters and steady state debt.

This scenario assumes that interest rates will stay low for the initial two years.

While the decline in demand creates downward pressure on inflation, depreciation of the yen and the increase in consumption taxes pull it up.

The improvement in emerging Asia’s current account stems from import suppression originating from the decline in demand prompted by to higher real interest rates. Similar to the fiscal consolidation scenario, with flexible exchange rates, import suppression is much less, dampening the impact on emerging Asia’s trade balance.

Prepared by Andrew Tiffin and Francis Vitek (SPR), and Kiichi Tokuoka (APD).

Vitek, F., 2010, -Monetary Policy Analysis and Forecasting in the Group of Twenty: A Panel Unobserved Components Approach,” IMF Working Paper 10/152.

Prepared by Akira Otani and Mitra Srobona (MCM).

Tomiyuki, Kitamura, and others, -Hybrid Japanese Economic Model: Quarterly Japanese Economic Model (Q-JEM)”, Bank of Japan Working Paper 09-J-06, Bank of Japan, 2009 (only in Japanese).

The Bank of Japan’s default rate function—outlined in Financial System Report, April, 2007—is used to estimate GDP-related losses in the loan portfolio. In addition, the three mega banks’ core operating profits are assumed to be the same as 2010.

Megabanks are assumed to aim at an 8 percent Tier I ratio.

We repeat the network analysis used in Chapter VI to simulate a funding shock restricted within the region.

Again, local banks will need to need sell some of their assets at fire-sale prices, and we assume a discount of 50 percent.

Prepared by Srobona Mitra (MCM).

See Annex 1 for the definition and methodology for calculation of exceedances.

Prepared by Tola Oni and Andrew Tiffin (SPR).

Segoviano, M., 2006, -The Consistent Information Multivariate Density Optimizing Methodology,” Financial Markets Group, London School of Economics, Discussion Paper No. 557; Segoviano, M., and C. Goodhart, 2009, -Banking Stability Measures”, IMF Working Paper 09/04.

Prepared by Phil de Imus, Andrew Tiffin and Francis Vitek (SPR).

See R, Lam (2011), -Bank of Japan’s Monetary Easing Measures; Are they Comprehensive and Powerful?,” Japan: Selected Issues Paper, forthcoming.

Vitek, F., 2010, -Monetary Policy Analysis and Forecasting in the Group of Twenty: A Panel Unobserved Components Approach,” IMF Working Paper 10/152.

This is likely an upper bound, as a larger part of the impact of LSAP tends to be on announcement.

A -carry trade” exploits opportunities presented by of low borrowing costs in one market combined with higher returns in another. Its success as strategy has long been a puzzle for economists, given that it violates the hypothesis of uncovered interest parity—the so called forward premium puzzle.

Prepared by Nagwa Riad (SPR).

Regulatory coherence is aimed at making the regulatory systems of the TPP countries operate more seamlessly and addressing so-called =behind-the-border’ issues that pose increasing barriers to U.S. business in trying to access foreign markets. The intent would be to establish oversight regulatory bodies such as the U.S. Office of Information and Regulatory Affairs (OIRA).

The next round (seventh) of negotiations on TPP will be held during the week of June 20, 2011, in Vietnam.

The latest version of GTAP is used (version 7); the base year is 2004.

Kawasaki, Kenichi (2010), -The Macro and Sectoral Significance of an FTAAP” ESRI Discussion Paper Series #244.

Wignaraja, Ganeshan (2011), ASEAN or TPP? Pathways Towards East Asian FTA Consolidation, presentation at the Fund in February 2011. His analysis assumes a TPP-11 which includes: Australia, Brunei, Chile, Korea, Malaysia, New Zealand, Peru, Singapore, the United States, and Vietnam. GTAP however does not cover Brunei, and is therefore not included in our analysis.

See Goldman Sachs Global Economics, 2011, -TPP and Its Positive Impact” Japan Economics Analyst, Commodities and Strategy Research, Issue No.11/03.

Other Resources Citing This Publication