8 Should East Asian countries return to a dollar peg again?

Transcription

1 Should East sia return to a dollar peg again? 8 Should East sian countries return to a dollar peg again? Eiji Ogawa The East sian currency crisis that began with the collapse of the Thai baht in July 1997 spread rapidly to other ssociation of South East sian Nations (SEN) economies and the newly industrialising economies (NIEs). Some East sian economies faced speculative attacks several times during the crisis. s a result, three East sian economies Thailand, Indonesia and Korea, requested financial support from the International Monetary Fund (IMF) in Even with such support, they experienced dire economic conditions in The East sian crisis had several distinguishing features. First, macroeconomic variables except for large current account deficits in Thailand, Malaysia and Korea were in a sustainable condition before the crisis. Neither budget deficits nor decreases in foreign reserves were found to be present, in contrast to past currency crises in Latin merican countries. lthough growth rates of export revenues had fallen dramatically in these countries since 1996, the large volume of capital inflows made the government complacent about the current account deficits (see IMF 1997 and 1998; Ito 1999). Second, the East sian currency crisis had contagion effects. Immediately after the Thai baht came under speculative attack and was floated on 2 July 1997, the crisis spread to the Indonesia rupiah, the Malaysia ringgit and the Philippine peso. s a result, all of the currencies were forced to adopt a floating exchange rate system. On 24 July, there occurred a currency meltdown after all suffered speculative attack (Lindgren et al. 1999). The Korean won faced a particularly severe attack immediately following speculative attacks on the Hong Kong dollar in October Third, the monetary authorities of East sian economies adopted a strongly weighted peg of their home currency to the US dollar. The de facto dollar peg system influenced both the current and capital accounts of the countries. It also tempted domestic borrowers and foreign lenders to ignore foreign exchange risks. The appreciation of the US dollar vis-à-vis the Japanese yen caused an appreciation of the East sian currencies under the de facto dollar peg system, which had worsened these countries accounts since Finally, the currency crisis occurred simultaneously with a financial crisis that 1

2 East sian Trade and Financial Integration: new issues caused some financial intermediaries to go bankrupt. The IMF (1997 and 1998) pointed out that fragility of the domestic financial sectors lay behind the financial crisis. In an effort to prevent further deterioration, the IMF required the governments of the supported countries to restructure their financial sectors as a condition for receiving financial support, as large capital inflows to these countries with inadequate financial institutions had brought about the financial crisis. These features of the East sian currency crisis indicate that the crisis was caused by a combination of worsening trade balances under the de facto dollar peg system, the rise and fall of capital flows under active global capital markets, and poor financial risk management within fragile domestic financial sectors. This chapter focuses on the role of the exchange rate system in the currency crisis with the aim of draining lessons. One such lesson is that the de facto dollar peg system is dangerous for these East sian economies that trade with not only the United States but also Japan, the European Union countries, and each other. This chapter examines recent movements in the exchange rates of East sian economies by estimating weights on the US dollar and Japanese yen in a currency basket for some East sian economies according to a method used by Frankel and Wei (1994). It is found that some of the countries have returned to a de facto dollar peg in recent years. The findings support the results addressed by Kawai and kiyama (2000) and McKinnon (2000). ttention is also given to the reasons why the monetary authorities have returned to a de facto dollar peg system. It is possible to point out some factors, including the US dollar as a nominal anchor, recent appreciation of the Japanese yen against the US dollar, and intra-regional trade relations and competitiveness. Focus is placed on the interdependence of exchange rate regimes of the East sian economies examined in this paper. Ogawa and Ito (2000) used a two-country model to analyse why the monetary authorities stayed with a dollar peg system. We discuss the possibility that coordination failure in exchange rate polices among the monetary authorities partly explain their decision to remain with the dollar peg system. DE FCTO DOLLR PEG SYSTEM S CUSE OF THE EST SIN CRISIS Exchange rate policies before the crisis ccording to the IMF s classification system, East sian economies adopted exchange rate arrangements other than the dollar peg system before the East sian currency crisis (IMF 1997). The exchange rate arrangements in Indonesia, Malaysia and Korea were classified as a managed float system. 2

3 Should East sia return to a dollar peg again? The exchange rate arrangement in the Philippines was classified as an independent float system. Thailand used a currency basket peg system. Hong Kong was the only country to adopt a dollar peg system. However, the classification differed from empirical analyses by Frankel and Wei (1994) and Kawai and kiyama (1998) with respect to the currency to which the East sian monetary authorities pegged their home currencies. Frankel and Wei (1994) estimated the weights placed on major foreign currencies in exchange rate policies during the period sian currencies (in terms of the Swiss franc) are regressed on the US dollar (in terms of the Swiss franc) and the Japanese yen (in terms of the Swiss franc) for various sub-periods in , using weekly data. Kawai and kiyama (1998) also produced estimates for the period between using the same method as Frankel and Wei. The results of estimation are summarised in Table 8.1. ccording to their empirical results, the weights placed on the US dollar were nearly equal to one for the Hong Kong dollar, the Korean won, the Indonesian rupiah and the Philippine peso. The weight on the US dollar was 0.91 or for the Thai baht. Moreover, Frankel and Wei (1994) showed that the Singapore dollar and the Malaysian ringgit had a coefficient of about 0.7. Thus, the estimation indicates that the monetary authorities of the East sian economies did in fact adopt a de facto dollar peg system. Table 8.1 Weights on the US dollar and Japanese yen in the exchange rate policies of East sian countries Frankel and Wei (1994) Kawai and kiyama (1998) Sample period: Sample period: Coefficient on Coefficient on Coefficient on Coefficient on US dollar yen US dollar yen Singapore dollar * Hong Kong dollar Korean won Malaysian ringgit Thai baht Philippine peso Indonesian rupiah Note:* coefficient on the SDR is Source: Ogawa, E. and Sun, L., 2001., How were capital inflows stimulated under the dollar peg system? in T. Ito and. O. Krueger, eds., Regional and Global Capital Flows: macroeconomic causes and consequences, Chicago: University of Chicago Press. 3

4 East sian Trade and Financial Integration: new issues The de facto dollar peg system contributed to the currency crisis by depressing the trade balances and stimulating capital inflows to the crisis countries prior to the crisis. The exchange rate movements had negative effects on the international trade competitiveness of East sian economies that had adopted the de facto dollar peg system. The real effective exchange rates of the Thai baht, the Malaysian ringgit and the Indonesian rupiah had been fluctuating without any trend towards appreciation in the early 1990s. However, they had been appreciating since May The de facto dollar peg system and the depreciation of the Japanese yen against the US dollar influenced the movements of the real effective exchange rates. Private capital inflows to Thailand, Indonesia and Korea increased in the 1990s surging noticeably in 1995 and The surge in capital inflows to Thailand was mainly caused by other investment, such as international bank loans. Portfolio investment flows to Korea were larger than international bank loans in 1993, but the opposite was true after Since private capital inflows reached their peak in 1996, international bank loans dominated capital inflows to these sian countries thereafter. Fluctuating trade balances We can consider whether the de facto dollar peg system caused the trade balances of the East sian economies to fluctuate by looking at results of empirical research by Ito et al (1998). In the empirical research, theoretical models were set up to estimate optimal weights on the US dollar and the Japanese yen in a currency basket, in order to stabilise trade balance fluctuations. The models had the following three characteristics. First, they had micro-foundation; that is, it was assumed that oligopolistic exporters maximised their profits while they competed with exporters of other countries. Thus, the export price was endogenously determined in response to the exchange rates. Second, effects of exchange rates on imports of parts (semi-finished goods) were explicitly introduced into the model. This reflected the cost aspects of the currency changes. Finally, optimality of the exchange rate policy in the model entailed minimising trade balance fluctuations. Therefore, it is possible to consider the issue by comparing our estimates of weights with actual weights. Table 8.2 shows the optimal weights for each country for Model (with a constant term in price equation (-1), and without a constant term (-2)) and Model B (with a constant term in price equation (B-1), and without a constant term (B-2)). 1 In general, estimates were between zero and one, except one subcase of the Philippines. Unfortunately, estimates were sensitive to the choice of model (-1, -2, B-1 or B-2) for each country. The estimates naturally varied across countries. It was pointed out that all of the estimated weights on the US 4

5 Should East sia return to a dollar peg again? Table 8.2 Optimal weights in a currency basket ctual weights Optimal weights Model -1 Model -2 Model B-1 Model B-2 US$ yen US$ yen US$ yen US$ yen US$ yen (%) (%) (%) (%) (%) (%) (%) (%) (%) (%) Thai baht Indonesian rupiah Korean won Singapore dollar Philippine peso Note: ctual weights are from Frankel and Wei (1994). Model -1 uses the coefficients estimated in the case of export price equations (Model ) with the constant and coefficients estimated in the export volume equations. Model -2 uses those in export price equations (Model ) without the constant and those in the export volume equations. Model B-1 uses those in the export price equations (Model B) with the constant and those in the export volume equations. Model B-2 uses those in the export price equations without the constant and those in the export volume equations. Source: Ito, T., Ogawa, E. and Sasaki, Y. N., How did the dollar peg fail in sia? Journal of the Japanese and International Economies, 12: dollar were smaller than the actual weights estimated by Frankel and Wei (1994). ccording to the estimates, in all countries, the weight on the US dollar should have been much smaller than the actual weights. The optimal weight on the Japanese yen is highest in Korea and Singapore. Since these countries are often thought to have an industrial structure and associated technological levels similar to Japan, their exports directly compete with Japanese products in Japan and the United States. Therefore, the results of the higher optimal weights on the Japanese yen look reasonable in the two countries. Stimulated capital inflows Ogawa and Sun (2001) empirically analysed how the de facto dollar peg system stimulated capital inflows to the crisis countries (Thailand, Korea and Indonesia) by focusing on the relationship between the de facto dollar peg system and the capital inflows to these counties. Capital flows were regressed on explanatory variables which included interest rates, foreign exchange risks, export growth rate, and rate of change in stock prices. Then an instrumental variable method was used to account for the way in which the instrumental variables were influenced by other variables. The ratio of capital flows to nominal gross domestic product was used to eliminate an increasing trend in capital flows. In addition, a simulation analysis was done by supposing that the monetary authorities had adopted a currency basket peg system instead of the de facto 5

6 East sian Trade and Financial Integration: new issues dollar peg system. It was assumed that estimated coefficients on the explanatory variables in the capital flow equations were unchanged even if the monetary authorities changed their exchange rate policy. 2 The estimated equations were used to conduct a simulation analysis of the capital inflows under the assumption that the monetary authorities had adopted a currency basket peg system in which the basket comprised of 50 per cent of the US dollar and 50 per cent yen, as opposed to the actual de facto dollar peg system, in which the basket consisted of 80 per cent of the US dollar and 20 per cent of the yen. In this case, fluctuations of the exchange rate of the home currency vis-à-vis the US dollar would have been doubled while fluctuations of the exchange rate vis-à-vis the yen would have been halved under the currency basket peg system. In addition, it was assumed that the foreign exchange risks would have altered in the same direction as the exchange rate fluctuations. Therefore, foreign exchange risks of the home currency against the US dollar would have been doubled while foreign exchange risks of the home currency against the yen would have been halved under the currency basket peg system. Simulated capital inflows under the currency basket peg system were compared with the capital inflows under the actual de facto dollar peg system. The means of simulated values were smaller than those of estimated values Table 8.3 Means and standard errors of estimated and simulated values Thailand Korea Indonesia Capital flows Other Portfolio and Other Portfolio Other investments other investments and other investments investments investments 1986Q1 Estimated Q1 value (0.0318) (0.0332) (0.0182) (0.0264) (0.0060) Simulated value (0.0558) (0.0633) (0.0251) (0.0393) (0.0053) 1990Q1 Estimated Q1 value (0.0118) (0.0086) (0.0095) (0.0178) (0.0036) Simulated value (0.0113) (0.0109) (0.0166) (0.0316) (0.0028) Note: Standard errors are given in parentheses. Source: Ogawa, E. and Sun, L., 2001., How were capital inflows stimulated under the dollar peg system? in T. Ito and. O. Krueger, eds., Regional and Global Capital Flows: macroeconomic causes and consequences, Chicago: University of Chicago Press (forthcoming). 6

7 Should East sia return to a dollar peg again? (Table 8.3). In particular, the simulated values of portfolio and other investments were significantly smaller than the estimated values in the case of Thailand. Thus, it is possible to conclude that the de facto dollar peg system had a stimulating effect on capital inflows to Thailand. The means of simulated values were also smaller than those of estimated values in Indonesia though not significantly different. Thus, the de facto dollar peg system also had a slightly stimulating effect on capital inflows to Indonesia. From the regression analysis of the actual capital inflows, it was found that responsiveness of capital flows to the foreign exchange risk against the US dollar was much larger than responsiveness of capital flows to the foreign exchange risk against the yen in the case of Thailand and Korea. The asymmetry in the responsiveness between foreign exchange risks against the US dollar and the yen would have decreased capital inflows under the currency basket peg system. In other words, capital inflows were stimulated more through the stable exchange rate against the US dollar under the dollar peg system. RECENT RETURN TO DE FCTO DOLLR PEG Movements of exchange rates for East sian currencies Despite the fact that the de facto dollar peg system was one of the causal factors of the East sian crisis, some of the East sian economies have returned to de facto dollar pegging. For example, the monetary authorities of Malaysia has officially adopted a dollar peg system. Figures show the movement of exchange rates of local currencies for SEN countries (Thailand, Indonesia, Malaysia and the Philippines) and three NIEs (Singapore, Korea and Taiwan) vis-à-vis the US dollar vis-à-vis the Japanese yen. Exchange rates of their home currencies vis-à-vis the US dollar fluctuated more widely during the currency crisis period from July 1997 to the end of Moreover, some of the countries (Thailand, Indonesia, Malaysian and Korea) experienced overshooting of their exchange rates during the currency crisis. The same movements occurred in their exchange rates vis-à-vis the Japanese yen during the currency crisis. Exchange rate movements had generally stabilised by However, we can find differences in fluctuations between their exchange rates vis-à-vis the US dollar and the Japanese yen; fluctuations against the latter have been more pronounced. 7

8 East sian Trade and Financial Integration: new issues Figure 8.1 Exchange rates of Thai baht Exchange rate Thai baht to US$ Thai baht to JPY 90 1/1/97 3/1/97 5/1/97 7/1/97 9/1/97 11/1/97 1/1/98 3/1/98 5/1/98 7/1/98 9/1/98 11/1/98 1/1/99 3/1/99 5/1/99 7/1/99 9/1/99 11/1/99 1/1/00 3/1/00 5/1/00 7/1/00 9/1/00 Source: Datastream Date Figure 8.2 Exchange rates of Indonesian rupiah 690 Indoneaian rupiah to US $ 590 Exchange rate /1/97 3/1/97 5/1/97 7/1/97 9/1/97 11/1/97 1/1/98 3/1/98 5/1/98 7/1/98 9/1/98 11/1/98 1/1/99 3/1/99 5/1/99 7/1/99 9/1/99 11/1/99 1/1/00 3/1/00 5/1/00 7/1/00 9/1/00 Indonesian rupiah to JPY Date Source: Datastream 8

9 Should East sia return to a dollar peg again? Figure 8.3 Exchange rate of Philippine peso Philippine peso to US $ Exchange rate Philippine peso to JPY /1/97 3/1/97 5/1/97 7/1/97 9/1/97 11/1/97 1/1/98 3/1/98 5/1/98 7/1/98 9/1/98 11/1/98 1/1/99 3/1/99 5/1/99 7/1/99 9/1/99 11/1/99 1/1/00 3/1/00 5/1/00 7/1/00 9/1/00 Date Source: Datastream Figure 8.4 Exchange rates of Malaysian ringgit Exchange rate Malaysian ringgit to US$ 120 Malaysian ringgit to JPY /1/97 3/1/97 5/1/97 7/1/97 9/1/97 11/1/97 1/1/98 3/1/98 5/1/98 7/1/98 9/1/98 11/1/98 1/1/99 3/1/99 5/1/99 7/1/99 9/1/99 11/1/99 1/1/00 3/1/00 5/1/00 7/1/00 9/1/00 Source: Datastream Date 9

10 East sian Trade and Financial Integration: new issues Figure 8.5 Exchange rates of Singapore dollar Exchange rate Singapore $ to US $ 100 Singapore $ to JPY 90 1/1/97 3/1/97 5/1/97 7/1/97 9/1/97 11/1/97 1/1/98 3/1/98 5/1/98 7/1/98 9/1/98 11/1/98 1/1/99 3/1/99 5/1/99 7/1/99 9/1/99 11/1/99 1/1/00 3/1/00 5/1/00 7/1/00 9/1/00 Source: Datastream Date Figure 8.6 Exchange rates of Korean won Exchange rate South Korean won to US $ South Korean won to JPY 90 1/1/97 3/1/97 5/1/97 7/1/97 9/1/97 11/1/97 1/1/98 3/1/98 5/1/98 7/1/98 9/1/98 11/1/98 1/1/99 3/1/99 5/1/99 7/1/99 9/1/99 11/1/99 1/1/00 3/1/00 5/1/00 7/1/00 9/1/00 Date Source: Datastream 10

11 Should East sia return to a dollar peg again? Figure 8.7 Exchange rates of Taiwan dollar Exchange rate Taiwan new$ to US $ Taiwan new $ to JPY 1/1/97 3/1/97 5/1/97 7/1/97 9/1/97 11/1/97 1/1/98 3/1/98 5/1/98 7/1/98 9/1/98 11/1/98 1/1/99 3/1/99 5/1/99 7/1/99 9/1/99 11/1/99 1/1/00 3/1/00 5/1/00 7/1/00 9/1/00 Source: Datastream Date Estimation of weights on the US dollar Next, we empirically analyse the weighting placed on the US dollar in the monetary authorities conduct of exchange rate policy. McKinnon (2000) and Kawai and kiyama (2000) used a method of Frankel and Wei (1994) to conduct a similar analysis. Their results indicated that correlation between East sian home currencies and the US dollar have recently returned to the high levels seen before the East sian crisis. We divide the sample period into sub-periods of six months when we use daily data in estimation, and into one-year sub-periods when we use weekly data. We estimate weights that the monetary authorities placed on major foreign currencies (the US dollar, Japanese yen, Deutsche mark and British pound) in their possible currency baskets during the period between January 1997 and September Each of the local currencies (in terms of the Swiss franc) is regressed on the major currency (in terms of the Swiss franc) for various subperiods during , with daily and weekly data. Regressions are made in terms of log differences because it is usual that levels of exchange rates are non-stationary in many cases. Log differences of exchange rates of a local currency vis-à-vis the Swiss franc are regressed on exchange rates of the major currencies vis-à-vis the Swiss franc. 11

12 East sian Trade and Financial Integration: new issues log e = a + a log e + a log e + a log e + a log e + ε t (8.1) home SF USD SF JPY SF DM SF BP SF Table 8.4 shows results of estimation of weights in a possible currency basket by using daily exchange rate data. In the case of the Thai baht, the weight on the US dollar was during the pre-crisis period from January to June The weight decreased during the crisis period from July 1997 to June Especially during the period from January to June 1998, weights on the US dollar were not significantly different from zero. However, the weight on the US dollar increased to one during the period from July to December Weights on the other currencies were extremely small over time. In the case of the Indonesian rupiah, the weight on the US dollar was before the crisis. The weight on the US dollar decreased during the crisis period from July 1997 to June The weight was not significantly different from zero. However, it increased to nearly one during the period from January to June 1999 and from January to September The Philippine peso has had a relatively stable relationship with the US dollar over time. The weight on the US dollar decreased to during the crisis period from January 1998 to June 1998, and while it was significantly different from zero. The weight on the US dollar increased to nearly one in 1999, though it was somewhat lower in The Malaysian monetary authorities placed a weighting of one on the US dollar before the currency crisis. During the crisis, the weighting was reduced, but returned to one after the crisis. The weight on the US dollar was exactly one in 1999 and 2000 because the monetary authorities officially adopted a dollar peg system after September The Singapore dollar had similar movements. The weight on the US dollar was during the pre-crisis period from January to June It decreased to during the period from January to June 1998, after which time it rose to almost one in 1999 and It is said that the monetary authorities of Singapore have conducted exchange rate policy by making reference to a currency basket. s a result, weights on the other currencies especially the Japanese yen were small but significantly different from zero. In the case of the Korean won, the weight on the US dollar was during the pre-crisis period from January to June It was decreased during the crisis, from July 1997 to June 1998, such that it was not significantly different from zero. However, the weighting has been increased to nearly one since July Weights on the other currencies were extremely small over this time, with the exception of the yen whose weighting was during the crisis period, from July to December The Taiwan dollar had a relatively stable and high correlation with the US dollar over the time period in question. The weight on the US dollar was nearly 12

13 Should East sia return to a dollar peg again? Table 8.4 Estimation of weights in a currency basket (daily data) Currency Period US dollar Yen DM British pound Thai Jan Jun *** 0.049*** baht Jul Dec ** Jan Jun Jul Dec *** Jan Jun *** Jul Dec *** Jan Jun *** Jan Sep *** ** Indonesian Jan Jun *** rupiah Jul Dec Jan Jun ** Jul Dec * Jan Jun *** 0.298* Jul Dec ** Jan Jun *** Jan Sep *** *** Philippine Jan Jun *** peso Jul Dec *** Jan Jun ** Jul Dec *** Jan Jun *** Jul Dec *** Jan Jun *** Jan Sep *** Malaysian Jan Jun *** ringgit Jul Dec ** 0.303* 0.602* Jan Jun * Jul Dec *** Jan Jun *** Jul Dec *** Jan Jun *** Jan Sep *** Singapore Jan Jun *** 0.095*** dollar Jul Dec *** * Jan Jun *** 0.209** Jul Dec *** 0.232*** Jan Jun *** 0.072*** 0.303*** Jul Dec *** Jan Jun *** Jan Sep *** Korean Jan Jun *** 0.049* won Jul Dec ** Jan Jun Jul Dec *** Jan Jun *** Jul Dec *** Jan Jun *** Jan Sep *** continued over... 13

14 East sian Trade and Financial Integration: new issues please define all (as no data, not applicable or zero) for all tables Currency Period US dollar Yen DM British pound Taiwanese Jan Jun *** dollar Jul Dec *** Jan Jun *** Jul Dec *** 0.099*** Jan Jun *** Jul Dec *** Jan Jun *** Jan Sep *** Notes: * significant at the 10 per cent level; ** significant at the 5 per cent level, *** significant at the 1 per cent level. Period [Jan Jun 1997]: 2/1/97 30/6/97 Period [Jul Dec 1997]: 2/7/97 31/12/97 Period [Jan Jun 1998]: 2/1/98 30/6/98 Period [Jul Dec 1998]: 7/2/98 31/12/98 Period [Jan Jun 1999]: 4/1/99 30/6/99 Period [Jul Dec 1999]: 2/7/99 31/12/99 Period [Jan Jun 2000]: 4/1/00 30/6/00 Period [Jan Sep 2000]: 4/1/00 15/9/00 log = + log + log + log + log + e home SF a0 a USD SF JPY SF DM SF BP SF 1 e a2 e a3 e a4 e ε t In doing regressions of exchange rates of a local currency on those of all the major currencies, we omitted variables that were significantly negative. Source: uthor s calculations. one, except for the crisis period from January to June During this time, the weight on the US dollar was Table 8.5 is abstracted results of the weights on the US dollar from the same estimations as Table 8.4. The standard errors of the coefficient can be interpreted as a measure of how precisely the monetary authorities targeted the relevant exchange rate. The standard errors were very small in all of the countries before the currency crisis, reflecting fact that they adopted a de facto dollar peg system. The standard errors have decreased in all of the countries after the currency crisis. In the case of Malaysia, the figure is zero because Malaysia has used the dollar peg system since September In the case of the Singapore dollar and Korean won, the standard errors have recently returned to almost the same level as in the pre-crisis period. We estimated weights in a possible currency basket, using weekly data of 14

15 Should East sia return to a dollar peg again? Table 8.5 Estimates of weights on the US dollar (daily data) Currency Period Coefficient Standard error Thai baht Jan Jun *** Jul Dec ** Jan Jun Jul Dec *** Jan Jun *** Jul Dec *** Jan Jun *** Jan Sep *** Indonesian rupiah Jan Jun *** Jul Dec Jan Jun Jul Dec * Jan Jun *** Jul Dec Jan Jun *** Jan Sep *** Philippine peso Jan Jun *** Jul Dec *** Jan Jun ** Jul Dec *** Jan Jun *** Jul Dec *** Jan Jun *** Jan Sep *** Malaysian ringgit Jan Jun *** Jul Dec ** Jan Jun * Jul Dec *** Jan Jun *** Jul Dec *** Jan Jun *** Jan Sep *** Singapore dollar Jan Jun *** Jul Dec *** Jan Jun *** Jul Dec *** Jan Jun *** Jul Dec *** Jan Jun *** Jan Sep *** continued over... 15

16 East sian Trade and Financial Integration: new issues Currency Period Coefficient Standard error Korean won Jan Jun *** Jul Dec Jan Jun Jul Dec *** Jan Jun *** Jul Dec *** Jan Jun *** Jan Sep *** Taiwanese dollar Jan Jun *** Jul Dec *** Jan Jun *** Jul Dec *** Jan Jun *** Jul Dec *** Jan Jun *** Jan Sep *** Note: bstracted results of weights on the US dollar from the same estimation as Table 8.4. Source: uthor s calculations. exchange rates (Table 8.6). The results of estimation with weekly data are somewhat different from those with daily data; weights on the US dollar using weekly data were somewhat smaller than those with daily data in the cases of the Thai baht, the Indonesia rupiah, the Philippine peso and the Singapore dollar. This corresponds to the results that weights on other currencies were somewhat larger. Some weights on the Japanese yen were significantly different from zero, though still small. In these currencies, the weights on the US dollar increased after the currency crisis, though they have not reached one. In the case of the Thai baht, the weight on the US dollar decreased from in the pre-crisis period to in the crisis period. The weight in the crisis period was not significantly different from zero. In the case of the Singapore dollar, the weight on the US dollar has increased since July 1998, reaching in In contrast, the weights on the US dollar exceeded 0.9 in recent years in the case of the Korean won and the Taiwan dollar. The weight on the US dollar has been one for the Malaysian ringgit because the monetary authorities have officially adopted the dollar peg system. Table 8.7 shows standard errors of weights on the US dollar, using weekly data. In the case of the Thai baht, the standard errors of weight (0.150 in 2000) on the US dollar after the currency crisis were smaller than those before the currency crisis (0.248 in the first half of 1997). The standard errors were zero in 16

17 Should East sia return to a dollar peg again? Table 8.6 Estimation of weights in a currency basket (weekly data) Currency Period US dollar Yen DM British pound Thai Jan 1997 Jun *** baht Jul 1997 Jun Jul 1998 Jun *** Jul 1999 Jun *** Jan 2000 Sep *** 0.249** Indonesian Jan1997 Jun *** rupiah Jul 1997 Jun Jul 1998 Jun * Jul 1999 Jun Jan 2000 Sep Philippine Jan1997 Jun *** peso Jul 1997 Jun Jul 1998 Jun *** Jul 1999 Jun *** Jan 2000 Sep *** Malaysian Jan1997 Jun *** ringgit Jul 1997 Jun Jul 1998 Jun Jul 1999 Jun *** Jan 2000 Sep *** Singapore Jan 1997 Jun *** 0.108*** dollar Jul 1997 Jun ** Jul 1998 Jun *** 0.129** * Jul 1999 Jun *** * Jan 2000 Sep *** 0.119* * Korean Jan 1997 Jun *** 0.131** won Jul 1997 Jun Jul 1998 Jun *** Jul 1999 Jun *** Jan 2000 Sep *** Taiwanese Jan 1997 Jun *** dollar Jul 1997 Jun *** Jul 1998 Jun *** 0.192*** Jul 1999 Jun *** Jan 2000 Sep *** Notes: * significant at the level of 10 per cent; ** significant at the level of 5 per cent; *** significant at the level of 10 per cent Period [Jan Jun 1997]: 8/1/97 2/7/97 Period [Jul 1997 Jun 1998]: 2/7/97 1/7/98 Period [Jul11998 Jun 1999]: 1/7/98 30/6/99 Period [Jul 1999 Jun 2000]: 7/7/99 5/7/00 Period [Jan Sep 2000]: 12/1/ /9/00 log e = a + a log e + a log e + a log e + a log e + ε t home SF USD SF JPY SF DM SF BP SF We omitted variables that are significantly negative when we made regression of exchange rates of a local currency on those of all the major currencies. Source:uthor s calculations. 17

18 East sian Trade and Financial Integration: new issues Table 8.7 Estimates of weights on the US dollar (weekly data) Currency Period Coefficient Standard error are these tables all your calculations? Thai baht Jan 1997 Jun *** Jul 1997 Jun Jul 1998 Jun *** Jul 1999 Jun *** Jan 2000 Sep *** Indonesian rupiah Jan1997 Jun *** Jul 1997 Jun Jul 1998 Jun Jul 1999 Jun Jan 2000 Sep Philippine peso Jan1997 Jun *** Jul 1997 Jun Jul 1998 Jun *** Jul 1999 Jun *** Jan 2000 Sep *** Malaysian ringgit Jan1997 Jun *** Jul 1997 Jun Jul 1998 Jun Jul 1999 Jun *** Jan 2000 Sep *** Singapore dollar Jan1997 Jun *** Jul 1997 Jun ** Jul 1998 Jun *** Jul 1999 Jun *** Jan 2000 Sep *** Korean won Jan1997 Jun *** Jul 1997 Jun Jul 1998 Jun *** Jul 1999 Jun *** Jan 2000 Sep *** Taiwanese dollar Jan1997 Jun *** Jul 1997 Jun *** Jul 1998 Jun *** Jul 1999 Jun *** Jan 2000 Sep *** Notes: bstracted results of weights on the US dollar from the same estimation as Table 8.6. Source: uthor s calculations 18

19 Should East sia return to a dollar peg again? the case of the Malaysian ringgit after the official adoption of the dollar peg system. In the case of the Singapore dollar, the Korean won and the Taiwan dollar, the standard errors have decreased, and returned to roughly their precrisis levels in From the empirical analysis, the results indicate that the weights on the US dollar returned to one at the daily level, while they have increased to almost one for some of the countries at the weekly level. In addition, we found that the weights on the US dollar have increased in most of the East sian economies. We need to consider why the monetary authorities might wish to return to a de facto dollar peg system. Some factors, including the US dollar as a nominal anchor, recent appreciation of the Japanese yen against the US dollar, and intraregional trade relations or competitiveness. COORDINTION FILURE IN CHOOSING N OPTIML CURRENCY BSKET SYSTEM two-country model There may be situations where the monetary authorities are forced to keep a dollar peg system instead of adopting a currency basket system, despite the effects experienced under a dollar peg system during the East sian crisis. Such situations are often the result of coordination failure. ssume that all the East sian economies are using a the de facto dollar peg system, and that each knows that it should adopt an optimal currency basket system in order to stabilise fluctuations in its trade balance. ssume also that firms within these countries are competitive in both Japanese and US markets. If one country switches to a currency basket system while the others keep the dollar peg system, the country with a currency basket system might face increased fluctuations in its trade balance. If the US dollar depreciates vis-à-vis the yen, the related appreciation of its home currency vis-à-vis the other currencies worsens the price competitiveness of its firms. On the other hand, if the US dollar appreciates vis-à-vis the yen, the related depreciation of its home currency vis-àvis the other currencies improves the price competitiveness of its firms. Thus, the monetary authorities of the country, by choosing to adopt a currency basket system have brought about greater volatility in the country s trade balance. t the individual country level, it is better for the monetary authorities to stay with the dollar peg system. Coordination failure has occurred. Ogawa and Ito (2000) used a game-theoretic framework of a two-country model to analyse the coordination failure in choosing an exchange rate system. 19

20 East sian Trade and Financial Integration: new issues We analyse possible coordination failures by comparing losses for the monetary authorities in two situations: 3 a situation where both of the monetary authorities adopt the dollar peg simultaneously, and a situation where the monetary authority of one country adopts an optimal currency basket peg while the monetary authority of the other country adopts a dollar peg. We express the above effects of exchange rates on the trade balances of countries and B in terms of rates of changes. ˆ ˆ Y ˆ $ ˆ BY ˆ B$ T = e + e e e (8.2) T ˆ = Be ˆ + B e ˆ B e ˆ B e ˆ (8.3) B BY B$ Y $ where T ˆ is the rate of change in trade balances, and ê is the rate of change in exchange rates. Coefficients and B are positive under the Marshall Lerner condition. Both of the monetary authorities are assumed to choose weights on the US dollar and the Japanese yen in a currency basket in order to stabilise the fluctuation of their own trade balances caused by changes in the exchange rates. Optimality of the exchange rate policy involves stabilising fluctuations in trade balances in terms of the US dollar under a currency basket peg system. We assume that the monetary authorities control the weights in a currency basket to minimise the squared rate of change in trade balances in terms of the US dollar, subject to the following equations: $ Y weˆ (1 ) ˆ + w e = 0 (8.4) B $ B Y weˆ (1 ) ˆ B + wb e = 0 (8.5) where w : a weight on the US dollar in a currency basket for country, w B : a weight on the US dollar in a currency basket for country B. We can derive the first order conditions for minimising their objective functions to obtain the following linear reaction functions: ( + ) w ( + ) w = (8.6) B 2 4 ( B + B ) w ( B + B ) w = B B (8.7) 1 2 B There is a unique equilibrium pair of optimal weights for countries and B because both of the policy reaction functions are linear functions. From Equations * * (8.6) and (8.7), we derive a pair of optimal weights ( w, w ) on the US dollar in B a currency basket, which stabilises the trade balances of countries and B simultaneously. If the monetary authorities of countries and B could, * * simultaneously, set w and w, respectively, trade balances would be stabilised B in both countries. However, it is not always guaranteed that the optimal weights 20

21 Should East sia return to a dollar peg again? for the both countries are a stable equilibrium. The condition for a stable equilibrium is 1+ 2 B3 + B4 < B1+ B (8.8) 2 In this case, a pair of weights converges to an equilibrium implied by the * * optimal weights ( w, w ). The weights for both countries should converge to B their optimal equilibrium weights. On the other hand, if 1+ 2 B3 + B4 > B1+ B (8.9) 2 * * a pair of optimal weights ( w, w ) is an unstable equilibrium. In this case, B * weights diverge from their optima once they leave the equilibrium point ( w, * w ). B Suppose that each of the monetary authorities of countries and B chooses its own weighting in order to stabilise its own trade balances, given the weights chosen by the other monetary authorities. The weights chosen by the monetary * * authorities should diverge from the optimal weights ( w, w ). Thus, the weights B on the US dollar increase and reach unity for both countries, provided that the weight is realistically constrained between zero and one. Both monetary authorities eventually adopt a full dollar peg system rather than the optimal currency basket peg system although they have been choosing weights in order to stabilise their own trade balances. Thus, if Equation (8.9) is satisfied, an optimal weight point is unstable. It is then difficult for the monetary authorities to alter their stance and adopt an optimal exchange rate policy. Next, we analyse whether the monetary authorities of countries and B can directly shift their exchange rate system from the dollar peg system to an optimal currency basket peg system. The direct shift to the optimal currency basket peg system depends on whether each the monetary authority can reduce trade balance fluctuations in trade balances under the optimal currency basket peg system as compared to the dollar peg system. Each monetary authority should care about fluctuations in trade balances in a case where it shifts to the optimal currency basket peg system while the other keep the dollar peg system. If both of the monetary authorities adopt the dollar peg ( w = wb = 1) simultaneously, fluctuations in country s trade balances are calculated: 2 2 ˆ 2 /$ ( 1) ( 1 3) ˆY T w= wb= = E (8.10) Suppose that the monetary authority of country adopts the above optimal * currency basket peg ( w = w ) while the monetary authority of country B adopts 21

22 East sian Trade and Financial Integration: new issues the dollar peg ( w B = 1). Fluctuations in trade balances for country are obtained in this case as: 2 2 ( /$ 2 * 1+ 2)( B1+ B4) ( 1+ 4)( B3 + B4) ˆY ( w= w, wb= 1) = ( 3+ 4) ( 1+ 2)( B1+ B2) ( 3 + 4)( B3 + B4) Tˆ E (8.11) When the monetary authority of country can choose to adopt the dollar * w = ) or the optimal currency basket peg ( w = w ), given that the monetary peg ( 1 authority of country B adopts the dollar peg ( w B = 1), the monetary authority of country compares trade balance fluctuations between the two options. The monetary authority of country compares Equation (8.10) with Equation (8.11). If fluctuations in the trade balance under the dollar peg (Equation 8.10) are less under the optimal currency basket peg (Equation 8.11), the monetary authority of country prefers the dollar peg to the optimal currency basket peg. The various exchange rate policy options for the two monetary authorities can be graphically Figure 8.8. Point D represents a situation where both monetary authorities adopt the dollar peg system. Point F represents a situation where the monetary authority of country adopts an optimal currency basket peg system while the monetary authority of country B keeps the dollar peg system. If fluctuations in trade balances at point F are larger than those at point D, the monetary authority of country should not change its own weight on the US * dollar from w = 1 to w = w. Furthermore, the monetary authority of country B should behave similarly to that of country because we assumed symmetry of the two economies. Thus, both of the monetary authorities should continue to peg their home currencies to the dollar if their trade balances fluctuate more widely under the optimal currency basket peg than under the dollar peg. t this time, they face a coordination failure, in that they are forced to adopt the dollar peg even though simultaneous adoption of the optimal currency basket peg minimises trade balance fluctuations. The need for necessary of international coordination If each of the monetary authorities of countries and B has a non-linear policy reaction function, there may exist multiple equilibria. In the case of Figure 8.9, point E is an unstable equilibrium. Once either of the monetary authorities chooses a weight in the neighborhood of point E, both of the monetary authorities will choose the weights which are diverging from the optimal weights represented by point E. In contrast, point F is stable. When either monetary authority chooses a weight in the neighbourhood of point F, both will gradually adjust their weights towards those represented by point F. However, the weights represented by point 22

23 Should East sia return to a dollar peg again? Figure 8.8 W B BB F 1 D W * B E 0 Source: Figure W * BB 1 D W titles and sources for these figures W ** B F W * B E 0 W * W ** 1 W Source: 23

24 East sian Trade and Financial Integration: new issues F are larger than the optimal weights represented by point E. In other words, the monetary authorities adopt the de facto dollar peg system. If we suppose that risk-averse monetary authorities choose their exchange rate system under uncertainty, coordination failure is more likely to occur. There is always uncertainty about future movements in the exchange rate of the US dollar vis-à-vis the Japanese yen. Suppose that the risk-averse monetary authority of one country switches its exchange rate system to an optimal currency basket system while the other monetary authority keeps the dollar peg system. The currency of the country that adopted the currency basket system would appreciate against the currencies of the neighbouring countries if the US dollar depreciated against the Japanese yen. Therefore, under such uncertainty, the risk-averse monetary authorities tend to adopt a strategy of waiting and observing the behaviour of the other monetary authorities. ll monetary authorities are likely to adopt a wait-and-see approach if they are risk averse. nd all will choose to keep the dollar peg system, which is the Nash equilibrium, although they know there is a better solution if they cooperate. Coordination among the East sian monetary authorities is necessary in order to shift from the Nash equilibrium to a cooperative solution. First, the monetary authorities should discuss what is their optimal exchange rate system and how to coordinate exchange rate arrangements to achieve it. Then, countries with similar optimal exchange rate systems should agree to an arrangement for creating a common currency basket amongst themselves, and conduct a kind of common exchange rate policy, making reference to this common currency basket. CONCLUSION The East sian currency crisis taught many economies the dangers of a de facto dollar peg system where those economies had diversified trade with the United States, Japan, the countries of the European Union and the region itself. Under the de facto dollar peg system, the movements of the US dollar against the Japanese yen served to worsen trade balances. Moreover, the de facto dollar peg system stimulated capital inflows to the crisis countries prior to the crisis. When we look at movements of the exchange rates of some East sian currencies after the crisis, we see that the exchange rates of the home currency vis-à-vis the US dollar have been stabilised while their exchange rates vis-à-vis the Japanese yen have continued to fluctuate. The monetary authorities of some countries have recently returned to the de facto dollar peg system used prior to the currency crisis. One of the factors explanations for the monetary authorities adoption of the de facto dollar peg system might be coordination failure in the exchange rate 24

25 Should East sia return to a dollar peg again? policies among the East sian economies. We can conclude that a currency basket system is good for the East sian economies in order to prevent another currency crisis in the future. However, the monetary authorities may be faced with coordination failure in choosing the optimal currency basket system. It will be necessary for them to work together in deciding on exchange rate policies that will achieve the optimal currency basket system. NOTES This paper was written during a visit as a visiting scholar to the Research Department of the IMF in September I wish to thank Taimur Baig, Warwick McKibbin, Naoyuki Yoshino, Hiroyuki Hino, Koichi Hamada and other staff at the IMF, and participants at both the NU Japanese MOF workshop (16 October 2000) and the NU Kobe University conferences for their useful suggestions and comments. I am also grateful for Kenichiro Kashiwase s research assistance. 1 We set up two competitive situations: the first is that each of the Japanese and US markets is modeled as a duopoly one where the domestic firm competes against the Japanese firm (Model ). The second is that each of the Japanese and US markets is modeled as a duopoly one where the domestic firm competes against each local firm in the markets (Model B). 2 The Lucas critique tells us that the change in the exchange rate policy might change the coefficients in the estimated equations. 3 Bénassy-Quéré (1999) and Ohno (1999) analyse pegging the US dollar as a coordination failure. REFERENCES Bénassy-Quéré,., Optimal pegs for East sian currencies Journal of the Japanese and International Economies, 13: Frankel, J. and Wei, S., Yen bloc or dollar bloc? Exchange rate policies of the East sian economies, in T. Ito and. O. Krueger, eds., Macroeconomic Linkage: savings, exchange rates, and capital flows, Chicago: University of Chicago Press, pp IMF, World Economic Outlook: interim assessment, Dec. 1997, Washington, DC: International Monetary Fund., World Economic Outlook, May 1998, Washington, DC: International Monetary Fund. Ito, T., Capital Flows in sia, NBER Working Paper No. 7134, Cambridge, 25