رابطه ی علی بین بهره و نرخ ارز در بحران ارز آسیا

This paper studies the causal relationship between interest rates and exchange rates in Indonesia, Korea, Malaysia and Thailand during the period bordering the 1997 Asian currency crisis to investigate the appropriateness of tight monetary policy in stabilizing exchange rates. We employ VAR models consisting of spot rates, forward rates and interest rate differentials to study the causal relations. In particular, we test for long-run causality as well as short-run causality by taking into account non-stationarity of the involved variables and the cointegrating relations among them. The test results show that except for some subsamples for Malaysia there is no evidence that interest rate differentials caused spot exchange rates at all horizons. Considering the ineffectiveness of high interest rates in stabilizing exchange rates and the high economic cost associated with keeping high interest rates for an extended time period, one may rightfully question the appropriateness of tight monetary policy during the Asian currency crisis.

One of the substantial controversies regarding the Asian currency crisis of 1997 has been whether tight monetary policy was effective in stabilizing foreign exchange rates during and in the aftermath of the crisis. According to conventional wisdom, high interest rates are essential in stabilizing the foreign exchange market in the middle of a currency crisis and in achieving reversal of currency undervaluation after a currency crisis. In the short-run, higher interest rates reduce capital outflows by raising the cost of currency speculation and induce capital inflows by making domestic assets more attractive. In the long-run, they improve current account balance by reducing domestic absorption. For this reason, tight monetary policy constituted an essential part of the IMF rescue package for Asian countries together with financial and corporate restructuring. This traditional view, however, has been challenged by several economists Feldstein, 1998, Furman and Stiglitz, 1998, Radelet and Sachs, 1998 and Sachs, 1998. They argue that tight monetary policy in Asia was either ineffective in stabilizing exchange rates or that it may have even exacerbated the situation. As was pointed by Bergsten, 1997, Fischer, 1998 and Krugman, 1998, among others, the Asian currency crisis is like a bank run where panicky foreign creditors lined up to withdraw their credit for fear of insolvency.1 When the solvency of a financial institution is at risk, it would be hard to stop withdrawal of deposits with higher interest rates alone. Likewise, when the solvency of a country is in doubt, capital flows are likely to be insensitive to interest rate differentials. Besides, higher interest rates and tight monetary policies in a country like Korea where the average debt to equity ratio of manufacturing firms was as high as 400% are likely to cause a chain reaction of bankruptcy, thereby further undermining the ability to service its foreign debt. Both the traditional and the revisionist views have some elements of truth in their arguments. Therefore, they need to be examined empirically. In response to this need, this paper aims at studying the causal relation between interest and spot exchange rates in Indonesia, Korea, Malaysia and Thailand by using the data from the period of the currency crisis. There have been a few attempts to empirically examine the effectiveness of high interest rate policy in the Asian crisis. Depending on the model specification, the frequency and the sample period of the data, these studies present mixed results about the effectiveness of high interest rate policy. Ghosh and Phillips, 1998, Kaminsky and Schmukler, 1998 and Goldfajn and Baig, 2002 used daily nominal interest and exchange rate data to estimate VAR models and to calculate impulse response functions. They were unable to find evidence that higher interest rates destabilized foreign exchange markets. Neither were they able to find significant negative correlation between interest rates and exchange rates except for some subsample periods. On the other hand, Dekle et al. (2002), using weekly data for Korea, Malaysia and Thailand, found that interest rates Granger-caused exchange rates during the crisis period. Their results, however, only provides a weak support for the traditional view because the estimated coefficients are too small. Park et al. (1999), using daily data to test causal relation between nominal exchange rates and interest rates in Korea, also found out that these two variables Granger-caused each other during the crisis period of September 1997–September 1998. However, they failed to find any causal relation between these variables in the pre-crisis and the post-crisis periods. In addition, Brailsford et al. (2006), using a VAR model that takes account of the currency contagion effect found that higher interest rates helped to support the exchange rates in South Korea, the Philippines and Thailand, but not in Malaysia. This paper has some new features, which are not necessarily shared by the aforementioned studies. First, we examine the causal relationship between interest rates and exchange rates with forward exchange rates being included as an auxiliary explanatory variable. According to the interest rate parity condition, the anticipated future exchange rate determines the expected return from the investment in foreign assets and thereby affects the current exchange rate. This implies that the expected future exchange rate should be included as an explanatory variable. The forward exchange rate used in our VAR specification captures the effect of anticipated future exchange rate. By contrast, most of the previous studies with the exception of Dekle et al. (2002)2 examine the bivariate VAR relation between current exchange rates and interest rates alone. But such a bivariate VAR approach may run the risk of committing a misspecification error by omitting other relevant variables. Second, we test causal relations between level interest and spot rates by taking into account non-stationarity of the variables included in the model and the cointegrating relations among these variables. For this, Toda and Phillips’ (1993; hereafter, TP) method and its extensions proposed in this paper will be used. Most previous studies employed VAR models involving differenced variables, which makes it difficult to interpret the results in relation to level variables. Third, we test for long-run causality between interest and spot exchange rates as well as short-run causality. The short-run causality which is equivalent to the well known Granger causality (cf. Granger, 1969) is concerned with prediction at the horizon one. But the long-run causality is about prediction at all horizons including infinity as introduced in Dufour and Renault (1998; hereafter, DR). When there are only two variables in the VAR, these two concepts are the same. But when there are more than two variables in the VAR model, they are not necessarily so. In a VAR model involving three variables y1t,y2ty1t,y2t and ztzt, for example, even when ztzt does not help predicting y1ty1t at the horizon one, ztzt may still do so at horizons higher than one by affecting y2ty2t. When ztzt helps predicting y1ty1t through the auxiliary variable y2t,zty2t,zt is said to cause y1ty1t indirectly. The concept of indirect causality was first introduced in Hsiao (1982) and further refined in DR. In our study, forward exchange rate will be the auxiliary variable y2ty2t, which allows us to test if higher interest rates could affect spot exchange rates indirectly through affecting expectation about future exchange rates. The rest of the paper is organized as follows. Section 2 specifies the model used to investigate the causal relation between interest and exchange rates. Section 3 reports some known results about causality tests and extends these results to the case of our interest. Section 4 reports the causality test results for Indonesia, Korea, Malaysia and Thailand. Section 5 contains summary and further remarks.