This action might not be possible to undo. Are you sure you want to continue?
Exchange rate depreciation and exports: The case of Singapore revisited WenShwo Fang
Feng Chia University
Stephen M. Miller
University of Connecticut and University of Nevada, Las Vegas
Working Paper 2004-45 December 2004
341 Mansﬁeld Road, Unit 1063 Storrs, CT 06269–1063 Phone: (860) 486–3022 Fax: (860) 486–4463 http://www.econ.uconn.edu/
but that exchange rate risk signiﬁcantly impedes exports.Abstract This paper revisits the weak relationship between exchange rate depreciation and exports for Singapore. bivariate GARCH-M model . exports. Singaporean policy makers can better promote export growth by stabilizing the exchange rate rather than generating its depreciation. exchange rate risk. In sum. Journal of Economic Literature Classiﬁcation: F14. The evidence shows that depreciation does not signiﬁcantly improve exports. F31 Keywords: depreciation. using a bivariate GARCH-M model that simultaneously estimates time-varying risk.
Introduction A traditional view expects that exchange rate depreciation improves exports. This paper argues that exchange rate risk provides another channel for the exchange rate to affect exports in Singapore. however. Broll and Eckwert (1999) note that the price of an option to export increases with risk. although exchange rate depreciation does not. Germany. except for the UK. and Bahmani-Oskooee and Kara (2003). and NIC exports. De Grauwe (1988) suggests that exporters might increase volume to offset potential losses. exists. 1 . Klaassen (2004) finds no effect on monthly bilateral US exports to the other G7 countries. In an interesting paper. since the collapse of fixed exchange rates in the early 1970s. Kroner and Lastrapes (1993) discover positive effects of conditional variance on exports of France. Arize (1995. For example. however. The probable effects of exchange rate risk received considerable attention. Weliwita et al.Exchange rate depreciation and exports: The case of Singapore revisited I. (2000). Pozo (1992) uncovers a negative effect of exchange rate risk on UK real exports to the US. employing data from a fixed exchange rate period. (2003) and Fang and Thompson (2004) find negative effects of exchange risk on US. G7. Asseery and Peel (1991) find positive relationships for multilateral exports. Arize et al. 1997). Junz and Rhomberg (1973) and Wilson and Takacs (1979). find that depreciation improves exports for developed countries. showing that exchange rate risk reduces exports. Ethier (1973) argues that exchange rate risk could lower exports due to profit risk. and Japan. Contrary evidence exists. McKenzie and Brooks (1997) report a positive risk relationship for Germany and the US. (1999). Little consensus regarding its effect on exports. Abeysinghe and Yeok (1998) discover that exchange rate appreciation does not adversely affect exports for Singapore because exports possess high import content. Chowdhury (1993). but negative effects for the UK and US. using data from a flexible exchange rate period. Arize et al. LDC.
We apply the bivariate GARCH-M modeling approach to Singapore to consider the effects of exchange rate depreciation and its time varying variance on exports.S. implies cheaper exports abroad and improves real exports. We convert the bilateral nominal exchange rate. if both affect exports and one is omitted. No research combines the two possible exchange rate effects together to analyze the relationship between exchange rates and exports in the previous literature. the real exchange rate (q). Engle et al. Seasonally adjusted real export revenue equals nominal export revenue in domestic currency deflated by the consumer price index (CPI).The effects of the exchange rate or exchange rate risk on exports individually may produce biased inference. Data and time-varying variances To assess the net effect of exchange rate depreciation and its risk on exports. This paper estimates simultaneously the effects of exchange rate depreciation and its risk. 1992). if exports and exchange rate risk correlate negatively (see Arized et al. An increase in the real exchange rate. II. and real exchange rate risk ( hq ). 1987 and Bollerslev et al. To provide evidence. Our methodology differs from the study of Abeysinghe and Yeok (1998) that uses OLS estimation with no exchange rate risk variable.g. 2003). Generalized autoregressive conditional heteroscedasticity (GARCH) models specify the relationships between means and variances (e. The effect of real exchange rate risk proves theoretically ambiguous. a depreciation.. on a monthly basis from January 1979 to October 2002. Real foreign income positively affects the demand for exports. Their specification may overestimate the effect of depreciation. defined as the Singaporean currency price of the 2 . we employ a nonstructural partially reduced-form approach of Rose (1990) and Klaassen (2004). where real exports (x) depend directly on real foreign income (y). we use bilateral exports between Singapore and the U.
and suggests studying the risk effect. Two reasons argue for the bilateral approach. All data come from the International Financial Statistics and Direction of Trade of the IMF. the ratio of bilateral exports between Singapore and the US to Singapore’s total exports is 15.53 percent while the average depreciation rate equals 0. can focus on the simple relationship between exchange rate changes and exports. accounting for a substantial share of Singaporean exports. Singapore experienced exchange rate depreciation and export growth. Klaassen (2004) finds that exchange rate risk in developed countries does not exhibit enough variability to determine its effect on exports. into a real exchange rate by multiplying the nominal rate by the ratio of the U. but Kurtosis statistics significantly exceed 3 at the 5-percent level. After selecting lag length by the AIC criterion. the Augmented Dickey-Fuller (ADF) test shows that ∆lxt and ∆lqt prove individually stationary [i. Statistical analysis of the variables identifies appropriate GARCH models for further analyses. dollar.093 percent. Valid inference in GARCH models requires stationarity in variables. We.S. First. The Jarque-Bera test rejects normality. The number of lags ( k ) affects the power of the test. The Ljung-Box Q-statistic tests for autocorrelation. Second. I(0)] series at the 5-percent level. then.. using bilateral exports avoids the asymmetric response of trade flows to exchange rate depreciation and its risk across countries. In our sample. for which more volatile exchange rate risk may exist. implying leptokurtic series with fat tails. The average rate of export growth equals 0. Non-normality and the fat-tailed nature suggest estimating GARCH models under the Student-t distribution. Both the mean and the standard deviation of export growth greatly exceed those of the rate of depreciation.S. In addition.U.3 percent over the sample period. CPI to the Singaporean CPI. using data on developing countries.e. Tsay (2002) 3 . Foreign income equals US industrial production with base year 1995. Skewness statistics for the growth rate of real exports ( ∆lxt ) and the growth rate of the real exchange rate ( ∆lqt ) cannot reject symmetry.
The estimated coefficients of the degree of freedom v are significant at the 5-percent level. To adequately capture the dynamic structure of the data. No evidence of autocorrelation emerges. but no autocorrelation in ∆lq t . The low Ljung-Box Q-statistics for the squared standardized residuals up to 6 lags show no remaining heteroscedasticity. implying the appropriateness of employing the GARCH(1. The higher coefficient of volatility persistence of ∆lxt relative to that of ∆lq t is consistent with the higher standard deviation of ∆lxt . we set k =5. in our sample equals 285.65. possess time-varying variances. The two variables. The estimates in the two variance equations are significantly positive. Ljung-Box Q-statistics for squared ∆lxt and squared ∆lq t suggest the possible presence of time-varying variance for the two series.1) models first to identify properties of the changing variances for ∆lxt and ∆lq t . we employ an ARMA process for both the mean and variance equations of the two variables. Since the exchange rate does not possess autocorrelation. The empirical bivariate GARCH-M model and estimation The following eclectic GARCH-M model provides the framework for assessing the net effect of 4 .1) for both ∆lxt and ∆lq t under the t-distribution. T . ∆lxt and ∆lqt . III. α 1 + α 2 = 0. we specify the mean equation of ∆lqt as a constant. suggesting that the AR(2) process achieves white noise. accordingly. The number of observations. using up to 6 lags. We estimate univariate GARCH(1.933 < 1 and β 1 + β 2 = 0. suggesting the use of bivariate GARCH models to examine the link between exports and exchange rate changes.suggests choosing k = ln(T ) . The Ljung-Box Q( k ) statistics for the standardized residuals of ∆lxt show no autocorrelations up to 6 lags. given the low Ljung-Box Q( k ) statistics for the standardized residuals of ∆lqt .807 < 1 show that each time-varying variance process is stable for ∆lxt and ∆lq t . Moreover. Ljung-Box Q-statistics indicate autocorrelations in ∆lxt .
t −1 + ε x .t −1 + β 2 ⋅ hq .t in equation (1) means that equations (1) to (7) constitute a bivariate GARCH(1.exchange rate depreciation and its risk on exports. and β i > 0 ensure positive conditional variances. This specification reduces the number of parameters and increases the degrees of freedom of model estimation. All parameters in equations (1) to (7) are estimated by maximizing the following log-likelihood function of the bivariate Student-t distribution: 5 .t .t . Now. ∆lqt ≡ 100 × ( ln qt .t equal conditional variances while hxq . The presence of hq . i =1 (1) (2) (3) (4) ∆lqt = e0 + ε q .t ⋅ hq . ∆ly t ≡ 100 × ( ln y t .t equals the covariance. v . and hxq .t −1 .t and hq . The constant correlation specification (Bollerslev 1990) is modeled through (7). hx .t ⎠ ⎛ hx .ln y t −1 ).ln qt −1 ).t −1 .t )′ ε t | Ψt −1 ~ Student − t (v) .t 2 hx . follows a bivariate Student-t distribution with degrees of freedom. The conditions that α i > 0 .1)-M model. ε t . where ∆lxt ≡ 100 × ( ln xt .t −1 + α 2 ⋅ hx . hq .ln xt −1 ). ε q .t ⎞ ⎟. hxq . conditional on the information set Ψt −1 available at time t − 1 .t .t . ρ xq equals the correlation coefficient of ∆lxt and ∆lqt .t Ht = ⎜ ⎝ hxq . The conditions that α1 + α 2 < 1 and β 1 + β 2 < 1 imply stable variances. (5) (6) (7) 2 hq .t = ρ xq ⋅ hx . 2 ∆lxt = a 0 + ∑ ai ⋅ ∆lxt −i + b ⋅ ∆ly t −1 + c ⋅ ∆lqt −1 + d ⋅ hq . ε t = ( ε x .t = α 0 + α 1 ⋅ ε x .t = β 0 + β 1 ⋅ ε q .
362 (0.082) 2 + 0.054) * (0.t = hq .153 Q2 (3) = 8.t (0.621∆lxt −1 1. t −1 (0. The estimation results are as follows:1 ∆lxt = − 0. The model explains changes in exports. * denotes significance at the 5% level.700hq .955 (2. When exchange rate uncertainty leads to profit risk.011) * + 0.711) + 0.5ln H t − ⎜ ⋅ ln ⎜ 1 + ⎟ ⎟ ⎟ v−2 ⎠ ⎝ 2 ⎠ ⎝ 2⎠ ⎝ 2 ⎠ ⎝ (8) where Γ (·) equals the Gamma function. the demand for exports falls (i. The reduced-form equation includes the depreciation rate and its risk as well as the rate of change of foreign income as explanatory variables.t = 0.t −1 (0.145 1 Q 2 ( k ) and Q 2 ( k ) are the bivariate Ljung-Box statistics (Hosking. The net effect on exports includes the exchange rate depreciation and its volatility. 6 .173 Q2(6) = 11.t −1 (0.431) * (0.277) − 0.104 ⋅ hx . .028) * − 0.072) 2 2 Q2(3) = 4. d < 0).050) * + 0. The statistical significance and sign of the estimated c and d coefficients in equation (1) provide a simple and straightforward test of the relationship between real export growth and exchange rate depreciation and its volatility.261) * hxq . v is degree of freedom.485 (0.t + 0.266∆lxt − 2 (0.880∆lyt −1 (0. Exchange rate volatility affects exports through exporters’ responses to perceived risk.229∆lqt −1 (0. then exchange rate depreciation improves exports.051) * ∆lq t = 0. ceteris paribus.864hR .154ε q .e.t −1 0.014) * 2 = + 0. 1980) for standardized and squared 2 standardize residuals for autocorrelations up to k lags.−1 ⎛v+2⎞ ⎛v⎞ ⎛ v + 2 ⎞ ⎛ ε t′ H t ε t ⎞ ln Lt = ln Γ ⎜ − ln ( v − 2 ) − ln Γ ⎜ ⎟ − 0. If the estimate of c exceeds zero.995 Q2 (6) = 22. then.104ε x .032 (0.692 (0.t −1 2.254hq.026) * v= 7.t ⋅ hq .150) * h x .666) * (0.
however. The degree of freedom of the t-distribution.968 for ∆lxt and 0. The two variance processes converge. Exchange rate depreciation exhibits the expected positive effect. a rather substantial 7 .75 implies that the range of potential monthly influences on export revenue calculated by (mean of hq .20%. Exchange rate risk possesses a significantly negative effect on exports. Volatility persistence equals 0. Thus. the hypothesis of using a standardized student-t distribution is not rejected at the 5-percent level.65 percent.854 for ∆lqt . but proves also insignificant. the mean value of conditional variance hq . Regarding the magnitude of the effect. That is.09%]. bilateral exports from Singapore to the US do not respond to US economic activity. but proves insignificant at the 5% level. The marginal effect of real US income (industrial production) on exports exhibits the expected positive sign. The estimated correlation coefficient between ∆lxt and ∆lqt equal 0. 1980) for the standardized and squared standardized residuals of ∆lxt and ∆lqt do not detect any remaining autocorrelation or conditional heteroscedasticity at the 5-percent level.104 that nearly equals the coefficient of 0.t in the bivariate GARCH-M model is 2. is significant.t × d ) equals -0.55.t ± standard deviation)× d covers the range [-0. estimated jointly with other parameters. The standard deviation of hq . The ceteris paribus average monthly effect of the risk on export revenue (mean value of hq . 1.t of 1. The bivariate GARCH-M model in equations (1) to (7) proves adequate for further inferences.Estimated coefficients in the two variance equations are positive and significant.102 calculated from the two series. The bivariate Ljung-Box Q2 (k ) statistics (Hosking.
the exchange rate risk effect dominates the depreciation effect in magnitude. The policy implications suggest that Singaporean authorities can elicit stronger export growth by ensuring a more stable exchange rate rather than by engineering its depreciation.5 percent. (1998) “Exchange Rate Appreciation and Export Competitiveness.S. The mean value (=0. This paper has revisited the weak relationship between exchange rate depreciation and exports in Singapore by using monthly data over the period of 1979-2002. and Yeok. leading to a negative net effect. T. leading to a negative net effect of exchange rate changes on export revenue. Gauging the net effect of the exchange rate movement against the sum of the two average effects. References Abeysinghe. C. we account for the time varying variance of the data. Conclusion Previous research that investigated the responsiveness of exports to exchange rate depreciation generally concluded that exports react increasingly to exchange rate depreciation.” Southern Economic Journal 62. Arize. In sum. Second. employ bivariate GARCH-M modeling technique to estimate the effects of exchange rate depreciation and its risk on exports.effect. Exports: An Empirical Investigation.2 IV. 8 . A. 2 While the exchange rate effect is insignificant. 34-43. T. its magnitude is also small.093) and the estimated coefficient (=0. the effect of exchange rate depreciation on exports is positive but insignificant.02 percent. since the mean growth rate of real exports equals just over 0. (1995) “The Effects of Exchange-Rate Volatility on U. as noted above. supporting the findings of Abeysinghe and Yeok’s (1998). the significant risk effect also dominates. 51-55.” Applied Economics 30. Unlike Abeysinghe and Yeok’s (1998) OLS estimates based on annual data of 1975-1992. time-varying real exchange rate risk exhibits a significant negative effect on exports of substantial magnitude. The Case of Singapore. L.229) of the depreciation rate imply that the average monthly effect equals 0. Third.
.” Econometrica 55.” Review of Financial Economics 6.” American Economic Review 63. (1987) “Estimating Time-Varying Risk Premia in the Term Structure: The ARCH-M Model. (1973) “International Trade and the Forward Exchange Market. (1997) “Foreign Trade and Exchange-Rate Risk in the G-7 Countries: Cointegration and Error-Correction Models. (2003) “Does Exchange-Rate Volatility Depress Export Flows: The Case of LDCs. 700-706. A. M. A. J. A. (1993) “Does Exchange Rate Volatility Depress Trade Flows? Evidence From Error-Correction Models. (1988) “Exchange Rate Variability and The Slowdown in Growth of International Trade. J.” Southern Economic Journal 66. A. (1991) “The Effects of Exchange Rate Volatility on Some New Estimates. C. 10-17. F. 5-59. D. Osang.” IMF Staff Papers 35. 9 . and Kara. A. Malindretos. 498-505. Ethier. 391-407. P. W. F. 178-185.Arize. Exports: Bahmani-Oskooee. and Peel. P. C. Lilien. Bollerslev. (1999) “Exchange Rate Volatility and International Trade.” International Review of Applied Economics 17. (2003) “Relative Responsiveness of Trade Flows to a Change in Prices and Exchange Rate. 7-19. U.” Journal of Econometrics 52. M. De Grauwe. 173-177..” International Advances in Economics Research 9. Engle. Arize. and Slottje. and Kroner K. M. D. and Eckwert.” Review of Economics and Statistics 72.” Journal of Business and Economic Statistics 18. 293-308. R. Arize. 63-84. Chowdhury.” Review of Economics and Statistics 75. K. T. (1992) “ARCH Modeling in Finance: A Review of the Theory and Empirical Evidence. B.. T. Asseery. R.” Economics Letters 37. C. R. 95-112. R. and Robins. Bollerslev. (1990) “Modelling the Coherence in Short-run Nominal Exchange Rates: A Multivariate Generalized ARCH Model. and Kasibhatla. Chou. J. O. T. (2000) “Exchange-Rate Volatility and Foreign Trade: Evidence From Thirteen LDC’s. 494-503. Broll. A. D.
147-165. D. (1999) “Real Exchange Rate Volatility and Sri Lanka’s Exports to the Developed Countries. Wilson P. 117-129. (1979) “Differential Responses to Price and Exchange Rate Influences in the Foreign Trade of Selected Industrial Countries. New York: John Wiley & Sons. H. H. Institutions and Money 7. and Takacs. Pozo. R. S. W. 47–63. R. (2001) "Exchange rates and the trade balance: the case of Singapore 1970 to 1996. Junz.S. D. and Thompson. Rose. 1978-96. F. E. and Lastrapes. Tsay. (1973) “Price Competitiveness in Export Trade among Industrial Countries. K. 817-839. 298-318. Wilson. F. (1980) “The Multivariate Portmanteau Statistic. J. 325-329. R. M.” American Economic Review.” Economics Letters 34. 267-279.” Journal of Economic Development 24. M.” Pacific Economic Review 9. (2004) “Why is it so Difficult to Find an Effect of Exchange Rate Risk on Trade?’ Journal of International Money and Finance 23. K. E. (2004) “Exchange Rate Risk and Export Revenue in Taiwan. and Tsujii. and Rhomberg. R. W. A. Kroner. W. Klaassen.” Journal of the American Statistical Association.” Journal of International Financial Markets. C. R. Ekanayake. and Brooks. S. 75. 271-275. 73-87. McKenzie. (1997) “The Impact of Exchange Rate Volatility on German-U.” Review of Economics and Statistics 74. Weliwita A. H. Papers and Proceedings 63. 10 . J.Fang. (1992) “Conditional Exchange-Rate Volatility and the Volume of International Trade: Evidence from the Early 1990s. (1993) “The Impact of Exchange Rate Volatility on International Trade: Reduced Form Estimates Using the GARCH in Mean Model. F. Trade Flow. (1990) “Exchange Rates and the Trade Balance: Some Evidence from Developing Countries.” Review of Economics and Statistics 61.” Journal of International Money and Finance 12. M.. and Tat." Journal of Asian Economics 12. Hosking. (2002) Analysis of Financial Time Series. D. 602-608. 412-418.
This action might not be possible to undo. Are you sure you want to continue?
We've moved you to where you read on your other device.
Get the full title to continue listening from where you left off, or restart the preview.