Stock Prices and Exchange Rates in a VEC Model-The Case of Singapore in the 1990s
Wu, Ying, Journal of Economics and Finance
This paper uses an error correction model to explore 1) the asymmetric effects of four different exchange rates on Singapore stock prices and 2) the effects' sensitivity to economic instability. Both the Singapore currency appreciation against the U.S. dollar and Malaysian ringgit and depreciation against the Japanese yen and Indonesian rupiah lead to a long-- run increase in stock prices for most selected periods of the 1990s; however, the effect associated with the U.S. dollar exchange rate has a sign reversal between the 1997-98 crisis period and the 1999-2000 recovery period. The influence of exchange rates on stock prices increases in a chronological order in the 1990s. (JEL E44, G15, F31)
The relationship between stock prices and exchange rate fluctuations is a controversial subject among financial economists though there is a dearth of literature on it. Gavin provided a modified Mundell-Fleming model to analyze stock market and exchange rate dynamics (see Gavin 1989), but the results of empirical studies on the subject have been mixed. Although some studies conclude that there is no long-run relationship between stock prices and exchange rates (Bhandari and Genberg 1989; Bahmani-Oskooee and Sohrabian 1992; Nieh 1996), others find the evidence of a positive relationship (Frennberg 1994; Choi 1995; Ajayi and Mougone 1996; Bahmani-Oskooee and Domac 1997).
According to macroeconomic theory, exchange rates and stock prices are positively related (if the exchange rate is expressed in units of the domestic currency per unit of foreign currency). A fall in the exchange value of a nation's currency stimulates exports and increases the income of export and import-competing industries, thus boosting the average level of stock prices. Nevertheless, the foreign-trade channel needs to be supplemented by the Fisher effect and financial theory in order to explain volatile asset prices and the resulting complex stock price-exchange rate relationship.
A positive relationship between the exchange rate and stock prices may result from a real interest rate disturbance: as the real interest rate rises, capital inflow increases and the exchange rate falls (appreciation); on the other hand, the theory of arbitrage suggests that a higher real interest rate reduces the present value of firms' future cash flows and causes stock prices to fall. However, under an inflationary disturbance, the exchange rate-stock price relationship could be negative: when inflation increases, the exchange rate rises because the domestic currency loses its value not only in terms of the goods and services but also in terms of foreign currencies; higher inflation expectations lead investors to demand a higher risk premium and demand a higher rate of return so that stock prices decrease. Although asset prices could exhibit deviations from the predicted stock price-exchange rate relationship, the consequent dynamic adjustments eliminate profit opportunities as predicted by efficient markets theory and thus restore equilibrium.
The focus of this paper is on exploring the existence of an equilibrium relationship between stock prices and exchange rates in Singapore's assets market and its sensitivity to different currencies and economic and financial conditions. Is there any pattern of asymmetry for Singapore-dollar exchange rates vis-ii-vis different currencies to influence stock prices? If the asymmetry exists, how is it sensitive to business cycle and financial market instability? The empirical investigation intends to extend our understanding of the asset price relationship embodied in efficient asset markets. The rest of the paper is organized as follows. The second section discusses the methodology used in the vector-error-correction model. The third section describes the data and the model setting. The fourth section summarizes the main findings. The fifth section concludes.
This section explains the methodology used to estimate the equilibrium relationship between stock prices and exchange rates. Consider a p-dimensional vector autoregressive model with Gaussian errors
where (Gamma)i = (I,...A^sub I^ - A^sub i^ ), and (i = 1,...,p - 1) and II = -(I - A^sub 1^-- A^sub p^),. The principal difference between model (1) and model (2) is that the time paths of co-integrated variables are influenced by the extent of any deviation from long-run equilibrium as well as by their separate self-feedback pattern plus stochastic shocks and exogenous variables. According to the Granger representation theorem, if II has a reduced rank r < k, then there exist k x r matrices such that II = (alpha)(beta), where a represents the speed of adjustment to disequilibrium while beta is a matrix of long-run coefficients. Thus, the term Beta'Y^sub t-1^, is equivalent to the error-correction term. Johansen's approach centers on estimating the matrix II in an unrestricted form, and then testing whether the restrictions implied by the reduced rank of II can be rejected.
The model consists of six variables: the logarithm of Straits Times price index of Singapore, Lgsti; the logarithm of Dow Jones Industrial Average Index, Lgdji; four bi-lateral exchange rates (all expressed as units of the Singapore dollar per unit of foreign currency) that link the Singapore dollar with the Malaysian ringgit, Indonesian rupiah, U.S. dollar, and Japanese yen, labeled respectively as Sgring, Sgindo, Sgus, and Sgyen. In terms of the focus of this study and the relative sizes of U.S. and Singapore economies, Lgdji is treated as an exogenous variable.
The Straits Times price index (STI) of Singapore came into being in 1999 and consists of 55 firms in various industries. It covers more business areas than its predecessor, the Straits Times Industrial Index (STII), which included only 30 blue-chip industrial and commercial stocks listed in the Singapore Stock Exchanges and was Singapore's most widely quoted stock market index until 1999.1 Exchange rates are measured in terms of different foreign currencies and in different phases of business and financial cycles. Two of the four exchange rates symbolize Singapore's international monetary relations with two world economic superpowers, the United States and Japan; the other two reflect foreign-exchange linkages with two of Singapore's neighboring countries, Malaysia and Indonesia. All five financial series are obtained from the database published by the. Primark Datastream International, Ltd.
This study divides the weekly data of the 1990s into four periods: high-growth period (04/03/91-01/25/95), pre-crisis period (02/01/95-06/25/97), crisis period (07/02/97-12/30/98), and recovery period (01/06/99-05/31/00). As depicted in Figure 1, Singapore's Straits Times Index exhibits a strong upward trend in the high-growth period accompanied by large currency appreciations vis-a-vis the U.S. dollar (17.5 percent), Malaysian ringgit (12.5 percent), and Indonesian rupiah (29 percent). The tight monetary policy in Japan after the burst of asset bubbles contributed to the strength of the yen, however, and the resulting depreciation of the Singapore dollar in turn helped Singapore's exports to Japan and thus its stock prices.
International trade has been the engine of growth for East Asia's expansion in the first half of the 1990s as well as in the previous two decades. However, beginning with the first quarter of 1995, trade slowed down and crashed later in 1996 as world trade growth fell and the real effective exchange rate appreciated; Singapore's export growth became negative in 1997.2 In the pre-crisis period, the momentum of high growth and the robust appreciation against the U.S. dollar came to a halt and the appreciation against the ringgit and rupiah also turned weak, as shown in Figure 1. The stagnant STI also coexisted with yen depreciation during this period when Japan implemented a large economic-stimulus package from 1995 to 1996.
In East Asia, the export slowdown emerged before the onset of the crisis, which intertwined with both the large current account deficits financed with unhedged short-term foreign borrowing positions and financial markets without adequate prudential regulations. The cyclical and structural vulnerability piled up and eventually led to the 1997-98 Asian financial crisis. In Figure 1, the shaded portion in each panel represents the crisis period. Even for Singapore, the least-hit economy in Southeast Asia, the STI plummeted by as much as 8.1 percent whereas the Singapore dollar depreciated vis-a-vis the U.S. dollar by as much as 23.7 percent during the crisis. In contrast, the Singapore dollar appreciated against the Malaysian ringgit (33 percent) and the Indonesian rupiah (82.5 percent) by a much larger extent than in the previous two periods because its two neighboring economies were more damaged by the financial turmoil. Finally, in the recovery period, stock prices soared again and most exchange rates fluctuated around their steady levels. The Singapore dollar has restored its depreciating trend vis-A-vis the Japanese yen as observed in the first half of the 1990s when the economy was growing rapidly.
Following the pattern of the data described above and the estimation procedure discussed in the preceding section, the empirical analysis proceeds in the following steps. In the first step, we estimate undifferenced VAR models of lags 2, 4, 6, and 8 and choose the length of lags using the Akaike information criterion and the Schwarz criterion. The results show that the two criteria uniformly favor the model with two-week lags. Next, all five non-exogenous variables are pre-tested for their order of integration in four different periods using the augmented Dicky-Fuller tests. The results suggest that all the series are difference stationary regardless of whether a drift term and/or a linear time trend is present in the autoregressive processes. It therefore follows that a co-integrating relationship may exist among the variables.
Having confirmed that the data generating processes are I(1) in the level form, we are further interested in determining whether there exists a long-run equilibrium relationship among stock prices and exchange rates, and if one exists, in identifying the relationship. The Johansen co-integration test is used for this purpose. We choose to include a linear trend in the data and a drift term in the co-integration equation. A trend is also incorporated in the co-integration equation in all but the recovery period. For all the four periods in the model, the Lambda^sub max^ test favors the hypothesis that there is one co-integration equation at the I percent significance level, as reported in Table 1. To further support the finding, the Lambda^sub max^ test is also used. The same conclusion is established by the Lambda^sub max^ test at either the I percent or 5 percent level.3
Thus, we start to construct the generic vector-error-correction model with two lags and one cointegrating equation as below.
The model is estimated using the "EViews" software for each and every four periods of the 1990s data.
Table 2 summarizes the estimates of co-integrating vectors (his) for four different periods in the 1990s. The results suggest that there is an asymmetry in terms of the equilibrium stock priceexchange rate relationship with respect to different countries' currencies. All the estimated coeffcients for the effect of the Singapore dollar exchange rate vis-a-vis the Malaysian ringgit are significantly negative and uniformly so across all the four periods.4 It can be inferred that inflationary shocks seem to have played a more important role than real interest rate shocks in the determination of the relationship between stock prices and the Singapore dollar-ringgit exchange rate. But the same finding does not hold for the relationship between stock prices and the Singapore dollarIndonesian rupiah exchange rate. Except for the estimate in the pre-crisis period, which is not significant by conventional criteria anyway, all the estimates have positive signs and are statistically significant. The finding that currency appreciation lowers the country's stock prices supports the predictions based on the foreign trade channel and on the real interest rate disturbances in the Fisher effect. Hence, based on the asymmetric effects of the exchange rate on stock prices between Singapore's two neighboring economies, we posit that the Singapore dollar-Malaysian ringgit exchange rate had more exposure to inflationary shocks than to either the real interest rate shocks or foreign-trade influences in the 1990s, and the opposite holds for the exposure of the Singapore dollar-Indonesian rupiah exchange rate.
A further asymmetry is apparent in the effects of internationally "hot" currencies, i.e., the U.S. dollar and Japanese yen, on Singapore's stock prices. Three of the four estimates for the effect of Sgus on Lgsti are significantly negative, in contrast to the estimates for the effect of Sgyen on Lgsti.5 Although two estimates for Sgyen are below conventional significance criteria, the two others for the second half of the 1990s are comparable with those for Sgus. In particular, in the crisis period, currency depreciation vis-a-vis the U.S. dollar lowers Singapore stock prices but currency appreciation vis-a-vis the Japanese yen leads to lower stock prices. Moreover, there is a sign reversal for the effect of Sgus on Lgsti between the crisis period and the recovery period, suggesting that in the postcrisis recovery period real economic health replaces inflation-expectation-based confidence as the key for the exchange rate to influence stock prices. In fact, in the recovery period, all the exchange rates except that of the Malaysian ringgit register significant positive impacts on stock prices. It follows that even after the crisis the exchange rate between the Singapore dollar and the ringgit, unlike others, remains more sensitive to inflation innovations than real sector disturbances. Finally, all the included trend variables in the co-integrating equations are significant and have correct signs.
Table 3 reports the vector-error-correction estimates for the (Delta)Lgsti equation in all the four periods. The significant negative values of the speed-of-adjustment coefficients show that the co-integration relationship between the STI and four exchange rates is well defined since any deviation from the long-run equilibrium leads to automatic self-correcting adjustments within the system. This finding justifies the vector-error-correction mechanism that Singapore stock prices change in response to the previous period's deviation from long-run equilibrium as well as in response to its own autoregressive terms and stochastic shocks. In the block of autoregressive coefficients in Table 3, the short-run response of the STI to the Singapore dollar-U.S. dollar exchange rate is clearly important during the crisis period and recovery period. Although the long-run effect of the exchange rate vis-a-vis the U.S. dollar undergoes a sign reversal, which implies that the strength of the real sector matters more to the stock performance than the inflation-based confidence factor in the long run, the short-run effects are significantly negative, indicating that the confidence factor remains a key variable in the short run. The significant short-run effects last two weeks with the effect in the first week larger than that in the second. In addition, the effect in the recovery period has much larger magnitude than in the crisis period. These properties reflect the nature of convergency of the system. In addition, the Dow Jones Industrial Average registers a consistently significant impact on the dynamics of the STI in the entire 1990s: the higher the level of Lgdji, the faster Lgsti increases. The significance of the estimated coefficients is further supported by the F test and the likelihood ratio (LR) test as reported in Table 4. In addition, considering the F and LR tests as the Granger causality test on the model that includes additional exogenous variables beyond a constant, it is observed that there exists a unidirectional causality from exchange rates to stock prices. Specifically, the STI can be well predicted by the Singapore dollar-Malaysian ringget exchange rate in the 1990s except in the crisis period, and by the Singapore dollar-U.S. dollar exchange rate since the crisis period.
The findings from innovation accounting support the results obtained from the co-integration relationship and short-run dynamic analysis! With the 10-week forecasting horizon, the results of variance decompositions are reported in Table 5. In the high-growth period, Lgsti explains over 93 percent of its forecast error variance, the preponderance of its own past values, while leaving about 2 percent for each of the Singapore dollar-U.S. dollar exchange rate and the Singapore dollarMalaysian ringgit exchange rate, respectively. In the pre-crisis period, Lgsti explains only 63 percent of its forecast error variance, and the Singapore dollar-ringgit exchange rate explains about 35 percent, more than half of what Lgsti can explain. The shares of the Singapore dollar exchange rates vis-a-vis the U.S. dollar and Japanese yen increase in the variance decomposition of Lgsti during the crisis period, the former accounting for 15 percent and the latter taking around 25 percent. This indicates the increased capital flow and its impact on exchange rates and stock prices in the financial-crisis period. The explanatory power of Lgsti's own innovations continues to fall in the recovery period to only 35 percent of its forecast error variance. During the period, both the exchange rate against the ringgit and that against the rupiah register discernible increases in their shares: 11 percent for the Singapore dollar-ringgit rate and 16 percent for the Singapore dollarrupiah rate. The largest share goes to the Singapore dollar-U.S. dollar exchange rate, achieving 35 percent of Lgsti's forecast error variance.
Figure 2 summarizes impulse responses of Lgsti to innovations in exchange rates. The pattern of movements in the Lgsti series matches most of what has been predicted by the co-integrating equation in each period. Particularly, the impulse response functions stay on the correct side of the zero axis as specified by the co-integration relationship; the position of impulse response function of Lgsti with respect to Sgus switches from the negative territory in the crisis period to the positive territory in the recovery period; the impulse responses of Lgsti to Sgring are consistently negative across the four periods.
This study explores the effects on Singapore stock prices of Singapore-dollar exchange rates vis-a-vis both the world's leading currencies and its neighbor countries' currencies in four different periods of the 1990s: high-growth period (4/03/91-1/25/95), pre-crisis period (2/1/95-6/25/97), crisis period (7/02/97-12/30/98), and recovery period (1/6/99-5/31/00). It appears that Granger causality runs only one way from exchange rates to stock prices. The co-integration analysis suggests that for most of the selected periods in the 1990s both the Singapore currency appreciation against the U.S. dollar and Malaysian ringgit and depreciation against the Japanese yen and Indonesian rupiah have positive long-run effects on stock prices. In contrast to the stock priceexchange rate relationship in the cases of the yen and rupiah, the stock price-exchange rate relationship in the cases of the dollar and ringgit seems to have more exposure to monetary shocks than to real shocks. The same pattern of asymmetry exists between the exchange rate effects of the U.S. dollar and those of the Japanese yen in the crisis period. Furthermore, there is a sign reversal for the effect of the exchange rate with the U.S. dollar on Singapore stock prices: a negative long-run effect of depreciation in the crisis period is replaced by a positive one in the recovery period although the negative short-run effect remains. The sign change suggests that in the post-crisis recovery period it is competitiveness of the real sector rather than just confidence of market participants per se that holds the key for the strength of stock prices.
The exposure of stock prices to exchange rates (endogeneity) increases in the following chronological order: high-growth period, pre-crisis period, crisis period, and recovery period. Beginning in the high-growth period in which stock prices are least endogenous, exchange rates gradually gain their explanatory power for variations in stock prices; the forecasting ability of exchange rates gets greatly improved in the second half of the 1990s when the economy is in or near the crisis compared to the first half of the 1990s.
1 Singapore Stock Exchange has over 450 internationally listed companies and is one of the major stock markets in Asia. The Singapore government has long been endeavoring to develop a regional financial center. Its consistently prudent monetary-fiscal policies have promoted sustained and non-inflationary economic growth and helped attract funds from well-established foreign banks, securities houses, and fund managers.
2 See World Bank (1998), p. 20.
3 There is only one ).. test result, the one for the pre-crisis period, which favors the hypothesis that there is more than one co-integrating equation. But, due to its low significance level of the test, the ) test is less dependable than the 4 test in this case.
4 Statistical significance of the individual coefficients in Table 2 is also confirmed by the likelihood ratio test, which is based on the unrestricted model and the reestimated restricted model. Statistical results are available from the author upon request.
5 The finding here is in contrast to Yu (1997), who argues that there is no significant connection between changes in stock prices and exchange rates in the Singapore market.
6 The system estimation and pertinent coefficient tests are conducted using EViews, indicating that Granger causality runs only one way from exchange rates to stock prices; therefore, to save the space, only the estimates and tests for (Delta)Lgsti equation are reported in the paper.
7 In this study, the magnitude of the correlation coefficient between residuals from the equations is rather small so that the ordering of the variables in innovation accounting is not important.
Ajayi, Richard A., and Mbodja Mougoue. 1996. "On the Dynamic Relation between Stock Prices and Exchange Rates." Journal of Financial Research 19: 193-207.
Bahmani-Oskooee, Mohsen, and Ahmad Sohrabian. 1992. "Stock Prices and the Effective Exchange Rate of the Dollar." Applied Economics 24: 459-464.
Bahmani-Oskooee, Mohsen, and Ilker Domac. 1997. "Turkish Stock Prices and the Value of Turkish Lira." Canadian Journal of Development Studies 18: 139-150.
Bhandari, Jagdeep S., and Hans Genberg. 1989. "Exchange Rate Movements and International Interdependence of Stock Markets." International Monetary Fund Working Paper WP/89/44 (May).
Choi, Jongmoo Jay. 1995. "The Japanese and U.S. Stock Prices: A Comparative Fundamental Analysis." Japan and the World Economy 7: 347-360.
Frennberg, Per. 1994. "Stock Prices and Large Exchange Rate Adjustments: Some Swedish Experience." Journal of International Financial Markets, Institutions and Money 4: 127-148. Gavin, Michael. 1989. "The Stock Market and Exchange rate Dynamics." Journal of International Money and Finance 8: 181-200.
Johansen, Soren, and Katarina Juselius. 1990. "Maximum Likelihood Estimation and Inferences on Cointegration-with Applications to the Demand for Money." Oxford Bulletin of Economics and Statistics 52: 169-210.
Johansen, Soren. 1991. "Estimation and Hypothesis Testing of Cointegration Vectors in Gaussian Vector Autoregressive Models." Econometrica 59: 1551-1580.
Johansen, Soren. 1995. Likelihood-based Inference in Cointegrated Vector Autoregressive Models. Oxford and New York: Oxford University Press.
Liitkepohl, Helmut, and Hans-Eggert Reimers. 1992. "Impulse Response Analysis of Cointegrated Systems." Journal of Economic Dynamics and Control 16: 53-78.
Nieh, Chien Chung. 1996. The Interrelationship among Macroeconomic Fundamental, Foreign Exchange Rates and Stock Prices. Dissertation, Rutgers University.
World Bank. 1998. East Asia: The Road to Recovery. Washington D.C.: The World Bank.
Yu, Qiao. 1997. "Stock Prices and Exchange Rates: Experience in Leading East Asian Financial Centres: Tokyo, Hong Kong and Singapore." Singapore Economic Review 41: 47-56.
* Ying Wu, Department of Economics and Finance, Franklin P Perdue School of Business, Salisbury State University, Salisbury, MD 21801, YXWU@ssu.edu. The author would like to thank three anonymous referees and Joachim Zietz, JEF editor, for helpful comments. A summer research grant by the Franklin P. Perdue School of Business, Salisbury State University, is gratefully acknowledged.…
Questia, a part of Gale, Cengage Learning. www.questia.com
Publication information: Article title: Stock Prices and Exchange Rates in a VEC Model-The Case of Singapore in the 1990s. Contributors: Wu, Ying - Author. Journal title: Journal of Economics and Finance. Volume: 24. Issue: 3 Publication date: Fall 2000. Page number: 260+. © Journal of Economics and Finance Jul 2008. Provided by ProQuest LLC. All Rights Reserved.