The impact of exchange rates on stock market returns: new evidence from seven free-floating currencies

ABSTRACT

This paper provides evidence of a significant exchange rate effect on stock index returns using data from seven selected countries practicing free-floating exchange rate regimes. This research uses parity and asset pricing theories, thus placing it within the monetary-cum-economics framework for international asset pricing. In this study, we apply a system of seemingly unrelated regression to control for unobserved heterogeneity and cross-sectional dependence. The findings constitute evidence of a statistically significant exchange rate impact on stock index returns across selected countries. These findings can be considered as falling under the arbitrage pricing approach of the international capital asset pricing model of Solnik who also used the parity-theoretical framework on exchange rate determination.

KEYWORDS:

• Exchange rates
• international asset pricing
• arbitrage pricing
• parity factors
• macroeconomic factor

JEL CLASSIFICATIONS:

• F23
• F31
• G12

Acknowledgements

This paper is benefitted from two group of anonymous reviewers and the editor’s constructive comments. We are grateful to the editor and the anonymous referees for their helpful and constructive comments on both revisions, which have improved the quality of the paper. The authors take full responsibility for remaining errors.

Disclosure statement

No potential conflict of interest was reported by the authors.

ORCID

Mohamed Ariff http://orcid.org/0000-0003-2757-9924

M. Ishaq Bhattihttp://orcid.org/0000-0002-5027-7871

Notes

1 A number of theories and models have been developed since the 1960s, for individual asset valuation purposes: Williams (1938) for bond valuation and Harvey, Liu, and Zhu (2016), provided a review on this perspective. Apart from an economy-wide impact, a segmented market impact also exists in the individual stock impact. Ariff and Khan (1998) considered each capital market as consisting of separately priced asset clusters. Accordingly, it is possible to model pricing behaviour away from the embedded practice of analysing individual stocks and towards macro-level markets (the latter was attempted by King (1966), who showed that 52 percent of share price changes are due to macroeconomic factors). Studies on aggregate level are few: De Santis and Gerard (1998) and Dumas and Solnik (1995).

2 In this revised version, the coefficients reported for all variables are from rechecking as pointed out during the review process. This variable was one of them that had a changed size.

3 We followed a generalization of the model to estimate the system of equations so that the standard errors of each equation are made to rely on the actual data by entering the mean equation while any generated regressor appear only as instruments from the first step estimation. This application led to significant improvements in our estimation by deriving standard errors, which are robust to some kinds of misspecifications.