The relationship between South Asian stock returns and macroeconomic variables

Abstract

This article investigates whether economic variables have explanatory power for share returns in South Asian stock markets. In particular, using data for four South Asian emerging stock markets over the period 1998–2012, the article examines the influence of a selection of local, regional and global economic variables in explaining equity returns; most previous studies that have examined this issue have tended to focus on only local and/or global factors. Important factors are identified by distilling the macroeconomic variables into principal components. Economic activities, real interest rates, real exchange rates and the trade balance represent local factors. Regional factors are represented by interregional trade and regional economic activity while global factors are represented by world financial asset returns and world economic activity. The vector autoregression results suggest that the South Asian markets examined are not efficient. Both local and regional factors can directly and indirectly explain Bangladeshi, Pakistani and Sri Lankan stock returns while the lagged returns of the Pakistani stock market and world economic activity can explain Indian stock returns.

Keywords:

• South Asian stock returns
• local economic factors
• international economic factors
• principal components analysis

JEL Classification:

• E44
• G12
• G15

Notes

¹ India also attracted the bulk of FDI inflows. For example, the World Bank (2013a) reported that India accounted for approximately 85.0% of South Asia’s FDI inflows in 2010.

² However, the APT, which was proposed by Ross (1976), fails to identify both the number and the nature of the relevant factors that are important in explaining returns (Dhrymes et al., 1985). The selection of relevant factors is therefore subjective and an unavoidable problem associated with this area of research (Fama, 1991).

³ PCA also helps to reduce the loss of degrees of freedom and overcomes any problems with multicollinearity due to the correlations between macroeconomic variables (Jolliffe, 1972).

⁴ To calculate these indices, S&P selects stocks based on market capitalization and liquidity, with the objective of capturing 80% of the total stock market capitalization in order to represent the trading reality of each market.

⁵ According to Alexander (2001), the variables should be made stationary before the PCA is applied, otherwise the first PC will be dominated by the input variable with the greatest volatility. Therefore, the growth rates of economic variables were used for the PCA.

⁶ In this article, the effects of the global financial crisis and financial liberalization on South Asian stock markets were investigated by including both regional and global economic variables in the analysis. Alternatively, the effects were also investigated by including two dummy variables representing the global financial crisis and the liberalization; however, their coefficients were insignificant. The results from this analysis are available from the authors upon request.

⁷ The interest rates and exchange rates used in the PCA are nominal rates. However, as the loadings for the interest rates and exchange rates obtained from the PCA have the opposite sign to the inflation rate or growth in MS, the local PCs are termed as ‘the real interest rate’ and ‘the real exchange rate’.

⁸ Hondroyiannis and Papapetrou (2001) argued that although macroeconomic activities affected the performance of Greek equities, a substantial proportion of the variation in returns was not explained by economic variables.

⁹ Mundell (1963) found that the real financial asset return was negatively correlated with real economic activity.