ریسک و بازده ویژه مورد انتظار در بازارهای مرزی: شواهدی از شورای همکاری خلیج فارس

We investigate the pricing of idiosyncratic volatility of seven frontier markets in six GCC countries. We find a significant (marginal) negative relationship between expected returns and lagged idiosyncratic volatility for individual stocks in Saudi Arabia (Qatar) but none in Kuwait and Abu Dhabi. However, when we estimate conditional idiosyncratic volatility either by EGARCH or AR Models, the relationship turns positive. Introducing unexpected idiosyncratic volatility as an explanatory variable to control for any unexpected returns uncovers the true relationship between expected idiosyncratic volatility and expected returns. The evidence turns out to be robust for return reversals and other control variables. Moreover, the pricing of idiosyncratic risk is less evident in higher country governance and seems to be unrelated to the degree of financial development.

In the realm of Capital Asset Pricing Model, CAPM (Sharpe, 1964 and Lintner, 1965) investors are well diversified and hold the efficient mean–variance market portfolio in equilibrium. The CAPM predicts a linear relationship between expected return and systematic risk but assigns no role to idiosyncratic risk. Moreover, the CAPM makes no predictions about the role of idiosyncratic risk in situations when investors are under-diversified. Merton (1987) extends the CAPM and advances a model that assumes investor under-diversification. Merton argues that if investors cannot, or do not, hold the market portfolio they will consider total risk in their asset allocation decisions and therefore require higher returns for stocks with higher idiosyncratic risks to compensate for imperfect diversification.2 Current empirical evidence on the existence of idiosyncratic risk premia in expected returns has not only produced mixed results but also been limited to developed or developing markets. Frontier markets such as those in the Gulf Cooperation Council (GCC) have been widely overlooked despite their growing importance in recent years.3 For example, Ang et al. (2009, hereafter, AHXZ) investigate the relationship between idiosyncratic volatility and expected returns in 23 countries. However, their sample is restricted to developed countries. Brockman et al. (2009) document international support to Fu (2009) findings on US traded stocks, using data across 44 markets, only two of which are frontier markets, Argentina and Pakistan. The authors show a positive relationship between conditional idiosyncratic volatility estimated by EGARCH Models and expected returns.

This paper takes on the important task of investigating the pricing of idiosyncratic volatility by deploying all available data of seven frontier financial markets in six GCC countries. We find a significant (marginal) negative relationship between expected returns and lagged realized idiosyncratic volatility for individual stocks in Saudi Arabia and Qatar but none in Kuwait and Abu Dhabi. However, when we pool stocks, the negative relationship is weak and not robust to return reversals. Controlling for momentum, liquidity, and leverage does not change our findings. Our findings clearly support the view that the low returns to high-lagged realized idiosyncratic risk is a global phenomenon, as it extends beyond the USA and the developed countries, to even frontier markets such as those in the GCC region.