하방위험을 고려한 포트폴리오 성과 평가

This study divides risk defined as mean preserving spread into loss risk and gain risk. I present risk-adjusted performance indicators based on their ratio and compare them with traditional performance indicators. The empirical results are summarized as follows. First, about 80% of the stocks listed on KOSPI market rejected normal distribution hypothesis through Shapiro-Wilk test. On the other hand, only 2% of those stocks rejected normal-inverse Gaussian(NIG) distribution that is known to capture skewness and kurtosis risk. Second, from cross-sectional correlation analysis between performance indicators and basic statistics of each stock, most of performance indicators found positive correlation with mean, positive correlation with skewness, and negative correlation with kurtosis. This is congruent with moment preferences of DARA(decreasing absolute risk aversion) type investors. Third, it was found that there existed momentum effect in the short-term daily KOSPI market because expected loss decreased systematically according to performance index. These results imply that general risk-reward ratio beyond mean-variance criteria should be developed for portfolio optimization and market efficiency. Post modern performance indicators based on the exact distribution will improve the efficiency of investment and financial market by providing more accurate information about risk-reward ratio.