Is skewness simply sufficient? Evidence from Monte Carlo simulation on asymmetric asset returns.

The purpose of this study is to explore the impact of skewness in asset return simulations and the effects of kurtosis on forecast precision. We use 9 years of daily returns of 30 stocks and run a Monte Carlo simulation to identify the forecasted returns based on Gaussian and skew normal distributions. We find that the term and precision do not have a relationship and that the use of the skew normal distribution does not improve the precision of the forecast; it in fact leads the kurtosis to drift to the undesirable direction. Further, persistent negative portfolio forecast errors show that both distribution types lead to significant underestimation of asset returns. The results suggest that simply apply designated skewness to normal distribution do not improve the quality of Monte Carlo simulation, and the fourth moment of realized distribution needs to be incorporated in asset performance forecast