Residual Maximum Drawdown in the Korean Stock Market

Abstract

Is there any predictability in the residuals from the pricing model? Does maximum drawdown contain unique information compared to other risk measures? We answer these questions through an empirical analysis of the South Korean stock market. The traditional pricing model calls the portion of the return that is not explained by the factors contained in the model residuals, and explains that the residuals have no information related to pricing. However, recently it has been demonstrated that the information extracted from the residuals has predictability power for the future returns. In this paper, the residuals are extracted based on the Korean stock market and the hypothetical residual-based price was generated in order to compute maximum drawdown. In our experiments, we compare the performances of residual maximum drawdown with those of residual volatility. Residual maximum drawdown exhibits clear predictability of both return and risk, while residual volatility is persistent only as a risk measure. The predictability of residual MDD remains significant for the risk-adjusted returns.