A study incorporating skewness in Expected Shortfall Estimation

Abstract: Expected Shortfall has become a prominent risk measure after the global financial crisis which hit the economy in 2007. This master thesis examines whether Expected Shortfall (ES) estimation gives better estimates when we incorporate skewness and the impact during turbulent versus tranquil period. This specific analysis scrutinized daily total returns (TR) of three Indexes: Standard & Poor 500 (S&P 500), US benchmark 10 year DS GOVT index (BMUS10Y), and S&P GSCI Gold Total Return - RETURN IND (GSGCTOT). The sample for the estimation was from Jan 2000 to end of Dec 2019, which embrace a turbulent as well as a tranquil period. The Value at Risk (VaR) and Expected Shortfall (ES) forecasts was done with different distributions and evaluated with back testing procedures. However, skewed t-distribution elucidated the main research purpose. The empirical results, gives the idea that ES estimates incorporating skewness helps by retrieving better estimates during turbulent period as well as during tranquil/normal period.