Comparing the Liquidity-Adjusted Expected Shortfall Models Over High and Low Liquid Stocks Portfolios: Empirical Results on Thailand Stock Market

Abstract: The stylized fact that stock markets are not perfectly liquid propels banks to incorporate liquidity risk in the risk metrics so that market risk can be managed properly. Disregarding liquidity risk can lead to an underestimation of overall risk and substantial losses. This is particularly true in emerging markets where illiquidity problem is more severe than in developed market economies. Liquidity risk can be measured by exogenous and endogenous variables, such as liquidity horizon, bid-ask spread and liquidity discount. This thesis attempts to evaluate the performance of liquidity-adjusted models incorporating exogenous and endogenous variables to estimate the expected shortfall (ES) over high and low liquid stocks portfolios in Thailand exchanges over the last 10 years. The evidences from model estimations, statistical inference and backtesting show that the regulatory liquidity-adjusted ES model, which uses liquidity horizons as liquidity adjustment, is sufficient for the banks holding low liquid stocks (small market capitalization stocks) over the evaluation period. However, all the liquidity-adjusted ES models in this study underestimate the losses on high liquid stocks (large and mid market capitalization stocks). One potential reason is that high liquid securities in Thailand have greater risk in terms of their quicker reaction to market changes and more transactions than low liquid securities, which can be seen in more volatile movements.