Accounting for nonnormality in liquidity risk

It is well-known that returns are not normally distributed. Liquidity costs, which measure market liquidity, are similarly nonnormally distributed, displaying fat tails and skewness. Liquidity risk models either ignore this fact or use the historical distribution to empirically estimate worst losses. We suggest a new, easily implementable, parametric approach based on the Cornish–Fisher approximation to account for nonnormality in liquidity risk. We show how to implement this methodology in a large sample of stocks and provide evidence that it produces much more accurate results than an alternative empirical risk estimation.