Estimation of Expected Shortfall

Our paper emphasizes on estimating extreme losses caused by heavy-tail behaviour of financial data by the indicator of expected shortfall as well as underlying the significance of using appropriate modeling methods. It is common that most research assumes financial data to follow a normal distribution, but this is generally not the case as many empirical data shows heavy-tail behaviour. This might lead to an underestimation of indicators in the sector of risk management. Trying to alleviate the potential harmful effects, we analyze the tail behaviour of empirical data by using normal distribution and power law distribution to fit it through both computing and direct calculation, and then estimate the expected shortfall through derived formula. Indeed, the accuracy and efficiency of power-law distribution is admitted by conducting backtesting on the estimation of expected shortfall of several chosen stock in a time period of two year from 2021 to 2023.