Multiple-Step-Ahead Forecasting of Value at Risk Based on Holt-Winters Exponential Smoothing Multiplicative Method

Abstract

In recent years, the Value at Risk as a useful tool has been able to assist investors and activists in the financial sector to estimate and forecast the amount of risk involved and its management greatly. In this paper, multiple-step-ahead method forecasted Value at Risk of the Auto and bank indices from March 2011 to September in 2016 with confidence levels of 95% and 99%‌. Holt-Winters Exponential Smoothing Multiplicative method that balances the model in level, trend, and season with three parameters is also considered as one of the most powerful members of the exponential smoothing family. To evaluate the accuracy of the aforementioned model, the Kupiec proportion of failure test, Christoffersen independence test and Conditional coverage test are used. Also, the proposed method and the classical method were compared with the Lopez and Blanco & Ihle loss function tests. The results show that Holt-Winters Exponential Smoothing Multiplicative method is reliable and accurate in predicting the risk factors in the stock indices.

Keywords

• Christoffersen Independence Test
• Holt-Winters Exponential Smoothing Multiplicative Method
• Kupiec Proportion of Failure Test
• Lopez Loss Function Test
• Value at Risk