Recently, great strides have been made in predicting volatility in the financial market. However, the so widely used GARCH model suffers from several problems, like the normality assumption which can lead to unreliable estimates and forecasts. In order to solve this problem, it is possible to introduce different new assumptions in the classical GARCH framework. In particular, the GED-GARCH model supposes that the innovation distribution follows a Generalized Error Distribution (GED). In this paper, after a brief theoretical discussion of the statistical model, we show empirically that it is possible to obtain better results using the GED-GARCH(1,1) instead of GARCH(1,1) based on the normality or t-student assumptions.
Improving Volatility Forecasts with GED-GARCH Model: Evidence from U.S. Stock Market
Massimiliano Giacalone;
2019
Abstract
Recently, great strides have been made in predicting volatility in the financial market. However, the so widely used GARCH model suffers from several problems, like the normality assumption which can lead to unreliable estimates and forecasts. In order to solve this problem, it is possible to introduce different new assumptions in the classical GARCH framework. In particular, the GED-GARCH model supposes that the innovation distribution follows a Generalized Error Distribution (GED). In this paper, after a brief theoretical discussion of the statistical model, we show empirically that it is possible to obtain better results using the GED-GARCH(1,1) instead of GARCH(1,1) based on the normality or t-student assumptions.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.
