Sector Volatility Prediction Performance Using GARCH Models and Artificial Neural Networks
This work addresses the problem of selecting appropriate volatility prediction models for different asset types, which is incremental as it provides specific guidance based on empirical comparison rather than introducing new methods.
This study compared the volatility prediction performance of artificial neural networks (ANNs) and GARCH models across stocks with low, medium, and high volatility profiles in the U.S. stock market from 2005 to 2020, finding that ANNs are best for low-volatility assets while GARCH models are superior for medium- and high-volatility assets.
Recently artificial neural networks (ANNs) have seen success in volatility prediction, but the literature is divided on where an ANN should be used rather than the common GARCH model. The purpose of this study is to compare the volatility prediction performance of ANN and GARCH models when applied to stocks with low, medium, and high volatility profiles. This approach intends to identify which model should be used for each case. The volatility profiles comprise of five sectors that cover all stocks in the U.S stock market from 2005 to 2020. Three GARCH specifications and three ANN architectures are examined for each sector, where the most adequate model is chosen to move on to forecasting. The results indicate that the ANN model should be used for predicting volatility of assets with low volatility profiles, and GARCH models should be used when predicting volatility of medium and high volatility assets.