Trade Versus Time Series Based Volatility Forecasts: Evidence from the Austrian Stock Market

Financial Markets and Portfolio Management, 2001, Vol. 15, pp. 500-515 (with Alfred Lehar and Martin Scheicher).

This paper compares commonly used predictors for the volatility of stock returns. The techniques studied are Moving Averages of squared returns, GARCH and Stochastic Volatility models, and the implied volatility. We perform this evaluation for the Vienna market, which has low liquidity compared to other exchanges in Europe, North America or Asia. We use a variety of econometric criteria to assess the forecasting performance. Our primary result is that the ranking of the models strongly depends on which criterion is chosen. Among the models we estimate, no clear winner emerges. The implied volatility is found to contain information which is absent in time series based forecasts. We discuss possible explanations for these results. Based on our findings we suggest practical consequences for the purpose of derivatives valuation and risk management.

button paper2