Predicting stock index volatility: can market volume help?

Abstract

This paper explores a number of statistical models for predicting the daily stock return volatility of an aggregate of all stocks traded on the NYSE. An application of linear and non-linear Granger causality tests highlights evidence of bidirectional causality, although the relationship is stronger from volatility to volume than the other way around. The out-of-sample forecasting performance of various linear, GARCH, EGARCH, GJR and neural network models of volatility are evaluated and compared. The models are also augmented by the addition of a measure of lagged volume to form more general ex-ante forecasting models. The results indicate that augmenting models of volatility with measures of lagged volume leads only to very modest improvements, if any, in forecasting performance.

The impact of trading volume, number of trades and overnight returns on forecasting the daily realized range

“…This striking lack of explanatory power of the number of trades and transaction volume is somehow surprising in the context of the broadly significant lead-lag correlations as described above. However, it is consistent with the results of some extant studies on the stocks' trading volume (Brooks, 1998;Fuertes et al, 2009).…”

An EGARCH-BPNN system for estimating and predicting stock market volatility in Morocco and Saudi Arabia: The effect of trading volume

“…In other words, volume modestly improves the forecasting accuracy of the BPNN. This finding is consistent with Brooks (1998) who used GARCH process to estimate and predict future volatility.…”

How to fly to safety without overpaying for the ticket

“…test show the significantly higher efficiency of neural network methods than NAÏVE. This is consistent with Hamid and Iqbal (2004) and Brooks (1998) who show significant economic benefits in using neural networks to forecast market volatility. Furthermore, when comparing the base hybrid prediction model combining GRU with GARCH (ALL-VARIATES) with the pure GRU model (NO-GARCH) the errors of the hybrid model are smaller than in the single models.…”