when the EGARCH(1,1) model is estimated without any exogenous variable. Once the
impact of news arrivals on conditional volatility is accounted for, we no longer observe
this effect. In addition, the level of volatility persistence declines from 0.982 (0.983) to
0.441 (0.425) for the S&P/ASX 200 Index (the SPI 200 Futures) after inclusion of the
number of news items variable in the conditional variance equation of the
EGARCH(1,1) model.^7 The log likelihood ratio test rejects the null hypothesis of
¼0 at the 1% level of significance. Thus, the inclusion of news variables in the
variance equation of the EGARCH(1,1) model has improved the overall goodness of
fit of the model.
We also examine the potential impact of trading volume on conditional volatility by
including lagged trading volume together with the news arrival variable in the condi-
tional variance equation of the EGARCH(1,1) model. Similar to Kalev et al. (2004), we
find that lagged trading volume has a positive effect on conditional volatility. This
finding is consistent with Blume, Easley, and O’Hara (1994), who emphasize the role
of trading volume in the price discovery process. More importantly, our results show
that the news arrival variable remains statistically significant after controlling for the
impact of lagged trading volume on volatility. For the S&P/ASX 200 Index, the
inclusion of lagged trading volume does not result in further reduction of volatility
persistence. In contrast, for the SPI 200 Futures, the level of volatility persistence
declines from 0.425 to 0.334 when the lagged trading volume is included together with
the news arrival variable in the EGARCH(1,1) conditional variance equation. A poten-
tial explanation for the different evidence regarding the level of reduction in volatility
persistence in index and futures markets may come from the observation that the trading
volume of SPI 200 Futures is based on the actual number of contracts traded, whereas
the trading volume of the S&P/ASX 200 Index is calculated by the dollar value of all
transacted shares of the constituent companies of the index. Alternatively, traders with
firm-specific information may prefer to trade primarily on the spot equity market, while
traders with market-wide information can also conduct their trades on index futures
markets (Subrahmanyam, 1991; Gorton and Pennacchi, 1993; Bessembinder, Chan, and
Seguin, 1996).
Table 12.5 reports the findings of the investigation of the news arrivals–volatility
relation in two subsample periods: October 1, 2003 to November 1, 2007 and November
2, 2007 to September 30, 2009. Consistent with the results for the whole sample period
presented in Table 12.4, the news arrivals variable is positively related to conditional
volatility in the two subsample periods for both the S&P/ASX 200 Index and the SPI 200
Futures. The inclusion of the news arrivals variable greatly reduces the level of volatility
persistence for both markets in both subsample periods. Further reduction in the level of
volatility persistence after the additional inclusion of lagged trading volume is only
observed for the SPI 200 Futures.
Overall, the results presented in Tables 12.3, 12.4, and 12.5 indicate that news arrivals
have a positive impact on conditional volatility. The level of volatility persistence is also
greatly reduced once the effect of news arrivals on volatility is accounted for. The
substantial reduction in volatility persistence, together with the evidence of strong
autocorrelation in news arrivals, as presented in Table 12.2, is consistent with the
Firm-specific news arrival and the volatility of intraday stock index and futures returns 281(^7) Volatility persistence does not disappear entirely, which is probably due to other private information arrival processes not
being accounted for in the number of news arrivals variable.