The E ects of Trading Activity on Market Volatility

In this paper we re-examine the question of the excessive implied persistence of volatility estimates when GARCH-type models are used. We consider ten actively traded US stocks and we con rm the already established result in the literature that, when volume traded is inserted in the GARCH(1,1) or EGARCH(1,1) models for returns, the estimated persistence decreases. Since we feel that volume is a ected also by within-the-day price movements and hence it is not weakly exogenous relative to returns, we suggest alternative proxies for trading activities. We assess that the di erence between the opening price and the closing price of the previous day accounts also for most of the persistence in the autoregressive conditional heteroskedasticity.