Essays on high-frequency financial econometrics

This dissertation is comprised of three chapters about research on high frequency financial data. In the first chapter, we apply the ACD-ICV method proposed by Tse and Yang (2012) for the estimation of intraday volatility to estimate monthly volatility, and empirically compare this method against the realized volatility (RV) and generalized autoregressive conditional heteroskedasticity (GARCH) methods. Our Monte Carlo results show that the ACD-ICV method performs well against the other two methods. Evidence on the Chicago Board Options Exchange volatility index (VIX) shows that it predicts the RV estimates. While the RV method is popular for the estimation of monthly volatility, its performance is inferior to the GARCH method. In the second chapter, we propose to model the aggregate trade volume of stocks in a quote-driven (specialist) market using a compound Poisson distribution. Trades are assumed to be initiated by either informed or uninformed traders. Our model treats trade volume endogenously and calibrates two measures of informed trading: relative frequency of informed trading and relative volume of informed trading. Empirical analysis of daily volatility estimates of 50 NYSE stocks shows that trades initiated by informed traders increase volatility, while trades initiated by uninformed traders reduce volatility. Yet trade volume and trade size have incremental information for volatility beyond that exhibited in trade frequency. In the last chapted, we propose to forecast the intraday value at risk (IVaR) for stocks using real-time collected transaction data. Transaction data thinned by price durations are modeled employing asymmetric autoregressive conditional duration models (AACD), and the IVaR is then computed using Monte Carlo simulation. Empirical analysis of the NYSE stocks shows that the AACD approach is better than the Dong, Duchesne and Pacurar (2009) and Giot (2005) methods.