The measurement and transmission of volatility in financial markets: Evidence from metal futures markets.

摘要

The measurement and forecasting of asset-price volatility plays a critical role in the study of financial markets. This dissertation verifies the importance of using the integrated volatility using Fourier transformation (IVFT) measure to estimate integrated volatility efficiently. Consequently, studies of volatility that ignore intraday returns series and the IVFT measure are likely to yield misleading conclusions. The IVFT measure and the information provided by high-frequency returns are valuable to a broad range of issues in financial markets. The dissertation provides strong evidence based on the multi-chain Markov switching (MCMS) model of the interdependence, but no comovements between, the three metal markets, which is critical information for portfolio management, derivative pricing and economic policy making.;The dissertation makes a comprehensive comparison of three volatility measures: daily absolute returns, cumulative intraday squared returns, and integrated volatility via Fourier transformation (IVFT). The comparisons are made using intraday futures price data for the time period 1999 through 2008 for three metal markets: gold, silver and copper, at four frequency intervals: 1 minute, 2 minutes, 5 minutes and 15 minutes. The forecasted volatility from a GARCH model is used as a baseline to evaluate the performance of the three measures of volatility. The principal findings of the study are: (A) using heteroscedastic root mean square error and loss function criteria, the IVFT measure better fits the GARCH predictions of volatility than either the daily absolute returns or the cumulative intraday squared returns measures. In addition to this, the goodness of fit of the IVFT measure to the GARCH forecast of volatility improves as the time frequency increases from 15 minutes to 1 minute. (B) Using a multi-chain Markov switching model, the study shows a spillover and interdependence between gold futures, silver futures and copper futures, but there is no comovement between the three metal futures markets during the study period. The distinguishing feature of this dissertation is providing evidence of an accurate measure of volatility using the Fourier transformation which is crucial for accurate forecasting of volatility. For risk and portfolio management, the dissertation provides useful results, including the fact that one of the three metal markets is sufficient as a hedge against inflation or reducing risk.