Volatility In The Natural Gas Market

Volatility in the natural gas market : the impact of high natural gas prices on American consumers : hearing before the Permanent Subcommittee on Investigations of the Committee on Homeland Security and Governmental Affairs, United States Senate, One Hundred Ninth Congress, second session, February 13, 2005, field hearing in St. Paul, Minnesota.

Using daily futures price data, I examine the behavior of natural gas and crude oil price volatility since 1990. I test whether there has been a significant trend in volatility, whether there was a short-term increase in volatility during the time of the Enron collapse, and whether natural gas and crude oil price volatilities are interrelated. I also measure the persistence of shocks to volatility and discuss its implications for gas-and oil-related contingent claims.

This papers tests a theoretical implication of the theory of storage. We examine the U.S. natural gas market over a period of extensive change in market structure brought about by regulatory action. We motivate and test the hypothesis that the change in market structure increased spot price volatility. The theory of storage implies that a shift to a higher volatility state leads to an increase in convenience yield and therefore, an increase in the use of storage. Using a switching ARCH model, which allows the ARCH parameters to be state dependent, we find evidence that volatility of natural gas prices increased with the change in market structure. We also find that the switch to a higher volatility state is associated with investment in additional storage capacity.

We examine the incremental power of a large set of key fundamental, financial, and macroeconomic variables for forecasting the volatility of natural gas futures prices. Among other results, we find that the option implied volatility (IV) significantly improves the performance of predictions regarding the future volatility of the natural gas market. We also identify several fundamental and macroeconomic variables (e.g., open interests, default premiums, and the return of the housing index) which are statistically-significant in a predictive regression even after including the option implied volatility. On the other hand, we do not find a substantial increase in the adjusted-R2 of the predictive regressions (including IVs) after adding significant fundamental variables. The small incremental predictive power of fundamental variables is interpreted as a sign of the efficiency of the options market. We also observe that the adjusted-R2s of predictive regressions are higher when the time-to-maturity of the futures contract increases. Our results can help users of natural gas volatility forecasts (e.g., producers, utility companies, and portfolio managers) make better informed and more efficient operational and financial decisions.