Essays in petroleum futures market, convenience yield, and long memory

grantor: University of TorontoThis thesis is a collection of three essays which address some empirical applications of long memory processes with specific interest in financial economics of energy futures market. The first essay 'Evidence of Long Memory in the Petroleum Market' studies evidence of long memory in the energy market using daily and weekly futures data. This essay concentrates on the question of interdependence between crude oil futures and the corresponding products. The empirical results provide strong support for long memory in the energy futures market. The cointegrating relations between crude oil and heating oil futures as well as crude oil and unleaded gasoline futures exhibit long memory, whereas the individual series are unit-root. The second essay ''Convenience Yield, Mean Reversion and Long Memory in the Petroleum Market'' analyzes convenience yields in the petroleum market. The focus of this essay is the behavior of the spot and futures prices over the long run. The implied convenience yield for petroleum and petroleum products is found to be driven by a nonstationary and mean reverting long memory process. The theoretical implication of this finding is established. It is discussed that this might be attributed to the fact that the market is expecting mean reversion in the spot prices. Furthermore, the volatility process and its relation with the mean process and the corresponding direction of causality have been studied in detail. The third essay ''Long Memory and Conditional Heteroskedasticity, A Monte Carlo Investigation'', unlike the first two, looks at the econometrics of the estimators of the long memory process. It evaluates performance of three methods of estimating the parameter of fractionally integrated noise: the exact maximum likelihood estimator ('MLE'), the quasi maximum likelihood estimator ('QMLE'), and the 'GPH' under different realizations for variance.Ph.D