Information transmission in energy futures markets
Since the mid 1980s the world oil price discovery process has been dominated by two crude oil futures markets: the New York Mercantile Exchange (NYMEX) and London's International Petroleum Exchange (IPE). To date considerable work has been done to scrutinize the degree to which these two markets price efficiently, but little with regard to the way the two markets interact. It is the first attempt, to our knowledge, to investigate the interaction of the two markets. Given that participants in these markets move with relative ease from one market to the other and usually take positions in both of them, prices of these two leading crudes are kept closely related to each other. It is of interest, therefore, to investigate the speed of information. transmission between IPE and NYMEX and, perhaps, identify which market is the true price leader. To carry out this empirical investigation, simultaneous and non-simultaneous trading sessions of IPE and NYMEX are examined separately. Interesting findingsare disclosed. Firstly, non-simultaneous trading sessions of IPE (IPE morning session) and NYMEX are analyzed with univariate and multivariate time series analysis respectively. In univariate analysis, spillover effects in mean returns are found in the IPE morning session from previous day NYMEX trading information, while no information transmission is found from IPE morning session to NYMEX same-day trading. In multivariate time series analysis with a larger data set, estimation using all data available suggests different results from that used in univariate analysis. However, closer analysis on sub-period estimation reveals consistent findings: the results from the first sub-period, which has the same observation data as in the univariate analysis, mirror those from univariate analysis; results from the second sub-period with extended data have a largely different behaviour from the first sub-period. It thus can be implied that the estimated results using all available information are averages of the behaviour of the two sub-periods. This changing behaviour from one sub-period to the next points to a possible structural break between the two sub-periods. Given that there are no significant political forces, such as "oil shocks", taking place during the period under investigation, the changing forces must be coming from the markets themselves. Secondly, the simultaneous trading session of IPE and NYMEX is examined to detect the temporal lead-lag relationship between the two futures markets using 5minute intervals. Results indicate a bi-directional relationship between the two, however the lead ofNYMEX futures is dominant within 5-minute intervals. Further analyses under major news effects both on the supply side and demand side reveal: (1) the two markets move closer when there are major US news events taking place, and IPE is more efficient in information incorporation when there are major news events both on the supply and the demand sides; (2) the lead ofNYMEX is stronger when there are major US events and that of IPE is stronger when there are major supply side events. Finally, intra-day trading activities of IPE are examined using the tick-by-tick transaction data. Empirical evidence from diurnal factor (intra-day seasonality), and from ACD model suggests that the patterns of IPE morning and afternoon durations are distinctively different from each other. These findings suggest that NYMEX has a large impact on IPE trading. Empirical findings in this thesis imply that NYMEX is a leader in the information incorporation process, but the extent of this leadership changes dynamically; under different news effects as well as different time periods. These results would impose significant challenges to regulators, in today's global market, to keep their market competitive as well as prudent. They should also benefit hedgers, who after taking into account their hedging implementation criteria such as liquidity, may be able to benefit from the faster information transmission ability of the leading market by directly taking hedging positions using the leading market contracts. The users most likely to benefit from the above findings are traders, who may be able to take arbitrage profits after taking into account trading costs, borrowing costs, etc.