The oil industry's response to new avenues in futures trading

Abstract

We examine the Cumulative Abnormal Returns (CARs) of petroleum, airline and investment banking firms to the announcement and initiation of trading for two new oil-related assets in 2006: West Texas Intermediate crude futures contracts traded via the Intercontinental Exchange and the American Stock Exchange's (AMEX) US Oil exchange traded fund (USO). In general, we find few significant changes, but the changes we find are marginally positive reactions related to the two new contracts. We also find evidence that firms which utilize derivatives benefit less than firms which do not. However, firms which trade derivatives (nonhedgers) have greater returns than nontraders (hedgers).

Acknowledgement

The authors thank Tim Heath for his help in data collection.

Notes

¹ The term ‘peak oil’ has come to be used to describe the point where the supply of oil will reach its zenith (Deffeyes, 2002).

² See Greenberger (2006) for a summary of the history of ICE and its regulation.

³ CBS’ 60 Minutes aired a segment January 2009 titled ‘Did Speculation Fuel Oil Price Swings?’ which includes the information that up to 70% of futures oil contracts are held by speculators.

⁴ The same 60 Minutes segment reports that Morgan Stanley had the capacity to store up to 20 million barrels of oil. Thus, investment banks not only trade oil-related contracts, but often have direct investment in the asset itself.

⁵ We note that the Liao et al. (2008) result is for ICE's conversion to electronic trading for Brent crude prices in 2004 and not for WTI crude prices in 2006.

⁶ We do note that 25 of the 32 γ_i coefficients are positive. Using a nonparametric sign test assuming a binomial distribution, the t-test for this number of positive reactions is 3.18 which is significant at the 0.01 level. Thus, there is evidence that the markets perceive these new products as value producing.

⁷ We also note that 19 of the 24 CARs are positive. Using a nonparametric sign test results in a t-test value of 2.86 which is statistically significant at the 0.01 level.