Spread Trading in Brent Crude Futures - A stochastic approach to intraday calendar spread trading
Abstract
In this thesis, we propose an intradaily spread trading strategy based on a stochastic process model. We go on to examine whether this strategy can be profitably applied in Brent Crude oil futures markets over the Jan-2015 to Apr-2018 period. For this purpose, tick-by-tick trading data of 63 unique Brent Crude oil futures contracts are used to construct intradaily data sets with 5-minute resolution. By considering the 9 most liquid futures contracts and by constructing 18 different calendar spreads for trading, we perform a thorough backtest of the intradaily trading strategy. Under optimistic assumptions, our strategy achieves a maximum Sharpe ratio of 4.3. Under conservative assumptions, however, Sharpe ratios are negative for all parameter choices. We conclude that intraday spread trading in Brent Crude futures based on the stochastic process model put forward in this thesis is not profitable. Although we show that such strategies may be highly profitable under optimistic assumptions, we emphasize that results are very sensitive to small changes in bid-ask spreads and the timing of trade execution. As these model parameters are difficult to estimate correctly without order book data, we conclude that a cautious approach should be taken when implementing these parameters in a backtest.