Research On Arbitrage Strategy Between Crude Oil Futures And Stock Indexes

Statistical arbitrage is a model-driven arbitrage strategy. When the asset price deviates its equilibrium price or predicted price, excess profit can be earned by building long-short positions.In statistical arbitrage, pair-trading is most commonly used. The idea behind pair-trading is the same with statistical arbitrage:when the portfolio price deviates from its equilibrium level, build long-short positions; when the portfolio price reverts to its equilibrium level, settle the positions. When building the portfolio, there must be two assets which should be highly correlated or have similar trends.By building long/short positions, statistical arbitrage becomes a market-neutral strategy, which is often used in hedge funds and also in markets where short position can be built easily. Since the number of financial assets which can be shorted is very limited and the short-mechanism is inadequate, statistical arbitrage cannot be conducted in large scale. Therefore, we would test our statistical arbitrage based on data from foreign markets.There are two key points of success of statistical arbitrage. One is to find two assets with high correlation:long the undervalued asset and short the overvalued asset. The other is to decide the buy/sell signals.Oil is very vital to one country's economy. The price turbulence of the oil will not only be reflected in market index but also be reflected in stock and index of oil-related enterprises. The trend of oil futures is closely related to oil itself. Meanwhile, oil future is price discovering. It can be deduced that the correlation of oil future and oil-related indexes will be very close to but not equal to one. Therefore, we will choose oil futures and oil-related indexes as the two assets to test the arbitrage strategy.In previous studies, there is a large number of articles focusing on the relationship with oil futures and oil indexes. But to build the arbitrage strategy based on this relationship is rare. Besides, articles focusing on the statistical arbitrage strategy are very common, including statistical arbitrage strategies used on Treasury bill, individual stock or agricultural products and so on. But this strategy is rarely used on oil, especially when the oil-related indexes have been categorized into separate industries and have been grouped into upstream indicators and downstream indicators.The oil-related indexes used in this paper include upstream indicators (S&P Oil& Gas Refining & Marketing index, S&P Oil & Gas Exploration & Production index, S&P Oil & Gas Equipment & Service index, S&P Oil & Gas Drilling Index and S & P Oil & Gas Storage & Transportation index), downstream indicators (Dow Jones Electricity Index, S&P Transportation Select Industry Index and S&P 500 Chemicals Industry Index) and market index (S&P 500 Index). We use these indexes to build our strategy.We will build unit-root test, co-integration test, Granger Causality Test and Error Correction Model to testify the equilibrium relationship between index and future. There are three strategies based on this relationship. First-the moving average strategy, trade will be activated when the price spread exceeds its historical average level to some certain extent measured by its standard deviation. Second-strategy based on co-integration relationship, the weight of the assets in the portfolio will be set according to their coefficient in the co-integration equation. Third-modified moving average strategy, the standard deviation will be calculated using GARCH model, since the deviations have been found to have clustering effect.The empirical results show:first of all, there are long-term steadily co-integrated relationships between oil-index and oil futures; second the modified moving average strategy is better than the original moving average strategy which is better than the co-integration based strategy. The modified moving average strategy profit result wins all corresponding buy & hold strategy result. There is one thing to be noticed, though the result of modified moving average strategy is quite outstanding, yet its trading frequency has been reduces, which increases the opportunity cost.