Statistical Arbitrage is a class of short-term financial trading strategies that employ mean reversion models, similar to a pairs trading or relative value strategy. This involves broadly diversified portfolios involving hundreds to thousands of securities. Holding periods can vary widely from incredibly short durations of more than fractions of a second, to a few days or even longer in rare instances. The strategy that profits from price discrepancies between identical or similar securities compared to the expected future value of the assets. Stocks, options, futures, FOREX commodities and CFD‘s are all products that can be used in this form of trading. Stat arb techniques can also be applied to merger arbitrage trading and other event driven investment strategies.
These strategies are characterized and supported by substantial mathematical, computational, and trading platforms, relying heavily on quantitative models to generate trading decisions. It involves data mining and statistical methods, as well as the use of automated trading systems. Stat Arb algorithms monitor financial instruments that are historically known to be statistically correlated or cointegrated, and any deviations in the relationship indicate trading opportunities. Statistical Arbitrage includes different types of strategies such as pairs trading, index arbitrage, basket trading or delta neutral strategies. These strategies vary depending on number, types, and weights of instruments in a portfolio and its risk taking capacity.
Like all trading strategies, statistical arbitrage has its risk. It depends heavily on the ability of market prices to return to a historical or predicted normal, commonly referred to as mean reversion. Even if the stocks used operate in the same industry they can remain uncorrelated for a significant amount of time due to both micro and macro factors. The costs of holding this positon may erode profitability even if the trade itself produces a positive return. Most statistical arbitrage strategies take advantage of high-frequency trading algorithms to exploit tiny inefficiencies that often last for a matter of milliseconds. Therefore, large positions in both stocks are needed often involving large amounts of leverage to generate sufficient profits from such minuscule price movements adding additional risk to statistical arbitrage strategies.
Statistical Arbitrage Example
Statistical arbitrage strategies are market neutral as they involve both a long position and short position taking advantage of inefficient pricing in cointegrated securities. A common example is to compare Coca-Cola and Pepsi. If a trader believes one is overvalued or undervalued, the trader can initiate a simultaneous long and short position in the stocks.
Statistical Arbitrage is often abbreviated as Stat Arb or StatArb and is commonly associated with algorithmic trading. For more information on this subject, please refer to Ernie Chan‘s book Algorithmic Trading: Winning Strategies and Their Rationale.
