Statistical arbitrage also called “Stat Arb” is an algorithmic trading strategy that uses quantitative analysis to exploit price patterns and inefficiencies across securities; either from the same asset class or across markets, to arrive at arbitrage opportunities.
Stat Arb evolved from pair trades, where stocks were matched and paired based on fundamentals, sector-specific instruments or due to market similarities. When the stocks moved out of sync with each other, the underperforming stock was bought and the other was sold/short with the expectation that they would converge over a period of time.
With the advent of arbitrage, statistical arbitrage too evolved to include
Cross Asset Arbitrage-
Comprises of a price discrepancy between a financial asset and its underlying security. Likewise, price inefficiencies between ETF’s and the underlying security are also considered as cross-asset arbitrage.
Example: Stock index futures and the underlying stocks that are included in the index.
Cross Market Arbitrage-
Exploits price inefficiencies in the same security across markets. Comprises of buying the asset in the market where it is undervalued and shorting the asset in the overvalued market.
- Statistical arbitrage relies heavily on the behavior of correlated asset classes to return to the mean, all other parameters remaining the same. If for some reason, the relationship between the two securities changes due to the various micro or macro-economic factors at play, the strategy will result in losses.
Example: In a statistical arbitrage strategy involving two stocks, if one of the company’s file for bankruptcy, the historical ratio between the stocks will be permanently altered.
- There is no predefined time period for the two securities that are out of sync to revert to the historically held mean ratio, and there is every possibility that the underlying securities will continue to remain uncorrelated for an extended period.
- The strategy comprises of going long in the underpriced security and short in the overpriced security at the same time. In many cases, the strategy involves a split second decision and a failure to execute them accurately and on time can sometimes lead to only one leg of the transaction being filled and the other leg open, resulting in a naked position, open to price risk.
- A statistical arbitrage strategy that was successful in the past may not develop in the future too. That’s because financial markets are continuously changing and evolving. In addition, the risk and/or reward involved in trading a security may or may not hold in the case of the other security as well, although they are from the same asset class.
- Generally, trading algorithms are used to identify statistical arbitrage opportunities and large positions are employed to generate substantial returns on tiny price variations. The cost of employing these systems and the capital required to manage these huge positions make it an unrealistic proposition for retail traders with small investments.