In a previous report, we discussed cross-sectional mean reversion strategies in equity markets. Pairs trading, which attempts to exploits a temporary mispricing between two securities with a stable relative price relationship, is another type of mean reversion strategy. In this report, we show how you can improve both the selection and trading aspects of a conventional pairs trading strategy.
Fundamental risk models help to identify profitable pairs
Pairs trading strategies typically look for co-integrated relationships between stocks belonging to the same country and sector/industry group. We believe there are superior means with which to capture the degree of “fundamental similarity” between stocks. For example, we show that utilizing a fundamental risk model to identify stock pairs significantly reduces divergence risk, and also improves the average return per pair.
News analytics overlay to further enhance pairs trading performance
Divergence risk increases in the proportion of idiosyncratic risk associated with a pair's constituent stocks. A news analytics overlay which helps to differentiate between price divergence due to news as opposed due to random price movements, significantly improves the performance of the trading strategy by reducing the number of non-convergent trades.
Beyond stock pairs
In looking for potential pairs candidates, we do not have to limit ourselves to stock pairs. We propose a novel method based on clustering and dynamic tree-cutting to systematically identify clusters of stocks as potential constituents for synthetic pairs trading strategies.
Mean reversion II - Pairs trading strategies highlights:
This Deutsche Bank research shows how to improve pairs trading with RavenPack’s news analytics. Their enhanced signal significantly reduces divergence risk and also boosts the average return per pair.
- The percentage of non-converged pairs dropped by about half from 15% to 7%.
- Average profit per pair increased from 2.3% to 2.8% for the European strategy, and from 1.6% to 1.9% for the US Strategy.
- Return distribution becomes more positively skewed.
Below is a pairs example for Sports Direct International plc and Dixons Carphone plc between November 2014 to January 2016.
Overall, the Sports Direct Intl price dropped by more than 40% over a period of two months. Clearly, trading the pair would have realized a loss. However, with access to a real-time news analytics feed, the loss could have been avoided by ignoring pair trades with price divergence supported by negative sentiment and abnormal news volume on either of the two companies (in this case, “Sports Direct Intl.”).Request White Paper