Using Copulas to Model Dependence in Pairs Trading
Summary
The document introduces copulas as a way to model how two stocks move together in pairs trading. Unlike distance and cointegration approaches, which focus on price gaps or long-run relationships, copulas combine each series’ marginal distribution with a model of dependence. This can represent asymmetric relationships and tail dependence, which may matter during unusually large market moves.
The described workflow fits candidate copulas to paired price data, compares fit statistics such as information criteria and log-likelihood, and uses a selected model to generate trading positions from thresholds. The module also supports sampling and plotting copulas, and names several pure and mixed copula families. These are descriptions of available methods and tools, not evidence of profitable performance. The introduction gives no empirical results or guidance on selecting pairs, setting thresholds, transaction costs, or managing risk, so any strategy still requires careful validation.
Key ideas
- Copulas model dependence between two series while accounting for their individual marginal distributions.
- Distance and cointegration methods do not explicitly model those marginal distributions.
- Tail dependence can represent relationships during extreme joint price movements.
- Candidate copulas can be compared using information criteria and log-likelihood.
- A fitted copula can be used with thresholds to generate pairs trading positions.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.