Cointegration
Cointegration describes a long run equilibrium relationship between two or more nonstationary time series. If two series are cointegrated, their spread tends to revert to a stable mean over time even if each series drifts on its own.
Why It Matters
Cointegration is used to identify pairs or baskets that move together over the long term. Short term deviations from the equilibrium can be traded with mean reversion strategies. This concept is central to many relative value approaches.
Testing Methods
The Engle-Granger two step method tests for cointegration between two series. The Johansen test can handle multiple series and identify the number of cointegrating relationships. These tests require careful data preparation and stationarity checks.
Trading Applications
Cointegration supports pairs trading and spread trading. A trader can go long the undervalued asset and short the overvalued one when the spread diverges. Position sizing often depends on the estimated hedge ratio.
Risks
Cointegrated relationships can break during regime shifts, corporate actions, or structural changes. Small sample sizes can produce unreliable estimates. Transaction costs and execution timing can also erode returns.
Conclusion
Cointegration provides a statistical foundation for relative value trading, but it requires robust testing and ongoing monitoring to remain effective.