What it is
Pairs trading takes offsetting long and short positions in two related instruments and profits when the spread between them narrows or widens, rather than from the direction of either leg. Done properly it removes most market beta and leaves a bet on a relationship.
How it works
Construction begins with selection, historically by distance metrics and now more often by cointegration tests such as Engle-Granger or Johansen, which look for a stationary linear combination rather than merely correlated returns. A regression supplies the hedge ratio, the spread is standardised into a z-score, and positions are opened at a threshold and closed at the mean or at a stop. The hedge ratio needs periodic re-estimation as the relationship drifts.
How traders use it
The appeal is that the strategy can work in any market direction and that its risk is defined by the behaviour of the spread rather than of the index. It generalises naturally to baskets, sector-neutral portfolios and broader statistical arbitrage, where the same logic is applied across hundreds of instruments at once.
Where it breaks down
Relationships break, and they break for reasons no amount of history contains: an acquisition, an index reconstitution, a regulatory change, a divergence in business model. Scanning thousands of candidate pairs for cointegration guarantees false positives through multiple testing unless the search is explicitly corrected for it. The short leg brings borrow costs, recall risk and dividend obligations, and because spread moves are small the strategy is usually leveraged, which turns a modest relationship failure into a serious loss.
Educational reference. This entry describes how a concept is defined and used. It is not investment advice, not a recommendation, and not a signal. Any rule you build from it should be tested with realistic costs before it is traded, and no historical result guarantees a future one.