Pairs Trading Limits and the Shift to Portfolio-Level Statistical Arbitrage
Summary
The article examines practical limits of traditional market-neutral pairs trading. Each trade consumes capital on two legs, incurs spreads and commissions on both, and may use capital on a fairly valued leg even when the opportunity is concentrated in the mispriced asset. These costs and buying-power demands can restrict how many selected pairs a trader can hold, leaving signals unused. Overlapping stocks across pairs also contain information that isolated pair trades do not aggregate.
Trading only the apparent mispriced leg could reduce costs and improve expected returns, but it adds market exposure and raises return variability. The proposed alternative is to flatten pair relationships into asset-level signals, combine signals across many pairs, and construct a portfolio that seeks to capture the mispricing while managing exposures and transaction costs. This broader statistical-arbitrage approach may support additional signals and risk models, but increases system complexity and data-engineering demands. The article outlines the trade-offs rather than providing portfolio construction rules or empirical performance evidence; its promised implementation details are deferred to a later installment.
Key ideas
- Traditional pairs trading uses capital on both legs even when the perceived mispricing may be concentrated in one asset.
- Trading two legs increases transaction costs and can limit the number of pairs held under buying-power constraints.
- Signals from overlapping pairs may reveal asset-level information that isolated pair trades fail to combine.
- Trading only the mispriced leg can reduce costs but increases market exposure and return variability.
- A portfolio-level approach aggregates asset signals and manages exposures across positions, at the cost of greater complexity.
Tags
This summary was written by Stratmill's research agent from the original; it is not a copy of the source.