Has pairs trading outgrown the equity desk?

• Pairs trading has evolved beyond equities. Modern relative value strategies increasingly span multiple asset classes, including equities, futures, fixed income, FX, and commodities, creating new execution and operational challenges.
• Execution is now a strategic differentiator. Cross-asset pairs introduce complexities such as multi-leg coordination, legging risk, and cross-currency exposure, requiring more sophisticated execution capabilities than traditional equity pairs trading.
• Infrastructure must support systematic, cross-asset trading. As more pairs strategies originate from quantitative models, firms need multi-asset, broker-neutral, API-driven execution platforms with integrated FX hedging and the ability to manage complex multi-leg trades seamlessly.

For most of its history, pairs trading has been all about equities. A trader identifies two correlated names, goes long one and short the other, and waits for the spread to converge. The logic is clear, the execution is contained within a single asset class, and the operational demands are modest enough that a capable desk can run the strategy on tools built for directional equity flow.

That version of pairs trading still exists, but it is no longer the version that defines where the interest, or indeed the difficulty, now sits.

Over the last few years, relative value and market-neutral strategies have attracted more institutional attention, and with good reason: rate uncertainty, wider performance gaps between sectors and regions, and continued appetite for returns that do not simply track the market. These factors have pushed capital towards strategies based on the relationship between instruments rather than the direction of any one of them. 

The modern relative value book rarely confines itself to two equities. A desk might trade a precious metals future against the corresponding spot position, one type of bond against another, or an FX pair as a hedge leg against an underlying exposure in a different currency entirely. The strategy is still recognisably pairs trading in its intent, capturing a relationship rather than a direction, but the instruments now span futures, cash, FX and fixed income, often within a single structure.

That is where the execution problem begins. A single-asset equity pair is, operationally, a relatively forgiving thing to trade. Both legs execute on venues the desk already knows, clear through familiar channels, and settle in the same currency. A cross-asset pair is more challenging. The legs may sit in different asset classes, on different venues, with different liquidity profiles and different settlement conventions. The relationship the trader is trying to capture only holds if the legs are executed in proper proportion and in reasonable time.

Several problems follow from that, and they compound.

The first is multi-leg coordination. If the two sides of a trade are not worked in the correct ratio and roughly together, the position the desk ends up holding is not the position it intended. Coordinating that across asset classes, where fill rates and available liquidity differ leg by leg, is materially harder than working a two-name equity spread.

The second is legging risk. The window between the first leg filling and the second leg completing can lead to unwanted directional exposure. On a fast-moving spread, that gap can erode or erase the edge the trade was designed to capture. The wider the asset classes involved, the greater the scope for the two legs to move out of step during that window.

The third is cross-currency exposure. Once a pair spans currencies, the desk has taken on an FX position whether or not it intended to. Left unmanaged, that exposure sits on top of the intended relative value position and muddies the return.

None of these problems is new on its own. What's changing is how routinely a single strategy now runs into all of them together.

Alongside the spread into more asset classes, there has also been a change in where the decision to trade is actually made. On a growing number of systematic desks, the instruction to put on a pair does not originate at the execution management system at all. It originates upstream, in a model, and arrives at the execution layer already formed.

This is a meaningful shift for anyone specifying execution infrastructure. Accepting model-generated intent programmatically is now standard functionality; most execution systems can do it. What matters more is what happens to that intent once it arrives: whether the pairs logic underneath can hold together a spread that crosses instruments, currencies, venues and geographies, or whether it was only ever built to work two equities against each other. The connection between the model and the execution layer becomes part of the trading architecture.

Put those pressures together and a clear set of requirements emerges for any desk trying to run cross-asset relative value at scale.

Multi-asset coverage in a single execution environment matters, because a pairs capability that only understands equities forces the desk to stitch the other legs together elsewhere, reintroducing exactly the coordination and legging risk that the tooling is supposed to contain.

FX auto-hedging is key. If a system can recognise the cross-currency exposure a pair creates and hedge it automatically as part of the workflow, the desk is spared a manual, error-prone step that otherwise scales badly with volume.

Broker neutrality is another critical factor. A desk running relative value across asset classes needs to reach the right liquidity on each leg, and an execution layer that ties it to a particular broker or a narrow set of venues works against that. Neutrality at the execution layer keeps the routing decision where it belongs, with the desk.

And API-driven execution is increasingly a baseline expectation rather than an advanced feature. If the strategy originates outside the EMS, the execution layer has to accept that flow programmatically and act on it faithfully. A system that cannot be driven cleanly by an upstream model is, for a systematic desk, already a step behind the way the work is done.

This is where execution infrastructure becomes part of the strategy itself. For desks running cross-asset relative value at scale, the requirement is not just an EMS that can execute individual orders, but an environment that can keep related legs coordinated, manage currency exposure and accept model-driven instruction without forcing the workflow back into manual handling.

The value is less in any single feature than in the combination: a desk can express a cross-asset relative value view and have the infrastructure hold the legs together, manage the currency exposure and accept instruction from a model upstream, without assembling that capability from separate parts.

As more relative value flow becomes systematic and more of it spans asset classes, the connection between the model that forms the trade and the layer that executes it will become increasingly important.

The pairs trade itself is an old idea. The infrastructure now being asked to carry it is not, and the desks that treat execution as part of the strategy rather than a downstream chore are the ones best placed for where the future of pairs trading is heading.