Alexander Barzykin on the Mathematics of FX Market-Making and Informational Risk

Market-making in foreign exchange presents a sophisticated balancing act, requiring complex optimisation that fundamentally revolves around managing asset inventory and establishing precise bid/ask pricing strategies.

Since FX trading predominantly occurs over-the-counter, dealers quote prices directly to clients or aggregators, establishing bilateral trading relationships rather than participating in a centralised market with a traditional order book structure. This OTC environment means FX dealers face client-specific risks that may be more pronounced than those encountered in centralised trading venues.

Alexander Barzykin, director in the foreign exchange, rates and commodities team at HSBC in London, explores how FX market-makers should navigate informational risk, with particular emphasis on adverse selection and price reading dynamics. The modelling framework he developed with Philippe Bergault, Olivier Guéant, and Malo Lemmel was recently published on Risk.net and formalises the theoretical principles underpinning HSBC‘s production systems.

Barzykin links the informational risk inherent in FX market-making to how dealers optimise their strategies to manage internal liquidity. He explains that when market-makers begin skewing prices, this generates supply-demand imbalance signals across the entire franchise, which certain participants can interpret, thereby creating information risk.

This optimisation methodology was originally introduced in a paper Barzykin co-authored with Robert Boyce and Eyal Neuman, published on Risk.net in April.

Adverse selection emerges when clients possess asymmetric information, whether through superior knowledge or latency advantages that may arise in delocalised markets.

Price reading represents a more nuanced challenge that proves difficult to identify. It describes how a dealer’s risk management inadvertently yet inevitably exposes information about its inventory, which algorithmic traders can subsequently exploit.

A takeaway of the paper is that adverse selection might not be so adverse after all

Barzykin emphasises that neglecting adverse selection and price reading considerations can result in substantial losses or rapid account depletion. He notes that when significant inventory positions are held and excessive information is revealed, prices can drift unfavourably, potentially leading to considerable financial damage.

Barzykin and his co-authors formulated their model as a stochastic optimal control problem, solving it through dynamic programming—an optimisation approach particularly suited for multi-period applications that Barzykin considers his primary analytical tool, having employed the same technique for internal liquidity optimisation in his April paper.

The model incorporates risks arising from adverse selection and price reading, factoring them into price perturbations that can be characterised through stochastic differential equations.

The resulting output provides a strategic framework for price skewing that optimises the inherent trade-off between risk management and information leakage.

One significant insight from the paper challenges conventional assumptions: adverse selection may not prove as detrimental as typically perceived. Barzykin illustrates this by referencing clients who display adverse selection or price reading behaviours based on long-term signals, suggesting that “you can rationally accept some adverse selection from this client if you can potentially use this information to risk-manage the rest of your franchise.”

While the model was conceptualised specifically for FX markets, its principles demonstrate potential applicability across all markets where information risk exists, particularly in OTC trading environments.

Barzykin identifies promising research directions, including the joint modelling of client and dealer optimisation problems, which would necessitate a game theory methodology.

Another avenue for future investigation concerns reputation feedback—the mechanism by which market-maker conduct, such as order rejection rates, shapes future client flow patterns.

Source link