Why has liquidity dynamics become increasingly important in FX, and what factors can influence liquidity over time?
As events in the Middle East unfolded in March 2026, institutional FX volumes surged across the major venues. Average turnover rose around 22% month-on-month, but headline activity masked substantial variation in available depth, transaction cost and markout profiles. Tradable pairs did not always behave as expected, for example, gold did not behave like a clean safe-haven asset. It first sold off as investors preferred USD liquidity, then later rallied as conflict intensity increased. Oil-related currencies were volatile and investor demand for USD fluctuated as risk changed. In this environment, spot ECN volume hit very high levels as dealers warehoused less risk and flow went to anonymous channels.
Bid-ask spreads widened, quote volatility increased and there were days when major LPs stepped back in metals and oil-correlated currency pairs. You could see the same risk-warehousing dynamic in the Treasury market, where two-year bid-ask spreads were roughly 27% wider in March than in February.
This is why liquidity dynamics matters more than ever to understanding how depth, spread, skew and rejection behaviour change through time and around events. Target internalisation rates now run upwards of 80% for major LPs, so much of the price a client sees reflects an individual LP’s moment-to-moment risk appetite rather than a visible central book. Periods of time that really matter such as economic releases or news headlines appear to keep getting shorter and more concentrated. Assumptions about who can be relied upon to be consistently available on the other side of an order, and at what cost, require ongoing assessment.
What benefits can knowledge of liquidity dynamics deliver for FX trading firms — many of whom may be looking to keep performance more consistent across market conditions?
With LPs aiming to internalise as much as possible, a client’s realised cost on any given day is shaped by a small set of LP relationships rather than by the market as a whole. Maintaining consistent liquidity means knowing how each LP behaves through different regimes and which counterparties can be trusted with different sizes and flow rates.
ECNs, brokers and clients that are able to measure the interaction of flow to liquidity at the microstructural level are best placed to maintain trading relationships in challenging markets. They see widening pricing, clusters of rejections after an event and counterparties whose skew tilts before flow moved against them. These changes in activity are normally framed in a negative light but behaviour and outcomes are not always obvious. An LP may widen a price but greatly reduce reject rates or round trip times. Having detailed analysis of flows, cost and impact can present positive opportunities to both LPs and clients that can be directly measured by pair to potentially suggest opportunities in related instruments.
In what ways are AI and next-generation analytical toolsets transforming liquidity dynamics?
It is no longer economically sensible to react to changes in flow or liquidity on monthly or quarterly reviews. Conventional TCA tends to completely average out the details that can have massive economic impact through surprisingly simple changes to order flow or pricing.
The quantity of data in FX can be overwhelming and there is a temptation to blindly apply AI to extract insights. We believe AI has a central role in analytics but that it will only be robust as part of a well defined, and most importantly trusted, quantitative framework. Firstly, data must be appropriately cleaned and labelled to appropriately sample the data so that features may be extracted and metrics computed. The bedrock of this area is machine learning which is fundamental in compressing large noisy datasets into something that AI frameworks can reliably analyse. Behaviours that matter tend to be short-lived and counterparty-specific: an LP that starts rejecting asymmetrically after losing flow, skews before a headline lands, or quietly halves or doubles a quote refresh rate.
The streams themselves also vary far more than headline spread numbers indicate. In our own tick-to-tick quote analysis there is no single archetype of an LP price stream, some are wide and firm with low reject rates, some flicker between narrow and wide but reject almost nothing; others are tight on screen and high on rejection. Each of these characteristics is best matched with a particular orderflow. None of this is visible without LP-level analytics at tick resolution, which is what FairXchange’s Quote Dynamics module does: tick-by-tick analysis of how each LP behaves around trades and events.
How is new technology enhancing transparency through advanced visualisation while enabling precise FX pricing adjustments based on order flow and volatility?
Transparency in FX has always been a visualisation problem as much as a data problem. The raw data has long existed but the difficulty is that LP behaviour is at least four-dimensional: timing, sizing, skew and rejection. Liquidity dynamics analysis begins to assess how these dimensions interact. March 2026 was a case in point: “spreads widened” misses the more interesting point that they widened asymmetrically by size, counterparty, currency group and even time of day. Visualisation lets a trader see those dimensions in the same analysis.
Traditional metrics such as daily volatility are of limited value when in real time we can process order flow and estimate very short-term volatility. We can then condition on different regimes enabling LPs to adjust spreads and skews to current conditions instead of relying on static settings. Clients can ensure that execution strategies are appropriate to the actual market conditions at the point of order generation. This transparency helps both sides: clients know that sharp flow will be instantly detected and LPs know that their quotes are being monitored at tick resolution in all market regimes.
What impact has the arrival of real-time analytics dashboards had on giving firms clearer views of liquidity depth, spreads, and execution quality across sessions?
For most firms the shift is still under way. Many still rely on monthly or weekly TCA, which is useful for governance and useless for immediate decisions on the trading desk. Even a move to daily dashboards allows depth, spread and execution quality become visible across Asian, London and New York sessions.
Trading in 2026 really underlined what that measurement gap costs. Firms relying on month-end TCA only saw LPs withdrawal in retrospect after the market had cycled through numerous risk-on, risk-off regimes; firms watching depth and rejection rates in real time saw it on the day and adapted. Liquidity monitoring becomes a daily workflow, surfacing short-lived spread & skew changes, rejection clustering and quote update behaviour.
What sort of actionable insights can be delivered by using next-generation technology to empower data-driven decisions?
Insights only matter if they translate into something a desk can act upon. Several areas benefit from direct action. Panel composition: a client observing an LP that underperforms in volatile sessions, or rejects asymmetrically after losing flow, can engage in immediate conversation with the LP rather than observing delayed metrics with a less focussed and out-of-date TCA report. Routing: orders can be steered to the counterparty actually clearing the size, not the one parameterised by last quarter’s results. Timing: understanding the periods around events when liquidity thins lets clients avoid them or work orders appropriately.
A byproduct of analysing liquidity dynamics provides insight into the performance and behaviour of order management and trading infrastructure. Consistent analysis of the interactions of flow and streams can readily surface latency issues and bottlenecks in trading infrastructure. The more timestamps that are collected the easier it is to reconcile the LPs perspective of a trade or reject to that of the client. As well as assisting LPs to price as they aim to, the analysis can surface client trading behaviour such as sweeping or persistently hitting quotes that is not actually intended.
What factors should influence a firm’s choice of suitable provider to partner with to help them capture the benefits that liquidity dynamics is now able to deliver?
The latest AI tools are able to build quite complex analytics with relative ease. Having confidence in them, managing history and maintaining reliability as analytics expand remains complex and fraught with difficulty. Data fidelity is extremely important: is the vendor capturing complete data? Are they able to handle bad data, dropped feeds and do they have the ability to reload and patch gaps? Does the vendor have a consistent, reliable framework for managing evolving and improving analytics, so that today’s results can be placed in historical context?
Consideration should be given to the additional reference data that vendors can provide. This may be truly independent reference data which is unpolluted by customer orderflow, consensus data or peer comparison.
Vendors should ideally be independent as a provider with no stake in the game with a set of LPs or platforms is only incentivised to provide the facts as they are. As the market evolves with new venues and liquidity providers it is then in everyone’s interests to integrate.
It is also worth considering whether vendors can offer multiple integrated products, so that firms benefit both from a simplified procurement process and from reduced compliance overhead. Risk frameworks such as DORA and the UK operational resilience rules are placing increasing scrutiny on each vendor relationship a firm holds. That due diligence should extend beyond the product itself to cover a vendor’s ongoing compliance strategy and funding stability.
Global coverage and integrations are becoming increasingly important as pricing and liquidity vary substantially by venue and time zone. Having pre and post-trade liquidity dynamics analysis that are consistent across venues, LPs and clients allows both sides of each transaction to talk the same language and genuinely match flow to suitable liquidity.
We’ve seen this first-hand with the introduction of consistent benchmarks and liquidity dynamics that will only extend this communication as the markets continue to evolve.