The new intelligence layer: Harnessing FX Options data for better trading decisions

FX options data is valuable because it reveals what market participants expect and how they are positioning for future currency moves, rather than simply showing what has already happened in the spot market.

Aspects of this data that are particularly useful include:

• Forward-looking information – options prices incorporate expectations about future exchange rate levels and volatility, which can provide clues about where traders think currencies could move over the coming weeks or months
• A measure of expected volatility – implied volatility from options prices shows how much movement the market expects, with a sharp rise a possible signal that investors anticipate greater FX uncertainty
• Insight into tail risks – options can reveal how concerned investors are about extreme currency moves with measures such as risk reversals showing whether demand is greater for protection against a currency rising or falling
• Evidence of hedging demand – corporate treasurers, asset managers and banks use FX options to hedge currency exposures so changes in option volumes and pricing can indicate where businesses and investors are becoming more concerned about FX risk
• Identifying market positioning – if options activity becomes heavily concentrated around particular strikes or maturities, it can indicate where investors are positioning or where significant hedging needs exist
• Early warning of market stress – options markets can sometimes show rising concern before it becomes obvious in spot FX (for example, a sudden jump in implied volatility or demand for downside protection can signal that investors are preparing for a significant move)

Options also reveal information that spot and forwards don’t provide. So while a spot rate indicates the current price and a forward rate the market’s agreed future exchange rate, options additionally tell market participants about the distribution of possible outcomes – including how much investors are willing to pay to protect against unusually large moves.

In addition, a CFO or treasurer can use options data to assess the cost of hedging future revenues, costs, debt or investments denominated in another currency.

Data drives analytics

FX options data is extremely valuable to options traders for their pre- and post- trade analytics and to FX traders of all types for insight into market expectations of future market moves, observes Chris Povey, executive director, head of FX options & NDFs at CME Group.

“High quality real-time and historic options market data has traditionally been hard to come by, particularly for the buyside,” he says. “Price discovery was driven by activity in the inter-dealer broker market, where data was limited only to participating dealers, delivered via a chat or GUI and generally not easily accessible for analytics.”

Electronification has fundamentally changed this landscape, with many participants now consuming options market data via API and applying increasingly sophisticated pre- and post-trade tools to extract useful insights.

This in turn has significantly increased the demand for high quality, transparent FX options market data, adds Povey.

“FX options data has evolved from execution reference into forward-looking intelligence that guides tactical decision making across the global FX ecosystem,” he says. “For example, real-time streaming prices in our short-dated options provide critical data for event-driven hedging for all types of FX traders, especially around elections, central bank meetings and major economic releases. Even those not trading options use the data to identify events that are anticipated to be market moving.”

Systematic buy-side firms have long used machine learning trained on CME’s FX options CLOB for signal generation, valuing its transparent and continuous real-time market data. However, the emergence of publicly available generative AI has significantly democratised the ability to extract actionable signals, which in turn has increased the demand for clean, reliable sources of high quality FX options market data.

Buy side firms also use CME Group’s data for viewing intraday moves across the FX volatility surface without relying on manual RFQs or indicative quote feeds. Systematic strategies inform their algorithmic execution systems, while macro strategies use it to gauge the competitiveness of pricing from their dealers. Historic datasets are critical for back-testing new options strategies. “With increasingly advanced AI powered models informing trading strategy and execution decisions, the source and quality of FX options market data has never been more important,” says Povey.

Critical management input

FX options data is one of the cleanest real-time reads on probability, risk appetite and market expectations but it is also a critical input for execution and risk management. In an era of frequent event-driven shocks, geopolitical tensions and policy pivots, the term structure and skew capture the expected range of outcomes and the asymmetry of those outcomes.

“What makes it more interesting now is that the market is growing rapidly but the data picture is still incomplete,” explains Chris Jackson, CEO & co-founder OptAxe. “FX options data is a mirror of the OTC market itself in that the data and expertise remain fragmented across individual institutions. There is no uniformity – data retrieval and interrogation for end-users is not straightforward. This is a challenge and opportunity because that data therefore becomes a critical differentiator for participants.”

The most relevant emerging use case for FX options data across different functions and FX workflows from OptAxe’s perspective is liquidity observation.

In a bilateral, OTC marketplace, liquidity is never transparent and the cost of obtaining liquidity is market footprint. The observable data is therefore a critical component of pre- and post-trade workflows but it is still largely absent for FX options.

“When [Microsoft CEO] Satya Nadella said customers essentially pay twice to access AI – once with money and once with something even more valuable, their proprietary knowledge – it really struck a chord with us,” says Jackson. “The same conflict exists in asking multiple LPs for an OTC FX option price.” 

In this scenario, participants pay both a financial cost and an information cost to obtain liquidity but only the first is routinely measured. Protocols such as book crossing or mid matching have existed in other asset classes to reduce the information cost but they only work if customers have real-time observable and reliable FX options data.

“That is interesting to us and to our clients because it leads to emergent and improved execution protocols in FX options,” says Jackson.

FX options data provides a window into what the market expects next, not just what has already happened. While spot market tells you where currency prices are today, the options market embeds information about volatility, future risks and investor positioning.

Metrics such as implied volatility, skew and positioning across strikes and maturities provide context that is difficult to find elsewhere, explains Oleg Shevelenko, head of pricing and execution product at FXGO. “As more options trading moves onto electronic platforms, the volume and quality of this data continue to improve, which makes it one of the sharper reads available on shifting market sentiment.”

Ever-expanding role

He notes that the role of options data has expanded well beyond the options desk. Traders use it to inform execution decisions and pricing in real time, while risk teams rely on it to identify concentrations of exposure building across currencies or maturities before they turn into a real risk. Portfolio managers are also making greater use of the data, particularly when assessing different hedging scenarios or modelling how a prospective trade could affect the wider portfolio.

“Rather than serving as a record of past activity, options data is increasingly part of the decision-making process itself.”

Shevelenko agrees that the increasing electronification of trading is changing the way FX options is obtained and utilised. Until relatively recently, much of the FX options market traded voice, so large parts of the trading process were difficult to capture in any meaningful way.

“It is worth noting that while the options segment is making meaningful progress towards electronification, it is not there yet,” he says. “But where electronic execution has taken hold – whether through RFQs, automated workflows or APIs – that activity is now creating structured data in real time.” 

The result is a much clearer picture of market behaviour before, during and after execution. Firms have better visibility into volatility, pricing dynamics and positioning than they did even a few years ago and that has improved both trading decisions and post-trade analysis.

As more of the options workflow becomes automated, from RFQ submission through to execution and post-trade confirmation, the data generated at each step becomes structured, time-stamped and machine-readable in a way that voice-traded flow simply cannot match.

“That is what makes electronification so consequential for analytics,” explains Shevelenko. “It is not just about faster execution, it is about creating a data trail that firms can actually learn from.”

In the meantime, AI is helping firms make better use of increasingly complex datasets. Options markets generate huge amounts of information across strikes, maturities and volatility measures and AI is well suited to identifying relationships that would be difficult to spot manually.

It is also proving useful in turning unstructured information – such as pricing shared through chat or other communication channels – into data that can be analysed alongside electronic trading activity. “This is an area of focus for Bloomberg where NLP is being used to streamline pricing information for FX instruments,” observes Shevelenko. “There is also a broader point about what APIs make possible in this context. When options execution is fully programmable, the feedback loop between data and decision-making becomes much tighter. A systematic fund can run a model, generate a signal, submit an RFQ, receive competing prices, execute and feed the outcome back into its risk system – all without manual intervention.”

AI helping traders

Shevelenko believes AI is most powerful when it is operating within that kind of closed loop, where the inputs and outputs are clean and structured rather than fragmented across voice, chat and email and that the real value is not replacing traders’ judgement but rather helping them identify meaningful shifts in market behaviour earlier and with greater confidence.

Jackson suggests that electronification is not just changing execution but creating new datasets. “A market that used to run on voice and chat generated substantial information but little of it was structured and reusable. Electronification in FX options should be seen as a data ownership opportunity, not just an automation opportunity.”

It is also bringing new entrants into view as more non-bank market makers enter FX options. Their approach to risk modelling and trading is very different to traditional banks, creating new and differentiated sets of data.

“The nuance is that without centralised clearing or a centralised tape there will always be opacity and competing data sets, so we need to make customers’ lives easier within the existing constraints,” adds Jackson. “We need to offer contextual and permissioned transparency.”

He describes genuine transaction cost analysis for FX options as an absolute minimum that will be achieved soon.

“What interests us now is the next generation of toolsets. Vol surface crossing, smart contracts, pre-trade capital pricing, tradable event weightings – there is more surface area to look at once you have the data. This is where the prediction markets have been so fascinating. Trading market products on events with implicit volatility and binary outcomes may shift to trading explicit event outcomes with freely observable probabilities. That will require re-engineering of risk models and adding new product sets, all of which will feed back into FX option pricing.”

The real value of AI lies in pattern recognition across a very high dimensional dataset: regime shifts, unusual volatility shapes, misaligned wings, response optimisation and the extraction of structured information from large, unstructured data sets. But that is not new, it has just been democratised. AI in that sense has dispersed knowledge across the market more evenly.

“In day-to-day workflows, we already see a lot of embedded AI tooling at the point of trade – chat bots, data retrieval, surface analytics, smart routing, etc,” says Jackson. “These are useful but feel very much like first generation use cases taking away manual workflows. Next generation use cases will be where the real ROI comes. Interrogating complex cross-asset vol surfaces, supporting trade selection, surfacing liquidity discreetly and building learning loops back into execution.”

Supportive intelligence layer

FX options data is becoming an intelligence layer that supports decision making across the wider FX ecosystem but Jackson reckons it is still far from perfect, noting that once equities data became standardised, machine-readable and relatively easy to obtain, it stopped being used only by traders and became an input into algo trading, TCA, smart-order routing, portfolio construction, market replay and trade surveillance.

“The risk ecosystem will continue to grow in complexity and breadth,” he says. “New information is available now from crypto, prediction markets, perpetuals and others and this has increased both the input and output from cross-asset risk assessments. FX options data is just one input of many in this process but it remains one of the harder ones to independently assess.”

“The layer will get richer as more of the workflow is captured electronically on regulated venues and that standardisation should compound in value. Standardisation should not be confused with the destruction of value, though. In equities, standardising the core dataset created the foundation for increasingly valuable analytics, execution tools and derived products and FX options can follow the same path.”

On the question of what separates firms that simply consume market data from those that systematically extract intelligence from it, Jackson recognises that there will always be asymmetry.

“A well-known market-maker has 25,000 top-end Nvidia GPUs in Finland operating at the cutting edge of latency, statistical modelling and research. The advantages in being able to consume, interpret and extract intelligence at scale are unprecedented for these types of organisations but there is a wide distribution of consumers of market data and AI is allowing them to treat their own workflow as a data-generating asset.”

Every enquiry, response, pass, execution and subsequent outcome can become part of a governed institutional dataset. According to Jackson, that workflow ownership is becoming one of the most strategically valuable assets in electronic trading and that is where the intelligence layer is added.

Options data was once largely confined to specialist desks but that is becoming less true as firms take a more integrated approach to managing FX exposure. Signals from options markets are now influencing decisions around hedging, execution and portfolio construction across a much broader set of teams.

According to Shevelenko, fragmentation across venues and data sources is a continued challenge. “The value of having data, liquidity and analytics in a single, integrated solution grows each time the market adds another venue or API. This is where APIs are quietly reshaping the intelligence layer in FX options.”

Making pricing accessible

Historically, the multi-dealer dynamic that gives buy-side firms pricing competition was only accessible if you had the operational infrastructure to manage multiple bilateral connections or multiple single dealer portals. A single multi-dealer UI or API framework simplifies that by broadening access to competitive pricing while simultaneously generating richer, more comparable data across dealers and execution styles.

“Over time, that data becomes the intelligence layer itself,” he adds. “Firms can see not just where they executed but how different dealers performed across market conditions, tenors and structures. As the underlying data becomes easier to access and integrate into existing workflows, it is evolving from a specialist resource into something that supports decision-making across the wider FX business.”

Shevelenko says the difference between firms that simply consume market data and those that systematically extract intelligence from it lies in how the data is used, noting that some firms treat market data primarily as a reference point for reporting, reconciliation or compliance while others build it directly into the way they trade, manage risk and allocate liquidity.

“In those organisations, data is not something that is reviewed after a decision has been made,” he continues. “It is one of the inputs shaping the decision itself. Capturing that data comprehensively matters too, including from voice workflows where trades are often booked after the fact with timestamps that do not reflect actual execution.” 

Shevelenko reckons firms that can wrap an electronic layer around their voice activity – preserving timestamps, spreads and counterparty data to the same standard as an electronic trade – are the ones that will get the most out of their TCA and dealer selection tools. “That depends on having the right technology but just as importantly it requires an organisation that sees data as a competitive advantage rather than an operational by-product.”