Brett Chappell: Who owns the bond trading workflow?

74

The Next Prize: Knowing What Traders Will Do Before They Do It

As ICE pursues MarketAxess and Tradeweb moves to acquire Trumid, the battle for fixed-income infrastructure is becoming a battle for trading intelligence.

What if your trading platform understood your execution habits better than you did? It already records which dealers you approach, which protocols you favor, and how quickly you respond to prices. Combine those observations with historical execution data, market conditions, and dealer performance, and a sufficiently sophisticated system could begin to anticipate how you might trade under different circumstances, perhaps even before you make the decision yourself.

That possibility introduces another dimension to the recent consolidation of fixed-income infrastructure. ICE’s proposed acquisition of MarketAxess, Tradeweb’s agreement to acquire Trumid, and LSEG’s planned purchase of Octus involve different businesses and commercial objectives. Then there is RealQ, bringing together Liquidnet Fixed Income and Neptune Networks in what I like to call the A&E (Axes & Execution) Channel. Together, these developments raise a common question: is the ultimate prize simply control over more of the trading workflow, or the intelligence generated by the decisions taking place within it?

THE VALUE OF KNOWING YOUR CUSTOMER
Financial markets have always placed a premium on understanding customer behavior. Experienced salespeople and traders know which clients favor particular instruments, when they tend to transact, and how their appetite for risk changes with market conditions. Much of that knowledge has historically resided in relationships, experience, and institutional memory.

What has changed is the ability to collect and analyze these observations systematically. Social media demonstrated the commercial value of converting everyday interactions into predictive behavioral intelligence. Institutional trading operates under different commercial and regulatory conditions, but the principle is relevant: understanding how customers make decisions may be as valuable as facilitating their transactions.

Every electronic bond trade leaves a trail of information extending well beyond the execution price. RFQs, dealer responses, rejected quotations, protocol choices, and timestamps reveal how traders seek liquidity and make decisions. Order and execution management systems provide additional context, while transaction cost analysis (TCA) helps establish whether those decisions produce satisfactory outcomes.

Consider a hypothetical credit trader at a large asset manager. Historical data might reveal that she uses RFQs for 70% of her transactions but relies more heavily on portfolio trading during periods of volatility. When her desk operates with reduced staffing, she might execute more quickly, narrow her choice of counterparties, or become less willing to negotiate. Now combine those observations with TCA, dealer responsiveness, market liquidity, and comparable transactions executed by colleagues. The resulting analysis could identify patterns that neither the trader nor her head of trading had previously recognized. More importantly, it could suggest alternative approaches tailored to the characteristics of the next transaction.

Predictive execution intelligence is already being commercialized. Tradeweb and MarketAxess, among others, use historical trading information and machine learning to inform counterparty selection and execution strategies. These tools primarily seek to predict execution outcomes rather than individual trader behavior. Yet they demonstrate how information generated by past transactions can inform future decisions. Extending that analytical capability to understanding how traders themselves behave represents another potentially valuable opportunity.

FROM DATA TO EXECUTION INTELLIGENCE
For a head of trading, the value goes beyond conventional TCA and peer comparisons. Understanding why execution strategies succeed or fail under different market conditions offers something more useful than another league table: evidence that can inform the next trading decision.

However, interpreting trading behavior requires more than identifying statistical correlations. A passive manager implementing an index rebalance faces different constraints from an active manager seeking exposure to an illiquid bond. A ten-minute execution is not necessarily superior to one taking twenty-five minutes. Transaction urgency, liquidity, information leakage, market impact, and the investment mandate all influence what constitutes a satisfactory result.

Dealer behavior adds another layer of complexity. Inventory, balance-sheet constraints, risk appetite, and quarter-end reporting pressures influence pricing and the willingness to commit capital. An algorithm that overlooks these factors may identify patterns without understanding their underlying causes.

That is where more sophisticated analytical models could add value. Instead of treating transactions as interchangeable, they can incorporate the circumstances surrounding an order and assess which execution approaches have historically produced favorable outcomes under comparable conditions.

A predictive system might recommend which dealers to approach, whether to use an RFQ or portfolio trading, or whether prevailing liquidity conditions justify delaying execution. Comparing subsequent outcomes against those recommendations creates a feedback loop through which the model can be tested, refined, and adapted to changing market conditions.

For the technology provider, this creates an opportunity extending beyond transaction processing. A platform that consistently delivers useful execution intelligence becomes more deeply embedded in the customer’s decision-making process. Its value increasingly depends not only on access to liquidity, but on the quality of the recommendations it provides.

The financial results offer another perspective on the changing economics of trading infrastructure. In the third quarter of 2026, MarketAxess reported 10% growth in information services revenue while total credit commission revenue declined 2%. Its average variable transaction fee per million dollars of credit traded fell 9%, partly reflecting changes in product and protocol mix. Although these figures do not measure behavioral analytics, they illustrate how information services can develop commercially alongside a transaction business facing pressure on fee yields.

FOUR PARTICIPANTS, FOUR INFORMATION ADVANTAGES
Owning more trading infrastructure does not automatically confer the greatest informational advantage. Trading venues observe activity within their own environments, including inquiries, counterparty responses, protocol selection, and transactions that never materialize. They can identify patterns in liquidity provision and execution behavior that are not visible in public transaction records.

Lawrence Peirson of RocketFin highlights the potential value of information that never appears in completed transaction records: “Missed trades may be the most valuable part. A traded trade tells you where the market cleared. A missed trade tells you where demand went unmet.”

EMS providers offer a different perspective. Andrew Mahoney of FlexTrade describes their role:

“An independent EMS acts as a cockpit for execution, aggregating liquidity and data from multiple sources to provide a unified view across venues, protocols, and counterparties. By sitting outside individual trading channels, it enables like-for-like comparison, workflow automation, and greater flexibility, reducing deterministic tasks and allowing traders to focus on judgement calls.”

The distinction extends beyond connectivity and execution choice. An EMS need not claim ownership of the intelligence derived from client activity, although access and usage rights depend on contractual arrangements and permissions. Independence alone is no guarantee of better information. Different venues, execution protocols, and voice channels capture trading activity in different ways, making the ability to reconcile those observations into a coherent analytical picture potentially more valuable than aggregation alone.

Asset managers possess another layer of information. Their internal systems connect execution decisions to portfolio objectives, investment mandates, risk limits, and subsequent investment performance. The platform may know how an order was executed, but the investment firm generally understands why the order existed.

Dealers bring a fourth perspective. Banks may observe customer behavior across multiple venues, bilateral inquiries, voice conversations, and longstanding trading relationships. They also understand their own inventory, balance-sheet constraints, and willingness to warehouse risk. That combination provides context that neither a venue nor an EMS necessarily possesses.

Consolidation creates opportunities to connect previously separate datasets, but access does not automatically confer the right to commercialize the resulting intelligence. Client confidentiality and regulatory obligations place limits on how trading information can be used. Competitive advantage therefore depends not simply on the quantity of data a provider possesses, but on its ability to turn legitimately accessible information into intelligence that customers value.

WHEN INTELLIGENCE BECOMES INFLUENCE
Fixed-income market participants operate under competing incentives. Asset managers seek efficient execution, dealers manage inventory and capital consumption, and venues compete for transaction activity. These interests often align, but artificial intelligence introduces another dimension to their interaction.

Consider a platform recommending a particular execution protocol. The recommendation may offer the best expected outcome for the customer while generating more revenue for the provider than an alternative route. Where those interests diverge, understanding which objectives the platform’s algorithms are designed to optimize becomes particularly important.

The distinction between predicting behavior and influencing it is critical. A system that recognizes how a trader responds to particular market conditions can adapt its recommendations accordingly, potentially improving execution quality, reducing information leakage, and identifying opportunities that might otherwise be overlooked. The question is whether those recommendations consistently serve the customer’s interests or are influenced by the provider’s commercial objectives.

This places greater importance on transparency and execution oversight. Investment firms retain responsibility for applicable best-execution obligations, whether decisions are made by traders or informed by algorithms. Understanding why a particular execution route was recommended, evaluating alternatives, and demonstrating execution quality remain essential.

THE NEXT COMPETITIVE FRONTIER
The consolidation of fixed-income infrastructure is often portrayed as a battle for control of the trading desktop. Yet the greater commercial opportunity may lie in what providers learn from the activity passing through their systems. As artificial intelligence becomes more sophisticated, the ability to translate trading behavior into predictive intelligence could prove as valuable as the execution services themselves.

Size alone will not determine who succeeds. Data quality, market expertise, contractual permissions, and the ability to interpret information in context will matter. Providers capable of delivering demonstrably better execution outcomes will have an advantage, particularly if their recommendations remain transparent and customers retain meaningful choice.

In my view, these developments invite us to reconsider where value is created in the bond trading workflow. The next competitive advantage may belong not to those who process the most transactions, but to those who learn the most from them. Owning the workflow provides access to the customer. Understanding the customer may ultimately be the more valuable asset.

©Markets Media Europe 2026

TOP OF PAGE