SOLVE, the provider of pre-trade data and predictive pricing for fixed income markets, has introduced a new Relative Value Analysis capability to its AI-driven municipal bond pricing platform, SOLVE Px. The enhancement is aimed at improving transparency and decision-making for municipal bond traders, portfolio managers, and risk teams.
The newly released feature allows users to visualise and assess changes in a bond’s relative value over time, enabling a more dynamic analysis of trading opportunities. It supports functions such as identifying price anomalies, spotting yield trends, and comparing bonds with similar characteristics, all using SOLVE’s AI-predicted pricing rather than traditional evaluated prices.
“Price discovery in fixed income is challenging, especially outside the most liquid or ‘on-the-run’ issues,” explains Eugene Grinberg, CEO of SOLVE, to TradingTech Insight. “Our initial focus was on aggregating foundational quotes data – bids and offers across hundreds of thousands of securities – capturing millions of quotes daily across all major asset classes: corporates, securitised products, municipals, syndicated bank loans, and more. But while quotes data is critical, many bonds don’t trade or even get quoted for months at a time. That gap led us to develop our predictive pricing models – initially for municipal bonds, and soon for corporates – using observable data to predict, in real time, where bonds should be priced, with a high degree of accuracy.”
The platform ingests data from over 300 buy- and sell-side clients, extracting quotes from unstructured messages using AI and natural language processing, to turn fragmented, conversational data into usable pricing insights, says Grinberg. “Given that many bonds lack regular quotes and trades, our AI models incorporate features from terms & conditions, comparable securities, and surrounding market data to price entire asset classes with precision, all in real time.”
Key tools include the ability to determine whether a bond is trading rich or cheap against its historical norms or peer group, filter securities for comparative analysis, and assess pricing signals across over 900,000 active municipal bonds. Processing over 20 million daily quotes, the platform aims to help users navigate this highly fragmented and fast-moving market.
“AI can feel like a ‘black box’, so the key to building trust is demonstrating performance,” notes Grinberg. “We help clients gain confidence by showing how closely our predicted prices align with actual trades, backed by months of back-testing across various interest rate and volatility environments. In addition, we publish a confidence score with each prediction – essentially a by-product of the model – which indicates how confident the AI is in that price. While the score is a more mathematical artifact, it should ultimately reflect the volume and quality of direct observations the model sees, whether from the bond itself or from comparable securities.”
The Relative Value Analysis feature was developed in response to growing demand from market participants for greater clarity in bond pricing, particularly during periods of volatility. SOLVE collaborated with clients to tailor the tool to real-world trading workflows, aiming to enhance pre-trade analytics and support faster, more informed decisions.
“With this foundational data, pricing entire asset classes with high accuracy, we’re well positioned to support various forms of relative value analysis,” states Grinberg. “Users can assess how a bond is performing relative to its own history, to other bonds from the same issuer, or to comparable securities. They can identify whether historical trends have shifted, potentially revealing opportunities to capture alpha. This analysis can also be scaled up beyond individual securities to evaluate entire sectors or market cohorts.”
Subscribe to our newsletter