About a-team Marketing Services
The knowledge platform for the financial technology industry
The knowledge platform for the financial technology industry

A-Team Insight Blogs

A New Era of Bond Trading: Leveraging Automation and Liquidity Score Splitting

Subscribe to our newsletter

By Vuk Magdelinic, CEO, Overbond.

Bond traders face a set of challenges not seen in most other markets for tradeable securities. There’s no comprehensive centralized record of fixed income trade price or size data because most bonds are traded over-the-counter. Additionally, many bond issues are small and trade infrequently or not at all. As a result, it’s often difficult for bond traders to assess the liquidity and volatility of the issues they’re trading.

In the U.S., TRACE provides a centralized record of historical trades in U.S. fixed income securities, but the data is available with a lag, and trades beyond a certain size are masked. In Europe, regulators are still working toward developing a consolidated tape. To close the data gap, many desks are turning to tools such as Overbond, which uses AI to aggregate data from multiple sources and generates a desk-level consolidated tape of near real-time prices for bonds.

This newly developed ability to generate more complete data is driving ever-greater automated bond trading capabilities. Traders are increasingly automating a portion of their trades, which is allowing them to respond to more queries with greater speed.  Currently, much of this automation is focused on the large volume of smaller-sized RFQs that desks process every day. This leaves traders with more time to focus on complex and larger trades, which increases their efficiency and profitability.

Traders who are allocating a portion of their trades to automation need a system to determine which trades can be fully automated and which trades require trader input. To aid with this decision, Overbond has developed AI-driven liquidity and price confidence analytics that auto-adapt to trade size and direction to assign liquidity and confidence scores to over 300,000 fixed income securities and 30,000 issuers.

The scores are derived from volatility and bid-ask spread components and are divided into three tiers according to their liquidity and pricing confidence relative to the universe of bonds with the same currency.

Tier 1 is a cluster of the most liquid bonds, which are  prime candidates for auto-execution. The bid/ask liquidity scores in this pool are made more accurate by excluding liquidity-based anomalies and applying a skew in certain market conditions.

Bonds in tier 2 are still in the best executable cluster but they have a slightly wider distance to mid than those in tier 1. Tier 2 is subdivided into three zones:  highly recommended, slightly recommended and not recommended. In tier 2, Overbond’s AI-driven pricing engine attempts to provide a best-executable price.

Bonds in tier 3 do not have the liquidity and pricing characteristics required for precise automated trading and are subdivided into slightly recommended and not recommended zones.

To generate liquidity and price confidence scores, Overbond AI uses innovative mathematical and statistical techniques such as case-adjusted cluster analysis.

Overbond uses six months of historical data to train its AI model. Incorporating settlement-layer data is key to enhancing the precision and auto-adjustment capabilities of the liquidity and confidence scores. For example, settlement-layer data is used to derive total market capacity, which is a measure of the portion of outstanding bonds of an issue that are available for trading on any given day. Total market capacity is then used to help derive the implied probability of execution for a desired trade size.

Settlement layer data is integrated from sources such as TRACE and through partnerships with Deutsche Börse/Clearstream and Euroclear, among others. Through these data sourcing partnerships, Overbond has access to settlement-level fixed income transaction data derived from close to 500 million transactions that Clearstream and Euroclear process annually across 100 currencies. This dataset represent the majority of EUR traded volume, giving traders the information they need to assess liquidity and volatility, allocate trades by level of automation, and increase efficiency and profitability.

Subscribe to our newsletter

Related content


Upcoming Webinar: Re-architecting the trading platform for interoperability, resilience and profitability

Date: 15 November 2023 Time: 10:00am ET / 3:00pm London / 4:00pm CET Duration: 50 minutes Trading platforms have come a long way since the days of exchanging paper certificates and shouting across trading floors, pits and desks in the early 2000s, but there is progress still to be made as firms strive to reduce...


Trading in the Cloud – Has Its Time Finally Come?

By Mike Powell, CEO, Rapid Addition. The structure of capital markets and how they are traded have, to a large extent, been shaped through the technology adopted by market participants and service providers over the past two decades. And while some firms remain cutting edge, many are left operating legacy infrastructure characterised by high cost...


ESG Data & Tech Briefing APAC

Join us in one of the greenest cities in the world as we bring together thought leading ESG specialists to explore how financial institutions are adapting to the evolving ESG regulatory and market infrastructure.


Institutional Digital Assets Handbook 2023

After initial hesitancy, interest in digital assets from institutional market participants has grown over the past three to four years. Early focus inevitably centred on the market opportunities presented by bitcoin and other cryptocurrencies. But this has evolved into a broad acceptance of a potentially meaningful role for digital assets in institutional markets. It’s now...