The limits of the electronicization of the bond market

Stay up to date with free updates

Simply sign up to Sovereign bonds myFT Digest — delivered straight to your inbox.

You may have noticed that we’re taking an above-average interest in bond market developments here, as trading is gradually migrating from phones to screens and subtly changing its nature in the process.

Analysts at Barclays have also been very good on the subject, particularly Zornitsa Todorova, the bank’s head of “FICC thematic research” – a title that practically seems like an excuse to do nonsense like this:

The digitization of the bond market over the past two decades has been an extremely important development, and yet data availability issues make tracking this development a considerable challenge, particularly for the European market. The way we try to address this issue in this report is by creating our own proxy for the future consolidated European bond band.

This tape is a transaction-level data set assembled using MiFiDII post-trade reporting records from February 2023 to March 2024, containing over 9 million transactions and over €23 billion in total volumes for both credit, as well as for installments. We then combined this European repository with data for the US market to paint a holistic picture of the global e-bond market.

This is from a new Barclays report on electronic bond trading by Todorova and Andrea Diaz Lafuente. The hard work of the data is a delight, but the conclusions in the title will not surprise anyone who has followed the subject.

Government bond trading is already highly digital, with about half of investment-grade corporate bonds traded electronically, and high-yield and convertibles are slowly catching on.

There are a lot of interesting statistics in the report that make the point, such as how it’s not just smaller bond trades that are being done electronically anymore:

. . . Computer algorithms now regularly block trade price requests (over €5 million in IG and over €1 million in HY). In fact, we estimate that in 2017 only 4% of credit block trading requests had at least one algo response, compared to 62% in 2023.

And that algo-to-algo trading becomes more of a thing:

The use of so-called automatic execution (auto-x), where investors execute trades based on a predefined set of criteria (eg price limits or number of responses), has taken the market by storm. In 2019, less than 23% of US Treasury transactions used auto-x functionality. By 2024, this figure has nearly tripled, with almost 62% of requests using auto-x.

Todorova and Lafuente see traditional telephone trading and e-trading as “strategic complements” – in other words, the former doesn’t just cannibalize the latter. When markets are a little crazy, having both is especially important, they argue.

But the main point of the report — entitled The human touch: why the bond market will continue to have a voice — is that there is probably a limit to how far electronization can go.

For example, while government bond trading is years ahead of corporate bonds in what Barclays poetically calls its “electronic journey,” a large minority of smaller, even algorithm-friendly, trades are still conducted over the phone.

In fact, it has generally hovered around 70 percent for several years because of what Todorova and Lafuente call the “three Cs” — many transactions still require banks to commit ccapital; the heterogeneity of bond markets means that many trades must be cnormal; and there is a rooted ca bond trading relationship community that is especially important when markets are chaotic.

This means that even with more adoption and better technology, there’s probably a limit to how far things can go:

. . . There is a limit to how much the market can ultimately rise, and based on our analysis, we see 80% as maximal threshold. This capability will vary across asset classes and transaction types, depending on how important the “three Cs” of voice trading are. Each approach has unique strengths and roles to play in financial markets – electronic trading offers scale, speed and efficiency, but voice brings the human touch and enables the management of complex transactions.

The coexistence of voice and electronic trading will play a key role in forming a flexible and resilient bond market ecosystem and enable investors to navigate different market conditions. With all the electronic evolution, we believe that bond market trading will never look like stocks.

This is a bit of a truism. Of course, bond markets will never look exactly like stocks for a multitude of reasons. And even in stocks, phones still play an important role in larger, more complex transactions.

But both the speed and scale of the shift to electronic trading have consistently surprised people for decades. It’s hard to say it won’t surprise people anymore.