MarketAxess Eyes Predictive Capabilities for Fixed Income Liquidity

As part of its drive to provide clients with more predictive insights into fixed income trading, MarketAxess is working on expanding its pre-trade automation capabilities, looking at aspects of its trading processes where it could better forecast the outcome of clients’ trading activities.

The move is in part a continuation of its experience with CP+, an AI-powered pricing engine that MarketAxess launched in 2017. For any bond that a client might be interested in trading, CP+ looks at a broad set of data points in real time and provides a prediction of what the next traded price will be. This predicted price then helps clients determine how the prices they are getting back from liquidity providers fit within the market context—a kind of a reference price.  

Gareth Coltman, global head of automation at MarketAxess, says the firm is now expanding its pre-trade prediction capabilities to look at other types of information that could help clients make trading decisions.

“What we are focused on now is broadening from just predicting the price, and starting to think about what other elements of that trading process could we help clients predict,” Coltman says.

He says MarketAxess can now predict how easy or likely it might be for a client to trade a particular bond. For example, the broker has developed the capability to predict for a bond how many responses or prices are likely to come back to a trader client when they send out a request for quote (RFQ).

If the bond turns out to have a very low likelihood of trading, the trader may have to consider a different bond, or go back and have a conversation with the portfolio manager about whether or not it is the right bond for them to be buying. In the case of a large order, a different approach could be taken by breaking it up into smaller pieces.

“This predictive analytic around tradability, and success in automation, is still a work in progress. It is something that we are exploring with clients, but it is not broadly live in the product,” Coltman says.

Automation Across Protocols  

MarketAxess’ pre-trade automation efforts are also designed to help clients navigate a growing number of trading protocols.  

Traditionally, electronic trading in the fixed income markets has been focused on relatively liquid credit and smaller sizes. As market participants have become more engaged in electronic trading, MarketAxess has to think about how to support trading of larger size trades and the ability to operate seamlessly across different markets. Coltman says the broker’s response has been to “think about how can we develop new protocols to make sure that we can support all of that demand, and make sure that whatever type of liquidity demand a client has, they can find a solution to that on MarketAxess.” 

The firm has been developing different types of trading protocols in recent years. One of them is Live Markets, an order book protocol it rolled out in the US this year. The protocol is geared towards all-to-all trading, connecting market participants directly to trade bonds without necessarily having a bank involved. Another new protocol, Mid-X, is built on top of its CP+ engine, and went live in Europe at the end of October.

“We have an ambition to roll this [Mid-X] out more broadly, and that will be a focus for us in 2021,” Coltman says, adding that both the Live Markets and Mid-X protocols ultimately will become global in terms of their product coverage. These sit alongside other protocols such as RFQ and axes. The latter protocol allows clients to post interest in a large block trade anonymously and negotiate.   

In the longer term, Coltman says MarketAxess will provide broader automation across all the protocols it supports, to help clients capture the trade workflow in an automated fashion. This could mean that when an order arrives on the trading platform, the platform would determine which type of protocol would work best for each client according to their needs. Protocols could be selected on the basis of whether the priority is cost-saving or tracking a specific benchmark.   

In order to wrap all these protocols in an “automation solution”, Coltman says, a prerequisite is having sophisticated data analytics. “Right now, that is where we are focused, on those kinds of pre-trade analytics,” he says.