The adage ‘build it and they will come’ wouldn’t seem to apply to the credit derivatives market.

Automated trading and execution of credit default swaps, both single-name and more widely traded indexes, is a rarity among buy-side firms. Yet experts point out that the ingredients for greater automation exist: the contracts are standardized, liquid, and can be traded on electronic platforms.

It’s a jarring contrast with the corporate bond market, where electronic trading and algorithmic execution are more commonplace.

“I don’t think any buy-side firm is at the stage where they are completely no-touch execution on index CDS,” says Laurent Samama, head of European flow credit trading at BNP Paribas. “It’s clear that the market structure allows it. The streaming is there, so it is technically possible to think of fully automatic execution on index CDS.”

There are doubts over the level of interest in changing the market, where request-for-quote protocols dominate investor activity. Regulation has increased transparency and pushed trading on to platforms, but the tried and trusted methods of voice trading are proving hard to dislodge.

A shift to execution on central limit order books (Clobs), which could unlock the spread of algo strategies, has not materialized.

Some see potential for automation in streaming protocols—in which dealers present constant executable prices on platforms like Bloomberg and Tradeweb—but others doubt the quality of execution they can receive on screens.

Representatives of platforms report a mixed outlook for algo execution. Charlie Campbell-Johnston, head of integration and workflow solutions at Tradeweb, says the firm has seen recent interest in CDS automation from real money firms.

“It makes sense to automate CDS. It has been surprising to us that we haven’t seen as much automation that we thought we would have in the past year, because it’s been available to do for quite some time,” he says.

Ravi Sawhney, global head of trade automation and analytics at Bloomberg, says clients have shown limited enthusiasm for extending the firm’s automatic trading tool, Rule Builder, to CDS. Those that have expressed interest in automation are able to use Bloomberg’s API offering, a plug-in that sits between the client’s own trading systems and the vendor’s.

“I expect demand for automation will increase over time, especially for more systematic types of CDS trading, for example during the roll periods. While we are likely to extend Rule Builder to CDS in the future, there’s just more appetite for automation in some of the other asset classes right now,” Sawhney says.

Voicing your intentions

Fifteen years ago, CDS was traded exclusively over the phone or electronic chat, known collectively as voice trading. But in the US, regulation enacted after the 2008 financial crisis attempted to push trading of liquid contracts on to more transparent venues, dubbed swap execution facilities (Sefs).

In February, 81% of the US dollar-denominated CDS volume and 82% of tickets reported to the Depository Trust & Clearing Corporation’s swap data repository passed through a Sef. These figures include interdealer activity as well as trades involving buy-side firms. The data doesn’t capture the whole market, as it includes large trades capped above certain notional values, but it offers one of the best publicly available insights into the CDS market.

The post-crisis Sef rules also included the requirement that venues offer Clobs in addition to request-for-quote (RFQ) trading protocols. In RFQ trading, a buy-side firm is required to ask at least three dealers for a tradeable price if the trade size falls under the block amount. Buy-side firms can also request streamed prices from their dealers.

In asset classes like FX or equities, Clobs are the gateway for automated trading. Standing orders allow executing algorithms to place orders and buy and sell securities automatically. The programs use the running Clob prices to derive buy/sell signals or break up large orders and hunt for liquidity across venues in an attempt to limit market impact.

The assumption is that Clobs would play a similar role in the credit derivatives market. But that has proved a step too far for most CDS traders.

“There had been attempts to launch Clobs, however moving from voice trading to purely electronic platforms was already a substantial shift for the market, and this protocol never took off in the CDS space,” says a CDS market structure expert.

Bloomberg, which accounted for 70% of CDS traded on Sefs in February according to FIA data, sees roughly a tenth of its CDS trading volume go through the order book.

Tradeweb users largely do not use the platform’s order book to trade CDS. Less than a fifth of Sef-traded CDS went through the platform in February.

Ice’s Clob is effectively dormant and does not record trades on a daily basis, Risk.net, a sibling publication of WatersTechnology, understands.

I look at fully automating RFQs as being on the very low end of electronic algorithmic trading. It’s almost like building a robotic arm to press the keyboard button for you

Alexis Blair, Aspect Capital

The lack of Clob trading has left RFQ as the trading protocol of choice for CDS markets. Alexis Blair, director of trading at systematic hedge fund Aspect Capital, says automating RFQ execution has its limits even if programmatic strategies can signal when to transact. The firm is looking at how it can auto-populate fields on platform interfaces, but execution would remain manual.

“To go beyond that and to automatically RFQ would be a lot of extra work for not much benefit for us,” Blair says. “I look at fully automating RFQs as being on the very low end of electronic algorithmic trading. It’s almost like building a robotic arm to press the keyboard button for you. It’s not very elegant.”

Blair estimates that CDS comprise less than 1% of the firm’s traded volumes.

Others doubt whether automated trading can bring best execution, especially for trades not required to be executed through RFQ. Michael Hattab, senior portfolio manager at asset manager LFIS, says that in the relationship-driven CDS markets it’s easier to achieve best execution through voice trading than by trading electronically.

“You have wider bid/offers on screens than what you can get by doing the execution by voice. When you ask dealers directly for the price, you get a better quote than what is shown on screen,” he says.

US regulation permits voice negotiation for swaps above block size before the transactions are executed and reported to Sefs. In Europe, the rule requiring RFQ users to request quotes from three dealers does not apply on multilateral trading facilities, a type of venue.

The concern over best execution is especially relevant when transacting in large sizes where traders worry that revealing their intention to too many dealers could move the market against them. US reporting requirements create an additional concern that trades will become public soon after execution, meaning dealer pricing may include a premium reflecting the possible challenge they’ll have hedging or offloading the position.

In the US, large notional swaps receive a delay from real-time public reporting between 15 minutes and 24 business hours after execution. The exact timing depends on various factors, including where trade was executed and whether the contract is subject to mandatory clearing. The impact of post-trade transparency remains a subject of debate.

“The transparency is not something that we’re typically used to in the fixed income world—it’s usually quite opaque. That makes traders that trade the index a little bit apprehensive, and if I’ve got any sort of size to do, they would much prefer that I engage them directly than put them in competition,” says Stuart Campbell, head of trading at BlueBay Asset Management

The rules also apply to interest rate swaps, but the size and liquidity of the market makes transparency less of a concern than CDS, says Campbell.

Given the small amount of CDS trading that BlueBay does and the individualized requirements from the firm’s various mandates, Campbell says automating CDS trading remains a low priority, especially compared to building similar capabilities for credit or FX trades.

Automatic drive

Still, some experts argue that Clobs are not essential for automated execution and that the preference for RFQ trading does not necessarily inhibit a certain level of automation.

Kevin McPartland, head of market structure and technology research at consultancy Greenwich Associates, points to the cash bond market as an example of where automation in a relationship-driven market has not depended on a Clob. CDS trading can mirror that development, he says.

“Lots can be done to automate RFQ trading. RFQs can be auto-generated and pick the right dealers to include based on past performance, dealers can auto-respond to RFQs using auto-pricers,” he says. “Also, moving from RFQ to request for stream and request for market lets the buy-side requester build up a virtual order book based on the prices dealer stream at that moment in time.”

In cash bond markets, investors can use a trading tool through platform MarketAxess to generate RFQs automatically and select the responses that meet previously defined criteria. The system comes with a low-touch and no-touch version, so traders choose their level of involvement.

The tool saw $124 billion in volume in 2020, a 60% increase from the year before. In the third quarter, 86 firms used the automated execution functionality, up from 64 in the third quarter of 2019.

This type of system is also available for CDS via Tradeweb’s auto-execution tool AiEX, which allows investors to automatically determine the best counterparties to put in competition when requesting quotes. But few firms are understood to be using the tool for CDS.

Indeed, prices streamed from dealers to clients may be the best way for automated CDS trading to take off. Over half of Bloomberg’s CDS Sef trading volume occurs through its request-for-stream protocol, in which investors ask dealers at least daily to send prices.

LFIS’ Hattab points out that changes made to standardize contracts over a decade ago by clarifying default scenarios and setting fixed coupons now mean that firms have years of consistent data to back-test trading algorithms.

Whether these initiatives are enough to nudge more buy-side firms towards automating CDS trading and execution remains to be seen.