Automation was central to a panel discussion on pre-trade transparency and front-to-back trading workflows last week at the Fixed Income Leaders Summit, with panellists noting that automation has different use cases and levels of success within the context of fixed income.
Panellists argued that automation can be applied in some instances such as rates, whereas in others ultimately clicks are just reduced with only certain functions of the trade automated. Automation was therefore described by the experts as being a spectrum.
“You have to understand that automation is not the same across various areas of fixed income,” said Ricky Goddard, head of credit trading at Schroders.
“In credit, automation effectively is just reducing the number of clicks on the trade. We’re not at the point where we’re going to have our OMS or EMS read for certain characteristics and then have them push orders forward. For rates that can happen, but for credit markets it will be an entirely different process.”
Some panellists did acknowledge that even on a high touch desk, high levels of automation can be achieved. This was noted as being useful particularly in times where smaller trades do not take precedence and pressure is on fulfilling larger trades.
“Even on a high touch desk, we can have a hybrid setup where we initially set a list of trades to be automated and the algo will look at whether it fills these conditions. If met, it will execute, otherwise it needs a human trader to interact,” said Anuj Thakur, lead, high touch fixed income trading at Nordea Asset Management.
Automation was described by some panellists as a “vicious cycle”, given that a lot of pre-trade data is a prerequisite to be able to automate.
Thakur added: “As automation has picked up on the sell-side, whether it’s algos, pricing, RFQs quite quickly or various other protocols, that generates more data, which gives confidence both to the buy- and sell-side on approximately where the mid is.”
With some benefits clear, panellists were hesitant to claim that full automation would be applicable in the fixed income context. However, some panellists noted that it’s best to never say never.
This was in part an acknowledgement to the already huge advancements of procedures through electronification, which previously may have not been believed to be possible.
“We do, however, need to get to the point where we’re able to really trust how we put that process in place and why we’re selecting the banks that we’re selecting. Also, are we putting certain trades that are automated all-to-all or some to a bespoke group of three to five brokers because the trade is potentially illiquid?” added Goddard.
“When I’ve looked at some of the automation protocols in place for products in the past, you can fully automate who you want to send certain products to because you have a fairly good idea of market share in those asset classes. You need to be able to tailor it and that all leads back to having good quality data.”
Goddard went on to add that he does not believe that we currently have the quality of data required to fully automate broker selection.
Still seeing the positives associated with automation in fixed income, Thakur added: “We want our traders to focus on those block trades. A lot of time can be saved if a lot of that flow gets automated where you’re automating 60-70% of the number of tickets which you’ve done manually. This will ultimately free up time to focus on block liquidity.”
Automation was a key discussion point in various panels at the conference, with many panellists highlighting the gains that new technologies are providing. As with any advancement though, it was agreed that adoption must rely of good data for gains to be actualised.