TradeTech: AI-enabled algo developments will drive price discovery

Decisions must be based on data, but you cannot ignore the human element, finds TradeTech panel.

A group of expert practitioners gathered together at TradeTech this week to discuss the complex developments in the field of AI-enabled algo developments, and how they can improve access to liquidity, enable better price discovery, and facilitate real-time market insights.

We all know the importance of data, and the crucial role it is now playing. But what role are these developments playing in the material improvement of the industry and its operations?

Joe Wald, managing director and co-head of electronic trading at BMO Capital Markets, emphasised that data has to achieve three elements: it must empirical, iterative, and collaborative.

“We need a philosophy that you move forward with across the entire platform,” he explained. “It has to be data-driven, it can’t be anecdotal anymore. All your decisions must be based on solid data. It has to have a framework where you can experiment, conduct B-tests, see what you’ve learned from the data over time. And it has to be a platform that allows for rapid customisation, so you can collaborate with your client to find what is successful. These three elements are key to achieving the right outcomes.

“There is a quote from The Kingsman that says ‘Manners maketh man’. For us, its ‘market structure maketh money’. There are places for our clients where we can trade, where we have an edge. There are ways to leverage the market structure in Europe that are evolving very rapidly. The ultimate and most tangible results in driving performance and execution, we’ve found, is in market structure research – how to find new tools and new ways of sourcing that liquidity.”

Dr Peter Ho-Spoida, vice president of data strategy at Deutsche Börse Market Data and Services, emphasised the need for increasingly sophisticated methods analysis in order to gain competitive advantage.

“We see increasing demand for data, but also demand for the right tools to analyse that data. An important issue is enablement – allowing clients to focus on their core expertise. Traditionally we have had many Tier 1 clients buying highly complex, large amounts of data. That gives lots of insights, but it needs specialised knowledge and lots of maintenance. So we have started to build infrastructure around it to allow clients to use this data – including a back-testing environment, or example.

“The level of noise here, compared to other areas in which AI is being applied, is very high. So you need a very disciplined approach.”

Daniel Mayston, head of electronic trading and market structure EMEA at BlackRock, warned that although data is important, it’s not the only consideration – and that the personal aspect should not be lost.

“From the buy-side perspective, we need to look at the bigger picture,” he stressed. “The electronic aspect is important, the data is important, but it’s not everything. In addition, most people think that automation and electronic trading means fewer people, but actually it can be the opposite – you have to prioritise resources to get things done, and the human aspect is as important as it ever was.”

Collaboration is also important. “You have to be on the same page as your sell-side counterparts, you still have an end goal,” said Mayston. “Our real proposition is that we have to solve something for our clients, our end investors. What really matters, ultimately, is execution quality.”

Jas Sandhu, global head of agency electronic solutions at RBC Capital Markets, agreed. “It might take different names, but the motivation of a trade is fundamental both internally and externally, and whoever is trying to help you achieve your goals has to understand that. Then you put numbers around it, add TCA, supplement with data.”

Another trend is innovation entering from alternate asset classes. “I’m really looking forward to the innovation coming to the exchange landscape, driven by the crypto asset class,” said Ho-Spoida. “Most of these cryptos are already using cloud-based technology, and this is now trickling into the conventional markets. Settlement for example – we have a big problem in the equities space, and this problem doesn’t exist in the crypto space.”

But the real battleground may not even be at the coalface, but behind the lines. “The war for talent in this space is real,” warned Sandhu. “Previously it was quants – now it’s coding, but its fungible, they can move between industries – the same coders can work for fintechs as can work for banks, and that makes the competition much more fierce.”