ION FX’s Eugene Markman discusses the challenges and opportunities of data and AI

Key Takeaways

• Data fragmentation might require industry consolidation
• AI will improve decision-making but will not call all the shots
• The need to be efficient with capital will drive FX automation

ION executive Eugene Markman and Colin Lambert from The Full FX recently appeared on Markets ConversatION podcast to share the key insights discussed at the Full FX Scandinavia 2024 conference in April.

Artificial intelligence and data were high on the agenda.

Markman, COO of ION Markets (FX), said that while you could program AI to analyze the data and answer your question now, in the future, AI will analyze the data and then give you the question. ‘To me, that’s going to be the big opportunity, and that’s when AI will change.’

The data challenge question

In terms of data, everybody has similar challenges that industry participants are solving at a micro and macro level. Everybody is doing the micro level, but it seems that many are doing the same work, which is inefficient and redundant, Markman said. ‘I think there’s an opportunity there for some consolidation and to have a few vendors who really become the industry norms.’

For Lambert, the bigger risk now is that data is becoming too fragmented and that consolidation will inevitably occur. How, though, is the overriding question. Is it going to be the law of the jungle where the weakest are just thrown off the side of the boat, or will it actually be through M&A? Will it be through more partnerships?

An important message from the conference is that data will be as much of a challenge to the industry as it is a godsend for people trying to build their businesses.

AI in workflows

Moving beyond data, AI was a recurring topic at the conference.

A key takeaway for Lambert was how AI changes workflow in financial markets generally, and specifically FX and rates, but he’s not convinced it changes how trading is done. While it’s definitely part of the workflow solution, there are a lot more risks ‘that I don’t think the industry is willing to take at this time’.

In Markman’s opinion, although AI has become a powerful tool, it will not be the machine that’s calling all the shots. ‘It’s not going to make all your decisions for you, but it could do a lot of work for you to free you up. So, you can have better information to make your own decisions and make better decisions because of that’.

Lambert said that the really good exponents of post-trade processing use technology for almost 100% of what they do. AI will be the driver in exception-based processing, especially with data. But he also underlined the risk that if there is an anomaly in the data, AI might ‘exacerbate the problem.’

Markman said that the great opportunity and change in the future will be AI analyzing data and coming up with the questions.

Liquidity beyond spot

For Lambert, there’s a real disconnect between the buy-side and the sell-side. Industry surveys indicate that liquidity is one of the biggest concerns for the buy-side.

‘If you look at the spot market, LPs are more aware than ever of their data and how their clients are behaving. And I think they’re using it in a much smarter fashion. Where we’re not really dealing with the challenges is beyond spot.’

Liquidity is a bigger issue in non-deliverable forwards (NDFs, currency derivatives contracts) and swaps.

Markman said that swaps make up more than 50 per cent of the market on a notional basis and are probably overlooked because they are not electronified.

‘As soon as you hit that critical mass, where you have a demand to automate, then everybody wants to stream it. Then everybody wants it to be a true electronic process.’

The main issues to solve in the swaps market are access to credit and capital usage. The necessity to be more efficient with capital will inevitably drive innovation and automation of FX swaps, Lambert said.

The FX Global Code as a living document

The Code, a set of global principles of good practice in the FX market, is widely considered a success and doesn’t need much changing.

However, Lambert said that there are a couple of areas where more clarity is needed and which should be a bit harsher, including enforcing “symmetric last look.” Confirming, firming and completing trades (or rejecting them) should be done as quickly as possible. ‘If you can accept a trade in one millisecond, then you should be rejecting your trade in one millisecond, perhaps two, not 20,’ he said

For Markman, the Code needs to be a living document, and change must occur at a rate that is acceptable to the market. We need to start thinking about how new developments, such as AI, should be included, and provide clearer for pre-hedging.

Best practices for tech providers

When asked if the Code should consider best practices for firms that provide critical infrastructure, the ION executive said most companies already follow standards like ISO and SOC 2. If they do not, it’s up to each to decide whether to take that risk.

The Code doesn’t necessarily need to create its own guidance, but perhaps adding suggestions like highly recommending SOC 2 as a guideline and using what already exists is the better approach.

If everybody thinks about technology in the same way, you don’t get to explore all the options or evaluate all risks. ‘There are things that kind of sit in the blind spot,’ Markman said.