Change is hard and, for some players, that resistance is slowing adoption of automation and AI.
If you’re spending time out the office at a conference, you want to be sure that you’re going to glean some wisdom from the thought leaders who populate the keynotes and panels. One of the reasons why the Fixed Income Leaders Summits are so successful and continue to grow is that they stay true to the name – they feature leaders worth hearing.
The line-up at last week’s Philadelphia event included former director of the CIA, US Army General David Petraeus. Also there were high profile C-levels from some of the biggest players in the space, who had plenty of knowledge to share and interesting things to say about the evolution of the market.
Dan Veiner, BlackRock’s Global Head of Fixed Income Trading, took on topics such as the inter-dealer market opening up and traders connecting straight to exchanges. He chose to sing the praises of sell-side banks who were still important to his firm and remained crucial as liquidity providers.
Buy-side participants were in high attendance, which makes sense, reflecting the way that the electronification of the market is heading. It’s the buy-side that wants to get more educated on electronic trading and a new wave of technologies.
No surprise that panel sessions explored automation, machine learning and AI for trading, topics that have been topping the agenda at trading events for the past year. The difference here, where the sector needs more innovation to electronify institutional-sized block trading, is any evidence of it being implemented at scale. I got the impression that there was a lot of talk but not much action.
One school of thought is that quant trading, which relies on mathematical computation and algorithms, make sense in asset classes like treasury bonds but less so in less liquid markets. Fixed income investments are often carried out in chat windows, dominated by inter-dealer brokers that are so large that it’s hard for them to change. There’s a feeling that a lot of legacy, and red tape, must be overcome before machine learning and AI can be considered game changers for this asset class.
Equity comparisons were a recurring theme. The absence of a fixed income trading equivalent to the Central Limit Order Book (CLOB) was also discussed. When an equities exchange publishes market data they do it in a CLOB, which provides members with updates on every bid and ask price for listed securities -- i.e., better transparency. In contrast, market data remains confidential between the parties involved in each transaction in traditional Request For Quote (RFQ) trading. Traders ask (via email, chats or phone) a known set of dealers to quote a bid/offer for specific securities -- i.e., less risk of a huge transaction going sideways for the dealer. The two seem to be worlds apart, so it’s hard for some to see how electronic trading can bring them together.
Even the relevance of TCA (Transaction Cost Analysis) is being questioned by a lot of buy-side trading companies when it comes to fixed income investments. While TCA is a mainstay of equities it is not necessarily a good fit for fixed income. Some speakers argued that measurements should be based more on performance than cost analysis.
This idea of expanding on existing analysis was mirrored by Craig Hurl, Director at Ontario Teachers’ Pension Plan, during a Greenwich Associates webinar Execution Quality: How the Buy Side Measures it; How the Sell Side Produces it. He described how the company is analyzing their broker’s performance metrics against their portfolio managers’ intentions, which informs how aggressive or passive its trading should be.
It will be fascinating to see how the market, and how it analyzes execution, continues to evolve --and if folks are singing from the same song sheet next year. If they are, it should at least gain a country twang because in 2020 the Leaders Summit moves to Nashville.
Learn more about Corvil Analytics for Fixed Income