October 2004 saw Lionel Messi make his debut for Barcelona.
October 2023 saw ipushpull make their debut at FILS in Barcelona.
Only time will tell which had the most significant long term impact...Vamos!
As another event draws to a close we reflect below on the key themes and hot topics discussed at the event.
The main stage at FILS on the first morning was focused on which protocols to choose when making trading decisions. While RFQs are heavily supported as a major source of price discovery, buy-side traders still feel they are giving away too much information when requesting, without getting the levels they want, and therefore called for a greater move towards RFM (Request for Market). In addition, some panellists felt that RFQs were becoming too customised and that, depending on the market and instrument, streaming prices are becoming more prevalent as technology has developed sufficiently to allow these to be addressable to given users. What wasn’t discussed is the continuous workflow of traders receiving instructions from portfolio managers via internal systems, and manually typing that into other external systems or chat sessions to multiple counterparts, and the ongoing price negotiation, which can be streamlined with greater interoperability and interconnectivity.
Other protocols being supported are algorithms to support streaming axes and automated RFQ levels, while others suggested that All-to-All had found its place during the pandemic and that this continued to find supporters, especially with the sell-side that could use their electronic systems advantage. The buy-side are looking to build out their capability, although perhaps participants should be cautious of a so-called algorithmic sweep! Interestingly, there was no mention of new guidance on the regulatory perimeter for trading venues that have been issued by ESMA and the FCA and what impact that may have on vendor solutions in particular.
Meanwhile, the growth of portfolio trading was highlighted, with the demand and liquidity previously seen in the US becoming more apparent in Europe. Perhaps it was not surprising to hear on the same panel about the further growth of Fixed Income ETF volumes, both trading activities helping to drive efficiencies through the use of basket execution and providing indirect transparency on the pricing of the component bonds. What could have been said is that, while workflows vary across products, the typical process is labour-intensive with traders pushing back and forth spreadsheets, with hundreds of line items, and lots of manual copy/pasting into correct formats for pricing via e-mails. The answer has to be the automation of data-driven workflow, which is recognised and mapped into the correct format so that pricing can be rapidly generated and eliminate keystroke risk.
An underlying theme that permeated a few panels and came to the fore on the Buyside Tech Wishlist panel stream at the end of the day was the subject of Axes. Including an Axe protocol in addition to RFQ, CLOB etc; was mentioned earlier and this morphed into the idea of tagging axes and making them more targeted by the end of day. It was made clear that the buy-side doesn’t want a list of 2,000 LEIs where only 100 are actionable. Half a dozen tags that provide information around things like security, price/yield, size, timeliness, trade terms and counterparty and whether that is principal/agency etc; are required. The idea of integrating these market axes into internal systems to understand the associated use of a protocol that should be used to action that information was also suggested.
One approach to this, currently in production with ipushpull at a leading UK bank, is the creation of an automated workflow to send targeted actionable bond axes, distributed to clients via chat. Publishing relevant axes with live prices in a chat window, which a client can auto-execute against a chatbot in a given size, triggers a full STP process to update downstream systems. Which sounds more compelling than a manual process, exchanging e-mails, and copying information across systems and applications into static chat windows!
Of course, you couldn’t have a financial markets conference without the vagaries of data management being included. From the perspective of too much data, finding the right data, tagging trades and colour using data, gathering data around the lifecycle of the trade from pre-trade to post-trade data and being consolidated in a utility. All of which led to discussions around the pros and cons of a consolidated tape and, “well, if it works in equities ….?” To the point where some felt that data is key and the protocol used is the last requirement, as without aggregating and mapping the data you can’t decide the protocol as you need to interrogate the data first and even overlay NLP. So, many have invested a lot in quants and predictive analytics to help determine what might be the best execution, including liquidity scores for liquidity providers. Suggesting that this shows further examples of where interoperability and interconnectivity of data are important.
Similarly, no conference could go by today without mentioning the use of Generative AI, Large Language Models (LLMs) or whatever generic AI term is the flavour of the month. One panellist provided a list of priorities to consider in this quest, ranging from fast and easy data retrieval, automated reporting, knowledge-based management and finding patterns in data for decision-making. More pointedly, one delegate suggested that this ranged from excitement, distrust, illiteracy and/or optimism, with everyone agreeing that they needed to learn more and understand how it fits with the rest of their technology stack. Many market participants are investigating how LLM projects can best be deployed in their own environment, searching for the efficient frontier of AI and current systems, that elusive combination of increased revenues and customer service alongside reductions in cost and risk. The ultimate goal for most institutions is greater productivity and increased efficiency by freeing up traders and salespeople to focus on things people do best, namely provide market insight, solicit opinions, discuss ideas and create value. Clearly, technology providers, particularly those that seamlessly provide satellite benefits to the core engine, will be in demand.
For all the talk about protocols, data, and new technology, the industry still needs to address its legacy silos and re-engineer its manual pre- and post-trade processes. However, with the collective will of the FILS attendees and a fair wind, there are reasons to be optimistic about the market structure of the future. This leaves two more general comments to end on. There is no single method of Best Execution. And voice trading is alive and well and traders still need to have real-life relationships in order to be successful! Plus ça change, plus c'est la même chose.
With deep technical and market expertise ipushpull is ideally placed to help our customers navigate the choice and integration of new technologies – especially around the use of AI and the pathway to identifying which approach is best suited to your business. Look out for our upcoming blog on what we’ve learnt from our LLM and AI work so far and how we work to create the best combination of tools –ranging from mapping and transformation modules, parsers, machine learning and AI.
Enjoying the networking evening with Matthew and Neil!