As outlined previously, Fixed Income dealers have been very slow to adapt, and even slower to adopt technology that their Equity and FX brethren have been using for decades. This is proven through the stark reality of the daily traded volume compared to overall market size, and more vividly in the glacial increase in electronic volumes. The clear failure point is large trades in the 99% of Corporate and High Yield bonds that are designated “not liquid”.
Dealers are wrestling with a human and technology cost-base that newcomer liquidity firms have solved in a trice. These newcomer firms do one thing exceptionally well, namely, provide a constant two-sided liquidity stream over multiple venues via API. Click-to-trade auto-execution is the name of the game here. But before any of that, these firms apply very straightforward client acquisition documentation and a swift integration process. Oh, and just for the record, there are very few traders or salespeople either, just a bunch of techies working remotely from low-cost locations.
They have deduced that by using low-cost processing and data analysis. It is possible to make a constant price stream in thousands of bonds consistently and profitably to numerous clients over multiple platforms. This new paradigm has begun to permeate the small order part of the volume spectrum where the larger dealers absented themselves a while ago. This still leaves the top tier dealers executing the large order part of the market, but in a resolutely analogue workflow using phones, spreadsheets, email and IB chat. With all this in mind, the key question is, can dealers find a more digital workflow for large order execution?
The answer is yes, but as was often heard on the bridge of the star ship Enterprise, “not as we know it”. Any trade in an illiquid bond is a challenge, no matter what the size. The main reason is the excessive cost for dealers of warehousing that relates to low volume turnover. The liquidity charge for dealers creates a tariff barrier that renders many bonds illiquid. Combine this with dealer’s relative inefficiency in finding the other side in such a trade and two things become clear. Firstly, dealer cost of capital unnecessarily compromises trading costs and secondly, finding matches for the other side can be efficiently performed using data and aggregation tools. Curiously it is the dealers who possess most of the required data. What they do not have is a protocol for distributing it and receiving qualified inquiries from their buy-side clients.
There is also a significant data quality issue. Talk to any buy-side trader and you quickly learn how they are inundated every day with a tsunami of runs, axes, indications and IoIs. Some of our clients receive seven thousand messages per day. Deciphering which are real and actionable is a thankless task. There is no simple or safe protocol to identify and contact the right dealer about a trade.
Furthermore, the variable data quality makes it almost impossible for a buy-side trader to determine or prove best execution, especially after a big move in the market which may have triggered some large portfolio adjustments.
So, in summary we have plenty of data, but it is not well structured, and the quality is extremely variable. There is also a marked lack of suitable protocol to identify the right dealers, monitor the execution cost performance in flight and finally prove best execution.
Dealers need an incentive to provide high quality and high integrity data and the buy-side needs a trading protocol that encourages inquiries. They remain at the centre of price formation as they receive an abundance of inquiries and inputs relating to the credits they trade. Their value to the buy-side is to disseminate this data in price form and provide the means to trade those credits by either taking principle risk or passing through the risk to another counterparty. Buy-side access to this data is the first part of the puzzle.
The first tricky piece of the jigsaw is trying to improve data quality and integrity. The best way of going about improving this is by segmenting dealer data into four types. Firstly, information is needed around the click to trade price and spread streams. Then data is required for indicative price and spread streams, as well as actionable axes and indications of interest. This all then needs to be underpinned with RFQ trading protocols.
This means that for pre-trade best execution analysis, a buy-side can use exclusively live, actionable data. A fund manager can also see actionable axes within this context. The other noise can also be considered, but never confused with the higher integrity data. Post-trade TCA also uses the same data but may also include competitive RFQ quotes. This is where slippage from screen quote can be analysed. The next piece around Portfolio Trading (PT). A PT application can radically improve the workflow of sending out all those spread sheets to multiple dealers to figure out whether the trade in question is viable. During execution a PT app tracks the execution performance and records the underlying data for TCA.
This is all possible by simply re-configuring dealer data into precise types. It will cut down the volume and improve the veracity of dealer data. It dramatically improves execution workflow efficiency. It makes TCA an effective activity rather than a box-checking exercise. Those that link these pieces together in the New Year will be that one step closer to solving the longstanding liquidity issues synonymous with Fixed Income markets.