September 28, 2022
Originally published by TabbForum - Click here to read the full article
What is the current role of fixed income technology and how do market participants see it evolving? Market structure consultant George Bollenbacher addresses these questions, explaining what makes the fixed income ecosystem so different than the other asset classes. In this article, Mr. Bollenbacher explores how technology is currently being used in fixed income, and what market participants on the buy side and sell side see on the horizon in its development.
Advanced technology has impacted all financial markets, increasing speed, choice, and the volume of data. But these changes have been slower in fixed income (FI) than in other markets like equities and currencies. As the world and the world’s FI markets begin to emerge from a global pandemic, it is appropriate to look at the current role of technology in those markets, and to see where it will lead the industry in the future. In researching this topic, I interviewed a number of representatives of leading tech firms that serve this market and buy-side firms, as well.
Since FI is somewhat behind other markets, like equities and FX, in the adoption of advanced technology, it is necessary to understand how this market is unique. The most salient factor is the disparity of issues within the market. Take liquidity. The market ranges from on-the-run Treasuries, where each issue can trade billions of dollars in daily par value, to seasoned municipal issues, where a single issue might not trade once in a year. In terms of trading, those securities couldn’t be further apart. In between are such categories as off-the-run Treasuries, investment grade corporates, mortgage-backed and asset-backed securities, and high yield bonds. Each has its own niche in the liquidity spectrum.
A second factor is the sheer number of individual issues. According to data provided by the CUSIP Bureau, there are over 2.3 million active FI CUSIPs (not including MBS pools), compared to about 475,000 equities (which number includes common stock, preferred stock, warrants, rights, mutual funds, ETFs and equity structured products). A market with almost five times the number of issues requires a fundamentally different approach to the use of technology.
A third factor is the terms of the securities themselves. They range from zero-coupon short-term securities with de minimis default risk (Treasury Bills) to securities with monthly interest and principal payments, significant default risk and an unknown average life (non-agency MBS) in the same portfolio. When combining the numbers of securities with the diversity just described, it is obvious FI is a completely different market than equities and FX.
As a result, the tech demands of the buy side cover a wide spectrum. One spectrum is the transaction sequence, as laid out below. The two rows are actually one sequence, but split into pre-trade (top row) and post-trade (bottom row).
It is important to note that allocation doesn’t happen for some, and maybe a majority of participants, such as hedge funds and mutual funds. But it is important for traditional asset managers. As evidenced by this diagram, it is virtually impossible for a single type of technology or algo to apply to all parts of the flow. So technology vendors often specialize in one or a few segments, often contiguous.
Another spectrum is security type and liquidity. Along that spectrum market testing and execution can take on a wide variety of forms. As a result of this disparity across several spectra, a simple question like “What is the biggest technology issue that bond buy-siders have currently?” can generate a wide variety of answers. From the perspective of a market venue, like BondsPro, David Parker, Head of MTS Markets International, says that the problem for buy siders is too many venues and too much data to process. For an aggregator like Trumid, Chief Revenue Officer Bryan Hankins says that the problem is aggregating and making sense of a large and disparate data universe. For an aggregator, distributor, and trading company like InspereX, President David Rudd says it’s automating a set of cumbersome processes around pricing and execution, and trusting the available data.
However, ultimately the buyside technology demands can be divided most effectively along liquidity lines. At the liquid end of the spectrum, which has most of the trading volume and most of the technology, but not most of the outstandings, technology needs tend to center on a proliferation of data, venues and possible counterparties, and on how to route and manage orders. In the illiquid segment of the market the problems center on valuation of issues that might not trade once in a year, and the handling of orders that could serve to move the market just by their arrival.
At the liquid end of the spectrum, the growing number and type of trading venues/counterparties worries everyone and is a problem to be solved. It actually affects buy-side participants at two points in the sequence: gathering the market data necessary for investment decisions, and market testing and execution of the order generated.
According to Adam De Rose, Director of Product Management EMEA SS&C’s Eze, there are now 164 destinations where the buy side could execute trades in FI securities, a number he attributes to John Greenan at Alignment Systems. And, according to Ganesh Iyer, Chief Marketing and Strategy Officer of IPC, all those venues/counterparties actually make for a quite fragmented market. And not all venues are available to any buy-side user. MultiLynq’s Co-Founder Patrick Scheideler sees the biggest issue in the credit markets as the growth of e-trading, advancing beyond the current capabilities of the technology most buy-side firms are using, and overwhelming the buy-side’s infrastructure. This will require a simplified approach on both sides of the market, he adds.
Tradeweb’s Managing Director, AiEX and Workflow Solutions, Brian Devers thinks that in the liquid sector there is still a big task around identifying a correct protocol and liquidity pool for each trade. He notes the promise of AI in this sphere, allowing slimmed down trading desks to handle an even larger flow. Other technologists, like 28Stone’s Founding Partner Tom Dolan, see this segment as somewhat less liquid than it seems. As a result, even in this market, he sees trading tools and strategies as essential to achieving something like best execution.
David Landisman, SVP and Sr. Director of FactSet, recommends traders use price aggregators to access market data, so as not to telegraph trading intentions while still planning a strategy. Once that data is accessed, Vuk Magdelinic, CEO of Overbond, sees an important role for smart order routing, including deciding which orders can be handled completely by algos and which require human intervention. It also helps determine the path to best execution in venue and trading protocol selection and reduces the number of child trades needed to execute the full desired size.
Another issue in the liquid space is that the FI market is still essentially a principal market. It is where the buy side generally faces each counterparty directly. That means that some, if not many, of the prices that a manager might see on a particular product are not useful to that firm if it doesn’t already have a trading relationship. Thus, one of the useful technologies in the liquid market is the screening of market data to indicate which prices are available to trade on and which are out of reach. Of course, anonymous venue prices are always available, as long as the manager is a member of, or has access to, the venue.
At the other end of the spectrum, the problem isn’t a diversity of data and venues. It’s getting current market information about a specific issue. Since this problem is as old as the markets themselves, and since bond mutual funds have to price every holding every trading day, many large buy-side firms have developed pricing matrices that ingest trade prices for a benchmark set of securities and derive values for everything they own. Or they can utilize commercial services like Bloomberg. For a bond market aggregator and trading company like InspereX, President David Rudd says the biggest problems in this sector are automating a set of cumbersome processes around pricing and execution, and trusting that a trader has all the available data, and that it’s valid.
Tradeweb’s Managing Director, Head of U.S. Cash Credit, Iseult Conlin sees a proliferation of data in more opaque and illiquid sectors of the fixed income market, which creates a need for information filtering, that removes what doesn’t apply and purifies what does, potentially allowing for greater use of EMSs in trading and transaction decision-making.
In addition, the transaction process in the illiquid sector is somewhat different from the liquid. In liquid securities, the executor (often the manager) scans the available markets and decides on 1) which execution strategy to use; and 2) which venues or counterparties to contact, often with a RFQ. In the illiquid sector the executor must make a strategy decision with very little market data on the security(ies) in question, and must depend on whatever data is available for comparable securities.
According to LTX’s CEO, Jim Toffey, new pre-trade liquidity metrics are available to help traders discretely identify trading opportunities in less liquid bonds and quantify the viability of these opportunities in the current market. A complication with inquiries such as RFQs is that they can move the market in that particular security before the executor even makes a decision. The new liquidity metrics allows traders to anonymously find natural contra liquidity while reducing information leakage.
To top it off, the trading volume in this market sector is significantly lower than in the liquid sector. Bonds tend to trade fairly frequently for a couple of months after issuance. Then the volume dries up gradually over the next year, until it can be close to zero for the rest of its life. The only time the volume picks up appreciably during the latter period is if the credit quality of the issue is perceived to change, especially if it is expected to decline. With these lower trading volumes, the buy side is less likely to justify an expenditure on advanced technology for this sector. As a result, the illiquids are much more difficult to automate, and we find far fewer technology vendors that specialize in it.
Another factor effecting technology use is two specialized types of trading: portfolio trading and spread trading. Portfolio trading occurs when a manager (or other market participant) trades a portfolio of securities at a single price, expressed as a percent of the total par value. Although this is a convenient way to trade a large number of securities, the portfolio must be broken up into individual securities, with their individual prices, for settlement. This type of trade has grown in volume with the growth of ETFs. Spread trades are usually done in a single security or a small portfolio. Still, they are executed at a yield spread to a specific security, usually Treasuries. Sometime after the trade is completed, the yield of the benchmark security is derived, and then the price of the traded security is determined.
Given the disparity of the FI markets, across several spectra, we can draw two conclusions: 1) advanced technologies can provide significant value throughout the transaction cycle; and 2) it can be extremely difficult to apply advanced technologies in this space. Let’s look at both aspects of this dichotomy.
First, from the value side. As the types of trades in the market multiply, and as the number of possible venues/counterparties expand, the first place to look is execution management. This encompasses two distinct choices: 1) the trading strategy, which is based on urgency, liquidity and market outlook; and 2) the execution location. Market pros often conflate those choices, but experience has shown that choosing the wrong strategy can undo the choice of location, and vice versa.
With the growth in the number of venues/counterparties, there has been a quantum change in two inputs, 1) the availability of market data (which can be needed even faster than TRACE provides it); and 2) the availability of executions for a particular manager. While humans can deal with this increase, there is no question that technology can add tremendously to the pre-trade process.
Algos that scan all the data sources for prices and volumes on all FI securities distill much more information more quickly than a human. This distilled information can be used to enhance the first step in our flow, market analysis. It can also screen for market data that comes from sources a manager can’t use, but that information can be utilized in managing an order to the venues/counterparties the manager can use. That kind of filtration and application of data is at the pinnacle of modern technology, and may not yet be readily available in the FI markets.
On the difficulty side, let’s return to the diversity problem. One aspect is comparing security characteristics in a 2.3 million node matrix. In order to move from a theoretical relationship, useful for daily pricing of a portfolio, to an actionable relationship, useful for executions, the algo has to incorporate many more data points and trend analyses. In many cases, managers don’t believe that algos can reliably detect the kinds of aberrations that they need to generate alpha, so they only utilize technology once the investment decision is made.
The second difficulty is the constantly changing nature of FI markets. Not only are new venues and counterparties appearing (and disappearing), but new trade types are evolving. As a result, any FI trading technology must evolve continuously, not a set application one can buy and run. Because systems like these depend on receiving data through APIs, they must adapt to data and API changes on the fly. It is difficult to design systems that intelligently adapt to these changes, so many vendors have to regularly update their interfaces, and the users of those systems have to integrate those changes into their own applications in a never-ending stream.
Finally, asset managers resist using advanced technologies for two reasons. The first is that it’s difficult to justify the costs of advanced technology, since the perception is that human management of the process is adequate. In addition, according to LTX’s Jim Toffey, screen real estate is tight, so integration between e-trading platforms and OMS/EMS providers is key to advancing technology adoption.
The second reason is less obvious – the natural resistance of those currently performing the job. We’ve seen this play out in many industries, like automotive, but it is less obvious, and less talked about, in an industry where there is a high premium on the creative intelligence of the work force.
What does the Buy Side think about some of these issues? One large manager said, “What little I would share for FI trading is that axe volume is a challenge. Too many brokers, too many, too frequent updates. Solutions to the axe problem are tech filtering/throttling/aggregation capability as well as asking/dictating to brokers to limit axe volume.”
Another manager said, “I think portfolio trading is still really interesting. It has gotten a lot of attention recently, a lot of hype maybe. How useful it is depends a lot on the end user – and there are certainly some drawbacks and risks that don’t get enough attention. We are also very interested in fixed income EMS solutions. It’s early days there and half the time I bring up FI EMS – the answer I get back is, ‘What exactly is a Fixed income EMS?’ I think on the technology side that solution is just getting started.”
Finally, in an article in Markets Media, Lindsey Spink, Co-Head of Global Fixed Income Trading, American Century Investments said, “the Street is less likely to provide tighter bid-offer markets, because they have less outlet to recycle risk. The result is that the actively traded universe shrinks to anything newly issued and larger-cap names, and the smaller, and ‘off the run’ bonds become increasingly illiquid. Liquidity has continued to migrate outside of traditional channels, which makes the need for new technology more urgent.” He also said, “Gaining access to new pools of liquidity is very difficult with legacy systems. Moving a large ship towards the standards of application programming interface (API), algorithms, and execution management system (EMS) is not simple. The fixed income market is not going to “equify,” but it is evolving with its own set of new standards and practices.”
So much for the past and the present, as the all-important question is what role will technology play in the future? Let’s begin with three years into the future. The responses were interestingly varied on technology’s role.
BondsPro’s David Parker expects dealers to move to an “electronic first” mentality around executions and for this pattern to migrate down the liquidity spectrum. He also expects investors to start using technology to set up and monitor hedges. David Landisman of FactSet expects electronic trading and venues to expand, and he sees regulation of that trading to increase. IPC’s Ganesh Iyer expects more participants in the dealer and venue roles, using more advanced technology, and more trading on all-to-all venues. Trumid’s Bryan Hankins expects human expertise to be combined with designing, utilizing and supporting algos, as opposed to algos simply replacing people. And 28Stone’s Tom Dolan expects a growth in venues and trading modalities to lead to disintermediation within the trade flow.
Click here to read the full article on tabbforum.com