In recent years, an unfortunate byproduct of decreased dealer balance sheet is the resulting problem of information overload in the fixed income market. The crux of this issue is the manual trading process utilized in varying degrees by most market participants. In the absence of out of the box solutions to solve this problem, many larger buyside firms have built their own tools to aggregate the millions of data points they receive daily and to uncover many more trades at favorable pricing.
Even as there is an increased need to solve for data overload, the rate at which fixed income has adopted technology to address this problem has been slow, though improving.
Why is that?
While there are many factors, cost is often cited. However, cost cannot be discussed in a vaccum without discussing return on that cost. For a buyside tool to solve the buyside's data issues, regardless of whether an internal build or the purchase of a third-party solution, the real question is: What is my quantifiable return?
We've never known-- until recently.
Lately, there has been an increase in buyside technology spend and in the article, "Institutions reported an average trading desk budget (for compensation and technology expense only) of $35.8 million, a level essentially flat from 2015." Further in the same article, it is estimated that in 2016, 65% of that budget went to comp, thereby leaving $12.5 million, or 35%, for fixed income technology costs.
Given that fixed income is comprised of illiquid, non-uniform assets, voice trading isn't going away anytime soon. According to this article, it’s not surprising that for the buyside, "...MiFID II compliance and other regulations, have been prime drivers of technology spending." At the same time, the article says that, "...[I]n order to maintain efficiency, desks need to have the latest and best technology at their collective fingertips to preserve alpha and meet retail investor demands to keep operating costs low and transparent."
So, if you need people to trade illiquid, non-uniform bonds and the remaining technology budgets are being eaten up by regulation, it is not a surprise that only the largest buyside firms have built these types of tools. Building software is not for the faint of heart and often requires lots of expertise and big budgets. A 20% increase, or $2.5 million, in this average technology budget ($12.5M) would not cover the expense to hire the right teams to build this software. It also detracts from the buysides' core proficiency: putting money to work.
While $12.5 million seems like a large expense for buyside firms for technology, there also is a hidden cost-- and it's not included in this $12.5 million, namely, opportunity cost. How large is this opportunity cost? Massive. In fact, according to Tabb Forum, "...Inadequate technology costs buy-side firms up to 50% of revenues...".
Cost vs. Return
In order to determine how large this opportunity cost is, we need to understand the return of one of these platforms. Until now, the return on an investment in one of these systems has been largely unknown, as large buyside firms have no particular incentive to share the data from their experience.
Recently, however, Alliance Bernstein has been getting a lot of press about their ALFA platform, which was built to try to address internally the information overload issue that plagued them. In conjunction with MarketAxess, one of the electronic trading platforms they employ to trade corporate bonds, AB has determined via MA's TCA tool that they are gaining 4.5 bps in benefit on the volume of bonds they trade.
Let me repeat this: For all Investment Grade (IG) Corporate bonds that AB traded in a year, their execution costs are estimated to be improved by a whopping 4.5bps owed to technology.
4.5bps is MASSIVE. How massive? Using AB's 4.5bps trading improvement due to technology, we can estimate the dollar amount of not having a similar solution will cost you.
Here's our example
Let's say that you trade only $5 million par amount of investment grade corporate bonds daily. With ~220 trading days in a year, that's $1.1B of par traded annually.
Let's also assume that the trading volume has an average duration of 7 years and that the price value of 1 bp (PV01) is ~0.0675. Multiplying 0.0675 X 4.5bps X annual trading volume leads to a potential cost saving of $3.3 million dollars annually for every $1 billion of high grade bonds traded.
Another way of looking at this is to realize that:
.0675 (PV01) X 4.5bps (execution improvent owed to FI Tech) X $1.1B (Annual trading Volume) = $3.3M (Missed Revenue)
In short, for every $1.1B in trading activity annually you are potentially giving up $3.3M in opportunity cost.
This is a significant amount of opportunity cost and it’s an annuity – it happens every year. Again, keep in mind that the average technology budget is about $12.5 million.
It gets worse...
As bad as these numbers are, it actually gets worse, for a few reasons:
- Volume traded: Most "average size" buyside firms trade well more than $5MM bonds daily. Want to see the opportunity cost for higher transaction volumes? Using the same formula as above and different par traded amounts, here is what buyside firms miss in opportunity due to a lack of technology tools:
- Bond liquidity profile: If you trade bonds that are more illiquid or more nuanced than IG, then it is conceivable that the execution improvements from a buyside workflow tool are much higher. We would suspect that for bank loans, high yield and distressed debt, many Municipals and MBS, the execution improvements are much larger.
- Investment style: AB manages mutual funds which is a uniform investment and benefits from potentially larger transactions. On the other hand, if you manage separate accounts (SMAs), it is highly likely that you are giving up much more than 4.5bps due to trade size, larger # of portfolios, portfolio customizations and differing cash needs of each portfolio.
- The curve: The example above uses a 7yr average duration. While we aren't predicting when rates will rise and the treasury curve will steepen, it's conceivable that it will steepen at some point in the not too distant future. Using the same inputs as above, a 10yr average duration on trading activity results in $4.3 million (vs $3.3 million) in missed opportunity costs on each $1.1 billion par amount traded annually. Thus, as the curve steepens, it will cost you even more.
What to do?
Clearly, there is value to buyside fixed income platforms that aggregate all the fragmented sources of data and route trades proactively, especially when done in real-time. The central question is at what cost? If you have millions of dollars to spend, the technological expertise in house and don't mind missing out on the additional alpha you otherwise could generate for an extended period, it may make sense to build on your own. Otherwise, your best bet is to consider an existing platform such as FixtHub.