I recently watched a video recording of a panel discussion held by the SEC’s Fixed Income Market Structure Advisory Committee (FIMSAC). The topic was Pre- Trade Price Transparency (PTPT). The Committee has been seeking information on the current status of PTPT, and what changes could be made to improve it. These are some of my thoughts.
The Driving Force Behind Establishing Pre-Trade Price Transparency
PTPT has been an industry (or at least, an SRO) goal for many years. The Municipal Securities Rulemaking Board (MSRB) published an informative brochure in 2018, “Milestones in Municipal Market Transparency- The Evolution of EMMA”, which tracked the efforts and accomplishments of the MSRB in providing information to the public.
In 2012, the SEC, in their Municipal Report, suggested two specific ways that retail investors could be made more aware of the prices that they should expect to receive when buying or selling bonds. One suggestion in the Report was to consider amendments to Regulation ATS to require an ATS with material transaction or dollar volume in municipal securities to publicly disseminate its best bid and offer prices and, on a delayed and non-attributable basis, responses to “bids wanted” auctions. Another suggestion was to consider rules requiring a brokers’ broker with material transaction or dollar volume in municipal securities to publicly disseminate the best bid and offer prices on any electronic network it operates and, on a delayed and non-attributable basis, responses to “bids wanted” auctions.
For scorecard watchers, it should be noted that several other recommendations from that same report, such as creating a Best Execution Rule, requiring Dealers to show their mark-up on some sales to customer, and establishing a Prevailing Market Price for bonds (showing where the mark-up originated from), have already been implemented. History would indicate that when the SEC makes a “recommendation”, it more often than not gets done.
Historic Difficulty in Providing Pre-Trade Price Transparency
PTPT has historically been a difficult goal. Unlike the stock market, where an investor has multiple venues for checking a price, the municipal market’s structure has not lent itself to the same pricing ability. As several panel members commented at the FIMSAC meeting, the pricing of any particular bond has faced several hurdles:
- The vast number of issuers (some 50,000, per the MSRB) and the even more vast number of securities that they issue (over 1 million).
- Municipal bonds, after their first issuance, tend not to trade actively. While there may be billions of dollars of bonds trading in the market every day, any one individual bond may not trade for months or years.
- The municipal market is bifurcated into fifty states, three territories, and the District of Columbia, each with their own taxation schedules. Retail investors tend to purchase bonds within their own state, and many times within their own geographical “domestic preference”.
For these and other reasons, the bond quotes in the secondary market for municipal bonds have historically been highly subjective, based on the opinions of traders who weighed risk not only on the reference data of a bond, but also on the firm’s ability to re-sell that bond, at a profit, to another customer.
The Rise of The Algorithms
The subjective opinions of traders are giving way to the icy calculations of algorithms, thanks to the vast amount of data now available in the market.
Some years ago, while strolling through the Vendor section of the FINRA National Convention, I was amazed at the number of times I overheard the topic of data, and the ways that data could be used to improve statement pricing and bond evaluations. While interesting to me at the time, it seemed that the vendors’ products mostly targeted a different section of the business than my own, which was, at the time, working at a Municipal Securities Broker’s Broker (MSBB).
Fast forward a few years, and I began to see how the use of this data was creeping into my business model. I would enter a bond into my company’s auction system, and be amazed to have a bid come back from a dealer within seconds. I knew that there was no way that a human trader could have done the normal work that traders do before bidding on a bond (checking the rating, seeing if there were already similar bonds in inventory that had or had not been selling well, checking EMMA to see if there were any other recent trades, etc.). It didn’t take long to realize that it wasn’t a human being on the other end of that electronic feed, but rather a computer that had been fed an algorithm, telling it what to bid for any bond with those certain characteristics. The algorithms were a bit sloppy in the first years of their use- they sometimes missed special features of a bond, usually a sinking fund or a recent rating change. But, over the years, the data has become increasingly detailed, and the algorithms have become extremely efficient.
These databases continue to expand as more information becomes available to them. For example, while all firms are required to report their trades to the EMMA system within fifteen minutes (meaning that this post-trade data is available to be built into pre-trade models), firms like Municipal Bond Information Services (MBIS) are also now gathering bid information from MSBBs on bonds that did NOT trade. This has vastly increased the amount of pricing information available, as only a small percentage of auction bonds actually trade.
The thousands of bond offerings sent out by dealers on electronic platforms and message systems are also providing sources of data to these information-gathering firms.
Where will this take the Municipal Market?
The pressure by the SROs to provide PTPT to retail investors will continue to push dealers toward the development of more and more accurate algorithms. Originally used by just one or two firms, they are now becoming more and more pervasive. This may be how we will arrive at Pre Trade Price Transparency for the retail investor; the data may soon provide a model (price) for every bond, and a retail investor will be able to enter a Cusip number into their firm’s web site and have a price returned showing exactly how that firm would sell (if the bond was actually available) or buy that bond at that moment. The only thing missing is the measure of competitiveness in the quote. It’s a market, but is it the best market?
To solve that issue, the SROs might eventually require the market to morph into a collective, single source where any customer could request and receive bids or offers on municipal bonds, receiving quotes from multiple market makers. The consolidation of MSBBs and the ease of using Alternative Trading Systems (ATSs) already points towards this possibility. And, while the sheer number of Cusips might prevent the publication of a daily quote on every bond, it is possible that the ability to receive immediate responses to individual requests is in sight.
I would add one word of caution as we move towards a market run strictly by data and models. The market has been “commoditized” like this once before, during a period when bond insurance offered by companies like MBIA, AMBAC, and FGIC was at its peak. Since all issues insured by these companies became AAA rated, the bid side for all of them was more or less the same. The need for the subjective opinions of local trading desks and traders decreased, as did their numbers. When the financial crisis hit, the insurance companies lost their ratings, and there was “no bid” for many issues. Suddenly some issuers were more equal than others.
During the next credit crisis, when a firm turns off its machines for fear of buying bonds based on a disintegrating model, the market will once again turn towards human subjectivity in order to find a price. Let’s hope that there are still some humans left in the business when that occurs.