February 25, 2016

Municipal Bond Pricing, Part II: What Trade Data Reveal—or Not

The second in a two-part series examines the use of EMMA trade data in establishing fair pricing and the problems that arise when trying to interpret the data

Pamela W. Peterson

The first part of this two-part article (published in the Fall 2015 issue of the Securities Litigation newsletter) discussed basic elements of fair pricing for municipal bonds, including the requirement of a fair price reasonably related to the market value of the security, and the challenges of finding that value in a thinly traded market. It also introduced the useful Electronic Municipal Market Analysis (EMMA) service, which provides access to data about municipal security trades and is made available by the Municipal Securities Rulemaking Board (MSRB). It ended with a discussion of how much a municipal securities dealer may charge as markup, markdown, or commission on a trade. An important question related to this discussion is this: How can one assess the fairness of that charge if it is not shown on a trade confirmation (by rule, commissions on as-agent trades are disclosed)?

If there’s recent trading history for a given CUSIP (the nine-digit alphanumeric identification number for a given security, including municipal bonds), it may be possible to estimate markups and markdowns. Here’s an example pulled from EMMA, using California general obligation bonds (with a September 2026 maturity, CUSIP 13063CER7).

Let’s look at the bottom three trades in this chart, which all appear to have happened at the same instant. One reasonable scenario—and I emphasize that this is merely a guess—is that, on April 29, 2015, the customer (Smith) of one broker-dealer wanted to sell a par amount of $25,000 of these bonds. The trading desk at that brokerage called around and found another broker who wanted to buy them. They dickered over a price. The second broker likely knew it had a customer (Jones) who might be interested. While Broker One was calling Smith back to confirm the price, Broker Two was calling Jones to confirm the price of that transaction. At 15:06, Broker One pulled the trigger, executing a trade with its customer at 119.092 and immediately sold the bonds to Broker Two at that price (Broker One may have been in an advisory customer relationship where it could not charge markdown, which would have meant a lower price to Smith). Broker Two could charge markup and promptly marked the bonds up by .100 and sold them to its customer at 119.192 (a markup of 0.084 percent).

But this entire scenario may be wrong. It’s just as likely that Broker Two’s client Jones was seeking to buy some of these bonds, and Broker Two called Broker One (let’s say they’d been co-underwriters not long ago), who knew that Smith needed to raise some cash for a sudden medical expense. Broker One then bought the bonds (without charging markdown) from its long-time customer Smith, sold them to Broker Two, and so on. Finally, note that there’s a second interdealer trade at the same moment, and two customer purchases 10 minutes later, at a presumed markup of .418, or 0.35 percent. There are a number of ways to interpret those facts as well.

So, as noted in the first part of this article, the picture delivered by EMMA is both intriguing and question-raising. Without knowing more about the nature of the customer-dealer relationships and who initiated the trade, one can’t be positive about what happened. However, one thing can be determined here: If the difference between the interdealer trade price of 109.092 and the various customer purchases is due to markup, all the resulting aggregate prices bear a reasonable relationship to the market value of the security as determined by a contemporaneous interdealer trade. It’s also worth noting that within the week prior to this set of trades, there had been an interdealer trade at 118.777, suggesting that the interdealer trade at 119.092 was also appropriate.

Finally, let’s hypothetically consider the situation if the last trades in this CUSIP had been days or weeks in the past. In that case, some of the difference between the last interdealer trade and the current customer transaction could be due to shifts in the interest rate environment generally. Say, for example, that a dealer bought some bonds with a 4 percent nominal interest rate, and in the last few weeks, interest rates have dropped by half a percent. The market value of those bonds would rise because they now have a better interest rate than can be obtained by buying newly issued bonds. The dealer, if it were doing an interdealer trade today, would price those bonds to account for that fact. Some critics regard this price appreciation as somehow being a “windfall” profit and argue that the dealer should start with its own cost—no matter what has happened to the market in the meantime—as the basis from which to calculate markup. Yet, they are unwilling to agree that, if interest rates move up, the dealer can still use its weeks-ago cost as a starting point for calculating markup. In other words, such critics argue that a dealer who holds bonds in inventory must give away market appreciation but “eat” market depreciation. No business can survive long under such fundamentally unfair strictures.

How is markup/markdown determined? Some brokerages impose a fixed transaction cost. Some have a range that’s determined by a “grid,” which usually takes into account the maturity of the bond and the credit quality of the bond. The private-client broker can agree with the client to take less than the maximum amount in that range but can’t exceed the maximum amount unless there are very special circumstances (and, usually, some kind of prior internal approval).

Why is there a “grid”? Why don’t all dealers just charge all municipal security trades a fixed percentage? There are several reasons. First, there’s a wealth of disclosure to be made on a secondary market retail trade, and the longer the customer holds the bond, the more likely it is that something could go wrong, triggering a “20/20 hindsight” lawsuit. Also, the lower the credit quality at the time of sale, the more likely it is that something can go wrong before maturity. Finally, even if all goes well, the longer the maturity of the bond, the longer the dealer must cope with the operational necessities and cost of distributing interest or handling any possible bond calls.

If one considers that a dealer is paid only once for assuming responsibility and legal exposure that may last for decades, a one-time charge of 2–3 percent of the purchase may not seem at all unreasonable. While the sophistication and abilities of customers vary immensely, the dealer has regulatory duties that do not vary. As far as the regulators are concerned, all retail customers are pretty much the same, yet one customer, a long-time veteran of the municipal bond market, may need only a few minutes to rapidly review the bond’s characteristics and may already have a sound appreciation of credit ratings and interest-rate trends, while a second customer may need patient, step-by-step analysis and explanations of a particular bond. The more general part of these explanations may be done several days prior, but all this still takes up a broker’s time, which is worth something. Servicing some customers can be quite time-consuming, particularly where a customer may discuss and reject numerous suitable offerings for quite specious reasons before choosing to buy a particular bond. Factors like these can affect whether an individual broker decides to charge a markup or commission that is the maximum or less than the maximum allowable under the dealer’s formulas.

When a customer sells a bond, rather than buying it, the dealer will pay slightly less than the market price. The difference is the markdown. Because the dealer usually has a lessened burden of disclosure (the customer is generally less concerned about the characteristics of a bond being sold), and because the dealer will no longer have to deal with the bookkeeping aspects of maintaining custody of an interest-paying security for its customer, markdowns are generally smaller than markups.

It is important to understand that the model implied above (determine an interdealer price, then determine a fair markup/markdown, and the aggregate must bear a reasonable relationship to the market price) is only one way to achieve regulatory compliance—there are others. For example, assuming that there are a number of recent, comparably sized trades with customers in a given, frequently traded municipal security, a dealer might also observe “same side of market” comparably sized transactions, set the aggregate price based on those transactions, and then compute how much of the aggregate price would be allocated to markup and markdown based on its internal rules. “Same side of market” means that if a customer is seeking to buy a security, the dealer must look at other trades where the customer bought, not at trades where the customer sold the security.

The protection of retail customers through full disclosure and discussion and a closer analysis of suitability is more costly for a dealer than trading with institutional customers. Generally speaking, such customers are sophisticated, self-directed “smart money” investors who meet asset or status thresholds and have agreed that they are capable of doing their own diligence and making their own decisions. This status doesn’t, however, excuse a municipal dealer from trading fairly with them. In addition, the vast majority of trades with institutional investors are likely to result in the securities being custodied somewhere other than with the executing dealer, typically with a prime broker or custodial bank. Overall, it’s more time-efficient and desirable for a dealer to transact with institutional accounts: There’s less long-term legal exposure, and there’s less ongoing housekeeping because the securities won’t be held at the dealer. These are some of the reasons why the per-bond markups and markdowns charged to institutional accounts are usually lower than those charged to retail customers. Municipal trading desks also follow the common practice that bulk purchasers of anything get some form of quantity discount.

Retail customers of municipal securities dealers absolutely have a right to expect that the securities they buy will be fairly priced, that they will get full disclosure about the benefits, features, and risks of those securities, and that the securities will be suitable for them. However, the information contained in these articles may help such customers or their counsel understand more about what the information shown on EMMA means and why it can be confusing. The EMMA website has a number of interesting educational links and videos that can also help investors and their advisors. Finally, investors can always ask their brokers for explanations of pricing and keep right on asking until they understand the answers.