Pre-Trade Market Activity: What Has Changed Since 2015?

The views expressed in this commentary are those of the author and do not necessarily reflect the views and positions of the MSRB.

Abstract

Since releasing a research report on pre-trade market activity in October 2018 (based on data from 2015), the Municipal Securities Rulemaking Board (MSRB) has obtained more recent quote data from the same two alternative trading systems (ATSs) with a significant presence in the municipal securities market and conducted an in-depth analysis for the period from June 1, 2018 through November 30, 2018. The analysis indicates that there was a significant increase in the amount of responses to Request for Quotes (RFQs, also known as “bid-wanteds”) and live quotes in the three-and-a-half-year timespan between 2015 and 2018. For RFQs, the preliminary analysis confirms the results from the prior analysis that the execution rate on an ATS platform was higher when more responses were received. For live offer quotes, the analysis indicates that live quotes increasingly provided a valuable pricing indicator to the market, even though a majority of live quotes only represented one (offer) side of the market and 22% of all trades (and 58% of inter-dealer trades) were executed on an ATS platform. Quoted offer prices may have become more visible to market participants, and more informative to execution prices for inter-dealer, customer buy and customer sell trades, as a result of increased quote provision and offer price competition.

The authors welcome feedback and suggestions on this report as well as recommendations on additional data and analysis that could be helpful to municipal market stakeholders. Please contact Simon Wu, MSRB Chief Economist, at [email protected] or 202-838-1500.

Introduction and Background 

In October 2018, the MSRB published a research study (the 2018 MSRB Report) on pre-trade market activity for municipal securities. Utilizing RFQ and live quote data provided by two predominant municipal securities electronic ATS platforms, in conjunction with trade data from MSRB’s Real Time Trade Reporting System (RTRS), the 2018 MSRB Report conducted an in-depth analysis covering a four-month period from February 1, 2015 through May 31, 2015 (“Phase I Period”). The main contribution of the 2018 MSRB Report to the existing literature was in analyzing whether the currently non-publicly available pre-trade information has implications for the broader market’s price discovery process.

This updated report seeks to assess how the pre-trade market changed over the three-and-ahalf-year period, and to determine whether quote information on ATS platforms continues to be valuable for price discovery purposes and provides value for investors and market participants.

Given the ever-changing dynamics in the municipal securities trading world, it is prudent to analyze newer data to provide further insights into the informativeness of ATS quotation data, as well as the latest composition of the market. For example, some of the existing ATS platforms appear to receive significantly more responses to RFQs and have significantly more offerings (CUSIP number) for live quotes than a few years ago. Furthermore, proprietary trading firms and algorithmic trading firms have become frequent users of ATSs in recent years. As a result, the MSRB obtained newer ATS data for the six-month period from June 1, 2018 through November 30, 2018 (“Phase II Period”) from the same two ATS platforms (“ATS 1” and “ATS 2”) prominent in the municipal securities trading market. This updated report seeks to assess how the pre-trade market changed over the three-and-ahalf-year period, and to determine whether quote information on ATS platforms continues to be valuable for price discovery purposes and provides value for investors and market participants.

Brief History of Pre-Trade Transparency 

Pre-trade information broadly includes quote data (bid-side and offer-side) signaling trading interests available on electronic platforms or through non-electronic means, new issue pricing scales, yield curves and indices, evaluated prices, trading in similar securities and other material disclosure information. For purposes of this report however, pre-trade information specifically refers to the narrower definition, which is the indication of size and price of prospective trading interest in specific securities. This includes responses to RFQs and live firm quotes of a specified size—that is, a commitment to buy or sell a specific quantity of a municipal security at a stated price.

Municipal Securities Market Structure and Electronic Trading 

The municipal securities market provides investment and trading opportunities for investors—both retail and institutional—and other market participants. It largely functions as an over-the-counter market, where investors place their orders with brokers, dealers and municipal securities dealers (collectively, “dealers”) directly. By purchasing municipal securities, investors are looking for a promise of income from interest payments—usually semi-annually—and the eventual return of the original investment, or principal. Other market participants, such as dealers seek trading profits by making a market for municipal securities and charging a spread (the difference between the bid and the ask for a security) and/ or a commission on trades with investors or other market participants. For example, after receiving a customer order to trade a municipal bond, dealers either execute the order by committing dealer capital (principal trades) or by searching for an intermediary in the market to facilitate the transactions (riskless principal trades or agency trades).

Market participants such as dealers, proprietary trading firms and institutional investors may prefer using an ATS to find counterparties for trading without broadcasting their trading position to the market.

As discussed in the 2018 MSRB Report, the advent of fixed income electronic trading venues changed the trading landscape of the last decade. The main functions of an electronic trading venue such as an ATS or some broker’s brokers are: 1) posting live quotes and soliciting RFQs electronically, 2) aggregating and consolidating quotes by price/yield and size and 3) electronic execution of a trade against posted quotes. Electronic trading may facilitate the management of dealer inventory and reduce counterparty search costs. ATS platforms also offer anonymity to participants that post quotes. As a result, market participants such as dealers, proprietary trading firms and institutional investors may prefer using an ATS to find counterparties for trading without broadcasting their trading position to the market.

For a detailed description of the municipal securities market structure, electronic trading venues and broker’s broker platforms, please refer to the 2018 MSRB Report.  

Pre-Trade Information 

The MSRB currently publishes certain pre-trade pricing-related information to the public, such as yield curves, municipal market indices and new issue pricing scales on its free Electronic Municipal Market Access (EMMA?) website, in addition to collecting and disseminating post-trade municipal securities data since 1995 through RTRS and its predecessor, the Transaction Reporting System (TRS). The MSRB, however, neither collects nor disseminates pre-trade information, such as quote data about the price and size of quotes for municipal securities signaling trading interests before a trade is executed. In addition, there is currently no central facility in the municipal market through which such pricing information is made broadly available to the public in a comprehensive manner (i.e., no national best bid and offer indicators as in the equity securities market). To the extent that pre-trade pricing information is available, it typically is provided by proprietary electronic networks but only to data subscribers, such as those operated by ATS platforms and some broker’s brokers, and occasionally through non-electronic venues. 

As discussed in the 2018 MSRB Report, most electronic platforms do not share pre-trade information (bids, offers, requests for quotes of a security or responses to a request) with the broader market, and this information is available only to ATS participants engaging directly with such venues or other proprietary data subscribers, who are predominantly financial professionals. In fact, not only is access to pre-trade pricing information limited to ATS data subscribers, information may be further restricted to a few market participants involved in some of those potential transactions, such as during the RFQ process. The level of live quote information disseminated could also be limited depending on each market participant’s willingness to share the information on some or all the bids and offers entered for a potential transaction.

The availability of this (pre-trade) information could improve pricing efficiency, investor confidence and market liquidity in the municipal market.

Potentially, this pre-trade information could provide investors, researchers, securities regulators and all other market participants with important bond pricing information currently only accessible to select market participants. Thus, the availability of this information could improve pricing efficiency, investor confidence and market liquidity in the municipal market. Since January 2012, there have been several regulatory developments on pre-trade transparency in the municipal securities market, both at the U.S. Securities and Exchange Commission (SEC) and at the MSRB. However, as of early 2020, there was no formal recommendation or rule proposal put forward by any regulatory agency that would make such pre-trade information available to the public at no cost.

For a detailed description of the recent regulatory developments in pre-trade disclosure and the academic research performed in the area of pre-trade information transparency, please refer to the 2018 MSRB Report.

Pre-Trade Analysis Data and Methodology 

As previously mentioned, to provide an apples-to-apples comparison, all Phase II Period results presented are limited to the same two ATS platforms captured in the 2018 MSRB Report for the Phase I Period, which are the two largest ATS platforms for municipal securities trading. Both ATS platforms voluntarily provided the MSRB with pre-trade and post-trade data, including RFQ (bids and offers wanted), response to RFQ, live quote and associated transaction data for a six-month Phase II Period. The MSRB performed an indepth analysis of ATS quote and trade data from the RTRS, comparing the results from the Phase II Period with the results from the Phase I Period. For the purposes of this research report, only secondary market trades in municipal securities are included in the analysis.

The RFQ data includes quantity and price information for each RFQ, RFQ responses and associated trades, if any, with nearly 1.2 million total requests and 6.4 million total responses during the six-month Phase II Period. The live quote data contain bidding and offering amount, bidding and offering price, and bidding and offering yield information, with nearly 240 million quote updates from the two platforms in the Phase II Period.

It is important to note that pre-trade quote data could also be available from other ATS platforms, venues designed for institutional investors, broker’s brokers, dealers and third party vendors. Dealers may have multiple offerings for an individual bond depending on where the quote is shown (i.e., an offer quote on an ATS versus an offer quote to a client). In particular, institutional investors may prefer other electronic platforms that tailor toward large block-size traders. The MSRB requested ATS data from specific ATS platforms in both 2015 and 2018 because of the significant amount of “retail-sized” trades ($100,00 par value or less) on those platforms, their prominent market shares during both timeframes, and their ability and willingness to voluntarily deliver a large amount of data quickly and efficiently.

Findings of Pre-Trade Data Analysis 

With the advantage of having the prior findings from the 2018 MSRB Report, the empirical analysis in this report focuses on the changes in pre-trade market activities over the course of the Phase I Period and Phase II Period, a three-and-half-year time span.

Market Share of Inter-Dealer and ATS Trades 

Before analyzing the ATS data, this section first compares the market share of different types of ATS trades in the Phase II Period to those in the Phase I Period. Table 1 presents the market share by number of trades and par value for customer purchase, customer sell and inter-dealer trades. The percentage breakdown by trade count did not change substantially from the Phase I Period to the Phase II Period. However, the breakdown by par value traded did change between the two periods, as the customer buy par value (and to a lesser extent, customer sell par value) gained at the expense of inter-dealer par value. This could be explained partially by an increase in published trades of commercial paper as a result of the reengineering of the MSRB’s RTRS starting on May 29, 2018. The reengineered RTRS allows a small percentage of previously non-published trades to be publicly disseminated, which are predominantly large-sized commercial paper trades. Since a vast majority of commercial paper trades are extremely large customer purchases by institutional investors, there has been an upward shift in the market share of customer buy par value after May 2018 as a result of the RTRS upgrade. 

Since many investors, especially retail investors, do not have access to ATS platforms, a vast majority of reported trades executed on ATS platforms are trades between dealers. Chart 1 examines the percentage of inter-dealer trades executed via an ATS and finds, overall, between 56% and 61% of inter-dealer trades were executed on all ATS platforms for every month from August 2016 through December 2018, including the Phase I and Phase II periods. On the other hand, the percentage of executions by par value fluctuated between 25% and 34%, suggesting that the average inter-dealer trade size on ATS platforms was smaller than the average inter-dealer trade size executed elsewhere. Chart 1 confirms that ATS participation in the overall inter-dealer market remained significant and steady through the end of 2018.

Volume of ATS Quote Data 

Since 2015, the volume of quote data on the two ATS platforms rose substantially. Table 2 compares the size of the pre-trade data between the Phase I and Phase II periods. While the monthly average number of trades on ATS 1 and ATS 2 in the Phase II Period was noticeably lower than the number of trades in the Phase I Period, with a 28% decline, the quote volume had risen substantially. The monthly average number of RFQs went up to 199,000 from 174,000, a 14% increase, and the growth rate for responses to RFQs was even higher, at 56%. Similarly, the number of live quotes also rose drastically, from 27.7 million live quotes per month in the Phase I Period to 39.3 million live quotes in the Phase II Period, a 42% rise.

Overall, the quote data size dwarfs the size of the reported trades through the RTRS, as the live quote data from a single ATS platform (ATS 1) during a six-month period in 2018 had more data points (nearly 162 million live quotes) than the number of reported trades in the MSRB RTRS database for its entire 15-year period in existence from January 2005 through December 2019 (144 million reported trades). In addition, there were a total of 4.9 million secondary market trades during the comparable Phase II Period, significantly less than the amount of live quote volume on ATS 1 and ATS 2.

Request-for-Quote Data 

As indicated in the 2018 MSRB Report, there is an information imbalance for quote data in the municipal securities market, as a vast majority of RFQs were solicited for bids while most live quotes were offer quotes. The imbalances still existed during the Phase II Period, where most of the RFQ data from the two ATSs were requests for bids, with only 0.3% of the data representing offers wanted (see Table 3), the same percentage as in the Phase I Period. This is not surprising as investors, especially retail investors tend not to liquidate their acquired municipal securities positions before maturity, resulting in significantly less live bid quotes posted by market participants on the platforms in comparison to live offer quotes, and therefore more RFQs soliciting bids whenever an investors does want to sell a position. On average, there were 9,400 RFQs per day across the two platforms during the Phase II Period, compared to 8,400 RFQs per day in the Phase I Period, a 12% increase.

By comparison, there was an average of 5.6 responses per each RFQ across the two platforms in the Phase II Period, with a median of five responses. The average and median number of responses were considerably higher than those during the Phase I Period, where the average was 3.9 responses and the median was three responses. The increase in the response rate is in line with the information gathered anecdotally from electronic platform operators, suggesting that more market participants are increasingly involved in the RFQ process. In addition, 5.7% of RFQs on the two platforms received no responses at all in the Phase II Period, a noticeable decline from 10.9% during the Phase I Period. The 2018 MSRB Report highlights that RFQ response information is not available to all market participants, unlike RFQ and live quote data, which are generally available to most of the subscribers on a platform. In some instances, an RFQ-requesting dealer or a dealer who provides a live quote may not allow all market participants on a platform to see the request or the live quote.

Table 3 also shows that, overall, there were about 255,000 trades on the ATS platforms associated with the 1,195,000 RFQs during the Phase II Period, which represents an aggregate trade-to-request ratio (number of trades divided by number of RFQs) of 21.4% on each requesting ATS platform. Therefore, the remaining 78.6% of all RFQs, or 940,000 RFQs, with an average of nearly 7,400 RFQs per trading day, did not result in a trade on the requesting platform. In the Phase I Period, the trade-to-request ratio was 24.9%, slightly higher than in the Phase II Period. This indicates that, despite the higher number of responses received for each RFQ, the execution rate on the platforms decreased in the Phase II Period as a higher number of RFQs and a much higher number of responses corresponded to a nearly equal number of trades (see the daily average numbers in Table 3). The decrease in trade-to-request ratio could be explained by some dealers choosing to disseminate RFQs for the same bond using multiple platforms simultaneously after the implementation of MSRB’s best execution rule in March 2016. As a result, the execution rate on a particular ATS platform could decline even though the real execution rate for a bond being solicited may not have changed. In addition, it should be noted that an RFQ could still result in a trade but not necessarily through the requesting ATS platform(s), as a portion of the “non-executed” RFQs could also be internalized by the RFQ-requesting dealer or be executed elsewhere.

The vast majority of these trades resulting from an RFQ process, or 91%, were still retailsized—100 bonds or fewer—during the Phase II Period, similar to the 91.2% in the Phase I Period.

Table 4 shows that for the vast majority of RFQs receiving at least one response on the two platforms during the Phase II Period, the trade-to-request ratio was 22.7%. Further, as an RFQ received more responses, it became more likely to result in a trade, with the trade-torequest ratio going up uniformly along with the number of responses received. The trade-torequest ratio ranged from 7.5% when an RFQ received only one response, to 56.7% when an RFQ received 20 responses and to 73.2% when an RFQ received more than 20 responses. The correlation between the trade-to-request ratio and the number of responses received is similar to what the 2018 MSRB Report found, though the 2018 MSRB Report included the results for only one of the ATS platforms.

Table 5 shows that of all the trades during the Phase II Period (on trades that had at least one response to a bid-wanted RFQ in the same CUSIP number on the same trading day), nearly 40% of those trades were customer sell trades, more than twice as many as customer buy trades. This ratio was the reverse of a normal day during the same period, where there were nearly twice as many customer buy trades as customer sell trades (see Table 1 above), suggesting that, as expected, there is a strong connection between bid-wanted RFQ activities and customer selling activities. The results in the Phase II Period were similar to the prior Phase I Period.

For those customer sell trades with the same-day responses to RFQs, Table 6 compares the highest bid response with the customer sell trade price and shows that the median difference (highest bid price – customer sell trade price) was zero in the Phase II Period, a 12-basis-point decline from the median in the Phase I Period. In addition, the dispersion of customer sell prices relative to the highest bid response noticeably tightened between the Phase I Period and the Phase II Period across nearly all percentile ranges. This implies that bid responses may provide a more useful pricing indicator for all customer sell trades as the number of bids received increased.

While the decline in the median difference between the highest bid response and customer sell price was in line with the declining customer transaction spread found in other recent MSRB studies, the fact that the median difference was zero would seem to be surprising though it is not entirely impossible and is likely explained by the following factors. First, some of the customer sell trades may be initiated by separately managed accounts (SMAs) affiliated with a dealer firm that typically charges a fee for managing the accounts based on the size of assets under management rather than a per-transaction markup. Second, it should be noted that while the market-wide customer sell trades were executed on the same day for the same CUSIP numbers as the responses to the RFQs, these trades might not be directly tied to the RFQ process. As mentioned previously, live quotes are visible to nearly all market participants who have access to ATS quotes, while responses to RFQs are only visible to requesting dealers but not to other market participants. Third, it is possible that the RFQ requesting dealers might have solicited additional responses from other ATS or broker’s broker platforms, and the actual best bid response received could have been higher than the best bid response from the two ATS platforms in this analysis. If this is true, the median difference between the actual best bid response and customer sell trade price (markdown) would have been higher than zero. Fourth, it is also possible that the requesting dealer might have considered that the best bid response price was too low based on other bond valuation criteria. In that case, the dealer could have internalized the order by buying the bond at a price that is higher than the best bid received by some amount, which may have resulted in the customer sell price being at or near the best bid response price. Lastly, when comparing the amount of markdown to the amount of markup, past research has indicated that the markdown amount for customer sell trades tends to be significantly lower than the markup amount for customer buy trades in municipal securities.

Live Quote Data 

As mentioned in the 2018 MSRB Report, the fragmented nature of the municipal securities market and the difficulty in shorting tax-exempt municipal securities are unique characteristics that present significant market challenges, such as discouraging dealer quotations in most municipal securities, as there is less economic incentive to provide quotes in securities that are infrequently traded. Dealers may believe that dealer capital is therefore better concentrated in a few highly traded municipal securities to maximize the marketmaking opportunity.

Unlike the RFQ data, Table 7 shows that more than 89.7% of all live quotes submitted were offer quotes and only 5.3% of all live quotes were bid quotes in the Phase II Period, with the rest of live quotes (5%, all on ATS 1) containing both bids and offers. By comparison, 97.7% of all live quotes were offer quotes and only 2.1% of live quotes were bid quotes in the Phase I Period, with a negligible 0.1% of live quotes being both bid and offer quotes. The continued imbalance in live quote submission for municipal securities, albeit less imbalanced in the Phase II Period than in the Phase I Period, likely is the consequence of investors, especially retail investors, executing a “buy-and-hold” strategy for municipal securities. Dealers may see less of a necessity to post a live bid quote when investors generally do not sell off their accumulated positions. This is further exacerbated by the sheer number of municipal CUSIP numbers outstanding and how unlikely an aged bond is to trade on any given day.

The number of CUSIP numbers quoted at the 10 a.m. snapshot went up from 47,300 on an average day during the Phase I Period to 77,700 during the Phase II Period, a 64.4% increase. Regardless, at any given moment, less than 8% of municipal securities had live quotes during the Phase II period out of approximately one million municipal securities outstanding.

Even for the municipal securities with live quotes, most of these securities only had one or two dealers offering at any given time. Similar to the 2018 MSRB Report, a snapshot was taken at 10 a.m. every trading day during the Phase II Period for both ATS platforms to derive the number of offering quotes at a given time. Table 8 shows:

• 78.5% of CUSIP numbers with live quotes on ATS 1 had only one dealer offering a quote;
• 94.2% of CUSIP numbers on ATS 1 had two or fewer dealers offering a quote;
• Likewise, 82.6% of CUSIP numbers on ATS 2 had only one dealer offering a quote; and
• 96% of CUSIP numbers on ATS 2 had two or fewer dealers offering a quote.

Compared to the percentages of CUSIP numbers with one or two dealers offering quotes on each ATS in the Phase I Period, the percentages in the Phase II Period did not change dramatically for ATS 1, with only a slight decrease across the board. However, for ATS 2, the percentage of CUSIP numbers with a single dealer quoting declined to 82.6% from 94.4%, with a near corresponding rise in the percentage of CUSIP numbers with two dealers quoting, to 13.4% from 5%. That said, having around 95% of all CUSIP numbers with only one or two dealer quotes on each ATS platform as recently as in late 2018 is unique to the municipal securities market, especially when compared with the active securities markets such as the equity market. As elaborated above, because of the vast number of securities in the municipal market and the relative high cost of shorting tax-exempt municipal securities, it is highly unlikely that dealers would offer live quotes for a majority of individual bonds. Dealers prefer offering bonds they own or have immediate access to buying.

Table 9 shows the median live quote size on the two ATSs was 20 bonds or $20,000 par value in the Phase II Period, which was exactly the same as the median live quote size in the Phase I Period. When isolating to live offer quotes at the time of a trade, the median offer size was $50,000 in the Phase II Period, compared to $35,000 in the Phase I Period, a noticeable increase. In comparison, the median trade size on the platforms was 25 bonds or $25,000 par value, while the average trade size was 52 bonds or $52,400 par value in the Phase II Period. The average and median trade size in the Phase II Period was comparable to the average and median trade size in the Phase I Period, though slightly more trades were considered retail-sized in the Phase II Period than in the Phase I Period (92% vs. 88.7%). In addition, the average trade size on the ATS platforms was much smaller than the average trade size for all municipal securities, which was about $274,000 par value during the Phase II Period, though it was in the range of a typical municipal bond retail-size trade, which is 100 bonds or less (or $100,000 par value or less).

Finally, to examine whether live offer quotes provide any indicative value to trade price regardless of whether a trade was executed on an ATS platform, market-wide trades were matched with live offer quotes at the time of a trade. Essentially, this analysis attempted to simulate what a dealer observed on the two ATS platforms at the time of a trade. Two methods of live offer quote snapshot analysis were used for the Phase II data: 

•  The first method (Method 1) allows live offer quotes on both ATS platforms to be carried over from previous trading days unless an offer quote was explicitly canceled; and
• The second method (Method 2) allows only ATS 1’s live offer quotes to be carried over from previous trading days unless a quote was explicitly canceled, while for ATS 2 only the same-day live offer quotes are incorporated and quotes submitted on prior days are assumed to be canceled or expired.

Previously, only Method 1 was used for the analysis of the Phase I Period. Therefore, Method 1 allows a direct comparison between the two periods, while Method 2 attempts to simulate a more accurate picture of the Phase II Period based on the differences in operations between the two ATS platforms. Depending on the method used, between 73% and 81% of all secondary market trades during the Phase II Period had a live offer quote on at least one of the two platforms at the time of execution, even though many of these trades were not executed on ATS platforms. By comparison, using Method 1 only, nearly 70% of all secondary market trades during the Phase I Period had a live offer quote at the time of execution. The increase in the percentage between the two periods was consistent with the substantial rise in the amount of live quotes.

For trades with at least one live offer quote at the time of execution, the median price difference between inter-dealer trades and best offer quotes (lowest offered price) residing on the two ATSs at the time of a trade was zero during both the Phase I and Phase II periods, regardless of the method used for filtering live offer quotes, as shown in Table 10. In addition, about half of all inter-dealer trades were executed within 25 basis points (0.25%, or $2.50 per bond) of a best offer quote in the Phase II Period when using Method 1. Furthermore, trade prices were also more clustered around the best offer quotes in the Phase II Period than in the Phase I Period, implying a further decline in the dispersion of trade prices for inter-dealer trades. For example, the spread between the 30th and 70th percentile range was 29 basis points in the Phase II Period compared to 44 basis points in the Phase I Period when using Method 1. When using Method 2 for the Phase II Period, which attempted to filter out potentially canceled or expired live offer quotes, the spread between the 30th and 70th percentile range was only 20 basis points. The shrinking price dispersion also exhibited in other percentile ranges, e.g., between the 20th and 80th percentile range and between the 10th and 90th percentiles. The fact that inter-dealer trades prices became more clustered around the best offer quote over the three-and-half year period suggests that live offer quotes may have become more visible to market participants and also more informative to market participants executing all inter-dealer trades, even if the trades themselves might not have been executed on the ATS platform offering the best quote as a result of increased quote provision and offer price competition. That live offer quotes provide a useful indicator for all inter-dealer trades can be further supported by the fact that over 90% of all inter-dealer trades were executed at no more than 55 basis points higher than the best offer quote (Method 2), including the 40% to 45% of inter-dealer trades that were not executed on an ATS platform.

Table 11 illustrates the difference in trade prices and best offer quotes ranked in percentiles for customer trades in both the Phase I and Phase II periods. For those trades with at least one live offer quote at the time of execution, when using Method 1 for the Phase II Period, the median price difference between the customer buy trade and the best offer quote was 48 basis points, while the median price difference between the best offer quote and the customer sell trade was 54 basis points. Compared to the Phase I Period, where the median difference in price between the customer buy trade and the best offer quote was 75 basis points and the median difference in price between the best offer quote and the customer sell was 73 basis points, the median spread between customer buy and sell price in the Phase II Period narrowed significantly to 102 basis points from 148 basis points. This finding is in line with MSRB’s recent research showing a steady decline of the spread between customer buy and customer sell trades in recent years.

When using Method 1, there seemed to be a near symmetry in price differentials between the best offer quotes and customer buy and sell trade prices at the median in both periods, though the customer sell price was slightly further away from the best offer quote than the customer buy price (54 basis points below vs 48 basis points above) in the Phase II Period. On the other hand, when using Method 2, which attempted to exclude live offer quotes submitted from prior days that were possibly canceled or expired, the near symmetry disappears at the median, with the customer buy price at 38 basis points above the best offer quote and the customer sell price further away at 68 basis points below the best offer quote. This is supported by the findings in Table 5 and Table 6 in the previous section where customer sell trades were found to be related to the bid responses from an RFQ process.

Similar to the results from the RFQ analysis and the inter-dealer analysis, Table 12 shows that the dispersion in price difference between customer trade prices and best offer quotes narrowed between the two periods, regardless of whether Method 1 or Method 2 was used. For example, in the range between the 10th and 90th percentiles, Table 12 illustrates that the dispersion of price differences declined from 282 basis points for customer buy trades and 346 basis points for customer sell trades in the Phase I Period to 252 basis points and 278 basis points, respectively, in the Phase II Period using Method 1, and to 246 basis points and 280 basis points, respectively, using Method 2. Other percentile ranges also exhibited an analogous pattern of decline in the dispersion of price differences between customer trades and best offer quotes. This result is consistent with the proliferation of live quotes between the two periods, which led to increased quote provision and offer price competition. It implies that live offer quotes may have become more visible to market participants, and also more informative to market participants executing customer buy and sell trades, even if the trades themselves might not have been executed on the ATS platform offering the best quote. Despite the unique feature of the municipal securities market where most quoted municipal securities only had a one-sided offer quote that was live and visible in both the Phase I Period and the Phase II Period, live quotes seem to have strengthened their indicative pricing value to the market.

Potential Future Research 

There has been a noticeable increase in the percentage of live quotes that were either bids only or bids and offers, from 2.3% in the Phase I Period to 11.3% in the Phase II Period. While the number of live bids still are greatly outnumbered by the number of live offers, it would be prudent to compare the market-wide trade price with the aggregate best bid responses to RFQs and best live bid quotes at the time of a trade in the future, assuming the market share of live bids continues to grow.

It is possible a few additional ATS platforms and/or other trading venues may continue to grow their market share and become more prominent in the municipal securities market in upcoming years.

Furthermore, it is possible a few additional ATS platforms and/or other trading venues may continue to grow their market share and become more prominent in the municipal securities market in upcoming years. Given the ever-changing market landscape and a continuing interest in pre-trade-related market structure issues, it may be prudent to solicit newer data from major market participants and monitor the trends in the pre-trade market in the future.

Finally, since the two ATS platforms are predominantly used for retail-sized trades, the findings in this report may not represent the trading and quoting behavior on other venues with mainly institutional-sized pre-trade information. It would be beneficial to examine data from an electronic platform primarily servicing institutional investors for comparison.

Conclusions

In summary, the analysis of the Phase II ATS quote data indicates a significant increase in the amount of responses to RFQs and live quotes during the three-and-a-half years between the Phase I Period and the Phase II Period. For RFQs, the preliminary analysis confirms the results from the prior analysis that the execution rate on an ATS platform was higher when more responses were received. Similar to responses to the RFQs, live offer quotes may have increasingly provided a valuable pricing indicator to the market, whether a bond was traded on an ATS or elsewhere. Even though a majority of live quotes represented only one (offer) side of the market and less than 22% of all trades were executed on an ATS platform, quoted offer prices may have become more visible to market participants, and also more informative to execution prices for inter-dealer, customer buy and customer sell trades in the market, as a result of increased quote provision and offer price competition by market participants.

Preliminary analysis confirms the results from the prior analysis that the execution rate on an ATS platform was higher when more responses were received.

We caution that the analyses in this report reflect market dynamics where only subscribers to an ATS platform could access pre-trade information. If some or all of the pre-trade information had been available to a wider market, the quoting and trading patterns observed in this report may not hold due to possible behavioral adjustments by market participants and changes in market structure and liquidity in reaction to broader quote dissemination. In addition, while the pre-trade quote data from the two ATS platforms represent a large portion of the market, it should be noted that there exists additional pre-trade information available on other venues, which is not captured in this report. This is particularly true for pre-trade information tailored toward institutional investors on other electronic platforms, as well as additional liquidity provision that may not be visible on any electronic platform. Finally, the Phase II Period in this analysis preceded the spring of 2020, a period of extreme market volatility as a result of the COVID-19 pandemic; the results captured in this report therefore may not represent what would be observed during a period of market stress.

Full References and Appendix can be viewed here.

The information and data in this document are provided without representations or warranties and on an “as is” basis. The MSRB hereby disclaims all representations and warranties (express or implied), including, but not limited to, warranties of merchantability, non-infringement and fitness for a particular purpose. Neither the MSRB, nor any data supplier, shall in any way be liable to any recipient or user of the information and/or data, regardless of the cause or duration, including, but not limited to, any inaccuracies, errors, omissions or other defects in the information and/or data or for any damages resulting therefrom. The MSRB has no obligation to update, modify or amend information or data herein or to notify the reader if any is inaccurate or incomplete. This document was prepared for general informational purposes only, and it is not intended to provide, and does not constitute,investment, tax, business, legal or other advice.