Thoughts on the TMPG White Paper
The Treasury Market Practices Group (TMPG) recently released its final white paper on automated trading in the US Treasuries market. The report describes the current state of electronic trading in the Treasuries market and points out that electronic trading can result in ‘greater operational risk’ and the possibility of ‘disruptive market practices and trading strategies.’
Regulators including the New York Federal Reserve Bank (NYFRB), which sponsors the TMPG, have become increasingly concerned about how automated trading affects liquidity in the US Treasuries market as a result of the ‘flash crash’ on October 15. In a speech at the annual primary dealer meeting on April 13, the head of the NYFRB’s markets group, Simon Potter, called the event ‘a highly unusual round-trip in yields’ and noted that ‘October 15 raises questions about the nature of liquidity in the Treasury market today’.
Backing Away #
Since liquidity reflects the fair price of an asset given available information, the sharp price movements that occurred on October 15th are notable. No amount of liquidity can ‘cushion’ a market re-pricing caused by new information, however the lack an event which would have been expected to cause a change in valuation assumptions suggests that a reduction in liquidity may have precipitated the knee-jerk price movements of October 15th.
During the flash crash, many participants including Primary Dealers, turned off their quoting systems to avoid transacting. Dealers play a critical role in price discovery in the bond market because trades are executed bilaterally between customers and dealers rather than on a central exchange. When dealers withdraw from the market, customers have fewer outlets for managing risk. Aside from putting customers (i.e., institutions representing investor money) in a bad position, this scenario is also concerning because of what it implies about market conditions for dealers. Since dealers form their prices based on observable information in the market, turning off quoting or otherwise backing away from transacting reflects a lack of confidence that they have captured a sufficient amount of information to provide a quote.
Information Asymmetries #
Liquidity providers are unwilling to quote prices when they believe that liquidity takers are in possession of private information that could substantially influence the price of an asset. Although private information is rarely distributed equally among market participants, markets require a rough equality between participants to function. Without basic equality of information, it would be difficult for buyers and sellers to negotiate prices, because buyers would reduce their bids to account for the possibility of sellers being more informed. Sellers, in turn, would be more reluctant to transact at prices below fair value. Many regulations in the equities market, such as Reg FD (‘fair disclosure’), aim to minimize these sorts of information asymmetries between participants, not just to prevent incidents of investors getting ripped off, but also to support the integrity and liquidity of the overall marketplace.
In the Treasuries market, ‘fair disclosure’ is enforced primarily through embargoed press releases by the Federal Reserve and other bodies that produce market-relevant economic reports. Although there have been notable breaches of this system, illegal insider trading in the US Treasuries market is rare. The scope of fair disclosure rules in markets has never been presumed to be static. As markets have evolved, there has been a continual need to revise enforcement and disclosure rules to sustain fairness in the marketplace. Today, granular controls on the release of information are needed to protect investors due to the speed at which participants can assimilate and transact upon new information.
Economic releases such as the Bureau of Labor Statistics’ (BLS) monthly employment report are macro events, where standards of fair disclosure are well-established and clear cut. However, in some fixed income markets, micro events, such as individual trades, are considered reportable to a public record. For example, in the corporate bond market, TRACE is widely credited with giving investors better access to pricing information for less liquid securities. By making it possible for investors who don’t see trade flows to trade with greater confidence against dealers who do, the overall size of the corporate bond market as well as the variety of issuers has increased dramatically since the inception of TRACE in 2002.
How Much Equality of Information? #
A key question of market structure is how much disclosure is appropriate to facilitate liquidity in a marketplace. Too much disclosure potentially drains the amount of capital available for intermediation, since liquidity providers need to be able to earn a return on providing liquidity. Conversely, too little disclosure can also reduce liquidity, preventing buyers from finding sellers or excluding certain participants from the market, resulting in overall economic loss.
Another variable is trading technology. The structure of markets is deeply rooted in the social context of how buyers and sellers negotiate. Electronic marketplaces not only formalize the norms of negotiation, but they also reduce the cost of transmission of information to zero. As a result, competition between human traders becomes more abstract, as data must be collected and analyzed, and execution algorithms must be used by liquidity providers to monetize smaller and smaller informational advantages. In highly electronic marketplaces, fleeting informational advantages constitute a last remaining source of pure alpha available to liquidity providers.
Conclusions #
Competition for granular trading opportunities creates operational risks for liquidity providers, who must participate in the market to capture these opportunities, but must also calibrate their participation to a desired level of risk. The TMPG notes:
Automated trading can occur at speeds that exceed the capacity of manual detection and intervention and therefore pose a challenge to traditional risk management protocols.
In the Treasuries market, although turnover (volume / issuance outstanding) is down nearly 70% in the past 10 years, the relative liquidity of the Treasury market is still supportive of automated trading. In the interdealer market, the share of trading executed by electronic market makers (non-Primary Dealers) is at least 50%. The prevalence of electronic market making is notable because electronic market makers compete against Primary Dealers using incomplete information. In the dealer-to-customer market, Primary Dealers not only see customer trades, but they also know who their trading counterparties are. For electronic market makers (which don’t have customers), the $300bn/day dealer-to-customer market is a dark pool, which means that they have to account for an extra degree of uncertainty when large flows are trading away from their field of vision.
Liquidity providers must make markets under conditions of uncertainty, but events such as October 15th highlight the need for closer scrutiny of how the market structure influences competition in the US Treasuries market. Since liquidity, particularly in Treasuries, relies on participation by diverse institutions around the world, some of the trends highlighted by the TMPG are concerning:
Narrow bid-ask spreads have also continued to drive the electronification of Treasury trading, affecting the DtC business model for many dealers, with some choosing to reduce their capital-intensive market-making activities, and incentivizing others to reduce costs via increased volumes and internalization of flows to achieve target returns on equity. These trends support the high concentration of trading activity among large market makers, with the top five dealers now accounting for more than 55 percent of DtC volume
At Direct Match, we believe that one of the lessons of the financial crisis is that markets are more exposed to shock when they rely on a few systemically important liquidity providers. Electronic trading has substantially lowered the search and information costs of transacting in Treasuries, allowing firms to provide more liquidity at a lower cost. This change in the market structure has occurred amid the backdrop of extended low rates and relative certainty about the forward path of interest rates. As monetary policy normalizes, asynchronous two-way investor flows have the potential to overwhelm the short-term supply of liquidity. We believe that the solution to this problem, in addition to providing more efficient ways of transacting, is to minimize informational inequalities between market participants by requiring disclosure of trades, thereby fostering a broader ecosystem of liquidity providers and improving confidence in prices.