Markus Baldauf, Joshua Mollner
Based on: Journal of Political Economy, Vol. 132, No. 5 (May 2024), pp. 1603–1641. DOI: https://doi.org/10.1086/727709
Most financial trading occurs over the counter (OTC)—that is, away from centralized exchanges like the New York Stock Exchange. Trading this way requires searching for a counterparty, usually a dealer who takes the other side of the trade. It is quite puzzling that when traders wish to buy or sell, they typically contact only a few dealers. For example, in swaps trading, the modal number of dealers contacted is only three, which is in fact the legally mandated minimum in that market.[1]
That OTC traders contact only a few potential counterparties is somewhat a conundrum—especially given the large sums involved—as the benefits of competition are typically portrayed as a bedrock of economics. After all, if you want to sell a house, the first rule is to create competition by attracting a large number of bidders. Or if you want to buy, say, home internet, you’re more likely to get a good deal if many firms are competing for your business. So why don’t OTC traders try to generate competition for their orders? Historically, OTC markets featured communication barriers: it was simply infeasible to cast a wide net by simultaneously contacting many dealers. One might think that these communication barriers fully explain this puzzle. However, in recent years it has become clear that a different explanation is needed. Indeed, technological advances—email, Bloomberg chat, and new electronic trading platforms—have all but eliminated communication barriers. Nowadays, with just the click of a button, any trader can announce to the entire world that she seeks to trade a large block of a given stock or bond, at a moment’s notice, 24/7. Yet traders do not tend to do this, even though they can. Why not? Industry insiders point to a different explanation for this puzzle: contacting each additional dealer increases information leakage, and with it, the risk of front-running.
To see how this works, consider a simple version of this problem. Suppose you are an institutional trader, perhaps a pension fund, and you wish to either buy or sell a security, perhaps an index credit default swap. To orchestrate your search for a dealer to trade with, you must answer two questions: How many dealers should you contact for a request for quote (RFQ)? And what should you tell them about your trade?
How many dealers to contact?
Contacting each additional dealer brings two intuitive benefits. It intensifies competition among the dealers for your order. It also improves the odds of locating a natural counterparty—a dealer who might, by virtue of an existing inventory position, be able to internalize your order (reducing or eliminating his need to hedge, together with associated hedging costs). Yet information leakage—in which your trading intentions become more widely known—is a countervailing force.
Information leakage can be costly because of what dealers do after your order. The dealer who wins your order will often seek to hedge it by conducting offsetting trades. Other dealers, whom you contacted but did not select as the winner, might anticipate this and front-run, i.e., trade ahead of, the winning dealer as he tries to hedge. For example, if they suspect that the winning dealer will be buying, then they might buy before turning around to sell after prices have gone up due to the winning dealer’s buying. This front-running exacerbates the winning dealer’s hedging costs. The cost is then ultimately borne by you, since dealers bake their anticipated hedging costs into what they quote for your order. You could, however, mitigate this front-running by contacting fewer dealers in the first place: fewer dealers would then be aware of your order, and dealers who are unaware cannot front-run the winning dealer’s hedging trades as effectively.
In situations where this risk of front-running looms large, the costs of information leakage can outweigh the benefits of competition, so that you would benefit from restricting the number of dealers that you contact. These considerations explain why OTC traders often request quotes from a puzzlingly small number of dealers—as in swaps trading, as mentioned above.
What to tell the dealers?
When asking for quotes, traders often have options regarding what information to reveal about their trading direction. At one extreme, you could disclose nothing about your trading direction by asking for a “two-sided quote,” that is, one quote for buying and another quote for selling. Of course, only one of those two sides will be relevant to you, but a dealer cannot initially know which—he learns your trading direction only if and when you accept. At the other extreme, you could fully disclose your trading direction by asking for a “one-sided quote.” You could also pursue intermediate strategies involving all sorts of partial disclosure.
It is in your best interest to disclose nothing. This rationalizes how, in many asset classes, asking for two-sided quotes has become common practice.[2] Again, the intuition is based on front-running: as you reveal less information, the other dealers cannot front-run the winning dealer’s hedging trades as effectively.
Implications for regulation and platform design
What lessons can be drawn from this analysis? Trading platforms should take these insights into account when designing their protocols. For example, many platforms currently require a trader to reveal both the size and side of her desired trade, effectively mandating an information policy of full disclosure. However, traders might prefer platforms that provide the flexibility to disclose less information. Similarly, many platforms require traders to reveal their identities when requesting quotes. Yet, for similar reasons, trading costs might be lower on average if quotes could be requested anonymously: without knowing the trader’s identity, other dealers would find it more difficult to guess if the trader seeks to buy or to sell, and hence more difficult to front-run the winning dealer. Although some trading platforms already allow flexible information policies and anonymous trading, many do not. Making these features more common might lead to healthier markets.
Regulators should also be aware of these insights when designing the rules that govern trading platforms. For example, swap contracts are increasingly traded on electronic request-for-quote platforms. The Commodity Futures Trading Commission, a major regulator overseeing these contracts, once proposed a rule to mandate that a minimum of five dealers be contacted before each trade. This proposal received considerable pushback from industry participants, who complained that the rule would exacerbate front-running and raise trading costs. Ultimately, the requirement was reduced to three. Even this weaker requirement is binding, since there is bunching at this threshold: as mentioned above, the modal RFQ contacts exactly three dealers. Weakening the requirement still further could be beneficial.
You might have thought that it would be a mistake for a trader to contact only a small number of potential counterparties—that the trader would have been better off had she created more competition for her order. And you might have thought that mandating a minimum number of contacts would save traders from errors like this. However, making only a small number of contacts is not necessarily a mistake: in OTC markets, the costs of information leakage and front-running can outweigh the benefits of competition.
[1] Riggs, L., E. Onur, D. Reiffen, and H. Zhu, “Swap Trading after Dodd-Frank: Evidence from Index CDS,” Journal of Financial Economics, 2020, 137 (3), 857–886.
[2] Risk.net, “Buy side using two-way prices in bid to hide trade intent,” August 2018, https://risk.net/5858751.
B.I. Ghert Family Foundation Associate Professor of Finance
University of British Columbia, Sauder School of Business
Associate Professor of Managerial Economics & Decision Sciences
Northwestern University, Kellogg School of Management