Streetwise Professor

Today’s post on SEFs led to a nice email exchange with Alex Yavorsky of Moody’s Investor Services.  It got me thinking about an aspect of transparency that has gone unremarked, to my knowledge.

In the mom-and-apple-pie treatment of transparency so beloved by Gary Gensler, Ken Griffin, the NYT, and others, improved transparency unambiguously increases competition.  With regards to OTC derivatives, the story is that lack of pre-trade transparency makes it necessary for customers to shop around for quotes.  Search costs can create market power: firms can charge higher prices because customers realize that even if they think the price is too high, they have to incur search costs to find a better deal, so they pay up in order to avoid the search costs.

There are models of monopolistic competition (e.g., some theory by Stiglitz from the ’80s, if memory serves) that generate this kind of result.  In these models, free entry generates zero profit but prices exceed the perfectly competitive price.  Transparency that reduces search costs reduces market power and moves prices closer to the competitive equilibrium.

But transparency doesn’t always work that way.  Indeed, greater price transparency can lead to less competitive outcomes.  In particular, transparency can facilitate collusion, and opacity can make it more difficult to collude.

The basic problem with collusion is the prisoners’ dilemma: each individual member of a cartel has an incentive to undercut the collusive price.   In order to deter this kind of conduct, colluders employ punishment strategies, cutting prices in response to a defection.  If a potential defector realizes that he cannot profit from his defection because others will respond immediately and slash prices, he can be deterred from cutting prices the first place.

But for these strategies to work, it must be necessary to detect defections.  Price transparency makes it possible to detect violations of the collusive agreement.  Forcing everybody to post prices publicly is a great way to support collusion: if everybody–including your supposed competitors–can see your price, they can see if you are defecting from a price setting agreement.  In contrast, if buyers and sellers can negotiate prices in private, it is much harder to detect defections.   Secret price cuts strike at the vitals of any collusive agreement.

The inane NYT article on OTC derivatives mentions, in a typically clueless way, the NASDAQ dealer collusion on spreads.  This actually illustrates the role that price transparency can play in facilitating collusion.  NASDAQ dealers posted their quotes on the NASDAQ system.  Any dealer who defected from the agreement to avoid odd-eighth quotes was immediately detected, and this spurred a variety of forms of retaliation–some revealed in rather entertaining fashion on recordings of traders’ phone lines.  Such collusion would have been much, much harder to sustain in the absence of an “automatic quotation” system that made price quotes transparent.

A fully transparent CLOB could–not necessarily would have, but could–result in the same outcome.  It would be a great way of coordinating pricing activities among OTC dealers.  In contrast, the supposedly scary dark OTC markets as they operate presently are very adverse to collusion.  Secret price cuts are quite easy in this environment.  (Moreover, since at present there are other terms of OTC deals, such as credit and collateral terms, that can be varied, it is possible to undercut a putative collusive agreement even while adhering to price terms by giving easy credit and collateral terms.  The increased standardization of deal terms as a result of the clearing mandate also would tend to make collusion easier.)

In the end, what made the NASDAQ market competitive was opening up access to the market.  Prior to the SEC’s order handling rules (put in place in 1997), dealers were under no obligation to submit customer limit orders to the NASDAQ system.  This limited the competition dealers faced in supplying liquidity.  The order handling rules required dealers to incorporate customer limit orders into their quote streams.  This led to the entry of new liquidity suppliers that dramatically narrowed spreads and reduced market maker profits.  It also encouraged the technological revolution–HFTs and algo trading–that has revolutionized liquidity supply.

What this suggests is that a transparent system without open entry is vulnerable to collusion, but that a transparent system with open entry and access can be extremely competitive.  Although we probably won’t find out given that the CFTC rule (as proposed) does not mandate a CLOB, it is interesting to conjecture how a swaps CLOB would work.  Would it be like the pre-order handling rule NASDAQ, or the post-rule NASDAQ?

That would depend on a couple of things.  One, it would depend on whether the swaps market is amenable to competitive liquidity supply from HFTs or algos the way that the equity market is, or the futures market is becoming.  Two, if it is, it would depend on whether the rules are conducive to entry of such liquidity suppliers and/or competition among platforms leads the CLOB operators to permit such access.

Differences  between the characteristics of swap transactions (size and frequency, most notably) makes me skeptical that equity-like or futures-like liquidity supply would be competitive in this market.  These are products for which big dealers may have a strong advantage in liquidity supply.

Insofar as the second issue is concerned, given the nature of CLOBs and the liquidity network effect, one CLOB would be likely to survive, or at least be extremely dominant.  The operator of this CLOB would have market power, which it can exercise by limiting entry (e.g., a CLOB operated by a group of dealers) or supercompetitive pricing.

This means that the CLOB rule would almost certainly have soon resulted in a battle royale over the terms and pricing of access to the CLOB.  This is seen again and again and again in network industries.  It has happened in the securities industry.  It has happened in the telephone industry.  It is happening as I write in the internet–today the FCC passed a net neutrality rule.

Suffice it to say that the implications of transparency and a CLOB would have been far, far more complex than the simplistic narrative advanced by Gensler et al.  It is by no means clear that mandated CLOBs and transparency would increase competition.  That result depends on a complex interaction between the products (which influences the economics of liquidity supply) and the rules governing the pricing of and access to CLOBs.

As an aside, it’s long been a pet peeve of mine that scholarly and policy debates on financial markets have largely ignored the extensive literature on competition and regulation in other network industries.  (I’ve just begun Spulber’s and Yoo’s book on network economics: it looks very good and its discussion of net neutrality echoes some of the arguments I’ve made regarding competition in OTC markets.)  My next book will be about financial market structure, and one of the themes of the book will be the close analogies between the policy issues in traditional network industries and in the financial markets.