Streetwise Professor

There is an old expression in the military: “Amateurs talk tactics. Professionals talk logistics.” There is a clear parallel in the financial markets; although the logistics of derivatives markets–clearing–attract little popular attention, they are matters of prime importance to market professionals.

The CBT-CME merger is bringing clearing into the public eye, however. This previously arcane issue has taken center stage in the debate over the the desirability of the merger. Former SEC chief economist (and academic) Larry Harris has opined that the CME’s ownership and control of clearing impedes competition in derivatives markets. He is seconded by PHLX president Meyer Frucher in his jeremiad against the merger. Moreover, (per a report in Crain’s Chicago Business) the Department of Justice is allegedly focusing its attention on the CME’s clearing operations.

The battle cry of Harris, Frucher, et al is “Fungibility!” (Not quite “Remember the Alamo!”) In essence, they argue that the clearinghouse is an essential facility, and by restricting access to the essential facility CME impedes competition for its futures contracts. If CME were to open access to its clearinghouse, it would allow other exchanges to offer contracts that were perfect substitutes for CME contracts, that is, were fungible with CME contracts. This would allow competition in futures trading, and in Harris’s phrase, “break the futures monopoly.”

There are several problems with this analysis. First, it implicitly presumes that the market for futures execution would be highly competitive, but for restricted access to the clearinghouse. However, due to the network effects of liquidity, this assertion is problematic. (I return to this issue later.)

Second, this argument is very retro. BC–Before Chicago (School)–in fact. It is a re-hash of the monopoly leveraging argument demolished by Chicago economists and legal scholars in the 1960s and 1970s. If the monopoly bottleneck is at the clearing level, the clearing operator would actually like to encourage competition in execution. This would increase the demand for clearing services by reducing the costs of execution services that are complementary to clearing. This would allow the clearing monopoly to charge a higher price for its services. Put differently, if clearing is a natural monopoly (as the essential facility/bottleneck arguments implicitly assume–and for which there is a reasonable basis), the operator of the clearing facility can extract the entire monopoly rent by pricing clearing services appropriately, and has no incentive to force a putatively inefficient organization of complementary activities, such as execution. If there are no scale economies in execution, or if there are scale diseconomies, integrating into execution actually reduces the clearing monopolist’s rents.

As the Chicago School scholars noted long ago, this reasoning leads to a presumption that there is an efficiency rationale for the vertical integration of clearing and execution. I set out the sources of efficiency in my Silos post from the summer. In a nutshell, integration eliminates double marginalization problems, and reduces transactions costs.

So what can be done? Merely requiring the CME to allow other exchanges to access its clearinghouse would not change things even if the merger’s critics are correct. The exchange can merely charge the monopoly price for clearing services. Thus, it would be necessary to regulate clearing prices and terms of access very intrusively in order to reduce the exchange’s market power. For anyone who doubts who messy this can be, just take a look at the experience of the regulation of access pricing and terms in telecommunications. This is an invitation to rent seeking and litigation (and rent seeking litigation), and will likely discourage innovation by the clearing operator. (I doubt that the DoJ has the power to impose such price and access regulations in any event.)

Similarly, attempts to increase competition in clearing by forcing the CME to eliminate its rule requiring contracts traded on the exchange to be cleared through the CME clearinghouse would be unlikely to have much of an effect. Due to the strong scale and scope effects in clearing, this function is (as argued by the fu(ngibility) fighters) a natural monopoly. Moreover, due to the specific assets and sunk costs associated with clearing, and the costs of coordinating a simultaneous defection of sufficient numbers of clearing customers to generate network economies, it is not likely to be a contestable one. Thus, the incumbent clearer would likely prevail over any entrant. And in the unlikely event that the entrant did prevail, users would be flung from the frying pan into the fire–from one monopoly to another. Moreover, this would entail additional costs–double marginalization and higher transactions costs.

So what about a more radical alternative? For instance, forcing the CME to divest its clearinghouse as a condition of the merger? In order to improve welfare, somehow it would be necessary to organize the (monopoly) independent clearing entity so that it would not exercise market power. One possibility is to organize the clearer as a user-owned cooperative that rebates revenues in excess of costs to its users. (The OCC and NSCC work this way). In theory, the cooperative could price its services at marginal cost, and fund its fixed costs through fixed assessments on its members. Alternatively, non-linear price schedules with inframarginal prices in excess of marginal costs and marginal cost pricing for the marginal unit would be efficient and cover fixed costs. Either alternative would lead to competitive pricing of clearing services.

Nirvana? Not so fast, chief. Two big problems.

First, the non-profit clearer could still exercise market power. How? By restricting access. My 1999 Journal of Financial Markets and 2002 Journal of Law, Economics, and Organization papers showed how non-profit cooperatives can exercise market power by restricting membership. In a nutshell, the big clearing brokers (e.g., the big banks and investment banks) could form a clearing utility, and impose unnecessarily burdensome financial requirements on membership. This would restrict entry into clearing services, and raise the price of these services above the efficient level. To the extent that the execution entity also has market power, this would result in double markups. Moreover, it would raise transactions costs (in the relationship between the clearing and execution firms.) Therefore, it would be necessary to pay close regulatory attention to the access and membership admission policies of this clearing entity. This is not a trivial task, by any means. Given the systemic implications of clearing for the stability of the financial system, financial standards for members are obviously desirable. Who is to evaluate the trade off between the effects of tighter/looser standards on financial stability vs. the pricing of clearing services? Any volunteers? Any confidence that the evaluation will be the right one? Not by me, that’s for sure.

Second, due to the network effects in trading I alluded to earlier (and which I have analyzed in great detail in my 2002 JLEO piece and several working papers) make it almost certain that even an exchange shorn of its clearing function can exercise market power. That is, the centripetal force of liquidity induces trading to concentrate on a single exchange. This confers market power on the incumbent exchange; it is very costly to coordinate a simultaneous defection of order flow sufficient to overcome the liquidity incumbent’s advantage.

This means that even if, through some miracle of economic organization, the clearing cooperative charged marginal cost prices and admitted the “right” number of members of the “right” kind, the exchange would still likely be a monopoly, and earn all the monopoly rent. That is, the clearing integration issue is a red herring. The CME is not a combination of a natural monopoly clearer with an execution facility that would be a near perfect competitor if shorn of its clearinghouse; it is a natural monopoly clearer merged with a natural monopoly execution facility. Indeed, as I argued in Silos, that is exactly WHY they are integrated. Integration addresses the problems posed by bilateral monopoly.

For evidence that an execution-only exchange with no clearing function can exercise market power, look at the London Stock Exchange. LSE has no clearing or settlement function. It gets clearing services through LCH.Clearnet. LSE has faced competition in the past–like Tradepoint–that failed miserably. It maintains a dominant market share in the trading of the stocks it lists. Despite all the flack that Deutsche Borse gets over its integrated clearing and trading, guess who has the bigger margins–LSE. It has the highest fees of any of the major world equity exchanges, and has higher margins (66 percent) than Euronext (60 percent) or DB (58 percent).

In sum, disintegrating clearing and trade execution is no silver bullet that will make futures markets perfectly competitive. As Posner, and Bork, and other Chicagoans who wrote on antitrust long ago noted, vertical integration should not be a focus of antitrust concerns. Vertical integration is almost always a way to economize on transactions costs or double markups. In my view, this is true in futures markets (and equity markets, for that matter). Vertical integration has long been the dominant form of relationship in futures execution and clearing, for exchanges large and small. This reflects transactional efficiencies, not monopoly leveraging.

If competition is imperfect in futures (and equities) markets, this imperfection arises from the strong scale, scope, and network economies in trading and clearing. At best, disintegrating clearing and execution will just allow the execution entity to capture the rents that arise from the imperfect competition. At worst, disintegration will increase transactions costs and the fees that participants pay (due to double markups). What market forces have joined together, let no man put asunder–without a compelling economic case for doing so. So far, the fungilistas have not done so. Indeed, in my view the compelling arguments are all on the other side.