A recent Wall Street Journal article (subscription required) discusses growing unease among brokerage firms over rising trading fees at NASDAQ and NYSE. European traders have expressed similar disquiet at rising fees on European bourses. The heady valuations of publicly traded exchanges, such as the CME and CBT, are predicated in part on the belief that they have pricing power that they will be able to raise fees in the future.
Is it plausible that exchanges have market power, and therefore can charge supercompetitive prices? In a nutshell: hell yes. There is little doubt that exchanges have considerable market power. Over the years I’ve written several articles on the subject setting out the arguments and evidence, but the essence of the story is rather straightforward. (This article sets out the argument in a non-mathematical way.) When some traders have better information than others, the relatively uninformed are better off when they concentrate their trading in a single marketplace. This creates a network effect. Traders like to trade where others trade. There are therefore natural monopoly elements to trading. (In the future I plan to blog about things that weaken the network effect and therefore facilitate competition between trading venues.) Moreover, it is hard to contest an incumbent monopolist’s position. (Cf., Eurex’s not-so-excellent US adventure, ditto EuroNextLIFFE.) Attracting liquidity to a new platform requires the simultaneous defection of numerous traders from the incumbent. This is a daunting coordination problem.
Hence, DTB/Eurex’s victory over LIFFE in Bund futures in 1998 is the exception rather than the rule. As I’ve shown in a working paper , moreover, that was something of a fluke for two reasons. First, DTB/Eurex had a less severe coordination problem because the German banks that were the source of large order flows also owned the exchange, so it was easier to build a base order flow to challenge LIFFE. Second, when DTB/Eurex had attracted enough volume to reduce the price impact disparity with LIFFE, it cut fees aggressively and LIFFE did not respond. If LIFFE had responded in kind rather than dilly-dallying for 3 crucial months, it might have hung on to a big chunk of the Bund business.
The CBT and CME learned from LIFFE’s errors and cut prices aggressively when faced with the prospect of entry; it’s always much more satisfying to learn from others’ screw ups than one’s own. Moreover, neither challenger could tap a block of affiliated order flow to build liquidity in contracts offered by the incumbent exchanges. As a result, the entrants weren’t able to repeat the Eurex coup on this side of the big pond.
When evaluating the prospectâ€”or realityâ€”of rising exchange fees, its important to remember that exchanges have always had market power, but they just exercised it in different ways. The traditional member-owned, not-for-profit exchanges did not exploit their market power by charging supercompetitive transaction fees. That would have cut into their members’ ability to make money by reducing the derived demand for brokerage and market making services. Instead, traditional exchanges exploited market power by limiting entry. With a few exceptions (primarily resulting from legislation and regulation rather than exchange choice) traditional member-owned exchanges limited the number of members. They also had rules that advantaged floor traders at the expense of others. These measures allowed the members to earn rents. As I’ve shown in a couple of articles, network effects make it possible for an exchange to choose a suboptimally small membership, raise trading costs, and yet face no credible competitive threat. Thus, exchange customers have historically paid supercompetitive prices for floor brokerage and floor market making services rather than supercompetitive exchange transaction fees.
Things are changing now that exchanges are moving away from the mutual non-profit (“MNP”) form. Investor-owned exchanges in particular have a new clientele to serve. Rather than catering to the interests of exchange members (typically floor traders or those who leased floor trading rights), some exchanges now respond to the interests of investors. For profit investor owned (“FPIO”) exchanges have no incentive to do things that inflate liquidity or brokerage costs. Indeed, they have the very opposite incentive. Higher profits to brokers and market makers reduce the derived demand for the services exchanges offer (namely, access to the trading engine.) As a result, they want to minimize barriers to entry by liquidity suppliers and suppliers of brokerage services. Exchanges will exploit the market power conferred by network effects by charging supercompetitive transaction fees.
High exchange fees are a more transparent way of exploiting market power. Market power wasn’t a secret to savvy folks in the old regime; I’ve talked to many upstairs traders, including some very successful ones, who routinely complained of being “robbed by the floor.” However, I imagine a lot of folks were unaware how limitations on entry and trading rules inflated trading costs. In-your-face increases in trading fees attract more attention, but the bottom line is that it’s just a different way of taking advantage of network effect-induced market power.
As an aside, I’ve been surprised for years at how little academic commentary and regulatory scrutiny has been directed at exchange membership limits. Economists since Friedman (in his dissertation on the medical profession) and Stigler (and probably beforeâ€”I just learned it from Stigler and Sam Peltzman) have noted that entry restrictions are the most straightforward way for regulators or professional organizations (e.g,. the AMA) to cartelize an industry and create rents for producers. To most economists, the use of entry restrictions is prima facie evidence of the inefficiency of a regulatory policy. Yet although I’ve read pretty much all I can get my hands on about exchanges, I can’t recall seeing anyone level similar criticisms against exchanges despite the fact that they have almost universally imposed such limits. Indeed, it’s not uncommon to hear exchanges lauded as the exemplars of perfect competition. ‘Tis a puzzlement. (Hat tip to anybody who can provide a cite contradicting my assertion regarding the absence of academic criticism of entry limits at exchanges.)
Is there any way of comparing the relative efficiencies of the old and new monopolies? Looking at outputâ€”trading volumeâ€”may provide some evidence. IPE experienced an appreciable volume jump when it closed the floor, which would suggest that the new model is more efficient than the old. Other comparisons of this type could be illuminating.
The moral of the story is that exchange market power inheres in the fundamental nature of trading, but that the means by which an exchange reaps the rents of market power depend on its organizational and ownership form.
It may be the case that the transparency of exchange fees has regulatory and anti-trust consequences. As exchanges become more like traditional firms, they may draw more anti-trust scrutiny. That will be, as Arte Johnson used to say, verrry interesting.