The debate over exchange competition and mergers–and the CME/CBT merger in particular–continues to percolate. I’ve commented on several aspects of this debate, but other remarks are in order.
Further my previous musings on elasticities of substitution between CBT and CME products. I’ve opined that they might be negative (due to spreading). I’ve also opined that even things like Treasury futures and Eurodollar futures are unlikely to be close substitutes due to the non-negligible basis risks. But upon further reflection, I believe there is another even more important factor.
Specifically, the demand for an exchange’s services is a derived demand. That is, futures traders demand a bundle of services–liquidity, brokerage, clearing, execution, etc. Of this bundle, liquidity costs (e.g., price impact costs) are probably by far the biggest. Exchange fees represent a relatively small fraction of the total cost of a trade.
This is exactly what gives exchanges pricing power. If the exchange’s fees represent say 5 percent of the total cost of executing a transaction, if an exchange doubles its fee, the cost of trading goes up by 5 percent. Thus, the demand for an exchange’s services is very own price inelastic. Hence the ability of exchanges to earn such nice margins–this inelasticity gives them considerable pricing power.
This also has important implications for cross price elasticities. For instance, if the CBT halved its trading fee on Treasuries in an attempt to attract business from the CME, the cost of trading Treasuries would fall by about 2.5 percent (again assuming that fees represent about 5 percent of the cost of a trade.) Thus, in this example, the cross price elasticity for CBT/CME services is about 2.5 percent of the cross price elasticity (taking ALL costs into account) between Treasuries and Eurodollars. Thus, if I am correct that the “gross” cross price elasticity (taking all costs into account) is small to begin with due to basis risks and spread trading, the cross price elasticity relevant for evaluating the competitive effects of a CME-CBT merger is far smaller. Even if the gross price elasticity is somewhat larger than I conjecture (and positive), the effect of the fact that exchange fees represent a relatively small fraction of total trading costs tends to make the relevant cross price elasticity small nonetheless. Small cross price elasticities drastically attenuates the “static” competitive effects of the merger of exchanges trading non-overlapping product sets.
There is one case where exchange fee cutting did appear to be decisive–Eurex vs. LIFFE in 1998. A couple of points. First, this involved competition in identical products–no basis risk or spreading considerations. This is not relevant in evaluating the CME-CBT merger. Second, LIFFE made key strategic errors that are unlikely to be repeated in the future. In particular, it did not respond to the Eurex price cuts, whereas such inaction is unlikely in the future. Anticipating a quick response to price cuts, an exchange is unlikely to initiate a round of price cutting. Third, due to the rough parity in other trading costs, the Eurex price cut was sufficient to induce tipping, meaning that the effect of the price cut was exaggerated. This is not likely to happen in competition between different products. For instance, a CBT price cut is not going to cause all volume to leave Eurodollars.
Thus, in my view, cross price elasticities between CBT and CME products are likely to be quite small, meaning that the “static” competitive implications of the merger are likely to be quite benign.
Now consider the impact of the merger on innovation. John Lothian opines that the merger would reduce competition to innovate. This is a complicated issue, but I respectfully disagree that innovation will suffer, or that if efforts to innovate do in fact decline, that this is necessarily inefficient.
I should note that there is little agreement among economists about the effect of market structure on innovation. Profitability is what provides the incentive to innovate. Reduction of competition through merger does not attenuate this incentive, and may strengthen it.
I don’t believe that a CME that absorbs the CBT will rest on its laurels, and feel no need to innovate. Indeed, the high CME stock price clearly capitalizes the stock market’s expectation of high cash flow growth. This growth is unlikely to result from volume growth in existing products alone, as healthy as this growth has been. Pleasing the god of the street will require growth from other sources–new products. The CME is like a pharmaceutical company that needs to keep its pipeline full. The stock market will punish the CME if it fails to do so. This will tend to concentrate management’s minds and keep them on the hunt for new products.
Moreover, even if one believes that more competitive market structures are more conducive to innovation, in the modern electronic trading era, in which foreign exchanges (absent regulatory barriers) can offer products in US markets (and vice versa), and in which global financial firms trade on exchanges around the world, the relevant market is the world market. Looking at the US market share of a combined CME-CBT entity is inappropriate in this regard. From a world-wide perspective, the increase in market share (or the Herfindahl index) resulting from the merger would not typically present anti-trust concerns.
It must also be remembered that the winner-take-all nature of exchange competition can mean that competition to create new products can be inefficient. The winner of the competition earns a rent. As in patent races or land rushes, competition to earn the rent can dissipate it, creating a social cost of approximately the same magnitude as the rent. Thus, competing firms can spend too much on innovation. (More innovation, or more expenses spent on innovation, are not always better. There can be too much innovation, or too much spent on trying to develop new products.) Perhaps counterintuitively (to most folks) softening competition can actually improve welfare.
Against this is a good point that Lothian raises. Specifically, nobody knows exactly what the right design for any futures product should be. Everybody agrees that credit derivatives are the Next Big Thing for exchanges. But nobody agrees on what the “right” credit derivative product design should be. Having more exchanges experimenting with different designs increases the likelihood that one of them will be close to optimal. Losing a “laboratory” through merger may reduce the odds of finding the Goldilocks design. Note, however, that exchanges can revise their contracts if they don’t appear well-suited to market needs. Also, at the risk of being repetitive, due to the rent at the end of the rainbow for the winning exchange, exchanges may spend excessively to find the product that is Just Right.
All in all, I am skeptical that the CME-CBT merger will appreciably reduce innovation. Moreover, given the nature of liquidity (and the consequent winner-take-all nature of exchange vs. exchange battles for market share in a given product type), dissipative (“excessive”) rent seeking competition between exchanges is a real concern. Given these circumstances, I do not believe that concerns about the impact of the merger on innovation can provide a firm basis for blocking it.