LCH.Clearnet has announced that three exchanges have expressed an interest in buying the heretofore independent CCP. Why does stuff like this happen on my vacation? I’ll just make a brief comment, and leave more for later.
Jeremy Grant posits that LCH is in play because of cross margining efficiencies. This conjecture is understandable, but fundamentally flawed. Yes, there are definitely cross margining efficiencies. More generally, there are economies of scope due to diversification effects that make it less costly to consolidate clearing for multiple asset classes in a single CCP. But this does not explain why exchanges would want to vertically integrate clearing and execution. Ownership is not necessary to exploit scope economies. Indeed, the two things are often in tension–and are in this instance.
Scope economies do not explain integration. Indeed, things cut the other way. To the extent that there are more scope economies in clearing than execution (which is quite plausible), in the absence of transactions costs the most efficient way to organize the business would be to form a single clearinghouse, with which multiple execution exchanges would contract to obtain clearing services. In fact, one of the costs of integration is that it often entails the sacrifice of scope economies in clearing.
Put differently, scope economies do not explain the pattern of ownership: the relevant question is why do exchanges find it optimal to own their own clearing facilities, rather than obtaining clearing services by contract. That would be the way to optimize cross margining, netting, and diversification economies. Such economies would be exploited maximally by having, say, LCH, clear just about everything, and sell its clearing services to myriad exchanges that offer execution services in different product classes. But that’s not the way the market is moving.
Thus, the fact that many exchanges are already integrated, and why more exchanges are seeking to integrate, must arise from some other economic force than scope economies. The most plausible explanation–and one that I advanced more than 5 years ago–is transactions costs. Due to double marginalization problems and the transactions costs that arise when complementary activities are supplied by single firms (or by a small number of firms that do not compete vigorously), it is often efficient to integrate these complementary activities. That is a basic lesson of Williamsonian Transactions Cost Economics (TCE).
In my view, that’s why exchanges like CME and DB have integrated, and why other exchanges have or are in the process of doing the same thing. This drives regulators and legislators nuts: they see such integration as a nefarious scheme to leverage monopoly. That’s bad economics–and it’s been known to be bad economics for going on 50 years. Instead, it is all about transactions cost economization. Regulators can try and fight it, and they will. If they succeed in their quest, however, they will only succeed at raising costs and reducing efficiency.
Insofar as the potential LCH deals with exchanges are concerned, SwapClear is presumably a major target in any deal, but also a major obstacle. I can’t see the banks that dominate SwapClear going along with a merger of LCH into NYSE-EuroNext or NASDAQ or LSE. Or if that happens, I would expect them to leave SwapClear and set up their own CCP, perhaps through ICE Trust or through a new, freestanding interest rate swap and repo CCP. (Going through ICETrust would exploit scale economies, so that would be a likely approach.)
I have more to say on this, and probably will. But it’s been a long day in the sun, so it will just have to wait.