Nirvana Is Just a Band
| Craig Pirrong |
Last week I wrote about one justification for exchange trading and clearing mandates in derivatives markets — the market power argument. This week I’ll examine another argument, and render a similarly skeptical verdict.
In a chapter of Restoring Financial Stability, Viral Acharya, Rob Engle, Steve Figlewski, Anthony Lynch and Marti Subrahmanyam argue that bilateral transactions in OTC derivatives markets involve an externality. Their argument is not stated that clearly, but FWIW here it is verbatim:
[A]ll OTC contracts . . . feature collateral or margin requirements, wherein counterparties post a deposit whose aim is to minimize counterparty risk. The deposit is marked to market daily, based on fluctuations in the value of the underlying contract and the creditworthiness of the counterparties . . . . The difficulty, however, is that such collateral arrangements are negotiated on a bilateral basis. Parties in each contract do not take full account of the fact that counterparty risk they are prepared to undertake in a contract also affects other players; indeed, they often cannot take account of this counterparty risk externality in an OTC setting, due to inadequate transparency about the counterparty’s positions and its interconnections with the rest of the market. While bilateral collateral arrangements do respond to worsening credit risk of a counterparty, such response is often tied to agency ratings, which are sluggish in capturing credit risk information and potentially inaccurate.
An externality means that some cost or benefit is not priced. By invoking the concept of externality Acharya et al (“AEFLS”) are asserting that something — a bad in this instance — isn’t priced. They are a very vague on just what this is, but here’s my interpretation of what they mean.
A firm that has already entered into financial contracts affects the risk exposure of its existing counterparties when it enters into new deals. A firm that has a large number of commitments outstanding can enter into additional contracts that substantially increase its riskiness, thereby harming the incumbent counterparties. The cost imposed on these incumbent counterparties isn’t, in this telling, priced.
This is, in essence, an asset substitution argument. And indeed, asset substitution is always a concern in any credit relationship, whether it is between a bank and a borrower, or derivatives counterparties.
But an understanding of the way these markets work suggests that the asset substitution problem is far less important than AEFLS suggest. One way of controlling this problem in lending relationships is to rely on short term debt, allowing repricing of old debt frequently and thereby limiting the gains from asset substitution.
In OTC derivatives markets, similar mechanisms are at work. The main dealers rely heavily on short term financing. Thus, attempts to engage in this type of behavior can lead to substantial increased financing costs, and indeed, a loss of access to short term funding altogether — this is a death sentence to a financial institution, as Drexel found out in the 1980s, Salomon almost experienced in the 1990s, and Bear and Lehman suffered during the crisis.
In fact, Diamond and Rajan argue that the “fragile” capital structure of financial intermediaries–the vulnerability to runs or other mechanisms that result in the rapid withdrawal of access to funding — are intended precisely to provide such discipline. Thus the supposed bug–the vulnerability of a big liquidity supplier to failure — is, in some respects, a feature. Without it, the opportunism that AEFLS worry about would be a greater concern.
OTC dealers also trade derivatives extensively with one another. Thus, changing creditworthiness can affect their ability to deal in these markets, and the pricing of these deals.
There are, therefore, mechanisms by which financial institutions internalize the costs associated with the riskiness of their dealings. Of course, these do not work perfectly. Of course, information is imperfect. But it is essential to note that myriad lenders and counterparties are continuously interacting with big OTC dealers in a variety of different markets. They are doing their own credit appraisals. They are voracious consumers of any bit of information about other players in the marketplace. The markets for funding and derivatives can serve to aggregate this private information in the prices and credit terms of derivatives trades, and in the funding costs that the institutions pay, thereby pricing (albeit imperfectly) counterparty risks.
Moreover, as Peter pointed out in an earlier post on O&M, it is necessary to make an explicitly comparative analysis when evaluating policy alternatives. The identification of an imperfection of the OTC market is not sufficient to conclude that it should be replaced by some other alternative that is inevitably imperfect as well.
In this regard, a comparison of the OTC market with the particular alternative that AEFLS advocate — central clearing–demonstrates that OTC markets arguably dominate clearing even on dimensions emphasized by AEFLS.
Most notably, clearinghouses almost NEVER condition margin levels on the creditworthiness of the member firms. If they do (as, for example, LCH.Clearnet does in limited circumstances) they rely soley on agency ratings. This reflects a variety of factors, including a lack of information to make such judgments. Moreover, it reflects the fact that it would be very difficult, in practice, for a clearinghouse to discriminate among its members. Such discrimination — which could have serious competitive effects — would expose the clearinghouse to intense influence activities. (Thus, clearinghouses are most likely to work effectively when the members are homogeneous, and heterogeneity can be an impediment to their formation.) This means that clearinghouses do NOT price counterparty risk in a discriminating way. In a cleared market, a major determinant of counterparty risk is in the public domain, and not priced. This is not conducive to an efficient allocation of trades among firms, or in the scale of trading activity.
In contrast, in the OTC market, dealers have a wide variety of sources of information about the creditworthiness of potential counterparties that they take into account when determining the terms on which they are willing to deal with these counterparties. That is, although their information is imperfect, big dealers almost certainly have better information on the creditworthiness of other dealer firms than a clearinghouse would. What’s more, the fact that any given firm is being evaluated by multiple counterparties and lenders means that (a) the private information aggregation mechanism noted before can work in the OTC market in a way impossible with a single evaluation by a clearinghouse, and (b) evaluation errors are somewhat diversified in the OTC setting, but with one monitor in a cleared market, if the clearinghouse gets it wrong the problem can be huge.
That is, counterparty information in OTC markets is imperfect, but almost certainly better than in a cleared market. Moreover, there are a greater array of mechanisms available to price counterparty risk based on this information: not just in margins, but in the pricing of deals (many deals explicitly adjust prices to reflect credit terms), the pricing of the financing that dealers require to operate, and in the sizes of transactions.
I’ve argued extensively in some working papers that as a result of these information advantages, contrary to the claims of AEFLS, counterparty risks are likely to be priced MORE accurately in a bilateral market than a cleared one. Clearing can offer some offsetting advantages, but the key point is that a trade-off is involved. In my view, it is highly likely that this trade off frequently favors bilateral, OTC dealings rather than central clearing.
Like so many advocates of clearing, AEFLS fall victim to the Nirvana fallacy. They point out imperfections in OTC markets, and do not carry out a comparative analysis of the relative imperfections of alternatives. What’s more, their analysis of the supposed failings of the OTC market is incomplete because they do not fully explore how market participants structure their transactions to deal with the putative market failure they identify.
They’re miles ahead, though, of Barney and Chris and Timmy! and Gary and Mary and the gang. At least AEFLS try to identify an alleged market failure. Even though they fall into the classic Nirvana fallacy, at least they point out a potential inefficiency, thereby permitting an rigorous exploration and analysis of this possibility. Which is way more than can be said for the Sorcerer’s Apprentices attempting to reorganize completely the largest and most complex financial markets in the world.
Cross posted on StreetwiseProfessor.