Streetwise Professor

November 11, 2010

One Member, One Vote?

Filed under: Clearing,Derivatives,Economics,Exchanges,Financial crisis,Politics — The Professor @ 9:50 pm

Frank-n-Dodd mandated clearing.  The CFTC is now promulgating a raft of rules regulating the clearing process.  One of these rules limits the ability of clearinghouses to determine to whom they offer clearing services, and how:

The Commodity Futures Trading Commission approved a proposed rule to force the world’s largest banks to offer swaps clearing services to smaller firms, a change that would promote competition in the $615 trillion market.

. . . .

“Swaps dealers or major swaps participants are prohibited from interfering with or attempting to influence decisions related to the provision of clearing or the acceptance of clearing customers,” according to the proposed rule. The potential regulation now enters a 60-day public comment period before any final CFTC vote.

I’ll defer comment now on this particular rule because it hasn’t been posted on the CFTC website yet, and want to read the full rule rather than going on just the news report.  But this does relate to a general issue that I’ve been thinking about lately: CCP governance.  So I’ll provide some general analytical observations on this subject that are likely to be relevant as the CFTC and SEC proceed on clearing issues.

I’ll focus on CCP membership and its implications for the way CCPs are run.  In particular, in addition to controversy about who should be allowed to open clearing accounts and who should have a say in who can do so, there is also controversy about who should be able to become a CCP member, and CCPs’ ability to set membership standards and to exclude some firms from membership.  This issue came up at the roundtable I participated in on 22 October.  There is a group of smaller banks and brokerage firms that are pressing to become CCP members, and to limit the ability of CCPs to set standards that could result in their exclusion.  This is likely to be a big issue going forward.

As I’ve written for years, there are difficult trade-offs here.  On the one hand, there is the issue of market power.  I’ve shown–in work going back into the 1990s–that when there are substantial scale economies, as is the case in clearing, it is possible for a CCP–even if it is member-owned, and a not-for-profit–to exercise market power by limiting membership.  That is, a suboptimally small “club” of intermediaries can get together and form a CCP, without fear of competition from another CCP due to the scale economies.

On the other hand, allowing open access can jeopardize the financial security of the CCP.  I want to focus on this issue.

A key point here is heterogeneity.  CCPs with heterogeneous members, in particular firms that vary in their financial strength, risk management capabilities, etc., face difficulties that CCPs with more homogeneous members do not.  These difficulties, in fact, can undermine the ostensible public purpose behind mandating CCPs: reducing systemic risk.

CCPs are firms that have to be governed and managed.  It is well known that heterogeneity increases the costs of governing cooperative organizations.  Henry Hansmann’s The Ownership of Enterprise provides numerous examples of this.  Conflicts between members are more likely, the more heterogeneous they are.  In particular, some members can attempt to use the organization to transfer rents from other members.  In my 2000 JLE paper “A Theory of Financial Exchange Organization” I explain a lot of the supposedly inefficient and cumbersome governance mechanisms in futures and stock exchanges as a rational way of mitigating rent seeking among the heterogeneous members of these exchanges.

Let’s take the case of a CCP in particular.  A CCP mutualizes risk, and therefore creates potential moral hazard problems.  The CCP can mitigate these problems by imposing capital requirements, and crucially, by choosing margin/collateral policies.

When CCP members are heterogeneous, in particular, when some are better capitalized than others, or when some incur higher costs to post collateral than others, the members will disagree on what the appropriate margins should be.  The better capitalized firms and those who incur lower costs to post collateral will prefer higher margins.  The more poorly capitalized firms and those who incur higher collateral cots will prefer low margins.

The well capitalized firms are concerned that since clearing makes the cleared products fungible, customers don’t care who they trade with.  Since risks are shared, every member is effectively a perfect substitute from the perspective of customers.  Thus, clearing tends to make it easier for less-well capitalized firms to get business.  But since default risk is mutualized, this reallocation of trading activity towards the less-well capitalized firms tends to force the better-capitalized firms to suffer more default losses.  One way (albeit, imperfect) to control this problem is to raise the costs that the less-well capitalized firms incur (relative to the costs of the better-capitalized ones).  Since collateral is typically costlier for the lower-cap firms, this can be done by charging relatively high margins.

Thus, I would expect that more highly capitalized firms would prefer higher margins than the less highly capitalized firms.  (Similar arguments can be made relating to other risk control policies that CCPs might adopt.)  The more heterogeneous the membership, the more acute the conflict between the members.

The CCP governance mechanism has to have some way of dealing with these conflicts, and making decisions.  Some decisions are going to be made by voting.  With one-member, one-vote kinds of systems, the median member will drive the outcome.  A CCP with a few big members and a larger number of small members may therefore tend to choose policies that favor the smaller members.  This is not likely to be optimal from a systemic risk perspective, and is also likely to result in larger members doing things to limit their exposure.  This in turn limits the risk sharing benefits that a CCP can offer.

Other kinds of voting rules can mitigate these problems, e.g., rules that are based on size or capitalization in some way, with bigger/better-capitalized members having more votes.  But smaller members are unlikely to be satisfied with merely belonging to a CCP but having to defer to the bigger members on key policy issues.  Hence if they pressure regulators and Congress to craft regulations that require CCPs to admit them, they are also likely to pressure regulators and Congress to craft regulations of CCP voting rules that give them some influence over CCP decisions.

Other decisions will be implemented by CCP management.  But CCP managers will be accountable to the members, and subject to influence by them.  Here again, voting rules will matter.  Moreover, Peltzman-Becker pressure group considerations suggest that managers, like politicians, will adopt policies that balance support among the various constituencies.  Median-voter type considerations again exist.  This means that CCPs with more heterogeneous members are likely to adopt policies that appeal to the median member.

In brief, when trying to determine the riskiness of CCPs, you have to consider how they are governed and managed.  Moreover, you have to recognize that governance and management depend on the composition of the membership.  And in particular, that CCPs with more heterogeneous members will behave differently than CCPs with more homogeneous ones.

One concern that is oft-expressed is that there will be a “race to the bottom,” with CCPs competing on margin and thereby compromising safety and creating systemic risks.  Given the economies of scale and scope in clearing, I doubt that the competition between CCPs will end up being all that intense, although there may be an initially period of intense competition that results in the market tipping to one or two dominant CCPs.

This analysis of CCP governance suggests another concern, and one that I think is more of a real worry.  Namely, that member heterogeneity can lead CCPs to adopt policies that benefit the median member, but which are not optimal from a systemic risk perspective.   Clearing inherently faces a moral hazard problem.  This moral hazard problem can create systemic risk.  Member heterogeneity makes it more difficult to control this moral hazard problem.  Internal pressures from riskier members–who can essentially pass their risks and the associated costs on to less risky ones–can lead CCPs to choose margin and other risk control policies that are not systemically optimal.  Internal pressures to reduce margins, rather than a race to the bottom between competing CCPs, is something to be concerned about.

One other historical note that illuminates another important cost of heterogeneity.  In Banking Panics of the Gilded Age, Elmus Wicker shows that with one notable exception (the Panic of 1873), the New York Clearinghouse was unsuccessful in dealing with financial panics and contagion.   (NYCH was a bank clearinghouse, not a derivatives clearinghouse, but there are important similarities.  Most importantly, NYCH had the ability to mutualize some risks.)  Wicker argues that conflicts between heterogeneous members were the main impediment in dealing with panics.  Although mutualization of risks would have mitigated panic (because the banks collectively were more likely to be solvent than any individual bank), mutualization transferred wealth from the stronger banks to the weaker ones.  The inability to overcome this distributive conflict stymied the ability of the NYCH to respond to crises in an effective way.   Thus, not only can heterogeneity increase the likelihood of a problem at a CCP (due to its effect on the severity of moral hazard problems), it can reduce the effectiveness of a CCP in dealing with a crisis situation.

This means that regulators should be extremely careful when crafting rules for CCP membership and voting policies.  Yes, there are legitimate grounds for concern that CCPs will use membership policies to exercise market power.  But there are also grounds for concern that forcing CCPs to admit more firms will lead to more heterogeneity, and that this can raise governance costs and jeopardize the financial safety of CCPs.  Similarly, regulation of governance and voting procedures to give more power to smaller members can also raise governance costs and jeopardize safety.  Inasmuch as the ostensible purpose of CCPs is to improve the stability of the financial system, it would be perversely ironic if political pressures induced regulators to adopt rules on CCP membership and governance that created systemic risks.  Tread carefully, people.

Print Friendly

3 Comments »

  1. […] This post was mentioned on Twitter by R, John Kiff. John Kiff said: Craig Pirrong: One member, one vote? Allowing open access can jeopardize the financial security of the CCP http://ow.ly/38yTa […]

    Pingback by Tweets that mention Streetwise Professor » One Member, One Vote? -- Topsy.com — November 12, 2010 @ 1:14 pm

  2. I can claim from personal experience and also based on checking with a number of my peers, that the claims of the mid-tier banks/brokers to have been excluded from clearing via large banks are simply false – as Head of Clearing of one G14 bank, none of them have called either my firm or the 5 of my peers with whom I’ve checked. We’d be delighted to have their business. That said, perhaps they object to the pricing or margin regime they may face, which is different from Futures for obvious reasons [risk, liquidity, capital costs].

    Allow me to elaborate. These brokers may seek to cross their trades, thereby running a market risk flat book and hence feel they should not be required to post margin. However, if either side of a cross fails to be cleared, which is possible in the OTC space, then the clearer acting for the broker has a problem ie it will have to post margin on the imbalanced positions, potentially intra-day, and so will ask the broker to overcollateralise its’ account to protect against such occurances. Scale this up for the volume of crosses and potential exposures, and suddenly you may arrive at a large “risk deposit” that the broker is required to provide. This is aside from the default fund contribution that may also have to accompany such imbalances that bizarrely is not charged on indirect participants but only on clearing members.

    Whilst this appears like an obstacle to deny access, it is actually prudent risk management in relation to “clients”/brokers who are typically thinly capitalised and whose volumes are disproportionate to their financial standing.

    As for heteogenous membership, this is definitely the key tension both with small v large members, in a similar fashion to the margin debate with indirect participants who would prefer that the risk burden was biased towards default fund contributions rather than initial margin [they only suffer the latter in funding terms].

    Notably, in contemplating lowering entry requirement, perhaps the regulators who are keen to lower entry requirements should underwrite the consequences of any failure associated with “newly eligible members” rather than pass this burden onto large banks, thereby avoiding increasing the dependency on what have been termed “Too big too fail” banks.

    Comment by John — November 12, 2010 @ 2:31 pm

  3. Has the prof read about Ambac’s bankruptcy plan?

    Comment by Surya — November 13, 2010 @ 12:07 pm

RSS feed for comments on this post. TrackBack URI

Leave a comment

Powered by WordPress