Streetwise Professor

CFTC Commissioner Jill Sommers gave a speech at a conference titled “Clearinghouses as Mitigators of Systemic Risk” that makes interesting reading.

As the title of the conference might suggest, Sommers was quite conversant in the clearing catechism:

Generally, a comprehensively regulated clearinghouse should help reduce systemic risks by facilitating the netting of transactions and by mutualizing credit risks. When a contract between a buyer and seller is submitted to a clearinghouse for clearing, the contract is “novated” to the clearinghouse. As such, the clearinghouse is substituted as the counterparty to the contract and then stands between the buyer and the seller. Clearinghouses then guarantee the performance of each trade that is submitted for clearing. Clearinghouses use a variety of risk management practices to assure the fulfillment of this guarantee function and to avoid losses at the clearinghouse. Moreover, the CFTC has closely monitored these risk management practices at the clearinghouse level and at the futures commission merchant level to ensure risk is properly being managed. The Commission has always recognized that the clearinghouse must protect itself from unnecessary risk, and if it failed to do so, large segments of the markets and market participants could be put at unnecessary risk as well.

Clearinghouses mitigate risk through the daily discipline of marking to the market price, at least once each business day, each position cleared through the use of prices that are independently determined by the clearinghouse. All gains and losses that arise as a result of the mark-to-market process are settled (i.e., paid and received) each day. Clearinghouses also require the daily posting of margin to cover the daily changes in the value of all positions as extra protection against potential market changes that are not covered by the daily mark-to-market. The methodology used by clearinghouses to calculate such margin requirements are subject to regulatory review and approval.

. . . .

With the mandated use of a regulated clearinghouse coupled with effective risk management practices, the failure of a single large trader, like AIG, would be much less likely to jeopardize all of the counterparties to its trades. I must stress, however, all risk is not eliminated, but it is substantially reduced. One of the lessons that emerged from the recent financial crisis was that institutions were not just “too big to fail,” but also too interconnected through non-transparent swaps that the institutions did not effectively manage. By mandating the use of central clearinghouses, institutions would become much less interconnected, mitigating risk and increasing transparency. Throughout this entire financial crisis, trades that were carried out through regulated exchanges and clearinghouses continued to be cleared and settled.

Needless to say, I’m a heretic.  Mandated clearing doesn’t necessarily increase netting opportunities, and even if it does, since it effectively reorders creditor priority it does not necessarily reduce systemic risk.  And if netting reduces costs so much, why didn’t firms adopt clearing voluntarily?  The “daily discipline of marking to the market price” is problematic if “the market price” is difficult to determine, due to illiquidity/lack of trading.  What’s more, regular posting of variation margin is not unique to CCPs, and is quite common in OTC deals.  Furthermore, the variation margin mechanism can be exactly what touches off or exacerbate market panics.  AIG is a terribly misleading example.  CCPs do not reduce interconnectivity; they just change the topology of the interconnections.  Furthermore, CCPs are tightly embedded in the financial marketplace, most notably in the credit system.  To pretend otherwise is foolish.

Sommers’s other remarks are more intriguing.  She recounts of the CFTC’s previous policies towards OTC derivatives and clearing in particular:

In July 1989, the Commission published its 1989 Swaps Policy Statement. This statement reflected the Commission’s view that most swap transactions, although possessing elements of futures or option contracts, were not appropriately regulated as such under the Act and Commission regulations. This was important because the Commodity Exchange Act required all futures contracts to be traded on exchange. The policy statement allowed these transactions to continue trading off-exchange by creating a “non-exclusive safe harbor” for transactions satisfying certain requirements. The safe harbor requirements included five elements: (1) individually-tailored terms; (2) absence of exchange-style offset; (3) absence of clearing organization or margin system; (4) the transaction is undertaken in conjunction with a line of business; and (5) a prohibition against marketing to the public.

In other words, prior to the CFMA of 2000, any attempt to introduce netting (item (2)) or clearing and margining (item (3)) into the OTC market would have run afoul of the Commodity Exchange Act’s requirement that all futures contracts must be traded on exchanges (designated contract markets); netted and/or cleared swaps would have been considered futures.

This should be remembered, but is typically not.  I’ve often wondered whether CFTC policy, and the CEA more generally, discouraged the evolution of OTC clearing, and channeled the industry along the path that Dodd-Frank has now declared an anathema.  We’ll never know for sure, but it must be recognized that during the formative years of the OTC market, the law and regulation effectively foreclosed–precluded–the development of institutions that Congress and the CFTC now deem essential for financial stability.  The irony.

Sommers adds some refreshing honesty about the monster that is Frank-n-Dodd.  Most notably, she discusses the swaps execution facility mandate:

An area that I am very interested in, as are many other market participants, is what the requirements of trading swaps on a SEF will be. The CFTC and SEC heard a great deal of concern about this issue from market participants at a joint roundtable on September 15th. CFTC staff has recently estimated that 30-40 entities will register as SEFs or designated contract markets. The designated contract market model is easy for the Commission and market participants to deal with. We know how they work, and market participants know how they work. SEFs are new, and the relevant statutory language is not very clear. When you read the new statutory language in conjunction with existing statutory language, it raises some questions.

Section 1(a)(34) of the CEA defines “trading facility.” It is a definition that the CFTC and market participants have been working with for years. In essence, a trading facility is a physical or electronic facility or system wherein multiple participants can execute or trade agreements by interacting and accepting bids or offers of multiple other participants. A many-to-many type model.

In defining what a SEF is in Section 721 of Dodd-Frank, Congress did not require that a SEF be a trading facility. Instead, Congress defined a SEF as a “a trading system or platform” in which multiple participants have the ability to execute or trade swaps by accepting bids and offers made by multiple participants in the facility or system, through any means of interstate commerce, including any trading facility, that (A) facilitates the execution of swaps between persons; and (B) is not a designated contract market.” (emphasis added).

So clearly, by this definition Congress intended that a trading facility could be a SEF, but that a SEF did not have to be a trading facility because Congress used the words “trading system or platform” to define SEFs. In my view, by introducing the term “trading system or platform” when we already have “trading facility” as a defined term means that Congress must have intended a SEF to be different from a trading facility.

Her confusion is palpable.  And understandable.  If the status quo ways of executing swaps were OK, Dodd-Frank would not have added  an SEF requirement.  If Dodd-Frank intended an SEF to be the same as a trading facility, it would have said so.  Thus, Frank-n-Dodd says what SEFs aren’t–the status quo or trading facilities–but doesn’t say what they are.  It is particularly confusing because some of the requirements that SEFs must meet–interaction between bids and offers submitted by multiple participants–characterize trading facilities.  So Congress essentially handed the CFTC an enigma and walked away.  One can feel for Sommers, her fellow Commissioners, the staff–and the industry.

She also mentions the confusion over the definition of swap dealer:

One of the areas that has caused a great deal of concern among many market participants is whether they will be classified as a swap dealer or major swap participant. Such a classification brings with it the imposition of capital, margin, recordkeeping and retention and business conduct rules, and rules for segregating customer funds. Currently six rule writing teams are working on this area. One is working on a series of definitions, including swap dealer and major swap participant, one is working on registration requirements, and others are working on business conduct standards, capital, margin, and segregation issues.

The CFTC’s Chairman has stated that initial estimates are that there could be in excess of 200 entities that will be swept up into the definition of swap dealer based upon current ISDA primary member status. Registering and regulating that many swap dealers will be a huge task for the CFTC and will have very important implications for those entities and their businesses. Comparatively, we now have about 127 FCMs that are registered with the Commission. In my view it is critical that we ensure that this new regulatory regime intersects seamlessly with our existing regime to ensure that registrants and their customers suffer no disruption in their business and hedging strategies. I have asked staff to estimate the number of entities that may be swept into the definition of major swap participant. Unfortunately, they have not yet been able to do so. This uncertainty raises concerns about ensuring a disruption-free transition, an issue that the Commission and staff will be closely monitoring.

Note, and beware.  If Gensler’s 200 number is even ballpark, CFTC will have responsibility over more than twice as many entities as is currently the case.  Moreover, these new entities are often larger, more complex, and certainly more diverse than the population of FCMs currently under its ambit.  Big hedge funds, supermajor oil companies, big money managers, and universal banks are all likely to be considered swap dealers, or major swap market participants.  These are very different animals, and far more complicated ones, than the FCMs.  Moreover, they are all different from one another in crucial ways, but they have to be shoehorned into one of two standardized categories.  This is unlikely to end well.

Sommers also notes that CFTC is tasked with establishing position limits, but doesn’t have the data necessary to do the job properly:

Complicating all of that are the provisions of Dodd-Frank. The CFTC and SEC have 360 days to issue regulations establishing swap data repositories to which swap data will be reported. The Commission has 180 days for energy and 270 days for agriculture to propose aggregate position limits across futures markets and equivalent OTC markets. The problem is, Dodd-Frank requires us to propose position limits months before a mechanism is in place for obtaining the necessary data from the OTC swap markets. In order to propose appropriate limits, we must know the size of these markets. Without the necessary data for OTC markets, we will not really have all the information we need to propose appropriate limits. But, the law is the law so we will propose limits and hope they are at such a level that they do not cause damage to these markets.

Sommers also indulges in hope over experience:

CFTC and the SEC. Historically, the SEC and CFTC have not been very successful proposing joint rules, even when directed by Congress to do so. This time, however, may actually be different.

Seeing will be believing.  And even if everybody has been playing well together so far, the tough decisions have yet to be confronted, and the affected interests have yet to play one agency against another, as will occur inevitably.  Add to that normal inter-agency rivalry, compounded by a profound change in relative authority, and it is pretty plain that serious disputes are inevitable.

Sommers’s speech is a good overview of the issues and challenges facing the CFTC.  Reading her speech and Frank-n-Dodd gives me the same feeling that I had on some Christmas mornings when tasked with assembling some rather complicated toy, and having to rely on directions originally written in Chinese, then translated into Korean, before being translated into English.  The main difference being that if I threw up my hands in frustration, there would be just one disappointed little girl.

I wish Sommers and the CFTC luck.  They’re going to need it.  And so will we.

Jeff Carter at Points and Figures wrote a nice post on my Cato clearing piece, for which I am appreciative. He is pretty much in agreement with me that mandates are a bad idea.  He is concerned with the command-and-control mindset inherent in Frank-n-Dodd, and supports the development of voluntary “Coasean” approaches.  Hear, hear.

Jeff does take issue with my characterization of the way that CCPs work today:

That meant that if you blew up, and caused major losses you could not afford, the rest of the clearinghouse members would have to pony up cash to cover the loss. If the loss were big enough to blow through all the cash that all the members had, the clearing house would go bankrupt. There hasn’t been a bankrupt clearing operation in the modern day, since 1900, history of trading that I know of.

Exchange clearinghouses are no longer mutually held, and have not been since the last major exchange went public in 2003. Instead, the clearing house takes out an insurance policy to cover potential problems. It also has lines of credit with several different bankers. This simple fact doesn’t change our agreement that government mandated clearing is a poor solution, but it does change the calculus of analysis a little.

I take issue with his taking issue.  Yes, exchanges are demutualized, but counterparty risk is still mutualized.  Take the CME, for instance.  This CME document, in the sections labeled “Default by a Clearing Member” and “Summary of Resources Backing Clearing” beginning on p. 11 demonstrates that the losses of default are borne by other clearing members.  First the CME dips into the guarantee fund–which the clearing members fund.  Then the CME can assess the clearing members.  (The CME also commits up to $100 million of its own capital.)  The assessments are limited to 275 percent of each CM’s original contribution.  The total from the CME contribution, the guarantee fund, and the assessment comes to about $7.5 billion.  The costs of default by a clearing member are thus shared among–mutualized by–other clearing members.

Jeff mentions insurance, but there is no third party insurance that pays in the event of a default at the major exchanges.  (There is no mention of insurance in the CME document.)

Jeff also mentions credit lines, and indeed CCPs have lines with major banks.  In the aftermath of Black Monday, 1987, exchange CCPs realized that they needed a more reliable source of liquidity.  On Black Monday, major banks (the clearing banks) balked at funding big CM margin calls; this is where Fed pressure and liquidity injection was important.  To avoid this problem going forward, the CCPs obtained lines of credit that they could call on to obtain liquidity, just as Jeff says.

It is important to note that these credit lines are just that.  If a CCP draws on the line, it–or more accurately, its members–still owes the bank the money.  The lending bank incurs a loss attributable to a CM default only if the CCP itself becomes insolvent.  That is, the credit line is a source of liquidity: it is not in the first instance a means of sharing the default risk with the bank.  Only if the CCP becomes insolvent do the lending banks suffer any default losses.

The CME document linked above makes that clear.  It calls the credit lines a “Temporary Liquidity Facility.”  It’s a source of liquidity to be used in extremis: it’s not a form of insurance.

There’s an irony, here, of course, and one that gives the lie to many of the claims in support of clearing.  Clearing mandate advocates, notably Gensler, talk about clearing reducing financial interconnectedness, and reducing the potential for contagion.  But clearing in times of stress demands ready access to liquidity.  This inevitably requires a connection between the clearing and banking systems.  It can’t be avoided.  If a clearinghouse runs into a big problem, and draws the credit line, it is possible that the CCP will not be able to pay it back.  In the event, the CCP problem would become a banking problem, and the lending banks would bear some of the default loss.  (Of course, to the extent that banks are clearing members, they will bear default losses too.)

The credit lines also expose the CCPs to bank credit.  If the bank extending the line is itself in financial difficulty, which may well be the case during the kind of crisis that would necessitate the CCP to call on its line, the clearinghouse may not actually obtain the needed liquidity.  So the contagion can work both ways.  From the CCP to the bank, or the bank to the CCP.

Again, the point is that clearing changes the topology of the network of connections among financial institutions: it doesn’t eliminate these interconnections.  One way or another, derivatives default losses are ultimately borne by major financial institutions, clearing or no.  No two ways about it.

S&P has put LCH.Clearnet on a credit watch for a potential downgrade.  The rating agency did so in response to NYSE Euronext’s decision to set up its own clearing operations, rather than use LCH’s.

This is a very useful reminder that CCPs do not make counterparty risk disappear, as certain people are wont to assert.  (You know who they are.)

But there is another important lesson here.  Nobody knows exactly how the market is going to evolve after the imposition of clearing mandates.  In particular, the structure that will evolve for the clearing business is subject to huge uncertainty.  As the NYSE-Euronext decision illustrates, and as the ICE decision to clear for itself before that,  and as did the decision of the SWX to clear for itself before that, clearing structure is a strategic choice for exchanges.  I wrote about this aspect of clearing ad nauseum before the financial crisis.  At the same time, there are strong scope economies to clearing across multiple instruments.   The push of strategic incentives to vertically integrate and the pull of incentives to exploit scope economies by using multi-exchange platforms like LCH make it very difficult to predict what the configuration of the market will look like.  This, in turn, makes it virtually impossible to understand what the systemic risks will be, and what measures will be necessary to mitigate the systemic risks inherent in clearing.

These systemic risks will depend on the topology of the clearing networks, and this topology will be determined by a complex interaction of incentives to integrate and dis-integrate.  Moreover, the topology is unlikely to be static, as the factors driving the costs/benefits of integration and disintegration will not be static either.  Thus, the interconnections in the system will be quite fluid, and as the financial crisis and the flash crash demonstrate, these interconnections can be the sources of systemic risk, if not understood and managed properly.  Force feeding clearing will only heighten the salience of these interconnections.  It will be a challenge to understand, let alone to manage, this system as would be necessary to realize the happy predictions of those who look on clearing as the cure.

Are the regulators ready for this?  Do the legislators who are enamored with clearing understand this?

So I don’t have to?  Because it’s getting tiresome.

But it has to be done, so here goes.

Jeremiah’s latest gurgling appears on the oped page of today’s WSJ.  It starts with a non-sequitur, and careens downhill from there.  Gensler tells a story about his role in the LTCM situation, and then claims that to prevent a recurrence, or a repeat of AIG, it is necessary to reduce the “cancerous interconnections” (Jeremiah Recycled Bad Metaphor Alert!) in the financial system by, you guessed it, mandatory clearing.

Look.  This is very basic.  Do I have to repeat it?  CLEARING DOES NOT ELIMINATE INTERCONNECTIONS AMONG FINANCIAL INSTITUTIONS.  At most, it reconfigures the topology of the network of interconnections.  Anyone who argues otherwise is not competent to weigh in on the subject, let alone to have regulatory responsibility over a vastly expanded clearing system.  At most you can argue that the interconnections in a cleared system are better in some ways than the interconnections in the current OTC structure.  But Gensler doesn’t do that.   He just makes unsupported assertion after unsupported assertion.

If you have any doubts about how interconnected a clearing system is with the banks, just look in detail at what happened on 19-20 October, 1987.

Don’t believe me?  Then consider what Ben Bernanke wrote as an academic in his “Clearing and Settlement During the Crash” (3 Rev. Financial Stud. 1990 at 133):

A prominent part of the institutional structure is the interconnection of the clearing and settlement systems with the banking system.  This interconnection exists at several points.  First, banks are operationally a part of the clearing process. Clearinghouses typically maintain accounts at a number of “clearing banks. Member FCMs are required to maintain an account at a minimum of one of these banks and to authorize the bank to make debits or credits to the account in accord with the clearinghouse’s instructions. This facilitates the settling of accounts and the making of margin calls. Note that the bank’s role may exceed simple accounting if, for example, it must decide whether to permit an overdraft on an FCM’s account.

Second, banks are a major source of credit, especially very short-term credit, to all of the parties, including the customers, the FCMs, and the clearinghouse itself. As was noted above, bank letters of credit can in some cases be used as initial margin. Customers and FCMs often rely on bank credit to facilitate the speedy posting of variation margin, and FCMs would typically have to turn to banks to finance payments made necessary by customers’ defaults or slow payment. In equity markets, banks are often the ultimate source of credit for the purchase of securities on credit.

Finally, it should be noted that while, in the conventional language, most margin postings and settlement payments are made in “cash,” these transactions are, of course, not really made in cash but by the transfer of bank deposits. Thus, the smooth operation of the financial market clearing and settlement system is based at all times on the presumption that the banking system is sound and can satisfy demands for withdrawals of funds.

“A prominent part of the institutional structure is the interconnection of the clearing and settlement systems with the banking system.”  Does it get any clearer?  (No pun intended.)

Ben, would you please drop the “Gentle Ben” demeanor and slap some sense into Gensler?  You actually know something about the subject.  You’re a former educator.  And somebody needs some educatin’.

And consider the implications of a dramatic increase in the scope of the clearing system, including the clearing of many products with unique tail/jump to default risks that have not been cleared before, on the magnitude of the interdependence between the clearing system and the banking and payment systems.  The potential for operational and financial gridlock in the face of a substantial price shock will be greatly amplified if clearing is greatly expanded.

Bernanke goes on to argue that the systemic centrality of the clearing system means that it is highly desirable for the Fed to serve as the “insurer of last resort” to prevent the failure of a clearinghouse or clearinghouses.  So much for Gensler’s assertion that clearing would “greatly reduce . . . the need for future bailouts.”

This is very serious business.  Very serious.  It deserves serious consideration of the real implications of the effects of a vast expansion of clearing.  That consideration must be predicated on an understanding of the real interconnections inherent in a clearing system, not on unsupported and unsupportable denials of the existence of such interconnections.

If the basis for the policies Gensler advocates, and which Congress seems hell-bent on implementing, is a belief that clearing does not entail an intricate web of interconnections (and potentially fragile interconnections) among financial firms, then they are policies built on lies.  And all that a policy based on lies will do is sow the seeds for the next crisis.