Alfred E. Newman Chooses an FCM: The Moral Hazard of Segregation
The next brawl on the Frank-n-Dodd fight card is over segregation of customer funds in the clearing model. From the FT:
The “risk” that is in the spotlight centres on how customer assets, such as margins required to cover potential losses on derivatives trades, are held.
Under the current futures model, “futures clearing merchants” (FCMs), which handle cleared derivatives hold these assets in pooled accounts. If one of the FCMs’ customers defaults, and causes the FCM to go under too, then the clearing house can tap into this pooled account – also called an “omnibus” account – to cover any losses that are not met by margin payments or the clearing house’s default fund.
For investors these accounts are an important safety net. And they want the pooled accounts to be replaced by a system of separate, or segregated, accounts. Then, if another investor defaults, their assets are safe.
“It is an extreme event, but it is not impossible,” said Richard Prager at BlackRock at the meeting organised by the CFTC.
. . . .
If regulators opt for segregated accounts, clearing houses won’t be able to tap into these in the event of a default. Instead, clearers say the current system of using omnibus accounts gives them an extra shield. The removal of this shield would mean they would have to find funds elsewhere, potentially driving up clearing costs for users.
Kim Taylor from CME Clearinghouse said the removal of omnibus accounts could require a 60 to 100 per cent rise in the default fund, the capital put up by the backers of clearing houses. Such a rise in capital costs could lead banks to “think twice” about being in the clearing business, she said.
Point number 1 (which everybody saying anything about clearing should repeat until they internalize it): a primary effect of any priority rule–and the difference between segregation vs. omnibus accounts is first and foremost a difference in who is in what loss position in default–is to shift risk around. Under the omnibus model, the customers of an FCM are at risk to its other customers and the FCM itself. If some customers default, and the FCM does not have the resources to cover this, under the omnibus system the clearinghouse can seize the entire customer margin held in the omnibus account to make whole those at other FCMs (including the other FCMs themselves) who are owed money. This means that non-defaulting customer funds at the defaulting FCM can be used to cover the losses of the defaulting customers at that FCM (though not the losses on the FCM house account). Under segregation, this can’t happen. But that doesn’t mean the losses go away: they are borne by the other clearinghouse members who don’t default. These firms are obligated to pay up to make those owed money whole.
In other words, with an omnibus account system, customer default losses are shared among FCM customers and members of the CCP. Under segregation, they are shared only among members of the CCP. The rule shifts losses around.
Note that this is somewhat ironic if the policy objective of a CCP mandate is to reduce the counterparty risk losses suffered by the large, systemically important financial institutions that are among the members of CCPs. Segregation rules force said firms to bear larger default costs than omnibus rules: losses are shifted in their direction. (And compared to bilateral markets, it can also be the case that large systemically important financial institutions will bear larger default losses under clearing than without it.)
In other words, just looking at the risk piece, the choice between segregation and omnibus is a choice between different sharing rules. Default losses are actually spread more widely under the omnibus system, as non-defaulting customers are in a potential loss position.
But whenever you look at insurance arrangements, you have to consider the incentive effects, and most notably, the potential for moral hazard. The crucial thing to understand here is that the segregation system is more vulnerable to moral hazard than the omnibus system.
This is true for at least a couple of reasons.
First, customer incentives to monitor FCMs are stronger with omnibus accounts. A customer realizes that if the FCM doesn’t monitor its customers appropriately, or lets them assume excessively risky positions, he is at risk of loss from the FCM’s carelessness. He therefore has an incentive to choose an FCM carefully, and to monitor that FCM, and potentially move to another FCM (“exit”) or raise hell with his current one (“voice”) if he finds that FCM is slacking in watching its customers.
Second, under segregation, customers don’t care about the riskiness of the FCM’s trading or financing. Remember a customer is in a loss position under an omnibus system when (a) there is a default in some customer accounts, and (b) the FCM can’t cover this loss. Condition (b) is more likely to occur, the riskier the FCM’s own trading, financing, etc. Therefore, a customer is more likely shy away from a high flying FCM that trades aggressively, or levers up to the max, or chooses fragile forms of leverage when under an omnibus system. Under segregation, the customer channels Alfred E. Newman: “What, me worry?” He will choose an FCM only on the basis of commissions, services, the quality of the Christmas party and golf outings, etc. This means that under segregation, customer accounts will be at riskier FCMs, on average, than under omnibus accounts.
More moral hazard means that default losses will be greater on average under segregation. And since non-defaulting CCP members are on the hook for the entire default loss with segregation, this means that CCP members have more capital at risk under segregation than with an omnibus system. That’s why Kim Taylor concludes: “Such a rise in capital costs could lead banks to ‘think twice’ about being in the clearing business, she said.” A bigger share of a bigger amount translates into a lot more capital.
Given this, it’s no surprise that the omnibus account system has had survival value. It is a way of controlling moral hazard. It reduces the insurance benefits of clearing (from the customers’ perspective), but it reduces the costs of providing this insurance. As is the case with virtually every insurance system, including the most prosaic automobile policy, you need something like a deductible to mitigate moral hazard. Insurance is almost never complete due to the perverse effects of moral hazard. Omnibus accounts play the role of mitigating some moral hazards in clearing.
I would hope that regulators ask why the omnibus system evolved, survived, and dominated before prescribing the sharing rule that CCPs adopt. (I’ve told one regulator exactly that.) The survival principle tells you something. Before changing what is, ask why it is. Believe it or not, there is very likely to be a good reason for it. And with omnibus accounts, that reason is control of risk.
I have to quote just one more thing from the FT article:
But the issue highlights the complexities – and costs – of reforms.
Well, no sh*t. Some people (sounds of throat clearing) have been saying that for literally years. If only the Sorcerers’ Apprentices had appreciated that before passing Frank-n-Dodd.
You make an excellent point in relation to whom losses are to be allocated.
By the same token, there is a moral hazard in obliging clearing members to provide default fund contributions [which are subject to loss mutualisation] for the risks of their clients. In this setting, clients pay “less” for their contributed risk by accessing the CCP indirectly than they would by becoming direct members. As such, a number of issues present themselves
– banks are effectively subsidising clients and clients can attain greater leverage by only partly funding their CCP risk, which is not an outcome one would assume regulators had in mind;
– customers are incentivised to push for lower initial margin at a CCP because they don’t bear the consequences of failure via a default fund contribution, a point which is overlooked by regulators in the Governance discussions. Already CCPs are competing on initial margin to attract market share by increasing default funds to compensate for lower initial margin;
– barriers to client portability are created because in a time of stress following a member default, the last thing a clearing member wants is to have to fund significant additional default fund sums as a consequence of on-boarding clients from a defaulting member.
With OTC clearing also requiring deeper pockets in terms of capital required by clearing members to support the activity [much higher than for Futures activity], it is also odd that Dodd-Frank has pushed this activity out of well-capitalised banks into thinly capitalised FCMs. These increased capital costs, which are unique to the USA presently, will inevitably lead to higher clearing costs in that market.
Comment by John — October 28, 2010 @ 4:48 pm
John–Thanks for your comment. I have an entire paper going into gruesome detail on exactly what you describe. It’s titled “Mutualization of Default Risk, Fungibility, and Moral Hazard: The Economics of Default Risk Sharing in Bilateral and Cleared Markets.” I can send you a copy if you’d like.
Thanks for your insightful comment. (All comments that agree with what I’ve written are self-evidently insightful 🙂
Interesting take. You give a little too much credence to customers ability to monitor risk at the FCM – although this “exit” was quite literally used by many when trading was floor-based, many accounts had the bulk of available cash withdrawn by check at any major threat, often by noon and certainly by 2:30. Usually this was on market related events such as a crash, but often it was due to a perceived threat from another customer. Fact is though, even then, this was an exit only available to the connected and the savvy. Not all were aware of threats and fewer still had the political pull at the FCM to demand, and receive, their funds immediately. Most would simply get a promise, a hard sell to relax or a threat to close the account with prejudice – something which negatively impacts the ability to open a new account elsewhere.
Today, one can move an account readily – but most have no clue as to what’s going on around them within the customer base, rumors of events leak slowly and often with vague and uncertain details. So yes, there once was a good reason for it, and why it worked. Those reasons are no longer valid.
What is true is that FCM’s are positioned to police each other – even if the customer can no longer do this. SEG accounts will, with absolute certainty, focus their attention on clearing risk – while defending the %99.9999 of the accounts utterly at their mercy who never stood a single chance to utilize any of the protections you argue have always been at their disposal.
Comment by JM — November 1, 2010 @ 2:39 pm
Hey The Professor,
I am a finance student from Berlin, Germany, and really interested in your paper. Could you send me a copy of it?
Thank you in advance!
Jakob
Comment by Jakob — November 5, 2010 @ 8:20 am
Hi, Jakob. Will do. I will send along a couple of other related things you might find of interest.
Hi Professor
I found your article fascinating. It’s interesting to see the other side of the ‘omnibus vs segregation’ argument. I also think that John had some very insightful thoughts. Would it be possible for me to get a copy of the paper you mention as well?
All be best.
Michael.
Comment by Michael — April 28, 2011 @ 8:00 am
[…] wrote about this last year in a post titled “Alfred E. Newman Chooses an FCM: the Moral Hazard of Segregation.” I argued that it is necessary to understand why futures exchanges would choose–and […]
Pingback by Streetwise Professor » Alfred E. Newman Wins! — June 15, 2011 @ 12:07 pm
Well, this argument of moral hazard sounds good but in theory, but it would be very hard in practical terms to determine if your FCM was a risky operation. There reports are incomplete and confusing. MF Global is a case in point. There is a little hope that the average customer could have known about the extent of risk being assumed by management. Ask an MF Global custoemer if they don’t wish for segregated accounts.
Comment by Stephen Kennedy — April 12, 2012 @ 4:15 pm