Pick Your Poison
This afternoon I attended the Working Group on Financial Markets at the Chicago Fed. The scintillating topic was segregation models. Really! It was scintillating! For a certain kind of geek, of which I am one.
It was scary, actually.
Segregation of customer money has been a subject of discussion post-Frankendodd, and I have written several posts on the subject. Segregation has received even more discussion post-MF Global (AKA Corzine-gate, but since he’s a made man I guess he will skate). The model that has been implemented for cleared swaps, and which (according to clearing maven John McPartland of the Chicago Fed) is likely to be adopted for cleared futures is LSOC-legally segregated, operationally commingled.
Back in the summer of 2011, I wrote that segregation could make the markets more fragile, because it would tend to reduce credit (mainly intraday credit) used to finance variation margin. This is important, because the markets depend on using credit to fund margin payments. If this credit freezes up, the markets will freeze up. Indeed, a cleared system works on such tight deadlines that an interruption of credit can be catastrophic. If margin calls aren’t met in a timely fashion, absent credit, margin payments don’t flow to those on the winning side of trades. When this happens, the clearing system breaks down. And in a Frankendodd world dependent on clearing, a breakdown in clearing means a meltdown in the markets.
Indeed, even delays of hours or even minutes in making margin payouts, or doubts that CCPs will make margin payments in a timely fashion, can be catastrophic. Almost exactly 25 years ago, on 19 October, 1987, the mere rumor that the CME would not pay out variation margin led to a run on FCMs and the CME clearinghouse that almost brought the market to a crashing halt.
One of the speakers at today’s event, Barclays’ Kevin Murphy, noted that under segregated models FCMs don’t have a lien on the collateral in customer accounts. Which means they won’t extend credit to customers because there is no collateral backing the loans. Murphy said that broker intraday credit is likely to be a thing of the past under greater segregation.
Think of the consequences on a day when markets move a lot. Those on the winning side are expecting to receive variation margin payments. Those on the losing side will be scrambling for cash to meet their VM obligations. Where will they get it? Not from their FCMs. From their banks? Uncertain-even if the customers arrange credit lines, banks can often find reasons to delay providing the credit, or not providing it at all.
This all means that there is a risk that VM owes will not be paid in time. With no credit being extended, or the amount of credit being sharply constrained, there is a serious risk that VM pays will not be made on time. If that happens, particularly during a period of market stress, all bets are off. Almost literally. People will fear that CCPs are insolvent, and there will be runs on them-people will liquidate positions to recover their margin money.
That would be very ugly indeed. Again, a cleared system is very tightly coupled, more tightly coupled than OTC markets. Tightly coupled systems are more prone to going non-linear, and failing catastrophically. Segregation increases the tightness of the coupling. Connect the dots.
Of course, markets don’t stand still. People will recognize the need for contingent liquidity, and make arrangements for it. But will those arrangements be as robust as the system we have now? There is serious room for doubt. If more robust alternatives are available, why weren’t they chosen before? I know this argument is not dispositive, given the coordination issues involved, but there are strong incentives to adopt better systems. We should be very leery of changes that implicitly assume that market participants with large amounts of money and risk in the game systematically choose wrong.
The whole move to clearing has been intended to wring credit risk from the derivatives markets. But every move to reduce credit risk-including moves to greater segregation-almost always involve creating greater liquidity risk.
If given a choice between credit and liquidity risk, I would choose credit risk every time .
In other words, you have to pick your poison: credit risk or liquidity risk. Yeah, credit risk may be like arsenic. But liquidity risk is more like cyanide.
I know which one I would choose.
I need to think about this a lot. (I know that’s not much of a comment, but there it is.) The problem for me is that the solution for a systemic credit event is essentially the same as the solution to a systemic liquidity event — Too Big Too Fail. I can choose my poison, but I’ll be given the same antidote.
Comment by Highgamma — October 16, 2012 @ 4:09 pm
Take your time: I’d be very interested in hearing your thoughts, Highgamma. You have put your finger on the fundamental trade-off.
Reading Gorton’s book on financial crises. Good historical perspective that helps understand the trade-off better.
[…] Segregating client money, it’s a decision. So, cyanide or […]
Pingback by Further reading | FT Alphaville — October 17, 2012 @ 12:35 am
You ignore the diverisifcation margin benefit that the FCMs have.
The total margin the FCM collects from its customers will practically always be more than it has to post to the exchange. This happens because some FCM customers’ positions will offset and so the FCM can net those out before it pays margin to the exchange.
Thus a large and so systematically important FCM will have a buffer (in addition to its own capital) to fund margin calls on volatile days.
Comment by mike — October 17, 2012 @ 3:04 am
I thought under LSOC,they were allowing for variation margin netting (though not initial margin netting). The liquidity risk comes from that requirement that changes in IM cannot be used to fund VM. Ie if you reduce your positions, creating an IM surplus on a customer account that money cannot now be used to fund VM. A seperate cash call has to be made resulting in a short term large potential liquidity requirement, between the funding of VM and withdrawal of IM.
Comment by EnquiringMind — October 17, 2012 @ 3:56 am
No, the choice is not between liquidity and credit. The choice is between liquidity and integrity. A relaxed segregation requirement will only start by providing liquidity. But things will change with time. If FCMs are allowed to take out loans for one purpose, they will probably take it out for other purposes–money is fungible, y’know. It’s like the distinction between hedging and speculating: almost impossible analytically, except in polar cases. Legal opinions will be forthcoming, and practice will erode. After awhile, they’ll treat their margin accounts like banks treat deposits. And the liquidity–as well as the credit–will be gone.
Comment by Ebenezer Scrooge — October 17, 2012 @ 5:51 am
Found Poetry:
“AKA Corzine-gate,
but since he’s a made man
I guess he will skate”
Comment by Jason L — October 17, 2012 @ 8:56 am
Not to be a cynic, but the intent of this model is to make it difficult for customers to post the required VM in times of stress. And at the point the FED will have no choice but to step in and provide the funds.
The 1987 incident to which you make reference was minimized because a major clearing firm, First Options, was owned by Continental Illinois Bank which in turn was owned by the US Government. So CINB had no problem meeting First Options margin calls. I’ve heard rumors that First Opt was able to lend money other clearing firms by executing conversions, reverse conversions and what not at very favorable levels.
Again, that was the decision of Ronald Reagan’s Treasury Department to “save” the markets. We are still paying the price for that decision.
Comment by Barry — October 17, 2012 @ 5:13 pm