Streetwise Professor

Ironically, in financial markets positive feedback loops are typically a very negative thing.  Self-reinforcing spirals in tightly coupled systems like the financial markets are usually downward spirals.

Margining and collateral mechanisms are the most important positive feedback mechanism in financial markets.  AIG and MF Global are two examples of how these mechanisms turn high flying financial firms into smoking holes in the ground.  They suffer losses, counterparties demand more collateral, that puts further stress on the firm, they lose customers, counterparties demand even more capital, and before you know it–impact with the ground at high speed.

Procyclicality of margin can work at the level of the market too.  We are seeing the first signs of that with respect to Italy today.  LCH.Clearnet SA (the French arm, and not RepoClear) raised margins on Italian government debt today.  This will tend to induce some sales of leveraged Italian bond positions, which will reduce prices further.  Similarly, increasing haircuts will tend to reduce liquidity, which will also tend to put downward pressure on prices.  Lower prices and lower liquidity put further pressure to raise haircuts.

So–lookout below!  The spiral has begun.  Keep eyes on RepoClear.  I imagine it will be under tremendous pressure to avoid jacking up repo margins.  And who knows, since the 450 basis point  margin raising trigger* is based on the spread over an average of German, French, and some other countries’ debt, maybe RepoClear won’t have to raise margins on Italian bonds because French yields are rising too, due in large part to French banks’ holding of $400 billion in Italian bonds.  The looking glass world of Europe . . . .

Another comment, following on the previous post.  There will be tremendous pressure on CCPs to accept riskier collateral in the clearing mandate world.  Sovereign debt has been a primary asset for margining purposes, and the assumption has been that this collateral is low risk.  But we now know that is a fiction.

Pressure to expand collateral pools, combined with the risk that said collateral can decline in value, and precipitously, creates another positive feedback mechanism.  Markdowns in collateral held at CCPs (or indirectly, via collateral transformation services) will induce fire sale liquidations of margined positions, or of other assets in order to raise cash to offset the decline in the value of collateral.

Geithner, Gensler and others have touted collateral as the elixir for all derivatives market ills.  No.  In some circumstances the elixir is really a deadly poison.  Positive feedback mechanisms destabilize financial markets, and collateral mechanisms are among the most important sources of positive feedback in these markets.  We are moving to a regime in which more collateral is being required, collateral mechanisms are being made more rigid and less flexible (e.g., reducing hypothecation) but the supply of safe collateral is contracting–or at least not growing commensurately with the demand.  This will be a major source of fragility in the system, and one that has been built into the system all in the name of making it more robust.  Another gift from the sorcerer’s apprentices.

* The term “trigger” is hardly comforting in this context.

Last week the Basel Committee released its revised proposal on capital requirements for exposures to CCPs.  It made a few tweaks to its earlier proposal, which had attracted considerable criticism.  The critics are not mollified, and some new criticisms have been directed at the new version.

I won’t deconstruct the new proposal in detail.  I’ll just focus on a few of the high level issues, all in the way of reiterating a broader point that I’ve raised previously.

One criticism of the new proposal is that it imposes punitive burdens on client clearing, and is therefore inconsistent with the stated goal of encouraging the movement of more derivatives transactions to CCPs.  Against this, it is claimed that the new rules facilitate portability of client accounts between clearing firms.  So, you won’t be able to find someone willing to clear for you in the first place, but you’ll be able to move your account easily.  Or something.

The allegedly punitive treatment of client clearing, combined with the capital and margin burdens placed on non-cleared trades, will tend to reduce end user hedging with derivatives.  This is clearly an undesirable outcome, and will induce end users to devise alternative means (and in particular, to adjust their capital structures) in order to achieve the same objectives as they currently do through derivatives.  These alternatives  are likely to be more costly than derivatives.  Moreover, they will still result in the creation of counterparty risk (or credit risk).  From a systemic perspective, it is not clear that the ultimate effect will be to reduce the amount of credit risk, or improve its allocation.

The new draft reduces slightly the capital charges for default fund contributions, but the treatment is still sufficiently tough to make it likely that CCPs will jack up margins to avoid incurring the high capital charges associated with exposures in excess of the CCPs “hypothetical capital.”  Maybe this is precisely the intent, because the idea that derivatives trades should be massively collateralized is a matter of faith among many regulators.  But demanding high collateral can lead to many adverse effects, including less hedging, the utilization of riskier assets to pay initial margins, the substitution of other forms of leverage, including fragile repo-based collateral transformation services, and the engineering of securities that can be used as collateral (either directly with a CCP or indirectly via a collateral transformation) but which are loaded with systematic tail risk.

The proposals retain the very crude distinctions between risk exposures.  Capital charges across exposures do not vary sufficiently to account for their different risk characteristics.

All of these issues point out the basic problem with Basel capital requirements.  These rules are essentially price controls.  They cannot–metaphysically–get the prices right.  The way the risks are priced via capital charges will affect resource and risk allocation decisions.  Those subject to the rules will have the information and incentive to figure out how to minimize their impact, and will tend to load up on risks that are underpriced and take too few of the overpriced risks.  Many of the endogenous responses to the capital charges will occur in unexpected ways and places, and the connection between these responses and the rules will often be highly indirect and difficult to identify.

But the big problem is that since everybody is facing the same distorted price signals, they will tend to make correlated mistakes.  This increases the vulnerability of the system.  That’s why Basel is faulty generally.  The CCP capital charges are just a particular example of that.  They are intended to reduce systemic risk, but they largely ignore how the system will respond to the rules, thereby creating new vulnerabilities.

Just step back and look at what is going on now in Europe.  One of the basic principles of Basel II was that sovereign debt was safe.  Another was that mortgages were safe.  Financial Crisis I exploded the latter belief.  Financial Crisis II is demolishing the former.  But the favorable capital charges accorded these investments encouraged the supply and demand for both.

If you haven’t noticed, this is not working out well.  The CCP capital charges will have far reaching effects on how much risk there is in the system, and where it resides. The necessary crudeness of the capital rules, their inability to price different risks in a discriminating way, and the endogenous behavioral responses that they induce not just in clearing but throughout the interconnected financial markets, will create new systematic errors and vulnerabilities that won’t work out to well either.

The conceit of these rules is that they are pricing systemic risk.  How is that even possible when those formulating rules do not–and realistically, cannot–consider fully the systemic responses to them?  Charging a price can make things worse if the price is wrong.  A financial Gosplan hardly inspires confidence, and methinks that’s exactly what Basel is.

As usual, Deus Ex Macchiato is worth reading:

One difficult criticism for those of us who believe that strong financial regulation is necessary – even if we don’t believe that the current structure is heading in the right direction – is that of rent seeking. The issue is that the stronger regulatory power is, the more it can be exploited. For me, the clearing provisions in Dodd Frank are a great example of this. The clearinghouses have been handed juicy profits by regulatory fiat.

Read the rest.  That’s an order.

Gretchen Morgenson is getting good reviews for her book Reckless Endangerment, which unlike most of the financial crisis genre casts considerable blame on Fannie and Freddie.  So bravo for that.

But sometimes she can be so clueless.  I mean really: how can she not understand the difference between credit default swaps (“complex swap deals it had struck with its trading partners”) and repo?  She goes on and on comparing MF Global to AIG, claiming both were brought down by CDS.  Then she goes into a riff on CDS written on Greece, and credit linked notes, and ISDA, and blah blah blah.

Gretchen: secured lending–repo–brought down MFG.  You could have written something interesting about the repo market, how it works, how it is a fragile form of financing that can bring down firms.  That would have been informative and useful.  But to flog the CDS horse when CDS aren’t even involved.  Sheesh.

David at Deus ex Machiatto turns his discerning and well-trained eye to the MF mess.  He basically concurs with my initial diagnosis that what brought down MF was increased haircuts on its repo trades.  He concludes:

It seems much more likely to me that Corzine et al. didn’t understand the repo haircut risk – a modern phenomenon – than that they were rogues. Or, to use my own slang, this is a cockup not a conspiracy.

This would indeed be a cockup of massive proportions, and an ironic one to boot.  Adjustment of repo haircuts is analogous to marking-to-market and variation margining in the futures world–which one of the world’s largest futures brokerages deals with on a daily basis.  Did they think that this wasn’t an issue in repo?  The very newness of the phenomenon should have put them on warning.

As David explains in his excellent and very readable book Unravelling the Credit Crunch, mark-to-market gives rise to risks throughout the life of the trade.  Included in these risks are risks to fundamentals–in this case, changes in the likelihood of default on Italian, Portuguese, etc. bonds.  But they also include changes in risk tolerance/risk aversion–the market price of default risk.  They also include liquidity risks: market prices change with liquidity conditions as well as fundamental conditions.  In a mark-to-market mechanism, these risks combine to give rise to funding risk: you have to be able to access liquidity in order to make the cash payments obligations caused by changes in market prices driven by any or all of these factors.

These considerations are second nature–hell, fifth nature–to anyone who trades futures with their rigorous mark-to-market and variation margining mechanism.  That is MF’s bread-and-butter, so it should have been particularly conscious of this issue.  Any trading strategy has to be made, and funding for it provided, based on an assessment and understanding of these risks.  Did Corzine really, truly not understand this was relevant for the repo trades in question?  That would be negligent beyond belief.  At the very least, he should have asked the question:  What are the funding risks over the life of this trade?  You should ask that about any and every trade you make.

In other words, Corzine’s trading strategy was either grossly negligent in its failure to take into account important risks (which seems to be David’s interpretation), or extremely reckless in its disregard thereof.

Cockup would be an understated description if Corzine did not ask this question, and take these risks into account.  Or perhaps he did, but thought they were worth running.  No, not a conspiracy, but an indication of a substantial willingness to take risk and/or a colossal deprecation of those risks.

To put it in NavSlang: the MFers really screwed the pooch.  $600+ million in customer funds are missing.  That represents about 11 percent of the segregated customer funds.  It will be fun and games for customers to recover their money.  Matt Leising reports that customers may wait years to recover if MF Global is sued, as is almost inevitable.

New emerging details suggest that right after waving goodbye to CME auditors last week, MF started dipping into the customer kitty.  ICE states that there was no segregation breach.  Perhaps ICE was going to audit this week.

Not surprisingly, the CME and CFTC are coming under considerable criticism for the MF collapse.  One suggestion in this article by Greg Meyer is that the CME Group is lax in its oversight because MF was one of its biggest sources of order flow, thereby presenting the exchange with a conflict of interest.

I disagree with this interpretation.  I’ve written a lot about self-regulation over the years, and the theme of that research is that self-regulation is well-suited for some things because incentives are strong and well-aligned, and ill-suited in others for which incentives are weak or poorly-aligned.  Market manipulation is the main example of the latter.  (I have an article in the JLE from 1995 that examines this in detail.)  Especially for a for-profit investor owned exchange, incentives for broker oversight are stronger.  Having a broker go bust, especially as ugly as MF went bust, is not good for business because it casts doubt on the reliability of FCMs generally.   Lower confidence in the reliability of brokers and the safety of customer funds reduces the demand to trade.  Yes, MF accounted for a decent amount of order flow, but there is vigorous competition to serve that flow, and it is in CME Group’s interest that it be served by reliable and trustworthy FCMs.  There is no upside for CME from lax enforcement here, and a whole lot of downside–which it is experiencing now.

There is no such thing as perfect, error free, real time auditing.  Moreover, regulators had been pressuring MF Global for some time, so it’s not like the firm completely escaped scrutiny.  The issue is that when firms like this go down, it tends to happen very rapidly.  They are inherently fragile.  Once the run begins, it tends to end rather quickly.  In MF’s case, it lost funding when haircuts on its repo’d Eurodebt went up because of (a) the decline in market value of the debt, and (b) its rating downgrade–which was in fact in large part a response to regulatory action.  The end came in a matter of days.

Another theme is that there was a breakdown in controls at MF Global.  Uhm, controls are intended primarily to keep underlings–like the Delta One trader at UBS–in line: they aren’t designed to, and are unlikely to, stop a determined CEO from taking on excessive risk.  That’s especially true when the CEO is somebody like Corzine.  (Who is, by the way, a disgrace to my alma mater.)  It’s the board that has to keep the CEO in line, and all too often the board is like the CEO’s petting zoo.  That appears to be the case here.

One last thing.  I noted that the collapse of MF makes the CFTC look foolish not because of its failure to oversee the firm, but because of its adoption of a rule regarding clearinghouse membership that MF lobbied for very hard.   (The NY Fed also doesn’t look too swift for approving MF as a primary dealer.)  Another big supporter of the $50 million capital rule adopted by the CFTC was Jefferies Group.  Jefferies was downgraded by a non-big 3 rating agency (Egan-Jones).  The company’s stock fell about 15 percent on the news; it has since recovered on positive comments from Meredith Whitney on CNBC.  This bears watching, and again highlights the tension between the goal of making CCPs safe and allowing easy access to CCPs.

*Obscure Supersuckers reference.  A h/t to the first to decode it.

FCM, investment bank, primary dealer, and wanna be Goldman MF Global declared bankruptcy today.  As is typical with financial firms, the end came quickly.  The firm’s problems metastasized quickly last week.  Another quarterly loss, a ratings downgrade, and worries about losses on trades in European government bonds led to the typical downward spiral of lost funding and lost customers.  These firms are very fragile.  When they fall, they usually don’t get up.

Ironically, MF’s FCM operation was acquired from  . . . REFCO, which cratered almost exactly 6 years ago, in October, 2005.  Will anyone dare to take on this cursed franchise?

Piecing things together, I surmise that the firm bought about $6 billion in Italian and other southern Euro bonds and repo’d them out.  Due to the decline in the prices of these bonds, and the firm’s deteriorating financial condition, the repo counterparties demanded higher haircuts.  The firm couldn’t come up with the cash, and in desperation maxed out its credit lines.  But that couldn’t stop the hemorrhaging.

The most likely explanation for all this is that the firm was already foundering, and CEO John Corzine tried to gamble on resurrection by ramping up the risk.  As is so often the case, this more often results in a more rapid descent to financial hell than resurrection.

One major irony is that MF Global was pushing hard for low capital requirements for members of OTC derivative CCPs.  How’s that idea looking now, GiGi?

Especially since suspicions are rife that MF has misused segregated customer funds, presumably to keep the resurrection gamble going.  Reports state that the firm is stonewalling regulators on turning over records, and that about $300 million in customer funds are missing.  This is one of the most egregious things a brokerage firm can do.

This is already a major fall for former Goldman CEO, US senator, and NJ governor Corzine.  He is also a major Obama fundraiser, who is presumably now a Nonperson.  If the suspicions about misuse of customer funds prove true, major fall won’t even come close to describing what lies in store for Mr. Corzine.  He will long to be merely a Nonperson.

Gillian Tett of the Financial Times is shocked! shocked! to learn that Frankendodd: (a) has dramatically increased the complexity of financial markets; (b) creates myriad opportunities for regulatory arbitrage; and (c) gives advantages to big, sophisticated, politically juiced players that can decode, game, and influence the regulatory process most effectively.

Yeah.  I can sympathize.  It’s not like anybody could have seen that coming.  Nosiree.

Starting with my earliest post-crisis writings on clearing, I have pointed out repeatedly that the supposed virtues of clearing–netting and greater collateralization–look far different when you examine the nexus of financial contracts as a whole, rather than focusing on derivatives alone.  A primary effect of netting and greater collateralization  is to increase the seniority of derivatives in bankruptcy.  But if you increase the seniority of one claim, the seniority of other claims is reduced.  Sure, you can make derivatives more secure, but at a cost of making other claims less secure.

The systemic implications of this are ambiguous.  The implicit presumption seems to be that derivatives are somehow uniquely threatening to the stability of the financial system.  I say “implicit” because the clearing pom-pom squad has never discussed explicitly the distributive aspects of reshuffling priority via regulation, so I don’t even know if they are even aware of it.

But it is not hard to imagine scenarios in which giving derivatives even greater advantages in bankruptcy (due to safe-harbor provisions they already have a lot) can cause a systemic problem, but through another channel.  For instance, derivatives counterparties of a firm teetering on the brink of failure may rest easy given their priority status, especially if the CCP is considered safe.  But other claimants may not be so serene.  For instance, consider money market mutual funds  that invest in the paper of the dodgy firm.  Realizing that they will be totally hammered in bankruptcy, the investors in the money market fund may run.  That would not be pretty.

It’s even more complicated than that, because given the new seniority rules, there is likely to be widespread alterations in capital structures of derivatives market participants, and repricing of those claims.  Who knows how that is going to play out, and what new fragilities will be baked into the new structures?

This point is finally dawning on people.  One person who caught on long ago is Harvard’s Mark Roe.  Today, he published an oped that makes the point clearly and forcefully. It is definitely worth a read.

Money quotes:

Equally problematic is the fact that, while the clearinghouse and its participants can net wins and losses to reduce risk for those inside the clearinghouse, the clearinghouse does not assuredly eliminate the basic risk facing the entire financial system. Often, it merely transfers that risk to creditors outside the clearinghouse.

. . . .

Whether the clearinghouse reduces systemic risk depends on the relative systemic importance of those inside and those outside the clearinghouse – AIG versus Citibank in this basic example – not on the clearinghouse’s capacity to reduce risk among its members. In this example, if Citibank is precarious and is as systemically vital as AIG, the clearinghouse has obscured that it has saved AIG only by transferring risk from the clearinghouse to Citibank, which then fails.

Much recent regulatory activity has focused on enabling, enhancing, and requiring clearinghouses for these kinds of financing arrangements. Yes, clearinghouses offer many benefits, including greater transparency, better pricing, and better regulatory focus, and we should try to make them viable. But regulators world have overestimated their overall benefit. Too much of what is justified as reducing systemic risk is really just offloading risk onto others.

Exactly.

Reading Mark’s piece  brought to mind Bernanke’s quoting of Puddin’ Head Wilson in his discussion of CCPs.  Bernanke said if you are going to put all your eggs in one basket, keep a very close eye on that basket.

One of my problems with the whole CCP debate is that too often the tendency is to treat derivatives as the only eggs that matter.  But if you pay attention only to the basket with the derivatives eggs, you will not pay attention to other ones that are as valuable, or moreso.

To change metaphors, target fixation all too often leads one to miss the bogey that just jumped your six, with fatal results.  The obsessive focus on derivatives, and clearing, is a form of target fixation.  It risks diverting attention from other potentially mortal dangers. You need to have a system-wide perspective to truly understand systemic risk.

Mark is generally negative on giving derivatives any priority.  He has written things very critical of safe harbors.  He and I have discussed this, and have agreed to disagree: I can see reasons for differentiating the treatment of derivatives from other claims.  But the important point is to understand that this is an issue, and one that has to be examined from a system-wide perspective, because seniority rules shift risk around.

Mark gets that, and makes serious arguments about what the right rules should be.  Unfortunately, all too many of those who make the decisions and the rules don’t appear to get it.

Like Tim Worstall, I am puzzled at all of the machinations in Europe to avoid a hard Greek default in order to avoid triggering CDS.  Apparently the concern is that payoffs on CDS would potentially jeopardize the financial health of banks that sold protection, leading to contagion.

Greek CDS represents about 1 percent of debt outstanding, with a notional amount outstanding of about $4 billion.

Look.  If European banks collectively cannot handle a $4 billion payout (which would be the maximum, assuming zero recovery), then their problems are even worse than feared.

To worry about the $4 billion CDS, rather than the $471 billion in outstanding Greek government debt, is beyond bizarre.  This suggests some weird sort of obsession among the Euros about derivatives, and CDS in particular.  You don’t have to look far for more evidence of this obsession.

And remember, CDS are zero sum.  Ever dollar paid by one party is received by the counterparty.

Moreover, many of the CDS contracts will have been entered by banks looking to hedge exposure to Greek sovereign debt, and private Greek debt.   Restructuring Greek debt in a way that leads to substantial writedowns but does not trigger CDS screws the hedgers worst: they suffer losses on their bonds, and their hedges do them no good whatsoever.

Punish the hedgers.  Wow.  That’s a really constructive policy.

And there are long-term consequences.  In particular, the effective abrogation of Greek sovereign CDS contracts will raise questions about the reliability of other outstanding sov CDS.  Many hedges are looking far shakier now, for if the Euros do backflips to avoid triggering Greek CDS payouts, just think of what they would do for Italy or Spain.  Since it is the most risk averse banks that would have decided to hedge, and since risk aversion depends in large part on the bank’s financial condition, this hurts the most risk averse (and likely weakest) banks most.  Less hedging will be done going forward, but risks will be transferred regardless–through the cash markets.  CDS prices will become less reliable (probably a feature rather than a bug to the Eurocrats).

Ironically, this creates a new speculative trade–a trade on the basis between CDS and the bonds–in which the parties effectively speculate on the political risk inherent in CDS.

There will be knock-on effects too.  If private contracts can be circumvented for political reasons in the case of sovereign CDS, the risk that other derivative contracts can be effectively nullified increases too.  The risk is probably greatest for other CDS–notably European bank CDS–but it exists for other types of derivatives too.  All because once you admit the principle that “payouts on derivatives contracts can spread contagion”, you can’t limit its application–or the risk that the principle will be applied–only to Greek CDS.  Just this once.  Pinky swear.

Looking at the whole European fiasco brings to mind Churchill’s quote about the United States always doing the right thing, after trying everything else first.  The Euros try about everything, but they don’t look like they are going to the right thing.  The exact opposite in fact.