Sunday, November 30, 2008

My "non-testimony" on the regulation of swaps and derivatives

I was recently asked to testify in a Senate hearing "to explore the role of financial derivatives in the current financial crisis, the current system for regulating them and recommendations for modifying the regulatory system".

This hearing was by the Committee on Agriculture, Nutrition, and Forestry, chaired by Senator Harkin of Iowa. That might seem like an unusual place to deal with the regulation of esoteric financial products, but this committee has oversight for the CFTC, because the first futures contracts were agricultural, and the CFTC has oversight on swaps and derivatives, because they have characteristics similar to futures.

I was not able to accept the invitation to testify at the hearing, but it was still worth it, because in his letter, Senator Harkin explained that "You were strongly recommended for your knowledge and insights on this subject by Warren Buffett in a conversation I had with him yesterday". That alone makes the letter worth framing. But although I could not testify, I exercised my first-amendment right to send along off-the-record comments (you can do it too). I will give excerpts from these comments here:

In academic theory, derivatives and swaps are intended to help "span the state space"; that is, they are intended to allow investors to efficiently hedge their specific risks, to more precisely meet contingencies of the market, and to mold their returns to meet investment objectives.

In practice, however, swaps and derivatives are often used for less lofty purposes. They are used to: avoid taxes (for example, total return swaps are used to take positions in UK stocks in order to avoid transactions-based taxes); take exposures that are not permitted in a particular investment charter (for example, index amortizing swaps were used by insurance companies to take mortgage risk); speculate (for example, the main use of CDSs is to allow traders to take short positions on corporate bonds); lever beyond an allowed level; and take risk off-balance sheet, where it is not as readily observed and monitored.

The more complex swaps and derivatives not only find ready demand by serving these functions, there are strong profit incentives for the investment banks to supply them. The more complex the instrument, the greater the chance the investment bank can price it to make profit, for the simple reason that investors will not be able to readily determine its fair value. And if they create a product that is exclusive to them, they can also charge a higher spread when an investor wants to trade out of it.

Viewed in an uncharitable light, derivatives and swaps can be thought of as vehicles for gambling; they are, after all, side bets on the market. But unlike the more common modes of gambling, these side bets can pose risks that extend beyond losses to the person making the bet. There are a number of ways the swaps and derivatives end up affecting the market:


Those who create these products need to hedge in the market, so their creation leads to a direct affect on the market.

Those who buy these instruments have other market exposures, so that if they are adversely affected by the swaps or derivatives, they might be forced to liquidate other positions, thereby transmitting a dislocation into other markets.

The value of some derivatives can have real effects for a company. For example, the credit default swaps are used as the basis for triggering debt covenants, so if the swap spread rises above a critical level, it can have an adverse effect on the company.

The use of derivatives and swaps can pull capital away from other productive uses. Because these are side bets, capital employed in these markets does not find an end-user.

Those who are writing the derivatives are in effect providing insurance to the buyers, but without any regulatory requirements. Often those writing these instruments are not in a well-capitalized position to pay out in the event that the option goes into the money.


In terms of regulation, here are some points to consider:

The regulators must know who owes what to whom in these markets. Right now there is no way to ascertain the effect on the swap and derivatives markets of a firm failing, because regulators do not know the web of counterparties to these instruments.

We should consider standardization of instruments and make the instruments simpler. As I point out in my book, complexity of financial instruments is one of the sources of market crisis.

We should consider having swaps and derivatives move as much as possible away from party-to-party into a clearing house. This would improve monitoring and provide for a lower risk of default.

There should be stronger oversight on the use of these instruments. Who is buying and selling them,and for what purpose. Perhaps require participants to specify if the instrument is being used for speculation or hedging. Regulators should also ascertain if the use is reasonable; are these instruments simply being used to move risk off-balance sheet or to take on positions that would have been prohibited if they were executed in another way.


[I have added responses to some of the comments]

17 comments:

  1. Rick, I bought your book almost a year ago. In Michael Lewis's recent article in the New York Times, he argues that the short selling of synthetic securitized mortgage bonds dramatically increased the impact of the failing subprime market.

    Felix Salmon on Looking for Alpha countered that Lewis was wrong because there wasn't that much demand for to short these bonds.

    Who do you think is right, if anyone?

    ReplyDelete
  2. Right. What European "counterparts" should do then in this business? I still contend that it's American made although of a European origin or where European had a kind of competitive advantage in early 2000.
    Should we "cancel" or regulate CDS and similar types of contracts at world level?
    Derivatives in Europe?

    ReplyDelete
  3. Rick,

    All excellent points, but I would add another.

    There is an important distinction between CDS written on top (supersenior) tranches, and those written on lower tranches. The former are essentially insurance against systemic risk, which no private party can credibly offer. As we are seeing (AIG), only the government can pay out on such insurance contracts. I conclude that only the government should be writing such insurance contracts, and the price charged can and should be a policy variable. The origin of the crisis was (in part) mispricing of this kind of insurance. The origin of the continued freeze is (in part) the absence of any credible counterparty writing such insurance.

    ReplyDelete
  4. I find it interesting, also, that the invitation came from Senator Harkin. Sen Harking cosponsored the Commodity Futures Modernization Act of 2000. It was this act that insured that swaps were not to be regulated, but were simply agreements between two entities.

    Perhaps Senator Harkin should have had this hearing before sponsoring the legislation that led us into this mess. Had I been invited to testify, I would have asked Tom Harkin if he could share the names of the individuals responsible for writing the legislation. I wonder if they represent and of the organizations currently receiving bailout money.

    ReplyDelete
  5. Is it true or not that AIG is "covering" and "insuring" primary european banks? That's why it was also saved?

    ReplyDelete
  6. As someone who ran a derivatives market-making operation for many years, I can attest to the accuracy of Mr. Bookstaber's observations. Far too much of the derivatives business is run as a kind of honey-pot for sharp dealing, be it tax avoidance, income statement manipulation, or good old-fashioned customer scalping. What was fascinating about Mr. Greenspan's blindspot was that he seemed to think 'proprietary financial models' were somehow worthy of protection from prying regulatory eyes. That was just bizarre.

    And Perry Mehrling's comment about the inappropriateness for private enterprise to sell 'insurance on systemic risk' will require some thought. Not sure I agree since that could have implications for sellers of literal catastrophe insurance too, and one could argue the merits of having the government happily sell insurance that encourages people to, say, build more houses on low ground in New Orleans. How about just actually regulating insurance sales, including CDSs, as they should have been in the first place?

    As to Senator Harkins' role then and now, we can only hope our leaders will be gracious enough to genuinely review earlier actions, eh? Certainly the self-serving behavior of Bob Rubin and Alan Greenspan nowadays is all too common and unproductive.

    Who testified before Harkins for the CFMA 2000?
    Well, for starters try ISDA, ISDA, ISDA.

    ReplyDelete
  7. Rick, not sure if you read the blog Information Arbitrage (www.informationarbitrage.com), but the tenor of your comments couldn't more closely match my own. Transparency. Simplicity. Centralized collateral management. Pushing the lion's share of the OTC derivatives market onto exchanges. This is the only way we can manage what has become a mind-bogglingly large, complex and interconnected series of obligations.

    Roger

    ReplyDelete
  8. Transparency and disclosures to reduce any asymmetry of information. Moreover there should be some central clearinghouse and standardization of contracts and markets' rules to avoid differences between countries.

    ReplyDelete
  9. Gentlemutt,

    It is of course possible that the government will misprice its credit insurance, just as (in retrospect) AIG did. If we can't get the price approximately right, we might go for regulation instead as second best. (Probably the government underprices flood plain insurance--I'm not sure why though.)

    ReplyDelete
  10. Rick, do you think that, just as I can't buy insurance coverage on anyone's life unless I have an insurable interest in it, maybe we should not let market participants buy financial derivatives "coverage" (the functional equivalent of insurance) unless they have an insurable interest? E.g. I can't buy life insurance on a stranger's life (for fear that this would give me an incentive to do him or her harm). But I can buy it on a family member (whom the law presumes, generally correctly, I love more than money). Maybe financial market participants should not be allowed to engage in derivatives trading unless they can prove that they are actually doing it to reduce a risk they are facing. The exception would be the insurance writers, like investment banks, who would be regulated and required to keep reserves. Thoughts?

    ReplyDelete
  11. Anonymous: You are basically describing the regulatory differences between "Hedgers," which have an external purpose in trading derivatives, and "Speculators," which do not.

    You don't want to exclude "Speculators" from derivatives markets because they can enhance both information and liquidity, making the derivatives markets more efficient.

    The last few years have taught us (again) that leverage, complexity, and skew in derivatives can hide a great deal of risk. However I think that central clearing of derivative contracts, coupled with the more rigorous collateral requirements already being instituted throughout the industry, would go a long way towards mitigating the risk of another meltdown.

    To extend a Bookstaber analogy: Where before anybody could build a nuclear reactor and acquire fuel from arbitrary numbers of suppliers, this would be like requiring all fuel to be purchased through a single middleman, and that middleman would ensure that no operator can possess more fuel than his reactors can safely contain.

    ReplyDelete
  12. Professor Mehrling makes a good point: There are some types of derivatives that amount to insurance against a market catastrophe. An individual market entity can't credibly offer such insurance if its collateral is bound up in the market. This would have been a valuable insight prior to the current crisis!

    However the bona fide insurance markets have already addressed this concern. For example, private insurers don't write flood insurance, and typically exclude "systemic" catastrophes like nuclear war, insurrection, etc. Meanwhile, entities that trade in catastrophe insurance (which includes centrally-cleared bonds) are very sensitive to counterparty quality and typically employ hefty posted collateral requirements to mitigate counterparty risk.

    ReplyDelete
  13. This is a nice, thought-provoking, thread.

    One area that seems unclear to me is how one defines 'systemic risk.' As long as AIG was the only pigeon at the poker table was the risk truly systemic? Who is the "definer?" A regulator? If so, the regulate-versus-sell-the-product issue seems a bit moot. Any of a number of approaches would probably work fine provided the political will to act is sufficient.

    As to owning assets that would be damaged in a catastrophe, Japanese insurance companies, amongst others, have handled this systemic risk for years, in their case with respect to earthquake risk on the main island. And our bond and stock markets (and our currency) have been beneficiaries.

    Defining "insurable interest" is an important point, which to my uninformed eye looks like a call for a regulatory role. Too bad that rule wasn't in place a few years ago --- that would have stopped the CDS market cold! But then again, that was part of the genius of the derivatives players, to migrate to the most weakly regulated turf, where we have 50 state-level insurance regulators and therefore no one really in charge. So we end up with Gov. Patterson of NY trying to help calm the markets as AIG finally folds under the weight of the scam. For all their many weaknesses, to the credit of the Chinese government they have leap-frogged the US in terms of a financial regulatory structure: they have Securities, Banking, and Insurance regulatory commissions nationwide, and are now politicking over the formation of a unified super-regulator. You can bet our US experience is being watched and debated.

    ReplyDelete
  14. A few remarks related to questions posed in the previous comments:

    On Michael Webster's question about the effect of short selling of synthetic securitized mortgages, I really do not know anything about that. But I would ask a question toward the statement that there was not much demand for these bonds, because I wonder how anyone can know that.

    On MG's question on whether regulation needs to be global, I think the answer is that it must be global, or at least be mirrored in all the major markets. We are going to need very tight discussion between ourselves, the ECU, UK and Japan for any new regulatory approach to be successful.

    On Anonymous's question on the need to prove economic value, I don't think anyone would go for that, even though in the extreme that is where the argument leads. It starts to sound like gun control for derivatives. But as a starting point, it seems there is no harm in trying to collect data that indicates who is using various derivatives for what purpose. I don't know how you get anyone to fill in the box that says "over-lever" or "avoid taxes".

    ReplyDelete
  15. It is clear to me that Federalist@Dec2 is a trader who believes that everything about financial markets must be tradable. Anonymous@Dec1 has a much better thought, though perhaps has not fleshed it out quite well enough.

    It is easy to understand why bond (in particular, but hardly exclusively, MBS) purchasers might want to insure their bonds. Among other possibilities, pension funds hungry for AAA-only investments could insure non-AAA products to act as if they were AAA. (Note that this hunger by pension funds for AAA was a significant driver of this current mess.)

    Now Federalist says that in order to provide accurate pricing, we have to allow specualtors to provide "both information and liquidity", but this assumes that either insurance products should be tradable, or that there would be insufficient interest in providing such a product by self-interested, legitimate, and fully qualified insurers to provide realistic pricing.

    The latter case first. This market is HUGE. There would be many competitors for this business were it not for the fact that unqualified and unregulated insurers (i.e., the investment banks) had not underpriced the market by ignoring prudent reserve practices. This is why there is (and should be) a separation between insurance companies and investment bankers. While they both deal in risk, their jobs are miles apart.

    But on to the tradabilty of bond insurance. In the case of an actual bond being insured, there's no reason why the insurance couldn't trade with the bond. The purchaser could continue to pay the premium, find another carrier, or could chose to assume the entire risk. Nothing special here.

    In fact, it is only when we trade without the underlying asset (as Anonymous points out) that we run into a problem. The technical (insurance) name is "anti-selection", and trust me, the very word makes insurers run for the hills. The reality is that when I insure an asset I possess, I insure it hoping I won't lose it. When I insure an asset I do not possess, I insure it hoping I will lose it. It doesn't take a genius to see that those two motivations are totally contradictory, and therefore the underlying risks must also be totally different. Futhermore, if the underlying risks are so different, the two products simply cannot be combined and expected to act under all conditions as if they were the same. Not only will this not happen, but their divergences would not be ratable. In other words, it wouldn't be insurance.

    Now the odds are that this will remain as Federalist suggests. We may very well end up keeping this hedger/speculator nirvana that Wall Street seems to believe it can't live without. If we do however, we should do all bond buyers (and especially pensions funds requiring AAA ratings) a big favor: We should never allow them to believe it is insurance they are buying with their bond. It never was insurance, it is not now insurance, and barring significant changes, it will never be insurance. To suggest otherwise in the future will hopefully be considered fraud.

    ReplyDelete
  16. I have a mini case study on swaps for whoever has the time to read... this is based on a true story which you can verify :) try to price the following swap - pay fixed aud, receive float usd, forward start in 15y, maturity 15y. calculate the breakeven fixed rate you would have to pay in aud. the answer has been around negative 80bp for more than a month now , negative meaning you would be receiving 80bp on the aud leg and also receiving usd libor for 15 years. so you receive with both hands (adds up to quite a bit), you pay absolutely nothing, and this has a PV of zero, what do you make of that (again, please check).
    and now, let's talk about efficient markets a little. how come nobody is pushing the price back into positive territory? the existence of this is not difficult to figure out, everybody in this market knows the situation, there has even been an article on it on totalderivatives. could there be hidden risks to this, like counterparty risk? counterparty risk as in not being able to fulfill your obligation to receive cash on both legs?? well, if it was me, you can take my word for it, I will definitely accept the cash. as for the coutnerparty's ability to hand you the cash at the time, this will come under collateralized trading agreements, you are fine there too.
    maybe this swap is not really tradeable, it's just implied from weird numbers on broker screens. not exactly.
    it is probably difficult to trade this particular swap as such, but you can make an approximate one doing maturity spreads on interest swaps and basis swaps. bid offer is wide and it doesn't take much volume to move this "market", but still, even if you end up receivning 70bp or 60bp instead of 80bp, doesn't sound too bad. so it can be done, there is only one risk to it, in my view. the same system that says that this is worth 0 now can also show it negative, depending where the underlying curves mvove. so the "breakeven" of this swap can move from -80bp to -100bp, in which case you take a "mark to market" loss on your position. but you will say "hang on, all you do is receive cash, how can this ever be a loss" - well, according to our beautiful set of rules, it can. and yes, the moment you have a "mark to market" loss on the position, somebody will force you to post collateral, to guarantee that you will in fact receive the cash... and it may be a very long time before you can show a "mark to market" gain on this and receive any bonus personally. so in the end my boss didn't agree to do it, particularly before the bonus season. we already have some exposure to this from some massive exotic derivatives positions (which are the cause of this distortion in the first place), at least I managed to keep that exposure and not "hedge" it by paying somebody both legs of the swap.
    I just want to use this example to show that the current set of rules encourage people to do stupid things and waste money (somebody else's money, of course). the market moved here because some us house allegedly pushed the basis swap in 30y to -150bp, in an attemt to bring the mark to market losses to this financial year, after a change in the management of the desk, so that the next year's PL looks better. I am sure you've seen this many times before, at least I have.
    Ultimately people hedge their bonuses, which are a function of accounting PL, and some accoutning rules need fixing really bad. this would never have happened if people had not not been allowed to recognize almost all of the PL of a 30y exotic structure on day one... anyway, this is a different discussion, some other time maybe.

    ReplyDelete
  17. Hi, Dr. Bookstaber. You're the hero of my spoof of Dr. Seuss's The Lorax about the derivatives market, "The Bookstabber." You can see it here, if you're interested: http://www.thenervousbreakdown.com/dcorrigan/2009/01/the-bookstabber/

    I hope your seussification is pleasing to you!

    ReplyDelete