Guns don't kill people.
People kill people.
With bullets.
From guns.
— Anonymous
FT Alphaville took gentle exception a couple of days ago to some remarks made by Professor Robert Merton—of Black-Scholes and Long-Term Capital Management fame—in this past Monday's interview in the Financial Times.
The "impish" former academic did say one or two things that could be construed as controversial, but the following is not one of them:
"The major advantages of using derivatives are that they are efficient in transferring huge amounts of risk. ... Derivatives are like anti-lock brake systems in a way—there is no question that they can make things safer, but only if people choose to use them that way. Often they don’t—they might choose, for example, to drive faster in worse weather. Often we have chosen to use these tools not to decrease risks but to increase the benefits of taking the same risks."
True, true, and true. But note his critical distinction that derivatives "transfer" risk. They do not eliminate it: they merely allow the separation, alteration, and recombination of risk inherent in securities and other financial instruments, and its subsequent repackaging and resale to other investors. Risk is incompressible: it cannot be eliminated, only transferred (for a price).
This is a point I think many people do not understand, or glide over if they do. All the wonderful risk transfer instruments that have been developed over the past several years—credit derivatives, total return swaps, collateralized debt and loan obligations, etc., etc.—have not eliminated one iota of risk in the global financial system. They have merely spread it around, presumably more efficiently, to the investors who want to hold it.
Some people argue that this broad-based, system-wide transfer of risk has indeed made the world's financial markets safer and better able to withstand shocks. For example, credit risk has been largely taken away from a limited number of structurally highly leveraged market participants—the commercial lending and investment banks—and distributed through CDOs, CLOs, and credit derivatives to a large number of hedge funds and other investors with an appetite for risk (and the accompanying yield).
All this is sensible, and probably true. But can we say that systemwide risk has indeed been reduced? To say that convincingly, you would have to argue that the hedge funds and others buying credit risk are somehow "better owners" of such risks; that in fact traditional balance sheet lenders were inefficient holders of credit risk, and hedge funds know how to price credit risk better than commercial banks. But is this true? I for one find it hard to believe that a collection of ex-Wall Street bond traders and fixed income quants—who are the guys buying this stuff for hedge funds nowadays—actually have superior credit skills than the green eyeshaded legions at JP Morgan, Citibank, and Bank of America who used to originate and hold corporate debt. After all, let's not forget that someone has to own that risk while the underlying credit obligation(s) are outstanding. Not everyone can trade it away when things turn sour.
Instead, I think a case could be made to support Merton's alternate hypothesis that drivers on the Credit Highway have indeed speeded up a little, confident in the belief that their anti-lock credit derivatives have reduced the risk of them crashing their own vehicle. (The historically low credit spreads and decreasing covenant protections for corporate debt in the face of dramatically increased volume for riskier credits certainly supports this contention.) But, if so, the analogy is imperfect, and may reflect a misunderstanding of the situation.
Unlike anti-lock brakes—which for a given level of speed and road conditions should reduce an individual driver's probability of a crash—the use of credit derivatives does not decrease the likelihood of an underlying security's default at all. (Remember, it has only transferred that risk to someone else: it has no preventive effect whatsoever.) Rather, it is as if a driver has sold off some or all of his individual "crash risk" to one or more of the other drivers on the same road. The driver who sold the risk is presumably better off (if slightly poorer), but what happens if the one who bought the "extra" risk crashes (i.e., ex post "mispriced" the purchased risk)? What happens if he crashes in front of our clever low-risk driver and causes a multi-car pile-up? After all, don't you think that the fact you could sell your crash risk to other drivers through the magic of Crash Transfer Derivatives would encourage more people to drive, and drive faster, in less safe conditions, than they would have otherwise?
Furthermore, just where is all this securitized risk going? Is it spreading out nice and smooth across hundreds of diversified hedge fund portfolios, so that no one risk event or connected series of risk events (e.g., Amaranth) could cause a pile-up on the Global Market Highway? Or is it collecting stealthily into large, concentrated, correlated pools of which their owners may or may not be aware (e.g., LTCM)? The former condition would indeed be a big improvement over the historical concentration of risk in a few large, highly levered commercial banks and broker dealers. The latter would be no improvement at all, and might be worse.
Only time will tell. And, in the meantime, I am afraid we will have to take the hedgies' word for it that they know what they are doing. I would feel better if I believed that, but I have seen too many examples of supposedly brilliant people acting like complete and utter horses' asses (Exhibit A: Professor Merton and his cohorts at LTCM) for me to rest easy. And, eight years after the implosion of LTCM, Professor Merton still has not learned enough about the functioning of markets to avoid saying claptrap such as this:
The causes of [LTCM's] collapse, though, are widely misunderstood, says Robert Merton. While some observers blamed events on the faith that the fund placed in financial models—founded on a belief in rational markets—Prof Merton says the real problem was the way that LTCM’s counterparties behaved.
When the $100bn fund started to suffer losses, the counterparties did not behave as proponents of finance science—or rational markets—predicted. Instead, they sold assets in a seemingly indiscriminate panic, triggering market swings more violent than anything Merton had expected.
Pace our Nobel laureate, there was nothing irrational about the behavior of LTCM's counterparties. The ones who sold outright sold fast and at almost any price because they knew or suspected that 1) LTCM had mammoth deteriorating positions on the same side of the market that were compounded by crushing leverage and 2) in the face of spiking volatility and vanishing liquidity they had better get out before the doors closed completely. The buyers on the other side of LTCM's positions knew the same damn thing, and just sat and waited for LTCM to completely collapse so they could swoop in and snap up its positions at fire sale prices. (That many of these selfsame counterparties and buyers happened to be LTCM's prime brokers—Wall Street's best and brightest—just makes the irony that much greater. Never let it be said that Wall Street is a pretty place.)
Notwithstanding the clever maths Professor Merton and his pocket protector set used to model continuous time finance processes in order to derive the simple and practical Black-Scholes model, real market participants bear very little resemblance to gas molecules diffusing in a jar. For one thing, I am not aware of any gas molecules that modify their behavior based upon the observed behavior of other gas molecules. Markets are not inanimate, robotic price discovery mechanisms that unconsciously seek out "rational" equilibria. They are composed of living, breathing, conscious agents with mortgages and demanding spouses who explicitly work to maximize their own advantage, especially in times of crisis.
Besides, nobody expects the Spanish Inquisition.
© 2007 The Epicurean Dealmaker. All rights reserved.
