The Paradox of Risk: How Limiting Risk Actually Increases Risk
or, the Unintended (Risky) Consequences of "Backstopping" Risk
  (August 12, 2008)

Risk is perhaps an overly broad word. The risk of being struck by lightning is low, we lower the risk of injury in an auto accident with airbags, and we "manage" financial risk with derivatives, swaps and other forms of portfolio insurance.

Do you drive carelessly because your auto is equipped with airbags? Perhaps not. But would you drive more cautiously if you were perched on the front bumper? If even the slightest collision would crush the driver's legs to pulp, I think it is safe to say we would all drive with a higher awareness of risk and with greater caution.

As a bicyclist, I am keenly aware that there is no barrier or insurance or protection between my thin frail skin and the pavement or the fast-moving ton of hard material that is a vehicle, or the steel of an open car door I might hit at 15 miles per hour.

You cannot assume anything on a bicycle, and you cannot lose control or awareness for even a moment without risk shooting straight up to undisguised danger. Thus if everyone drove like they were on a bicycle, there would be far fewer accidents and injuries.

But it is easier by far to protect ourselves with airbags and dashboards and all the rest. If vehicles were truly safe, and our driving cautious, how is it that 39,000 people lose their lives in vehicle accidents every year in the U.S. and hundreds of thousands of others are injured?

The "risk," we are informed, is statistically low: "only" 21 Fatalities per 100,000 Licensed Drivers (practically zero, eh, except that it adds up to 40,000 people killed each year).

Even more deceptive is the 1.4 Fatalities per 100 Million Vehicle Miles Traveled statistic; my goodness, only 3 accidental deaths in 200 million miles driven? That sounds absurdly unlikely--yet 40,000 people losing their lives is a non-trivial number of needless deaths.

What statistics do not adequately describe, of course, is that most of those accidents occured in high-risk settings in which the drivers' focus and/or ability was impaired, even as they reckoned risk was managed/limited by the equipment, their safe driving record, etc.

In other words, the somewhat inebriated gent who slips behind the wheel on a dark rainy night senses the heightened danger; but reassured by the fact he's never been in a fatal accident, by his car's airbags, by the low statistical odds of getting killed, etc., he roars off into the unlit darkness. The odds of an accident in these conditions are much higher than the average listed in statistical abstracts, yet they are glossed over by the apparent "low odds" of the drive ending badly.

This is the Paradox of Risk: the more risk is apparently lowered, the higher the risk we are willing to accept. If the drunk wheeled off into the darkness on a bicycle, he probably wouldn't get far; he'd more than likely lose his balance and end up scuffed and sore but very much alive in a ditch. The "real risks" of navigating a dark road while impaired by alcohol are very much exposed on a bicycle.

Ironically, the real danger arises when the driver is reassured by "safety" features and past experiences which appear to "backstop" or limit risk.

Let's now turn to the financial "management" of risk which now threatens to destroy the U.S. financial system and Federal government finances. You have undoubtedly read that Fannie Mae and Freddie Mac are "backstopping" the risk of a mortgage meltdown in the U.S.

Here's how "backstopping" works. Let's say I am a mortgage lender/originator and I want to write a jumbo loan on a property which is, well, at risk of declining in value. Gee, that will make it tough to sell this mortgage if I reveal this risk.

Even worse, the borrowers are a little shaky. If their income drops even a few dollars they wouldn't qualify to buy a doghouse, never mind this expensive home. Again, I have a problem: I certainly don't want to hold this risky mortgage on my books, but how can I sell it for anywhere close to 100% value once I disclose the risks of future foreclosure and loss?

Answer: sell it to Fannie! Yes, Fannie and Freddie will buy any mortgage which isn't stamped "fraudulent" in day-glo orange letters--although they were happy to do that for the past 5 years. Nonetheless, they will buy any risky mortgage I originate, thus "backstopping" my own risk.

Ask yourself: if I had to keep this risky mortgage on my own account, would I approve it? Heck no. I accept the higher risk because I can shove it onto Fannie, which has "backstopped" risk throughout the entire mortgage market.

As a result, risk is piling ever higher. Perversely, this backstop doesn't decrease risk of a systemic meltdown--it ramps it up wildly. All of us lenders are now encouraged (or at least freed) to write "iffy" mortgages because we have a willing buyer who is heedless of risk.

So who's backstopping Fannie and Freddie? The U.S. taxpayer, of course. If all those iffy loans go bad, then the taxpayer picks up the tab to "recapitalize" Fannie and protect its bondholders from any losses.

Some time ago I introduced the concept of "choice architecture" which is basically a structured analysis of the incentives built into a system.

The choice architecture of "backstopping" mortgages is horrendously perverse. As a lender, I have no incentive to carry any risk myself, and every incentive (those huge fat delightful origination fees) to write risky loans and then dump them on the garbage pile which is taxpayer-backed Fannie and Freddie.

There is another form of pernicious systemic risk: assuming the "100-year flood" can't happen every 5 years or so. I have from time to time highly recommended The Misbehavior of Markets and now is the perfect time to check this book out, for the author, fractal pioneer Benoit Mandelbrot, explains in simple mathematical ways how "Modern Portfolio Theory", i.e. the management of risk, is based on a faulty conception of risk and statistical chance.

In a nutshell: while modern portfolio management is statistically based (all those "standard deviations" you always see referenced in quantitiative analyses), the markets behave fractally. Fractals are known as the geometry of chaos, for they describe how seemingly stable systems can quickly, and unpredictably, degrade into chaos.

This was brilliantly illustrated over the past 2 years, when the vast majority of market players and pundits saw no source of risk. They literally could not figure out where risk could even appear, except as a "100-year flood"--some unexpected event no one could foresee.

But as Mandelbrot explains, "100-year floods" actually occur with startling regularity in all markets. Put another way: you simply cannot massage/manage all risk away with fancy statistical models and trillions of dollars in derivatives, just as you cannot eliminate fatal accidents with airbags and collapsing bumpers.

In other words, all you're really doing is masking the risk--you're not eliminating it. And in hiding the real risk, you are lulling the market participants into a pernicious choice architecture in which their willingness to take riskier and riskier actions is rewarded and encouraged, while caution is punished.

This is how you get a total systemic collapse of the entire choice architecture. And by this I mean not just the financial system and Freddie and Fannie, but the "backstop" provided by the taxpayers: I mean the default of the Federal government as risk skyrockets to heights none of the blind practitioners of modern portfolio theory were able to conceptualize, anticipate or stop.

NOTE: contributions are acknowledged in the order received. Your name and email remain confidential and will not be given to any other individual, company or agency.

Thank you, Claude M. ($103), for your stupendously generous (and completely staggering) donation to this site--and for your interest in Weblogs & New Media: Marketing in Crisis. I am greatly honored by your support and readership.

For more on this subject and a wide array of other topics, please visit my weblog.


All content, HTML coding, format design, design elements and images copyright © 2008 Charles Hugh Smith, All rights reserved in all media, unless otherwise credited or noted.

I would be honored if you linked this wEssay to your site, or printed a copy for your own use.


consulting   blog  fiction/novels   articles  my hidden history   books/films   what's for dinner   home   email me