The 1% Panic

Our financial models were only meant to work 99% of the time.

The Panic of 2008 is a crisis of trust. Investors don’t trust the value of bad debts enough to offer market-clearing prices. Banks don’t trust one another to stay in business long enough to do business together. And there’s definitely no trust that Washington can avoid creating costly new moral hazards as it attempts to bail out the system.

But the most paralyzing loss of trust may be in Wall Street’s system itself: How did the smartest people at the best banks running the most sophisticated financial models fail to forecast the collapse of mortgage-related securities? How did this unpredicted collapse devastate the system? And most of all, can we ever again trust the financial models on which value is supposed to be determined?

These questions matter because despite the current crisis, modern finance has delivered enormous benefits, from explaining to investors why they should diversify their investments to the creation of mutual and index funds. Related innovations helped financial institutions speed capital to its best use, fund new businesses and accelerate global prosperity. In other words, financial engineering worked beautifully — until suddenly it didn’t.

So what happened? Financial models take logic and historical data into account, but it’s now clear that these elegant models have a serious weakness: They can’t cope with illogical and uneconomic factors. Washington’s insistence for years on artificial subsidies for mortgages through Freddie Mac, Fannie Mae and other programs led to a loud “Does not compute!” that is still rocking the financial system.

[Information Age]

Here’s how ill-conceived regulation poisoned the system. Until recently, bank CEOs and regulators slept well at night thanks to a financial model developed in the 1990s called “value at risk” or VaR. It assesses historical variances and covariances among different securities, informing financial institutions of the risks they’re taking. By assessing risk factors across all securities, VaR can compare historical levels of risk for given portfolios, usually up to a 99% probability that banks would not lose more than a certain amount of money. In normal times, banks compare the VaR worst case with their capital to make sure their reserves can cover losses.

But VaR can’t account for extreme unprecedented events — the collapse of Barings in 1995 due to a rogue trader in Singapore, or today’s government-mandated bad mortgages bundled into securities that are hard to value and unwind. The “1% likely” happened. And because the 1% literally didn’t compute, there was no estimate of the stunning losses that have occurred.

Yale mathematician Benoit Mandelbrot pointed out the shortcomings of the VaR model in his “The (Mis)behavior of Markets,” published in 2004. He noted that bell curves work for, say, disparities in the height of people. In markets, instead of flat tails of rare events at either end of the bell curve, there are “fat tails” of huge upsides and huge downsides. Markets are more complex than the neat shape of bell curves.

Last year’s bestselling nonfiction book had a similar theme. In “The Black Swan,” former trader Nassim Nicholas Taleb pointed out that extreme outcomes are actually common, warning that financial engineers — “scientists,” as he calls them — ignore these unlikely outcomes at their peril. But today’s credit panic was not entirely unpredictable. Mr. Taleb was prescient in writing, “The government-sponsored institution Fannie Mae, when I look at their risks, seems to be sitting on a barrel of dynamite, vulnerable to the slightest hiccup. But not to worry: Their large staffs of scientists deemed these events ‘unlikely.'”

Likewise, the financial engineers at once high-flying hedge fund Long-Term Capital Management thought they had taken all risks into account, but the Russian financial crisis of 1998 blew their model. Last week the former general counsel of LTCM, James Rickards, reflected on how an incomplete VaR model undermined his firm. “Since we have scaled the system to unprecedented size, we should expect catastrophes of unprecedented size as well,” he wrote in the Washington Post. “We’re in the middle of one such catastrophe, and complexity theory says it will get much worse.”

Global markets and new financial instruments are indeed complex. This complexity led to a fragility that made government meddling in markets more dangerous than ever before — creating the 1% likely disaster. The good news for VaR and similar models is that the free market alone would not have allowed the bubble of subsidized mortgages, but the bad news is that it’s far from clear that Congress has learned from the current crisis to pursue policy goals in ways that don’t distort the fundamentals of markets.

Now the regulators trying to fix the damage in the financial system must also try to avoid more 1% likely crises. Transparent steps that restore market efficiency are better than complex, ad hoc policies that postpone market solutions. These programs should be judged on whether they make the financial models function better or function not at all. As we’ve learned, there’s not much room in between.

Advertisements
Post a comment or leave a trackback: Trackback URL.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s

%d bloggers like this: