Capitalmind
Capitalmind
Actionable insights on equities, fixed-income, macros and personal finance Start 14-Days Free Trial
Actionable investing insights Get Free Trial
Commentary

Of Quants And Mad Traders

From The Daily Beast, Quants, Derivatives and the Myth of the Rogue Trader: (HT: @ashenoya)

Just what did these sorcerer’s apprentices do all day? The City veteran tries to explain to a non-math-speaker. Let’s say you want to develop a financial product that will protect investors with huge amounts of money against various contingencies in the market. “The people who build the models are the technical guys, and those are called the quants,” says the veteran—“quant” being short for quantitative analyst, as you’ll recall, although the math can be reminiscent of quantum physics in its incomprehensibility. “Everything is very mathematical in derivatives, and it is pure math, even though we say it is applied mathematics,” the veteran says.

Then you have “the structure guys,” as the City veteran calls them. They’re the ones who actually put together the financial products that are based on the models developed by the quants; the City veteran was one of them. After the structure guys come the traders, who buy and sell the various elements in the packages the others have devised and assembled. “The traders manage the risk,” the veteran says, “and they tend to have similar backgrounds to the quants.” Very broadly speaking, the quants work in the financial equivalent of an ivory tower to develop models for instruments they think will work; the traders report the results from the trenches, and the structure guys figure out solutions and improvements. And finally there are the salespeople, who may not know much about the math, but know how to talk a persuasive bottom line.

The danger in this juxtaposition of mathematical theory and multiple billions of dollars should be evident to anyone who has ever heard of Murphy’s Law. And yet it may not be so obvious to traders in the middle of the action as they implement the quants’ models. In fact, their world can be simultaneously very high pressure and almost mystical. “The richness of math is in the abstraction,” El Karoui told a science publication for women in 2008, after the crash. “It allows you to take a step back from reality, and that gives us the freedom to think the way we want. Mathematicians display a great deal of imagination.”

The issue of using an imaginary world versus a real one prompts statements (largely from big banks) that a market is "not reflecting real value", when positions go against them. Buyers often complain that speculators have, by short selling, taken stocks way below where they should go. Yet they don’t complain when valuations get way ahead of themselves (like in the dot com boom, or the real estate boom, or in India, the stock market madness of 2007-08).

And selling complex products to customers who should know better amounts to mis-selling in the end. Because the salesmen hardly understand the complexities enough to warn their customers about the inadequacies of the models, of the underlying assumptions and so on. From selling what’s best for the customer, they move to selling what’s best for themselves at the cost of the customer.

Remember actuarial-funded-insurance policies that IRDA banned in 2007? These products offered nearly 100% allocation (no charges!) but hid the charges under a complex formula which penalized early withdrawals with charges mentioned through flowcharts. The model might actually work, but it doesn’t make sense to sell it to someone who trusts the insurer to provide him a good product instead of nickel-and-dime methods to steal his money.

The essential problem in such products is the fact that risk is hidden behind complexity, and transferred to someone who doesn’t understand. In the stock market, I might create a quantitative trading method and thus buy a set of stocks that I think are ridiculously low priced. That’s okay, because I have no obligation to disclose my methods to the seller; I have no fiduciary relationship in which he trusts me to get him the right price. And if things go wrong, I assume the risk. The problem would only come if I created a hedge fund, and invested a customer’s money in this kind of model when they are unaware of the risk, and if I say anything short of "you could lose nearly all your money", I would be lying. But in reality, that’s what happens, people lie.

The secondary issue is that quants fall in love with their models even when it’s divorced from reality. When real estate prices fell in the west, the traders who had carefully modelled their derivative bets saw the basic assumption in their models collapse: that one fall wouldn’t trigger another in a different part of the world. When prices fell in a synchronized manner, their thought process was: this doesn’t make sense, and it’s even more out of whack than before, so let’s double our bets. Now that does result in huge wins at some point, but it all comes down to nought when the market really misbehaves.  

Like radioactive material, quant models are only useful if handled carefully.