Risk is not a number

(May 2007) Although using models of expected returns to guide portfolio construction is the starting point for any investor forgoing the certainty of a T-bill return, “the equations cannot always describe the true nature of competition and combat” in the securities marketplace, warns doyen of investment management studies Peter Bernstein. Nevertheless, “theory matters even though theory does not work in the strictest sense of the word — but few theories in the social sciences do.”

Bernstein, speaking at the annual conference of the CFA Institute in New York this week, was basing his comments on Hungarian-born mathematician John von Neuman’s game theory. “Investing,” he says, “is a giant von Neuman game,” with every player vying against the others in combat and competition, struggles that have an impact on forecasted market valuations. The quiet isolation of the solitary financial theorist gets tested in the real world of thousands of actors making buying and selling decisions.

“Let’s face it, your expected return may be a neat number based on wise calculations,” he explains, “but your actual return will be what some investor in the future will pay you for your assets. The enemy is us and that makes it a lot more difficult to quantify,” he says, making it dangerous to “view risk as a number” that represents volatility. Everyone focuses short term risks, he suggests, even though daily volatility does not signify a permanent impairment of capital — and more ominously, “we do not know what the future holds.”

Liquidity-event risk

Indeed, as if to confirm, Nobel Prize-winner Robert Merton, a pioneer in options-pricing theory, explained to the conference his experiences with ill-fated Long-Term Capital Management in 1998, when the star-powered, highly leveraged hedge fund faced a liquidity crisis. “You think you’ve looked at every angle,” he recalls. While using derivatives, LTCM was not using “overnight money” but instead was “all termed out; that was the prudent thing if you have an illiquid position.”

But since the counterparties, because they were locked in on the other side, couldn’t liquidate their side of the bargain, they did something else. Positions were marked to market every day, and the counterparties marked in their own favour to create a cushion. For LTCM, that meant the daily net asset value went down. It “created a correlated series of events that had a feedback on NAV I never thought about,” Merton says.

He made his comments after defining the chief peril in the derivatives markets: liquidity-event risk: “it’s outside the playbook,” a situation where deer freeze before headlights, and investors, because they don’t understand the event, pull money from the market.

“The theory of finance revolves around the notion that risk is equivalent to volatility: beta, standard deviation and so forth,” says Bernstein. While many criticize this formulation, Bernstein acknowledges that “volatility does get you in the gut.” Still, measuring risk and quantifying it as a number is not the whole game. “For long-term investors, with strategies oriented toward buy and hold, short-term volatility is either irrelevant or an opportunity to upgrade and revise a portfolio,” he argues. But, harkening back to his catchphrase, “we do not know what the future holds,” Bernstein avers a passion for diversification “as an explicit statement of our ignorance of what the future holds.”

Capital ideas

Game theory is one way to approach the implementation of the “capital ideas” that Bernstein suggests have shaped investors’ approaches to investing. They include the efficient markets hypothesis, mean-variance optimization, the options pricing model and the Capital Asset Pricing Model. These underlie his new book, Capital Ideas Evolved.

Not all such ideas have worked in practice. This is something that behavioural finance theorists have attacked. Bernstein cites career psychologist Daniel Kahneman, who, despite his lack of training in economics, won a Nobel Prize: “the failure in the rational model is in the human brain it requires,” one reason that alpha — excess returns against an index — can still be found in capital markets that even behaviourists accept, at least as a benchmark, as efficient.

For conspicuous failure to get market rationality and investor logic to line up, there’s the Capital Asset Pricing Model, for which William Sharpe and Jack Treynor won a Nobel Prize in economics. “CAPM has failed most miserably in empirical tests,” Bernstein notes. “Yet it is the cornerstone of a growing number of active strategies today.” Moreover, “CAPM, much scorned empirically, today is more alive and well than any skeptic could have foreseen.”

Breakthrough of 1952

That’s because “risk is at the heart of all investment decisions.” Bernstein says the investing world after 1952 is very different from what it was before. That’s the year Nobel laureate Harry Markowitz laid out the mean-variance framework: the risk–reward trade-off. Today it is the starting point for active investment managers, and “diversification is gospel.”

Sharpe, says Bernstein, argues that “too many practitioners, and a large number of the business school professors from whom they learn their trade, tend to forget that all asset pricing models — this is a very important point — all asset pricing models are about expectations. Now I’m quoting him: ‘how in the world can you measure expectations, which look forward, not backward? Hence, it is dangerous to think of risk as a number.'”

There are other factors that defeat a distillation of risk into a set of numbers and defy the thesis of efficient markets. “We know well that markets as a whole go through booms and busts. How can we deal with that?” Bernstein asks. “The bottom line: we can’t.” Markets show a micro-efficiency in pricing; for example, Robert Shiller and Jeeman Jung have discovered that individual stocks are priced relatively efficiently, based on simple factors such as dividend yield. Such analysis may not apply to the whole market.

More than that, financial crises are endogenous: they are generated within the financial system.

With LTCM, Bernstein argues, “models were built on the markets they knew, the markets they had experience with, the markets they had data on.” But the moment they started to invest on the basis of those models, Bernstein says in an application of Heisenberg’s uncertainty principle, “they thought of everything everybody in this room could think of, and then some more.” However, “when they opened up for business and entered into the market, the markets that they knew, the markets that they had data on were no longer the markets they were operating in because they were there.”

That leads to a basic principle for Bernstein, a shift away from risk measurement to risk management. “Risk means that more things can happen than will happen,” he argues, “hence my preference for diversification.”

Filed by Scot Blythe, Advisor.ca, scot.blythe@advisor.rogers.com.

(04/04/07)