Do Our Financial Models Still Work?
Tuesday, April 27, 2010 / 2:30 pm - 3:45 pm


Glenn Yago, Executive Director, Financial Research, Milken Institute

Aaron Brown, Risk Manager, AQR Capital Management; Author, The Poker Face of Wall Street and A World of Chance
Colin Camerer, Robert Kirby Professor of Behavioral Finance and Economics, California Institute of Technology
Stacy-Marie Ishmael, Reporter, Financial Times
Myron Scholes, Nobel Laureate, 1997; Chairman, Platinum Grove Asset Management
Bruce Tuckman, Director of Financial Markets Research, Center for Financial Stability

Economic models were a casualty of the subprime collapse. How could the frameworks used by Wall Street's tech whizzes discount the possibility that home prices might fall? When markets collapsed, ostensibly uncorrelated asset classes fell in tandem and liquidity dried up.

A panel of financial economists contended that economic models are necessarily imperfect depictions of reality. But models failed miserably this time because quants too often built models in isolation, without a broad understanding of markets. The models' failure also revealed two fundamental problems — their inability to account for liquidation values and the dynamics that result when traders make similar bets.

Even though financial models proved inadequate, society's need for them has only increased. "We need more skill, more modeling and more technology, not less, to run our modern financial institutions," said Myron Scholes, who won the Nobel prize in 1997 for a new way to value derivatives.

Bruce Tuckman, who heads financial markets research for the Center for Financial Stability, said models have two main blind spots. One is that values given to portfolios often fail to account for the difficulty of liquidating them. This is especially the case with large portfolios. Whereas the price of a marginal trade in a $10 billion portfolio of assets can be determined, the ability to unload a huge portfolio at marginal prices doesn't exist. If liquidation values were taken into account by regulators, institutions with larger portfolios would be required to have higher capital ratios than they do, Tuckman said.

A related modeling problem is the inability to assess risks related to "crowded trades." That's the dynamic that occurs when a lot of investors are in exactly the same trade and exiting at the same time. "Tails in distributions in financial markets are created by crowded trades," Tuckman said. "We have to model this in trading."

The subprime meltdown underscores the essentially chaotic essence of financial markets. Models require continual calibration; risk cannot be avoided. Yet that presents a conundrum to policymakers keen to instill confidence in consumers and financial markets. "Politicians are saying we'll be safe after passing financial reform, but this is actually the last thing we want them to say," Tuckman said.