Let the numbers do the talking

Rice statistician's strategies inform Real World Markets conference

Mike Williams, Rice University news staff

There's no sweet vindication in the nation's current economic straits for James Thompson, who could be justified in saying, "I told you so."

But there are lessons to be learned, especially for those who intend to invest in a bear market.

Thompson, Rice University's Noah Harding Professor of Statistics, will host the First Eubank Conference: Modeling Real World Markets. Named for Rice trustee emeritus J. Thomas Eubank, the conference will be held March 23-24 at Rice's McMurtry Auditorium in Duncan Hall. Preregistration is required for the free conference, which is open to all and will address investment strategies that do not depend on the efficient market hypothesis, or EMH.

Thompson has been a chief critic of EMH for decades, and he blames money managers' religious adherence to the theory -- which he said is taught as gospel at business schools -- for many of the nation's troubles.

EMH is the concept that stock markets assimilate new information quickly enough that the current trading price of a security tends to be an accurate reflection of its real value. Such a self-correcting mechanism should make it difficult for investors to beat the market, but Thompson and other nonbelievers said EMH ignores the kind of deep research, including computational, that can help investors make real gains. It also allows the kind of folly that, as Thompson wrote several years ago, "frequently required government interventions of great complexity."

 "This business we're in right now was eminently avoidable," said Thompson, who doesn't hesitate to let his curmudgeonly side out when talking about the economy. His work in statistics over nearly 40 years at Rice has addressed subjects ranging from corporate process control to cancer and AIDS research.

Thompson said the financial pressures of running two wars has made the current crisis particularly acute.

"I really think that's what got us. In the fullness of time, the price of housing - - which is cyclical -- would have gone back up. The values would have largely been recaptured," he said. "But the $3 trillion cost of the war is an enormous hit. It's like a surcharge on the federal government of 15 percent a year."

Thompson and several colleagues loosed a particularly vitriolic attack in 2006 with the publication of a paper titled "Nobels for Nonsense" in the Journal of Keynesian Economics, which laid blame for the derivatives collapse on the Black-Scholes-Merton option-pricing model -- "the granddaddy of all the derivative formulas" -- that won a Nobel Prize for its authors in 1997.

"Alan Greenspan (former chairman of the U.S. Federal Reserve) was very angry when Warren Buffett called derivatives 'economic weapons of mass destruction.' Greenspan thought that options were terrific. I think that's been proved to be wrong," said Thompson.

Greenspan was mistaken, he said, in organizing the $3.5 billion bailout of the failed Long-Term Capital Management in 1998, a hedge fund that had both Robert Merton, a Harvard professor, and Myron Scholes of Stanford listed as advisers. The similar collapse of Enron a few years later "was too large for even Chairman Greenspan to make disappear," Thompson and his colleagues wrote in "Nobels for Nonsense." The flaws of Black-Scholes-Merton, they wrote in another article, "gave encouragement to accounting firms to do bizarre things," and government pressure on lending institutions to increase home ownership by selling subprime mortgages "to persons who had no reasonable hope of sustaining them" exacerbated the problem.

None of the efficiency theories stands up to analysis of the long-term historical data, said Thompson, who teaches that very kind of analysis to Rice students who will not necessarily become financial wizards but instead plan careers in physics, engineering, chemistry and other professions. "A lot of people who get their degrees in theoretical physics are smart enough to think they're never going to make a living that way," said Thompson, noting students last spring routinely made up to 15 percent gains in their experimental portfolios. "We're trying to teach something different from the classical finance course taught in business schools. We're trying to let the data speak to us and to form our models from the data, as opposed to saying, 'This is our model, and the data had better conform to it,'" he said. "This is essentially a portfolio design course, and it just takes long positions. That is to say, it just buys and sells stocks. It doesn't short stocks or any of those nice, round, jolly games that people like to do." Thompson said vindication will come when his kind of course is taught in finance departments and business schools. "It's nice to know what the truth is, but it's even better if you're enabled to share it."