MADISON, Wis. - Credit Union National Association plans to include an equity portfolio using game theory - a mathematical field honored in last year's Nobel Memorial Prize in Economics - in its $250 million money purchase and 401(k) plans.
ANB Investment Management & Trust Co., Chicago, will manage the portfolio, using a stock selection technique it developed from a problem in game theory known as "winner's curse."
"Winner's curse is one of the main problems people try to solve in using game theory," said Neil R. Wright, ANB chief investment officer, who developed the game theory-based portfolio.
Winner's curse can apply to the stock market because in bidding contests, "winners tend to win in bidding because they are overly optimistic about their estimate" of the upside potential of a stock.
"You win because you miscalculated," Mr. Wright said. "The outlier" - the one with the winning bid - "has a tendency to be incorrect. That means the best estimate tends to be in the middle of the bids," rather than at the extreme.
"But it you want to stay in business, you have to bid low (or high, depending on the situation)," Mr. Wright continued. "That's the winner's curse."
"An active manager faces winner's curse," Mr. Wright added. "More often than not, he will be wrong," because he will buy too high. "The consensus tends to be right on price."
Mr. Wright developed his strategy to take advantage of the winners' curse problem.
CUNA is one of ANB's first clients for the new strategy, which already was in development when the Nobel prize was awarded in October to three games theorists.
CUNA will include the portfolio in each of the four lifestyle investment options it intends to add in April to the two plans, said James C. Cox, senior financial officer.
The association plans to blend six ANB portfolios to create the four lifestyle funds of varying degrees of expected risk and return, from conservative to aggressive.
The six portfolios will consist of four equity index funds - Standard & Poor's 500, an enhanced S&P 500, S&P MidCap 400, and international - an intermediate bond fund and a money market fund.
It is the enhanced S&P 500 index fund that uses game theory.
CUNA also intends to retain the five existing investment options, which are run by seven money managers.
Mr. Wright manages the strategy around the "winner's curse" problem, using it to select which stocks to overweight or underweight in the portfolio.
The enhanced S&P index portfolio will make up 20% of the equity allocation in each of CUNA's four lifestyle portfolios.
Through game theory, Mr. Wright said, the object is to try to enhance the return of the S&P 500 index without adding to the risk.
The portfolio uses 350 stocks from the S&P 500, selected by using the winner's curse decision rules he developed.
To start, Mr. Wright eliminated utilities and foreign stocks, narrowing his S&P 500 universe to about 450 stocks. Then he uses the winner's curse screen to eliminate about 100 stocks from the remaining 450 in the S&P 500 universe. It then selects 150 stocks to overweight in the portfolio and another 200 stocks to hold at market weightings.
"It's not easy to use game theory" in investing, Mr. Wright said.
Mr. Wright explains how he applies winner's curse to select the stocks to place in ANB's enhanced index fund.
"When the price range of a stock is extremely large, that means there is a lot of uncertainty about how the company will do," he said. So to buy the stock, a manager would have to bid up at some extreme from the consensus.
"When the price range of a stock is narrow, that means the consensus is pretty close." So a manager buying this stock wouldn't have to bid much more than the consensus.
Over time, the price of a stock will move from a large range to a narrow range, a movement he calls the cycle of expectations.
"This cycle of expectations - the times when the range in prices is large and the times when they are narrow - reverts to the mean," he said.
Mr. Wright looks at two sets of variables - fundamental analysis about the companies and technical analysis about the trading - that seem to be highly correlated with the ranges of expectations.
Among the fundamental variables are the change in inventory, sales, employment and research-and-development expenses,
The variables that apply to each stock (e.g., inventory doesn't apply to McDonald's) are equally weighted.
The technical variables include changes in trading volume.
"All I'm trying to do is to see how market participants are biasing stock prices," Mr. Wright said. "If I see the trading range is wide or narrow, it tells me what stocks to buy or not buy without looking at the fundamental data of each stock."
Mr. Wright said he studied the work of John Nash, a mathematician and visiting research collaborator at Princeton University, Princeton, N.J., who was one of the three game-theory academicians awarded the Nobel prize. "We all studied the Nash equilibrium in college," he said, although he wasn't as familiar with the work of the two other co-Nobel laureates, John C. Harsanyi, professor emeritus in economics at the University of California at Berkeley, and Reinhard Selten, professor of economics at Rheinische Friedrich-Wilhelms-Univeristat in Bonn, Germany.
