Two Behavioral Finance Exercises
Dale Rude
University of Houston
The Stock Picking Simulation
Background. One of the most difficult topics to teach is the efficient market hypothesis and its implications for stock picking. Students are often convinced that they have an effective stock picking method or know someone who does. This simulation is designed to allow students to observe the range of returns generated by a various levels of knowledge. It shows that a high return relative to a benchmark is inadequate to establish stock picking ability.
The Simulation. Annual returns (3/1/01 to 2/28/02) for the S&P 500 have been obtained and placed in the spreadsheet. For the simulation, 100% has been added to each return to make them all positive. Then the returns are normalized (by raising them to the 1/8 power) using a Box-Cox transformation. After the student specifies a desired correlation of predictive accuracy, the spreadsheet devises an error term for each stock. This is added to the normalized return to create a simulated judgment. By clicking on a "button," the simulation is rerun producing a new set of simulated judgments for the specified correlation.
The instructor can easily update the returns in the future. In addition, the spreadsheet can be modified for use with different samples of companies.
Procedure. Students specify level of error in the simulation by specifying a correlation between simulated judgments and the actual returns. The simulation then generates a set of simulated judgments and devises portfolios of the top 25, 50, 100, 150, 200, and 250 stocks, reporting the average return for each. The student repeats the process, revealing what percentage of portfolios at each accuracy level that beat the average S&P500 return. This demonstrates ambiguities in assessing stock-picker accuracy because the returns are highly variable for low correlations.
Benefit. Students observe that returns can vary widely for a given level of knowledge just due to chance. The net effect is to make them more skeptical of claims by mutual fund companies, newsletter writers, and investment press about their stock-picking efficacy.
To obtain copies of these exercises, go to my web site If you have any questions, contact me via email at or by phone at 713.743-4673. This stock picking simulation is new this semester. The version available on my web site is a beta version.
The Illusion of Market Stability Exercise
Background. The illusion of market stability is a naïve belief that the market will exhibit low variability and stable predictive relationships for the foreseeable future. This exercise offers students a chance to create their mental picture what the distribution of returns for the S&% 500 look like and to compare their distribution to the actual distribution.
The Spreadsheet. The Excel spreadsheet contains the year 1928-2001 in column A. The actual S&P 500 returns for those years are “hidden” in column Z of the spreadsheet.
Procedure. Using an Excel spreadsheet, students record their estimates of the annual returns for the S&P 500 from 1928-2001. After doing so, they retrieve the actual S&P 500 returns from column Z of the spreadsheet. Next, they compute mean, standard deviation, and autocorrelation (lagged one year) for estimated-returns and for the actual returns. Comparison of these sets of numbers reveals that students typically a) underestimated average return and risk/variability and b) overestimate the degree of autocorrelation. The practical implications of this is that they are likely to avoid the stock market because of underestimated returns, to overweight stocks when they do invest in stocks because risk is underestimated, and to overestimate their ability to forecast the market's future.
Benefit. Students directly compare their "mental picture" of the stock market's operation to actual data, learning that the market is more difficult to predict than they thought.