The Sober Outlook of Competent Investors – A framed-field Experiment in Times of Volatile Markets

Link al seminario:

The paper explores the results of a framed-field Web experiment where N=141 Swedish participants provided forecasts for the price of four DJIA stocks 3-month ahead, decided how much they purchase or sell from each stock assuming 50% of their investment budget is initially invested, and provided confidence statements regarding the accuracy of their price predictions. The experiment was run in times of strong market uncertainty due to the pending Ukraine war and increased inflationary pressures. The VIX fear index stood at 30 on March 15th 2022 when the questionnaire was first administered, but decreased to a level of 21 by March 27th when the last participant took the experiment. The four prediction stocks exhibited an average decrease of -5.4% along the 3-month investment intervals. The sample consists of private investors (N=80), finance-industry professionals (N=33), and finance students (N=28), and almost 2/3 of the participants rate their market knowledge at 5 or more in a 1-7 Likert scale. To avoid the endogeneity concerns that complicate the joint analysis of price forecasts, trading decisions, confidence statements and potentially endogenous controls, we take a data-driven approach, searching for the competence indices CMPT that exhibit the strongest correlation with the returns on trading (RoT). Our best CMPT is a linear combination of four variables: an indicator for professional industry job, a market familiarity index extracted from the participants’ ratings of their general market knowledge and their familiarity with the specific stocks of the experiment, the forecast-confidence statements, and the age of the participants. The best-CMPT correlation with the RoT is 0.57 and buy/sell portfolios that copy the trades of the highest CMPT decile while reversing the trades of the lowest CMPT decile show mean (three-months) returns of about 2.7%, taking the costs of borrowing into account. The CMPT index links with more pessimistic predictions for the investments stocks, selling rather than buying stocks, and generally more skeptical beliefs regarding the predictive value of historical prices and professional analysis. The strong-competence participants are aware of their superior skills, but do not exhibit the overprecision that characterizes the sample in general.

Keywords: Expertise, Framed-field Experiments, Overconfidence, Trading decisions