Past stock prices mislead investors
Investors often decide how to invest based on past stock prices. In fact, however, future and past returns do not depend on each other. If investors are made aware of this error, they change their investment behavior. This is shown in a study by the team of economist Prof. Dr. Christine Laudenbach, member of the Cluster of Excellence ECONtribute: Markets & Public Policy at the University of Bonn. The study was published in advance as an „ECONtribute Discussion Paper“.
The researchers analyzed data from around 2000 customers of a major German online bank. In a first step, the team asked participants about their assumptions regarding the stock market. Respondents were shown the twelve-month return on the German share index (DAX) at six different points in time over the past 50 years and had to estimate the average return over the following twelve months.
The result: Around 70 percent of participants believed in a so-called negative autocorrelation. This means they assumed that the stock market would always move in the opposite direction. If stock prices had risen in the past twelve months, for example, they would fall in the following twelve. A quarter of respondents, on the other hand, expected share prices to increase consistently. They would therefore continue to rise if they had risen in the past twelve months (positive autocorrelation). These beliefs were also reflected in the purchasing behavior of respondents within the past five years. Those believing in a negative autocorrelation were more likely to buy when yields were falling.
Biased perception causes different investment behavior
The DAX actually increased by an average of about 8.5 percent in each time period surveyed. „Recent rates have no predictive value for the future,“ explains Christine Laudenbach, professor of finance at ECONtribute at the University of Bonn. Therefore, it does not make sense to base purchasing decisions on past prices, she says.
To correct the distorted perception, the researchers showed the actual average return to half of the respondents and explained that past returns cannot be used to predict future returns. In a follow-up survey, those affected were more likely to say they expected prices to rise – regardless of how prices had performed previously. Compared to the control group, those who received instruction and had previously assumed that returns would move in the opposite direction bought significantly fewer shares during the coronavirus stock market crash in February and March 2020.
Understanding investment behavior better
The study contributes to a better understanding of households‘ perceptions of the stock market and their consequences on investment behavior. „Our results show that investment mistakes are relatively easy to fix and households actually adjust their behavior based on new information,“ Laudenbach says.
The study furthermore suggests that especially educated people assume a negative autocorrelation. According to Laudenbach, it will therefore be particularly interesting in the future to take a closer look at the factors that determine the different attitudes toward the stock market.
About ECONtribute: Markets & Public Policy
ECONtribute is the only Cluster of Excellence in economics funded by the German Research Foundation (DFG), supported by the Universities of Bonn and Cologne. The Cluster conducts research on markets at the intersections of economy, policy, and society. Its aim is to apply a new approach to understanding markets and market failures in times of social, technological, and economic challenges, such as increasing inequality, global financial crises, and digitalization.
Contact for scientific information:
Prof. Dr. Christine Laudenbach
ECONtribute, University of Bonn
phone: +49 228 73-60199
Press and Communication:
Junior Science Editor
phone: +49 221 470 89965
phone: +49 221 470 7258
Christine Laudenbach, Annika Weber & Johannes Wohlfart: Beliefs About the Stock Market and Investment Choices: Evidence from a Field Experiment. ECONtribute Discussion Paper; https://www.econtribute.de/RePEc/ajk/ajkdps/ECONtribute_128_2021.pdf