|Forscher:||Matthias Blonski, Paul Gortner|
|Kategorie:||Financial Markets, Experiment Center|
Topic and Objectives
The project is an extension of project 11223 (“Non-Standard Preferences and Financial Decision Making”). In winter 2016, we conducted additional experiments to answer a question raised by the original project. In the original project, we conducted experimental asset markets with different levels of social information and found that the information condition strongly correlates with a measure of risk aversion. Since the risk measure was taken after market experience, market experience possibly influences choices under risk. This result is non-causal and could simply reflect a sampling error. However, it raised the hypothesis whether social information within a market influences risk aversion. A nascent literature examines the impact of social influence on choices under risk.
Therefore, the hypothesis warranted an additional experiment where a test for a causal relationship between risk aversion and market experience is administered. In order to do so, risk aversion was elicited both before and after subjects traded under different market conditions. However, contrary to the original findings, there was no correlation between treatment and risk aversion. Additionally, risk aversion also did not change after being exposed to different market environments. Hence, there was also no causal relationship between social aspects of the trading environment and risk aversion. Therefore, this extension was not pursued further in a separate working paper, but rather the additional results were used to expand and strengthen the results of SAFE WP No. 67 (“Peer Effects and Risk Sharing in Experimental Asset Markets”).
- Salient social comparisons in experimental asset markets reduce risk taking.
- In line with theoretic predictions portfolios are more similar when social comparisons are possible.
- The reduction of risk in treatments compared to treatments without social comparisons is driven by the desire not to earn less than peers.