The stability of financial risk preferences under changing probabilities : an experimental study

Abstract

Is your risk preference stable? After Arrow and Pratt defined the method to measure risk aversion, many people try to find out what are their risk preferences. For a rational consumer, his or her risk preference should be stable in similar situations, but some researchers have found out that risk preferences are not stable in many cases. They claimed that risk preferences are diverse between men and women, or between youngsters and elders. And risk preferences are not stable according to the time either. Furthermore, if you use different eliciting method or different survey questions, the risk preferences you get may differ as well.

The problem we are going to solve in this thesis is the stability of risk preferences under different probabilities. We know that people’s risk attitudes vary in different fields, such as financial situations, health decisions and career scenarios, etc. Here we are going to focus on risk preferences in the financial field. To conduct the research, we design an experiment which contains five lotteries whose winning probabilities are different. From the responses of all the subjects, we can estimate their equivalent values for those lotteries and elicit their risk aversion.

According to the experimental results, we find out that people’s risk preferences are unstable in these five lotteries. People show up a significantly different risk attitude when facing the lowest-winning-rate lottery. And this kind of instability is affected by the gender and age as well. We find out that the risk preferences of women and elders are more stable than those of men and youngsters respectively. These results can be applied to the policies about revising the pricing method of financial instruments, marketing strategies of financial companies, designing and selling financial instruments, etc.

Since people’s risk preferences are not stable especially when they face the extreme probabilities, we can guess the reason is that people usually overweight low probabilities. Then in the future research, we can also study the effect of probability distortion in this experiment, to find out how it affects the stability of risk preferences in a gaining situation.

A restriction in this experiment is that the sample cannot necessarily represent the other population groups in the market because students are supposed to be young and bold comparing with others. In the future study, to generalize this research in a real market, we can ask real investors in the financial markets to participate in this experiment. And considering the overconfidence problem in this experiment, in the future, we can still study its effect on people’s risk preferences when people are making financial decisions.