Non-Technical Summary:
Risk preferences are an important component of human decision making. A person’s willingness to take risks, such as purchasing risky assets, starting a business or breaking the law, can greatly determine the lifetime economic and social trajectory of that person. An important question is whether risk preferences are stable over the life-course and whether any instability is deterministic (i.e. driven by observable changes in the person or her environment) or simply measurement error. Related to this is the question of whether changes in risk preferences are transient or permanent. This matters for policy makers because it provides information on whether policy changes could have indirect effects on risk preferences, how long these effects might last, as well as whether preferences are malleable and could be shaped to meet desirable social objectives.
I make two main contributions to the literature. First, I study whether risk preferences change in response to common life events (changes in finances, health shocks, parenthood, property crime and bereavement) and whether there are important dynamics in the response function. More specifically, I explore whether the response to risk preferences is stronger closer to the event date and decreases over time. Second, I test for mechanisms between life events and risk preferences (i.e. why do risk preferences change in response to life events?).
The data for my study are from the Household, Income and Labour Dynamics in Australia survey. The main risk preference instrument I use is stated willingness to take financial risks. My estimation strategy controls for the fact that those who experience life events may be innately more/less risk averse. I find that risk preferences do respond to certain life events. People tend to be more willing to take risks after an improvement in finances and are more risk averse after a worsening in finances, birth of first child or death of a child or spouse. Importantly, these effects are more pronounced closer to the event date and tend to disappear over time, supporting a model of preference stability in which variation over time is at least partially deterministic but preferences are mean reverting. The mechanisms underlying these risk preference dynamics are explored. I find limited support for the hypotheses that changes in total expenditure, mental health or mood explain the results. Instead, emotional stability is found to be an important moderator – responses to life events are generally stronger for those with low emotional stability – implying that emotions play an important role.