Risk attitude and preference
This is a great primer on risk (spanning from the St. Petersburg paradox to prospect theory and current risk assessments)
Abstract (Via Weber)
Citizens of Western countries are asked to make an increasing number of decisions that involve risk, from decisions about how much and how to save for their old age to choices among medical treatments and medical insurance plans. At the same time, uncertainty about choice outcomes has gone up as the result of ever faster social, environmental, and technological change. Accuracy in predicting what choices people will make, at least in the aggregate, is an important determinant for the success of public policy interventions. In addition, corporate and public policy often tries to influence and modify people’s choices in the face of risk and uncertainty, for example, getting people to save more of their income or getting women to invest in less conservative instruments. Understanding the processes that underlie risky decisions and the drivers of risk taking is critical to both agendas.
Introduction (Via Weber)
To introduce the reader to how risk preference and risk attitudes have been modeled, this paper starts out with a historical overview of normative risky choice models from philosophy and economics [Expected Value (EV) and Expected Utility (EU) models] and from modern finance (Risk–Return models). EV and EU theory both assume that (1) people view risky options as distributions of possible outcomes, (2) outcomes (either their objective value or their subjective utility to the decision maker) are discounted as a function of how likely they will occur, (3) these discounted values are integrated over all possible outcomes to provide a measure of the value of each risky option, and (4) the option with the greater overall value is chosen. Risk–Return models, on the other hand, assume that risky options get represented as a trade-off between the first and second moment of their distributions of possible outcomes, and that the value of a risky option increases with its first moment (EV or average outcome) and decreases with its second moment (variance or the degree of unpredictability of its outcomes). These models lie at the root of rational choice theory and thus describe how risky decisions ought to be made within economic definitions of rationality. What they have in common is that individual differences in risk preference are described by a single parameter, which is typically described as the decision makers’ risk attitude.
Miguel’s Favorite Insights (Via Weber)
Because risk taking is context- and process-dependent, people’s willingness to take risks and the multiple determinants of their risk taking described above need to be assessed in context- and process-sensitive ways. There is no single best way in which risk attitude can be assessed or risk taking predicted. EU-based lottery choice methods (e.g.,55) predict risk taking mostly for contexts similar to the assessment method, i.e., for other monetary lottery choices. Without a highfidelity match between the assessment tool(s) and the nature of the situation for which one is trying to predict or modify risk-taking behavior, researchers as well as practitioners will continue to find low accuracy for their predictions and limited success for their interventions. Situationalmatching of assessment tool(s) to situation should be done on both concrete and abstract characteristics.
Another important abstract feature of risk-taking situations is the distinction between static and dynamic risks. Much real world risk taking is incremental and dynamic, involving sequential risk-taking with feedback, from taking risks in traffic to risky substance (ab)use. Risk taking in such dynamic contexts is typically not predicted by static assessment tasks, like one-shot lottery choices that are not resolve until the end of the assessment.60 If the risk taking to be predicted is dynamic, dynamic task assessment tools like the Balloon Analog Risk Task61 or the diagnostically richer Columbia Card Task (CCT; see62) should be employed. The CCT provides an assessment of risk taking in either an emotionally engaging (hot) and more analytic (cold) context and also yields a measure of the complexity of information use.