Behavioral finance is useful only if it can be applied to help people make better decisions. This chapter offers reflections on the good, bad, and ugly of practical applications of behavioral finance in a commercial banking setting. It explores the difficulties of nonexperts experimenting with behavioral finance, and how effective applications require a unique mix of expert knowledge and the ability to effect change through a business. Principles of good applications of behavioral finance are also presented, with information on how to start using behavioral finance within an organization. The importance of senior management’s acknowledging that behavioral finance practitioners do not necessarily know the correct answer and that they will need to use randomized control trials to learn is also discussed.

Behavioral finance is useful only if it can be applied to help people make better decisions. This chapter offers reflections on the good, bad, and ugly of practical applications of behavioral finance in a commercial banking setting. It explores the difficulties of nonexperts experimenting with behavioral finance, and how effective applications require a unique mix of expert knowledge and the ability to effect change through a business. Principles of good applications of behavioral finance are also presented, with information on how to start using behavioral finance within an organization. The importance of senior management’s acknowledging that behavioral finance practitioners do not necessarily know the correct answer and that they will need to use randomized control trials to learn is also discussed.

This study extends experimental tests of (cumulative) prospect theory (PT) over prospects with more than three outcomes and tests second-order stochastic dominance principles (Levy and Levy, Management Science 48:1334–1349, 2002; Baucells and Heukamp, Management Science 52:1409–1423, 2006). It considers choice behavior of people facing prospects of three different types: gain prospects (losing is not possible), loss prospects (gaining is not possible), and mixed prospects (both gaining and losing are possible). The data supports the distinction of risk behavior into these three categories of prospects, Further, probability weighting and diminishing sensitivity of utility as predicted by PT are observed. Loss aversion is, however, less pronounced, except for choices where one prospect is degenerate. The data suggests that the probability of losing may be relevant for loss aversion.

An understanding of an investor's risk tolerance is essential to determining investment suitability and yet there is no universally agreed definition of what risk tolerance is. The authors discuss the need for a holistic approach to measuring risk tolerance that provides a consistent framework for constructing efficient multi-asset allocation and ensuring a suitable aggregate level of risk and return, given the investor's personality and financial circumstances. Risk tolerance is only stable when considered as a personality trait, which can be measured effectively using holistic psychometric scales. Attempts to measure risk tolerance at lower levels of granularity are unstable, they accentuate, rather than mitigate short-term behavioural distortions and cannot be meaningfully aggregated to provide an overall level of risk that is appropriate given the investor's risk capacity. This implies that the recent move towards mental accounting, goal-based, assessments of risk tolerance is misguided and exposes institutions to risks associated with providing unsuitable advice.

Abstract In comparing risk attitudes across individuals we usually aim to determine whether one can reliably measure a difference in risk attitudes between two individuals, the magnitude of that difference, and what factors should be controlled for to ensure comparisons are meaningful. These comparisons critically depend on the method of risk attitude measurement and elicitation used, and a clear understanding of what the modeled risk attitude represents. Individual risk attitudes are fragile and very specific to domain and framing effects, and thus comparisons should always be made within the same elicitation method. Depending on the measurement method, comparisons can be made on relative or absolute scales. In general, the more precise the measurement, the more sensitive it will be to slight changes in assessment. Psychometric measures are the simplest and most robust, but allow only for ordinal comparisons. Certainty equivalent and utility models of choices can give more precise, but noisy, cardinal estimates of individual risk aversion. The purpose and need for precision of the comparison will often drive the method a researcher chooses for eliciting risk attitudes, and advice is given for researchers seeking to design or analyze such studies.

A behavioral condition of loss aversion is proposed and tested. Forty-nine students participated in experiments on binary choices among lotteries involving small scale real gains and losses. At the aggregate level, a significant proportion of the choices are in the direction predicted by loss aversion. Individuals can be classified as loss averse (28 participants), gain seeking (12), and unclassified (9). A comparison with risk behavior for binary choices on lotteries involving only gains shows that risk attitudes vary across these domains of lotteries. A gender effect is also observed: proportionally more women are loss averse. In contrast to the predictions of comonotonic independence, the size of common outcomes has systematic influence on choice behavior.

A new definition of loss aversion is proposed and tested. Thirty-one students participated in experiments on lotteries involving small scale real gains and losses. At the aggregate level, approximately 60% of the choices are in the direction of loss aversion. The analysis at the individual level shows that, compared to loss seekers, more than twice as many are loss averters, the remaining subjects being unclassified. Comparing these results with risk behavior involving gain only lotteries shows a strong polarization effect: when loss outcomes are introduced, a majority of subjects shift from unclassified risk attitude in the domain of gains towards loss aversion, but some exhibit loss seeking. A strong gender effect is also observed. Proportionally more women are sensitive to losses. There is statistical evidence that in these binary choices the sign of common outcomes has influence on choice behavior, which is in contrast to the predictions of comonotonic independence, the core principle of rank-dependent utility theories.