|
John Grable, PhD
Michael Roszkowski, PhD
In a recent Wall Street Journal (WSJ) article (September 22, 2017), Professor Mier Statman criticized risk-tolerance questionnaires, contending: “The truth is that the questions advisers ask often don’t measure what they purport to.” While we applaud his attempt to alert advisors of the need to probe and educate their clients rather than accepting a risk tolerance score blindly, we must respectfully disagree with Statman’s conclusions about how to best assess investors’ risk tolerance.
Risk tolerance can be conceptualized in various ways, with many definitions and an even larger number of measurement techniques. Statman advances one particular point of view, typically espoused by economists, rather than presenting indisputable and universally accepted principles about how an investor’s risk tolerance should be assessed.
Specifically, we disagree with his evaluation of the diagnostic value of personality traits and related attitudes in the process of profiling an investor’s risk tolerance. The value of a profile is determined by its accuracy. When profiling, characteristics associated with risk tolerance can be enlisted to differentiate between risk takers and risk averters, even though these factors by themselves do not constitute risk tolerance. In a sense, these characteristics can serve as symptoms of risk tolerance. A case in point is the use of measures of confidence and regret.
As Statman acknowledges, there is a very extensive list of studies documenting a positive relationship between self confidence and risk tolerance. However, Statman implies that if high-risk tolerance is accompanied by overconfidence, somehow this invalidates the risk tolerance score. This is a misconception because confidence and risk tolerance frequently go hand-in-hand. Typically, persons who are confident tend to be more risk taking than individuals who lack self-confidence. Knowledge of investors’ self-confidence therefore allows one to predict their behavior when facing risk. Statman further contends that the correlation between risk tolerance and regret is “close to zero.” Is Statman’s assertion correct? We analyzed data from over 400,000 people and found a correlation of .39 between self-assessed risk tolerance and a question Statman claims is a measure of regret. Either the question is not measuring regret or there is, in fact, a relationship between regret and risk tolerance. In either case, because answers to this question correlate with risk tolerance, this question can be legitimately used to infer risk tolerance.
At the root of Statman’s contention that risk-tolerance questionnaires often lead to misguided investment strategies is the notion that a score based on a combination of personality characteristics and attitudes known to be related to risk tolerance leads to an invalid indication of someone’s willingness to engage in a risky financial endeavor. Instead, Statman argues for testing risk tolerance using traditional economic procedures. We want to stress that there is little evidence to show that the economic approach works particularly well. In fact, a number of studies show that personality and attitude measures do a much better job of predicting what happens in the real world. In the end, what really matters to advisors is how well a risk-tolerance assessment tool predicts future behavior, not the purity of the questionnaire. Risk profiles need to predict behavior rather than explain choices in laboratory experiments.
|