The Econometrics of Risk

Taring All Investors with the Same Brush? Evidence for Heterogeneity in Individual Preferences from a Maximum Likelihood Approach

“…To the best of our knowledge literature on preferences in financial markets provides no direct evidence (Holt and Chaves (2001), Guiso and Paiella (2008)) or mixed results at best to the question, by how much preferences are connected to personal characteristics and govern trading decisions of individual investors (Barsky et al (1997), Foucault et al (2011)). However, studies from the field of experimental economics indicate that some personal traits such as age and gender might be connected to the tendency to hold stocks (Dohmen et al (2009)).…”

“…The demographics file contains information of marital status, gender, age but also on current reported financial wealth and income as prompted by the bank and reported in euro. As in our dataset information on portfolio volume was reported as ranges, we calculated them as mean wealth and income levels (Holt and Chaves (2001)) using mid-range values. Expected utility models are denoted as EU T , Rank-dependent Utility is denoted as RDU .…”

Is Risk Aversion Really Correlated with Wealth? How Estimated Probabilities Introduce Spurious Correlation

“…Thus common estimation techniques that allow for heterogeneous risk aversion parameters across socio-economic variables, but not heterogeneous risk perception (e.g., Bar-Shira, Just, and Zilberman 1997;Cicchetti and Dubin 1994; and virtually all studies examining wealth and risk aversion) may misinterpret the underlying behavior. Holt and Chavas (2001) suggest that structural estimation is underutilized in measuring properties of risk aversion. Antle (1987), Chavas and Holt (1996), and Saha, Shumway, and Talpaz (1994) each use a structural approach to measure risk parameters.…”

“…This form has become popular because it is simple, and produces risk aversion coefficients that appear as simple linear coefficients of variance measures (e.g., Bar-Shira, Just, and Zilberman 1997;Chavas and Holt 1990;Pope and Just 1991). Two primary drawbacks, noted by Holt and Chavas (2001), to reduced form estimation are that (1) it can be difficult to infer how changes in socio-economic variables affect risk aversion measures, and (2) one must create some estimate of the individual's perceived risk from various choices. Estimating perceived risk may sound simple at first blush.…”

“…Such econometric studies use either structural estimation based on a parametric utility model or reduced form estimation. Holt and Chavas (2001) outline many of the drawbacks of econometric estimation of risk aversion parameters and in particular of reduced form estimation. Reduced form estimation relies on estimating a Taylor series approximation of the expected utility function around the mean wealth level.…”