TY - JOUR

T1 - Eliciting risk preferences and elasticity of substitution

AU - Burgaard, Johan

AU - Steffensen, Mogens

PY - 2020

Y1 - 2020

N2 - Risk aversion and elasticity of intertemporal substitution (EIS) are separated via the celebrated recursive utility building on certainty equivalents of indirect utility. Based on an alternative separation method, we formulate a questionnaire for simultaneous and consistent estimation of risk aversion, subjective discount rate, and EIS. From a representative group of 1,153 respondents, we estimate parameters for these preferences and their variability within the population. Risk aversion and the subjective discount rate are found to be in the orders of 2 and 0, respectively, not diverging far away from results from other studies. Our estimate of EIS in the order of 10 is larger than often reported. Background variables like age and income have little predictive power for the three estimates. Only gender has a significant influence on risk aversion in the usually perceived direction that females are more risk-averse than males. Using individual estimates of preference parameters, we find covariance between preferences toward risk and EIS. We present the background reasoning on objectives, the questionnaire, a statistical analysis of the results, and economic interpretations of these, including relations to the literature. Funding: The authors gratefully thank the company Ipsos for funding this survey.

AB - Risk aversion and elasticity of intertemporal substitution (EIS) are separated via the celebrated recursive utility building on certainty equivalents of indirect utility. Based on an alternative separation method, we formulate a questionnaire for simultaneous and consistent estimation of risk aversion, subjective discount rate, and EIS. From a representative group of 1,153 respondents, we estimate parameters for these preferences and their variability within the population. Risk aversion and the subjective discount rate are found to be in the orders of 2 and 0, respectively, not diverging far away from results from other studies. Our estimate of EIS in the order of 10 is larger than often reported. Background variables like age and income have little predictive power for the three estimates. Only gender has a significant influence on risk aversion in the usually perceived direction that females are more risk-averse than males. Using individual estimates of preference parameters, we find covariance between preferences toward risk and EIS. We present the background reasoning on objectives, the questionnaire, a statistical analysis of the results, and economic interpretations of these, including relations to the literature. Funding: The authors gratefully thank the company Ipsos for funding this survey.

KW - Certainty equivalents

KW - Constant equivalents

KW - Nonrecursive separation

KW - Questionnaire

KW - Subjective discounting

UR - http://www.scopus.com/inward/record.url?scp=85098561495&partnerID=8YFLogxK

U2 - 10.1287/DECA.2020.0415

DO - 10.1287/DECA.2020.0415

M3 - Journal article

AN - SCOPUS:85098561495

VL - 17

SP - 314

EP - 329

JO - Decision Analysis

JF - Decision Analysis

SN - 1545-8490

IS - 4

ER -