"Serving the energy market" - http://www.erasmusenergy.com
Risk-aversion and willingness to pay in choice experiments
http://www.erasmusenergy.com/articles/163/1/Risk-aversion-and-willingness-to-pay-in-choice-experiments/Page1.html
Mehdi Farsi
Mehdi Farsi, Senior Scientist and Lecturer
Center for Energy Policy and Economics (CEPE)
Department of Management, Technology and Economics
ETH Zurich
 
By Mehdi Farsi
Published on 02/15/2008
 
Keywords: choice experiment, willingness to pay, risk aversion, energy efficiency, housing
Published in:
Publication year: 2007

This paper extends the linear utility model commonly used for estimating the willingness to pay for non-market goods to a non-linear model with decreasing marginal utility. The proposed approach relaxes the assumption of constant rate of substitution between income and non-market commodities, an assumption which can be especially restrictive in cases when the non-market good is a luxury commodity or a new good whose benefits are not completely known. The adopted non-linear formulation can therefore accommodate risk-averse behavior with respect to nonmarket goods particularly when the non-market attributes are measured by discrete variables. The proposed models have been applied to data from a choice experiment for energy efficiency measures in apartment buildings. The econometric specification is based on a fixed-effect logit model. The results suggest that ignoring consumers’ risk-aversion toward new non-market goods could lead to an underestimation of the marginal willingness to pay. However, consistent with previous studies the non-linear effect of income does not have a considerable effect on the estimation results.

Risk-aversion and willingness to pay in choice experiments

This paper extends the linear utility model commonly used for estimating the willingness to pay for non-market goods to a non-linear model with decreasing marginal utility. The proposed approach relaxes the assumption of constant rate of substitution between income and non-market commodities, an assumption which can be especially restrictive in cases when the non-market good is a luxury commodity or a new good whose benefits are not completely known. The adopted non-linear formulation can therefore accommodate risk-averse behavior with respect to nonmarket goods particularly when the non-market attributes are measured by discrete variables. The proposed models have been applied to data from a choice experiment for energy efficiency measures in apartment buildings. The econometric specification is based on a fixed-effect logit model. The results suggest that ignoring consumers’ risk-aversion toward new non-market goods could lead to an underestimation of the marginal willingness to pay. However, consistent with previous studies the non-linear effect of income does not have a considerable effect on the estimation results.