Nonparametric Estimation of Utility Functions

点击翻译按钮获取中文摘要

It is commonly assumed that consumers' demand functions are generated by the maximization of some utility function. Detailed knowledge of utility functions is necessary for the prediction of consumer demand under different environments and for the assessment of the welfare implications of policy changes. The purpose of this project is to develop nonparametric methods for estimating utility functions from consumer-demand observations. The methods are nonparametric because they do not require the specification of a specific functional form. This is important because this approach can eliminate misspecification errors leading to faulty predictions. The computational complexity of estimating nonparametric demand functions requires the use of a supercomputer, and this award will provide access to a supercomputer for developing and testing nonparametric estimation methods.