Homothetic Efficiency: Theory and Applications

We provide a nonparametric revealed preference approach to demand analysis based on homothetic efficiency. Homotheticity is widely assumed (often implicitly) because it is a convenient and often useful restriction. However, this assumption is rarely tested, and data rarely satisfy testable conditions. To overcome this, we provide a way to estimate homothetic efficiency of consumption choices. The method provides considerably higher discriminatory power against random behavior than the commonly used Afriat efficiency. We use experimental and household survey data to illustrate how our approach is useful for different empirical applications and can provide greater predictive success.

本文提出一种基于齐次效率（homothetic efficiency）的非参数显示偏好（nonparametric revealed preference）方法，用于需求分析。齐次性（homotheticity）常被（往往隐含地）假定，因其是一种简便且颇具实用价值的约束条件。然而，该假设极少得到检验，且实际数据极少能满足其可检验条件。为解决这一问题，本文提出了一种估算消费选择齐次效率的方法。相较于常用的阿菲亚特效率（Afriat efficiency），该方法对随机行为的甄别能力显著更强。本文通过实验数据与家庭调查数据，阐释了该方法在不同实证场景中的应用价值，且可实现更优的预测效果。