UNCOVERING BASIC WANTS USING THE ROTTERDAM AND AIDS MODELS: THE US HOUSEHOLD ENERGY CONSUMPTION CASE

Abstract

Economists inflate the explanatory power of measurable variables such as price and income to explain demand. Using only quantifiable variables is very attractive since it makes it easy to construct mathematically consistent and well expressed models. However, since Lancaster (1971), economic awareness has increased to such a degree that latent, hardly observable and/or measurable variables may bring more insight to the demand analysis. Two approaches compete to introduce such variables: an economic approach originally developed by Lancaster and Becker (1965), and a statistical approach. For Lancaster, beyond quantities of goods demanded, the characteristics of goods are what shape consumer utility and consequently determine its choice. This approach is theoretical and largely based on economic intuition. Few empirical studies using Lancaster have been successful so far. The second, purely statistical approach, considers the possibility of transforming observed data to obtain the "basic wants" that truly affect consumer choice. This approach, known as the Preference Independence Transformation (PIT), has so far been applied only in a few studies using the Rotterdam model frame. The PIT was certainly deduced through mathematically thorough and consistent analysis to uncover the basic want, denoted as T-goods. We intend to revisit the PIT under the Rotterdam framework to uncover the basic goods. Alongside, we implement --for the first time--an independent transformation that eliminates the Slusky interdependencies from the Almost Ideal Demand System (AIDS) setting. We will refer to it as the Slutsky Matrix Independent Transformation (SMIT). Regarding our purpose to check if the two techniques identically define the basic goods, the findings were not conclusive. As a result, we further the analysis by introducing a possibility to unveil the basic wants using US household data.