In recent years several regression-based decomposition methods have been developed in order to identify the main determinants of socioeconomic inequality of health. In this paper we present a new regression approach that decomposes the correlation between socioeconomic conditions and health outcomes more directly than has been done so far.
The method can be applied to both rank-dependent and level-dependent indicators of socioeconomic inequality of health. The response variable of our model measures the overall performance of individuals in the health and socioeconomic domains, and is regressed on a set of explanatory variables using OLS. The core of our composite response variable consists of the product of an individual’s health outcome and the rank or level the individual attains in the socioeconomic distribution, depending on whether a rank-dependent or level-dependent indicator is decomposed. This simple reformulation of the indicator does not require the explanatory variables to be exclusively related to either health or the socioeconomic variable, but allows for a combined relationship. Regression results are described in terms of the marginal effects of the explanatory variables, but also in terms of their logworths or importance values. We illustrate our method by means of an empirical study using Australian health and income data. We compare our decomposition results to those obtained by other methods, such as the recently proposed recentered influence function (RIF) regression approach.