Title: Approximate Bayesian inference in a semiparametric copula model of income and wealth
Authors: Anna Stelzer - Vienna University of Economics and Business (Austria) [presenting]
Abstract: The aim is to evaluate economic well-being by means of estimating the joint distribution of income and wealth. While both income and wealth are of crucial interest individually, analysing them together draws a fuller picture of people's economic possibilities. Recent studies investigate the dependence structure between income and wealth by using copula models, which allow estimating both marginal distributions as well as the joint distribution separately. The tail behaviour of the joint distribution is of special interest, as it indicates whether the relationship between income and wealth for individuals in the lower part of the distribution is different from the ones in the top part. Measures of dependence, both overall and in the upper as well as in the lower tail of the distribution, are estimated in a flexible way. Applying a Bayesian approach to a semiparametric copula model and using HFCS data for several Euro Area countries, inference on summaries of the dependence structure can be obtained by only partially specifying the model. By using this approach, choosing a complete copula structure and the risk of misspecification can be avoided while still answering important questions about the joint distribution of income and wealth.