The ``missing rich'' in household surveys: causes and correction approaches

This paper presents a survey of causes and correction approaches to address the ``missing rich'' problem in household surveys. ``Missing rich'' here is a catch-all term for the issues that affect the upper tail of the distribution of income: undercoverage, sparseness, unit and item nonresponse, underreporting and top coding. Upper tail issues can result in serious biases and imprecision of survey-based inequality measures. A number of correction approaches have been proposed. A main distinction is between those that rely on within-survey methods and those that combine survey data with information from external sources such as tax records, National Accounts, rich lists or other external information. Within each category, the methods can correct by replacing top incomes or increasing their weight (reweighting). Correction methods can be nonparametric and parametric. This survey aims to help researchers choose appropriate correction strategies and design robustness tests.