Multiple Imputation for Poverty Rate Estimation from Rounded Income Data

• Activity Number: 283
• Type: Topic Contributed
• Date/Time: Tuesday, August 5, 2014 : 8:30 AM to 10:20 AM
• Sponsor: Survey Research Methods Section
• Author(s): Hans Kiesl*+ and Jörg Drechsler
• Companies: Regensburg University of Applied Sciences and Institute of Employment Research
• Keywords: multiple imputation ; poverty rate ; heaping

Abstract

In surveys on income, respondents tend to round their answers with unknown degree, resulting in "heaps" in the distribution of the observed values. In this talk we illustrate the substantial impact that rounding can have on important measures derived from the income variable such as the poverty rate. To obtain unbiased estimates, we propose a two stage imputation strategy that estimates the posterior probability for rounding given the observed income values at the first stage and re-imputes the observed income values multiple times given the rounding probabilities at the second stage. A simulation study shows that the proposed imputation model can help overcome the possible negative effects of rounding. A slight overestimation of the sampling variance of non-parametric poverty rate estimators is outweighed by a significant bias reduction. We also present empirical results based on the household income variable from the German household panel study "Labor Market and Social Security".