The Impact of Alternative Imputation Methods on the Measurement of Income and Wealth: Evidence from the Spanish Survey of Household Finances

Abstract

The goal of this paper is to emphasise the importance of the way of handling missing data and its impact on the outcome of empirical studies. Using the 2002 wave of the Spanish Survey of Household Finances (EFF), I study the performance of alternative methods: listwise deletion, non-stochastic, multiple and single imputation based on linear-regression models, and hot-deck procedures. Using descriptive statistics of the marginal and conditional distributions of income and wealth and estimating mean and quantile regressions, listwise deletion brings imprecise and biased estimates, non-stochastic imputation underestimates variance and dispersion and hot deck fails to capture the potential relationships among survey variables.