Nonresponse Bias in Inequality Measurement: Cross-Country Analysis Using Luxembourg Income Study Surveys

Abstract

Objective: This study evaluates the bias to inequality measurement from survey nonrespondents. Methods: Sixty-six Luxembourg Income Study surveys for 38 middle- and high-income countries, encompassing some 900,000 households, are used to derive estimates of the Gini coefficient for countries and selected world regions. Results: Household nonresponse typically biases national Ginis downward by 1–8 percentage points. The Gini for North America appears robust to nonresponse, rising by a mere 0.34 percentage points to 44.72 when a correction is applied. The Gini for the European Union Single Market is sensitive to nonresponse, rising by 5.52 percentage points to 37.63. The OECD-wide Gini rises by 5.93 points to 44.58. Conclusion: These results appear consistent with one another and with prior evidence. They suggest that, across countries and selected world regions, household nonresponse biases the Ginis downward by 1–8 percentage points. © 2020 by the Southwestern Social Science Association