Who misreports welfare receipt in surveys?

Abstract

We match survey and administrative data and determine the extent of misreporting on welfare receipt. In our data, 10.5% of German welfare recipients under-report and 1% over-report benefit receipt. The analysis shows that particularly households who are close to the labour market, without children, and with relatively high household incomes and savings are prone to under-report their welfare receipt. This information is important for the study of transfer programmes based on survey data.

Keywords:

• under-reporting
• over-reporting
• survey response
• welfare use

JEL Classification:

• I32
• C81

Notes

¹  We lose about one third of our sample when estimating the interviewer-fixed effects model. As a robustness test, we also estimated the interviewer-fixed effects model using the Stata ‘asis’ option which does not omit observations for which the interviewer-fixed effects cannot be identified. The results are qualitatively similar to those presented and available upon request. The main results also hold up, when we reestimate model 2 with the sample used in column 3.