Welfare receipt misreporting in survey data and its consequences for state dependence estimates: new insights from linked administrative and survey data

Journal for Labour Market Research

Table 5 Regression of underreporting on socio-demographic and household characteristics: marginal effects

	Marg. effect	Std. error
Age in years	− 0.0019***	0.0004
Male	0.0036	0.0081
Education (Casmin, categories 1–3)	0.0187**	0.0059
Health restrictions	− 0.0632***	0.0084
Migration background	0.0066	0.0092
Partner	0.0273**	0.0080
Child < 2 years	− 0.0525*	0.0189
Child 2 to 3 years	− 0.0458*	0.0157
Child 4 to 6 years	− 0.0006	0.0110
Child 7 to 16 years	− 0.0256**	0.0086
Living in East Germany	− 0.0109	0.0098
Local unemployment rate (annual average)	− 0.0031*	0.0015

Marginal effects on the probability to underreport UB II benefits based on the sample of 7102 respondents who could be identified as recipients in the administrative data (613 Underreporters, 6489 respondents who reported correctly). Standard errors clustered at the individual level. ***/**/* denote statistical significance at the 0.001, 0.01, 0.05 levels
Source: PASS_0614_v1 & LHG V08.01.00 – 201504