Non-take-up of unemployment benefit II in Germany: a longitudinal perspective using administrative data

Many households eligible for means-tested benefits do not claim them. While most empirical models of benefit non-take-up use only contemporaneous covariates, this paper links the German Panel Study Labour Market and Social Security to administrative records, which provide households’ benefit receipt and earnings histories even outside survey participation and allow a correction of misreported receipt. Eligibility for Unemployment Benefit II is simulated with the microsimulation model GETTSIM. Findings indicate that a longer history of benefit receipt is associated with a higher probability of claiming, whereas higher past earnings and recent earnings gains are associated with a lower probability. Including these histories noticeably changes the estimated conditional associations of conventional covariates.