Unemployment Duration in the United Kingdom: An Incomplete Data Approach

For the evaluation of policy reforms numerous governments use, among other sources, administrative social security data. Although this data is large and contains detailed information about policy measures, it inherits several limitations due to the administrative process of generating data. This paper explores the implications of missing interval information in data from the UK (JUVOS) for the analysis of unemployment duration. Variants of the JUVOS are used by the labour administration and the research community as an important source for the analysis of unemployment. While previous work has mentioned the relevant data limitations, they were not taken into account in the empirical approaches. The econometric analysis in this paper shows that competing implementations of unemployment duration in the data yield partly unstable empirical result pattern even in presence of a huge number of observations.