Westminster Applied Evaluation Workshop

A series of seminars organised by Policy Studies Institute and
the Centre for Employment Research

session six: Thursday 6 May, 5.30pm - 7.00pm

Partial Identification and Data Limitations
in Empirical Labour Market Research

Dr Ralf Wilke
University of Nottingham

Abstract

We consider the relevance of identifying assumptions and data limitations for the precision of empirical labour market research. In particular, we focus on estimation of the effect of unemployment benefits on unemployment duration using linked administrative individual data from Germany. These data are similar to the Work and Pension Longitudinal Study (WPLS) and contain daily information on employment periods and periods of benefit receipt. The lecture summarises the main findings of a series of papers and includes the following topics:

• Defining unemployment duration in register data.
• How relevant are missing or unobserved periods for the precision of estimated policy effects?
• The non-identifiability of the competing risks model.
• How important is the assumption of independent censoring for estimation results?
• What can we say then? Identifiability of the sign of a policy effect.
  

Venue:

Churchill Room 2/58, HMRC, 100 Parliament Street, SW1A 2BQ

Further information

Please contact Stefan Speckesser.