Abstract
Many popular estimators for duration models require independent competing risks or independent censoring. In contrast, copula-based estimators are also consistent in the presence of dependent competing risks. We suggest a computationally convenient extension of the copula graphic estimator to a model with more than two dependent competing risks. We analyse the applicability of this estimator by means of simulations and unemployment duration data from Germany. We obtain evidence that our estimator yields nice results if the dependence structure is known and that it is a powerful tool for the assessment of the relevance of (in-)dependence assumptions in applied duration research.
| Original language | English |
|---|---|
| Pages (from-to) | 359-376 |
| Number of pages | 18 |
| Journal | Journal of the Royal Statistical Society. Series C: Applied Statistics |
| Volume | 59 |
| Issue number | 2 |
| DOIs | |
| Publication status | Published - 1 Jan 2010 |
Bibliographical note
We thank Simon Lee for contributing several important ideas to our work and Melanie Arntz, the Joint Editor and two referees for helpful comments.Funding
Wilke is supported by the Economic and Social Research Council through the grant 'Bounds for competing risks duration models using administrative unemployment duration data' (RES-061-25-0059). This work uses the Institute of Employment Research employment subsample (IABS 2001-R01) of the Research Data Centre at the Institute. The Institute of Employment Research does not take any responsibility for the use of its data.
Keywords
- Archimedean copula
- Dependent censoring
- Duration of unemployment
