Alternative Approaches for the Analysis of Duration Data

Abstract

In the last chapter we considered the construction of explanatory models for the probability of leaving unemployment built on the hazard rate concept. Clearly, application of hazard rate models is not the only way to analyze the impact of socio-economic characteristics on unemployment duration. In this chapter we consider three important alternative approaches. Though not only applied to unemployment duration data but also to unemployment stock data we begin with a look on time—series regression analysis of aggregate unemployment. This serves to demonstrate the type of specific complications that can be expected to arise if aggregate instead of individual unemployment data are analyzed. In chapter IV.2 we inspect an example for “structural” models of individual unemployment duration that aim at separate identification of variables affecting the reservation wage. Finally, chapter IV.3 demonstrates how inference on search behavior and exit from unemployment is possible without statistical modeling, but by computing exact elasticities of reservation wages and re—employment probabilities for individual unemployed.