Beschreibung
Wage mobility in a labor market describes the chances but also the risksof an individual to move between wage categories. Wage mobility is usually
measured on an aggregate level for the whole labor market and little work is
done to capture unobserved heterogeneity across individuals. The present study
tries to obtain some results on unobserved heterogeneity for the Austrian Labor
Market.
For each individual, wage mobility is described through a first order Markov
chain which is heterogenous across the individuals. We will compare two
approaches for capturing unobserved heterogeneity. The first approach is based
on using a finite mixture of Markov chains models moving with different
speed. In this approach it is assumed that within each hidden group no more
heterogeneity is present. This leads to a rather large number of groups which
are not easily interpreted. The second approach is based on using finite mixtures
of multinomial-Dirichlet distributions. In this approach it is assumed that
within each hidden group heterogeneity in wage mobility is still present which
may be described through a Dirichlet distribution with an unknown dispersion
parameter.
Practical implementation in R is based on combing a Bayesian approach
with Monte Carlo simulations based on Markov chains.
| Zeitraum | 15 Juni 2006 → 17 Juni 2006 |
|---|---|
| Ereignistitel | UseR 2006 |
| Veranstaltungstyp | Keine Angaben |
| Bekanntheitsgrad | National |