Font Size:
Shrinkage of time-varying effects in panel data models
Last modified: 2023-06-09
Abstract
We consider regression models for panel data with time-varying effects in a Bayesian framework. We implement shrinkage of regression effects
and the process variances of the effects to distinguish between effects that are practically zero, constant or time-varying via shrinkage priors. Longitudinal dependence is taken into account by including
a subject specific random factor with weights that may also vary over time.
The model is applied to analyse panel data on annual incomes of mothers
returning to the job market after maternity leave
and the process variances of the effects to distinguish between effects that are practically zero, constant or time-varying via shrinkage priors. Longitudinal dependence is taken into account by including
a subject specific random factor with weights that may also vary over time.
The model is applied to analyse panel data on annual incomes of mothers
returning to the job market after maternity leave