Bayesian unit root test for panel data

The present paper studies the panel data auto regressive (PAR) time series model for testing the unit root hypothesis. The posterior odds ratio (POR) is derived under appropriate prior assumptions and then empirical analysis is carried out for testing the unit root hypothesis of Net Asset Value of National Pension schemes (NPS) for different fund managers. The unit root hypothesis for the model with linear time trend and linear time trend with augmentation term is carried out. The estimated autoregressive coefficient is far away from one in case of linear time trend only so, testing is not executed but in consideration of augmentation term, it is close to one. Therefore, we performed the unit root hypothesis testing using the derived POR. In all cases unit root hypothesis is rejected therefore all NPS series are concluded trend stationary.