Your Cart

Your cart is empty

Home  >  Volume 46 (May, 2018)

13. SENSITIVITY OF PRIOR INFORMATION ON THE ESTIMATION OF HETEROGENEOUS DYNAMIC MICRO PANEL MODEL by J. O. Iyaniwura, A. A. Adepoju and Y. O. Akinlade Volume 46 (May, 2018 Issue), pp101 –106
Sale price: $5.00
Quantity:

SENSITIVITY OF PRIOR INFORMATION ON THE ESTIMATION OF HETEROGENEOUS DYNAMIC MICRO PANEL MODEL

J. O. Iyaniwura, A. A. Adepoju and Y. O. Akinlade

Department of Statistics, University of Ibadan, Ibadan, Nigeria

Abstract

The key feature of Bayesian estimation is typically based on sensitivity of choice of prior which determines the inference made about posterior estimates. The prior density and the likelihood function are crucial elements of any analysis, and both must be fully specified for estimation to proceed. This paper extends the importance of prior information under estimation of Bayesian approach to heterogeneous panel data models with lagged dependent variable. It considers the fundamental issues of statistical inference of a random coefficients formulation using Bayesian approach. Noninformative and Informative priors are considered in this study and posterior results are based on 10,000 replications with N = 20 and T = 5. 

Theoretical findings are accompanied by extensive Markov Chain Monte Carlo experiments, which show that the estimator perform well so long as the dimension of N>T. The results show closed (estimates)/values of the parameters which establish the sensitivity of prior in the estimation.

Keywords: Heterogeneous effect, Dynamic panel data, Hierarchical Bayesian Inference, Informative and Noninformative prior, Posterior Simulation (Gibb sampling)

click here to download full Abstract
Close
Loading...