A BAYESIAN-ANALYSIS OF AUTOREGRESSIVE TIME-SERIES PANEL-DATA

We describe a Bayesian hierarchical model to analyze autoregressive ti me series panel data. We develop two algorithms using Markov-chain Mon te Carlo methods, a restricted algorithm that enforces stationarity or nonstationarity conditions on the series and an unrestricted algorith m that does not, Two examples show that restricting stationary series to be stationary provides no new information, but restricting nonstati onary series to be stationary leads to substantial differences from th e unrestricted case. These examples and a simulation study also show t hat, compared with inference based on individual series, there are gai ns in precision for estimation and forecasting when similar series are pooled.