Some Aspects of Pooling Time Series and Cross Section Data: I. Linear Models with Two Random Components

摘要

The paper considers some aspects of the problem of pooling time series and cross section data. Specifically, we analyse, from a Bayesian viewpoint, linear regression models with two random components one of which is autocorrelated. The problem of making inferences about the parameters when the autocorrelated component is stationary is first discussed. The analysis is illustrated by a numerical example showing that the time series component of the data can exert a strong influence in determining the posterior distribution of the regression coefficients. The results are then generalized to non-stationary and explosive models. Finally, the case of pooling several linear models is considered and a possible application to seasonal series is indicated. (Author)