Trends and cycles in non-stationary panel models

摘要

This paper utilises Bayesian approach to extract latent common trends and cycles of non-stationary panel data. I develop a Markov Chain Monte Carlo (MCMC) algorithm to explore the highly dimensional posterior distribution of the panel model. Numerical simulation shows that the Bayesian approach based on this algorithm is effective at both estimating the elements of regression coefficients and error variance matrix and extracting latent components. To illustrate the potential of the approach, the study applies the method to investigate quarterly metropolitan housing prices and daily dot-com stock prices. The empirical results show the stronger the long-run growth the higher the cyclical volatility.