Characterizing economic trends by Bayesian stochastic model specification search

We extend a recently proposed Bayesian model selection technique, known as stochastic model specification search, for characterising the nature of the trend in macroeconomic time series. In particular, we focus on autoregressive models with possibly time-varying intercept and slope and decide on whether their parameters are fixed or evolutive. Stochastic model specification is carried out to discriminate two alternative hypotheses concerning the generation of trends: the trend-stationary hypothesis, on the one hand, for which the trend is a deterministic function of time and the short run dynamics are represented by a stationary autoregressive process; the difference-stationary hypothesis, on the other, according to which the trend results from the cumulation of the effects of random disturbances. We illustrate the methodology for a set of U.S. macroeconomic time series, which includes the traditional Nelson and Plosser dataset. The broad conclusion is that most series are better represented by autoregressive models with time-invariant intercept and slope and coefficients that are close to boundary of the stationarity region. The posterior distribution of the autoregressive parameters, estimated by a suitable Gibbs sampling scheme, provides useful insight on quasi-integrated nature of the specifications selected.