The authors address the issue of estimation and inference in dependent non-stationary panels of small cross-section dimensions. The main conclusion is that the best results are obtained applying bootstrap inference to single-equation estimators, such as fully modified ordinary least squares and...

Non stationary panel models allowing for unobservable common trends have recently become very popular. However, standard methods, which are based on factor extraction or models augmented with cross-section averages, require large sample sizes, not always available in practice. In these cases we...

We develop a sieve bootstrap range test for poolability of cointegrating regressions in dependent panels and evaluate by simulation its performances. The test seems to have good size and power properties even with small cross-sections, moderate time samples, and low heterogeneity.

A bootstrap algorithm proposed by Psaradakis (2001) for hypothesis testing in I(1) regressions is discussed and shown to be valid only under the null hypothesis. A simple correction making the procedure valid under both the null and the alternative hypothesis is proposed.