Testing for Unit Roots in Semi-Annual Data

We consider the problem of testing for unit roots at the zero and seasonal frequencies in time-series data which are recorded semi-annually. The proposed methodology follows that of Hylleberg et al. (1990) and Beaulieu and Miron (1993) for quarterly and monthly data respectively. The non-standard asymptotic distributions for the single and joint tests are derived, and various percentiles of the finite-sample distributions are tabulated. Monte Carlo simulation is used to investigate the powers of the tests, and we illustrate their application to several semi-annual economic time-series.