Abstract:
The UK unemployment is examined by means of ARFIMA models using Sowell's [1992] estimation procedure. A model-selection strategy based on diagnostic tests on the residuals, along with likelihood criteria is adopted to determine the correct model specification. The results suggest that the UK unemployment is well described as an ARFIMA model, with the order of integration fluctuating between 1 and 2. Thus, the standard approach of first differences leads to series with long memory behaviour.