Regional economic forecasting models without cointegration outperform traditional methods.
The study tested if regional and national industry outputs are connected. They found little evidence for this connection in the data. Forecasting models that don't assume this connection work best for predicting regional economic trends.