King’s Men Fiddle With U.K. Forecasts: David G. Blanchflower

It is time to reveal a dirty little insider’s secret. It isn’t hard to make time-series forecasting models produce wildly different results.

Estimates have a tendency to whizz around all over the place in the face of small changes in assumptions, data used and time period covered. And errors can be huge, especially at turning points, so carefully considered judgment matters. Economic forecasts can be massaged, so independence is vital.