We compare model forecast error statistics with forecast error statistics of professional forecasts. We look at a standard sticky-prices–wages model, concluding that it delivers too strong a theoretical forecastability of the variables under scrutiny (…)