"How is model-based macroeconomic forecasting possible?"

There's an interesting discussion of economic forecasting over at Worthwhile Canadian Initiative. The post and the responses raise some points I've been wondering about myself. It's good to see people inside the field addressing them.