Why Economic Forecasting Is an Art, Not a Science
Shilling rejects the hope that a large econometric model can accept inputs and automatically deliver a reliable economic forecast. Models often require extensive manual intervention, so forecasting ultimately combines indicators, history, policy analysis, and judgment.
- Large economic models can produce nonsensical results
- Repeated manual adjustments undermine the ideal of automatic forecasting
- Leading indicators and policy conditions still matter
- Precise quantification cannot replace judgment
“forecasting in my view as an artist not a science.”
“They simply do not work. They blow up, They produced nonsensical.”