Global economic modelling requires rethink

Subject

Methodological changes in economic forecasting.

Significance

Forecasting failures over the past decade have accelerated methodological changes, moving away from point estimates towards a larger use of forecast ranges and alternative scenarios. More risk indicators have been added to forecasters' watch list, developing early warning signals. Given estimates for key economic determinants, many other variables' outcomes are predictable based on these key factors, thanks to the importance of dominant drivers and the slow-moving nature of some underlying trends. Forecasters are looking for better ways of predicting and assessing combinations of events that push outcomes beyond 'normal boundaries'.

Impacts

• In normal conditions, short-term forecasting methods are broadly reliable, but predictions based on 'turning points' are imprecise.
• Even when underlying trends are shifting, this process is usually slow, enabling forecasters to keep forecasts largely on track.
• Prediction and analysis of 'catastrophe' or 'regime-changing' events are treated differently from standard forecasting.