The ability to predict economic activity is a vital skill for governments, central banks and business. Forecasters use a wide range of techniques from judgmental methods that involve the expertise of individual forecasters to models based on modern economic theory like dynamic stochastic general equilibrium (DSGE). The goal is to produce a forecast of national output that will be useful for business plans and financing decisions.
The largest errors in forecasts typically occur around the beginning and end of recessions as the linear time series models used by many forecasters break down. Non-linear models that account for seasonal patterns, time lags and asymmetric information perform better during these periods.
Forecasting is often highly subjective. A forecaster’s personal theory of how the economy works determines which indicators he or she will pay attention to, which can lead to biased projections. The tendency to avoid risk in the face of public pressure can also bias forecasts, as forecasters may be too conservative to make bold projections that would damage their reputation.
Long-range forecasts are more difficult than short-range forecasts because demographic factors have a larger impact on economic activity. For example, the age composition of the population changes consumption patterns and the demand for goods and services. In addition, there is a need to anticipate technological developments that may change the way consumers use goods and services. Forecasters have to weigh the impact of these changes with the fact that they can only know what is happening now, not what will happen in the future.