IMF Staff Papers, Volume 49, No. 1

IMF Staff Papers, April 18, 2002

Source details

Presents details of a symposium on forecasting performance organized under the auspices of the IMF Staff Papers.
Challenges the assumption that the forecaster's sole goal is to predict the actual outcome as accurately as possible.
Notes that private sector forecasters may herd rather than reveal true forecasts because of incentives.
Observes that public sector forecasts may be distorted for different reasons; forecasts associated with IMF programs are often the result of negotiations between IMF staff and country authorities and may be more accurately viewed as goals or targets rather than pure forecasts.
Emphasizes components of the standard theory of time series forecasting:
- Choice of an information set.
- Choice of a cost function.
- Evaluation of forecasts in terms of the average costs of the forecast errors.
States that including more relevant information in the information set should generally lead to better forecasts.

Incentive structures can lead private forecasters to herd rather than to report true beliefs.
Program-associated forecasts (e.g., IMF programs) may function as negotiated targets rather than objective forecasts.
Forecast evaluation rests on the specification of cost functions and the calculation of average forecast-error costs.
Expanding the information set is generally acknowledged to improve forecasting performance.

Subjects: Consumption taxes, Exchange rates, Foreign exchange, Labor taxes, Purchasing power parity, Real exchange rates, Taxes
Keywords: Baltics, consumption tax, Consumption taxes, East Asia, Eastern Europe, Europe, Exchange rates, Global, group mean, IMF projection, Labor taxes, panel money demand estimation, private sector forecast, program projection, Purchasing power parity, Real exchange rates, SP, standard deviation