A Simple Forecasting Accuracy Criterion Under Rational Expectations: Evidence From the World Economic Outlook and Time Series Models

IMF Working Papers, June 1, 1992

Source details

Canonical URL: A Simple Forecasting Accuracy Criterion Under Rational Expectations: Evidence From the World Economic Outlook and Time Series Models

A simple criterion based on the properties of the forecast error is presented to evaluate the accuracy of forecasts.
The efficiency conditions of an optimization problem are used to show that under rational expectations the standard statistical conditions are necessary, but not sufficient to ensure efficiency.
The criterion is applied to examine the accuracy of the World Economic Outlook projections of growth and inflation for the seven major industrial countries.
Time series models are estimated and the efficiency of the World Economic Outlook projections relative to a benchmark time series model is examined.
A number of empirical tests suggest that the year ahead projections of growth and inflation in the World Economic Outlook are unbiased after 1982.

Develops a forecasting accuracy criterion derived from properties of the forecast error.
Uses efficiency conditions from an optimization problem to assess implications under rational expectations.
Compares standard statistical conditions against the derived criterion, finding standard conditions are necessary but not sufficient for efficiency.
Employs estimated time series models as benchmark comparisons to the World Economic Outlook projections.

Application focuses on growth and inflation projections for the seven major industrial countries.
Empirical tests indicate the year ahead projections of growth and inflation in the World Economic Outlook are unbiased after 1982.
Examines efficiency of WEO projections relative to benchmark time series models through multiple empirical tests.

Subjects: Economic forecasting, Economic sectors, Economic theory, Industrial sector, Inflation, Prices, Production, Production growth, Rational expectations
Keywords: forecast error, Industrial sector, Inflation, inflation error, least squares, models of inflation, Production growth, q statistic test, random walk, Rational expectations, Theil inequality statistic TS, time series forecast, time series model, WP