An economic forecast is a prediction of the value of an economic variable. The economic variables that can be predicted include gross domestic product, one or more interest rates, and various measures of employment, unemployment, and price inflation. Because of the variety of economic variables that can be predicted, a key challenge is to evaluate the relative quantitative accuracy of alternative forecasts.
The simplest approach is to compare the mean or median of individual forecasts from professional economists. Such forecasts are typically published by organizations such as the Survey of Professional Forecasters (SPF) or Blue Chip Indicators.
In more advanced cases, econometric models are employed to attempt to model the behavior of economic variables over time. The models can range from simple time series analysis to estimated dynamic stochastic general equilibrium (DSGE) models.
Regardless of the particular model that is used, the process of making an economic forecast begins with gathering information inputs and assumptions. The historical values of these inputs and the relationships between them are then determined by means of regression analysis or other appropriate methodology.
These relationships are then incorporated into a mathematical model that is run over the forecast time period. The model results are then reported in a format that provides the forecast values and other related data and commentary. Various methods have been tried to improve the accuracy of economic forecasts including nonlinear models, Bayesian techniques, and time series analysis. A recent literature has also highlighted that it is important to consider the likelihood of stochasticity in the economy as a result of which some variables may not behave in linear ways over time.