A economic forecast is an estimate of the performance of an economy at some future time. It may be in terms of a real measure, such as GDP, or in a proxy, such as employment.
Economic forecasting methods are widely varied, and a large literature exists on the topic. The common thread is that a given methodology assumes that patterns of historical behavior or relationships between contemporaneous variables can predict the performance of a related variable, such as a stock market price, or that some combination of these can predict the outcome of a particular event.
Often, these methodologies are employed because the person making the forecast lacks detailed knowledge of why one variable behaves in a particular way or simply does not have the time to attempt such an analysis. The fact that many of these methods require a statistical characterization of the variable in question also makes them attractive, since they can be readily applied by those who do not have the expertise to understand or apply a more analytical method.
This does not mean that the quantitative accuracy of a forecast is easy to assess, though. There is a large literature that explores how to evaluate the performance of a forecast and, in particular, it is well known that the earliest estimates are often quite inaccurate. Moreover, the nature of an economy’s predictive context is often more complex than that of large-body physical motion, so that a prediction of a quantity at a point in the past can appear wildly overestimated when compared with what ultimately turned out to be the case at that point in time.