Economic forecast is an attempt to predict the future state of an economy using past patterns and relationships between current variables. The concepts behind this methodology transcend specific models or even the particular underlying relationships used to make them work, as they are broadly applicable to any time series that exhibits some sort of statistical pattern. For example, stock market prices are commonly predicted by examining the relationship between the price of certain stocks over a period of time and the rate at which these prices rise or fall.
These forecasts are often made by government agencies and economists employed by private companies, as they serve to help determine which fiscal and monetary policies to implement or change. Many rational people look upon the forecasts of these organizations with a healthy dose of skepticism, however, as there is always room for bias, no matter how well conceived a model may be.
For instance, the CPI is a lagging indicator that looks at the cost of goods and services over a time period in order to see if they are increasing or decreasing. If the cost of living increases, this could indicate inflation and if they decrease it may signal deflation. However, a large body of research shows that the average linear effect between two time series isn’t enough to explain the dynamics of the real world and can yield large errors when applied to recessions.
For this reason, non-linear models and a broad range of other cutting edge time series modeling techniques have been found to be useful in forecasting economic activity. The literature is vast and constantly evolving with innovations being proposed at a rapid pace.