Perhaps it is the different jargon used in different disciplines, not sure. But for some reason, the terms ‘predictions’, ‘forecasts’ and ‘projections’ are frequently used interchangeably.
There should be at least some distinction, here is what I entertain:
- The word ‘predictions’ should be reserved for situations where the future is already here. For example, given the characteristics of a property, we predict its value in the market. The property is here, and it has market value already. In that sense, the future is here. Statisticians also call those ‘fitted values’.
- The word ‘forecasts’ should be reserved for situations in which the future is not here yet. We forecast the earnings of a company for the next quarter.
- The word ‘projections’ should be reserved for situations where the future is not yet here, but we have some influence regarding it. If you are sitting in the board of the company, you don’t forecast the earnings of that company for the next quarter, you project them. You have a clear plan where those earnings should come about, based on profits based on revenues based on your company’s operation. The members of the Fed project the policy rate, they do not forecast them. The main reason they project rather than forecast is that they have control over it.
There, confusion removed.
There’s already a term for #1: nowcasting.