Projections vs. Forecasts
Projections vs. Forecasts
On this blog, I regularly use the terms “projection” and “forecast.” Unfortunately, I use them rather interchangeably. There is an important difference between them and I’ll explain it in this blog post, with the hope that I better sear the information into my own grey matter.
A forecast is the best estimation of the actual future events that will occur. I forecast winegrape prices. I develop indices that combine the most important influences on winegrape prices and I use these to try to forecast the actual behavior of prices into the future. I produce an average of the expected results that becomes the forecast.
Because this is a forecast, I can then determine the range of outcomes that are possible and the probability that those will or will not happen for scenario and sensitivity analysis. I can further calculate the likelihood that I got everything wrong or that one of my indices is invalid.
A projection, on the other hand, is just an illustration of how things either (a) would be in the future, given certain assumptions or (b) might be, according to a guess. The important difference is that the projection is not actually aiming for a most likely or average outcome. It’s either an uninformed guess or based on assumptions that are not, according to the data, most likely to reflect reality.