When meteorologists predict the weather for a month, they often use one of the so-called naïve forecasts. What the average spend during June for the last five years – the counterpart of the climate of the region and the basis for the naïve forecast – stands as a first-cut forecast.
A second type of naïve forecast is persistence, which would predict spending based on the average of the preceding months. All else being equal, it assumes the pattern will continue. A model developed to predict something needs to perform better than both of the naïve forecasts.
A third method, which is not described in David Orrell, Apollo’s Arrow: the Science of Prediction and the Future of Everything (Harper 2007) at 150, looks at the trend line of preceding months. The (persistence) average smoothes out and hides change; a trend line suggests the direction of change and therefore gives more information.
A fourth method disregards the past and focuses on data from current events. It tries to build a budget from the ground up, although no one can do that without some awareness of what has happened in the past. In essence, it places less value on extrapolation and more on context.