In a few recent articles, The Economist has tackled the issue of economic predictions and forecasts: how to make them more accurate, timely, and reliable, or how to nowcast rather than forecast. Amidst more accolades for data series and newfangled statistical measures, an interesting point resulted from their investigation: the failure of all economic predictions—especially of government forecasting—to intuit a coming recession. This is mainly due, they reckon, to the fact that most of the economic growth models—used by the IMF, for example—“overestimate the extent to which the future will look like the recent past.”
Had The Economist taken the analysis a step forward, a deeper problem would have been revealed: the issue lies not in overestimating, but in assuming in the first place that the future will look like the past at all. Mises explained this best:
If the future were merely a continuation of the trends that prevailed in the past, it would not be uncertain and we would not then be in need of any forecasting. But as this is not the case, what is called economic forecasting is merely guesswork.
Professional forecasters blame their much talked about failures on the fact that the figures available are insufficient and reach them too late. This apology misses the point. However complete and recent statistical information may be, it always remains information about the past and does not assert anything about the future (Mises 1990, 128-9).
Surely, there are numerous other problems with the current forecasts of economic activity, which have deep roots in erroneous economic theories. But no matter how correct one’s theory is, the future remains fundamentally unpredictable—particularly from a quantitative point of view. It is less unpredictable for the entrepreneur who, in the whirlpool of the market, can make use of the best available tool for penetrating the fog of the future: economic calculation. Everyone else outside the market, from theorist to (especially) policy maker, is best employed in humbler endeavors.
