You, Too, Can Forecast

May 14, 1956

In the issue of Jan. 30, Newsweek carried an amusing and instructive article on how far wrong the economic forecasters went at the end of 1954 in predicting what was going to happen in 1955. The moral was that forecasting is not a science.

This is completely true. This column, in fact, has gone farther, and has contended (see Newsweek, Jan. 9, 1956; Jan. 3, 1955; Nov. 22, 1948) that not only is economic forecasting not an exact science yet, but from the very nature of its subject can never be. Ludwig von Mises, in an excellent article in National Review of April 4 last, concisely explained why this is so. Economics, because it deals primarily with human choices, decisions, and actions, cannot be a “quantitative” science. Statistics have a limited value because they necessarily refer to the past only. Knowledge of the future can never be firmly assured by charts or curves “extrapolating” some recent trend. And it is merely of negative help to the speculator or businessman even to guess the future correctly unless his correct guess is shared only by a few. For in the economic field expectations concerning the future themselves help to determine that future.

Yet when all this has been said, all of us must forecast. For we are daily forced to act on some expectation regarding the future. Therefore, a recent carefully compiled and sensibly reasoned study of 76 pages, “Forecasting in Industry,” published by the National Industrial Conference Board, can be read with profit, as much for the negative as for the positive lessons that it has to teach.

As the study declares in a foreword: “Forecasting is a rudimentary and essential element of business. An appraisal of future prospects, whether based on study, judgment or hunch, is inherent to financial planning, appropriating capital, timing purchases, setting production and inventory levels, directing sales. . .”

The study then goes on to discuss the various methods of forecasting actually used in business. Among these are: (1) The “Jury of Executive Opinion Method.” This is merely a process of combining and averaging the views of top executives. It has its merits, though “in some companies the process amounts to little more than group guessing.” (2) The “Sales Force Composite Method.” This is more of a “grass roots” approach, but it also has its disadvantages, one of which is that salesmen often tend to be either too optimistic or too pessimistic according to immediate conditions. (3) Next we come to the use of various statistical methods, including correlation analysis, trend and cycle projections, and mathematical formulas. These have the advantage of being “measurable” and “objective,” and of “quantifying” their conclusions. But they are full of traps for the unskilled and unwary.

Take correlation. As the Conference Board’s study admits: “Good correlations occasionally result from coincidence rather than from logical cause and effect or from any direct relation between the two factors. . . . The great disadvantage of correlation analysis is the danger of relying too heavily upon the statistical method and a tendency when using it to neglect independent appraisal of future events.” But isn’t this a way of admitting that it isn’t so much statistics or mathematical relationships that enable us to forecast the economic future as sound prior reasoning that leads us to select the right statistics?

An instructive case in point is the American Radiator Co. It discovered years ago that there was “a good correlation between the F.W. Dodge figures on residential building contracts awarded and sales of its plumbing products.” It found that residential contracts led its sales by about four months. But, as homes always need plumbing, it was obviously common sense, and a prior knowledge of cause and effect, that led company officials to look for this correlation. To sum up, statistics may often be helpful, but may also be treacherous. They can never take the place of shrewd judgment, good sense—and a large measure of good luck.

And now that you know all the secrets, you too can forecast.