On a Method of Statistical Business-Cycle Research; a reply (to Keynes)

Also published in: A.H. Hansen and R.V. Clemence (Eds), Readings on Business Cycles and National Income, Allen & Unwin, London, 1953, pp. 341-354, in: Adrian C. Darnell (Ed), The History of Econometrics Vol. 2, Edward Elgar, Aldershot, 1994, pp. 214-227 and in: Dale J. Poirier (Ed), The Methodology of Econometrics Vol. 1, Edward Elgar, Aldershot, 1994, pp. 58-71

In the Economic Journal of September 1939, p. 558, Mr. Keynes discusses the method of statistical business-cycle research used in Vol. I of the League of Nations publication of which I am the author.1 Mr. Keynes has serious objections and numerous questions. Although part of these objections and questions have been answered in the second volume, which has appeared recently, there remain a number of points which I think it is worth while to discuss separately. I shall follow Mr. Keynes's argument exactly in the order in which he gave it. 2. To begin with, Mr. Keynes formulates a number of conditions which, in his mind, must be fulfilled in order that the method of multiple correlation analysis may be applied. With the formulation he gives on p. 560-viz. that " the most he may be able to show is that, if they (i.e., certain given factors) are verm causce, either the factors are not independent or the correlations involved are not linear, or there are other relevant respects in which the economic environment is not homogeneous over a period of time-I find myself only partly in agreement.