LAWS OF NATURE IN ECONOMIC AFFAIRS 225

tions in activity, and adds that these fluctuations are peculiar to countries
organised on a business basis, that they appear in all such countries, that
they tend to develop the same phase at nearly the same time in different
countries, that they follow each other without intermissions, that they are
affected by all sorts of non-business factors, that they represent pre-
dominant rather than universal changes in trend, and that, while they vary
in intensity and duration, the variations are not so wide as to prevent our
identifying different cases as belonging to a single class of phenomena.'1

In the intellectual history of an industrial Western Society the pheno-
menon of trade cycles had been discovered empirically from direct social
observation before it had been confirmed statistically from patterns dis-
cernible in collections of data.2 The earliest known description of it had
been given in A.D. 1837 by a British observer, S. J. Loyd, alias Lord
Overstone (vivebat A.D. 1796-1883), in the light of experience in Great
Britain since the time of the Napoleonic Wars;3 and there was nothing to
suggest that this empirically demonstrated relativity of the phenomenon
to the social milieu in which it presented itself was not one of its intrinsic
features. In a book first published in A.D. 1927 an American student of
business cycles declared his belief that 'the characteristics of business
cycles may be expected to change as economic organisation develops'.4
On the basis of 'business annals' compiled by another American scholar,
W. L. Thorp,5 from non-statistical evidence for the economic history of
the West in its Industrial Age, a third American scholar6 had descried,
within the longer Time-span illuminated by this less precise but more
widely ranging kind of information, a secular tendency towards a pro-
longation of the wave-length of 'business cycles' of the shortest kind. In
the light of these data, F. C. Mills had calculated that the mean wave-
length of a short cycle was 5-86 years in the early stages of industrializa-
tion, 4/09 years in a subsequent stage of rapid economic transition, and
6*39 years in a succeeding state of relative economic stability; and, in
the judgement of the first of the three American scholars just cited, ex-
pressed in a book first published in A.D. 1927,' 'there can be little doubt
that the average duration of business cycles has undergone secular changes
in the countries for which Thorp has compiled the longest records'. In a
book published in A.D. 1939 a German economist, in whose belief 'inno-
vation' was 'the outstanding fact in the economic history of Capitalist
Society'8 and was at the same time the cause of the cyclic fluctuations in

1 Mitchell, W. C.: Business Cycles, the Problem and its Setting (New York 1930,
National Bureau of Economic Research, Inc.), pp. 458-9.

a See Hawtrey, R. G.: 'The Monetary Theory of the Trade Cycle and its Statistical
Test*, in The Quarterly Journal of Economics, vol. xli (Cambridge, Mass. 1927, Harvard
University Press), p. 471. 'No statistical finding can ever prove or disprove a proposition
•which we have reason to believe by virtue of simpler and more fundamental facts
(Schumpeter, J. A.: Business Cycles (New York 1939, McGraw-Hill, a vols.), vol, i, p. 33)-

3 See Hawtrey, ibid., pp. 471-2. + Mitchell, op. cit., p. 413- ,

s Thorp, W. I*.: Business Annals (New York 1926, National Bureau of Economic
Research, Inc.).

6 Mills, F. C., in The Journal of the American Statistical Association, December 1936,

7 Mitchell, W. C.: Business Cycles, the Problem and its Setting (New York 1930,
National Bureau of Economic Research, Inc.), p. 415.

8 Schumpeter, op. cit., vol. i, p. 86. The same authority, in op. cit., vol. i, p. 223,
defines Capitalism as 'that form of private property economy in which innovations are

B 2015. EC I