INFLUENCE OF ECONOMIC CONDITIONS 173
Certain limitations of the method used in constructing these
composite indices will, of course, be recognized. For one thing,
they may convey an exaggerated impression of the homogeneity of
cyclical fluctuations in the given country. The averaging of
several series obscures differences which may not be altogether unimportant
 in their effects on migration tendencies. Furthermore,
the use of moving averages in estimating the trend, in some instances,
such as toward the end of a depression period which is followed by a
rapid recovery, results in extending the computed depression period
beyond the time when in absolute terms the several economic
phenomena are beginning to show signs of recovery. For example,
in 1879, the production of pig iron in Germany was 2,227,000 tons,
as compared with only 2,148,000 in 1878, but because this increase
is less than the computed increment to the trend line from 1878 to
1879, the movement in the cycle curve from 1878 to 1879 appears as
a decline for pig iron.
This tendency may account for occasional discrepancies between
the evidence presented by the cycle curve and the descriptive accounts
 of variations in business conditions. Inasmuch as the cycle
curves for migration and the industrial composites are computed by
the same method, occasional minor discrepancies between interpretations
 which rest on the unadjusted statistics and those based
upon the cycle curves do not necessarily affect the validity of
comparisons between the cycle curves.
International Similarities in Business Cycles (Chart 40)
A comparison of the cycles of economic conditions in the United
States, the United Kingdom, Germany, and Italy reveals that the
fluctuations in the first three countries show marked similarities,
although the agreement is by no means complete or invariable. In
each of these three countries there was a boom in the early seventies
followed by a decline in the middle part of the decade which continued
 well toward the end. Likewise, each experienced a boom in
the early eighties, a decline near the middle of the decade, and a
recovery in the late eighties or in the first years of the following
decade, then a depression in the middle nineties, and a recovery
again at the turn of the century. In each, a depression appears in
1904, a marked rise culminating in 1907, and a sharp decline in 1908,
followed in a year or two by the beginning of a recovery which,
during the remainder of the period prior to 1914, is not broken in
any of the three countries by a reaction as severe as that of 1908.