IMMIGRATION PRIOR TO 1890
the probable maximum and minimum limits of the typical lag. An
examination of the curve of imports from 1820 to date (on both the
calendar year and fiscal year basis after 1870) and a comparison
with the curve for total immigration, indicates that Immigration
probably lags behind imports a few months, the period evidently
being nearer six months than a full year and possibly less than six
months.
A similiar serutiny of the pig iron and immigration curves from
1854 to date reveals a similiar tendency for some laginimmigration
of an apparent length of less than one year, as evidenced by the
fact that, in about half of the instances, the troughs and peaks are
reached in the same year by the immigration and pig iron curves,
and, in the other half, the annual immigration movement reaches
the corresponding maximum or minimum a year later than pig iron.
It remains to examine the question of lag more closely with the
aid of quarterly and monthly data and by numerical computation
of the allowance for lag which gives the highest degree of correlation
between immigration and industrial activity.
Numerical Computation of the Typical Lag.
In the usual statistical terminology, the arithmetic evidence of
the lag giving the closest correspondence between the curves may
be expressed by the following summary:
Correlation coeffcient when the gwen number of years
Period compared lag is imputed to immigration as compared with pig iron.
No lag One year Two years
1857-1914 +.64 +.35 —_ 3
1857-1919 +.51 +.27 —
1872-1914 +.78 +.48 J
This numerical interpretation of the statistical evidence, in terms
of the coefficient of correlation, indicates that the closest relation-
ship is found between pig iron and immigration when the years
between the Civil War and the Great War are selected for study and
it is assumed that such lag as may exist is less than one year.
COMPARISONS BASED UPON QUARTERLY DATA:
1868-1889
Male Immigration.
Quarterly immigration data are available by sex beginning with
the quarter ending September 30, 1856. For reasons previously
5
