COST OF LIVING INDEXES COMPARED 105
for July. Since then the average difference has been less
than 59.
The difference between the two series for the cost of living
as a whole 1s repeated for the separate items. Thus, prior
to 1918, increases for clothing and for fuel and light were
greater in the National Industrial Conference Board series
than in that of the Bureau of Labor Statistics and lower
afterwards until the end of 1924; rent increases in the Na-
tional Industrial Conference Board series have always been
more and sundries cost increases less, than in the Bureau
of Labor Statistics series. The comparative trends of the
two sets of lines for the separate items are shown in Charts
4a to 4k inclusive.?
It was pointed out in Chapter I that certain fundamental
considerations connected with the making of index numbers
might affect the results. In succeeding chapters these
principles were illustrated in descriptions of the basic
budgets and the methods of collecting and combining current
price data for certain of the more important cost of living
index numbers. In order to determine whether or not there
are fundamental differences between the National Industrial
Conference Board series and the United States Bureau of
Labor Statistics series, these circumstances must be carefully
considered. Although both series of numbers relate to the
standard of living of working class families, the basic bud-
gets may differ as to (a) the period to which the consump-
tion weights relate, and (b) the goods and services used as
samples. The method of collecting and combining the
, 1Although the National Industrial Conference Board number is recorded as an
increase above July, 1914, and the Bureau of Labor Statistics number as the increase
above average prices in 1913, these bases are practically identical, and the series
may be compared as they stand. This is because, as already explained in Chapter
IL, the difference between the average retail price level in 1913 and July, 1914, as
measured in a balanced cost of living budget was due almost exclusively to an
Increase in food prices, and this is taken account of in the series of the National
Industrial Conference Board, since the Bureau of Labor Statistics retail food price
Index number, unconverted to a July, 1914 base, has been used by the Board as a
measure of food prices in its series. To all intents and purposes, therefore, the
Increase reported by the Bureau of Labor Statistics since 1913 exaggerates only
slightly, if at all, what would have been the increase if the seties had been reduced to
a July, 1914 base for comparison with the increase reported by the National Indus-
trial Conference Board since July, 1914.
2 The figures from which these charts are drawn are given in Table 1 (p. 30) and
Table 4 (pp. 66-67).