MALE‐FEMALE EARNINGS DIFFERENCES: A FURTHER ANALYSIS*
| Published date | 01 March 1976 |
| Author | B. Chiplin,P. J. Sloane |
| Date | 01 March 1976 |
| DOI | http://doi.org/10.1111/j.1467-8543.1976.tb00039.x |
British
Journal
of
Industrial Relations
Vol.
XIV
No.
1
MALE-FEMALE EARNINGS DIFFERENCES:
A FURTHER ANALYSIS*
B.
CHIPLINT
AND
P.
J.
SLOANES
IN a recent issue of this journal' we attempted to assess the relative importance of
unequal payment and unequal employment distribution in explaining the gross
male-female earnings differential in Great Britain by making use
of
the
1971
New Earnings Survey (N.E.S.). Assuming levels of employment remain unaltered
this can be assessed by two simple standardising procedures. First, one may
assign to females in their existing occupations the level of male mean earnings
within each occupation. Secondly, one may redistribute females to accord with
the 'expected' occupational distribution on the basis of the proportion of the sexes
in the total labour force. On this basis, the evidence suggested that unequal pay-
ment was far more important than unequal occupational or industrial distribution
in explaining the gross earnings differential between the two groups. Using the
same methodology similar results were obtained on an industrial basis by AddisonZ
for four European countries (Germany, Italy, Holland and Belgium). Since
the earlier British analysis was conducted a more detailed occupational
breakdown has become available in the New Earnings Survey and the purpose of
this note is to analyse whether or not
this
changes the earlier conclusions and to
consider some possible qxplanations of the observed results.
In
1973
the occupational classification in the N.E.S. was based for the first
time on the recently introduced list of Key Occupations for Statistical Purposes
(KOS) which in turn was derived from the new Classification
of
Occupations and
Directory of Occupational Titles (CODOT). Under the occupational classification
used in the
1970, 1971
and
1972
N.E.S.s there were sixteen main groups and
189
occupations, whilst under the new coverage there are eighteen main groups and
over
400
individual occupations (derived from the
3,500
occupations under
CODOT). It is estimated broadly that about
80
per cent of the total working popu-
lation are to be found in the KOS classification. In addition to the above change
the
1974
N.E.S. (like that conducted in
1970)
included questions on earnings by
age and this allows estimates of the effect on the gross earnings differential of age
differences between the sexes. Table
1
gives the earlier results obtained from the
1971
survey for comparative purposes and the new estimates derived from the
1973
and
1974
surveys. It can be seen that the broad conclusions of our earlier
paper that the earnings adjustment considerably outweighs that relating to oc-
cupations still hold for both
1973
and
1974
at the finer level
of
occupational dis-
aggregation that the N.E.S. data now allow. Thus at the minimum the earnings
adjustment accounts for no less than
63
per cent of the gross amount. However,
the relative importance of unequal earnings and unequal employment distribution
varies both according to the type of N.E.S. data used and the year in which the
calculations are made.
A
priori
the expectation would
be
that the occupational ad-
justment would be less important in the type A case where the occupational
classification is relatively broad. However, finer occupational categories, which
provide both male and female earnings data, cover a much smaller proportion of
*We are grateful to
Alun
Reynolds for calculating the figures contained
in
Table
1
and for
general research assistance and to the Social Science Research Council for a research grant which
has enabled
us
to study
this
area
in
some detail.
t
Lecturer
in
Department of Industrial
Economics,
University of Nottingham.
$
Professor
in
Department of Economics and Management, Paisley Technical College.
77
To continue reading
Request your trial