UNIONS AND THE DISTRIBUTION OF EARNINGS
| Published date | 01 July 1982 |
| DOI | http://doi.org/10.1111/j.1467-8543.1982.tb00094.x |
| Author | David Metcalf |
| Date | 01 July 1982 |
UNIONS AND
THE
DISTRIBUTION
OF
EARNINGS'
DAVID
METCALF"
AT the height
of
the 'winter
of
discontent' (Feb. 1979) the joint statement issued by the
TUC and Government stated that 'The trade union movement acknowledges there are
problems in sectional collective bargaining between competitive groups in
our
society
and not every one can improve his standard
of
living relative to someone else'. This
impact of unions on the distribution of earnings has long been a matter
of
concern. It is
now possible to add a little evidence
to
the concern. This paper analyses pay dispersion
among covered and uncovered workers and the impact of unions on pay differentials
by sex, race, occupation and industry. It seems that union activity in the
UK
narrows
the pay structure by occupation and by race but widens the inter-industry wage
structure. The data refer to individuals receiving pay. Non-earners such as rich
wealth-holders who choose not to work and the unemployed are excluded.
EARNINGS
DISPERSION AMONG
COVERED
AND
UNCOVERED WORKERS
Earnings dispersion
-
measured
by
the coefficient
of
variation
-
among male
workers covered by collective agreements is substantially lower than the dispersion
among uncovered workers. Table
1
shows that in 1978 in each
of
eight socioeconomic
TABLE
1
Collective agreement coverage and pay dispersion. males
1978
Socioeconomic
group
NON-MANUAL
Managerial
Professional
Intermediate
Junior
MANUAL
Foremen
Skilled
Semiskilled
Unskilled
11
8525
8750
9990
7953
4880
2347
I
13705
6871
Percent covered by
(National agreemenr only)
plus (National
plus
supplemenrary agreement)
29.1
54.1
55.3
59.0
52.9
67.3
61.9
68.8
Coefficient
of
variarion
of
weekly earnings
covered uncovered
workers
workerc.
0.378 0.636
0.345 0.475
0.293 0.423
0.294 0.402
0.303 0.338
0.290 0.350
0.29
1
0.369
0.304 0.392
Source:
M Gregory and
A
Thomson" Tables
1
and
4
and appendix.
Nore:
The coefficient
of
variation refers to actual weekly earnings
not
the logarithm
of
weekly
earnings. Actual earnings are likely
to
be more highly skewed than logarithm earnings.
Also
Gregory and Thomson do not provide information on higher moments
or
whether the
distribution
of
the covered sector is more highly skewed than that
of
the uncovered sector.
groups the coefficient
of
variation
of
weekly pay is lower for covered than uncovered
workers. This reflects either or both
of
two forces. First, covered workers or jobs may
be more homogeneous than the uncovered. Second, collective bargaining policies
pursued by unions and employers may have reduced the dispersion among covered
*
Professor
of
Economics Kent University
163
To continue reading
Request your trial