A MODEL OF COLLECTIVE BARGAINING FOR U.K. AND U.S. MANUFACTURING: A COMPARATIVE STUDY
| DOI | http://doi.org/10.1111/j.1467-8543.1979.tb00963.x |
| Author | Davinderf Singh,C. Glynh Williams |
| Published date | 01 November 1979 |
| Date | 01 November 1979 |
RESEARCH
NOTE:
A MODEL
OF
COLLECTIVE BARGAINING FOR
U.K.
AND
U.S.
MANUFACTURING: A COMPARATIVE STUDY
DAVINDER
SINGH~
AND
C.
GLYN
WILLIAMS$
A
collective bargaining model developed by Tylecote for U.K. manufacturing wage
changes appeared in the March
1975
issue
of
this journal.’ The model provides
a
remarkably good fit for the period
1957-70
and its success provides a reason for
extending it
to
other contexts, in this case, to the U.S. manufacturing sector.
Tylecote tests three equations, designating them as the ‘core’
of
the theory, the ‘full’
theory, and the ‘agnostic’ model. The equations and the description
of
the variables are
given below.
The
‘core’
of the theory
W,
=
a,
+
a2
Cr
+
a3
L
+
a,
S,.
The ‘full’ theory, apart from
R
Wm
=
a1
+
a2
Cr
+
a3
L
+
a,
S,
+
a5
E,‘+
a6
Wm.
The agnostic’ model
Wm
=
a1
+
ua
Cr
+
a,
L
+
a,
S,
+
a5
E,
+
a6
W,
+
a,
R.
Where:
W,
=
The percentage increase
in
the average hourly earningsof male manual workers over twenty
years of age in minimum list leading industries, over the period
1954-70;
C,
=
The percentage of the total salesof each industry made by the five largest firms in
it,
in 1963;
S,
=
The ratio
of
total net output, minus total earnings
of
manual workers, to total earnings
of
manual workers, in 1963;
L
=
The ratio of total sales to total earnings
of
manual workers in 1963;
E,
=
The average number
of
male manual workers over twenty years old
per
establishment;
W,
=
The average hourly earnings, per male manual worker over twenty,
in
1954;
R
=
The percentage change
in
employment
of
manual workers, 1954-71.
In our attempt to fit Tylecote’s equation for United States data, we were unable to
obtain data on all variables which exactly resembled his for British manufacturing?
Whenever differences existed, the following proxies were computed:
(1)
Our
data refer
to all production workers rather than adult males;
(2)
We have used four-firm rather
than five-firm concentration ratios;
(3)
Tylecote weighted
his
variables by the average
number
of
workers
per
establishment, whereas we weighted
ours
by the average firm
size
of
the largest eight firms
(to
derive these weights we calculated average firm
size
by
multiplying production worker man-hours by the concentration ratio and then dividing
it by four.
The results for the ‘core’ of the theory model are shown in Table
1.
We have also
reported in column
(2)
the unweighted coefficients
of
the wage equation, because the
unweighted equation
is
superior to the unweighted.3 All the coefficients for Britain are
more highly significant than those
for
the United States. However, the major difference
liesin the value of
Rz,
0.85
for Britain as opposed to
0.10
or
0.09
for the United States.
By the
Rz
criterion, the results do not improve significantly for the ‘full’ theory or the
‘agnostic model (see Table
2).
As in the case
of
the ‘core’ theory, the majority
of
the
variables are statistically significant, but the overall regression fit remains unsatisfac-
tory. In addition there is
one
interesting observation.
For
the United States the coeffi-
cient on the initial wage equation is negative and statistically significant at the
-01
level.
This supports the belief that the low wage industries received a larger percentage wage
increase than the high wage industries, thus causing the wage dispersion to narrow for
the
1958-72
period. However
no
such pattern is obvious for Britain.
t
Murray State University, Kentucky.
$
Professor
of
Economics, University
of
South Carolina.
386
To continue reading
Request your trial