The Impact of Industrial Relations Legislation on British Union Density
| DOI | http://doi.org/10.1111/j.1467-8543.1990.tb00360.x |
| Date | 01 July 1990 |
| Author | Jeffrey Pelletier,Richard Freeman |
| Published date | 01 July 1990 |
British Journal
of
Industrial
Relations
28~2
July
19900007-1080
$3.00
The
Impact
of
Industrial Relations
Legislation on British Union
Density
Richard Freeman
*
and Jeffrey Pelletiert
The unionized share of the work-force changed markedly in the UK
between the
1970s
and
1980s.
In the
1970s
density rose steadily, making
the
UK
the most heavily organized large
OECD
country.
In
the
1980s,
by
contrast, density
fell
by
1.4
percentage points per annum
-
a faster drop
than
in
the rapidly de-unionizing USA
or
Japan. What explains this turn-
around
-
the severe recession
of
the
1980s?
shifts in the composition of
employment from unionized manufacturing
to
services? the Thatcher
government's industrial relations legislation?
In this paper we investigate these questions with a quantitative analysis
of
1945-86
changes in British union density. In contrast to studies that
concentrate on cyclical determinants of unionism (Bain and Elshiekh
1976,
Carruth and Disney
1988,
Booth
1983),
we focus on industrial
relations legislation. We develop an index
of
the favourableness of labour
laws to unionism and relate
it
to changes in density
in
time-series regres-
sions that control for inflation, unemployment and the manufacturing
share
of
employment, among other variables. As a further test, we de-
velop an analogous labour law index for Ireland, whose industrial rela-
tions system
is
similar to the UK's and which experienced a similar severe
1980s
recession but did not pass new laws to weaken unions; and we
contrast changes in density between the countries with differences in
industrial relations law. Our major finding is that the Thatcher govern-
ment's labour laws caused much
of
the
1980s
fall in British union density.
We present the evidence
for
this claim
in
three stages. Section
1
lays
out
the facts
of
changing union density
in
the UK and Ireland and exa-
mines
structural explanations
of
the
UK
changes. Section
2
discusses the
1980s
UK
labour laws and develops an index
of
their likely impact on
unionism. Section
3
presents our econometric analysis of the
UK
time-
series data.
'Haward University and the National Bureau
of
Economic
Research
tUnited Food and Commercial Workers
142
British Journal
of
Industrial Relations
1.
CHANGES IN UNION DENSITY
While at first blush measuring the change
in
union density would appear to
be a simple matter of obtaining relevant counts, it is in fact more complex,
for two reasons. First, there
is
a range of plausible choices
of
numerator and
denominator
in
any union density statistic. In the numerator one can use
union membership
or
workers covered by collective bargaining
-
figures
that differ, because some unions do not win recognition and some workers
reject unions even in organized workplaces. In the denominator one can use
employment,
or
a subset thereof,
or
the labour force. Each measure yields a
somewhat different picture of levels and changes in union representation,
as
Kelly
(1987)
has stressed.
For
the
UK,
collective bargaining coverage
exceeds the proportion
of
workers who are unionized, which in turn exceeds
the union proportion of the labour force because few unemployed unionists
maintain membership.
'
By dividing cyclical union employment by the less
cyclical labour force, density based on the labour force shows greater
changes over the cycle than density based on employment. Between
1979
and
1985,
for example, the ratio
of
union membership to the labour force
dropped by
11.1
points compared with a
7.8
point fall in the ratio
of
union
membership to employment.'
As
we are concerned with secular changes in
density, we concentrate on membership/employment. Regardless
of
the
measure, however, there is
no
gainsaying that density fell markedly in the
1980s.
The second problem in measuring union density relates
to
the member-
ship statistics themselves. In the UK (and in Ireland), membership data are
obtained from unions, which differ in their methods
of
counting: some
unions include retirees, some are slow to drop from their books workers who
are unemployed and do not pay dues, some have computerized data files
while others do not, etc. (Walsh
1985;
Eurostat). In the
UK
some unions
exaggerate membership to maintain a high profile
on
a national
level,
biasing upward the recorded density and potentially minimizing declines in
density. In Ireland the Registrar
of
Friendly Societies obtains membership
from recognized unions, some of whom report irregularly, and excludes
other labour organizations, creating potential errors
in
the data though with
no
obvious bias in trends3 These problems notwithstanding, we calculated
densities from official membership figures from
1945
to the
1980s
for both
countries, and obtained the pattern shown in Figure
1:
increases in density in
the UK
in
the late
1960s-1970s
followed by sharp drops in the
1980s
and a
gradual trend upwards in Irish density. These patterns are the phenomenon
of concern to this paper.
Potential Causes
of
Change
One widely cited cause
for
the drop in British density is the loss of
manufacturing jobs that characterized the first Thatcher term in office.
To continue reading
Request your trial