THE IMPACT OF UNIONS ON PRODUCTIVITY
| Author | A. H. Barnett,J. T. Addison |
| Date | 01 July 1982 |
| Published date | 01 July 1982 |
| DOI | http://doi.org/10.1111/j.1467-8543.1982.tb00093.x |
THE IMPACT
OF UNIONS
ON
PRODUCTIVITY
J.T.
ADDISON*
and
A.H.
BARNETT**
I
INTRODUCTION
The twin issues
of
productivity (and growth) and union impact on resource
allocation are among the least satisfactory areas of mainstream economic research.
Recently, a new and highly controversial literature associated with ‘Harvard School’
analysts has exploited deficiencies in the basic neoclassical theory
of
production to
argue that unions are,
ceteris paribus,
associated with higher productivity
(vulgo:
unions cause higher productivity). While the studies in question may be interpreted as
restating the important hypothesis that ‘internal organisation matters’, their primary
emphasis has been empirical and has focused upon unionism’s
net
impact on
productivity.
Interestingly, this new and largely empirical literature was anticipated in this
Jburnal
by Pencavel
(1977)’,
who pointed to a significantly
negative
impact
of
unionisation on productivity in British coal mining,
1900-13.
Pencavel
(p.
145)
offered
the conjecture that his particular estimate of the output
loss
attributable to unionism in
British coalmining was ‘a lower bound on that prevailing in British industry today.’
Given the unsatisfactory state
of
productivity and growth theory and the deepening
controversy surrounding unionism’s component contribution,’ the time is ripe for a
review of the evidence and clarification
of
the theoretical and measurement issues
involved. This is the basis
of
the current paper.
While welcoming the Harvard School literature because of its emphasis on the
regularisation
of
trading relations
and
its substitution
of
systematic study for anecdotal
evidence and casual empiricism, we shall nonetheless argue that the empirical
evidence is less clear cut than the authors suggest and the theoretical content of the
underlying model somewhat obscure. Our suggestion is that unionism
is
an
endogenous variable, not determining productivity
in those cases where
a
significantly
positive effect
of
unionism is obtained
but. rather, being jointly determined with
productivity.
The plan of the paper is as follows. First. the basic ‘theory’
of
unionism and
productivity is discussed. Second, the empirical evidence with
a
bearing
on
the issues
is
evaluated. Third, some wider theoretical and empirical observations are entered.
Finally, the threads
of
the preceding arguments are drawn together.
I1
THETHEORY
The concept
of
productivity in econoniics and industrial relations literature is
plagued by vagueness and misconception. This
is
notably evidelit in literature dealing
with the effects of unionisation on labour productivity. In the interests of clarity,
therefore, we shall begin with an elementary discussion of the general causes
of
union-nonunion labour productivity differentials.
Differences
in
the
measured
marginal
or
average product of labour can be
categorised under two headings; (i) those attributable to differences in relative wages
or
prices (including wage induced technical differences), and (ii) those generated hy
*Associate
Professor
of
Economics, University
of
South
Carolina
““Assistant
Professor
of
Economics, University
ol
South
Carolina
14.5
146
BRITISH JOURNAL
OF
INDUSTRIAL RELATIONS
differences in the form
or
parameters
of
production functions.
Of
these two
categories, only the latter represents differences in the ability
of
a given labour force to
generate output. In general, unionisation will affect both wages and production
functions. Unfortunately, the effects of wage differentials on measured productivity
confound attempts to estimate the magnitude and direction
of
true productivity
differentials induced by unionisation.
To
be more specific, although the effects
of
unions on wages is still subject to doubt,
much empirical evidence suggests that unions increase wages for union members
relative to their nonunion counterparts. Other things equal, higher wages in unionised
plants will reduce the quantity of union labour demanded. This occurs for two reasons.
First, higher wages increase the cost of production and, therefore, reduce the quantity
of output associated with profit maximisation. Second, higher wages encourage
managers to substitute capital for labour. Hence,
if
the marginal product
of
labour
falls with additional employment, then measured marginal and average product will
rise with higher wages-not because labour is more productive in any meaningful
sense, but simply because there is less
of
it
relative
to
capital.
Additionally. higher wages in unionised plants may induce employers to seek out
‘better quality’ workers (i.e. workers with more human capital) and, other things
equal, better quality workers generate more output. However. as with other wage
effects, this does not mean that union workers are more productive because they are
unionised. Rather, they are more productive because higher wages
for
unionised
workers cause a reallocation of higher quality workers from nonunion to union jobs.
Apart from wages effects, which do not represent an increase in the ability
of
a given
labour force to generate higher output,3 there may be X-efficiencies and inefficiencies
associated with organisation which do generate true productivity
differential^.^
The
traditional economic literature has tended in the main to emphasise the
inefficiencies
introduced by unionisation (Rees,
1963)’.
These inefficiencies stem from union
imposed work rules and other
job
regulatory practices that limit managerial flexibility.
The resulting output losses have conventionally been added to those emanating from
the relative wage effects
of
unionism to provide
a
static
measure
of
the welfare losses
associated with unionism.’
However, recent empirical work has challenged the conventional approach in
arguing that unions produce X-efficiencies through the expression of a ‘collective
voice’ and that these efficiencies more than offset any union induced inefficiencies.
In
short, unions produce
net
X-efficiencies.
Obviously, whether unionism’s impact on productivity is negative
or
positive is an
empirical question. Theory can however guide us in the construction
of
empirical
models. Perhaps most importantly, empirical studies must distinguish between true
productivity differentials attributable to union induced differences in production
functions and those differences attributable
to
management responses to
union-nonunion wage differentials.
These issues will be explored below. Meanwhile, let us amplify the theory
underpinning the new view
of
unionism retailed by Harvard analysts. The basic model
is contained in Freeman’s
(1976)’
labour market application
of
Hirschman’s
(1970)’
exit-voice dichotomy.’ By providing workers with a ‘voice’ at the workplace, as an
alternative to ‘exiting’
(or
quitting), it is argued that unions can indeed affect positively
the working
of
the economic system. More specifically, the efficiencies resulting from
the expression
of
a collective voice can offset the inefficiencies resulting from a higher
union wage.
Let us briefly itemise the advantages claimed for the collective voice mechanism
before delving
a
little more closely into the theory. First, unions can induce managers
to alter methods of production and adopt more efficient policies. Second, unions
collect information about the preferences
of
all workers, thus permitting the firm to
To continue reading
Request your trial