ARE LABOUR MARKETS COMPETITIVE?—A CASE STUDY OF ENGINEERS*
| Author | J. D. Mace,G. C. G. Wilkinson |
| Date | 01 March 1977 |
| Published date | 01 March 1977 |
| DOI | http://doi.org/10.1111/j.1467-8543.1977.tb00068.x |
British Journal
of
Industrial Relorions
Vol.
XV
No.
1
ARE LABOUR MARKETS COMPETITIVE?-A CASE STUDY
OF
ENGINEERS*
J.
D.
MACE? and G. C. G. WILKINSONS
THE neoclassical theory of wage determination provides a convenient framework
for the analysis of labour markets. The literature contains
no
conclusive refuta-
tion of the neoclassical theory but, on the other hand, corroborative evidence in
favour of it is usually both weak and indirect. The theory has in fact proved ex-
tremely difficult to test.’ However, most of the evidence that has
so
far been
analysed relates to blue-collar workers.2 It has been maintained that white-collar,
and especially professional, labour markets tend to be characterised by more
information, greater mobility, less unionisation, and general training and skills
appropriate to a wider variety of employers than is found in bluecollar
market^.^
In other words, we expect neoclassical theory to be particularly applicable to pro-
fessional labour markets.
Our
aim in this paper is to examine that presumption with reference to the
labour market for engineers. If the neoclassical theory
is
to claim any generality,
it should stand up particularly well for this market. We begin with an explanation
of the methodology employed in this study. We then consider the notion of the
competitive labour market and identify specific predictions of the neoclassical
theory of labour markets. These predictions are examined in the light of
our
data.
In the final section of the paper, we discuss the implications of
our
findings and
conclude, however tentatively, that the evidence we have collected
on
the labour
market for engineers is not consistent with the predictions of competitive wage
theory.
METHODOLOGY
Studies of the demand and supply
of
engineers have sometimes disregarded
neoclassical economics altogether, choosing instead to work within the framework
of ‘manpower ~lanning’.~ Nevertheless, the bulk of the work
on
engineering
labour markets has been carried out within the neoclassical framework. Examples
of the latter are provided by Blank and StigleP and W. Lee Hansen6
in
America,
and
V.
A. Richardson’ in this country. Richardson examines changes in the
salaries of British engineers relative to other categories of workers in the
U.K.,
and concludes that there was
no
shortage of engineers in the
1960s.
He justifies
this method of examining labour markets for shortage by asserting that the U.K.
engineering labour market is highly competitive.8
A rather more ambitious study along similar lines was undertaken by Layard
and
other^.^
This
study employed sophisticated econometric methods to discover
*
This paper forms part of a S.S.R.C.-sponsored research programme
on
the match between
education and occupation, carried out at the Research Unit in the Economics of Education at the
University of London Institute of Education. We are greatly indebted to Professor Mark Blaug,
head of the unit,
for
his constructive criticism of this paper, and to Susan Taylor for assisting in the
collection and analysis of data. We are
also
grateful for useful comments from members of the
Lon-
don School of Economics Research Seminar in the Economics of Education, and an anonymous
referee.
t
Lecturer in the Economics of Education University of London Institute of Education.
$
Economist at the British Steel Corporation, formerly Research Consultant at the Economics
of
Education Research Unit, University of
London
Institute of Education.
1
2
BRITISH JOURNAL
OF
INDUSTRIAL RELATIONS
whether the relationships derived from
a
neoclassical model can
in
fact be obser-
ved. The findings were almost totally negative. The reasons for this failure pre-
sumably stem either from errors in the data, or from the mis-specification of the
model, or from the inappropriateness of the neoclassical model to engineering
labour markets. Whatever the explanation, the failure of the Layard study to yield
statistically significant results suggests the need for
a
more fundamental examina-
tion of the engineering labour market at the micro level. Such an approach rules
out the use of published data, which besides being highly aggregated and notori-
ously unreliable, are also not available for many of the relationships that need to
be examined.
There appears to be a presumptive case for employing
a
method
of
analysis
that does
not
rely
on
aggregate data and that is not rigidly tied to one set of theo-
retical preconceptions. This presumption
is
reinforced when we take account of
two problems of labour markets that have been recognised
in
virtually every
stan-
dard textbook. The first is the existence
of
imperfections, such
as
elements of
monopoly and imperfect information, and the second
is
the difficulty of inter-
preting wage data in the face
of
the non-pecuniary costs and benefits of different
occupations, More recent theoretical work has emphasised the latter problem of
interpreting wage data. Becker," for example, has argued that observed wages
and marginal products will diverge when specific training is undertaken. The pro-
blem of investment in training, either by individuals or by firms, has been further
examined in the work of Mincer" and Rosen.'*
The concept of the 'internal labour market' has posed further questions for the
interpretation of wage data.13 Many firms fill vacancies by means of internal pro-
motion, only recruiting workers from the external labour market at one or two
'ports
of
entry', usually at the bottom of the occupational hierarchy. In con-
sequence, workers within the firm develop vested rights in their jobs, which
soon
come to be taken for granted by both labour and management. An 'internal
labour market' may, therefore, be defined as 'an administrative unit
.
. .
within
which the pricing and allocation of labour is governed by a set
of
administrative
rules and procedures', and not by market forces.I4 Where insulated labour mar-
kets develop, wage rates are determined with respect to groups and not with res-
pect to individuals as in neoclassical theory. Further doubts about the validity of
the neoclassical paradigm are to be found in the work of Spence15 and Thurow."
It is for these reasons that we have come to believe that what is needed now to
carry the argument further is
to
eschew aggregate, published wage data and
instead to collect individualised data for well-identi fied firms.
The argument for purposively collected data is strengthened by a practical
consideration. To secure insights into the hiring practices of firms, as well as to
collect data
on
the employment history of engineers, it is necessary to secure the
cooperation of firms. A few we approached were not prepared to take part in our
study at all. Others were prepared to provide
us
with
a
non-random sample of
engineers that, in their view, represented the manpower structure of the firm, and
to allow
us
to
interview both engineers and management, but they were under-
standably reluctant to reorganise their personnel records
so
as to allow
us
to draw
a genuinely random sample of employees. We were driven, therefore, by prac-
tical circumstances to adopt the social-survey approach
of
sociology, rather than
the statistical approach now employed in most economic research."
We would, however, be the first to admit that there are certain problems associ-
ated with
a
case study investigation. For example, 'noise', roughly defined as
the
influence of non-random systematic forces not captured in the theory, makes the
small case study
a
rather unsatisfactory technique for examining economic for-
ces." The law of large numbers favours the use of random sample techniques.
Nevertheless, aggregation may often conceal more than
it
reveals, particularly in
To continue reading
Request your trial