EMPLOYMENT DISCRIMINATION AND THE EMPLOYER TASTES MODEL
| Published date | 01 June 1984 |
| DOI | http://doi.org/10.1111/j.1467-9485.1984.tb00471.x |
| Date | 01 June 1984 |
| Author | JOHN SHOREY |
Scottish
Journalo/Polirical
Economy,
Vol.
31,
No.
2,
June
1984
Q
1984
Scottish
Economic
Society
EMPLOYMENT
DISCRIMINATION
AND
THE
EMPLOYER
TASTES
MODEL
JOHN
SHOREY*
University
College,
Card@
I
INTRODUCTION
The conventional economic model of discrimination, due largely to Becker, is
based upon employer tastes and competitive markets, and predicts wage
discrimination. Workers of identical productivity are paid different wage rates
to perform identical tasks. The model predicts that under certain circum-
stances the labour force may be segregated across firms, but it does not predict
the segregation within firms across different jobs which is observed in the
labour market. More important, the model implies that discrimination
is
merely a transitory phenomenon, for in the long run competition between
employers in product markets drives out discriminating firms. Only when we
allow for non-competitive behaviour in product markets, or severe trans-
actions costs and lags,
do
we arrive at a model whose predictions are consis-
tent with the persistence of wage differences between similar workers observed
in the real world.
In this paper we develop a model
of
employment discrimination using the
employer tastes approach. Non-competitive behaviour is built into the model
from the very outset as a central feature of the labour market. Equally
productive workers are systematically allocated to different grades within the
firm, in which case they earn different wages,
or
to jobs within the same grade
with different performance requirements. The model predicts both segregation
across jobs and an uneven distribution of disadvantaged labour across the
labour market. Discrimination persists
in
the long run because there are firms
who, though they perhaps operate in quite competitive product markets,
possess monopsony power in the labour market in large part arising from the
stance of trade unions and the tastes for discrimination of workers.
Our
approach therefore brings together several different strands of analysis from
the present literature on discrimination.
*With thanks
to
Caroline
Joll,
David Metcalf, the referees and editor
of
this journal.
Date
of
receipt
of
final manuscript:
14
November
1983.
157
158
J.
SHOREY
EMPLOYER
TASTES
AND
WAGE
DISCRIMINATION
Let
us
first briefly outline the conventional employer tastes model of wage
discrimination from which the present analysis draws its inspiration.' In that
model, there are two groups of workers,
A
and B, with identical productive
abilities supplying labour to market
i2.
Let
us
assume the existence
of
J
competitive employers in that labour market, each with a utility function of the
form
:
au
au
an
8~
Uj=U(~,B)j
->O-
j=l
...
J
where
n
is profit and B
is
the number of
B
workers employed. Employers
experience disutility from, having
B
people in their workforce. Implicit in this
utility function is an indifference map representing a trade off for each firm
between profit and B employment at each level of utility. The slope of an
indifference curve, the marginal rate of substitution, is defined
as
the
discrimination coefficient (db). This measures the extent
of
the taste for
discrimination and varies across firms.
If
we assume that the first order partial
derivative of the utility function with respect to
B
employment increases with
the number
of
B employed
so
too does db and the indifference curves are non
linear.
Within the conventional wage discrimination model, the
J
utility maximis-
ing firms, each facing a given market price for its product,
P,
employ that
number
of
B
workers as part of their optimal total employment level that
equates the marginal rate of substitution between profits and
B
employment
with the absolute difference in market wages paid to
A
and
B
workers,
(WA
-
W,). Each firm acts
as
$the wage of B workers were WA(l +dbj). B workers
are only employed by a discriminatory firm if
W,
<
W,.
Specifically, firm
j
hires that proportion of
B
workers as part of its desired labour force such that,
for given
WA
and
W,,
w,
=
W,+dbj
=
I'q
where
q
is the marginal product
of
identical
A
and
B
labour. The firm thus
hires that number of B workers that, at the margin, compensates it in lower
'
The employer tastes model
of
discrimination was developed in Becker
(1971)
and extended in
particular by Arrow
(1972).
A
considerable literature has grown out
of
this analysis;
of
an
empirical kind, e.g. Haessal and Palmer
(1978),
and
of
a critical kind, e.g. Marshall
(1974).
It
has
also stimulated the development
of
alternative models
:
of an informational nature, e.g. Aigner and
Cain
(1977);
of monopsony, e.g. Madden
(1974);
and
of
a radical kind, e.g. Reich
(198
1).
The
A
B distinction here embraces sex, age, colour, race and religious differences. This is not to
deny that certain aspects
of
any analysis
of
discrimination might be more relevant
to
certain
differences than to others, nor that special factors will be relevant to each case. We believe our
analysis has general significance
so
we retain this
useful
and recognised convention.
Linearity
of
the indifference curves gives the result mentioned in the introduction; firms will
employ either an all A
or
all B labour force according to the strength of their tastes, is., it produces
segregation across firms.
EMPLOYMENT
DISCRIMINATION
159
wages for the disutility it experiences from their employment. Each firmj has a
certain demand for B labour at each wage differential reflecting its particular
tastes for discrimination.
The final equilibrium wage differential between
A
and
B
labour in market
i
depends upon the supply of, and the demand of
B
labour. The latter in turn
depends upon the level and distribution of tastes for discrimination across the
J
firms. The larger the
B
group,
ceteris
paribus,
the larger must be the wage
differential between
A
and
B
workers to provide sufficient compensation to
increasingly more discriminatory firms to achieve
full
employment of the
B
labour force. Most firms, in equilibrium, employ both
A
and B people
:
only the
most discriminatory firms hire few
or
zero B people. The analysis implies that
those firms which employ
a
high proportion
of
B
labour, those with
low
tastes
for discrimination, have lower labour costs for identical output levels than
other, more discriminatory firms.
The above analysis can, of course, be applied to all sub labour markets, each
of
which contains labour
of
a different productive capability, different
B
supply
conditions, and employers with different tastes and market positions. The
labour market as a whole is thus made up of various categories of labour
within each of which
A
and B workers are paid differently, the difference
varying with the conditions governing each sub market.
We now turn to consider a model of employment discrimination based upon
the employer tastes approach. But before we do
so
let
us
elaborate on two
other central elements of the model. First, let us identify the main features of
the market assumed in the analysis. It is evident from the wage discrimination
model that discrimination and rigorous competition, as envisaged in
economic theory, are incompatible. The economy in the employment
discrimination model thus differs noticeably from the neo-classical
conception, in particular in having important non competitive elements
in the labour market. Second, let us try to identify somewhat more precisely
what we mean by tastes for discrimination
for
this has a significant bearing
upon the structure of the model.
I11
THE
LABOUR MARKET
The principal characteristics of the labour market analysed in this paper are
(1)
there are two sectors to the market. One sector is made up
of
firms within
as follows
:4
4The kind
of
labour market we describe here has been the subject of much discussion in the
literature recently. There is as yet
no
accepted single model of the dual labour market. Key ideas in
this approach are to be found in Doeringer and Piore (1971) and Piore (1975). There
is
as yet
no
decisive evidence on these ideas, e.g. Osterman (1975), McNabb and Pschacharopoulos (1981).
Dual labour market ideas are assessed in Cain (1975). Our analysis is one possible, plausible
synthesis of existing ideas which follows Wachter (1975), rather than Edwards (1979), in
emphasising the efficiency aspects of internal labour markets in the face of peculiar technological
and institutional constraints.
160
J.
SHOREY
which there are structured internal markets; the other sector is composed
of
firms with unstructured internal labour markets. Each firm in the structured
sector displays an extensive, discontinuous, hierarchical and, to all intents
and purposes, rigid structure of jobs, both manual and non manual. Jobs
within the firm are arranged in grades, each grade requiring a different but
significant amount of both general and specific skills of workers who are to fill
it. Each job in fact requires a clear minimum productive performance of
workers if the production unit as a whole is to operate efficiently. The structure
of jobs, their content, their relationship to one another and the productive
requirements associated with them are determined by the peculiar advanced
technology used to produce the output of these firms. The distribution ofjobs
across productive requirements is positively skewed. Jobs in the unstructured
sector of the labour market by contrast require much less human capital, and
less specific human capital in particular. They do not have the same rigid
performance requirements of the workforce.
(2)
every grade in those firms with structured internal labour markets has
a specific wage attached to it. The wage that any single worker receives in that
sector therefore depends directly and primarily upon the job that he
or
she
performs and not upon his or her personal productive capabilities. The
internal labour market is managed through collective bargaining,
so
unions
play a key part in allocation decisions within the firm. Wages are set according
to the productive requirements laid down for each job, and negotiated
standard labour input rewards. In practice, therefore, individual wages and the
wage structure are fixed by agreed job description and job evaluation criteria.
There is a formal system of selection and training for higher grades which gives
preference to internal candidates. The structure of wages within each firm will
be different. Each structure will be stable, sensitive only to radical shifts in
technology. Unions accept the influence of technology on job definitions and
performance requirements because of the higher income levels this system
produces; because such methods of wage determination reduce conflicts
within the workforce; and whilst the internal labour market is closed to
outsiders
so
guaranteeing job security and significant income growth pros-
pects. Aquiescence to the influence of technology in job definitions and to rigid
rewards schemes also affords some protection against demand variations. This
supplements the limits to hiring and firing set by the fixed costs of labour
possessing specific human capital. Wages in the unstructured sector on the
other hand are determined predominantly by the unconstrained market forces
of supply and demand. There is much more variability of wages and
employment through time, and greater similarity of wages across the different
firms within the sector.
(3)
we assume that there
is
a discontinuous distribution of labour skills
that parallels the distribution of grades within firms. Each skill level is a
minimum requirement for a certain grade. Workers at each skill level have
varying degrees of individual productive capability above the minimum
required of that skill level, though below that required of the next, higher skill
level. Each worker can be said to have a degree of skill and, within that skill, a
EMPLOYMENT DISCRIMINATION
161
degree of quality. For some workers in the structured sector, there will be an
incentive to offer a higher productive performance than the productive
requirement associated with the grade they
fill
in the firm. To display such
quality is a means of signalling superior talents to the employer and in turn of
enhancing the workers’ chances of receiving training and promotion.
However, for most workers there is no incentive to work beyond the agreed job
requirements. Indeed it
will
be part
of
their union’s task to see that this does
not happen to preserve the existing bargain on job evaluation criteria.
(4)
firms in the structured sector will face a distribution of productive
capabilities offered by their workforce skewed negatively compared to the
distribution ofjobs across productive requirements. Despite this firms will not
attempt to push wages down below collectively agreed levels. Wages higher
than perhaps seem necessary are part
of
the bargain with labour in return for
which firms are assured stability
in
internal labour relations, low turnover, and
continuity of effort (which would be difficult to guarantee by other means,
given the importance of specific human capital).’ Continual revision of
wages, and thus
of
the internal wage structure,
is
to be avoided because of the
conflict it produces. Linking wages to the technology through jointly agreed
job
description and evaluation formulae is in the circumstances an efficient
solution to the firm’s complex labour management problem. The excess supply
of good people provides the firm with an immediately available source of
competent labour to
fill
higher level grades as and when vacancies arise.
A
stable structured internal labour also provides the firm with an efficient means
of
screening promotion candidates.
(5)
workers in the structured sector do better than those in the un-
structured sector. Wages across the board are significantly higher for
comparable general skills, jobs more secure, specific training available and
promotion more accessible. We assume the structured section to be a
significant part of the total labour market, though it must be the case that its
size will be limited by the demand for the products produced by such
technology and the capital intensity of that technology. There is every
incentive therefore, for workers to want to enter and remain within structured
firms. This implies low turnover out of, as well as within the sector. Workers
who hold jobs in its structured sector earn an element of rent arising from the
technological distinctiveness
of
the structured sector.6 It also implies the
existence
of
a queue for jobs,
a
queue over which unions have some measure
of
control through collective bargaining.
Tastes
for
discrimination
An individual’s choice across any set of goods is always influenced to some
6This begs the question
of
course as to whom amongst the workforce at each grade this rent
accrues. Clearly screening by firms will assure that higher quality workers will generally succeed.
(Human capital has been found empirically to be important in determining access to the good jobs
sector, e.g. Rosenberg
(1980).)
There is also an element
of
luck involved, family connections, class,
location and
so
on. The point
of
the present paper
is
of course to explain the poor showing
of
B
people.
For a full analysis of this important argument, see Williamson
et
al.
(1974).
162
J.
SHOREY
degree by social attitudes and pressures. Employers’, or to be more precise
managers’, choices across different categories of labour are fundamentally
influenced by social attitudes and pressures. Though there are competing and
changing ideas about the nature and role of different groups of people in our
society, there are currently certain dominant ideas detrimental to the interests
of specific groups; women, blacks, racial minority groups, etc. These ideas
exist as a set of prejudices and stereotypes as to the contribution such groups
make to the community’s economic, social and political life. They exert
considerable influence upon the workplace.’
There is much still to be learnt about how these ideas arise, persist, change
and how exactly they influence people’s behaviour. The thesis underlying this
paper is that
of
group closure.* Conflict arises in society because of
competition for scarce resources. In the workplace this appears as competition
for good jobs (including those of managers). Groups who have long been
established
in
the labour market, and particularly those in the structured
sector, have an interest in excluding others, in monopolising opportunities, and
in putting others at a disadvantage in competition. Such self interest is reinforced
by fears of unfair competition, unfamiliarity with members of the disadvan-
taged groups in the workplace, and also perhaps culture clashes. Group closure
is
made possible by the visibility of disadvantaged groups. It cuts across class
lines. Widespread, continuous labelling and stereotyping unites and binds the
dominant group, and stabilises social relations to the extent that the dominant
ideology affects the disadvantaged groups themselves. The ideas supporting
group closure are mutually reinforcing, and in time self fulfilling in that disadvan-
taged workers actually begin to acquire undesirable market characteristics.
The disutility that workers and employer experience from the employment
of
B
wor‘kers may appear as a distaste for physical proximity but this is
symptomatic of a much deeper antipathy towards any
B
worker who disturbs
the accepted distribution of jobs and thus the accepted distribution
of
resources, status and power. Employers are not only subject to the dominant
ideology and as individuals keen to protect their positions, they are also as
employers aware that
B
group penetration can cause internal problems.
Employers’ tastes for discrimination are thus founded upon a community of
social values and self interest.
Let us now consider a model
of
employment discrimination that brings
together employer tastes for discrimination, segmentation
of
the labour
market and group closure as the rationale for discrimination. In Section
TIT
we
concentrated upon the structured sector because we now want to consider the
implications of assuming that employers in the structured sector
of
the labour
market have tastes for discrimination. We first discuss the hypothetical short
run
;
a period of time before all market agents can
fully
adjust to the presence
of
discriminatory behavoiur.
Thevaluesystem
upon
which labour market discrimination
rests
originates in the wider social
setting. It has origins therefore in historical events (e.g, colonialism) and the state’s
own
activities
in
the
past
(e.g., male
only
suffrage).
See Parkin
(1979).
EMPLOYMENT DISCRIMINATION
IV
163
EMPLOYMENT
DISCRIMINATION
IN
THE
SHORT
RUN
We continue to assume that each employer's utility function can be written
as equation
1.
However, we assume that the marginal rate
of
substitution
between profits and
B
employment is constant for each firm with respect to the
size
of
its
B
workforceg The rate of substitution, however, differs between firms.
Each firm is thus assumed to receive a given but different amount of disutility
from employing
B
members
of
its
workforce.
Given its tastes for discrimination firm J, as in the wage discrimination
model, employs
B
labour only if appropriate compensation is available. In the
present employment discrimination model such compensation comes in the
form of a marginal product in excess of the wage rather than in the form
of
a
lower wage. Firmj appoints
B
labour at any grade only if their productivity is
greater than the wage set for that grade by an amount no less than the firms
discrimination coefficient dbj.
So
at every grade
:
The wage paid to
B
workers is the same as that paid to
A
workers and equal to
the required productivity for each grade.
So
for grade level
x
we have
:
B
workers at grade and skill level
x
must produce more than the productivity
requirement
q"
so
Since there is a distribution of qualities across
B
workers at each skill level,
some workers are able to remain at grade
x.
However, there are others of lower
qualities who cannot offer sufficient compensation to the firm in the form of
productivity
q"
+
db,. They must, therefore, move down the job hierarchy to
the next lower grade
z.
Here their higher skill endowments perhaps allow them
to offer a productive performance of
q"
+
dbj and thus sufficient compensation
to
the firm. If not, they must move further down the job hierarchy.
B
workers
are thus assigned to grades in the job structure some
way
below what their skill
capacities warrant and below identical
A
workers. The number of steps
is
determined by the strength of firm
j's
tastes for discrimination.
B
labour as a group is displaced down the job hierarchy through the entire
structured sector. Each
B
worker must, of course, maintain his or her higher
productivity at all times otherwise they would be moved down. They must be
"Linearity
of
the indifference curves
does
not have the importance to the employment
discrimination model it
has
for
the wage discrimination model.
It
merely influences the size
of
the
B
workforce in any firm.
164
J.
SHOREY
more productive than the
A
workers operating at the same grade. They must
display a higher quality, work to higher standards and meet higher expec-
tations. This may mean they face higher output targets but it may also mean
that they are required to display higher standards of workmanship, less
absenteeism and industrial activity, better discipline, greater flexibility in task
assignment and
so
on.
B
workers however, have no more reason than
A
workers to work beyond the minimum required of them at grade
x.
So
they perform at
q”
+
db,.
If the distribution of
A
and
B
workers across productive capacities is similar,
the process of displacing the latter downwards in firm
j
shifts the
B
wage
distribution to the left against that of
A
labour. This could mean that certain
grades at the top of the job hierarchy have only
A
workers in them, the number
of such grades again depending upon the size of the firm’s tastes for
discrimination. It could also mean that shortages arise in certain jobs at a
given level
x
if more
B
workers are moved down the job hierarchy from
x
than
come down from a higher level to
x.
However if, as we have assumed, there are
more workers at the top of the distribution of productive capacities than there
are jobs, there is no reason why the firm should face such shortages. Indeed
discrimination acts as a convenient rationing device. There is no reason to
believe therefore, that firm
j’s
unit costs are higher because of its dis-
criminatory behaviour.
At the lowest grades within the structured labour market
of
firmj, some
B
workers are not able to meet the productivity requirement necessary to
compensate their employers. They cannot move down the job hierarchy far
enough to offer adequate compensation because they already
fill
the lowest
grades. They will not, therefore, be employed in firmj, and must move to work
in a less discriminatory structured firm
or
move to the unstructured sector.
The numbers
so
affected again depend upon the size
of
dbi. Overall, structured
firms contain a higher proportion of
A
workers than does the labour market as
a whole, and unstructured firms have a higher proportion of
B
workers than
the labour market as a whole,
ceteris
paribus.
In the model different firms have different tastes for discrimination,
so
each
firm has a different proportion of
B
workers
in
its total workforce and at each
grade. Identically productive
B
workers must display different productive
qualities in different firms, though they may earn the same wage.
B
workers
with the same skills earn different wages as differences in productive
capabilities
or
tastes for discrimination force some workers to move further
down the job hierarchy.
Promotion as well as entry procedures will work rather differently for
B
workers than for
A
workers. Any
B
worker at grade
x
attempting to move up
to grade
y
must prove to the firm that his
or
her productivity is at least
qy
+
db,
whereas
A
workers need only prove a productive potential of
qy.
However,
since
B
workers have an existing performance of
qx
+
db, against
qx
for
A
workers the two groups will face the same selection criteria in incremental
terms. Employment discrimination does not, therefore, reduce the
incentive
to
B
workers to acquire additional skills, either specific
or
general, nor, if
EMPLOYMENT
DISCRIMINATION
165
compensation is available, for the firm to offer them promotion and training.
Incremental rates of return to
A
and
B
workers are unchanged and thus still
the same. On average
B
workers will however, be less successful in meeting
promotion criteria to the extent that individual ability constraints apply
sooner.
Having
A
and
B
workers operating alongside one another at the same grade
doing the same job, paid the same wage, but subject to different working
standards would pose logistic problems for the
firm
in its dealing with
B
labour
and outsiders. In practice, however, there are always a number of job slots at
each grade in the firm, each with its own peculiar job definition.
Discrimination, therefore, usually means that
B
workers are assigned to
certain jobs at each grade and
A
workers to others. There are different
productivity requirements attached to the two kinds of jobs, but the same
wage. In this form, employment discrimination is far less visible than wage
discrimination would be and indeed is very difficult even for
B
workers to
identify effectively, supported as it always is by a negotiated task definition and
paid according to standard job evaluation procedures. It is difficult to prove
that jobs with very different, complex task definitions are in fact technically of
the same grade, and that the
B
workers' jobs involve higher requirements
without earning a correspondingly higher wage. There thus emerges a mixed
labour force in the firm but segregation by jobs. In the extreme case this
produces two distinct hierarchies ofjobs, one for
A
workers and another for
B
workers, the latter being shorter than the former according to the value of db,.
Our analysis implies that:
(a)
A
and
B
workers at the same grade in any structured firm earn identical
wages, but within any firm,
B
workers often earn less than
A
workers of
identical productive capabilities. The average wage for
B
workers
in
the
structured sector is below that
of
A
workers." The relative wage of
B
workers
in the labour market as a whole is even lower because many
B
workers are
relegated to the lower paying unstructured sector." For the same reason
B
workers will be more prone to unemployment and job instability.
It is clear that employment discrimination in this model is conceptually
different from wage discrimination. Even where they produce the same result,
wage differentials, they do
so
by very different mechanisms. Under employ-
ment discrimination equally productive workers are not employed at the same
grades. Furthermore, employment discrimination for many
B
people does not
produce wage differentials. They work at the same grade as equally productive
A
people but under more demanding working conditions.
(b) employment discrimination results in many
B
workers operating at
lower grades than are normally associated with the skills they possess. The
structured sector is thus characterised by significant overqualification of
B
lo
Effectively then, as Rumberger and Carnoy
(1980)
confirm
empirically, the structured sector
The inferior position
of
certain
B
groups
of
workers in terms
of
both wages and employment
of the dual labour market itself becomes segmented.
has been well documented, e.g. Department of Employment
(1974).
166
J.
SHOREY
workers. Outside the structured sector there are many
B
workers with equal,
if
not superior qualifications to
A
workers in the structured sector but excluded
from that sect or.
(c) the model predicts that each firm operates with a mixed labour force.
Only certain grades at the top of the job hierarchy are filled by
A
workers
alone. As long as
B
workers accept their lower assignment and can establish
and maintain their differential productivity, firms are compensated ap-
propriately and are indifferent when hiring or promoting into any grade
between identical
A
or
B
labour. This implies that it is harder for
B
workers on
average to enter the structured sector but that turnover of
B
people inside the
sector will be as low as for
A
people. Internal mobility measured in terms of
frequency or distance per move to very similar for
A
and
B
workers, though
B
people
of
course on average start lower down.’ Employment discrimination
shortens the occupational ladder for
B
workers, puts them continually behind
equivalent
A
workers and
so
generates differential lifetime
experience^.'^
discrimination takes the form of
B
workers operating at higher
productive requirements at each grade in the job hierarchy than are negotiated
for jobs performed by
A
workers at the same level and paid the same wage.
Discrimination is possible because of the non competitive manner in which
wages are determined and jobs defined within firms with structured internal
labour markets, along with the fact that exceptional treatment of
B
workers is
possible.
(e) the model predicts that discrimination in the structured sector in the
short run increases the firms profits. Firmj pays each
B
worker less than his
or
her value to the firm; the firm gets more output than it pays for. This
conclusion
is
in marked contrast to the prediction from the wage discrimi-
nation model that discrimination raises the firm costs and, therefore, reduces
profits. In fact in the model of employment discrimination the more
discriminatory the firm is the higher are its profits. Employers are thus able to
satisfy their tastes for discrimination unrestricted by the minimum profit
constraint usually placed upon them by owners.
(f)
and crucially, the model implies that discrimination can arise only
because the local trade union is party to the firm’s discriminatory treatment of
B
people. Unions are powerful agencies in the structured sector. Employers
have to concede genuine gains to the union in the form of wages, security, job
definition, promotion, and firing and hiring procedures to secure labour’s
continuing co-operation. Discrimination exists because unions do not force
similar conditions for
B
people.
A
local union’s policy towards disadvantaged
workers, following the arguments of Ashenfelter
(1972),
depends
inter
alia
(d)
Evidence that overqualification exists and that
B
workers are concentrated in the
unstructured sector
of
the labour market is readily available, e.g. Hakim
(1978).
Econometric
analysis to isolate the existence
or
otherwise
of
overqualification
of
B
workers is only in its first
stages. Results
so
far are not clear cut. See Kamilich and Polachek
(1982).
l3
For
empirical evidence consistent with these predictions see Leigh
(1976).
14Stewart
(1983)
finds that between
75
per cent and
100
per cent
of
the average unexplained
differential in earnings between blacks and whites in his study is due to differences in occupational
attainment.
EMPLOYMENT DISCRIMINATION
167
upon the direction given by local leaders, the numerical importance and
leverage
of
B
workers in the union, and the influence
of
national union policy
and officials. It also depends upon the opportunity
for
securing gains by
discriminatory practices, which in turn depends upon the degree to which the
local union controls the labour supply. When we consider the position of the
local union in
our
model there is every reason to anticipate that union policy
will disadvantage
B
workers
:
the local leadership works on the shop floor and
is thus subject to the ideology of the workplace and close control by the
membership; there will be strong pressures for exclusion policies from the
A
membership because the technology of the structured sector, as already noted,
offers considerable rewards to workers employed there.
B
workers have not
historically been a significant element in the labour force of the structured
sector nor in their unions; local unions in the structured sector do not rely on
national unions to sustain their bargaining power and
so
are not constrained
by national policies, even if these were openly hostile to discrimination
;
local
unions have considerable influence
upon
the
firm’s
labour decisions due to the
strength
of
collective bargaining and the superior skill level. It seems clear then
that the local union will pursue the interests
of
the majority of its members by
conspiring with the employer in the firm’s discriminatory
practice^.'^
EMPLOYMENT
DISCRIMINATION
IN
THE
LONG
RUN
In the conventional employer tastes model of wage discrimination,
discrimination cannot continue in the long run (a period of time in theoretical
analysis when economic agents are able
to
adjust fully to market signals and
incentives). Competition in product markets eliminates discriminatory be-
haviour.16 Any firm in market
i
which discriminates fails to minimise its costs,
leaving non discriminatory existing firms and potential entrants with a
significant cost advantage. Long run adjustment takes two forms. First, there
is entry/expansion of firms with low tastes for discrimination who plan to
employ a high proportion
of
B
workers, to cut costs and to underprice more
discriminatory firms in product markets. Second, and simultaneously,
B
There is a
lot
of casual empiricism in the literature supporting the proposition that unions
have not defended the interests
of
B
workers vigorously: see hell
et
al.
(1981). On the other hand
econometric analysis (Ashenfelter (1972), Nickel (1977)) suggests that on average the B/A wage
ratio is
higher
where unions are stronger, i.e. if there is discrimination, there is less where
unionisation is important. However, econometric analysis also suggests that the
B/A
differential
is
lower
in the case of craft unions and the union leverage is stronger in local as opposed to national
negotiations. (Ashenfelter (1972), Mulvey (1976)).
I6
As pointed
out
by Goldberg(1982) and others, this result depends formally upon expressing
discrimination as
a
preference
against
the B group rather than
as
a
preferencefor the
A
group (i.e.
rather than as favouritism or nepotism).
168
J.
SHOREY
workers find their search for less discriminatory employment more produc-
tive. They are able to improve their position by quitting their current
employment for a job in one of the expanding firms. Both the emergence of
more jobs in less discriminatory firms and mobility by B workers are essential
to the demise
of
discrimination.
As
long as any discrimination exists, there is
the incentive for less discriminatory firms to enter and/or grow, and thus scope
for profitable labour mobility. To survive, firms with a taste for discrimination
must reduce their demand for
A
labour to reduce their costs, which means the
wage of B labour rises. The process only ceases when the wage differential
disappears. Firms who persist in practising wage discrimination are driven out
of business by competitive reductions in product prices. Wages for
A
and
B
workers are eventually equalised.
A
small number of employers with strong
tastes for discrimination might survive this process. Assuming there are
sufficient
A
workers around, such employers employ only
A
workers. The
labour market might thus become partially segregated across firms, but labour
is everywhere paid the same wage for identical abilities.
Of
course, working labour markets contain many features which conspire
seriously to dampen (if they were ever to exist) the vigorous competitive
pressures always assumed in simple economic models. Non-competitive
behaviour is in reality no aberration once one recognises the strictures of the
competitive ideal. Including non-competitive product markets, rigidities and
frictions or non-maximising behaviour in the model means that high cost
discriminating firms are not driven out. The long run in the model looks much
like the short run and discrimination would be one mechanism by which firms
expend their monopoly profits. Capital market imperfections, and inequalities
in the distribution of wealth in particular, work to prevent the entry of less
discriminatory employers, most notably of course of B people themselves.
Nevertheless, the incentive and thus the basis for competition in product
markets exists.
Under the system of employment discrimination the less discriminatory
firms in any market
i
and iess discriminatory potential entrants are at a cost
disadvantage to existing discriminatory firms, because the latter secure greater
output from B workers than they pay for. This discourages the creation ofjobs
in less discriminatory firms. If discrimination is
so
widespread that there are
discriminatory firms at the margin earning low profits, the only way that a
non-discriminatory firm might be able to survive entry would be to adopt a
discriminatory policy, something presumably that it would find disagreeable.
Much will depend, therefore, on how much discrimination is necessary for a
new entrant to achieve a profit rate that will satisfy its
investor^.'^
Ceteris paribus,
the competitive process would in time encourage firms at
the margin earning low profits, presumable less disciminatory firms,
to
become more discriminatory. They would copy the job assignment, job
The entry
of
non-discriminatory firms
is
particularly difficult because the capital intensity
of
production methods in the structured sector will work to limit entry by any new operator.
In
addition,
no
doubt, established firms in this sector have considerable advantages arising out
of
their experience
of
specialising in the peculiar high technology products this sector produces.
EMPLOYMENT DISCRIMINATION
169
definition, hiring and promotion practices
of
the more discriminatory firms.
There is also a financial incentive for the more discriminatory firms to become
even more discriminatory. In either case it would not be possible to increase
profits by increasing the number
of
B
workers employed at the expense of
A
workers. Unions in the structured sector will close down this option. Through
collective bargaining they will see to it that a customary low maximum ratio
of
B
to
A
workers is established and maintained in each grade. Firms' search
for profits would, therefore, have to be directed towards increasing the
surplus on individual
B
workers, with the inevitable consequence that
B
employment would fall as
B
people are displaced to the unstructured sector.
Thus in the long run the number and quality
of
opportunities for
B
workers in the structured sector would diminish. Employment discrimi-
nation, unlike wage discrimination, does not actively foster the seeds of its
own destruction, quite the reverse in fact. These are pressures for discrimi-
nation to escalate.
What is there to prevent this escalation?
As
each firm increases the size
of
the surplus from each
B
worker, the size of the surplus generating
B
workforce
falls. In principle, therefore, there must be some limit to the extent of
discrimination
;
some optimal combination of surplus and size unique to each
firm. However, there are economic and social factors that inhibit escalation of
discrimination long before this point. In the first place, firms are restricted in
their ability to exploit
B
labour by the quit threat of
B
workers. Extending the
surplus on
B
employment too far could
so
precipitate quits that the firm
reduces its total profit. In the circumstances, however, the threat of mobility is
unlikely to be a very powerful one. The mobility of labour is always impaired
by search costs, information gaps, transactions costs, psychic costs and some
degree of risk aversion. Mobility for many workers discriminated against in
the structured sector would mean a move to the less desirable unstructured
sector." Any worker employed in the structured sector has less scope for
mobility than other workers. The acquisition of specific skills, which are
rewarded only in the worker's current firm, means that a move
to
a less
discriminatory
job
could significantly harm the worker's financial position.
Furthermore, the impediments to the mobility of
B
workers could be even
stronger than for other workers simply because of the economic and social
position of the people who make up that group. For example, married women
are less mobile geographically because
of
their husbands' jobs, because of the
secondary importance attached by the household to their market wage, and
because of their inferior access to household transport. Minority racial groups
have fewer reserves of wealth, perhaps lack extensive knowledge of markets
and tend to be geographically concentrated and thus restricted in their access
to jobs. Mobility is, therefore, a limited restraint upon discrimination.
Nevertheless,
if
employment discrimination were to lead the firm to ex-
propriate virtually the entire rent
B
workers derive from access to the
'*
Opportunities for mobility to the unstructured sector fall significantly when demand as a
whole contracts because this sector
is
so
sensitive to demand conditions. This implies that
structured
firms
have greater scope far discriminatory practices in recessions.
170
J.
SHOREY
structured sector, quits into the unstructured sector and out of the labour
market would become important.”
Second, the amount of individual surplus that firms can extract from
B
workers by employment discrimination depends upon the nature of the
available technology. There are limits
to
the amount that jobs at the same
grade can be differentiated and higher productive requirements attached to
some and not to others.
Third, and most important perhaps, there are limits to the acquiescence of
trade unions to the exploitation of
B
workers. Total acquiescence would
undermine group unity, question union leaders credibility as worker rep-
resentatives and of course, pose the question to all union members about their
own security.
B
workers will be antagonised into overt opposition and the
A
dominated union will, like the employer, not want to see the emergence of B
unions. Conceding to the firm the ability to exploit
B
labour is part of a
bargain in return for which
A
workers receive certain benefits. There will be
limits to the scope for such bargaining.
Finally, we would argue that whilst workers and employers derive some
disutility from the employment of
B
workers they will also derive disutility
from excessive and blatant abuse of the
B
groups. The value system that
sustains the tastes for discrimination will also set limits to such tastes.
No
clear
cut boundary exists. Different firms reach different outcomes according to
local conditions.
The extent
of
discrimination within any firm in the short run depends upon
the strength of employers’ tastes for discrimination. In the long run
it
depends
much more on the tastes for discrimination of workers as management and
union decide through collective bargaining, the firm’s labour policy. Given the
dominance of common social values and the fact that for once this is not a zero
sum game for the two parties, they are unlikely ever to come to open conflict
on this issue. Once instituted, discriminatory practices will be reinforced and
become less obtrusive as
A
and
B
groups become increasingly dissimilar by
virtue of the powerful cumulative forces highlighted in the segmentation
literature.
Several points emerge from our long run analysis. First, formally we have
constructed our model around the employer tastes approach because it is
employers who enact discriminatory policies and employer participation is
essential. But it is apparent that worker tastes for discrimination are equally
important to the persistence of discrimination.20 In practice either the union
or the firm could be the instigator of a move towards more
or
less
discrimination through the collective bargain. There
is
thus scope here for
19There are several models in the literature that focus upon the monopsony position
of
firms
facing upward sloping labour supply curves due to information gaps. Because of differences
in
information, search costs, and perspectives on job rewards, the supply curves
of
A
and
B
workers
have different elasticities. See Madden
(1974),
Gordon and Morton
(1974),
Frank
(1978).
As
is
apparent above we recognise the pertinence
of
such arguments, but we doubt that they alone
explain very much of the discrimination we observe.
2o
Models
of
employee discrimination exist in the literature. They tend
to
generate segregation
rather than discrimination.
EMPLOYMENT DISCRIMINATION
171
distinguishing different firms and periods of time. But it is the consistency
between employer and employee tastes that is central to our approach. It
follows that the distinction between demand and supply side factors has little
of its normal significance here. It is not possible to categorise discrimination as
a demand side phenomenon due to employer tastes or as a supply side
phenomenon due to market imperfections that produce monopsony. It is the
union
of
demand and supply sides through the common value system of
employers and employees that underpins discrimination.
Second, in that our model embraces monopsonistic firms it shares with
conventional monopsony models of discrimination the idea that firms
maximise subject to non-competitive labour supply constraints. But the
present model is based upon utility and not profit maximisation. Firms have
no effective monopsony power over
A
workers. What potential they have in
this respect is counterbalanced by collective action. Firms only have
monopsony powers over B workers. This arises out of the technological
characteristics
of
the structured sector and persists because
of
the dis-
criminatory stance of unions.
Third, discrimination persists because it benefits both employers and the
dominant group amongst the workforce of the structured sector.” Limiting
the employment opportunities of B workers, imposing higher job require-
ments and creating separate inferior job structures satisfies the common desire
to protect established positions and accords with the perceived inferiority
of
B
labour. Total segregation across firms or exclusion from the structured sector
is unnecessary. Furthermore simultaneously, the surplus from
B
labour
satisfies the requirements of owners and provides material gains for managers
and workers.
Fifth,
B
workers in the structured sector are exploited in the sense of
receiving rewards less than their productive contribution. They are neverthe-
less better off than equivalent
B
workers in the unstructured sector. Unions
play an active or passive part in this exploitation but
it
is a crucial part given
the nature of the employment relationship. It amounts to exploitation by
proxy. The economic literature sometimes distinguishes between discrimin-
ation as behaviour based upon tastes that involves economic agents in
financial disadvantages and exploitation as behaviour based upon self interest
that involves pecuniary advantages.
No
such distinction is warranted here,
discrimination and exploitation are inseparable.
Sixth, our analysis concentrates entirely upon the structured sector. This
does not mean that discrimination is absent from the unstructured sector of
the labour market, merely that if it does exist it requires a different form
of
analysis. The value system prejudicial to
B
workers prevails in un-
structured firms but the factors supporting discrimination in the long run must
be different, because firms have no monopsony power in the unstructured
This distinguishes
our
approach from that of the radical literature where the gains accrue
to
employers
only
and discrimination
is
part
of
a much wider strategy to divide the workforce and
thus control (exploit) workers as a whole: See Franklin and Resnik
(1974).
172
J.
SHOREY
sector, the technology is simple, unions are not assertive and
B
workers appear
in significant numbers.’
’
Our analysis has the following policy implications.
(a) equal pay legislation would be ineffectual since workers at the same
grade are paid the same.
(b) equal opportunities legislation is likely to involve slow tortuous
policing of firms’ job assignment practices which in the end might have little
overall effect. Job evaluation and the comparison of productive requirements
across jobs is extremely difficult for outsiders to accomplish sati~factorily.~~
(c) subsidising additional jobs for
B
workers in discriminating firms, even
if
it got past
A
workers, would be a mixed blessing.
B
employment in structured
firms might rise but the additional workers would be subject to discrimination.
Most of the new jobs would arise at the lower end of the job hierarchy. The
same arguments would apply to policies
to
increase the training opportunities
for
B
workers.
(d) expanding the structured sector as a whole would increase the number
of
B
people with good jobs. But this is hardly something that governments can
legislate. It would depend upon introducing new employers with low tastes for
discrimination. One avenue open to the government of course would be to
expand its own activities. There are difficulties here too unfortunately. Much
of
the State’s activities are not in the structured sector, the State is itself
constrained by union activities, and the positive discrimination required to
make a real impression would not impress
A
workers or their
union^.'^
(e) the government could vigorously pursue policies to shift the value
system that underpins discriminatory behaviour. The State’s influence on
many
of
the agencies
of
socialisation, particularly the mass media, make this
feasible. The past suggests that reforms primarily come about as the result of
vocal campaigns to gain the support of the leaders of the interested groups
through appeals to liberal sentiments rather than as a result of the
mobilisation of mass support from the disadvantaged groups.
So
inevitably
change brought about by this means will be slow; many workers lose by
reforms and liberal sentiments are often far removed from the workplace.
(f)
mobilisation of
B
workers would be difficult and of limited value in
countering employment discrimination. B workers are concentrated in the
unstructured sector where union organisation is severely hampered. Such
workers can have little impact upon the structured sector from which they are
22
This would point towards segregation in the unstructured sector. There is empirical evidence
however that discrimination as well as segregation exists, e.g. Rumberger and Carnoy (1980).
23
There is a good deal of evidence relating to existing legislation indicating just how difficult
such policing would be, e.g. Snell
et
al.
(1981). This is not
to
say that such legislation
is
unimportant. It is a vital element in undermining the value system that supports tastes for
discrimination at
work.
24
Following Anderson and Wallace (1975) the most attractive policy the government couid
follow has nothing directly to do with discrimination.
It
is the case that the desire to exclude
B
workers from good jobs,
B
workers’ inability to resist and thus the level of employment
discrimination increase with falling economic activity. Maintaining growth in the economy is
therefore essential in aiding disadvantaged
groups.
EMPLOYMENT DISCRIMINATION
173
excluded.
B
workers are in fact segmented by the market. On the other hand
mobilisation
of
B
workers in the structured sector is stultified by the numerical
weakness of the
B
group there.
(g) a commitment to all of the above policies will be necessary to reduce
employment discrimination. However, it seems from
our
model that the most
effective action requires a change in union behaviour. Local unions play a
critical role in the continued exercise of monopsony power by employers.
Unions operate permanently inside the firm and alone could effectively police
anti-discrimination rules. Unions have bargaining power in structured firms.
It follows from
our
earlier discussion of the origins
of
union policy that moves
against discrimination, either self determined
or
inspired by external in-
fluences are unlikely. It is necessary for the government, political parties and
the trade union movement, despite the inherent weakness of their position,
nevertheless to encourage
B
workers to join local union activities and to take
part in their decision-making hierarchy. Such encouragement will require
financial support for training, child care etc. There may well be a case for going
further to actively foster the growth of separate
B
group representation, inside
or outside existing unions. There would no doubt be problems here.
Separation could actually result in open conflict which, given the long
standing weakness of
B
workers in the structured sector, would not be in their
best interests. However, it is unlikely that until such moves begin local
A
dominated unions will alter their stance.25
VI
CONCLUSIONS
The employer tastes approach has been used here as
a
basis for a model
of
employment discrimination. The model incorporates a labour market that has
important features distinguishing it from the competitive paradigm.
For
many
jobs, wages are determined first and foremost by the technological charac-
teristics of the job, not by the characteristics
of
the workers that fill them.
Wages and performance requirements are laid down for every such job.
Worker co-operation is essential if the technology is to be operated efficiently.
Unions provide this co-operation in return for good working conditions for
A
workers. However, they do not negotiate for
B
workers with the same
conviction given a deep rooted desire for closure amongst the majority
of
their
membership.
B
workers are, therefore, systematically exploited by the firm to
the benefit of both employers and
A
workers.
Several insights into the nature of labour market discrimination emerge
from this analysis. First, discrimination is much more likely to emerge as
employment than wage discrimination simply because the former does not
foster harmful competitive responses from non-discriminatory firms, it
benefits employers and the majority of workers, and it is far less visible.
"There
is
a growing literature
on
the problems and possibilities
of
greater involvement
by
B
workers
in
union activities,
e.g.,
Coote and Kellner (1980).
I74
J.
SHOREY
Second, discrimination is likely to be a marked feature
of
particular firms-
those with structured labour markets. The technology used by such firms plays
a vital part in producing discrimination. Third, employment discrimination
has important implications for disadvantaged workers arising from the fact
that they are concentrated in the unstructured sector.26
As
well as receiving
lower pay than
A
workers in the structured sector they will accumulate less
capital, experience less job satisfaction and exercise less control over their
working lives. Fourth, employment discrimination provides a convenient
rationing device for firms facing a job distribution more positively skewed
across productive requirements than the labour distribution. Removing
discrimination will necessitate the adoption of an alternative rationing device.
Many
B
workers will remain in the unstructured sector. Fifth, discrimination
in the workplace is one
of
the products
of
a socially wide set of prejudices. The
basis and nature of discrimination cannot therefore, be fully appreciated
except within a social context. However, our analysis indicates the crucial role
played
by
local unions in the persistence of discrimination and therefore, of the
need for greater unionisation in its widest sense, amongst
B
workers.
26Concentration of B workers in the unstructured sector has the crowding effect
of
the
literature (pushing down wages in that sector, for
A
and B people, below their level in the absence
of any discrimination) only if closure produces a reduction in the employment
of
B
workers in the
structured sector that is not accompanied by an equivalent increase in A employment.
REFERENCES
AIGNER, D.
J.
and
CAIN,
G.
G.
(1978). Statistical Theories of Discrimination in Labor Markets,
Industrial and Labour Relations Review,
vol. 30, pp. 175-187.
ANDERSON
and WALLACE (1975). Public Policy and Black Economic Progress:
A
Review
of
the
Evidence,
American Economic Review,
vol. 65, pp. 47-52.
ARROW,
K.
(1972). Models
of
Job Discrimination, and Some Models
of
Race in the Labour
Market, Chapters 2 and 6 in
A.
H. Pascal (ed.),
Racial Discrimination in Economic Life,
pp.
83-102,187-203. Lexington, Mass, D.C.
:
Heath.
ASHENFELTER,
0.
(1972). Racial Discrimination and Trade Unionism,
Journal
of
Political
Economy,
vol. 80, pp. 43-63,
BECKER,
G.
(1971).
The Economics
of
Discrimination,
2nd edn. Chicago: University
of
Chicago
Press.
CAIN,
G. G.
(1976). The Challenge of Segmented Labor Market Theories to Orthodox Theory:
A
Survey,
Journal
of
Economic Literature,
vol.
14, pp. 1215-1257.
Coom,
A.
and KELLNER,
P.
(1980).
Hear This Brother.
New Statesman Report
No.
1.
London.
DEPARTMENT
OF
EMPLOYMENT (1974).
Women and
Work;
A
Statistical Suroey.
Manpower Paper
DOERINGER,
P.
B.
and
PIORE,
M.
J.
(1971).
Internal Labour Markets and Manpower Analysis.
EDWARDS, R. (1979).
Contested Terrain.
London
:
Heinemann.
FRANK, R. H. (1978). Why Women Earn Less: The Theory and Estimation
of
Differential
FRANKLIN,
R.
J
and RESNICK,
S.
(1974).
The Political Economy
of
Racism.
New
York:
Holt,
GOLDBERG, M.
S.
(1982). Discrimination, Nepotism and Long Run Wage Differentials,
Quarterly
GORDON, M. M. and MORTON, T.
E.
(1974). A Low Mobility Model
of
Wage Discrimination with
HAESSAL, W. and PALMER, J. (1978). Market Power and Employment Discrimination,
Journal
of
No.
9. London
:
HMSO.
Lexington, D.C.: Heath.
Overqualification,
American Economic Review,
vol. 68, pp. 360-373.
Rinehart, Winston.
Journal
of
Economics,
vol. 97, pp. 307-3 19.
Special Reference to Sex Differentials,
Journal
of
Economic Theory,
vol. 7, pp. 241-253.
lfuman Resources,
vol. 13, pp. 545-560.
EMPLOYMENT DISCRIMINATION
175
HAKIN,
C.
(1979).
Occupational Segregation.
Research Paper No.
9.
London: Department
of
Employment.
KAMLICH, R.
F.
and POLACHEK,
S.
W.
(1982).
Discrimination
:
Fact or Fiction? An Examination
Using an Alternative Approach,
Southern Economic Journal,
vol.
49,
pp.
45U61.
MCNABB,
R.
and
PSCHACHOROPOULOS,
G.
(1981).
Further Evidence
on
the Relevance
of
the Dual
Labor Market Hypothesis
for
the U.K.,
Journal ofHuman Resources,
vol.
16,
pp.
442-448.
MADDEN,
J.
F.
(1975).
Discrimination-A Manifestation
of
Male Market Power, in C. B. Lloyd
(ed.),
Sex Discrimination and the Division
of
Labor,
pp.
146174.
New York: Columbia
University Press.
MARSHALL, R.
(1974).
The Economics
of
Racial Discrimination
:
A Survey,
Journal
of
Economic
Literature,
vol.
12,
pp.
849-871.
MULVEY,
C.
(1976).
Collective Agreements and Relative Earnings in U.K. Manufacturing in
1973,
Economica,
vol.
43,
pp.
419-428.
NICKELL,
S.
(1977).
Trade Unions and the Position
of
Women in the Industrial Wage Structure,
British Journal oflndustrial Relations,
vol.
15,
pp.
192-210.
OSTERMAN, P.
(1975).
An Empirical Study of Labour Market Segmentation,
Industrial and Labour
Relations Reoiew,
vol.
28,
pp.
508-523.
PARKIN,
F.
(1979).
Marxism and Class Theory:
A
Bourgeois Critique.
London
:
Tavistock.
PIORE, M.
J.
(1975).
Notes for a Theory of Labor Market Stratification, in Edwards,
R.
C., Reich,
M. and Gordon, D. M. (eds),
Labor Market Segmentation.
Lexington, D.C.
:
Heath.
REICH, M.
(1981).
Racial Inequality.
New Jersey
:
Princeton University Press.
ROSENBERG,
S.
(1980).
Occupational Standing and the Dual Labor Market,
Industrial Relations,
vol.
19,
pp.
34-49.
RUMBERGER,
R. W. and CARNOY, M.
(1980).
Segmentation in the
U.S.
Labor Market:
Its
affects on
Mobility and Earnings
of
Whites and Blacks,
Cambridge Journal ofEconomics,
vol.
4,
pp.
SNELL,
M.,
BLUCKLICH,
P.
and POVALL,
M.
(1981).
Equal Pay and Opportunities.
Department
of
Employment Research Paper
No.
30.
London:
HMSO.
STEWART,
M.
B.
(1983).
Racial Discrimination and Occupational Attainment in Britain,
Economic
Journal,
vol.
93,
pp.
521-541.
WACHTER, M.
(1974).
Primary and Secondary Labor Markets: A Critique of the Dual Approach,
Brookings Papers
on
Economic Activity,
vol.
3,
pp.
652-680.
WILLIAMSON,
Q.,
WACHTER, M. and HARRIS,
J.
(1974).
Understanding the Employment Relation
:
The Analysis
of
Idiosynchratic Exchange,
Bell Journal
of
Economics and Management
Science,
vol.
6,
pp.
250-278.
117-132.
To continue reading
Request your trial