Editors' Preface to Macmillan Studies in Economics

The rapid growth of academic literature in the field of economics has posed serious problems for both students and teachers
of the subject. The latter find it difficult to keep pace with
more than a few areas of their subject, so that an inevitable
trend towards specialism emerges. The student quickly loses
perspective as the maze of theories and models grows and the
discipline accommodates an increasing amount of quantitative
techniques.
'Macmillan Studies in Economics' is a new series which
sets out to provide the student with short, reasonably critical
surveys of the developments within the various specialist
areas of theoretical and applied economics. At the same time,
the studies aim to form an integrated series so that, seen as a
whole, they supply a balanced overview of the subject of
economics. The emphasis in each study is upon recent work,
but each topic will generally be placed in a historical context
so that the reader may see the logical development of thought
through time. Selected bibliographies are provided to guide
readers to more extensive works. Each study aims at a brief
treatment of the salient problems in order to avoid clouding
the issues in detailed argument. Nonetheless, the texts are
largely self-contained, and presume only that the student has
some knowledge of elementary micro-economics and macroeconomICS.
Mathematical exposition has been adopted only where
necessary. Some recent developments in economics are not
readily comprehensible without some mathematics and statistics, and quantitative approaches also serve to shorten what
would otherwise be lengthy and involved arguments. Where
authors have found it necessary to introduce mathematical
techniques, these techniques have been kept to a minimum.
The emphasis is upon the economics, and not upon the quantitative methods. Later studies in the series will provide analyses
of the links between quantitative methods, in particular
econometrics, and economic analysis.

MACMILLAN STUDIES IN ECONOMICS

General Editors: D. C. RowAN-and
Executive Editor: D. W. PEARCE

G.

R. FISHER

Labour Economics

J. E. KING

 J. E. King 1972
All rights reserved. No part of this publication may be reproduced or
transmitted, in any form or by any means, without permission.

First edition 1972

Reprinted 1983
Published by
THE MACMILLAN PRESS LTD
London and Basingstoke
Companies and representatives
throughout the world
ISBN 978-1-349-15469-2

ISBN 978-1-349-15467-8 (eBook)

DOI 10.1007/978-1-349-15467-8

The paperback edition of this book is sold subject to the condition that it shall
not, by way of trade or otherwise, be lent, re-sold, hired out, or otherwise
circulated without the publisher's prior consent in any form of binding or
cover other than that in which it is published and without a similar condition
including this condition being imposed on the subsequent purchaser.

Contents
Preface

Acknowledgements

11

1 Labour Markets

13
13
14
15
19

Labour market models

Institutional forces in the labour market
The intemallabour market of the firm

2 The Firm's Demand for Labour

3 Labour Supply

Trade union wage policy

The economic effects of unions

24
24
29
31
33
35
35
39
40
43
43
45

6 The Functional Distribution of Income

50

7 The Size-Group Distribution of Income

55

8 Unemployment

62
70

The individual's supply decision

Labour mobility
The local labour market revisited
The supply of labour to the firm

4 Wage Differentials

Occupational differentials
Inter-industry differentials
Geographical differentials

5 The Economics of Trade Unions

Bibliography

Preface
Labour economics, like most other branches of the discipline,
involves the study of market behaviour. Its subject-matter
includes choices made by firms about the amounts and types
of labour that they buy, and the wages and salaries that they
pay; choices made by individuals and families about their jobs;
the economic behaviour of trade unions and its consequences;
and the wider issues of wage structures of all kinds, the distribution of incomes from employment and their relation to incomes,
from other sources, and macro-economic features of wage and
employment behaviour. The present study, however, is
necessarily selective. Wage inflation is the subject of another
study in this series, and is not dealt with directly here; but some
of the themes developed with respect to the theory and operation of labour markets are, indirectly, at the root of recent
controversies about inflation. Institutional aspects of the labour
movement (issues of trade union structure and collective
bargaining, etc.) are, for reasons of space, barely mentioned.
Likewise, the technical approach to bargaining theory has been
omitted; an adequate coverage of such analysis would require
a separate study of considerable length.
Thus the present study cannot claim to be comprehensive.
Rather its scope depends on a personal selection of those areas
oftheory and empirical research which are thought to be central
to the understanding of labour markets in advanced capitalist
economies. The emphasis throughout is on the criticism of
traditional or 'neo-classical' theory in the light of modem
research, rather than on its exposition. Hence the very scanty
treatment of the marginal productivity principle itself, con9

doned if not excused by the wealth of existing analyses, some of

which are mentioned in the text.
Labour economics is increasingly seen as relevant in the
wider context of economic theory as a whole (e.g. the revival
of interest in distribution theory as a factor in economic growth
models), and some significant social and economic problems
(among others inflation, poverty and unemployment). At the
same time the defects of traditional labour market theory are
increasingly recognised. The aim of this study is to examine,
briefly and selectively, some of the theoretical issues which have
emerged.

10

Acknowledgements
My colleagues John Hillard, Paul Langley and Jim Taylor
made many helpful and discerning comments on earlier drafts
of this study. Dr John Corina ofSt Peter's College, Oxford, did
the same and, as my original teacher, stimulated and encouraged my interest in labour economics.
All remaining defects of style and content are entirely my
own.
J. E. K.

11

Labour Markets

LABOUR MARKET MODELS

Neo-classical economists (e.g. Hicks [33]) treated labour markets as close approximations to the perfect markets which are
described in most elementary textbooks, and which are exemplified by stock exchanges and international financial and
commodity markets. Labour is assumed homogeneous, information and mobility are costiess, and atomistic competition
prevails between large numbers of buyers and sellers.
Kerr suggests that the 'natural' market for labour is quite
different: 'The average worker has a narrowly confined view of
the market, and, in addition, is not an alert participant in it'
([42] p. 281), a conclusion which is supported by exhaustive
empirical research (e.g. Reynolds [74]).1 Traditional models
allowed for 'frictions' in the operation of labour markets, but
substantially underestimated their significance. Moreover, employers typically exercise a certain degree of control over the
market, through the tacit collusion which Adam Smith was the
first to notice ([88] p. 75). Nor is labour homogeneous, even
within broad occupational groups.
These considerations cast doubts on attempts to analyse
labour markets as a matrix of perfect sub-markets, classified by
occupation and geographical location, in each of which a single
wage rate prevails and 'within which workers are willing to
move and do move comparatively freely from one job to
another' (Kerr [42] p. 278). These two aspects of the labour
market - concerned respectively with wage determination and
with the distribution of jobs - need not, therefore, coincide.
Consider an example: the local labour market (LLM) for a
1 And distinguishes the average worker from the stockbroker,
banker or international commodity speculator!

13

particular type of labour was traditionally seen as one cell of

the labour market matrix, acting as both a wage market and a
job market, as defined above, in which normal arbitrage
operations guaranteed the observation of the law of equal
price. l Assuming transport costs within an LLM to be negligible,
labour of a particular type would fetch the same price throughout the market. Differences in wages between firms for the same
type of labour would be eliminated as workers move from lowwage to high-wage firms.
In practice it is impossible to define the LLM as the complex
of jobs for which the same wage is paid (empirical evidence is
discussed below, pp. 31-3). Even as ajob market the LLM is
extremely difficult to define. From the point of view of employers, it may be taken as the catchment area from which he
recruits his labour. For the worker, it is the area within which
he is prepared to work without moving house. There is no
reason why these two definitions need give the same geographical area, except perhaps in the case of a small city
surrounded by hundreds of miles of desert! Goodman [29]
provides an operational definition of the LLM as that area in
which firms look for workers and in which most residents work.
Here again, nothing is implied about wage uniformity.
INSTITUTIONAL FORCES IN THE LABOUR
MARKET
A satisfactory model of the labour market must do more than
allow for market frictions. It is not realistic to assume that each
firm is an atomistic competitor, without influence over the
labour market and thus without a wages policy in the way in
which the pure competitive producer can have no price policy.
Nor is it true that all workers bargain individually with their
employers, to the exclusion of trade union activity. Thus:
The institutional market is distinguished by the substitution
of institutional rules for frictions as the principal delineator
1 If the market for a particular commodity is a perfect one, every
unit sold will fetch the same price.

14

of job market limits; of managerial and leadership comparisons for physical movement as the main basis of wage
markets; and of policies of unions, employers and government for the traditional action of market forces as the more
significant source of movements ... the job market no longer
alone sets the upper and lower bargaining limits for wage
determination ... the single price does usually exist but as a
consequence of policy and not the operation of market
forces. (Kerr [42] pp. 282-3)
We shall consider the impact of trade unions below; for the
present, the implications for the firm are more relevant.
THE INTERNAL LABOUR MARKET OF THE FIRM
In a totally casual or structureless labour market 'there is no
attachment except the wage between the worker and the
employer. No worker has any claim on any job, and no
employer has any hold on any man' (Kerr [41] p. 101). Such,
for example, was the market for harvest labour in the California
of the 1940s, analysed by Fisher [24]. Kerr argues that 'structure enters the market when differential treatment is accorded
to the "ins" and the "outs" , ([41] p. 101). We shall see that
this process makes each large firm a labour market in its own
right.
For many occupations there is no multi-employer labour
market at all in the traditional sense:
The first thing one notes in examining present-day hiring is
that much of it is conducted on a non-occupational basis, i.e.
specific occupational experience is not essential in the background of applicants for many industrial jobs, particularly
for the mass of semi-skilled machine operators and assemblers.
Most technological changes place a premium on speed,
dexterity and adaptability rather than on experience and
skill. (Raimon [67] p. 181)
This is perhaps less a function of technology itself than of the

15

diversity of processes, products and jobs in modem manufacturing industries. In any case, it implies that workers are
trained by individual firms. (Internal training is of course also
significant for skills which are not specific to the firm, e.g. most
maintenance skills.) Walter Oi observed in a pathbreaking
article [63] that firms' training costs, along with the unavoidable costs of hiring and firing, mean that labour can no longer
be treated simply as part of the firm's variable costs. He argued
that labour is a quasi-fixed factor of production.
The phenomenon of labour hoarding - the retention in
depressed periods of workers who would, in atomistic labour
markets, be dismissed - is entirely consistent with this argument. So is the universal finding, reported by Lester [49], that
non-probationary workers are never in practice subject to
competition for their jobs from outsiders; generally, firms do not
consider replacing established employees with new applicants,
even if the latter are prepared to accept lower wages. The
seniority system, whereby higher-grade jobs are filled by
internal promotion rather than by hiring from the outside, is
another example of the importance of the firm's internal labour
market. The system may be strengthened by pressure from
trade unions, but its prevaknce in the U.K., where such
pressure is weaker and less systematic, suggests that this is not
the sole cause.
For the typical non-craft production worker the 'port of
entry' (Kerr [41]) into the firm's internal labour market is at
the bottom of the job ladder, those lowest-paid jobs for which
outside recruitment is important. The worker progresses by
internal promotion on the basis of ability and seniority, but his
external mobility is hindered since he will have to start again
at the bottom of the job ladder in another firm. For craft and
low-level white-collar jobs the labour market is different.
In the latter case, internal promotion is again important, with
the port of entry located in the lowest clerical rather than
manual jobs. Craftsmen, however, are more able to move
horizontally (i.e. at the same occupational level) between firms,
and the port of entry is training itself-above all apprenticeship - or the membership of a union in crafts where the preentry closed shop ([86] p. 160) is prevalent.

16

For the great mass of manual workers, however, it is true

that 'Each company employment office is really a distinct
market for labour. The employed worker is attached basically
to a company, rather than to an industry or occupation'
(Reynolds [74] p. 42). This conclusion is reinforced when
another important neo-classical assumption is relaxed. Reynolds' study is just one of a number which have shown that
workers' information about wages and working conditions in
other firms is thin and patchy. Information is not costless, and
indeed the search for labour market information can be treated
as a form of investment (Stigler [91]).
The system of wage payment used by the firm is also relevant.
Payment-by-results (PBR), in various forms, is widespread in
manufacturing in most industrialised countries. In an excellent
recent survey, the National Board for Prices and Incomes [61]
found that 42 per cent of workers in U.K. manufacturing
in 1961 were paid by results, and that a further 31 per cent
were employed in plants where some workers were so paid.
Precisely because of the heterogeneity of job content within
industries, uniformity of PBR systems within an industry is
decidedly the exception rather than the rule, and this again
emphasises the firm as a labour market in its own right.
Even where attempts are made to standardise PBR systems
throughout an industry, these may be thwarted by the pressure
exercised by trade unions at the level of the workplace (Greenwood [30]). The fragmentation of collective bargaining in
what the Donovan Report called the 'informal system of
industrial relations in the U .K.' [86] is further evidence of the
irrelevance of industrial (even occupational?) labour markets.
The Report's endorsement of company-wide productivity
agreements was based less on the abstract arguments presented
above than on its recognition that large firms increasingly
regarded their wage policies as too important to be determined
by industry-wide negotiations through employers' associations.
(Remember that the atomistic firm of neo-classical theory does
not have a wages policy!)
One final 'institutional' point must be made. The use of job
evaluation [62] is increasingly common as a means of determining the firm's internal wage structure. Job evaluation

17

techniques provide a solution to a problem which ought not to

exist in terms of conventional theory: the structure of wage
differentials for the range of jobs performed within the firm.
Their use both implies that everything is not determined by the
external market, and reinforces the isolation of the internal
market, since industry-wide schemes are again relatively
uncommon. And since the techniques used are not scientifically
accurate, and the results of evaluation often differ significantly
between firms, the fragmentation of labour markets is carried
one stage further.
We have stressed the importance of the firm and the internal
labour market deliberately. It has important implications for
the study of labour supply and mobility, and for the evolution
of wage differentials of all kinds. Less obvious is the additional
dimension it adds to the analysis of labour markets: since the
'quality' or occupation of its workers its partly its own decisionvia internal promotion and training - the firIJl's hiring and
promotion standards become an economic variable. Job content, an important factor in the demand for labour, is not
determined solely by technology.
But firms do not exist in isolation from one another, any more
than labour markets. A major part of what follows will concern
the interrelationships between firms and their labour markets,
internal and external, and with the environment in which both
firms and workers operate. First, however, the firm's demand
for labour must be considered in a more rigorous theoretical
framework.

18

The Firm's Demand for

Labour

The marginal productivity principle has been at the root of

labour economics for a very long time. It is used both to
determine the firm's demand for labour, and to form a bridge
between the product and labour markets. As will be seen later,
it has also featured in the analysis of the distribution of income
and, to a lesser extent, of the causes of unemployment. A brief
outline of the principle itself will be given, followed by a
consideration of some of the criticisms made against it.
The firm's demand for labour is seen as determined by the
combination of a technological relationship - the production
function - and a market relationship - the demand for the
product. The former gives the marginal physical product of
labour, holding the level of all other factors of production (above
all capital) constant. The latter gives the marginal revenue
derived from small variations in a firm's sales. Multiplying the
marginal physical product by marginal revenue gives the
marginal revenue productivity of labour (MRP), the addition
to total revenue produced by the employment of one extra unit
oflabour, and net of changes in other costs (e.g. raw materials)
incurred. By equating MRP to the marginal cost oflabour, the
firm is fulfilling one of the conditions necessary for it to maximise its profits.
In the long run all factors of production are variable, and the
firm is faced with a choice between different factor combinations, all of which produce the same level of output. The firm
aims at the minimum cost for each level of output, and this will
be obtained by the application of the marginal productivity
principle to each factor. The equilibrium combination of
factors is arrived at in much the same way as consumer equilibrium is reached in indifference curve analysis.
19

Assume that there is pure competition in the firm's product

market, so that the demand curve facing it is perfectly elastic.
Then price equals marginal revenue, and MRP equals marginal
physical productivity times price, the value of the marginal
product. Similarly, if there is pure competition in the market
for the firm's labour, the supply curve of (assumedly homogeneous) labour to the firm will be perfectly elastic, and the
marginal cost oflabour will equal the wage. Thus, the firm will
be in equilibrium where the value of the marginal product
equals the wage.
In a sense this is only the principle by which the firm decides
its employment; it is not a theory of wages, since the wage is given
to the firm under the assumption of pure competition in the
labour market (see Cartter [13] chaps 1-2). However, we shall
see when looking at the distribution of national income
between the factors of production that the marginal productivity principle can be converted into a theory of wages with
no great difficulty.
Suppose that the firm is a monopolistic competitor in its
product market. Its demand curve is then downward-sloping:
an increase in its sales now leads to a lower price. But the price of
the intra-marginal units also falls, since we assume that the firm
is unable to discriminate between its customers. In this case
MRP is less than the value of the marginal product, because of
the reduction in revenue from the intra-marginal units (Chamberlin [14] chap. 8); this, however, makes little difference to
the broad implications of the marginal productivity principle.
What are these implications for the behaviour of the firm?
The curve relating MRP to the firm's employment of units
of labour - the firm's demand curve for labour - will slope
downwards whatever the type of product market competition,
since the marginal physical product is assumed to fall as more
labour is added to a given amount of capital. (This is the 'law
of variable proportions' or, less precisely, the 'law of diminishing
returns'.) The lower the wage rate, the more labour the firm
will employ;1 the effect of impure competition is only to make
1 The MRP curve may have an initial upward-sloping section;
but if the firm's marginal cost oflabour curve is horizontal, it is easy
to see that this section cannot contain a possible equilibrium for the
20

the downward tendency of the MRP curve even stronger, since

marginal revenue falls faster than price when the latter falls.
Nor is the tendency for the firm to substitute capital for labour
when the relative price of the latter rises in any way undermined
if product market competition is imperfect.
It would be a mistake to assume that a monopolistic competitor must also be a monopsonistic competitor, facing a rising an4
not a horizontal supply curve. If a dozen large firms produce
different brands of toothpaste, they are certainly monopolistic
competitors. If they are all located in London, and employ
only grades of labour not, specific to the toothpaste industry,
they are likely to be pure competitors for their labour. Conversely, if the whole of Norfolk is farmed by one large firm, it
will be a monopsonist with a rising labour supply curve, but
will still sell its wheat on the purely competitive world market.
Labour markets are fragmented on a geographical basis in a way
in which product markets usually are not, and in the absence
of skills specific to the industry substantial product market
power in itself implies nothing about a firm's labour market.
In any case, only minor revisions in the marginal productivity
principle are required if the firm's labour supply curve is
imperfectly elastic. The marginal cost of labour is now higher
than the wage, since increased wages must by assumption be
paid to the firm's existing labour force rather than merely to
new recruits. Thus on two counts the necessary condition that
MRP equals the marginal cost of labour no longer implies that
the wage is equal to the value of the marginal product. But the
negative relation between wages and employment, and the
substitution principle, are still valid.
The inapplicability of marginalist price theory for oligopoly
situations, though still disputed, is now generally recognised.
The implications for the firm's demand for labour have
received much .less attention. Attempts have been made to
apply the kinked product demand curve, with its discontinuous
marginal revenue curve, tq the labour market (Bloom [4]).
firm (any more than profits are maximised where a downwardsloping marginal ,cost curve intersects a horizontal marginal
revenue curve in the pure competitor's product market).

21

The result is a discontinuous MRP curve, in which the wage

can vary over a certain range with no short-run effect on
employment (though in the long run factor substitution
applies). Reynolds [75] argues that the main effect of a wage
increase in this range will be to make management more 'cost
conscious', involving more efficient organisation rather than a
direct reduction in labour inputs. If labour is semi-fixed, and
overheads at least semi-variable (Andrews [2]), even afall in
the product demand curve and hence in the MRP function may
lead to improved efficiency, thus shifting the MRP curve back
again! Reynolds suggests that short-run cost and product
demand curves are interdependent. A simpler conclusion might
be to reject the whole concept of a determinate product demand
curve (and hence MRP function) in oligopoly, since 'all other
prices' will not remain constant when one oligopolist changes
his, and he will realise this.
In a famous and controversial article, Lester [47] attacked
the marginal productivity principle, not in terms ofits analytical
deficiencies, but because the results it predicted failed to occur
in practice. The replies of a sample of U.S. manufacturing
firms to a questionnaire on the determinants of their employment decisions were interpreted as inconsistent with marginalism. By far the most important factor cited was product
demand, with the level of and changes in wages given far less
prominence. Those firms with plants in both the South and the
North denied that the South's relatively lower wages had
induced them to use more labour-intensive techniques in the
South. When questioned as to their reactions to an increase in
the relative Southern wage level, most firms gave as much
stress to improving methods and efficiency as to introducing
labour-saving machinery, and slightly more stress to improved
sales efforts than to changing price or product quality. Almost
none thought in terms of a reduction in output. Lester showed
that a sharp decline in the North-South differential between
1937 and 1941 in the footwear and clothing industries had
failed to prevent employment in the South from rising faster
than in the North. Within the South employment had risen
fastest in just those firms where wages rose most. Lester concluded that employment tends to vary 'simply and directly'

22

with product demand, and that 'For many manufacturing

concerns it is not feasible, or would prove too costly, to shift the
proportion of productive factors in response to changes in wages,
in the manner suggested by marginal analysis' ([47] p. 82).
In a stinging reply, Machlup [54] stressed the defects of the
questionnaire method, and of Lester's formulation of his
questions. More significantly, he attacked Lester's interpretation of his findings. The importance placed on 'product
market demand', he argued, was entirely consistent with
marginalist theory, and was in any case predictable since
demand tends to fluctuate more sharply and more often than a
firm's wage level. In answer to the question on the effects of a
wage increase, a significant proportion of the respondents
mentioned the introduction oflabour-saving machinery. A substantial proportion also referred to price and quality adjustments, and Machlup argued that these would lead indirectly to
lower employment via reductions in output. Machlup concluded that the wage-employment relation, and the substitution
principle, remained substantially unimpaired.
Powerful though it is, Machlup's critique is unsatisfactory.
He ignores the statistical evidence of the effects of increases in
Southern wage levels, and the rationalisation given by Lester's
respondents in terms of a general increase in efficiency. This
implies that the MRP curve is unstable; a move along it (wage
increase) shifts the curve as increased efficiency means a higher
MRP at every level of output (see p. 22 above). Machlup
himself denies any objective existence to the MRP curve, 'since
the raw material for the calculations could not come from any
records or documents, but merely from respondents' guesses of
a purely hypothetical nature' ([54] p. 548).
'Hypothetical' demand curves based on 'guesses' are a long
way from those of traditional theory. A tentative conclusion is
that the marginal productivity principle is not very useful in
explaining short-run situations, at least in oligopolistic markets.
For substantial wage changes the predicted results will no doubt
occur,but this is so close to tautology that it cannot be deemed
a very useful conclusion. Small adjustments at tke margin are at
best doubtful. We shall return to this point again, notably in
analysing the economic effects of trade unions.

23

3 Labour Supply
The problems of labour supply are multi-dimensional. They
involve firstly the decisions of individuals (or families) whether
to seek work or not, and how long to work each week or year.
These decisions, in aggregate, determine the supply of labour
to the whole economy; this is a product of the 'participation
rate' - the proportion of the relevant part of the population in
the labour force, i.e. working or seeking work - and hours
worked per week or year. Secondly, individuals must decide
what sort of work to do, determining the supply of labour to
specific occupations. Thirdly, they must decide for whom to
work, determining the supply oflabour to particular firms. The
first and third of these decisions will be discussed in this
section; the second is discussed below (Section 4) in the context
of occupational wage differentials.

It must be stressed that not all of the population above the

age of compulsory education are continuously employed.
Some are self-employed. Some, at any point in time, are unemployed but actively seeking work. Others are not even
participants in the labour force, owing to retirement, marriage,
continuing education, disability, etc. Of this last group, some
are occasional labour-force participants, sometimes seeking
employment, sometimes not: these constitute the 'secondary
labour force'. Movement into and out of the labour force, and
within the labour force from unemployment into employment,
may well be quantitatively as significant as the movement
between jobs of those constantly in employment.
THE INDIVIDUAL'S SUPPLY DECISION
Assume that labour supply decisions are taken in a purely
individual context, without reference to the family as a whole

24

(a simplification which will be dropped later). Assume further

that work and leisure ('non-work') are the only ways of spending one's time; alternatively, that a fixed part of one's time
must be spent satisfying certain basic physiological needs (e.g.
eating and sleeping), but that this is not influenced by economic
considerations. The individual must thus choose between work
and leisure, and it is reasonable to assume that relative prices
will play some part in his decision. In fact indifference analysis
can be used under these assumptions in exactly the same way
as in the theory of consumer behaviour. The opportunity cost
of leisure is the wages forgone, and thus the wage rate determines the slope of the price line facing the individual. Superimposing a set of indifference curves, his equilibrium hours of
work can be obtained.
This formulation permits the partial solution of a hoary old
problem in labour economics, the backward-sloping labour
supply curve for the individual. Robbins [78] was the first to
show that such authorities as Pigou and Knight were mistaken
in their belief that a wage increase, because it led to an increase
in the individual's real income, inevitably implied a reduction
in his hours of work because of the highly income-elastic
demand for leisure. As Robbins pointed out - without the
benefits of modern terminology! - this is to confuse the price
effect with the income effect. The effect of a change in the
relative prices of two goods is, of course, the sum of two parts:
the income effect and the substitution effect. Even if the income
effect is negative, the price effect need not be if the (necessarily
positive) substitution effect is strong enough.
A wage increase means both a rise in income and a rise in
the relative price of leisure (since the price of leisure to the
worker is the wage forgone). 'Leisure is not an inferior good',
and so the income effect implies that the worker will take more
leisure, and hence work fewer hours. But he will also tend to
substitute work for leisure, since leisure is now relatively more
expensive. (See Gilbert and Pfouts [28] for a mathematical
presentation in terms of the Slutsky equation of demand
theory.) Which is the stronger effect is an empirical question,
and not one which can be settled by a priori argument. And the
case for a backward-sloping supply curve to a firm or industry

25

is even weaker, as perversities in the shape of individuals' supply

curves may well disappear in the aggregation process.
The indifference curve technique has been widely used in the
analysis of hours of work (Perlman [65]). It can provide at
least a partial, supply-side, explanation of the prevalence of
overtime working in the U.K. Typically, a higher rate ('time
and a quarter', 'time and a third', etc.) is paid for overtime
hours. Because the increase applies only at the margin, the
income effect will be dominated by the substitution effect, and
the premium will normally be sufficient to induce the worker
to supply extra hours of work.
More formally, the sharp rise in the relative price of leisure
at the margin'induces a move to a higher indifference curve via
the supply of overtime hours, as shown in Fig. 1. In the absence
of overtime premia, the price line facing the worker is CED, its
slope OD/OC giving the hourly wage rate. The worker's

~-------!!:----~-------=~2~~c--Hours of
Leisure_

FIG. 1

-Work

leisure

equilibrium position is at E, where he works CA hours per day.

If a higher rate is offered for hours worked in excess of CA, the
price line becomes CEF, and the worker moves to a new
equilibrium position at G on the higher indifference curve 12,
working AB additional, overtime hours.
In practice, workers are unable to vary their working week
as they please. The standard working week is generally enforced as a minimum, with disciplinary sanctions against
absenteeism. It may also serve as a maximum, for example

26

because of trade union opposition to overtime working. In the

latter case, institutional restrictions on working hours may
induce workers to 'moonlight' by taking second, part-time jobs.
But, as Perlman shows [65], willingness to moonlight or work
overtime does not entail that a worker is underemployed
working the standard week, because of institutional constraints.
The reverse might even be true: a worker who would have
preferred, at the normal hourly wage, to work less than the
standard week may still work overtime or moonlight in
response to higher marginal rates of pay.
Why should firms be willing for their workers to work overtime? The alternative is the expansion of the labour force,
hiring more workers to work the standard week. Apart from
the incurring of extra 'quasi-fixed' costs, if the supply curve to
the firm is upward-sloping this will involve the firm in paying
wage increases to its existing workers as well as to the new
.entrants; blatant wage discrimination is rarely practicable. It
is often preferable to discriminate in a more subtle way by
paying more for marginal (overtime) hours only. Moreover, the
withdrawal of overtime in periods of slack demand for labour is
far easier than the direct imposition of wage-cuts. For workers,
too, overtime has advantages not easily expressed in terms of
formal analysis. Substantial evidence exists (Whybrew [100])
that much overtime in the U.K. is 'bogus', involving little
increase in effort or output but rather its dispersion Qver a
longer working week. This replaces increased pay for the
standard week, which employers may resist for reasons already
stated. Furthermore, the refusal of overtime without any
increase in hourly effort is often a cheap and effective substitute
for all-out strike action as a means of putting pressure on an
employer.
The usefulness of traditional indifference analysis can be
challenged on a more strictly theoretical level. Iflabour supply
decisions are made on a family and not on a purely individual
basis, strange results may occur. An increase in the husband's
hourly wage may induce him to work longer hours and his wife
to give up her job. Here the 'family income effect' is on the
wife's labour supply: if we looked only at the husband's
reaction, the income effect would appear very weak. The crude

27

division of time into work and leisure is also debatable. Leisure

is not the only substitute for paid employment (Mincer [58]).
Housework is not employment in the normal sense of the term,
but lavour-saving gadgets bought out of a wife's earnings from
ajob are a substitute for domestic work and thus relevant to her
decision to supply labour to the market. Commuting time is
another factor to be considered in this context (Perlman [65]).
Two-dimensional diagrams have a very restricted range of
applicability to such problems.
An institutional minimum working week means that the
decision to participate in the labour force becomes a problem
in its own right. A continuous trade-off between leisure and
paid employment is no longer possible, the choice being
between working the standard minimum hours or not working
at all. Indeed, as Hunter argues [35], the whole concept of an
income effect cannot be meaningfully applied to such a situation, since the initial income level, where no work is done, is
zero. From the viewpoint ofJamily income this may cease to be
a problem, but the 'lumpiness' of the participation decision
remains. Analysis in terms of marginal adjustments is no longer
realistic.
Empirical studies of the shape of labour supply curves are
necessarily undertaken at a high level of aggregation, but all
rest on some set of assumptions about individual behaviour
similar to those presented above. Over the last century weekly
and annual hours of work have fallen in all industrialised
countries, while real earnings have risen. Cross-section studies,
taken at a point in time, also suggest that working hours vary
inversely with earnings: skilled men tend to work fewer hours
than labourers, arid piece-workers fewer than time-workers.
Unfortunately, this cannot be taken as evidence that the income
effects of higher earnings outweigh their substitution effects.
Feldstein [23] has shown that most statistical attempts to apply
the simple economic model fail to prove anything about the
shape of supply curves. In technical language an identification
problem exists: a regression line relating a number of observed
combinations of earnings and working hours may show:
1. A backward-sloping supply curve traced out by shifts in
the demand curve.

28

2. A single demand curve, and several upward-sloping

supply curves (i.e. the regression line is a demand curve!).
3. The intersection of several demand curves with an equal
number of upward-sloping supply curves. In this case, the
regression line has little economic significance. It is illustrated in Fig. 2, where RR is the regression line.

Hours worked

FIG. 2

A similar problem appears to exist in attempts (e.g. Douglas

[18]) to relate participation rates to family income levels.

LABOUR MOBILITY
Although each firm's internal labour market can be considered
as a sub-market in its own right, this does not mean that there
are no links between these sub-markets. An expression of these
connections, and the most important criterion for the existence
of an articulated multi-firm labour market, is the movement of
workers between firms. Thus mobility provides a channel for
competition in the labour market; it also provides one line of
approach to a problem neglected until now: the nature of the
supply of labour to the firm.
The complexities of labour mobility are worth stressing (see
Hunter and Reid [36]). Actual movement between employers

29

is only one aspect. The ability to move - e.g. the possession of

relevant skills which are not specific to anyone firm - and the
willingness to move are perhaps even more important. It can
be argued that potential mobility exerts a greater influence on
the behaviour of labour markets and of wage structures than
does actual movement. (Movement from job to job will ali:lo
take place within the firm, suggesting the concept of disguised
mobility, which may be significant but will not of course be
reflected in statistics of movement between firms.)
More complications follow. Mobility (potential or actual)
from unemployment to employment, or from non-participation
to participation, is also relevant. This is especially so in view of
the evidence (Reynolds [74]) that most workers are firmly
attached to the jobs they hold, and that the long-term unemployed and the secondary labour force provide the great
bulk of those who are 'in the market' at any time. It is useful
to distinguish between changes in 'labour-force status', as just
described, and changes in job status reflected in movement
between employers. Related to this is the common distinction
betwe~n voluntary and involuntary mobility. Economists often
speak of workers changing their jobs in search of higher wage
or non-pecuniary benefits. This implies a conscious decision on
the part of job-changers which cannot be expected from those
who have been dismissed rather than quitting their jobs
voluntarily.
Even if restricted to voluntary and actual changes in job
status, labour mobility is hardly a simple affair. De Wolff [17]
provides a summary of a mass of evidence on types of job
changes. Taking voluntary and involuntary mobility together,
about one-third ofjob changes involve no change in occupation
or industry; of the remainder, changes in industry are far more
frequent, and are almost a necessary, though certainly not a
sufficient, condition for a change in occupation.
By definition, all actual mobility involves geographical movement, since a change in employer is the criterion by which
mobility is measured. But obviously short-distance mobility is
far more common than moves over long distances. Most job
changes thus occur within a local labour market (LLM), and
only these changes will be considered in what follows.

30

THE LOCAL LABOUR MARKET REVISITED

Derek Robinson has succinctly summarised the actual state of
affairs in local labour markets: 'If LLMs are to be described in
one word, they are "chaotic". Ifin two words, they are "bloody
chaotic". Despite the current easing of censorship, the editor
refuses to allow me to describe them in three words' ([79] p. 39).
His was the first British study to supplement a mass of American
evidence on the huge dispersion between firms in LLMs in
earnings within specific occupations. This dispersion is huge in
terms of anything that might reasonably be termed 'frictional'
or considered a result of minor market imperfections. For most
manual occupations, it is the exception rather than the rule for
the highest-paying firm in an LLM to pay less than 50 per cent
more, per hour, than the lowest-paying firm.
Note that this does not imply that mobility is limited. On the
contrary, both crude turnover rates and quit rates in manufacturing range up to 200 per cent of the labour force per year.
Job changes are not lacking; they simply do not produce the
results expected of them. Nor is there any evidence to suggest
that non-pecuniary benefits vary inversely with wages, so that
net advantages between firms are equalised despite large wage
dispersion. For fringe benefits, a major component of nonwage advantages, the reverse is probably true: high-wage firms
tend to have above-average fringes.
Perhaps wage dispersion does indicate an imperfect market.
Is it not at least possible that there is a long-run tendency to
perfect equilibrium, in which the law of equal price will eventually apply? If there is such a tendency, it is yet to be discovered. A recent American study suggests that intra-occupational wage dispersion in 85 LLMs narrowed very little
between 1954 and 1968 (Buckley [12]). A sophisticated neo-:classical economist might argue that the tendency to long-run
market perfection, though present, is obscured by changes in
parameters and shifts in functions. But in the long run we are all
dead: 'equilibrating forces' may be fighting a (losing) battle in
a dynamic world, but if the result is chaos, then chaos is
equally suitable as an explanation!
Alternative explanations for wage dispersion are not in fact

31

difficult to find. Job information is neither limitless nor costless,

nor do workers seem to seek it in the most rational way.
Typically, information about employment opportunities and
prospects in other firms comes through friends and relatives or
through direct and largely random approaches to firms. This
is not entirely an indicator of irrationality on the part of the
worker; it is also conditioned by employer hiring practices
(Reynolds [74] pp. 105-7). Firms often prefer to hire workers
who apply directly to them, so that many jobs are never
advertised or referred to government placement agencies.
Employers discourage workers from 'shopping around' for
jobs, offers of work usually being made on a 'take it or leave it'
basis (Lester [49]). The practice of hiring at the bottom and
promoting from within places institutional checks on labour
mobility, as we have seen. The existence of 'non-poaching'
agreements between firms, and the customary immunity of nonprobationary workers from outside competition for their jobs,
further reinforce the balkanisation of local labour markets.
Leaving aside the stark facts of wage dispersion, these
empirical findings themselves imply that labour markets are
highly imperfect. Labour mobility appears largely random,
with conscious decisions to maximise the net advantages from
employment playing a relatively minor part even in voluntary
job changes. If these conclusions are justified, the role of wage
differentials and changes in wage structures in the allocation of
labour is small. Inasmuch as mobility does have a rational,
non-random basis, the existence ofjob opportunities may be far
more important than wage differences in allocating labour
between firms, industries and occupations.
Before considering the implications for the supply of labour
to the firm, the validity of these empirical conclusions must be
examined. Rottenberg [83] criticised the methodological basis
of this argument. He suggested that it was more relevant to test
the predictions of orthodox theories of the labour market, than
to test their abstract assumptions. (Regarding the latter, he
insisted that rational job choice does not require perfect information and complete certainty, nor explicit (:alculation by
workers.) The predictions of neo-classical theory, he argued,
were easily verified in practice by the net tendency of workers

32

to move in the direction it implied, for example from low- to

high-wage regions.
Note that the predictions which Rottenberg imputes to his
theory do not have the degree of precision which earlier
economists asserted. That mobility is, on balance, towards
higher net advantage does not imply that advantages are
equalised. We have already seen that the law of equal price is
emphatically not verified in LLMs. Indeed, it is difficult to see
how eq~alisation could ever be verified, since non-pecuniary
factors are very difficult to quantifY. In his reply to Lampman's
[46] critique of his article, Rottenberg concedes this very
point:
The statement that net advantages are equal in all employments is derived logically from the convention that workers
maximise net advantage, employers maximise returns, and
all are free to pursue their objectives, in the same way that
the statement that two straight lines can intersect only in one
point is derived logically from the convention that only one
straight line can be drawn through any two points. Neither
is a prediction: their proofs lie in logical and not in empirical
experience. ([84] p. 639)
Rottenberg has, in fact, met the same fate which befell Machlup
in the latter's defence ofmarginalist price and wage theory: by
refining a simple theory with ever greater sophistication, he
has emptied it of all empirical content.
THE SUPPLY OF LABOUR TO THE FIRM
What does all this suggest about the supply of labour to the
firm? Since the early 1930s imperfectly elastic labour supply
curves have been generally accepted as a logical conclusion
from the existence of 'imperfect competition' between firms in
LLMs. To recruit more of a particular type of labour, an
employer must pay slightly more. (This cannot apply to
oligopsonistic markets, where a few large buyers of labour
recognise their interdependence; there the firm's supply curve
L.E.-B

33

is either kinked or indeterminate, like the oligopolist's demand

curve.) Various conclusions have been drawn regarding the
prevalence of the 'exploitation' of labour, and the enhanced
scope for trade union bargaining (Bloom [4]).
Reynolds [76] was the first to point out the internal inconsistencies of the upward-sloping curve. (For example, ignorance
alone cannot be a sufficient condition for an upward-sloping
curve, unless it can be dispelled by wage increases alone; this
does not seem likely.) He suggests that in all except fullemployment situations perfectly elastic supply is more realistic.
But the evidence on wage dispersion is equally relevant: if
workers do not endeavour to maximise net advantages, it is
difficult to see what the firm gains by offering a higher wage
rate; and if they do, only those firms at the bottom of the wage
ladder should find any difficulty in expanding their force at
their own going wage rate. Recruiting labour from the external
LLM is in any case only one way of overcoming a labour
shortage (D. Robinson [80]). For all except the lowest-skilled
jobs, an expansion of internal training and promotion, coupled
perhaps with a relaxation of job requirements, may be equally
effective. Overtime working in the short run, and productivity
bargaining in the long run, are also possible solutions within the
internal labour market. Just as the internal labour market may
be flexible enough to provide a solution, so it may be possible
to extend the frontiers of the LLM, e.g. by operating part-time
or evening shifts for married women, or providing transport for
workers from other areas. If all else fails, a firm can opt out of
the LLM altogether, moving its plant or subcontracting work.
These alternatives are especially important in oligopsony
situations, for they minimise the danger of retaliation by
competitors, which may be a serious deterrent to direct wage
increases.
On a number of grounds, the horizontal labour supply curve
isoa good working hypothesis for most firms. It is as well that
this is so. To reconcile an upward-sloping curve with the widely
accepted horizontal short-run marginal cost curve would be an
almost insuperable task. Here at least labour economics can
make a small contribution to micro-economics as a whole.
34

4 Wage Differentials
We have seen that any realistic analysis oflabour markets must
allow for both heterogeneity of labour and market imperfections. Indeed, one of the major defects of the simple marginal
productivity approach is that it says very little about differentials in wages between different types of workers in different
labour markets. To this problem we now turn.
The most comprehensive treatment of the subject (Reynolds
and Taft [77]) deals with five types of differential. Two of
these - interpersonal and inter-firm - have already been discussed. What follows is concerned with occupational, interindustry and geographical wage differentials. The fragmentation of labour markets is most obvious along occupational
lines. Mobility between occupations tends to be more impeded,
by 'market' and 'non-market' forces, than mobility between
industries and even regions. (At least as a generalisationmovement from floor-sweeper to labourer is clearly less difficult
than mobility of floor-sweepers from northern Scotland to
London.) Differences in pay between occupations form a large
part of the differentials between industries, and industry
differentials account for a major part of regional differentials.
These rather pragmatic considerations determine our order of
treatment.
OCCUPATIONAL DIFFERENTIALS
Adam Smith ([88] pp. 112-23) suggested five reasons for
differences in pay between occupations. Firstly, the 'agreeableness' of the job: net advantages, not merely cash payments,
must be considered. Secondly, the degree of constancy of
employment, the liability to unemployment in different occupations. Thirdly, the 'probability of success' differs between

35

jobs, being less for barristers than for solicitors, for example.
Both these factors suggest that discounted expected earnings
over a lifetime are what is involved in rational job choice,
rather than pay at anyone point in time. Fourthly, 'the trust
to be reposed in the worker', or responsibility requirements:
goldsmiths in Smith's day, workers maintaining expensive and
delicate machinery in our own, are well paid on this account.
Lastly, the 'cost of learning the trade', comprising tuition fees
and earnings forgone while undergoing training. (In modern
terms this is 'human investment'; cf. Becker [3].)
Abstracting from the first four factors, the fifth provides us
with a simple long-run theory of occupational wage differentials. Assume that workers bear all the costs of investing in
themselves. If natural ability is irrelevant, training open to all,
and all non-pecuniary advantages identical, the long-run
supply of every occupation will be perfectly elastic: pay
differentials between jobs will equate the rates of return on the
various amounts of human investment required. These assumptionsmay be modified - e.g. by allowing for variations in the
suitability of workers for positions of responsibility - without
substantially altering the conclusion that differences in pay
depend largely on differences in the level of education and
training required for different jobs. The demand for each
occupation has no impact, and relative shifts in demand will
have no effect on differentials.
An alternative theory of occupational wage differentials can
be developed by reversing one or two of these crucial assump..
tions. Again assume that non-pecuniary factors are identical,
and that Smith's second and third elements are inoperative, so
that the risks of failure and unemployment do not differ
between occupations. Suppose, however, that natural ability is
all-important, that part of the labour force is born as sheep and
part as goats, and that their abilities cannot be altered by
education or training. Then, since the supply of each is perfectly inelastic, their earnings will differ only because of
differences in the strength of demand for their services. If goats
are in greater demand, they will be paid more, and earn a rent
equal to their differential over the sheep. Changes in this
differential will occur only because of changes in relative

36

demand. (For an exposition of both models, see Reder

[72].)
It is unnecessary - and unwise - to rely solely on genetic
differences to derive a theory of occupational wage differentials. The division of society into non-competing groups may
occur on other grounds: for example, social prejudice (the lack
of Cockney ambassadors) or inequality in educational opportunity. Both factors may be eliminated in the long run as
occupational supply curves become much more elastic. The
consequent narrowing of differentials suggests that the goats
were only earning quasi-rents. A process of this sort will often
occur if training periods are long and changes in demand
cannot easily be foreseen: the relatively high earnings of
computer staff in the 1960s are an example. Demand factors
tend to dominate occupational differentials in the short run,
and underlying supply factors (above all human investment) in
the long run.
We know (Reynolds and Taft [77]) that there was a secular
decline in occupational wage differentials in both the U.K. and
the U.S. over the last half-century. This is true in general
between both manual and white-collar workers (Routh [85]),
and skilled and unskilled manual occupations. There is also
some evidence that differentials have moved counter-cyclically,
tending to rise in depressions and fall in booms. (Ozanne [64]
denies this tendency, but his data are drawn from one firm
only.) How far are these trends explicable in terms of our
simple models, and what alternative explanations can be
offered?
Part of the reduction in differentials must be attributed to
the massive and general expansion in state education. This,
with increased equality of access to human investment, has
reduced the rate of return to it, partially breaking down the
barriers between non-competing groups. In the U.S. the cessation of mass immigration of unskilled labour was an additional
factor. It has also been suggested (Keat [39]) that, as incomes
rise, the non-monetary factors in a job become increasingly
important to the worker, encouraging him to prefer lower-paid
but more enjoyable work. If occupational supply curves are
imperfectly elastic even in the long run, shifts in relative labour

37

demand may have affected differentials. Modem technology

is often assumed to have stimulated the substitution of less
skilled for more highly-skilled labour, and this would tend to
reduce differentials.
Reder [68] suggests that the loosening of hiring standards in
booms increases the supply of skilled labour relatively to that of
unskilled, and the reverse process operates when aggregate
demand is low. He explains the failure of differentials to widen
in some slumps by the existence of a conventionally accepted
'social minimum' standard below which the lowest-paid are
not allowed to fall. Partial support for his thesis comes from the
tendency for the relative severity of unemployment of lowlyskilled workers to increase in depressions. Oi's [63] theory of
labour as a semi-fixed factor leads to similar conclusions via
differential fluctuations in labour demand.
One big problem is the evidence - clear for the U.K., less so
for the U.S. - that major reductions in occupational differentials have been sudden, coming during and immediately after
the two world wars, and breaking periods of long-run stability
which in some cases lasted for centuries (Phelps Brown and
Hopkins [11]). But changes in education, and in relative

demand under the influence of technical progress, have been

much more gradual. This gives rise to an identification problem, for these periods of major change have coincided with
rapid iriflation'(Perlman [66]) and the growth of mass unionism
among lower-skilled workers (Turner [97]). Either or both
might be held responsible for narrowing differentials independently of broader economic factors. Routh [85] suggests that
occupational differentials are inherently stable because of social
conservatism, being displaced only by major shocks. A compromise position is possible: changes in the relative supply of and
demand for different occupations play only a permissive role,
requiring specific institutional forces to bring them into operation. Reynolds and Taft [77] make a similar point in arguing
that American differentials were too high, on simple economic
criteria, at the tum of the century, and that union pressure has
produced a more optimal occupational wage structure.

38

INTER-INDUSTRY DIFFERENTIALS
Assume, with Reder [69], a perfect labour market with
homogeneous labour in each occupation. All workers in each
occupation are thus paid the same. Average wages can differ
between industries only because of differences in the 'skillmix' : industry A's average wage will exceed industry B's if and
only if A has a larger proportion of highly-skilled workers. If
inter-industry differentials do exist, they will change only if
there are changes in occupational wage differentials, or differential changes in the skill-mix. (A's percentage wage advantage
over B will widen if, for example, its skill-mix improves while
B's remains the same.)
Reder relaxes the assumption of perfect occupational labour
markets with the postulate that the short-run supply curves of
specific occupations to specific industries are upward-sloping.
To expand its labour force, A must pay each occupation a
premium over its earnings in B. Thus in the short run changes
in inter-industry differentials will be positively correlated with
relative changes in employment between industries; in the long
run this relationship disappears as equilibrium is restored.
Evidence on this hypothesis is conflicting, and testing raises
important analytical problems (Perlman [65]).
Other empirical evidence strengthens our doubts as to the
validity of either the long- or the short-run version of Reder's
theory. The following have been found, although not unanimously, to be relevant to the structure and evolution of interindustry wage differentials: trade union activity (Lewis [51]);
product market concentration (Weiss [99]) and the average
size of plants; profitability and the level and rate of growth of
productivity (e.g. Brown [8]). All these factors are inconsistent
with Reder's simple models. Even productivity differentials
present problems, for if labour mobility is perfect and the law
of equal price prevails, either long-run marginal productivities
are equal in all industries or equilibrium conditions are absent.
The influence of trade union activity will b~ explored below.
At least part of the contribution of product market concentration to inter-industry differentials seems to be explained by the
higher-paying firms having 'the pick of the market', selecting

39

the ablest members of each (heterogeneous) occupational

group (Weiss [99])'. The role of plant size and profits may be
explained in the same way, though in view of our knowledge of
substantial market imperfections even within localities this is
unlikely to be the whole answer.
It is unclear how far these factors influence the course of
inter-industry differentials over time. One would not, for
example, expect highly concentrated or well-unionised industries to have a long-term tendency to pull away from the others
in the earnings league. The actual trend of inter-industry
differentials is still disputed, though it seems that the ranking
of industries by average wages has been relatively stable in the
long term, while the percentage dispersion has probably
narrowed (Cullen [16]). (Absolute cash dispersion is unlikely to
have narrowed; a 50p weekly differential, huge in 1900, is
probably within the margin of statistical error today!)
The forces of inertia can easily be exaggerated, and major
short-run fluctuations ignored; in one U.K. depression (19204) money-wage reductions ranged from zero to 60 per cent.
Likewise there is still little agreement generally on the factors
influencing changes in the inter-industry wage structure. This
would perhaps be no great loss, except that wage statistics are
generally collected on an industry basis (in the U.K. to the
exclusion of almost all others).
GEOGRAPHICAL DIFFERENTIALS
One reason for the plight oflow-wage areas is the concentration
of low-paying industries that they have collected (Ministry of
Labour [60]) - the industry-mix. This in turn is partly a reflection of regional variations in the skill-mix. Low-wage regions
are weak in their endowments of highly-trained, well-educated
labour. Regional wage differentials may thus be seen as the
sum of two elements: an excess of low-paying jobs, due to
unfavourable skill- and industry-mixes; and a tendency for
relatively low earnings in each job. The typical research
procedure has been to standardise for one or more of these
explanatory variables, and to treat the residual differential as
40

reflecting exploitation in the technical sense of unequal pay for

equal work.
These techniques have been widely used to analyse the
North-South differential in the U.S. Fuchs [26] obtains a very
crude approximation of the effects of regional differences in theskill-mix by assuming that the South's stock of human capital
was the same as that of the rest of the U.S., and found that
one-third of the unadjusted wage gap between the South and
the national average was thus explained. A further third was.
found, by a similar procedure, to be due to the smaller average
size of Southern towns and cities (reflecting, perhaps, greater
employer monopsony power in small labour markets). Much
of the residual third seems due to racial discrimination: not
only do Blacks form a major part of the Southern labour force,
but racial discrimination in employment is more severe in the
South than in the U.S. as a whole.
Fuchs and Perlman [27] used a similar procedure in showing
that the South's inferior industry-mix was a major cause of the
differential. They estimate that the South's relative wage level
will rise only slowly, for the industries which it is attracting
have been precisely those with a relatively low average wage in
the U.S. as a whole. It is not clear how this finding can be
reconciled with Scully's [87] report that the capital-labour
ratio ~ and hence, ceteris paribus, labour productivity - is now
above the national average in the South. This refers only tophysical capital; Scully found that the South's deficiency in
human capital was the single biggest factor in the region's
continued wage disadvantages, with racial discrimination
running it a close second.
All researchers are agreed that the South's relative earnings.
level has slowly improved since 1900, a finding which is.
repeated in fragmentary evidence for low-wage areas in the
U.K. (Knowles and Robertson [44]). It is uncertain how far
this has been due to the in-migration of capital predicted by
theory and verified by Scully's results, and how far to institutional factors like the extension of unionisation to the South, and
Federal minimum wage legislation. (In Britain the emergence
of industry-wide collective bargaining seems to have been the
single most important factor.) Confusion is increased by the

41

dispersion of regional differentials between industries: some

Southern industries pay more than their Northern counterparts,
and some very much less (Lester [50]). This evidence, confirmed
by a rudimentary examination also for regional differentials in
Britain, serves to underline the severe limits to our knowledge
of wage differentials generally.

42

5 The Economics of Trade

Unions
TRADE UNION WAGE POLICY
Strictly speaking, it is possible to assess the economic effects of
trade unions without any consideration of their aims and
motives. But any serious attempt to put union activity into an
economic context must also concern itself with what unions try
to do. We shall deal with this question before discussing what,
if anything, they have actually achieved.
The classic account of trade union wage policy is that of
Dunlop [21]. Assuming that unions were rational economic
units, he argued that they must be maximisers, like firms and
households are assumed to be. With a downward-sloping
labour demand curve, maximisation of the wage rate implies
that all but one of the union's members will be unemployed!
Maximising employment implies that the wage rate may be
very low. The total wage bill emerged by default as the most
probable maximand. Wages and employment must be considered together by the union, and the downward-sloping
labour demand curve is at the heart of Dunlop's analysis.
It is difficult to apply this simple model to oligopolistic firms,
or indeed to any realistic bargaining situation where the denial
to the other side of the relevant information -like the true shape
of the demand curve - is an essential feature. Moreover, it is
the long-run effects of wage increases on employment which
ought to concern the union, and traditional labour demand
curves are drawn on explicitly short-run assumptions.
None of these objections was basic to Ross's criticism of
Dunlop [82]. Ross stressed the nature of unions as organisations,
with goals in their own right distinct from those of the membership. Observation of American unions suggested to Ross that

43

the interests of union leaders might differ substantially from

those of the rank-and-file members around whom the Dunlop
model was centred. The behaviour of unions was not consistent
with maximisation: the use of 'coercive comparisons' in bargaining, the rarity of wage discrimination, and the prevalence
of coalition bargaining involving several unions in defiance of
the Marshallian 'importance of being unimportant', all implied
an interest in the survival of the union as an organisation, and
the continuing control of the leaders over it, rather than any
attempt to maximise the members' wage bill. Nor was the
connection between wage changes and inverse movements in
employment obvious and direct enough to be considered as
Dunlop thought it to be. In short, the basis of union policy was
poiitical (in the broad sense) and not economic.
The defects of treating unions as monopoly sellers of labour
are numerous and obvious (Mason [55]). The existence of
economic constraints on the policies of unions and their leaders
cannot, however, be doubted. Precise estimates of demand
elasticities are clearly neither possible nor necessary; the
employment effects of very large wage increases are certainly
taken into account by unions. Much of union policy on nonwage issues can be interpreted (Reder [70]) as an attempt to
forestall these effects: 'featherbedding' and 'restrictive practices', l enforcement of apprenticeship and entry controls,
seniority rules and other forms of intra-union stratific$ltion, and
1 The economic implications of 'featherbedding' deserve more
consideration than is possible here. ('Restrictive labour practices'
are described in some detail in the Donovan Report ([86] chap. vi)
and in the research paper referred to there.) Interpreted in one way,
these phenomena suggest that the firm's production function is no
longer a purely technological datum: both the possible factor combinations and the resulting levels of output are constrained by
institutional rules, and thus become variables rather than parameters.
On the other hand, firms are rarely optimisers in the sense that they
use the most efficient techniques and make the most efficient investment and output decisions open to them. Evidence abounds in the
reports of the National Board for Prices and Incomes; it is significant
that the Board was established to operate a productivity, prices and
incomes policy. Thus it can be argued that the 'shock effects' of
union pressure force firms closer to optimal positions (cf. Lester

44

direct intervention in the product market, e.g. in tariff campaigns, are all examples. The use of coercive comparisons is not
without economic significance, for the union must decide with
whom comparisons are to be made, and how far they can safely
be pushed when the adverse employment effects can be large.
Organisational pressures on union leaders have strong economic
roots.
Reder [70] has suggested an alternative, though qpt entirely
contradictory, explanation of unions' concern with non-wage
issues. They may be more an attempt to ration scarce jobs - the
'ethics of the queue' - than to maximise anything. Union job
controls are typically slackened in booms: overtime is viewed
with less disfavour, the use of 'dilutees' on skilled work less
strongly opposed, and membership restrictions less firmly
adhered to. Nor do unions always push their bargaining power
to its limit in booms, as evidenced by the continued existence of
a high level of job vacancies in highly organised sectors,
implying that the wage is insufficiently high to clear the market.
This suggests an important distinction: the ~ffects of unions may
reveal only part of their power, some of which will be held in
reserve until conditions deteriorate. Reder concludes that the
actual impact of unions on their members' wages will be small,
except perhaps in slumps.
This view of unions as regulatory agencies rather than
monopoly sellers is echoed in modern developments in industrial
relations theory (Flanders [25]). It is usually held in conjunction with an interpretation of the empirical evidence on the
wage effects of trade unions that suggests that these have been
small. To this empirical question we now turn.
THE ECONOMIC EFFECTS OF UNIONS
One approach is to treat collective bargaining as a tax and
study its incidence (Bronfenbrenner [6]). If union members
[47]); that these effects outweigh the institutional constraints mentioned above; and that the net contribution of unions to this aspect
of resource allocation is probably a beneficial one.

45

benefit, who loses, and how? The answer may, of course, be no

one: market forces may operate unaffected by trade unionism,
all appearances to the contrary. Or potential losses to nonmembers may be offset by gains: union pressure may accelerate
productivity growth ('the economy of high wages') or increase
consumption, aggregate demand, and hence the level of
national income.
If, however, income distribution is shifted in favour oflabour,
profits will suffer. If collective bargaining generates inflation,
fixed-income recipients will be the losers. If neither distribution
nor the price level is affected, and national income is not
boosted, any gains by union members will be at the expense of
other workers (and indirectly at that of society as a whole if
total output is actually reduced). Only this third, allocation aI,
aspect of the incidence of successful trade unionism will be
discussed here. Possible repercussions on distributional shares
will be analysed below, and inflation is outside the scope of this
study. It must, however, be remembered that redistribution of
income in favour oflabour as a whole, rather than redistribution
within the working class, is the avowed object of trade unionism.
There have been two starting-points for the analysis of the
resource allocation implications of unions, depending on one's
views as to the nature of typical labour markets. The assumption of perfect markets in the absence of unions is crucial to the
first approach; we shall adopt it at first and then relax it,
fundamentally altering our conclusions by so doing.
Assume, with Rees [73], homogeneous labour (e.g. a single
occupation) sold in a perfect market. Each man receives the
same hourly wage, which just equates the perfectly inelastic
supply with market demand. Suppose that part of the occupation - in one region or industry - forms a trade union which
demands and receives a higher wage than this competitive
equilibrium level. Then employment will fall in the 'union
sector', and the perfectly inelastic supply to the non-union
sector will rise, pushing down the wage level there.
In Fig. 3, Df}, and Dn are the labour demand curves for the
union and the non-union sectors; they are linear, parallel and
sum to give Dt, the demand curve for the occupation as a
whole. In the union sector, wages rise from 'W"o to Wf}" and employ-

46

ment falls by AB. Since D", is the MRP curve for this sector, the
area beneath it represents its total revenue product, which has
fallen, as a result of the wage increase, by ABFE. In the nonunion sector, supply increases by AB ( = CD) to AS; wages fall to

s
On

w,,1-----\

ABC 0

Employment

FIG. 3

Wn and employment rises by CD, giving an increase in output

of GHDC. Assuming linear, parallel demand curves, the reduction in total output (GHDC-ABFE) is shown by the shaded
area in the diagram; this is, approximately, the sectoral shift
in employment multiplied by one-half the wage differential
between the two sectors.
Rees then makes an ambitious attempt to estimate the size
of this area, and hence the 'output costs' of trade unionism, for
the U.S. In this he draws on Lewis' work on the effects of trade
unions on the inter-industry wage structure [51]. By a series of
very careful comparisons between the wages of union members
and non-members in otherwise similar labour markets, Lewis
estimated the average effects of unionism as follows: for each
percentage point difference in the percentage of workers who
are union members - the 'union density' - earnings are increased by between 01 per cent and 02 per cent. From this
Rees calculates the net loss in output caused by union 'distor-

47

tions' of the inter-industry wage structure at about one-seventh

of 1 per cent of U.S. GNP in 1957. This follows from the relatively small wage effects indicated by Lewis, who also found,
interestingly, that only in major depressions could the effects of
unions on relative wages be considered really large.
An opposite viewpoint has also emerged (Reynolds and Taft
[77]; Kerr [40]). If there are significant labour market imperfections, the intervention of unions may improve resource
allocation in the sense that wage differentials conform more
closely to those which would be found in a perfect market (see
also Kerr [42]). Thus successful implementation of the 'common rule' can end wage discrimination against individuals
within the firn:t, and reduce unjustified wage dispersion in local
labour markets. It was argued above that union pressure may
have been responsible, in both Britain and the U.S., for a
narrowing of occupational differentials which was probably
overdue on strictly economic criteria; this in itself would affect
inter-industry differentials. In Britain the major reduction in
geographical differentials between 1914 and 1920 probably
reduced significantly the exploitation of workers in isolated
areas. Some of the reduction in the u.s. North-South differential may be attributed to the extension of unions to the South.
Some qualifications are necessary at this point. Even if
unionism is a 'good thing' in the very technical (and long-run)
sense of changing wage structures in the direction of competitive
equilibrium, its initial effects may be rather different. Union
membership has not grown evenly, so that, for example, the
large nineteenth-century differential of the skilled worker may
have been maintained or increased by a union movement
which had little impact on the unskilled. The importance of the
(temporary) emergence of labour aristocracies is strengthened
by evidence that the initial wage effects of unionism are greater
than its permanent influence on the wage structure.1
1 This is only part of the story. A complete theory of labour
aristocracies would have to explain the structure and growth of the
trade union movement, in terms of the interaction between organisations and, inter alia, the pressures of the labour market. To take one
relevant example, the emergence of a 'craft' may be due only partly
to technology-determined 'skill requirements' (see Turner [98]).

48

Once again, the average union member would deny that his
gains, if any, were at the expense of other workers. Moreover,
both supporters and critics of trade unionism would agree that
its impact on non-wage issues has been at least as important, in
every sense.

49

6 The Functional Distribution

of Income
The analysis of the functional distribution of income entails the
acceptance of a rather traditional view of the factors of production. The Ricardian definition of land is unsatisfactory, since
land is neither 'original' (since it incorporates capital investment) nor indestructible. The difficulties of defining and
measuring capital are many and well known. Recently it has
been realised - or rather rediscovered - that labour inputs also
involve investment; the concept of human capital is still far
from fully digested in distribution theory. Independent proprietors, including farmers, clearly use all three of the traditional productive factors, giving rise to imputation problems.
They also represent a source of a fourth, very troublesome,
factor - entrepreneurship. Convenience and common sense suggest,
as they have done to most modern statisticians, a simple distinction between labour and property incomes.
If we accept the marginal productivity approach to factor
pricing, it seems at first that no distribution problem can exist.
Every unit of every factor is simply paid the value of its marginal product. Unfortunately a problem remains unless the sum
of the marg~nal products is exactly equal to the total product.
If it is not, then either there is a surplus which must somehow
be distributed, or the claims of the factors cannot be met. This
is the famous 'adding-up problem'. Two solutions have been
presented, both in the early years of this century, by Wicksell
[101]. By the use of Euler's theorem, it can be shown mathematically that if the aggregate production function - itself a
tricky concept - is linear homogeneous, the sum of the marginal
products is exactly equal to the total. In economic terms, this
implies constant returns to scale. This seemed unduly restrictive;
the production function is determined by technology, and it can

50

only be by accident that it should take this particular form.

Wicksell was able to prove that the adding-up problem also
disappeared so long as each firm was in purely competitive
long-run equilibrium. The reason is intuitively obvious: the
firm will then operate at the bottom of its long-run average cost
curve, and this (infinitesimal) flat section of the curve is exactly
analogous to constant returns to scale.
Remove the assumption of pure competition, and this argument provides no solution. The wage no longer equals the value
of the marginal product, and the firm no longer attains equilibrium at the lowest point of its long-run average cost curve. A
related problem concerns the disputed role of 'pure' profit, as
a return to entrepreneurship as opposed to capital. The entire
theory of profit is in a very uncertain state, and the traditional
position of the entrepreneur in distribution theory has been
subjected to severe criticism (Joan Robinson [81]).
Hence the simplification already mentioned: the division of
national income into labour and property (or capital) shares.
Attempts to provide empirical verification for the marginal
productivity theory of distribution are invariably in these
terms. What is now known as the Cobb-Douglas production
function was first suggested by Wicksell as one case of the linear
homogeneous function necessary to satisfy Euler's theorem. In
its simplest form, with output (r), labour (L) and capital (K),
and constants a and b, the function is r = KaLb. 1 Simple calculus
reveals that the respective marginal products of capital and
labour, arj aK and arj aL, are arjK and brjL, and their shares
in national income a and b. Data for output, capital and labour
were obtained in both time-series and cross-section form, for
states over a period of time and for industries within a nation
at a point in time. Regression yielded values for a and b which
not only resemble the actual shares of the two factors, but also
add up almost exactly to one (Douglas [19]).
Criticisms at this attempt at verifying neo-classical distribu1

In this, later, form the Cobb-Douglas is homogeneous of degree

(a+b). If a+b= 1, the degree of homogeneity is unity and the

adding-up problem is solved. Douglas adopted this form rather than

the question-begging form Y=KaLa-l, which assumes that the
function is linear homogeneous.

51

tion theory are best surveyed by a sympathiser (Bronfenbrenner

[5]), and by a strong critic (Phelps Brown [9]). There is an air
of circularity about it, since it is only on the assumption that
the factors arc paid their marginal products that a and b become
their shares in national income. This of course assumes competitive equilibrium in both labour and product markets, which
is what is under scrutiny. Phelps Brown [9] argues that
researchers have not identified a production function at all, let
alone proved that it is the determining factor in income distribution. On a time-series basis, the Cobb-Douglas 'can describe
the relations between the historical rates of growth of labour,
capital and output, but the coefficients that do this do not
measure marginal productivity' ([9] p. 551). The cross-section
studies he finds equally inconclusive.
Another, still explicitly micro-economic, approach is that of
Kalecki [38]. He adopts a neo-Marxist definition of profits as
including all salaries and all other overheads, assumes horizontal short-run variable cost curves and a 'mark-up' pricing
procedure, and arrives at a formula which expresses the wage
share in national income as an inverse function of the degree of
monopoly (the ratio of price to prime cost) and the relative
price of raw materials. Theoretically there is much to be said
for the model; its explicit recognition of departures from pure
competition and marginalist pricing procedures was overdue.
Less satisfactory is the vague distinction between wages and
salaries (or 'productive' and 'non-productive' labour) and the
way in which the degree of monopoly - reflected in the size of
the mark-up component of price - is itself left undetermined.
Kalecki's model suggests a bargaining-power theory of distribution, which is plausible and yet empty because nothing is
said about the nature and sources of bargaining power.
Kalecki's own attempts at empirical testing failed; since no
evidence was available for the degree of monopoly, he reversed
the direction of causality and derived changes in the latter
from well-documented movements in the labour share in
national income. Yet the general approach can be valuable in
careful, non-econometric hands (Phelps Brown [10]).
One reason for the continuing interest in relative factor shares
is its relevance, as a major determinant of effective demand,

52

for macro-economic theory. The rich tend to save a larger

proportion of their incomes than the poor, so that aggregate
consumption and saving are strongly influenced by income
distribution. Conversely, the level of national income may
determine relative factor shares. A pioneer in the macroeconomic theory of distribution was Kaldor [37]. He abstracted
from changes in real national income by assuming full employment. Then, if investment is given and the propensity to save
out of profits is greater than that out of wages, the distribution
of income will be determined solely by the two savings propensities. Crudely, if planned investment exceeds planned saving,
the resulting inflation will increase the profit share in national
income and thus raise ex post saving to the required level.
One empirical test for periods of full employment found that
the core of Kaldor's model- the different savings propensities
of workers and capitalists - provided no better explanation of
actual changes in distribution than the assumption that they
were equal (Reder [71]). Moreover, a theory which largely
ignores the behaviour of firms and individuals at micro-level
cannot be considered definitive. Despite these criticisms, the
macro-economics of distribution must be part of any complete
analysis; in formal terms alone, the marginal productivity
approach is insufficient and requires the introduction of some
information about aggregate demand (Solow [89]).
If theoretical work on distribution is in a state of flux, the
same is no less true of empirical work. Not long ago, the labour
share was seen as a constant no less fixed than 'the velocity of
light and the incest taboo' (Solow [90] p. 618). This result
could be supported by almost any theory. The Cobb-Douglas
approach necessarily involves the constancy of relative shares,
since a and b are mathematical constants. If relative shares do
change, we must explain the consequent changes in a and b,
and hence the shifts in the function itself, the stability of which
would no longer provide a theory of distribution! More sophisticated neo-classicists showed that, whatever the production
function, very large changes in the relative quantities oflabour
and capital would be required to shift distribution to any significant degree (Bronfenbrenner [7]). Kalecki arrives at a rather
unconvincing explanation of a stable wage share in terms of

53

offsetting variations in the degree of monopoly (which falls in

booms), and the relative price of raw materials (which rises).
Stable investment-income ratios and constant consumption
propensities would give the desired result in terms of Kaldor's
theory.
However, the evidence suggests that relative shares have not
been constant. In the last two or three decades the labour share
in national income has risen in most Western countries, and the
increase is not always 'eliminated' when allowance is made for
the rising proportion of employed (and corresponding decrease
in self-employed) members of the occupied population (Heidensohn [31]). The labour share in the U.K. is much higher than
it was a century ago (Feinstein [22]), and Kravis' careful
research suggests the same trend in the U.S. [45]. It is true
that, in both cases, this can be explained in terms of the vastly
increased ratio of capital to labour in the economy and the
consequent, larger, increase in the relative marginal productivity of labour as opposed to capital (in technical terms, an
elasticity of substitution between capital and labour ofless than
unity).
But the changes took place, in the U.K. at least, in the form
of two major 'faults', during and immediately after the two
world wars and breaking up periods of general stability
(Feinstein [22]). We thus have exactly the same problem as we
encountered when trying to explain the decline in occupational
wage differentials (see above, pp. 37-8). The increase in the
labour share could equally have been due to inflation, to state
intervention or to increased trade union pressure. None of the
theories which has been described provides a very convincing
solution to this problem.

54

7 The Size-Group Distribution

of Income
Estimates of the relative shares of capital and labour are only
one, rather unsatisfactory, way of describing the degree of
income inequality. Individual incomes, even of employees, may
include elements of both wages and property incomes, and the
dispersion within each category of income may also be large.
We shall ignore the distribution of property incomes - with the
significant exception of 'human capital' - observing only that
this will depend on the distribution of wealth, savings propensities and attitudes to risk-taking, and inheritance taxation and
social customs (Reder [72]). We shall deal only with the
distribution of earnings.
Why is a separate theory of earnings dispersion necessary,
when occupational wage differentials have 'already been
analysed? Labour is not homogeneous, nor equally well paid,
within occupations. Possibly only 25 per cent of U.S. earnings
dispersion is attributable to differences in average earnings
between occupations, the remainder being due to intraoccupational dispersion (Lydall [53]). The well-known difficulties of obtaining satisfactory occupational definitions reinforce the need to analyse the distribution of individuals'
earnings.
Ceteris paribus, part-time workers earn less than full-timers,
and those liable to unemployment less than those with secure,
stable jobs. Women tend to earn less than men, juveniles less
than adults, and farm-workers less than workers in other
industries. Lydall's 'standard distribution' eliminates these
influences by restricting the coverage to full-time adult male
workers outside agriculture. The resulting distribution is positively skewed, and leptokurtic or hump-shaped. The distribution for the top 20 per cent of incomes approximates to Pareto's

55

law: N = AX-a, where A and a are constants and the expression

gives the number of incomes (N) above a given income (X).
Apart from the top 20 per cent and the bottom lO per cent of
incomes, the distribution is normal in the logarithm of income.
The shape of the distribution, in particular the skewness, has
been found even when the standardisation procedure is less
rigorous than Lydall's.
These empirical findings, which apply across a wide range of
countries and over substantial time-periods, provide an interesting starting-point for theoretical work, especially in view of the
paradox, to be discussed below, that earnings should on certain
assumptions be normally distributed. Another finding worth
consideration is the positive correlation, up to late middle age,
between age and earnings, and the tendency for the dispersion
of earnings to be greater in the higher age-groups. (This itself
contributes to the overall dispersion of earnings, which is lower
within a specific age-group than for men of all ages.)
The paradox is this: earnings are assumed to depend on
productivity, which depends on ability; but psychological
testing suggests that ability is normally distributed, and this is
difficult to reconcile with the skewed, lognormal distribution of
earnings. Solutions take two forms: the addition of further
explanatory variables, and a denial that ability is normally
distributed. Taking the second argument first, Mayer [56]
argues that the measurement of ability is an essentially arbitrary process, and the resulting distribution reveals more about
the structure of the test than the dispersion of abilities. Compare
a test involving the solution of ten differential equations with
one requiring the addition of two and two: the observed distribution of abilities will be rather different! Mincer goes further
and suggests that the explanation of earnings in terms of abilities
is circular: 'Stating facts about distributions of productivities
comes dangerously close to stating facts about distributions of
earnings. Both sets of facts require explanation' ([59] p. 6). It
may be more appropriate to reverse the argument, so that the
dispersion of earnings acts as an indicator of the distribution of
economically relevant abilities, and not vice versa.
It is in any case possible to arrive at a skewed earnings
distribution without any assumption about the distribution of
56

ability, in terms of a stochastic process theory. The role of

chance is summarised by Mayer:
Chance effects can lead to a lognormal distribution if the size
of an income increase, or the probability of its occurrence, is
a function of the previous size of the variable. For example,
if we take a group of people and in successive periods give
$1 to individuals selected at random, we will eventually
approach a normal distribution. But if, instead of giving the
fortunate individuals $1 each, we give them, say, 1 per cent
of their previous income, we can get a lognormal distribution.
([56] p. 189)
It can be argued that this bold adaptation of the Central Limit
Theorem is a gesture of despair; that chance effects 'explain'
only that residual dispersion for which there is no other explanation. The corollary is that the importance of stochastic theories
will decline as more strictly economic theories of earnings
dispersion are developed. However, this statistical approach is
important in that it illustrates a more general point of wide
application: if two elements, both of which are normally
distributed, combine multiplicatively to determine men's earnings, the resulting distribution of earnings will be positively
skewed. This would be the case if, for example, the two
elements were positively correlated.
To clothe the bare bones of this highly abstract proposition,
consider some possible applications. Lydall [53] suggests that
'general educated ability' is a mixture of genetic ability, energy
or motivation, and environmental factors. The latter include
education itself, the distribution of which is known to be highly
skewed. But even if all elements are normally distributed and
are independent of each other, if they combine multiplicatively
rather than additively, then general educated ability (and
hence earnings) will be positively skewed.
Mayer [56] postulates a positive correlation between a man's
ability and his 'scale of operation', measured by the number of
subordinates he controls and/or the value of production he is
responsible for. Given normally distributed or positively
skewed abilities, the distribution of productivities and hence of

57

earnings will be positively skewed, and this conclusion is

reinforced if a superior is able to pass on his ability to his
subordinates. It is further reinforced if the employer has more
reason to discriminate between workers of different abilities, the
larger the scale of operation. Reder [72] discusses this related
concept of 'sensitivity' .
This principle may explain the 'responsibility premium' in
occupational wage differentials (see above, p. 36). Men may
rejoice in responsibility and yet be well paid for it; demand
factors overcome the supply side. Lydall explains the 'Pareto
upper tail' in his standard distribution in a similar way: in
hierarchical institutions it is an organisational- as much as an
economic - necessity to pay the man at the top more than his
subordinates. Evidence for these two related arguments is
found in the far stronger relationship between managerial
salaries and company turnover than between salaries ~nd profits,
and in the importance given to responsibility in job evaluation
schemes.
Mincer [58] has suggested that the shape of the overall
earnings distribution and rising dispersion with age can be
explained in terms of investment in human capital. (This term
is often abbreviated to 'education',which can be misleading as
it ignores all post-school training.) Education involves a private
cost, at least in terms of the scholar's forgone earnings, and a
return in higher future earnings. (For a similar analysis of onthe-job training, see Becker [3] chap. 1.) In very simple terms,
a man's earnings in periodj are

Tj=Xj+

j-1

L rtCt-Gj

t=o

where Xj is the earnings of 'raw uneducated ability', rt and Ct

respectively represent the private rate of return to education
and the level of investment in education in each period. Cj is
the - zero or negative - contribution to earnings from education
undergone in the current period (i.e. the earnings forgone).
Current earnings are thus composed of a base element, Xj,
which may be ignored, the costs of current education, Cj, which
are assumed zero after formal education ceases, and the returns
58

to the individual's accumulated human investment, rC for

short.
From our original statistical argument, and ignoring Xi, a
skewed distribution of earnings emerges even if rand Care
uncorrelated and both normally distributed, simply because
they combine multiplicatively. And since rC accumulates with
age, the skewness of the earnings distribution will also increase
with age. It is in any case probable that the distribution of C
is itself skewed; such is the evidence from the education
statistics. Furthermore, if rates of return to education reflect
differences in ability - as assumed by university selection
procedures! - rand C will be positively correlated and the
skewness of the earnings distribution further increased.
A more general economic, rather than stochastic, approach
to the problem has been made by Tinbergen [95], in terms of
the degree of 'tension' or mismatching between the distribution
of job requirements and the distribution of skills and other
relevant characteristics in the working ,population. Despite its
generality, and conceptual (if not mathematical!) simplicity, it
is open to a major criticism: Tinbergen specifies his job and
skill characteristics as parameters, whereas in any long-run model
of income distribution they must surely be regarded as variables.
Job requirements are flexible in response to differential wage
changes (Reder [68]; above, p. 38); and the human capital
approach pivots on the assumption that different types of skill
are, in the long run, similarly wage-elastic.
Each of these theories, or some combination of the elements
of each, can explain the almost universal shape of the distribution of earnings. Can we also explain differences in earnings
dispersion, both between countries and within individual
countries over time? Lydall~s careful research reinforces conclusions drawn in many previous studies, that income dispersion
is greater, the poorer the country (excluding certain Eastern
European countries where dispersion was, at least until
recently, less than would be predicted from relatively low per
capita incomes). Since the dispersion of education is also
greatest in poor countries, this is quite consistent with the
human capital approach.
Similarly, the apparent reduction in income inequality in most

59

advanced Western countries in the present century may be

partly attributed to an increase in both the volume of, and
equality of access to, education. In the V.S. there seems to have
been a sharp increase in income equality between 1939 and
1949, which may have been due to an equally major extension
of public education in the previous decade, magnified by onthe-job and military training (with civilian spin-offs) in the
war (Miller [57]). In Britain, income inequality again seems to
have fallen sharply in the 1940s, following major educational
advances in the inter-war period.
Two cautious conclusions must be stressed. While the results
of cross-sectional, inter-country comparisons of income inequality show differences large enough to be beyond dispute,
the same cannot be said of intertemporal studies. One's view of
the significance of increased egalitarianism depends largely on
one's optimism as to the validity of the statistics! Titmuss'
celebrated critique [96] of the British data applied mainly to
the many opportunities of tax evasion open to recipients of
property incomes, and thus is only indirectly relevant here
(though one of the defects of current work on size-group
distribution is precisely its failure to integrate labour and
property income, suggesting the continuing need for some
theory of functional shares). Thatcher [93] found that the
dispersion of earnings of manual workers in the V.K. had
scarcely changed between 1886 and 1966, despite all the
evidence of a sharp narrowing in occupational wage differentials (see above, pp. 37-8). It is certainly rather strange to find
Miller [57] vouching for the accuracy of V.S. data on the
grounds that 'people don't lie to the tax-man'.
Secondly, it is not clear that increasing the general level of
education will inevitably and always lead to a reduction in the
inequality of earnings. If education and ability are at present
positively correlated, the swamping of the human capital
element in earnings by a vast expansion of education which
lowers the private rate of return may increase inequality by
stressing differences in abilities more sharply than before.
Meritocracy may not be the answer (Reder [72]). We may be
at the stage where compensation for innate and environmental
disadvantages is necessary, involving a negative correlation

60

between education and ability (LydaU [53]). This is becoming,

nominally at least, a major feature of anti-poverty programmes,
and appears more promising than the initiation or improvement
of minimum wage legislation, the employment effects of which
are still disputed and possibly large (Hildebrand [34]). An
attack on income inequality may thus also serve to reduce any
structural element in unemployment (see below, Section 8).
And so it should, as the inclusion of unemployed and partly
employed workers greatly increases the dispersion of incomes.
The principle of hierarchy in large organisations has been
seen as one factor in the skewness of the whole earnings distribution (Mayer [56]), or the Pareto tail of extremely high incomes
(Lydall [53]). It is uncertain how far high incomes in leading
positions in such organisations are merely institutionalised rent
payments which could be eliminated without any loss of
productive efficiency. In a really radical attack on inequality,
attention to property incomes and the very poor should not
obscure the possibility of democratising bureaucracies.

61

8 Unemployment
Post-war complacency about unemployment was dispelled in
the U.S. by a failure, in the relatively prosperous late 1950s and
1960s, to achieve tolerably full employment, l a problem which
now appears to be facing Britain too. Unemployment is also
seen as a major factor in poverty and income inequality, and
its nature and causes are once again far from a dead issue.
Traditionally, a fourfold classification of unemployment has
been adopted. Seasonal unemployment is self-explanatory,
relatively unimportant in Britain, and easily 'eliminated'
statistically. Frictional unemployment (Uj) results from the
immobility caused in the short run by imperfections in labour
markets. Structural unemployment (Us) is caused by more
fundamental maladjustments between labour supply and
demand. Cyclical 'unemployment varies with the level of
aggregate demand; modern' growth theory suggests a secular
variant, 'growth-gap unemployment', resulting when the
economy's growth potential exceeds its actual rate of growth.
Both this and cyclical unemployment can be termed demanddeficient unemployment (Ud).
Assuming Uj relatively constant at an 'irreducible minimum'
level, the recent American controversy centres on the alleged
increase in Us since the early or mid-1950s. For orthodox
Keynesians, any excess of unemployment over the bedrock Uj
level is to be explained largely in terms of deficient aggregate
demand. Thus something approaching full employment can be
achieved by manipulating aggregate demand. The 'Keynesians'
(e.g. Heller [32]) do not deny that the structure of unemployment is unbalanced, being higher, for example, among teen1 In the U.S. this generally means about 4 per cent of the labour
force; this would be somewhat reduced if British definitions of
unemployment were used, but it remains, in European terms, very
high.

62

agers, Blacks, and unskilled and poorly educated people

generally. This is explained, given the level of aggregate
demand, in terms of the 'shape-up' analogy: for a given
number of jobs, employers vary their hiring standards so as to
select the best available workers, leaving those of inferior
quality (or the wrong colour) unemployed. But jobs would be
found for large numbers of such groups as demand expanded.
The apparent success of the 1964 tax cut gives strong support
to their case.
Some 'structuralists' remain unconvinced (Killingsworth
[43]), observing that Keynes' was a short-run theory of employment, with such factors as the structure of demand and the state
of technology held constant. But changes in these factors form
the core of the structuralist case. The expansion of the tertiary
at the expense of the secondary sector, the exodus from agriculture, the development of automation, have all altered the
structure of the demand for labour in favour of well-educated,
highly-skilled and white-collar workers, at the expense of the
rest, whose fate is seen in the increasing imbalance in the
incidence of unemployment. Killingsworth denies the power
of the market as the 'great homogeniser' of labour; expanded
demand will not create jobs on any scale for those groups with
the highest unemployment rates, for market substitution is
simply not strong enough. He explains the apparent success of
the tax cut in terms of several distorting factors, notably the
Vietnam war. Further evidence is seen in the increase over the
last two decades of the white-collarfblue-collar wage differential, and in differentials for workers with differing degrees of
educational attainment. The cure for unemployment thus lies
in direct action to correct the structural imbalance of the labour
market, for example by massive retraining programmes.
The controversy appears in a rather different light once a
precise definition of the different types of unemployment is
attempted. Lipsey [52] suggests the relevance of the Phillips
curve relating unemployment and the rate of change of
money-wage rates, which reveals the trade-off between unemployment and inflation. His.modified Phillips curve - using
the rate of change of prices rather than wage rates on the x-axis
- is shown in Fig. 4.

63

PC'

So
5

.c:.
u

PC

Ir-_~

.g l'
0.

'0
cP

, c,
r ,
,

,,
...

b
Unemployment (%)

..

...... .......

PC
......... pc#

FIG. 4

Superimposed is the policy-makers' indifference curve (II),

showing various equally-valued combinations of unemployment and inflation. These assume that price increases and
unemployment are substitutes for each other, with an increasingly large reduction in unemployment needed to justify each
successive percentage point rise in the rate of inflation.
An initial unemployment rate of Ob can be reduced by
demand expansion to Oa with a stable price level,1 and to Oc at a
still acceptable inflationary cost. (E is equivalent to a consumer's equilibrium position; given the objective possibilities
shown by the Phillips curve, it is the best position attainable by
the policy-makers.) Thus Ud, = be. Unemployment can be
further reduced only if the Phillips curve is shifted back to the
left, e.g. by a retraining programme to elimjnate a mismatching
1 In the long run.

64

due to skill differences in the structure of the demand and supply

of labour. Otherwise the inflationary costs are unacceptably
high. But retraining is also expensive, and only a shift to PC'
may be considered worthwhile on cost-benefit grounds. This
gives a new equilibrium position (E') on a higher indifference
curve (1'1'), with a#reduction of unemployment of cd, and a
lower rate of inflation. Lipsey defines U, = cd, and U/ = Od,
which is again the irreducible minimum level of unemployment given the cost of retraining and allied policies.
Lipsey's classification depends on the preferences of policymakers and on the existence and stability of the hotly disputed
Phillips curve. He ignores job vacancies (unsatisfied demand),
concentrating solely on unemployment (unsatisfied supply).
Nor is his approach easily testable: a necessary condition for a
rise in U, (given the policy-makers' preferences) is a rightward
shift of the Phillips curve, but since the curve may shift for many
other reasons, this is hardly a sufficient condition. Two interesting implications may, however, be noted: U, is unlikely ever to
be large - it is a maximum of Oc in Fig. 4 - and its existence and
size are quite independent ofthe composition of unemployment.
The second approach to the disaggregation of unemployment
owes its inspiration to Dow and Dicks-Mireaux [20], and is
elaborated by Perlman [65]. It starts with a simple definition of
excess supply: the total supply oflabour (SL) is th~ sum of those
employed (N) and those seeking work (the unemployed, U).
Total demand (DL) comprises employment (N) plus jobs
offered (unfilled vacancies, V). Hence excess supply is
SL-DL=(N +U) -(N + V) = U- V

which is equivalent to demand-deficient unemployment Ud.

Since Uti. cannot meaningfully be negative, when there is
excess demand for labour all unemployment (seasonally adjusted)
is either frictional or structural. In symbols:
(1) Ud,=SL-DL= U - V~O
(2) Ud,=Ofor U-V~O
thus, since by assumption U = Ud, + U, + Uf,
(3) U = U, + U, when U - V ~ O.
Dow and Dicks-Mireaux used the general term 'maladjustment' to denote the causes of non-demand-deficient unemployL.B.-Q

65

ment, which they defined as UB + Ut as above. Before discussing

the possibility of separating Us and Uf, we must see how Ud, and
'non-Ua' may be measured.
Perhaps the major limitation to this approach is its dependence on the existence of reliable job vacancy data. Such data
are available only quite recently for Britain, and not at all for the
U.S.; in any case, since the reporting of vacancies is not compulsory, and agencies other than state employment bureaux are
used by firms hiring workers, reported and 'true' vacancy data
may diverge. At best we have an ordinal index showing directions and rough orders of magnitudes of changes in the level of
vacancies, rather than an accurate cardinal indicator. (Similar
considerations apply to unemployment data: cyclical variation
in participation rates implies a varying reserve of 'hidden
unemployment', and labour hoarding also operates to undermine the accuracy of published unemployment statistics - see
Taylor [92].)
Ignoring these problems, Thirlwall [94] suggests some simple
techniques for distinguishing Ud, and non-Ud,. Since by definition Ua = 0 when U = V, the points in time where total unemployment equals total vacancies indicate the level ofnon-Ud,;
and a rising trend would suggest rising UB' Regression of U
against V, or (U - V), produces the same result.
This second approach has obvious advantages over the first.
It does not depend on anyone's preferences, nor on the suspect
Phillips curve (though it will be suggested below that this
advantage is more apparent than real). Its implications are
equally interesting. Thirlwall [94] suggests an average postwar total UB + Ut for Britain of about 16 per ,cent, again a
relatively low figure. Once more, information about the composition of unemployment is insufficient to allow any classification into Ud, and non-Ua.
This latter point is easily proven. Assume two sectors, A and
B, with equal and constant labour forces of 1,000 (LA =L B
= 1,000). These might represent two different industries, or
regions, or skilled and unskilled labour. Take an initial arbitrary combination of employment, unemployment and job
vacancies, and define Ut as that part of total unemployment
which disappears in the short run as a result of normal labour
66

mobility. (This does not imply that U, ever becomes zero, only
that different individuals will be frictionally unemployed in
each time-period.) Consider the following two cases:

Casel
Employment
N A = 980
N B = 900
Unemployment UA=LA-NA=20 UB=LB-NB=100
Vacancies
VA = 100
VB = 20
Here total unemployment is 120 (6 per cent of the labour force).
The ratio of unemployment in the two sectors is 5: 1. Total
vacancies equal total unemployment, so that by definition all
unemployment is non-Ua. By the end of the period, the 20
A-unemployed men take 20 of the A-vacancies and 20 of the
B-unemployed take the 20 B-vacancies. This leaves 80 Avacancies and 80 B-unemployed; since Ua is zero, and UJ is, on
our earlier definition, 40, it follows that U, = 80, or 4 per cent
of the labour force.
Case II
Employment
N A = 980
N B = 800
Unemployment UA=L A -NA=20 UB=LB-NB=200
Vacancies
VA=20
VB=O
Here total unemployment is 220 (11 per cent of the labour
force). The ratio between the two sectors is 10: 1. By the end of
the period, the 20 A-unemployed have filled the 20 A-vacancies
awaiting them; but the 200 B-unemployed remain unemployed,
there being no vacancies for them. Thus Ua=200, U/=20 (the
number of vacancies. filled in the short run by labour mobility)
and UII is zero. The two cases are illustrated in Fig. 5.
Compare case II with case I: total unemployment is much
higher, and its incidence between the two sectors is much more
unequal. But at the end of the period, when all frictional
unemployment has been eliminated, there is no structural
unemployment. In case I the opposite is true: there is no Ua,
the remaining unemployment all being structural. An interesting conclusion follows: U8 exists only where there are equal
numbers of unfilled vacancies, of which there are none in case
II. Generally, the higher is the overall level of unemployment,

67

120

120

{'~ ---4

60
40
20

A
U

100

Case I

Ud

80
60
40

20
A
0

CaseR

FIG. 5

the less likely it is that U, will be high: 'That there is an

excessive number of square pegs, which cannot fit round holes,
contributes nothing to a given level of unemployment, so long
as there is also a surplus, however small, of round pegs' (Perlman
[65] p. 173). Thus it is misleading, as some economic historians
have done, to describe Britain's unemployment in the 1930s as
predominantly 'structural'. In the absence of vacancy data in
substantial detail, both schools of thought in the contemporary
U.S. debate are defeated by inadequate information!
It can still be argued that unemployment can be increased,
and given structural undertones, because of wage and price
rigidities. Hildebrand [34] suggests that the influence of shifts
in the structure of labour demand has been small compared to
that of 'unrealistic' wage increases for lowly-skilled workers.
Given a sufficiently large wage differential, employers will have
an incentive to substitute unskilled for skilled workers (turning
A-vacancies into B-vacancies in the example above), or to
reduce the differential by training (turning the B-unemployed
into A-workers). High unemployment rates among vulnerable
social groups may be less a function of either deficient aggregate
demand or ofstructural shifts in demand, than evidence that these
types of labour are being priced out of the market. This is a very
sanguine view of the power of market substitution, and ignores
the recent evidence of rising skill and education differentials.
68

It is likely that the wage reduction necessary fully to employ the

young, unskilled and poorly educated would be much greater
than anything politically or socially acceptable (quite apart
from effects on aggregate demand which might increase
unemployment in other categories of the labour force).
There is still a basic defect in the above analysis. The logic of
case II situations suggests the need for an expansion in aggregate demand. There is no point in retraining coalminers as
computer programmers if there are no jobs available in either
occupation! But suppose that demand is increased, but extra
jobs become available only for programmers: the result will be,
in Lipsey's terms, an unacceptable rate of inflation, while the
overall level of unemployment will remain the same. We shall
have converted case II into case I, and demand-deficient into
structural unemployment! Agnostic solutions - 'more retraining and demand expansion' - fail to say how much of each
is required.
It may be apt to end a rather pessimistic survey of labour
economics with two further unsolved problems in the theory of
unemployment. One is the relation between frictional unemployment and the level of aggregate demand; far from
staying constant whatever aggregate demand, U, probably
varies with it, but which way is still uncertain (Corry and
Laidler [15]). Secondly, the definition of full employment as
that level of unemployment where the supply of and demand
for labour are, in aggregate, equal, is unsatisfactory. Perlman
[65] observes that a position where there are 13 million
unemployed and 14 million vacancies is, on this definition,
(over-)full employment!

69

Bibliography
[1] R. G. D. Allen, Mathematical Ana[ysis for Economists
(Macmillan, London, 1960).
[2] P. W. S. Andrews, Manufacturing Business (Macmillan,
London, 1949).
[3] G. Becker, Human Capital (National Bureau for Economic
Research, N ew York, 1964).
[4] G. Bloom, 'A Reconsideration of the Theory of Exploitation', Quarter[y Journal of Economics (1940-1); reprinted in
A.E.A., Readings in the Theory of Income Distribution
(Allen & Unwin, London, 1950).
[5] M. Bronfenbrenner, 'Neoclassical Distribution Theories',
in]. Marchal and B. Ducros, The Distribution of National
Income (Macmillan, London, 1968).
[6] M. Bronfenbrenner, 'The Incidence of Collective Bargaining', Southern Economic Journal (Apr 1958).
[7] M. 'Bronfenbrenner, 'A Note on Relative Shares and the
Elasticity of Substitution', Journal of Political Economy
(June 1960).
[8] D. G. Brown, 'Expected Ability to Pay and Interindustry Wage Structure in Manufacturing', Industrial
and Labour Relations Review (Oct 1962).
[9] E. H. Phelps Brown, 'The Meaning of the Fitted CobbDouglas Production Function', Quarter[y Journal of
Economics (Nov 1957).
[10] E. H. Phelps Brown, 'The Long-term Movement of Real
Wages', in J. T. Dunlop (ed.), The Theory of Wage
Determination (Macmillan, London, 1966).
[11] E. H. Phelps Brown and S. Hopkins, 'Seven Centuries of
Building Wages', Economica (Aug 1955).
[12] K. Buckley, 'Intraoccupational Wage Dispersion in
Metropolitan Areas, 1967-8', Month[y Labour Review (Sep
1969).
70

[13] A. M. Cartter, The Theory of Wages and Employment

(Irwin, Homewood, Ill., 1959).
[14] E. H. Chamberlin, The Theory of Monopolistic Competition
(Oxford D.P., 1966).
[15] B. Corry and D. Laidler, 'The Phillips Relation: A
Theoretical Explanation', Economica (May 1967).
[16] D. E. Cullen, 'The Inter-Industry Wage Structure,
1899-1950', American Economic Review Gune 1956).
[17] P. deWolff, Wages and Labour Mobiliry (O.E.C.D., Paris,
1965).
[18] P. Douglas, The Theory of Wages (Kelley, New York,
1934).
[19] P. Douglas, 'Are There Laws of Production?', American
Economic Review (Mar 1948).
[20] J. C. R. Dow and L. M. Dicks-Mireaux, 'The Excess
Demand for Labour: A Study of Conditions in Great
Britain, 1946-56', Oxford Economic Papers (Feb 1958).
[21] J. T. Dunlop, Wage Determination under Trade Unions
(Blackwell, Oxford, 1950).
[22] C. Feinstein, 'Changes in the Distribution of National
Income in the D.K. since 1860', in J. Marchal and
B. Ducros, The Distribution of National Income (Macmillan,
London, 1968).
[23] M. Feldstein, 'Estimating the Supply Curve of Working
Hours', Oxford Economic Papers (Mar 1968).
[24] L. Fisher, 'The Harvest Labour Market in California',
Quarterly Journal of Economics (Nov 1951).
[25] A. Flanders, 'Collective Bargaining: A Theoretical
Analysis', British Journal of Industrial Relations (Mar
1968).
[26] V. Fuchs, Differentials in Hourly Earnings by Region and Ciry
Size, Occasional Paper 101 (National Bureau for Economic Research, New York, 1967).
[27] V. Fuchs and R. Perlman, 'Recent Trends in Southern
Wage Differentials', Review of Economics and Statistics
(Aug 1963).
[28] F. Gilbert and R. Pfouts, 'A Theory of the Responsiveness of Hours of Work to Changes in the Wage Rate',
Review of Economics and Statistics (May 1958).

7I

[29] ]. F. B. Goodman, 'The Definition and Analysis of Local

Labour Markets: Some Empirical Problems', British
Journal of Industrial Relations (July 1970).
[30] ]. Greenwood, 'Payment-by-Results Systems: A CaseStudy in Control at the Workplace with a National
Piecework Price-List', British Journal of Industrial Relations (Nov 1969).
[31] K. Heidensohn, 'Labour's Share in National Income: A
Constant?', Manchester School (Dec 1969).
[32] W. Heller, 'Employment and Manpower', in S. Lebergott (ed.), Men Without Work (Prentice-Hall, Englewood
Cliffs, N.J., 1964).
[33] ]. Hicks, The Theory of Wages, 2nd ed. (Macmillan,
London, 1963).
[34] G. Hildebrand, 'Some Alternative Views of the Unemployment Problem in the U.S.', in]. Stieber (ed.),
Employment Problems of Automation and Advanced Technology
(Macmillan, London, 1966).
[35] L. C. Hunter, 'Some Problems in the Theory of Labour
Supply', Scottish Journal of Political Economy (Feb 1970).
[36] L. C. Hunter and G. Reid, Urban Worker Mobility
(O.E.C.D., Paris, 1968).
[37] N. Kaldor, 'Alternative Theories of Distribution', Review
of EcoT/omic Studies, xxm (1955-6).
[38] M. Kalecki, The Theory of Economic Dynamics (Allen &
Unwin, London, 1954).
[39] P. C. Keat, 'Long-run Changes in Occupational Wage
Structures, 1900-1956', Journal of Political Economy
(Dec 1960).
[40] C. Kerr, 'Wage Relationships: The Comparative Impact
of Market and Power Forces', in]. T. Dunlop (ed.), The
Theory of Wage Determination (Macmillan, London, 1966).
[41] C. Kerr, 'The Balkanisation of Labour Markets', in E.
Wight Bakke (ed.), Labour Mobility and Economic Opportunity (Wi1ey and M.I.T. Press, New York, 1954).
[42] C. Kerr, 'Labour Markets: Their Character and Consequences', American Economic Review, Papers and Proceedings (1950).
[43] C. P. Killingsworth, 'Full Employment and the New

72

Economics', Scottish Journal oj Political Economy (Feb

1969).
[44] K. G. J. C. Knowles and D. J. Robertson, 'Differences
between the Wages of Skilled and Vnskilled Workers,
1880-1950', Bulletin oj the Oxford Universi9' Institute of
Statistics (Apr 1954).
[45] I. Kravis, 'Relative Income Shares in Fact and Theory',
American Economic Review (Dec 1959).
[46] R. Lampman, 'Comments: On Choice in Labour
Markets', Industrial and Labour Relations Review (July
1956).
[47] R. Lester, 'Shortcomings of Marginal Analysis for WageEmployment Problems', American Economic Review (Mar
1946).
[48] R. Lester, Communication, ibid. (Mar 1947).
[49] R. Lester, 'A Range Theory of Wage Differentials',
Industrial and Labour Relations Review (July 1952).
[50] R. Lester, 'Diversity in North-South Wage Differentials
and in Wage Rates within the South', Southern Economic
Journal (Jan 1946).
[51] H. Gregg Lewis, Unionism and Relative Wages in the United
States (Chicago V.P., 1963).
[52] R. G. Lipsey, 'Structural and Demand-deficient Vnemployment Reconsidered', in A. Ross (ed.), Employment
Policy and the Labour Market (California V.P., Berkeley,
1965).
[53] H. Lydall, The Structure oj Earnings (Oxford V.P., 1968).
[54] F. Machlup, 'Marginal Analysis and Empirical Research', American Economic Review (Sep 1946).
[55] E. Mason, 'Labour Monopoly and All That', Proceedings
of the Industrial Relations Research Association (Dec 1955).
[56] T. Mayer, 'The Distribution of Ability and Earnings',
Review of Economics and Statistics (May 1960).
[57] H. P. Miller, Rich Man J Poor Man (Crowell, New York,
1964).
[58] J. Mincer, 'Labour Force Participation of Married
Women: A Study in Labour Supply', in National
Bureau for Economic Research, Aspects of Labour Economics (N.B.E.R., Princeton, 1962).

73

[59]
[60]
[61]
[62]
[63]
[64]
[65]
[66]
[67]
[68]
[69]
[70]
[71]

[72]

[73]

74

J.

Mincer, 'The Distribution of Labour Incomes: A

Survey with Special Reference to the Human Capital
Approach', Journal oj Economic Literature (Mar 1970).
Ministry of Labour, 'Effects of Regional Employment
Structures on Average Earnings', Ministry of Labour
Gazette (Mar 1969).
National Board for Prices and Incomes, Payment by
Results pystems, Cmnd 3627 (H.M.S.O., London, 1968).
National Board for Prices and Incomes, Job Evaluation,
Cmnd 3772 (H.M.S.O., London, 1968).
W. Oi, 'Labour as a Quasi-fixed Factor', Journal oj
Political Economy (Dec 1963).
R. Ozanne, 'A Century of Occupational Wage Differentials in Manufacturing', Review of Economics and Statistics
(Aug 1962).
R. Perlman, Labour Theory (Wiley, New York, 1969).
R. Perlman, 'Forces Widening Occupational Wage
Differentials', Review oj Economics and Statistics (May
1968).
R. L. Raimon, 'The Indeterminateness of Wages of
Semi-skilled Workers', Industrial and Labour Relations
Review (Jan 1953).
M. W. Reder, 'The Theory of Occupational Wage
Differentials', American Economic Review (Dec 1955).
M. W. Reder, 'Wage Differentials: Theory and Measurement', in National Bureau for Economic Research,
Aspects oj Labour Economics (N.B.E.R., Princeton, 1962).
M. W. Reder, 'The Theory of Trade Union Wage
Policy', Review of Economics and Statistics (Feb 1952).
M. W. Reder, 'Alternative Theories of Labour's Share',
in M. Abramowitz et al. (eds), The Allocation oj Economic
Resources: Essays in Honour oj B. F. Haley (Stanford U.P.,
1959).
M. W. Reder, 'A Partial Survey of the Theory ofIncome
Size Distribution', in L. Soltow (ed.), Six Papers on the
Size Distribution of Income and Wealth (National Bureau for
Economic Research, New York, 1969).
A. Rees, 'The Effects of Unions on Resource Allocation',
Journal of Law and Economics (Oct 1963).

[74] L. G. Reynolds, The Structure of lAbour Markets (Harper,

New York, 1951).
[75] L. G. Reynolds, 'Towards a Short-run Theory of Wages',
American Economic Review (June 1948).
[76] L. G. Reynolds, 'The Supply of Labour to the Firm',
Quarterly Journal of Economics (May 1946).
[77] L. G. Reynolds and C. Taft, The Evolution of Wage
Structure (Yale U.P., New Haven, 1956).
[78] L. Robbin~, 'On the Elasticity of Demand for Income in
Terms of Effort', Economica (June 1930).
[79] D. Robinson, 'Myths of the Local Labour Market',
Personnel (Nov 1967).
[80] D. Robinson, Wage Drift, Fringe Benefits and Manpower
Distribution (O.E.C.D., Paris, 1968).
[81] Joan Robinson, 'Euler's Theorem and the Problem of
Distribution', Economic Journal (Sep 1934).
[82] A. Ross, Trade Union Wage Policy (California U.P.,
Berkeley, 1948).
[83] S. Rottenberg, 'On Choice in Labour Markets', Industrial and lAbour Relations Review (Jan 1956).
[84] S. Rottenberg, 'Reply', ibid. (July 1956).
[85] G. Routh, Occupation and Pay, 1900-1960 (National
Institute of Economic and Social Research, London,
1964).
[86] Royal Commission on Trade Unions and Employers'
Associations (Donovan Commission), Report, Cmnd 3623
(H.M.S.O., London, 1968).
[87] G. Scully, 'Interstate Wage Differentials: A Crosssection Analysis', American Economic Review (Dec 1969).
[88] A. Smith, The Wealth of Nations (Methuen, London,
1961).
[89] R. Solow, 'Distribution in the Long and Short Run', in
J. Marchal and B. Ducros, The Distribution of National
Income (Macmillan, London, 1968).
[90] R. Solow, 'A Sceptical Note on the Constancy of Relative Shares', American Economic Review (Sep 1958).
[91] G. J. Stigler, 'Information in Labour Markets', Journal
of Political Economy (Oct 1962).
[92] J. Taylor, 'Hidden Unemployment, Hoarded Labour

75

[93]
[94]

[95]
[96]
[97]
[98]
[99]
[100]
[101]

and the Phillips Curve', Southern Economic Journal (July

1970).
A. R. Thatcher, 'The Distribution of Earnings of Employees in the U.K.', Journal of the Royal Statistical Socie~,
series A, part 2 (1968).
A. P. Thirlwall, 'Types of Unemployment: With Special
Reference to "Non-Dema:ld-Deficient" Unemployment
in the U.K.', Scottish Journal oj Political Economy (Feb
1969).
J. Tinbergen, 'A Positive and a Normative Theory of
Income Distribution', Review of Income and Wealth (Sep
1970).
R. Titmuss, Income Distribution and Social Change (Allen &
Unwin, London, 1962).
H. A. Turner, 'Trade Unions, Differentials, and the
Levelling of Wages', Manchester School (Sep 1952).
H. A. Turner, Trade Union Growth~ Structure and Policy
(Allen & Unwin, London, 1962).
L. Weiss, 'Concentration and Labour Earnings', American Economic Review (Mar 1966).
E. Whybrew, Overtime Working in Britain, Royal Commission on Trade Unions and Employers' Associations,
Research Paper No.9 (H.M.S.O., London, 1968).
K. Wicks ell, Selected Papers in Economic Theory (Allen &
Unwin, London, 1958).

Note: Published too late to be incorporated into the main

bibliography, Derek Robinson (ed.), Local Labour Markets and
Wage Structures (Gower Press, London, 1970) is a major contribution to the problems considered in Sections 1 and 3 of this
study, as is D. 1. MacKay et al., Labour Markets under Different
Employment Conditions (Allen & Unwin, London, 1971). See also
P. B. Doeringer and M. J. Piore, Internal Labour Markets and
Manpower AnalYsis (Heath, Lexington, Mass., 1971).

76

MACMILLAN ECONOMICS BOOKS

FOR THE STUDENT

Economics
SOVIET ECONOMIC POWER
ITS ORGANIZATION, GROWTH AND CHALLENGE

Second Edition
Robert W. Campbell
An examination of the historical background of Russia's
present economic development, the current performance of her
economy, and the methods by which this performance has been
achieved.
Paper 125 SBN 333 01872 9

THE OBJECTIVES OF MACRO-ECONOMIC POLICY

Ajit K. Dasgupta and A. J. Hagger
An advanced text in Keynesian macro-economics built around
the four objectives-internal balance, external balance, price
stability and a high growth rate-of advanced countries. For
third-year and postgraduate students.
495

SBN 333 123964

MODERN ECONOMICS
J. Harvey
A textbook of current economic theory for professional examinations and business studies, presenting. economics as a
method of thought, not a mere body of knowledge.
150 SBN 333 10401 3
MODERN ECONOMICS STUDY GUIDE AND
WORKBOOK
J. Harvey and M. Johnson
This Study Guide and Workbook draws attention to the
important points upon which the student should concentrate

77

his study, tests his understanding of the basic concepts, and

indicates how economic theory is applicable to the practical
decisions of the firm and government. While it has been
designed particularly to accompany Modern Economics, it can
be used with similar texts which cover the same subjectmatter.
90p SBN 333 10533 8
INTRODUCTION TO MACRO-ECONOMICS
J. Harvey and M. Johnson
A clear and concise treatment of macro-economic theory
intended to bridge the gap between general coverage in basic
textbooks and the advanced treatment of more specialised
books.
Paper 160 SBN 333 12511 8
THE GENERAL THEORY OF EMPLOYMENT,
INTEREST AND MONEY
J. M. Keynes
Lord Keynes wrote in his preface: 'This book is chiefly addressed
to my fellow economists. I hope that it will be intelligible to
others. But its main purpose is to deal with difficult questions
of theory, and only in the second place with the application of
this theory to practice.'
Paper 75p SBN 333 00942 8

Economic Growth
OUTPUT, INFLATION AND GROWTH
AN INTRODUCTION TO MACRO-ECONOMICS

D. C. Rowan
A first-year undergraduate course in macro-economics with
emphasis primarily on theory; economics is regarded as a
discipline with measurable and quantifiable concepts.
Paper 225 SBN 333 10187 1
78

PLANNING AND GROWTH

T. Wilson
'Mr Wilson has brought together a collection of essays which
are marked by robust common-sense and a deep understanding
of the complex relationship of economics and politics.'
- Times Literary Supplement
Paper 75p SBN 333 04384 7

Business Economics
THE ECONOMICS OF 'MANAGERIAL' CAPITALISM
Robin Marris
In this stimulating book, the author has attempted to replace
the old notion of the businessman as one primarily interested
in profit, with the original thesis that the profit motive has
given way to the concept of growth.
Paper 105 SBN 333 08704 6
BUSINESS FINANCE AND THE CAPITAL MARKET
K. Midgley and R. G. Burns
An up-to-date textbook that examines the Capital Market
from the point of view of both supply and demand. It provides
first-class instruction for all students reading for professional
examinations, C.N.A.A. and other degrees, and H.N.D. and
H.N.C. Business Study courses.
Paper 1'50 SBN 333 10410 2

79