THE IMPACT OF INDUSTRIAL RELATIONS PRACTICES ON

EMPLOYMENT AND UNEMPLOYMENT

David Marsden *
Centre for Economic Performance, London School of Economics, Discussion
Paper No. 240, May 1995

Abstract

The rules and institutions of collective bargaining are widely held to have an
adverse effect on employment and thus on unemployment. These views are
analysed, and it is argued that many industrial relations institutions provide
a much greater degree of flexibility for firms than is often realised. It is often
forgotten that collective agreements involve the agreement of employers who
are thus able to influence the rules that are fixed. Nevertheless, there are many
areas where collective rules conflict with the needs for firms to raise
productivity and to adapt to market changes.

The paper then examines a number of the sources of flexibility in such areas
as low inflation bargaining structures, industry agreements, workplace
cooperation, and mechanisms for youth inclusion. It also reviews some of the
developments in the area of flexibility agreements, and recent reforms of
collective bargaining systems in a number of OECD countries, all features
likely to increase the ability of firms to adjust in these countries.

It concludes with a number of policy recommendations for further reform of

industrial relations systems including the need to combine some form of
macro-level coordination with greater flexibility at the enterprise level.

This paper accompanies a similar paper by the same author on management

practices and unemployment.

This paper was produced as part of the Centre's

Programme on Industrial Relations.
 THE IMPACT OF INDUSTRIAL RELATIONS PRACTICES ON
EMPLOYMENT AND UNEMPLOYMENT

David Marsden *

Table of contents
Page

1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
2. Bargained pay and work rules and unemployment . . . . . . . . . . . . . . . . 2
3. Sources of flexibility in industrial relations systems . . . . . . . . . . . . . . . . 26
4. Recent changes favouring more flexibility . . . . . . . . . . . . . . . . . . . . . . . 44
5. Policies to increase adaptability. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
End notes. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 63
References. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 66

* Department of Industrial Relations and Centre for Economic

Performance. This paper, and its twin on management practices and
unemploment, has been prepared with the financial support of the OECD as
a part of its project on unemployment. Thanks are due to Scott Kelly, LSE, for
excellent research assistance. © OECD 1994 The impact of industrial relations
practices on employment and unemployment. Reproduced by permission of
the OECD. The views expressed in this paper are those of the author and do
not necessarily represent those of the OECD.
 THE IMPACT OF INDUSTRIAL RELATIONS PRACTICES ON
EMPLOYMENT AND UNEMPLOYMENT

David Marsden

1. Introduction

Concern about the effects of industrial relations practices on employment and

unemployment has been voiced in many OECD countries. In Europe, it has
been associated with the idea of 'Eurosclerosis': that employers' freedom to
hire and use labour has become excessively restricted by collective agreements
and labour laws. In Australia the influence of central arbitration procedures
on firms' ability to adjust to changing markets has been hotly debated, and lies
behind some of the recent reforms to promote enterprise bargaining. In Japan,
as domestic markets open up and overseas ones are threatened by the high
Yen, a number of writers have questioned how much longer Japanese firms
can afford to offer employment security to their core workers as a part of the
deal with the unions for workplace cooperation.1 In the United States, the
success in boosting employment opportunities for less skilled and lower paid
workers has been widely held up as one of the achievements of labour market
deregulation.

This chapter deals with those practices which are associated with collective
bargaining and other forms of collective employee representation, such as
works councils which, strictly speaking, usually 'co-decide' and participate
with management rather than bargain. 'Joint regulation' embraces both
collective bargaining and worker participation, and cover both the regulation
of pay and benefits, and industrial government 2 , that is, the joint management
of employment relations within the enterprise. Although negotiated and legal
rules differ in many respects, for the purposes of this chapter, they will often
be treated as analogous unless the difference has some direct impact on
employment outcomes.

Because union strength has declined during the last decade in many countries,
and there is growing evidence that many practices associated with joint
regulation might be introduced even in a non-union environment, a second
chapter is devoted to the impact of employee management practices on
unemployment. This focuses on policies management has been adopting
unilaterally rather than through joint channels.

The impact of industrial relations, and more generally, 'joint regulation'

practices on unemployment is complex and controversial. It has been common
to distinguish between 'substantive' and 'procedural' rules, and broadly
speaking their effects on employment differ. Substantive rules refer to those
specifying particular outcomes, such as the rates of wages to be paid to a
particular class of workers, or the number of weeks notice to be given before
lay-offs. These have commonly been contrasted with the outcomes that would
be expected under a régime of competitive labour markets. The latter relate
more to the process of industrial government in which management and
employee representatives jointly manage important areas of employee relations
and are influential over the process of workplace cooperation. They play an
 -2-

important role in adapting substantive rules, such as those on pay, to the local
circumstances of each firm. The impact of procedural rules is not easily
analysed within the standard competitive model of labour markets because
their main function is that of reducing transaction costs. Thus, whereas many
of the substantive rules associated with collective bargaining are deemed to
cause lower levels of employment and enterprise efficiency, the procedural
rules, and the institutions regulating them, on both the employer and
employee side, may improve cooperation, and thus indirectly, help jobs.

This chapter starts by examining some of the substantive rules commonly

associated and the evidence for their effects on employment in OECD
countries. It then looks at some of the compensating flexibilities which are
introduced by procedural rules. It will also be argued that some procedural
rules are dependent on substantive rules: for example, that employment
security may be a precondition for flexibility in applying job classifications.
Equally, some substantive rules would not be viable without appropriate
procedures: for example, industry-wide pay rates cannot function without
some flexibility at the enterprise level. However, what may be necessary for
institutional reasons may not be sufficient for labour market purposes. The
chapter looks at forms of flexibility which exist within the current frameworks,
and at recent changes introduced by employers and worker representatives to
increase adaptation to changing labour and product markets. It concludes with
an analysis of some policies which could help mitigate employment effects of
current joint regulation practices, and which could help improve enterprise
competitiveness.

2. Bargained pay and work rules and unemployment

There are strong theoretical reasons for expecting that when collective
bargaining raises the cost of employment above the competitive level, it will
have an adverse effect upon the distribution of employment and
unemployment. 3 First, because the demand for labour depends upon that for
the goods and services it produces, bargaining up the pay of a particular
category of workers above the competitive level is likely to cause employers
to hire less of it. The reaction may not be immediate because it takes time to
reorganise production and to invest in alternative methods, but in the long
run, a higher price will lead to less of that category being hired.

Second, fixing either standard rates of pay irrespective of individual

performance, or compressing wage differentials raises the price of less efficient
workers within a particular job category, and reduces incentives for the more
efficient and the more skilled. The latter may depress the organisation's
performance, and thus its sales, and potentially employment.

Third, collective bargaining, especially at enterprise level, by giving

 -3-

disproportionate weight to the interests of incumbent workers, may establish

hiring and lay-off provisions which reduce job opportunities for unemployed
workers. Fourthly, collectively agreed work rules may restrict labour
utilisation and slow organisational change, thus causing the firm to be less
competitive. Finally, certain kinds of bargaining structure may be more prone
to inflation, and thereby force governments to run their economies with higher
levels of unemployment that would otherwise be necessary.

2.1 Wage structure effects on employment

Collective bargaining may affect the distribution and levels of unemployment

where it causes wage differentials to deviate from the levels that would
prevail in competitive labour markets. Notable examples among these are the
union/non-union differential; policies for particular categories of workers,
such as the low paid and young workers; and solidarity wage policies.

a) Union/non-union differentials.

Compared with individual bargaining in a competitive labour market,

collective bargaining over wages enables a group of workers to negotiate a
higher wage, the size of the increase and the amount of employment foregone
depending on the sensitivity of employers' demand for labour to changes in
its price (its elasticity). If the elasticity is low, a relatively large increase in the
price of labour will have only a small effect on jobs so that the union will face
a favourable trade-off between pay and jobs, than if labour demand is highly
elastic. Labour demand will be less elastic, giving a union a more favourable
trade-off between pay and jobs, if consumers cannot easily switch to other
products; if employers cannot easily substitute other categories of labour or
capital; if other categories of workers are in a weak bargaining position
because they cannot afford to withhold their labour; and, under some
conditions, if the pay of the bargaining group represents only a small
proportion of the employer's total pay bill. The first of these depends
primarily upon the degree of competition in product markets, discussed later.
The second, substitution, may be limited by restrictive employment rules. The
third raises the possibility of wage gains at the expense of other categories of
labour. The fourth, the 'importance of being unimportant', does not always
hold (v. Hicks, 1963), and is of less relevance to employment levels because
of the relatively small numbers of workers directly involved.

The standard analyss of the micro-level employment impact of union

bargaining divides a particular labour market into covered and non-covered
sections, and traces the effects on each. Raising the wage above the
competitive level in the covered section causes employers to reduce the
number of jobs on offer, and so displaces workers into the non-covered
section, where the increased labour supply will depress wages. The crowding
 -4-

effect in the non-covered section may depress pay levels to the point at which
some workers withdraw from the labour force, and employment falls; some
may search for longer in the hope of a better paying job, thus causing
unemployment to rise; or if the wage cannot fall to the market clearing level,
perhaps because of a minimum wage rule, then unemployment will also rise.

Estimating the size of such employment effects requires information on the

size of the relevant wage differential and of the demand elasticities. This
would give the size of the potential employment displaced from the section
covered. Full estimation of unemployment effects would need additionally to
take account of the ability of the non-covered section to absorb displaced
employment, and the possibility that additional workers would be drawn into
the industry or occupation in the hope of gaining employment in the covered
part.

Average union wage mark-ups across an economy appear to be fairly modest

in many countries, except for the United States. Recent estimates for union
members by Blanchflower and Freeman (1992) using a similar range of
controls 4 across countries show the typical union wage effect varies between
about 5% in Switzerland to about 10% in the UK, and 20% in the United States
(Table 1). 5

Table 1. Union wage effects and bargaining coverage in selected countries.

Union/ Dispersion of earnings Union % of workers

wage Union Non-union density collective
% % %
United States 22 58 81 16 18
UK 10 53 74 39 47
West Germany 8 43 64 32 90
Austria 7 43 60 46 98
Australia 8 56 65 40 80
Switzerland 4 46 85 27 53

Source: Blanchflower and Freeman (1992) Tabs.3 and 4.; density and coverage: OECD
Employment Outlook 1994, p. 173.
Note: density and coverage rates adjusted for % of workers excluded from collective
bargaining (eg. managerial employees in some countries).

Because union-non-union differentials generally change only slowly over time,

one may presume employers have fully adjusted to them, so that their
employment effects depend upon the long-run labour demand elasticities, and
the degree of coverage. Such labour demand elasticities tend to be fairly low
(-0.15 to -0.5, Hamermesh 1986, 1993), and are lower for adult males than for
women and young males. Multiplying these by the wage mark-ups would
give an approximate idea of the displacement effect on the covered section:
between 3% and 10% in the US, and between 1.5% and 5% in the UK on
 -5-

Blanchflower and Freeman's estimates, but considerably lower in the other

countries. The potential unemployment effects in both the US and the UK
would however be mitigated by the relatively smaller size of the covered
sections in these countries, and the ability of the larger non-covered sections
to absorb displaced workers.

However, such national averages are not a reliable guide to bargaining effects
because labour demand elasticities vary among occupations, and because
different kinds of institutional arrangements enable bargaining to be more or
less effective. Information on the latter depends very much upon the existence
of suitable surveys of institutional arrangements, such as the UK Workplace
Industrial Relations Surveys (WIRS). Using the 1980 and 1984 WIRS, Stewart
(1987) and (1991) found that, in 1980, the presence of a closed shop enhanced
semi-skilled workers' earnings, with a differential of about 5% if there was a
post-entry closed shop, and about 10% if a pre-entry one. 6 The effect was
slightly smaller for skilled workers, and it appeared to increase slightly
between 1980 and 1984. Later work by Metcalf and Stewart (1991), using WIRS
1984 data, found that in plants without a close shop, union recognition had
only a statistically significant impact on pay compared with non-union plants
if membership density was above 95%. Then the differential rose to 7-10%.
Similarly, the presence of a post-entry closed shop had little effect on pay
unless membership density was above 95%. In contrast, the presence of pre-
entry closed shop led to a 17-19% differential over workers in non-union
plants.

In many OECD countries, collective bargaining at the enterprise level builds

upon industry level agreements which set minimum rates of pay for each
position in the industry's job classification. British evidence suggests that
coverage only by an industry agreement adds less to the union mark-up than
coverage by both an industry and an enterprise or plant agreement. The
biggest mark-up occurs when workers are covered by an enterprise or local
agreement only (Gregory and Thompson 1981). Recent evidence for Italy
shows that the presence of local (establishment level) agreements and the
degree of local union strength raise wage levels (Lucifora, 1993). 7

Thus, the main employment effects of pay agreements are likely to be

concentrated where there are local agreements, and local union membership
is strong.

The other limitation on national averages of bargaining mark-ups is that the

size of the mark-up may depend upon product market competition. If the firm
is a monopolist, its demand for labour is likely to be less elastic than the
average for all firms, and the wage gain will be met by sharing product
market rents. For the UK, Stewart (1990) found the mark-up attributable to
collective bargaining was dependent upon whether the establishment faced a
competitive product market: the mark-up was 8-10% where there was
 -6-

significant product market power, and non-existent where there was none.

The dependence of the bargaining mark-up on institutional structures and on

product market competition raises doubts about the importance of resulting
employment effects in the future. In the UK, the two institutions most
associated with the mark-up were pre- and post-entry closed shops, both of
which have now been abolished. Secondly, increased competition in many
product markets as a result of the globalisation of markets, and the removal
of barriers within the European single market, should reduce the level of rents
firms could gain from a dominant market position.

b) Sectoral wage agreements.

Unlike the type of bargaining which has been common in the US and the UK,
which is largely focused on the enterprise level, and which has attracted most
theoretical and econometric work, the commonest form in most European
Union countries is that of industry or sectoral agreements which provide a
framework for subsequent negotiations at plant and enterprise level. These
tend to produce higher rates of coverage by agreements than in either the US
or Britain and so pose a rather different problem (Table 2). The first relates to
pay, and the second to the uniformity they may impose across an industry.

Table 2. Bargaining systems and coverage in selected countries.

Country Industry Coverage of Rate of Type of min Direct state

bargain- collective union- wage intervention
% %
Belgium Yes 45 77 Agreed nat. min. No
Denmark 83 86 Sectoral agt. No
France Yes 80 12 National legal min.
Germany Yes 100 40 Sectoral agt. No
Great Britain 55 46 Abolished 1993/4 No
Greece 25 Agreed nat. min.
Ireland 57 58 Legal sectoral min No
Italy Yes 100 63 Sectoral agt.
Luxembourg 50 National legal min.
Netherlands Yes 70 30 National legal min. No
Portugal 58 30 National legal min.
Spain 61 16 National legal min.

United States No Federal minimum No

Source: Coverage: CERC 1991, p25., unionisation, OECD Employment Outlook

1991, EIRR.

Blanchflower and Freeman's study showed a comparatively small average

 -7-

union pay differential in some of these countries, and so one might,

misleadingly, suppose that the employment effect would be small. However,
because such agreements generally set minimum rates of pay to apply across
a whole industry, the more relevant effect might be on wage dispersions and
on different types of firms, illustrated in Chart 1. For there to be some
compression of the lower tail of dispersions, the minimum would need to be
set above the rate what would prevail in the weakest firms. The dynamics of
industry bargaining suggest that this should be so. Although the employers'
organisation might seek to satisfy the ability to pay of all of its members, the
union would be unlikely to agree to the minima being dictate by the marginal
firms. If they did, workers in the most prosperous firms would probably find
their ability to negotiate held back too much, and so have no incentive to
remain within the industry union. They might also favour wage compression
for organisational reasons (see below: §3.6).

Chart 1. Illustration of the potential effect of industry minima upon the

earnings distribution in a particular industry.

The effects of such agreements on wage dispersions have not been widely
studied, but some fragmentary evidence suggests their potential effects. In
Sweden, Hibbs (1990) showed the LO union's solidarity wage policy between
1970 and 1980 was one of the chief factors responsible for reducing the
dispersion of wages among contract groups, and among blue collar workers.
However, the Swedish centralised bargaining system was not typical of
continental European models in that it fixed the size of the kitty to distributed
at lower levels, and often prescribed how it should be distributed between
different categories. It did not set minimum wages along the lines of the
 -8-

French, German or Italian models.

In the absence of analyses of the effects of bargained minima one might look
at studies of the effects of minimum wages on pay. The French minimum
wage (SMIC) is in many ways analogous to an industry minimum in terms of
its functioning, as was the British sectoral minimum wage system under the
wages councils (abolished in 1993). For France, Bazen and Martin (1991) found
that a 1% increase in the real SMIC would raise real youth earnings by 0.4%.
Since young workers are closest to the minimum, this would suggest
compression of the dispersion. In the UK, Kaufman (1988) found that the
minimum had no effect on average male wages in the sector, and could
therefore have only had a very small effect on dispersions of their earnings),
but for women, a 1% rise would increase their earnings by 0.07% to 0.15%,
combining, according to the authors, a reduction in the dispersion and a
'ripple' effect on higher paid women in the sectors. None of this evidence
provides a direct test of how far industry agreements compress the lower end
of the wage dispersion, but the results are suggestive.

The other aspect of industry agreements is that they usually consist of two
elements: how jobs should be classified, and what rates of pay should attach
to individual categories. The two are logically related as one cannot fix rates
of pay without defining the units to which they are attached. These constrain
employers as concerns both work organisation, as the classification system
imposes a degree of uniformity across firms, and rates of pay. Such uniformity
can be ill-suited both to firms using atypical technologies or organisation
patterns. IBM's withdrawal from the German metal industry bargaining
machinery on the ground that much of its activity had shifted from
manufacturing to computer services provides an illustration of these
problems. 8

c) Union bargaining and the pay of young and unskilled workers.

During the 1960s and 1970s, unions in many countries followed policies that
favoured the pay of low paid workers, including young workers and the
unskilled. The reduction of wage dispersions and skill differentials in many
countries in Table 3 affected young and unskilled workers.
 -9-

Table 3. Summary of the main changes in wage structure 1970-1988.

Country. Dispersions Occupations Industries

Male/female

1970s 1980s 1970s 1980s 1970s 1980s 1970s

1980s
France - + (84) - + (86) - + (84) -
-
FR Germany - - a ) = - (<84) + + -
-
Italy -- +/- a) -- + ( 8 4 ) - - + -
?
Japan + a) + a)
Netherlands ? ++ ? na ? ? - -
-
Sweden - + (83) - - + + ?
?
United Kingdom - ++ - ++ _ + - -
+
U n i t e d States- ++ +/- + + + =
-

Key: + indicates increase; - decrease; = not much change; +/- changes in different
directions depending on categories; double sign major changes; n.a not available. Where
the turning points in trends are known to deviate markedly from the turn of the decade,
they are indicated.
Sources: Marsden (1990) except: a) OECD Employment Outlook 1993, and Katz et al. 1993.

There were many reasons for this such as national incomes policies the unions
supported which favoured larger increases for the low paid, a desire to ensure
new entrants joined their respective unions. During the 1960s, this may also
have been driven by market forces as tight labour markets ensured employers
competed hard to recruit young workers. Nevertheless, with the break in
conditions, in some countries youth pay continued on its upward trend, and
in others, a ratchet effect meant that it did not fall back to earlier levels. For
the UK, evidence of a switch from excess demand to excess supply in youth
labour markets at the end of the 1960s was provided by Merilees and Wilson
(1979), and Wells (1983) showed that youth rates of pay did not respond for
several years after the switch. In several other European Union countries,
notably, France and Italy, collective agreements did not distinguish between
youth and adult rates, a practice which did not appear to harm youth entry
in the 1950s and 1960s, but which appears to have damaged their access to
jobs from the 1970s (Marsden and Ryan 1986 and 1991).

Raising youth and unskilled relative pay, and sustaining enhanced levels once
the market has turned is likely to be particularly risky for these categories.
Layard (1982) and Wells (1983) found that demand was more sensitive to
relative wages for young workers than for adults, especially male adults.
Layard (1982) for manufacturing blue collar workers 1948-77 obtained demand
 -10-

elasticity estimates of -1.3 and -0.3 respectively for young males and females,
with -1.6 and -0.4 for adult females and males. For the US, work by Freeman
and Anderson obtained similar sized demand elasticities for young, and adult
males, and for adult women. Hamermesh's (1986) review including other
countries supported their results, with somewhat higher estimates for the US.

By skill, Nissim (1984) estimated higher demand elasticities for unskilled than
for skilled labour for UK engineering, ranging from -1.1 for skilled, to -2.3 for
unskilled workers. 9 Hamermesh (1986) reported that whereas skilled labour
generally emerged as a complement to capital, unskilled labour was a
substitute. This, combined with a greater likely substitutability between youths
and unskilled adults, would make unskilled employment particularly
vulnerable to increases in their relative pay that were not demand-related.

There is thus prima facie evidence that, in the absence of accompanying

policies to raise skill levels, the raising of youth and unskilled relative wages
during the 1970s contributed to their employment difficulties.

d) Bargaining and 'rate-for-the-job' rules.

Another source of the reduced wage dispersions associated with collective

bargaining is the use of 'rate-for-the-job' rules, that a common rate of pay
should apply to all workers in a given occupation or category irrespective of
individual differences in productivity. Such rules might be deemed to raise the
relative cost of employing less efficient workers, and to reduce incentives for
the more efficient within the same job categories.

One important reason for the prevalence of such rules is that they are easy to
monitor and to enforce. If a collective agreement were to allow management
to vary the rate of pay in line with its valuation of a worker's productivity, the
union, and its members, would have no easy way of gauging whether the
employer was respecting the agreement or not. Since employers are better
informed about current and likely future sales, there is an information
asymmetry which gives them a strong bargaining advantage over individual
workers.

Another important reason for such rules is that in large firms, the
administrative cost of establishing individually variable pay is high. Therefore,
unless the variability of work demands within a particular job category is
great, which it often was not under tightly supervised and tayloristic patterns
of work organisation, pay differentiation is unlikely to be economic.

Such rules may limit job openings for less efficient workers as employers are
likely to filter them out in the recruitment process. If rate for the job rules are
motivated primarily by union policy or by employers' administrative
 -11-

convenience, then it is likely that such workers will be penalised.

One area where rate for the job rules have proved particularly damaging is in
that of youth employment. By insisting on all workers being paid the same
rate for the job, young workers, whose productivity is lower than that of
adults have been penalised in several European countries. Garonna and Ryan
(1986) identified two policies: one of regulated inclusion, under which a
special status is created for young workers so that they may be hired at a
special trainee or youth rate of pay; and one exclusion, under which no such
discounts apply. The conditions for regulated inclusion are discussed in a later
section (§3.5), but exclusion generally arises because adult workers fear
substitution by cheap young workers unless there is some way of regulating
their use by employers. They attribute the small size of youth differentials in
Italy and France, as compared with Britain and Germany, to this effect, and
hence the lower levels of youth employment there compared with countries
where regulated inclusion is practised.

2.2 Reduction of incentives

Some policies pursued through collective bargaining may damage employment

by reducing incentives. One instance of this is by means of rate for the job
rules. Another is by policies to reduce skill or other pay differentials. Both
factors have generated a good deal of practical action. One reason for the
popularity of 'merit pay', or performance related pay policies is the belief that
they enable management to motivate employees to achieve higher standards
of performance (Marsden and Richardson, 1994). The compression of wage
differentials, on the other hand, was one factor in the revolt of skilled workers
and supervisors against their own unions in Italy in 1980, and a factor in the
breakdown of central bargaining in Sweden.

The theory of 'efficiency wages' has gone furthest in attempting to explain the
reason employers might seek pay differentiation within job categories, and in
particular why they might pay individuals above the competitive market rate
for their particular skills. This becomes particularly important where jobs
demand a high degree of employee discretion so that the quality of work
cannot be easily monitored. As a result, employers depend upon how
conscientiously workers do their jobs. It is argued that they can ensure higher
quality work either by increasing the cost of dismissal to the workers, or by
inducing better motivation. According to the first, employers pay above the
competitive market clearing rate in order to strengthen the sanction of
dismissal. According to the second, by paying more, they induce a feeling of
reciprocity among the workers so they feel obliged to work better (see Akerlof
and Yellen, 1986).

There have been a number of attempts to apply the theory to wage data (eg.
Dell'Aringa and Lucifora 1990, Lang and Leonard, 1987) but they provide at
 -12-

best indirect tests of the potential impact of rate for the job rules, so it is
difficult to gain any idea of their possible magnitude.

A major difficulty with the theory, and thus with the feasibility of alternatives
to rate for the job rules, except for higher management, is that it is very
difficult to measure employee output to the satisfaction of both management
and employees. Bishop (1987) cites evidence that the correlation between the
ratings of the same employees by different supervisors is quite low. Hence,
employee confidence in the fairness of evaluations is likely to be low, and the
potential for causing demotivating jealousies among staff is high. Thus
employers may be deterred by the risk that the extra productivity gained by
motivating high performers may be less than that lost because of conflicts over
evaluations. Therefore, for most employees, the arguments about the efficiency
loss due to rate for the job rules are finely balanced.

Pay differentials for skill and responsibility are widely held to play an
essential part in the functioning of an effective labour market. It is worth
distinguishing job movements between firms from those within the same
internal labour market. Among the former, it seems that once workers have
entered an occupation, it seems that wages play only a small role in
reallocating labour among firms and between industries (OECD 1965, Mackay
et al. 1971). Although workers with occupational skills are somewhat more
likely to change firms than those with enterprise specific ones, many
employers seek to retain their more skilled labour whatever the type of market
to which it is integrated. In contrast, the position is different for new entrants.
In contrast, studies of new entrants suggest that supply is sensitive to relative
starting wages, especially among young males (eg. Zabalza et al. (1979) UK
teachers, and Freeman (1971) US college trained engineers). Thus, union
policies compressing differentials for skill and responsibility are most likely
to affect mobility among the latter category.

As concerns the motivation, and readiness to accept promotion of incumbent

workers, the effects of reduced differentials are harder to gauge. If their
external mobility is reduced by firm-specific skills and other costs of job
changing, internal pay relativities may be more important. Anecdotal evidence
suggests some sensitivity. In both Britain and Italy, in the late 1970s, when
differentials for skilled and supervisory staff had been compressed, many
employers complained that they could not pay their skilled and managerial
staff adequately. The revolt of the skilled and supervisory staff at Fiat against
the policies of the main Italian unions which had promoted greater equality
in October 1980 was generally seen as a manifestation of the same problem.
However, in both countries this evidence remains ambiguous between
inadequate levels of relative pay and feelings of unfairness caused by their
recent reduction.

How important such questions are may depend in part on the type of
 -13-

management structure adopted. It will be argued below (§3.6) that certain

kinds of structure, notably those placing a greater emphasis on cooperation
and team work, may require smaller occupational differentials than more
hierarchical models.

2.3 Hiring and firing restrictions

In a recent study, Büchtemann (1993) distinguishes between job and

employment security at the enterprise level, the former relating to workers'
probability of retaining employment in their current job, and the latter, to
retaining a job with their current employer. Permanency of employment in a
particular job is clearly much more restrictive than in one's current firm since
the latter allows the possibility of redeployment to other work.

In practice, most collective agreements which protect workers in the event of

lay-offs relate to employment rather than job security. Even then, employment
security is usually treated as an objective to sought rather than a guarantee.
Employers are expected to use what measures they can, including retraining
and redeployment, before, finally, using permanent lay-offs (Büchtemann 1993,
Gennard 1979, and 1986, Yemin 1982).

Precise estimates of the costs of laying workers off are difficult to ascertain
because regulations vary in their nature from country to country, and because
they may be enforced with varying degrees of severity. Indeed, where lay-offs
depend upon agreeing appropriate measures with worker representatives, as
is the case in Germany and France, the cost will vary according to the degree
of cooperativeness forthcoming from workforce representatives.

A second consideration is that in many countries, employment security

provisions have not been imposed by legislators upon reluctant employers, but
have been agreed initially between employer and employee representatives.
Such was the case in France where the 1969 iron and steel agreement, and
national inter-industry agreement paved the way for subsequent legislation
(Reynaud 1969). In Germany, the basis of much of the employment security
provision applicable during enterprise rationalisation (consultation, retraining,
redeployment, and lay-offs) was established in a wave of industry agreements
on rationalisation between 1964 and 1968 (Delamotte 1971). These preceded
the enhanced powers given to works councils in this domain by the 1972
Works Constitution Act. Hence, the law cannot be considered to be
independent of union and employer preferences in this domain.

Employer agreement to employment security provisions does not preclude

possible adverse effects upon employment. First, there may be significant side-
effects experienced by other groups (externalities). Secondly, many such
agreements were made were made in the 1960s when employment problems
 -14-

were different. Thirdly, whatever the possible compensating advantages

employers may gain, many perceive the provisions as a disincentive to further
hiring.

Although agreed provisions on lay-offs may reflect the trade-offs between the
objectives of employers and employee representatives, not all parties are
equally represented. Typically, large firms and their employees are more
influential respectively in employer and worker organisations. Large firms
may find it easier to anticipate potential redundancies and to manage
providing employees with advance notice, and because they are engaged in
a wider range of activities, may find it easier to redeploy staff to other
activities, than would small firms. Among employees, more senior, skilled,
male workers are frequently better represented in works councils and unions
than are young unskilled workers, women, and ethnic minorities (Fürstenberg
1969).

In fact, allowance is frequently made for such interests in lay-off provisions.

In several countries firm size thresholds exempt the smallest firms, and length
of service thresholds often exempt employees with less than two years service.
As a result, a great many of the employees laid off receive no compensatory
payments, and of those that do, only a very small proportion receive large
payouts. Such factors, according to Büchtemann (1993), help to explain the
generally small employment effect detected for employment security measures
(the one major exception being the study by Lazear, 1990).

However, Flanagan (1993) argues that, in European Union countries, the

persistently high levels of unemployment, as compared with Japan and the
US, appear to result from greater reluctance among European firms to hire
new labour. Hiring rates declined sharply in Europe in 1970s and 1980s
(OECD, 1986), and European firms began to adjust to external shocks by
varying hours of work more than employment. This, he argues, is tentative
evidence that 'insider' interests have come to weigh more heavily than those
of 'outsiders' in economic adjustment. The evidence, however, is tentative.

The reluctance to hire is a particularly serious problem, if Flanagan's argument

is correct, because it explains the greater concentration of unemployment on
a smaller number of individuals, but who remain unemployed for a much
longer time than in the US. In addition to the human cost, such workers see
their skills and future ability to work decay, and so there a decline in the
effective labour force available.

Employment protection measures may also affect the distribution of

employment conditions. If the exemptions from employment protection
provisions make it cheaper for certain types of firm to hire certain kinds of
labour, then the effect may not be to cause lower employment, but rather a
form of segmentation in employment. The emergence of this kind of dualism
 -15-

has been suggested by Atkinson and Meager (1986). However, the radicalism
of their predictions, and the extent to which firms were consciously using this
model in the UK has been contested strongly by Pollert (1988) and Hakim
(1990) using labour force survey and WIRS follow-up data. 10 Nevertheless, the
three sets of findings would be consistent with some employers responding
in a piecemeal fashion to the increased cost of laying off certain categories of
workers. In addition, because employment security provisions in the UK are
felt to be among the least stringent in Europe (see below), such tendencies
may be less developed there than elsewhere. In Spain and France employers
have made widespread use of measures to facilitate hiring on short-term
contracts. 11 Often these are followed by hiring on a standard employment
contract so they have served as part of the induction and screening process.
Since one of the greatest costs in recruitment is that of selecting the wrong
person, the advantages of induction contracts to employers should not be
underestimated. Special employment contracts for young workers have been
widely used indicating their popularity among employers. Especially in
countries where standard rates of pay apply to workers of all ages, such
measures enable young workers to be hired at more economic rates without
threatening the integrity of collective agreements and employers' salary
systems.

In European countries, the big wave of employment security agreements and

laws occurred between the mid-1960s and the mid-1970s at a time when
unemployment was rising, but low by current standards, and employers and
unions were keen to deal with problems of technical change and
rationalisation of production. The overriding problem then was to persuade
workers to accept changes in production organisation which would threaten
the future need for their skills (and semi-skills). Thus, employers were paying
for greater flexibility and acceptance of change. Once set up, such schemes are
difficult to change. So it is possible that even though employers had agreed
these procedures, they have ceased to be appropriate.

Many European employers, according to a 1989 opinion survey, regard the

difficulty and cost of laying people off as a factor reducing their willingness
to hire new labour (European Commission 1993: 180) 12. There was
considerable variation among countries, there being, in 1989, relatively small
percentages of employers claiming the costs were restrictive in the UK, and
above average percentages in France, Spain, the Netherlands and Italy. More
detailed figures available for 1985 showed that one third of industrial
employers believed that inadequate flexibility of procedures for hiring and
laying off was a 'very important' factor in their not hiring additional labour.
Of employers in the larger countries, more than one third felt it was
'unimportant' in only Germany, the Netherlands and the UK (Nerb 1986: 74).

Although these are opinions, and the firms were not asked to specify for
which categories of labour these remarks applied, they are given some
 -16-

additional force by the inverse correlation between the EC's index of

employment protection and the rate of dismissals, and the positive correlation
with use of early retirement. The former implying that the measures
discouraged lay-offs, and the latter that they encouraged substitution of
alternative job-reduction measures 13 (European Commission 1993: Ch. 7).

2.4 Labour utilisation and competitiveness

Collective agreements and unwritten workplace custom can exert a

considerable influence on patterns of labour utilisation, and thus can favour
or hinder productivity growth. In the short-run, such practices may increase
employment, as indeed has sometimes been their purpose, but this is only
sustainable, in a competitive economy, in the medium term if those concerned
accept lower pay. Even this result may not be sustainable in the longer run as
new generations of skilled workers find the better pay of other firms more
attractive, so that the average quality of the workforce involved gradually
declines.

In an economy in which adjustment costs are negligible, such a result may

have few repercussions beyond those workers and those firms bound by such
practices. However, this is rarely the case, and it can take a considerable time
for a national economy to recover from the loss of a major industry which has
failed to remain competitive. The catastrophic loss of market share by UK-
based automobile firms during the 1970s had far-reaching effects on
component suppliers, research and development, not to mention those whose
firms supplied goods and services to car workers.

Thus, one has to consider agreed practices which inhibit the efficient use of
labour and organisational change because of their likely indirect effects upon
employment and unemployment in the longer term. Labour utilisation in the
firm can be considered under two main headings: (i) constraints arising from
rules or commitments to using labour in a particular way which lead to
under-utilisation of current workforce skills; and (ii) limitations on firms'
ability to modernise their human resource management practices.

In a famous study of productivity and unions in the late 1940s, Zweig (1951)
warned that it was relatively easy for economists to define 'restrictive labour
practices' in terms of lost productivity or increased production cost. However,
at the end of his detailed field study of such practices in five major sectors of
British industry, he concluded that there was little unanimity among managers
in different firms as to whether they were harmful or simply, often, the price
to be paid for doing other things effectively. For example, the craft skill
demarcations which confine the execution of certain tasks to those who have
completed the appropriate apprenticeship was often seen as a natural
consequence of the way vocational training was organised.
 -17-

There have been many famous cases of monopolistic labour practices which
have restricted production, and these have attracted much publicity in certain
industries, such as the media. But to rely on such cases runs the risk of
reasoning from anecdotes. An alternative approach is to look at the sorts of
issues that have arisen in flexibility discussions and negotiations between
unions and employers in a number of countries. The efforts employers have
been prepared to put into such deals is an indication of the gains they expect.

Recent examples much discussed in the flexibility negotiations of the 1980s:

- US: reforming seniority rules and related seniority districts
- UK: reducing craft and other occupational job demarcations
- Many countries: revising out-of-date job grading systems
- France and Germany: easing restrictions on working time arrangements
- Italy: increased management control over recruitment, selection,
training, appraisal and upgrading;

Seniority rules

Seniority rules, by the late 1970s, were identified by many US firms as an

obstacle to efficient production. For many years before, use of seniority rules
in lay-offs and in subsequent re-hiring had served a number of useful
functions. It enabled the firm to retain more experienced employees, and to
minimise its loss of human capital investment during recession. It provided
a degree of predictability to employees which reduced the danger that those
with the most marketable skills would leave in anticipation of lay-offs. It
helped preserve morale because seniority was an easily observable criterion
so that management could be seen to be behaving fairly rather than arbitrarily.
Finally, as long as production systems were fairly stable, and work
organisation based on narrow, routine, tasks applying seniority rules even for
some job progression did little harm to productivity. However, such
production systems quite suddenly became obsolete in the late 1970s because
of changes in product markets and in production technology. In the
automobile industry, for example, Japanese producers managed to provide a
wider range of models, with a greater variety of accessories, and to change
models more quickly and cheaply, so that it became necessary for the US
producers to be able to redeploy labour more quickly between jobs, and to
broaden the range of tasks undertaken by skilled workers.

Seniority rules became a major obstacle to redeployment for two reasons. First,
new production methods broke the continuity of job progression from one job
to another. Training as well as experience became necessary. Secondly,
seniority rules imply the prior existence of a group or category of workers
who can be ranked by seniority, what was commonly called a 'seniority
district'. Movement of workers from one district to another would normally
disrupt the working of seniority. If it were transferable, then the seniority
 -18-

ranking in the host district is disrupted, if were not, then workers have no
incentive to accept redeployment. Katz (1985) follows the introduction of more
flexible arrangements in the US automobile industry and the removal of many
seniority rules, as a part of the recovery programme of the major firms in that
sector. Kochan, Katz and McKersie (1986) provide similar evidence for a wider
range of sectors stressing the transformation of US industrial relations in the
process, with a shift in the focus of union action from job control and seniority
to acceptance of joint labour-management and 'quality of working life'
programmes. Kochan, Katz and Gobeille (1983) provide some evidence that
union involvement in QWL programmes across a sample of GM plants had
a favourable impact on productivity, at the very least, because it appears to
have enabled management to carry out some of the organisational changes it
was seeking. McKersie (1990) has argued that many firms have moved away
from seniority towards new concepts of employment security based on more
flexible deployment of labour, and broader job grading systems (discussed
below: §4.4).

The other feature of seniority rules, brought out by the studies of Gouldner
(1954) and Crozier (1964), is that they reduce management's control over job
allocations, and so reduce the rewards at the disposal of line management in
order to gain workforce cooperation.

There is no systematic evidence on the distribution of seniority rules across

OECD countries. Evidence from France, Germany, and the UK suggest that
their formalisation into collective agreements in the US was without parallel
in other countries. The studies of Crozier (1964), of Sainsaulieu (1965), and of
Eyraud (1983) in France suggest that any such rules were informal, and were
not recognised by higher levels of management (which they were in the US,
hence their inclusion in collective agreements). Evidence from the German
steel industry again suggests that seniority rules in job progression were at
best informal and confined to small groups, and were not transferable, and
probably not recognised by higher management (Bosch and Lichte 1982). In
the UK, seniority has been one criterion in lay-offs, but its use in job
progression less common (if only because other (craft) principles have been
used in job regulation, Eyraud et al. 1990).

Craft and job demarcation rules

Craft and job demarcation rules which limit management's freedom in

allocating work have gained most prominence in the UK, and figured
prominently among the labour practices targeted by management in the
flexibility negotiations of the 1980s (Atkinson and Meager 1984, Marsden and
Thompson, 1990, MacInnes 1987). In their survey of reported flexibility
agreements, Marsden and Thompson (1990) found that about half included
clauses relating to improving the deployment of labour, with about one third
specifically covering skill demarcations.
 -19-

As with the seniority rules of the US, skill demarcation rules do not
necessarily cause loss of productivity. Apart from the risk to capital equipment
and to safety when work is undertaken by inappropriately trained workers,
under the persistent conditions of skill shortage that have characterised UK
labour markets, strict demarcations could ensure better skill utilisation (Prais,
1981). Nevertheless, during the 1980s, the removal of a number of particular
kinds of job demarcation attracted employers. First, the economic crisis faced
by a number of major firms gave management the opportunity to gain
acceptance for changes which had been delayed during the 1970s. Secondly,
as in the United States, manufacturing management faced the same need to
adopt more flexible production systems and to raise quality, both of which
transformed previously acceptable working practices into obstacles to
efficiency. New technology blurred the boundaries between established
electrical, mechanical, and hydraulic skills such that rigid skill demarcations
led to poor utilisation. Similarly, as competitors made more use of preventive
maintenance, teams of skilled maintenance workers able to work flexibly
proved more economical than previous practices. Recent work by Richardson
and Rubin (1993) on the work practice concessions made in the working time
negotiations in the British engineering industry suggests that this kind of
change has continued, enabling further productivity improvement in the firms
concerned.

Evidence on the existence of such rules outside the UK is fragmentary. Jürgens

et al (1993) reporting an international comparative study of the automobile
industry could find no evidence of such practices in Germany, and the US.
Nor is there evidence of such practices in Japan (Dore 1973, Aoki 1988). In
France, Eyraud (1983) reported no evidence of British style demarcations, but
d'Iribarne (1988) has argued that French firms may be characterised by a rule
of non-interference in each other's work between levels in the firm's hierarchy,
and between different functions. For management to interfere in the work of
another group outside times of crisis, he argues, would offend their sense
'honour' by casting doubt on their competence.

Job grading systems

Job grading systems have been another major area of change of working
practices pressed by employers during the 1980s in order to facilitate easier
deployment of labour. This was an important part of the workplace level
changes in both the United States (eg. Katz 1985, Osterman 1988), and also in
the UK (Marsden and Thompson 1990). It has also figured widely in
workplace changes in France (Eyraud et al 1989), and in Germany (Schudlich
1991).

Although apparently very technical, job classifications lie at the heart of the
employment relation, and in many countries (with Japan as a notable
exception), they provide the vital link in the 'wage-work bargain' by enabling
 -20-

each category of jobs within the enterprise to be priced. In most industrial

relations systems, there are complex agreements dealing with how workers
should be paid when they are temporarily allocated to other work. Should
they retain their initial rate of pay even if the new work is less demanding or
less skilled? If so, for how long? And so on. Clearly, the finer the detail of the
job classification, the more likely it is that even a small change in the work
done will require some variation in the rate of pay. In decentralised shop floor
bargaining systems, such as that in Britain, it was not uncommon for even
minor redeployment, perhaps to another department, to require first
negotiation between first line management and either the worker concerned
or his or her shop steward. The broader job categories introduced in the US
and in the UK have generally sought to give management greater freedom to
redeploy labour as changing circumstances demanded.

Similar changes have also been sought by French employers, moving away
from job classification systems based on work posts, which made transfer from
one post to another, and adaptation to new skills often very difficult. They
have sought to move to more flexible systems based on job evaluation (critères
classants), although often the influence of the old system has remained strong
(Lanfranchi 1988, Eyraud et al. 1990).

Working time arrangements

Working time arrangements have also been targeted by employers in order to

improve the utilisation of labour and other resources. The main concern of
employers has been to obtain working time régimes which are better adapted
to short-term fluctuations in demand, and to make it easier to deal with
unforeseen problems; to deal with peak loads during the year; and to enable
fuller utilisation of expensive capital equipment. This suggests three main
aspects of working time: variability in the length of the working day;
variability in the pattern of weekly work and of holidays; and changes in the
pattern of shift working. At the same time that employers in many countries
have wanted to achieve greater flexibility in working time, unions have
wanted to negotiate shorter working hours thereby opening up the possibility
of a trade-off.

A fixed length for the working day can lead to underutilisation and raise costs
in two ways: customer demands do not always arrive predictably, and some
jobs may require specific blocks of time so that they can only be fitted into a
fixed working day if they are started early on (eg. train drivers' journeys).
Overtime working provides some flexibility, albeit at a cost, but in some
countries, there are legal or agreed restrictions on the amount of overtime that
can be worked in a given period (eg. in France, the law allows no more than
an average of 46 hours worked per week over a twelve week period Lyon-
Caen and Pélissier, 1988).
 -21-

Employers in both France and Germany have been concerned to get the law
and collective agreements to concentrate on fixing limits to working time over
longer periods than one week. Annual working time, for example, would give
employers the chance to increase weekly hours during peak periods, and then
to offset these by shorter hours in off-peak periods. This would also avoid the
cost of paying for overtime hours. They have also been interested in altering
shift-working patterns and to increase the use of weekend working in order
to extend the time during which capital equipment can be run, or particular
services provided.

Legal regulation of working time, for example, setting limits on weekly hours
and overtime, can be unduly rigid because of the need to impose a general
rule which cannot be easily adapted to take account of local or changing
circumstances. The general rule may be made more flexible by allowing
exemptions, for example, to allow for certain services where there a long
periods of inactivity during the day, and to adapt conditions to different
categories of workers. It is also often possible to obtain approval from a labour
inspector for exceptional treatment. However, such adaptation multiplies the
number of exceptions and increases the burden on the enforcement agency,
and on firms which may not be aware of the law's complexities. For this, and
other reasons, successive French governments have sought to increase the role
of collective agreements on the ground that they are more capable to adapting
to the needs of particular industries while at the same time providing an
acceptable level of protection to employees. The negotiations have, however,
been fraught with difficulty, not least because the unions are often too weak
to provide effective control over enterprise negotiations, and so have been
reluctant to lose the protection of the law. Nevertheless, during the 1980s, and
particularly after the Auroux laws, a large number of agreements were
reached at both industry and enterprise level, many of which dealt with
aspects of the organisation of working time (Ministère du Travail 1989). The
employers appear to have used negotiations on the reduction of working time
in order to gain concessions for more flexible patterns of utilisation (Boulin
and Taddéi, 1989).

In Germany, although the impulse for negotiation on working time has come
from the unions' campaign for reduced weekly working hours, the employers
have used the negotiations to press for their own changes such that shorter
working time for each worker comes at the price of more flexible working
time arrangements (Bosch1990).

The different emphasis in such negotiations among countries highlights the

close relationship between these practices and the type of job regulation
practised there. The contents of flexibility and concession agreements are
strong evidence of employers' views of practices that impede productivity
since they are prepared to pay for the changes. However, the surveys are few
and the data fragmentary. Other evidence of pronounced national differences
 -22-

in the obstacles to flexibility can be found in opinion surveys of employers.

One such survey from the mid-1980s (Elbaum 1987) showed that French and
Belgian firms were much more likely to see insufficient flexibility in hiring and
lay-offs as a reason for not increasing employment as compared with Britain.
In contrast, British firms were more likely to see rationalisation and new
technology as a reason. Use of temporary workers, and inflexible working time
arrangements were also of more concern to Belgian and French, than to British
firms. 14

One systematic comparison of Britain and France highlighted the greater

interest of British employers in increased flexibility of job demarcations and
labour deployment, and relatively little interest in working time arrangements.
In contrast, flexibility agreements in France placed much greater emphasis on
working time arrangements, and correspondingly less on job control issues.
This appeared to be due to the greater emphasis on the defence of
occupational based jobs supported by occupationally based union organisation
in Britain. In France, in the absence of strong principles around which to
organise workplace job control, unions focused instead on obtaining general
rules likely to attract a broad coalition of support within the enterprise and
within the branch. They therefore focused more on general wage increases for
all, and general working time rules, and built their organisation around the
defence of these. Hence, when employers came to seek productivity enhancing
changes, the sticking points tended to be those where change affected certain
practices which were fundamental to the pattern of job regulation: in Britain,
skill demarcations, and in France, working time arrangements (Eyraud et al.
1990).

2.5 Bargaining structure, inflation, and unemployment.

High and variable rates of inflation, by shortening time horizons for

investment, depress investment and growth (Leigh-Pemberton 1992), and thus
are bad for jobs. Thus, collective bargaining structures which make inflation
more likely could be deemed potentially harmful to employment. This is all
the more likely as rising inflation is likely to cause governments to adopt
restrictive economic policies.

Concern about whether some types of collective bargaining structures are

more prone to inflation than others has a long history. Hicks (1955) suggested
that the interaction between the claims of different bargaining groups might
be prone to generate inflation. For example, suppose that, because of skill
shortages, skilled workers covered by local bargaining were able to gain a pay
increase, whereas those whose pay was fixed by an industry agreement could
not. Adjusting the pay of the latter, which would be justified on grounds of
scarcity or fairness, might only be possible if the whole structure of pay rates
in the industry were jacked up. Thus, Hicks argued, seeking to maintain full
 -23-

employment with this kind of bargaining structure would be inflationary.

More recently, Nickell (1990) has argued that a combination of skill shortages
and a fragmented bargaining structure in Britain could prevent from reducing
unemployment without boosting inflation.

Adjustment to economic changes and shocks which imply adaptation of pay

levels may also be impeded by fragmented bargaining structures. Under such
conditions, no single group might be willing to allow its pay to decline in
response to the new conditions out of the fear that it would be the only one
so to do. The collective situation resembles that of the 'prisoner's dilemma' in
which the expectation or fear of a lack of reciprocity by the other parties
discourages any one party from being the first to make concessions. Because
no group is prepared to risk taking a cut in nominal pay, and given the
fragmentation, no one is prepared to risk taking the lead, all refuse, and
adjustment falls on employment rather than wages. As in the previous
example, government attempts to maintain employment then translate into
inflation.

During the inflationary periods, persuading pay bargainers to moderate their

claims has encountered similar problems. Aubrey Jones (1973), former head
of the National Board for Prices and Incomes, argued that 'wage-wage'
inflation had come to dominate as each group fought to avoid being left
behind by the others.

The degree of fragmentation of bargaining structures is critical. Extreme

fragmentation may leave bargainers with little market power so that agreed
rates correspond closely to their competitive levels. However, it becomes more
of a problem as groups increase in size and acquire more bargaining power
(Crouch, 1985). A number of writers have argued that such 'prisoner's
dilemma' problems are particularly common where bargaining groups are
sufficiently large to exert a degree of market power, and to compete with each
other for prestige or membership, but in which no one group is sufficiently
large to provide stable leadership.

Clearly, such processes can only take effect if employers have the ability to
pay which depends partly on monetary policy, and partly on other income
recipients, and income for other purposes being squeezed. Tightening
monetary policy will restrain inflation by curbing demand, and thus
employment, but it may not always be the employment of those with most
bargaining power.
 -24-

3. Sources of flexibility in industrial relations systems

The picture so far in one in which collective bargaining appears to have a

generally negative effect upon unemployment as compared with competitive
labour markets. Use of bargaining power is likely to distort competitive wage
structures and lead to some workers being excluded from jobs they would
accept. Restrictive working practices, even if originating in rules set by
management, slow productivity growth, and harm the competitiveness of the
firms concerned. And the bargaining structures common in most industrial
countries, being neither wholly decentralised nor wholly centralised, appear
to do badly in terms of higher than necessary inflation and unemployment.

Against this rather negative balance sheet, a number of more positive features
of collective bargaining must be set before any final conclusions may be
drawn for policy. Many of these features are often neglected in the discussion
of institutional influences on unemployment. They suggest both a number of
positive benefits from working through collective bargaining and other forms
of employee representation, and by examining some of the less visible forms
of flexibility contained within collective institutions and how they have been
adapted in recent years, it is possible to identify some areas of reform which
could help combat some of their unemployment effects.

3.1 Role of employers in setting up IR practices

The discussion of bargained rules often proceeds as if they were imposed

unilaterally by unions on unwilling employers, and as if rules of a similar
nature would not be adopted by employers even in the absence of such
pressures. In practice, collective bargaining usually brings together powerful
organisations on both sides, each capable of inflicting heavy costs on the other.
Therefore, one must consider that a good proportion of the outcomes reflect
a compromise between the objectives of employers and worker
representatives. Indeed, one of the historic arguments advanced to justify
workers' collective action is that it helps to redress the imbalance of power
that would otherwise exist in labour markets between firms, especially large
ones, and individual workers. Moreover, even in largely non-unionised sectors
where collective bargaining has little direct influence, one can often find the
same characteristics of inertia in wages and employment with respect to
labour market fluctuations. Some of these issues will be dealt with in greater
detail in the chapter on management practices, and so will not be considered
further here. However, because employment security regulations have
attracted so much attention in the discussion of unemployment, and especially,
the apparent reluctance of European employers to hire new labour, it is
worthwhile delving deeper into the role of employers in this issue.

As illustrated in the earlier discussion of employment security agreements, in

several countries, such as France and Germany, collective agreements between
 -25-

employers and unions at both sectoral and national level both preceded
legislation in this area, and provided a model for the type of provisions to be
included. A clue as to the underlying logic of such agreements. Far from
conceding simply to worker demands for greater employment security, as
might be the case in pay deals, the agreements were mainly concerned with
promoting organisational change and rationalising production, both of which
would have serious employment consequences. In the steel industry, which
pioneered such agreements in France, and in other mass production industries,
a great many workers had skills that were not easily transferable, and would
be rendered obsolete by the changes. The same applied to middle management
employees whose skills lay primarily in their previous work experience and
their detailed knowledge of their particular organisation and its work norms.
With the potential loss of jobs and skill obsolescence, workers' morale could
easily collapse, and with it, the ability of the firms to implement their change
programmes. Management had also to ensure that once the first wave of
changes was complete, the employees remaining would work effectively and
willingly.

Because of the likelihood of further changes, management could not just

introduce change unilaterally, and expect those workers remaining to work
hard simply out of gratitude that they had survived. It was therefore
necessary to set up a framework to discuss the implications of change, and
agree the measures needed to minimise the cost to employees. In some OECD
countries this took the form of ex ante agreements about procedures, while in
other, reliance was placed upon customary methods for working out
arrangements as the need arose (Gennard 1979).

It might be objected that such arguments are valid for voluntary but not for
legally based employment security rules. It is possible that legislation sets
more costly provisions than would agreements. Nevertheless, in many
countries, employment law has been used to set minimum conditions
sufficiently low to enable worker representatives and employers to agree
locally the arrangements which suit their needs best. The law has often been
modeled on existing agreements, particularly as concerns consultation
procedures.

3.2 Bargaining structures that avoid wage competition

It was argued earlier that certain kinds of pay bargaining structure were likely
to harm employment levels because governments would need to restrain their
inflationary tendencies by running the economy below capacity (§2.5). One
solution to the problems of competitive pay bargaining, and to the reluctance
of bargaining groups to make economically necessary concession under
fragmented bargaining has been to propose greater coordination by means of
centralised bargaining.
 -26-

Blyth (1979), intuitively, and Calmfors and Driffil (1988), more formally, have
argued that centralised bargaining systems should be less prone to inflation
than more fragmented ones. The Calmfors and Driffil argument rests on the
interaction of two opposing processes. As bargaining units increase in size
from the plant or enterprise to the central level, their bargaining power
increases. However, as the coverage of collective agreements increases, so it
becomes harder to escape the negative consequences of individual settlements,
and easier to benefit from any sacrifices they might make in the general
interest (Olson 1971). The latter force serves to moderate unions' use of
bargaining power, and enables bargainers to take a longer term and more
moderate view. Bargainers who take a moderate line are also less prone to see
their position being undermined by rival groups taking a much more militant
one.

However, for bargaining groups with an intermediate degree of centralisation,

bargaining power is only partly moderated by their limited inclusiveness.
Groups which seek to opt for non-inflationary increases can find themselves
in a classic prisoner's dilemma as they cannot control the behaviour of rival
bargaining groups. Individual unions have only limited incentives to pursue
moderation, and individual employers have little incentive to resist such
claims if facing down a strike will result in loss of market share and possibly
only temporary advantage in costs. Hence, economies in which bargaining
structures are neither wholly decentralised nor highly centralised can be very
vulnerable to wage inflation.

The impact on employment may come through several channels. Proneness to

spurts of acceleration of inflation may discourage investment. It may also force
governments to adopt restrictive macro-economic policies which use
unemployment in order to restrain bargaining pressures.

There is some evidence supporting this view. Between the 1970s and 1980s,
countries with centralised bargaining systems, and those with highly
decentralised ones tended to score better both on inflation and employment
(Paloheimo 1990). However, the indicators of bargaining structure used are
rather subjective and the results rather prone to changes in the rankings of
countries by degree of centralisation (eg. Soskice 1990). Moreover, the degree
of influence that central organisations within any given country can, or seek
to, exert over bargaining outcomes at lower levels varies over time, as does
the degree to which strong union organisations use their power aggressively
or cooperatively.

3.3 Flexibility within industry agreements

The earlier discussion of industry bargaining (§2.1) focused on its action in

compressing wage structures. However, in practice most industry bargaining
systems set minimum rates of pay, or some analogous indicator, and leave
 -27-

considerable room for additional negotiation (although not always bargaining)

at the enterprise and establishment levels. It might be objected that such
flexibility always involves additional payments by employers, however, this
would be to misunderstand the objectives of employer representatives at
industry level. There have been occasions when employers have complained
that what was agreed at industry level was reopened for additional
negotiation at the plant and enterprise level. 15 However, the target for
employer organisation negotiators is to obtain a settlement such that the
combined outcome of the industry and company level negotiations accords
with its members' business plans. Public statements to this effect are rare for
the obvious reason that they would then become the minimum target which
the union could sell to its members. Nevertheless, in the days of the heyday
Swedish centralised system, calculations of expected wage drift from plant and
enterprise bargaining were common. Thus, one can expect the likely effect of
subsequent discussions over pay to have been already taken into account in
deciding what employers can accept at the industry level. Indeed, if that were
not the case, many employers would probably resign from their organisations
(as occurred with the flight of small and medium-sized firms from the German
metal industry employers organisation in 1993 over the wage convergence
agreement).

A less misleading term to describe industry bargaining might be that of

'articulated bargaining' (Giugni 1965) in which the industry level sets a
framework for discussions at the local level in order to adapt the agreement
to the conditions of different firms and different localities. The basic principle
is that lower levels of negotiation should not contravene what has been
already agreed at a higher level. The degree of freedom left to local
negotiators can of course vary, depending on how closely the parameters for
lower level agreements are set, and how much bargaining room is left by
substantive terms agreed at the industry level.

Although there is a lot of institutional information pointing to the scope for

the terms of industry-level agreements to be adapted to the conditions of
individual enterprises, the most telling evidence is to be found in the small
number of studies of the pay mark-up, or 'wage gap'16 , determined at
enterprise and plant level. Some information on this is available for Germany,
Italy, France, and Sweden.

Germany

Teschner's (1977) study of works council wage policies in metal, chemical,

textile and tobacco industries found that the wage gap between industry
minima and average gross earnings varied between 5% and 40% among
individual plants. The size of the gap was correlated with regional
unemployment, sector, and the type of payment system in operation. The
wage gap was also higher for skilled than unskilled workers with an
 -28-

unweighted mean across plants of 23% as against 17%, suggesting that

employers have the scope at enterprise level to adjust skill differentials
modestly according to their individual needs.

More recently, Meyer's (1991 a and b) survey of 103 small and medium-sized
firms in Niedersachsen and Baden-Würtemberg found an average mark-up
over the industry minima of 14% among firms paying above the industry
minimum. In his survey, 80% of the firms applied a wage agreement. Of these,
15% paid only the rates set in the industry agreement. The others, on average,
paid 14% above the industry rate. The main variables associated with this
being: presence of a works council; number of vacancies; level of training; and
(inversely) the percentage of women. Among these small and medium sized
firms, there appeared to be little difference in mark-up by size.

These two surveys suggest that economic pressures at the enterprise, notably,
recruitment and skill levels, do indeed affect the size of the wage gap, and are
evidence of some capacity of pay levels set at the industry level to adapt to
local pressures.

Italy.

In Italy, the scope for enterprise bargaining and pay supplements unilaterally
decided by the employer have varied over time, but were greatly squeezed by
the indexation system in force from 1977. In 1976, according to
Federmeccanica data, the wage gap between gross earnings and the minima
set in national industry agreements in the metal industry was 28%, this being
the result of local bargaining and employer supplements. It remained steady
until after the reforms of 1984 and 1984, climbing to 43% in 1990 (Asap 1993,
Appendice Statistica). Data for Assolombarda give similar, but slightly lower
estimates of the wage gap for metal industry employers in the Milan area: up
from 26% in 1983 to 37% in 1991.

The scope for enterprise bargaining and employer supplements was squeezed
by the combination of flat rate indexation and unanticipated inflation in the
late 1970s and early 1980s (de Luca 1987), and was only stabilised by the
agreements of 1983 and 1984 reducing the degree of indexation. The
Assolombarda data show the main change to be the decline in the share of
pay determined by indexation, down from 44% in 1983 to 39% in 1991 (Asap
1993, Appendice Statistica III).

The same sources show a steady increase in the wage gap for the Milan metal
industry in 1991, from 15% for unskilled to 23% for the highest skilled, and
from 14% for junior clerks to 93% for the highest level of managers still
covered by collective agreements (Asap 1993).
 -29-

One can only surmise the scope for adaptation by individual firms as data are
not shown at this level. However, de Luca (1987) records quite large
differences between industrial sectors. For 1984, about 25% in textiles and
clothing, about 45% in chemicals and pharmaceuticals, and about 30% in the
metal industry. In addition, the Assolombarda data show that, in 1991,
individual premia added 18% to industry minima: ranging from 2% for blue
collar workers and 17% for white collars, to about 40% for managers covered
by collective agreements. This suggests that employers have a significant
margin with which to respond to labour market and motivational needs.

France.

In France, the gap between industry rates and enterprise pay levels is often
considerable, and has increased in recent years. Eyraud, et al. (1989: p.159 ff.)
found that average basic pay 17 in their sample of plants was considerably
above the industry agreement minima. In the metal industry firms, lower paid
manuals were paid 20-60% above above the industry minima, about 20% when
blue collar workers were in the majority, and up to 60% when white collars
were in the majority. Higher paid white collar workers were also well above
the industry minima for their grades. So, establishment level factors provide
a considerable mark-up, and also alter the structure of pay, particularly for the
larger firms and market leaders which made up the bulk of their sample.
Interestingly, the wage gap was not uniformly greater for more highly
classified workers in the plants studied, suggesting a margin for adapting
wage differentials within the enterprise. The authors concluded that the main
importance of industry agreements was as a benchmark for annual pay
increases, and a guide for wage structures.

In France, industry agreements also leave individual firms with a good deal
of freedom to develop their own policies on other remuneration issues such
as fringe benefits. A study by the CERC (1987) identified four distinct types
of policy which appeared to be pursued whatever the industry agreement
concerned.

Type 1. firms where employment is regulated by a set of stable rules;

mostly large firms able to develop their own practices in agreement
with their unions; they gain freedom from sectoral agreement
because they can do what they like provided benefits exceed the
branch minima;
Type 2: firms with organised flexible management: relative weakness of
institutional constraints;
Type 3: firms with unorganised flexibility: industry agreements are
relatively unrestrictive, and there are no enterprise agreements;
Type 4; construction industry: low level of local bargaining activity, low
pay, but relatively high fringe benefits financed at sectoral level.
 -30-

The authors concluded that the diversity of firms' practices within sectors
highlighted the relatively weak influence of sectoral agreements on firms'
practice in France.
 -31-

Sweden.

Sweden has had a different model whereby central bargaining did not specify
minimum rates of pay for industry and local levels, as in France, Germany,
and Italy, but rather set the aggregate increase in the wage bill for each
industrial contract area. This 'wage pot' or 'kitty' could be distributed as local
bargainers decided, and its size was usually determined in relation to the
growth of the total wage bill (Hibbs 1990). At various times, the central
agreement also determined set norms for its distribution, for example, that the
lower paid should get special increases. When in force, the solidarity wage
policy imposed an equalising effect by specifying how the 'wage pot' or 'kitty'
was to be distributed at lower levels. 18 Apart from that period, the model
provided a good deal of room for adaptation to local conditions, and has
indeed been introduced as a part of the government's strategy to decentralise
pay determination in the public sector (Schager,1993).

Overall, industry agreements in many EU countries provide a good deal of

flexibility at the enterprise level, and it is one that has been increasing with the
greater use of additional enterprise level agreements. The main employment
problem would seem to lie with the marginal firms and unprofitable sections
of industrial bargaining groups discussed in the first part of this paper.

Because of the need for trade union cohesion (§3.6), it seems unlikely that
these margins of flexibility in industry agreements avoids problems for
marginal and temporarily weak firms. Two problems will need to be discussed
later: whether there are ways of making industry agreements more responsive
to the needs of firms going through a temporary crisis, and whether they can
be linked with certain active labour market policies.

Japan

The Japanese case offers an interesting variant in which pay bargaining takes
place at the enterprise level between management and enterprise unions.
However, the enterprise unions are grouped into industry federations which
play a critical role in coordinating bargaining strategies among their members.
In addition, the national union and employers' confederation coordinate the
activities of their constituents at the national level. To assist coordination, and
to avoid problems of some groups holding out until after the others have
settled in the hope of gaining a better increase, with all the problems this
entails for obtaining a first settlement, bargaining is concentrated in a short
period in the 'spring offensive' (Dore 1987).

Additional scope for individual firms to adjust their own pay increases is
provided by the annual bonus which can represent the equivalent of several
months base salary (Nohara 1992). This is decided at company level once its
financial results are known, and may also be the subject of bargaining, but
 -32-

outside the spring offensive. According to Hashimoto and Raisian (1988) the
average bonus for manufacturing varies roughly in similar proportion to
production (a 10% change in production producing an 11% change in the
bonus) thus affording some variability in employers' labour costs. In 1978,
average bonus stood at roughly 20% of total benefits. Although that would
give an average margin of variation of only ±2% of benefit costs, it may
provide more scope for variation among firms.

3.4 Agreed frameworks and cooperative exchange

There are three broad arguments concerning positive effects of worker

representation on productivity which mitigate some of the efficiency losses
discussed earlier. The first concerns the 'union-voice' argument of Freeman
and Medoff (1984), the second, an analysis of the reasons for restrictive work
practices, and the third, the conditions for 'cooperative exchange' within the
enterprise.

The union-voice argument

The union-voice argument stresses the value of the information management

obtain when they learn about worker grievances and their likely causes. When
acting individually, workers may express their grievances to management
('voice'), they may quit the firm ('exit'), or they may grin and bear it. If they
quit, they will probably take with them valuable information about the causes
of their discontent, which could be due to too heavy supervision, bad working
conditions, badly organised work, and so on. They will also take with them
any investment management has made in their skills, and replacing them may
prove costly. In addition, the firm's most marketable employees are those most
likely to quit. These costs may be avoided if management can discover the
cause of discontent, but employees express their 'voice' individually, they risk
being seen as trouble-makers. Moreover, their colleagues, who stand to gain
from the improved conditions may prefer to keep their heads down, and 'let
Harry do it'. On the other hand, collective representation can protect those
expressing grievances on behalf of their colleagues. Thus Freeman and Medoff
(1984) argue that unions are necessary to provide 'voice', and that by this
means, management can learn about the causes of employee discontent and
avoid the resulting costs.

Management has some alternative channels for employee voice at its disposal,
for example, surveys of employee opinions, and encouraging first line
management to be more active in discovering the causes of employee
discontent. Although these can provide valuable information, they are not
independent of management. Hence, employees may still fear victimisation,
particularly if first line management is also responsible for performance
appraisal and merit awards.
 -33-

The evidence that union representation raises productivity has been

extensively reviewed. Data problems are severe. The great majority of studies
in the United States indicate a positive effect on productivity, but a negative
effect on other variables of company performance, such as profits (Belman
1992, Addison and Hirsch 1989). One point to emerge from some studies has
been that whether or not union-employer relations are cooperative or
conflictual matters, the negative effects being associated with the latter
(Belman 1992). Those in Britain paint a more mixed picture.

Restrictive work practices and 'low trust'

Restrictive work practices are often motivated by defensive rather than

monopolistic considerations. Sometimes they are embodied in formal
agreements, but more often they are informal customs developing within
individual workplaces, and sometimes extending across many firms. Brown's
(1973) study of workplace custom in British engineering firms showed that
often the sources of customary rules were in fact management decisions and
tolerances. In most work environments, informal workplace custom plays an
important part in regulating the relations between management and workers
because it is almost impossible to make explicit all the tasks workers have to
undertake in exchange for their pay. Coase (1937) argued this was why the
employment contract usually contained the presumption that the employer
would have the 'right to manage' and to direct what work was to be done
subsequently. However, management's authority to detail work exists only
within certain limits of what management and workers consider a fair and
acceptable exchange. But because the detailed tasks to be undertaken are not
specified, the exact limits of management's authority cannot be made precise.
In addition, firms have to cope with changing markets and changing demands,
some of which cannot easily be forecasted, so management has to be able to
vary the demands made of its workers if the firm is to survive.

In such an environment, custom plays a very important part in regulating the

exchange between management and workers, in determining what are the
legitimate demands management can make, and beyond which limits workers
may legitimately refuse additional requests. In some environments, a great
deal of workplace custom is regulated by small group bargaining, as was the
case in the British engineering firms studied by Brown. In the United States,
often such customary rules had to be incorporated into legally binding
collective agreements, although the day-to-day interpretation would still rely
heavily on informal agreement. In the French firms studied by Bernoux (1972),
Crozier (1964) and Sainsaulieu (1988) such custom was more fragile, relying
on implicit understandings between supervisors and work groups, but were
not recognised by senior management. However, supervisors could not work
without them, and so had to connive with their work groups.

If such rules are part of the normal give and take between management and
 -34-

workers in adapting to the changing needs of firms, why should they

sometimes become obstacles to change, and why should they be any more
restrictive than the terms of any contractual agreement?

One important reason is that workers (or management) come to fear that the
other party is seeking covertly to alter the terms of the agreement in their
favour. If workers were confident that management would faithfully report the
changing value of their output and pay them accordingly, there might be less
need to insist on easily monitored 'rate for the job' rules. Equally, if they were
confident that the employer would seek to minimise redundancies, and not
use lay-offs to get rid of those who have presented grievances, they might
insist less on strong seniority rules in lay-offs and in bidding for internal
vacancies. Similarly, restrictive job demarcation rules which go beyond the
needs of technical competence provide skilled workers with a buffer against
employment insecurity.

This is not to deny that monopolistic motives may also be present, but these
are often limited by the need for the support of other groups of workers.
Brown (1973) stressed that shop stewards' ability to defend work customs that
were advantageous to one group depended upon their ability to convince
other work groups that they were acceptable. His study cited cases in which
shop stewards supported management against the claims of some groups,
albeit based on custom, that threatened the good faith of their relationship
with management and their ability to sustain the support of other groups.
D'Iribarne (1988) similarly stresses that the stability of customary relations of
non-interference in each others' work ('honour') he found in French firms
depended also upon willingness to moderate claims that threatened other
groups. 19 There is less likely to be sympathy and support from other groups
for one which defends a customary rule for monopolistic goals than for one
which does so to defend its livelihood.

One reason why such rules appear to be less restrictive in German and
Japanese firms is that employment relations function at a higher level of trust
between workers and their employers, at least for their core workforces. With
a high degree of employment security, provided in the German case by widely
recognised skills and fairly long job tenures, and in the Japanese one by the
implicit understanding of company-based employment and basic income
security, many of the defensive measures discussed earlier become
unnecessary. Added to that, the methods of workforce representation, based
on works councils in Germany and strong enterprise unions in most large
Japanese firms, remove the need for workers to organise their own job level
defensive practices. In both countries also 'rate for the job' rules play a smaller
role, job rotation playing an important part in training, and in Japan, the link
between seniority and pay and the bonus system outweigh the influence of
current job on pay (Nohara 1992).
 -35-

'Cooperative exchange'

'Cooperative exchange' as a source of greater productivity has come to the fore

as a result of research on work organisation Japanese and German firms.
'Organisational slack' or 'X-inefficiency' is widespread in all organisations, and
much of the information needed to reduce this lies with employees at or near
the job level. Under the low trust conditions which prevail in many firms, and
which are encouraged by tight managerial control, and an insistence on
contractual obligations, workers have little incentive to share this information
with management. If anything, they have every incentive to use it to make
their own jobs easier, and to withhold it from management in case it is used
as a reason for reducing their pay or increasing their workloads Mottez (1966).

In organisations with stable markets, management may be able to reduce its

dependence on such information by dividing jobs up into small units along
Taylorist lines, and using close control to ensure work of the right quantity
and quality is done. But even this form of organisation is vulnerable. Crozier's
(1964) study showed how even in tightly organised bureaucratic organisations,
management is unable to predict and plan for all eventualities, and that
unanticipated problems easily create bottle-necks in the organisation which
then provide a basis for small group bargaining power. Drawing on their
research on Japanese and east Asian firms, Koike and Inoki (1990) argue that
the degree of uncertainty attached to all markets means that job knowledge
inevitably eludes management control in reality, thus making them heavily
dependent upon workers' job knowledge for efficient organisation.
Management might respond to this uncertainty by increasing buffer stocks, but
this also would tend to reduce the organisation's efficiency. In a low trust
environment, management is encouraged to reduce further workers' scope for
discretion, and this generates a vicious circle (Fox 1974).

Although sharing such information could enhance the firm's performance, in

a low trust environment, such cooperation is discouraged by the fear, or
expectation, that it will not be reciprocated. Sharing information reduces one's
own bargaining power and puts one in a position of weakness vis-à-vis the
other parties. Workers could lose some of their control over work rhythms, or
even their jobs. Management could find itself paying for productivity which
never materialises. A prisoner's dilemma problem emerges, and the gains from
mutual cooperation remain elusive.

Most of the strategies for cooperation which deal with the parties concerned
in isolation prove to be unstable, including those based on 'reputation' and the
use of 'tit-for-tat' strategies. Cooperative behaviour early on could, for
example, be designed to trick the other party. In 'tit-for-tat' strategies, in which
one party keeps withdrawal of cooperation in reserve in order to punish the
other party should it not cooperate, the final round of the game is always
problematic. As a result, cooperation in every round is problematic. 20
 -36-

The main lesson of the 'prisoner's dilemma' model is that it is each party's
expectations as to the other's behaviour which are critical. Even if one wishes
to cooperate, but one is worse off if cooperation is not reciprocated and one
suspects that the other will not, then it is tempting to refuse cooperation.
Given the dependence of cooperation on expectations, then each party
fulfilling its commitments becomes very important. Not to do so could be a
signal that one's intentions have changed, and that in the future cooperation
will not be forthcoming. In practice, worker representatives and employers
know that economic life is full of uncertainty, and that not all commitments,
however well intended, can be met. The problem lies in distinguishing
genuine cases of force majeure in which good faith has been kept from those
in which faith has been broken. The latter would signal changed behaviour
and therefore alter the expectations of the other party.

If cooperative exchange is to take place, there is a need for guarantees and

sanctions that apply from outside. One solution is for employers and workers
to form coalitions with others of their own kind. These have the power to
inflict more severe sanctions on individuals for failure to cooperate, but
equally, to conserve their power, they need to be able to police the activities
of their own members. Here representative organisations of employers and
workers have a critical role to play in evaluating and encouraging action in
good faith. In addition to sanctions, which may not always be very effective,
they can monitor critical information affecting the ability of local parties to
fulfil each others' expectations.

Such would seem to be one of the important functions of the large amounts
of information exchanged between German and Japanese employers and their
respective worker representatives both at the industry and at enterprise levels.
It would also be one of the reasons why German and Japanese unions have
not had to rely upon the forms of 'job control' unionism found commonly in
the US and the UK, and why field research has found so little evidence of the
workshop level restrictive labour practices discussed earlier in this paper.

These three related arguments, of union-voice, eliminating defensive practices,

and cooperative exchange suggest important compensating gains for the
presence of forms of collective worker representation. However, their impact
is likely to be greatest in those firms where cooperation leads to greater
productivity. They would appear critical where work involves the application
of high levels of skill and a reasonable level workers' judgement. They appear
less critical in highly routine types of work.
 -37-

The problem of 'low trust cooperation'

One important counter-argument to the above is that management and

workforce may decide to 'live and let live', giving the appearance of
cooperation because management refrains from anything which would disturb
its relationship with the union. Thus, management might tolerate many
restrictive labour practices because the cost, in terms of likely conflict, would
be too high. Something of this kind appears to have occurred with the failure
of the 'Donovan reforms' in Britain and with the increased powers of factory
councils in Italy during the 1970s. In both cases, increased powers for workers
and their representatives led to a collapse in productivity as the new
institutions failed to generate active cooperation. 21 Instead, they led to a rise
of worker expectations which could not be met, and were the foundation for
a period of combative rather than cooperative industrial relations. In Sweden
too, the additional representational rights obtained by workers during the
1970s were associated with increased militancy rather than increased
cooperation (Kjellberg 1992).

These observations show that the frameworks for cooperative exchange and
collective voice are necessary but not sufficient conditions for cooperative
relations. The contrast with the German case is revealing. There, pressures for
greater powers for shop stewards, and for direct union-enterprise bargaining
were strongly resisted by employers who maintained and encouraged the
existing system of works councils. This suggests that employers have a key
role in the development of cooperative relations. Although they may be
supported by an institutional framework which protects against non-
cooperative action, the parties have to build their cooperative practice
themselves. Employers and worker representatives demonstrate to each other
their commitment to cooperation by the amount they are prepared to invest
in it, and this encourages further cooperation. In the German case, during the
1970s, management refused to negotiate with shop stewards, which would
have undermined their relationship with the works councils, and the unions
and works councils generally kept stewards in a subordinate role.

In contrast, in Britain, Italy and Sweden, many of the additional powers came
from legislative intervention or were gained against employer opposition. It
was therefore difficult for employers to present their action within the new
institutions as the first step in a process of cooperation and so begin to win the
trust of workers and their representatives. As a result, many of the new
powers were used as a platform for additional gains at the expense of
employers rather than as the beginning of a new era of cooperation, as many
of the architects of the reforms had hoped.

3.5 Cooperative frameworks for youth inclusion.

There is widespread evidence that youth access to jobs is reduced by high

 -38-

levels of youth pay relative to adults. Yet, in some countries, unions and
employers are able to agree low youth rates of relative pay whereas in others
they are not. Contrasting six European Union countries, Garonna and Ryan
(1990) identify two distinct modes of youth employment: exclusion, and
regulated inclusion. Given the persistently high levels of young people's
unemployment in Europe, the reasons for this merit analysis.

Marsden and Ryan (1986 and 1990) argue that the problem lies in the
acceptability or otherwise of special low rates of youth, and particularly
trainee, pay relative to adult skilled workers. Young and adult workers always
have potentially opposed interests because the latter, if engaged on low rates
of pay, become potential substitutes. Ideally, their lower productivity and need
for training and experience should be compensated by correspondingly lower
rates of pay. However, there is always a potential problem that employers will
take advantage of the lower pay, but fail to deliver on the training so that
young workers become a source of cheap, substitute, labour. 22 Apprenticeship
regulations provide one means by which the status of young trainees can be
clearly identified, and the monitoring of training received becomes easier for
adult workers and their representatives. Under such circumstances, skilled
adult workers are more likely to agree to special apprentice pay rates, and
other kinds of wage for age system.

In the absence of such provisions, and where induction training is largely

informal and on-the-job, the provision of training and useful work experience
is harder to monitor. Hence, adult workers have there been less willing to
accept special trainee and youth rates of pay, and have been more insistent on
rate for the job rules which apply standard rates of pay to all workers,
whatever their age or training status. Thus, among the countries Marsden and
Ryan examined, Belgium, France, and Italy all fell into the latter category with
small or non-existent discounts on youth pay and youth exclusion was the
rule. In contrast, in Britain and Germany, apprenticeship rules made low
youth rates for trainees more acceptable, and thus facilitated greater youth
access to employment: a system of regulated inclusion.

Both this and the previous example suggest that the impact of collective
bargaining institutions upon employment depends less upon their existence
or otherwise, but on what kinds of policy they pursue. Adversarial relations
may be compatible with good overall economic performance under
decentralised bargaining, and with relatively simple production processes
which are less subject to the dysfunctions of tight management control.

3.6 Substantive rules underlying cooperation

There are two main areas in which workplace cooperation may require
substantive rules: a degree of employment security may be necessary to ensure
cooperation in job flexibility; and the various institutions of workplace
 -39-

cooperation, and cooperative working itself, may call for smaller occupational
pay differentials than would be the case under competitive markets.

Employment security and job flexibility

One the biggest problems in achieving cooperative exchange is reassuring

workers that sharing information with management may threaten their own
jobs. Passing on ideas for more efficient practices, and helping train fellow
workers so that they can work more flexibly or undertake some routine
maintenance involve the surrender of a measure of individual bargaining
power, and may enable the employer to dispense with certain jobs. Workers
may take a long term view of such issues if they believe they will still be
employed by the firm when the benefits start to accrue.

Cooperative exchange and wage hierarchies

It has been common to contrast two types of management organisation: one

in which it exerts central control and plays the role of a specialist coordinator;
and one in which a greater part of the coordinating activities are undertaken
by work teams sharing information, and management, as a specialist
coordinator plays a smaller part. Studies of Japanese management indicate that
positions of authority are less sharply delineated than in many western firms,
and that greater reliance is placed upon coordination through teams (Aoki,
1988). Comparing British and Japanese software engineers, Lam (1994) has
shown that work roles are much more sharply delineated in the UK whereas
in Japan overlapping work tasks between jobs, and systematic job rotation
leave much more coordination done between work teams rather than through
specialist coordinators.

Where firms use specialist coordinators, there are strong arguments for
concentrating rewards on those in these key positions (eg. Alchian and
Demsetz 1972). Substandard performance in these roles is likely to damage the
whole organisation. In contrast, in an organisation with more widely diffused
responsibility, such roles are less critical, and so require less rewards.

In contrast, wage differentials may be smaller in an environment of

cooperative exchange. With extensive cooperative exchange and information
sharing, coordination tasks are less concentrated on key positions, and greater
reliance is placed upon consultation and joint coordination. Because less
hinges on the decisions of a few critical individuals, there is less reason to tilt
the reward structure towards management. Indeed, it may be that team
oriented work structures favour greater equality of rewards because of the
greater mutual dependence.

There are, therefore, some reasons to expect cooperative exchange to work

 -40-

better with smaller wage differentials between managerial and other

employees than with the scale of differentials needed under the control
oriented models. Thus, the reduction of wage differentials associated with
some kinds of union action discussed in the first section may harm incentives
only when there is a control-model of management in operation.

Encompassing bargaining coalitions and wage differentials

Policies to reduce wage inequalities have been common when unions have
been closely involved in general fights against some national emergency, such
as the control of inflation, and when they have sought to develop
'encompassing' structures which embrace a large percentage of their potential
membership. 23 Examples of such policies can be found in those of the British
and Italian unions supporting their countries' incomes policies of the latter half
of the 1970s, and in the long term policies of the unions in Germany, Sweden,
and a number of the other countries with 'neo-corporatist' bargaining systems.

The association between wage solidarity policies and strong central

coordination in wage bargaining may be explained by the organisational needs
of the union groups leading such policies. All unions, even strongly
centralised ones, are, in effect, coalitions of many different groups of workers,
some of which would have considerable bargaining power on their own while
others are much weaker. The problem facing the union leadership is that of
holding the coalition together, knowing break-aways by powerful members are
always possible.

The main argument for the beneficial macro-economic impact of centralised

or coordinated bargaining by encompassing union and employer organisations
is that it is easier for such groups to achieve moderation in the interests of
longer term growth than under more decentralised bargaining (Calmfors and
Driffil 1988). This entails that powerful groups of workers forego their short-
term advantage and sacrifice possible wage gains, whereas weaker groups will
do relatively better. Managers in prosperous firms might also wish to pay
their workers more in order to facilitate organisational change or use of more
performance related pay. In the short-run, at least, groups with greater
bargaining power have to sacrifice their selfish interest in that of some wider
goal. However, such groups are also those which could most easily prosper
outside the encompassing union or employer association.

The problem therefore is how to ensure that the coalition holds together. The
only bonds would seem to be either long-term self-interest, which may seem
very abstract when compared with the immediate and concrete gains from
breaking away, or some moral goal. Hence the importance for encompassing
unions to have wider goals such as solidarity, greater equality, or social
justice. These provide a moral justification for the self-restraint by stronger
 -41-

groups. It is notable that both the British and Italian unions justified their
wage equalisation policies in the fight against inflation by the idea that those
on higher pay had to help the lower paid who suffered disproportionately
from price rises. It is also notable that the Swedish unions promoted greater
wage equality as one of the goals of central bargaining (Hibbs 1990) and that
the German unions, although not stressing equality to the same extent, have
long stressed the importance of skill and its associated rewards, as something
which can be achieved by all workers, rather than individual performance.

Thus it seems likely that neo-corporatist arrangements, whether as a response

to a particular emergency, as in Britain and Italy, or whether a long term
organisational strategy, as in Sweden (until recently) and Germany, contains
an inherent bias towards reducing wage differentials, and as a by-product,
reducing individual wage incentives.

4. Recent changes favouring more flexibility

The processes discussed so far have already generated some response from
firms, employers' and workers' organisations, and these may provide clues as
to policies whose extension might be encouraged in other firms, industries, or
countries.

4.1 Reforms in industry bargaining systems.

Among the biggest problems facing industry bargaining systems in recent

years has been that of developing sufficient flexibility at the enterprise and
establishment levels without destroying the benefits accruing from a stable
system of multi-employer regulation, such as limiting tendencies towards
competitive pay bargaining. Adaptation has been needed to give firms more
leeway to adapt to increasingly volatile international markets and faster
moving competition. It has also been needed to enable management to carry
through a number of organisational reforms. These two factors, greater
variability in costs, and greater freedom for management to adapt work
organisation have been among the key factors behind the recent movement to
greater enterprise autonomy in bargaining in some countries (Katz 1993,
Windolf 1989). This section examines some examples of attempts to adapt
bargaining institutions in this direction, and some of the motives driving them.
 -42-

Germany:1993 agreement for 'hardship clauses'

One of the most significant developments in German industry bargaining in

recent years has been the introduction of special 'hardship clauses' in industry
agreements to ease the employment consequences of the convergence of wage
rates between former East and West Germany. 24 The most famous case was
the metal industry agreement of May 1993, but similar clauses have also been
negotiated in other industries such as printing, textiles, and clothing (Bispinck
R, and WSI-Tarifarchiv 1993). The provisions are intended to enable firms
facing severe financial difficulties to opt out of the industry agreement for a
limited time to enable them to avoid lay-offs and to carry out restructuring.

The hardship agreements are concluded between individual firms and their
works councils, and have so far involved such concessions as delayed
implementation of the industry pay agreement and possible payment of lump
sum compensation to the employees at the end of the period. By December
1993, fifteen such agreements had been reported in the metal industry in
Saxony (EIRR 12.1994).

Although the circumstances of the agreement were exceptional, as were the

strains on industry bargaining imposed by reunification, and the resulting
employment adjustments, the idea appealed to a number of employers and
commentators who stressed the utility for extending the idea to all industry
agreements, and across all Germany. 25 In fact, the agreement on hardship
clauses is fairly cautious, and builds in considerable guarantees to maintain
the integrity of industry bargaining.

If individual firms and their works councils could unilaterally exempt

themselves from the provisions of industry agreements, the latter would very
quickly lose their force. Too restrictive a rule would also harm industry
bargaining as the employers were faced with increasing discontent from small
and medium-sized firms who complained that they could not remain in
business under the agreements in force, and were therefore leaving the
employers' organisation. 26 Equally, union members might question the value
of organisations that whose policies were pricing them out of jobs. Thus,
deciding the conditions under which hardship clauses may be invoked is
critical to the survival of industry bargaining. The compromise reached left the
decision with the parties to the industry industry agreement: the industry
employers' organisation and the union. Should they fail to reach agreement,
there is provision to set up a joint commission, but if that fails to agree, then
the industry agreement should be applied.

It remains to be seen whether the new opt out clauses will be effective, either
in giving enough flexibility to enable some firms to survive or in halting the
erosion of membership of the employers' organisations and of that of the
coverage of industry agreements.
 -43-

Italy: reform of collective bargaining system

Industry bargaining in Italy has undergone two major reforms during the past
ten years: the first reducing the coverage of the indexation provisions of the
Scala Mobile in 1983 and 1984, and the second, involving the final abolition of
indexation and the agreement, in July 1993, to establish a new system of
articulated bargaining.

The impact of the first change has been to increase the scope for adjusting pay
at the enterprise level, and to enable employers to reverse the compression of
pay differentials that occurred as a result of flat-rate indexation. The wage
gap, discussed earlier, increased as a result of these reforms, and occupational
wage differentials were markedly increased (Ministero del Lavoro 1987, ch 4).

The agreement reached in July 1993 sought a new constitution for collective
bargaining with national level discussions on incomes policy, sectoral
agreements setting pay for two years and other issues for four, supported by
a system of enterprise or territorial agreements of four year duration. The aim
of the sectoral agreement is to set pay within a forum which enables
coordination with macroeconomic performance, while leaving room at the
enterprise level for pay to reflect changes in productivity, quality, and other
elements of company performance. The system should be articulated in the
sense that the higher levels set a framework for negotiation at the lower levels.
In particular, company level negotiations should be bound by what has been
agreed at the sectoral level. At the enterprise level, a form of works council
should be established with two thirds of its members directly elected by
employees, and one third by union members (Protocollo, 1993, EIRR 236, 9.93).
Enterprise level representation was agreed in December 1993 (EIRR 241, 2,
1994) and the first sectoral agreement, in the chemical concluded in the
chemical industry, was concluded in March 1994. Shortly after its formation,
the first government of the new, reformed, electoral system reaffirmed its
support for the new industrial relations in system.

The new system has been developed from an analysis of the ills of the former
system. Apart from reducing wage differentials, the former indexation system
had also greatly limited the scope for collective bargaining at both sectoral and
company level. New initiatives at the sectoral level and change at company
level could only be 'funded' by adding to the inflation payments. By
substituting national level discussions for indexation, it becomes possible for
industry-level negotiators to make allowances for likely pay supplements
negotiated at enterprise level, and so keep the overall total increase in line
with macroeconomic objectives. Reforming work place representation with a
mixture of individual and union-based representation is intended to overcome
the declining support for the previous workplace system, and thus to increase
the effectiveness of workplace employee negotiators.
 -44-

France: growth of enterprise bargaining

After several government led attempts to encourage greater depth of collective

bargaining during the 1970s, the Auroux laws introduced in 1982 appear to
have given a considerable boost to bargaining activity at the enterprise level.
In the late 1970s, industry agreements outnumbered enterprise agreements by
about two to one. However, by 1986, about the number of enterprise
agreements had more than tripled to about 6,000, outnumbering industry
agreements by six to one (Eyraud and Silvestre 1991).

This growth appears to have favoured two developments. First, it is likely to

have contributed to the faster growth of average earnings over industry
minima after 1986 (Lanfranchi and Afsa, 1990). Secondly, it appears to have
been associated with the increased use of merit increases, which by 1991,
affected nearly 60% of employees, and represented about one third of the
annual pay increase. According to the Ministry of Labour survey, the most
widely used form, that of mixed general and merit increases, has been the
subject of enterprise bargaining in the majority of firms applying it (Ministère
du Travail, 1992). In view of the concentration of bargaining and of merit
increases in larger firms, it is likely that such negotiations affect a still larger
proportion of employees.

Sweden: decentralisation of model.

The dismantling of central bargaining in Sweden began in earnest in 1983 with

the break-away of the metal industry employers, and was consumated in 1990
when the SAF decided to end central bargaining and confine its activities to
coordination of industry bargaining. Steps in a similar direction were been
taken in the Swedish public sector with the decision to promote local
bargaining over central pay fixing with the 1985 reforms designed to promote
bargaining at sub-sector and agency level (Schager 1992).

The decentralisation in Sweden appears to address a number of problems.

From the early 1980s, coordination among the key actors was increasingly
difficult. The classic model of central bargaining had been dominated by the
bargaining partnership between LO, the main private sector blue collar union,
and SAF, the private sector employers' association. With strong shared
interests in success of the export sector of the economy, and an ability to
assume wage leadership, the two organisations had been able to agree pay
increases consistent with continued export competitiveness (Martin, 1984). By
1980, this partnership had lost its dominance over other bargaining parties,
notably in the public sector and among white collar workers, and any wish LO
had for wage moderation was challenged by more militant demands from the
other unions, and an era of wage competition set in (Kjellberg, 1992).
 -45-

Apart from losing confidence in the ability of central bargaining to coordinate

effectively, private sector employers found a number of its features
increasingly constraining. The wage equalisation policy had squeezed the
differentials of skilled blue collar workers, and opened the gap between them
and white collar workers, resulting in recruitment and retention difficulties
which could only be resolved by agreeing additional payments locally. This
added to the strains on the central bargain. Finally, employers needed more
autonomy at the local level to develop new production methods and new
patterns of work organisation, and found established pay structures too
restrictive (Myrdal, 1991). Thus the increasing pressure from private employers
for greater pay autonomy arose from breakdown of control of central pay
bargains, and new needs at the local level.

The decentralisation to industry bargaining would appear to be reflected in the

increased differentials among industries in Sweden during the 1980s (although
some movement in this direction was also in progress during the 1970s, Table
3). In the public sector, Schager shows that the decentralisation of public sector
pay determination coincided with an increase in the dispersion of earnings
among public servants from 20% to 28% between 1985 and 1990. Over the
same period there was only a small increase in the private sector.

Australia: enterprise bargaining

Australia has traditionally had a relatively centralised system of pay fixing

through a system of national and industry awards by its industrial tribunals
which set minimum rates of pay. Industry, and often enterprise bargaining,
could translate these awards into effective pay, negotiate additional
supplements ('overawards'), and deal with other conditions of employment.
It has been widely argued that the award system leaves too little room for
adjustment to local factors. Even though the minima may follow competitive
pay pressures, for which there is conflicting evidence (Niland, 1986), they
would do so for a tribunal's whole jurisdiction, irrespective of individual
employers' ability to pay. Negotiated overaward payments provide additional
flexibility, but, according to Plowman (1986), this is not always the case.

Green and Macdonald (1991) cite survey evidence from the 1991 AWIRS27
supporting the view that the award system had removed responsibility for
confronting certain issues from workplace level unions and management.
Particularly in large workplaces, many managers reported employee related
factors as barriers to change, and among those cited, the most important were
unions, lack of resources, and government regulations.

Thus, the shift towards enterprise bargaining fostered by the 1991 National
Wage Case represents an attempt to give enterprise level bargainers greater
scope over pay in order to agree new working practices and new
 -46-

arrangements for workplace representation.

4.2 Changes in wage structures likely to mitigate employment effects.

One of the effects of union and government action on pay that is likely to
reduce employment reviewed earlier was the reduction of wage differentials,
particularly as concerned lower paid and less skilled workers. During the
1970s, in many OECD countries, wage differentials were compressed, often
under the influence of such policies. This movement could be seen in most
dimensions of wage structures: dispersions, as well as differentials in earnings
by skill, industry, and by gender (Table 3).

In contrast, during the 1980s, much of the earlier movement of compression

was reversed (Marsden 1990, Katz et al.1993). There has been considerable
debate over the causes (for a review see OECD 1993). Important among these
have been the strong demand for educated labour induced by technical
change, increasing faster than supply in the US and the UK (Bound and
Johnson 1992, Bean et al. 1990), and high unemployment, especially among the
unskilled (Pencavel 1994).

However, more important in the present context is whether institutional

arrangements have allowed sufficient flexibility for the changes to take place.
Exploring these may give some clues as to ways of making bargained and
administered wage structures more responsive to labour market pressures.

A number of changes in bargaining arrangements in various countries can be

cited as permissive factors enabling changes in wage structures:

France: decline of industry bargaining and greater importance assumed by

enterprise bargaining during the 1980s; from the mid-1980s, less
government support given to SMIC which ceased to boost the relative
earnings of low paid workers after about 1984 (Katz et al. 1993); severe
public sector pay limits from 1982 (Meurs 1992);

Germany: limited government policies to 'deregulate' labour markets and

to limit unions' ability to strike (Bosch and Sengenberger 1989);

Italy: reassertion of the 'right to manage' in large enterprises from the late
1970s (Becchi and Negrelli, 1986); government led negotiation of partial
de-indexation in 1983/84 (dell'Aringa and Lucifora 1990), and its
abolition in 1992;

Netherlands: decline of industry bargaining and growth of enterprise

bargaining during the 1980s (Visser 1990);

Sweden: the breakdown of central pay negotiations from 1983 when the
 -47-

metal industry employers broke away, and their permanent

abandonment decided by the employers from 1990; 1985 public sector
pay reforms seeking more decentralised pay fixing (Schager 1993);

UK: union concessions in the early 1980s decided against a background of

closures, restructuring and heavy job losses; general retreat from
industry to single-employer bargaining (Brown and Wadhwani, 1990);

US: union concessions in large unionised plants (Kochan, Katz and

McKersie (1986)); government policy not to update minimum wage
with inflation;

Thus, in many OECD countries, there have been changes in institutionally

regulated wage structures during the 1980s which would seem to favour
employment. The persistently high levels of unemployment in many countries,
and especially those in EU countries, suggests there may be need for further
change. However, management's needs within the enterprise, and the need for
a stable framework for cooperation may set a limit on how much wages
should respond to external movements in labour markets.

4.3 Spread of 'merit pay' and individualised increases.

'Merit pay', or performance related pay decided by the employer according to

management assessment of an employee's performance, has been spreading
fairly rapidly, particularly in Britain, France, and Italy among the West
European countries. It has been widely regarded as a new source of flexibility
in remuneration systems for employers enabling them to respond better to
external labour market pressures, and to use pay in order to motivate staff. It
has also been associated with a move away from pay systems and tight
supervision developed for repetitive jobs under mass-production to newer
forms of effort management (Schudlich 1991). As with a number of other
issues discussed in this paper, its effect on unemployment is more likely to be
indirect, by giving enterprise management more local influence over pay, and
more freedom to adapt remuneration systems to meet new management
needs.

Survey evidence from France, Italy and the UK, indicates that its use is now
quite widespread, but that it accounts for a greater share of total earnings
among managerial than among other staff. For other staff, its use has
increased, especially among white collar staff, but the percentage of total pay
resulting from it has remained limited.

As discussed earlier, the spread of merit increases in France has been

associated in part with the growth of enterprise bargaining, and the
increased tendency for enterprise management to use pay as a part of its
 -48-

personnel policies rather than rely on industry agreements. Among the

reasons reported by the French Ministry of Labour for firms using
individualised pay in 1986 were the need to promote organisational change
in an environment constrained by the elaborate job and pay classification
systems, to reduce absenteeism, and to restore skill differentials 28.

Similar movements have been under way in Britain and Italy with
increased use of discretionary bonuses and merit pay systems (Della Rocca
1989, Negrelli (1991).

How far the spread of merit pay has contributed to greater pay responsiveness
to markets is not clear. It may simply give management more autonomy over
pay for the purposes of local enterprise management. It may establish a closer
link between individual pay and individual performance, but there are
considerable difficulties for management to measure the performance of
individual employees accurately (see §2.2 above). It may also enable
management to allocate more pay to highly skilled or able employees who
might otherwise leave. One indicator of change is the growth of earnings
dispersions within occupations which might signal the decline of rate-for-the-
job systems. Data available for France, Sweden and the UK show an increase
in intra-occupational dispersions during the 1980s. In France, the movement
began from the mid-1980s (CERC, 1989: 38). In Sweden the movement began
in the early 1980s, and in the UK, from the late 1970s (Marsden 1989).

4.4 Flexibility agreements.

Bargaining over changes in working practices has been primarily associated

with 'job control' unionism as distinct from models based more on
participation through elected councils (Delamotte 1971). Thus, concession
bargaining and flexibility agreements concerned with changes in working
practices such as seniority or job demarcations have been primarily associated
with 'job control' unionism. As Delamotte observed in his study of
productivity agreements of the 1960s, one may contrast cases where
negotiation fixes on changing specific working practices in exchange for more
pay from those where a works council may have to agree general issues
related to changes in working practices, but dissociated from pay bargaining.
Like the earlier wave of productivity agreements, concession and flexibility
negotiations have provided a means by which management can gain support
for change through collective bargaining.

There is no systematic data source for such agreements, so it is not possible

to assess directly how much change takes place outside such agreements, and
there has been little evaluation of how far the reported changes were actually
delivered, nor of their subsequent effects on productivity.

A number of case studies in the United States reveal a significant role for such
 -49-

agreements in helping to restructure firms and restore their lost

competitiveness, notably in the automobile industry (Kochan et al 1986; Katz,
1985). In the UK, there was a wave of flexibility agreements in the 1980s, often
concluded against a background of severe economic difficulty. Marsden and
Thompson (1991) argued that they were genuine agreements (as opposed to
unions acquiescing in management decisions), and that there was tentative
evidence that they, and related bargained concessions, had contributed to the
productivity surge of the early 1980s. Productivity concessions in exchange for
pay or other benefits such as shorter hours have continued (Richardson and
Rubin 1993) although it seems that the productivity concessions have been
more far-reaching than in the early 1980s.

4.5 Unionism and new management methods.

Finally, an important question is whether existing patterns of employee

representation have impeded or facilitated the introduction of new
management methods, such as lean production, flexible specialisation, team
working, and total quality management. This is issue is dealt with in more
detail in the chapter on management practices and unemployment,
nevertheless, some preliminary considerations are in order. There are two
main reasons why a number of writers have argued that these methods are
incompatible with current patterns of employee representation: they are often
alleged to be based on individualised relations between employees and
management; and they involve a greatly increased right to manage which is
incompatible with joint regulation.

This dichotomy has been refined for intellectual debate, but in practice, in
several countries there is evidence that unions, workplace representatives, and
local management have been able to reach working compromises. The reason
is probably the same as that behind the flexibility concessions agreed in Britain
and the United States: neither workers nor management have many illusions
about the survival of non-competitive firms. In practice, lean production has
been widely adopted in Japanese firms that were unionised, and from there
spread to North American and European producers. In Italy, the survey
reported by Negrelli (1991) showed individual negotiation over aspects of
labour deployment and flexibility developing within a framework of collective
rules for employee representation and adjustment of the production process.
In another survey, or northern Italian firms, Regalia and Ronchi (1989) found
that practices such as quality circles and management-employee
communication programmes were positively correlated with local union
presence, although there had been some conflict over introduction of quality
circles. They also reported that they had been widely used for purposes of
work reorganisation and increased flexibility. In Germany, in the automobile
industry, lean production, team working, and new skill concepts have been
dealt with by enterprise management and the works councils, and the metal
workers' union, IG Metall, has been active in promoting its own concepts of
 -50-

team working. The main conflict has not been so much over the use of new
working methods as over the powers of works councils to oversee such
processes as staff appraisal, and whether the emphasis of merit pay should be
upon skill, which can be externally measured, or management assessment of
an employee's performance (Huber and Lang, 1993).

Such examples can be multiplied (see for example reports on individual firms
in EIRR), but they show only that in many firms, management and employee
representatives have been able to introduce new concepts of employee
management and work organisation by adapting rather than abolishing pre-
existing forms of employee representation. There is no necessary contradiction
between them, although it would not have been possible to introduce these
changes in the face of militant or ideological worker opposition.

Without detailed case studies, it is not possible to evaluate whether the

adoption of new methods would have been faster or slower without employee
representatives being involved, nor can one establish whether employee-
management relations would have been more or less stable afterwards.
Management may often be able to introduce changes unilaterally against
workforce opposition, but if it cannot maintain a good working relationship
subsequently, the hoped for productivity and quality may not be forthcoming.

5. Policies to increase adaptability.

At the microeconomic level, the main analytical conclusions to this paper may
be summarised by looking at the effects of collective employee representation
on three broad groups of workers. For the less skilled, lower paid groups,
raising their rates of pay by collective action, in the absence of suitable
training policies, is likely to cause greater unemployment. For the highly
skilled and educated, reducing pay differentials may reduce incentives to
train, and, perhaps more important, to take the risks associated with exercising
initiative and responsibility. For the middle groups, those with blue and white
collar technical and intermediate professional skills the outcome is less clear.
For such workers, both employer and employee often have considerable sunk
costs in the employment relationship: detailed knowledge of the employer's
business, skills learned on the job, and so on, such that both parties incur
considerable costs if the relationship ends. For this group, motivation, and the
quality of the exchange between employer and employee is the vital ingredient
to good economic performance.

There is something of a trade-off between policies for these three broad

groups. Unskilled and semi-skilled employment may benefit from greater
wage flexibility, and flexibility of individual incentives may encourage
performance of the managerial group. But encouraging cooperative relations
between workers and management, and among workers in the middle group,
 -51-

may militate against both wage flexibility and high managerial incentives.
Flexible working and information sharing needed for high productivity and
quality among the intermediate group requires an environment of cooperative
exchange.

Cooperative working may require greater equality of rewards than the right
to manage, more market-oriented, model. First, it has been argued (§3.4 above)
that cooperation needs a stable framework, and that when this is provided by
the parties involved, it relies upon coalitions of interests within encompassing
groups. Because the existence of such groups usually entails a loss of
individual bargaining power by the strongest groups, some wider moral or
ideological purpose is needed to motivate them. In the post-war years, this has
mostly been provided by solidarity and egalitarian sentiments. Secondly,
cooperative working in the workplace, where people work flexibly within
teams implies a less sharp division between tasks of coordination and
organisation and those of execution. Where there is a strong polarisation it
makes sense to concentrate rewards on those in the coordinating role because
their performance is critical. Without such polarisation, cooperation depends
more on sharing information and responsibility so the need for exceptional
rewards is weaker. Indeed, it might even harm cooperation.

There is therefore a tension between the policies which favour greater

employment opportunities for the unskilled and low paid and incentives on
the one hand, and those favouring cooperative exchange on the other. In the
short-run, the first might have a swifter impact on employment, but in the
longer run, cooperation could prove more effective by promoting the general
level of efficiency of firms, and thus their position in national and international
markets. We return to some of these issues in the chapter on management
practices.

a) Coordination and the need for a macroeconomic forum

The main function of central coordination of pay bargaining is to enable the

parties to tackle problems that are more difficult to deal with in a more
decentralised way, and to achieve outcomes that might not otherwise be
possible. The difficulty of achieving rates of pay increase consistent with full
employment when bargaining groups were large but nevertheless fragmented
was one such issue. Traxler's (1993) survey suggests that coordination rather
than a having central bargain is the most important factor in restraining group
pressures. When coordination of the main parties broke down in Sweden, so
did central bargaining. Countries with behind the scenes coordination, such
as Japan and, to a lesser extent, Germany have fared more successfully on
inflation and unemployment in recent years than those with central bargains.

Coordination has to work by gaining consent rather than by constraint so that

 -52-

the main problem is that of communicating bargaining pressures and

economic problems among the parties involved, and of forming a consensus
of what goals are realistic. Coordination in the Japanese case is interesting
because of the main organisational strength of the unions lies at the company
level, and although the industry federations play a key role in coordination of
bargaining, they have little in the way of coercive power (Dore 1987). In the
run-up to the annual 'spring offensive' there is intense consultation sharing
information on the problems to be resolved during the intensive period of pay
bargaining. In this way, the problems facing individual enterprises can be
given as much weight as economy-wide considerations. Thus, the system
avoids the fragility that came to characterise central bargaining in Sweden, and
which led to its collapse.

By its very nature, coordination does not dictate a single institutional

structure. In some countries this has worked through central bargaining, but
both Germany and Japan have achieved similar outcomes without there being
central bargains. Indeed, as Traxler observed, in neither country is
coordination institutionalised, being mostly behind the scenes.

There may indeed be considerable advantages to coordination over central

bargains as the latter may often become transformed into the minimum level
any industry or enterprise level worker representative can accept without
losing the confidence of his or her membership. Equally, it is difficult for
central bargainers to be too specific about how much is gained centrally, and
how much at lower levels without appearing to the rank and file to sacrifice
their interests to those of some abstract notion of the 'wider economy'.

Apart from discussing pay in the coordination process, a number of other

more qualitative issues can be broached. For example, the encouragement of
different kinds of pay innovation, or working time régime could be discussed,
as could the approach to retraining, lay-offs, or of how to deal with the
problems of the long-term unemployed. These might then take more substance
at the industry or enterprise level. Especially suitable for discussion at the
peak level are issues where individual firms would find it difficult to act in
isolation. For example, it is widely argued that firms offering employment
security in isolation will be penalised by attracting poor quality workers who
stand the greatest risk of being fired by other workers (Büchtemann 1993).

b) Industry level pay negotiations

Industry bargaining provides one means by which national level coordination

may be translated into concrete outcomes, but two conditions are needed for
this to be successful. First, industry bargainers need to be involved in the
coordination process; and secondly, the agreement itself has to leave enough
room for there to be effective enterprise level negotiations on local issues. The
 -53-

potency of the competition between rival groups excluded from the peak level
negotiations in Sweden between LO and SAF, and their unwillingness to be
bound by agreements which represented the needs of the industrial sector
illustrate the problems arising in the first case (Kjellberg 1992). The pressures
for bargaining decentralisation highlighted by Katz (1993) and Windolf (1989)
arising from the need for enterprise level flexibility underline this.

Industry bargaining generally sets minimum rates of pay for whole industries
in a particular region, and usually play an important role in determining
annual pay increases. Although the industry bargaining systems discussed in
this paper leave considerable room for adjustment at the enterprise level, there
being a wide and varying range of wage gaps, this does not always meet the
needs of firms that enter into severe employment crises. This is particularly
a problem for small and medium-sized firms which, although efficiently run,
may still face extreme difficulties as a result of the loss of a major order, or an
exchange rate shift suddenly wiping out a once profitable market. In such
cases, industry minima may be too restrictive, and there is a strong case for
the development of 'hardship clauses' to enable the firm to weather the storm.
Although it is still early to evaluate the outcome of the German experience,
first reports suggest that the guarantees built into the 'hardship clauses' have
given flexibility while maintaining the overall integrity of industry agreements
which provide a critical element in overall bargaining coordination there.

The other major role that can be played by industry agreements is to provide
a framework for negotiations at enterprise and plant level to enable existing
work rules to be adapted to the new patterns of management and production
organisation. The importance of trust was stressed earlier for there to be
cooperative exchange in the workplace. The introduction of organisational
change can be extremely important to enable firms to respond to the increased
competition flowing from the removal of tariff and other trade barriers. Yet
a workforce on the defensive could prove a serious obstacle to the efficient
working of such practices. If management's goal in reorganising is to innovate
and raise efficiency, and not to destroy existing forms of employee
representation, then it will be important to be able to signal these intentions.
An agreed framework at the industry level, and agreed means of supporting
existing representation could provide a strong signal of commitment to
working within agreed procedures.

c) Enterprise level policies.

Many of the working practices commonly cited as causing inefficiency have

been reduced or removed during the 1980s, and the influence of 'rate-for-the-
job' pay systems has declined introducing a greater degree or variability
according to individual skills and performance. Such practices, however, are
never removed once and for all because, as argued above, they are born of
 -54-

workplace custom and workplace conflicts. As a result, the problems of

maintaining flexible patterns of work organisation and elements of pay
flexibility in the workplace are continuous rather than discrete. Workplace
custom does not cease to exist when the rules are changed. Rather, the
substantive content of what is accepted as fair and reasonable changes.
Therefore, the problem is one of maintaining the atmosphere in which change
was agreed, and of maintaining a consensus on the purposes which have to
be met. This would rather highlight the importance of maintaining cooperative
exchange.

This does not have to be tied to pay flexibility, but there are a number of areas
in which some of the recent developments in pay could help promote
cooperation. It was argued earlier that at least one of the reasons for the
widespread acceptance of 'rate for job' pay rules was that in an environment
of mutual suspicion, the application of such rules could be easily monitored.
Similar factors have laid behind the difficulty of using both merit pay and
profit sharing. In a study of the UK tax office merit pay scheme, Marsden and
Richardson (1994) found that one of the most important factors which turned
general support among the employees for the principle of relating pay to
performance was distrust of the way management was applying it. In the case
of profit related pay, the French government found that the only criterion that
was acceptable to the unions was that of declared profits because only this
could be effectively monitored (Camerlynck and Lyon-Caen 1982).

Many production systems have evolved beyond the point at which workers
produced a highly standard output with little scope for quality. This change
has undermined many of the traditional individual output-based incentives,
a point recognised in the debates on payment systems in all the major EU
countries (eg. Schudlich 1991, Della Rocca 1990). Equally, relating pay to
average experience, as is done by length of service payments is unpopular in
many administrative-type jobs because of the automaticity of the pay whatever
the quality of effort. 29 Indeed, both kinds of system are forms of rate for the
job: output based incentives being based often on a form of piece price; and
length of service increments corresponding to steps towards the rate for the
experienced worker at the top of the scale.

Provided it is linked to a suitable system of performance appraisal, and the job

offers enough scope for individual improvement, then 'merit pay' offers a
means of encouraging better productivity while at the same time diluting the
effect of 'rate for the job' principles. It is a system that has been spreading
rapidly in the British public sector, and in France in both public and private
sector employment especially in terms of the numbers of workers covered if
not in the amounts of money individuals receive (CERC 1989). There has as
yet been little evaluation of the effectiveness of merit pay in improving
workers' motivation. A recent study tax office staff in the UK suggested that
the scheme in force there could indeed be demotivating, but that the main
 -55-

cause of its failure lay in its perceived unfairness (Marsden and Richardson
1994), and notably in the suspicion that management had placed a ceiling on
good evaluations for cost reasons irrespective of how good individual
performance was.

This points to an important feature of 'merit pay' systems, that they are part
of a wider wage-effort bargain, and so cannot easily be insulated from pay
bargaining. In addition, markets reward individual merit only in so far as it
meets customer demands ( a question of outputs) whereas individual workers
tend to judge their work from the quality of the inputs. The difficulty of
measuring performance to the satisfaction of all parties, and the problem of
establishing the amounts of money to be used for merit awards indicate the
need for joint regulation.

Profit sharing schemes also break away from rate for the job rules, and
introduce a modest element of variability into pay in response to the firm's
performance. Recent work reviewed by Blinder (1990) and by FitzRoy (1992)
indicates that it can indeed improve performance. A frequent theoretical
objection is that the impact of any one individual's effort on profits is too
small to be noticed, and that there will be tendency towards free riding as
those who put in minimum effort will get the same as those who work much
harder. This would indeed be a problem if the employment relation were
short-term, but because it is still usually long-term (with median job durations
of ten or more years), there is ample scope for peer group pressure to limit
'free riding'. Thus, profit sharing can turn to advantage the very continuity of
employment relations that makes norms of fairness so powerful.

Productivity bargaining, and flexibility deals are another important category

of instruments for use at the enterprise level. Although discredited in Britain
in the late 1960s on account of phoney deals designed to evade incomes policy
restrictions, during the 1980s and 1990s in Britain, management has been much
more successful in using them as a vehicle for change. Although it is not
customary in most continental countries to link pay increases to changes in
specific working practices there have been other developments. For example,
in France, recently, some firms have negotiated pay increases conditional upon
attainment of some specified improvement in productivity (Eyraud et al. 1990).
These examples could well be generalised by means of a frame agreement.

However, 'rate for the job' rules are a product of low trust relations between
employers and employees. They have been a powerful principle in collective
bargaining because they are easy for workers and their representatives to
monitor. Merit pay, profit sharing, and productivity change depend upon a
much greater amount of trustworthy information being available from the
employer. For that reason, the information needs to be subject to joint control.

d) Training and product market competition

 -56-

Two final policy areas include training and increasing product market
competition. Training emerges in two important contexts: the problems of the
unskilled unemployed, and that of labour market bottlenecks. Policies to foster
cooperation among the middle section of the workforce, it was suggested, tend
to favour smaller occupational wage differentials and greater equality of
rewards, but these are harmful for the jobs of the unskilled. Providing greater
opportunities for improving one's own skill levels, and for replacing obsolete
skills could ease the policy dilemma in this area.

Bottlenecks in labour supply, it has been argued, pose major problems to

fragmented bargaining systems because they distort wage structures between
those who can improve their earnings and those whose pay is locked into
organisational or bargained pay structures. Recently, Nickell (1990) has
stressed this as one of the major obstacles to the UK achieving low inflation
and full employment.

Product market competition may have much to contribute to easing any

distortions in wage structures arising from collective bargaining. Stewart's
(1990) evidence for the UK highlighted the importance of rent sharing in the
union mark-up. Thus, increased product market competition could do much
to ease some of the employment displacement problems associated with
monopolistic pay bargaining.

It could also play an important part in the realm of restrictive practices.

Analysing the impact of the employment law reforms of the 1980s in the UK
which had done much to reduce the institutional power of trade unions,
Brown and Wadhwani (1990) concluded that the most important factor had
been the competitive pressures on management to take the initiative in
promoting organisational change. In a study of the UK automobile industry's
labour adjustments in the early 1980s (Marsden et al. 1985), the authors also
concluded that the depth of the commercial crisis facing the UK-based
manufacturers, and the government's announcement that it would not act as
employer of last resort had contributed greatly to the reorganisation of labour
relations in that industry.

Finally, product market competition may contribute to promoting cooperative

exchange because of the pressure on management to tackle sources of
organisational slack and to promote information sharing. In so far as these are
a source of increase efficiency, they put pressure on management to avoid 'live
and let live' environments. They may also put pressure on worker
representatives towards greater economic realism as to what can be achieved
for their members.

Thus, changes in bargaining structures and industrial relations systems to help

employment need to be supported by appropriate labour market and product
 -57-

market policies.
 -58-

End notes.
. For an extremely clear statement of this argument see the survey of Japan
1

in the Economist magazine (9.7.94).

2
. Chamberlain and Kuhn (1965).

. This does not apply in labour markets dominated by a single major

3

employer. Under such conditions of a single buyer, or of a cartel among local

employers, the wage could be bargained up to the competitive level with no
loss of employment. With the great increase in transport, the classic case of
monopsony, that of the company town, is less and less common. Perhaps the
most likely cases of monopsony now are those where workers' travel to work
range is very limited, for example, by the demands of child-care.

. These were: experience, schooling, gender, part-time status, marital status,

4

and industry.

. These results are broadly consistent with those for the US reported by
5

Freeman and Medoff 1984, and Lewis 1986, and for the UK by Gregory and
Thompson (1981), Blanchflower and Oswald (1988), and Metcalf (1990).
6
In a post entry closed shop workers are expected to join the union after the
.

employer has hired them. In the pre-entry closed shop, workers have to be
members of the appropriate union before they can be hired. The pre-entry
kind was made illegal in 1990.
7
The coefficient on membership density was 0.07, and on the presence of an
.

agreement, 0.03, for log earnings.

8
. Handelsblatt 3.7.92.
9
.The own wage demand elasticities were respectively for skilled, semi-skilled
and unskilled engineering workers: -1.06, -1.76, and -2.31. The substitution
elasticities were: skilled/semi-skilled, 2.63; skilled/unskilled, 1.68, and semi-
skilled/unskilled, 3.75.
10
Part of the discrepancy of findings may result from the use of different
.

kinds of survey. Atkinson and Meager based their conclusions on interview

with about 70 commercial organisations with more than 500 employees across
a range of sectors. Their questions were qualitative: whether certain practices
were in use for at least some categories, and whether they had increased, or
the firm was intending to increase their use. The idea that these changes
corresponded to a conscious strategy was the authors' own conclusion, and
was not supported by Hakim's use of a follow-up to the WIRS survey.
Observing business 'strategies' however is not easy, especially if one believes
that they tend to emerge from an accumulation of decisions taken as the
 -59-

opportunity arose. Thus, if employment security incentives were pushing firms

in one particular way, an overall, coherent, picture could emerge from their
successive employment policy decisions without there being any explicit grand
strategy. Pollert's critique suggests that Atkinson and Meager's model was not
accurate as a general description of employment policies of the majority of
firms in the middle 1980s, but it remains possible that they had captured a
direction of movement among key firms.

.Percentage of employees on fixed term contracts in 1991: 32% in Spain, 10-

11

16% in Denmark, Greece, Portugal, and France (Employment in Europe 1993:

181).
12
. Employers were asked in 1985 and in 1989 to assess the importance of
'insufficient flexibility in hiring and shedding labour' as a reason for 'not being
able to employ more people' (Employment in Europe 1993, ch. 7).
13
In some countries, such as France, use of early retirement has been heavily
.

subsidised by the pension funds, and so have proved a relatively cheap

measure for employers, although the cost to the pension funds and from
consequent reductions in the labour force are considerable.
14
. For more results see: European Economy no. 27 March 1986.
15
. Such complaints were common among Italian employers during the 1970s.
16
Defined as the amount by which gross earnings exceed the industry minima
.

as a percentage of the latter.

17
That is: excluding premia such as 13th month, intéressement, and length of
.

service bonuses.
18
1964 saw the first 'low wage pot': a common flat rate increase for every
.

worker; wage drift amounts; cost of living adjustments, flat rate; and low
wage adjustment amounts (Hibbs 1990 p.62 ff.).

. At a more general level, this argument finds some support also in game
19

theory. Maynard-Smith (1982) shows that the expectation that one will escalate
any conflict that threatens one's own territory (because that it something one
cannot afford to lose) produces a stable pattern of behaviour whereas the
conventional choice between aggressive ('hawk') and non-aggressive ('dove')
strategies does not.
20
This is the 'reverse induction' argument. The 'tit-for-tat' strategy implies one
.

should start by cooperating in the first round, and only refuse cooperation in
the round following one in which the other party failed to cooperate. Thus not
cooperating is used as a retrospective punishment. However, this sanction
 -60-

does not work for cooperation in the last round of a finite game, so both
parties fear the other will not cooperate. Since mutual non-cooperation means
that both parties lose the gains from cooperation, the threat of punishment for
non-cooperation also loses its effect. Therefore cooperation is also problematic
in the penultimate round, and so on, back to the first round (Dasgupta 1988).
21
The failure of the Donovan reforms to produce the gains in productivity and
.

low inflation that their architects had intended is demonstrated by Metcalf

(1989). Batstone (1988) reached similar conclusions reviewing the failure of the
reforms to generate the kind of cooperative practices intended, and problems
raised by increased expectations. A special issue of Rassegna Sindacale (86/87,
Sept-Dec. 1980) was devoted to the high expectations generated by the Italian
factory councils and which they proved unable to meet.
22
.Ryan (1989) takes the argument further in the context of public workplace
training schemes for young workers.

. 'Encompassing' groups are those which include a large percentage of the

23

members of a particular group.

. In 1991, the metal industry employers and union had agreed a staged
24

convergence between the basic wage rates fixed in industry-regional

agreements between former East and West Germany by April 1994, and
convergence for other benefits over a slightly longer period. Several factors
explain why the employers agreed to the convergence. Herr Stümpfe, then
president of the metal industry employers, argued that wage convergence
could not await productivity convergence because of the flight of skilled
labour from east to west in search of high pay and which was slowing
productivity convergence (die Zeit, 21.6.1991). It is also likely that the large
employers, who could better cope with convergence, were better represented
in the employers' association than small employers, hence the discontent of
small and medium employers (EIRR 221, 6.1992, and Pleiten, Pech und
Pannen, die Zeit,14.5.93).
25
.Mundorf H., in the Handelsblatt 16/17.7.93, and Barbier H, in the
Frankfurter Allgemeine Zeitung, 18.5.93, cited in (Bispinck R, and WSI-
Tarifarchiv 1993).
26
On leaving the employers' organisation they would continue to be bound
.

by the agreement until its expiry. Only after then would they cease to be
bound by it.
27
. Australian Workplace Industrial Relations Survey.
28
Bilan de la négociation collective en 1988, Ministère des Affaires Sociales et
.

de l'Emploi, Paris.
 -61-

29
.The Efficiency Unit (Next Steps) reported that 70% of civil servants
approved of the idea of rewarding good performance with better pay, and a
recent survey of Inland Revenue staff found that 57% thought that
performance pay was good in principle (Marsden and Richardson 1991).
References.

Addison J.T, and Barnett A.H. (1982) The impact of unions on productivity.
British Journal of Industrial Relations, Vol XX, No. 2, July 1982, pp.145-162.

Addison J, and Hirsch B. (1989) Union effects of productivity, profits, and

growth: has the long-run arrived? Journal of Labor Economics, 7:1, pp. 72-105.

Akerlof G, and Yellen J. (eds.) (1986) Efficiency wage models of the labor
market. Cambridge University Press, Cambridge

Alchain A, and Demsetz H. (1972) Production, information costs, and

economic organisation. American Economic Review, December, pp. 777-795.

Asap (1993) 1992 Rapporto sui salari, Franco Angeli, Milano.

Atkinson J. and Meager N. (1986) Changing working practices: how companies

achieve flexibility to meet new needs. National Economic Development Office,
London.

Bartels W. (1992) Gruppenarbeit 'beim Daimler- kein Modegag, sondern eine

neue Philosophie der Arbeit. Die Mitbestimmung, vol. 38, April pp. 50-52.

Batstone E. (1988) The reform of workplace industrial relations: theory, myth

and evidence. Oxford University Press, Oxford.

Bazen S, and Martin J. (1991) The impact of the minimum wage on earnings
and employment in France. OECD Economic Studies, No. 16, Spring, pp. 199-
221.

Bean C, Bernhloz P, Danthine J-P, Malinvaud E. (1990) European labour

markets: a long-run view. Centre for European Policy Studies, CEPS Paper no.
46, Brussels.

Becchi Collidà A, and Negrelli S. (1986) La transizione nell'industria e nelle

relazioni industriali: l'auto e il caso Fiat. Franco Angeli, Milano.

Belman D. (1992) Unions, the quality of labor relations, and firm performance.
In Mishel L, and Voos P. eds. pp. 41-107.

Bernoux P. (1972) Les OS face à l'organisation industrielle. Sociologie du

Travail, 72:4, pp. 410-36.

Bishop J. (1987) The recognition and reward of employee performance. Journal

of Labor Economics, 5:4, Part 2, Oct. pp. S36-S56.

62
Blanchflower D, and Freeman R. (1992) Unionism in the US and other
advanced OECD countries, in Bognanno M, and Kleiner M. Labor market
institutions and the future role of unions. Blackwell, Cambridge Mass.

Blinder A. S. ed. (1990) Paying for productivity. Brookings Institution,

Washington.

Blyth C. (1979) The interaction between collective bargaining and government

policies in some OECD member countries. in Collective bargaining and
government policies. OECD, Paris, pp.63-103.

Bosch G, and Lichte R, (1982) Die Funktionsweise informelle Senioritätsrechte

(am Beispiel eine betriebliche Fallstudie), in Dohse K, Jürgens U, and Russig
H. eds. Statussichering im Industriebetrieb: alternative Regelungssätze im
internationalen Vergleich, Campus Verlag, Frankfurt.

Bosch G, and Sengenberger W. (1989) Employment policy, the state and the
unions in the Federal Republic of Germany, in Rosenberg ed. The state and
the labor market, Plenum, New York.

Bosch G. (1990) From 40 to 35 hour. Reduction and flexibilisation of the

working week in the Federal Republic of Germany. International Labour
Review, 129:5, pp. 611-628.

Boulin J-Y, and Taddéi D. (1989) Les accords de réduction-réorganisation du

temps de travail: négociations et conséquences économiques. Travail et
Emploi, pp. 33-50.

Brown W, and Wadhwani S. (1990) The economic effects of industrial relations

legislation since 1979. National Institute Economic Review, No. 131, Feb, pp.
57-70.

Büchtemann C. (1993) Introduction: employment security and labor markets.

Büchtemann C. ed. pp. 3-66.

Büchtemann C. ed. (1993) Employment security and labor market behavior:

interdisciplinary approaches and international evidence. ILR Press, Ithaca.

Calmfors L, and Driffill J. (1988) Centralisation of wage bargaining. Economic

Policy: a European forum. No. 6, April, pp. 14-61.

Camerlynck and Lyon-Caen (1982) Droit du travail. Dalloz, Paris.

Carniti P. (1988) I salari in Italia negli anni ottanta. Rapporto della

Commissione Carniti. Marsilio Editori, Venezia.

63
Carniti P. (1988) I salari in Italia negli anni ottanta: Rapporto dell
Commissione Carniti. Marsilio Editori, Venezia.

CERC (1989) Les Français et leurs revenus au tournant des années 1980,
Documents du CERC no. 94, 3rd quarter, Documentation Française, Paris.

Coase R. H. (1937) The nature of the firm. Economica, November, pp. 386-405.

d'Iribarne Ph. (1989) La logique de l'honneur: gestion des entreprises et

traditions nationales. Seuil, Paris.

Dasgupta P. (1988) Trust as a commodity. In Gambetta D. ed. Trust: making

and breaking cooperative relations. Blackwell, Oxford.

de Luca P. (1987) Differenziali retributivi e slitamento salariale per qualifica

e livello: i dati delle indagini Federmeccanica e Assolombarda. In Asap 1986
Rapporto sui salari, pp. 201-238, Franco Angeli, Milano.

Delamotte Y. (1971) Les partenaires sociaux face aux problèmes de

producitivité et d'emploi. OECD, Paris.

Dell'Aringa C, and Lucifora C. (1990) Wage determination and union

behaviour in Italy: an efficiency wage interpretation. In Brunetta R, and
Dell'Aringa C. eds. Labour Relations and Economic Performance, Macmillan,
London.

Dell'Aringa C. (1992) Industrial relations and the role of the state in EC

countries. In, Marsden D. W. ed. Pay and employment in the new Europe,
Edward Elgar, Aldershot.

Dell'Aringa C. (1974) Egualitarismo e sindacato: l'evoluzione des differenziali

retributivi nell'industria italiana. Vita e Pensiero, Milano.

Della Rocca G. (1991) Produttività senza partecipazione nella diffusione degli

incentivi in Italia. In Della Rocca G, and Prosperetti L. eds. Salari e
produttività, Franco Angeli, Milano.

Dore R. (1987) Taking Japan seriously: a confucian perspective on leading

economic issues. Athlone, London.

Elbaum M. (1987) Qu'attendent les entreprises de la flexibilité de l'emploi?

Une comparaison France, Belgique, Royaume-Uni. Economie et Statistique, no.
197, March, pp. 41-52.

Eyraud F, and Silvestre J. J. (1992) La contrattazione del salario nel settore

privato in Francia. In Asap, 1991 Rapporto sui salari, pp. 85-120, Franco

64
Angeli, Milano

Eyraud F, Marsden D. W. and Silvestre J. J. (1990) Internal and occupational

labour markets in Britain and France. International Labour Review, vol 129,
no. 4, pp. 501-517.

Eyraud F. (1983) The principles of union action in the engineering industries

in Great Britain and France: towards a neo-institutionalist analysis of IR.
British Journal of Industrial Relations, Vol XXI, No. 3, Nov. 1983, pp.358-378.

Ferner A, and Hyman R. eds. (1992) Industrial relations in the new Europe.
Blackwell, Oxford.

FitzRoy F. (1992) Wage Structures, Employment Problems and Economic

Policy in Europe, in Marsden ed.

Flanagan R. J, Soskice D. W, Ulman L. (1983) Unionism, economic stabilization

and incomes policies: European experience. Brookings Institution, Washington
D.C.

Flanagan R. J. (1993) Hiring behavior and unemployment. In Büchtemann C.

ed.

Fox A. (1974) Beyond contract: work, power and trust relations. Faber,
London.

Freeman R. (1971) The market for college-trained manpower. Harvard

University Press, Cambridge, Mass.

Freeman R.B, and Medoff J.L. (1984) What do unions do? Basic Books, New
York.

Fürstenberg F. (1969) Workers' participation in management in the Federal

Republic of Germany. Bulletin of the International Institute for Labour Studies,
June, pp. 94-148.

Garonna P, and Ryan P. (1986) Youth labour, industrial relations and

deregulation in advanced economies. Economia e Lavoro, 20, pp. 3-19.

Giugni G. (1965) Recent developments in collective bargaining in Italy.

International Labour Review. Vol. 91, No.4, April, pp. 273-91.

Gouldner A. W. (1954) Patterns of industrial bureaucracy: a case study of

modern factory administration. Free Press, New York.

Gourevitch P, Martin A, Ross G, Bornstein P, Markovits A, and Allen C. (1984)

65
Unions and economic crisis: Britain, West Germany, and Sweden. George
Allen and Unwin, London.

Green R, and Macdonald D. (1991) The Australian flexibility paradox. Journal

of Industrial Relations, December, pp. 564-585.

Gregory M, and Thompson A, (1981) The coverage mark-up, bargaining

structure and the distribution of earnings in Britain 1973-78. British Journal of
Industrial Relations, 19:1, March, pp. 26-37.

Hakim C. (1990) Core and periphery in employers' workforce strategied:

evidence from the 1987 ELUS survey. Work, Employment and Society, 4:2,
June, pp. 157-188.

Hamermesh D. (1985) Substitution between different categories of labour,

relative wages and youth unemployment. OECD Economic Studies, No. 5,
Autumn, pp. 58-90.

Hamermesh D. (1986) The demand for labour in the long run. In Ashenfleter
O, and Layard P. R. G. eds. Handbook of labour economics. North Holland.

Hamermesh D. (1993) Labor demand. Princeton University Press, Princeton,

New Jersey.

Hibbs D. A. (1990) Wage compression under solidarity bargaining in Sweden.

Economic Research Report No. 30, Trade Union Institute for Economic
Research (FIEF), Stockholm.

Hicks J. R. (1955) The economic foundations of wages policy. Economic

Journal, vol. LXV, no. 259, December, pp. 389-404.

Hirsch B. (1992) Firm investment behavior and collective bargaining strategy.

In Bognanno and Kleiner eds.

Hirschman A. O. (1970) Exit, voice and loyalty: responses to decline in firms,

organisations and states. Harvard University Press, Cambridge Mass.

Huber B, and Lang K. (1993) Tarifreform 2000: Forderungskonzepte und

Verhandlungsstände im Bereich der Metallindustrie. WSI Mitteilungen,
12/1993, pp. 789-797.

Jones A. (1973) The new inflation. Penguin, Harmondsworth.

Jürgens U, Malsch T, and Dohse K. (1993) Breaking from taylorism: changing

forms of work in the automobile industry. Cambridge University Press,
Cambridge.

66
Katz H, Kochan T, Gobeille K.R. (1983) Industrial relations performance,
economic performance, and QWL programs: an interplant analysis. Industrial
and Labor Relations Review, Vol 37, No. 1, Oct. 1983, pp.. 3-17

Katz H. (1993) The decentralisation of collective bargaining: a literature review

and comparative analysis. Industrial and Labor Relations Review, 47:1, Octob
er, pp. 3-22.

Katz H. C. (1985) Shifting gears: changing labor relations in the US automobile

industry. MIT Press, Cambridge, Mass.

Katz L, Loveman G, and Blanchflower D. (1993) A comparison of changes in

the structure of wages in four OECD countries. Centre for Economic
Performance, Discussion Paper 144, London School of Economics.

Kaufman R. J. (1989) The effects of statutory minimum rates of pay on

employment in Great Britain. Economic Journal, 99: 398, December, pp. 1040-
1053.

Kjellberg A. (1992) Sweden: can the model survive? In Ferner A, and Hyman
R. eds. pp. 88-142.

Kochan T, Katz H. C, McKersie R. B. (1986) The transformation of American

industrial relations. Basic Books, New York.

Koike K, and Inoki T. eds. (1990) Skill formation in Japan and Southeast Asia.
University of Tokyo Press, Tokyo.

Lanfranchi N, and Afsa C. (1990) Le rôle normatif des conventions collectives

de branche. Travail et Emploi, 45:3

Lanfranchi N. (1988) L'application des classifications dans quelques branches

conventionnelles. Travail et Emploi, no. 36/37, June-September, pp. 71-83.

Lang K. and Leonard J. S. eds. (1987) Unemployment and the structure of

labor markets. Basil Blackwell, New York.

Layard R: Youth unemployment in Britain and the US compared., in Freeman

R. B, and Wise D. A. (1982) The youth labor market problem: its nature, causes
and consequences. NBER, University of Chicago Press, Chicago.

Leigh-Pemberton R. (1992) The first Bank of England Lecture given at the

London School of Economics, 12.11.92

MacInnes J. (1987) Thatcherism at work: industrial relations and economic

67
change. Open University Press, Milton Keynes.

Mackay D. I, Boddy D, Brack J, Diack J. A. and Jones N. (1971) Labour

markets under different employment conditions. George Allen and Unwin,
London.

Maguire M. ed. (1993) Pay flexibility in the public sector. OECD, Paris.

Marsden D. W, and Richardson R. (1994) Performing for pay? The effects of

'merit pay' on motivation in a public service. British Journal of Industrial
Relations, 32:2, pp. 243-262, June.

Marsden D. W, and Richardson R. (1994) Performing for pay? The effects of

'merit pay' on motivation in a public service. British Journal of Industrial
Relations, 32:2, pp. 243-262, June.

Marsden D. W. (1976) Aspects of company organisation and industrial

democracy: a discussion paper. Social Science Branch, Research and Planning
Division, Department of Employment, London. August. mimeo, 65 pages.

Marsden D. W. (1990) The flexibility of wages and changing economic

conditions. International Labour Office, Geneva.

Marsden D. W. ed. (1992) Pay and employment in the new Europe. Edward
Elgar, Cheltenham.

Marsden D. W.1981 Vive la Différence: pay differentials in the 1970s in Great

Britain, West Germany, France and Italy. Department of Employment Gazette,
July.

Martin A. (1984) Trade unions in Sweden: strategic responses to change and

crisis. In Gourevitch et al. (1984).

Maynard Smith J. (1982) Evolution and the theory of games. Cambridge

University Press, Cambridge.

McKersie in Ambrosini M, et al. (1990) Transforming U.S. industrial relations.

Franco Angeli, Milano.

Merrilees W. J, and Wilson R. A. (1979) Disequilibrium in the labour market

for young people in Great Britain. Manpower Research Group Discussion
Paper no. 10, University of Warwick.

Metcalf D. (1989) Water notes dry up: the impact of the Donovan reform
proposals and Thatcherism at work on labour productivity in British
manufacturing industry. British Journal of Industrial Relations, vol. 27 no. 1,

68
March, pp 1-32.

Metcalf D, and Milner S. eds. (1993) New perspectives on industrial disputes.

Routledge, London.

Metcalf D, and Stewart M. (1991) Closed shops and relative pay: institutional
arrangements or high density? Oxford Bulletin of Economics and Statistics,
52:2, pp. 503-516.

Metcalf D. (1990) Trade unions and economic performance: the British

evidence. In Brunetta R, and Dell'Aringa C. eds.

Metcalf D. (1993) Industrial relations and economic performance. British

Journal of Industrial Relations, 31:2, June, pp. 255-283.

Meyer (1991a) La contrattazione salariale in Germania. ASAP, 1991 Rapporto

sui salari, Franco Angeli, Milano.

Meyer W. (1991b) Übertarifliche Bezahlungen im Produzierenden Gewerbe:

Ergebnisse einer Befragung. Mimeo, Universität Hannover

Ministère du Travail (1992) L'individualisation des salaires en 1991. Premières

Informations, No. 291, 11.8.92, Service des Études et de la Statistique, Ministère
du Travail, de l'Emploi et de la Formation Professionnelle, Paris.

Ministère du Travail, de l'Emploi et de la Formation Professionnelle (1989)

Bilan 1988 de la négociation collective, Documentation Française, Paris.

Ministero del Lavoro (1987) Rapporto '87. Ministero del Lavoro e della
Previdenza Sociale, Roma.

Mishel L, and Voos P. eds. (1992) Unions and economic competitiveness. M.

E. Sharpe, New York.

Mottez B. (1966) Systèmes de salaire et politiques patronales: essai sur

l'évolution des pratiques et des idéologies patronales. Centre National de la
Recherche Scientifique, Paris.

Myrdal H-G. (1991) The hard way from a centralised to a decentralised

industrial relations system. Jacobi O, and Sadowski D. eds. Nomos Verlag,
Baden-Baden. Also SAF Doc. Nr. 159, Stockholm.

Negrelli S. (1991) Relazioni industriali e relazioni interne nelle piccole, nelle

medie e nelle grandi imprese. In Treu T. ed. Le attività: contrattazione
collettiva, consultazione/concertazione; contrattazione politica. Giappichelli,

69
Torino.

Nerb G. (1986) Problèmes de l'emploi: opinions des chefs d'entreprise et des

travailleurs. Economie Européenne, 27, March, European Commission,
Brussels.

Nickell S. (1990) Inflation and the UK labour market. Oxford Review of

Economic Policy 6:4, pp. 26-35.

Niland J. ed. (1986) Wage fixation in Australia. Allen and Unwin, Sydney.

Nohara H. (1992) Le syndicat d'entreprise et le micro-corporatisme au Japon:

acteurs, compromis et dynamique temporelle. Paper delivered at the
Université d'Eté, Université de Paris I, organised by B. Gazier.

Olson M. (1971) The logic of collective action: public goods and the theory of
groups. Harvard University Press, Cambridge Mass.

Organisation for Economic Cooperation and Development (OECD) (1986)

Flexibility in the labour market. OECD, Paris.

Paloheimo H. (1990) Between liberalism and corporatism. In Brunetta R, and

Dell'Aringa C. eds.

Pencavel J. (1994) British unemployment: letter from America. Economic

Journal, 104:424, May, pp. 621-632.

Plowman D. (1986) Developments in Australian wage determination 1953-83:

the institutional dimension. In Niland J. ed. (1986).

Pollert A. (1988) The 'flexible firm': fixation or fact? Work, Employment and
Society, vol.2, no. 3, pp. 281-316, September.

Pratten C. (1976) Labour productivity differentials within international

companies. Cambridge University Press, Cambridge.

Protocollo (1993) Protocollo sulla politica dei redditi e dell'occupazione, sugli

assetti contrattuali, sulle politiche del lavoro e sul sostegno al sistema
produttivo, Vita Italiana, Documenti e Informazioni, Roma.

Reynaud J-D. (1969) La convention sociale de la sidérurgue lorraine. Droit

Social, No. 4 April, pp. 219-227.

Richardson R, and Rubin M. (1993) The shorter working week in engineering:

surrender without sacrifice? In Metcalf D, and Milner S. eds.

70
Sainsaulieu R. (1965) Pouvoir et stratégie des groupes ouvriers dans l'atelier,
Sociologie du Travail, 65, pp.162-171.

Schager N-H. (1993) An overview and evalation of flexible pay policies in the
Swedish public sector. In Maguire ed.

Schudlich E. (1991) Traditionelle Leistungspolitik in der Sackgasse. WSI

Mitteilungen, No. 3 1991, pp. 181-188.

Silvestre J.J. ed. (1992) Les systèmes de rémunération dans le secteur public:
une comparaison France, Grande-Bretagne, Italie. ILO, Geneva.

Stewart M. (1987) Collective bargaining arrangements, closed shops, and

relative pay. Economic Journal, 97: pp. 140-156.

Stewart M. (1990) Union wage differentials, product market influences and the
division of rents. Economic Journal, 100: 403, December, pp. 1122-1137.

Stewart M. (1991) Union wage differentials in the face of changes in the

economic and legal environment. Economica, 58: pp. 155-72.

Stiglitz (1984) Theories of wage rigidity, paper presented to the Centre for
Labour Economics, London School of Economics, April.

Teschner E. (1977) Lohnpolitik im Betrieb: eine empirische Untersuchung in

der Metall-, Chemie-, Textil-, und Tabakindustrie. Campus Verlag, Frankfurt.

The Economist (9.7.94) Oriental renaissance: a survey of Japan.

Traxler F. (1993) The level and coverage of collective bargaining: a cross-

national study in patterns and trends. Mimeo, University of Vienna, 45p.

Visser J. (1990) Continuity and change in Dutch industrial relations, in Baglioni

G. and Crouch C. eds. European industrial relations: the challenge of
flexibility. Sage, London.

Wells W. (1983) The relative pay and employment of young people.

Department of Employment Research Paper 42, HMSO, London.

Windolf P. (1989) Productivity coalitions and the future of European

corporatism. Industrial Relations, 28:1, Winter, pp. 1-20.

Zabalza A, Turnbull P, and Williams G. (1979) The economics of teacher

supply. Cambridge University Press, Cambridge.

71