The Economics of Imperfect Labor Markets
Tito Boeri
September 2010
Tito Boeri and Jan van Ours (2008)
The Economics of Imperfect Labor Markets
Princeton University Press
Chapter 3. Unions and Collective Bargaining
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
1/1
Unions – What Are We Talking About?
Unions: What are we Talking About?
Unions typically bargain over all aspects of an employment
contract: wages, working hours, overtime pay, fringe benefits,
employment security, and health and safety standards.
Voluntary membership organizations: workers will only join a
union if it is profitable to do so.
First unions in the UK (18th Century) as craft organizations
providing mutual insurance to their members; later, in the 19th
century, industrial unions representing workers in semiskilled
positions; since the beginning of the 20th century national
organizations with political role.
Involved in collective bargaining with employers.
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
2/1
Unions – Measures and Cross Country Comparisons
Union density (% of all workers)
Denmark
France
Germany
Italy
Netherlands
Spain
UK
US
1960
57
20
35
25
42
–
40
31
1980
79
18
35
50
35
–
51
22
2000
74
8
25
35
23
17
30
13
2006
69
8
21
33
20
15
28
12
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
3/1
Unions – Measures and Cross Country Comparisons
Union membership in OECD countries
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
4/1
Unions – Measures and Cross Country Comparisons
Unions – presence and influence
Union density (only active members?)
Coverage of collective bargaining
Dichotomy between unions’ influence and presence: “excess
coverage”
Centralization of bargaining (formal level)
Coordination of unions (informal level – implicit)
Wage share
Strikes
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
5/1
Unions – Measures and Cross Country Comparisons
Measures – turn of the 20th century
Denmark
France
Germany
Italy
Netherlands
Spain
UK
US
Workers
in firm
joining
employer
association
(% ) (1)
48
74
72
40
79
72
54
-
Workers
covered by
collective
agreements
(% ) (2)
52
75
80
81
79
67
35
13
Workers
joining
trade unions
market
sector
(% ) (3)
68
10
25
36
19
16
19
10
Excess
coverage
(2)-(3)
-16
65
55
45
60
51
16
3
Central.
2
2
3
2
3
3
1
1
Coord.
4
2
4
4
4
3
1
1
Centralization and coordination: 1 to 5 (low to high)
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
6/1
Unions – Measures and Cross Country Comparisons
Strike activity – days lost per 1000 workers
Denmark
France
Germany
Italy
Netherlands
Spain
UK
US
1988–93
35
80
15
246
14
589
87
61
1993–99
258
92
4
103
24
253
20
39
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
7/1
Unions – Theory
Union membership & union bargaining
Union membership
Worker joins a union if the wage-employment combination with a
union membership is higher than without a union membership
Power of unions is higher with steep (inelastic) labor demand
curve: tradeoff wage – employment
Union bargaining
Most theories of union behavior take membership as given and
concentrate on collective bargaining.
The latter is modeled in three different ways:
Monopoly unions
Right-to-manage
Efficient bargaining
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
8/1
Unions – Theory
Monopoly union model
A union is the sole “seller” of labor
Union sets wages unilaterally maximizing the expected utility of a
representative worker (median member) subject to the labor
demand of the firm.
The firm reads off the employment level corresponding to w.
No bargaining takes place. Decision applies to all workers
(“closed shop”).
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
9/1
Unions – Theory
Monopoly union
Wages
Union indifference curves
w*
IC2
IC1
Labour demand curve
n* = n(w*)
N° of workers
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
10 / 1
Unions – Theory
Right-to-manage model
Union and firms bargain over any surplus.
Nash-bargaining: max of product of surplus of workers and firms
weighted by respective bargaining strengths (β and (1 − β)).
Gains as surplus over fall back option.
For the firm, the fall-back option is zero.
For the union member it is the reservation wage, w r .
Bargaining is on the wage
Employment = on the demand curve
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
11 / 1
Unions – Theory
“Right to manage” outcomes
w
β-1
wu
0<β<1
wm = w r
β-0
Ld
L
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
12 / 1
Unions – Theory
Efficient bargaining: over wage and employment
Firms: highest iso-profit curves
Unions: highest utility curve
Bargaining only over wage generates non-efficient outcome
Bargaining over wage and employment: efficient outcome
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
13 / 1
Unions – Theory
Efficient bargaining: Labor demand curve & isoprofit
curves
Wage
π2
π1
D
π0> π1 > π2
Peaks
π0
Labor demand curve
Employment
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
14 / 1
Unions – Theory
Efficient bargaining: Isoprofit curves & Union utility
curves
Wage
Both better off
M
iso-profit curve through
monopoly union point M
D
Employment
area: both unions and firm
are better off
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
15 / 1
Unions – Theory
Efficient bargaining – contract curve
w
Indifference curve of the unions
β=1
wu
A
w*
β=0
B
Isoprofit
L
d
L
BA = contract curve
BA vertical = strongly efficient contracts
BA upward sloping = “featherbedding” (overstaffing)
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
16 / 1
Unions – Theory
Strikes
A strike may occur if employers and unions do not reach an
agreement
Strikes are costly, they shrink the surplus over which bargaining
occurs
When perfect information it is irrational to strike
“Hicks paradox”: both parties would be better off without a strike
Strike due to imperfect information about firms’ financial situation
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
17 / 1
Unions – Theory
Endogenous union membership
Decision to join a union: depends on policies of unions.
Generally sponsor egalitarian wage policies: high-skilled workers
no incentive to join unions.
High wage demands: low-skilled workers are crowded out of their
jobs.
More successful in recruiting among medium-skilled workers.
Under excess coverage, free-rider problem: why should workers
pay union dues if they are covered in any event?
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
18 / 1
Unions – Theory
The free-rider problem
Membership decision: Cost-benefit analysis.
I join if costs of membership (fees, time) are smaller than benefits
(wages, security).
How unions solve their free rider problem:
Externalities – reputation for “good societal values”.
Provision of exclusive services to members: on-the-job training,
retirement and tax counseling.
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
19 / 1
Unions – Empirical Evidence
Effects of unions on wages
“Union wage gap”: estimated via regressions of wage equations of
the type
ln(wi ) = α + βm Mi + X 0 i γ
where
Mi is a dummy variable denoting membership of a trade union (it
takes value 1 when the individual is member of a trade union and 0
otherwise)
X is a vector of personal characteristics affecting wages, such as
age, educational attainment, and tenure on the job.
Denoting by w u and w n mean wages of union and non-union
members respectively, the estimated union wage gap is given by
the parameter βm
wu − wn
≈ ln(w u ) − ln(w n ) = βm
wn
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
20 / 1
Unions – Empirical Evidence
Results
Union wage gap (βm ) between 3 and 19% in the UK, 5 to 20% in
the US.
In countries with excess coverage, it is meaningless (no
counterfactual).
Problems also in countries with no excess coverage:
endogeneity: self-selection into unions in industries with high
surplus
measurement error: not easy to collect information on pay
spillovers: bargaining position of non-union members may improve
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
21 / 1
Unions – Empirical Evidence
Effects of de-unionization on US wage distribution
(box 3.4)
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
22 / 1
Unions – Empirical Evidence
Research on Unemployment and bargaining level
Macro empirical literature estimating employment and
unemployment equations
In countries with excess coverage problematic
The macro performance of an economy with both high bargaining
coordination and high unionization is, ceteris paribus, superior to
that of an economy with low coordination and low unionization
Lacking co-ordination, intermediate regimes offering the worst
performance (“hump-shaped“ relationship)
Serious measurement and endogeneity problems
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
23 / 1
Unions – Empirical Evidence
Real wage (unemployment)
A hump-shaped relationship
III
II
I
Low
Intermediate
High
Bargaining
Notes:
I. The effect of internalization of negative externalities
II. Hump-shaped relationship with small foreign trade
III. Hump-shaped relationship with large foreign trade
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
24 / 1
Unions – Policy Issues
Policy Issue: Do Unions Increase Efficiency?
The good and the bad face of unions.
Good face: “Exit–voice”: union give workers an option of voicing
problems, instead of exiting the firm when they are unhappy.
Bad face: Rent extraction – unions stronger in industries with no
product market competition.
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
25 / 1
Unions – Policy Issues
Unions and Product Market Competition
70
60
members as % of employees
50
40
30
20
10
0
FRANCE
GERMANY
Public Utilities
Public Administration
ITALY
SPAIN
Manufacturing
Source: fRDB survey
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
26 / 1
Unions – Policy Issues
Policy issue: Should bargaining be decentralized?
Trade-off between internalization of spillover effects (and
bargaining costs) and capacity to adapt to economic shocks.
High level of bargaining: internalization of spillover effects,
“right-to-manage” model – no “efficient” bargaining.
Low level of bargaining: “efficient” bargaining is possible –
performance-related pay
Also effects on workers incentives, motivations, hence productivity.
Problem of frequency of bargaining too: staggered contracts.
Can performance-related pay reduce frequency of bargaining?
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
27 / 1
Unions – Why Do Unions Exists?
Why do unions exist?
Because they are popular among some socioeconomic group.
The fast aging of the median union member in some countries
suggests that unions may be caught in a vicious circle of aging
membership and reduced attractiveness among the young and
active population. New firms start often without unions.
The share of retirees among union members is increasing
everywhere. This means that unions increasingly favor older
people in intergenerational conflicts, for example, in the design of
public pensions.
Unless unions solve this intergenerational problem, they may be
heading for the grave.
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
28 / 1
Unions – Review Questions
Review Questions
1
What are the pros and cons of the various measures of the
strength of labor unions provided by the literature?
2
Why are unions stronger in industries where there is less
competition in product markets?
3
Why is a right-to-manage bargaining system inefficient?
4
Why do unions pursue egalitarian wage policies?
5
How does competition affect efficient bargaining?
Source: Tito Boeri and Jan van Ours (2008), The Economics of Imperfect Labor Markets, Princeton University Press.
29 / 1