Labor Unions

Albrecht Glitz
Department of Economics, Universitat Pompeu Fabra

June 2023
What is a Labor Union?

• Labor union: organisation of workers whose objective is to

improve the pecuniary and non-pecuniary conditions of
employment among their members.

• Unions bargain with employers about various aspects of the

employment contract:
▶ Wages, hours of work, fringe benefits, labor turnover, job
satisfaction, promotion

• Bargaining can be more or less centralised.

• Unions also represent a political force in democratic countries.

• Union membership and union power can differ substantially.

A Brief Bit of History

• Labor unions had historically a significant impact on the labor

market.

• People have strong and opposing opinions about unions.

• Many sometimes violent conflicts during the 20th century.

• Generally speaking, unions are not as important today as they

used to be, especially in Anglo-Saxon countries.
Union Membership and Coverage

Trade Trade
Union Union
Union Union
Coverage Coverage
Density Density
1980 2016
Sweden 80% 82.5% 67% 90.0%
United Kingdom 51% 72.5% 24% 26.3%
United States 22% 26.0% 13% 12.0%
Italy 50% 82.5% 34% 80.0%
Japan 31% 27.5% 17% 16.7%
Spain 7% 62.5% 14% 73.1%
France 18% 82.5% 8% 98.5%
Union Membership in Spain

• Unions are legal only since 1978.

• Low union membership.

• Given the legal framework, there are no important advantages

for union members.

• Bargaining at several levels (national, by sector, by region, etc.).

• Political affiliations.

• Major unions:
▶ UGT (Unión General de Trabajadores)
▶ Workers’ Commissions (Comisiones Obreras, CCOO)
Union Membership in the United States

• Bargaining is very decentralised.

• Most workers are non-union and not covered by collective

agreements.

• Most empirical studies have used US data.

▶ Findings hard to generalise.

• But many countries are moving towards more decentralisation.

Union Membership in the United States, 1900-2016
Determinants of Union Membership (I)

• Benefits
▶ Negotiate compensation with the firm.
▶ Regulate arbitrary behaviour of employers.
▶ Influence undesirable working conditions.

• Costs
▶ Union dues.
▶ Time spent in union activities.
▶ Lost wages due to strikes.
▶ Constraints on individual freedom of workers.
Determinants of Union Membership (II)

• Demand for unionisation: workers are more likely to support

unionisation when:
▶ The union organiser can promise a high wage and a small
employment loss ⇒ this depends on the elasticity of the firm’s
labor demand.
▶ When costs for joining the union are low.

• Supply of unionisation depends on:

▶ Cost of organising the workforce
▶ Legal environment
▶ Management resistance to collective bargaining
▶ Firm’s excess rents
The Decision to Join a Union

The budget line is given by AT , and the worker maximizes utility at point P by working h∗
hours. The proposed union wage increase (from w ∗ to wU ) shifts the budget line to BT . If
the employer cuts back hours of work to h0 , the worker is worse off (utility falls from U to U0
units). If the employer cuts back hours only to h1 , the worker is better off.
Why Has Union Membership Declined?

• The structure of the labor market has been changing since the
1960s.

• Blue collar workers are less prevalent.

• Jobs have shifted to right-to-work states in the US.

• There has been a marked increase in labor force participation

rates of women.

• Workers’ demand for union jobs has declined.

• Firms have become more resistant to unions.

Unions’ Goals

• Advance the welfare of their members.

▶ Procedural objectives
• Workers’ participation in managing the workplace
• Job assignment, overtime, safety, worker discipline, etc.
▶ Increasing compensation
• Wages
• Pensions, health insurance, vacations etc.
Constraints

• Employers are on the other side of the table.

• Employers want to be able to operate successfully.

• Downward-sloping labor demand.

▶ Higher compensation implies a lower demand.

• Position and elasticity of the firm’s labor demand curve.

▶ Unions’ ability to raise their member’s wage will be strongest
in rapidly growing industries with inelastic labor demand.
Two Models of Union Behaviour

1. The Monopoly-Union Model (⇒ inefficient)

2. The Efficient-Contracts Model

The Monopoly-Union Model (I)

• The union wants to maximise its utility which depends on both

wages and employment.

• The union is dealing with a profit-maximising firm that cannot

influence the price of output.

• The firm’s downward-sloping labor demand curve can be viewed

as a constraint on union behaviour.
The Monopoly-Union Model (II)

• The union has an effective monopoly on the sale of labor to the

firm.
⇒ The union sets the price of its products (the wage) and the
firm looks at the demand curve and determines how many
workers to hire.

• Some workers will lose their job as a result of the union’s wage
demand.
⇒ It is not surprising that unions get more utility when labor
demand is inelastic.
The Behavior of Monopoly Unions

A monopoly union maximizes utility by choosing the point on the firm’s labor demand curve,
D, that is tangent to the union’s indifference curve. The union demands a wage of wM (up
from the competitive wage w ∗ ) and the employer cuts back employment to EM (down from
the competitive level E ∗ ). If the demand curve were inelastic (as in D ′ ), the union could
demand a higher wage, experience fewer employment cuts, and receive more utility.
The Monopoly-Union Model (III)

• Wage-employment solution is inefficient because unions reduce

the total value of labor’s contribution to national income.

• The efficiency loss as a fraction of national income can be

approximately computed as (see figure on next slide):
′
Efficiency loss 1
= × wU − wN
wN
× E1 E− E1 × E1 wN H
× National Income
National Income 2 1 H | {z }
| {z } | {z } |{z}
% union- % decline Fraction of Labor’s share
nonunion in employment labor force of national
wage gap in union sector that is unionized income

• The last worker hired in the non-union firms would have a

greater productivity if he had been hired in the union sector.
• The value of labor’s contribution to national income would
increase if some workers were reallocated across sectors.
Unions and Market Efficiency

In the absence of unions, the competitive wage is w ∗ and national income is given by the sum
of the areas ABCD and A′ BCD ′ . Unions increase the wage in sector 1 to wU . The displaced
workers move to sector 2, lowering the non-union wage to wN . National income is now given
by the sum of areas AEGD and A′ FGD ′ . The misallocation of labor reduces national income
by the area of the triangle EBF .
The Efficient-Contracts Model

• The monopoly-union model leads to inefficient outcomes.

• The firm and the union could make a deal that makes at least
one of them better off without making the other worse off.

• Concepts used: isoprofit curve and contract curve.

Isoprofit Curve

• Set of wage-employment combinations that give the firm the

same level of profits.

• It has an inverted-U shape.

• Lower isoprofit curves are associated with higher profits.

The Labor Demand Curve and the Firm’s Isoprofit Curve

If the wage is w0 , the firm maximizes profits (and earns $100, 000) by hiring 100 workers. If
the employer wants to hire 50 workers and maintain profits constant, it must reduce the wage.
Similarly, if the employer wants to hire 150 workers and maintain profits constant, it also must
reduce the wage. The isoprofit curve, therefore, has an inverse-U shape. Lower isoprofit curves
yield more profits.
The Contract Curve (I)

• There are many off-the-demand-curve wage-employment

combinations that are beneficial to both the union and the firm.

⇒ All points where the union’s indifference curves are tangent to

the firm’s isoprofit curve.

• These wage-employment combinations are pareto optimal.

⇒ Deviations from a particular deal on the curve can improve the
welfare of one party only at the expense of the other.

• Locus of efficient contracts.

Efficient Contracts and the Contract Curve

At the competitive wage w ∗ , the employer hires E ∗ workers. A monopoly union moves the
firm to point M, demanding a wage of wM . Both the union and firm are better off by moving
off the demand curve. At point R, the union is better off, and the firm is no worse off than at
point M. At point Q, the employer is better off, but the union is no worse off. If all bargaining
opportunities between the two parties are exhausted, the union and firm agree to a
wage-employment combination on the contract curve PZ .
The Contract Curve (II)

• The contract curve lies to the right of the demand curve.

⇒ For any given wage, an efficient contract leads to more
employment than would be observed with monopoly unionism.

• Therefore, the employer-union relationship is not characterised

by the union demanding a higher wage and the firm responding
by moving up the demand curve.
▶ Efficient contracts imply that unions and firms bargain over
both wages and employment.
▶ An upward sloping contract curve implies a degree of
overstaffing (relative to the employment level at a
non-unionised firm paying the competitive wage).
Strongly Efficient Contracts

• If the contract curve is vertical, the deal struck between the

union and the firm is strongly efficient because the unionized
firm is hiring the competitive level of employment.

• Under such a situation, the union captures some of the firm’s

fixed-size rents (since output and revenues are constant here).

• Wage-employment combinations on an upward sloping contract

curve are efficient in the sense of exhausting all bargaining
opportunities between the employer and the union; they are not
efficient in an allocative sense.

• Only under strongly efficient contracts do firms hire the “right”

number of workers, and there is no deadweight loss to the
national economy.
Strongly Efficient Contracts: A Vertical Contract Curve

If the contract curve PZ is vertical, the firm hires the same number of workers that it would
have hired in the absence of a union. The union and firm are then splitting a fixed-size pie as
they move up and down the contract curve. At point P, the employer keeps all the rents; at
point Z , the union gets all the rents. A contract on a vertical contract curve is called a
strongly efficient contract.
Evidence on Efficient Contracts
• The process of collective bargaining narrows down the
possibilities to a single point on the contract curve.
• Point chosen depends on the bargaining power of the two
parties involved.
▶ Economic conditions
▶ Legal environment
▶ Ability of unions to provide financial support to its members
• Empirical research tries to determine whether unions and firms
indeed reach an efficient contract or whether unions behave as
monopoly-unions.
▶ Estimating regressions that relate employment in union firms
to the union wage and to the competitive wage in the industry.
• Available studies seem to indicate that wage-employment
outcomes in unionised firms do not lie on labor demand curve.
Strikes

• A strike occurs when neither party is willing to give in when

negotiating.
▶ In 2019, there were only 27 work stoppages involving 1,000 or
more workers in the United States.
▶ In Spain, there were 726 strikes in 2018, involving an average
of 466 workers each.

• Because strikes are costly due to lost production time, strikes

shrink the amount of rents over which the parties negotiate.

• When parties have good information about the costs and likely
outcome of a strike, then it is irrational to strike.
▶ The fact that “irrational” strikes occur is known as the Hicks
Paradox.
The Hicks Paradox: Strikes Are Not Pareto Optimal

The firm makes the offer at point RF , keeping $75 and giving the union $25. The union wants
point RU , getting $75 for its members and giving the firm $25. The parties do not come to an
agreement and a strike occurs. The strike is costly, and the poststrike settlement occurs at
point S; each party keeps $40. Both parties could have agreed to a prestrike settlement at
point R ∗ , and both parties would have been better off.
Strikes and Asymmetric Information

• Strikes can be optimal if workers are not well informed about

the firm’s financial status (if there is asymmetric
information).

• Since the union will experience losses during a strike, it will

reduce its demands over the duration of a strike.

• A firm knows that the union will moderate its demands over
time.

• A firm incurs costs during a strike, so it will choose a strike

duration that maximizes the present value of profits.
The Optimal Duration of a Strike

Unions moderate (decrease) their wage demands the longer the strike lasts, generating a
downward-sloping union resistance curve. The employer chooses the point on the union
resistance curve that puts her on the lowest isoprofit curve (thus maximizing profits). This
occurs at point P; the strike lasts t periods and the poststrike settlement wage is wt .
Empirical Facts on Strike Activity

• Strikes are more likely to occur and last longer the higher the
wage demand.

• Unions tend not to make high wage demands during periods of

high unemployment.

• Strikes are more frequent when real wages are growing slowly or
during periods of inflation.

• Strikes are more likely when a firm has a more volatile stock
value.
Strike Activity in the United States, 1967-2010
Strike Activity in the United States, 1967-2010
Strike Activity in Spain, 2000-2018

1200 6,000,000

1000 5,000,000

800 4,000,000

600 3,000,000

400 2,000,000

200 1,000,000

0 0
2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018
Strikes Participants Days not Worked

Source: Ministerio de Trabajo y Economı́a Social

Union Wage Effects
• It is difficult to calculate the effect of unions on wages since we
often do not know what a worker would earn outside the union.
• Optimally, we would want to compute what a given worker i
would gain from working in a union job:

wUi − wNi
∆i =
wNi

and then compute the so-called union wage gain:

PN
∆i
Union wage gain = i=1
N
• Because wNi - the missing counterfactual - is not observed, we
instead tend to rely on the calculation of the union wage gap,
the % difference between average union and non-union wages.
The Union Wage Gap

wU − wN
• Union wage gap: D = .
wN
• The union wage gap, adjusted for differences in socioeconomic
characteristics, has fluctuated greatly over time, but since the
1970s it has hovered between 15 and 20 percent.

• In 2016, the average union wage gap was estimated to be 14.5

percent.

• The union wage gap overestimates the union wage gain,

because a typical worker in a union job will be more productive
than workers in a non-union job (due to skimming).
Wage Gap Between Union and Non-union Workers, 1920-2016
Threat and Spillover Effects

• Selection bias is not the only reason for why the union wage
gap is a potentially poor proxy for the union wage gain.

• The existence of a union sector has two side-effects on the

non-union sector.
▶ Threat effects involve non-union firms offering higher wages
to reduce incentives of workers to unionize (the threat effect
leads the union wage gain to be underestimated).
▶ Spillover effects result when workers unemployed in the union
sector enter the non-union sector, thus increasing the supply of
labor and decreasing non-union wages (the spillover effect
leads the union wage gain to be overestimated).
Unions and Wage Dispersion

• The dispersion of wages in the unionized sector is 25 percent

less than the dispersion of wages in the non-unionized sector.

• Unionized firms offer a lower payoff to education than

non-unionized firms.

• Unions flatten the age-earnings profile.

• Unions reduce sensitivity of their members’ wages to the

business cycle.
Unions and Fringe Benefits

• The fringe benefits package received by union workers (health

and life insurance, vacation and sick days, pensions, bonuses...)
is generally worth more than the package received by non-union
workers.

• The “union compensation gap” may be 2 or 3 percentage

points higher than the union wage gap.
Non-wage Effects of Unions

• Unions give workers an option of voicing problems through a

formal grievance procedure, instead of exiting the firm when
they are unhappy (the exit-voice hypothesis).

• This implies that worker turnover should be lower in unionized

firms.

• As labor turnover declines, firms’ productivity increases.

• Productivity rises but not enough to cover the increased labor

costs, leading to lower profits.
Stock Market Returns Before and After the 1999 Representation
Election at National Linen Service Corporation