LABOR MARKETS: UNIONS, DISCRIMINATION, AND
DR. RICHARD GEARHART
LABOR SUPPLY AND DEMAND
Unions are a collection of workers that negotiate wages, fringe benefits (health
insurance, pension, …), and working conditions provided to them by the employee.
This process is known as collective bargaining.
The theory is that a group of workers has more power than each worker,
Before unionization, workers have no bargaining power.
They are such a small part of the market, they can only receive w*. This is a perfectly
competitive labor market.
Assumes that each worker is replaceable by other workers (i.e., this is a job that doesn’t require
significant education or skills).
With a union, the labor supply (LS) curve is upward sloping.
Or, the more workers there are in the union, the more power they have, and the higher the
wage that they can earn
For unions, do higher wages lead to higher productivity (more produced by each worker)?
There are some reasons why this may be the case:
Workers shirk less with higher wages (less time spent on FaceBook, on Reddit, …). For fear of losing
this job and going to a lower paying job.
Less employee turnover (fewer workers leave the job for other jobs). Reduces training costs (how
to use the computer system, …) for the employer. More experience at the job typically leads to
With fewer employees, firms higher more capital (machines and technology) to pair with the
workers. Capital is highly productive, which makes the workers paired with the capital highly
Why has union membership, in the United States, declined over time?
Increase in information. Because of salary websites like Glassdoor.com, perhaps individual
employees are better able to negotiate what they think a fair wage is.
Move away from manufacturing towards service. Manufacturing and resource extraction are
dangerous, time consuming, unhealthy jobs. Service sector jobs (fast food, retail) are not, so the
need for employee protection is less.
Politicians enacting laws (minimum wage laws, OSHA laws, overtime laws) have decreased the need
for these protections.
So, why is there such a large difference in union density between the U.S. and Europe?
How unions are formed.
In many other European countries, if workers want to unionize, they simply need to reach a
threshold to unionize (i.e., they need 60% of workers to agree to be unionized, and they will
have formed a union).
In the United States, if workers want to form a union, there is an election. Workers lobby to vote
“yes” for the union, while the firm lobbies to vote “no”. This may exert pressure.
How effective are modern day unions?
The union wage premium (how much more you earn because you join a union) is 5-8%.
Go from earning $100 a week to $108 a week.
Unions compress the earnings distribution.
Less wage differences between different workers (i.e., the wage difference between a CEO and a fry
Newer unions in the US have had no impact on employment, output, worker productivity, or
Much of the union wage premium is from unionized public sector workers.
LABOR MARKET DISCRIMINATION
Labor market discrimination is when different workers are paid differently, but not
based on education, skills, training, or length of experience on the job.
In other words, discrimination occurs when people are NOT paid based on workrelated factors.
Discrimination occurs when some trait, that is largely not a choice by the
individual, leads to lower pay (gender, race, sexual orientation, religion, …).
There are two types of discrimination.
Statistical discrimination: when you discriminate based on the population average.
Before you hire a worker, you do not know how productive they will be, or how good of a worker.
You therefore base your expectations on observable features of that worker.
While bad (this tends to prevent upward mobility by the “discriminated” groups), it is based on
How can I get an education, and raise the population average, if I cannot get the jobs that will pay
for the college education?
Example 1: why are women paid less than men?
The probability of a male getting pregnant is strictly 0.
The probability of a female getting pregnant may be 0, but could be positive.
At older ages, women are more likely to be in ill health than males (more hospitalizations, more
sick leaves, etc.).
Statistical discrimination would therefore say that, each year, the probability of a female
missing work is higher than the probability of a male missing work. Pay them less.
Example 2: in terms of a college education, non-Hispanic whites are more likely to have a
college education (29%) compared to both Hispanics (14%) and Blacks (17%).
Suppose that you hire a worker at age 18, and your benefits include a paid college education.
You would like this worker to achieve the college education by age 30.
Based on statistical discrimination, you would be less likely to hire the Hispanic or Black
individual, as they are less likely (ON AVERAGE) to achieve a college degree.
Biased Discrimination: discrimination that occurs because a particular individual does not
like some non-choice attribute of a worker.
I will not hire a female because I do not like women in the workforce.
I will not hire a Methodist because those religious beliefs are antithetical to mine.
THIS is the bad discrimination. This is not based on any belief OTHER than dislike/hatred.
Statistical discrimination is, at the very least, based on a set of objective facts.
With both types of discrimination, workers in the disadvantaged group will be hired ONLY if
they agree to be paid less.
Does discrimination pay in a competitive labor market.
Based on a seminal article by Gary Becker.
The only way that a discriminatory firm will hire workers they discriminate against is if they pay
the workers less.
HOWEVER, it is unlikely that all firms discriminate against these workers. They will therefore
earn higher wages elsewhere, and will leave.
Discriminatory firms therefore lose high productivity “minority” workers.
BUT, they lose more than this. “Minority” shoppers will not want to shop at this firm, so they
also lose business revenues.
A study of female lawyers found that 55% of the wage difference (between them and male
lawyers) was due to the fact that women had worked less (childbirth, childcare, …).
Childless women earn no more than women with children.
Thus, other 45% is largely biased discrimination.
“Are Emily and Brendan More Employable Than Lakisha and Jamal? …”
Sent out IDENTICAL resumes to firms. Only difference was the name.
White-sounding names received 1 call for every 10 resumes sent out. Black-sounding names
received 1 call for every 15 resumes.
White-sounding, high quality resumes, got more callbacks than white-sounding, low quality resumes.
Black-sounding, high quality resumes got no more callbacks than black-sounding, low quality
Found in EVERY industry (even for federal contractors and large employers).
We have both statistical discrimination and biased discrimination.