Labor Market Discrimination

1
Labor Market Discrimination
2
Discrimination?

• It is found that of 32 new hires, only 3 are
  female.
• The Tiki Room only hires Asian female wait staff.
• Jim, who is a member of a minority group, gathers
  information about salaries of other persons in
  his firm with the same years of experience and
  education. He finds that their average salary is
  20 higher than his.
• Female doctors average lower incomes than is true
  for male doctors.
• Males pay higher insurance for automobiles than
  is paid by females when both are under 26.
• When males and females pay identical amounts into
  their pensions, males average less total pension
  payment..

3
Discrimination?

• On average, people who are considered beautiful
  are more likely to be hired and make more money
  on average than others.
• Only 14 of the American male population is
  taller than six feet. Fifty-eight percent of
  American CEOs are taller than six feet. On
  average males taller than 6 earn hundreds of
  thousands of dollars more over their lifetime
  than vertically challenged males.

4
Labor Market Discrimination

• Discriminate to make a difference in treatment
  or favor on a basis other than individual merit
• Discrimination
• the act, or practice or the instance of
  discriminating categorically rather than
  individually
• prejudiced or prejudicial outlook, action or
  treatment

5
Labor Market Discrimination

• Two separate ideas treatment vs prejudice
• Labor Market Discrimination focuses on TREATMENT
• How can we measure treatment?
• Result or process?
• Different outcomes
• Different opportunities to achieve the same
  outcome
• Timing When does discrimination begin or end?

6
Labor Market Discrimination

• Definition of LMD the valuation in the labor
  market of personal characteristics which are not
  related to productivity
• Emphasis Different outcomes in the labor market
  FOR OTHERWISE equal groups
• Differences in productivity are assumed to come
  from PRE-MARKET conditions and are NOT considered
  LMD

7
Labor Market Discrimination

• What does this definition of discrimination miss?
• Many differences in productivity are due to
  unequal access to things that improve individual
  productivity (education, experience,training).
• Expectations of labor market returns will
  influence individual investments that affect
  their productivity.

8
Legal Definitions of Discrimination

• Disparate Treatment Contains LMD treatment
  that might not lead to different economic
  results
• Disparate Impact The use of hiring criteria
  that lead to different results for protected
  groups even though they may not, on the surface,
  appear to be discriminatory

9
Statistical Discrimination

• Due to uncertainty, employers use characteristics
  of an individual to predict her productivity on
  the job
• education
• experience
• marital status
• gender
• race

10
Statistical Discrimination

• Even though the characteristic predicts
  accurately for the GROUP, it may not predict
  accurately for the individual in the group. This
  is especially true when we are considering
  historical characteristics of a group. E.g.
  womens turnover characteristics. When this
  occurs we say that the employee experiences
  statistical discrimination

11
Statistical Discrimination

• Statistical discrimination does not come from
  prejudice, but rather from a problem in
  predicting performance under uncertainty.
• Statistical discrimination typically will occur
  when different groups bring different
  characteristics to the labor market that make
  their PREDICTED performance different. The
  prediction can be right on average, but easily
  wrong for an individual.
• Profiling in its various forms is an example of
  statistical discrimination. When should it be
  banned?
• Driving while black?
• Terrorist identification?
• Insurance?

12
Statistical Discrimination Using Preemployment
Tests

• There are two other types of statistical
  discrimination that we should worry about
• biased tests
• unbiased, but inefficient tests
• Test bias A test under predicts for one group.
  Why?
• Test inefficiency A test predicts correctly on
  average, but does a much poorer job of predicting
  for one group as opposed to another.

13
Unequal test results ? Bias

• Many people believe that unequal test scores for
  different groups mean that the test is biased.
  This is a misunderstanding of what bias means.
• A test for bias must contain both test results
  AND performance results. A test would only be
  biased against a group if it under predicts the
  performance of that group
• Almost all tests that are now used to predict
  education performance and performance on the job
  have been shown to be unbiased.

14
Test Inefficiency

• Groups 1 2 have the same mean test score and
  performance.
• The test predicts more accurately for Group 1
• The ex post performance of Group 1 will be better
  than Group 2

15
Test Inefficiency

• Inefficient tests can lead to employees from one
  group appearing to be poor workers. This is due
  to the ability to pick good workers from the
  other group.
• Any characteristic that is used in hiring can act
  this way. Hobbies, dress, actions, etc.
• Since employers can choose better from one group,
  you will see that the probability of anyone from
  the other group being hired is low

16
Government Programs

• Civil Rights Act of 1964
• race,color,religion,sex,national origin
• age and disability added later
• EEOC established for class action suits
• not retroactive
• seniority systems are not discriminatory even
  though they yield discriminatory outcomes
• disparate treatment versus disparate impact

17
Government Programs

• Griggs vs Duke Power 1971
• Set the standard for plaintiffs burden showing
  disparate impact
• Requires an affirmative defense must show
  business necessity
• Comparable worth?
• Fed Contract Compliance Program 1965
• requires affirmative action program
• De facto quotas?