Labour Market Discrimination: Mcgraw-Hill/Irwin Labor Economics, 5 Edition

Download as pdf or txt
Download as pdf or txt
You are on page 1of 18

10/7/2009

Chapter 9

Labour Market
Discrimination

McGraw-Hill/Irwin
Labor Economics, 5th edition Copyright © 2010 The McGraw-Hill Companies, Inc. All rights reserved.

10 - 2

Introduction

• Labour
abou Market
a et Discrimination
sc at o iss another
a ot e possible
poss b e reason
easo that
t at
might explain wage dispersion.
• Discrimination occurs when participants in the marketplace
(e.g. employers, employees, customers) take into account such
factors as race and sex when making economic exchanges.

• More elaborate definition:


Discrimination occurs when there are different earnings and
employment opportunities across equally skilled workers
employed in the same job because of workers’ race, gender,
national origin, sexual orientation, age, religion, ‘beauty’ etc.

1
10/7/2009

10 - 3

9-1 Race and Gender in the Labour Market

Some US & international data on ggender and racial differences in


skills and labour market outcomes:
- Men earn more than women and whites earn more than nonwhites.
- But differences in educational attainment between whites and nonwhites
and differences in the proportion of part-time versus full-time work clearly
generate earnings differentials and explain some of the observed
differences.

For additional NZ data (some were shown for earlier topics) see, for
example:
- Department of Labour (2009), Labour Market Statistics 2008, Wellington.
- Statistics New Zealand (2003), Human Capital Statistics, Wellington.

10 - 4

9-2 The Discrimination Coefficient

Gary Becker’s neoclassical ‘taste discrimination model’!


• Taste discrimination – a concept that essentially translates the notion of racial
prejudice into the language of economics.
- Assume two types of workers earning competitive hourly wages: Black
workers earn wage wB and white workers earn wage wW.
• Instead of black or white workers, you could substitute ‘male or female’, ‘white
or Asian’, ‘catholic or protestant’, ‘gay or straight’, ‘beautiful or ugly’, and
many others!

- Assume the employer is prejudiced against black workers and this prejudice
enters his/her utility function as disutility. Even though it costs wB dollars to
hire one person-hour of black labour, the employer will act as if it costs
wB(1+d) dollars, where d is positive and called the discrimination
coefficient.
- Racial prejudice causes employers to blindly perceive the costs of hiring
blacks as higher than their true costs.

2
10/7/2009

10 - 5

Taste discrimination ctd.

• There can also be a kind of ‘reverse discrimination’: Nepotism.


- Ass
Assumeme some employers
emplo ers prefer to hire black workers.
orkers
- The employers utility-adjusted cost of hiring a black worker is
then less than otherwise: wB(1-n) dollars; where n is a positive
number called the nepotism coefficient.
- In this case racial prejudice causes employers to blindly perceive
the costs of hiring blacks as lower than their true costs.

B k ’ model
Becker’s d l can be
b applied
li d in
i many different
diff circumstances.
i
- The discrimination coefficient ‘monetises’ prejudice, regardless of
whether the source of the prejudice is the employer (employer
discrimination), the employee (employee discrimination), or the
customer (customer discrimination).
- In a sense, the labour market has to generate a compensating differential
to compensate prejudiced persons for their utility loss or gain!

10 - 6

9-3 Employer Discrimination

• Assume q = f (EW + EB). Other inputs than labour are


neglected.
neglected
• Both types of workers are assumed to be perfect substitutes and
equally productive. The last worker hired (of whichever type) has
MPE.
• If the employer does not discriminate, s/he will higher whichever
type of worker is cheaper!
- For example, if wB < wW , the firm will hire only black
workers (up to the point where wB = VMPE). See Figure 9-1.
• In short, if blacks and whites are perfect substitutes and their
wages differ, employers have a segregated work force (an
implication of the Becker model).

3
10/7/2009

10 - 7

Figure 9-1: The Employment Decision of a


Firm That Does Not Discriminate
Dollars

If the
h market-determined
k d i d
black wage is less than the
white wage, a firm that does
not discriminate will hire
only blacks. It hires black
wB workers up to the point
where the black wage
equals the value of marginal
VMPE
product of labour, or E*B.
EB Employment

10 - 8

Employer discrimination ctd.

If the employer does discriminate:


• Assume the employer acts as if wB equaled wB(1+d).
• Decision rule in this case:
- Hire only blacks if wB(1+d) < wW
- Hire only whites if wB(1+d) > wW
• In this case, too, employers will have a segregated work force!
We end up with ‘allall white
white’ and ‘all
all black
black’ firms.
firms
- In our case with wB < wW, non-prejudiced employers hire
only blacks, very prejudiced employers hire only whites.
See Figure 9-2.
• The prejudiced employer will hire too few workers (whites or
blacks)!

4
10/7/2009

10 - 9

Figure 9-2: The Employment Decision of a


Prejudiced Firm
Dollars Dollars

wW
wB(1+d1)

wB(1+d0)

wB

VMPE VMPE

EW E l
Employment
t E0B E* Employment

(a) White Firm (b) Black Firm


Firms that discriminate can be either white firms (if the discrimination coefficient is very
high) or black firms (if the discrimination coefficient is relatively low). A white firm hires
white workers up to the point where the white wage equals the value of marginal product.
A black firm hires black workers up to the point where the utility-adjusted black wage
equals the value of marginal product. Firms that discriminate hire fewer workers than
firms that do not discriminate.

10 - 10

Employer discrimination and profits

• Important insight of Becker’s model: Discrimination does not pay.


- By hiring the wrong colour workers and/or the wrong number of
workers, a prejudiced employer pays a wage above the competitive
wage and does not hire enough workers, resulting in lower output
and profits than without discrimination! See Figures 9-2 & 9-3:
• Highest profit where d = 0 and firm hires E*B black workers.
• Small d: Still ‘all black’ firm, but fewer workers hired & lower profit.
• High
g d: ‘All white’ firm,, veryy low pprofit.

• Far-reaching implication: In competitive markets, discriminating


employers should eventually disappear or abandon discrimination
(the zero economic profit condition of competitive markets implies
that discriminating firms cannot generally persist in the long-run).

5
10/7/2009

10 - 11

Figure 9-3: Profits and the Discrimination


Coefficient
Dollars
Discrimination reduces
profits in two ways. Even
πMAX
if the discriminatory firm
hires only black workers,
it hires too few workers.
If the discriminatory firm
πW hires onlyy white workers,
Black
Firms
White
Firms
it hires too few workers at
a very high wage.
0 dW Discrimination
Coefficient

10 - 12

The Black-White Wage Ratio in the


Labour Market
• Use decision rule of discriminating employer to derive market demand curve for
black workers.
• Assume first that all employers discriminate (all d are positive).
• Think in terms of the black/white wage ratio: (wB/wW)
• Compare black wage with white wage (Figure 9-4)(Note: These are the actual,
not the utility-adjusted wages). Assume the black-white wage ratio falls further
and further: At first no black workers get hired etc. As it falls below a certain
level (R), more and more get hired etc.

• Some features of the equilibrium in Figure 9-4:


- It usually
ll occurs bbelow
l th the point
i t where
h theth wage ratio
ti equals
l 1,
1 i.e.
i
employer discrimination generates a wage gap between equally skilled
black and white workers. In a sense, black workers must ‘compensate’
prejudiced employers!
- Black workers are matched with employers who have the least prejudice.
- Note that if some firms discriminate in favour of black workers (nepotistic
firms), and the supply of black workers is small, the wage ratio could end
up greater than 1, even if most firms discriminate against black workers!

6
10/7/2009

10 - 13

Figure 9-4: Determination of Black/White


Wage Ratio in the Labour Market

Black-White
Wage Ratio
If the black-white wage ratio is very high,
no firm in the labour market will want to
S
hire blacks. As the black-white wage ratio
falls, more and more firms are
compensated for their disutility and the
(wB/wW)′ demand for black workers rises. The
equilibrium black-white wage ratio is
1
given by the intersection of supply and
R D′
demand, and equals (wB/wW)*. If some
firms prefer to hire blacks, they would be
willing to hire blacks even if the black-
black
white wage ratio exceeds 1, shifting the
(wB/wW)* demand curve up to D′. If the supply of
D blacks is sufficiently small, it is then
0 N Black
possible for the black-white wage ratio to
Employment exceed 1. Minority workers may benefit
from ‘enclave economies’.

10 - 14

Employer discrimination ctd.

• Some criticisms/shortcomings of the employer taste model of


di i i ti
discrimination:
- Most large firms have divorced ownership from control, i.e. the
model assigns discriminatory tastes to personnel department
practices rather than individual bosses.
- The model does not explain where discrimination comes from!
Only some of the symptoms, but not the root causes, of
discrimination are addressed.
- It addresses wage discrimination, but employment discrimination
may be the more important result of prejudice!
• Employers offer a structured hierarchy of jobs and put some types
of workers into a low level with low pay, despite them being as
productive as other workers in higher levels of the hierarchy.

7
10/7/2009

10 - 15

9-4 Employee Discrimination

• Assume: Whites do not like working with blacks (but blacks do


nott mind
i d working
ki withith whites);
hit ) whites
hit andd blacks
bl k are perfect
f t
substitutes in production.
• Whites act as if their wW is equal to wW(1-d).
• Assume a non-discriminatory (i.e. colour-blind) profit-maximising
employer: The firm will NOT pay a compensating wage
differential to whites. The firm will hire the cheapest workers
((either blacks or whites)) and ppayy them their MP!
- The firm will employ either only blacks or only whites (completely
segregated workforce).
- Employee discrimination does not generate a wage differential between
equally skilled black and white workers.
- Employee discrimination does not affect the firm’s profitability. Market
forces will NOT diminish this type of discrimination over time!

10 - 16

9-5 Customer Discrimination

• If ‘majority’ customers are prejudiced (i.e. discriminate against) ‘minority’


workers,
k th
their
i perceived
i d price
i off a goodd sold
ld by
b minority
i it workersk isi utility-
tilit
adjusted with the discrimination coefficient. They act as if p equaled p(1+d)
and buy less of the good.

• Firms might get around this by putting ‘majority’ workers in customer


contact jobs. Competition for ‘minority’ workers might then result in equal
wages for equally productive majority and minority workers.

• When a firm cannot hide minority workers, customer discrimination can


have an adverse impact on their wages: The firm has to charge a lower price
for the good to compensate the prejudiced customers for their disutility. The
minority workers have to compensate the firm for this by accepting a lower
wage.

8
10/7/2009

10 - 17

9-6 Statistical Discrimination

• This form of discrimination is based on treating an individual on


th bbasis
the i off membership
b hi in i a group andd knowledge
k l d off that
th t group’s

average characteristics.
• An individual is not seen as an individual, but as an ‘average
person’ of the group s/he belongs to.
• It arises because an employer cannot perfectly predict the
worker’ss true productivity,
worker productivity and therefore also uses ‘group
group infor
infor-
mation’.
- Example: Equally qualified etc. men and women apply for a job, but
employer’s experience suggest many women leave work for child-bearing.
Employer wants the worker to stay long-term and hires the male.
- Many examples outside the labour market, e.g. in the insurance market.

10 - 18

Statistical discrimination ctd.

• A simple model of the impact of statistical discrimination on


wages:
- White and black applicants have their individual test scores T.
- If the Ts are perfectly correlated with people’s productivity, people
would get offered a wage equal to their MPs.
- Assume (realistically) that the Ts are imperfect. In that case, the
employer also wants to take into account the average T (e.g. Ť) for the
group the worker belongs to so that the expected productivity is a
weighted
i h d average off bboth:
h

w = α T + (1 - α ) Ť (9-4)

- α measures the correlation between the individual test score and true
productivity. Extreme cases: α = 1 or α = 0.
- Figure 9-5.

9
10/7/2009

10 - 19

Figure 9-5: The Impact of Statistical


Discrimination on Wages
Dollars Dollars

White

Black

T* Test Score T− Test Score


(a) Whites have higher average score (b) Both have same mean score, but Test is
better predictor for white workers

The worker’s wage depends not only on his own test score, but also on the mean test score of workers
in his racial group. (a) If black workers, on average, score lower than white workers, a white worker
who gets T* points earns more than a black worker with the same score. (b) If both groups have the
same mean test score but the test is a better predictor of productivity for white workers, high-scoring
whites earn more than high-scoring blacks, and low-scoring whites earn less than low-scoring blacks.

10 - 20

Statistical discrimination ctd.

• Figure 9-5b:
- The black worker’s wage is mostly set on the basis of the group
average, the white worker’s wage is mostly set on the basis of
his/her own test score.
- Blacks that scored low on their individual test actually benefit
from statistical discrimination (relative to high scoring blacks)!

• Should the employer use group averages ((‘profiling’)?


profiling )?
- Policy issue: Should there be ‘race or gender norming’ of test
scores?

10
10/7/2009

10 - 21

9-7 Experimental Evidence on Discrimination

• Very interesting material.


material Students read if interested!

• How to measure discrimination when it is illegal?

• Review of some studies using labour market experiments to test


whether discrimination does exist.
- Mail outs of fake resumes to test for discrimination because of a
person’ss name (names often reveal ethnic background).
person background)
- Fake job applicants going to job interviews etc.

10 - 22

9-8 Measuring Discrimination

• One possible definition of discrimination: The difference in


mean wages between
b t various
i groups off people
l (the
(th ‘raw’
‘ ’ wage
differential). But this measure would likely be incorrect!
• More appropriate to compare the wages of equally skilled
workers: Adjust the ‘raw’ wage differential for skill differences.
• Economists try to do this by estimating earnings regressions.
- Assume two g
groups
p of workers (e.g.
( g males and females)) and
decompose the raw wage differential (see equation 9-7).
• Oaxaca Decomposition – a decomposition of the raw wage
differential into a portion related to skill differences and a
portion related to labour market discrimination (see eq. 9-8).
- The intuition of this decomposition is shown in Figure 9-6.

11
10/7/2009

10 - 23

Figure 9-6: Measuring the Impact of


Discrimination on the Wage
Dollars

Men s Earnings
Men’s The average woman has s-F
Function
years of schooling and earns
w-F dollars. The average man
w −M
w*F has s-M years of schooling and
earns w-M dollars. Part of the
w*F
Women’s Earnings wage differential arises
αM
Function because men have more
w −F schooling than women. If the
αF average woman was paid as if
she were a man, she would
earn w*F dollars. A measure of
s− F s− M Schooling
discrimination is then given by
(w*F − w-F).

10 - 24

Measuring discrimination ctd.

• Does the
h OOaxaca Decomposition
i i really
ll measure
discrimination?
- Depends on whether all dimensions of the skill differences between
groups have been measured and controlled for.
- That is unlikely (e.g. quality of education, motivation, drive?). The
results of the decomposition can therefore always be criticized.
• This is an easy argument for people who do not think that there is
discrimination!
- But why do quality of education, motivation & drive differ? If they
differ due to discrimination (not by the employer, but by ‘society’),
they should NOT be controlled for in the decomposition!

12
10/7/2009

10 - 25

9-11 Policy Application: Determinants of the


Female-Male Wage Ratio

• Table 9-6: “Discrimination in the Oaxaca sense” accounts for


most of the wage gap between men and women.
women But this does
not take into account that on average, women spend a smaller
proportion of their potential work years in work.
• The Mincer-Polachek hypothesis (see footnote 39):
1. Human capital (HC) is more profitable the longer the payoff period.
HC acquired by men has longer pay-off period. Women take time-off
(e.g. because of children). Women have shorter HC pay-off periods
q
and therefore acquire less HC.
2. Women’s HC depreciates during their breaks from working life.

• Is this HC story correct? Problem of causation: Does women’s


lower work attachment lead to less HC acquisition and lower
wages, OR do lower wages for women (possibly due to
discrimination) lead to weaker work attachment?

10 - 26

Determinants of the Male-Female


Wage Ratio ctd.

• Occupational crowding: The occupational segregation between


men and women in the labour market (see Table 9-7).

• The occupational crowding hypothesis is a discrimination-


based explanation of the segregation of women into particular
occupations where the return to education is lower (i.e. ‘female
jobs’ tend to pay lower wages).
The hypothesis assumes that the observed segregation of men and
women into particular types of jobs is intentional (due to ‘society’
or ‘social climate’, not necessarily due to male employers!)

13
10/7/2009

10 - 27

Determinants of the Male-Female


Wage Ratio ctd.

• But the human capital model provides an alternative


explanation
l ti off occupational
ti l crowding:
di Women
W are better
b tt offff if
they enter occupations in which their skills do not deteriorate
during the years they spend in the household sector. This
maximizes their lifetime earnings!
• Affirmative action programs (p. 405).
• Comparable
Co p b e wo worth pprograms
og s – these
ese bbreak
e thee link between
be wee
wages and labour market conditions. Demand & supply effects
no longer apply. Jobs are evaluated and rated (‘job scores’), e.g.
‘mechanic’ (men’s job) = ‘receptionist’ (women’s job).
Can have large impact on female-male wage ratio at the cost of
economic efficiency.

10 - 28

9-12 The Economics of Specialisation:


Household Production

• The neoclassical labour


labour-leisure
leisure choice model is very flexible
and adaptable! It can be used to analyze household production
by including many forms of nonmarket work, including work
around the home, under the category ‘leisure’.

• Why do some household members specialize in the market


sector and other members specialize in the household sector?
- Allocation of time to various activities (paid work, household
work, leisure time etc.) varies by gender and marital status.
• For example, women usually allocate many more hours to the
nonmarket sector than men. Classic example: difference in time
spent doing housework.

14
10/7/2009

10 - 29

Household Production

• Consider Jack and Jill, a married couple (a two-person


household).
• Let’s adapt our diagram: Assume they want to maximize their
utility by consuming both goods bought in the market (market
goods) and goods produced in the household (HH goods).
• Assume they each have a total of 10 hours per day for paid and
non-paid
p work. How will theyy allocate their time between the
two activities?
• “The household production function tells us how much
household output Jack and Jill can generate for any given
allocation of time.”

10 - 30

Household Production

• The textbook makes the somewhat sexist assumption that Jill is


more productive in terms of household production, i.e. she is
assumed to produce more $ worth of output in household
production per hour than Jack ($25 versus $10 worth of HH
output per hour).
• Assume Jack’s (market) wage rate is $20 per hour, Jill’s is $15
per hour.
• We can draw their individual and combined budget lines
(Figure 9-12).
• Their household opportunity set is greater than when Jack and
Jill were not married (since each can now specialize in the
sector where they are relatively more productive).

15
10/7/2009

10 - 31

Figure 9-12a&b: Jack and Jill –


Unmarried Opportunity Sets
Market Market
Goods ($) Goods ($)

200

150

100 Household 250 Household


Goods ($) Goods ($)

10 - 32

Figure 9-12c: Budget Lines and


Opportunity Frontier of Married Couple
At point E, Jack and Jill allocate all
their time to the household sector. If
350 G they wish to buy market goods, Jack
gets a job because he is relatively
more productive in the labour
market, generating segment FE of
F
200 the opportunity frontier. After he
Jack’s uses up all his time in the labour
150
market Jill then gets a job,
market, job
Jill’s
generating segment GF of the
E′ frontier.
E
100 250 340 350 Household
Goods ($)

16
10/7/2009

10 - 33

Figure 9-13: Which Point on the Opportunity Frontier will


the Household choose, or Who Works Where?

Market Market Market


Goods ($) Goods ($) Goods ($)

P
U P

U
P

U
Household Household Household
Goods ($) Goods ($) Goods ($)

(a) (b) (c)

10 - 34

Division of Labour in the Household

• The Household will choose the point that gets it on the highest
indifference curve. Three possible solutions:

- (a) The indifference curve U is tangent to the opportunity frontier


at point P. Jill specializes in the household sector and Jack
divides his time between the labour market and the household.

- (b) Jack specializes in the labour market and Jill divides her time
between the labour market and the household.

- (c) Jack specializes in the labour market and Jill specializes in the
household sector.

17
10/7/2009

10 - 35

Division of Labour in the Household


• “Higher wage rates create incentives for specializing in the
market sector!”
• “Greater aptitude for producing output in the household sector
creates incentives for specializing in that sector!”
- See Borjas, p. 411, Figure 9-14 (not shown in class).
• But: Changes in developed countries in recent decades:
- Increase in women’s real wage means that it is more common that
women earn as much or more than men (although on average they still
earn less).
)
- Plus: Technological changes in household production imply that
productivity differences between men and women have declined.
• Result: Tendency for more men to stay at home and specialize
in HH production.
End of Chapter 9

18

You might also like