4. Anti-Discrimination Policies Flashcards
Discrimination
The valuation in the market place of personal characteristics of the worker that are unrelated to worker productivity
Measures and cross-country comparisons
All OECD countries have a legal and institutional framework to fight labour market discrimination
Whether these frameworks are effective depends on:
-how difficult it is for workers to take action
-how hard it is to prove discrimination
-how severely violations by employers are punished
Discrimination theories
Competitive labour markets: taste-based discrimination
1) by employers
2) by co-workers
3) by customers
Imperfect labour markets: three theories
1) Monopsony-based discrimination
2) Statistical discrimination
3) Occupational crowding
Taste-based discrimination - colleagues
Discrimination against women can arise from male co-workers
Model predictions:
- in the presence of women and prejudiced men, firms hire either men or women - the workforce will be segregated
- if employers are not prejudiced, there will be no gender wage gap even if all men are prejudiced
Taste-based discrimination - customers
Customers’ willingness to pay Pw for a product may depend on the presence of female workers in its production process
Model predictions:
Taste-based discrimination - customers
Customers’ willingness to pay Pw for a product may depend on the presence of female workers in its production process
Model predictions:
-since firms pay workers according to their marginal product - in an environment with customer discrimination - women will have a lower wage
-Firms will be segregated
-No within-firm discrimination as firms hire only men or only women - this is persistent in the LR
Comparing taste-based discrimination
SR:
-workforce will be segregated irrespective of whether the taste-based discrimination is related to employers, co-workers or customers
LR:
Employer’s discrimination -> competition drives out prejudiced employers
Co-worker’s discrimination -> does not involve a gender pay gap
Customer’s discrimination -> persistent in the LR
Imperfect labour market - monopsony
Robinson (1933) - in a monopsony, employers may have more monopsony power over women than they do over men
We assume:
-labour market for men is fully competitive
-labour market for women is monopsonistic
Imperfect labour market - statistical discrimination
Employers’ have imperfect information about the true productivity of individual workers -> individual test scores
Perceived productivity q of individual i in group j amounts to the weighted average of the individual test score and the group test score
Two types of statistical discrimination:
-different perception of group productivity but no difference in the weight attached to individual test scores
- individual test score of one group is valued more than the test score of one other group - even though average productivity is perceived to be the same
statistical discrimination - Occupational crowding
Discrimination that arises when some groups of workers are restricted in their entrance to certain occupations
Barriers to entry could come from unions, customs, or self-selection (‘sorting)
Pay gaps are occupation specific - not gender specific:
-within industries there is no earning differential between men and women
However women are more likely to work in lower paying industries due to occupational crowding
—-> on average, women earn less than men
Model can also explain wage discrimination if men working in ‘female jobs’ do not move to ‘male jobs’ in response to wage differential. This could happen because:
-preferences
-mobility costs
Carruthers and Wanamaker (2017)
Wage data from 10 Southern US states - taking into account differences in school quality, years of schooling, local characteristics
Pay gap is related to differences in school quality:
-school resources varied across races in first half of 20th century
- in the South in 1930 - annual spending per black student was $10, for white was $60
-with equal schooling quality, estimate is that inequality would have been reduced by 30% - 50%
Field experiments
Audit studies: fictitious candidates introduced among real ones, differing only for characteristics of interest
Issues with this approach:
-expensive
-difficult to obtain adequately diversified fictitious candidates
-fictitious candidates could exhibit different attitudes towards the interview - as they know they are not really competing for the job
Correspondence studies: introduction of fictitious CVs and cover letters - with the aim of studying call-back rates -> easier to tailor fake candidate’s on paper
Issues with this approach:
-call-back rates could be influenced by unobservable characteristics
-if real candidates know which firms discriminate more and avoid them - discrimination can be overestimated
Ban the Box (BTB) policies
Under these policies, employer’s do not ask about an applicant’s criminal history
Agan and Starr (2018)
Correspondence study: fictitious applications sent to entry-level jobs posted online on behalf of young males before and after BTB introduced
Result:
increased racial difference against black applicants in call-back rates after BTB
Goldin and Rouse (2000) - impact of blind auditions
Auditions at US orchestras: at different moments in time, “blind rounds” were introduced (candidates could audition, concealing their identity)
Question was whether auditioning in a blind audition increased odds of being hired
Result:
women experience 65% increase in probability of being hired (no difference for men)
-One third of the increase in the proportion of females among new hires can be attributed to the introduction of blind auditions
