Affirmative action refers to policies that try to reduce persistent under-representation of historically disadvantaged groups in employment, education, or contracting. Economically, the idea is that unequal opportunity, discrimination, and network effects can keep talented people out of positions they could fill productively.
Why Economists Study It
Affirmative action is usually discussed as a response to market and social frictions rather than as a stand-alone moral slogan. Economists focus on questions such as:
- whether discrimination distorts hiring or admissions
- whether unequal access to schooling, networks, or information keeps qualified candidates from applying
- whether representation today changes investment in skills and expectations tomorrow
That makes the issue dynamic. A policy can affect not only who is selected now, but also who prepares to compete in the future.
Common Policy Forms
Affirmative-action policies vary by jurisdiction, but common forms include:
- targeted outreach and recruitment
- broader evaluation criteria rather than narrow screening metrics alone
- contractor or institutional diversity requirements
- mentoring, bridge support, or retention programs that reduce attrition after selection
A Simple Representation Measure
One descriptive statistic compares a group’s share in outcomes with its share in the qualified or applicant pool:
[ \text{Representation ratio}_g = \frac{\text{share of hires or admissions in group } g}{\text{share of qualified applicants in group } g} ]
A ratio below 1 indicates under-representation relative to the pool. The next question is why that gap exists and which mechanism a policy is trying to change.
Main Trade-Offs
The debate is not only about fairness. It is also about mechanism and design.
A policy may improve matching if discrimination had been excluding qualified candidates. It may also change incentives for preparation, mentoring, and application behavior over time. But poorly designed rules can create legal conflict, stigma, or measurement problems if the target group or benchmark is not defined carefully.