Affirmative Action in Centralized College Admissions Systems -- by Nano Barahona, Cauê Dobbin, Sebastián Otero

TEXT START: We study the consequences of affirmative action in college admissions.

The Dissection

The paper measures how affirmative action reallocates scarce college seats in Brazil. It separates direct gains for targeted students from direct losses for displaced students, then traces the cascading reallocations through the centralized admissions system. Its central finding is that targeted students gain roughly twice as much in graduation outcomes as displaced students lose, but the policy displaces about 30% more students than it directly benefits. The aggregate result is therefore a modest net gain built on substantial redistribution.

The Core Fallacy

The paper’s weakness is not primarily econometric. It is temporal and structural. It treats program quality, graduation, and educational matching as durable proxies for productive viability. Under the Discontinuity Thesis, that assumption is unstable: P1 makes cognitive labor increasingly automatable, P2 prevents institutions from preserving large-scale human-only economic domains, and P3 destroys mass access to economically necessary work.

The study optimizes the allocation of seats on the labor ladder while ignoring whether the ladder still leads to durable employment, wages, and consumption. Its “net gain” may be real inside the admissions model and nearly irrelevant to the terminal economic system.

Hidden Assumptions

• Higher-quality programs will continue to generate materially better economic outcomes.
• Graduation remains a meaningful route to productive participation rather than a credential in an oversupplied labor market.
• Human cognitive skill will remain complementary to AI rather than substituted by it.
• Graduation gains and losses are adequate measures of welfare, excluding ownership and control of productive AI capital.
• The admissions system operates within a stable labor market rather than a collapsing wage-consumption circuit.
• Redistribution among students can produce systemic improvement without changing who owns the machines performing the valuable work.

These assumptions convert a question of structural power into a question of seat allocation.

Social Function

Classification: partial truth and transition management.

The paper reveals a genuine mechanism: centralized affirmative action generates secondary displacement, and the political cost is larger than a simple beneficiary-versus-loser comparison shows. That is useful accounting. But its practical function is narrower: it helps administrators ration access to increasingly scarce educational credentials while preserving the fiction that improved allocation within the credential system is equivalent to preserving broad economic opportunity.

It is a map of who gets a better position in the queue, not an escape from the queue.

The Verdict

This is a competent analysis of redistribution inside a shrinking human-capital pipeline. It demonstrates that affirmative action can improve targeted students’ measured outcomes while imposing broader displacement costs. It does not address the decisive question under the Discontinuity Thesis: who controls the AI capital that makes most cognitive labor unnecessary?

Without ownership, control, or indispensable operational leverage, a better degree is only a higher-grade claim on a disappearing wage system. The paper improves the rationing mechanism. It does not prevent the system from dying.