In 1970, women were less than 5 percent of the players in the top five US symphony orchestras. Over the next two decades, most orchestras began auditioning candidates behind a screen, so the jury heard the playing without seeing the player. By the late 1990s, women made up about 25 percent. Two economists later set out to measure how much of that rise the screen explained.
The veil of ignorance is John Rawls’s thought experiment for testing fairness: choose the rules of a society, or any group, without knowing which position you will hold in it. Because you could end up anywhere, you have a reason to pick rules you could accept from every seat, including the worst one.
A screen did what the veil asks you to imagine
Claudia Goldin and Cecilia Rouse studied real audition records. In their NBER working paper, they found that the screen increased by 50 percent the probability that a woman advanced out of certain preliminary rounds. It also raised, “by severalfold,” the likelihood that a woman won the final round. The switch to blind auditions explained between 30 and 55 percent of the increase in women among new hires.
The published version in the American Economic Review is careful. “Some of our estimates have large standard errors and there is one persistent effect in the opposite direction,” the authors wrote, but “the weight of the evidence suggests that the blind audition procedure fostered impartiality in hiring.”
That caveat matters, and so does the mechanism. The jury did not become more virtuous. It lost the information that let bias operate. Double-blind peer review and blind résumé screening use the same move.
Rawls built the veil to take self-interest off the table
Rawls introduced the veil in his 1971 book A Theory of Justice. He built on social contract thinkers such as Locke, Hobbes, and Rousseau, who argued that legitimate rules come from agreement. Rawls added a condition to make the agreement fair: the people choosing must not know their own place in the result.
He called this setup the original position. As the Claremont Review of Books summarizes, the parties know nothing of their personal characteristics, natural abilities, social status, or idea of the good life. They do know general facts about economics and psychology. Rawls treated it as a model for testing principles, never as a description of how real people decide.
Rawls argued that people in that position would avoid gambling with their basic life prospects. They would choose:
- Equal basic liberties for everyone.
- Fair equality of opportunity for positions and offices.
- The difference principle: inequalities are allowed only if they benefit the least advantaged.
The test works on any scale. A pay structure, a layoff rule, or a team’s on-call schedule can each be checked by asking whether you would accept it without knowing your role.
| Question | Ordinary bargaining | Behind the veil |
|---|---|---|
| What do I know about my position? | Everything | Nothing specific |
| Whose interests do I weigh? | Mine and my group’s | Every affected position |
| What outcome do I fear most? | Losing my advantage | Landing in the worst seat |
| What rules tend to win? | Rules favoring the strongest bargainer | Rules with a decent floor |
The experiments show it changes how people reason
The veil began as philosophy. It now has an experimental record.
- Wealth. Michael Norton and Dan Ariely asked a nationally representative US sample to pick a country to live in, knowing they would be randomly assigned to one of its wealth quintiles. About 92 percent chose a Sweden-like distribution over the actual US one, a result that held across gender, income, and 2004 vote (Perspectives on Psychological Science).
- Moral dilemmas. Across 7 experiments with 6,261 participants, Karen Huang, Joshua Greene, and Max Bazerman found that people who first asked what they would want without knowing who they would be made more impartial, greater-good choices, including a real donation decision (PNAS, 2019).
- Scarce ventilators. In two preregistered COVID-19 studies, veil reasoning shifted preferences toward saving younger patients, even among older participants (study).
- AI rules. In five studies with 2,508 participants, people who did not know their own position more often chose principles for an AI assistant that prioritized the worst-off (Weidinger et al., PNAS 2023).
These results show the veil changes how people reason. They do not show that it hands anyone a finished policy.
Veil questions for everyday decisions
| Decision you face | The veil question to ask | What usually changes |
|---|---|---|
| Setting pay bands or equity | If I were randomly assigned to any role here, would I accept this structure? | Compresses the top-to-bottom spread |
| Hiring and promotion | Can I remove names, schools, and photos from this stage? | Fewer identity-driven judgments |
| Layoffs or budget cuts | Would I endorse this rule not knowing which team I’m on? | Rules over ad hoc exceptions |
| Remote-work or on-call policy | Would I accept this as a junior and as an executive? | Protects the weakest position |
| Rationing scarce resources | Would I accept this if I might be any patient in the queue? | Explicit, transparent criteria |
| Deploying an AI system | Would I accept this system if I were the person it scores worst? | Guardrails for the worst-off |
For a fuller method, the ethical decision-making guide adds stakeholder and consequence checks. In meetings, the veil question works well as a round in group decision-making, because it gives people a shared reason to set their own position aside.
Where it already shapes policy
- Organ allocation. Bioethicists at Santa Clara University’s Markkula Center argue that planners “ignorant as to whether or not they will ever need a transplant or retransplant” would favor giving scarce organs to those most likely to gain long-term survival. US policy separately rules out allocating by sex, ethnicity, or socioeconomic status (OPTN ethical principles).
- Social insurance. Private insurers sort people into risk pools, which is the knowledge the veil removes. Mandatory social insurance is a way to choose the system before you know your own risk (Graetz and Mashaw).
- Shared resources. In a tragedy of the commons, each herder overgrazes because someone else will. Behind the veil, no one knows which herder they will be, so limits look better. Game theory explains why that agreement is hard to reach in practice.
When the veil backfires
- It gives direction without a blueprint. Critics note that the veil stays abstract and gives little concrete guidance on specific allocation questions (review). Use it to test options, then do the detailed design separately.
- It assumes caution. Rawls’s parties avoid risk. The economist John Harsanyi argued that rational people behind the veil would instead maximize average welfare, which leads to different rules.
- It can miss groups. Some critics argue the framework centers individual rights and underweights group disadvantages. Defenders point to fair equality of opportunity as a partial answer.
- It can dress up a decision already made. If you run the veil only after picking an answer, it becomes a rationalization. The self-serving patterns in cognitive biases explain why that is so tempting.
Questions
Who came up with the veil of ignorance?
John Rawls, in A Theory of Justice (1971). It is the central device of his “original position,” building on earlier social contract theory.
Is it the same as the original position?
Closely related. The original position is the imagined choice situation. The veil is the condition inside it that hides each chooser’s own place.
Can I use it at work?
Yes. Before approving a rule, ask whether you would accept it from any affected role. Where you can, remove identifying details from the decision itself, as the orchestras did.
Your 5-minute veil test
The rule or decision in front of me:
Every position it affects (list them):
The worst seat to land in:
Would I accept this rule from that seat?
What identifying detail could I hide from the decision?
One change that would make the worst seat acceptable:Run one pending decision through this test today, and save your veil questions as a reusable prompt in Save Effort.