Should companies have obligations beyond maximizing shareholder value?
Yes, but the council splits on whether obligations within the current ownership structure can ever be enough.
Friedman and Hayek hold that executives spending shareholders' money on social goals are acting without a mandate. If society wants obligations, legislate them. Rawls and Sen counter that shareholder primacy systematically fails workers: it externalises costs onto those with the least power to absorb them. Roosevelt's record in 1933 shows that legislated constraints can stabilise a system markets were destroying on their own.
The split that argument cannot close is structural. Luxemburg argues that every reform leaving ownership intact teaches capital to absorb the reform and continue. Rawls and Sen stop short of that conclusion, accepting constrained capitalism as improvable. Neither side can be ruled wrong on the evidence alone.
Confidence summary: Strong convergence on the failure of pure shareholder primacy in practice; deep and irreducible split on whether any reform short of changing ownership can hold.
1. The core argument
The most clarifying moment in this deliberation is not the familiar Friedman objection but Roosevelt's concession. He did not claim the New Deal changed who owned productive assets. He claimed it stopped the system from destroying itself. That distinction matters enormously: it means even the council's strongest defender of democratic intervention accepts that the ownership relationship survived intact. The question then becomes whether obligations layered on top of that relationship are durable improvements or absorbed irritants.
Rawls and Sen supply the affirmative case: shareholder primacy fails distributive justice by design, externalising costs onto workers and communities who lack the power to refuse them. Friedman and Hayek supply the procedural objection: unelected managers imposing social priorities on other people's capital are exercising political power without a mandate. Roosevelt supplies the historical counter: markets, left unconstrained, produce crises that democratic governments must then repair. Luxemburg supplies the structural question none of the others fully answer.
2. How each member frames it
Milton Friedman does not simply defend profits. He defends a specific theory of authorisation. The corporate executive is an agent; the shareholders are principals; spending their money on social goals the executive chose is a form of taxation without representation. What the card left out is Friedman's actual concession: he does not oppose corporations meeting social obligations when those obligations are legislated. His target is the voluntarist version, the manager who appoints himself a social reformer. He would accept a statutory minimum wage or a carbon tax far more readily than he would accept a CEO deciding unilaterally to pay above market rates for community benefit. That distinction is sharper than his critics usually acknowledge.
Friedrich Hayek presses further into the mechanism. Where Friedman's objection is political, Hayek's is informational. A board instructed to balance the interests of workers, communities, and shareholders must weigh claims that have no common unit. Prices already perform that aggregation across millions of decisions; stakeholder mandates override the signal without replicating it. The candid limit Hayek rarely states explicitly: his argument assumes competitive markets are functioning well enough to generate prices that genuinely reflect social costs. Where monopoly power, information asymmetry, or uninternalised externalities distort those prices, the epistemic case for trusting markets becomes circular.
John Rawls shifts the axis from efficiency to justice. The corporation is not an island; it is part of the basic structure of society, the set of institutions whose rules shape life chances before anyone makes a choice. On his difference principle, inequalities in that structure are only legitimate if they improve the position of the least advantaged. Shareholder primacy fails that test not occasionally but structurally: ownership is concentrated, costs are pushed outward, and the workers bearing those costs have no standing in the governance that generates them. Rawls stops short of demanding worker ownership; he accepts that well-regulated capitalism can satisfy his principles. That stopping point is precisely where Luxemburg challenges him.
Amartya Sen grounds the argument in what people can actually do and be, not in abstract distributions. A firm posting strong returns while exposing workers to chronic health hazards has not created value; it has redirected costs onto bodies that cannot refuse them. The capability lens is unsparing about what profit measures and what it does not. Sen's deeper move, which the card could not carry, is that capability destruction is often invisible in aggregate statistics: GDP rises, inequality indices move modestly, but the actual freedom of specific workers to stay healthy or participate in civic life contracts. That invisibility is not accidental; it is a structural feature of what shareholder returns measure.
Franklin D. Roosevelt is the council's empiricist. The New Deal was not a theory; it was a response to collapse. Wages too low to sustain consumer demand, financial instruments sold on fraud, workers with no protection against age or injury: these were not peripheral failures but self-defeating dynamics built into the system's operation. The Securities Exchange Act and the Social Security Act imposed obligations markets would never have generated. His candid limit is Sen's challenge in reverse: reform saved the system without changing who owned it. Whether that counts as a solution or a postponement depends on whether one believes the structural relationship between owners and workers is itself the source of the problem.
Rosa Luxemburg refuses the terms of the other five. The council's debate has largely been about what obligations corporations should accept within the existing ownership structure. She argues that framing leaves the decisive question unasked. Capital absorbs reform; the history of labour law, environmental regulation, and corporate social responsibility codes can be read as a series of concessions that stabilised exploitation rather than ending it. Her position is not that reform harms workers in the short run. It is that reform that does not touch ownership teaches capital to adapt and continue. Whether that is historical law or a contingent pattern is the line on which she and Rawls genuinely disagree.
3. Where the council agrees
The most surprising point of agreement is that pure laissez-faire has no serious defender here, including Friedman and Hayek. Both accept that legislated obligations are legitimate; neither argues that corporations should operate without any external constraint. That convergence rules out one major option before deliberation even begins.
Beyond that, all six members accept that externalising costs onto workers and communities is a real phenomenon, not a theoretical abstraction. They disagree sharply about whose job it is to correct it and how, but none of them contest the baseline finding. Roosevelt's record in 1933 is treated as evidence, not ideology, by members who otherwise disagree profoundly: markets left entirely unconstrained produced a crisis severe enough to threaten the system itself. The council also agrees, implicitly, that the current governance debate is not purely academic: the gap between shareholder returns and wage growth over the past forty years gives the distributive critique concrete purchase.
4. Where the council splits
The line is ownership. Rawls, Sen, and Roosevelt accept that constrained capitalism, with legislated obligations, redistributive taxation, and enforceable worker protections, can satisfy the demands of justice or at least functional stability. Luxemburg argues that every reform leaving the ownership relationship intact is ultimately absorbed: capital learns to accommodate the constraint without surrendering the structural advantage that generates exploitation. Neither side is wrong on the evidence alone. The reform camp points to genuine improvements in worker welfare under social democratic regimes. Luxemburg points to the long-run tendency of those improvements to stall, erode, or be traded away when capital can exit. The disagreement is partly empirical and partly about which time horizon one considers decisive.
5. For a policymaker to decide on
The council cannot resolve this choice: whether to pursue legislated obligations on existing corporations (stronger labour protections, mandatory stakeholder reporting, carbon pricing) within the current ownership structure, or to redesign governance itself through worker representation on boards, mandatory profit-sharing, or expanded cooperative ownership. The first is more immediately achievable; the second addresses the structural critique Luxemburg and, partly, Rawls raise. Which you choose depends on whether you believe durable reform is possible without touching who governs the enterprise.