The missing principal: who decides what boards optimise?

Who decides what boards are asked to optimise?

That question lies behind almost every debate about corporate governance, climate transition and fiduciary duty. Yet it is almost never asked directly.

Every organisation optimises an objective function. Sometimes it is explicit: earnings per share, total shareholder return (TSR), return on equity. More often it is embedded, in remuneration, analyst expectations, index construction, stewardship policies and promotion incentives. Boards do not simply maximise profits. Boards maximise whatever their institutions reward.

Directors, in other words, do not choose the objective function they optimise. Someone, or something, chooses it for them.

Corporate governance has spent decades asking whether directors may look beyond profit. That question is exhausted. The harder question is institutional, and it is the subject of this article.

Corporate governance is the design of the objective function.

Everything that follows explains why. This is not a commentary on company law, nor another contribution to the environmental, social and governance (ESG) debate. It is the sketch of a framework, a theory of fiduciary governance.

In our recent five-part illuminem series, The Broken Ballot, we argued that shareholder voting cannot fully reveal shareholder welfare. The reason is simple. A ballot asks shareholders to say yes or no to questions the company has chosen to put. But what shareholders actually care about, their savings, their health, their jobs, the climate their children will inherit, cannot be reduced to yes or no. No vote can carry that information into the boardroom. Someone has to fill the gap, and that someone is the board, exercising judgment. Because shareholders cannot say everything through the ballot, directors cannot avoid deciding on their behalf. The question therefore changes from "who decides?" to "how should they decide?", and, ultimately, to "what should they be asked to optimise?"

The law sets boundaries, not objectives

I begin, briefly, with the law, not because it is the obstacle, but because so many assume that it is.

English law establishes the legal space within which judgment may be exercised. It does not prescribe the objectives that judgment should pursue.

Section 172 of the Companies Act 2006 does not require directors to maximise short-term profit or share price. It requires them to act, in good faith, in the way they consider most likely to promote the success of the company for the benefit of its members as a whole, having regard to long-term consequences, employees, suppliers, customers, the community, the environment, business reputation and fairness between members.

That is not a mechanical instruction to maximise a number. It is a statutory architecture for judgment.

Parliament, indeed, deliberately refrained from ranking these competing considerations. The statute tells directors what to consider. It refuses to tell them how those considerations should be weighed. That refusal is not an omission. The weighing is the judgment, and Parliament left it where it belongs, with the board.

ClientEarth v Shell is often misunderstood in this respect. The High Court's refusal to allow the derivative claim to proceed was not a statement that climate risk is irrelevant to directors' duties. It was a statement that the weighing of climate risk against other commercial considerations is, in ordinary circumstances, a matter for directors' good-faith commercial judgment. Courts review process, honesty and integrity. They do not replace the board's commercial judgment with their own. The case therefore demonstrates not judicial deference for its own sake, but constitutional respect for the institutional role of the board.

Even where the law now compels consideration of climate effects, as the Supreme Court held in Finch v Surrey County Council for the downstream emissions of an oil development, it compels the consideration, not the conclusion. That decision binds public authorities under the environmental assessment regime, not company boards; and even there, once the emissions have been assessed, the weight to give them remains a matter of judgment. Public law and company law share the same architecture: the law ensures that the right questions reach the decision-maker. It declines to answer them.

Law therefore sets the boundaries of discretion. It does not determine how discretion is used.

That is where the real governance problem begins.

The missing principal

The principal in modern corporate governance has become strangely invisible. Who actually instructs the board? Beneficiaries? Pension trustees? Asset managers? Proxy advisers? Index providers? Stewardship teams? In theory, shareholders are the principals. In practice, ownership is mediated through a long chain of agents, each optimising its own measurable proxy.

Half a century ago, Michael Jensen and William Meckling framed corporate governance as the problem of aligning managers with shareholders. This article departs from, rather than rejects, that tradition. The modern problem is no longer aligning a manager with an owner. It is aligning a long chain of fiduciaries with the welfare of the ultimate beneficiaries at its end.

The result, today, is not shareholder democracy but institutional noise.

Oliver Hart and Luigi Zingales help clarify what is missing. Their distinction between shareholder value and shareholder welfare shows why maximising the market value of a company is not necessarily the same as maximising the welfare of its shareholders. Shareholders are people. They are also workers, citizens, consumers, taxpayers, parents and pension beneficiaries. They may rationally care about clean air, stable employment, climate resilience and public health, even where those interests are not fully reflected in a company's share price.

Hart and Zingales therefore transform the fiduciary question. It is no longer "how should companies maximise value?" but "whose welfare should fiduciaries seek to advance?"

For diversified long-term investors, the answer is not sentimentality. It is arithmetic.

A pension fund does not own one company. It owns a slice of the economy. Pollution generated by one portfolio company does not disappear. It returns as healthcare costs, insurance losses, lower productivity, agricultural damage, political instability and sovereign risk. What looks like an externality at company level may be internalised at portfolio level.

Firm-level value maximisation and portfolio-level welfare maximisation are different objective functions.

What we choose to measure

This is where Charles Seaford's work on wellbeing becomes the pivot of the argument. If Hart and Zingales identify whose welfare matters, Seaford explains why measurement determines optimisation. Organisations optimise what they count. Societies get more of what they measure. So long as boards are rewarded for earnings, TSR and short-term capital market performance, they will produce those outputs, whatever the statute books say about communities, employees or the environment.

Change the measure, and the optimisation changes.

The architecture of the argument is now visible. Section 172 supplies the legal authority. Hart and Zingales identify the beneficiary. Seaford identifies the objective. And Sir Andrew Likierman, to whom I return below, explains how judgment itself can be exercised well.

The great corporate governance debate has always been framed as a question of constraints. What duties should directors owe? What conduct should the law prohibit? Yet mature systems of governance do not depend primarily upon prohibitions. They depend upon institutions that align incentives with purpose. The central challenge is therefore no longer to constrain directors but to design the objective function within which fiduciary judgment is exercised.

Stewardship as constitutional governance

Pension funds occupy a unique constitutional position within modern capitalism. They are neither regulators nor corporations. Yet through mandates, stewardship priorities, voting policies, manager selection and remuneration expectations, they are uniquely positioned to reshape the objectives corporations ultimately pursue.

Corporate governance, seen this way, begins to resemble constitutional government. Investment mandates are the constitutional settlement, defining the purposes for which power is conferred. Stewardship is the legislature, translating the preferences of millions of beneficiaries into standing instructions. Executive remuneration is secondary legislation, the delegated rules that determine behaviour in practice. The board is the executive. And the courts perform judicial review: they test process and good faith, but they do not govern the enterprise.

The analogy is more than decorative. Modern corporate governance resembles constitutional government more than contractual governance. Neither constitutions nor company law prescribe every decision. They establish institutions, allocate authority and create mechanisms through which judgment is exercised legitimately.

Stewardship, on this view, is not merely engagement. It is constitutional governance.

The missing task is therefore to design institutions that reward judgment rather than merely financial output.

Judgment, not just outcomes

Sir Andrew Likierman explains how judgment itself can be exercised well. Judgment is not an ineffable gift. It is a discipline: the capacity to take in relevant information, test evidence, decide whom to trust, weigh uncertainty, recognise trade-offs and reach a decision that can be explained and defended.

That understanding maps closely onto English law. Courts do not ask whether directors predicted the future correctly. They ask whether directors acted honestly, in good faith and within the proper scope of their discretion. A governance system serious about judgment should do the same. It should evaluate the quality of decision-making, not merely last year's financial result. Board evaluation, director selection and stewardship assessment could all be reoriented around a single question, how well does this board decide?, a standard which, in substance, English courts have applied for a century.

Why AI makes this urgent

Artificial intelligence (AI) makes this urgent.

AI systems are formidably powerful optimisers, and they are indifferent to objectives. They will optimise profit, welfare or pollution with equal facility. The moral responsibility therefore moves upward: away from the optimiser, and towards those who specify what is to be optimised. An AI-assisted board rewarded for quarterly earnings will maximise quarterly earnings with unprecedented efficiency. A board rewarded for long-term beneficiary welfare would optimise something different.

The danger is not that machines will suddenly exercise fiduciary judgment. They cannot decide what should matter. The danger is that institutions will hand them the wrong objective.

AI does not solve the governance problem. It exposes it.

Four generations of governance

This suggests a four-generation evolution in corporate governance.

Generation 1 — Friedman
"What should companies maximise?"

Generation 2 — Section 172
"What may directors consider?"

Generation 3 — Hart & Zingales
"Whose welfare matters?"

Generation 4 — Fiduciary governance
"How should institutions reward fiduciary judgment?"

The fourth generation is the subject of this framework: mandates, metrics and incentives redesigned so that boards are rewarded for exercising fiduciary judgment consistent with long-term shareholder welfare.

Each generation answered one question and left another unasked. Friedman told us what companies should maximise; Section 172, what directors may consider; Hart and Zingales, whose welfare matters; Seaford, what institutions should measure; Likierman, how judgment is exercised. The question that remained, the one this framework exists to answer, is how institutions should be designed so that fiduciary judgment is consistently exercised in pursuit of long-term shareholder welfare.

That is the real work.

How that redesign works in practice, how beneficiary preferences are elicited and priced, how mandates are drafted to authorise welfare objectives without abandoning fiduciary rigour, and how remuneration can reward resilience rather than reported earnings, is deliberately reserved for the second article in this series.

Corporate governance has spent half a century debating whether directors may look beyond profit. English law answered that question years ago. The next generation of governance will be judged not by the duties it imposes, but by the institutions it creates, institutions capable of translating the welfare of millions of long-term beneficiaries into the objective functions that boards, and increasingly artificial intelligence, are asked to optimise.

The defining question is no longer what fiduciary judgment permits. It is how society chooses to reward it.

illuminem Voices is a democratic space presenting the opinions of leading Sustainability Thought Leaders, their views do not necessarily represent those of illuminem.

The world needs sustainability knowledge. At illuminem, no interest group or shareholder can influence our work. Thank you for supporting our mission to make high-quality and independent sustainability information free for all. Every contribution helps. Thank you for donating today.