The returning principal

This is article 2 of 2 in The Principal series. Here is article 1.

Somewhere in Britain tonight a nurse is finishing her shift. Part of this month's pay has gone quietly into a pension fund, as it does every month, and with it travels a question that nobody has ever thought to ask her: what should that money want?

The first article in this series argued that corporate governance comes down to the design of the objective function – the set of goals, measures and rewards that tells a board what counts as success. English law already allows directors to look well beyond short-term profit. What stops them is not the law but the machinery around them: boards pursue whatever their institutions reward, and at the moment nobody is deliberately deciding what that should be. Between the nurse and the boardroom stretches a long chain of trustees, consultants, asset managers and proxy advisers, each translating her interests into whatever its own performance happens to be measured by. By the time her money reaches a company, her voice has gone.

This article asks who might restore it, and answers: the pension funds themselves. Not because they are nobler than other investors, but because of what they are.

Why this has become urgent now

Why has this become urgent now, rather than twenty years ago? Four things have changed. Ownership has concentrated, so that a handful of pension and sovereign funds now hold slices of the entire listed economy, and the costs companies push onto society land back inside their own owners' portfolios. Climate change has made those costs large enough to move whole economies rather than single firms. The value of companies has migrated from factories and machines to things the old financial measures were never designed to capture: people, ideas, trust. And artificial intelligence (AI) has arrived, ready to pursue whatever goals we set it with an efficiency no board has ever commanded. Any one of these would strain the old settlement. Together they mean that the objective function has become the most consequential document in capitalism, at exactly the moment when nobody is drafting it.

Why pension funds are the natural draftsmen

Three things make pension funds the natural draftsmen. They are permanent: a pension fund cannot sell its way out of the future, and a member of twenty-five will be drawing her pension half a century from now, in whatever economy, climate and society then exist. They are universal: a large fund owns a slice of everything, which means the pollution of one company in its portfolio comes back to it as healthcare costs, insurance losses and lost productivity in a hundred others. There is no elsewhere to push the damage. And at the end of the chain stands a human being – the nurse again – who needs both an income in retirement and a world worth retiring into.

These are not pieties; they are incentives. Permanence removes the escape route that lets other investors sell before the consequences arrive. Universal ownership means that costs dumped on society come straight back onto the fund's own books. And the duty runs to real people rather than to an abstraction. Nor is any of this a thought experiment. Japan's Government Pension Investment Fund, the largest in the world, describes itself in precisely these terms, as a universal owner investing across generations. Norges Bank Investment Management publishes what it expects of companies on climate, workforce and boards, and votes against those that ignore it. When the Dutch civil service fund ABP decided in 2021 to sell its holdings in fossil fuel producers, it gave the views of its own members among its reasons. Something is stirring.

What the chain currently carries

Yet for most funds the chain still looks like this:

1. Beneficiaries
2. Trustees
3. Investment consultants
4. Asset managers
5. Proxy advisers and stewardship teams
6. Boards

At every link, welfare is translated into something easier to measure: tracking error against a benchmark, quarterly manager performance, league tables, voting statistics, total shareholder return (TSR). A human life goes in at one end; a demand for earnings comes out at the other. Another code or another disclosure will not mend this. What has to change is what the chain carries.

The first article described the arrangement in constitutional terms – mandates as the constitution, stewardship as the legislature, remuneration as the delegated rules, the board as the executive, the courts as judicial review. Working down that constitution, five things need redesigning, and none of them requires an Act of Parliament.

Five things need redesigning

Begin with the beneficiaries. Our earlier series, The Broken Ballot, showed why voting cannot carry their voice: a ballot invites yes or no on questions someone else has framed, while the things people care about will not reduce to yes or no. The remedy is not silence but better instruments – surveys built around real trade-offs, panels of members given the time and information to deliberate, sensible defaults with genuine opt-outs. Above all, preferences need prices attached. Ask the member directly: would you give up a quarter of one per cent of expected return for a portfolio that supports a stable climate? Some will say no, and that too is worth knowing. When the largest field study of its kind, published in the Review of Financial Studies, put a binding version of that question to Dutch pension members, two-thirds chose more sustainability even though they expected it to cost them – and the fund then did what its members had asked. Preferences with a price on them can be acted upon; preferences merely assumed cannot.

Then write the answer into the mandate. A typical mandate instructs a manager to beat a benchmark by a margin over rolling three-year periods, and that single sentence, repeated across thousands of mandates, is the real objective function of modern capitalism – one that nobody whose welfare it governs ever chose. Redrafting it is the constitutional moment. A mandate grounded in what members have actually said would require competitive long-term returns together with the things they have asked for: a resilient climate, natural capital, decent work, a stable financial system. This is not charity with other people's money. It is the end of guessing.

Stewardship then stops being a voting service and becomes a legislature. Parliaments do not govern by dividing on other people's motions; they give standing instructions. A stewardship policy built on welfare mandates would tell boards what the members want more of – resilience, less exposure to stranded assets, investment in the workforce – and would judge them by the question the first article placed at the centre: how well does this board decide? English courts have judged directors that way for a century, looking at the honesty and quality of the process rather than the luck of the result. Funds should do the same, and escalate when the judgment is poor rather than when the luck is.

Measurement comes fourth, and needs care, because measurement is where good intentions usually die. Boards are not algorithms, and the purpose of welfare metrics is to sharpen their discussions, not to settle them. A board might sensibly be assessed on a handful of dimensions it can actually reason about – its financial performance, its people, its use of nature, its capacity to innovate, the trust it commands – with the weights a matter for honest argument rather than false precision. What matters is restraint. Goodhart's law, that a measure which becomes a target stops being a good measure, has wrecked every governance scorecard yet devised, and always for the same reason: the scorecard tried to do the deciding. Metrics should work the way evidence works in a courtroom – indispensable to the judgment, never a substitute for it.

And none of this is speculation. In 2016 two modest British funds, the Church of England Pensions Board and the Environment Agency Pension Fund, built exactly such a scorecard. The Transition Pathway Initiative, housed at the London School of Economics, grades companies on how well they manage the climate transition and how credible their carbon path is. More than a hundred investors, with over forty trillion dollars behind them, now use it; the Church fund votes against the chairmen of companies that score badly; and when the scorecard was turned into an investable index, the fund moved six hundred million pounds into it and the New York State retirement fund followed. In the Netherlands, Detailhandel let its members bind it. Every instrument this article proposes has been used somewhere, in earnest, with real money. What is missing is not the technology of instruction but the habit.

Last, pay for it. Remuneration is the delegated legislation of this constitution, the detailed rules that actually govern behaviour, and at present it legislates almost entirely for earnings per share, TSR and the annual bonus. Nobody should be surprised at what results. Tie a serious fraction of executive reward – a quarter, say – to the welfare dimensions above, and choose and keep asset managers for the quality of their stewardship rather than their tracking error alone, and the chain begins to reward at every link what the beneficiary values at its end. Judgment finally has something working for it.

Notice what these five changes have in common: none of them takes a single decision away from a board. They surround its judgment with better questions, better information and better reasons, and leave the deciding where English law has always left it, with the directors. The point of governance design is not to decide for boards but to make good judgment the rational choice.

A precondition: funds must govern themselves

Honesty requires an admission before the objections. Most pension funds are not yet equipped for this role. Trustee boards are often part-time and thinly resourced; consultants dominate decisions they do not own; committees turn over faster than the horizons they are supposed to steward; and funds herd, because straying from the pack is a career risk while failing with it is not. Consolidation into larger, professionally governed pools – already under way from Australia to Britain's local government schemes – is not a detail of this argument but its precondition. It comes first. A principal that cannot govern itself cannot govern capitalism.

Four objections

Four objections deserve straight answers, and the largest first. Doesn't all this trade prosperity for virtue? That assumption underlies the litany now recited at industry gatherings across Europe – sustainability as bureaucratic ballast, the transition as 'ideology', green and growth as natural enemies. The evidence runs the other way. The firms that have invested most in energy efficiency and in their environmental and social performance are, more often than not, the more competitive ones; the loudest complaints tend to come from those whose real handicap is inefficiency, defended by louder lobbies. The trade-off looks fearsome at quarterly frequency and shrinks at the horizon a pension fund actually inhabits, because over decades the expensive strategy is not transition but delay: assets stranded, capital written off, heat that withers harvests and labour productivity alike. When those Dutch members voted for sustainability even though they expected it to cost them, the expectation itself was probably too pessimistic – at portfolio horizon the cost of an orderly transition is modest and uncertain in sign, while the cost of a disorderly one is neither. A fund that treats green and productivity as enemies has simply chosen the wrong timescale. And resignation is not a strategy: a fund that shrugs at the transition betrays the very generations it exists to pay.

Doesn't fiduciary duty forbid all this? No. The duty is owed to the members, not to a benchmark, and in Britain the Law Commission said as much in 2014: trustees may take account of non-financial concerns where the members share them and the returns are not significantly harmed. The survey establishes the first condition; the portfolio arithmetic protects the second. The objection assumes the one thing this framework removes – ignorance of what members actually want.

Won't the chain refuse to transmit? Sometimes, yes. Benchmarks, cautious consultants, competing managers and regulation all create friction, and pension funds are well placed to reshape objectives, not guaranteed to. But position counts. Funds sit at the top of the chain: they write the mandates, hire the consultants, choose the managers and set the terms. When the largest of them move, standards tend to follow – the histories of index investing and of climate disclosure both say so.

And won't the members disagree with one another? Of course, as voters do, and nobody concludes from that that countries should go ungoverned. Deliberation, sensible defaults and honest disclosure of the trade-offs are how collective preferences are legitimately formed. A fund that asks stands on firmer ground than a fund that assumes.

The machines

Which brings us back, at the end, to the machines. AI has no view about objectives; it will pursue profit, welfare or pollution with the same relentless competence, which is why it exposes the governance problem rather than solving it. But turn that indifference around. Once mandates, measures and pay define an objective worth having, the same machines will pursue the welfare of beneficiaries with all the efficiency they currently devote to quarterly earnings. They will amplify whatever constitution we write. Better, then, to write it deliberately, and soon.

Nothing here needs legislation. The instruments exist – a survey, a mandate, a stewardship policy, a scorecard, a pay scheme – and every one of them, as we have seen, has already been used by real funds with real money. What is needed is for the institution that stands closest to the nurse to remember whom it serves, and to say so, in writing, down the whole length of the chain.

The principal of modern capitalism was never abolished; it was mislaid somewhere between her payslip and the boardroom. Capitalism does not lack intelligence. It lacks instruction – and only the pension fund stands close enough to her to give 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.