BP — The Strategic Stack

Executive summary

In February 2020, Bernard Looney committed BP to net zero by 2050 and a forty-percent cut in oil and gas production by 2030. In February 2025, his successor Murray Auchincloss withdrew the production target, raised oil and gas capital expenditure to $10 billion a year, and reduced renewables spend. Helge Lund, the chair who backed both moves, announced his departure in April 2025. Elliott Management, having taken a stake of approximately five percent in the weeks before the reset, is now public in pressing for further divestment of renewables. As of April 2026, the reset has not resolved. The coalition can still move.

This is the case the Strategy Room has been building towards. It is also the case that exposes the limits of any single strategic framework when applied to a live boardroom decision under contested ownership. This whitepaper sets out the discipline we have been calling The Strategic Stack — a method of composing six analytical lenses in a specific order so that each constrains the reading produced by the one beneath it.

The six lenses, from input to foundation:

1. Stakeholder theory (Freeman, 1984) names who is in the room.
2. Mendelow's matrix (1991) weights the room by power and interest.
3. Agency theory (Jensen and Meckling, 1976) versus stewardship theory (Donaldson and Davis, 1991) asks how the board is trying to govern that weight.
4. The ESG framework (UNGC, 2004; ISSB, 2023) is the measurement language in which the governance is assessed.
5. The Balanced Scorecard (Kaplan and Norton, 1992) is the instrument that puts that measurement into executive pay.
6. Shareholder primacy versus stakeholder capitalism (Friedman, 1970; Business Roundtable, 2019) is the ideological foundation the other five rest on.

The argument of this paper is that the reset is not a single decision. It is a stack — a chain in which Elliott used Mendelow's power to impose an agency-theoretic governance shift, legitimised by a rewritten ESG taxonomy, operationalised through a rebalanced Balanced Scorecard, and justified by a return to shareholder primacy. Each move supported the next. The Notes treated the lenses one at a time. The stack is what the Notes could not show.

We close with a predictive frame: six conditions under which the reset holds, six under which it reverses, and the 2027 proxy season as the moment the argument is forced to settle. The frame is offered as a scorecard chairs and NEDs can use against their own decisions, not only against BP's.

1. The case, dated

Bernard Looney took the chief executive's office at BP on 5 February 2020. Seven days later, on 12 February, he committed the company to net zero across operations and across the full value chain by 2050, with a forty-percent cut in oil and gas production by 2030. Capital expenditure was to tilt towards renewables, hydrogen, and electric charging. The pledge was the most ambitious commitment by any oil major to that date, and it was made in a register — purpose, transition, coalition — that other oil majors did not adopt.

Looney left in September 2023. Not over the strategy. Over undisclosed personal relationships at work that the company concluded he had not declared as required. Murray Auchincloss, his chief financial officer, stepped up.

For a year, the pledge held on paper. Strategic change rarely moves on the day a chief executive arrives. Capital commitments already in flight do not stop on instruction. Internal consensus takes months to rewrite itself. The twelve-month interval between Looney's departure and the reset was not indecision. It was institutional inertia, which is the form most strategic change actually takes.

On 26 February 2025, Auchincloss announced what the company called a fundamental reset. Capital expenditure on oil and gas was raised to $10 billion a year. Renewables spend was cut. The forty-percent production-reduction target was withdrawn. Distributions were prioritised. The strategic horizon shortened.

Three contextual events are material. First, in late 2024 and the early weeks of 2025, Elliott Management took a stake in BP of approximately five percent and became public in pressing for further divestment of renewables. The five-percent stake mattered because of what came with it — a credible threat of escalation, a reputation for winning campaigns, and the urgency of a fund that does not hold for a decade. Second, on 4 April 2025 Helge Lund — chair since 2019, who had backed both the original pledge and its reversal — announced his departure. The succession process is live as this paper is written. Third, the UK Labour government, in office since July 2024, has watched the reset with declared interest in the energy-security implications but has not yet moved any of the three policy levers — windfall tax, licensing, infrastructure friction — that would shift its position from interested observer to active counterparty.

As of April 2026, the reset has not resolved. The coalition can still move.

2. The six lenses, in order

We build the stack from the ground up. The directional convention matters. A board encounters the lenses in the order in which it is forced to use them. The board sees the stakeholder list before it sees the weighting. It sees the weighting before it sees the governance theory. It sees the governance theory before it can specify the measurement language. It builds the scorecard last, after the other four are settled. And only the rare board, under sustained pressure, is forced to articulate the ideological foundation it has been operating on without knowing.

A caveat before we begin. Read in retrospect, the six moves at BP look engineered. In real time they were not. Boards rarely stack these lenses cleanly. The stack is how we make sense, after the fact, of a collision — between an activist's agency-driven demand and a board's stewardship-driven history — that felt, in the room, like a series of separate judgements. Retrospective coherence is not the same as real-time clarity. What the stack gives us is the first. What it earns us is the right to attempt the second.

2.1 Freeman — who is in the room

R. Edward Freeman, in Strategic Management — A Stakeholder Approach (1984), defined a stakeholder as any group or individual who can affect, or is affected by, the achievement of the firm's objectives. Affect or be affected. The phrase is wider than the firm's chart of legal obligations. It pulls in employees, customers, suppliers, capital providers, regulators, communities — and, as later scholarship made explicit, the natural environment and future generations.

Donaldson and Preston, writing in 1995, made the distinction that still matters. Stakeholder theory is three theories in one. Descriptive — a claim about who the firm actually has relationships with. Instrumental — a claim that firms which take those relationships seriously produce better returns. Normative — a claim that those relationships carry moral weight, whether or not they are useful. Most boards, when they adopt the language, mean the instrumental version. Freeman's theory, at its foundation, is normative.

Freeman's list, applied to BP in February 2025, is long. Elliott Management. Legal and General Investment Management. Norges Bank Investment Management. Nest, the UK workplace pension. Vanguard. The activist filers — Follow This, ClientEarth. The workforce of approximately sixty thousand employees, internally divided between refining communities at Gelsenkirchen, Whiting, and Castellón and renewables teams in Iberia and the North Sea. The UK government. Suppliers. Communities at every operational site.

What Freeman's lens supplies is the input to the stack. It does not say what happens next. It says who cannot be defined out of the room. The board that re-weights a stakeholder's claim to near-zero — without being able to articulate the ethical justification for that re-weighting — has not practised stakeholder theory. It has managed the public-relations costs of shareholder primacy in stakeholder-theory language. The distinction is the test.

2.2 Mendelow — power and interest, plotted twice

Aubrey Mendelow's 1991 power–interest matrix takes Freeman's list and forces a board to prioritise. Two axes — structural power, observed interest — and four quadrants. Key players (high–high). Keep satisfied (high–low). Keep informed (low–high). Minimal effort (low–low). The matrix does what Freeman's list cannot do — it forces a board to prioritise, in language the board can defend.

The instructive move is to plot the matrix at two points and look at what has changed.

February 2020. The key-players quadrant is populated by the major index-weighted holders — LGIM, Vanguard, Norges Bank — and by BP's own board. Keep satisfied — the UK government, European regulators, sovereign-wealth holders indexed but passive. Keep informed — ClientEarth, Follow This, transition-aligned NGOs, student activist organisations. Minimal effort — retail holders, short-cycle hedge funds uninterested in a fifteen-year thesis.

February 2025. The matrix has redrawn itself in two directions at once. Elliott has taken a five-percent stake and become a key player overnight. Climate-aligned index holders have moved down the power axis — not because their shareholdings shrank, but because the new strategic story is no longer written in a language their stewardship teams can defend internally. Follow This has moved up the interest axis but remains low on power. The retail-holder segment has become quietly more powerful through proxy aggregation. The UK government, under Labour since July 2024, has become a higher-interest actor — though, on Mendelow's axes, not yet a higher-power one. A government moves up the power axis through three instruments. A windfall tax. A change to licensing conditions on oil and gas blocks. Regulatory friction applied to pipeline or port infrastructure. Labour has, so far, used none of the three.

Two maps. Same company. Five years apart. No stakeholder has left the room. Each has moved.

What Mendelow adds to Freeman is this. Auchincloss could read the 2025 map and know that acting against the pledge would cost him less from the current key-player set than holding to the pledge would cost him from Elliott. That is not a moral claim. It is a readable one. The matrix let the decision be read.

The critique that holds. The matrix treats power and interest as independent — which empirically they are not. Interest without power tends to acquire power, through proxies, through coalitions, through regulators. A climate-aligned group at keep-informed in 2020 can become a key player by 2030 through a single legislative act, a single court ruling, or a single index methodology change. Mendelow gives a map of today. It does not give a map of trajectory. The board that reads only today loses.

2.3 Agency theory versus stewardship theory

Jensen and Meckling's Theory of the Firm (Journal of Financial Economics, 1976) frames the firm as a nexus of contracts. The shareholder is the principal. The manager is the agent. The two have different interests. The manager will, unless constrained, shirk, empire-build, consume perks, and resist risks the shareholder would rationally take. Governance is the architecture of constraint. Incentive alignment through pay. Monitoring through independent boards. Threats of removal through the market for corporate control. Agency theory has produced most of modern governance — separation of chair and chief executive, majority-independent boards, annual director elections, executive pay tilted towards equity, clawbacks, say-on-pay.

Donaldson and Davis (1991) propose the counter-theory. The same manager, observed through a different psychological frame, is not an agent seeking private utility. The manager is a steward — psychologically aligned with the long-term success of the firm, identifying with its purpose, motivated by achievement rather than by extraction. Surveillance, on this theory, is the problem. A board that distrusts its executive structurally produces the executive behaviour it fears. Stewardship governance gives the chief executive more autonomy, longer horizons, broader scorecards, and a chair who holds ground rather than polices the agenda.

The empirical literature is long and mixed. Muth and Donaldson (1998) find stewardship-designed boards outperforming agency-designed boards over long horizons in capital-intensive and high-research-and-development sectors — precisely the sector profile of integrated energy. The agency-theoretic evidence is stronger on short-horizon returns around chief-executive transitions and activist campaigns. Real boards rarely choose one theory cleanly. They are one theory on some questions, and the other on others.

BP under Looney was being run on a stewardship thesis. The pledge assumed an executive team capable of stranded-asset judgement across fifteen years and a board willing to hold the shareholder line while the thesis matured. The chair, Helge Lund, came from stewardship-style oil and gas — Statoil, BG Group — and had publicly backed the pledge as a long-horizon commitment.

BP under Auchincloss is being run on an agency thesis. The reset is a response to the market for corporate control — Elliott's stake, the risk of further activism, the concern that a continued stewardship pose would invite a takeover bid. The pay design shifted. The scorecard shifted. The strategic horizon shortened. Lund, the same chair, approved both moves.

The same chair approving two opposite governance theses five years apart is not incoherent. It is common. Most boards hold one theory until an external event forces the other. The test of the board is whether the theory it adopts under pressure is the one the situation requires — or the one the loudest stakeholder prefers.

2.4 The ESG framework — the measurement language

The Principles for Responsible Investment, launched in 2006, today represent more than $120 trillion in signatory assets under management. The ESG apparatus that has built up since the 2004 Who Cares Wins report is not one framework but several — MSCI, Sustainalytics, ISS ESG, Refinitiv, the Global Reporting Initiative, the Sustainability Accounting Standards Board, the Task Force on Climate-related Financial Disclosures, the International Sustainability Standards Board (consolidating through standards S1 and S2 from 2023 onwards), the European Union's Corporate Sustainability Reporting Directive (applying from 2024), the UK's Sustainability Disclosure Requirements, and the SEC's climate disclosure rule, still on hold in US courts as of April 2026.

The point of the letters is not what they prescribe. They do not prescribe. The point is that they give a board a shared language for non-financial performance — a measurement grammar that allows a pension fund in Oslo, an activist fund in New York, a regulator in Brussels, and an NGO in Nairobi to argue about the same company using the same variables. What they disagree about is the values attached to the variables. The variables, increasingly, are shared.

BP's pre-reset scorecard, held from 2020 through 2024, measured a distinctive bundle of variables. Scope 1, 2, and 3 emissions, reported under TCFD and subsequently ISSB S2. Transition capital expenditure as a percentage of total. Renewables operating capacity, in gigawatts, against a 2025 interim target. Methane intensity. Safety metrics tied to Deepwater Horizon obligations. Water intensity. Board diversity. Executive pay equity.

The reset rewrote the scorecard. The forty-percent oil-cut target — removed. The transition-capex target — reduced from a commitment to a trajectory. The renewables-operating-capacity goal — reset to a lower number. Methane intensity — retained. Safety — retained.

The three major rating agencies moved within a month. MSCI downgraded BP's ESG rating from AA to A. Its methodology weights industry-relative transition opportunity and forward-looking climate-scenario exposure. Sustainalytics raised BP's ESG Risk Rating from medium to high. Its methodology weights unmanaged present-day ESG risk. The two agencies asked different questions of the same reset and got different answers — the inconsistency Bebchuk and Tallarita have built their critique around. ISS ESG withdrew BP's climate-leadership classification.

The ratings moves were material in a specific sense. LGIM's internal climate policy uses the MSCI rating as one input into its engagement and divestment decisions; the downgrade triggered a formal engagement process. Norges Bank's ethics council, which operates on its own criteria, opened a preliminary review. The ratings did not make any decision. They made decisions legible to the actors whose decisions mattered.

What the BP case demonstrates is the counterpoint to the standard ESG critique. The framework failed to prevent the reset. It also failed to hide it. The variables were still measured. The ratings still moved. The legitimation BP could have sought in ESG language was not available in April 2025. The framework is not a policy. It is a vocabulary. And a vocabulary's job is to make disagreement sayable.

2.5 The Balanced Scorecard — strategy made operational

Kaplan and Norton's The Balanced Scorecard — Measures That Drive Performance (Harvard Business Review, January–February 1992) made the observation that financial metrics measure the past — return on equity, earnings per share, total shareholder return. They tell the board whether last year's strategy worked. They cannot tell the board whether this year's is working. For that, a board needs measures that lead, not measures that lag. Four perspectives, balanced — financial, customer, internal process, learning and growth. For each, objectives. For each objective, measures. For each measure, targets. For each target, initiatives.

The methodological origin is worth pausing on. Mobil's North American Marketing and Refining business in 1994 had lost money for years under a scorecard tilted to financial metrics. Kaplan and Norton rebuilt the scorecard around customer and internal-process measures. Within three years, Mobil NAM&R was the most profitable business in its segment. The Mobil case and the Chemical Bank case that followed were the empirical backbone of the 1996 book.

BP has used a Balanced Scorecard design for executive pay for more than a decade. Four dimensions, weighted, reported in the annual remuneration report, tied to the long-term incentive plan. Trace the weight changes year by year:

• 2021 remuneration report, after Looney's pledge — learning and growth carried approximately thirty percent of the scorecard, with transition-capability metrics explicit.
• 2023 report — the weight dropped quietly to around twenty percent.
• 2024 report — it dropped again.
• 2025 report, the first under Auchincloss — operating cash flow and distributions carried more weight, taken together, than the entire learning-and-growth dimension.

The reset did not begin on 26 February 2025. It began in the remuneration committee two years earlier. When the scorecard shifted, the strategy shifted. What the executive was being paid to deliver changed before any capital decision was announced. The board that can read the scorecard can read the strategy twelve months before the strategy is written.

The critique that holds — Goodhart's Law. Any measure that becomes a target ceases to be a good measure. The BP case itself supplies the pattern. By 2023, the transition-capex figure was being counted toward the scorecard target in cases where projects had slipped out of the reporting year. Renewables-capacity figures aggregated pipeline capacity alongside operating capacity — flattering the number without producing the gigawatt-hours. The numbers satisfied the scorecard without producing the operating reality the scorecard was designed to encourage. Paul Niven (2006) is the most operationally candid account of where scorecards fail.

2.6 Shareholder primacy versus stakeholder capitalism — the foundation

Milton Friedman's The Social Responsibility of Business is to Increase its Profits (New York Times Magazine, 13 September 1970) is the intellectual foundation of half a century of shareholder-primacy governance. The argument is not that the firm has no other obligations. The argument is that the chief executive's obligation runs first to the firm's owners, and that any other use of the firm's resources constitutes a spending of other people's money. Social goals are for elected governments. The firm's job is to make money lawfully.

The legal scaffolding is older. Dodge v. Ford Motor (Michigan Supreme Court, 1919) — a corporation is organised and carried on primarily for the profit of the stockholders. Revlon v. MacAndrews (Delaware Supreme Court, 1986) — once a sale of control is inevitable, the directors' duty is to maximise the price received by shareholders. eBay v. Craigslist (Delaware Chancery, 2010) — a non-profit purpose is not a lawful corporate purpose unless rationally related to shareholder value. Shareholder primacy is a doctrine with three levels — economic, legal, cultural — each reinforcing the others.

The counter-tradition has a shorter but denser history. Freeman (1984). Colin Mayer's Prosperity (2018). The British Academy's Future of the Corporation programme. The Business Roundtable's Statement on the Purpose of a Corporation (19 August 2019), signed by 181 chief executives of major American corporations. The World Economic Forum's Stakeholder Capitalism Manifesto (Davos, 2020).

The wave broke on its first major test. March 2021. Emmanuel Faber, chief executive of Danone, signatory of the Roundtable position and architect of Danone's entreprise à mission conversion, was removed by his board under activist pressure from Bluebell Capital and Artisan Partners. Faber's successor quietly rolled back parts of the mission commitment. Bebchuk and Tallarita (Cornell Law Review, 2020) published a paper arguing that the Roundtable signatories had made no observable governance changes consistent with the statement. The wave, they argued, was rhetoric. The doctrine underneath had not moved.

The 2020s have been the return of the primacy argument. Inflation. Rising interest rates. Private-equity liquidity pressure. Activist campaigns at scale. Larry Fink's BlackRock letter series, which had led the purpose language between 2018 and 2022, has steadily re-centred on returns since 2023 — after the Texas Comptroller's 2022 divestment listing, Florida's near-$2 billion divestment, West Virginia following, and 19 state attorneys general writing jointly to Fink challenging the firm's ESG positioning. The architect of stakeholder-capitalism language proved, under pressure, subject to the same power dynamics as the companies he had been writing to.

Looney's February 2020 pledge was recognisably stakeholder-capitalism language. Auchincloss's February 2025 reset is the clearest major-company test this decade of whether the primacy doctrine has reasserted itself. The reset language is almost entirely returns-based. Capital discipline. Distributions. The phrase used at the investor day — "back to basics" — is not accidental. It is a deliberate signal to a primacy-theoretic audience that the firm is, again, theirs.

Neither ideology survives sustained examination as a clean position. Shareholder primacy, properly understood, contains stakeholder thinking — the long-term shareholder is served by healthy stakeholder relationships. Stakeholder capitalism, properly understood, still needs a residual claimant and a way of resolving trade-offs. The useful question is not which ideology a board holds in theory. It is which one is load-bearing at the moment of a hard trade-off. And at that moment, most firms reveal a position they did not know they held.

The position held at the moment of the trade-off is the position the firm holds. Everything else is a press release.

3. The stack as discipline

The lenses are not six answers to the same question. They are six layers of one answer.

Stack them in order, and each constrains the readings produced by the one beneath it. Freeman names the room. Mendelow tells you where the weight is. Agency-versus-stewardship tells you how the board is trying to govern that weight. ESG is the language the governance is being measured in. The Balanced Scorecard is what puts that language into pay. And shareholder primacy versus stakeholder capitalism is the prior choice underneath all of it — the question of whose interests the scorecard ultimately serves.

The five instrumental lenses above the foundation assume an answer to a question they cannot ask themselves. The sixth lens is that question. Not another layer of analysis. The layer the others rest on.

Apply the stack to the BP reset:

• Freeman — Elliott, LGIM, Norges Bank, Nest, Follow This, ClientEarth, the workforce in three regions, the UK government, communities at named sites. The room.
• Mendelow — Elliott key player; LGIM and Norges Bank moved off the upper-right quadrant; Follow This high-interest, low-power; the UK government high-interest, middling-power. The weight.
• Agency vs stewardship — Auchincloss's reset is agency-theoretic in its responsiveness; Looney's pledge was stewardship-theoretic in its horizon. The governance theory.
• ESG — pre-reset scorecard supported the pledge; post-reset taxonomy supports the reset. MSCI down, Sustainalytics' risk up, ISS ESG climate-leadership withdrawn. The language.
• Balanced Scorecard — pay-design weight migrated from learning-and-growth to financial across 2021–2025. The instrument.
• Primacy vs stakeholder capitalism — "back to basics". The foundation.

Read the chain forward. Elliott used Mendelow's power to impose an agency-theoretic governance shift, legitimised by a rewritten ESG taxonomy, operationalised through a rebalanced scorecard, and justified by a return to shareholder primacy. Each move supported the next. The reset is not a single decision. It is a stack.

The Notes treated the lenses individually. The stack is what they could not show.

3.1 Why six and not three

A reasonable objection to the stack is that six lenses is more than a board needs. Three would do. Stakeholder theory, the Balanced Scorecard, and the primacy-versus-capitalism question — the input, the instrument, the foundation. Why insist on the middle three?

Because the middle three are where strategy is actually contested. Mendelow weights what Freeman lists; without that step, the board treats every stakeholder as equal, which it has neither the time nor the legitimacy to do. Agency-versus-stewardship gives the board the language to describe its own design — what its committee structure is for, what its pay design assumes about the executive, what its surveillance posture says about its trust. ESG is the vocabulary in which the governance theory is rendered legible to the rest of the room. Strip those three out and the board has stakeholder ambition at the top, executive metrics at the bottom, and no traceable chain between them. The middle three are how the chain is built.

Six is also the number at which the stack begins to repay the cost of using it. Three lenses can be held in working memory simultaneously. Six requires a structured composition — which is why most boards never attempt it, and why the stack matters.

3.2 What the stack does that the lenses cannot

A single lens diagnoses a decision. Two lenses, used together, can disagree — and the disagreement is its own diagnostic. Six lenses, stacked, produce a chain in which agreement is not informative and disagreement is. When all six say the same thing, the case is over-determined and the strategic position is clear. When two adjacent lenses contradict, the contradiction is the early signal that the position the board holds in private is not the position the strategy document claims.

The BP reset, viewed through the stack, contains one major adjacent-lens contradiction. Mendelow says Elliott was the binding constraint on the decision. Freeman says Elliott's claim, however structurally powerful, is not ethically privileged over the claims of the workforce or the climate-aligned capital. The board's reading resolved the contradiction by accepting Mendelow's reading and downgrading Freeman's. That is the choice. The other four lenses then ratified it. But the foundational choice was not made at the top of the stack. It was made at the second layer.

4. The predictive frame

Predictive frames in governance are not forecasts. They are conditions — falsifiable propositions that allow the analyst to look at the same case six months later and say what has held and what has not. The frame below sets out, lens by lens, what would have to be true for the BP reset to hold, and what would have to be true for it to reverse. It is offered as a scorecard chairs and NEDs can use against their own decisions, not only against BP's.

4.1 What would have to be true for the reset to hold

Freeman. The downgraded stakeholders remain passive — particularly the transition workforce and the climate-aligned capital. No coalition forms among LGIM, Norges Bank, Nest, and the campaigning shareholders that exceeds the 25% blocking threshold on a material resolution.

Mendelow. Elliott retains its stake. No counter-activist takes the opposite position publicly. The UK government continues to use none of its three available levers — windfall tax, licensing, infrastructure friction.

Agency vs stewardship. The incoming chair is selected from an agency-aligned shortlist. The remuneration committee chair, in particular, signs the next LTIP design without resigning.

ESG. The rewritten reporting taxonomy is accepted by major indices without asterisk. No regulator opens an inquiry on the reporting transition. The MSCI rating stabilises at A or recovers; Sustainalytics' risk classification does not migrate further.

Balanced Scorecard. The 2026 financial metrics are met. Operating cash flow guidance is delivered. Distribution targets are hit.

Ideological. The wider market continues to treat the Business Roundtable 2019 statement as rhetoric, not duty. The 2027 proxy season produces no purpose-language voting threshold.

4.2 What would have to be true for the reset to reverse

Freeman. A downgraded stakeholder acts. The most plausible vector is LGIM filing a resolution, or a Labour-aligned UK government taking the first of its three available policy steps.

Mendelow. A second activist fund takes the counter-position publicly, splitting the upper-right quadrant. The most plausible vector is a purpose-aligned activist — a successor to Engine No. 1's 2021 ExxonMobil model — building a one-to-two percent stake and forcing a board seat.

Agency vs stewardship. The new chair, or the audit committee chair, refuses the agency-theoretic pay design. A nomination-committee selection that produces a chair from a stewardship-style background — a sovereign-fund alumnus, a long-tenure FTSE 100 chair from a capital-intensive sector — would be the structural signal.

ESG. The CSRD, the ISSB, or the FCA downgrades BP's reporting classification. Or — a different signal of the same kind — a major index methodology change that removes BP from a flagship climate index.

Balanced Scorecard. The 2026 financial metrics miss. Operational disappointment erodes the legitimacy of the reset within months of its first anniversary.

Ideological. Purpose-language returns as a voting threshold, not a slogan, at the 2027 proxy season. A twenty-percent-plus vote against a director on stakeholder grounds is the structural signal.

4.3 The 2027 proxy season

The argument lands at the 2027 proxy season. By then, three things will be observable that are not observable now. First, the chair-succession decision. Second, the 2026 full-year operating result and how the scorecard reads against it. Third, the voting pattern across the major institutional holders on any director election or remuneration vote.

If the reset holds, the season will read as quiet. If the reset reverses, the season will read as the moment the alternative coalition formed. The frame above lets the reader convert the 2027 season into a one-page diagnostic.

5. The board-ready scorecard

The frame above, expressed as a scorecard. Print it on one side of A4. Read your own board's decisions against it.

The scorecard is not a forecast. It is a reading instrument. A chair using it should expect the lenses to disagree — the case is more legible when they do. Disagreement among lenses is the early signal that the position the board holds is contested at a level the board has not yet had to articulate.

6. Implications for governance practice

The BP case is a public case. The question of how to read it is also a question about how to read every board's own decisions. Six implications.

Implication 1 — the stack is the unit of analysis, not the lens. A board that uses framework promotion — picking PESTLE for one decision and Porter for the next — is not practising strategy. It is borrowing language. The discipline is composition, not selection. A board using the stack should be able to say which lens it leant on for which judgement, in which order, and how the lenses constrained each other.

Implication 2 — the scorecard is the strategy, twelve months ahead. The BP reset began in the remuneration committee in 2023, not in the boardroom in 2025. The pay-design changes preceded the strategic reset by approximately twenty-four months. A chair reading their own scorecard year-on-year, lens by lens, will read their own strategy before the strategy is written.

Implication 3 — the chair succession is the deepest signal. The succession decision after a reset reveals the governance theory the board has actually adopted, in a way no investor-day announcement can. The agency-aligned shortlist and the stewardship-aligned shortlist will not overlap. The choice is the answer.

Implication 4 — disagreement among ESG raters is the signal, not the noise. Regulators are catching up — the FCA has been consulting on rating-provider oversight since 2023. But for a chair reading the case in real time, the divergence is the early indicator that the rater methodologies are picking up different facets of the same decision. A rating downgrade by one agency and a risk upgrade by another are not contradictory. They are the case showing up in two different lenses simultaneously.

Implication 5 — stakeholder theory is a test of articulation, not of inclusion. Most boards say they consider their stakeholders. The test is what the board says when it has chosen to downgrade a stakeholder. If it can articulate the ethical justification — without retreating to "the shareholders pay our fees" — it is practising the theory. If it cannot, it is managing the public-relations costs of the theory it has actually rejected.

Implication 6 — the ideological foundation is reached last, and most. Most boards never have to articulate the layer underneath the other five. The ones that do are the ones operating under sustained pressure. When the moment comes, it is too late to hold the conversation in committee. The chair who has rehearsed the ideological position privately, before the moment forces it, will hold ground better than the chair who has not.

Implication 7 — time horizon is a governance variable, not a strategy variable. The BP case turns on the difference between a fifteen-year transition thesis and a thirty-six-month activist thesis. A board that does not declare its time horizon as a board variable — and hold it consistently across pay design, scorecard, and chair-tenure expectation — is a board that has implicitly delegated the variable to whichever stakeholder pulls hardest at any given moment. The horizon, like the foundation, has to be held in advance. It cannot be improvised under pressure.

7. Three propositions

A reading. The six anchor texts are: Friedman, The Social Responsibility of Business is to Increase its Profits (1970); Freeman, Strategic Management — A Stakeholder Approach (1984); Jensen and Meckling, Theory of the Firm (1976), with Donaldson and Davis (1991) as counter-pole; the IFRS Foundation's ISSB S1 and S2 (2023); Kaplan and Norton, The Balanced Scorecard (1992); and Bebchuk and Tallarita, The Illusory Promise of Stakeholder Governance (Cornell Law Review, 2020). Six texts, two hours of reading, and the stack is in place.

A question for your board. If you applied the six lenses to the decision in front of your board this month — not in sequence, but as a stack — where would the lenses contradict each other, and which contradiction would you have to resolve first?

A position to take. The strategic position your board holds is not the position written in the strategy document. It is the position the board reveals at the moment of a hard trade-off. The stack is how you find that position before the moment arrives. The chair's job is to make sure the position the board holds in private is the position the board can defend in public.

Annexes — the seven companion Notes

Each of the six lenses, and the synthesis, is treated at length in a companion short-form Note in the Strategy Room. The Notes present the same material in a register designed for a single-sitting watch. The whitepaper is the consolidated read. The Notes are how the argument circulates.

Annex A — Note 09 · Freeman's Stakeholder Theory. A 7:30 walkthrough of Freeman's 1984 argument and the Donaldson–Preston (1995) decomposition into three theories — descriptive, instrumental, normative. Applied to the BP reset.

Annex B — Note 10 · Mendelow's Stakeholder Matrix. A 4:55 explanation of the 1991 power–interest matrix, with BP plotted twice — February 2020 and February 2025. The motion of stakeholders between quadrants is the analytical content.

Annex C — Note 11 · Agency Theory vs Stewardship Theory. A 6:40 contrast of the two theoretical traditions, with 3G Capital and Unilever-under-Polman as the contrast pole. BP under Lund-Looney as stewardship; BP under Lund-Auchincloss as agency.

Annex D — Note 12 · ESG Framework. A 6:00 history of the measurement language since the 2004 Who Cares Wins report, with the BP rating moves of February 2025 as the worked case.

Annex E — Note 13 · Balanced Scorecard. A 6:15 account of Kaplan and Norton (1992), Mobil NAM&R (1994), and the year-by-year migration of weight in BP's executive scorecard from learning-and-growth into financial across 2021–2025.

Annex F — Note 14 · Shareholder Primacy vs Stakeholder Capitalism. A 6:20 treatment of the two ideological poles — Friedman, Dodge v. Ford, Revlon, eBay v. Craigslist; against Freeman, the Business Roundtable, Faber and Danone, Bebchuk and Tallarita.

Annex G — Note 15 · BP — The Strategic Stack (Synthesis). A 9:35 stack of the six lenses, with the predictive frame as the closing.

The Notes are the entry point. The whitepaper is where the entry leads.