We separated the roles. We told ourselves that fixed the problem. It didn’t.
Walk into almost any governance conversation and you’ll hear the same article of faith - split the chair and the CEO, and you protect shareholders from concentrated power. Independence is good governance.
The world has largely agreed.
- Globally, 76% of countries with one-tier boards now require or encourage separating the two roles - up from just 44% in 2014. [1]
- In the UK, the debate is over around 95% of the FTSE 100 separate the roles, the legacy of Cadbury (1992) and Higgs (2003). [2][3]
- In the US, combined chair/CEO roles (“CEO duality”) fell from 47% of the S&P 500 in 2020 to 42% in 2025, while independent chairs rose from 33% to 39%. [4]
- In Australia, the ASX Principles recommend the chair be independent and not the CEO - and the ASX 200 complies almost universally. [5]
Case closed?
Not even close.
The uncomfortable evidence
Here’s what doesn’t make the press release, decades of research have failed to show that separating the roles reliably improves performance. [6]
The academic verdict is inconclusive. Splitting the roles improves the optics of governance. It does not consistently improve outcomes.
One study of FTSE 350 firms through the 2007–09 crisis found something close that tighter compliance with the UK Governance Code was negatively associated with firm survival, while having an insider CEO improved the odds. [7]
The companies ticking the most boxes were not the ones most likely to survive.
When the split actually works
The point isn’t that separation is pointless. Done with a genuinely independent, capable chair, it is powerful.
Apple is the textbook case. From 2011, Tim Cook ran the company as CEO while Arthur Levinson chaired the board as an independent director. Over that period Apple became the most valuable company on earth - and built one of the most admired succession plans in business. From September 2026, Cook moves to executive chairman and John Ternus becomes CEO, with Levinson shifting to lead independent director. [8] Even as Apple tilts toward a more combined structure, it deliberately preserves an independent counterweight. The safeguard is the function, not the title.
Commonwealth Bank is Australia’s version. After the AUSTRAC scandal and the Hayne Royal Commission, an independent chair, Catherine Livingstone, and a new CEO, Matt Comyn, rebuilt the country’s largest bank from the boardroom out - a remedial plan agreed with the regulator, seven new independent directors, and a wholesale culture reset. [9] Separation worked because the chair engaged, challenged, and held management to account.
And combination can work too. Jamie Dimon has been both chair and CEO of JPMorgan since 2005 - exactly the structure codes warn against - and the bank just posted record profits and an all-time-high share price. [10] But notice the asterisks: analysts openly price a “Dimon premium,” with one estimating the stock could fall ~5% the day he leaves, and JPMorgan has confirmed its next CEO will not also be chair. [11] Duality worked because of the person, not the structure.
When it fails - on both sides of the Pacific
Now the other column. And it’s long.
Combined roles, no brake. Boeing held the roles together through the 737 MAX crisis: 346 lives lost, more than US$18bn in losses. [12] In the US, Wells Fargo’s John Stumpf was chair and CEO while the bank opened millions of fake customer accounts; regulators fined it, he resigned, and only then did the board split the roles and appoint an independent chair. [13]
Separation on paper, capture in practice. This is the trap. Tesla has a separate chair - Australian Robyn Denholm - and a CEO, Elon Musk. Yet a Delaware court voided Musk’s earlier pay package, finding the board’s process “deeply flawed” and lacking independence given directors’ close ties to him. In November 2025, shareholders nonetheless approved a new package worth up to ~US$1 trillion, with about 75% support. [14] The titles were separated; the independence was not. Theranos had a board of statesmen - Kissinger, Shultz, Mattis - and not one understood diagnostics. They never tested the technology. Investors lost the bulk of ~US$700m. [15]
Founder power that overwhelms the board. And sometimes the structure is impeccable and irrelevant. In 2024–25, both WiseTech and Mineral Resources had independent directors and a separate chair - and in both, a dominant founder prevailed anyway. At WiseTech, four independent directors including the chair walked out in February 2025, briefly wiping ~A$10bn. [16][17] That isn’t really a structure problem. It’s a personality problem - and it’s the subject of Part 2 of this series.
In every one of these failures, the org chart was beside the point. The chair either couldn’t, or wouldn’t, challenge power.
What we should actually be measuring
Shareholders keep pushing on structure. US proposals for an independent chair jumped 113% in early 2023 - yet average support has been stuck around 28–35% for a decade. [18] Investors sense the structure matters, but aren’t convinced it’s the answer. They’re right.
The variables that actually predict whether a board spots the iceberg don’t fit neatly into a listing rule:
- Does the chair have the competence to understand the business well enough to challenge it?
- Does the chair have independence of mind - not just the independence box ticked - to say no?
- Is power so concentrated in a founder that the board’s formal rights are theoretical?
- Does the boardroom actually welcome dissent, or quietly manage it away?
None of these appear on the cover of a governance report. All of them decide whether the company survives its next crisis.
The provocation
Separating the chair and CEO is necessary. It is nowhere near sufficient.
Apple and CBA show what the split delivers when the chair is genuinely independent and engaged. Tesla, WiseTech and Theranos show what it delivers when they aren’t: nothing. The reform we’ve spent twenty years perfecting is a proxy - and we keep auditing the proxy instead of the thing itself.
The org chart is the easy part. The hard part - a chair with the skill and experience to challenge a powerful CEO, and a board with the literacy to know when they should - can’t be mandated. It has to be chosen.
So if your board has split the roles that is good, but now answer the harder question.
Could your chair actually stop your CEO - and would they?
Because once the boxes are ticked, the real risk isn’t the structure. It’s the people sitting inside it - the steamrolling CEO, the chair who won’t let go, the two egos who’d rather fight than yield. That’s Part 2: “My Way or the Highway.”
Written from an Australian board and governance perspective. Agree, disagree, I’d genuinely like to hear it.
References & further reading
- OECD, Corporate Governance Factbook 2025 - separation of chair and CEO across jurisdictions (76% require or encourage, up from 44% in 2014).
- The D&O Diary / Millstein Center (2009) - FTSE 350 combined-role prevalence (~5%); academic data showing ~95% of the FTSE 100 separate the roles.
- Cadbury Report (1992); Higgs Review (2003); UK Corporate Governance Code 2024 (Financial Reporting Council).
- The Conference Board / ESGAUGE, CEO/Chair Leadership (2026), via the Harvard Law School Forum on Corporate Governance - S&P 500 duality and independent-chair trends.
- ASX Corporate Governance Council, Principles and Recommendations (Recommendation 2.5).
- Goergen, Limbach & Scholz-Daneshgari on CEO duality and shareholder value - summarised in IE Insights, “CEO Duality: For Better and for Worse.”
- Journal of Corporate Finance (2015) - UK Corporate Governance Code compliance and FTSE 350 firm survival through the 2007–09 crisis.
- Apple newsroom and reporting (April 2026) - John Ternus to become CEO on 1 September 2026; Tim Cook to executive chairman; Arthur Levinson (independent chair since 2011) to lead independent director.
- Commonwealth Bank newsroom (2022) - Catherine Livingstone’s chairship (2017–2022), the Remedial Action Plan, board renewal and culture reset under CEO Matt Comyn.
- JPMorgan Chase Q1 2026 results and SEC filings.
- CNBC and Yahoo Finance (2026) - JPMorgan succession; the “Dimon premium”; confirmation the next CEO will not hold the chair.
- The New York Times reporting on the Boeing 737 MAX; Boeing disclosures.
- Reporting and regulatory findings on the Wells Fargo unauthorised-accounts scandal (2016) and the subsequent separation of the chair and CEO roles.
- Delaware Court of Chancery rulings on Elon Musk’s 2018 Tesla pay package; CNBC (Nov 2025) - shareholders approve a new package (~US$1tn) with ~75% support; Tesla SEC filings (Chair Robyn Denholm).
- Wharton, Darden and Yale SOM case analyses of Theranos governance.
- Reporting on WiseTech board resignations, February 2025 (Sydney Morning Herald, Fortune, Startup Daily).
- Reporting on Mineral Resources governance and the Chris Ellison succession, 2024–25 (The Nightly, Mining.com).
- ISS Corporate Solutions (2023), via the Harvard Law School Forum on Corporate Governance - independent-chair shareholder proposals and support levels.
Statistics reflect the most recent figures available at the time of writing (2025–2026). Private-market figures are indicative rather than exhaustive.
