The operating chair

Every board that buys advice takes a second decision hidden inside the first: whose vantage the advice is written from. Two chairs produce documents that look alike. Only one owns the consequence — a P&L that must close, a payroll that must clear, a licence to operate that a regulator can suspend. The observer's chair owns the document: it analyses, benchmarks and recommends, and when the recommendation fails it files a lesson learned. The question is not whether observers are intelligent.

Our position, written into how this firm works, is that decision advice — a recommendation that commits someone else's payroll, balance sheet or licence — is only trustworthy when it is authored from the operating chair, by people who have owned that class of consequence. Analysis can be produced honestly from the observer's chair. Advice cannot, and the two should never travel unlabelled in the same document.

The claim is deliberately narrow. We do not argue that operators are always right; operators fail conspicuously, and the liability record proves it. We argue that the legal architecture, the governance codes and the judgment research all point at the same seat, and that a board can test cheaply which chair its advisers write from.

Start with where the law puts the consequence. Under the OHADA Uniform Act on commercial companies — the company law a business in Abidjan, Dakar or Lomé actually lives under — liability for management fault is personal and attaches to the seat.

The governance codes make the same allocation from the other direction: they refuse to let accountability travel with delegation. A board may buy all the advice it likes; the codes return the consequence to the boardroom unopened.

The judgment research explains why the seat matters for the quality of the advice, not merely its fairness. The operating chair is structurally the seat that receives the feedback: the missed payroll, the covenant call, the regulator's letter arrive there and nowhere else. An observer studies such events; the operator has been corrected by them.

Assembled, the three point one way: the law fixes consequence to the seat, the codes forbid it to leave the boardroom, and the research says trustworthy judgment is built by consequence feedback. Our position, graded as such: advice authored from the observer's chair escapes both the liability architecture and the learning loop, so a board should treat it as analysis until the author's owned consequence is shown.

The distinction stops being abstract in our corridor. A European holding board authorising capital, covenants or an exit for an operating company in Abidjan or Dakar is deciding about seats it can name: a managing director whose signature engages the company under OHADA law and whose personal exposure under Article 161 is not theoretical; a payroll cleared in CFA francs under the BCEAO's currency arrangement; a licence held from a francophone regulator. Most of the advice arriving at that board is written a long way from those seats. The observer's question is which structure is optimal; the operating question is which failure this payroll can absorb, and it produces different sections — thresholds rather than scenarios, exit conditions rather than upside cases, the named seat that acts when the assumption breaks.

Evidence cards

Operators carry documented biases of their own — sunk commitment, survivorship, proximity — and observer analysis remains the correct instrument for many questions. Where this firm has not owned the class of consequence at stake, the claim is graded down and the grade is printed.

Sources and limitations

Sources and limitations. The institutional facts in this note rest on the text of the revised AUSCGIE (2014), the King IV report and the UK Corporate Governance Code 2024, each cited with its caveat; the judgment condition rests on Kahneman and Klein (2009), a peer-reviewed synthesis that quantifies nothing about advisory work. The step from those anchors to the operating-chair rule is the firm's position, graded SEG-2: an interpretation consistent with the cited architecture, not a measured result — we cite no dataset comparing outcomes of operator-authored and observer-authored advice, because we found none meeting our standard. The boundary is explicit: the position covers decision advice only; analysis, benchmarking and description are legitimately produced from the observer's chair, and operators carry documented biases of their own. This note is not valid as legal advice on AUSCGIE liability, nor as a claim that any named class of adviser is negligent.

1. The Chair requires every adviser to state in writing what consequence the author has personally owned — P&L, payroll or licence — and when. If the answer is a methodology rather than a seat, reweight the document as analysis.

2. The Company Secretary labels analysis and advice separately in every board paper before it is circulated. The two are checked differently: analysis against sources, advice against the author's owned consequence. A paper that blends them is asking not to be checked.

3. The Audit Committee Chair records in the minutes what the adviser loses if the recommendation fails. Fee structures that survive failure are a reading instruction.

4. The managing director who carries the Article 161 class of liability reads the paper before the board votes. If the consequence-bearing seat has not seen the advice, the board is choosing on that seat's behalf with the observer's eyes.