Two Names, One Signature: Decision Rights for Firms That Cannot Afford a Tie

Most partnership deeds in two-principal advisory firms are drafted at the point of maximum optimism. Two people who have worked alongside each other for years, who trust each other's judgement and who, at the moment of signature, struggle to imagine a genuine impasse, commit to a document that deals carefully with profit share, capital, retirement and restrictive covenants — and says almost nothing about what happens at 9am on the Monday after the first irreconcilable disagreement.

That is less carelessness than a category error. The deed is treated as a record of an agreement already reached, rather than as a mechanism for reaching agreements that have not yet been tested. In a firm of thirty, disagreement is a governance problem to be managed. In a firm of two, it is an existential one.

The statutory default assumes an odd number

Partners who leave the point unaddressed do not fall into a vacuum; they fall into the Partnership Act 1890. Section 24 provides, among other things, that every partner may take part in management, that differences arising as to ordinary matters are decided by a majority, and that no new partner may be introduced without the consent of all existing partners. For a firm of three or five that is workable enough. For a two-name firm it is close to unworkable: a majority of two does not exist, so the statutory fallback produces a deadlock at precisely the moment a decision is required. Limited liability partnerships fare little better under the default provisions of the LLP Regulations 2001, which replicate much of the same logic.

The remedy is rarely a change of legal form. It is a decision architecture, written down and agreed while the relationship is still comfortable.

Begin with a decision register, not a clause

The practical starting point is a single schedule — sometimes called a decision register or a matters matrix — that lists every category of decision the firm actually takes and assigns it a route. Hiring and dismissal above a salary threshold. Taking on or exiting a lease. Drawing down or extending borrowings. Admitting a lateral principal or a salaried partner. Changing the fee model or the lock-up policy. Accepting a client that sits outside agreed risk appetite, or one that would create a conflict with an existing mandate. Settling a professional negligence claim. Altering the profit-sharing key.

The register forces the partners to confront the decisions that have historically been resolved by whoever felt most strongly at the time. It also exposes the categories where the pair have genuinely different instincts — which is the point.

Reserved matters and thresholds

Reserved matters should be short, clearly defined and unanimous. A schedule that reserves twenty items is not a governance document but a recipe for paralysis. Most two-principal firms operate comfortably with six to ten, anchored to financial thresholds that are reviewed annually against fee income rather than fixed in perpetuity. A capital commitment above a stated sum, a client engagement carrying fees above a percentage of turnover, or any borrowing should ordinarily qualify.

Thresholds matter as much as the matters themselves. Without them, the reservation covers everything and delegation becomes a fiction; with them set too high, the firm drifts into decisions that one principal would never have approved had they been consulted.

Delegation, and the problem with casting votes

Below the reserved threshold, decision-making should be delegated — normally to a managing principal for a defined term, with the scope recorded in a short side letter or in the deed itself. Rotation of that role, whether every two or three years, reduces the risk that delegation hardens into hierarchy in a firm that is meant to be equal.

The casting vote is the mechanism most often reached for and most often regretted. It looks tidy on paper and behaves badly in practice: it converts every close call into a win and a loss, and it gives one principal a permanent structural advantage that the other will come to resent. Where an odd-numbered decision body is possible — a three-person management committee with an independent chair, for example — it is generally a better answer than a tie-break clause. Where it is not, a casting vote should be expressly confined to ordinary course matters and excluded from anything touching remuneration, capital or the identity of the firm's clients.

Escalation ladders

A workable deed sets out a sequence for matters that cannot be agreed, and gives each rung a timetable. In outline: the principals exchange short written positions within five working days; failing resolution within a further ten, the matter is referred to mediation before a named body such as CEDR, with costs shared equally; and if mediation fails, the dispute goes either to expert determination, where the question is valuation or accounting, or to arbitration or the court, where it concerns legal rights. Specifying the forum in advance removes one further decision from the deadlock itself.

Deadlock triggers and the pre-agreed exit

The clause that most often goes missing is the one that defines deadlock and attaches a consequence. A serviceable definition runs along these lines: a reserved matter has been tabled twice within a stated period and has failed to obtain the required unanimity on both occasions. At that point the buy-sell is engaged on pre-agreed terms.

Those terms should settle four questions. First, valuation method — commonly fair value determined by an independent expert, with the basis stated expressly: maintainable earnings multiple, pro-rata share of net assets, or a formula tied to fee income and lock-up. Ambiguity here is what turns an orderly separation into litigation. Second, whether any minority or marketability discount applies, which for an equal two-name firm it usually should not. Third, payment mechanics: a completion payment, deferred consideration over two or three years, and security for the balance if the departing principal is leaving over a genuine disagreement rather than retirement. Fourth, treatment of work in progress, client retention payments and the allocation of any liabilities notified after completion.

Mechanisms such as Russian roulette or a sealed-bid shoot-out have their place, but they suit firms of near-identical financial strength. Where one principal is materially better capitalised, a first-mover buy-sell can be used as a lever rather than a resolution, and the deed should be tested against that scenario before signature. It is also worth confirming with advisers how any payment will be characterised for tax purposes, and whether the firm's regulator requires notification of a change in control or approved persons.

A checklist before the next partner meeting

• Is there a written decision register covering hiring, capital, borrowing, client acceptance and remuneration?
• Are reserved matters limited in number and anchored to financial thresholds reviewed annually?
• Is delegation to a managing principal documented, time-limited and revocable?
• Does any casting vote exclude remuneration, capital and client matters?
• Is the escalation ladder timetabled, with a named mediator and a specified forum?
• Is deadlock defined with precision, and does it trigger a buy-sell automatically?
• Is the valuation method, discount position and payment schedule agreed in advance?

None of this is a comment on the strength of a particular partnership. It is simply an acknowledgement that a two-name firm has no tie-breaker but the one it writes for itself.