When a document requiring execution begins circulating inside a company, the question of who may sign it is usually resolved by asking a person rather than by consulting a record. The person asked ordinarily sits on the finance side, and the answer comes quickly and accurately, drawn not from institutional memory but from personal recollection of who has signed what over the years. This arrangement produces no friction in daily operations — it reduces friction — which is precisely why no one records it as a deficiency. In the same company, a signature circular issued three years earlier remains in circulation, a general power of attorney granted for a particular project stays in force long after that project has closed, and the divergence between the list of authorised signatories held at the bank and the board's most recent authority resolution appears on nobody's agenda.
At the review table this pattern surfaces on the first day. The diligence side does not ask about the delegation regime as such; it requests the schedule of contracts binding on the company, then reconciles the execution pages of those contracts against the underlying authority instruments. The picture produced by that reconciliation points, almost invariably, in the same direction: authority resides in one person on paper and in another in practice, the scope of a power of attorney fails to cover the transaction actually executed under it, or the grant remains valid while the monetary threshold set at the date of grant sits far below the size of transactions now being signed. The question the company has never put to itself is the first question diligence asks, and delay in answering is itself logged as a finding.
What is described as a delegation regime is not one document but a set of records required to corroborate one another: the board resolution on representation and binding authority, the authority registered with the commercial registry, notarised powers of attorney, the mandate held at each bank, and the internal delegation matrix. The dispersal of that set arises less from neglect than from the circumstances in which authority is typically conferred. Grants are made under urgency — the founder is abroad, a tender deadline is approaching, a bank requires an instrument by the following morning — and authority conferred under urgency is drafted broadly rather than narrowly, since a narrow drafting carries the risk of a second delay. The later narrowing of a broadly drafted grant, however, never reaches anyone's agenda.
The mechanics of that asymmetry are straightforward. Withdrawing an authority risks being read as a signal of diminished trust in the individual holding it, and therefore carries an immediate, personalised cost, whereas the cost of leaving the authority in force is deferred, diffuse, and at the moment of decision entirely abstract. A decision-maker weighing those two costs against each other and selecting the second is behaving rationally over the short horizon. The difficulty lies not in the shortcut itself but in its persistence after the conditions have changed: as the company grows and as transaction sizes and counterparty counts increase, the same grant begins to carry a materially different risk, while the instrument conferring it remains the instrument drafted years earlier.
A second drift runs parallel to the first, opening between formal authority and earned standing. In most companies the person named in the instrument is not the person who in fact makes the decision; the signature records the decision rather than constituting it. At modest scale this separation produces no difficulty and may well add speed. Once authority and decision-making diverge, however, accountability diverges with them: when a transaction goes wrong, an institutionally unowned space opens between the signatory's account — that instructions were given — and the decision-maker's account — that no signature was applied. The reviewing party reads that space not from individual transactions but from the mismatch between the delegation matrix and the organisational chart.
The finding reaches valuation through deal architecture rather than through headline price. A contract executed outside the authority chain, even where subsequent ratification is legally available, adds a line to the conditions-precedent list; and where ratification requires counterparty consent, it hands that counterparty negotiating leverage it did not previously hold. Due authorisation is a standard head in any representations and warranties package, and every contract entering the disclosure schedule against that head extends the seller's survival period, drives the escrow percentage upward, and hardens the triggers in any earn-out structure. Where warranty and indemnity insurance is in play, underwriting will frequently carve this head out of policy coverage altogether, and risk left outside the policy is balance-sheet risk directly.
The second channel is the calendar. Reconstituting the authority chain — a board resolution, a registry filing, updated bank mandates, and where required the ratification of affected contracts — looks short when set out on a timeline, but because the steps run sequentially and each depends on the processing speed of a third institution, it can push closing back by a meaningful fraction of a budget cycle. A deferred closing is not merely lost time; it reopens the validity periods of financing commitments, the fee structures attached to them, and the reference periods governing price adjustment mechanisms. Price, more often than not, moves in that reopening rather than in the original negotiation.
In financed projects the channel is more direct still. Signature authority at the project company level is fixed as a schedule to the credit agreement, with drawdown requests, progress certificates, and engineer's certifications tied to named individuals, and the currency of that list becomes a covenant head in its own right. Where a dispersed delegation regime at the parent is carried across into the project company unchanged, a drawdown package returned for want of a valid signature translates directly into delay in the construction programme. The lender's response at that point is typically procedural rather than penal, but the procedure itself — a supplementary instrument, an additional approval, a further verification round — permanently reduces the velocity of every subsequent drawdown.
The structure that neutralises this tendency has four components. The first is the anchoring of the delegation matrix to monetary and risk thresholds rather than to job titles; "the general manager signs" describes a habit rather than defining an authority, whereas "procurement contracts below a stated amount require a single signature, those above it two signatures and notification to the board" is a rule capable of being tested. The second is a single-source authority inventory in which the basis, scope, monetary threshold, commencement date, expiry date, and revocation record of every grant sit in one table, with open-ended grants treated as the exception. The third is the limitation of powers of attorney by scope, duration, and purpose. The fourth is the reconciliation of the internal matrix against the registry, banking, and contractual surfaces at a defined cadence, since absent that reconciliation the five surfaces age independently.
BEIREK's intervention in this area does not begin with the drafting of an authority policy; it begins with the extraction of the existing authority inventory. Every power of attorney in force, every registry entry, every board resolution, and every list of authorised bank signatories is consolidated into a single table, after which the execution pages of recently signed contracts are read against that table; the resulting schedule of divergences is a list of risks to be prioritised rather than a list of items to be corrected. The delegation matrix built on top of it is calibrated to the company's actual distribution of transaction sizes — thresholds derived from two years of executed volume rather than from a theoretical grid, since a matrix set otherwise will either be breached continuously or bind nothing at all.
The second layer is cadence. The authority record is maintained at the point of proposal rather than at the point of approval; every grant carries an expiry date and a named owner, and that owner is not the person exercising the authority. Three indicators are read at regular intervals: the proportion of transactions executed outside the matrix relative to total transactions, the average lag between a board resolution and its registry filing, and the number of powers of attorney that have expired without formal revocation. The continuity dimension is tested through a single question — who deputises for the deputy — since whether a secondary authority is defined in advance for the moment the primary signatory is unavailable indicates, on its own, whether the regime has been built around a person or around a role.
What determines a company's valuation is more often the demonstrability of performance being reproducible independently of its founder than the performance itself; the delegation regime is the plainest and earliest-legible form of that demonstration, because where the answer to who may bind the company concentrates in a single individual, every remaining indicator of institutional maturity is read in that individual's shadow.