Release capital in stages and measure what justifies the next line

The operating record should let a reviewer follow a garment from production to sale, collection and debt service without changing denominator or period along the way. GFT proposes shift-level records of good output, rejects and rework, weekly review of collections and stock, and monthly reconciliation of cash flow, cash available for debt service, debt-service coverage, minimum cash, covenant headroom and remaining funding availability. These cadences are recommendations for management control, not universal lender requirements. Contractual reports should follow the executed agreements and use definitions that reconcile to the operating record.

Every metric needs a compact dictionary: unit, numerator, denominator, period, owner, source system, cut-off and reconciliation. Good output should mean accepted units under a stated quality rule, not all units that crossed a machine. Collections should identify the invoice, payer, amount, value date and deductions rather than rely on a sales forecast. Stock should distinguish raw material, work in progress, finished goods, rejected items and goods already allocated to a customer. Cash should distinguish unrestricted balances from controlled accounts and amounts reserved for a defined obligation.

The monthly review should begin with physical and commercial reconciliation before calculating coverage. Accepted output connects to dispatch records, dispatch to invoices, invoices to collections and deductions, and collections to bank statements. Operating payments connect to the same period's production and stock movements. Only then does the finance team calculate cash available for debt service under the stated definition. New borrowing and sponsor injections remain financing flows, not operating repayment capacity. A growing bank balance can therefore coexist with weak operating cash if its source is another capital injection.

The proposed dashboard in Exhibit 63 places factory signals and financing consequences beside one another. It does not populate performance values because none has been observed for this proposed venture. Its purpose is to force each reported number to carry an owner and a reconciliation path. Where a measure is unavailable, the report should say so and identify the decision affected. A blank or estimated field should not be converted into an apparently precise zero.

Management should review exceptions in the same cycle as the headline measures. A sudden yield improvement may reflect a changed acceptance rule. Faster collections may reflect one prepayment rather than repeatable buyer behaviour. Lower stock may indicate an impending material shortage. The reviewer should ask what operational event explains each movement and which independent record supports that explanation. Trend commentary without this challenge can make the dashboard look disciplined while concealing a definition change.

The manual performance-management system is therefore part of the financing infrastructure, not a temporary inconvenience before digitisation. It establishes the identifiers, responsibilities and reconciliations that a later digital layer can automate. Software cannot repair an unclear acceptance boundary or an unrecorded adjustment. GFT's recommendation is to test the manual chain during commissioning, retain correction history and introduce automation only where it preserves those controls. The resulting record serves operations first and financier reporting through a controlled extraction, rather than maintaining two inconsistent versions of performance.

A strong debt-service ratio cannot compensate for an unsafe or environmentally unmanaged line. GFT proposes a monthly review of relevant resource use, safety events, worker concerns and environmental and social action-plan progress beside the cash dashboard. The precise indicators should follow the actual process boundary and risk assessment. IFC's Performance Standard 1 includes management programmes, organisational capacity, emergency preparedness, stakeholder engagement, and monitoring and review within the management system. A financing report should show those activities as continuing responsibilities, not as documents completed at close.[S22]

Each action needs an owner, due date, budget, closure evidence and, where necessary, verification of effectiveness. Closing a training action means more than recording attendance if the remaining risk concerns whether supervisors apply the procedure. Closing a metering action means more than installing a device if shared loads remain unreconciled. Where an action is late, the report should state the operational and cash consequence and identify whether the delay affects a capital gate. Essential corrective work should remain funded in the downside case.

Use-of-proceeds reporting follows a separate evidence chain. ICMA's Green Bond Principles connect green issuance to projects with environmental benefits and emphasise reporting that permits proceeds to be tracked. That supports an allocation control from the financing facility to eligible invoices and outstanding financed shares. It does not turn an allocation into evidence of cash-flow performance or proof that every factory activity is green.[S15] The green register and the credit dashboard should reconcile where they share an invoice, while retaining their distinct purposes.

Opened green-loan guidance distinguishes estimated from actual impact and calls for clear disclosure of data scope and limitations. GFT would carry that distinction into every capital gate. Before operation, an engineering estimate can support appraisal if its assumptions remain visible. After operation, measured values should replace the relevant estimate through a documented bridge rather than overwrite the original file. Where actual impact cannot yet be calculated, the report should retain that limitation instead of presenting forecast benefit as delivered benefit.[S17]

Sustainability-linked reporting requires its own determination. Opened loan guidance connects verification to the dates or periods that can change financial or structural characteristics, and calls for the report to describe procedures, criteria and limitations. The financing administrator should therefore use the verified result specified in the agreement, not an informal dashboard extract. A missed target, an unavailable result and a qualified verification conclusion are different states with consequences that must come from the executed documents.[S18]

Carbon claims need an explicit accounting boundary. The GHG Protocol Corporate Standard addresses a corporate emissions inventory and states that it is not the method for quantifying offset credits; it also does not itself prescribe how verification should be conducted. GFT would therefore keep corporate emissions, product footprints, avoided emissions and any credit revenue as separate records. None should enter the cash case or an expansion announcement merely because another environmental indicator improved.[S25]

Development outcomes require the same restraint. IFC's AIMM guidance distinguishes expected development impact from additionality and describes monitoring against expected outcomes over time. GFT's proposed report would separate financed outputs, observed operating results and wider outcomes that require an attribution argument. It would not issue itself an institutional impact score or claim that local employment and market effects are caused solely by the first line.[S26] This status discipline gives the capital committee a credible choice: an assured result can support the decision within its scope, while an estimate remains a reason to test and measure.

A conditional recommendation is useful only if readers can see what would reverse it. GFT proposes a decision memorandum that records the approved scope, capital at risk, evidence accepted, exceptions, downside liquidity, responsible roles and the conditions for the next review. It should state why the decision is proceed, hold, redesign or stop, and what new fact would move it to another state. This makes disagreement inspectable and prevents a later narrative from replacing the original basis of approval.

Decision authority should be equally visible. The memorandum should identify who may accept evidence, who may release funds and who may approve a waiver or redesign. A reviewer should not waive the very condition that protects their own delivery responsibility without the required independent authority. Any waiver should retain the original condition, the reason for departure, the additional risk accepted and the funding consequence. Where documents do not permit a waiver, management should treat the condition as binding rather than create an informal exception in the dashboard.

The illustrative path in Exhibit 65 keeps evidence review separate from the capital decision. Proceed authorises only the defined release. Hold preserves the boundary pending a specified remedy. Redesign returns scope, funding or operating assumptions to appraisal. Stop limits further discretionary exposure and initiates the agreed protective actions. A sponsor funding shortfall or unresolved safeguard can therefore change the decision even when the factory's technical report is favourable.

The refresh register should be published internally with the recommendation. Before any transaction, it should cover lender products and eligibility, interest and foreign-exchange quotations, local law, permits, tax, insurance, environmental baselines, supplier terms and buyer contracts. Each item needs an owner, source date, expiry or review trigger and a model destination. A changed quotation should update the cash model and gate decision, not remain in a correspondence folder. A changed law or permit requires qualified review rather than an editorial assumption.

External reporting should use the same status controls. Approved finance is not drawn cash, installed capacity is not accepted output, estimated impact is not measured impact, and an operating result is not automatically a wider development outcome. Public statements should identify the evidence boundary and avoid implying institutional endorsement. Confidential commercial records can remain protected while the decision team retains a complete audit trail.

GFT's closing recommendation is therefore conditional. Fund a bounded first stage completely. Match the route to readiness. Release against accepted evidence. Manage cash, production and safeguards in one review cycle. Expand only when repeated records support the incremental liability, and stop or redesign when a critical assumption fails. The method does not promise a lender, a subsidy or a second line. It defines what must be true before any of them can responsibly support the next decision.