What a Development Finance Institution Tests That a Commercial Lender Does Not

Every Commercial Test Still Applies, and Then a Second Set Begins

A development finance institution is a lender. It assesses repayment capacity, tests cover across the tenor, examines security and sets covenants, and it does all of that at least as rigorously as a commercial bank — frequently more so, because its tenors are longer and its exposure therefore runs further into a future nobody can forecast. Borrowers who expect a softer credit assessment because the institution has a development purpose have misread it entirely. The commercial tests are the entry condition, not the assessment.

What distinguishes the institution is what happens after those tests are passed. A DFI, a multilateral lender or an impact-linked fund operates under a mandate that is not simply to lend profitably, and that mandate generates a second screen the borrower must also clear. Would this project have proceeded without us, and if so what are we adding? What are the environmental and social consequences, how have they been assessed, and are they being managed to a standard the institution can stand behind? What will change in the world as a result of this money, how will that be measured, and against what baseline? How will the money be spent, and through what procurement? Each of those is capable of stopping a proposal that a commercial lender would have approved without hesitation.

The consequence is a category of outcome that borrowers find genuinely confusing: a proposal that is comfortably bankable and is nonetheless declined, sometimes with an observation that the applicant should approach a commercial lender. That is not a criticism of the project. It is very close to a compliment about it — a project that can raise commercial finance on reasonable terms is, for many development institutions, a project they should not be crowding out. Understanding that inversion is the beginning of preparing for this reader.

This article sets out the tests that are additional: additionality itself and why it is assessed first, environmental and social safeguards and what categorisation determines, results frameworks and the discipline of measurable indicators, procurement standards and where they bite, what a long tenor does to a sensitivity analysis, and the dynamics of co-financing where a DFI sits alongside commercial lenders. Where our transaction advisory practice prepares documentation for this audience, it is built to carry both screens, because a document that carries only the commercial one is a document that will be assessed against criteria it never addressed.

The Opening Question Is Whether the Project Needs This Institution at All

Additionality is the proposition that the institution's participation changes something — that the project either would not have proceeded without it, or would have proceeded on materially worse terms, at smaller scale, later, or with weaker standards. It is assessed first because it is dispositive: an institution whose mandate is to address gaps that private capital does not fill has no business displacing private capital that was willing to act. Where additionality fails, nothing downstream is reached.

Borrowers commonly answer this question badly, and the characteristic failure is to describe the project's merit rather than the institution's contribution. A submission that explains how important the project is, how many people it will employ and how much it will produce has answered a question about impact, not about additionality. The question is counterfactual: what happens in the world where this institution declines? If the answer is that the borrower approaches a commercial bank and obtains similar terms with modest additional effort, the additionality case has failed however excellent the project is.

A credible answer is specific about the constraint being relieved, and constraints come in recognisable forms. Tenor is the commonest: commercial markets in many places will not lend at the duration infrastructure requires, and a project whose economics only work over fifteen years cannot be financed by a market that stops at seven. Currency is another — local-currency debt at a tenor that matches local-currency revenue may simply be unavailable, and the alternative is an unhedged mismatch the project cannot carry. Risk perception is a third: a first-of-its-kind project, a first mover in an untested regulatory regime, or a borrower in a market where commercial lenders have no appetite at any price. Scale is a fourth, where the required ticket exceeds what any available lender will take alone.

There is also a non-financial form that is frequently the strongest available and is systematically under-argued. An institution can be additional through what it requires rather than through what it lends: safeguard standards a project would not otherwise have adopted, governance arrangements imposed as conditions, technical assistance, or a demonstration effect that opens a market to other financiers. Where the money genuinely is available commercially but the standards would not have been, that is additionality, and it should be argued explicitly rather than left implicit. In the mandates our advisers run for a development-finance audience, the additionality case is drafted before the financial model is finalised, because a structure that undermines it — a tenor no longer than the commercial market would offer, on terms the commercial market would match — is worth knowing about while it can still be changed.

The honest version of this argument is also the durable one. An additionality case that overstates the constraint invites an obvious challenge: the institution knows its own market, frequently better than the applicant does, and a claim that commercial finance is unavailable is straightforwardly testable. Where some commercial finance is available but not on the tenor or at the scale required, saying exactly that — and quantifying the gap — is far stronger than asserting a total absence that the reader can disprove from their own knowledge.

Environmental and Social Assessment Determines the Timetable Before It Determines Anything Else

Development institutions apply environmental and social standards as a condition of finance, and the standards are published rather than proprietary, which makes this the most knowable part of the entire assessment. The IFC Performance Standards on Environmental and Social Sustainability comprise eight standards, addressing in turn the assessment and management of environmental and social risks and impacts; labour and working conditions; resource efficiency and pollution prevention; community health, safety and security; land acquisition and involuntary resettlement; biodiversity conservation and the sustainable management of living natural resources; Indigenous Peoples; and cultural heritage. The Equator Principles are a separate framework, voluntarily adopted by financial institutions for determining, assessing and managing environmental and social risk in projects, and its published categorisation sorts projects into three categories — A, B and C — reflecting the magnitude of the potential environmental and social risks and impacts. Which framework applies, and how, is a matter for the institution and the transaction; what matters here is that both are public documents an applicant can read in advance.

Categorisation is the step with the largest practical consequence, because it determines the assessment work required and therefore the timetable. A project assessed as carrying potentially significant adverse impacts requires a correspondingly thorough assessment process, and that process has an elapsed duration which is frequently measured in seasons rather than weeks — baseline studies may need to observe conditions across a full cycle, and consultation cannot be compressed without undermining the thing it is for. A project with limited and readily addressed impacts requires proportionately less. Applicants who discover their likely category late find that the financing timetable was decided by the assessment timetable all along, and that no amount of commercial urgency compresses it.

The scoping question is therefore worth asking at the outset rather than after a structure has been committed to. Land acquisition and physical or economic displacement, effects on Indigenous Peoples, impacts on critical habitat or cultural heritage, significant labour influx, and material pollution loads are each capable of moving a project into a more demanding category and each is knowable early. Where our advisers support a development-finance submission, the likely scope of assessment is established before the financing timetable is set, because a timetable built on a commercial assumption and then confronted with a full assessment process is a timetable that fails publicly.

Two misconceptions are worth correcting because they cause real damage. The first is that safeguards are a compliance formality to be satisfied by a report. They are a management requirement: the institution is concerned with whether impacts will actually be managed through the life of the project, which means an assessment must produce a management plan with named responsibilities, resources and monitoring, and that plan becomes a covenanted obligation. A high-quality assessment attached to a project with no capacity to implement it is a recognisable and unpersuasive submission.

The second is that safeguards are purely a cost. Frequently the assessment surfaces something the financial model should have carried anyway — a resettlement obligation, a water constraint, a community commitment, a closure liability — and surfacing it before financial close is considerably cheaper than discovering it during construction. Where an assessment changes the project cost, that is information the model needed. Our financial due diligence team reflects assessment-driven obligations in the base case rather than treating them as contingencies, because a covenanted obligation is not a contingency and modelling it as one misstates the cover ratio the institution will test.

Impact Asserted in Narrative Is Not Impact; It Has to Be Measurable and Baselined

A development institution reports to its own shareholders or members on what its capital achieved, which means it needs from each project a set of indicators it can aggregate and defend. That requirement lands on the borrower as a results framework: a small number of indicators, each with a definition, a baseline, a target, a frequency of measurement and a stated source of data. It is the part of a submission most often treated as an afterthought, and it is the part most likely to generate obligations that persist for the entire life of the facility.

The discipline is unfamiliar to most borrowers and it is worth being concrete about what it excludes. Employment supported is not an indicator; jobs, defined as full-time equivalent positions on the payroll of the project company, measured annually from payroll records, against a stated baseline, is an indicator. Improved access is not an indicator; connections added, defined and counted from a stated system, is. The test is whether an independent person, given the definition and the data source, would arrive at the same number. Anything that fails that test will be renegotiated later, at a point where the borrower has less leverage and the reporting obligation has already been covenanted.

Baselines are where frameworks most often collapse, and the failure is usually one of sequencing. An indicator without a baseline cannot demonstrate change, only a level — and where the baseline has to be established after financial close, it is frequently established badly, from records nobody kept for that purpose. Establishing baselines while the assessment work is being done, rather than after commitment, costs almost nothing and prevents a class of reporting difficulty that recurs annually thereafter. Our valuation team builds baselines into the framework at the point the indicators are defined, because an indicator whose baseline is to be determined later is an indicator that will be argued about later.

Attribution is the intellectually harder problem and it rewards honesty. A project rarely causes an outcome by itself, and a framework that claims sole credit for a change with many contributing causes will not survive scrutiny by an institution that reads these routinely. The defensible position is to measure what the project directly produces, to state clearly what is contributed rather than caused, and to resist indicators whose movement depends mostly on factors outside the project. A modest framework of indicators that genuinely track the project is worth considerably more than an ambitious one that will embarrass everyone at the third annual report.

Where finance is explicitly impact-linked, the framework stops being a reporting obligation and becomes a pricing term: margin steps, availability of tranches or other economic consequences attach to indicator performance. At that point every definitional weakness becomes a financial exposure. An ambiguous denominator, an indicator sensitive to a factor the borrower does not control, a measurement frequency that cannot actually be met — each is now capable of costing money, and each is far cheaper to fix in drafting than in a subsequent waiver. Where our advisers support impact-linked structures, the indicator definitions are tested against the question of what happens if the indicator moves for a reason nobody intended, because that is the case that arises.

The Institution Cares How the Money Is Spent, Not Only Whether It Comes Back

A commercial lender is largely indifferent to how a borrower procures, provided the cost is realistic and the contractor can perform. A development institution frequently is not, and the reason follows directly from its accountability: it is answerable for whether public or concessional capital was spent competitively and transparently, and that answerability extends into how contracts were awarded. Procurement requirements therefore commonly attach as conditions, and they attach to the borrower's conduct rather than only to the borrower's numbers.

The requirement that catches applicants most often is retrospective. Major packages awarded before the institution's involvement — a construction contract placed with a long-standing partner, equipment ordered from an affiliate, a design consultancy appointed without competition — may need to be revisited, evidenced or in some circumstances re-tendered. An applicant who has efficiently locked in their supply chain before approaching the institution may have created exactly the problem they thought they were avoiding, and unwinding it is expensive in a way that lengthens the timetable rather than merely costing money.

Related-party contracting deserves separate attention because it is common in owner-managed and group structures and is rarely thought of as a procurement issue at all. Where the contractor, the equipment supplier or the operator is connected to the sponsor, the institution is looking at both price and process: whether the terms are arm's length, and whether there was a genuine competitive alternative. A related-party award at a defensible price still raises the process question, and a submission that addresses the price while ignoring the process has answered the easier half. Documenting the comparison at the time of award — what alternatives were considered, on what terms, and why the related party was selected — is straightforward while the decision is being made and close to impossible to reconstruct afterwards.

Eligibility restrictions add a further layer that is easy to miss. Institutions commonly maintain exclusion lists, sanctions screening and eligibility criteria that apply to counterparties as well as to borrowers, and a supply chain assembled without reference to them may contain a party that cannot be paid from the facility. Screening major counterparties early is inexpensive; discovering the problem after award is not. In the mandates our advisers run for this audience, procurement requirements are established and applied from the point the institution is first approached, because the requirements attach to conduct from that moment and cannot be satisfied retrospectively by explanation.

A Long Tenor Is the Product, and It Makes Every Assumption Work Much Harder

Long-tenor debt is frequently the specific thing a development institution provides that the commercial market will not, and it is often the whole of the additionality case. It is also the reason the credit assessment is harder rather than easier. A seven-year facility is exposed to seven years of forecast error; a twenty-year facility is exposed to twenty, and the later years are exposed to conditions nobody can meaningfully forecast — regulatory regimes that will be rewritten, technologies that will be superseded, competitors who do not yet exist, and demand patterns that will have shifted for reasons unavailable today.

The consequence for the financial model is that a base case built to commercial standards is frequently only half-built for this audience. Applicants extend the forecast by continuing the trend, which produces a model whose first seven years are carefully reasoned and whose remaining years are arithmetic. An experienced reader identifies this immediately, because the later years show no discontinuities at all: no reinvestment, no contract expiry, no margin compression from competitive entry, no maintenance capital cycle. A twenty-year projection in which nothing structural happens after year seven is not a forecast, it is an extrapolation, and it will be treated as one.

The specific items that should appear in later years are knowable and are usually omitted. Major maintenance and replacement capital expenditure falls on a cycle that a long forecast must cross, and crossing it typically produces the minimum cover year. Contracts expire and are renewed on terms the market will set at the time rather than the terms in the original agreement — and an offtake contract that expires in year twelve of a twenty-year facility means the last eight years are merchant risk that no amount of contractual comfort in the early years addresses. Technology risk, regulatory change and competitive entry all sit in the same period. Where our financial due diligence team extends a base case to a development-finance tenor, the later years are built with those discontinuities in rather than smoothed over, because it is precisely the years a commercial lender never had to think about that this lender is exposed to.

Sensitivity analysis has to change shape as well, and this is where the analytical work is genuinely different. A commercial sensitivity moves a variable and reports the effect on cover. Over a long tenor the more informative question is when the project becomes vulnerable, and to what: at what point does the contracted position run out, at what point does the maintenance cycle coincide with the amortisation profile, and how are those two related. Presenting cover as a schedule with the minimum year identified — the discipline the companion article on credit committees sets out — matters more here than anywhere, because a twenty-year average is almost meaningless while a twenty-year minimum in year fourteen is the entire assessment.

None of this makes a long tenor a disadvantage. It is the reason the project is financeable at all, and a sponsor who has built the model to withstand the duration is presenting a stronger case than one who has extended a commercial model and hoped. The additional work is concentrated and specific: identify the discontinuities, model them, and show the reader where the schedule is tightest and why it survives.

When a Development Institution Sits Beside Commercial Lenders, Its Standards Travel

Development institutions frequently finance alongside others — commercial banks, export credit agencies, other development institutions, and sometimes the sponsor's own equity partners. Mobilising private capital is for many of them an explicit part of the mandate rather than a convenience, which means the co-financing structure is itself part of what is being assessed: a facility that brings commercial lenders into a market they would not otherwise have entered is demonstrating additionality in a form the institution can report.

For the borrower the practical effect is that standards travel across the whole financing. Where a development institution participates, its environmental and social requirements, its procurement conditions and its reporting obligations commonly apply to the project, not merely to its own tranche — and commercial co-lenders, who would not have imposed those requirements themselves, inherit the benefit of them. A sponsor who expected the safeguard obligations to be proportionate to the development institution's share is usually mistaken, and the misunderstanding surfaces at documentation, late.

The lender-of-record and agency arrangements add mechanics worth understanding in advance. Where one institution acts as lender of record and others participate, or where a common terms agreement governs several facilities, the borrower is dealing with a structure that has its own internal decision rules: which decisions require unanimity, which a majority, how a waiver is sought, and whose consent is needed to amend what. Those rules determine how quickly the borrower can obtain an answer to anything, and they are frequently the binding constraint on timetable during construction — a point at which speed matters most.

The consequence for preparation echoes the consortium problem in commercial lending, with an additional dimension. Documentation must satisfy the most demanding participant on each dimension, which will not be the same participant on every dimension: the commercial bank may be strictest on cover and security, the development institution on safeguards and procurement, an export credit agency on content and eligibility rules of its own. A file prepared to the highest applicable standard on each axis clears the group; a file prepared to the average clears nobody, and discovers the binding constraint one institution at a time. In the mandates our advisers run on co-financed structures, the requirements of each participant are mapped before drafting so the document is built once, which is the only version of this that does not cost a cycle per lender. A network of 80+ specialist consultants operating across 22 countries is what makes that mapping practical, since the participants are rarely all in one jurisdiction.

What the Screen Is Asking, and What Evidence Answers Each Question

The table below maps the questions a development-finance screen puts to a proposal against the evidence that answers each one. It is not a checklist for any particular institution — mandates, published policies and application differ, and the specific institution's own published requirements always govern. It is a reference for the shape of the preparation, and the useful way to read it is as a set of questions to ask of a draft before it is filed rather than a set of sections to write.

The pattern worth noticing is that almost every row is answered by evidence rather than by argument. That is the through-line of this reader: the commercial case is argued, and the additional screen is evidenced. A submission that argues its additionality, asserts its impact and describes its safeguards has produced narrative where the institution needed documents, and the resulting query list is long, slow and entirely avoidable.

Bankable Is the Entry Condition, Not the Argument

The reorientation this reader requires is the sharpest of any in this series, because it inverts the instinct a borrower brings from every other financing conversation. In commercial lending, the stronger the standalone case the better the reception. Here, the standalone case has to be strong enough to be creditworthy and the argument for participation has to be about something else entirely — what is not otherwise available, what would not otherwise be done, and what will be measurably different as a result. A submission that maximises the first while leaving the second implicit has optimised for the wrong screen.

In practice the preparation is specific and mostly front-loaded. Establish the additionality case before the structure is fixed, so a structure that undermines it can still be changed. Scope the environmental and social assessment before the financing timetable is set, because the assessment duration governs the timetable and no commercial urgency compresses it. Define indicators with baselines while the assessment work is being done, rather than reconstructing baselines afterwards. Establish procurement requirements before major packages are awarded, since that requirement reaches backwards. Build the model to the full tenor with its discontinuities in. And where others are financing alongside, map every participant's requirements before drafting rather than discovering them one institution at a time.

None of that is proprietary and none of it requires an adviser. The frameworks are published — the IFC Performance Standards and the Equator Principles are public documents, and the institution's own policies generally are too — and the discipline they imply is available to any sponsor willing to read them before rather than after committing to a structure. Where our transaction advisory practice prepares documentation for this audience, these are the questions the document is built to answer, and they are offered here as observations about how the screen works, drawn from work delivered by a network of 80+ specialist consultants across 22 countries. A sponsor who applies them without instructing anyone has been served correctly, which is the only test this article is trying to pass.