I have sat in rooms where documents were approved that should not have been. And in rooms — perhaps more difficult to recall without some discomfort — where genuinely strong proposals were turned away because the document carrying them had failed to do its job. Thirty years working at the intersection of institutional finance, sovereign advisory, and cross-border policy will teach you many things. The one I did not expect: the quality of the writing almost never decides the outcome. What decides it is something that happens long before a word is written.
I am not making a case for bad writing. Far from it. But I have watched too many capable people — technocrats, economists, senior officials, experienced project teams — pour months into a proposal, only to have it fail at the evaluation stage for reasons that had nothing to do with the quality of their thinking. The thinking was often excellent. What let them down was the document. There is a difference between those two things that is worth sitting with for a moment.
This, in my experience, is the single most misunderstood thing about high-stakes institutional documents. The people who matter — the review committee member, the investment officer, the ministerial adviser, the multilateral evaluator — are not reading your document the way a curious person reads a book. They are moving through it looking for reasons to hesitate. Their professional disposition is defensive. Their job, stated or not, is to find what could go wrong if they approve what you are asking them to approve.
I have watched this dynamic play out across four continents and several dozen approval processes. It does not change much. A document that is well-written but structurally disordered gives a scanner nothing to hold. They pass through it searching for the logic — where is the problem defined, where is the evidence, where is the risk acknowledged and addressed — and when they cannot locate these things without effort, they do not read more carefully. They form a judgement. It happens quickly. And it is very hard to reverse.
The writer's task, then, is not to impress the evaluator. It is to make their scanning effortless. To ensure that what they are looking for is precisely where they expect to find it, expressed in the register their institution recognises, supported by the grade of evidence their standards require. Achieving that — consistently, across different institutional cultures — is a great deal harder than it sounds.
The gap between a document that passes and one that fails is almost always structural, not stylistic. Most authors do not discover this until after the rejection.
I have reviewed funding applications declined by multilateral development banks despite technically sound proposals and demonstrably strong project fundamentals. In nearly every case, the failure was architectural. The document had been assembled rather than designed — sections written in sequence as the work progressed, content added as it became available, the argument left to find its own shape rather than given one from the outset.
An assembled document carries the evidence of its assembly. The executive summary does not reflect the body. The financial analysis appears before the contextual framework that makes it legible. The risk section — typically written last, when the team is exhausted and the deadline is close — sits awkwardly at the end rather than being integrated throughout the argument where it belongs. An experienced evaluator senses this unevenness immediately, even when they cannot name exactly what is wrong.
A designed document works from a different starting point. It begins with the reader, not the writer. With the question of what the evaluator needs to understand, in what order, supported by what quality of evidence, to move from professional scepticism to considered confidence. The structure is not a container for the content. It is the argument itself. Change the structure and you change what the document is capable of achieving.
There is a register that institutional documents must occupy — considerably narrower than most writers appreciate. Too formal and the document reads like a legal instrument: impersonal, airless, signalling nothing about the human judgement behind it. Too conversational and it reads like advocacy — which, in institutional contexts, is almost universally received with suspicion. Advocacy signals that the author wants something. Evaluators are trained, over long careers, to be wary of authors who want things.
The register that works is what I would call analytical confidence. The tone of someone who has examined the evidence seriously, is honest about its limitations, and has arrived at a considered position. Not certainty — certainty is itself a warning signal in environments where genuine complexity is understood. What evaluators trust is rigour that has been transparent about what it does not know.
I once worked on a government policy document — a white paper for a ministry in the Gulf — where the original draft had been written with the unmistakable tone of a sales pitch. Every claim was superlative. Every projection was optimistic. Every risk had been minimised or explained away. The ministry had sound reasons for its confidence; the underlying policy was genuinely strong. But the document read as though it had something to conceal. The revision was not about inserting caveats for their own sake. It was about making visible the interrogation the authors had actually conducted — which, once on the page, made their conclusions far more persuasive than the original had ever managed to be.
A techno-economic feasibility study for an infrastructure project of any scale will be read by an engineering directorate, a financial underwriting team, an environmental compliance desk, and a regulatory authority — each with a different threshold of evidence, a different professional vocabulary, and a different definition of what constitutes an adequate answer. A document calibrated to satisfy the engineers may lose the financiers entirely. A document written for the regulatory desk may frustrate the technical reviewers before they reach the sections that matter to them.
This is one of the genuinely hard problems of institutional writing, and it is almost never solved by producing separate documents for separate audiences. It is solved by architecture. By structuring the document so that each audience can locate what they need without being required to work through what they do not. Executive summaries that actually summarise — not restate. Technical annexures that do not interrupt the main argument. Financial projections presented in the language of the investment committee rather than the project delivery team.
Holding multiple institutional perspectives simultaneously is not a writing skill. It is an institutional literacy skill. It comes from having occupied the roles on both sides of these documents over a long career. There is, in my experience, no shortcut to it.
Every institution carries a set of unspoken questions — the concerns that experienced evaluators bring to a document without articulating them, which a well-prepared author has already addressed before they are raised. In MDB funding applications, these tend to gather around additionality, financial sustainability, and the demonstrated capacity of the implementing agency. In policy white papers, around implementation feasibility and political economy. In executive memoirs and leadership narratives, around the distance between what is claimed and what can be independently corroborated.
A document that answers only the questions it has been asked has been written for submission. Not for approval. Submission is an administrative act — the form is filed, the requirement is met, the process is satisfied. Approval is a decision. And decisions require that doubt has been addressed, not merely that the paperwork is in order.
The most reliable sign that a document is genuinely approval-ready is that a careful reader, upon finishing it, finds their objections have been anticipated. Not argued away — anticipated. Arguing away an objection signals defensiveness. Anticipating it signals that the author understands the institutional environment well enough to know what questions will be asked before they are asked. That signal, quiet as it is, changes the relationship between document and evaluator considerably.
When documents fail, the instinct is to revise. To improve the language, strengthen the evidence, tighten the argument at the sentence level. In most cases, this is insufficient. A document with a structural failure cannot be edited into soundness. It needs to be redesigned — which is a harder conversation to open, particularly with a client who has spent months building something and is, quite reasonably, attached to what they have produced.
I have had that conversation more times than I can comfortably count. It is never easy. But the alternative — a polished version of a document that will still fail — serves no one. Least of all the client.
Senior editorial work at this level begins with diagnosis. Identifying where the architecture has come apart. Why the tone has drifted into the wrong register. Which evaluating body is being underserved and why. What the document is not saying that it must say. The writing, once those questions have honest answers, is the more straightforward part of the work.
It is also, by that point, the most satisfying.
Sudhir Kumar Rao is the principal consultant at SKR Editorial Services, a senior editorial consultancy serving governments, development finance institutions, and C-suite principals across India, the GCC, Singapore, and Sub-Saharan Africa. Enquiries may be directed to skr@skreditorial.com or via the contact page.