Most RAPs fail on evidence, not intent

Resettlement cost is usually the line item that moves a project's economics most and is estimated with the least evidence. A feasibility study that treats displacement as a percentage assumption rather than a counted population is the origin of most RAP budget overruns. Early screening — how many households, what tenure mix, which categories of loss, what the land market actually does in that district — is cheap to do and expensive to skip.

The alternative chosen here determines the size of every obligation that follows. Comparing options on engineering cost alone, without the resettlement footprint of each, front-loads a liability nobody priced. This is also where avoidance and minimisation — the first requirement of PS-5 and ESS-5 — is either done or permanently foreclosed.

The ESIA and the RAP are separate instruments on entangled timelines, and sequencing them wrongly means either the RAP is written against a design that changes or the ESIA is submitted without the social baseline that justifies its conclusions. Traceability between the two is what a reviewer checks when the numbers in each do not agree.

The cut-off date is the single most contested fact in a resettlement, and it is only defensible if you can show when it was set, how it was disclosed, and who was enumerated before it. A census and socio-economic survey captured on paper months later cannot answer any of those questions. This is the moment the evidentiary record either begins or does not.

Full replacement cost is the rule most often applied incorrectly — depreciation deducted where it should not be, transaction costs omitted, rates set once and never revisited as a multi-year project runs through inflation. An entitlement matrix that maps loss type and tenure to package is what keeps eligibility consistent across thousands of households instead of decided case by case.

Stakeholder engagement under ESS-10 is judged on record, not recollection. Which document was disclosed, in which language, at which meeting, attended by whom. The disclosure register is usually the first thing an auditor asks for and the thing most often reconstructed after the fact — which is visible, and damaging.

Money moves to thousands of people, often through mobile money, often to people without formal identification, sometimes to the wrong household member. Reconciling what was owed against what was paid — and being able to evidence it per person years later — is where the gap between a compliant RAP and a documented one usually opens.

A grievance mechanism is assessed on whether it is accessible, whether it escalates, and whether it closes cases within stated timeframes. Most fail on the third. A log that cannot show acknowledgement times and resolution rates by category is not a mechanism a lender will accept, whatever the design document says.

Compensation is a transaction; livelihood restoration is a programme with its own activities, targets and monitoring surveys, running long after the payment clears. It is the most commonly under-resourced part of a RAP and the most common finding at completion audit.

The completion audit asks a question the project cannot prepare for retroactively: show that outcomes were achieved and that the process reached everyone. Indicators defined at the end measure whatever data happens to exist. Indicators defined at the start, and captured as the work happens, measure what was promised.