What is a Resettlement Action Plan? A practitioner's guide

A Resettlement Action Plan is the document a project produces when it takes land, and taking land displaces people. It sets out who is affected, what each of them loses, what they are owed for it, how that figure was reached, who delivers it and by when. Every other description of a RAP — a safeguard instrument, a lender deliverable, a compliance document — is downstream of that.

The part practitioners learn late is that a RAP is judged twice. Once as a plan, before implementation, when a lender's safeguards team reviews whether the design meets the standard. And once as a record, at completion audit, when someone tests whether what the plan promised actually reached the people it named. Plans routinely pass the first review and fail the second, and the reason is almost never bad intent. It is that the evidence needed at audit was never captured while the work was happening.

When a RAP is required

The trigger is displacement, not land purchase. A project that buys land from a willing seller at market price on an open market has not triggered involuntary resettlement. A project that acquires land using expropriation powers — or under the credible threat of them — has, regardless of whether the transaction looks voluntary on paper.

Both of the standards most RAPs are written against split displacement into two kinds, and a project can cause either or both:

• Physical displacement — relocation, loss of residential land, or loss of shelter.
• Economic displacement — loss of land, assets, or access to assets that leads to loss of income sources or means of livelihood, with no relocation involved.[2]

Economic displacement without physical displacement is the case projects most often miss. A transmission line that takes a strip of a farmer's field, or a road that severs a trader from their customers, has displaced those people economically. They are entitled under the standards, and they belong in the RAP.

Thresholds vary by lender. The African Development Bank, for example, requires a full Resettlement Action Plan where a project displaces more than 200 people, and an abbreviated plan below that.[3] A project financed by several lenders inherits the strictest requirement in the set, not the average.

What a RAP has to contain

The structure is broadly settled across standards, and a reviewer will expect to find each of these:

• A census of affected people and an inventory of assets, anchored to a declared cut-off date that fixes who is eligible.
• A socioeconomic baseline — because livelihood restoration is defined as restoration relative to pre-project conditions, and unmeasured conditions cannot be restored to.
• A legal framework review, reconciling national expropriation law against the lender standard, and stating explicitly which applies where they diverge.
• An entitlement matrix — one row per category of loss, setting out eligibility, entitlement, valuation basis, responsible party and timing.
• Valuation methodology, evidencing that compensation reaches full replacement cost.
• A grievance redress mechanism that affected people can actually reach and use.
• Consultation and disclosure records, showing what was disclosed, to whom, when, and in what language.
• An implementation schedule and budget, tied to the works programme, since compensation must be complete before displacement.
• Monitoring, evaluation and a completion audit provision.

The entitlement matrix is the part a reviewer reads first. It is the only place where eligibility, adequacy and timing can be checked against each other on one page.

Why the second review is the hard one

The scale of what goes wrong here is documented. A 2015 investigation coordinated by the International Consortium of Investigative Journalists examined World Bank–financed projects between 2004 and 2013 and estimated that 3.4 million people were physically or economically displaced by them, finding that the Bank had frequently failed to follow its own resettlement rules.[4] The response across the sector since then has been a steady tightening of what counts as evidence — not of what the standards aspire to.

That tightening is the practical challenge. Consider what a completion auditor asks for: proof that this household received this amount, on this date, calculated on this basis, before they were displaced, and that any grievance they raised was resolved. Each of those is a record that had to exist at the moment the event happened. None can be honestly reconstructed two years later from a spreadsheet that has been edited by fourteen people with no change history.

This is why RAP delivery is increasingly a data problem wearing a compliance costume. The standard has not changed much in a decade. The evidentiary burden has.

Who actually produces one

In donor-financed infrastructure, the RAP is usually contracted to an external environmental and social consultancy, frequently bundled with the ESIA and ESMP as a single deliverable. The implementing agency signs and pays; the lender clears the deliverable but does not procure the tooling. That three-party structure explains a durable feature of this work: the consultancy builds a bespoke database for each project, uses it for the duration, and abandons it at handover.

The cost of that pattern lands at completion audit, when the project needs the audit trail and finds a folder of spreadsheet versions instead.

Where to go next

If you are building one now, the entitlement matrix is the fastest place to see whether your scope is complete — a matrix missing tenants, employees of affected businesses, or common property resources is missing eligible people, and it is easier to see that in a table than in a narrative.