What a lender reviewer checks in a RAP, in the order they check it

A reviewer reading a resettlement action plan is not reading it as a document. They are looking for internal contradictions, because a plan is a set of numbers that have to agree with each other, and where they do not agree the reason is usually instructive. The sequence below is common to review under the IFC and World Bank standards and, through the Equator Principles, to most commercially financed projects as well.[5]

Knowing the sequence they follow is useful for two reasons: it is faster to self-review in the same order, and most findings are avoidable by anyone who has done so.

1. Do the population figures agree?

The number of affected households appears in the executive summary, the census chapter, the entitlement analysis, the budget and the monitoring framework. In a surprising share of plans these are four different numbers.

Where they differ legitimately — households versus persons, physically versus economically displaced, surveyed versus eligible — the plan should say so at each point. Where it does not, the reviewer's next question is which number the budget was built on, and that is a harder conversation.

2. Does the entitlement matrix cover everyone in the census?

The reviewer takes the categories of affected person identified in the census and checks each against a row in the matrix. Tenants, sharecroppers, employees of affected businesses, users of common property and people without any recognisable claim are the categories most often present in the census and absent from the matrix.[1]

3. Is the valuation basis stated and benchmarked?

For each entitlement, where does the number come from? A district schedule cited without a replacement-cost benchmark is the most common single finding, because the standards define compensation adequacy independently of national rate schedules.[2] Depreciation deducted from structure values is the second.

4. Does the budget reconcile to the entitlements?

Multiply the affected quantities by the stated rates and compare to the budget. The reviewer does this arithmetic, and it frequently fails — usually because the budget was drawn from a feasibility estimate and never revised after enumeration.

They will also look for the lines that are habitually omitted: transitional support, livelihood restoration beyond training, grievance operating costs, monitoring through to completion audit, the implementing unit's own costs, and contingency at a level that reflects the variance actually observed in comparable programmes.[4]

5. Is the cut-off date fixed, disclosed and evidenced?

One date, stated once, consistently throughout. The reviewer wants the disclosure evidence, not the assertion, and will check that the date precedes the enumeration it governs rather than following it.

6. Does the schedule pay before it displaces?

Compare the payment schedule to the construction schedule. Where possession precedes payment completion for any section, the plan has documented a non-compliance in advance, and no amount of narrative around it will help.[3]

The most productive self-review is to read the plan as though looking for the sentence that contradicts another sentence. That is what the reviewer is doing.

7. Is the livelihood restoration programme an outcome commitment?

A programme consisting of training with no measurement of income afterwards is an activity plan. The reviewer looks for the baseline income data it will be measured against, the target, the period, and the budget that funds it after construction finishes.

8. Is the grievance mechanism operable?

Named receiving points, defined response times, an appeal route that is not the same office as the first instance, a record structure, and a commitment that using it does not prejudice access to courts. And crucially: is it operating before compensation offers are made, rather than being established once complaints arrive?

9. Can the monitoring framework be computed?

Take each indicator and ask where its baseline value comes from. Indicators referencing income, food security or livelihood status require baseline data that must already have been collected. Where it was not, the indicator is decorative and the plan has no way to demonstrate restoration.

10. Who implements, with what capacity?

The reviewer looks for named institutional responsibility, staffing, and evidence about throughput — how many payments this agency has processed before, and how long they took. An institutional analysis that describes structure without capacity does not answer the question being asked.

The pattern behind all ten

Every item above is a consistency check between two parts of the same document. That is the whole method, and it is available to the drafting team at no cost. The plans that pass review cleanly are not the ones written most elegantly; they are the ones whose numbers were reconciled before submission.