ESIA and RAP sequencing: why the order these are done in decides the cost

The Environmental and Social Impact Assessment and the Resettlement Action Plan are usually procured together, in one consultancy contract, with overlapping delivery dates. That arrangement is efficient on paper and creates a dependency problem in practice: the RAP needs outputs the ESIA has not produced yet.

Where the two run genuinely in parallel, the RAP team makes assumptions about the project footprint, and every assumption that turns out wrong is re-enumeration, re-valuation and re-consultation — paid for twice.

The dependency, stated plainly

A RAP cannot begin its census until it knows who is affected. Knowing who is affected requires a defined footprint. Defining the footprint is an ESIA output, arrived at partly through the alternatives assessment.

That last point is the one most often lost. PS-5's first requirement is to avoid displacement, and where unavoidable, to minimise it through alternative project design.[1] Avoidance is a design decision, made during the ESIA's alternatives analysis. If the alignment is fixed before anyone has counted who it displaces, the standard's first requirement has been skipped — and no amount of careful compensation work afterwards restores it.

Displacement counts are an input to the alternatives assessment, not a consequence to be measured after the route is chosen.

A workable sequence

1. Scoping and preliminary screening. Establish whether displacement is likely at all, and at what order of magnitude, across candidate options.
2. Alternatives assessment with displacement as a criterion. Enough reconnaissance-level data on each option to compare them — not a full census, but a defensible estimate of people and assets affected.
3. Footprint confirmed. The point at which the RAP's real work can begin.
4. Census and asset inventory, fixing the cut-off date.
5. Valuation, entitlement design, RAP drafting, informed by the ESIA's socioeconomic baseline rather than duplicating it.
6. Disclosure and consultation on both instruments, which is also where their consistency gets tested publicly.

What goes wrong when they are compressed

The census is run on a provisional footprint

A corridor that shifts by fifty metres after enumeration creates two populations: people enumerated who are no longer affected, and people affected who were never enumerated. Both are expensive. The second is also a compliance problem, because those households now have no cut-off-date evidence and no baseline.

Two baselines, two answers

The ESIA collects socioeconomic baseline data. So does the RAP. Run separately by different teams with different instruments, they produce different numbers for the same communities — and a reviewer reading both documents will notice. Livelihood restoration is measured against a baseline, so it matters which one is authoritative.

The fix is a single baseline instrument, designed once, covering both purposes, with the RAP's additional depth built into it rather than bolted on afterwards.[3]

Consultation fatigue

Two teams consulting the same communities weeks apart, asking overlapping questions, degrades participation and credibility. By the third visit attendance falls, and the project's consultation record starts to show exactly the thinning that a reviewer reads as disengagement.

Where the standards attach

ESS-5 obligations sit with the borrower under an agreed commitment plan, while PS-5 binds the client through loan covenants — and under EP4, through the wider network of financial institutions that have adopted the IFC Performance Standards as their reference framework.[4] In all three cases the resettlement instrument has to be consistent with the impact assessment that scoped it. Contradictions between the two documents — different affected-population figures, different baselines, different footprints — are among the easiest findings a reviewer can make, because they require only reading both.

Procuring it better

If the two are bundled into one contract, the terms of reference should sequence them rather than parallelise them: RAP census milestones conditional on footprint confirmation, one shared baseline instrument, one consultation programme covering both, and a deliverable schedule that admits the dependency instead of pretending it away.

That is a scheduling decision made at procurement, months before either team mobilises, and it is considerably cheaper than the re-enumeration it avoids.