Comparing international resettlement standards: IFC PS-5, World Bank ESS-5 and AfDB OS-2

1. Why the comparison matters operationally

A transmission line or road corridor in East Africa is commonly financed by a syndicate. Where lenders bring different safeguard regimes, the project does not get to choose among them: it must satisfy each, which in practice means satisfying the most demanding on every dimension. Understanding where the standards diverge is therefore not an academic exercise but the basis for scoping a compliant instrument.

The comparison is also relevant to the growing number of projects financed under the Equator Principles, which import the IFC standards for projects in non-designated countries and thereby extend PS-5's reach well beyond IFC's own portfolio.

2. Eligibility without legal right

All three regimes accept the central proposition that entitlement to assistance does not depend on legal title. This is the single most consequential feature of international standards relative to most national expropriation law, which typically compensates recognised interests and nothing else.

The standards are consistent in distinguishing those with formal legal rights, those with claims recognisable under national law or capable of becoming so, and those with no recognisable claim at all. The third category — occupants without any legal basis — is entitled to resettlement assistance and compensation for improvements and structures, though not for the land itself. The practical significance is that an informal settler on public land, who under national law may be removed without payment, is under all three regimes entitled to assistance sufficient to restore their livelihood.

Divergence appears less in the categories than in how firmly each regime requires the cut-off date to be established and disclosed, since the cut-off is what makes the eligibility determination administrable at all.

3. Replacement cost

All three require compensation at replacement cost and are explicit that this differs from market value as ordinarily assessed. The recurring points are that depreciation should not be deducted from the cost of replacing a structure, that transaction costs and fees associated with acquiring a replacement asset form part of the compensation, and that valuation should be established with reference to what it actually costs to obtain an equivalent asset in the local market.

The practical divergence is procedural rather than substantive. Where regimes differ is in expectations about who determines replacement cost, how the methodology is disclosed, and whether an independent check on the valuation is required. Projects most often fall down not because they applied the wrong concept but because the methodology by which rates were derived was never documented in a form a reviewer could interrogate years later.

4. Livelihood restoration

Convergence is weakest here, and it is where completion audits most often record findings. All three regimes require that displaced people's livelihoods be restored, and the more recent formulations set the bar at improvement or at least restoration to pre-displacement levels. The obligation is expressed as an outcome, not as a set of activities.

An outcome obligation implies measurement over time, which distinguishes livelihood restoration from every other part of a resettlement instrument. Compensation can be evidenced by a payment record; restoration cannot. It requires a baseline of income and its sources, a defined period of observation, and monitoring able to distinguish a household that has recovered from one that has been paid.

The gap between the standards' ambition here and typical practice is wide, and it is a budgeting gap as much as a technical one. Livelihood programmes run for years after the construction contractor has demobilised and the project's attention has moved on.

5. Monitoring, completion and external review

The clearest divergence across the three regimes is evidentiary. All require monitoring; they differ in the extent to which external, independent verification is expected, and in what constitutes acceptable evidence that obligations were discharged.

The direction of travel across all three has been toward stronger external review, driven substantially by accountability-mechanism findings and by external scrutiny of cases where displaced populations were documented as worse off after projects that reported compliance. The institutional lesson those cases taught was not that standards were too weak but that self-reported compliance was insufficiently verifiable.

6. Conclusion

For a project scoping a multi-lender instrument, the practical guidance is straightforward: substantive requirements have converged enough that a well-designed RAP will satisfy all three on eligibility and valuation, while evidentiary and monitoring expectations remain the dimension on which regimes most differ and on which projects are most often found wanting. Scope the evidence architecture to the most demanding regime in the syndicate, because retrofitting it is not possible.