Cash compensation or land for land: what the outcome evidence supports

1. The stated preference and the observed one

International standards express a preference for land-based compensation where the displaced derive their livelihoods from land, on the reasoning that a land-based household given money has been converted from a producer into a consumer of savings unless the money is successfully reinvested in productive capacity. The preference is qualified rather than absolute — it applies where land-based livelihoods are affected and where replacement land is available.

Recorded practice diverges sharply. Across most programmes the great majority of affected households take cash, and the RAP records this as an informed choice exercised freely. The record is accurate as far as it goes. What it omits is the structure of the choice.

2. How the choice is actually posed

The two options are rarely commensurate at the moment of decision. Cash is specific: an amount, a date, a payment mechanism. Replacement land is a commitment to identify a parcel of comparable productive capacity at a location to be determined, subject to availability, with a timetable that the project cannot guarantee because it depends on a willing-seller market the project does not control.

A household choosing between a known sum available within months and an unspecified parcel available at an unspecified time is not choosing between cash and land. It is choosing between certainty and a promise from an institution it has no reason to trust, made by a project that will demobilise. Under those conditions the cash election is rational and tells us very little about what the household would prefer if both options were equally concrete.

This matters for compliance as well as for outcomes. Where the land option is offered in terms that no reasonable household would accept, the resulting universal cash election does not evidence that a land-based alternative was genuinely made available.

3. What happens to a lump sum

The pattern reported across contexts is consistent enough to be treated as the base case. A substantial share of the payment is absorbed within the first year by obligations that predate the project and by claims from within the extended family that the recipient has limited standing to refuse. Consumption expenditure rises and then reverts. Where land is bought, it is often smaller, further away, or of lower quality than what was lost, because the compensation was assessed against the acquired parcel while the purchase occurs in a market that has already repriced.

The gendered dimension is well documented: payment to a single registered claimant, usually male, transfers to that individual an asset that previously supported a household, and the household's claim on it is thereafter informal.

None of this indicates that money is the wrong instrument. It indicates that a single large transfer to a recipient with no history of managing one, in a market that knows the transfer is coming, is a weak mechanism for reconstituting productive capacity.

4. Conditions under which each performs

Cash performs adequately where three conditions hold: the loss is partial rather than total, so the household retains a productive base; a functioning market exists in which the replacement asset can actually be bought at something near the compensated rate; and the recipient has prior experience of monetary savings or is supported to acquire it. Where the loss is total and the market is thin, cash is a transfer that terminates a livelihood and finances a period of consumption.

Land-based resettlement performs where replacement land of comparable productive capacity genuinely exists within the affected community's operating radius, where the transfer is completed before dispossession rather than after, and where the receiving area's own population has been engaged so that the transfer does not simply relocate the conflict. Its failure mode is delivery: land promised and not delivered, or delivered years later, or delivered as marginal land that was available precisely because nobody was farming it.

The corollary is that the choice should be assessed as a capacity question about the implementing institution and the local land market, not as a matter of respecting individual preference. A project that cannot deliver land within the schedule should not offer it as an option; a project that offers cash where the market cannot supply replacement land should treat the accompanying livelihood programme as the substantive obligation rather than as an adjunct.

5. Instruments that sit between the two

Practice has developed intermediate forms whose evidence base is thinner but whose logic addresses the failure modes directly. Staged payment — a first tranche at possession and later tranches tied to the acquisition of a replacement asset — reduces the lump-sum problem at the cost of administrative burden and of paternalism that has to be justified. Assisted purchase, where the project brokers replacement parcels and verifies title before releasing funds, addresses the thin-market problem where the project has the capacity to run it.

Each of these transfers work from the household to the implementing institution, which is the reason they are proposed less often than they are discussed. They are also the only mechanisms that respond to what the outcome evidence actually says.

6. Implications for instrument design

First, record the basis on which the choice was offered, not merely the choice made — the specificity of each option at the moment of election is the fact a reviewer needs and the fact projects do not capture. Second, treat universal cash election as a diagnostic signal rather than as a result. Third, size the livelihood restoration programme against the share of households whose entire productive base was acquired for cash, because that group carries essentially all of the programme's impoverishment risk and is identifiable at baseline.