Cash is the easiest form of livelihood restoration to deliver and the hardest to make work. A payment clears an obligation on the company’s books the day it is made. On the household’s side it starts a clock. School fees, a funeral, a motorbike, a debt to a relative, a bad season. Within a year the money is usually gone and the field, the pasture or the fishing ground it stood in for is still gone too. The family has been compensated and is poorer than before. Both statements are true, and the second is the one that comes back as a grievance.
The international standard already knows this. IFC Performance Standard 5 puts cash last in its order of preference for economically displaced people. It says in plain terms that cash compensation alone is frequently insufficient to restore livelihoods. The evidence from mining regions agrees. Studies of displaced households in Ghana find resettlement policies that widened the poverty gap rather than closing it. Work on rutile mining communities in Sierra Leone finds sustained impoverishment rather than improvement. None of this is news to anyone who has sat in a resettlement village three years after the payments. The houses are there. The fields, the incomes and the goats often are not.
What follows is a practitioner’s account of what restoration beyond cash requires, and how to know when it has been achieved.
Why cash fails as a restoration tool
A livelihood is a system, not an asset. A hectare of land produces food, a surplus to sell, fodder, a place to keep animals, collateral for informal credit and a claim on the community’s respect. A cash payment prices the hectare and replaces none of the rest. The household has to rebuild the system itself, in a land market that may not exist. Meanwhile the payment is spent on the emergencies that arrive whether or not a project has. In villages I have worked with, the first large expense after a payment is usually a funeral, a wedding or a debt, not a plot of land.
The payment also goes to one person, usually a man, usually the one whose name was on the census form. The livelihood it replaced was often run by several people, and a share of it by women who never see the cheque. Michael Cernea’s Impoverishment Risks and Reconstruction model has named the consequences since 1997: landlessness, joblessness, food insecurity, loss of access to common property. Cash addresses the first only if replacement land can actually be bought, and it addresses the other three not at all.
None of this means cash has no place. The standard allows it where livelihoods are not land-based, or where the land taken is a small fraction of a viable holding. It also allows it where active land and labour markets exist and people use them. Those conditions are the exception in most mining regions. A project that treats them as the default has chosen the method that is quickest for itself and least likely to work for the people it displaces.
What restore or improve actually requires
The objective in the standard is to improve, or at least restore, the livelihoods and standards of living of displaced people. Paragraph 28 turns that into an order of preference that every livelihood plan should follow rather than reinterpret. For land-based livelihoods, replacement land of at least equivalent productive potential and location comes first, as a matter of priority. For livelihoods based on natural resources, the first option is continued access, then access to an equivalent alternative, with collective arrangements where the resource was collectively used. Only when neither is possible do credit, training, cash and employment enter, and then in combination.
Two further rules give the plan its teeth. Paragraph 29 requires transitional support for as long as it reasonably takes to restore income, production and living standards, which is a test, not a calendar. And paragraph 25 says the mitigation of economic displacement is complete only when people have received their entitlements and have been given adequate opportunity to re-establish their livelihoods. A plan that ends when the budget line is exhausted has not met either rule. The IFC’s 2023 handbook on land acquisition and resettlement builds an entire module around this, and the reason is simple: most plans stop too early.
Land for land, done properly
Replacement land is the measure that works when it is done well and generates the bitterest disputes when it is done badly. The difference is in the details the household notices and the plan often does not. Is the soil comparable, or is it the plot nobody wanted? Is there water in the dry season? How far is it from the house, the school and the market, and who bears the cost of that distance every day? Is tenure secure, with documentation in both spouses’ names, or is the family farming somebody else’s land on a promise?
The only way to answer those questions is to walk the replacement plot with the household before the offer is made, in the season that matters. Then let them refuse it. A refused plot is information. A plot accepted under pressure is a grievance with a delay on it. Where the project cannot find comparable land nearby, it should say so early and move to the next option with the household. Presenting a poor plot as if it were equivalent only delays the dispute. The valuation questions I set out in Compensation at Replacement Cost apply to the replacement land too. Transaction costs, land preparation and the first seasons of lower yield on new soil all belong in the account.
When land is not available
Projects that cannot offer land usually offer some mix of jobs, training and small-enterprise support. Each can work. Each has a characteristic way of failing. Jobs fail at the construction-to-operations cliff. A household that traded its field for a labourer’s wage during construction loses the wage when the contractor demobilises, and has no field to return to. Training fails when it prepares people for work that does not exist within reach of the village. Enterprise support fails when it funds a hundred identical kiosks in a market that could sustain ten.
The correctives are practical. Tie any employment offer to the operational phase, not the construction peak, and be honest about how many operational jobs there will be. Train for demand that has been checked, including demand from the project’s own supply chain. Support enterprises with markets, working capital and the boring things that keep small businesses alive, such as bookkeeping and payment terms the project actually honours. And treat all of it as transitional, with the restoration test applied to the household’s income at the end, not to the number of people who attended the course.
Restoration also has to be distinguished from development. A community fund, a clinic or a scholarship scheme is a benefit to the community. It is not restoration of the livelihood a specific household lost, and it should never be counted as such. I make the same argument about the difference between mitigation and shared benefit in From Mitigation to Enhancement. Both matter. Confusing them lets a project fund a school and leave a family hungry.
The households that fall through
Every resettlement has households that the standard plan does not fit. Elderly people who cannot farm a new plot. Women-headed households whose land was never registered in their name. People with disabilities. Households in the standard’s third eligibility category, who used land they had no recognised claim to and are entitled to assistance but often receive none. Seasonal users who were absent at the census.
These are the households that appear in the grievance register two years later, and they are the ones for whom cash fails fastest. The plan needs targeted measures for them from the start. That means a vulnerability screen in the census and a named entitlement that fits their circumstances. It also means a monitoring flag, so their progress is tracked separately rather than lost in an average. The standard is explicit that skills training, credit and jobs must be equally available to women and adapted to their needs. In practice that means asking women what they need in their own meetings, not reading it off the household head’s form.
Measuring restoration honestly
A livelihood plan without a restoration test is a compensation schedule with a longer name. The test needs a baseline, taken before displacement, of household income, production, food security and assets. It needs household-level tracking against that baseline, not community averages, because averages hide exactly the families who are failing. It needs monitoring that continues until the target is met, with affected people consulted during the monitoring, as paragraph 14 requires. And for larger or more complex displacement it needs the external completion audit in paragraph 15. That audit belongs at the point where people have had adequate opportunity to restore their livelihoods, not on a date fixed at the start.
The settlement that closed the backlog of more than 2,000 claims I mediated in Mozambique paired individual payments with collective measures. Both sides had come to understand that cash alone would not hold, and that whatever was agreed had to be checkable afterwards. That combination, individual entitlement plus collective restoration plus a way to verify both, is the shape of a plan that survives its own implementation. The building of durable community benefit I describe in Community Development That Lasts starts from the same principle.
Map the plan before the first payment is made
The Livelihood Restoration Plan Roadmap sets out the sequence in five phases and 15 steps. The phases run from baseline and eligibility, through measures in the standard’s order of preference and land before alternatives, to transitional support and the restoration test through to completion audit. Each step is scored In place, Partial or Absent. Read it phase by phase. An Absent in the first phase means the plan cannot know what it is restoring. An Absent in the final phase means it cannot know when it is finished, which is the failure behind most of the resettlement grievances that reach a mediator years later. Download the Livelihood Restoration Plan Roadmap
Restoration is a negotiation, not a disbursement
A company cannot restore a livelihood to a household that has not been part of designing the restoration. The measures that work are all products of a conversation: land the family has walked, access it has agreed, training for work it can reach. The conversation has to be one both sides trust. Where that conversation never happened, the payment was the plan, and the plan failed on the day the money ran out.
This is why livelihood restoration belongs inside a structured, independently facilitated process rather than a disbursement schedule. The Social Accord Architecture treats restoration measures as commitments in the agreement, with owners, dates and a monitoring rhythm carried into the Stewardship phase until the restoration test is met. Dual accountability holds both sides to it. The company is accountable for delivering land, access and transitional support that actually restore income, and for measuring that honestly. The community is accountable for engaging with the measures and for using the agreed track when a household falls behind. Livelihoods and investments are protected together, because a project that leaves displaced families poorer carries that fact through every year of its operating life.
One thing to do on your next plan. Before any payment is approved, ask for the restoration test in writing: what will be measured, against which baseline, for how long, and who decides when it is met. If nobody can answer, the plan is a payment schedule. To design a livelihood restoration process or to work through a resettlement that has stalled, write to me at thomas@thomasgaultier.com.