How to build an entitlement matrix that survives review

The entitlement matrix is the one table a reviewer reads first and a field officer reads every day. It states, for each category of loss, who is eligible, what they receive, on what valuation basis, and under what condition the entitlement is discharged. Everything else in a RAP explains or implements it.

Most matrices are adapted from a previous project. That is reasonable — the structure is stable across projects — but it is also how a matrix acquires rows for losses that do not occur here and loses rows for losses that do.

The four columns that do the work

A matrix can carry more columns, and many do. Four are load-bearing:

• Loss category. Not the asset — the loss. Losing a house you own and losing a house you rent are different rows with different entitlements, even though the structure is the same building.
• Eligible person. Defined by relationship to the loss and by tenure, not by name. The register names people; the matrix defines classes.
• Entitlement. What is provided, stated specifically enough that two officers would calculate the same amount.
• Basis and reference. How the figure is derived, and the provision of the standard or national law it satisfies.

That last column is the one most often omitted and the one that makes the matrix defensible. A rate with no stated derivation is an assertion, and at review it will be treated as one.

Start from the losses, not from the template

Build the loss list from the census and the asset inventory before opening any previous matrix. Walk the footprint category by category: land by tenure type, structures by use, crops and trees by species and maturity, businesses by whether they own or occupy, tenants, employees of affected businesses, access to resources held in common, graves and ritual sites, and public assets belonging to no household at all.

The categories most often missing from an inherited matrix are consistent enough to check for directly: agricultural tenants and sharecroppers, employees who lose work when a business is displaced, common property access, and losses that occur only during construction — a severed access track, a market cut off for a season.[1]

Eligibility follows the three categories

The standards distinguish those with formal legal rights, those with claims recognisable under national law or capable of becoming so, and those with no recognisable legal claim at all.[1] The third category is entitled to resettlement assistance and to compensation for structures and improvements, though not for the land itself.

In matrix terms this means a row cannot say compensation at replacement cost and stop. It has to split by tenure, because for one class the entitlement includes land value and for another it does not, while both receive assistance. A matrix that omits the split reads as though informal occupants get nothing, which is both wrong and the fastest way to acquire a finding.

If a row cannot be applied to a real household in the register without a further decision by whoever is holding the form, the row is not finished.

Valuation basis, stated per row

Different losses take different bases and mixing them silently is a common defect. Structures take replacement cost with no deduction for depreciation.[2] Land takes replacement cost measured against what an equivalent parcel actually costs locally, including the fees and transfer costs of acquiring it. Crops take the value of the lost harvest, or in the case of perennials the cost of re-establishment plus lost yield to maturity. Businesses take lost net income for a defined period plus the cost of moving.

Each row should name its basis and its source: a district schedule, a rate study commissioned for the project, a quantity surveyor's build-up. Where a statutory schedule is used and it falls below replacement cost, the matrix should say how the gap is bridged rather than leaving the reader to discover the shortfall by arithmetic.[3]

Allowances are not a residual category

Disturbance allowances, transitional support, moving assistance and vulnerable-household support are usually appended at the end of the matrix as flat sums. They are also the entitlements most likely to be challenged, because a flat sum implies a calculation nobody did.

Two disciplines help. State what the allowance is intended to cover — three months of subsistence at the local wage rate, or the cost of hiring transport for a household's effects over the relevant distance — so the figure has a derivation. And state the trigger: what a household must do or show for it to become payable, and by when.

Make it operable

A matrix that cannot be executed produces exceptions, and exceptions are where compliance is lost. Before finalising it, run three or four real records from the census through it end to end — a titled owner with crops and a grave, a tenant with a business, an informal occupant with a structure — and see whether the entitlement can be computed from the register fields you actually collected.

Where it cannot, the fix is upstream in the enumeration instrument, and it is considerably cheaper to make now than after the cut-off date has passed and re-enumeration is no longer available.[4]

Version it

The matrix will change: rates are updated, a loss category is added after disclosure, an allowance is revised. Each change alters entitlements that may already have been communicated to households, and a completion audit will ask which version applied to a given payment.

So the matrix needs a version number, a date, an approval, and a record of which households were assessed under which version. Projects that keep this in a single spreadsheet overwritten in place cannot answer that question, and the answer is not reconstructable afterwards.