The grievance plan and the grievance log are different documents

Most resettlement projects can produce a grievance log. Far fewer can produce the document that explains why the log is structured the way it is — which tier a valuation dispute goes to, how many days a sub-county committee has before a complaint escalates without the complainant asking again, and what authority the district tier holds that the village tier does not.

That second document is the grievance redress management plan, and its absence is not cosmetic. Without it the tiers exist informally, escalation depends on a complainant knowing to press, and a reviewer has no basis to judge whether the log reflects a designed mechanism or a series of ad hoc decisions written down afterwards.

Four tiers, four different jobs

A mechanism spanning village, sub-county, district and national levels is common in Uganda because it matches the local government structure. The number of tiers matters far less than whether each one has a distinct authority. Two tiers that both "try to resolve it" are one tier with extra paperwork.[1]

• Village or parish. First contact, no fee, local language, complaints received verbally or in writing. Most complaints should end here. This tier fails when the community was never told who sits on the committee or where to find them.
• Sub-county. Delegated authority to approve remedies up to a stated value. This tier fails when escalation depends on the complainant re-filing rather than on an automatic timeout.
• District. A different technical capability — valuation review, contested eligibility, access to the Chief Government Valuer. This tier fails when it duplicates the sub-county rather than bringing something the lower tier lacks.
• National or project. Authority to commit project funds and instruct contractors. This tier fails when it only ever sees complaints that skipped district — which means district was never functioning as a tier at all.

Authority has to be financial, not just procedural

"The sub-county committee will attempt resolution" describes an intention. "The sub-county committee may approve remedies up to a stated value" describes an authority. Without a financial limit per tier, every remedy above a token amount escalates regardless of the structure, and the tiers become a queue rather than a filter.

One reference number, intake to closure

The most common finding in a grievance mechanism review is not a missing tier. It is that each tier keeps its own spreadsheet, so a complaint's intake record at village level cannot be matched to its resolution at district level without asking someone who remembers the case.

The fix is structural rather than procedural: one reference number issued at intake and carried through every escalation, so the full history — who received it, what was tried, how long each tier took, what was decided — is one query rather than four people's recollections.

What a lender checks first

Two things, both simple to state and commonly missing. Whether escalation is automatic on timeout rather than dependent on the complainant. And whether the external route — an independent accountability mechanism, an ombudsman, the courts — is disclosed at intake rather than buried in an annex.[2] A mechanism that presents itself as the only avenue is not a grievance mechanism; it is a waiver collected under another name.

The complaint that must never enter the plan

Every tier should be trained to recognise a disclosure of gender-based violence or sexual exploitation and divert it immediately to a separate, confidential channel — never to log it, discuss it in committee, or investigate it within the general mechanism. That exclusion belongs in the plan explicitly, because the default behaviour of a well-run GRM, applied to a GBV disclosure, causes serious harm.