What makes a grievance mechanism work: accessibility, escalation and the closure problem

1. Three properties, unequally difficult

A grievance mechanism is assessed on three properties. Accessibility asks whether a person with a complaint can lodge it without cost, literacy, travel or fear of retaliation. Escalation asks whether a complaint that is not resolved at first instance moves upward rather than circulating. Closure asks whether cases reach a determination within the timeframes the mechanism itself has published.

These are not equally difficult. Accessibility is largely a design problem and can be solved by multiplying intake channels. Escalation is a governance problem, solvable by specifying levels and authorities. Closure is an operational problem that recurs every day for the life of the project, and no amount of design work discharges it.

2. Why accessibility is over-solved and under-measured

Mechanisms typically offer several channels: a written form, a phone line, committee meetings, suggestion boxes, and increasingly a digital route. The design documents describe these fully. What is rarely established is which channels are actually used, by whom, and which populations are not represented among complainants at all.

Absence of complaints from a group is ambiguous evidence. It may mean the group has no grievances; it may mean the mechanism is unreachable for them. Distinguishing the two requires disaggregating intake by channel and by complainant characteristics, and comparing that distribution against the affected population. Mechanisms that do not do this cannot tell whether their accessibility provisions worked.

3. The closure problem

Closure is where mechanisms fail, and the failure has a characteristic shape. Cases are logged. Some are resolved quickly. A residue accumulates — the contested valuations, the eligibility disputes, the claims requiring a decision nobody at project level is empowered to make — and that residue ages without moving.

Reporting typically obscures this, because the natural summary statistic is a resolution rate, and a resolution rate computed over all cases ever logged improves mechanically over time as easy cases accumulate in the numerator. A mechanism can report an improving resolution rate while its oldest and most serious cases sit untouched.

The statistics that make the problem visible are different: age distribution of open cases, time-to-acknowledgement and time-to-closure as distributions rather than averages, and closure rates disaggregated by category. A mechanism reporting only aggregate resolution is not being dishonest so much as measuring the wrong thing.

4. Recording as a precondition

Each of these measures depends on information captured at intake and maintained through the case: when it was received, through which channel, from whom, in which category, when acknowledged, what was decided, when closed, and whether the complainant accepted the outcome. That last field is the one most often absent, and without it a mechanism cannot distinguish resolution from abandonment.

A grievance log that records only complaint text and a status field cannot answer any of the questions on which the mechanism will be assessed. This is the ordinary case, and it is why grievance performance is so often reported qualitatively — not because the mechanism failed, but because it was never instrumented to demonstrate that it worked.