Monitoring indicators for resettlement: inputs, outputs and the outcomes nobody measures
Abstract
Resettlement monitoring systems reliably report what was done — payments made, meetings held, grievances logged — and rarely report whether anything was achieved. This paper examines the indicator hierarchy in resettlement monitoring, the structural reasons outcome measurement is displaced by output measurement, and the design requirements for an indicator set capable of detecting failure rather than confirming activity. It argues that indicators defined at project closure necessarily measure whatever data exists, and that specifying them before enumeration is the only point at which the choice is genuinely open.
1. The hierarchy and where reporting stops
Monitoring distinguishes inputs (budget committed, staff deployed), outputs (households compensated, training sessions delivered, grievances logged), outcomes (income restored, livelihoods reconstructed, disputes resolved to complainants' satisfaction) and impacts (poverty status relative to a counterfactual).
Resettlement reporting concentrates almost entirely at the output level. This is not laziness. Outputs are generated by the process itself as a byproduct of doing the work, available immediately, unambiguous and cheap. Outcomes require returning to households, asking questions whose answers are contested, and waiting long enough for effects to appear.
2. Why outputs displace outcomes
Three forces push reporting toward outputs. Timing: outputs are available during implementation when reporting obligations fall, while outcomes appear after the project's reporting cycle has closed. Attribution: an output is unambiguously the project's, whereas an income change may reflect rainfall, prices or policy. And incentive: outputs are almost always favourable, since a programme that paid everyone it enumerated has a good number to report, and that number says nothing about whether payment restored anyone.
The combined effect is a reporting record that is accurate, complete and silent on the question the standards actually ask.
3. What an outcome-capable set requires
Outcome indicators impose four conditions that output indicators do not. A baseline measurement of the same variable, or there is nothing to compare against. Stable household identifiers, or the same households cannot be revisited. A defined observation period extending past construction. And disaggregation by sex, vulnerability status and loss category, since an aggregate that improves can conceal a subgroup that has collapsed.
The last is worth emphasising because aggregate reporting is the norm and distributional failure is the characteristic failure of resettlement. A programme in which most households recover and a minority is destitute reports well in aggregate and has failed precisely the people the standards exist to protect.
4. Timing of definition
The practical conclusion is narrow and important: indicators must be specified before enumeration, not before reporting. Indicators defined at closure can only be constructed from data that happens to exist, which means they will be output indicators, because those are what the process generates unbidden.
Specifying indicators first inverts the relationship — the baseline instrument is then designed to serve the indicators rather than the indicators being derived from whatever the baseline captured. This is the single decision that most determines whether a completion audit can be answered with evidence or with assertion.
References
- [1]Performance Standard 5: Land Acquisition and Involuntary Resettlement. International Finance Corporation, 2012.
- [2]Guidance Note 5: Land Acquisition and Involuntary Resettlement. International Finance Corporation, 2012.
- [3]Handbook Module 5: Livelihood Restoration and Improvement. International Finance Corporation, 2023.
- [4]Good Practice Handbook: Land Acquisition and Involuntary Resettlement. International Finance Corporation, 2023.
- [5]Environmental & Social Issues Update. Office of the Compliance Advisor/Ombudsman (CAO), 2023.