Karuma hydropower: managing a linear project beside a major works site

This is a public-record case study. It uses the project context named in the title to examine a RAP problem; it does not claim access to a project register, confidential settlement or unpublished audit. Project documents and current lender requirements should be checked before relying on it.

Thesis

Karuma demonstrates why complex infrastructure should not be managed through a single undifferentiated “PAP count.” A hydropower scheme can produce several legally and operationally distinct impacts: permanent land acquisition, temporary construction occupation, transmission corridors, access restrictions, damage to crops or structures, restrictions on natural-resource use and, depending on the project design and assessment, wider economic effects. IFC PS5 is deliberately impact-based: it covers physical displacement and economic displacement and applies where land acquisition or restrictions on land use produce significant adverse impacts. Guidance Note 5 further recognises that significant impacts on land, assets or access can require PS5-type treatment even where the formal mechanism is not straightforward acquisition. World Bank ESS5 similarly distinguishes physical and economic displacement and requires the resettlement instrument to be proportionate to the impacts. The analytical error is to place all affected people into one denominator and then declare the RAP “90% complete” because 90% of names have been paid. Such a metric can conceal unresolved temporary occupation, restoration obligations, transmission impacts or livelihood losses. The stronger legal and evidentiary model is an impact register: each affected interest should carry a unique identifier linking the person or entity, tenure or use right, asset, impact type, entitlement, valuation basis, payment, mitigation obligation, responsible party, deadline, grievance and verification evidence. This is not bureaucratic excess; it is what makes a complex project auditable. Authorities: IFC PS5 paras. 1, 7–9 and Guidance Note 5; World Bank ESS5 and Guidance Note; Uganda Constitution Art. 26(2); Cernea's impoverishment-risk framework.

The legal issue in this case

Karuma raises the legal classification problem created by a mixed project footprint. Permanent acquisition, temporary occupation, transmission connections and access restrictions can trigger different compensation, assistance and reinstatement duties. A single RAP category can obscure the legal consequence of duration and control over the affected asset.

The project question

The factual question is whether the record follows each impact as the design changes. A dam project may move an access track, add a contractor camp or revise a transmission route after the first census. The issue for counsel is whether those changes triggered screening, re-disclosure, re-survey, budget revision and a fresh decision for newly affected people.

What the scholarship still needs to establish

The research gap is change control in African hydropower RAPs. Published analysis usually describes the approved project, not how social commitments change when engineering changes. A scholarly study could compare design revisions with revised affected-person lists and determine whether the social assessment was adaptive or merely backfilled.

Evidence a lawyer would request

Request the design baseline, alignment revisions, contractor land records, temporary-occupation agreements, asset photographs, reinstatement inspections and grievance trends by project component. The case is credible only if the reader can distinguish a permanent entitlement from a contractor obligation and see who owned closure of each.

Legal frame: Uganda

The legal starting point is Article 26(2)(b) of the Constitution: compulsory acquisition or possession for a public purpose requires prompt payment of fair and adequate compensation before the taking. The Land Act, Cap. 227 supplies the tenure and compensation context, including the disturbance allowance in section 77, while the Land Acquisition Act, Cap. 226 supplies the acquisition procedure.

Those provisions do not by themselves answer every lender-standard question. Eligibility of people without formal title, replacement cost without inappropriate depreciation, livelihood restoration, meaningful disclosure and a project-level grievance mechanism must be analysed under the applicable financing and safeguard instruments as additional obligations. Exact statutory wording and current amendments should be checked against the authoritative text before legal reliance.

The authorities below are starting points for verification. They are not a substitute for checking the consolidated law, regulations, cases and project agreements applicable to the specific acquisition.

• Uganda Constitution and legislation portal ↗
• Uganda Land Act, Cap. 227 ↗
• Uganda Land Acquisition Act, Cap. 226 ↗

The setting

Karuma illustrates the mixed footprint of a large Ugandan energy project: permanent works, access roads, transmission connections, worker activity and temporary construction areas can affect different people at different times.

The RAP problem

The risk is building one undifferentiated caseload. A person affected by a tower, a temporary access road or a reservoir-related change may need a different measurement, entitlement and reinstatement test from a household that relocates permanently.

What the record should preserve

The project record should identify the impact geometry, duration, responsible contractor and closure evidence for every affected asset. Spatial versioning matters: a revised alignment must not silently overwrite the survey that supported the original consultation.

Reading the case through the standards

The legal and safeguards question is not whether a project can produce a compensation schedule. It is whether the schedule can be connected to a lawful acquisition process, a fair valuation method, an eligibility decision and an outcome that the affected person can actually experience. In Uganda, Article 26 of the Constitution places prompt, fair and adequate compensation before compulsory taking; the Land Act adds the tenure and disturbance-allowance context. Across the lender standards, the inquiry is wider still: people without formal title, economic displacement, replacement cost, livelihood restoration, consultation and grievance access all need a place in the project record.

That does not mean that every project is governed by Ugandan law or that a lender standard replaces national procedure. It means the case has to be read at the intersection of the applicable regimes. A title search may be legally necessary and still be an incomplete census. A signed agreement may prove that a document was executed and still be weak evidence that the amount was understood or that a livelihood was restored. A grievance may be closed administratively and still reveal a repeated design failure. Credibility comes from showing which proposition each document proves and which proposition requires a different kind of evidence.

What a serious case analysis would test

The first test is the impact boundary. Compare the engineering footprint with the social footprint: parcels, structures, crops and trees, but also access, customers, employees, common resources, seasonal users and people who depend on an affected household. The second is the decision chain. For each person or asset, can a reviewer move from identification to measurement, valuation, entitlement, approval, payment and remedy without relying on an unexplained spreadsheet change? The third is time. Notice, payment, possession, relocation, reinstatement and livelihood recovery are different events, and a project that reports only the earliest completed event will overstate its performance.

The fourth test is distribution. Aggregate completion figures can hide delayed claimants, women whose interests were recorded under another household member, tenants without title, vulnerable people who could not use the complaint channel, or communities waiting for a shared commitment. A credible monitoring sample should therefore be designed to find the difficult cases, not only to confirm the median case. It should preserve the reason an exception was made, who approved it, what the affected person received and what evidence supports closure. This is also where corruption risk becomes an evidence question rather than a rhetorical accusation: unexplained overrides, duplicate claims, missing measurements, payment mismatches and suppressed complaints are control signals that require investigation, not automatic conclusions of misconduct.

Research gap and practical implication

Public project material usually tells us what an instrument promised and, sometimes, what was paid. It rarely follows the same household from the baseline through implementation and into a measured outcome. That is the significant research gap running through East African RAP practice. A stronger study would combine the public legal and project record with anonymised household interviews, geospatial change, grievance trends and livelihood indicators. It would distinguish documented fact, reported allegation, project assertion and independent finding. It would also state what cannot be known from the available record.

Sources and limits of the public record

The article’s cited materials establish the standards, legal context or public accountability framework; they do not necessarily establish every factual proposition about the named project. A scholar should separate a primary legal rule, a lender requirement, a project disclosure, a reported allegation and an independently verified finding. That separation is not pedantry: it is what prevents a case study from laundering an assertion into a fact. It also identifies the next research step, whether that is a court record, a disclosed RAP, a monitoring report, a household interview or a project response.

• Performance Standard 5: Land Acquisition and Involuntary Resettlement ↗ — International Finance Corporation, 2012.
• Good Practice Handbook: Land Acquisition and Involuntary Resettlement ↗ — International Finance Corporation, 2023.
• The Land Act, Cap 227 (as amended by the Land (Amendment) Acts 2004 and 2010) ↗ — Ministry of Lands, Housing and Urban Development, Republic of Uganda, 1998.

For practitioners, the implication is immediate: design the register as a chain of evidence before the first payment. Give every affected person and asset a stable identifier. Keep versions of the footprint and entitlement decision. Store the source of each rate and the reason for each exception. Record consultation in a way the affected person can use, protect sensitive grievances without making them invisible, and assign every livelihood or reinstatement commitment an owner, date, indicator and closure document. A case study is useful when it changes those controls, not when it merely provides a memorable project name.

Questions that test the thesis

• Which impacts are permanent and which expire when construction ends?
• Can every asset and claimant be traced to a mapped project component?
• What proves that temporary land and access were restored to an agreed condition?