Before the next Thwake promise, show the cost of reaching the tap

A dam can be close to completion while the service it is meant to provide remains some distance away. That is the question Kenya should ask about Thwake. The public needs to know not only when a reservoir will hold water, but what remains to be built, financed and operated before that water reaches households, farms and businesses. A completion percentage may describe a structure accurately and still leave the most important part of the public bargain unexplained.

The African Development Bank’s recent account of a mission to Thwake is useful precisely because it identifies unfinished tasks. The mission took place in September, and the bank’s October statement discusses financing conditions, contractor mobilisation, environmental and social issues, and grievances from people affected by construction. It also records the government’s January 2027 target for impounding water. These details show a project moving through several connected obligations, rather than a single switch waiting to be turned on.

The bank separately announced additional financing in June. Approval of financing is an important step, but it should not be confused with disbursement, payment to a contractor or delivery of a completed asset. Nor should financing for one phase be presented as proof that every downstream component is fully funded. My argument is that the government should publish one reconciled account showing those distinctions. It would allow supporters and critics of the project to argue from the same set of facts.

That account should start with a map of what each phase is expected to deliver. The dam wall, water treatment, distribution, irrigation and power facilities are not interchangeable achievements. A household interested in a reliable water connection may reasonably ask why a large completed structure has not yet changed daily life. The answer may lie in the sequence of works rather than a failure of the dam itself. But that sequence must be explained before expectations are raised, not after they are disappointed.

The next question is cost. A useful public statement would separate the original budget, approved changes, amounts committed, amounts disbursed and amounts spent. It should identify the currency and the date used for any conversion. Figures presented in different currencies or covering different phases can sound contradictory even when they describe different things. A reconciliation would prevent honest accounting differences from becoming needless suspicion, while making genuine unexplained increases harder to hide.

This is where public-debt scrutiny must become more exact. Additional borrowing is neither proof of misconduct nor proof of responsible completion. It may reflect necessary engineering changes, inflation, delays, a revised scope or poor initial planning. Those explanations have different implications. The state should identify which applies, provide the supporting documents and explain the alternatives considered. A borrowed amount is a fact to investigate, not a verdict that can be reached by attaching an angry adjective to it.

There is also a difficult decision that deserves to be discussed openly. Once substantial money has been spent, abandoning a project can leave the public with liabilities and little benefit. Yet the fact that money has already been spent cannot justify every further commitment. The question is what the additional expenditure is likely to achieve from this point forward, compared with feasible alternatives. An independent assessment should address that question instead of treating completion as self-evidently wise or self-evidently wasteful.

Such an assessment needs to include the cost of operating the eventual service. A completed asset requires maintenance, skilled staff, monitoring and arrangements for repairs. If these costs are omitted from the political story, the public can be invited to celebrate construction without understanding the continuing obligation. An honest project case should explain who will operate each component, how recurrent costs will be met and how performance will be measured after the construction teams leave.

The people affected by the works must also be visible in that account. The bank’s mission report notes grievances involving homes affected by nearby construction and stresses timely compensation and monitoring. Those concerns should not be treated as a footnote to engineering progress. A household dealing with damage or a disputed claim is part of the project’s real cost. Its unresolved position does not disappear because the wall is nearly finished or because a ceremony has been pencilled into a calendar.

At the same time, the public should avoid turning every outstanding claim into an established debt of a particular amount. Some claims may require investigation or resolution through a defined legal process. The demand should be for a clear register of categories, a fair determination and a traceable remedy. Personal information can be protected while publishing aggregate progress. What matters is that affected people can find their own case and that the responsible agencies can be held to a timetable for action.

The government should also distinguish promises made by different institutions. A ministry may set a construction target, a lender may approve a financing package and a contractor may propose a work programme. These statements are related, but they are not identical commitments. A single public progress report should reconcile them and identify the assumptions on which each depends. If a deadline relies on a disbursement condition being met, readers should be told that condition in understandable terms.

This would improve the treatment of delay. Some delays may be avoidable; others may arise from necessary safety work or a legitimate dispute. Publishing the reason allows the public to distinguish the two. Quietly replacing one date with another creates the impression that deadlines are ceremonial. A revised schedule should preserve the previous commitment, explain the change and show the effect on cost and service delivery. The historical record is part of accountability, not an embarrassment to be edited away.

There is a role for Parliament and auditors beyond asking whether the project will finally be finished. They should examine the relationship between the financing agreements, the approved scope and the benefits being promised. Are downstream obligations included in the public cost estimate? Have changes been independently evaluated? Is the operating arrangement credible? Those questions do not require legislators to become dam engineers. They require them to insist that the relevant engineering and financial judgments are available for scrutiny.

A public dashboard could help, but only if it links to actual evidence. A green indicator beside the word progress is not an explanation. The useful material would include dated milestones, expenditure categories, outstanding conditions and summaries of independent inspections. Where disclosure must be limited, the reason should be specified. The aim is not to publish every technical drawing. It is to give citizens enough information to understand whether the state’s latest promise is supported by a workable plan.

Critics should meet the same standard. A claim that funds were stolen requires evidence of the transaction and the wrongdoing alleged. A claim that the whole project should be abandoned requires consideration of what abandonment would cost. Scrutiny loses force when it moves directly from frustration to a conclusion that has not been demonstrated. The most effective challenge is often narrower: this increase lacks an explanation, this condition remains unmet, or this promised service has no disclosed financing arrangement.

That discipline is particularly important in debates about odious debt. A controversial project or an unpopular government does not automatically settle the legal character of a borrowing obligation. The evidence, applicable rules and potential remedies must be examined separately. Kenya needs the capacity to investigate questionable commitments without making unsupported declarations that expose the public to further uncertainty. A strong case is built from agreements, authorisations and the use of funds, not from the intensity of a political speech.

Supporters of Thwake should welcome that approach. If the additional work is necessary, the financing is justified and the service plan is sound, disclosure strengthens the case for completion. It helps distinguish a defensible public investment from the suspicion that gathers around repeated announcements. Confidence should be earned by explaining the difficult parts as well as the encouraging ones. The presence of unresolved work is not itself a scandal; refusing to account for it is a different matter.

The immediate request is therefore practical. Before the next major promise, publish the remaining work, the financing status of each component, the arrangements for affected households and the route from stored water to an operating service. Give each milestone a responsible institution and a dated update. Report what has changed since the last statement. That would turn a large national project into a sequence of commitments that citizens can follow and assess.

Benefits should be tested against disclosed assumptions as carefully as costs. If the investment case depends on future demand or on another component being completed, the assessment should say so. That would allow an independent reviewer to examine the basis of the forecast instead of being asked to accept an encouraging outcome as inevitable.

Thwake should eventually be judged by the useful service it delivers and the full cost at which it delivers it. A reservoir, an inauguration and a financing approval are milestones along that route. They are not substitutes for the destination. The government owes the public an account that reaches beyond the dam wall, through the distribution system, and all the way to the person waiting for water. That is where the investment’s promise becomes a lived result.