The DA candidate puts water at the centre of his bid for eThekwini. Municipal records support the urgency, but expose the harder questions: what can be collected, what can be repaired, and how residents will know. Durban’s water crisis has an accounting counterpart. The city’s 2024/25 oversight report records R2.91 billion in water losses and R24.43 billion provided for doubtful consumer debts . One figure describes a costly failure in the water system. The other warns against assuming that money owed to the municipality will arrive when an infrastructure programme needs it. That is the central test of Haniff Hoosen’s pitch for eThekwini. The Democratic Alliance’s mayoral candidate has promised R10 billion for the water network , funded through spending cuts, debt recovery, grants and a shift in priorities. The ambition addresses an essential service. Its credibility depends on a funding schedule and a delivery system that can withstand more scrutiny than an election speech. In his conversation with Truth Report, Hoosen presents water as the first task of a prospective administration and proposes a public account of progress by his hundredth day. He links dependable services to Durban’s tourism prospects, small businesses and the ability of households to pay municipal bills. These are useful commitments to hold him to, provided that promises of investment are kept separate from promises of immediate results. A crisis that needs precise numbers Hoosen’s 10 September water speech cited more than 13,000 active leaks and losses exceeding 65% . Those figures should be attributed and dated. They are not interchangeable measures, and neither should be carried forward indefinitely as the current position. The municipality has since offered a different account. In a statement dated 21 October , it repeated Mayor Cyril Xaba’s announcement that reported leaks had fallen from 13,000 to 3,900 during an intervention programme. That is material counterevidence to a picture of complete inactivity. It remains the municipality’s account of its own performance. There is also a dispute about how that performance can be checked. The South Coast Sun reported councillor André Beetge’s challenge to the figures , including his complaint that daily outage information was not publicly available. His objection does not establish that repairs never happened. It does explain why residents need an accessible register showing which faults were closed, when repairs were completed and whether supply was restored. A reduction in reported faults cannot, by itself, establish an equivalent reduction in lost water. A small leak and a large burst count as one incident each; reporting practices and duplicate complaints can also affect totals. Measuring the volume entering and leaving a network answers a different question from counting repair jobs. The distinction is visible in the eThekwini Water Security Dashboard dated 30 January 2026 . For December 2025, it reported non-revenue water of 60.9%, with water losses of 47.5%. The broader measure includes authorised use that is not billed, such as some communal supplies and firefighting. Water losses themselves include both physical leakage and commercial losses, such as meter inaccuracies and unauthorised consumption. This is more than a dispute about vocabulary. Replacing a broken pipe, correcting an inaccurate meter and properly accounting for a communal supply require different interventions. A credible recovery programme must identify which problem it is paying to solve. Debt recovery is a funding plan only when the cash arrives Hoosen’s proposal to redirect spending and pursue outstanding accounts has a clear appeal. In answers published by the Sunday Tribune , he connected the pipe programme to an aggressive recovery drive and the removal of non-essential expenditure. But a debt balance is a claim on future payment, not a bank balance available to contractors. The 2024/25 oversight report records a R24.43 billion impairment provision because recovery of those consumer debts was doubtful. It also records management’s explanation that unemployment, affordability and other pressures constrain payment. This does not mean every impaired account is unrecoverable. It means the promise requires a realistic estimate of collections, with dates and contingencies. The practical questions follow. How much of the proposed R10 billion would come from additional collections, rather than revenue already needed for daily operations? Which grants would support which projects? What would be deferred if collections fell short? And what provision would protect households that genuinely cannot pay? The same discipline must apply to expenditure cuts. An identified saving becomes useful when it releases money that can lawfully and practically be reassigned. A list of unpopular activities is not yet a costed replacement programme. Voters should be able to see the projects, funding sources and procurement milestones to