About and method | The Ledger SA

Sources

We only use public records: Stats SA, National Treasury, the Auditor-General, the Department of Water and Sanitation, the Electoral Commission, SAPS and each metro’s own notices, budgets and reports. Every chart names its source and links to it. Where we group or calculate, we say how.

Wards and place names

Ward numbers and boundaries are those used in the 2021 local elections (the Municipal Demarcation Board’s 2020 delimitation). The 2026 local elections use new boundaries, which may differ.

Suburb and place names are Stats SA’s Census 2011 sub-places and main places. Stats SA has not yet published Census 2022 place names; we will switch when it does. A place is linked to every ward that holds at least 10% of its area, so a suburb can appear in more than one ward.

Election results come from the Electoral Commission (IEC) by voting district and are added up into wards. The 2024 elections did not use wards, so 2024 voting districts are placed in 2021 wards using the IEC’s 2024 voting-district boundaries.

The ward boundaries, the IEC’s voting districts and the Census 2011 place boundaries are the official releases. We took them from SA-Maps, a public GitHub repository that mirrors those files, at commit e79972d7d65283877df3101a24270f5e1415d579, because the official download portals were not reachable from where we build the site. We check the mirror against the IEC’s own results: its 742 metro ward numbers match the IEC’s one for one.

The 2026 wards are the Municipal Demarcation Board’s 2026 ward delimitation, taken from the same SA-Maps open-data copy at the same commit, because the Board’s own site was not reachable from where we build. We will switch to the Board’s own download when it is. Their 2026 ward numbers match the IEC’s 2026 voting-district file in all eight metros.

Ward population is Stats SA’s Census 2022 ward estimate (SuperWEB2, Ward Estimates 2022), copied by hand because SuperWEB2 cannot be read automatically. It uses the wards of the 2021 elections. We never estimate a population for a 2026 ward. Where a 2026 ward is the same area as one or more whole 2021 wards, the page shows the official count of those wards (allowing 1% for slivers where the two boundary files were drawn separately). Every other 2026 ward page lists the 2021 wards it overlaps, with their official counts. Stats SA has not published Census 2022 services (water, refuse, toilets, lighting) by ward, so those are shown for each metro only.

Water and power notices are linked to wards by the places they name. We do not guess. A name is linked only when it matches an official place name exactly: Stats SA’s Census 2011 sub-places and main places (and, once loaded, the Municipal Demarcation Board’s or the metro’s own list of suburbs). A name that matches no official place, or more than one, stays unmatched. The one exception is a short list of rules a person has set from the city’s own usage: we count Malukazi as Umlazi X, the name eThekwini uses in its notices.

So a ward’s notice count can be lower than the number of notices that actually mention it. For example, Census 2011 splits Zwide in Nelson Mandela Bay into Zwide 1 to 4, so a notice that just says “Zwide” is not linked to a ward.

How we check figures

We publish figures exactly as the official source reports them. Automated checks look for figures that cannot be right: spending more than one and a half times its budget, a negative amount where that is impossible, a figure more than 25% away from another official figure for the same thing, a jump of more than three times (up or down) from the year before for amounts of R50 million or more, or a ratio outside its possible range.

A figure that fails a check is still shown, with an asterisk and a footnote: “As reported to [source]. This figure is inconsistent with [reason] and may be a reporting error.” Such figures are never used in rankings, headlines or “biggest” lists. Every figure links to its exact source, so you can check it yourself.

Money

Audit outcomes are the Auditor-General’s, as published by National Treasury’s Municipal Money for 2020-21 to 2023-24, and from the Auditor-General’s 2024-25 local government report for 2024-25. Treasury’s categories match the Auditor-General’s metro counts for the years where both are published.

Unauthorised, irregular, fruitless and wasteful expenditure comes from Municipal Money up to 2023-24 and from the Auditor-General’s 2024-25 report for 2024-25; earlier years are checked against the Auditor-General’s own figures. The data behind the financial health ratios come from Municipal Money (National Treasury). The ratios follow MFMA Circular 71: cash coverage is cash and short-term investments, less overdraft, divided by a month’s operating costs excluding depreciation, impairment and losses on disposal (Treasury’s dataset has no line for unspent conditional grants, so they are not deducted); the collection rate is opening debtors plus billed rates and service charges, less closing debtors and bad debts written off, divided by billed revenue; creditor payment days are trade creditors divided by credit purchases (bulk purchases, inventory, contracted services, operating costs and capital spending) times 365; capital budget spent is audited capital spending on new, renewed and upgraded assets divided by the adjusted capital budget; repairs and maintenance are divided by the value of property, plant and equipment and investment property.

Spending lines (catering, consultants, legal fees, overtime, travel, security, advertising, interest on late payments, councillor pay) and water and electricity losses are read from each metro’s audited annual financial statements, for the municipality itself and not its entities. The exception is water and electricity losses where a metro’s utility company runs the service; those are marked. Metros label these lines differently, so for each statement we record which line is used, and the figure links to its page. Where a metro does not show a cost separately, we say so rather than estimate it. Councillor pay is also checked against Treasury’s dataset.

Where a metro’s audited statements do not show one of these lines separately, we try other official sources in this order and use the first that has the figure: other notes in the same statements; the following year’s audited statements, where they restate the line for the year before (marked “as restated in [year] statements”); the line as the metro reported it to National Treasury in its monthly mSCOA returns (Section 71, the sum of the twelve months, unaudited), marked “As reported to Treasury (unaudited)”; Treasury’s budget schedules; the metro’s annual report, marked “annual report”; and, for interest and penalties, the Auditor-General’s report. Treasury publishes mSCOA lines per metro only for 2023-24 and 2024-25. We match them by the mSCOA item names: catering services, legal advice and litigation, overtime, security services, interest on overdue accounts, travel and subsistence, and advertising, publicity and marketing. Consultants cannot be matched, because the same item names are also used for outsourced services. Treasury’s travel and advertising categories are wider than most statements’ lines (they include staff own-transport claims, and recruitment and tender notices), so we say so next to those figures. Where both exist, the audited figure is shown, and it gets an asterisk if it differs from the Treasury return by more than 25%; travel and advertising are not checked this way, because the two cover different things. Cape Town’s figures for overtime, travel and advertising before 2023-24 come from the cost containment table in its annual report, which is unaudited. Where the only copy of a metro’s statements is a scanned image (Mangaung’s municipality-only statements for 2021-22), we read it with text recognition (OCR) and mark the figures “read from scanned document”; the same checks apply. Where the statements print a line as “-” (nil), we show R0 with the note “shown as ‘-’ (nil) in the statements”; if the metro’s Treasury return for the same year is not zero, the R0 gets an asterisk. If no source has the line, it shows “Not reported separately by the metro”.

Per resident figures divide by Stats SA’s mid-year population estimates (Mid-year population estimates 2026, district estimates 2002 to 2026, published 30 Jul 2026). A financial year runs from July to June; we use the mid-year estimate of the year it ends in, so 2024-25 uses the 2025 estimate. Per-day figures divide the year’s total by the days in that financial year.

Three kinds of data

Official figures come from government and statutory bodies and are labelled “Official: [source]”. Independent figures come from named, credible non-government sources and are labelled “Independent source: [name]”. Community reports from residents, when we add them, will be labelled “Community reports, unverified”. We never mix them, and only official figures feed headlines, rankings and image cards.

Headlines and image cards

Each page opens with one finding from official figures, usually how many metros meet one of National Treasury’s norms. A figure that failed our checks is never used in a headline, a ranking or an image card; the metros left out are named. Every big number and comparison chart can be shared as an image (square, and 9:16 for stories) that carries its source line. “Last updated” is the date the figures on that block last changed on this site.

What the words mean

Distribution losses: Water or electricity a metro buys or produces but never bills: leaks, theft, illegal connections and faulty meters.
Non-revenue water: All water that earns no money, including water given free to poor households. Close to, but not the same as, water lost.
Treasury norm: A target National Treasury sets for every municipality in MFMA Circular 71, so metros can be compared on the same scale.
Capital budget: Money set aside to build, renew or upgrade things that last for years: pipes, roads, substations, clinics.
Adjusted budget: The budget after the council’s mid-year changes. Spending is measured against this one.
Repairs and maintenance: Spending to keep existing roads, pipes and buildings working. Treasury’s guideline is about 8% of what those assets are worth each year.
Property, plant and equipment: Everything physical a metro owns and uses: networks, roads, buildings, vehicles. Its value is the yardstick for maintenance.
Property rates: A yearly tax on property, based on its municipal valuation. It is the metro’s main tax.
Collection rate: How much of what the metro billed residents and businesses in the year was actually paid.
Money owed to the metro: Unpaid bills for rates, electricity, water, sanitation, refuse and other charges, from households, businesses and government.
Older than 90 days: Bills unpaid for more than three months. The older a debt, the less likely the metro is to collect it.
Fuel levy share: Part of the national tax on petrol and diesel, which national government shares with the metros.
Infrastructure grants: Money national government gives metros for a set purpose, such as housing land and services or public transport. Unspent money can be taken back.
Income tax assessed: The personal income tax SARS worked out from the tax returns of people who live in the metro. It is paid to national government, not to the metro.
Own revenue: Money a metro raises itself: property rates and charges for electricity, water, sanitation and refuse removal.
Transfers received: Money national and provincial government gave the metro in the year: the equitable share for running costs and free basic services, and grants for set purposes such as infrastructure.
Unemployment rate: Stats SA’s official rate: people of working age with no work who looked for work in the past four weeks and could start, as a share of everyone working or looking. It leaves out people who have given up looking.
Grants: Money national government sends to metros, mostly for building infrastructure and free basic services.
Audit outcome: The Auditor-General’s verdict on a metro’s annual financial statements and how it followed the law.
Cash coverage: How many months the metro could keep paying its running costs from the cash it holds.
Creditor payment days: How long, on average, the metro takes to pay its suppliers. The law expects 30 days.
Unaudited: Reported by the metro but not yet checked by the Auditor-General.
mSCOA returns: The monthly figures every municipality must send National Treasury, using one standard list of accounts.
Restated: A figure the metro corrected in the following year’s statements.
Ward ballot: In local elections, the vote for the ward councillor. The candidate with the most votes wins the ward.
PR ballot: In local elections, the vote for a party. It decides how many council seats each party gets overall.
Voting district: The area served by one voting station. The Electoral Commission counts and publishes results for each one.
Turnout: Votes cast (valid and spoilt) as a share of registered voters.
Energy availability factor (EAF): The share of Eskom’s power-station capacity that could actually run over a period, after planned maintenance, breakdowns and other outages. Higher is better.
Council seats: A metro council has one seat per ward plus the same number again (roughly) for party lists, shared out by the PR vote.
Ml a day: Megalitres a day. One megalitre is a million litres, about half an Olympic swimming pool.
Water-use licence: The national Department of Water and Sanitation’s permission for Rand Water to take water from the Vaal system: a permanent amount, and a higher temporary one. Rand Water shows each metro’s share as a level in Ml a day.

Neutrality

We are not affiliated with any political party, candidate, government body or campaign, and we do not endorse anyone. We report election results as published by the IEC and do not publish predictions or projections.

Corrections

If we get something wrong, we fix it and add a dated note to the story explaining what changed. Tell us at [email protected].

Who we are

We publish without bylines so the numbers stand on their own. Our method, sources and corrections are always public.

Funding

Independent and community-funded. We're self-funded for now, and as we grow we'll rely on support from readers. We never take money from political parties, candidates or campaigns.