⚠ This is an earlier IHS Markit release vintage than the 2011–2021 series above. Kept separate rather than merged since the two vintages disagree slightly on overlapping years (statistical re-estimation, not error).

⚠ This table is from StatsSA Census 2022, a different data source and methodology to the IHS Markit municipal estimates used elsewhere on this page (e.g. total population 4.04M here vs 3.65M IHS estimate for 2021 above). Both are legitimate but not directly comparable — kept separate by design.

⚠ This table has been carried forward essentially unchanged across at least 8 consecutive Annual Reports (2016/17 through 2023/24) — the Department of Water and Sanitation has not issued fresh Blue Drop scores since 2014/15 except for a partial 2023 re-assessment shown below under updated area definitions. *Some source columns transpose Bronkhorstspruit/Bronkhorstbaai population and treatment figures between report-years; both values are shown where they differ.

⚠ Several years show two slightly different figures depending on which Annual Report is the source — both retained where they conflict (*marked).

⚠ "Total households" denominator is genuinely unstable across report-years — footnoted in source as "to include informal settlements" but the included scope appears to change between reports, which is why MSL+ percentage does not move smoothly even though the MSL+ sub-total grows steadily. Treat the % column with caution; the household sub-totals are more reliable. *One 2024/25 source table shows a likely typographic outlier of 14,213,161 for this cell (10x too large) — discarded in favour of 1,421,316.

⚠ Same household-denominator instability noted in the water table above applies here.

⚠ Ga-Rankuwa's estimated lifespan rose from 2 years (2022 estimate) back to 8 years (2025 estimate) — likely reflects an approved extension or methodology change rather than data error, but the swing is large enough to flag.

⚠ *2019/20 surplus/deficit appears as both a small operating surplus (R63.257m, before capital transfers) and a large deficit (-R4.32bn, including capital transfers and contributions, per the Budget vs Actual statement) depending on which statement line is read — this is not a contradiction once capital transfers are accounted for, but the two presentations should not be confused. The COVID-affected year (2019/20 Annual Report) describes an accumulated surplus decrease of R2.805bn for that year specifically.

⚠ This table is sourced from the 2023/24 Annual Report's own "Table 65: Financial Ratios of the City" (pp.272–273) and uses the City's current indicator set, which is not the same definitional basis as the "Financial Position Ratios — GROUP, 2016 to 2019" table above (e.g. that table's audited-financial-statement-based "Collection rate" of 88–93% versus this table's current-indicator-based "Collection rate ratio" of 90%). Net debtor days and Repairs and maintenance as a percentage of property, plant and equipment appear to use consistent definitions across both tables and can reasonably be read as one continuing series (66→79→68→63→69 days in 2016–2019, then 67→83 in 2022/23–2023/24); the remaining rows should be read as a separate, more recent indicator set rather than a direct extension.

⚠ This is the single largest data-quality issue found across all nine years of source reports. Reading down the "own year" column gives one trend; reading the "prior year restated" column from the following year's report gives a different number for the same year (e.g. 2022 net assets: R36.586bn per the 2021/22 report's own figure vs R31.361bn per the 2022/23 report's restated comparative — a R5.2bn gap. Similarly 2023: R32.306bn own-year vs R29.617bn restated, a R2.7bn gap). This is consistent with genuine, audited prior-period error corrections and transfers of functions between entities under common control, both of which are mentioned explicitly in the source statements — it is not believed to be a transcription error on this portal's part. The Overview and Financial charts on this portal use the "own year" column throughout precisely to avoid double-counting a restatement as if it were one smooth trend.

⚠ *2019/20 Total Assets/Liabilities reconstructed as Net Assets + Liabilities where the source page was truncated; verify against full statement if precision matters for this specific year.

⚠ 2021/22 and 2022/23 show an identical R113,457,210 maintenance figure — likely carried forward rather than independently re-measured.

⚠ *Several years carry two figures depending on source report; both retained where they genuinely differ. †One source table embeds running-total arithmetic ("3367 + 26482 = 29849") directly in the cell — preserved as-is rather than silently resolved.

⚠ Infringement notice counts vary by orders of magnitude between years — almost certainly different metric definitions, not a real decline.

⚠ *Firefighter counts differ between report-years for the same year, likely due to inconsistent inclusion of admin/support staff.

⚠ Figures decline sharply and are not smoothly comparable — reports appear to mix "actual" and "estimate/target" values inconsistently. The 2016/17 narrative separately cites an EMS division total of 423 staff and 70 ambulances, which does not match this table's per-year figures for that year. Treat as directionally indicative only.

⚠ The 10 libraries maintained in 2023/24 and the 10 maintained in 2024/25 are entirely different facilities — these appear to be rotating annual maintenance programmes rather than the same buildings recurring, so they should not be read as "no progress" on the earlier year's list.

⚠ Top and senior management headcounts dropped sharply (23→8, 827→300) while skilled technical roles grew — this looks like organisational restructuring rather than only attrition; worth verifying against governance/structural change records if precision matters.

⚠ 2021/22 column appears to undercount severely relative to every adjacent year — almost certainly a different/incomplete extraction in that specific report rather than a genuine 70%+ clinic closure. Treat with caution.

⚠ "14,8" for new business licence applications is ambiguous in the source — likely a formatting artifact for either 148 or 1,480. Preserved as-printed.