Municipal finances in crisis: Johannesburg and Cape Town in historical comparison

It is widely acknowledged that the City of Johannesburg is in crisis (Harrison, et al. 2025). On an almost daily basis now, news articles and opinion pieces provide evidence of service delivery breakdowns, political and institutional dysfunction, and corruption, and in turn mounting community frustration. Frustration has led to communities choosing to fix service failings themselves rather than wait in vain for City maintenance, or taking to the streets in protest. In the second week of September 2025, the residents of Westbury, Coronationville and Ivory Park violently protested their neighbourhoods’ perennially dry taps as leaking infrastructure again failed to match water supply to rising summer demand (Evans, 2025).

While the fact of a crisis, and its symptoms, are plain for all to see, its precise contours, and its underlying causes and dynamics, are not fully visible. This interactive visualisation looks to shine more light on the mechanics of the crisis in the City of Johannesburg by comparing the state of its finances to that of the City of Cape Town for the period 2008/09 to 2025/26.

The data for this interactive data visualisation has been assembled from the Annual Reports and audited Annual Financial Statements (AFS) for both the City of Johannesburg and the City of Cape Town for the financial years 2008/09 through 2023/24.¹ All figures for these years are original budget, adjusted budgets and audited actual amounts as they appear in the AFSs. Where amounts have been restated for previous financial years, notably those that appear in the Summary of Financial Performance table in the AFS, these restated figures are used as the final and most historically accurate. For the financial years 2024/25 and 2025/26 the two Cities’ Annual Budget Book for 2025/26 provided the most recent as yet unaudited figures, and here the adjustment budget amounts were used in lieu of audited actuals. In some instances, as explained below, the Audited Financial Statements do not include crucial details, and so the analysis relied on the statements of prior-year audited figures as they appear in the Annual Budget Book. .

Operating revenue: budgets versus actual

Municipal budgets are divided into two parts, operating budgets and capital budgets. Capital expenditure is money spent on developing new infrastructure such as water, electricity and local road networks. The money used for this capital expenditure comes in from loans raised or bonds issued by the municipality, conditional grants provided by other spheres of government, or cash retained from where operating revenue exceeds operating expenditure.

Operating expenditure includes: employee costs; repairs and maintenance of infrastructure that has already been developed; purchases of electricity and water from bulk service providers such as Eskom and (in the case of Johannesburg) Rand Water²; the ‘finance costs’ of redeeming loans or bonds for infrastructure development; money set aside for bad debts resulting from consumers not paying for services consumed; and other more minor items. A municipality’s operating revenue is derived mainly from property taxes (rates); service charges (where residents, businesses and other spheres of government pay tariffs for services consumed); and an annual transfer from the national fiscus known as the equitable share of nationally raised revenue.

In municipal budgeting and accounting the money budgeted and actually raised for capital always precisely matches the expenditure. The trick is on the operating budget, where healthy finances require that a municipality: (a) ensures that revenue projected is actually realised at the end of the year, (b) keeps expenditure tightly within budget; and (c) does not allow actual expenditure to exceed actual revenue, resulting in a deficit on the operating budget.

An accounting nuance here is that municipalities are required to report as operating revenue any grants provided by other spheres of government that are destined to be used for capital expenditure, but the revenue is only recognised at the point at which the funds have been spent as capital. In the overall Statements of Financial Performance that appear in the AFS, total operating revenue is counted as inclusive of capital transfers recognised. While this is proper accounting practice, these capital transfers recognised – having already been spent as part of the capital budget – do not actually cover operating expenditure, and so excluding them from the picture of operating revenue arguably gives a truer sense of the state of the municipality’s finances.

Figure 1 compares the budgeted and final actual operating revenue for the City of Johannesburg (CoJ) and the City of Cape Town (CCT) over the years 2008/09 to 2025/26, excluding capital transfers recognised. The figures are budgeted and audited actuals as recorded in the AFS, except for the last two financial years where the 2025/26 budget is used. An important note here is that the ‘actual revenue’ is revenue recognised as either received or owed by customers at the end of the financial year, and not all revenue actually received on payment: some billings may not be realised and are eventually written off as bad-debts.

Three key points emerge from the figure. First, up until 2014/15 actual revenue recognised at the end of the financial year closely matched the original budgeted revenue in the City of Johannesburg. And in three of the six financial years audited actual revenue exceeded original budgets. Thereafter, actual revenue routinely dropped well below budgets (illustrated by the red shading between the budget and audited outcome lines). This suggests that prudent budgeting gave way to less realistic revenue expectations in the year before the 2016 municipal elections, with this situation worsening through the subsequent period of coalition governments. By contrast actual revenue exceeded budgeted revenue in almost every year in the City of Cape Town (as illustrated by the green shading between the budget and audited outcome lines), indicating far more credible budgets.

Second, revenue expectations were dramatically higher than actual in the years of the COVID-19 pandemic and the period of subsequent loadshedding. In 2021/22 the gap widened to R4,7 billion and stayed at R4,3 billion in 2022/23. Indeed, a closer analysis (see figure x below) indicates that underperformance on energy revenue is the main cause of budget targets not being reached. This affected both cities but Johannesburg much more severely – in 2022/23 the gap between audited outcome and budget fell to less than a billion Rand in Cape Town, but nonetheless stayed positive.

Third, with loadshedding diminishing as a factor in the last quarter of 2023/24, actual revenues surged for Johannesburg. Yet it still remained below that anticipated for the year by over R4 billion. Revenue expectations were then significantly tempered in the financial year just concluded, increasing by only R1 billion between 2023/24 and 2024/25. While adjusted budgets project the final figures exceeding budget, it remains to be determined whether audited actuals will see even these diminished expectations being realised. Incongruously, budgeted revenue has again spiked by almost R8,5 billion in the current 2025/26 financial year, raising questions about financial prudence in light of the historic trend, current economic conditions, and prevailing willingness of consumers to pay.

Figures 2 and 3 look at the main revenue sources of the two cities in more detail.

Figure 2 examines budgeted revenue from property rates for both Johannesburg and Cape Town over the period 2008/09 to 2025/26. Again, the figures to 2023/24 are from the audited financial statements and for the last two years projected from the current budget. Intriguingly, Johannesburg performs as well as, if not better than, Cape Town in budgeting accurately for this revenue stream. For the financial years 2020/21, 2021/22 and 2022/23 Cape Town saw audited actual revenue realised from property rates fall short of expectations by several hundred million Rand. By contrast, Johannesburg’s annual budgets for property rates have accurately predicted rates revenue recognised in most of the last fifteen financial years, and in a number of years the audited outcome has exceeded expectations. However, in comparative perspective, Johannesburg faces a major challenge with revenue projected from service charges, as shown in Figure 3.

Figure 3 illustrates that actual service charges revenue matched original budget expectations by the City of Johannesburg until 2011/12. Following this, the audited outcome of revenue recognised at the end of the financial year began to fall further and further behind budget projections. It is noteworthy that the change in 2012/13 and following does not seem to have been an overoptimistic expectation in the original budget. That line seems consistent with previous years. Rather the widening gap is due to a falloff in revenue actually recognised by the end of the financial year. In 2017/18 the gap surged to almost R4 billion, and in both 2022/23 and 2023/24 it had grown to over R4,6 billion, primarily driven by South Africa’s loadshedding crisis. However, by contrast the lines marking Cape Town’s budgeted and final actual recognised revenue from service charges have remained consistently close. Revenue recognised exceeded budgets in all years except 2016/17 and 2017/18, when Cape Town faced its Day Zero water crisis, and in 2022/23 when it too was clearly knocked by loadshedding.

What are the overall effects of successive years of lost revenue from service operations? Figure 4 compares the two cities on the cumulative impact of underrecovery. When each year’s gap between budgeted and audited actual revenue from rates and service charges is added up it is evident that by the end 2023/24 Johannesburg was cumulatively short some R25,3 billion on what it had originally budgeted. By contrast Cape Town had cumulatively underrecovered just R29 million by 2023/24.

Again it needs to be noted that this is not the full extent of the gap, since audited actual revenue recognised is not actual income received, with many consumers who have been billed for services received then failing to pay. At present the City of Johannesburg’s payment rate on rates and service charges stands at around 85%. So many billions more have been lost on bad debts.

Operating expenditure: budgets versus actual

Figure 5 compares the operating expenditure budgets and the actual expenditure of the Cities of Johannesburg and Cape Town, between 2008/09 and 2025/26.

In the three years 2008/09, 2009/10 and 2010/11, Johannesburg’s actual audited total operating expenditure exceeded the original budgets. The situation stabilised thereafter and over the last decade and a half actual expenditure has quite closely matched – and for most years has fallen below – the original budget projections. While it is positive that actual expenditure has in general not exceeded budgets, weaker operating expenditure relative to intention is not inherently good, and may reflect an inability to spend effectively, including on crucial objectives such as repairs and maintenance. So the comparatively wide gap between budgets and actuals in the City of Cape Town in the years 2015/16 through 2020/21 is no less a concern than its surge of actual expenditure in excess of budget in 2021/22.

However, the most noteworthy, and concerning trend in the graph is the spike in actual audited operating expenditure in Johannesburg in 2022/23 (some R1 billion over budget) and 2023/24, (more than R1,5 billion over budget). In 2024/25, Johannesburg budgeted for a far smaller increase in total operating expenditure than in previous years. It remains to be seen whether final actual operating expenditure was contained within this lower limit. Unwisely perhaps, the 2025/26 budget anticipates a further 7% growth in operating expenditure, from an estimated R75,4 billion to R80,6 billion.

It is worth exploring the main drivers of operating expenditure growth over the years. The three panels in Figure 6 compare the two cities on actual operating expenditure for five main items: employee costs; bulk purchases of water and electricity; infrastructure repairs and maintenance; finance costs (the capital redemption and interest the cities must pay on loans from the private sector for capital expenditure); and debt impairment (money set aside for bad debts on unpaid accounts for services).³ Figure 6a and 6b give the actual expenditures on these items for the City of Johannesburg and City of Cape Town respectively, with the starting value in 2008/09 and 2024/25 shown. The exception is repairs and maintenance, for which no figure is available for either city for 2008/09, and so the 2009/10 actual is used.

Figure 6c then illustrates growth over the period to 2024/25 by indexing all numbers to the same starting point of 100 for 2008/09 (with the exception of repairs and maintenance since no data was available for either city for 2008/09, so the curve is indexed to 100 for the following year 2009/10). This approach allows for growth in each item to be measured on a comparable basis. The red line shows inflation, also indexed to 100. Put simply, the value of R100 in June 2009, adjusted for inflation, would now be R216,6 in June 2025 – a 216,6% increase.

Total audited actual expenditure in the City of Johannesburg grew from R19,7 billion in 2008/09 to R80,7 billion estimated in 2025/26. This was a 409% increase, roughly double inflation. Total audited actual expenditure in Cape Town increased from R14 billion to 71,3 billion, a 509,3% increase.

.

Reports in the media have highlighted the perceived excessive growth in the costs of salaries of municipal employees. For example a recent article in a national newspaper headlined with the claim: “Startling truth behind the collapse of SA’s cities … SA’s major cities are running out of cash because they have become employment agencies instead of hives for the delivery of services.” (Marrian, 2025). In our two cities the evidence is more complicated. The graphs show that employee costs in both cities certainly grew much faster than inflation over the period 2010-2025 but, when viewed in comparison, it was not the expenditure item with the largest increase over the period. It is also worth highlighting that Johannesburg's salary bill growth was not greater than Cape Town’s. Employee costs in Johannesburg grew from R5,1 billion actual in 2008/09 (26% of total expenditure) to R21,6 billion budgeted in 2025/26 (27% of total). This was an increase of 424,4%. The curve does suggest that the City’s salary bill rose much more steeply than before when coalition governance started in 2016/17, and has only recently begun to taper off. Cape Town’s employee costs increased 457,6% over the same period, rising from R4,6 billion to R20,9 billion. However, because other expenditure items grew even faster the employee costs share of total expenditure dropped from 33% to 29%.

Weighing more heavily on both cities’ budgets was the growth in bulk purchases and debt impairment. In Johannesburg, bulk purchases of water and electricity increased from R5,4 billion actual in 2008/09 (28% of total operating expenditure) to R27,4 billion budgeted in 2025/26 (34% of total) – an increase of 504,3%. These are administered costs imposed by bulk providers Rand Water and Eskom over which the City has little control. In Cape Town, bulk purchases increased from R2,9 billion (21% of total operating expenditure) to R17,9 billion (25% of total) over the same period. While there was a slight dip in the curve in 2017/18 – the year of the city’s Day Zero water crisis – the overall increase in bulk purchases was 623,3%, three times inflation, and the largest of any of the five expenditure items.

Debt impairment is provision set aside for billed amounts not paid by consumers and needing ultimately to be written off as bad debts. In Johannesburg, debt impairment increased from an actual R1,5 billion in 2008/09 (7% of the operating budget) to a staggering R8 billion (10% of budget) budgeted for 2025/26. This was an increase of 547,5% since 2008/09. The curve clearly shows that the big increase in debt impairment came in 2019/20, when COVID-19 shutdowns and the resulting economic collapse saw both households and businesses struggle to pay bills. In this year provision for debt impairment surged to 13% of the operating budget.

Cape Town has fared moderately better on debt impairment, but has also shown signs of strain. Although it started from a lower base of R903 million actual in 2008/09 (6% of total operating budget), which dropped still further to R684 million in 2009/10 (4% of budget), debt impairment grew to R3,2 billion budgeted in 2025/26 (now 5% of budget). Although less than half of Johannesburg’s current provision, this represented a 356% increase since 2008/09. As in Johannesburg, Cape Town’s impairment provision spiked in 2019/20, to 7% of operating budget. The comparable curves in debt impairment across the two cities suggests that cost containment here is not simply a matter of sound financial management – both cities have been impacted by the same structural factors of societal inability and unwillingness to pay for services consumed.

As costs on other fronts, notably bulk purchases and debt impairment, have ballooned, budget space for infrastructure development and maintenance has been squeezed. In Johannesburg, provision for repaying loans and bonds stood at R1,3 billion actual in 2008/09 (6% of operating budget). It grew just 209,7% – lower than the total inflationary increase – to stand at R2,6 billion (just 3,3% of total budget) in 2025/26. While almost flat since 2017/18, there has been a further clear drop in this expenditure item in the 2024/25 and the current 2025/26 budget. Cape Town saw the same trend but with a sharp divergence in the latest budgets. Here, R408 million was spent on finance costs in 2008/09 (3% of budget). Growth on this item was then well below inflation to 2023/24, when R828 million was spent (just 1,5% of budget). However, budgeted finance costs have increased dramatically to R1,4 billion in the last two years, though this is still only 2% of budget. This reflects Cape Town’s until recently very conservative capital spending, discussed in more detail below.

It has become an article of faith in South Africa’s municipal finance that spending on repairs and maintenance needs to be much higher than it is currently, and at least 8% of the value of the municipality’s total property, plant and equipment, and investment property (PPE). However, it should be recognised that repairs and maintenance are not a directly counted cost that can be precisely determined on a city’s statement of financial performance. Rather it is indirectly counted as the ‘outcome’ of other expenditure on the basis of estimating what share of the salary bill is made up of workers who spend their time repairing and maintaining assets, the cost of contractors, the purchase of necessary materials (inventory consumed) and so on. In addition, where this spend simply maintains the useful life of an asset it is counted as repairs and maintenance, but where it expands the capacity of an asset or unlocks new capacity, it is recounted – off the operating budget – as capital. Furthermore, if more money is spent on new assets the denominator increases, and the outcome percentage decreases.⁴ So the estimated outcome is not a precise science. Johannesburg’s spend on repairs and maintenance currently hovers around 5% of PPE, while Cape Town’s is approximately 7.5%. It is clear from the curves that Johannesburg's spend on operations and maintenance increased very sharply until 2014/15 but then tapered off and hardly grew until 2023/24. However, the 2025/26 budget indicates a dramatic improvement, with spend on this item estimated at R6,3 billion for the current financial year. Cape Town has shown a similar trend, with spend on maintenance actually declining for several years after 2017/18, before rapidly increasing after 2021/22.

Operating revenue vs expenditure

Figure 7a shows the actual operating revenue relative to actual operating expenditure for the City of Johannesburg and City of Cape Town for each year between 2008/09 and 2025/26. Where operating revenue exceeds operating expenditure the cities have realised an operating surplus, and the difference is shaded in green. Where expenditure exceeds revenue the city has an annual operating deficit, with the difference shaded in red. Figure 7b shows the actual operating surpluses or deficits for each year when expenditure is subtracted from revenue.

.

The City of Cape Town has achieved fairly consistent operating surpluses over the full period shown, with the exception of 2014/15 and, understandably, the 2019/20 financial year when COVID-19 hit. In the City of Johannesburg, operating revenue exceeded operating expenditure for most years until 2016/17, when the arrival of a coalition government coincided with an unexpected deficit because of a spike in expenditure. The situation recovered thereafter, and unlike Cape Town Johannesburg showed a surplus in 2019/20. However, over the last three years the situation has deteriorated considerably, with deficits of almost a billion Rand in 2021/22, R3,3 billion in 2022/23, and R3,7 billion in 2023/24. While a surplus was projected for the financial year ending June 2025, the final outcome is awaited.

Figure 8 provides a little more texture by considering revenue versus expenditure for Johannesburg’s two core trading services, energy and water and sanitation. Only the picture for Johannesburg is given since directly comparable figures for Cape Town are not available. The graph is clear that revenue exceeded expenditure in Johannesburg’s water and sanitation entity, Joburg Water, across the entire period. However, electricity utility City Power presents a far less positive picture. Reasonably healthy surpluses were returned to the City of Johannesburg as sole shareholder until 2015/16. Then in 2016/17, the first year of coalition government, expenditure spiked relative to revenue, and the surplus narrowed. In the subsequent years both expenditure and revenue fell drastically and surpluses were negligible. In both 2022/23 and 2023/24 – the two years of most severe loadshedding – City Power showed a large deficit. It remains to be seen whether the projected surpluses for 2024/25 and 2025/26 will be realised.

Capital budgets

In comparison the two cities show mixed performance, though in overall terms it is fair to say that Cape Town demonstrates tighter and more credible budgeting, and the ability to keep expenditure within the available revenue. However there is one area where Johannesburg clearly outperforms Cape Town, even while performance in both cities is weak.

Figures 9a and 9b compare the actual capital expenditure compared to budgeted capital expenditure, for the City of Johannesburg and the City of Cape Town, over the period 2008/09 to 2025/26. A well-run city will have a growing capital budget and capital expenditure close to 100% of budget, reflecting that plans to develop infrastructure are being realised.

The most noteworthy trend shown in the graph is that capital spending in both the City of Johannesburg and the City of Cape has stagnated over the last decade. It is anomalous that in a city with Johannesburg’s infrastructure development and replacement needs, capital spend was considerably higher in nominal terms in 2014/15 and 2015/16 than it is currently. Cape Town’s capital spend barely grew over the decade from 2012/13 to 2022/23, although projected spending for the last and current budget years are now dramatically higher.

The most noteworthy aspect of both city’s capital budgets is the year-on-year inability to spend what was originally intended, as shown in Figure 10 and 11. In the City of Johannesburg actual capital expenditure exceeded or closely tracked capital budgets between 2008/09 and 2013/14. However over the last decade actual capital expenditure has lagged behind budgets. Spend against budget averaged 83% between 2014/15 and 2023/24, although it dropped (perhaps understandably) to a low of 52% during the COVID-19 period of 2019/20. The large excess of spend over budget in 2023/24 is simply due to an accounting quirk. As explained in the City’s Annual Report: “The overspending is due to the implementation of GRAP 13 Leases, which states that an entity will recognize an asset and corresponding liability if the lease meets the requirements of Finance lease Par 06. Therefore, the overspending of Capex is due to accounting treatment of leases rather than actual expenditure.”

It is noteworthy that capital expenditure against budget performance in the City of Cape Town has been worse than Johannesburg, until very recently. Capital spend in Cape Town has averaged 81% between 2014/15 and 2023/24.

Figure 12 shows the long term effects of this underspending of capital budgets. If all the years of underspending are added together the residents of Johannesburg cumulatively lost out on some R6,4 billion of infrastructure development between 2008/09 and 2024/25. In Cape Town the tally was a staggering R22,9 billion.

.

.

Notes

1. An earlier version of this analysis used data from South Africa’s National Treasury Municipal Finance Database (https://municipaldata.treasury.gov.za/), supplemented with the reports consolidated as excel files from data submitted by municipalities in terms of Section 71 of the Municipal Finance Management Act (MFMA) (Act 56 of 2003) (s71 reporting). However this data was not consistent with information appearing in the Annual Financial Statements of the two cities, and as of September 2025 was last updated for the 2022/23 financial year. Furthermore, at the time of writing, s71 reporting was not accessible on the National Treasury website. The analysis was therefore rebuilt using only data from the two Cities’ own reporting and budgets.
2. Rand Water abstracts water from raw water sources, mainly the Vaal Dam, treats it, and sells it at a ‘bulk supply’ tariff to municipalities in Gauteng and across a wider supply area stretching beyond the provincial boundary. Municipalities then reticulate this water from their reservoirs to residents across their jurisdiction, ideally billing them for their consumption. Note that Cape Town does not have an equivalent bulk water supplier, and the municipality itself purifies the water it distributes to residents. Cape Town’s bulk service costs are therefore almost exclusively for electricity.
3. These five items make up some 81% of total operating expenditure in Johannesburg and some 70% in Cape Town.
4. See p110 of the City of Cape Town’s 2024/25 budget book: "NT MFMA Circulars 55 and 70 set the ratio of operational repairs and maintenance to asset value (write down value of the municipality’s property, plant and equipment (PPE)) at 8%. The ratio outcome is 7.5% for 2024/25, which is slightly below the NT benchmark of 8%. The lower ratio outcome is due to an increase in investment in new assets, which will not require immediate repairs and maintenance."

References

Evans, J (2025). ‘Scarily predictable’ — Joburg water crisis reaches boiling point as tyres burn and rubber bullets fly. The Daily Maverick, 11 September 2025. https://www.dailymaverick.co.za/article/2025-09-11-scarily-predictable-joburg-water-crisis-reaches-boiling-point-as-tyres-burn-and-rubber-bullets-fly/

Harrison, P., Gotz, G., Nunez Carrasco, L., Seedat, R. (2025). Johannesburg’s problems can be solved – but it’s a long journey to fix South Africa’s economic powerhouse. The Conversation Africa, 9 June 2025. https://theconversation.com/johannesburgs-problems-can-be-solved-but-its-a-long-journey-to-fix-south-africas-economic-powerhouse-256013?utm_source=clipboard&utm_medium=bylinecopy_url_button

Marrian, N. (2025). Startling truth behind the collapse of SA’s cities. Business Day, 6 June 2025. https://www.businesslive.co.za/bd/politics/2025-06-06-startling-truth-behind-the-collapse-of-sas-cities/