Section 01
Introduction
Johannesburg is South Africa’s most important city. It is home to the largest number of people and constitutes the largest economy in southern Africa. It also has a great many challenges, some of them existential: high levels of unemployment and poverty, political instability, large numbers of households in informal housing, low levels of density resulting in considerable economic inefficiency… the list is long and it would take even a high-functioning administration considerable time to ameliorate, much less address, all of Johannesburg’s many challenges. As a result, the city’s infrastructure is decaying, leading to chronic service delivery failures across the portfolio of its activities; its planning and administrative functions are poor; a large fraction of households is not paying for services; corruption has become endemic in municipal structures… again, the list is long.1
Given all the challenges, charting a way forward for improving the quality of services, investing in infrastructure and containing costs for households and businesses will have to be at the centre of the work of the next administration. This report seeks to provide some guidance to policymakers about the current state of the City’s finances and the extent to which the level and composition of both the city’s income and expenditure might facilitate or impede the process of addressing Johannesburg’s many challenges. It does this by:
- Setting out the structure of Johannesburg’s budget, including sources of its income and the nature of its spending
- Identifying key trends in income and expenditure patterns over the past decade
- Identifying some key sources of risk to the sustainability of the City’s finances, including the build-up of both defaulting debtors and unpaid creditors
- Identifying a seemingly impossible assumption about the City’s capacity to manage costs embedded in the most recent budget.
Key headlines
To anticipate some of the conclusions presented below, highlights of the analysis include:
- In 2024/25, the last year of audited figures, the City spent R79.9 billion with accounting revenues of R81.1 billion, leaving an accounting surplus of R1.2 billion.
- Most of the City’s revenues derive from selling electricity and water, and levying rates on residents’ properties. Most of its expenditure is on the purchase of electricity and water, and the salaries of ~34 400 employees.
- The City’s revenues in 2024/25 are the equivalent of about R19 000 per adult resident or R40 600 per employed person
- The City’s cashflows in 2024/25 roughly matched the accounting surplus (R1.8 billion v R1.2 billion), but that was the result of net borrowing of over R3 billion, absent which the cashflow deficit would have been R1.7 billion. In unaudited 2025/26 financials, cashflows were negative (a deficit of R3.1 billion), though debt repayment exceeded borrowings.
- The City is solvent as assets far exceed its liabilities, but most of its assets are hard to value and illiquid, making this a somewhat academic metric.
- Spending and revenue have increased by over 7.7% and 6.8% a year for the past eleven years, which is far faster than inflation, which ran at about 4.8% a year over the period. In inflation-adjusted per capita terms spending and revenues are similar to levels seen in the mid-2010s.
- While the data are difficult to parse, the slow pace of economic growth in Johannesburg means that the City’s revenues and spending account for a larger share of the economy now than ten years ago and residents now pay a larger share of their income in rates and service charges. This raises questions about the sustainability of the current trajectory, since the City’s operations must be affordable to its residents. This is reinforced by the observation that the City’s inflation-adjusted revenues have grown five times faster than the rate of growth of employment in the local economy, a differential that is unsustainable.
- A major shift is underway in the composition of the City’s revenues as electricity sales decline, both as a result of loadshedding and as a result of residents’ defection from the grid.
- Expenditure growth in the City has been driven by employee costs, debt impairment costs and spending on bulk water.
- Although the City’s balance sheet reflects that outstanding debts owed by customers has risen at about over 9% a year between 2014/15 and 2024/25, this grossly understates the challenge the City faces because that figure is net of provisions against bad debt. The underlying numbers show that unpaid bills have risen at the rate of nearly 17% a year from 2015, rising from R15 billion to over R71 billion in that time.
- By 2024/25, the rate at which unpaid debts were accumulating amounted to 15 to 17% of all the City’s billings, reflecting a massive challenge of sustainability.
- Although the City’s employee-related costs have risen by more than 9% a year, five percentage points faster than inflation, this appears to be driven by a rise in the number of employees rather than a rise in average remuneration. The composition of the City’s establishment appears to have changed over time, however, with the in-sourcing of previously outsourced services. Since the early 2020s, average remuneration has increased at 10% a year. Care should be taken with these numbers as there appear to be a number of problems with the City’s reporting of personnel-related data.
- In inflation-adjusted terms, investment spending by the City has fallen by roughly 50% over the past ten years. In real per capita terms, investment spending has fallen by 70%.
- The medium-term budget frameworks adopted in 2025 and 2026 make a number of extremely optimistic assumptions about trends in revenues, cashflows and costs, all of which generate unprecedently large surpluses. The plausibility of these assumptions is highly questionable, however, and the projections in the 2025 medium-term framework have already proved to be badly wrong.
Section 02
A note of the financial data used
This report is intended primarily to assist a new administration to quickly get to grips with the financial dimensions of Johannesburg’s crisis. It is also intended to help inform the public about the nature of the challenges the City faces. The goal is to try to make the issues as accessible as possible. By their nature, however, analysis of public finances is highly technical in character. We have sought to make the report as accessible as possible, but there are limits to how accessible it can be without becoming unduly pedantic for its primary audience.
Analysts must make some important choices when seeking to describe the state of the City’s finances (and note that, when capitalised, the “City” refers to the municipality itself and its entities, not the geographic city). Chief amongst those choices is which financial data to use: the City produces a wealth of budget documentation each year, with most budgets including comparable figures from one or two previous years, as well as a budget for the current or forthcoming year, and projections for the next two years. These figures are frequently updated. Budgets can be adjusted in-year through a formal adjustment budget, and actual revenues and spending may depart from budgeted numbers (even after adjustments are approved) for a wide variety of reasons, most of which are benign. This is one important reason to rely on audited financials when conducting an analysis of a government institution’s finances. Another is that unaudited figures are more likely to contain mistakes. This is evident, for example, the so-called s71 reports that municipalities and metros file with National Treasury, and which National Treasury publishes on its website. Figures relating to Johannesburg’s cashflows on the Treasury website for 2019 and 2020 clearly contain uncorrected errors or did so when last accessed in early 2026.
Having said that, it is worth noting a number of surprising deficiencies in the financial data contained in the City’s budgets over the past decade. The most confounding is that some of the documentation is simply absent from the City’s website. This does not mean that no budget was passed or that information is being deliberately withheld. It means only that some documentation had not been uploaded by whomever is supposed to administer the website. While gaps in the availability of audited figures for all years under review were eventually addressed, there remained more significant challenges.2 One of these is that the published data on the number of people employed by the City are seriously deficient and, likely, unreliable, with inexplicable and seemingly implausible changes from one year to the next. Sometimes, the data are not published at all, and, in at least one case, the data tables from one year appear to be simply reproduced in the next year without being updated at all. A degree of skepticism about the personnel numbers reported here is, therefore, warranted.
Similar points could be made about other kinds of data—such as the purchase and sale of electricity and water, the numbers for which in the City’s annual financial statements are sometimes quite different from those published in City Power and Johannesburg Water annual reports. In this regard, some of the price data in the annual reports of City Power and of the City seems to imply a rate of increase in unit prices that is much lower than actually occurred, and which seems to us to be wrong. It is beyond the scope of this exercise to reconstruct accurate numbers, but we have sought to use as consistent a set of data as possible, while recognising that it likely contains errors.
All the data used in this report come from audited financial statements of the City of Johannesburg (using the group financials rather than the metro council in order to include the consolidated results for the City’s entities).
Engagement with the 2026/27 budget
The first draft of this report was completed in April 2026, before the adoption of the 2026/27 budget by the City. The adoption of that budget was controversial because it flew in the face of a letter sent to the City by the Minister of Finance and leaked to the media by Helen Zille. The Minister’s letter expressed serious concerns about the sustainability of the City’s budget, concerns that were amply supported by the analysis in our draft and which we continue to share. The budget adopted by the City, however, did nothing to address those concerns. Indeed, the updated numbers in the budget for the 2026/27 year reflected concerns we had flagged in the draft of this report. Below, we will engage with these issues in a more detail. We have also updated our analysis of the three-year medium-term forecast for the 2027/28 and 2028/29 presented in the new budget, as well as how these figures depart from those presented in the 2025/26 budget.
Section 03
Outline of the report
The next section introduces the City’s budget, including how much money it raises as revenue and how much it spends, as well as some details on how the budget is allocated to different functions and activities. It also distinguished between accrual and cash-based income and spending. It also provides some details about the City’s outstanding debts and the money it is owed by customers/residents. Finally, it seeks to provide a sense of scale by looking at levels of revenue and spending in relation to the size of the city’s population and economy.
Having laid that foundation, the rest of the report focuses on (i) historical trends in income and spending as well as assets and liabilities, and (ii) an examination of the latest budget’s projections for the medium term, and, in particular, a deeply implausible assumption that is embedded in those projections.
Section 04
The latest audited figures: 2024/25
To understand the budget and the historical trends in revenues and expenditure, it is important to begin with a basic understanding of the main elements of the consolidated budget of the City of Johannesburg.3 To do this, we begin with the latest financial year for which audited figures are available—2024/25. The essential facts are that the City’s revenues (including capital subsidies provided by national government) ran to R81.1 billion, while its expenditure was R79.9 billion, resulting in an accounting surplus of R1.2 billion, a figure that expands to about R1.6 billion after a large tax credit on the operating deficit is accounted for. As we will see, however, this accounting surplus is not the same thing as a cash surplus.
The City raised most of its revenue by selling electricity (R21.3 billion) and water (R9.8 billion), and collecting rates (R17.7 billion). Together these accounted for 60% of all revenue (including the capital subsidy). Other significant sources of income were service charges for wastewater management and solid waste management (R7.3 billion and R2.9 billion), and a variety of transfers from national and, to a lesser extent, provincial government, running to R11.4 billion.
City expenditure is dominated by City Power (R24.4 billion or 30% of the total) and City Water (R17.4 billion or 21%), with other major cost centres including Group Finance (9.3%), Public Safety (8.1%) and Pikitup (5.2%).4 The City’s management of its entities accounts for 2.1% of spending, while the City Manager (2.3%) and the Transport department (3.3%) also have substantial budgets. Health and the Johannesburg Roads Agency each account for just under 2% of spending. Most Votes are much smaller than these, with half of all Votes having budgets of less than R1 billion in 2023/24.5
A different way of expressing expenditure patterns shows that the main categories of spending by the City in 2024/25 were employee-related costs (R20.7 billion) and bulk purchases of electricity (R19.2 billion) and water (R8.6 billion). Another major expense was contracted services (R6.9 billion), while the City also incurred R3.4 billion in interest charges while writing off R9.5 billion in unpaid debts.
In 2024/25, the City had 34 453 employees.6 With employee-related costs running at R19.5 billion in that year, average remuneration of the City’s employees was just over R600 000, which was 20% higher than the average for personnel employed by national and provincial government in 2024/25. It should also be noted that the figures for the number of City employees and for remuneration spending exclude “contracted services”, a category that would include some “outsourced” workers whose functions may once have been performed by in-house employees, or which might be performed in-house in other metros.
While there is little detail about the breakdown of personnel costs between higher- and lower-ranked officials (i.e. better- and worse-paid employees), budget documentation records that the City Manager earned just under R3.4 million in 2023/24 but was not the highest paid employee as the chief executive officer of Pikitup earned R3.6 million, including bonuses. The head of Johannesburg Property Company also earned R3.5 million.7
Councillors’ salary costs are recorded separately from employee costs. They ran to just over R190 million, with the executive mayor earning R1.6 million.
Per capita revenue and expenditure
To contextualise the commitment of resources to the City’s functions, it is worth thinking about revenues and expenditures per person in the City. Doing so is complicated, however, by the fact that we do not have annual official updates of population figures for the City, though StatsSA says that the Census count of the full population in 2022 was 4.8 million. Since 2015, however, StatsSA has been reporting Quarterly Labour Force Survey (QLFS) data for each of the metros. It is possible, therefore, to calculate the level of spending per working age adult (i.e. people aged between 15-65) and per employed person. As reflected in Table 1, in 2024/25 the City raised about R19 000 of revenue per adult or R40 600 for every person who was employed and who lives in the City. It should be noted that about 25% of these revenues were transfers from national government (the equitable share, conditional grants, and the fuel levy), so it is somewhat more accurate to say that the City itself raises revenue equivalent to R13 500 per adult or R30 000 per employed adult.
In relation to the City’s employment numbers of 34 453 in 2024/25, there were about 123 adult residents for every employee of the City, or 57 employed people for every City employee.
Cashflows
Discussion of revenues and expenditures reflects an accounting definition of both terms that is distinguishable from actual cash flows. Thus, unpaid rates bills are a form of revenue in an accounting sense since the revenue is genuinely accrued as “earned” to the City even if it has not yet been paid. To the extent that residents do not pay their rates and service charges, however, the City’s cash receipts will be lower than revenues. (Conversely, if the City were to receive payment for debts that accrued from unpaid revenues of previous years, cash receipts could, conceivably, be greater than revenues). Similarly, the writing-off of unpaid bills is reflected as an expenditure even though no cash actually flows out of the City’s accounts when this is recorded in the financials. If the City borrows money, on the other hand, this is an inflow of cash, but it is not reflected as revenue, but as balance sheet transaction (cash on hand—an asset—increases along with debts owed to the lender).8
The distinction between accounting incomes and cash receipts is critical when thinking about the viability of a City’s budget because it is cash inflows (not revenues) that are the binding constraint on how much the City can spend. Indeed, ex post, spending is always equal to or lower than cash inflows (not revenues) unless it starts out with money in the bank (in which case this might be spent down to zero before the constraint kicks in).
Naturally, there will be a relationship between these different accounting concepts—cash receipts and revenues. There are also times when it is a positive necessity for organisations to have very large differences between cash payments and revenues, whether to build up cash balances to be spent down later or to spend borrowed money in excess of revenues as new infrastructure is built up. But any material difference between the net cash and net income positions that is sustained over a long period must be considered a sign that its position is not commercially/fiscally sustainable. In this regard, sustained negative cashflows are of an existential challenge for municipalities: if cash outflows exceed cash inflows, they are bound to get into deepening trouble because municipalities can only spend the cash in their bank accounts (whether from customers, lenders or national/provincial government). If outflows exceed these combined inflows, they will not be able to meet all their obligations as soon as their bank balances hit zero.
According to its cashflow statement, in 2024/25, the City’s receipts totalled just over R78.2 billion, of which R34.8 billion (45%) came from service charges, R16.9 billion (22%) came from property rates. Transfers from national government accounted for 15% of its receipts, while it also borrowed R8.5 billion (11%). Payments made by the City were about 2% less than cash receipts at R76.3 billion, with 50% (R38.2 billion) accounted for through payments to suppliers and 27% (R20.6 billion) accounted for by payments to employees. The City invested about R8.3 billion (11% of payments) in new assets. Overall, there was a cash surplus of about R1.9 billion.
Operational and investment-related cashflows
A useful distinction is cash receipts and payments for operational purposes versus cash receipts and payments for investment purposes. Broadly speaking, borrowings should fund investment rather than operations, since it is investment in new assets that generates the basis for future operational cash receipts. Borrowing to fund operational deficits, while not always unjustified, is unsustainable in the long run because, by its nature, such spending (unlike investment spending) does not produce the assets needed to increase future operational cashflows and repay the loans used to fund the spending in the first place.
In this regard, the audited cashflow figures for 2024/25, show that the City was, indeed, running an operational surplus, some of which was being used to fund investment, as reflected in the table below, which shows that the City generated over R7.4 billion in cashflows from operations. In addition, it raised about R3.2 billion in net cashflow from financing activities. This permitted investments spending of R9 billion, and allowed for a cash surplus of R1.8 billion.
The mere fact that the City generated positive operational cashflows, and that, overall, net cashflows were marginally positive does not automatically mean that the cashflow position is sustainable. Operational payments, for example, include payments to suppliers. If the City withholds payment, these outflows are lower than they should be, which increases the size of the operational cash surplus, but at the expense of increased commercial debts. We return to this issue below.
Comparing different measures of financial performance
As noted above, according to the City’s income statement (or statement of financial performance), the City of Johannesburg recorded an accounting surplus of about R1.2 billion in 2024/25, a figure that rose to R1.6 billion after taking account of a tax credit. Its cashflow statement, however, reflects a surplus of about R1.9 billion (which also implies that the City added to its cash balances over the course of the year). In many ways, this looks like a budget that is reasonably balanced: not only are revenues more or less in line with expenditures, but cash receipts are more or less in line with cash payments. These sound like outcomes that a budget official would be proud of. There are, however, grounds for significant concern once one considers that about R14.2 billion of the expenditure recorded on the income statement is for depreciation (R4.9 billion) and debt impairment (R9.3 billion). These are non-cash items in that, while recorded as expenses, they do not require the City pay out any cash. On the face of it, therefore, the net balance of the income statement should imply positive operational cashflows of something like R15 billion. Indeed, even this understates the matter: in 2024/25, the debts the City owed to its suppliers (and, to a lesser extent, SARS) increased by R3.6 billion. Again, everything else being equal, this should increase operational cashflows because the City has allowed unpaid debts to increase compared to the year before. Offset these adjustments with a few adjustments in the opposite direction (e.g. because the City took in more deposits from new customers than it paid out), and one might expect a surplus of nearly R17 billion on the operational cashflow statement in 2023/24. Instead, the actual figure is R7.4 billion. The principal cause of this disjuncture is the very large write-off of debt (R9.5 billion in 2024/25), an ongoing, chronic issue in the City’s finances to which we will return.
The City's balance sheet
The City’s balance sheet as of the end of the 2024/25 financial year reflected R119.6 billion in assets compared to R60.6 billion in liabilities, with the remainder reflected as R59 billion in “community wealth”. The City is, therefore, solvent. Some care needs to be taken with this conclusion, however, since 74% of the City’s assets (over R88 billion) are in the form of property, plant and equipment of which over R50 billion is the valuation of infrastructure assets (mostly roads, water, sanitation and electricity-related infrastructure). These assets are both hard to value and extremely illiquid (in the sense that it would be very difficult to monetise the assets), which makes the valuation of the assets on the City’s balance sheet somewhat academic.
Of the R60.6 billion in liabilities, R22.8 billion were financial liabilities to lenders and unpaid bills owed to service providers (R28.5 billion), particularly Eskom. The City was also owed R14.6 billion by its customers. In and of themselves, these numbers do not tell us much and can only be interpreted after we look at underlying trends over time, to which we return below.
Assets and liabilities per capita
The City’s R119 billion in assets is the equivalent of over R28 000 per adult resident or just under R61 000 per employed resident. The equivalent numbers for liabilities are R14 250 and just under R31 000, respectively. In effect, every adult in Johannesburg owes the City’s creditors R14 250 but has R28 000 in City-owned assets to pay off that debt (albeit that it is extremely unlikely that value could be realised if it had to be).
Section 05
The budget over time
A budget reflects a range of political, economic and statutory considerations. Looking at a budget for a single year provides valuable insights into the priorities and structure of activities of the institution. But those insights can be greatly enriched by providing historical context: how does this budget relate to the budgets of previous years? What has changed? What has not? Is the current budget shaped by the stresses and pressures arising from decisions made in the past? Does the budget mitigate those stresses and pressures, or does it reinforce and deepen them?
This section looks at spending trends and seeks to identify key issues that have shaped more recent decisions. It also asks whether the budget is becoming more or less sustainable. To do this, we start by charting the key trends in revenue, spending, cashflow, assets and liabilities. For the most part, we do this in nominal terms (i.e. we do not adjust for inflation), but each section includes an analysis of these trends in per capita and inflation-adjusted terms. In this regard, SARB figures show that consumer prices in urban areas rose by an annual average of almost 4.8% between 2015 and 2025, so any annual average rate of increase reported here that is higher than 4.8% is rising faster than inflation, while any increase below that level means it is rising less quickly than inflation. Similarly, population growth—as measured by the rate of growth of the population of working-age adults provided in StatsSA’s QLFS data for the metro—has averaged just under 2% a year since 2014. Putting these two facts together means that any nominal growth rate of less than 6.8% a year means the value is falling in inflation-adjusted per capita terms; while any increase in excess of 6.8% a year means that the value is rising in inflation-adjusted, per capita terms.
It is important to say that we have no reliable figures for economic growth in Johannesburg, which would be extremely valuable in this context, since a key metric of a City’s financial sustainability is the extent to which its finances represent an increasing burden on residents. This is not merely a matter of the absolute cost of the services it provides, but also of the extent to which these costs are rising more or less quickly than the capacity to pay for them. This may be rising as infrastructure improvements support productivity growth, or it may be falling if decaying infrastructure is undermining productivity growth. In the absence of credible figures for local economic activity, this cannot be fully assessed. We do, however, measure changing revenues and spending against changes in employment levels in the City to demonstrate that there are legitimate concerns about the sustainability of the existing path.
Spending: 2015-2025
As reflected below, between 2014/15 and 2024/25, revenue rose by an annual average of 6.8% a year, rising from R41.8 billion in 2014/15 to around R80.3 billion in 2024/25. Expenditure growth was somewhat faster, at 7.7% per year, rising from R38 billion in 2014/15 to just under R80 billion in 2024/25. Over that period, the City budgeted for a surplus averaging around R3 billion a year, though this number has varied quite widely, and has tended to fall over time, particularly after 2018/19. The City recorded an accounting deficit in 2022/23.
As noted in the preceding section, revenues and expenditure in the income statement are not always matched by actual receipts and payments that are reflected in the cashflow statement. Importantly, between 2014/15 and 2024/25, both income and cash receipts grew at about the same rate (6.8% a year), but expenditure (as reflected on the income statement) grew much faster than cash payments (7.6% a year versus 6.5% a year). As we will see below, much of this was driven by the rapid increase in non-cash expenditure items, notably debt impairment figures, which rose by over 9% a year for the period under review. Another notable result from this analysis is that, if we exclude both new borrowings (from cash receipts) and the repayment of debt (from cash payments), the difference between the rate of growth of cash payments and cash receipts widens, reflecting an increasing reliance on borrowings to fund City activity. Again, we return to this below.
The composition of revenue and expenditure
Between 2014/15 and 2024/25 aggregate revenue has grown at 7.2% per year (excluding capital transfers), but there have been quite significant differences in the rate of growth of different components of the City’s revenue. The fastest growth is for wastewater charges (10.3% a year), solid waste management (9.1% per year) and water (8.8% per year). Revenue from property rates grew at 8.7% a year, while electricity (which remains the single largest source of revenue) has grown at the slowest pace (5.2% per year). These divergent growth rates mean that the composition of the City’s revenues has changed significantly: in 2014/5, electricity sales accounted for 34% of the City’s revenues; by 2024/25, that had fallen to 28%, while the contribution of property rates rose from 20% to 23%.
What is going on with electricity revenues?
One of the most important trends to note from the above is the decline in the importance of the sale of electricity as a source of revenue (and spending) for the City. This is driven by a decline in electricity demand, not a decline in its price.
Using data from City Power’s annual reports, we can see that the number of units of electricity purchased by the City each year has fallen by an annual average of 2% from 12.4 trillion kWh to 10 trillion kWh, while the number of units sold has fallen at a similar rate from 8.7 trillion kWh to 7.1 trillion kWh. In fact, this understates the rate of decline in electricity sales: taken from the highpoint of electricity sales in 2015/16 (9.7 trillion kWh), electricity sales have fallen by 4% a year. Throughout the period, between 20% and 30% of electricity was “lost” (i.e. it was purchased by the City but was not sold to customers), with about one-third of that amount being attributed to technical losses and two-thirds to non-technical factors (i.e. illegal connections).
The decline in the purchase and sale of electricity is a result of a range of factors: rising levels of loadshedding (which, by definition, means less electricity is both purchased and sold by the City), rising prices and slow economic growth (which reduce the underlying demand for electricity), and consumers’ defection from the grid and adoption of self-generation solutions. Some of these trends can be elaborated using data from City Power, which show that, while the number of City Power customers fell by around 8% between 2014/15 (414 000 customers) and 2023/24 (383 000).9 This may be a reflection of economic stagnation, but also suggests falling levels of service delivery. It is highly unlikely that the combined number of households and businesses fell as much as 8% over that period, since that would imply significant outflows of people from Johannesburg, when in fact the opposite appears to have happened. In any event, StatsSA estimates the population of adults in Johannesburg is growing at about 2% a year. It is possible, however, that some fraction of new households is not being connected to the grid, especially if household formation disproportionately occurs in informal settlements or as backyard rentals. Another partial answer to the puzzle of this data is a second important trend: the transition of residential customers (who are billed retrospectively) to prepaid customers. It is also possible that an increasing proportion of the city’s residents live in areas serviced by Eskom directly, rather than by City Power. This does not seem likely, though, given that the City has been taking over Eskom customers’ accounts (albeit that process is not evident in any increase in the number of City Power customers).
The City’s push to transition its customers to prepaid solutions is a response to rising numbers of unpaid electricity bills. While this was always a problem, it is clear that Covid-19 had an enormously detrimental impact on the ability and/or willingness of City Power’s customers to pay their bills: before 2020/21, the City recorded about R4 billion in unpaid electricity bills, more than a third of which was current debt (i.e. it was less than 30 days old); by 2024/25, unpaid electricity bills amounted to over R11.4 billion, of which more than half had been outstanding for more than a year. In fact, electricity debt that had been unpaid for more than a year increased more than 30-fold since 2018, rising from about R200 million to over R6 billion.
Importantly, outstanding electricity debt (the growth of which is moderated by the fact that so many customers are on prepaid meters) is only a portion of total debt owed to the City by residents, businesses and government institutions. We will return below to the rising level of debt owed to the City by its residents.
Is spending growth exceeding the rate of growth of the city’s economy?
While StatsSA doesn’t provide estimates of the rate of growth of Johannesburg’s economy, in a report commissioned by CDE as part of the Johannesburg Matters project, PWC, using data from a number of independent providers, estimated that Johannesburg’s economy grew at a fraction over 1% a year in real terms between 2014 and 2024, which would be the equivalent of 6% a year in nominal terms. Over the same period, the City’s revenues grew at a rate of 6.7% in nominal terms, which implies that, in inflation-adjusted terms, the City’s revenues were growing almost twice as fast as the local economy (1.7% a year v 1%). Compared to spending growth, the gap is even greater: between 2014/15 and 2023/24, spending grew at 7.7% a year, or 2.7% a year in inflation-adjusted terms. That is almost three times higher than the 1% a year in real growth estimated for the local economy.
Another way we can look at this question is by comparing the trajectory of the City’s revenues to the trajectory of employment in the City, which is a proxy for the rate of economic growth in Johannesburg. In this regard, StatsSA’s QLFS data for metros reflects an economy that has stagnated since (at least) 2015, with employment growth running at 0.3% a year, far less than the 2% a year growth in the working-age population in the city.
The fact that the City’s revenues grow at 1.7% a year and its spending grows at 2.7% a year in real terms in a period in which employment grew at only 0.3% a year is bound to be unsustainable unless income growth for those who are employed and who pay rates and buy services is growing quite strongly.10
If this is not the case, the sustainability of the budget will depend heavily on whether the City is able to secure an increasing share of nationally raised taxes. However, the sustainability of this latter option depends on the rate of economic growth of the economy as a whole. Given that the economy has grown substantially slower than 2% a year, however, it is mathematically impossible for subsidies to Johannesburg to increase indefinitely to make up the gap between the rates and service charges that the City can raise locally and its increasing spending needs.
Consider, therefore, trends in inflation-adjusted revenues of the City in comparison to the rate of growth of the adult population. In inflation-adjusted terms, annual revenues per working-age adult have been basically stagnant at around R19 000 (in 2024 rands) since 2014/15. Because employment growth has been so slow, however, in inflation-adjusted terms, revenues per person employed have risen by an average of 1.5% a year, from R35 000 in 2014/15 to nearly R41 000 in 2024/25 (in 2024 rands).11
The fact that inflation-adjusted revenues per employed person in Johannesburg have risen at 1.5% a year is rendered more significant when one considers that StatsSA’s Quarterly Employment Statistics (QES) data shows that average real remuneration in formal sector employment has risen by just 0.6% a year over the same period. If real wages in Johannesburg have risen at the same rate as those of the country as a whole—QES data is not available at sub-national level—then it implies that the City’s revenues were taking an increased share of households’ income (measured in real terms) in 2023/24 than they were in 2014/15. Nor is this share insignificant: at an average of R40 000 per employed person per year, Johannesburg’s revenues count for over 10% of the average worker’s annual remuneration, which the QES estimates at just under R350 000 per year in 2023/24. It is perhaps worth noting that most of the City’s revenues are services—electricity, water, sanitation, waste management—so it would not be accurate to characterise these payments as taxes. It should also be pointed out that these averages conceal a great deal of variability given the vast differences in property values and services utilisation.
If revenues raised by the City (i.e. the invoices it sends out for services and rates payments) rise more quickly than the number of people employed and the rate increase in real wages, then the burden placed on residents must be rising, albeit that much of the responsibility for this lies with faster-than-inflation growth in the cost of water and electricity, over which the City has limited control, apart from ensuring that losses are minimised. The rise in the financial burden on consumers, however, is likely to be a key reason debt owed to the City has ballooned, as we will see below.
The drivers of spending growth
Overall, spending grew at a rate of 7.7% a year between 2014/15 and 2024/25, with total spending rising from just under R38 billion to nearly R80 billion. That growth has been driven by faster-than-average expenditure growth in employee-related costs (9.2% a year) and contracted services (10.3% a year). Spending on bulk purchases of water and electricity, which accounts for a large fraction of total spending, grew at only 3.9% a year. The spending category “other” is dominated by the City’s depreciation allowance, and it grew by 7.5% a year. Critically, the charge for debt impairment has risen at over 9% a year.
The effect of these differential rates of growth means that there has been a noticeable change in the profile of spending in the City, with a larger share allocated to contracted services. This accounted for 7% of spending in 2014/15, but rose to 10% in 2023/24 before falling back to 9% in 2024/25. Employee costs have also increased as a share of spending: from 23% to 26% of the total. The share of total spending accounted for by electricity purchases, by contrast, fell from 33% to 23% over the period under review.
The City is owed a lot of money, and it's been rising fast
According to the City’s audited balance sheet, it was owed over R12 billion by its customers at the end of 2024/25. That figure is about 150% larger in 2014/15, which means it rose by an average annual rate of 9.3% over that period, nearly double the rate of inflation. Nevertheless, it actually dramatically understates the reality of the growth of non-recovery of rates and service charges. This is because the figure reflected in the balance sheet as an asset is net of the City’s provisions against unrecoverable debt. If those provisions are included, the rise in unpaid bills is much more dramatic: between 2014/15 and 2024/25, gross debt to the City rose by nearly 17% a year (i.e. 12 percentage points faster than inflation) from R15.3 billion to over R71 billion. Over the same period, aggregate provisions against unrecoverable debt rose by over 19% a year, from just over R10 billion to nearly R60 billion. Thus, the doubling in net debt owed to the City as reflected on its balance sheet is the residual after a much, much faster rise in gross debt (which has increased almost five-fold) and the rise in provisions against unrecoverable debt (which has increased almost six-fold).
The City’s financial statements provide a great deal of detail about the composition of this debt. They show that household debt is the largest component of gross debt and has also grown the fastest—by over 18.1% a year from R8.4 billion to R44.6 billion. The smallest component of the debt belongs to national and provincial government entities, but it, too, has grown at 17% a year, rising from R400 million to R2 billion. The remainder—debt owed by businesses—has grown at 14.5% a year from R6.5 billion to over R25.2 billion. Provisions against unrecoverable debt (not shown in the graph) have grown fastest for household debt (at over 20% a year) and slowest for debt owed by government agencies (under 9% a year).
It hardly needs to be said that the rate at which unpaid bills are accumulating is wholly unsustainable, and it represents a critical challenge for the City’s administration. Indeed, scaled against revenues generated from the sale of services and property taxes, by 2024/25, the annual increase in gross debt amounted to the equivalent of 15% of the City’s billings to residents/customers. Effectively, the City does not expect to receive any income from about 15% of the revenue. Indeed, if we add to this the increase in the annual increase in unpaid bills that has not been written off, the figure rises to 17% (depending on how one measures the City’s income).
Importantly, the transition to prepaid metering in electricity does not appear to have made a significant difference to the rate of growth of outstanding payments: unpaid electricity bills have grown at the rate of 16.7% a year, which is slower than the rate of growth of unpaid water bills (20.2% a year), but faster than the rate of growth of unpaid property taxes (12.1% a year).
The City’s liabilities are significant, but are falling in real terms
At the end of 2024/25, the City had nearly R60.6 billion in liabilities. Of these, 50% (R30 billion) were financial liabilities (i.e. borrowings) and another R28.5 billion (47%) were trade payables (i.e. debts owed to providers for goods and services), much of which was owed to Eskom and Rand Water.12
The rest of the City’s liabilities was accounted for by provisions (largely for funds received for conditional grants but not yet spent).
Overall, at 4.7% a year, the rate of growth of liabilities has been slower than the rate of growth of the City’s income and expenditure, which suggests that debt accumulation is not in itself unsustainable. Within the portfolio of liabilities, however, the rise in trade payables—i.e. debts owed to suppliers—has risen fast, at 9.1% a year since 2014/15. Importantly, much of that growth was driven by an exceptionally large rise in trade payables in 2023/24 (R25.4 billion) relative to the previous year (R18.2 billion), which was itself substantially larger than the figure in 2021/22 (R14.5 billion). Notably, the City’s previous budget book, adopted in 2025, anticipated that trade payables would decline by 30% between 2023/24 and 2024/25. The fact that the audited figures for 2024/25 reflect a further increase in trade payables demonstrates how severe the challenge is, and that the City (and all its stakeholders) need to keep a very close eye on this dynamic. We return to issues related to the latest budget below.
Have Johannesburg’s lenders started to turn away?
Between 2014/15 and 2024/25, receipts and payments have grown at similar rates, rising from about R40 billion to about R76 billion. Over that time, the City has run a net cash deficit of R1.6 billion, having received R603 billion and spent R605 billion. If we exclude both borrowings and debt repayments, the net cash deficit over the period under review increases to R9.5 billion, which is one reason why the City’s liabilities have increased over that period.
Disaggregating the City’s net cashflows between operational, investment and financing make the City’s challenge clearer: it shows that in the late 2010s, the City generally borrowed substantially more than it repaid outstanding debts. This changed in the early 2020s; the City borrowed about the same amount as it repaid in debt each year. By2022/23, however, the City was not able to borrow more than it repaid: in that year and the next, the City repaid R4 billion more than it was able to borrow. That was reversed in 2024/25, when the City raised two new loans, one from DBSA for R2.5 billion and one from the French development agency, AFD, for a similar amount.
In many—perhaps even most—circumstances, repaying more debt than one borrows is a positive, as it did in 2022/23 and 2023/24. But this is only true if the effect is that one’s balance sheet improves. In the case of the City, however, this did not happen. Instead of borrowing money from lenders, the City appears to have passed the problem onto its suppliers (principally Eskom), with debts to suppliers rising over the period in which the City was paying back debt to lenders. Indeed, it is pretty clear from a comparison of the income statement and the cashflow statement that the City would have borrowed more money if it could have.
In the chart below, we show that in the period under review, the gap between accounting revenues and cash receipts has generally grown larger over time, with revenues almost always exceeding actual cash receipts. Thus, in 2023/24, the City’s accounting revenues were R9.7 billion greater than its actual cash receipts, though this fell to R3 billion in 2024/25. This suggests either that budgeting processes have become increasingly disconnected from reality, or that a structural change is underway, in that the City’s residents may be less able or willing to pay rates and service charges or that the City’s lenders are no longer willing to close the gap between what the City expects to receive through rates and service charges. This is also reflected in the widening gap between the City’s expenditure and its payments, which reached R7.1 billion in 2023/24. It is hard not to conclude that, therefore, had it been able to do so, the City would have borrowed more money than it did to bridge the gap between revenues and receipts, and raise payments to something closer to expenditure.
Finally, there is another way in which we can show that something unusual is happening in the City’s budget in recent years. This is by comparing the numbers presented for the 2024/25 budget in May 2025 (which was near the end of the financial year, when an adjustment budget was tabled by the City), and the figures reported in May 2026 for the financial year 2024/25.
When the 2024/25 adjustment budget was presented in May 2025 (just two months before the end of the financial year), the City reported that it would raise R77.6 billion in revenue that year, and would spend R75.4 billion, leaving a surplus of R2.2 billion. In the audited financials reported a year later, however, revenue for the year was R400 million higher but spending was R4.4 billion higher, with the result that a projected surplus of R2.2 billion had become a deficit of R1.9billion. Among the reasons for these changes: a R5.5 billion increase in expenditure on bulk purchases and a R3 billion decline in services-related revenues. In the cashflow statements, the audited figures (tabled in 2026) for both receipts and payments were significantly lower than the figures presented in the adjustment budget presented in 2025 (see Table 2). By far the biggest adjustments in the cashflow statement between the presentation of the budget in 2025 and the pre-audit figures presented in 2026 was in relation to operational payments, which were reported to be almost R8 billion lower in the pre-audit outcomes than had been expected in the adjusted budget, with payments to suppliers falling by over R8.5 billion. The figure for receipts from operations was also considerably lower in the pre-audit outcome than in the adjusted budget.
The City’s employee-related costs are growing faster than its revenues.Between 2014/15 and 2023/24, employee-related costs rose more quickly than revenues, particularly cashflows generated from customers (i.e. rates and service charges), with employee related costs rising by 9.2% a year on average, and customer-generated cashflows rising at 7% a year. Over that period, employee-related costs rose from 33% of the value of cashflows generated by customers to 43% in 2023/24 before falling back to 40% in 2024/25. Clearly, a rise of this kind cannot continue indefinitely without very serious implications for the sustainability of service delivery. Indeed, the fact that employee costs have risen so quickly is likely to have been an important driver of declining service quality in the City since there is less and less funding available for non-personnel spending.
It is important to note three important qualifications to the analysis presented here:
- Changes in employee costs are largely determined by decisions taken through collective bargaining procedures, with South African Local Government Association representing employers. This means that the City does not have full control over factors meaningfully impacting on its cost-of-employment. That said, while it cannot affect employment costs associated with particular staffing grades affecting most of its employees, it can impact the profile of staffing grades (albeit this is also partly determined through collective bargaining, and which is a key factor in the dispute over the decision taken in 2025 that would see an increase in total employment costs—see box below).
- The City can also determine how many people to employ directly, and whether or not it should outsource services. In this regard, it might be worth noting that spending on “contracted services” has grown 10.3% a year between 2014/15 and 2024/25, from R2.7 billion to R6.9 billion, having hit a high of R7.2 billion in 2023/24. It is not possible to know exactly what proportion of this spending goes to wages (as opposed to other input costs) or how many people work essentially full-time for the City through these arrangements. It is apparent, however, that if we simply aggregate employee costs and the costs of contracted services, the rate of increase in spending between 2014/15 and 2024/25 is actually higher than the 9.2% a year measured for employee-related costs in the City itself. It should be noted that National Treasury guidance to the effect that a municipality’s budget should allocate no more than 40% of expenditure to employee-related costs may be a source of some perverse incentives, in this regard.
- One way in which the City can control employee costs is by capping the rate of growth of remuneration for its most senior employees, something that is not subject to collective bargaining. In this regard, however, there does not seem to be any consistency in the way that the City reports executive remuneration, so it is not possible to assess whether that has grown more or less quickly than employee costs as a whole.
Employee numbers
Between 2016/17 and 2024/25, total employee costs rose by 9.2% per year on average, from just under R10 billion to R20.7 billion. This increase is a consequence of two drivers: the number of City employees rose by an annual average of about 4.2% from 24 620 to 34 453, while average remuneration rose at just under 6% a year from R400 000 to R595 000 (see Table 4). (The reader is reminded, however, that there are clear problems with the reporting of personnel numbers in City budget documentation—as noted earlier—so the analysis presented here must be treated with caution.
In effect, in other words, the City’s salary bill has risen in inflation-adjusted terms by a little more than 5% a year, but almost all of that is accounted for by the increase in staff numbers, with average salaries rising at less than 1 percentage point faster than inflation over the period as a whole. Critically, however, and as is apparent in Table 4, in 2020/21, the City’s personnel establishment expanded dramatically (by 27%) and average remuneration fell by 16%. This may reflect the policy decision to in-source previously outsourced services, which would have brought a large number of relatively low-wage employees into the City. If we look at wage increases since then, the average remuneration has risen by more than 10% a year.
Over the period 2016/17-2023/24, employment levels in the City have risen faster than the City’s population (4.2% a year compared to 1.8% a year), both of which have risen more quickly than the number of people in Johannesburg who have jobs (1.3% a year). It is also notable that the City’s employment expanded during Covid, precisely when overall employment in Johannesburg fell. Presumably, this was in response to the specific challenges being confronted in that period. By 2024, however, employment levels in the City had fallen significantly relative to their peak. This fall was driven by a decline in permanent employees, whose numbers fell by nearly 20% between 2023 and 2024, from over 35 000 to close to 29 000. It is possible, however, that this reflects problems in the data, rather than real changes in the City’s personnel numbers.
A note on the salary agreement with SAMWU
The DA has taken the City of Johannesburg to court to prevent the implementation of an agreement with SAMWU that it says will cost the City R10.3 billion over three years, with significant payments due in 2026 and 2027. This issue was also raised by the Minister of Finance in the letter he wrote to the City, and which was leaked by Helen Zille, where he noted,
[Y]ou have committed the City into a financial obligation that is not possible to fulfil. Of concern is that this decision is a direct transgression of MFMA and budget and reporting regulations. You are requested to indicate what measures will be implemented and how this violation will be arrested or reversed. You are hereby directed to stop proceeding with the implementation of this illegally signed agreement that has the potential to destroy the sustainability of the City of Johannesburg beyond this term of Office as well as the negative impact on the national economy at large. Secondly, not only is your Adjustments Budget not funded in terms of Section 18 of the MFMA, but you very well know the City cannot afford this agreement.
There is very little public information about the nature of the agreement, the calculation of the R10.3 billion, and the carry-through costs on the City’s budget. What appears to be at stake is the following:
- A collective agreement between SALGA and SAMWU in 2015 established a process for determining whether posts in municipal personnel establishments had been graded appropriately, and agreeing that employees who had been employed in posts that were graded below the actual skill content of the job, would be regraded and their salaries would be retrospectively increased.
- The implementation of this process has been contested, and SAMWU has disputed the extent to which the City has corrected its members’ job grades.
- In the run up to the G20 in 2024, the City reached an agreement with SAMWU to fully implement the wage curve adjustments to avoid a strike that would impact on the hosting of the G20.
There is no publicly available data on how the figure of R10.3 billion in additional costs was calculated, but the DA has said that the agreement stipulates a payment of R1.2 billion to R2 billion by March 2026, a further R5 billion or R6 billion in July 2026 (which would be in the next financial year), and a further tranche in 2027. To the extent that this can be confirmed with publicly available documents, it can be confirmed that the City has proposed an adjustment budget that would increase the budget for employee costs in 2025/26 (i.e. for the period including March 2026) by R1.35 billion, with the increase funded by cuts elsewhere in the budget, notably a cut of more than R2 billion to spending on repairs and maintenance.
Even assuming that the full R10.3 billion is paid over the next 18 months, it is not possible to estimate what the carry-through implications would be for employee costs in the City because a large (but unknown) fraction of the R10.3 billion will be for backpay. It is worth noting that R10.3 billion is about 50% of the current budget for employee costs, so the additional funding needed to implement the agreement with SAMWU is a remarkably large fraction of annual employee costs.
The infrastructure-spending squeeze
One of the key trends over the past decade is that, measured in inflation-adjusted terms, spending on infrastructure, and, in particular, on investment in new infrastructure has fallen dramatically. That is illustrated in the charts below, which show (on the left), that investment spending (as reported in the cashflow statement) has fallen from around R10 billion in 2014/15 and 2015/16, to between R6 billion and R8 billion in subsequent years. If we adjust for inflation, however, the fall is significantly greater, with investment spending in 2014/15, when measured in 2024 rands, the equivalent of nearly R16 billion in 2024/25. This means that investment spending has fallen by around 50% in real (i.e. inflation-adjusted) terms in that period. In per capita terms, investment spending has fallen by close to 70%.
Overall, spending on new investment exceeds the annual depreciation costs, which is as it should be in a city that requires growth. Care should be taken in thinking about this result, however, since spending on new investment may be understated if new infrastructure costs less (or provides more infrastructural services to users). Conversely, the costs of new investment may overstate the value of those investments if productivity declines and/or if corruption- or extortion-related costs rise.
The graphs also show that, over the same period, spending on repairs and maintenance has climbed somewhat faster than inflation, rising, in inflation-adjusted terms, from the equivalent of around R2 billion in 2014/15 to R3.8 billion in 2023/24, before falling back to R3.2 billion in 2024/25.
Section 06
What is in the budget for the next three years?
The first draft of this report was completed in April 2026, a few months before the City tabled a new budget for the period 2026/27 to 2028/29 in May 2026. That draft was premised on data contained in the City’s Budget Book tabled in May 2025 for the period 2025/26. Because the new budget tabled in May 2026 did not provide audited figures for 2024/25, the final draft has not updated the analysis of budget trends because of a decision taken at the outset to rely only on audited figures because of the sometimes quite significant adjustments that are made when the annual financial statements are concluded. We do, however, need to update our analysis of the City’s budget going forward. This is because of changes in the figures for 2024/25 between the adjustment budget tabled in May 2025 and the “pre-audit outcome” figures tabled in May 2026 (as noted above), because of changes to the proposed numbers for 2026/27 and 2027/28, and because the new budget now includes projections for 2028/29.
What we initially concluded about the three-year budget for 2025/26 to 2027/28?
Our initial assessment of the three-year budget for 2025/26 to 2027/28 concluded that the City was projecting that its finances would be on a considerably more sustainable path than had been the case over the preceding 10 years. This was largely because, while revenues were projected to grow at 7.5% a year for the three years (roughly in line with previous rates of growth), expenditure growth was budgeted to slow significantly to 5.1% a year. The result was a projection of rapid increases in the City’s surplus, which would rise to R13 billion in 2027/28. We believed, however, that the assumptions underpinning these projections were dubious.
The May 2025 budget also projected that net cashflows would turn very strongly positive in the following three years. In particular, the budget projected that cash receipts would grow at 6.2% a year over the next three years (roughly in line with previous rates of growth), but that cash payments would flatline in 2025/26 and 2026/27, before increasing modestly in 2027/28. The rate of growth of cash payments projected in May 2025 was far, far lower than had been the case in previous years and relied on holding payments to suppliers essentially flat.
These projections did not seem plausible when we reviewed the May 2025 budget numbers, and a review of the numbers tabled a year later, in May 2026, confirms that was the case for 2025/26. At the same time, however, the May 2026 budget continues to rely on implausible assumptions about how income and expenditure will evolve over the next three years, albeit that the degree of implausibility is much reduced.
What the May 2026 budget projects for the future
In the May 2025 version of the budget, the City projected that accounting surpluses would rise from R700 million in 2023/24 to R5.9 billion in 2024/25, R8.2 billion in 2025/26 and then over R12 billion and R13 billion, respectively, in the next two years. The May 2026 version of the budget demonstrates how unrealistic those projections were: for 2024/25, the budget surplus is estimated in pre-audit outcome figures at R1.2 billion, about R4.7 billion lower than originally projected. The surplus for 2025/26 is now projected at R7.7 billion (just R500 million less than the previous year), while the projections for 2026/27 and 2027/8 are far lower than previously projected. Given the revisions between May 2025 and May 2026, however, it is hard to know how much credence to attach to these projections.
Similar points can be made with respect to cashflow projections: the budget presented in May 2025, projected very large cash surpluses, which would rise from zero in 2024/25 to R11.4 billion in 2027/28. The May 2026 budget is far less optimistic, but its projections, were they to materialise, would reflect an unprecedented level of financial performance.
As was the case in the May 2025 budget, the May 2026 budget projects a very sharp slowdown in the rate of increase in cash payments that the City will make between 2024/25 and 2028/29, albeit that the slowdown is not quite as dramatic as initially projected and starts a year later (in 2026/27 rather than 2025/26). Again, it is far from clear how this is to be achieved. It is true that cash payments projected for 2024/25 in May 2026 are lower than those projected in May 2025, but this seems to be a result of the City’s failing to pay Eskom what it was owed. This is not a sustainable strategy, however, and it is one that passes the burden of financial difficulties from the City to Eskom.
What has happened to spending on maintenance and repairs?
In the course of reviewing the May 2026 numbers, a potentially important anomaly was identified in the repairs and maintenance data that warrants scrutiny by policymakers and, possibly, the Auditor-General. It needs to be stressed that these figures are for 2024/25 and are described as “pre-audit outcome” figures, which means that the issue may be explained by an accounting error. The issue is evident in the graph: spending on repairs for “machinery and equipment” in 2025/26 (i.e. the current financial year) is vastly out of line with historical precedents and future budgets, and this is at the expense of spending on the repair/maintenance of almost everything else, especially roads and electrical infrastructure.
As noted, this might simply reflect an accounting error that will be corrected when the budget is audited. On the other hand, given the prevalence of corrupted tenders, such anomalous behaviour should be carefully reviewed.