Energy From Eskom - Charging More for Less

Remarkably, In the same week that Statistics South Africa confirmed the economy contracted 0.2% in the second quarter of 2026, the National Energy Regulator of South Africa gazetted Eskom’s proposed 8.83% retail tariff structure for 2027/28. Direct Eskom customers would pay from 1 April 2027; municipal customers from 1 July. The average unit price would move from 240.28c to 261.50c per kilowatt-hour. Public comment closes on 2 October. A hearing follows on 8 October. A final decision is due on 26 November.

The percentage is not a new ask dreamed up this month. It is year three of the sixth Multi-Year Price Determination, inflated by NERSA’s own earlier miscalculation of Eskom’s generation regulatory asset base. After a High Court-ordered redetermination, the regulator granted Eskom an extra R54.7 billion. Of that, R23 billion is to be recovered in 2027/28, lifting the originally approved 6.19% increase to 8.83%. The legal process is tidy. The economic context is not.

A Tariff Decision against a shrinking industrial base

Q2’s contraction ended six quarters of growth. Mining fell 3.0%, manufacturing 1.8% and trade 1.9%. Those are not abstract sector codes. They are the loads that once underwrote Eskom’s coal fleet: smelters, mills, foundries, mines and the logistics around them. Household spending still eked out 0.4%. Investment did not. July inflation was 4.3% against a 3% Reserve Bank target. An 8.83% electricity increase is therefore roughly double current CPI, levied on an economy that has just gone backwards.

Eskom’s own books show why this is more than another administered-price complaint. In the year to March 2026 it sold 178 terawatt-hours — 6.2% less than the year before, and the lowest annual volume since 2000. A decade earlier it sold about 214 TWh. Industrial sales collapsed 22.5% in a single year, a loss of 9.7 TWh, driven largely by ferrochrome hardship. Mozal entered care and maintenance in March. Residual grid demand through 2026 has run 8 to 10% below the prior year. Customers who can leave are leaving: Eskom put rooftop solar at 7.7 GW by year-end; the transmission company put it near 9.1 GW by June. For the first time in more than a decade Eskom admits to 2–3 GW of surplus. That is not planned headroom. It is what remains when a utility prices itself out of the market it was built to serve.

Over ten years the arithmetic is brutal. Sales volumes down about 17%. Electricity revenue up about 118%. The average tariff from roughly 76c/kWh to about 242c/kWh. Last year revenue still rose 4.1% because the standard tariff increased 12.74% even as volume fell. This is the textbook death spiral: fixed costs loaded onto a shrinking captive base; the next increase making the next exit more rational; industry closing, cutting shifts, or building its own plant.

Why the Price–income Gap Keeps Widening

South African real incomes have not tracked electricity prices for years. The gap is not a mystery of global commodity markets. It is a domestic cost structure that regulation has repeatedly socialised onto the invoice.

Start with labour, the line item management can control and has chosen not to shrink. Gross employee benefit expense reached R50.4 billion in FY2026, from R45.4 billion the year before and about R35 billion two years earlier. Headcount at year-end was 43,274 and still rising. Eskom’s own disclosure puts average total remuneration near R942,000; calculations on the full benefit pool land closer to R1.16 million. Solidarity’s bargaining-unit average is nearer R38,150 a month, which means middle management and professional grades — not only Exco — pull the average up. Incentives have been restored at scale: a R5.1 billion short-term incentive obligation, R1.6 billion in production bonuses, about R3.5 billion in overtime. Exco cost R85.1 million. The chief executive was paid R12.2 million; the finance chief R13.5 million. NUM and Solidarity accepted a further 7% a year for three years from July 2026. Skilled operators must be paid. A growing workforce, multi-year above-inflation settlements and large incentive pools, while selling the least electricity in a quarter-century, are a political settlement with labour, not a productivity settlement with customers.

Then there is the taxpayer. The Debt Relief Act packaged R254 billion. Eskom received R64 billion in FY2025 support and R80 billion in March 2026, converted or being converted to equity. External debt excluding the shareholder loan remains R356 billion. Cumulative state support since 2008 runs to hundreds of billions. The minister declared an end to bailouts in October 2025; the R80 billion still arrived six months later because it was already legislated. FY2026 profit of R30.3 billion after tax is real on the income statement. It is also a profit taken after a debt holiday, a 12.74% tariff and a collapse in diesel burn — OCGT spend fell from R17.7 billion to R7.1 billion — not after a productivity revolution.

Waste has not vanished with national generation stability. Municipal arrears sit near R120 billion. Non-technical losses were estimated at 13.1 TWh. Local networks still fail when the national system does not. Load reduction persists in parts of Gauteng. The energy availability factor improved to 65.16% for FY2026 and has since run closer to 68% — better than the crisis years, still short of a healthy thermal utility. Winter 2026 was loadshedding-free. That matters. It does not entitle the utility to treat a R50 billion wage bill and a shrinking sales book as someone else’s problem.

A regulator structurally deaf to demand destruction

NERSA’s mandate, applied through the Multi-Year Price Determination, is a cost-plus machine. Once a regulatory asset base and an allowable revenue are accepted, the tariff becomes the plug that fills the gap. The ERTSA process now under consultation does not even reopen the quantum. The regulator will only test whether Eskom’s proposed schedule of fixed charges and energy rates collects revenue already approved. That is lawful. It is also economically tone-deaf.

The consultation will not ask whether Eskom should still recover depreciation on plant that demand no longer needs. It will not ask whether a R50 billion wage bill is consistent with 178 TWh of sales. It will not treat negotiated pricing for smelters — concessionary deals around 62c/kWh for ferrochrome — as an admission that the posted tariff is already above the price that keeps industry alive. It will not treat rooftop solar and wheeling as a verdict on Eskom’s offer. Eskom’s response to lost sales is more fixed charges, network charges on solar users, wheeling fees, data-centre hunting and even a Bitcoin-mining pilot. That is how a utility behaves when it has priced itself out of the market and still has a political mandate to remain whole.

What a Purposeful Rethink Would Look Like

A serious reset would stop asking the remaining captive customer to fund yesterday’s costs. First, separate the efficient grid from the inefficient generation stack. Transmission and distribution should be a regulated wires business. Generation should compete. Customers who can leave will leave anyway; policy should stop pretending otherwise.

Second, stop socialising the wage and incentive settlement through the kilowatt-hour. If labour stability is a national objective, put it on the fiscus explicitly. Do not hide it in the tariff. Third, price industrial power to keep load, not to punish remaining load. The smelter concessions already concede the point. Fourth, treat surplus capacity as a signal to close or convert plant, not as a reason to raise prices on a smaller base. Fifth, treat municipal debt as part of the same failure. R120 billion unpaid is why honest paying customers are overcharged.

The Bill the country can no Longer Pay

South Africa has just posted a 0.2% quarterly contraction. Mining and manufacturing — the loads that made a large coal system rational — are going backwards. Eskom is selling less power than at any time this century, at prices that have more than tripled in a decade, with a workforce that is still growing and a payroll that has swollen by tens of billions. Generation reliability has improved. The business model has not.

An economy shrinking at 0.2% can no longer afford the opulence and arrogance of a power entity that keeps charging its budget and keeps delivering no change in current. The 8.83% gazetted this week does not save Eskom. It buys another year of the same model: recover yesterday’s costs from whoever cannot yet leave. Industry has already voted with its switchgear. The regulator is still counting the invoice.