Electricity and logistics constraints on South African growth

Electricity nameplate capacity describes what equipment is designed to produce under specified conditions. Available output depends on maintenance, breakdowns, fuel, weather, and network access. Electricity that cannot reach a load center does not provide the same productive service as a reliable supply at the factory connection.

For businesses, interruption frequency, duration, and predictability all matter. A short outage can damage a continuous production process or require lengthy restart procedures. Scheduled interruptions may allow some adjustment but still impose costs. Two firms receiving the same annual energy volume can therefore face different production losses.

Backup generation and storage alter the transmission. A firm with an effective backup system can maintain production while paying higher costs and tying up capital. Observing stable output at that firm does not establish that unreliable public supply has no economic effect. The burden may appear in operating margins, investment composition, or prices instead.

Small firms may be less able to finance backup equipment or relocate operations. Aggregate industry output can consequently conceal a shift toward businesses with more resilient balance sheets. An infrastructure assessment should examine business entry, closures, and the distribution of adaptation costs as well as the output of surviving establishments.

The SARB April 2026 review, Box 5, decomposes potential growth using capital, labor, and total factor productivity. It identifies weak investment and infrastructure inadequacies among the factors constraining productive performance, while noting that eased electricity and logistics constraints likely supported a recent productivity improvement. This is a model-based diagnosis, not a directly observed infrastructure multiplier.

Potential output is estimated rather than measured in the same way as an electricity meter reading. Different models and revisions can produce different levels. The useful implication is that persistent constraints can reduce the output attainable without generating inflation, not that the published estimate precisely identifies the loss at every factory.

Temporary improvements and durable capacity should also remain separate. A period of fewer interruptions can permit higher utilization of existing equipment. Sustained reliability can then change investment decisions and encourage new capacity. The second effect takes longer and depends on confidence that service improvements will persist.

Demand can limit both effects. A firm with weak sales may not increase output after electricity improves, yet it can still save backup costs or reduce delivery risk. Measuring only production volume would miss that benefit. Conversely, a rise in output after service improves is not entirely attributable to the network if foreign demand strengthens at the same time.

Transnet's 2025 financial statements, directors' report, record rail volumes of 160.1 million tonnes for the year ended March 2025, compared with 151.7 million a year earlier. The same discussion identifies locomotive availability, network conditions, security incidents, and maintenance as continuing constraints. Higher throughput therefore coexisted with unresolved limitations.

The report also distinguishes rail performance from ports and pipelines. A recovery in one network should not be described as complete recovery of logistics. Freight routes are connected: rail delivery can be constrained by terminal handling, while efficient ports can remain underused if inland access is unreliable.

Throughput itself combines supply and demand. A larger tonnage may reflect improved capacity, stronger commodity production, changed routing, or a different cargo mix. To identify the operational contribution, the analyst needs scheduled and completed services, cancellations, turnaround times, equipment availability, and evidence of customers unable to obtain capacity.

Revenue adds a price dimension. Higher tariffs can raise receipts without increasing physical service. A utility's financial improvement and the competitiveness of the firms using its network are consequently related but distinct outcomes. Both require assessment; neither can replace the other.

Using the two reported annual rail volumes, the change is 8.4 million tonnes. Dividing by the earlier 151.7 million gives growth of approximately 5.54 percent, consistent with the report's rounded 5.5 percent.

This is an actual-period reconciliation, not a forecast. It does not show how much GDP the increase generated or whether all unmet demand was served. Those claims would require cargo detail, alternative transport use, prices, and production evidence.

A shift from road to rail can raise rail tonnage without an equivalent increase in total goods produced. It may still reduce costs, congestion, or environmental burdens. Conversely, more output transported over the same constrained network can increase delays. The economic result depends on service quality and the counterfactual, not only the aggregate volume.

Freight composition should remain visible in the recovery account. Moving an additional tonne of bulk minerals is not operationally equivalent to handling another container or a finished vehicle. Routes, loading equipment, service frequency, and customer requirements differ. A national tonnage increase can therefore coexist with poor service on an important manufacturing corridor.

The same distinction affects an exporter's inventory decision. Better reliability can permit lower buffer stocks and faster collection even where annual production is unchanged. The released working capital is a commercial benefit that a volume-only comparison misses. A South African industrial assessment should examine delivery variance and inventory days alongside average transit time, then distinguish these cost and liquidity effects from additional output. This provides a fuller account of the return to network improvement.

A producer needing both reliable power and export logistics cannot necessarily substitute one for the other. Better electricity can permit more production while a rail bottleneck limits shipments. More freight capacity can be underused if processing plants cannot operate. The return to relieving one constraint depends on the condition of the other.

This complementarity explains why isolated capacity announcements can overstate the near-term growth response. A new generation project needs a usable connection; a freight corridor needs terminals, rolling stock, and reliable operations. The last missing component can determine whether earlier expenditure becomes productive service.

The location of industry matters as well. A national improvement may have little immediate effect on a firm connected to a weak municipal distribution network or a poorly served transport route. Sector averages should be supplemented with geographic and network information before estimating exposure.

Municipal services introduce further dependencies through water, roads, distribution infrastructure, and local administration. These can constrain production even when national electricity availability improves. A business assessment should identify its actual service providers and routes rather than assign the national average to every establishment.

A firm deciding whether to expand evaluates expected reliability over the life of the asset. One favorable month is insufficient for a long-lived project. Credible maintenance, network investment, and operating arrangements can change the perceived risk and therefore the investment hurdle.

Financing conditions can nevertheless prevent an otherwise useful response. High borrowing costs, uncertain sales, or limited collateral can hold back expansion after infrastructure improves. The absence of immediate investment is therefore not proof that the service improvement has no value. It may indicate that a second constraint remains binding.

Private adaptation also changes the composition of investment. Spending on diesel generation or excessive inventories can protect production but provides a different return from spending on new productive machinery. A recovery should examine whether firms can redirect resources from coping with unreliable services toward expanding output or improving efficiency.

Public network expenditure requires similar scrutiny. Maintenance can preserve more effective capacity than a new asset poorly connected to the system. Capital expenditure totals should be divided between renewal, rehabilitation, and expansion, then connected to service outcomes. A larger budget is an input to recovery, not its completed result.

A convincing diagnosis combines orders, industrial production, utilization, electricity reliability, freight service, and business costs. Where orders are strong and service access limits delivery, infrastructure remains a binding constraint. Where service improves but orders remain weak, the immediate production response may be limited even as costs fall.

Evaluation should preserve those mixed outcomes. A business can gain resilience without raising output; a utility can raise throughput while still failing some customers; a sector can recover while smaller firms remain excluded. An aggregate growth rate cannot resolve these differences on its own.

South Africa's infrastructure recovery should therefore be judged by dependable service and the productive choices it enables. The evidence reviewed here supports recognition of measurable rail improvement and a plausible easing of some supply constraints, while retaining the separate roles of demand, network complementarity, and investment confidence. That is the basis for assessing how much further operational progress can translate into sustained production and income.