Electricity costs get a 10-year forecast you cannot use

The Revised Electricity Pricing Policy would require NERSA to publish a 10-year electricity price forecast. For most South African businesses that forecast will change nothing, because they cannot say what electricity costs them per site, per vehicle or per job. A price path you cannot apply to your own numbers is trivia.

Electricity and Energy Minister Kgosientsho Ramokgopa set out the proposals at a media briefing on 18 August. SAnews reported him putting the rise in tariffs at about 977% over a period stretching back to 2007, which is the trigger he gave for the rewrite. Mining Weekly reports the same remark as over 900%. It updates the 2008 policy and sits alongside the unbundling of Eskom and the Electricity Regulation Amendment Act, 2024.

Where this sits in the process, since the coverage reads as though it is already law. Cabinet approved the policy for public comment in July. Miningmx, citing Bloomberg, and Mining Weekly both put the gazette on Friday 21 August, with comment open until 20 September.

As at the morning of 19 August that gazette has not appeared. Nothing below is in force, and the numbers in it are proposals.

What is actually in it

Four things matter to a business buying power rather than selling it.

• A 10-year price forecast from NERSA. The stated reason is capital planning. Ramokgopa's example was heavy industry computing a return on investment where electricity is a large input cost.
• Unbundled, cost-reflective tariffs, so hidden costs cannot be buried inside a single rate.
• Municipal bad debt comes out of the tariff. Ramokgopa put the current recovery at between 1% and 2.5% of what you pay, charged to paying customers because others do not pay. Under the policy that is not allowed.
• Negotiated Pricing Agreements widen. Today the instrument triggers when a firm is in distress. The proposal adds a second route for industries that are not in distress but would accelerate growth or employment in priority sectors.

Alongside that, Eskom and municipalities would have to let customers buy from alternative suppliers while charging transparent, non-discriminatory network fees. Bilateral deals between a generator and an off-taker get a rule set. That is the wholesale market taking shape.

The relief is smaller than the headline

Take the bad debt line, because it is the only immediate number here. Multiply your own monthly electricity bill by 1% and by 2.5%. That range is what the policy proposes to stop charging you, and it is what a year of this reform is worth on your bill if everything else holds still.

Everything else is not holding still. Eskom's directly supplied customers took an 8.76% increase on 1 April, and customers buying through a municipality took 9.01% on 1 July. Against those, one to two and a half percent is a rounding correction, not a saving. This policy is about certainty and structure, not about a cheaper bill next quarter.

Certainty is worth having. It is worth more to a business that can price it.

The forecast is only worth what your own numbers are worth

A 10-year price path lets you answer questions like these. What does a 12% real increase over three years do to margin on the Nelspruit run? Which of our six sites should get solar first, and what is the payback at forecast tariffs rather than today's? Should we take a Negotiated Pricing Agreement, and what would we have to prove to qualify?

Every one of those needs a cost per site, per vehicle or per job. Almost no operator with 40 to 200 staff can produce it, and not because anyone is lazy. The meter readings sit with whoever collects them. The job cards sit in a different system, or in a book. The electricity account lands in the general ledger as one number against one supplier every month.

Three sources, never joined. So the business knows exactly what it spent on power last month and cannot tell you which work that power did. Ask which of two branches is the expensive one to run per unit produced, and the honest answer is usually a guess dressed up as a view.

What this is really measuring

That gap is not an electricity problem. It is the same gap that stops you knowing your true cost per delivery, per callout or per production hour. Power is simply the line item that just got a decade of public forecasting attached to it.

Fixing it is unglamorous and finite. Meter readings get captured against a site. Jobs get captured against a site. Both land in the same place as the invoice, and the ledger picks up an allocation rather than one lump.

That is a discovery audit finding one bottleneck, and then one workflow fixed properly. It is how Zorah works, because the alternative is a business that buys a system for every question it cannot answer.

What to do on Monday

Take last month's electricity invoice for one site. Then take the list of jobs, shifts or production runs that site did in the same month. Try to divide the one by the other.

If you can do it in ten minutes, this policy is genuinely useful to you and you should read the gazette when it lands. If you cannot, the 10-year forecast will be reported, discussed and forgotten, and your power bill will keep arriving as a single number nobody can argue with.