Long time frames and dodgy numbers justify worry about nuclear power’s cost
Keith Bryer’s article “Anti-nuke campaigners never let facts intrude on a good tale” (Business Report, June 6) was somewhat misdirected.
The issue I have with nuclear power is not about safety. The fact is that we will never know enough about the safety issue unless Eskom agrees to publish the health records of the workers at Koeberg, Pelindaba and Vaalputs.
It would help also if the City of Cape Town would conduct a peer-reviewed, independently funded study on rates of childhood leukemia in Atlantis and Mamre. Until this is done, the only honest scientific answer is that there is not enough evidence one way or another in the public sphere regarding the effect of radiation produced during the normal operations of a nuclear power in South Africa.
In this sense the nuclear issue is not about science but about governance. The public right to know what is being done with taxpayers’ money is not being met by sufficient information.
The safety issue has been fully dealt with by organic farmer and poet Wendell Berry, who observed: “Dangerous accidents do happen in nuclear power plants. Officials and experts claim that accidents can be foreseen and prevented, but accidents are surprises by definition. If they are foreseen, they do not happen.”
Any technology that depends for its safety on the notion of human infallibility is making wildly over-optimistic assumptions. This fact alone makes me reluctant to entrust them with almost R1 trillion of taxpayers’ money.
My chief objection to nuclear power is precisely on the cost issue.
So far, a technology has not been chosen. Thus, an earlier tender issued by Eskom in 2008 for nuclear power did not stipulate the type of reactor required.
As a result, the two bids received were not comparable and a tender was not awarded. Both bids were more for than double the cost for nuclear energy set out in the nation’s 20-year electricity plan, the Integrated Resource Plan (IRP) version one, and more than 50 percent higher than the cost set out in the final, post-consultation IRP II.
In other words, from 2008 to early 2010 nuclear costs were stated to be twice as much by nuclear power vendors than by the state. By early 2011, after a multitude of submissions by civil society, even the state admitted that it had undercosted nuclear power by about 40 percent and included a new estimate. The new estimate remained substantially lower than the two tenders submitted by suppliers.
Yet the amount of nuclear energy planned in the IRP II remains unchanged, despite the variation in costs. International experience rates the cost of constructing nuclear plants at about $6 000 (R63 500) per kilowatt hour.
The South African energy planning process calculated this cost as $3 000 in 2010 and $4 300 in 2011 – on what basis remains a mystery.
The second problem with assembling a budget for nuclear procurement is that the costs of waste disposal and decommissioning are of the same order as the cost of construction, but are beset by large uncertainties. For example, in the 2007/08 annual report and accounts of nuclear power generator British Energy, it was estimated the cost of decommissioning its eight plants was £9.4 billion (R167bn) and the cost of disposing of the spent fuel was £5.5bn.
Although we can estimate the order of magnitude, the actual costs are affected by the choice of technology. As such, it is important to include these costs in a tender since they deeply affect the final choice of bidder. So far these costs have not been included in South African energy planning.
As far as the cost of waste disposal is concerned, it has to be borne in mind that these costs have to be borne for a minimum of two and a half centuries before the waste can safely be neglected. Even a very small error in calculation can lead to very large divergences across this timespan. Under conventional accounting procedure, liabilities that must be met in the future should be discounted.
Effectively, this means that a sum of money (or assets of that value) is set aside now and it is assumed that money will earn interest and grow to meet the liability.
So if a liability of $105 falls due in a year and an interest rate of 5 percent can be earned, the discounted value of the liability is $100 because in one year it will grow sufficiently to meet the liability. In the short term this sounds a sensible procedure. But over longer periods, the operation of compound interest means that growth rate of the money is unpredictable. For example, a sum invested for 100 years earning an interest rate (net of inflation) of only 3 percent, will grow 19 fold.
But say that the costs of waste disposal are estimated at 50 percent of the cost of construction, and end up actually being 53 percent of construction. Over 250 years this means the discounted value (the sum of money we need to set aside today to cover the costs of waste disposal) easily doubles.
What happens if, as is the case in Germany and Japan, the interest rate is negative? It would mean we have to set aside more money now than will be required in the future.
This example demonstrates the point that the really difficult part of nuclear energy planning is that the amount it is going to cost is not knowable. Calculating the net present value of a 250-year expenditure would require that we could foresee the interest rate and the inflation rate for the next 250 years. But we are citizens, not soothsayers. Anybody who tells you they can predict these costs is talking through their hat. To any suggestion that we should give the approval anyway and trust officials to prevent any unreasonable cost overruns, I have but one word: Nkandla.
* Dr Yvette Abrahams works in the department of women and gender studies at University of the Western Cape and with Electricity Governance Initiative South Africa.