South Africa’s Electricity Choice (Part 1): Why Ramaphosa must kill the nuclear dream

Deputy President Cyril Ramaphosa has indicated his determination to improve the economy and attract more investment to South African shores in every one of his speeches since his election as ANC president in December 2017. But to succeed in his mission to put the country back on the economic track once he becomes president, he will need to act firmly and decisively to close the chapter on the Jacob Zuma administration’s obsession with nuclear power.

Introduction

The South African government’s obsession over the past decade with nuclear power may finally come to an end with a new president and administration committed to restoring good governance and securing reliable electricity supply at least cost to support economic growth and development^ii,iii. The inherent complexity of nuclear procurement, financing, and especially construction, means no new nuclear generated electricity could flow for at least a decade even with immediate procurement, and then only at prices well above other electricity options^iv. The risk that the project might collapse as unfinanceable^v,vi, or through corruption, creates further uncertainty around South Africa’s electricity choices.

Some South African advocates^vii,viii,ix assert that nuclear costs will prove affordable for five reasons:

vendors are eager to sell;

vendor financing might come at low cost and long debt tenors;

higher prices or subsidies are worth paying for such a “strategic input” as nuclear electricity;

South Africa is “ideally positioned” to help satisfy the African continent’s “huge appetite for nuclear energy”; and

Western countries’ “severe over-regulation and a culture of being overly risk averse”, over-reacting to the ostensibly innocuous events at Chernobyl and Fukushima, caused “severe cost and schedule overruns” that South Africa, presumably with a more sensible risk perception, could avoid.

Whether capital markets would validate these assertions or view them as signs of disconnection from technical and market realities is a testable empirical question.

The country’s latest two flirtations with modern nuclear power have not gone well. First, plans for a home grown “pebble bed” modular reactor (PBMR)^x were abandoned in 2010 after 12 years and $1.8-billion^xi had been wasted – much of it after the project was acknowledged to be a non-starter (as Eskom presumably knew when its US utility partner pulled out in 2002). Now rarely mentioned, the PBMR found no customers or investors, but its advocates’ ambition and influence live on.^xii

In 2008, perhaps slightly chastened by the unfolding PBMR episode, the government rejected as “uncompetitive and unaffordable” bids to build two light-water reactors totalling 3,2 gigawatts (GW or billion watts). The two undisclosed bids were reportedly more than twice government’s expectation^xiii. Despite such prices, the rejected nuclear vendors, which didn’t include Russia^xiv, failed to thrive: Areva became insolvent and is now being bailed out for €5-billion by French taxpayers^xv, while Westinghouse sold its nuclear business to the now junk-rated Toshiba, and is currently in bankruptcy proceedings^xvi.

Despite these setbacks, the South African government’s ambition mushroomed into a plan three times bigger than in 2008 – 9.6 GW of nuclear power – even though each GW will now cost more because real costs have risen by more than one-fourth^xvii while the rand has lost half its value. With the national debt downgraded to junk-bond status^xviii, “affordability” can hardly have improved.

Electricity planning

Government’s nuclear policy consistently cites as its analytic basis the Integrated Resource Plan (IRP) for Electricity 2010 to 2030, which was updated twice in 2010 after public consultation, promulgated 25 March 2011, and formally published 6 May 2011^xix. It called for 13% nuclear capacity and 21% solar and wind capacity in 2030. The IRP’s analysis found new nuclear power not economically optimal, even at real dollar costs far below today’s (though 40% above initial assumptions) and assumed to fall by a further 3%/y (they instead rose by roughly 5%/y).

The IRP nonetheless committed to 9.6 GW of nuclear power – as a matter of “policy” despite its inconsistency with the National Development Plan^xx. This “non-negotiable” nuclear policy has been enunciated consistently by President Zuma and his energy ministers from 2010^xxi, even as its stated rationale has steadily evaporated. That rationale was chiefly “to account for the uncertainties associated with the costs of renewables and fuels”. As we’ll see, South African renewable energy costs were then discovered in the marketplace and proved to be far lower than empirical nuclear costs, with the spread between them continuing to widen; renewables, just like nuclear power, displace the same fuels whose uncertain prices were of concern; and renewables meet all the other stated policy objectives as well as or better than nuclear power.

A 2013 draft IRP update^xxii, though somewhat flawed^xxiii, was posted on the Department of Energy (DoE) website in November 2013^xxiv, soliciting public comment by February 2014 for consideration “in preparing a final draft IRP 2010 Update which (was to) be submitted to Cabinet for final approval by March 2014”. That updated draft was required, said the DoE’s website, because “there have been a number of developments in the energy sector” since 2010 and “the electricity demand outlook has changed markedly”.

The 2013 draft update strengthened the previous finding that nuclear new-build lacked economic rationale. It concluded that “in 2015, even if demand is high and there is no prospect of shale-gas power plant,…if nuclear costs exceed $6,500/kW, then the (nuclear) procurement programme should be abandoned.” Specifically, the draft found new nuclear plants clearly uneconomic if they got costlier, or renewables got cheaper, or demand growth slackened. In fact, all three strongly occurred. The plan indicated new nuclear to be unneeded for 15 to 25+ years – or ever if its capital costs rose another 12%. They have since risen much higher than this, with costs in some countries exceeding $6 500/kW^xxv,xxvi while renewables cost less and (using the best U.S. data) performed one-third better^xxvii,xxviii than the 2013 draft update had assumed. Two years on, the gap between nuclear and renewable costs had widened further^xxix,xxx.

However, the IRP 2013 was never finally approved and the parliamentary energy portfolio committee chair was reported at the time saying it “will never see the light of day”; indeed, it was never gazetted^xxxi and can thus be officially ignored.

By 2016, many of the 2013 draft update’s data had become even more antiquated and a decision was made by the DoE to produce a further IRP update, which was gazetted on 25 November 2016 for public consultation^xxxii. In the base (or reference) case, nuclear energy appears only from 2037, but after questions posed formally to the department by the Ministerial Advisory Council on Energy (MACE), it became clear that the modellers had placed artificial constraints on the amount of solar or wind energy that could be built in any one year. Once these constraints were removed, nuclear energy did not appear in any of the optimal least-cost scenarios^xxxiii.

The government ignored rational warnings and requested Eskom to expedite the procurement of nuclear power. However, it soon started butting up against increasingly vocal challenges from various quarters. In 2016, civil society organisations launched a court challenge to government’s continued preference for nuclear power. On 26 April 2017, the Western Cape High Court not only set aside a number of international nuclear co-operation agreements, including that with Russia, but also invalidated the ministerial determination (made in terms of section 34 of the Electricity Regulation Act) to procure nuclear electricity. The court found that the National Energy Regulator of South Africa (NERSA) had concurred with this ministerial determination (as required in law) but had done so without adequate public consultation. In effect, the ruling forced government to restart the procurement process from scratch, including producing a new update of the IRP and a new official ministerial procurement determination, but only after the concurrence of NERSA is obtained, and this only after adequate public consultation^xxxiv.

During the last quarter of 2017, the DoE asked the IRP modellers (based in Eskom) to produce another updated IRP to reflect lower actual electricity demand and the latest comparative energy prices. The DoE may have noticed the global publicity about plummeting renewable energy prices, now so low in many good sites, without subsidies, that they rival or undercut just the operating costs of many existing nuclear or fossil-fuelled power stations, let alone the total cost of power from new ones^xxxv,xxxvi.

The resulting new electricity demand forecasts of the modellers are almost certainly still too high, as are their assumptions of renewable energy costs. In a number of emerging economies in 2017, we have seen solar and wind auctions deliver prices that are half those registered in the last renewable energy auction in South Africa held in November 2015, after which Eskom had stalled further bids by refusing to sign PPAs ^{xxxvii,xxxviii,xxxix}. And the DoE’s nuclear cost assumptions seem much lower than many recent contracts, such as the Rosatom deal for up to $21bn in Egypt^xl.

But even with these conservative assumptions, nuclear energy is not picked in any of these new South African modelling scenarios, other than one where artificial constraints are placed on how much solar and wind energy can be built and where additional carbon budget limits are imposed. Even in this extreme scenario, nuclear energy would only be required after 2039.

These outcomes are consistent with the IRP update undertaken by the DoE in 2016 under the direction of the Ministerial Advisory Council on Energy and with the Council for Scientific and Industrial Research’s (CSIRs) model which used the same PLEXOS software and assumptions as the DoE in 2016.^xli A few years back, the National Planning Commission asked South Africa’s premier university-based energy modelling group – the University of Cape Town’s Energy Research Centre – to run an independent model. They too came to the same conclusion^xlii.

These outcomes were not welcomed by the nuclear section in the DoE and the modellers were asked to run a scenario where 9,6 GW of nuclear power stations are “hard-wired” or forced into the model. Fortunately, the modellers also calculated the additional wholesale cost of electricity in this scenario – R800-billion over 20 years at net present value^xliii.

The optimal, least-cost mix recommended by all these models includes solar and wind energy complemented by pumped storage, hydro-electricity and gas turbines or engines, on top of legacy coal fired powered stations.

South Africa’s Energy Minister decided to ignore the conclusions of his modelling team and presented a revised IRP2017 to Cabinet on 6 December 2017, which still included the procurement of 9,6 GW of nuclear power but over a longer period to reflect lower electricity demand.

South Africa’s Constitution requires administrative action to be lawful, reasonable, and procedurally fair. Almost certainly, civil society organisations will go back to court to challenge the rationality of the latest IRP and any actual procurements that might follow.

The IRP should be a living planning document that is regularly revised and updated as necessitated by changing circumstances. At the very least, it should be revised by the DoE every two years and should provide a rational and informed base for power procurement and investment decisions. The way in which DOE has continued to manipulate and adjust least-cost, optimal plans to incorporate nuclear power reduces stakeholder confidence that Cabinet is fully informed about the very fast-moving competitive landscape in electricity^xliv. Just within one year (2016), levelised world costs for new onshore wind power fell by 18%, for offshore wind by 28%, and for utility-scale photovoltaics (PV) by 17%^xlv, while low bids fell 37% for Mexican PV and 43% for EU offshore wind^xlvi. Yet South Africa’s Cabinet is still apparently relying on data using 2010 cost data if not earlier – myopia that would not pass muster in any private-sector boardroom.

Shifts ever more adverse to nuclear new-build have accelerated on every front. The global nuclear enterprise is slowly dying of an incurable attack of market forces^xlvii. Financial distress stalks vendors, with cascading insolvencies spreading in the past two years. Construction cost and delays keep rising worldwide^xlviii. As a Frost & Sullivan analysis found in 2014^xlix, whatever is driving South Africa’s nuclear push, it’s not economics. And as financial consultant Dirk de Vos concluded^l, “The nuclear industry can only exist when the state underwrites its cost. Countries that procure their power through democratic, transparent and market-based methods are not building new reactors.”

All this raises several questions: what could emerge from the “full, transparent and thorough”^li cost-benefit analysis previously requested by the ANC national general council in 2015^lii,liii, and from the Constitutionally required transparent competition demanded by the National Union of Mineworkers^liv? What salient evidence was absent when Cabinet formally launched the nuclear procurement programme based on the gravely outdated 2010–11 analysis? What major decision factors have changed since the failed 2008 nuclear plan? We explore these questions in the context of the two most dynamic competitors – renewable generation and efficient use – then discuss policy risks. DM

In Part 2, we explore renewable energy.

Physicist Amory Lovins is a renowned, leading world energy expert. A former Oxford don, honorary US architect, Swedish engineering academician, adviser to business and government leaders for 44 years in over 65 countries including South Africa, he has won many of the world’s top energy and environment awards, received 12 honorary doctorates, taught at 10 universities, and written 31 books and over 625 papers. Time named him one of the world’s 100 most influential people; Foreign Policy, one of the 100 top global thinkers

Anton Eberhard was inspired to undertake his PhD in the field of energy and development in 1979 after reading Amory Lovins’ seminal publications on renewable energy and energy efficiency. He has recently been elected to the rank of Emeritus Professor and Senior Scholar at the University of Cape Town after 35 years of research, teaching and policy advocacy in energy and sustainable development in Africa.

Photo: A file picture dated 11 September 2006 shows the Units 5 and 6 of the Bulgarian nuclear power plant during a sunrise in the town of Kozloduy. EPA/VASSIL DONEV

Footnotes: