Developments in the South African energy debate

We recently published a set of three briefs on the South African Government’s policy on nuclear power. The Department of Energy has now published a draft new Integrated Resource Plan, whose purpose is to provide the basis for long-term planning in the provision of electricity. This brief aims to provide a succinct summary of the latest developments and, more specifically, to draw attention to the most important issues that have arisen in this process.

We recently published a set of three briefs on the South African Government’s policy on nuclear power. The Department of Energy has now published a draft new Integrated Resource Plan, whose purpose is to provide the basis for long-term planning in the provision of electricity.  This brief aims to provide a succinct summary of the latest developments and, more specifically, to draw attention to the most important issues that have arisen in this process.

Public consultation on the draft Integrated Resource Plan

Following the approval by Cabinet on 2 November 2016, the Minister of Energy released draft documentation relating to a new Integrated Energy Plan (IEP) and Integrated Resource Plan (IRP) on 22 November 2016.  The IEP addresses the future energy landscape in general, whereas the IRP focuses specifically on electricity.  Both are long-term planning documents.  More attention so far has been devoted to the IRP, particularly because of the emphasis placed by Government (and Eskom) on the importance of a substantial expansion of nuclear power within the overall electricity provision forecasts to 2050.

The draft updated IRP’s base case envisages a new build programme of power generating capacity over the period 2021 to 2050, with the following components:





Capacity (MW)


% of total

Solar PV (energy produced by the exposure of photovoltaic cells to solar radiation)

17 600



37 400


Demand Response (programme established to motivate changes in electric use by end-use customers in response to changes in the price of electricity over time, or to give incentive payments designed to induce lower electricity use at times of high market prices or when grid reliability is jeopardized)



Nuclear (first new plant only to start generating electricity in 2037)

20 385


OCGT (Open cycle gas turbine)

13 332


CCGT (Combined cycle gas turbine)

21 960


Coal PF (pulverised fuel)

15 000


Inga (hydro-electric power from the Congo River)

2 500



128 677


Currently, total installed capacity is 51 000 MW.  Much of this will be obsolete by 2050, but with the new capacity to be added, the draft updated IRP envisages a base case with a total installed capacity in 2050 of 136 000 MW.

The Minister announced that the next step was for public consultation to be held during December 2016 and January 2017, culminating in Government adjustments in March 2017 and the promulgation of a final IRP, once it is approved by Cabinet.

The first public consultation in this process was held in Johannesburg on 7 December 2016.  From the presentations by a number of institutions and companies, it became clear that there were certain major issues that require attention.  These issues are summarised as follows:

  Limited notice for preparation of comment

Several presenters emphasised that a period of just over two weeks from the Minister’s announcement to a public consultation event is insufficient for the preparation of substantive and meaningful comment.  The comment was made that it is not clear why the process is being conducted with such urgency.  The Energy Intensive User Group of Southern Africa (EIUG), for instance, felt that significant electricity generation from the new power stations at Medupi, Kusile and Ingula would lead to a surplus capacity for a period, so that there is no need for an urgent investment decision for further generating capacity in the short term.  

Constraints placed on annual new-build of renewable energy and outdated cost estimates for renewables

The CSIR pointed out that in the draft updated IRP, limits had been placed on the annual capacity to build wind and solar PV and that these limits are not technically or economically justifiable.  Such constraints had not been placed in the draft updated IRP on any other power technology.  In addition, the CSIR pointed out that the cost assumptions for solar PV and wind are far too high, representing an excess of 50% and 30%, respectively, above the latest South African tariffs in those sectors.  These current tariffs reflect a cost for solar PV and wind of R0.62/kWh, whereas by comparison, the cost assumption in the draft updated IRP for new nuclear power is R0.97/kWh[1]. 

On the other hand, the nuclear cost estimate is derived from construction costs in Asia, in a study commissioned by the Department of Energy, which felt that cost assumptions derived from nuclear construction in the West were too high for its purposes.  This nuclear cost assumption is based on the USD/ZAR exchange rate in January 2015 of R11.55.  The depreciation of the Rand since that date (currently about 16%), has not been taken into account here and would increase the cost in Rand terms.

If no constraints are placed on new build capacity for solar PV and wind and their costs in the model are reduced to reflect actual current market conditions, the CSIR model shows that there is no place at all for a nuclear option in the forecast of new power plant construction up to 2050.  The forecast electricity production in 2050 would have the following composition in an unconstrained model:  18% solar PV, 52% wind [2], 16% gas, 3% hydro, 12% coal and a negligible peaking requirement (although the installed peaking capacity would amount to 13% of total installed capacity, to act as a safety net in case of need.).  The CSIR model also shows that by 2050, its optimised scenario would be R90 billion (in 2016 Rand) per annum cheaper than the draft IRP base case of R580 billion, which includes a nuclear component.

It is important to note in this context that the draft updated IRP and the CSIR model rely on identical software platforms.  Any changes in the inputs to the model are therefore clearly identifiable.  It is not a situation where attempts may be made to explain away conflicting results by reference to different model methodologies.

In a later presentation, the Chairperson of the Ministerial Advisory Council on Energy (MACE) Working Group, reported to the public consultation meeting that the Working Group (a group of industry experts appointed by the Minister on 16 September 2016) had submitted a memorandum to the Minister on 31 October 2016 with its findings, which were: 

  • the base case should be least-cost and the annual new-build limits imposed on solar PV and wind should therefore be removed;
  • assumed costs for solar PV and wind should be revised downward to align them with actual South African tariffs; and
  • cost differences between the least-cost unconstrained base case and alternative scenarios must be reported, to enable a value for money case to be assessed. 

None of these recommendations was taken into account in the draft updated IRP, and no reasons for ignoring them have so far been provided.

During the general discussion at the end of the presentations, the Department of Energy’s representative stated that it had maintained certain constraints on renewable energy new-build as a starting point.  It was not taking a position that this could not be changed. Although no reasons were provided in the draft updated IRP document for the placing of limits on renewable construction, in a statement dated 30 November 2016, the Minister declared  that “renewable energy cannot be unconstrained …  This is because there are network constraints that will limit the extent to which renewable energy can be connected to the electricity distribution grid.”

The Department’s representative added that there was nothing sinister about the constraints.  He said that he was “half-sold” on the suggestions for the constraints to be removed, but that the Department needed to reflect and take this issue back to its principals.  In response to a question on why grid adjustments are not also necessary for new nuclear power stations, the Department’s representative said that a grid analysis should be undertaken to show the effect of new renewables.  He added that with nuclear you know where it is going to be built, whereas with renewables, this is not the case. 

  CSP Technology

Concentrating Solar Power (CSP) relies on the concentration of the sun's energy through large mirrors and utilises that concentrated energy to produce steam to drive a conventional steam turbine for electricity generation.

A common theme from CSP providers who made presentations was that they were dismayed at the exclusion of their technology from the IRP.  In particular, they pointed out the advantage of the CSP technology in providing flexibility to the electricity grid as a result of its energy storage capabilities (compared to Solar PV, which they said does not have such a storage capability).


Similarly, dissatisfaction was expressed by representatives of industry which have a co-generation capacity (for example, the Paper and Pulp Manufacturers Association) that this technology is not considered in the IRP.  (Co-generation refers to technologies that operate together for the simultaneous generation of electricity and heat, in a process that is generally more energy-efficient than the separate generation of electricity and heat.)

Cost overruns and delays in nuclear

The point was made by more than one speaker that the nuclear sector has the highest risk of cost and construction-time overruns.


It was also pointed out by participants that Eskom cannot be allowed to be both a player and decision-maker and according to one participant, this displayed a situation of monopoly abuse.  (Eskom did not make any presentations to the gathering.)

Some comments

Given the flurry of activity and the divergent approaches on this topic over the past month, it is necessary to try to make sense of the whole situation.   

In the first place, Eskom’s attitude needs to be understood and addressed.  It is Government’s function to do this, since Eskom is a public utility which is wholly-owned by the state.  In a recent article on (published on 22 November 2016, on the same day as the Minister published the draft updated IRP), Eskom’s head of generation (and now, acting CEO), Matshela Koko made the following comments:  firstly, Eskom is placed under financial pressure by the cost of independent power producers, and secondly, wind and sun power cannot economically replace sources such as gas and nuclear for a reliable supply.   

However, his comments on the financial burden that renewables place on Eskom do not contain any comparative cost analysis.  It is not sufficient to complain about renewable costs and at the same time, not to include a comparison of any kind with the cost of non-renewable supplies such as nuclear.   He also points out that these renewable costs are much higher than Eskom’s short run marginal cost of applying its current surplus capacity.  This argument assumes that these marginal costs (clearly of a temporary nature, arising from current surplus capacity) will continue forever.  The huge upfront cost which applies to any nuclear power plant construction is not referred to in this context. 

Eskom’s second point, namely that “intrinsic limitations mean that wind and sun can never economically replace sources such as gas and nuclear for large-scale, continuous, reliable supply”, takes no notice of recent studies (such as that of the CSIR), which show that the electricity grid can be powered mainly by renewables, provided that it is supplemented by peaking facilities (for example, gas).  Whilst it is clear that such peaking facilities would be more expensive to run on a continuous basis than most other power sources, they firstly do not require a massive capital outlay such as in the case of the nuclear option and secondly, unlike nuclear facilities, they can be turned on and off at will at short notice.  In addition, the potential for cost overruns and delays in the building of nuclear power stations is very substantial, as any analysis of global nuclear power plant construction will show.  The same is true, incidentally, for conventional coal power stations such as Medupi, which has suffered from massive cost overruns and delays under Eskom’s own construction management. 

Eskom has recently also stated that it would be going to the market to seek estimates for the building of nuclear power stations, but that this was initially only a price discovery procedure, since if the nuclear project is too expensive, it would not be pursued.  Eskom is correct that this is the only way to obtain a firmer idea of what nuclear power would cost. But at the same time, a level-headed assumption would indicate that it would not only be too expensive (on a comparative basis), but would also carry substantial and unquantifiable additional risks of cost-overruns and delays.

Regarding the tempo at which the updating of the IRP is being conducted, it remains completely unclear why there seems to be such a haste in trying to finalise an important policy document, instead of preparing a properly considered policy through an initial process of discussions between interested parties, before publishing a draft for public comment.  The quality and brevity of the documentation prepared by the Department of Energy in this process only supports the perception that too much haste and too little reflection has gone into the whole project so far.

How can a sensible conclusion be reached?

It is clear that there are a number of major issues to be resolved between the conflicts in the apparent policy of the Department of Energy, Eskom’s declared preferences and the findings of independent institutions. 

In any final IRP, it is assumed that the Department of Energy will have to set out what the lowest cost option is, and if it does not want to adopt that option, what the reasons are for a variation.  In addition, some of the protagonists seem to forget that the current debate is aimed at determining long-term policy, and that it serves no purpose in limiting any specific technology because of short-term difficulties that are experienced at present in integrating it into the grid.  After all, the base case put forward by the Department of Energy envisages new nuclear power only by 2037.  Surely, within the space of 21 years, there is enough opportunity to do whatever is necessary to develop the grid for what seems to be the cheapest option by far?

The process going forward will be closely watched by entities outside of Government and Eskom.  If the process, carried through in such obvious haste, culminates in a long-term energy plan which does not represent a rational or defensible outcome, litigation will be inevitable, thwarting the development of a coherent energy policy and jamming electricity capacity development for years to come. 


Anton van Dalsen
Legal Counsellor


[1] The CSIR presentation may be found on:, together with copies of other presentations made at the Johannesburg public consultation.
[2] South Africa has a very high potential for renewable energy development, compared with other countries