World Bank South-Africa-Eskom-Investment-Support-Project (2022) EISP (106p)

Public Disclosure Authorized
Document of
The World Bank
FOR OFFICIAL USE ONLY

Report No: ICR00005338

IMPLEMENTATION COMPLETION AND RESULTS REPORT
ON A LOAN

IN THE AMOUNT OF US$ 3750 MILLION
TO THE

ESKOM HOLDINGS LIMITED

GUARANTEED BY THE REPUBLIC OF SOUTH AFRICA

FOR THE

ESKOM INVESTMENT SUPPORT PROJECT

February 28, 2022


Energy and Extractives Global Practice
Eastern and Southern Africa Region
This document has a restricted distribution and may be used by recipients only in the performance of
their official duties. Its contents may not otherwise be disclosed without World Bank authorization.
CURRENCY EQUIVALENTS
(Exchange Rate Effective March 26, 2021)

Currency Unit = South African Rand (R)
R 15.01 = US$1
R 1 = US$0.066

ESKOM’s FISCAL YEAR
April 1 – March 31

ABBREVIATIONS AND ACRONYMS

AfDB African Development Bank
BESS Battery Energy Storage System
BSP Battery Storage Program
CAPEX Capital Expenditure
CCGT Combined Cycle Gas Turbine
CSP Concentrated Solar Power
CPF Country Partnership Framework
CPS Country Partnership Strategy
CTF Clean Technology Fund
DCCSF Development and Climate Change Agenda: A Strategic Framework for the World Bank Group
DME Department of Mineral Resources and Energy
EAF Energy Availability Factor
EIRR Economic Incremental Rate of Return
EISP Eskom Investment Support Project
ERR Economic Rate of Return
ERSP Eskom Renewables Support Project
ESKOM Eskom SOC Holdings Limited
FGD Flue Gas Desulphurization
FIDIC Fédération Internationale Des Ingénieurs‐Conseils (International Federation of Consulting
Engineers)
FIRR Financial Internal Rate of Return
GDP Gross Domestic Product
GEF Global Environmental Facility
GHG Greenhouse Gas
GoSA Government of South Africa
HSE Health, Safety, and Environmental
ICR Implementation and Completion Results Report
IRP Integrated Resource Plan
IPP Independent Power Producer
ISR Implementation Status and Results Report
LNG Liquefied Natural Gas
M&E Monitoring and Evaluation
MCPP Medupi Coal‐Fired Power Plant
MET Medupi Execution Team
MTPY Million Tons Per Year
NERSA National Electricity Regulator of South Africa
NPV Net Present Value
O&M Operation and Maintenance
OEM Original Equipment Manufacturer
PAD Project Appraisal Document
PCLF Planned Capacity Loss Factor
PDO Project Development Objective
PFMA Public Financial Management Act
PPPFA Preferential Procurement Policy Framework Act
PV Photovoltaic
REIPPP Renewable Energy Independent Power Producer Programme
SAPP Southern Africa Power Pool
SDR Safeguards Diagnostic Review
SVC Social Value of Carbon
TA Technical Assistance
UCLF Unplanned Capacity Load Factor
UCS Use of Country Systems
URS User Requirement Specifications
WACC Weighted Average Cost of Capital
WTP Willingness to Pay

Regional Vice President: Hafez M. H. Ghanem
Country Director: Marie Francoise Marie‐Nelly
Acting Regional Director: Ashish Khana
Practice Manager: Julia M. Fraser
Task Team Leader(s): Franz Gerner, Frederic Verdol
ICR Main Contributor: Joel J. Maweni
TABLE OF CONTENTS
DATA SHEET ............................................................................................................................1
I. PROJECT CONTEXT AND DEVELOPMENT OBJECTIVES ........................................................ 7
A. CONTEXT AT APPRAISAL ........................................................................................................... 7
B. SIGNIFICANT CHANGES DURING IMPLEMENTATION (IF APPLICABLE) ...................................... 20
II. OUTCOME ...................................................................................................................... 22
A. RELEVANCE OF PDOs .............................................................................................................. 22
B. ACHIEVEMENT OF PDOs (EFFICACY) ........................................................................................ 25
C. EFFICIENCY ............................................................................................................................. 30
III. KEY FACTORS THAT AFFECTED IMPLEMENTATION AND OUTCOME ................................ 37
A. KEY FACTORS DURING PREPARATION..................................................................................... 37
B. KEY FACTORS DURING IMPLEMENTATION .............................................................................. 40
IV. BANK PERFORMANCE, COMPLIANCE ISSUES, AND RISK TO DEVELOPMENT OUTCOME .. 44
A. QUALITY OF MONITORING AND EVALUATION (M&E) ............................................................. 44
B. ENVIRONMENTAL, SOCIAL, AND FIDUCIARY COMPLIANCE ..................................................... 45
C. BANK PERFORMANCE ............................................................................................................. 50
D. RISK TO DEVELOPMENT OUTCOME ........................................................................................ 53
V. LESSONS AND RECOMMENDATIONS .............................................................................. 54
ANNEX 1. RESULTS FRAMEWORK AND KEY OUTPUTS ............................................................ 59
ANNEX 2. BANK LENDING AND IMPLEMENTATION SUPPORT/SUPERVISION ......................... 68
ANNEX 3. PROJECT COST AND FINANCING PLAN ................................................................... 71
ANNEX 4. EFFICIENCY ANALYSIS ............................................................................................ 74
ANNEX 5. BORROWER, CO‐FINANCIER AND OTHER PARTNER/STAKEHOLDER COMMENTS ... 98
The World Bank
Eskom Investment Support Project (P116410)
DATA SHEET
BASIC INFORMATION
Product Information
Project ID Project Name
P116410 Eskom Investment Support Project
Country Financing Instrument
South Africa Investment Project Financing
Original EA Category Revised EA Category
Full Assessment (A) Full Assessment (A)
Organizations
Borrower Implementing Agency
ESKOM Holdings SOC Limited ESKOM
Project Development Objective (PDO)
Original PDO
The project development objective (PDO) of the Eskom Investment Support Project for South Africa is to enhance
its power supply and energy security in an efficient and sustainable manner so as to support both economic growth
objectives and South Africa's long-term carbon mitigation strategy.
PDO as stated in the legal agreement
The objective of the Project is enabling the Borrower to enhance its power supply and energy security in an
efficient and sustainable manner so as to support both economic growth objectives and the long‐term carbon
mitigation strategy of the Guarantor.
Page 1 of 102
The World Bank
FINANCING
Original Amount (US$) Revised Amount (US$) Actual Disbursed (US$)

World Bank Financing
3,750,000,000 3,159,610,781 3,159,610,781
IBRD‐78620
Total 3,750,000,000 3,159,610,781 3,159,610,781
Non‐World Bank Financing
0 0 0
Borrower/Recipient 4,736,770,000 9,208,000,000 9,208,000,000

MULTIPLE DONORS 350,000,000 77,000,000 77,000,000
Bilateral Agencies
2,361,250,000 3,257,000,000 3,257,000,000
(unidentified)
Foreign Multilateral
2,663,800,000 2,427,000,000 2,427,000,000
Institutions (unidentified)
Total 10,111,820,000 14,969,000,000 14,969,000,000
Total Project Cost 13,861,820,000 18,128,610,781 18,128,610,781
KEY DATES
Approval Effectiveness MTR Review Original Closing Actual Closing

08‐Apr‐2010 31‐May‐2010 30‐Jun‐2014 31‐Oct‐2015 30‐Jun‐2021
Page 2 of 102
The World Bank
RESTRUCTURING AND/OR ADDITIONAL FINANCING
Date(s) Amount Disbursed (US$M) Key Revisions

11‐Dec‐2014 2033.48 Reallocation between Disbursement Categories
07‐Jul‐2015 2300.73 Change in Results Framework
Change in Components and Cost
Change in Loan Closing Date(s)
Change in Financing Plan
Reallocation between Disbursement Categories
Change in Legal Covenants
Change in Implementation Schedule
07‐Dec‐2018 3027.61 Change in Results Framework
Change in Safeguard Policies Triggered
Change in Legal Covenants
Change in Procurement
Other Change(s)
27‐Dec‐2019 3073.61 Change in Results Framework
Change in Procurement
25‐Jun‐2020 3108.43 Cancellation of Financing
30‐Jun‐2020 3108.43
KEY RATINGS
Outcome Bank Performance M&E Quality

Moderately Unsatisfactory Moderately Unsatisfactory Modest
RATINGS OF PROJECT PERFORMANCE IN ISRs
Actual
No. Date ISR Archived DO Rating IP Rating Disbursements
(US$M)
01 18‐Jun‐2010 Satisfactory Satisfactory 0
Page 3 of 102
The World Bank
02 28‐Mar‐2011 Satisfactory Satisfactory 361.61

03 01‐Dec‐2011 Satisfactory Moderately Satisfactory 569.14
04 11‐Jul‐2012 Satisfactory Moderately Satisfactory 968.61
05 12‐Apr‐2013 Moderately Satisfactory Moderately Unsatisfactory 1134.62
06 25‐Oct‐2013 Unsatisfactory Unsatisfactory 1332.85
07 08‐May‐2014 Unsatisfactory Unsatisfactory 1465.53
08 02‐Dec‐2014 Unsatisfactory Unsatisfactory 2033.48
09 08‐Jun‐2015 Unsatisfactory Unsatisfactory 2292.00
Moderately
10 30‐Dec‐2015 Unsatisfactory 2460.15
Unsatisfactory
Moderately
11 10‐Jul‐2016 Unsatisfactory 2639.77
Unsatisfactory
Moderately
12 18‐Jan‐2017 Unsatisfactory 2786.60
Unsatisfactory
Moderately
13 26‐Jul‐2017 Unsatisfactory 2857.10
Unsatisfactory
Moderately
14 30‐Mar‐2018 Unsatisfactory 3027.61
Unsatisfactory
Moderately
15 02‐Nov‐2018 Moderately Unsatisfactory 3027.61
Unsatisfactory
17 11‐Nov‐2019 Moderately Satisfactory Moderately Unsatisfactory 3073.61
19 17‐Aug‐2020 Moderately Satisfactory Moderately Unsatisfactory 3108.43
Moderately
20 12‐Feb‐2021 Moderately Unsatisfactory 3131.79
Unsatisfactory
Moderately
21 03‐Aug‐2021 Moderately Unsatisfactory 3131.79
Unsatisfactory
Page 4 of 102
The World Bank
SECTORS AND THEMES
Sectors
Major Sector/Sector (%)
Energy and Extractives 100
Renewable Energy Biomass 2
Renewable Energy Geothermal 2
Energy Transmission and Distribution 80
Renewable Energy Solar 2
Renewable Energy Wind 2
Other Energy and Extractives 12
Themes
Major Theme/ Theme (Level 2)/ Theme (Level 3) (%)

Private Sector Development 30

Jobs 30
Job Creation 30
Urban and Rural Development 60

Urban Development 30
Urban Infrastructure and Service Delivery 30
Rural Development 30
Rural Infrastructure and service delivery 30
Environment and Natural Resource Management 10

Climate change 10
Mitigation 10
ADM STAFF
Role At Approval At ICR
Regional Vice President: Obiageli Katryn Ezekwesili Hafez M. H. Ghanem
Country Director: Ruth Kagia Marie Francoise Marie‐Nelly
Director: Inger Andersen Ashish Khanna
Page 5 of 102
The World Bank
Practice Manager: Subramaniam Vishwanathan Iyer Julia M. Fraser
Task Team Leader(s): Reynold Duncan, Pankaj Gupta Franz Gerner, Frederic Verdol
ICR Contributing Author: Joel J. Maweni
Page 6 of 102
The World Bank
I. PROJECT CONTEXT AND DEVELOPMENT OBJECTIVES
A. CONTEXT AT APPRAISAL
Context
1. At the end of 2008, the Government of South Africa (GoSA) requested the World Bank to
participate in financing, alongside the African Development Bank (AfDB) and other agencies, a 4,800 MW
coal‐fired power plant that was already under construction at Medupi in the country’s Limpopo Province.
South Africa was experiencing acute power shortages, which peaked in early 2008 and, together with a
global financial crisis, had serious adverse impacts on the economy, including a 1.7 percent gross domestic
product (GDP) contraction in the first quarter of 2008,1 a record decline of about 26 percent in mining
sector output, job losses, and business closures.
2. Annual GDP growth had averaged a steady 4.1 percent during the previous decade (1999–2008)
driven in part2 by increasing electricity consumption. Electricity demand had grown faster (> 6 percent per
year) than GDP. South Africa’s high growth of electricity demand3 arose from the capital‐intensive nature
of its manufacturing and mining industries; the country’s rapid urbanization; and the successful
electrification program, which had increased access from 34 percent in 1993 to about 81 percent in 2007.
3. The growth in electricity demand was, however, not accompanied by significant investments in
new generation capacity. As a result, reserve margins declined from 15 percent to about 6 percent by the
end of 2007. By early 2008, peak demand exceeded generation capacity, leading to severe shortages and
load shedding. The private sector had not stepped in to compensate for the lack of public sector
investment. This was for a variety of reasons, including perceptions of high risks given the slow pace of
structural and policy reforms and low electricity prices in South Africa. In addition, traditional investors in
the power industry had withdrawn from emerging markets following the bankruptcy of the US energy
company, Enron, in 2001.
4. In 2001 the GoSA had, based on a policy decision, prohibited Eskom from adding new generation
capacity in the expectation that the private sector would do so. By 2004/05, when it became apparent
that the anticipated private sector response was not forthcoming and power deficits would occur around
2007, Eskom started planning for new capacity generation expansion. In 2005, Eskom’s Board of Directors
approved a procurement and contracting plan for the construction of the Medupi Coal‐fired Power Plant
(MCPP). The original plan was to finance the project through international commercial debt and Eskom’s
internal cash generation. However, the 2007/08 global financial crisis reduced Eskom’s access to
international commercial debt. At the same time, Eskom’s internal cash generation was reduced due to
1 According to a 2009 study by Deloitte, nearly 1 percent of the reduction in GDP could be attributed to electricity shortages.
2 A record of sustained macroeconomic prudence and a supportive global environment and quality infrastructure, including
adequate and reliable electricity supply, had contributed to that growth.
3 According to the Project Appraisal Document (PAD) (page 10) GDP per unit of electricity consumed in South Africa in 2006 was
only 60 percent that of the average upper‐middle‐income country, 40 percent of Brazil, and 30 percent of Mexico; in fact, it was
at par with the average low‐income country.
Page 7 of 102
The World Bank
the macroeconomic downturn. Therefore, Eskom and the GoSA sought funding from the World Bank and
other international financial institutions.4
5. The project’s economic and developmental rationale was clear. The concern was that power
shortages would hold back the economy’s recovery, which was critical for job creation, especially in the
mining, manufacturing, and informal sectors. Although the average economic growth during the
preceding decade was reasonable, the benefits were not widely shared, and there was a sense of
increasing inequality and non‐inclusion. The then current GoSA’s development strategy placed more
emphasis on shared and inclusive growth than had been the case in previous administrations. The
availability of adequate, reliable, and affordable electricity, among other infrastructure services, was
considered an essential link between growth and broader economic and social development. The impact
of providing electricity access was demonstrated by studies conducted by the Department of Mineral
Resources and Energy (DMRE) in 2008/09 in the Limpopo, KwaZulu Natal, and Eastern Cape provinces.
These studies pointed to significant benefits among the electrified communities, including improvements
in educational outcomes for women and girls, income generation opportunities, and job creation.
6. The economic and developmental rationale of the project was not just confined to South Africa
but to the entire subregion, many of whose countries (Botswana, Namibia, Eswatini, and Zimbabwe) were
dependent on electricity imports from South Africa. These countries were severely affected by South
Africa’s power shortages of 2008. In addition, because the South African economy was about two‐thirds
of the Southern African subregional economy, slow growth in that country affected the whole subregion,
particularly through trade flows and labor remittances.
7. The GoSA had responded to the energy crisis by assigning high priority to electricity generation
capacity expansion in the short term. Eskom prepared a five‐year US$50 billion program to increase the
power generation capacity by about 5,000 MW or 12.5 percent of the existing capacity. The MCPP was
included as part of the least‐cost Integrated Expansion Plan. The analysis of alternative baseload
generation options indicated that coal was the only feasible option in the short term. The alternatives to
coal were renewable energy sources (wind power and concentrated solar power [CSP]), hydro, diesel, and
combined cycle gas turbine (CCGT) fueled with liquefied natural gas (LNG). None of these, however, could
be developed in the short term to meet the generation capacity expansion needs because (a) wind was
not suitable for baseload operations due to its intermittency; (b) hydropotential was limited; (c) while CSP
had great potential and could be used for baseload operations, the technology had not yet matured, and
funding was not available to cover its incremental costs above thermal options; and (d) diesel was
prohibitively expensive.
8. Having decided that coal was the only feasible alternative for meeting power demand in the short
term, Eskom reviewed alternative coal power plant technologies in 2006.5 The alternatives were
subcritical, supercritical, ultrasupercritical pulverized coal; fluidized bed combustion; and integrated
gasification combined cycle. The choice of technology was determined based on the criteria of efficiency,
carbon emissions, maturity of technology, and operational risks. Supercritical technology has higher
4 The AfDB’s Board approved a loan of US$2.63 billion equivalent on November 25, 2009. The AfDB loan was provided to finance
some of the already committed and procured contracts for the MCPP, specifically the boiler and turbine contracts.
5 At appraisal, the World Bank undertook an extensive review of alternatives, as documented in the economic analysis section
and efficiency analysis, annex 4.
Page 8 of 102
The World Bank
efficiency (35–38 percent) and lower usage of coal per unit of electricity produced (lower emissions)
compared to subcritical technology under similar ambient conditions. Therefore, supercritical was a
better alternative to subcritical technology in use by Eskom at the time. Fluidized bed combustion, while
more advanced and less polluting, had no boiler size units in the range needed by Eskom. Ultra‐
supercritical technology, while more efficient than supercritical (40–42 percent), had no significant
operating track record at the time and had higher capital costs than supercritical technology. Integrated
gasification combined cycle also had no operational track record and was, therefore, considered risky.
Hence, supercritical technology was selected as the most viable option for both Medupi and Kusile power
plants.
9. However, despite the clear developmental impacts of the project, three critical issues had to be
considered in the World Bank’s decision to support the project: (a) the MCPP would burn coal, the biggest
emitter of greenhouse gas (GHG) emissions per unit of energy produced of all other electricity sources;
(b) implementation of most of the project components had already begun and that of the main
component, the MCPP had reached an advanced stage; and (c) the World Bank had no prior involvement
in the sector6 and therefore the Eskom Investment Support Project (EISP) would become the first large‐
scale project supported by the World Bank in South Africa and in its energy sector.
10. Coal‐fired power generation. The World Bank’s decision on whether to finance a coal‐fired power
plant had to be consistent with the institution’s commitment to supporting the global agenda on climate
change and with its energy sector policies. Although the World Bank adopted a definitive policy position
on financing coal in July 20137, after the EISP was approved in 2010, its commitment to promoting low
carbon investments was clear from earlier documents. In 2007, the World Bank produced the ‘Clean
Energy for Development Investment Framework: The World Bank Group Action Plan, March 28, 2007’.
This report included clean coal as a low carbon technology for power generation. The following year, the
World Bank produced a technical report, ‘Development and Climate Change Agenda: A Strategic
Framework for the World Bank Group (DCCSF)’. This report laid out strict criteria8 for the World Bank’s
support of coal‐fired power plants.
11. While the World Bank’s climate change policies for the energy sector were evolving, South Africa
had also taken serious steps to address climate change concerns, including by (a) developing a climate
change response strategy (2004) outlining an array of mitigation and adaptation measures; (b) committing
to implement economywide emissions targets by 2020 and 2025 under the Copenhagen Accord (2009);
(c) assigning, as part of its response to the energy crisis, high priority to energy efficiency, investments in
renewable energy, and the use of economic and regulatory instruments to stabilize GHG emissions in the
medium term and reduce them in the long term; and (d) approving at the Cabinet level (2008) a long‐term
mitigation scenario, which envisaged a gradual shift away from coal toward nuclear and renewable energy
with a view to ensuring that carbon emissions from all sources, including electricity generation, would
peak during 2020–25 and plateau for a decade before declining thereafter. The country had taken other
6 The World Bank’s only involvement was as an implementing agency of a small Global Environmental Facility (GEF) grant—
Renewable Energy Market Transformation.
7 “Towards a Sustainable Energy Future for All: Direction for the World Bank Group Energy Sector”. 2013
8 The criteria included demonstration of the developmental impact, including improving energy security, reducing power
shortages, or increasing access for the poor; provision of assistance for development of low carbon energy sources; optimization
of energy sources, including energy efficiency and conservation; non‐availability of funding support from donors to cover the
incremental costs of other viable technologies; and an approach for incorporating environmental externalities in project analysis.
Page 9 of 102
The World Bank
steps pointing to a strong commitment to low carbon growth such as the approval of renewable energy
feed‐in tariffs by the National Electricity Regulator of South Africa (NERSA) on October 29, 2009, to
facilitate private sector investment in clean energy, and the decision by Eskom to use supercritical or
ultrasupercritical technology for new coal power plants.
12. Based on the GoSA’s commitment to pursue a low carbon growth strategy for its energy sector, a
broader scope was designed to include renewable energy plants (comprising a pilot CSP plant and wind
power plant) and a renewable energy development and energy efficiency technical assistance (TA)
component.
13. The World Bank assessed the consistency of the project with the DCCSF report and concluded that
it met all the following criteria: (a) demonstrated developmental impact including improving overall
energy security and reducing power shortage or access for the poor; (b) provision of assistance to identify
and prepare low carbon projects; (c) optimization of energy sources by considering the possibility of
meeting the country’s needs through energy efficiency (both supply and demand) and conservation; (d)
full consideration of viable alternatives to the least‐cost (including environmental externalities) options
and when additional financing from donors for their incremental cost is not available; (e) design of the
MCPP to use the best appropriate available technology to allow for high efficiency and therefore lower
GHG emissions; and (d) incorporation of environmental externalities in project analysis.
14. An external panel of experts was appointed in October 2009 to review and advise the World Bank
on the application of the DCCSF criteria. The panel concluded that “South Africa is facing an immediate
shortage of electric power that has already crippled its economy. Hence, as a transition strategy in the
near term we accept that it is necessary to build additional coal‐fired electric power units. But this must
be coupled to a longer‐term strategic shift to an economy based upon a low carbon energy supply.”
15. In 2013, the World Bank published a new energy sector policy—'Towards a Sustainable Energy
Future for All: Directions for the World Bank’s Energy Sector’. The policy stated that the World Bank would
provide financial support for greenfield coal power generation projects only in rare circumstances and
cited, as examples, countries with no feasible alternatives to coal for meeting basic needs and lacking
financing for coal power generation. The policy endorsed the use of the criteria established by the DCCSF
used by the World Bank in assessing the consistency of the EISP with the World Bank’s policy position on
financing coal only in rare circumstances.
16. World Bank involvement in the EISP started after project implementation had already begun.
Implementation of the project started in December 2005 when Eskom’s Board of Directors approved a
procurement and contracting plan for the construction of the Medupi and Kusile power plants. There was
an urgency to move ahead swiftly with the projects given that power shortages were expected to occur
around 2007.
17. By 2009, when the World Bank started preparing the project, its implementation was well
advanced. The project design and implementation arrangements were decided; contracts for about 24
out of 38 packages for the MCPP, including for the 2 largest packages for the boilers and steam turbines,9
had already been awarded; procurement of other packages was in progress; and Eskom had prepared an
9 These two packages constituted more than 50 percent of the total estimated cost of the MCPP.
Page 10 of 102
The World Bank
Environmental and Social Impact Assessment and Environmental Management Plan and on September
21, 2006, received a conditional approval from the Department of Environmental Affairs and Tourism to
proceed with the construction and operation of the MCPP.
18. The World Bank undertook several due diligence reviews, including with the aid of external
independent third parties, regarding the acceptability of the project design and implementation
arrangements, procurement, and environmental and social safeguards for its financing support. This
specifically involved the following due diligence work:
(a) Project design and implementation arrangements. The World Bank engaged an
independent US consulting firm (the technical due diligence consultant), to advise if Eskom
had used good engineering practices as a basis for the design of the power plant; possessed
the engineering, procurement, and construction management capability to successfully
complete the project; and had the operation and maintenance (O&M) capability to safely
operate the plant while meeting the country’s regulatory environmental requirements. The
consultant’s review concluded that (i) the design of the plant was based on Eskom’s user
requirement specifications (URS) which met or exceeded performance, reliability, and
availability requirements of similar coal‐fired power plants in operation or under
construction at the time; (ii) overall plant design, steam cycle optimization, and plant layout
were consistent with current engineering design practices and that the MCPP’s performance
was expected to be similar to state‐of‐the‐art supercritical power plants elsewhere; and (iii)
Eskom had assembled a capable team to execute the construction of the MCCP. The
consultant noted, however, that supercritical power plants required well‐trained and skilled
staff and that the size of the client’s O&M staff was small relative to the size of the plant.
Because detailed engineering investigation of the boiler design was undertaken by the
contractors, investigation of its adequacy was the responsibility of Eskom’s project
management team with the support of their embedded technical engineer, PB Power of the
United Kingdom.
(b) Contracts awarded and in progress. In addition to conducting technical due diligence,
consultant assessed whether Eskom had followed good procurement practices, obtained
competitive prices, and complied with both its own policies and procedures and South
Africa’ regulations. The consultant reviewed only the main contracts for boilers and steam
turbines and indicated that the balance of the contracts awarded up to that point were
reviewed by Ernst & Young at the request of Eskom. Ernst & Young’s findings confirmed that
the processes used were fair and consistent with Eskom’s project guidelines. The World Bank
also engaged an individual expert to review all the awarded MCPP contracts that had already
been signed and were proposed for World Bank financing.10 Both the technical due diligence
consultant and the individual expert concluded that, overall, the procurement process was
transparent, fair, and consistent with Eskom’s policies and procedures and with the South
African regulatory requirements. The two most consequential deviations between the World
Bank’s procurement guidelines and the process followed by Eskom that were identified by
10 For the MCPP, the World Bank was asked to finance about US$1.8 billion of contracts that had already been awarded or
bidding was in progress. Outstanding procurement for only about US$0.6 billion was to be conducted following the World
Bank’s policies and procedures.
Page 11 of 102
The World Bank
the individual expert were (i) the absence of provisions on fraud and corruption and audit
rights in the contract documents and (ii) the requirement for public agencies, including
Eskom, to seek local content and skills development targets as key evaluation criteria in their
tenders. The requirement for use of local content and skills targets in contract bid
evaluations was in accordance with the Accelerated and Shared Growth Initiative (ASGI‐
SA).11 The technical due diligence consultant also benchmarked the plant costs and found
them (excluding flue gas desulphurization [FGD] unit) to be comparable to the US market
and to the lower end of the European market, but above the cost of plants constructed or
under construction in India and China, based on an average industry schedule of four years
from order to operation.12 Based on the overall positive findings by both consultants and
Eskom’s agreement to retrofit the contracts with fraud and corruption and audit rights
clauses, the World Bank Board granted the management’s request for a one‐time policy
exception to allow the financing of contracts which had been awarded following South
Africa’s procurement policies and procedures and Eskom to retain the local content and skills
development provisions in its tendering for the remaining contract packages.
(c) Environmental and social safeguards. On the safeguards side, most environmental impact
assessments and environmental management plans had similarly been completed when the
World Bank started preparation of the EISP. South Africa was previously selected to pilot the
use of country systems (UCS) for safeguards. A small project—Development, Empowerment,
and Conservation of the iSimangaliso Wetland Park and Surrounding Region Project funded
by the Global Environment Facility (approved in 2009)—was the first project to use country
systems for safeguards in South Africa. The World Bank conducted a Safeguards Diagnostic
Review (SDR), which confirmed the appropriateness of using country systems for the EISP.
19. First large‐scale project with new client and new operating environment. The third factor that
weighed heavily on the World Bank’s decision to support the project was the large scale of the project
with a new borrower and an unfamiliar operating environment. The World Bank, on one hand, and the
GoSA and Eskom, on the other hand, were not familiar with each other’s policies and procedures for
project implementation. The World Bank was unfamiliar with South Africa’s legal and regulatory
framework as it related to the energy sector and the application of the framework to environmental and
social safeguards and procurement. The World Bank had no knowledge and experience of Eskom’s project
management capabilities. Eskom had no recent experience with implementation of a project of the scale
of the EISP, its most recent power plant having been commissioned about 15 years before.
20. The World Bank’s analysis and due diligence carried out with the support of third parties
concluded that the EISP was consistent with the World Bank’s existing policies on climate change, Eskom
had the capability to implement and operate the MCPP (despite Eskom not having implemented a project
of the scale of the MCPP in more than a decade), the activities carried out to that point, including the
plant design, environmental and social safeguards, and procurement were conducted in accordance with
Eskom’s sound policies and procedures, South Africa’s regulatory framework, and good international
11 ASGI‐SA is an initiative and policy of the GoSA, initiated in 2004, to accelerate economic growth by 6 percent and halve poverty
and unemployment by 2014.
12 No similar plants had been or were under construction in Africa. Those in India and China were not representative because
they were based on local South East Asian suppliers.
Page 12 of 102
The World Bank
practices. The positive outcomes of the due diligence assessments were reinforced by the global ranking
of Eskom as a premier power utility organization at the time, and by the strong reputation that South
Africa’s legal and regulatory framework enjoyed globally.
21. Given the positive outcomes of the due diligence the Bank’s decision to support the EISP was
based on: (a) the EISP’s perceived strong developmental impacts; (b) the conclusion that coal was the only
feasible alternative to address the problem of power shortages in the short term, and the selection of
supercritical technology by the Eskom as the best available option; (c) the GoSA’s commitment to low
carbon development of the electricity sector as demonstrated by its agreement to include renewable
energy and energy efficiency in the project design; (d) the opportunity for the World Bank, through its
participation in the EISP, to strengthen the GoSA’s strategy for low carbon growth of the electricity sector
by both financing and leveraging financing by other development partners of renewable energy and
energy efficiency components, including coverage of incremental costs of CSP; and (e) the opportunity
for the World Bank to share knowledge and experience with Eskom and the GoSA on renewable energy,
thus facilitating its uptake.
22. In addition, as lender of last resort, the World Bank’s support to cover the large financing gap for
Eskom and the GoSA during a global financial crisis was considered critical. It was expected to signal to
other potential financiers the credibility of Eskom’s investment program. The project was also considered
a first critical step in establishing a long‐term engagement with an important development partner in the
region.
23. Having decided to support the EISP the World Bank committed substantial human and financial
resources throughout the project cycle from preparation through completion. This included the
involvement of senior World Bank management in consultations with officials at the highest levels of the
South African Government during the project preparation process and at various junctures during its
implementation, and the intensive work carried out by staff to ensure adequate due diligence during
preparation and the implementation support provided until project completion.
24. The World Bank Group’s Country Partnership Strategy (CPS) for South Africa was prepared in 2008
for FY08–12. Although the CPS focused primarily on knowledge sharing and technical support, it was
designed to flexibly respond to emerging demand‐driven client needs. This was the case with the EISP,
which was incorporated into the World Bank’s support strategy through the CPS Progress Report prepared
in parallel with the project and issued in FY10. 13
Theory of Change (Results Chain)
25. The project was aimed at addressing the problem of inadequate generation capacity causing
electricity shortages which, in turn, were reducing economic growth and employment and slowing down
implementation of the rural electrification program. This program was instrumental in addressing the
problem of inequality and non‐inclusive growth through the creation of economic opportunities and jobs
and the delivery of better educational and health outcomes. In addition, the project was to support the
objective of reducing South Africa’s high carbon intensity of energy production by (a) adding 4,800 MW
13 The Project Concept Note was reviewed by a World Bank‐wide Operations Committee meeting on September 21, 2009.
Page 13 of 102
The World Bank
of generation capacity based on more efficient supercritical technology14 compared to conventional
subcritical technology, (b) implementing two renewable energy plants (Sere wind power plant and
Upington CSP), (c) undertaking demand management measures and improving the efficiency of existing
coal plants, and (d) enabling the development of additional renewable energy capacity by the private
sector.15
26. The results matrix (PAD, annex 3) provides data on the process through which project inputs were
to be converted into outputs and translated into intermediate and outcome indicators, with the latter
being set at the level of the desired outcomes. From these data the Implementation Completion Report
(ICR) has derived a logical framework or theory of change mapping three groups of activities (construction
of the MCPP, construction of renewable energy power plants, and implementation of renewable and
energy efficiency activities) to their relevant outputs, outcome indicators, Project Development
Objectives (PDOs), and long‐term goals as shown in figure 1.
27. The theory of change shows that successful implementation of project activities was expected to
provide the following outcomes: increased supply and security of energy provided efficiently and in a
sustainable manner and contributions to economic growth and the country’s low carbon mitigation
strategy. Several critical risks were identified, and risk management measures were developed (PAD, page
41). The main risks and the corresponding management measures were as follows:
(a) The Borrower’s lack of familiarity with the World Bank processes and policies. Because this
was the first Investment Financing Project to South Africa and Eskom after a long hiatus,
there was concern about the World Bank’s processes and procedures, especially regarding
loan conditionalities. In addition to the due diligence undertaken at appraisal, the risk was
to be managed through country dialogue around the CPS engagement, open knowledge
sharing on World Bank policies and procedures, and timely quality responses to client needs.
(b) Reputational risk to the World Bank for supporting a coal‐fired power plant given its public
commitment to the global climate agenda. To mitigate this risk, a communication strategy
was designed targeted at stakeholders in South Africa and the international community with
clear messages about (i) the developmental impact of the project for both South Africa and
the subregion; (ii) South Africa’s commitment to low carbon development as evidenced by
its international commitments under the Kyoto Protocol and Copenhagen Accord and its
own long‐term mitigation scenario ; (iii) Eskom’s choice of supercritical technology for the
MCPP; (iv) the inclusion of significant low carbon renewable energy investments and energy
efficiency as integral parts of the EISP; (v) Eskom’s commitment, included as a covenant in
the Loan Agreement, to install the FGD units at the MCCP to ensure compliance with South
Africa’s emission standards; and (vi) approval of feed‐in tariffs for renewable energy in the
efforts to mobilize private sector investment in renewable energy.
(c) Adequacy of measures to mitigate SO2 emissions from the MCPP to ensure consistency
with the World Bank’s environmental health and safety guidelines for new thermal power
plants and South Africa’s proposed emissions standards for new plants. The proposed
14 The MCPP was expected to discharge less carbon emissions compared to subcritical technology which was the dominant
technology in Eskom’s portfolio of coal‐fired power plants.
15 This was an acknowledged contribution of the project but was not part of the Results Framework.
Page 14 of 102
The World Bank
mitigation measure was for Eskom to retrofit each unit of the MCPP with an FGD unit. A
schedule for commissioning of the FGD units was included as a covenant in the Loan
Agreement.
(d) Reduced impact on low carbon development that could result from delays in attracting
private sector investment in renewables. The project was expected to play a catalytic role
in attracting these investments through the demonstrative effects of its renewable power
plants and the provision of associated transmission infrastructure.
(e) Sensitivity around the UCS on the first large investment in the country. To manage the risk,
the World Bank reviewed South Africa’s legal and regulatory framework and Eskom’s policies
and procedures for safeguards. The World Bank found that South Africa and Eskom had a
robust system in place for management of environmental and social safeguards, except for
a few areas that are described in section III B (Environmental, Social, and Fiduciary
Compliance), including the related risk management measures.
(f) Reputational risk related to procurement. Allegations of fraud and corruption could arise in
relation to contracts awarded before the World Bank’s participation in the project. If such
cases arose, the World Bank reserved the right to refuse to finance such contracts and cancel
the related amounts from the loan funds or obtain reimbursement if the disbursements had
already been made on the contracts. In fact, the previous contact awarded to Hitachi (before
the World Bank got involved in the project), who then subcontracted the fabrication of boiler
pressure parts to DB Thermal Company, did result in a fraud case after the World Bank had
started participating in the project and resulted in large additional costs and schedule impact
on the first three units.
28. The theory of change assumed that the above risks would be effectively managed according to
the identified risk management measures. However, some of the risk mitigation measures, such as the
FGD installation for sulfur dioxide emissions, were not as effective or implemented as timely as
anticipated.
Page 15 of 102
The World Bank
Figure 1. Theory of Change at Appraisal
ACTIVITIES OUTPUTS OUTCOME INDICATORS PDO

PDO L‐T GOALS
CONSTRUCT  4, 764 MW INSTALLED MW

MEDUPI POWER MEDUPI PLANT CAPACITY AS A %  ENHANCE
PLANT  2,224 KM OF PEAK DEMAND POWER SUPPLY
TRANS. LINES AND ENERGY
SECURITY
 Energy supply from  EFFICIENTLY AND
ENERGY SUPPORT L‐T
BUILD RE PLANTS RE MW CAPACITY SUSTAINABLY
 SERE WIND SUPPLIED FROM  SUPPORT LOW CARBON
 CCCons • 100 MW WIND
 UPINGTON CSP RE CARBON MITIGATION
• 200 MW CSP
MITIGATION
STRATEGY
STRATEGY
 SUPPORT
ENERGY 64 KM MAJUBA CO2 EMISSIONS ECONOMIC
RAIL SIDING AND DISCHARGED GROWTH
EFFICIENCY & TA
COAL PLANTS TA YARD (KG/KWH) OF
RE & EE TA RE & EE TA ELECTRICITY
PLANNED ACTIVITIES INTENDED RESULTS
Project Development Objectives (PDOs)
29. As stated in the Loan Agreement: “The objective of the Project is enabling the Borrower to
enhance its power supply and energy security in an efficient and sustainable manner so as to support both
economic growth objectives and the long‐term carbon mitigation strategy of the Guarantor.” 16 In the
PAD, it is stated as “The Project Development Objective (PDO) is to enable Eskom Holdings to enhance
power supply and energy security in an efficient and sustainable manner so as to support both economic
growth objectives and the long‐term carbon mitigation strategy of South Africa.”
30. The PDO had four parts as follows: (a) to enhance its power supply and energy security; (b) in an
efficient and sustainable manner; (c) to support [both] economic growth objectives; and (d) the long‐term
carbon mitigation strategy of the Guarantor.
31. The first and second objectives (enhancing power supply and energy security in an efficient and
sustainable manner) were at project level; the third, a long‐term high‐level objective; and the fourth, both
a project‐level and long‐term objective. These objectives were to be achieved by (a) investing in
supercritical coal‐fired and renewable energy power plants to increase system generation capacity (MW)
by about 12.5 percent, (b) increasing the amount of coal transported by more efficient rail than road, and
(c) implementing both demand‐ and supply‐side efficiency measures.
16 There are slight differences in the wording of the PDO between the PAD and the Loan Agreement, but these are more semantic
than substantive.
Page 16 of 102
The World Bank
32. The PAD did not explain the meaning of the second objective (…in an efficient and sustainable
manner…) and how it was to be achieved. By reviewing various parts of the PAD (including figure 6, page
28), the ICR has interpreted the ‘efficient’ part to refer to (a) the project’s contribution to sector efficiency
using efficient supercritical coal‐fired power plant technology at the MCPP, (b) the partial shift of coal
transportation from road to rail which was expected to result in reduced unit transportation costs (US$
per ton of coal transported), (c) the TA provided to support Eskom’s efforts to reduce the heat rate of the
existing coal fleet by 1 percent on average, and (d) the TA to support Eskom in scaling up the solar water
heater and time of day metering. Arguably, ‘efficiency’ could have been defined to also include an efficient
project schedule of four years to completion consistent with power sector industry standards. This aspect
is, however, captured in the efficiency evaluation through the economic rate of return (ERR) analysis and
is not dealt with through the evaluation rating of efficacy.
33. The ‘sustainable’ part of the objective is assumed to refer to (a) the increased use of domestic
renewable energy sources—from both the wind and solar power plants which were part of the project
and from other renewable energy investments enabled by the project (b) the environmental benefits and
the reduction in carbon emissions that were expected to result from the partial shift from road to rail coal
transportation for the Majuba Power Plant and (c) the slowing down of Eskom’s carbon emission due to
the use of supercritical compared to subcritical technology.17
34. Support for carbon mitigation strategy. The causal relationship between the operation of
renewable energy plants constructed under the project and the carbon mitigation objective through
reductions in carbon emission was clear and direct and was quantified and so was the effect of the partial
shift of coal transportation from road to rail. Improvements in efficiency upgrades of the existing coal‐
fired power plants were also to contribute to emissions reductions. The ICR interprets the project as
contributing to both project‐specific impacts and the country’s high‐level and long‐term low carbon
mitigation strategy.
35. Support for economic growth objectives. Key macroeconomic variables such as growth rates,
fiscal balances, and employment creation had shown a close association to the level of energy provision.
The projects aimed to improve the provision of power for the economy to support viability of the mining;
manufacturing; and micro, small, and medium enterprises (MSMEs) sectors that are the engines for
economic growth and employment creation. In 2008, the rate of increase in access to electricity (critical
for employment creation in informal sectors) slowed down due to power shortages. However, the
objective of supporting economic growth, a high‐level objective was not specified in the project at
appraisal.
Key Expected Outcomes and Outcome Indicators
36. At appraisal, the outcome indicators were specified for two of the four expected outcomes as
follows: (a) for enhanced security of supply ‐ installed capacity as a percentage of peak demand and the
GWh (amount) of energy generated from renewable energy sources and (b) for contribution to carbon
mitigation ‐ carbon emissions discharged per unit of electricity (per kWh). No indicators were specified
for the project’s expected contribution to the outcomes on efficient and sustainable supply and on
17 Since the conclusion reached was that coal was the only alternative for developing the required generation capacity quickly
supercritical technology was superior to subcritical technology from an efficiency and environmental perspective.
Page 17 of 102
The World Bank
support to economic growth objectives. Table 1 presents the indicators and associated targets as at
appraisal and after project restructuring.
Table 1: Expected Outcomes, Indicators, and Targets at Appraisal and after Restructuring
At Appraisal As Restructured
PDO Outcome Indicator Target Outcome Indicator Target
Enhanced power Installed capacity Installed capacity at Generation capacity 4,800 MW
supply and energy (MW) as a percentage 124.6 percent of installed and
security of peak demand peak demand (2015) commissioned at
Medupi (MW)
Energy supply from 520 GWh per annum Generation capacity 100 MW
renewable energy (2015) installed and
(GWh) commissioned from RE
(MW)
Enhanced power No indicators or targets were set at appraisal or during restructuring
supply and energy
security in an
efficient and
sustainable manner
Support to low Carbon emissions Carbon emissions Direct CO2 emissions 238,000
carbon mitigation discharged per unit of discharged per unit avoided under the MT
strategy electricity (kg/kWh) of electricity ‐ 0.950 project (in metric tons)
kg/kWh (2015)
Support to economic
growth objectives of No indicators or targets were set at appraisal or during restructuring
the Government
Note: RE = Renewable energy.
Components
37. The project had three components: (a) Component A for construction of the 6 x 800 MW Medupi
Coal‐fired Power Plant and associated transmission lines, (b) Component B for investments in renewable
energy power plants, and (c) Component C for investments and TA for in energy efficiency and renewable
energy development.
38. Total project costs were estimated at US$13.86 billion18 and the actual costs were US$18.1 billion.
Detailed project costs and financing plans as at appraisal and at loan closure are provided in annex 3. The
following provides a summary of the project cost breakdown by component and IBRD financing.
(a) MCPP (Component A) (Estimated cost19 US$12,048 million, actual cost US$17,474 million;
estimated IBRD Loan US$3,040 million, actual IBRD disbursed US$2,866 million). This
component comprised a 6 x 794 MW coal‐fired power plant based on supercritical
technology.
18 Total estimated project costs included US$10 million to finance the front‐end fee on the IBRD Loan (about US$9 million) and a
Clean Technology Fund (CTF) management fee (about US$1 million). These costs are not included in project component costs.
19 The total cost estimates were inclusive of development costs, contingencies, and financing costs.
Page 18 of 102
The World Bank
(b) Renewable Energy Power Plants (Component B) (Estimated cost20 US$1,227 million, actual
cost US$252 million; estimated IBRD Loan US$260 million; actual IBRD Loan US$23 million;
estimated CTF US$350 million21, actual CTF US$77 million22). The component comprised a
100 MW Sere wind power plant and a 100 MW CSP plant at Upington. The Sere wind power
plant was prepared as a first phase of a two‐phase 200 MW project in the Western Cape
Province. The first phase under this project was expected to generate about 219 GWh when
completed. The Upington Power Plant was to be implemented on a design, construct, and
commission basis in the Northern Province. It was to be implemented on a pilot basis to
provide benchmarks for cost and performance on a utility scale basis in a region that has
abundant soler resources.
(c) Energy Efficiency Investments (Component C) (Estimated cost US$576 million, actual cost
US$402 million; estimated IBRD financing US$441 million; actual financing IBRD Loan
US$270 million). The component consisted of both sector investments and TA. Included in
this component were three subcomponents. The first was construction of the Majuba Rail,
which would be more cost‐efficient and environmentally better than road for coal
transportation to the Majuba Power Plant. The project also included a new rail yard layout
for faster coal off‐loading. The second subcomponent was TA for assessing opportunities for
coal‐fired power plant efficiency improvements to support Eskom’s objective of reducing
the average heat rate of its fleet by 1 percent by 2012. The third subcomponent provided TA
to support implementation of the Upington CSP and provide technical, financial, and legal
advisory services to Eskom to develop domestic or cross‐border renewable projects.
39. The financing plan at project appraisal included US$350 million in loans from the Clean
Technology Fund (CTF) of which US$250 million was to be channeled through the World Bank and US$100
million through the African Development Bank. The Grant Agreement (TF 10690) for the US$250 million
was later linked to the Eskom Renewables Support Project (P122329) approved by the Bank on October
27, 2011, and the financing was used for activities under both the EISP and the ERSP. About US$ 34.9
million was disbursed under the EISP by loan closing date.
40. Disbursement of the IBRD Loan totaled US$3,159 million compared to the committed amount of
US$3,750 million. The difference of US$591 million was canceled. The first cancellation of US$320 million
was made after the June 2020 project restructuring indicated that the loan would not be fully disbursed
by loan closure, and the final cancelation of the balance of US$271 million was made on November 8,
2021, after the expiration of the disbursement period. The breakdown of the total undisbursed funds
compared to appraisal estimates was as follows: MCPP (Component A) ‐ US$183 million, Renewable
Energy (Component B) ‐ US$237 million, and Energy Efficiency Investments (Component C) ‐ US$171
million.
20 The costs for the component were inclusive of development costs and associated transmission lines required to connect the
plants to the Eskom grid and to allow other renewable projects in the area to be connected as well.
21 US$249 million of the CTF was allocated to the CSP component and later to the Battery Energy Storage System (BESS)
subcomponents, which replaced CSP during the 2018 project restructuring, but the BESS could not be implemented during the
life of this project.
22 Of which US$35 million was channeled through the World Bank and US$42 million through the African Development Bank.
Page 19 of 102
The World Bank
B. SIGNIFICANT CHANGES DURING IMPLEMENTATION (IF APPLICABLE)
Revised PDOs and Outcome Targets23
41. The project underwent five restructurings, all of which took place at Level 2, that is, they were
approved by the Regional Vice‐President or Country Director. None of the restructurings involved
revisions to the PDOs, but three of the restructurings on July 7, 2015, December 7, 2018, and December
27, 2019, included changes to the outcome indicators, target values, and due dates. Table 2 provides a
summary of the project restructurings and their rationale.
Table 2. Summary of Project Restructurings
Date Restructuring Rationale for Restructuring
Paper Document
December 11, RES17080 Rationalize and simplify disbursement categories, reallocate funds
2014 among categories, and update disbursement schedule
July 7, 2015 RES19126 Rationale was to update several parameters in view of huge actual
delays foreseen to completion. The updates included a closing date
extension of about 50 months and revisions to the implementation
schedule, project costs and financing plan, project outcome indicators
and targets, disbursement schedules, loan allocation among
disbursement categories, economic analysis, and legal covenants.
December 7, 2018 RES32580 The reasons for the restructuring were to replace CSP with the battery
energy storage system (BESS) and allow more time for
implementation of interim measures to address SO2 emissions
exceedances before installation of the FGD units at the MCPP. The
Results Framework was amended to include the MW storage capacity
of the BESS as an intermediate outcome indicator in place of the
former MW capacity indicator of CSP.
December 27, RES39035 The rationale for this restructuring was to extend the loan closing
2019 date by six months to allow completion of the components that had
fallen behind (the transmission lines for the MCPP, the Majuba Rail,
and the BESS component) and support Eskom’s restructuring road
map. The changes included revisions of target end dates to the new
loan closing date.
June 25, 2020 RES42516 This restructuring was to cancel US$320 million24 that was projected
to remain unused by the loan closing date. The amount arose from
projected savings on the transmission lines and the Majuba Rail
components and US$100 million previously allocated for the
restructuring of Eskom, which was no longer feasible to complete
within the remaining time before loan closure.
23 A split evaluation was not carried out because (a) there was no change in project objectives; (b) the scope and ambition of the
operation did not change; and (c) according to the Project Restructuring Paper dated July 7, 2015, the change in the indicators
and related targets was done to “clarify the PDO outcomes” or to “better capture the project’s achievements.”
24 A further US$270 million of undisbursed funds was canceled on November 8, 2021, following expiration on October 31, 2021,
of the grace period for disbursements.
Page 20 of 102
The World Bank
Revised PDO Indicators
42. During the project restructuring on July 7, 2015, both indicators for enhanced power supply and
energy security were revised from (a) installed capacity (MW) as a percentage of peak demand to
generation capacity installed and commissioned at Medupi (MW) and (b) energy supply from renewable
energy (GWh) to generation capacity installed and commissioned from RE (MW). The indicator for carbon
mitigation was revised from carbon emissions discharged per unit of electricity (kg/kWh) to direct CO2
emissions avoided under the project in metric tons. The revisions to the PDO indicators are summarized
in table 1. According to the Restructuring Paper of July 7, 2015, the changes were made to “clarify the
PDO outcome indicators” (page 5) or “to better capture project achievements” (page 47). The
intermediate indicators related to the MCPP and renewable energy25 that had originally been stated as
added generation capacity were revised and restated in terms of implementation progress (for example,
percentage of construction completed). During the next project restructuring on December 7, 2018, the
target for the outcome indicator associated with the renewable energy plants was further revised to drop
the contribution from CSP, which was being replaced by the BESS. The third restructuring, which affected
the Results Framework, was undertaken on December 27, 2019, and aligned the target dates for achieving
indicator target values with the loan closing date.
Revised Components
43. There were a few but significant changes to the project components. The main change was the
replacement, during the project restructuring on December 7, 2018, of the Kiwano26 CSP plant with a grid‐
scale Battery Storage Program (BSP)—later referred to as the battery energy storage system (BESS).
44. The decision to substitute the CSP plant with the BESS was taken after long procurement delays27
and after Eskom received non‐responsive bids for CSP. CSP was conceived as a learning pilot when there
was no private sector involvement in renewable energy in South Africa. The World Bank, by packaging
renewable energy as an integral part of the EISP and providing technical support to the GoSA, facilitated
a substantial uptake of renewable energy, especially under the successful Renewable Energy Independent
Power Producer Programme (REIPPP).28 Thus, by 2018, the private sector had taken interest in CSP
technology and was about to launch a CSP of its own.29 There was, therefore, no compelling reason for
Eskom to implement a CSP anymore. The BESS, on which there was no private sector initiative, was a more
interesting alternative to CSP as it was expected to provide energy storage capacity to facilitate the
integration of other renewable energy plants, especially private sector plants to the grid. The BESS was to
be implemented in two phases with the first phase being funded by the IBRD Loan under this project.
25 The added generation storage capacity expected from the BESS was included at the level of an intermediate indicator.
26 The original name at appraisal was Upington but was later renamed.
27 There were multiple reasons for the delayed procurement of CSP, including (a) long time taken for the financing partners to
agree on a procurement strategy and for all the partners to agree with Eskom on the technology to be adopted and (b) almost
one‐year delay in getting the National Treasury’s decision on a waiver of the Preferential Public Procurement Policy, and so on.
28 By the end of 2018, about 4,000 MW had been installed under the REIPPP in South Africa (Restructuring Paper, Report No.
RES32580 of December 7, 2018). This has increased to about 6,000 MW as of mid‐2021.
29 Around the time the EISP was prepared, the World Bank had started piloting CSP in other countries, including in Morocco where
the project was structured as a public‐private sector operation and turned out to be successful.
Page 21 of 102
The World Bank
Other financing was to be provided by concessional loans under the World Bank’s Eskom Renewables

Support Project (ERSP) (P122329), approved October 27, 2011; CTF; and AfDB.
45. During the December 27, 2019, restructuring, a further change was made to the description of
the investments in renewable energy and energy efficiency component to incorporate support to Eskom
in implementing unbundling reforms. A total of US$100 million was reallocated for that activity. However,
the allocation was canceled during the last project restructuring on June 25, 2020, because, with only one
year left before loan closing and without any specific proposal for implementation, there was no realistic
prospect that the funding would be used.
46. Component A for the MCPP also included about 2,244 km of 765 kV and/or 400 kV transmission
lines (and associated substations) to evacuate power from the power plant to the electricity grid. The
scope of the transmission lines activities was designed with broad network flows and stability
considerations in mind and was not limited to the needs for integrating the MCPP. The scope was later
revised due to internal cash constraints within Eskom to a level necessary to effectively integrate the
MCPP to the grid while maintaining technical requirements according to the grid code. Hence, the lengths
of the lines were revised to about 1,020 km during the 2019 project restructuring. The revised lengths of
the lines were included in the Results Framework as an intermediate indicator.
Other Changes
None
Rationale for Changes and Their Implication on the Original Theory of Change
47. The theory of change remained substantively the same after project restructurings as it was at
appraisal, but all three outcome indicators were revised as shown in table 1 and discussed above. The ICR
supplements the efficacy evaluation (section II B) by using alternative indicators on the project’s
contribution to the system energy generation.
II. OUTCOME
A. RELEVANCE OF PDOs
48. As described at the beginning of the ICR, the World Bank’s preparation of the EISP started in 2009
following a request by the GoSA for financing support for urgent implementation of the MCPP. The project
was the most feasible option in the short term for addressing a severe energy crisis, which was adversely
affecting economic growth and employment and posed a serious threat to economic recovery after the
2007/08 global financial crisis.
49. The EISP was prepared during the period of the FY08–12 World Bank CPS for South Africa. The
CPS had the dual objectives of supporting the Government’s national growth and development programs
and collaborating with regional partners on key regional development issues. The World Bank’s support
strategy was organized under two pillars for (a) Urban and Rural Development focusing on urban and
municipal development, land reform, agriculture, private sector development, environment, and
Page 22 of 102
The World Bank
infrastructure and (b) Regional Integration through outward investment, regional communities, and
knowledge sharing. Improved service delivery was a cross‐cutting theme.
50. The FY08–12 CPS was prepared before the 2008/09 global financial crisis at a time when South
Africa had easy access to finance and as a result consisted entirely of demand‐driven advisory services
and analytics and TA. The EISP was, therefore, not part of the lending program under the original CPS.
When the power crisis emerged, the World Bank assessed that addressing power shortages was critical
for South Africa’s economic recovery and the whole Southern Africa subregion. Ending power shortages
was also essential for broader developmental outcomes as increased access to reliable and affordable
electricity would stimulate entrepreneurship and job creation in the informal sector. Thus, the PDO was
supportive of the CPS’s overall objective of supporting national growth and development programs. A CPS
Progress Report, dated March 1, 2010, was prepared in parallel with the EISP. The progress report30
included and described, in some detail, the EISP activities and objectives under the Urban and Rural
Development pillar’s infrastructure sub‐objective. The results matrix (Progress Report, page 26) stated
the issues to be addressed by the EISP as the backlog of unmet investment needs, which had precipitated
an energy crisis and the high costs of infrastructure due to the international financial crisis. The outcomes
to be delivered by the EISP were identified as increased power generation capacity and a gradual shift to
a low carbon trajectory.
51. The World Bank’s subsequent CPS for FY14–17 reaffirmed the importance of the then ongoing
EISP stating explicitly that the “IBRD strategy will be centered on knowledge and technical cooperation….
the implementation of the ongoing lending program in energy and the environment.” The CPS was
anchored in the Government’s National Development Plan’s objectives of eliminating poverty, reducing
inequality, and improving job creation. The CPS’s particular focus was organized around three
engagement areas to (a) promote increased competition and improved business environment for
sustainable growth, (b) strengthen the performance of MSMEs and skills development to support job
creation, and (c) improve the infrastructure investment framework and selected infrastructure services.
Under the three pillars were included eight program areas, including energy which fell under the pillar of
promoting investments.
52. The FY14–17 CPS further confirmed, as a continuing priority, support to South Africa’s
collaboration and cooperation with partners on the Southern Africa Power Pool (SAPP) and the planned
energy schemes such as the Inga Hydropower Program in the Democratic Republic of Congo and the Kudu
gas‐based power generation in Namibia, among others. These were all highlighted as strategic priorities
for the SAPP countries in the EISP PAD. The CPS results matrix confirmed (pages 29 and 30) the energy
sector developmental issues addressed by the EISP, and the CPS outcomes as described above and in the
PAD.
53. The FY14‐17 CPS was revised in the Performance and Learning Review of November 2016 at which
time it was also extended by one year through FY18. There was a three‐year gap before the current
country program (FY2022‐26) was approved on July 22, 2021. The first two years gap was due to the need
to align the policy dialogue with the electoral cycle of the May 2019 national elections and to build
consensus with the new leadership. The additional one‐year gap arose due to the Covid ‐19 pandemic.
30 The CPS was used flexibly to allow program priorities to be shaped through an annual business planning exercise led by the
GoSA.
Page 23 of 102
The World Bank
54. The current CPF is structured around three strategic focus areas to (a) promote increased
competition and improved business environment for sustainable growth, (b) strengthen MSME and skills
development to support job creation, and (c) improve the infrastructure investment framework and
selected infrastructure services. Energy is covered through the first and third focus areas as follows:
(a) Objective 1.2: Greater climate change resilience and environmentally sustainable
investments in selected sectors. The emphasis is on supporting South Africa in
implementing its Integrated Resource Plan (IRP) (2019) to achieve a secure and sustainable
energy mix by decommissioning and/or repurposing Eskom’s old coal‐fired power plants,
installing battery storage facilities to support grid integration of renewables, and increasing
the share of renewables in the electricity generation mix.
(b) Objective 3.2: Improved infrastructure services by selected state‐owned enterprises. This
could include advisory services for the design and implementation of sector reforms;
financing of initiatives selected by the GoSA; and targeted support to Eskom on topics such
as unbundling, debt management, and technical, operational, and financial deficiencies,
which contribute to power shortages.
55. The CPF’s represents a continuation of the focus on enhancing energy supply and security, albeit
with more emphasis on climate change‐related initiatives and sector reforms and reduction of
institutional, operational, and financial deficiencies faced by Eskom. The emphasis on renewable energy
and decommissioning/repurposing of Eskom’s old coal‐fired power plants represents progress in the
manner in which South Africa plans to meet future energy needs with clean energy based on the maturing
of the renewable energy technologies and reductions in costs that make them competitive with fossil fuel
fired power generation. Energy security remains a key objective of the GoSA as demonstrated by its
inclusion as one of the eight priorities in the GoSA’s Economic Reconstruction and Recovery Plan
announced on October 15, 2020.
56. The overall assessment of relevance of the PDO is based on three considerations: (a) the
alignment of the PDOs with the World Bank’s CPS/CPF for South Africa at project closure; (b) the country
context for the project, that is, whether the project’s objectives were outcome oriented and appropriately
pitched for the development status and capacity of the country; and (c) the World Bank’s historical
experience in the country and sector.
57. Regarding alignment, the project, on one hand, and all three strategy instruments (FY08–FY,
FY14–17, and FY22–26), on the other hand, identified the problem of power shortages and high carbon
intensity of electricity generation as significant developmental issues. The project activities and objectives
are specifically described in CPS FY08–12 (as updated by the Progress Report in 2010) and CPS FY14–17.
58. The World Bank had no historical experience in the energy sector in South Africa31 to provide a
reference point for setting the operation’s objectives. However, due diligence carried out at appraisal by
31 Except for a small GEF grant implemented by the World Bank as executing agent for GEF—Renewable Energy Market
Transformation (US$6 million).
Page 24 of 102
The World Bank
the World Bank team and by third parties indicated that the objectives were set at the right level and
were within the capacity of Eskom to achieve, although the latter turned out to have been an inaccurate
assessment during project implementation. Further, the objectives—increasing power supply and energy
security, supporting economic objectives, and the GoSA’s carbon emissions mitigation strategy—were
outcome oriented, although there were some deficiencies in the selection of indicators.
59. Based on these factors the relevance of the PDOs is Substantial. This is because there is strong
alignment of the EISP objectives with both the CPF at project closure (FY14‐17, extended to FY2018) and
the FY22–26 CPF which came into effect three weeks after the loan closed.
B. ACHIEVEMENT OF PDOs (EFFICACY)
Assessment of Achievement of Each Objective/Outcome
Objectives 1 and 2: Enhancing power supply and energy security in an efficient and sustainable
manner
60. Objectives 1 and 2 are combined for evaluation purposes because they share a common results
chain, that is, the same activities (the MCPP, renewable energy, TA for development of renewable energy
and energy efficiency, and the Majuba Rail components) were expected to support achievement of the
two objectives.
61. The assessment of efficacy is based on the outcome indicators and targets as formally revised
during the project restructuring in 2015 and 2018 because the revisions did not change the scope or
ambition of the project to warrant a split evaluation. In addition, the evaluation incorporates other
metrics that strengthen the assessment of the project’s contribution to enhanced power supply and
energy security beyond the installation and commissioning of additional generation capacity. These
include energy availability factors (EAFs) and energy generated by the power plants completed under the
project and capacity savings accruing from demand‐side management measures.
The MCPP’s contribution to power supply and energy security
62. The revised outcome indicators and targets for these two objectives were generation capacity
installed and commissioned at Medupi (target of 4,800 MW) and generation capacity installed and
commissioned from Sere wind power plant (target of 100 MW). The total additional generation capacity
target of 4,900 MW (4,800 MW from the MCPP and 100 MW from the Sere wind power plant) was
installed and commissioned.
63. The completion and commissioning of the MCPP units was considerably delayed. The MCPP was
implemented over 11 years (2010–2021)32 compared to the estimated implementation period of five
years at appraisal. Table 3 compares the estimated commissioning dates at appraisal to the actual dates.
32 The last unit to enter commercial operation (Unit #1) was commissioned on July 31, 2021, one month after loan closure. One
section of the associated transmission lines will only be completed by 2022, resulting in an overall implementation period of 12
years for the project.
Page 25 of 102
The World Bank
Table 3. Estimated and Actual Commissioning Years for the MCPP
Capacity Commissioned in MW33
Unit # 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 Total
According to the PAD

1 – – – 800 – – – – – – 800
2 – – 800 – – – – – – – 800
3 – – 800 – – – – – – – 800
4 – 800 – – – – – – – – 800
5 – 800 – – – – – – – – 800
6 800 – – – – – – – – – 800
Total according to the PAD 800 1600 1,600 800 0 0 0 0 0 0 4,800
Actual

1 – – – – – – – – – 794 794
2 – – – – – – – 794 – – 794
3 – – – – – – – 794 – – 794
4 – – – – – 794 – – – – 794
5 – – – – – 794 – – – – 794
6 – – – 794 – – – – – – 794
Total actual – – – 794 – 1,588 – 1,588 – 794 4,764
64. The first and last unit (Unit #6 and Unit #1) to enter commercial operation did so three and six
years later than estimated at appraisal, respectively. The key reasons for the delays were plant design
shortcomings (including basic and latent defects), poor performance by contractors, project management
constraints on the part of Eskom given the scale of the project, slow decision‐making by the GoSA
agencies, disruptive labor, community disputes, and the impact of the COVID‐19 pandemic in 2020. These
are discussed in detail in section III (Factors Affecting Project Implementation) and their impacts through
increased costs and delayed revenues are captured in the economic analysis (section II C).
65. Not only did the project experience long delays, but serious design and construction defects were
encountered. These and subsequent O&M inefficiencies and lack of spare parts affected the performance
of the plant resulting in low plant reliability and outputs. The primary defects were first noticed after
commissioning of the first unit (Unit #6) in 2015. The main problems were with the boiler components—
too small furnaces, air heaters, pulverizers, gas side erosion, and pulse‐jet fabric filters (PJFF)—all of which
had the effect of reducing energy availability and increasing unplanned capacity loss factors (UCLFs).
66. A program of remedying the defects was instituted, which comprised agreement on design
changes between Eskom and the contractor followed by optimization, testing, and implementation of the
changes on a unit‐by‐unit basis, evaluation of the outcomes, and eventual rollout to all units, if the
changes were successful. In the case of unsuccessful design changes, the options were to rework the
design changes and repeat the cycle with the contractor or for Eskom to do it alone or with another
contractor. A cost sharing arrangement was worked out between Eskom and the boiler original equipment
manufacturer (Mitsubishi Hitachi Power System ‐ MHPS and later Mitsubishi Heavy Industries ‐MHI). The
33 The estimate of 800 MW per unit at appraisal was the nominal gross output, whereas the actual capacity added of 794 MW
per unit was the final contracted gross output based on the contractors’ detailed heat rate calculations. Therefore, the difference
between the two figures does not indicate a deviation between expected and actual output.
Page 26 of 102
The World Bank
process was first applied to Unit # 3 in 2020 and the successful design changes were rolled out the same
year to Units 1, 2, 4, and 6 and to Unit #5 in 2021. Implementation of remedial measures has meant taking
the units out of operation for 75–90 days at a time. Implementation of the plan agreed between Eskom
and the contractor is continuing and according to Eskom it will take four to five years (or longer) to cure
the remaining latent defects.
67. The combined impact of the latent defects on the downtime of the equipment for implementation
of curative measures and the unplanned downtimes due to sudden plant failures or operational
inefficiencies34 (arising from shortages of spare parts and skilled staff) have resulted in high planned and
unplanned capability loss factors and low EAFs, as shown in table 4.
Table 4. Energy Availability and Capacity Loss Factors for the MCPP35
2016 2017 2018 2019 2020 2021
Energy availability factor (EAF) 59 76 77 60 65 58
Planned capacity loss factor (PCLF) 24 18 13 11 5 18
Unplanned capacity loss factor UCLF) 17 7 10 29 30 24
Source: EISP Closing Report, page 31.
68. The plant’s EAFs have varied between 58 percent and 77 percent with an overall declining trend.
The EAFs compare to an international standard of 92 percent for similar power plants, which is also
Eskom’s own URS for the plant. Even if the URS target is adjusted for defects (5.4 percent) and tube failures
(1.7 percent), according to Eskom’s own calculations, the expected EAF of 85 percent would still be higher
than the current EAF of about 58 percent in 2021. In addition, the plant’s EAFs are comparable to much
older power plants in the system and lower than the newer Kusile and Matimba power plants with EAFs
of about 70 percent and 88 percent, respectively.
69. The combined generation loss from the UCLF and the PCLF should be 8 percent based on Eskom’s
URS and international benchmarks for similar power plants. However, the UCLF alone accounted for about
30 percent in 2020. The PCLF, which accounts for generation losses due to downtime for preventative
maintenance, was also high, ranging from 5 percent to 24 percent. Eskom understands that, in the
immediate term, reductions in the UCLF are possible and plans to achieve them by: (a) increasing the
number of qualified O&M staff, (b) strengthening operational procedures to ensure safety of people and
equipment, and (c) undertaking strategic procurement of spare parts and materials to ensure their
availability when needed. This should proceed in parallel with expeditious measures to remedy the plant
defects.
70. Normally, on multiunit projects, the availability statistics improve as experience is gained from
unit to unit. However, due to the embedded latent defects primarily with the boiler design and Eskom’s
34 In August 2021, an explosion was reported at Unit #4. The incident occurred during repair work by Eskom staff during which
appropriate procedures were reportedly not followed.
35 EAF is the ratio of the amount of energy generated during a given period to the reference electricity generation (or expected
maximum output) during the same period. A typical industry standard is 85 percent after the first year of operation (shakedown
period) for a coal‐fired power plant; PCLF = ratio of planned electricity losses during a given period to the reference electricity
generation (or expected maximum output) during the same period. A typical industry standard is 10 percent; UCLF = ratio of
unplanned energy losses during a given period to the reference electricity generation (or maximum expected generation) during
the same period. A typical industry standard is 5 percent.
Page 27 of 102
The World Bank
inability to fix the defects on each subsequent unit (due to schedule and no defect fixes available in time),
the impact of the latent defects remained embedded on almost all units except portions of the last two
units.
71. Transmission lines. The associated transmission lines totaling 1,020 km as revised during project
restructuring (December 2019) were substantially completed by project closing date, except for two
sections for a total of 101km on which works were in progress at the time of project closure. As of the
beginning of February 2022, the Medupi‐Borutho 400 kV (Section A) (24 km) had been delayed pending
clearance of defects and of trees that needed to be trimmed. Challenges were encountered due to the
lack of access in the absence of the landowner from the country and commissioning was expected by
March 30, 2022. Regarding the second section, the Masa‐Ngwedi 765 kV about 61 percent of tower
erection and 93 percent of stringing and regulation was complete at the beginning of February 2022.
Completion of the transmission line was expected in May and energization in August 2022. The delayed
completion of these transmission lines does not affect the evacuation of power from the MCPP.
72. The actual energy generation by the MCPP remains substantially below the expected design
outputs and industry benchmarks, and energy output increases are dependent on the curation of latent
defects and on improvements in operational efficiencies, which will take several years to implement.
Nonetheless, the outcome indicator target for generation capacity has been installed and commissioned,
the plant contributes a significant 15 percent of the system energy generated (about 20, 000 GWh net
output)36 and is the lowest carbon emitter in Eskom’s fleet of coal‐fired power plants. Without the plant’s
contribution to total system energy generation power outages would be much worse. In addition, the
availability of the MCPP has provided Eskom with the flexibility to decommission/repurpose some its
inefficient and high carbon emitting coal power plants.
Renewable energy’s contribution to power supply and energy security
73. The Sere wind power plant was commissioned on March 31, 2015, slightly behind the original
schedule. The average annual net output is 329 GWh, which is higher than the 217 GWh estimated at
appraisal as the plant is operating at about 38 percent capacity factor compared to 25 percent estimated
at appraisal. Thus, the plant has exceeded appraisal expectations by contributing about 50 percent more
output than expected. An additional 200 MW was expected to be added by the BESS program. The BESS
program encountered multiple implementation problems, including delays in obtaining investment and
regulatory approvals, an unsuccessful first bid process leading to a revision of the implementation
approach, and a retendering that is currently in progress. Implementation of the BESS program is
continuing with financing by the ERSP CTF funding of phase 1 if restructured, the AfDB, and the New
Development Bank.
Investments and TA for energy efficiency and renewable energy development component’s contribution to
power supply and energy security
36 Other metrics considered for evaluation of the objective of increased power supply and energy security included the incidence
of blackouts before and after the project came into operation and the ratio of installed capacity to peak demand. These have not
been used because of the (a) problem of attribution for systemwide indicators; (b) irregular pattern of power blackouts, and (c)
discrepancy between the prevailing high ratios of installed capacity to peak demand (100 percent) and recurrent power shortages
due to system wide equipment failures and other operational challenges.
Page 28 of 102
The World Bank
74. Implementation of efficiency upgrades of existing coal‐fired power plants based on studies
financed by the project was expected to contribute to improved supply‐side operational efficiencies.37 By
2013, Eskom had completed efficiency upgrades at three coal power stations (Hendrina, Kendal, and
Matimba) and design studies for seven power plants. Eskom reported cumulative total capacity savings of
4, 557 MW against a target of 3,171 MW from supply‐ and demand‐side efficiency measures. Further
efficiency upgrades of more power plants appear to have been constrained by limited opportunities for
taking units offline.
All components’ contributions to efficiency and sustainability
75. Efficiency and sustainability. Eskom’s Carnot thermal efficiency target for the MCPP is 38 percent.
Based on structured field testing for almost all units, the MCPP’s actual thermal efficiency is slightly higher
and compares favorably to 30–35 percent for the rest of the subcritical coal‐fired power plants in Eskom’s
portfolio. When Medupi was developed, it was expected to meet incremental demand of at least 4
percent per year. However, given the much lower demand growth in recent years, the plant is displacing
older coal‐fired power plants, which are less cost‐efficient. Because the plant is the lowest carbon emitter
in Eskom’s coal fleet, its displacement of other coal‐fired power reduces Eskom’s total emissions—hence
its contribution to sustainability. As noted, the efficiency upgrades of some of the existing coal‐fired power
plants and demand‐side efficiency measures yielded about 4,557 MW in total capacity savings. The coal
transportation cost reduction has not been realized and neither have the expected environmental
benefits that were expected to result from a partial shift of coal transportation from road to rail38. The
Majuba Rail is now set to enter commercial operation in 2022 beyond the project’s closure on June 31,
2021.
76. Overall, the efficacy of objectives 1 and 2 is deemed Modest in light of: (a) the renewable energy
activities fully achieved their capacity (100 MW Sere wind power plant) and energy expectations; (b) the
technical assistance activities made substantial contributions to the PDO; and (c) the MCPP delivered fully
on potential capacity (4,764 MW), but only partially on actual energy output (significantly low availability
energy factors relative to both Eskom’s target and international benchmarks, and significant unplanned
outages that account for 28 percent of time on average over the past three years in particular constrained
the ability of the project to fully contribute to the energy security). Yet, despite these shortcomings and
the challenges of addressing latent defects the MCPP is contributing a significant 15 percent of the power
system generation, thus helping to alleviate power shortages, is the lowest carbon emitter in Eskom’s
fleet of coal power plants and is enabling Eskom to start work on decommissioning/repurposing its most
inefficient coal power plants. Thus, on balance, considering both the shortcomings and these positive
aspects and the substantial contributions of the Sere wind power subcomponent, the demand and supply
37 Eskom did not use the US$20 million of the IBRD’s original allocation to support efficiency upgrades of but used its own
resources. Part of the funds were reallocated to support Eskom’s unbundling reforms, but this activity could not be implemented
before the loan closing date, in consequence of which this reallocation was canceled.
38
In December 2019 a fire started on the incline conveyer belt system from the rail offloading facility, thereby, disrupting coal
supplies by existing rail line and shifting the load to road transportation. The repair work was undertaken and the
first Majuba coal train (since the fire in December 2019) successfully offloaded in early October 2021
Page 29 of 102
The World Bank
side efficiency measures, the overall efficacy rating for objectives 1 and 2 (enhancing power supply and
energy security in an efficient and sustainable manner) is deemed Modest.
Objective 3: Supporting the GoSA’s economic growth objectives
77. No indicators or targets were provided for this high‐level objective. The underlying objective in
developing the project was to increase availability of adequate and reliable electricity so that industry,
manufacturing, and MSMEs would grow and create jobs and access would be expanded to the remaining
20 percent of the population without access. Therefore, this objective is assessed with reference to the
success of the project in increasing energy availability to the customers. Because the objectives of
enhancing power supply and energy security in an efficient and sustainable manner are rated Modest for
efficacy, so is the rating of the objective of supporting the GoSA’s economic growth goals.
Objective 4: Supporting the GoSA’s long‐term carbon mitigation strategy for South Africa
78. The aim of this objective was to support the reduction in the carbon intensity of electricity
generation. Initially, achievement of this objective was to be measured by a reduction in the carbon
emissions discharged per unit of electricity with a target of 0.950 kg per kWh by project closure. The
indicator was replaced by the volume, in metric tons of CO2 reductions by project end (with a target of
238,000 MT). According to the Results Framework, the actual reduction in CO2 emissions was 315,330
MT. However, the achievements would have been even higher had the BESS and the Majuba Rail
components become operational by project closure or soon thereafter and had the Medupi plant’s EAFs
been higher resulting in displacement of more output from less‐efficient coal‐fired power plants.
79. Given that the target was exceeded a rating of High would be expected. However, because of the
non‐completion of two other activities (the BESS and the Majuba Rail) that were expected to contribute
to this target the objective is rated Modest.
Justification of Overall Efficacy Rating
80. The overall efficacy rating is Modest based on a Modest rating for Objectives 1 and 2 (enhancing
power supply and energy security in an efficient and sustainable manner), a Modest rating for Objective
3 (supporting the GoSA’s economic growth objectives), and a Modest rating for Objective 4 (supporting
the GoSA’s long‐term carbon mitigation strategy for South Africa). The detailed justification for the rating
of the three individual objectives (with Objectives 1 and 2 rated together as one) is provided above.
C. EFFICIENCY
81. At project appraisal, the project’s economic and financial analysis was conducted for the following
four subprojects: (a) Medupi Coal Power Project, (b) Sere Wind Power Project, (c) Kiwano Concentrated
Solar Power Plant, and (d) Majuba Rail Subproject.
82. The main aspects covered by the economic analysis at appraisal were assessments of (a) whether
additional base load capacity was required and whether the Medupi Power Project represented the least‐
cost option for meeting that requirement and what its economic incremental rate of return (EIRR) would
be; (b) the economic viability of the Sere wind power and Kiwano CSP projects using the EIRR and net
Page 30 of 102
The World Bank
present value (NPV) analysis, primarily; and (c) the economic viability of the Majuba Rail Project based on
the EIRR and NPV analysis.
83. Eskom’s least‐cost expansion plan had demonstrated that the MCPP was the least‐cost alternative
option for meeting South Africa’s demand for electricity in the short term. The World Bank’s economic
analysis confirmed this by comparing Medupi to alternatives such as renewable energy, diesel generation
and LNG, and imports from the region. The analysis indicated high shadow prices of carbon as a switching
value for alternatives to the MCPP except for the Inga Hydropower Project in the Democratic Republic of
Congo, which showed a switching value of US$7 per metric ton of carbon. The incremental costs of
renewable energy capacity needed to substitute the MCPP would be high given the intermittency nature
of renewables and beyond Eskom’s capacity to finance even with some support from carbon finance and
concessional sources. Although an LNG option having lower carbon emissions than the MCPP and Inga
would have been a competitive option, neither could be implemented speedily to deliver the required
energy in the short term. Moreover, the huge uncertainties over the Inga hydropower project expansion
in the Democratic Republic of Congo and the need for an additional high voltage direct current
transmission corridor through Angola and Namibia would have been unacceptably risky.
84. The World Bank Inspection Panel. In 1993 the World Bank’s Board of Directors established “The
Inspection Panel” – an independent complaints mechanism for people and communities who believe that they
have been or are likely to be adversely affected by a World Bank‐funded project. The Inspection Panel (IP)
responds to complaints from project‐affected people alleging harm because the Bank has not complied
with its environmental and social policies and procedures. It focuses on the Bank's actions and does not
investigate individuals or borrowers, nor does its mandate extend to allegations related to procurement
or corruption. On April 6, 2010, the IP received a ‘Request for Inspection’ related to the project. The
request covered a wide range of issues39, including that the World Bank had not adequately analyzed
alternatives to coal. The IP issued its findings in a report (No.64977‐ZA) dated November 21, 2011. The
World Bank Management response dated March 12, 2012, stated that the PAD had gone beyond the
identification of the “least cost” options for meeting electricity needs, thereby satisfying requirements for
analysis of alternatives in projects of this type.
85. The above description of the work carried out to analyze the alternatives to the MCPP confirms
the ICRs’ agreement with management’s response to the Inspection Panel. The IP’s findings related to
safeguards matters and the manner in which they have been handled during project implementation are
discussed in Section IV: Bank Performance, Compliance Issues and Risk to Development Outcome,
Subsection B: Environmental, Social and Fiduciary Compliance.
86. A summary of the economic returns at appraisal is shown in table 5. The aggregate NPV is US$15.9
billion and the ERR is 22.7 percent.
39 In addition to the issue of alternatives to coal the other issues covered in the Inspection Panel request included: impacts on
health due to air quality, water resources, livelihoods, cultural heritage and practices, influx of laborers, involuntary
resettlement, and associated coal mining. The other issues are evaluated in the sections on safeguards and other impacts.
Page 31 of 102
The World Bank
Table 5. Economic Returns at Appraisal
Component ERR (%) NPV @10 Percent (US$, millions)
A ‐ Medupi Coal 24.0 15,727
B ‐ Wind 10.7 9.2
B ‐ CSP 1.8 –266
C ‐ Majuba Railway 19.0 437
Project as a whole 22.7 15,907
Source: PAD, annex 9, table 14.
87. When the net increases in GHG emissions are considered, based on a carbon valuation of US$29
per ton of CO2,40 the economic returns of the project were found to decrease. The bulk of the impact is
attributable to the Medupi Coal Project, whose returns are negatively affected as it is the only carbon
emitting project component. The economic returns of the wind, CSP, and Majuba Railway Project
components all increased slightly (table 6). When GHG emission damage costs were included, the
aggregate NPV reduced from US$15.9 billion to US$12.9 billion (using 10 percent as the discount rate).
Table 6. Economic Returns, including CO2 Valued at US$29/ton
Component ERR (without GHG) ERR (with GHG) Change in EIRR NPV
(%) (%) (US$, millions)
A ‐ Medupi Coal 24.0 21.5 –2.5 12,585
B ‐ Wind 10.7 14.1 3.4 54
B ‐ CSP 1.8 4.0 2.2 –206
C ‐ Majuba Railway 19.0 19.6 0.6 457
Project as a whole 22.7 20.5 –2.2 12,890
88. As stated in section I, the CSP Project was replaced by a BESS pilot project during the December
7, 2018, project restructuring. However, implementation of the first phase, which was to be supported by
this project, could not start before the project closed on June 30, 2021. Therefore, no economic or
financial analysis was conducted for the BESS in this ICR as it will now be implemented under separate
financing.
89. The revised economic analysis at project closure has incorporated the following new
developments in economic analysis of power projects and other information that was not available at the
time of project appraisal: (a) the World Bank’s guidance note on the shadow price of carbon that was
issued in 2017; (b) the World Bank’s Energy Practice guidance on economic analysis of power sector
investment projects issued in 2016 and the OPCS41 guidance on economic analysis issued in 2013, which
require the calculation of NPV without externalities, with local externalities and with both local and global
externalities; and (c) research data on the valuation of local environmental costs of air pollution published
by the World Bank in 2015.
40 As noted in annex 4 of the PAD, in the year of appraisal this value was already much higher than that used in other World Bank
energy sector projects. The mandatory World Bank guidance had not been issued at the time.
41 OPCS = Operations Policy and Country Services.
Page 32 of 102
The World Bank
Table 7. Revised Economic Returns at Project Closure
Baseline With Low SVC With High SVC
NPV NPV NPV
Component ERR (%) (US$, millions) ERR (%) (US$, millions) ERR (%) (US$, millions)
Medupi 12.9 17,744 12.5 15,485 12.2 13,333
Sere Wind 17.1 220 19.4 275 21.6 331
Majuba < 0 –185 <0 –184 < 0 –184
Whole project 12.8 17,779 12.5 15,576 12.2 13,480
Note: SVC = Social value of carbon.
90. The project’s overall ERR is re‐estimated at 12.8 percent compared to the appraisal estimate of
22.7 percent. This significant reduction of the project’s economic returns is attributable largely to those
of the MCPP—its largest component. The reduction in the ERR for the MCPP is due to a capital expenditure
(CAPEX) increase of some 25 percent, an extended contract period (about six years delay), and failure to
meet expected energy generation in the early years of operation up to project closure.
91. It is noted that the impact of GHG emissions is small, in part because GHG emissions in the early
years of the MCCP have decreased (as plant has replaced coal generation from old inefficient coal plants).
Even at the high value of the SVC (according to World Bank guidelines), the decrease in the ERR is just,
from 12.9 percent to 12.2 percent, and the impact of local environmental damage costs based on
estimates from studies published after appraisal is also relatively small—less than a 0.1 percent decline in
the ERR.
92. The economic returns of the Sere wind project (17.1 percent) have increased significantly over
the 10.7 percent at appraisal. This is due to much higher energy generation than estimated at appraisal
(329 GWh per year compared to 219 GWh), an efficient construction schedule of three years instead of
the expected four years, and lower capital costs than the PAD estimate.
93. At appraisal, the economic returns of the Majuba Railway Project were estimated at 19 percent.
At project completion, the returns are significantly lower, though the extent of the decline is due to three
problems. The first is the inordinate delay in project completion, with start of operations now only
expected by 2022. The second is the smaller differential between the existing rail freight, and the now
expected Majuba Rail rate, from R 140 per ton at appraisal to the now estimated rate of only R 45 per ton.
When combined with the dramatic decrease of the coal use at Majuba, stated by Eskom to now be just 8
million tons per year (mtpy) instead of the appraisal expected volume of 13 mtpy, the economic returns
are negative. The forecasted reduced coal consumption at Majuba is based on Eskom’s production plans
which consider, amongst other factors, declining or stagnant system demand, a decreasing electricity
market share for Eskom as the penetration of renewable energy increases, and Majuba’s competitiveness
vis‐a vis other Eskom power plants in the system42.
94. At 14 mtpy, Eskom’s financial internal rate of return (FIRR) of the Majuba Rail Project component
is a highly satisfactory 14.1 percent, and an ERR of 8.9 percent, substantially above the 6 percent default
discount rate suggested by the latest World Bank guidance. At 8 mtpy, the leveraged Eskom FIRR drops
42 Eskom states that energy availability factors for its entire fleet of coal fired power plants have been declining, including that
of the Majuba Power Plant which stood at 72 percent in 2021, accounting for a coal burn of 12.5 mtpy.
Page 33 of 102
The World Bank
from 69.5 percent estimated at appraisal to just 5.5 percent. This is unsatisfactory because it is below
Eskom’s weighted average cost of capital (WACC). The ERR at this utilization of Majuba is negative.
Assessment of Efficiency and Rating
95. The efficiency assessment is based on both the ERR and overall efficiency of project
implementation. The efficiency of implementation assesses the efficiency of resource use in project
implementation.
96. Economic analysis. The economic analysis concludes that although the project’s overall ERR of
12.8 percent is much lower than the appraisal estimates of 22.7 percent, it remains above the 10 percent
hurdle rate used at appraisal and is substantially above the World Bank’s guidance discount rate of 6
percent. The main reason for the decrease in economic returns is the problems at the MCPP: cost
overruns, protracted delays, and poor operational performance in the first few years of operation. Hence,
the MCPP’s revised ERR is much lower at 12.9 percent compared to the appraisal estimate, while the
Majuba Rail revised ERR is negative compared to 19.1 percent at appraisal. Only the Sere wind power
plant was completed below budget and has an ERR above the appraisal estimate—17.1 percent compared
to 10.7 percent. The fourth project, the BESS, is still at the bidding stage and will be implemented under
separate financing and has therefore not been included in the economic analysis.
97. Efficiency of resource use. The assessment of efficiency of use of resources compares the unit of
generation costs (US$ per kW) of the MCPP and Sere wind power plant to industry comparators. The cost
per kW for the MCPP is about US$4,000 compared to an industry comparator of US$3,50043 in 2021
nominal prices. The comparator costs include wet FGD whereas the MCPP cost are without FGD. Adjusting
for the FGD would mean a significant cost difference between the two, thus confirming the inefficient
implementation of the MCPP. The Sere Wind Project’s capital cost is about US$2,700 per kW compared
to industry comparator of US$2,600 per kW.44 Thus, the project was implemented efficiently with
outcomes comparable to industry standards.
98. In addition to the unit cost of the MCPP, the implementation delay of about six years involved
substantial additional resources by both Eskom and the World Bank that are not captured in the revised
ERRs, including the costs of staff time and other resources.
99. The project is rated Modest for efficiency for the following reasons: (a) while the overall ERR
remains robust and above the cost of capital, two subprojects were not completed–one is deferred for
completion under another operation and the second one will only be completed next year with an
expected negative ERR and (b) the unit costs for one of the two completed subprojects (the MCPP) are
significantly higher than the industry comparator for similar plants, while the wind project is competitive
relative to comparators.
43 Adapted from the new coal‐fired power plant performance and cost estimate, Sargent and Lundy LLC 2009.
44 Adapted from the 2019 Cost of Wind Renewable Energy Review, National Renewable Energy Laboratory.
Page 34 of 102
The World Bank
D. JUSTIFICATION OF OVERALL OUTCOME RATING
100. The outcome of the project is rated Moderately Unsatisfactory because the PDO’s relevance to
the World Bank Group’s CPF at project closure (FY14‐17, extended to FY2018) and the FY22–26 CPF which
came into effect three weeks after the loan closed is deemed Substantial, their efficacy is rated Modest
because they were only partly achieved, and the project’s efficiency was below expectations for the sector
and, therefore rated, Modest.
101. The detailed justifications for each of the three performance dimensions (relevance, efficacy, and
efficiency) have been described in the relevant sections above. It is important to note the dominant
impact of the MCPP in the outcomes of three of the four objectives (that is, enhancing power supply and
energy security (Objectives 1 and 2) and supporting economic growth (Objective 3), which are all rated
Modest. While the MCPP has achieved its outcome indicator target of generation capacity installed and
commissioned, is making significant contributions to the system generated energy, is the lowest carbon
emitter among Eskom’s coal‐fired power plants, and is providing Eskom the flexibility to
decommission/repurpose its most inefficient coal power plans, its below‐expected energy outputs and
substantial defects, which will take many years to address, have lowered the project’s overall efficacy
rating to Modest. In addition, it has contributed to an efficiency rating of Modest due to its longer
implementation duration and higher costs per kW than comparator plants in other countries.
E. OTHER OUTCOMES AND IMPACTS (IF ANY)
Gender
102. At project preparation, it was noted that the rate of electricity connections had decreased in prior
years due to the shortages of power and yet studies conducted by the DMRE in 2008/09 in the Limpopo,
KwaZulu Natal, and the Eastern Cape Provinces pointed to significant benefits among the electrified
communities—including opportunities for productive uses of electricity and better education and health
outcomes for women and girls. Although no specific statistics are available, in line with global experience,
women and girls benefit when support, such as that provided by Eskom to schools and health centers, is
made available. In addition, during project restructuring in July 2015, an indicator of project direct
beneficiaries, of whom 50 percent would be female, was introduced in the Results Framework. From a
baseline of 35,608, the total number of beneficiaries was expected to increase to 4,584,283 by December
31, 2019. The actual number of beneficiaries as of July 20, 2020, was 3,600,000.
Institutional Strengthening
103. This was the World Bank’s first major project with Eskom. The donor partnership in support of
Eskom was acknowledged in Eskom’s ICR as an effective method for the exchange of knowledge and
experience. The large scale of the project, while challenging, also offered many opportunities for
development of skills within Eskom in project management and contract management, monitoring and
evaluation (M&E), and stakeholder engagement with communities.
104. Although Eskom witnessed the departure of many staff who had benefited from the experience
on the project, the benefits remained largely within the economy as many moved to private sector jobs.
Page 35 of 102
The World Bank
105. In addition to Eskom, contracting opportunities were made available to local businesses and many
succeeded. In addition, Eskom offered entrepreneurial development programs in the project communities
and supported schools and clinics with equipment and supplies, thus helping strengthen local institutions
for broader development.
Mobilizing Private Sector Financing
106. The CSP component was to pilot CSP technology with a view to provide learning for subsequent
private sector development of this technology. However, due to the delays in the CSP Project, the private
sector started working on the technology before the pilot could be implemented, which in part
contributed to Eskom’s decision to abandon the pilot. By the time Eskom started on the BESS in 2018,
about 4,000 MW of renewable energy capacity was already online under the REIPPP. Some of the BESS
sites were to be located around REIPPP plants to motivate more private sector investment in renewable
energy by making it a dispatchable resource through battery storage. The World Bank’s participation in
the EISP was instrumental in supporting the GoSA’s development of renewable energy, particularly
through the REIPPP, through which renewable energy capacity increased to about 6,000 MW as of 2021
compared to almost zero in 2009 when the World Bank started its involvement in South Africa’s energy
sector. The energy supply enabled directly through this project and indirectly through the World Bank’s
technical support to the GoSA has also facilitated private sector investments in industries that depend on
electricity for productive activities.
107. In addition, the building of transmission lines around the Sere wind power plant is also to
facilitate private sector investment in renewable energy by facilitating the integration of their plants to
the grid.
Poverty Reduction and Shared Prosperity
108. The EISP’s footprint on poverty reduction and shared prosperity in the project area and the wider
national community is significant even though there are no income metrics designed to demonstrate this.
The significance of the project’s impacts can be inferred from the (a) numbers of people employed by the
project disaggregated by gender, age, and geographic areas of origin; (b) entrepreneurial development
initiatives spearheaded by Eskom and contractors, including the contractor and supplier development
programs supported by Eskom; and (c) socioeconomic initiatives by Eskom.
109. Employment impacts. When the employment statistics were first collected in 2013, the number
of contractor employees at site totaled about 16,797, of whom about 43 percent were from the project
area (Lephalale and Waterberg), 66 percent were youth, and 7 percent were female. The numbers
gradually decreased as the project neared completion and by December 2020, 2,949 contractor
employees remained, of whom 56.0 percent were from the project area, 73.5 percent were youth, and
11.2 percent were female. Between 2013 and 2018, Eskom, contractors, local authorities, and others
formed the Medupi Leadership Initiative to support an exit strategy for demobilized workers. The Medupi
Leadership Initiative enhanced the employment prospects of the demobilized workers and other people
in the local communities by providing training and support. Contractors contributed more than R 24
million to build and equip training centers and about 4,616 (that is, more than Eskom’s 3,071 target for
local skills development) were trained of whom about 60 percent were from the project area – Province
of Limpopo.
Page 36 of 102
The World Bank
110. Entrepreneurship development. Through the Local Enterprise Development Program, 144 local
business owners were trained, including 116 at Eskom’s Contractors Academy and another 28 at
IKUKASA45. The Medupi Project, in collaboration with Mitsubishi Hitachi Power Systems Africa, established
the Lephalale Business Development Center operated by Black Umbrella, a nonprofit organization. The
center closed due to financing constraints but may reopen in future.
111. Socioeconomic initiatives. Through its corporate social investment program, Eskom made
substantive contributions to the well‐being of people in the project area by investing about R 145 million
that benefited about 80,000 people. Investments in social infrastructure and services delivery included a
mobile clinic, equipping and expansion of six clinics, pediatric treatment services to more than 6,000
children annually, and extensive training in 31 schools covering about 620 staff members. Beneficiaries
continue to benefit from the established infrastructure even though Eskom’s funding has now ceased with
the completion of the project.
112. On complaints made to the Inspection Panel regarding the project’s impacts on livelihoods and
poverty reduction the Panel had found that the links between the MCPP and poverty were not
systematically addressed. It also found that the project’s impact on agriculture and tourism was difficult
to predict. As described above, Eskom made substantial efforts to support local infrastructure
development, especially to schools and clinics, and worked with contractors and other industry to support
entrepreneurial development in the project area. Eskom also supported wastewater treatment capacity
building in Lephalale.
Other Unintended Outcomes and Impacts
113. There have been several unintended impacts in the areas surrounding the project area, including
increased levels of traffic, dust, and noise, especially in the nearby Marapong area. The impacts were
generally due to the development of the areas and were not necessarily related to the project. Eskom
helped relieve pressure on infrastructure services such as wastewater treatment plant in Lephalale, which
Eskom has supported through capacity building, and through support to social facilities and services
(schools and health clinics). Eskom also built its own waste treatment plant at the Matimba power station
to relieve the pressure on the Lephalale plant.
III. KEY FACTORS THAT AFFECTED IMPLEMENTATION AND OUTCOME
A. KEY FACTORS DURING PREPARATION
114. The project’s MCPP component was largely prepared before the World Bank was invited by the
Government to participate in its financing. Initially the Government’s interest was in external financing
support for the MCPP and the Kusile coal‐fired power plant. The need to meet the World Bank’s criteria
for financing a coal‐fired power project influenced the project’s eventual structure and composition.
However, because the MCPP was already under implementation, there was limited scope for changing
the project’s design, major contracts that had been awarded, or Eskom’s implementation arrangements
that were already in place as major changes would have disrupted the GoSA’s plan for meeting urgent
45 A local entrepreneurs development center
Page 37 of 102
The World Bank
power needs. Nevertheless, the World Bank appraised the project following its policies and procedures
and conducted due diligence which resulted in some changes on environmental and social safeguards and
procurement. The ICR’s assessment of quality at entry reviews the standard issues normally covered by
ICRs in this section with due attention to the due diligence conducted during project preparation.
(a) Realism of project objectives. The project objectives were clearly defined and articulated in
the PAD and Loan Agreement. They were appropriately targeted at addressing two key
developmental issues—acute power shortages and the carbon intensity of electricity
production in South Africa. The only shortcoming was pitching the project’s support to the
economy—a high‐level objective—as a project objective.
(b) Project design. The project design was reasonable in terms of the structure of components,
operational logic, and sequencing of activities, except for the expectation that an unproven,
risky, and costly technology—CSP—would contribute to project outcomes. Including the
activity as a pilot was consistent with the World Bank’s efforts to explore the potential for
CSP, which it was pursuing in other countries such as Egypt and Morocco.46 As a learning
pilot, the activity’s outcome could have been limited to lessons learned. The technical design
of the MCPP is a separate specific issue, which is addressed in detail in the subsection on
factors that affected the project during implementation.
(c) Results Framework and monitoring plan. The Results Framework had three outcome
indicators and eight intermediate outcome indicators and clearly indicated the data
collection instruments, responsibilities, frequency of reporting, and use of the collected
data. All the indicators had baseline values and targets. Eskom was to provide the data on
all three outcome indicators through its annual reports and the data on intermediate
indicators through either its annual reports or the project’s quarterly progress report.
However, a significant shortcoming was the misalignment of the indicators with the
objectives in that they would measure system performance regarding the ratio of installed
capacity to peak demand and regarding carbon intensity of electricity production.
(d) Adequacy of risk and mitigation measures identification. A thorough identification of risks
and mitigation measures was carried out during the project preparation process. These are
summarized in section I under theory of change. The identification of risks and mitigation
measures may have missed some important ones that could have been picked up from the
due diligence analysis, as described in the next paragraph.
(e) Technical due diligence. The technical due diligence consultant hired by the World Bank to
assess the quality of project preparation concluded that the MCPP project design was
conducted in accordance with current engineering practices and that the plant’s
performance could be expected to be comparable to similar state‐of‐the‐art projects. In
retrospect, serious design issues which later plagued project implementation and operation
were missed. The use of the ‘virtual’ design from another plant (Majuba) in a different
location with defunct building/design codes and some of the early design‐related
implementation problems were known at the time. This could have resulted in an early
identification of design risks and mitigation measures. The consultant assessed Eskom’s
46 The Morocco CSP was designed to be implemented under public‐private partnership arrangements, which provide for sound
risk sharing arrangements between the public and private sectors and satisfactorily achieved its outcomes.
Page 38 of 102
The World Bank
capability to be appropriate for the implementation and operational needs of the MCPP.

However, an important qualification was made regarding the size of the staffing for the
operational phase, which the consultant considered to be small for the size of the plant. This
could have been flagged as an important risk and appropriate risk mitigation measures put
in place for the implementation phase.
(f) Procurement due diligence. The consultant who carried out the technical due diligence also
assessed whether the procurement that had already been carried out followed Eskom’s
policies and procedures and complied with South Africa’s regulations. The conclusion was
that not only had the procurement processes complied with Eskom’s and the country’s
policies and regulations, but they were consistent with international good practices. The
processes were fair and transparent and were carried out with due regard to economy and
efficiency resulting in costs comparable to similar projects in the United States and the lower
end of the European market for similar projects. The consultant noted that the bidding was
carried out at a time when the market for coal power plants was largely sold out (2006),
resulting in less competition for the boiler and turbine packages. As a result, the same
contractors for the boiler and turbine packages had also been awarded contracts for the
same packages for the Kusile coal‐fired power plant resulting in the largest world record
contract award and load on the original equipment manufacturers (OEMs), which should
have been flagged as an onerous duty which could, and did, cause schedule and design
weakness. Assessment of the capacity of the contractors to implement both sets of contracts
concurrently may not have been adequately detailed. A detailed review of all the contracts
proposed for World Bank financing also concluded that the bidding was carried out in
accordance with principles of fairness and transparency and with due attention to economy
and efficiency. As a result of the due diligence, the Board agreed, on an exceptional basis,
for the World Bank to finance contracts that were awarded without following the World
Bank’s guidelines and to the application of South Africa’s local content requirements to
future contracts under the project.
(g) Environmental and social safeguards due diligence. The World Bank selected South Africa
as a pilot for the UCS for environmental and social safeguards. The EISP was to be the second
project to use country systems in South Africa under the World Bank’s pilot program.47 The
World Bank conducted an SDR to assess the equivalence of South Africa and Eskom’s
systems for safeguards to the World Bank’s operational policies and procedures, assess their
acceptability for application to the EISP, and agree with the Borrower on gap‐filling measures
for addressing any identified deviations. The SDR was applied to four of the World Bank’s
safeguards policies that were applicable to the EISP: Environmental Assessment (OP/BP
4.01), Natural Habitats (OP/BP 4.04), Physical and Cultural Resources (OP/BP 4.11), and
Involuntary Resettlement (OP/BP 4.12). These were the same policies that were assessed
for the iSimangaliso Project. The SDR concluded that there were few minor gaps or
ambiguities in the language of the South African legal framework with respect to these four
policies. However, the SDR further concluded that the gaps were fully addressed and
internalized through Eskom’s own policies and practices, except for the involuntary
47 UCS was first applied in South Africa to the GEF‐funded Development, Empowerment, and Conservation of the iSimangaliso
Wetland Park and Surrounding Region Project, approved in 2009.
Page 39 of 102
The World Bank
resettlement which lacked a legal requirement for the preparation and public disclosure of
resettlement action plan and/or frameworks as well as documentation of the impacts of
resettlement. The SDR also supported the installation of the FGD units at Medupi for the
plant to meet the country’s SO2 emission limits.
115. Thus, project preparation was detailed and covered all important aspects, but some significant
issues related to the project objectives/design/Results Frameworks and risks were not adequately
addressed. Regarding the project design, (a) the formulation of the PDO could have excluded the high‐
level objective of economic growth, (b) the contribution of the pilot component for CSP to the project
outcomes could have been knowledge generation and lessons learned, and (c) the original outcome
indicators could have been designed to better target project‐specific rather than systemwide impacts.
Regarding risks, the due diligence had flagged inadequacy of Eskom’s staffing plans for operation of the
plant, but this was not captured in the risk identification and mitigation measures for proactive
management during implementation. It also appears that the implementation of the MCPP in parallel with
the Kusile Project (a large project of the same size as the MCPP) with the same contractors for the large
contract packages (boilers and turbines) did not trigger a risk alert and a performance mitigation strategy
during the World Bank’s project preparation. Finally, implementation issues related to the technical
design and initial construction48 were already evident during the World Bank’s project preparation phase
but do not appear to have been adequately factored into the implementation support strategy.
B. KEY FACTORS DURING IMPLEMENTATION
116. Other than the Sere wind power plant, which was completed almost on time and within budget,
all other activities were considerably delayed and some, such as installation of the FGD units at the MCPP,
the Majuba Rail, and the transmission lines, are yet to be completed.
Factors Subject to the Government and Eskom’s Control
117. The main factors which affected implementation were as follows:
(a) MCPP plant design shortcomings. When it became apparent that power deficits would
occur around 2007, Eskom started planning for new capacity generation expansion. Driven
by the urgency, Eskom used a virtual design based on the Majuba power plant, which was
constructed about 10 years earlier. Thus, the MCPP civil works and plant designs, various
outdated codes and standards, and some of the procurement and initial construction were
based on the Majuba power plant. When subsequent investigations and site excavations
indicated the need for deeper foundations and reinforcement for the turbines and boilers,
some of the foundation work that had already been done had to be excavated and modified.
This set back the implementation schedule by about one and half years and increased
project costs. On the equipment side, the boiler contractor had fabricated and shipped
components to the site and structural construction of the boilers had started by the time the
Bank began preparing the project. When Eskom finally modified their designs and
specifications to meet the specific needs for the MCPP, the contractors had to modify their
48 This observation relates to the virtual design of the plant, which resulted in some civil works structure for the equipment and
steel structures for the boilers having to be pulled down and replaced.
Page 40 of 102
The World Bank
steel designs and some vendors had to do the same and resubmit their bids. Some of the
steel structures that had been erected for Unit #6 had to be pulled down and replaced by
modified designs. Eskom continued to refine its specifications and request contractors for
equipment design and quality modifications. The changes caused significant cost increases
and schedule delays.
(b) Eskom’s contracting strategies. The MCPP had numerous contract packages—more than
400 at peak for the MCPP and associated transmission lines. This posed a huge contract
management challenge for Eskom, which did not have an owner’s engineer for the MCPP.
Instead of an independent owner’s engineer, Eskom had opted to establish Medupi
Execution Teams (METs), in which it integrated personnel provided by a consulting firm for
various contract packages/activities. The model did not work as effectively as expected.
Eskom suffered from a continuous loss of experienced experts to the private sector, which
exacerbated the challenges of effective construction and contract management.
(c) Performance of contractors and subcontractors. A local subcontractor to the boiler
contractor did not follow the required weld procedure specifications in the manufacture of
the boiler tubes to be able to handle the high temperatures and pressures of the supercritical
boiler.49 This caused thousands of boiler welds and major steam separator vessels and
connections to be replaced in Units #6, 5, and 4, resulting in about nine months delay. A
World Bank supervision mission’s visit to the project site in June 2013 summarized the root
causes of the boiler problems as inadequate: (i) early source inspection and thorough review
of welding quality assurance records; (ii) prequalification of contractors and their
subcontractors for local pressure part fabrication; and (iii) training and supervision of local
contractors and subcontractors, which is essential to ensure the success of localization with
quality outputs. There were also problems with the boiler protection system and the control
and instrumentation equipment by another contractor, which failed most of their factory
acceptance tests and site acceptance tests. The contractors had used their own
specifications instead of Eskom’s and were unable to improve despite Eskom’s intervention
over nine months. The problem with the control and instrumentation equipment had the
impact of delaying other contracts in the schedule, thus contributing to further overall
delays. Eskom had to bring in another contractor to partner with the original contractor for
the boiler protection system and the equipment to be remanufactured. This ended up with
a different control design and instrumentation and OEM on each of the three plants
groupings (Units #6, 5, and 4 and Units# 3, 2, and 1), which has taken years to meld together.
On other activities, contractual performance issues resulted in a rebidding of the remaining
works on three sections of transmission lines. Similarly, several contractual problems
delayed completion of the Majuba Rail yard, connection of the yard to the new rail line, and
commissioning of the new coal train service.
(d) Plant defects at the start of plant operation. Soon after commissioning of Unit #6 boiler
plant, defects started to emerge one by one. Due to contractor’s resistance, it took Eskom
more than two years to get them to negotiate a resolution. Hence, execution of the agreed
49 Procedures are written according to specific code and are laboratory tested and certified by an international independent
authority.
Page 41 of 102
The World Bank
remedial plan started only in 2020. A bespoke agreement was reached on October 6, 2020,50
between Eskom and the contractor on how to implement remedial actions and on a 50/50
percent formula for cost sharing pending the resolution of responsibilities through contract‐
level discussions or eventually through arbitration, if necessary.51 In the meantime, further
delays arose due to the sale of the boiler contractor’s shares in Medupi to their holding
company, which took time to settle the obligation for the defect issues.
(e) Regarding the boiler defects, the admitted root cause of all the defects as studied by Eskom,
is the boiler furnace, which is too small for the combustion of South African coal. The South
African coal exhibits a delayed combustion characteristic (well known in the industry), which
requires a larger furnace to reduce gas temperatures to the other heat transfer surfaces.
Most boiler OEMs with experience knew about the South African coal needs and the normal
practice to pretest the coal in the boiler OEM testing facilities to develop proper design
calculations. The MHPS had claimed to Eskom that they in fact had done their early design
due diligence but, if so, the small furnace should not have occurred. Compounding all this
was that the MHPS subcontracted the basic boiler design to Steinmueller and the air heater
design and PJFF design to Balke‐Durr (both now defunct). These were major subcontracts,
which require engineering compatibility and are difficult to manage and may have been one
reason for the design mismatches for the unique South African coal. Furthermore, though
the World Bank requested confirmation from Eskom and the contractor that the quality of
South African coal had been considered in the detailed design of the boilers, no
documentation was ever provided to the World Bank. Had this been done, the too small
furnaces could have been identified to at least institute some design remedies. Now, all 12
units at Medupi and Kusile have the same small furnace—a situation that cannot be rectified.
(f) Governance. During project preparation, opposition to the project by some civil society
groups had raised concerns about governance in the energy sector. During the project
implementation period, there were successive turnovers of senior staff in key oversight roles
in the project. These changes at senior executive levels, combined with the departure of
many technical experts after the project was approved, weakened Eskom’s capacity to
implement the project. Unfortunately, the departure of skilled personnel to the private
sector is continuing and constitutes a risk to sustainability to efficient operation of the new
power plant. External auditors started issuing an audit qualification on Eskom’s financial
statements in relation to some irregular expenditures—a development that added to the
perception of a governance crisis in Eskom. The audit qualification is issued every year until
the related contracts expire or the National Treasury decides to condone the contracts as
requested by Eskom.
(g) Legal and regulatory framework. Delays in decision‐making by the GoSA agencies
significantly delayed project implementation. The key decisions related to approvals
required under the Public Financial Management Act (PFMA) for contract awards; waivers
on the application of the Preferential Public Procurement Policy Act (PPPA) (2000); and
licensing and permitting decisions, including environmental water use licenses. In several
cases, awards of contract had to await approvals under the PFMA. Regarding the PPPA, an
50 The contractor had started implementing some of the remedial actions on the first three units.
51 The contractor attributes some of the problems to Eskom’s O&M (including staffing) deficiencies.
Page 42 of 102
The World Bank
agreement was reached with the GoSA for World Bank‐financed contracts to be exempted
from the provisions of the act. However, from December 7, 2012, exemptions were ceased
unless special waivers were granted.52 The impact was to put on hold bidding for several
packages on which procurement had not commenced (CSP, the second phase of the Medupi
transmission lines, and Majuba Rail). The request for a waiver was launched with the
National Treasury in December 2012, but the waiver was granted 11 months later in
November 2013.
(h) Environmental and social. Resettlement and right‐of‐way issues contributed to the delays
for both the transmission lines and the Majuba Rail components. The resettlement issue for
two families that has been outstanding for more two years is only now nearing final
resolution.
Factors Subject to the World Bank’s Control
118. Adequacy of supervision. The World Bank’s supervision inputs were generally adequate. The
World Bank supervised the project at least twice a year and had team leaders and other senior staff based
in Pretoria for most of the project’s duration. Both senior World Bank management and the project team
engaged Eskom and high‐level GoSA officials on many occasions to discuss implementation issues and
remedial actions for the delays. However, the effectiveness of the supervision efforts was not enough to
achieve resolution of the main project implementation issues. This was, perhaps, because in the early
years, there was a hesitancy by Eskom to be more open to the World Bank on the progress and status of
the main issues. In addition, familiarity with each other’s policies and procedures took time to develop as
did the working relationship. Thus, the issue of Eskom’s capability to implement a project of such
complexity in the face of continual staff changes at the top and at technical levels remained unaddressed.
119. Adequacy for reporting. Reporting on the project progress, issues, and proposed remedial actions
was of high quality internally within the World Bank and from Eskom to the World Bank. Joint missions
with the AfDB were complementary and generally targeted the right issues.
Factors outside the Control of the Government and Eskom
120. The World Bank team estimated that labor disputes and bad weather combined delayed project
completion by more than a year. For example, during February and March 2013, there was industrial
action at Medupi over wages, benefits, and issues of post‐construction transition support for labor. A
partnership agreement was reached between Eskom, labor unions, and the main contractor and yet by
July of the same year, further unrest had erupted.
52 Without the waivers the bidding would not have complied with the World Bank’s policies. For example, the PPPA would have
certain structures for transmission lines to have 100 percent local content.
Page 43 of 102
The World Bank
IV. BANK PERFORMANCE, COMPLIANCE ISSUES, AND RISK TO DEVELOPMENT OUTCOME
A. QUALITY OF MONITORING AND EVALUATION (M&E)
M&E Design
121. The project M&E system agreed with Eskom during project preparation comprised quarterly
progress reports, an annual report, and quarterly interim unaudited financial statements. The
intermediate and outcome indicators were to be monitored through either the quarterly progress report
or the annual report. The borrower established a Medupi Environmental Monitoring Committee, which
met every quarter and had minutes of its meetings distributed to key stakeholders. The World Bank had
an observer status at these meetings. Reports and the responsibilities for data collection are provided in
the Results Framework (annex 3 of the PAD). During the project restructurings, changes were made to the
Results Framework, especially in the July 7, 2015, restructuring when major changes were made to the
outcome and intermediate outcome indicators. These were later incorporated into Eskom’s quarterly
progress reports.
122. In addition to reporting on the indicators, the quarterly progress reports were also intended to
highlight implementation issues and challenges as well as progress of various components. Eskom was
responsible for preparing both reports but liaised with the DME regarding the indicator for renewable
energy supply. The Eskom project coordinator was responsible for overall project monitoring and
preparation of reports.
123. The World Bank also produced biannual Implementation Status and Results Reports (ISRs) to
inform management of the status the project and issues requiring their needing their attention. A version
of the ISR was publicly disclosed. Management and the project team provided regular updates to the
Board on the status of the project.
124. Overall, the design of the M&E system had significant shortcomings regarding monitoring
indicators (two out of three—installed capacity as a percentage of peak demand and carbon emissions in
kgs per unit of electricity), which were initially set to measure systemwide performance instead of
measuring impacts that could be unambiguously attributed to the project. These were subsequently
revised during restructuring and better aligned to outcomes but set at an intermediate position in the
results chain.
M&E Implementation
125. According to Eskom, in the initial stages of the project, there was a lack of clarity regarding
reporting expectations from Eskom. This was later clarified, and a common reporting format was agreed
for the World Bank and other financiers. The project files indicate that regular reports were submitted to
the World Bank.
Page 44 of 102
The World Bank
M&E Utilization
126. The quarterly progress reports provided the basis for the implementation support missions
discussions with Eskom. The use of a common report format facilitated the joint missions by the World
Bank and other financiers and appear to have helped in focusing all parties on the key issues including
project delays, cost increases, and latent defects at the MCPP. However, as noted earlier, the action plans
were not always implemented timely, if at all, and some problems remained unresolved for long periods.
Justification of Overall Rating of Quality of M&E
127. The design, implementation, and utilization of the M&E system is rated Modest. The Modest
rating is because there were significant shortcomings in the design of the system indicators and in its
utilization. The shortcoming on the M&E utilization was because of the limited impacts of the reporting
in addressing key problems that arose during project implementation.
B. ENVIRONMENTAL, SOCIAL, AND FIDUCIARY COMPLIANCE
128. This project was based on the UCS for safeguards. The two safeguard policies triggered at project
appraisal were OP/BP 4.0 (Piloting the Use of Borrower Systems to Address Environmental and Social
Safeguards Issues in Bank‐Supported Projects) and OP/BP 7.50 (Projects on International Waterways).
OP/BP 7.50 was not eligible for the UCS, and the project needed to follow the World Bank’s policy rather
than use the country systems. Four of the five safeguard policies relevant to the project were eligible for
the UCS. These were Environmental Assessment (OP/BP 4.01), Natural Habitats (OP/BP 4.04), Physical and
Cultural Resources (OP/BP 4.11), and Involuntary Resettlement (OP/BP 4.12). The Indigenous Peoples
(OP/BP 4.10) policy was later triggered during the restructuring of the project in December 2018 due to
the presence of indigenous peoples in one site proposed for the BESS.
129. An SDR was carried out to determine equivalence of the four policies with the World Bank’s
systems and the acceptability of the country systems with respect to these policies and determine gaps
between country systems and the World Bank’s policies and agree with Eskom on gap‐filling measures.
The SDR reviewed the South Africa legal system as it related to the four policies proposed for the UCS and
all the environmental impact assessments and environmental management plans that were prepared for
Medupi, Kusile, the Sere wind power plant, and the Kiwano CSP. The SDR concluded that the South African
systems were fundamentally equivalent to the objectives of OP/BP 4.0 with respect to the three
environmental policies and partially equivalent with respect to the policy on Involuntary Resettlement.
130. The main specific safeguards issues identified by the SDR and the environmental and social
assessments and the actual outcomes during implementation are shown in table 8.
Table 8. Comparison of Safeguards Issues at Appraisal to Implementation Outcomes
Issue Identified during Project Preparation Implementation Outcome
Environmental
Management The Department of Environmental Affairs Air emissions standards were promulgated by
of air quality (DEA) was planning to issue emission South Africa on March 31, 2010, requiring the
regulations for existing power plants. Thus, at existing power plants to meet specified
the time of appraisal, the national legal minimum emissions standards by April 2015
Page 45 of 102
The World Bank
framework was uncertain on the emission and then meet more stringent emissions
limits that would apply to the project. standards by April 2020.
The main issue for Medupi is SO2 emissions
for which the current allowable limit is 3,500
mg per Nm3 while its daily emissions are in
the 3,000–4,000 mg per Nm3 range.
In February 2014, Eskom applied to the DEA
for a postponement of compliance time
frames. The DEA approval required Eskom to
meet the existing plant standard for SO2 of
3,500 mg per Nm3 in 2020 and the new plant
standard of 500 mg per Nm3 starting in 2025.
On March 27, 2020, Eskom and other parties
operating similar equipment (e.g., Sasol)
were granted a higher limit of 1,000 mg per
Nm3 starting in 2025 by the DFFE53. However,
Eskom has since applied for still a higher limit
of 4,000 mg per Nm3 by 2030.
Eskom will be unable to comply with the new
standard of 500 mg per Nm3 without
installing the FGD units.
Retrofitting each MCPP unit with an FGD was Eskom submitted a program for FGD
assessed to be the most effective method of installation as required by the Loan
complying with emission standards and was Agreement.
agreed with Eskom in the Loan Agreement.
Due to delayed completion of the MCCP, the
Agreement between the World Bank and Loan Agreement was first amended in July
Eskom according to Loan Agreement for (a) 2015 to require (a) Eskom to provide the
Eskom to submit to the World Bank a World Bank with a program for installation of
program for installation of the FGD units by the FGD units by June 30, 201354 and
June 30, 2013, and to implement it implement it thereafter; (b) the first unit to
thereafter; (b) the first unit to be installed no be installed no later than six years after
later than six years after commercial commercial operation of the unit and
operation of the unit and March 31, 2018; September 30, 2021; (c) remaining units to be
and (c) remaining units to be installed installed sequentially thereafter such that the
sequentially thereafter such that the final unit final unit would be installed no later than
would be installed no later than December December 31, 2025.
31, 2021.
At project closure, Eskom submitted an
action plan indicating that the first FGD unit
53 Eskom has stated that the limit of 1,000 mg per Nm3 starting in 2025 was a regulatory change and was not a result of an
application for exemption by the company.
54 When the project was restructured on July 7,2015 Eskom had partially met the loan requirement on submission of a program
for installation of FGD units and the loan amendment did not revise the due date.
Page 46 of 102
The World Bank
will be commissioned in the first quarter of
2026 and the last one in the first half of 2027.
The World Bank is committed to monitoring
implementation of the FGD to completion.
Available water supply was not adequate to The Mokolo‐Crocodile Water Augmentation
meet the power plants’ needs at full capacity Project (phase 2) water is expected to deliver
with the operation of all the FGD units. Eskom water, including for Medupi, by 2026 when
was to seek an allocation from the the first FGD unit is expected to come on
Department of Water Affairs under the stream. If a delay appears likely, a study on
planned Mokolo‐Crocodile River Water contingency plans will be carried out in 2022–
Augmentation Project. 2023. In the meantime, Eskom is awaiting
issuance of the water use license by the
Department of Water Affairs.
Disposal of Ash in the form of coal waste was to be
solid waste disposed of in engineered ash storage
facilities with the minimum amount of water
necessary to avoid dust formation.
Physical and Graves found on the route for the Majuba Eskom handled the issue in accordance with
cultural Rail the provisions of the South Africa Heritage
resources Act, which was found to be consistent with
the World Bank’s OP/BP 4.04 on Natural
Habitats.
Projects on Water for the MCPP was to be sourced from Notification requirements were met at the
International tributaries of the Limpopo river, which project preparation stage (2010) and no
Waterways borders South Africa and Zimbabwe and flows concerns have since been expressed.
through Mozambique and therefore riparian
states had to be notified.
Natural Several sensitive natural features such as Managed through adjustment of rail line
Habitats springs and seeps found on the Majuba Rail routing of the Majuba Rail
route. Eskom analyzed alternative alignments for
transmission lines based on inputs from
public consultations to avoid critical natural
habitats areas.
Social Absence of regulatory requirement for Eskom had audits carried out on all
Safeguards preparation and public disclosure of resettlement actions that had occurred and
resettlement action plans and/frameworks. publicly disclosed a Resettlement Policy
No requirement in South Africa’s regulations Framework (2009) for future activities under
for implementing agencies to document and the project.
publicly disclose the impacts of their
resettlement actions. Eskom’s practice of acquiring land on a willing
buyer‐willer seller basis at market prices was
confirmed.
For the MCPP, Eskom purchased two game
farms and only one laborer had to be
relocated at the farm owner’s expense.
Page 47 of 102
The World Bank
For transmission lines, Eskom continued its
practice of selecting corridors that avoid or
minimize the need for relocation of
households or farm structures and avoid
adverse effects on livelihoods or economic
activities and carrying out land valuation to
determine appropriate compensation for
acquisition of right‐of‐way for construction
and maintenance.
No major issues were encountered during
implementation, except for some delays in
obtaining environmental licenses.
131. The 2011 Inspection Panel (IP) Report identified a number of what it considered as shortcomings
related to safeguards. These “shortcomings” and the manner in which they were addressed in this ICR
were as follows:
(a) The IP found that the equivalence analysis under the SDR did not adequately identify the gaps
between the World Bank’s policy requirements and South Africa’s legislation and that the
acceptability analysis did not assess the capacity and implementation track record of key
regulatory institutions. This ICR concluded that no significant issues arose regarding South
Africa’s regulatory framework for safeguards except for the regulation of SO2. The DEA has
appropriate frameworks and tools for management of both project‐specific and cumulative
impacts.
(b) The IP found shortcomings in the assessment of air quality issues and development of
responsive measures in that no consideration was given to impacts of future projects in the
area. However, the ICR noted that: (i) in anticipation of further residential and industrial
development in the region around Medupi on 15 June 2012 the DEA Minister declared the
Waterberg–Bojanala Priority Area (WBPA) as the third National Priority Area (DEA, 2012a)
for pollution control after the IP had issued its report on November 21, 2011 – a declaration
that enabled more extended scrutiny and monitoring; (ii) the World Bank closely monitored
Eskom’s compliance with emissions regulations, particularly for particulate matter and SO2;
and (iii) Eskom complied with the regulations for SO2, but only because it received an
exemption and will continue to seek exemptions until it has installed the FGD. Regarding
particulate matter, limits were exceeded on a regular basis but have come down recently.
(c) The IP found inadequate consideration of the project’s direct, indirect, and cumulative
impacts on availability and quantity of surface and groundwater resources and that
consideration of impacts was not based on a risk‐averse approach. The IP determined that
the World Bank’s assessment of the project prior to Board approval was too narrowly
focused on ensuring that Medupi had a reliable source of water supply and that impacts
related to the plant were managed, but paid insufficient attention to the potential impacts
on quantity and quality of surface and groundwater resources for other users and the
environment and in particular was not satisfied that these impacts were adequately
Page 48 of 102
The World Bank
addressed in the documentation related to MCWAP Phase 1 and Phase 2, the expansion of
the operations of the Grootegeluk Mine, and additional river‐bed sand excavation from the
Mokolo River for Medupi. The ICR noted, however, that the ESIA had determined the
required water use licenses and the relevant conditions, and that during project
implementation Eskom had hired various consultants, including the Institute for Ground
Water Studies of the University of Free State, to monitor surface and underground water
resources. The monitoring revealed that underground water tended to be widely variable
and of generally poor quality due to natural causes rather than project‐level activities. Eskom
also has a long‐term plan for monitoring surface water flows in the Medupi area.
132. At project closure, the two main outstanding safeguards issues were implementation of Eskom’s
action plan for installation of the FGD units at the MCPP, starting with the first unit in 2026 and completing
the installation of all units by 2027 and completion of arrangements for the resettlement of two families
affected by the Majuba Rail Project. The World Bank is committed to monitoring implementation of these
actions until they are satisfactorily completed.
133. Overall, safeguards issues have been well managed and rated Moderately Satisfactory in the last
ISRs.
Fiduciary
134. Financial management was rated Moderately Satisfactory for most of the project implementation
period and remained so at project closure. The only issue is the continuing audit qualification of Eskom’s
financial statements. The basis for the audit qualification is that the auditors are unable to confirm
whether the accounts fully record all irregular expenditures associated with contracts that they had
previously deemed irregular. The audit qualification has been issued for the past four years and will cease
once the relevant contracts have been closed or the National Treasury has condoned them. Eskom has
been pursuing condonation with the National Treasury, but the latter has not yet issued a decision on the
matter. There were no overdue audit reports at the time of loan closure.
Procurement
135. The bulk of the contracts were already awarded before the World Bank was invited to participate
in financing the project. The World Bank carried out due diligence on the awarded contracts to ensure
that they were awarded in a transparent manner and in accordance with the principles of economy and
efficiency. The World Bank’s Board gave a one‐time exception from application of the World Bank’s
procurement policies and guidelines to allow these contracts to be financed by the World Bank given the
urgent need to proceed with the project. The World Bank’s fraud and corruption and right to audit
provisions were incorporated into all such Medupi‐related contracts. It was also agreed that outstanding
consultant contracts were to be awarded in accordance with the World Bank’s guidelines.
136. For the contracts that were bid after the World Bank’s decision to participate in project financing,
there were several procurement issues. There were delays in bidding and contract awards due to the
requirements to meet the GoSA’s regulatory approvals under the PFMA and PPPA. For example, on the
Majuba Rail Project, the issuance of bidding documents for the main civil works contract was delayed
while the PFMA prenotification approval was awaited, and the contract award had to also wait for about
Page 49 of 102
The World Bank
three months for approval under the PFMA. There were a few cases of non‐responsive bids (transmission
lines, one package on the Sere Wind Project, and CSP) that resulted in retendering, except for CSP, which
Eskom decided to drop. Procurement for CSP was affected by the challenge of coordination among the
five financiers to agree on a procurement strategy and on procurement documents consistent with each
agency’s policies and procedures, delayed granting of a waiver under the PPPA, and receipt of only two
non‐responsive bids.
137. There was one case in which the World Bank concluded that the selection of an owner’s engineer
on the Majuba Rail Project had not been brought to successful conclusion for reasons beyond Eskom’s
control. An Iranian‐South Africa joint venture was selected for the contract award, but the issue arose
that Eskom would not be able to make payments to an Iranian firm due to international sanctions against
Iran. As a result, the World Bank was not able to finance the contract but concluded that there were
legitimate reasons to agree to Eskom’s request for reallocation of the loan proceeds.
138. Following completion of all procurement activities, the procurement rating at project closure was
rated Moderately Unsatisfactory due to the outstanding procurement of the BESS.
C. BANK PERFORMANCE
Quality at Entry
139. The World Bank was requested to support financing of the project after Eskom had already started
implementing it and, therefore, the engineering design was completed for the MCPP, several contracts
were awarded, and some environmental impact assessments were completed while others were in
progress.
140. The World Bank had to conduct due diligence on processes that had already been completed to
determine eligibility for World Bank financing. These included (a) due diligence of contracts that were
awarded for the MCPP, which accounted for about 58 percent of financing allocated to all contracts under
this component; (b) seeking of an independent review of the technical design of the plant; and (c)
conducting of an SDR to determine the equivalence of South Africa’s country systems to the World Bank’s
polices and their acceptability for use with the project.
141. There were some apparent shortcomings in the preparation of the project as follows:
(a) PDO, design, and Results Framework. There were some shortcomings in the PDO in that it
conflated project‐level objectives with high‐level objectives. Specifically, the project’s
contribution to economic growth objectives related to a higher‐level outcome to which the
project’s contribution could not be easily assessed objectively. The project design included
a CSP activity as a pilot but set the expectation that it would contribute to the objective of
enhanced power supply and energy security and reduction of carbon emissions, which was
ambitious for a technology that had not reached maturity at the proposed scale. The original
project indicators were set to measure systemwide outcomes rather than project‐specific
outcomes (installed capacity as a percentage of peak demand and carbon intensity of
electricity). The indicators, as revised at restructuring, measure project‐specific impacts but
require use of supplemental measures to properly capture outcomes.
Page 50 of 102
The World Bank
(b) Assessment of Eskom’s implementation capacity. The World Bank’s assessment, supported
by the due diligence by an external consulting firm, concluded that Eskom had adequate
capabilities to manage and operate the MCPP. Several facts suggest that there were risks in
this assessment as follows: (i) Eskom had not implemented a project of such scale for at least
a decade; (ii) the MCPP was being implemented in parallel with Kusile, another equally large
project; (iii) the external due diligence had noted that the planned staffing for the
operational phase was small relative to the size of the plant; and (iv) Eskom’s decision to
manage construction and contract management through METs, which included personnel
from an external project management consulting team instead of having in independent
owner’s engineer. These factors should have triggered a risk identification and mitigation
measures. The World Bank could have insisted on the need for Eskom to retain an
independent owner’s engineer, especially given that the MET model had not been used in
recent times on a project of this scale and complexity.
(c) UCS. The SDR concluded that (i) the MCPP did not meet ambient air quality emission
standards under the South African regulations without installation of the FGD equipment
and (ii) there was an absence in the South African regulatory framework and Eskom’s policies
and procedures of a requirement to prepare and publicly disclose resettlement action plans
and/or frameworks and for implementing agencies to document their evaluations of the
impacts of resettlement actions. Appropriate gap‐filling measures were identified and
agreed with Eskom. With hindsight, the specification of the FGD (a complex and water‐
intensive technology) as the only option for ensuring compliance with SO2 emissions
standards could have diverted attention from other easier‐to‐implement solutions. Because
the South African regulations on air emission were changing, it was not clear against which
emission standards the MCPP was to be assessed and whether compliance with South
African regulations would guarantee alignment with the World Bank’s Environment, Health
and Safety Guidelines (EHSG). The World Bank could have considered seeking agreement on
the MCPP complying with the higher of the South African emission standards and the World
Bank’s EHSG.
142. The World Bank’s overall performance at entry was Moderately Unsatisfactory as there were
significant shortcomings regarding (a) the PDOs, project design, and Results Framework; (b) the
assessment of Eskom’s implementation capability and the design of risk and mitigation response
measures; and (c) the management strategy for ensuring that the project would comply with the World
Bank’s standards for EHSG even if South African regulations became lax or exceptions were granted to
Eskom (as happened in practice). While extensive preparation work was undertaken by the World Bank,
including various due diligence assessments, the identified shortcomings were significant and had adverse
impacts on the overall project outcome.
Quality of Supervision
143. Several features of the supervision work show a strong implementation focus. However, other
aspects reflect the problem of overcoming the legacy of the project’s preparation deficits. Both aspects
are summarized as follows:
Page 51 of 102
The World Bank
(a) Overall adequacy of supervision inputs. The PAD contains a credible implementation
support plan (PAD, annex 6), which detailed the timing of supervision missions, skills
composition, timing for the midterm review, and proposed budget allocation. The project
had three task team leaders over a period of about 11 years,55 two were based in South
Africa for an extended period. The task team leader at project preparation had two tenures
during the implementation period. The core team remained reasonably stable during the
task team leader transitions. As the implementation issues increased, adjustments were
made to augment the team by adding (i) a senior procurement specialist based in South
Africa in addition to the lead procurement specialist based in Washington, DC; (ii) a senior
energy specialist to focus on the Medupi transmission lines activities; and (iii) a specialized
engineering consultant to provide technical solutions to implementation problems. The
continuity in task team leadership and stability of the core team membership helped ensure
a continuity of approach on issues from the World Bank’s side, including overall good
cooperation and problem discussions for solutions with the AfDB.
(b) Supervision of safeguards. The World Bank’s supervision of safeguards was Satisfactory.
Substantive annexes detailing status of safeguard issues and actions agreed with Eskom or
proposed by the World Bank for follow‐up were standard features of Aide Memoires.
(c) Project reporting. The project supervision records—Aide Memoires, ISRs, ad hoc
management memos, and presentations—indicate high‐quality project implementation
support. ISRs were filed regularly at intervals of six to nine months.
(d) Coordination with the AfDB on the MCCP. The AfDB financed the major boiler and turbine
contracts for the MCPP and, therefore, the World Bank worked closely with them to present
a uniform approach with Eskom. Initial problems arose from the use of virtual instead of site‐
specific designs; these were known at the time the World Bank prepared the project.
However, it was only until boiler defects were discovered in Unit #6 in 2015 that the issue
started to get focused attention from the World Bank. Had the World Bank and the AfDB
coordinated a risk assessment based on what was known at the time of the World Bank’s
preparation of the project, and when the boiler issue was first detected remedial actions
could, perhaps, have been identified and implemented before the installation of Unit #6 was
completed or before other units were completed.
(e) The World Bank’s proactivity. Based on quarterly progress reports submitted by Eskom, the
World Bank team regularly discussed action plans for addressing problems encountered
under the various components at different times. Eskom’s acceptance of the World Bank
team’s concerns and recommendations on various implementation issues was initially low,
but this improved over time as both parties became more familiar with each other’s
operating procedures and practices. Apart from the issue of the MCPP defects, progress was
slow regarding implementation of the FGD, CSP, BESS, and Majuba Rail.56 All these activities
were delayed, in part due to delays in getting approvals under the PFMA and the Preferential
Procurement Policy Framework Act (PPPFA) No.5 of 2000 or from the National Energy
Regulator of South Africa. The World Bank team engaged at high levels within the GoSA on
these issues, but problems were not fully resolved. Cancelation of any or all these
55 From Board approval, excluding the preparation period.
56 A project that had an estimated ERR of about 19 percent is now ending up with a negative return due to excessive delays.
Page 52 of 102
The World Bank
components, which were an integral part of the justification for the World Bank’s support of
a coal‐fired power project, would have undermined the rationale for the World Bank’s
support for the project. Invoking the legal remedy of suspension of disbursements was
considered due to Eskom’s noncompliance with the obligation to submit plans to the World
Bank for implementation of the FGD. However, the World Bank’s senior management was
unwilling to approve this measure, which would have been immensely disruptive to the
implementation of the ongoing contracts and to the World Bank’s long‐term engagement in
the sector.
(f) Eskom’s implementation capacity. The World Bank team witnessed the problems faced by
Eskom in effectively managing construction, contract administration, and transitions to
commercial operation for individual units at Medupi. This was due to the limited resources
and few available experts relative to the size of the tasks. As the problems increased, it
became evident that the model of integrated consulting services in the METs was not
effective and/or Eskom did not have enough qualified and experienced staff relative to the
tasks. Eskom was continuing to lose experienced staff to the private sector. The World Bank
could have insisted on Eskom hiring an owner’s engineer to coordinate all contracts on the
MCPP in concert with the embedded engineer—if not at the beginning of the project at least
when problems were mounting during project implementation.
Justification of Overall Rating of Bank Performance
144. Overall World Bank performance is rated Moderately Unsatisfactory. This is based on a
Moderately Unsatisfactory rating for both quality at entry and supervision. Performance at entry is rated
Moderately Unsatisfactory as there were significant shortcomings regarding; (a) PDOs, project design, and
Results Framework; (b) assessment of Eskom’s implementation capability and the design of risk and
mitigation response measures; and (c) management strategy for ensuring that the project would comply
with the World Bank’s standards for EHSG even if South African regulations became lax or exemptions
were granted to Eskom (as happened in practice). The World Bank’s supervision inputs were generally
adequate with reasonable continuity of task team leadership, adequate and timely monitoring and
reporting, understanding of the defects with suggestions made (some of which Eskom acknowledged and
complimented the World Bank on), and appropriate team skills composition. The World Bank’s overall
performance during implementation is Moderately Unsatisfactory, as there were significant shortcomings
regarding the lack of early coordination with the AfDB to investigate the causes of the initial boiler issues;
the slow pace at which the World Bank was able to gain the position of trusted adviser to Eskom; low level
of success in resolving the problem of slow decision‐making within the GoSA’s agencies, which impeded
implementation of several project activities; and the lack of impact of the World Bank’s advice for the
resolution of Eskom’s limited project implementation capabilities relative to the scale and complexity of
the MCPP.
D. RISK TO DEVELOPMENT OUTCOME
145. Risks to development outcome refer to the risks that the outcomes that have been achieved at
project closure may not be sustained. The project has partially achieved its key objectives of enhancing
power supply and energy security and supporting the country’s carbon mitigation strategy, but significant
Page 53 of 102
The World Bank
risks remain that the outcomes may not be sustained or enhanced by improving performance of the
MCPP. The risks are both strategic and operational.
146. The continual loss of expertise to the private sector during the project implementation period is
a limiting factor to Eskom’s ability to efficiently operate, not only the MCPP, but other plants in its fleet.
Thus, at a strategic level, Eskom requires a strong plan for hiring and retaining specialized staff for the
operation and maintenance of its facilities. Eskom is also focusing its efforts on implementing a plan for
curing the MCPP’s latent defects in parallel with measures to reduce UCLFs, which would, in turn improve
EAFs. To realize this goal Eskom needs a strategic plan to determine the allocation of responsibilities and
costs for fixing the plant defects between itself and the contractor. Execution of the strategy needs to
continue and reach completion as quickly as possible to enable the MCPP to generate energy at levels
consistent with its design purpose. Without implementation of effective solutions, the risk exists that
with the age of the plant, the EAFs my fall even further below Eskom and industry targets.
147. At the operational level inadequacies of procurement and chain supply management procedures
to ensure the availability of critical spare parts and financial constraints are limiting Eskom’s maintenance
of Eskom’s power system as a whole, including the MCPP, thus contributing to the power outages. Eskom
if fully aware of these risks and is working to develop solutions.
148. Despite these serious risks to development outcomes of the EISP, the World Bank’s engagement
in the sector over the past 11 years has helped strengthen working relations with client partners in the
sector and in the GoSA and has built a strong foundation for further cooperation. As a result, the GoSA
and Eskom have shown interest in working with the World Bank on new initiatives, such as the
decarbonization/repurposing of existing old coal‐fired power plants, and on broader sector and Eskom‐
specific reforms.
V. LESSONS AND RECOMMENDATIONS
149. The preparation and implementation of the EISP offers a large number of lessons for South Africa,
the World Bank and other interested stakeholders. The ICR reflects only the main lessons and discusses
them under several categories as described below:
Decisions by financiers to participate in projects involving technical and policy complexity
(a) Balancing trade‐offs in the face of conflicting policies is essential for reaching credible
decisions: The EISP posed a dilemma to the World Bank whether to support a coal‐fired
power plant at a time when it had just committed to the global agenda on climate change
and its policy position on financing coal generation plants was evolving. The costs of
participating in the financing of a carbon emitting project were carefully weighed against
the perceived developmental benefits of the project to South Africa and its regional
neighbors, and the potential for the World Bank to influence the development of clean
energy in South Africa through this engagement. Use was made of external experts to
ensure that management decision was based on objective analysis of the costs and benefits,
thus lending credibility to the final decision to finance the EISP.
Page 54 of 102
The World Bank
(b) Rushed preparation of projects identified as urgent can lead to larger problems than those
they are designed to solve: The Medupi and Kusile projects were rushed responses to
counter the looming power crisis due to poor generation capacity planning. As a result,
short‐cuts, such as the use of a “virtual design” from another plant implemented more than
ten years before were used for the MCPP and these turned out to be unsuitable for the
project. The project that was rushed to deliver energy in five years was completed only
eleven years later and even then, is delivering power at levels well below targets and
industry benchmarks. The World Bank and other financiers should undertake more in‐depth
due diligence analysis on “urgent” projects than usual to ensure that normal standards of
preparation adequacy are reached and that implementation timeframes are feasible.
(c) The World Bank should seek changes in project design and project management
arrangements, if suggested by due diligence, before committing to participating in
ongoing operations. When the Bank was invited to participate in the MCPP Eskom had
already decided on project design and project management arrangements. The due
diligence produced generally satisfactory conclusions, but also identified specific issues
such as the small size of Eskom’s project operations staff relative to the size of the project.
The World Bank was aware of the limitation of Eskom’s project management arrangements,
and specifically the decision not to use a FIDIC engineer. However, the Bank did not insist
on changes to these aspects at project preparation, perhaps because Eskom had already
committed to the arrangements. However, the purpose of due diligence is to provide
information for financiers to decide whether to participate or not and/or to seek changes
needed for participation. In this case compromising on these essential conditions for
successful project outcomes contributed to the eventual implementation and outcome
shortfalls.
Lessons on procurement and contract management
(d) Complex projects with numerous contract packages present difficult coordination
challenges even for the most sophisticated borrowers and more so when a new
technology is involved and there is an urgency to implement the project. Using an owner’s
engineer according to Fédération Internationale Des Ingénieurs‐Conseils (FIDIC) conditions
of contract is advisable in such circumstances and normally provides better results. An
alternative approach to multiple contracts is a turnkey approach, which is less demanding
on the owner’s project management capabilities. This approach is preferred by US utilities
and in many countries. The MCPP had about 38 main contracts and more than 400 contracts
at peak when subcontractors are included. Instead of an owner’s engineer, Eskom’s
approach was to set up METs comprising the company’s staff and embedded personnel of
a project management firm retained for the purpose. The model did not work well as was
expected partly because the embedded consulting personnel had no formal authority
equivalent to that of a FIDIC engineer and because of the lack of adequate qualified and
experienced staff in Eskom.
(e) A basic criterion for selection of contractors, among several others, is to ensure that the
contractor has adequate capacity to carry out the specific task in question after
considering the contractor’s total commitments. Eskom was implementing the MCPP in
Page 55 of 102
The World Bank
parallel with another large plant of the same size—4,800 MW Kusile coal‐fired power plant.
Thus, Eskom was implementing two plants for a total capacity of 9,600 MW (12 x 800 MW
units). Some of the contractors for the main contracts, such as the contractor for the boiler
package, were awarded contracts for all the 12 units—an unusual practice in the power
industry with a huge load on the contractor’s ability to perform, which, in fact, turned out
to materialize with some inferior design and poor response performance. This unusually
heavy workload for a single contractor probably accounted for the contractor’s
performance problems on the MCPP’s boiler package.
(f) Detailed knowledge of industry design standards and specifications for equipment
through own staff and/or through experienced external engineering consultant services
is critical for ensuring that specifications are consistent with the industry’s capacity to
deliver. Eskom prepared detailed bidding/contract specifications with the objective of
reducing quality problems based on many years of special experience with coal‐fired plants.
However, some OEMs have standard designs for their equipment, which are either
expensive or not possible to change. The case with the instrumentation and controls was
an example, where the OEM just refused to comply with Eskom’s specifications. Eskom then
canceled the contract, causing huge delays and a mix of controls in the plant to this day.
Other contractors, especially local ones, realized that they could not meet Eskom’s
specifications and did not bid, leading to few bids being received for some packages. In such
cases, Eskom could have requested waivers under the Public Procurement Preferential Act
on local content requirements and used international bidding procedures. A further problem
was that due to shortage of experienced staff and complexity of Eskom’s project
management arrangement, some deviations from specifications were not detected early,
which made making changes difficult and costly and contributed to delays.
(g) Preferential treatment of local contractors to build local capacity and employment is not
always the best option. This should be assessed for each project depending on the urgency
and complexity of the project as trade‐offs may be required between a focus on building
local capacity and skills and one on speedy completion and quality. This would result in
avoiding the impacts to both the project and the contractors who lost resources, some even
facing bankruptcy. Local preferential treatment also requires that local contractors are
properly trained in relevant disciplines—bidding, project and contract management, and
financing.
Lessons from implementation experience
(h) Understanding and securing agreement with the borrower and government agencies on
the timelines for critical decisions on their part is essential for successful project
implementation. Even though the World Bank had no prior experience of working with
Eskom and the South African Government, it would have been possible to review the
processes and timelines for key approvals or authorizations needed for implementing the
project activities, such as those required under the PFMA or for licensing of renewable
investments by NERSA. Specific decision timelines could then have been negotiated with
the GoSA agencies for key activities and incorporated in implementation action plans.
Page 56 of 102
The World Bank
(i) Complex projects involving multiple financing agencies require substantial coordination
to among the financiers to ensure common approaches on implementation issues and
consistent advice to the Borrower. The coordination of efforts works best when financiers
take a holistic view of the project and are able to share views on all project components and
not confine their inputs to own‐financed activities. This approach enables optimizing of use
of expertise available within the financiers’ teams to address implementation issues. Had
the World Bank and the AfDB coordinated a risk assessment based on what was known at
the time of the World Bank’s preparation of the project, remedial actions could perhaps
have been identified and implemented before the installation of Unit #6 was completed.
(j) For large, complex operations requiring a long‐term engagement between the World
Bank and the borrower, a multi‐phase programmatic approach may be tailored to the
borrower’s capacity. This approach provides the borrower a degree of certainty on
availability of financing while simultaneously providing flexibility for adaptation as
conditions change. While this is a good approach, it was not feasible in this case given that
almost all the major contracts had already been awarded and the project was under
implementation when the World Bank was invited to participate in its financing.
(k) A long‐term engagement in a sector and at the country level beyond individual projects
can provide benefits to both the World Bank and the country through continual
generation of new developmental initiatives. Despite its implementation problems and its
moderately unsatisfactory outcome, the EISP has provided the platform to engage with the
borrower, leading to new initiatives such as those on decarbonization, sector, and Eskom
reforms. In contrast, the application of remedial measures to enforce compliance with the
EISP loan covenants would have possibly led to disengagement and loss of opportunities for
cooperation in these new sector development initiatives.
Public‐Private Partnerships
(l) Carefully prepared and packaged public sector owned renewable energy projects can
produce results comparable those of IPPs of the same vintage. The Sere wind power plant
was constructed efficiently at lower cost than the budget due to careful micro‐siting, sound
contracts packaging, and use of an independent owner’s engineer. However, the CSP was
designed as a pilot technology that was not fully matured, with high capital costs and a low
ERR, thus requiring concessional funding. An important lesson from the CSP experience is
that implementing a new technology even on a pilot basis requires careful planning and
should not be attempted on a tight schedule.
Use of country systems
(m) The UCS is important for building capacity in the borrowers’ systems to improve
effectiveness of public expenditures, but the UCS is not appropriate when the regulatory
system is subject to frequent changes and/or involves substantial discretion for
compliance exemptions and/or postponements. At the time of project appraisal, emission
standards applicable to the MCPP were not available. The DEA promulgated the emissions
standards on March 31, 2010 (just a week before Board approval of the EISP), requiring the
Page 57 of 102
The World Bank
existing power plants to meet specified minimum emissions standards by April 2015 and
then meet more stringent emissions standards by April 2020. On June 5, 2013, Eskom
publicly disclosed its intention to request postponement or exemptions for compliance for
some of its power stations, including Medupi, indicating that it would be unable to comply
with SO2 emissions in 2015 or 2020. The current standard which Eskom should comply with
is 500 mg per Nm3 starting in 2025. However, Eskom applied for and was granted a higher
limit of 1,000 mg per Nm3 in April 2020 but has since applied for a still higher limit of 4,000
mg per Nm3 by 2030. The lack of constancy in emission standards made it difficult during
implementation for the World Bank staff to know the applicable emission standards for the
MCPP. Eskom has noted that this needs to be understood in the overall context in which
emission standards are relatively new and have evolved with very little change in South
Africa. While this is true for Medupi it is also equally true that the 3,500 mg per Nm3, has
not changed due to Eskom’s request for postponement of due date as the company could
not meet the 2015 or 2020 dates. A more fundamental issue is the potential for large
deviations from the World Bank’s HSE Guidelines when country systems allow wide
discretion for regulatory exemptions and/or postponements. Given the potential for large
deviations between country systems and the World Bank’s HSE Guidelines, it may be
difficult to establish equivalence between the two in which case the UCS may not be
appropriate for this safeguards policy.
Page 58 of 102
The World Bank
ANNEX 1. RESULTS FRAMEWORK AND KEY OUTPUTS
A. RESULTS INDICATORS
A.1 PDO Indicators
Objective/Outcome: Enhance power supply and energy security
Formally Revised Actual Achieved at
Indicator Name Unit of Measure Baseline Original Target
Target Completion
Generation capacity installed Megawatt 0.00 4,800.00 4,800.00 4,000.00

and commissioned (Medupi)
31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021
Comments (achievements against targets):
While all the units were commissioned by loan closure on June 30, 2021, except for the last unit which was commissioned in July 2021, the main shortfall
was the significantly lower than expected energy outputs and the substantial defects which will take time and additional resources to resolve.
Target Completion
Generation capacity installed Megawatt 0.00 100.00 100.00 100.00

and commissioned from
renewable energy (Sere Wind 31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021
Farm)
Page 59 of 102
The World Bank
Target was fully achieved.
Objective/Outcome: Support longterm carbon mitigation strategy
Target Completion
Direct CO2 emissions avoided Metric ton 0.00 238,000.00 238,000.00 315,330.00

under the project (per year)
31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021
The target was exceeded.
A.2 Intermediate Results Indicators
Component: Medupi Power Plant
Target Completion
Direct project beneficiaries Number 0.00 3,600,000.00 5,500,000.00 4,234,100.00
31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021
Page 60 of 102
The World Bank
Female beneficiaries Percentage 50.00 50.00 50.00 50.00
The target was not met due to delays in commissioning the plant and due to its low energy availability factors.
Target Completion
Quarterly Medupi Yes/No No Yes Yes Yes

Environmental Monitoring
Committee meetings held 31‐Mar‐2009 31‐Oct‐2015 30‐Jun‐2021 30‐Jun‐2021
and minutes distributed.
Fully achieved.
Target Completion
Medupi construction Percentage 0.00 100.00 100.00 99.95

progress rate towards
completion 31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021
Page 61 of 102
The World Bank
Not fully achieved due to delayed completion and COD of the last unit.
Target Completion
Number of units Number 0.00 6.00 6.00 6.00

synchronized
31‐Mar‐2009 31‐Oct‐2015 30‐Jun‐2021 30‐Jun‐2021
Fully achieved.
Target Completion
Transmission lines completed Kilometers 0.00 1,020.00 1,020.00 941.90
31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021
Not fully achieved due to the delayed completion of two transmission line sections.
Indicator Name Unit of Measure Baseline Original Target Formally Revised Actual Achieved at
Page 62 of 102
The World Bank
Target Completion
Transmission lines progress Percentage 0.00 100.00 100.00 95.88
rate towards completion
31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021
Not fully achieved due to delayed completion of two sections. The two sections are scheduled for commissioning in March and May 2022.
Target Completion
Number of units handed over Number 0.00 6.00 6.00 5.00

to commercial operations
31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021
Not fully achieved because COD was for the final unit delayed until July 2021.
Component: Renewable energy (Sere Wind Farm and Battery Storage)
Target Completion
Number of wind turbines Number 0.00 46.00 46.00 46.00
Page 63 of 102
The World Bank
erected 31‐Mar‐2009 31‐Oct‐2015 30‐Jun‐2021 30‐Jun‐2021
Fully achieved because plant was completed in 2015/16.
Target Completion
Progress rate towards Percentage 0.00 100.00 100.00 100.00

completion (Sere Wind Farm)
31‐Mar‐2009 31‐Oct‐2015 30‐Jun‐2021 30‐Jun‐2021
Fully achieved.
Component: Support for low carbon energy efficiency comps., comprising the Majuba Railway for coal transportation & TA prog. for energy efficie ncy
Target Completion
Progress rate towards Percentage 0.00 100.00 100.00 97.50

completion for Majuba rail
project 31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021
Page 64 of 102
The World Bank
Not fully achieved and now scheduled for completion in 2022.
Target Completion
Number of studies for power Number 0.00 7.00 7.00 7.00

plant efficiency
improvements completed 31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021
Fully achieved.
Target Completion
Project generation capacity Megawatt 0.00 3,171.00 3,171.00 4,557.00

savings
31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021
Target exceeded.
Indicator Name Unit of Measure Baseline Original Target Formally Revised Actual Achieved at
Page 65 of 102
The World Bank
Target Completion
Majuba coal transportation Text 93.00 20.00 20.00 190.00
cost
31‐Mar‐2009 31‐Oct‐2015 30‐Jun‐2021 30‐Jun‐2021
Not achieved.
Target Completion
Energy Storage Capacity Megawatt 1.00 200.00 200.00 0.00

Installed and Commissioned
27‐Nov‐2018 31‐Dec‐2020 30‐Jun‐2021 30‐Jun‐2021
Not achieved. Implementation has not started and activities will likely be supported under another project.
Page 66 of 102
The World Bank
B. KEY OUTPUTS BY COMPONENT
Objective/Outcome 1: Enhance power supply and energy security
1. Generation capacity installed and commissioned at Medupi
Outcome Indicators
2. Generation capacity installed and commissioned from Sere Wind Power
1. Medupi construction progress
2. Transmission lines constructed
Intermediate Results Indicators 3. Transmission lines progress rate towards completion
4. Number of units handed over for commercial operation
5. Progress rate towards completion of Sere wind farm
Key Outputs by Component 1. 4,764 MW generation installed and commissioned at Medupi
(linked to the achievement of the Objective/Outcome 1) 2. 100 MW generation capacity added at Sere Wind Power
Objective/Outcome 2: Support long‐term mitigation strategy
Outcome Indicators 1. Direct CO2 carbon emissions avoided under the project
1. Progress rate towards completion for Majuba rail project
2. Number of studies for power plant efficiency improvements completed
Intermediate Results Indicators 3. Number of wind turbines
4. Progress rate towards completion of Sere wind farm project
5. Energy Storage Capacity Installed and Commissioned
1. Seven power plant efficiency studies completed
Key Outputs by Component
2. 46 wind turbines installed
(linked to the achievement of the Objective/Outcome 2)
3. 100 MW wind capacity added at Sere
Page 67 of 102
The World Bank
ANNEX 2. BANK LENDING AND IMPLEMENTATION SUPPORT/SUPERVISION
A. TASK TEAM MEMBERS
Name Role
Preparation
Reynold Duncan Lead Energy Specialist/Co‐Task Team Leader
Pankaj Gupta Lead Financial Specialist/Co‐Task Team Leader
Suman Babbar Consultant and Project Finance Advisor
Sandeep Mahajan Lead Economist
Elzbieta Sieminska Lead Procurement Specialist
V.S. Krishnakumar Regional Procurement Manager, Africa Region
Mark Walker Chief Counsel, Africa Region
Edith Ruguru Mwenda Sr. Counsel
Mohammed Bekhechi Lead Counsel, Environment
Charles Di Leva Chief Counsel, Environment
Juan Gaviria Sector Leader
Pierre Pozzo de Borgo Lead Transport Specialist
Harvey Himberg Consultant/ Environmental Specialist
Harvey van Veldhuizen Lead Environmental Specialist
Thomas Walton Consultant/ Environmental Specialist
Frederick Edmund Brusberg Lead Social Development Specialist
Mudassar Imran Sr. Economist
Mustafa Zakir Hussain Sr. Infrastructure Finance Specialist
Andrey Gurevich Financial Analyst
Ahmad Slaibi Young Professsional
Heather B. Worley Communications Officer
Sarawat Hussain Senior Communications Officer
Karan Capoor Senior Financial Specialist
Gert Van Der Linde Lead Financial Management Specialist
Modupe Adebowale Senior Financial Management Specialist
Page 68 of 102
The World Bank
Miguel Oliviera Finance Officer
Andrew Asibey Senior Monitoring and Evaluation Specialist
Armando Araujo Consultant/ Procurement Advisor
Aman Sachdeva Consultant/ Procurement Advisor
Peter Meier Consultant/ Power Economics
Richard Bullock Consultant/ Railway Economics
Vicktor Loksha Consultan/CTF Financing
Susan Shilling Program Assistant
Jemima Harlley Program Assistant
Rita Ahiboh Program Assistant
Supervision/ICR
Franz Gerner, Frederic Verdol Task Team Leader(s)
Reynold Duncan Task Team Leader
Paivi Koljonen Task Team Leader
George Daniel, Chitambala John Sikazwe Procurement Specialist(s)
Tandile Gugu Zizile Msiwa Financial Management Specialist
Mabel Nomsa Mkhize Procurement Team
Mary C.K. Bitekerezo Social Specialist
Fowzia Hassan Senior Operations Officer
Knut J. Leipold Procurement Team
Blessing Manyanda Procurement Team
Petrus Benjamin Gericke Lead Transport Specialist
Jorge Luis Alva‐Luperdi Counsel
China Chhun Program Assistant
Brandon Enrique Carter Environmental Specialist
Ines Perez Arroyo Energy Analyst
Zandile Portia Ratshitanga Senior External Affairs Officer
Gina Cosentino Social Specialist
Lindiwe Dube Procurement Team
Page 69 of 102
The World Bank
Mokgabo Molibeli Team Member
Gabriel Tafirenyika Ngorima Social Specialist
Johanna Martina Whitfield Environmental Specialist
Bruce Paley Engineering Consultant
Peter Meier Power Economist Consultant
Joel J. Maweni ICR Lead Author/Consultant
B. STAFF TIME AND COST
Staff Time and Cost
Stage of Project Cycle
No. of staff weeks US$ (including travel and consultant costs)
Preparation
FY09 9.075 91,587.34
FY10 128.173 1,228,881.54
Total 137.25 1,320,468.88

Supervision/ICR
FY10 12.248 117,717.06
FY11 75.853 602,775.75
FY12 72.609 535,561.99
FY13 54.920 480,270.80
FY14 52.494 458,369.55
FY15 55.220 491,601.27
FY16 45.415 424,231.34
FY17 52.685 462,521.82
FY18 54.083 553,291.11
FY19 74.234 713,375.00
FY20 84.760 739,165.47
FY21 72.790 522,517.90
Total 707.311 6,101,399.06
Page 70 of 102
The World Bank
ANNEX 3. PROJECT COST AND FINANCING PLAN
Table 3.1. Projects Costs by Component as at Appraisal Compared to Actual Costs
Project component Estimated costs at Actual costs at Actual/Estimated
appraisal (US$ m) completion (US$ m) total costs to
completion57 as %
of estimated costs
at appraisal
Medupi Power Plant58 12,048 17,474 145.0
Renewable energy plants of which:
Sere wind power (100M) 445 252 56.0
Kiwano CSP 782 ‐ 0.0
1,227 252 20.6
Energy Efficiency Investments of which:
Majuba Rail 546 402 73.7
TA for coal power plant rehabilitation 20 ‐ 0.0
TA for development & implementation of 10 ‐ 0.0
domestic & cross border RE projects
576 402 69.8
IBRD front end fee and CTF management fee 10 ‐59 0.0
Totals 13,861 18,128 130.8
57 Costs at completion include some elements of estimated costs for activities that were not yet completed and for correction
of defects at the MCPP.
58 Note: The World Bank financing of the Medupi Power Plant included the main civil works, associated transmission lines for
integrating the plant to the grid and power plant equipment (other than the main boiler packages which were financed by the
African Development Bank). The other plant equipment included electrical equipment and switchgear, dust and coal handing
equipment.
59 Included in Medupi Financing
Page 71 of 102
The World Bank
Table 3.2. Project Financial Plan as at Appraisal Compared to Actual/ Revised Plan at Completion
Estimated at Appraisal Actual/Revised costs to
Component/Financier (US$ m) completion (US$ m)
Medupi Power Plant
World Bank Loan 3,040 2,866
African Development Bank ‐ 2,418
China Development Bank ‐ 1,410
BNP (Coface) ‐ 708
CACIB (Coface) ‐ 26
KfW (Hermes) ‐ 982
Eskom 4,083 9,063
Unidentified per PAD Annex 4,925 ‐
Total Medupi Power Plant 12,048 17,474
Renewable energy of which:
Wind Power
African Development Bank Loan ‐ 9
African Development Bank CTF ‐ 42
World Bank Loan 110 23
World Bank CTF 100 35
AFD ‐ 130
Eskom 135 13
Unidentified per PAD 100 ‐
Total for wind power 445 252
CSP
IBRD 150 ‐
IBRD CTF 249 ‐
Eskom 383 ‐
Total for CSP 782 ‐
Total Renewable energy plants 1,227 252
Energy Efficiency Investments of which:
Majuba Rail
IBRD 411 270
Eskom 135 132
Total Majuba Rail 546 402
TA for Coal Plants Rehabilitation
IBRD 20 ‐
Page 72 of 102
The World Bank
Estimated at Appraisal Actual/Revised costs to
Component/Financier (US$ m) completion (US$ m)
TA for dev. and implementation of domestic and
cross border RE projects
IBRD 10 ‐
Total Energy Efficiency Investments 576 402
IBRD front end fee and CTF management fee 10
TOTAL PROJECT COSTS & FINANCING
REQUIREMENT 13,861 18,128
Table 3.3. Summary of World Bank Financing
Components Amount at Actual at Project Percentage of
Approval Closing Approval
(US$, millions) (US$, millions)
Medupi power plant 3,040 2,857 93.9
Renewable energy (Sere Wind Farm and Battery 8.8
260 23
Storage Program)
Support for low carbon energy efficiency 61.2
comps., comprising the Majuba Railway for coal
441 270
transportation and TA program for energy
efficiency
Front‐end fee (IBRD) and manageemnt fee (CTF) 9 9 100.0
Total 3,750 3,159 84.2
Page 73 of 102
The World Bank
ANNEX 4. EFFICIENCY ANALYSIS60
4.1 Key Assumptions for the ICR Economic Analysis
1. Electricity demand growth. The Department of Energy’s 2009 IRP estimated an annual demand
growth of 2.3 percent for the following five years, which included the impact of expected price increases
and demand‐side management measures. Indeed, much of the rationale for Medupi was grounded in the
extended power cuts observed in 2008 and the perceived need to avoid debilitating power cuts in the
short to medium term. By 2019, the demand should have been around 290 TWh, but instead it was only
about 208 TWh. There are several reasons for the decline in demand including the sharp tariff increases
of 2010–2012, apparent greater success of demand‐side management measures than had been
anticipated, and dramatically lower GDP growth than the 4 percent annual growth rate expected in the
2009 IRP.
2. Benefit valuation. At appraisal, benefits were assessed at willingness to pay (WTP), based on
diesel self‐generation, consistent with the assumption that Medupi would be incremental. However, with
demand lower than expected, for the first few years the MCPP has been non‐incremental, replacing
existing coal generation at less efficient coal projects. While this has reduced the baseline benefits, it has
increased the benefits associated with coal project externalities (both GHG and local air emissions).61
3. The appraisal economic analysis was US dollars at constant 2010 prices. The ICR re‐estimation of
economic returns is again at this price level for consistent comparisons to the PAD appraisal estimates.
4. Externalities. At appraisal, the World Bank had not yet issued guidance for the SVC. The World
Bank team assumed a value of US$29 per ton CO2. The World Bank has since issued a guidance document
(Shadow Price of Carbon in Economic Analysis: Guidance Note, November 12, 2017), which is applied to
the economic analysis in this ICR. At appraisal, there was considerable uncertainty regarding the valuation
of local externalities and extrapolation of United States and European studies to South Africa, and hence
only the GHG emissions impacts were monetized and included in the calculation of economic returns.
Since appraisal, several South Africa‐specific studies on the valuation of local environmental damages of
coal generation and coal mining have been published, whose results have been now reviewed in detail
and their results used in the revised ICR economic analysis.
5. World Bank guidance documents. The World Bank’s Energy and Extractives Practice
Department’s new guidance on economic analysis of power sector investment projects issued in 2016,62
and the Operations Policy Department’s guidance on economic analysis, issued in 2013,63 now require
quantification of all relevant externalities and the calculation of NPV without externalities, NPV including
local externalities, and NPV including both local and global externalities. That guidance is followed in this
60 The annex summarizes the detailed economic and financial analysis report prepared for the ICR, available in the project files.
61 The background report discusses this in more detail: the economic returns based on non‐incremental counterfactuals (both
CCGT‐based on LNG and renewables), does reduce the ERR by 2–3 percent, but the returns are still above the opportunity cost
of capital, or the discount rate of 6 percent based on recent World Bank guidance (see footnotes 52 and 53).
62 World Bank. 2016. Guidelines for Economic Analysis of Power Sector Investment Projects, GEEDR.
63 World Bank. 2013. OPSPQ, Investment Project Financing, Economic Analysis, Guidance Note.
Page 74 of 102
The World Bank
ICR. New research on the valuation of local environmental costs of air pollution was published by the

World Bank in 2015, whose methodology is also followed in this ICR.
6. Impact of COVID‐19. Starting in April 2020, COVID‐19 depressed consumption and generation of
electricity. Demand has since improved but remains low and uncertainties remain as to whether the low
demand will persist and for how long. The ICR concludes that even if low demand persists into 2022 and
2023, the MCPP is unlikely to be negatively affected because the MCPP and the Kusile power plant are
the most efficient coal‐fired power plants in the power system. They will, therefore, displace less‐efficient
plants under conditions of low demand.
7. Discount rate. The choice of discount rate is of central importance to the economic analysis as it
determines the hurdle rate for the ERR; that is, the threshold for meeting an efficiency rating of
satisfactory. At appraisal, the discount rate was set at 10 percent.64 This value was chosen for consistency
with the long‐term mitigation study prepared for the DME in 2003. Several other discount rates have been
used in South Africa over the years, including (a) 12 percent by for calculating the feed‐in tariff for
renewables in 2009 and its most recent calculation (in its MYPD4 decision) of the real WACC (as would
apply to economic analysis conducted at constant prices) is 7.1 percent and (b) guidance of 8 percent
issued by Department of Energy for the 2010 IRP.65 The guidance noted that the “..discount rate should
be set at a real (after inflation) rate of 8%, sensitivities should be calculated at 3% and 13%.”
8. The World Bank published its guidance on discount rates in 2016.66 This argues that for economic
analysis, the discount rate should be anchored in welfare economics: application of the Ramsey formula
under the assumption of a value of 2 for the elasticity of marginal utility of consumption results in a
discount rate of twice the rate of per capita economic growth. The guidance notes that across all World
Bank countries, long‐term average per capita GDP growth rates have been around 3 percent, which value
is suggested as a default (with a corresponding discount rate of 6 percent).
9. However, while South Africa’s per capita income growth ranged between 1 and 4 percent in 2000–
2010, the rate has dropped sharply since then and has been negative since 2015. Even under the
assumption of a 4 percent GDP growth rate into the future, given a population growth rate of 1 percent
(that has fallen from over 2 percent in the 1980s to the current 1.3 percent), the per capita GDP growth
rate would be around 3 percent, which calls for a discount rate of 6 percent. This has important
implications for judging the ERR of the EISP (and of Medupi in particular). Under some counterfactuals,
the ERR of Medupi is in the range of 8 to 10 percent—so lower than the appraisal threshold, but
comfortably above the World Bank default recommendation of 6 percent value and above the NERSA
WACC‐based rate of 7.1 percent, or the IRP 2010 rate of 8 percent.
10. Valuation of global externalities. At the time of appraisal in 2009, there was no World Bank‐wide
guidance for the SVC or for the discount rate to be used. Both were left to the best judgement of the
project economist and task team leader. For the SVC, a constant value of US$29 per ton was used and for
64 PAD, Economic and Financial Analysis (annex 8).
65 http://www.energy.gov.za/IRP/irp%20files/IRP%20Parameter%20‐%20S3%20Discount%20Rate.pdf.
66 World Bank, OPSPQ. 2016. Discounting Costs and Benefits in Economic Analysis of World Bank Projects.
Page 75 of 102
The World Bank
the discount rate, 10 percent.67 For the ICR re‐estimation of economic returns, the SVC values in the
guidance document of 2017 were used.68 The NPVs and ERR are presented for both the low and high
valuations in this guidance.
11. Opportunity cost of water. Water consumption at Medupi is a major concern, for which reason
dry cooling rather than evaporative (wet) cooling is the chosen technology. Water shortage is also a major
reason for the delay in adding the FGD. Clearly, the present financial cost of water to Eskom does not
reflect its economic opportunity cost. At the time of appraisal, the financial cost of water to Eskom was
R0.3 per m3. For sake of conservative assumption, at appraisal a value of R 20 per m3 was used in the
economic analysis of Medupi, though not further justified in the PAD.
12. After the PAD issuance, a June 2010 study by Louw and van Schalwyk69 estimated the marginal
value of water to agriculture at R 20 per m3, which matches the value used in the PAD. For the ICR, the
detailed report assessed the opportunity cost of water, including consumptive water use in coal mining
(also unlikely included in the financial cost of water to Eskom), at R 24 per m3, based on photovoltaic (PV)‐
powered, reverse osmosis desalination plus transportation cost from the Mozambique coast.
13. Local environmental impacts. Power generation from fossil fuels gives rise to a range of local

environmental damage costs that need to be included if the true social cost of generation is to be
estimated.70 The key damage costs are (a) impacts associated with fossil fuel extraction, (b) damage costs
of air pollutants, (c) heavy metal emissions, and (d) ash disposal at coal plants.
14. As noted, the local environmental externalities were not quantified for the PAD table of economic
flows, on grounds that the preliminary results of an ongoing study had concluded that these local
externality costs were small when compared to the climate change damage costs. Moreover, there was
no local data available other than simple extrapolations of damage cost estimates in the European Union
and the United States. These were not considered sufficiently credible or reliable to be applied to actual
conditions in South Africa. When that study was finally published in July 2010, with damages estimated as
costs per kWh, this judgement was confirmed.71
15. However, while damage costs expressed as US$ per kWh may be appropriate for high‐level
assessment in an IRP, they are of limited value at the project level. If one is to assess the benefits of highly
efficient supercritical coal compared to the older coal projects in the Eskom fleet (that suffer further from
low efficiency due to age) or the benefits of pollution control devices such as the FGD, the relationship
67 See background report for further assessment of the consistency of SVC damage cost estimates and the discount rate used
for the cost‐benefit analysis. The guidance values are at constant 2017 US dollars and must be converted back into 2010
constant prices (at which level they appear as somewhat lower values than when at 2017 prices).
68 World Bank. 2017. Shadow Price of Carbon in Economic Analysis: Guidance Note.
69 Louw, D. B., and H. van Schalkwyk. 2010. “The Impact of Transaction Costs on Water Trade in a Water Market Allocation
Scheme.” Agricultural Economics Research 40 (4): 780–793.
70 For coal projects using open cycle and evaporative cooling, there are also a range of issues related to thermal effluents which
have potentially damaging impacts on fisheries and local ecosystems, particularly in shallow bays and estuaries. These are not
of concern at an inland dry‐cooled project such as Medupi.
71 Edkins, M. H. Winkler, A. Marquard, and R. Spalding‐Fecher. 2010. External Cost of Electricity Generation: Contribution to the
Integrated Resource Plan 2 for Electricity. Energy Research Center, University of Cape Town.
Page 76 of 102
The World Bank
between emissions and technology is paramount. Obviously, if the FGD reduces SOx emissions by 90
percent but does little to reduce NOx, it follows that
 Air emission damage costs must be considered on an individual pollutant level—separately
for PM10, NOx, and SOX—and
 Emissions must be estimated based on the pollution controls being applied and, on the
population, affected: emissions from uncontrolled diesel self‐generation in densely
populated urban areas (as in the counterfactual) will be much higher per kWh of benefit
than from a project in a relatively remote area with a tall stack (as at Medupi).
16. The economic and financial analysis prepared for this ICR has conducted a detailed examination
of the South Africa‐specific studies over the last decade that have attempted valuation of local air
pollution damages from coal projects, including all the criteria pollutants and heavy metal emissions,
particularly mercury. The literature reviewed includes studies by Greenpeace, the World Bank
Environment Practice, the 2019 IRP, and the peer‐reviewed academic literature relevant to South Africa.
An important study is that of the Kusile project, which is similar in design to the MCPP, prepared by the
University of Pretoria for Greenpeace Africa.72 Table 4.1 shows the results of this review of externality
valuations, as used in the ICR economic analysis for the MCPP.
Table 4.1. Summary of Externality Calculations
R¢/kWh US¢/kWh
Water use in mining 0.475 0.063
Water use at Medupi 1.35 0.18
Health damages from coal combustion (Kusile Greenpeace study) 0.7 0.09
(Edkins Study) 1.35 0.18
Mercury 0.1 0.013
Coal mining 0.293 0.039
Source: EISP ICR, Economic and Financial Analysis Background Report, November 2021.
17. Opportunity cost of coal. At appraisal, the opportunity cost of coal to Medupi was based on a
netback calculated of the Richards Bay South African export price, adjusted for coal quality and domestic
transportation costs. This 2009 forecast, actuals, and the latest, October 2021 World Bank forecast is
shown in figure 4.1.73
72 University of Pretoria. 2011. The External Costs of Coal‐Fired Power Generation: The Case of Kusile. Report to Greenpeace Africa.
73 The World Bank coal price forecast is for Australian coal, FOB Newcastle. The price of South African coal closely tracks that of
Australian coal.
Page 77 of 102
The World Bank
Figure 4.1. World Bank Forecast for Coal
Source: World Bank Commodity Markets Outlook, October 2021.
18. By October 2021, following the sharp increases in 2021, natural gas and coal prices are expected
to decline in 2022 and fall further in 2023, as demand growth eases (especially outside Asia) and
production and exports increase, driven by the United States. Further price spikes are likely, however, as
inventories remain low, and production is not expected to materially increase until 2022. More
importantly for the economic analysis of the MCPP, the present price spike in international coal prices is
matched by a corresponding price spike in oil (and gas): what matters is the relative price of coal to oil (or
gas).
19. As shown in table 4.2, domestic coal prices are still significantly below that of export coal, though
the differential has narrowed since appraisal.
Table 4.2. Domestic and Export Price of Coal
Source: South African rand per ton from Department of Energy, South African Price Statistics, 2018.
Medupi prices from Eskom.
Page 78 of 102
The World Bank
20. The Eskom price forecast for Medupi coal for FY21 is R 505 per ton, which value is used in the
financial analysis and escalated to domestic inflation.
4.2 Economic and Financial Analysis of the Sere Wind Power Plant
21. Economic returns at appraisal. The following assumptions were applied to the economic analysis
of the Sere wind power plant:
 Benefits: US¢17.5 per kWh based on a weighted average consumer WTP
 Transmission and distribution losses of 9.6 percent in 2010 increasing to 10.7 percent by
2015 and 10.7 percent thereafter
 Installed capacity of 100 MW load factor: 25 percent, with annual energy at 219 GWh
 Counterfactual for GHG emission calculations: Emission factor of 1.03 kg CO2 per kWh
(based on the Eskom emission grid factor according to the Eskom 2009 annual report)
 Additional transmission CAPEX of US$7 million, plus US¢2 per kWh for transmission and
distribution cost74
 Value of carbon at US$29 per ton constant
 CAPEX of US$205 million (or US$2050 per kW)
 O&M costs at 1.26 percent of CAPEX increasing by 1 percent per year in real terms
 Economic life: 20 years.
22. Under these assumptions, the PAD baseline ERR was assessed at 10.9 percent, increasing to 14.4
percent when GHG emission benefits were considered (table 4.3).
Table 4.3. Sere Wind Farm ERR at Appraisal (at constant 2010 prices)75
74 No basis for these assumptions was provided in the PAD.
75 At the time of the midterm review in 2015, the economic returns were re‐estimated at 11 percent based on a 34 percent load
factor and capital costs of US$293 million. The details of this recalculation could not be found in the available archives.
Page 79 of 102
The World Bank
23. Table 4.4 shows a comparison of the actual costs with those estimated at time of appraisal. While
the appraisal estimates shown in this table are at constant 2010 prices, to which contingencies and
interest during construction (IDC) have been added, the actuals are in nominal dollars. When the actual
disbursements are converted back to constant 2010 prices, the total nominal cost (of US$227 million)76
recalculates to US$210 million. This is just slightly higher than the US$205 million estimated in the PAD.
In short, the Sere wind farm project has been efficiently implemented.
Table 4.4. Project Costs and Financing
Project Component Appraisal Actuals Percentage of
Estimate Appraisal
Wind turbines and generators, towers, and other equipment 199 195 98
Civil works, roads, and buildings 18 4 24
132 kV lines and substations 58 8 14
Owner’s engineer 20 5 26
Owner’s development costs 15 15 97
Contingencies, 10% of base costs 25 0 —
IDC (excluding Eskom’s balance sheet financing costs) 18 16 91
Total cost of funding 353 243 69
Source: Eskom, Project Close‐Out Report: Sere Wind Farm, 2016, page 53.
24. The revised assumptions for the re‐estimation of economic returns for the ICR are as follows.77
 OPEX:78 Based on the actual reported cost of R 320 per kW per year.
 GHG emissions reduction valuation: Now based on the World Bank 2017 Guidance.
 GHG emission factor: Use of the published grid Eskom emissions factor overestimates the
emissions under the counterfactual used to assess benefits—based on oil self‐generation
the emissions reduction is 0.6 kg per kWh, not 1.05 kg per kWh.79
 CAPEX: Nominal expenditure US$227 million.
 Construction disbursements: Three years (based on actuals).
 Energy: Actuals to 2019, thereafter at average of the first four years.
 Benefit estimate: Based on the revised estimate of WTP (as discussed in section 2.3).
 Transmission losses: At actuals, then 10 percent into future years (the actual 2018 rate is
reported at 9.7 percent.
25. Economic returns at completion. The economic flows at completion are shown in table 4.5. The
ERR before externalities has increased from 10.9 percent at appraisal to 18.3 percent (NPV from US$9
76 Subtracting IDC and financing costs of US$16 million from the total US$243 million.
77 The source for actual costs and generation is Eskom, Project Close‐Out Report: Sere Wind Farm, 2016; generation in subsequent
years from the Eskom annual reports,
78 OPEX = Operating expenses.
79 One could make the case that at the margin, wind power would reduce the most inefficient coal project, so an emissions factor
of around 0.9 kg per kWh. But that too would be inconsistent with the counterfactual used for benefit valuation.
Page 80 of 102
The World Bank
million to US$96 million). The ERR, including GHG emission reduction benefits, has increased from 14.1
percent to 21.1 percent at the low value of the SVC and to 23.4 percent at the high value.
Table 4.5. Economic Returns at Project Completion
26. The 10 percent hurdle rate is reached in 2023 (after eight years of operation)—but already by
2022 at the low SVC and by 2021 at the high SVC (after six years of operation).
Figure 4.2. ERR versus Year
27. The reasons for the higher economic returns are many. The change is calculated for each input
assumption, as shown in table 4.6.
Page 81 of 102
The World Bank
28. Higher‐than‐expected energy and efficient construction are the two main explanations for the

increase in economic returns. The only negative factor is the decline in benefit valuation, given the
decrease in international oil prices over the last decade.80
Table 4.6. Explanation of the Increase in ERR at ICR
No. NPV ERR (%) Change in
(US$, millions) ERR (%)
1 At appraisal As described in the PAD 16 10.9
2 Transmission Slightly lower than expected 17 11.0 +0.1
and distribution
losses
3 Energy Average energy now 329 GWh, 90 15.9 +4.9
higher than 219 GWh at appraisal.
The actual plant factor in the first
four years of operation is between
36 and 38 percent.
4 Benefit Slightly lower over the 20 years of 64 14.9 –1.0
valuation assumed life
5 OPEX Based on actuals, somewhat lower 73 15.5 +0.6
than expected at appraisal
6 Transmission Incremental transmission cost 74 15.7 +0.2
costs evaluated on actual cost of service
analysis (fraction of total cost
attributable to transmission and
distribution)
7 At project More efficient construction (3 98 18.3 +2.6
completion years rather than 4 years), lower
CAPEX
8 Impact of Sharply lower oil prices between 96 17.5 –0.8
COVID‐19 2020 and 2030 due to expected
pandemic‐induced global
economic contraction
29. The impact of the COVID‐19 pandemic has reduced the expected returns because of the lower oil
price forecasts, which affects the cost of diesel self‐generation that underpins the WTP: row (4) illustrates
the impact of the lower oil prices in the October 2019 commodity markets outlook oil price forecast;
however, in row (8) the latest forecast of October 2020 that has lower oil prices in 2020–2029 is used.
30. The ERR that includes GHG emission benefits is significantly higher due to the SVC estimates now
stipulated in the November 2017 Guidance (and also because the energy generation is much higher). At
appraisal, the inclusion of GHG emissions reduction benefit increased the ERR by 3.5 percent (NPV by
US$39 million). At project completion, the ERR increases by 5.1 percent at the high SVC (NPV by US$65
80 Sere started generation in 2016, when the average oil price had fallen to US$51per barrel from its peak in 2014 of US$104 per
barrel. In 2017, the oil price had fallen to US$43 per barrel, so in that year the average benefit was just US¢16per kWh, lower
than the US¢17.5 per kWh assumed at appraisal.
Page 82 of 102
The World Bank
million) compared to the low SVC. Clearly, the wind project has delivered substantial GHG emission
reduction benefits.
31. The Sere wind farm appraisal avoided the widespread overestimation of the wind resource
observed in the first cohort of private wind farms implemented in the so‐called round 1 of the REIPPPP.
32. Operational performance. The Sere wind farm has performed well, perhaps because of the O&M
contract awarded to an experienced operator (Siemens). A good measure of performance is availability.
The South African average, as reported by the Lloyds Register study, is 97.6 percent (almost all of which
are independent power producers [IPPs]) and that of Sere is marginally above that average at 97.7.81 In
other words, the operational performance of the Eskom‐owned Sere is on par with the performance of
IPPs commissioned at the same time.
33. Financial analysis. Because the Sere wind farm is an Eskom project, which is not implemented by
a separate Eskom business unit or otherwise ring‐fenced, there are no project financials that identify
revenue at the plant gate. At appraisal, the financial revenue was based on the tariff under the renewable
energy feed‐in tariff regime, estimated at R 1.21 per kWh in 2010, and on NERSA's October 2009
announcement that it would allow the real WACC (12 percent) plus the rate of inflation (6 percent) to be
covered, allowing an 18 percent FIRR: the R 1.11 per kWh was assumed to rise at the rate of inflation.
34. But no wind projects were financed on this basis, because shortly after the PAD finalization, the
Government announced that the system would be replaced by auctions. Four rounds between 2012 and
2015 enabled around 2,570 MW of wind projects and a gradually decreasing price.82
35. The finances of the Sere wind farm were ring‐fenced, are clearly satisfactory, sustainable, and

obviously much greater than estimated at appraisal. To yield the same 18 percent return on Eskom’s
equity investment, the necessary (internal accounting) tariff is 50 percent lower than estimated in the
PAD. Thus, the main explanations can be summarized as follows:
 Lower cost of debt than expected
 Lower capital cost than expected
 Much greater energy generation than expected.
4.3 Economic and Financial Analysis of the Medupi Power Project
36. The Medupi Coal Project accounts for the major share of the EISP. The major issue for the re‐
estimation of economic and financial returns is the high construction cost overrun and associated delays
and following synchronization of the units, the difficulties experienced with achieving the planned
operational performance targets. The cost overruns are summarized in table 4.7. These are in nominal
terms and include the impact of inflation over the 16 years of construction (from start of construction in
2008 to 2022): increases in real terms are much smaller.
81 As reported in the Project Close‐Out Report in the first year of operation.
82 This is discussed further in the detailed economic and financial analysis report. That report presents a detailed cash flow
assessment of the Sere wind farm which is assessed on the same basis as an IPP.
Page 83 of 102
The World Bank
Table 4.7 Medupi Cost Overruns
PAD PAD ECTC Overrun
US$m ZARb ZARb
Medupi EPC 10,000 75
other costs 797 6
10,797 81 132.2 63%
IDC 1,251 9 44.6 374%
total 12,048 90 176.8 96%
IDC increase over CAPEX 12% 12% 34%
Source: PAD, Annex 5: Project Costs. Actuals according to Eskom. Exchange rate R 7.5 = US$1.at project
appraisal date
Note: ECTC = Estimated costs to completion, EPC = Engineering, procurement, and construction

37. The CAPEX overrun of 63 percent is at the high end of the World Bank’s experience with large
generation projects: comparable cost increases have been observed only in large hydro projects.
However, the increase in IDC of 375 percent over the appraisal estimate is unprecedented. IDC estimates
at appraisal typically account for an additional 10–15 percent over CAPEX for projects with long
construction time (as are typical for hydro projects). However, based on Eskom’s presentation, the share
of IDC at completion is 34 percent.
38. The current estimate of CAPEX is R 132.2 billion, in nominal terms, excluding the FGD component
that was included in the PAD CAPEX estimate. This is unchanged over the estimate at the time of the 2015
midterm review.
39. Given the various still unresolved issues and disputes, the R 9.2 billion allowance for contingencies
and liquidated damages for 2024 has been retained—though normally contingencies no longer appear in
ICRs at project completion. The estimates of economic and financial returns are to that extent
conservative.
40. Economic returns. The returns are calculated in three steps:
(a) Without externalities and without the FGD system: ERR 13.5 percent
(b) Including local externalities, but for the sake of conservative calculation excluding the
benefits from the avoided environmental costs of self‐generation (indoor air pollution, diesel
self‐generation, and kerosene for lighting): ERR 13.2 percent
(c) Revised estimate with the FGD systems in place, ERR 12.9 percent. This is the estimate that
may be compared to the PAD economic return of 24 percent.
41. The main risk to the maintenance of satisfactory economic returns is the failure to reach the EAF
targets (figure 4.3). The importance of reaching the EAF target as soon as possible to maximize the
economic returns is obvious: the later the EAF is reached, the lower the returns.
Page 84 of 102
The World Bank
Figure 4.3. Sensitivity of Economic Returns to EAF
Table 4.8. ICR Economic Returns (at constant 2010 prices)
Page 85 of 102
The World Bank
42. As noted, the counterfactual for the economic analysis assumed that the benefits were
incremental and assessed at the estimated WTP. However, this is only one of several possible
counterfactuals and methodologies for assessing economic benefits. The other options are as follows:
 Application of a power systems planning model. This is the preferred methodology, in
which the economic benefits of a project are assessed by comparison of the NPVs of the
least‐cost plan with and without the candidate project. Certainly, in current practice, a
power systems investment project of this size would not be considered by the World Bank
without a power system planning study. The World Bank now has its own in‐house Electricity
Planning Model that would be used for this purpose where the client does not have
equivalent modeling capability. However, Medupi was clearly identified in the least‐cost
plan in the 2008 IRP.
 WTP based on the tariff. In theory, the tariff is the lower bound of the economic benefit to
consumers because it ignores the additional consumer surplus under the demand curve.
 Comparison with the next best project delivering equivalent benefits. Where a power
system planning model is not available, this is the most frequently used counterfactual.
43. WTP based on tariff. An analysis of the economic returns based on the tariff was presented in the
PAD. Unsurprisingly, this assessment found that returns were negative at the average tariff as it existed
in 2009. Substantial increases were approved by the regulator—24.5 in FY11, 28 percent in FY12, and 25.9
percent in FY13 (nominal), which would result in an ERR of 2.47 percent. If there were further increases
in FY14 and FY15 by 12.5 percent and 1 percent above inflation thereafter, the appraisal analysis found
that the ERR increases to 8.1 percent (all before adjustments for externalities). The point of this
assessment was to emphasize the importance of raising Eskom tariffs.
44. However, the reality was different: while tariffs in US dollar terms indeed doubled between 2009
and 2012—as expected—thereafter tariffs in nominal US dollar terms remained largely unchanged (figure
4.4), implying a decrease in real terms.
Figure 4.4. Actual versus Expected Tariffs US¢/kWh
20.0
18.0
16.0
14.0
12.0
Usc/kWh
10.0
8.0
6.0
4.0
2.0
0.0
2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
PAD assumption 10.8 12.1 13.6 14.5 15.4 16.4 17.5

Actual 2.5 2.5 2.6 2.3 3.1 4.4 5.5 6.1 6.1 5.8 5.3 5.1 6.1 6.2 6.0
PAD assumption Actual
Page 86 of 102
The World Bank
45. Non‐incremental counterfactual. Although no ERR calculations were presented for alternatives,
the PAD did examine a range of levelized economic cost calculations for alternative generation options to
confirm that Medupi constituted the least‐cost option. Table 4.9 shows the levelized cost comparisons in
the PAD, which includes an estimate of the carbon price necessary to justify the alternative in question.
The next best alternative to Medupi (other than Inge‐III) is CCGT‐LNG‐, which was competitive with
Medupi at a carbon price of US$105 per ton CO2.
Table 4.9. Levelized Cost and Carbon Shadow Prices
Source: PAD, annex 9, figure 3 and table 1.
46. The detailed background report re‐estimated the economic returns based on two alternative
counterfactuals. Leaving aside the Inge‐III hydro alternative, which could not reasonably be assumed to
have been implemented to the same timetable, even with the delays encountered at Medupi, the next
best alternative would have been gas‐fired CCGT‐based on LNG. This showed an ERR of 8.8 percent (albeit
with the benefit of hindsight on LNG prices over the last decade). Against an all‐wind counterfactual, the
ERR was estimated at 9,.2 percent. Both provide economic returns above the 6 percent discount rate.
Impact of GHG Emissions
47. Eskom’s CO2 emissions have decreased, not increased, over the last decade. Energy from Medupi
and Kusile is displacing coal‐fired generation from the remainder of Eskom’s coal fleet, projects whose
GHG emissions are much higher than for these two modern, supercritical projects (figure 4.5).
Page 87 of 102
The World Bank
Figure 4.5. Eskom CO2 Emissions
Source: Eskom annual reports.
48. Table 4.10 shows the calculations of GHG emission valuation on economic returns. The ERR
decreases from 12.9 percent to 11.5 percent at the low value of SVC and to 10 percent at the high value
of SVC. Under the more realistic counterfactual that reflects the displacement of coal generation at
existing coal plants, the reduction in ERR is smaller—to 12.2 percent under the low SVC and 11.6 percent
at the high SVC value.
Table 4.10. Impact of GHG Emissions on Economic Returns
49. Financial returns. At appraisal, the return on Eskom’s equity was estimated at 18 percent.
50. At project completion, the following assumptions are used for the assessment of financial returns:
 2021 tariff increase of 15 percent, as per the recent court ruling
 Subsequent tariff increases at the rate of domestic inflation (that is, constant in real terms)
Page 88 of 102
The World Bank
 EAF:70 percent
 Cost of Eskom balance sheet finance: 11.5 percent, principal repaid over 15 years starting in
2023.
51. The reduction in FIRR to 12.3 percent (table 4.11) over the appraisal estimate of 18 percent may
be seen as remarkably small given the much higher‐than‐expected burden of IDC, construction delays,
and CAPEX overrun. The are several reasons for this.
 In the absence of a commitment fee (on undisbursed amounts), there is no direct penalty
for delay. Delays in issuing construction contracts also result in delays to disbursements.
 Because the estimate in appraisal already assumed a 50 percent CAPEX overrun in its
estimate of FIRR, the actual CAPEX overrun in excess of this figure is quite small.
 Interest costs are significantly lower than expected.
Table 4.11. Financial Returns at Project Completion
Page 89 of 102
The World Bank
52. Table 4.12 provides a step‐by‐step explanation of the lower FIRR. By far, the greatest impact is
the failure of NERSA to increase tariffs, as was expected at appraisal.
Table 4.12. Impact of Assumptions on Eskom FIRR
NPV Equity Change
(R, billions) IRR (%) in IRR
(%)
1 At appraisal As described in the PAD 129.0 21.4 —
2 Exchange rate Actual depreciation much higher than 131.0 19.9 –2.2
appraisal
3 Interest rate Actuals (1–3%) lower than estimated at 144.0 21.7 1.8
appraisal (4%)
4 Tariff Actuals significantly lower than forecast (with –3.7 9.5 –12.2
2021 increase at 5%)
5 Latest tariff ruling Actuals +15% increase in 2021 8.3 10.9 1.4
6 Construction –0.8 9.9 –1.0
delay
7 Actual CAPEX Because both delayed and high contingency in 29.9 12.7 2.8
actual PAD CAPEX assumption, small net benefit
disbursements
8 Capital transfer Small decrease in return is the cost of the 26.0 12.3 –0.4
from US$ to R hedge: actual FOREX depreciation lower than
hedge breakeven
‐ ICR estimate 26.0 12.% —
53. Risk assessment. The downside risk is primarily that associated with the tariff adjustments. The
latest tariff increase improved the financial returns at Medupi, but the track record is that significant tariff
increases are often followed by several years of small increases. The upside is in further improvements to
the EAFs. At 80 percent, the ERR increases to 13.6 percent, and the NPV doubles from R 24.9billion to R
45.5 billion.
54. Reality check. This general conclusion of financial sustainability can be confirmed by a simple
calculation, as illustrated in table 4.13. From the total completed financial cost, the team subtracts
principal payments already made to the end of 2021 and assumes that the entire remaining cost is
recovered over 20 years at 11 percent. At 70 percent EAF, the total cost is R 1.1 per kWh, which is below
the 2021 tariff.
Page 90 of 102
The World Bank
Table 4.13. Reality Check
55. This is the upper bound of the tariff requirement because it is known that some part of the
remaining debt is financed at lower interest rates. Table 4.14 shows a sensitivity analysis of the tariff
requirement as a function of the EAF and the interest rate. At the assumed current tariff of R 1.1 per kWh,
the green cells illustrate the combination of interest rate and EAF that have lower requirements (and
imply financial sustainability). The unshaded cells indicate a revenue shortfall. At an 80 percent EAF, the
cost of finance could be as high as 15 percent to stay within the sustainable zone.
Table 4.14. Sensitivity Analysis: Required Tariff as a Function of EAF and the Interest Rate
Risk Assessment
56. The downside risk is primarily that associated with the tariff adjustments. The latest tariff increase
improved the financial returns at Medupi, but the track record is that significant tariff increases are often
followed by several years of small increases. The upside is in further improvements to the EAF. At 80
percent the ERR increases to 13.6 percent, and the NPV doubles from R 24.9billion to R 45.5 billion.
Conclusions on Medupi
57. An excellent project with high economic returns at appraisal is now merely a good project with
satisfactory returns. While economic returns have halved from 24 percent to around 12 percent, this is
still comfortably above the 10 percent hurdle rate assumed at appraisal, and under the new guidance for
Page 91 of 102
The World Bank
discount rates in economies with slow per capita income growth, which suggests a lower 6 percent
discount rate, the project remains good.
58. The main explanations for the decline in returns are the
 25 percent increase in CAPEX (at constant prices);
 Extended construction delays; and
 Failure to meet expected energy generation in the early years of operation (unsatisfactory
EAF).
59. The main risk to the economic returns is the failure to quickly achieve the EAF target. The 70
percent target used for the ICR estimate should be regarded as an absolute minimum: international best
practice suggests that such supercritical projects should have EAFs of at least 85–90 percent. Indeed,
increasing the EAF at Medupi will have especially positive environmental impacts, because this additional
generation displaces generation and emissions from old coal projects whose efficiency is low and whose
emissions per kWh are much lower. This has already been observed in Eskom’s aggregate GHG emissions,
which have declined over the past five years.
4.4 Economic and Financial Analysis of the Majuba Rail Project
60. At appraisal, the economic and financial returns of the Majuba Rail Project were good, with a 42.9
percent (leveraged) FIRR to Eskom. The corresponding economic return was estimated at 19 percent.
Page 92 of 102
The World Bank
Table 4.15. Financial Returns at Appraisal
61. Table 4.16 shows the key assumptions for the economic and financial analysis of the Majuba
Railway, at appraisal and at project completion.
Table 4.16. Comparison of Assumptions, Appraisal versus Actual/ECTC
At Appraisal At Project Completion Change
Construction time Years 4 11 +7 years delay
Commercial operations date 2015 2022 +7 years delay
CAPEX R, millions, 3,531 4,904 1,372 (39%)
nominal
Coal consumption at Majuba mtpy 13 8 –5
2024 freight differential versus old R/ton 120 45 –75
rail
Page 93 of 102
The World Bank
At Appraisal At Project Completion Change

Results
Economic returns
ERR 19% Negative –19%
Financial return to Eskom equity 43% 5.5% –37.5%
62. Quite obviously, with a CAPEX overrun of 39 percent (in Rand terms) and a 13‐year construction
time with 7 years delay, the economic and financial returns will be seriously affected. Moreover, with coal
volumes required at Majuba estimated to be almost 5 mtpy lower than expected at appraisal and the
freight rate differential between the Majuba and existing rail routes narrowed from R 120 per ton to R 45
per ton, it is hardly surprising that economic returns under these conditions are negative. The financial
return to Eskom of 5.5 percent is Moderately Unsatisfactory because it is below Eskom’s WACC. As shown
in table 4.17, total coal shipments to Majuba over the past decade have been around 14 mtpy. As
communicated by Eskom, the 8 mtpy of coal now expected to be used as Majuba appears low for a
relatively new project whose efficiency would normally be expected to be much higher than many of the
older Eskom coal projects. However, according to Eskom’s forecasts electricity demand will remain
relatively stagnant in the next ten years and Eskom’s market share will continue to decrease as the share
of IPP generated renewable energy increases. These factors will reduce coal burn across the board and
so will declining plant performance and operational inflexibility of some plants, Majuba included.
Table 4.17. Coal Shipments to Majuba (million tons/year)
2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020
By rail 5.04 6.88 8.19 7.95 7.94 7.97 7.66 7.68 7.34 6.66 5.22
By truck 6.73 6.73 5.89 5.57 6.50 6.62 6.68 6.62 9.03 10.08 8.62
Total 11.77 13.61 14.08 13.52 14.44 14.59 14.34 14.30 16.37 16.74 13.84
Source: Eskom.
Page 94 of 102
The World Bank
Table 4.18. Economic and Financial Returns, Majuba Rail Project, Eskom Coal Forecast of 8 mtpy
Page 95 of 102
The World Bank
63. The new Majuba Rail link has a capacity of about 21 million tons, more than the demand for
Majuba alone. The Majuba Rail yard can handle the additional capacity and Eskom is now looking at
options to use this additional capacity to transport coal to other power stations in the area such as Tutuka.
It is therefore likely that even if the Majuba coal requirement is only 8 mtpy, significant additional benefits
would accrue to Eskom by the general rationalization of Eskom’s coal supply in the region. If it is assumed
that an additional 7 mtpy can benefit from the line with a saving of 75 percent of the cost differential
between the new and old routes, then the FIRR to Eskom changes from negative to 9 percent.
64. In summary, as shown in table 4.19, there remains high uncertainty over the economic and
financial returns of the project. The negative economic and financial returns based on the 8 mtpy estimate
may well be overly pessimistic.
Table 4.19. Scenario Assessment, Majuba Rail Project
Majuba Counterfactual Eskom ERR, ERR + ERR +
Coal FIRR (%) Constant GHG @ GHG @
Demand 2010 Prices Low SVC High SVC
(mtpy) (%) (%)
High Majuba coal demand 14 Rail and road 14.1 8.9 9.3 9.6
Eskom coal demand 8 Rail 5.5 Negative Negative Negative
+transportation credit 8 Rail 9.0 4.0 4.2 4.3
Eskom coal demand 8 50% rail, 50% road 13.1 7.0 7.4 7.7
4.5 Inspection Panel Findings
65. An Inspection Panel investigation report issued in 2011 in response to objections from affected
stakeholders noted several issues in the PAD. These issues, and the relevant ICR evaluations, can be
summarized as follows:83
(a) The PAD may have underestimated the costs associated with the damage from CO2
emissions (458).84 This was true but the much higher values of the SVC to be used for World
Bank projects were not issued until 2015, and the US29per ton CO2 value used was high for
its time. This issue has been addressed by the World Bank’s new guidelines for the SVC. The
NPVs and ERRs are now routinely presented for the stipulated range of values: this is no
longer left to the task team leaders and project economist to determine (as it was in 2009
at the time of appraisal).
(b) The comparisons with other technologies were based on current costs, whereas the costs
of CSP, wind, and PV are continuing to decline significantly relative to coal (459). Of course,
that was true, but the additional power was needed at the time, and hence the relevant
costs of CSP, wind, and PV alternatives would have been those that would have been faced
at that time.
(c) The economic analysis valued neither the local environmental damages nor the benefits
from the avoidance of fossil fuels used in the absence of electricity from the project, (496,
497). In fact, including monetization of local costs and benefits—even based on the
83 See economic and financial analysis background report for detailed rejoinders.
84 These are references to the numbered paragraphs of the Panel’s investigation report.
Page 96 of 102
The World Bank
unreliable valuations available at that time—would have made little difference to the ERR
and NPV. Their exclusion at appraisal was correctly judged to be immaterial compared to
other uncertainties likely to have much greater impact (as was indeed the case with CAPEX
and construction delays). However, the Inspection Panel’s criticism that the attempt at
valuation should have been attempted remains valid, and one can say with confidence that
no major fossil‐fueled project would today be presented to the Board in the absence of a
valuation of local externalities based on research and new World Bank guidelines on health
impact valuation.
(d) The opportunity cost of water required for the expansion of the Grootegeluk mine was not
included in the analysis (499). The ICR reanalysis shows that when this is expressly included
as a line item in the table of economic flows, the impact on returns is negligible. However, it
would have been better for this to have been expressly included (and which is now in any
event required under the new 2016 Guidance on Economic Analysis of Power Sector
Projects)
(e) The risk analysis contained an inadequate assessment of domestic and trans‐boundary
externalities (504). The main trans‐boundary issue is indeed GHG emissions, but this issue
has been addressed by the World Bank’s new Guidelines for the SVC. The NPVs and ERRs are
now routinely presented for the stipulated range of values. This is no longer left to the task
team leaders and project economist to determine (as it was in 2009 at the time of appraisal).
4.6 Conclusions
66. The ICR reanalysis shows that
 When all the World Bank guidance and regulations, now in force, are applied, the project
still meets the hurdle rate for economic returns and financial returns that exceed Eskom’s
WACC and
 The monetized impacts of local environmental impacts, or the opportunity cost of water,
based on reliable South Africa‐specific studies now available, would have made no material
difference to the economic benefits at appraisal.
Page 97 of 102
The World Bank
ANNEX 5. BORROWER, CO‐FINANCIER AND OTHER PARTNER/STAKEHOLDER COMMENTS
The World Bank is grateful to Eskom for supporting preparation of the ICR by providing data and

comments on the draft report. Eskom’s comments were discussed with the World Bank and changes
were incorporated in the final report except in a few areas which are summarized below. These
comments are cross referenced to the relevant paragraphs in which these issues are discussed in the ICR.
The Bank is also grateful to the AfDB for their comments on the draft ICR which have been incorporated,
including the main ones which are also shown below.
Comments by Eskom
ICR page, Related text Eskom’s Comments

para. ref.
42, para. Furthermore, though the World Bank The coal quality issue was known and included
117 (e) requested information from Eskom and at time of contract. This was communicated
the contractor that the quality of South during supervision missions to the lenders. This
African coal had been considered in the is covered in section 2.3.2. ‘High Reheater
detailed design of the boilers, no Temperatures’ in the AfDB PCR Supplemental
documentation was ever provided to the Report Rev.0.7 and in more detail when
World Bank. Had this been done, the too discussing latent defects. It would be ideal if
small furnaces could have been there was consistency across reports.
identified to at least institute some
design remedies. Now, all 12 units at
Medupi and Kusile have the same small
furnace—a situation that cannot be
rectified.
46, para. On March 27, 2020, Eskom and other This statement still sends an incorrect message

130 parties operating similar equipment that Eskom and others like Sasol were granted
(e.g., Sasol) were granted a higher limit new limits. The new limit is not specifically for
of 1,000 mg per Nm3 starting in 2025 by Eskom and the likes of Sasol, but a change in
the DFFE. However, Eskom has since the MES themselves and therefore applicable
applied for still a higher limit of 4,000 to all in South Africa.
mg per Nm3 by 2030.
48, Eskom complied with the regulations for Eskom had not requested ‘exemption’, in the

para.131 SO2, but only because it received an case of Medupi, an application was made only
(b) exemption and will continue to seek in terms of SO2 postponement from April 2020
exemptions until it has installed the until March 2030 at 4000mg/Nm3 and
FGD. thereafter to meet the new plant MES of
1000mg/Nm3.
Page 98 of 102
The World Bank
ICR page, Related text Eskom’s Comments

para. ref.
55, Complex projects with numerous This paragraph implies that Eskom did not have
para.149 contract packages present difficult a FIDIC engineer on the project, which is
(d) coordination challenges even for the incorrect because there is a FIDIC engineer.
most sophisticated borrowers and
more so when a new technology is
involved and there is urgency to
implement the project. Using an
owner's engineer according to FIDIC
conditions of contracts is advisable in
such circumstances and normally
provides better results.
57‐58, para. The current standard which Eskom This statement is not correct as the change
149 (m) should comply with is 500mg per from 1,000mg was not an application by
Nm3 starting in 2025. However, Eskom, but a regulatory change that was made
Eskom applied for and was granted a by the Department of Forestry, Fisheries, and
higher limit of 1,000mg per Nm3 in the Environment (DFFE). Eskom’s application
April 2020 but it has since applied for
was for a limit of 4,000mg. It must be put in
a still higher limit of 4,000mg per context that the emission standards in South
Nm3 by 2030 Africa are relatively new and therefore have
evolved, with very little change. It has always
been 3500 for existing plant (like Medupi) and
500 for new plant and only the 500 changed to
1000.
94 – 96, Concerns on projected Majuba coal Background
Annex 4, burn based on FY2022 Corporate Plan
paras. 62‐ i. The World Bank has raised concerns on the
64 projected coal burn for Majuba over the
next 10‐year period detailed in the
document named “EISP ICR P116410 (DM
Final Version_11_17_2021). World Bank
has funded the Majuba rail project with a
view of burning about 14 MT of coal per
annum. However, the latest 10 Year
Production Plan dated 21 October 2020V5
which was used for FY2022 Corporate Plan
projects the average of 8 MT coal burn per
annum.
Page 99 of 102
The World Bank
Majuba projected coal volumes
ii. Production plan is the integrated plan which
takes into account demand, reserve
requirements, plant parameters for existing
and new build power plants, primary energy
cost, constraints as well as other external
supply such as IPPs and Imports. The main
objective of the production plan
optimization is to meet demand at least cost
taking into account all known constraints.
iii. Over the past 11 years, electricity energy

demand has decreased from 239 TWh
achieved in FY2009 to 219 TWh achieved in
FY2021. At the same time, the energy
contribution from Independent Power
Producers has increase from 0 TWh to 13.5
TWh. The decline in the Eskom sales is
expected to further decrease going forward.
Furthermore, Eskom plant performance has
deteriorated from about 85% in FY2009 to
about 64% in FY2021. Majuba has also
contributed to this decline in performance.
Majuba’s EAF was 87% (12.5 mtpy) and only
achieved 72% (12.8 mtpy) last financial year
(FY2021).
iv. With the stagnant electricity demand, high
penetration of renewables and inflexible
power plant, it is expected that most of the
power plants will drop production in the
near future. Demand volume, demand
profile as well as system availability dictates
how units are dispatched to meet hourly
demand. Ideally one would like to use the
cheaper stations to the maximum and only
utilize the more expensive ones to meet the
peaks if necessary. This would work if the
units were flexible enough to be taken off
when not required and only synchronized
when required for energy. Our reality is that
we have predominantly baseload/load‐
following units which means even if you only
need a unit, for maybe 3 hours in a day, you
Page 100 of 102

The World Bank
are forced to keep that unit running for the
full 24 hours even during the periods when
you don’t need it. The upside is that you
save on the start‐up costs and there is
reduced wear on units by not starting and
stopping often.
v. Majuba is not among the cheapest stations
in the fleet and it’s EAF is projected to drop
from 77% (FY2022) to 68% (FY2031) which
will have further impact on coal burn. This is
further compounded by inflexibility of the
power plant and also the drop in the Eskom
market share from 88% in FY2022 to 66% in
FY2031. This is subject to all assumptions
being met.
Page 101 of 102

The World Bank
Comments by the AfDB
(a) Our understanding is that there is sufficient water for Medupi at full operation plus
operations of first 3 FGD units. The water supply shortage risk will only affect the last three
FGD units if MCWAP2 is not completed on time.
(b) AfDB queried the use of the word exception rather than exemption regarding Eskom’s
application for exemptions from regulatory SO2 limits.
Page 102 of 102

World Bank South-Africa-Eskom-Investment-Support-Project (2022) EISP (106p)

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

World Bank South-Africa-Eskom-Investment-Support-Project (2022) EISP (106p)

Uploaded by

Copyright:

Available Formats

Public Disclosure Authorized

Original Amount (US$) Revised Amount (US$) Actual Disbursed (US$)

Borrower/Recipient 4,736,770,000 9,208,000,000 9,208,000,000

Approval Effectiveness MTR Review Original Closing Actual Closing

Date(s) Amount Disbursed (US$M) Key Revisions

Outcome Bank Performance M&E Quality

02 28‐Mar‐2011 Satisfactory Satisfactory 361.61

Regional Vice President: Obiageli Katryn Ezekwesili Hafez M. H. Ghanem

Country Director: Ruth Kagia Marie Francoise Marie‐Nelly

Director: Inger Andersen Ashish Khanna

Practice Manager: Subramaniam Vishwanathan Iyer Julia M. Fraser

Task Team Leader(s): Reynold Duncan, Pankaj Gupta Franz Gerner, Frederic Verdol

ACTIVITIES OUTPUTS OUTCOME INDICATORS PDO

CONSTRUCT  4, 764 MW INSTALLED MW

Other financing was to be provided by concessional loans under the World Bank’s Eskom Renewables

capability to be appropriate for the implementation and operational needs of the MCPP.

Generation capacity installed Megawatt 0.00 4,800.00 4,800.00 4,000.00

Generation capacity installed Megawatt 0.00 100.00 100.00 100.00

Direct CO2 emissions avoided Metric ton 0.00 238,000.00 238,000.00 315,330.00

Direct project beneficiaries Number 0.00 3,600,000.00 5,500,000.00 4,234,100.00

31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021

Female beneficiaries Percentage 50.00 50.00 50.00 50.00

Quarterly Medupi Yes/No No Yes Yes Yes

Medupi construction Percentage 0.00 100.00 100.00 99.95

Number of units Number 0.00 6.00 6.00 6.00

Transmission lines completed Kilometers 0.00 1,020.00 1,020.00 941.90

31‐Mar‐2009 31‐Dec‐2019 30‐Jun‐2021 30‐Jun‐2021

Indicator Name Unit of Measure Baseline Original Target Formally Revised Actual Achieved at

Number of units handed over Number 0.00 6.00 6.00 5.00

Number of wind turbines Number 0.00 46.00 46.00 46.00

erected 31‐Mar‐2009 31‐Oct‐2015 30‐Jun‐2021 30‐Jun‐2021

Progress rate towards Percentage 0.00 100.00 100.00 100.00

Progress rate towards Percentage 0.00 100.00 100.00 97.50

Number of studies for power Number 0.00 7.00 7.00 7.00

Project generation capacity Megawatt 0.00 3,171.00 3,171.00 4,557.00

Indicator Name Unit of Measure Baseline Original Target Formally Revised Actual Achieved at

Energy Storage Capacity Megawatt 1.00 200.00 200.00 0.00

Total 137.25 1,320,468.88

Totals 13,861 18,128 130.8

Total 3,750 3,159 84.2

ICR. New research on the valuation of local environmental costs of air pollution was published by the

13. Local environmental impacts. Power generation from fossil fuels gives rise to a range of local

28. Higher‐than‐expected energy and efficient construction are the two main explanations for the

35. The finances of the Sere wind farm were ring‐fenced, are clearly satisfactory, sustainable, and

PAD assumption 10.8 12.1 13.6 14.5 15.4 16.4 17.5

PAD assumption Actual

At Appraisal At Project Completion Change

The World Bank is grateful to Eskom for supporting preparation of the ICR by providing data and

ICR page, Related text Eskom’s Comments

46, para. On March 27, 2020, Eskom and other This statement still sends an incorrect message

48, Eskom complied with the regulations for Eskom had not requested ‘exemption’, in the

ICR page, Related text Eskom’s Comments

iii. Over the past 11 years, electricity energy

Page 100 of 102

Page 101 of 102

Page 102 of 102

You might also like