Economy-wide Implications of Policy and Uncertainty in the Power Sector of South Africa: A Linked Modelling Approach June 2014

Tara Caetano, Britta Rennkamp and Bruno Merven Energy Research Centre, University of Cape Town in Collaboration with UNU-WIDER

Overview

Background on South Africa and policy/uncertainty landscape

Description of modelling framework

Nuclear case study

Future work

Background Electricity in South Africa

90% generation from coal

large emitter of greenhouse gases, particularly CO2 (± 80% of total)

Improving access instead of increasing capacity - constrained supply

Low real price - rising by about 300% over last 5 years

Consideration of energy policy: Integrated Resource Plan/Integrated Energy Plan

environmental sustainability

depleting low cost coal reserves

cost competitive alternatives

Important element of growth strategy → growth, employment and welfare

Price impact

Investment

Other: e.g. ability to localise (how does this fit in with other policies)

Policy Options

Policy Options and Uncertainty Uncertainty

Commitment to a Nuclear

Program

CO2 Price/tax level

Commitment to support a Gas

Infrastructure program

Commitment to support

Renewable Program

Open economy to electricity

imports from the region

(generated from hydro/gas)

Cost of Nuclear (R/kW) and risk of delays and overruns

Economic growth (and demand for electricity)

CO2 Price/tax level

Global energy commodity prices

Availability and cost of shale and other gas resource (still under exploration)

Future cost reductions on RE

Whether regional projects materialise

Motivation for Linked Energy-Economy-wide Models

Need tool that can measure the macro- and socio-economic impacts of Energy

Policy

Available tools:

Detailed bottom-up energy sector models

Economic models

But existing models approaches are inadequate

Economic Model (CGE type): over-simplification of the energy system

Optimization Energy System Models: no/little economy and energy system feed-back

We choose the linked iterative approach over full integration:

Full inter-temporal integration constrains the level of detail

Stakeholders like to see detail they can relate to

Electricity Sector Model: SATIM-el

Inter-temporal bottom-up partial equilibrium optimisation model of South Africa’s energy sector (Energy Research Centre) SATIM-el: South African TIMES Model - Electricity Sector

Optimisation problem Minimize the sum of all discounted costs over the planning horizon subject to constraints

and system parameters Costs include capital costs, operating costs and taxes (e.g. CO2 tax) Constraints: electricity demand, resource limits, reserve margin, policy targets System Parameters: load curves, existing stock of power plants, new power plant options, fuel

price and availability Other: discount rate, taxes, etc.

SATIM-el:

SATIM Calibrated and parameterised in line with recent Integrated Resource Planning Report (update 2013)

20 time-slices, annual periods to 2040

Economy-wide Model: e-SAGE General equilibrium model of South African economy (SAGE, UNU-WIDER)

Recursive dynamic country-level economy-wide model

eSAGE: detailed electricity sector

Comprehensive representation

62 industries

49 products

9 factors of production

14 representative households

Energy treated as an intermediate input (Leontief)

Simplified energy-saving investment behaviour, which allow sectors of production to reduce

energy intensity in response to increasing energy prices constrained by the rate of investment

in the sector

Upward sloping labor supply curves for less-educated workers

“Putty clay” capital and endogenous capital accumulation

Fixed current account with flexible real exchange rate

Savings-driven investment

e-SAGE-SATIM-el Iteration Process e-SAGE

SATIM-el

• Electricity demand • Electricity production mix by technology/fuel

• Electricity price

• Power plant construction expenditure schedule

SAGE

2010 2020 2030 2050

SATIM

2007

SAGE

SATIM

SAGE

Ite

rati

ve

co

up

led

ru

ns

Committed Forecast

SATIM

TC

TT (IRP)

2010

2020

2030

Emulating the Planning (IRP) process

Nuclear Case Study Initial work done for the IAEA

South Africa has a clear commitment to nuclear power

Risk of cost and delay

Overnight costs range between US$ 5800 and US$7000 per kW

Hickley Point currently estimated around US$8000 per kW

Lead time between 7 and 12 years (although there are outliers)

Availability of renewable energy, gas and regional imports

REIPPPP coming in under budget and ahead of schedule

Shale gas potential in SA and gas fields in the region

Hydropower developments

What are some of the socio-economic implications of nuclear power?

Scenarios Base remains heavily-reliant on coal

3 Nuclear scenarios

Optimistic case: overnight cost of US$5800

Higher cost: overnight cost US$7000

Nuclear delays: simulated delay of 5 years (lead time 12 years)

Renewable target of 50% renewables by 2040

0

100

200

300

400

500

600

2010 2030 2040 2010 2030 2040 2010 2030 2040 2010 2030 2040 2010 2030 2040

Base OptimisticNuclear

Nuclear HigherCost

Nuclear Delays RenewableTarget

Ele

ctri

city

Su

pp

ly (

TWh

)

Electricity Supply Breakdown for Scenarios

Imported

Diesel

Gas

Waste

Wind

Solar

Hydro

Nuclear

Coal

Electricity supply around 500 to 530 TWh in 2040

Some demand response from CGE

Impose a reserve margin of 15%

Dispatch model needed to account for the transmission cost for nuclear versus renewables

Investment and Prices

The total investment cost of the base case is just over R1 trillion for the period until 2040

Nuclear scenarios:

- Optimistic costs R2 trillion

- Higher cost R2,25 trillion

- Delays actually the least because of 180 TWh of nuclear supply opposed to 245 TWh

The renewable target scenario totals at R1,4 trillion, substantially less than the nuclear scenarios attributed to the high reliance on gas generation options.

Electricity price

Lowest under the base case at 72 cents/kWh; Highest under nuclear delays at 98 cents/kWh in 2040

The under-supply of electricity is driving up the price

0

20

40

60

80

100

120

2007 2012 2017 2022 2027 2032 2037

Ele

ctri

city

pri

ce

(ce

nts

/KW

h)

Average Electricity Price Projection

Base Case

OptimisticNuclearNuclearHigher CostNuclearDelaysRenewableTarget

0

50

100

150

2007 2012 2017 2022 2027 2032 2037

An

nu

al c

ost

s (R

and

bil.

)

Annual Electricity Investment Cost (after interest on debt payments)

Base Case

OptimisticNuclear

NuclearHigher Cost

NuclearDelays

RenewableTarget

Emissions Base case emissions from the electricity sector more than double from

429 Mt of CO2 in 2010 to 856 Mt of CO2 in 2040.

Nuclear scenarios reduce emissions by around 300 Mt in 2040.

Slightly less for the renewable energy target scenario (625 Mt in 2040)

Larger share of coal-fired generation in the 2040 capacity mix

Room for more

0

100

200

300

400

500

600

700

800

900

2007 2012 2017 2022 2027 2032 2037

Tota

l CO

2 e

mis

sio

ns

(Mt)

Total Co2 Emissions to 2040

Base Case

Optimistic Nuclear

Nuclear Higher Cost

Nuclear Delays

Renewable Target

Jobs and Welfare

Trade-off between high investment cost and economic growth (savings-driven investment)

Even burden on households

Expected more of a price effect

Electricity employment increased by similar amounts for nuclear and renewables (18000 and 17000)

Nuclear delays = decreased investment demand for electricity and increased employment

Conclusions

The higher cost scenario increased total investment demand by about

US$25 bn

Nuclear delays caused an escalated electricity price

Burden experienced by both households and firms

Employment increased by the same margin for the electricity sector in the

renewables case as well as the nuclear case

The indirect job loss was substantially lower for renewables

Around 100 000 more jobs were created

All scenarios take South Africa closer to its Copenhagen pledge

There is more room for reductions in the renewable energy scenario

Future Work Unbundling the household price effect

Further work on labour markets

The issue of financing has to be addressed

How will this be financed? Pressure on the fiscus?

Implications of electricity supply shortages

Quantifying the risk

Expansion of the transmission network for nuclear versus renewables

Decommissioning of nuclear power

Costs and process

Nuclear waste

Sites, process and cost

Thank you

tara.caetano@gmail.com http://www.erc.uct.ac.za

Sectoral growth Given the savings-driven investment closure we know

that an increase in the investment allocated to the

electricity sector will have a slightly contractionary effect

on the rest of the economy.

Overall annual GDP growth remains at around 3% for all

scenarios, with the renewable target scenario having the

least contractionary effect on the economy (3,1% annual

GDP growth compared to the 3,14% in the base case).

The nuclear higher cost scenario has the largest effect

on GDP

The effect of nuclear investment on sectoral growth tells

an interesting story by changing the structure of the

economy.

The impact on the mining sector is the most

pronounced, The move away from coal-fired generation

is shown by the mining sector shinking slightly, in

realation to the base.

Metals, water distribution and construction are also

bear a higher burden due to the investment in nuclear

power.

This picture could change if there were a localisation

plan modelled along with the investment in nuclear

power. However, until the details of the localisation plan

are know, we are unable to simulate it.

Analysis 1: Impact of CO2 Prices

Two sets of scenarios tested at three CO2 concentration levels:

650, 550 and 450 ppm

1. Optimistic 2. Pessimistic

Nuclear Overnight Cost ($/kW) Lead time (years)

5800 7

7000 12

RE cost reductions Optimistic Pessimistic

Domestic Natural Gas yes no

New Hydro Imports from the region

yes no

Global Prices from Paltsev (2012)

650: CO2 Price -> ~$10/ton

550: CO2 Price -> ~$20/ton

450: CO2 Price -> increasing: ~$70/ton in 2030 and >$100/ton in 2050

Data set from: Sergey Palstev data set on global commodity prices for a no policy and 3 global stabilisation targets (Paltsev, S. (2012) 'Implications of Alternative Mitigation Policies on World Prices for Fossil Fuels and Agricultural Products', UNU-WIDER Working Paper No. 2012/65, www.wider.unu.edu)

0

20

40

60

80

100

120

140

160

180

20

10

20

15

20

20

20

25

20

30

20

35

20

40

20

45

20

50

Oil

Pri

ce $

/bb

l

Oil Price

No Policy

650

550

450

0

2

4

6

8

10

12

14

16

18

20

20

10

20

15

20

20

20

25

20

30

20

35

20

40

20

45

20

50

Gas

Pri

ce $

/tcf

Gas Price

No Policy

650

550

450

0

20

40

60

80

100

120

140

20

10

20

15

20

20

20

25

20

30

20

35

20

40

20

45

20

50

Co

al P

rice

$/t

on

Coal Price

No Policy

650

550

450

0

20

40

60

80

100

120

20

10

20

15

20

20

20

25

20

30

20

35

20

40

20

45

20

50

CO

2 P

rice

$/t

on

CO2 Price

650

550

450

Results: Electricity Production in Optimistic and Pessimistic

0

100

200

300

400

500

600

2010 2030 -650

2030 -550

2030 -450

2040 -650

2040 -550

2040 -450

Ele

ctri

city

Pro

du

ctio

n (

TWh

)

Optimistic

Imported

Gas

Wind

Solar

Nuclear

Coal

0

100

200

300

400

500

600

2010 2030 -650

2030 -550

2030 -450

2040 -650

2040 -550

2040 -450

Ele

ctri

city

Pro

du

ctio

n (

TWh

)

Pessimistic

Imported

Gas

Wind

Solar

Nuclear

Coal

0

10

20

30

40

50

60

70

80

90

100

20

07

20

12

20

17

20

22

20

27

20

32

20

37

Ele

ctri

city

pri

ce (

cen

ts/K

Wh

)

Electricity Price

No Policy

650

550

450

Optimistic

0

20

40

60

80

100

120

140

2007 2012 2017 2022 2027 2032

An

nu

al c

ost

s (R

and

bil.

)

Investment in Power Sector

No Policy

650

550

450

Results: Socio-Economic Impacts (optimistic)

-1.0%

0.0%

1.0%

2.0%

3.0%

4.0%

5.0%

6.0%

20

14

20

15

20

16

20

17

20

18

20

19

20

20

20

21

20

22

20

23

20

24

20

25

20

26

20

27

20

28

20

29

20

30

GD

P L

oss

Re

lati

ve t

o R

efe

ren

ce

GDP Loss Relative to Reference

650

550

450

0.00 0.10 0.20 0.30 0.40 0.50 0.60 0.70

AGRICULTURE

INDUSTRY

Mining

Manufacturing

Food processing

Textiles and clothing

Wood and paper products and…

Petroleum products

Chemicals

Non-metal minerals

Metals

Machinery

Vehicles and transport equipment

Other manufacturing

Other industry

Electricity

Water distribution

Construction

SERVICES

Trade and hotels

Transport and communication

Financial services

Business services

Government services

Other services

Average Sectoral GDP loss 2010-2030 for 450 case (Optimistic)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

20

14

20

15

20

16

20

17

20

18

20

19

20

20

20

21

20

22

20

23

20

24

20

25

20

26

20

27

20

28

20

29

20

30

Job

Lo

ss R

ela

tive

to

Bas

e (

mill

ion

)

Job Losses Relative to Reference (million)

650

550

450

0.225 0.230 0.235 0.240 0.245 0.250

Poor (0-50)

Non-poor (50-100)

Middle (50-90)

Top (90-100)

Drop in per capira consumption growth (%)

Drop in per capita consumption growth (2010-2030)

Analysis 2: Nuclear Program: 10GW by 2030? Green Barley Cases

4 Cases:

Case Nuclear Cost/Lead Time

RE Costs Domestic Gas Regional Hydro

1. Worst case for Nuclear – no early program (free)

High (pessimistic)

Low Yes Yes

2. Best case for Nuclear – no early program (free)

Low (optimistic)

High No No

3. Worst case for Nuclear – imposed early program (forced)

High (pessimistic)

Low Yes Yes

4. Best case for Nuclear – imposed early program (forced)

Low (optimistic)

High No No

(pessimistic)

(optimistic)

(pessimistic)

(optimistic)

GDP Loss Relative to Unforced Nuclear (Free)

-0.2%

0.0%

0.2%

0.4%

0.6%

0.8%

1.0%

1.2%

20

15

20

16

20

17

20

18

20

19

20

20

20

21

20

22

20

23

20

24

20

25

20

26

20

27

20

28

20

29

20

30

GD

P L

oss

re

lati

ve t

o "

Fre

e"

GDP Loss relative to "Free"

worst

best

550 - Scenario

-0.2%

0.0%

0.2%

0.4%

0.6%

0.8%

1.0%

1.2%

20

15

20

16

20

17

20

18

20

19

20

20

20

21

20

22

20

23

20

24

20

25

20

26

20

27

20

28

20

29

20

30

GD

P L

oss

re

lati

ve t

o "

Fre

e"

Worst

Best

450 - Scenario

-10

10

30

50

70

90

110

130

150

2015 2020 2025 2030

An

nu

al E

xpe

nd

itu

re (

Ran

d b

il.)

Worst - Free

Best - Free

Worst - Forced

Best - Forced

Annual Expenditure on Power Plants

-10

10

30

50

70

90

110

130

150

2015 2020 2025 2030

An

nu

al E

xpen

dit

ure

(R

and

bil.

)

Worst - Free

Best - Free

Worst - Forced

Best - Forced

Annual Expenditure on Power Plants

Outstanding Issues and other Current and Future Work

Improve integration:

i/o coefficients in eSAGE better aligned to SATIM

SATIM to take account of changes in Capital and Labour costs

Linking the full sector model: to improve energy consumption behaviour

of sectors other than electricity

More comprehensive analysis of uncertainty via Expert-Elicitation, Monte

Carlo and Stochastic Programming

Other considerations: Water constraints, spatial aspects of demand and

resource, non-dispatchability of RE techs

More detail on Renewables Cost Reductions

Source: IRP update 2013, department of energy, government of South Africa

Exports Imports

Total supply

Labor Capital Inputs (incl.

Energy)

Sector

output

Factory

Supplier 1 Supplier n

Output 1 Output n

Domestic

supply

Warehouse

Households

Government

Investment

Intermediates

Supermarket

Traders

Freight transport

Consumption linkages

Production linkages

Overview of e-SAGE

CES LEO

LEO

CES

CET

CES

LES LEO LEO

Energy as an intermediate input

Labor Capital

Inputs

Sector output

LEO

Inputs

Input 1

Energy

Input n

LEO

CES

Energy-saving investment behavior

Change in energy inputs per unit of output based on energy

prices

Energy product input coefficient (ioij) falls when…

Energy prices (pi) rise (provided there is some new investment)

New investment share (sj) is positive (provided the price rises)

Governed by a response elasticity (ρ)

𝑖𝑜𝑖𝑗,𝑡+1

𝑖𝑜𝑖𝑗𝑡= 1 − 1 −

𝑃𝑗𝑡

𝑃𝑗,𝑡−1

−𝜌

∙ 𝑠𝑖

Macro closure rules

Upward sloping labor supply curves for less-educated workers

“Putty clay” capital and endogenous capital accumulation

Fixed current account with flexible real exchange rate

Savings-driven investment

Distinguish between electricity and non-electricity sector investment

Electricity investment differentiated by subsector (esp. import content and job

creation)

Government borrows abroad to pay for investment (gradual interest and principal

repayment)