TECHNICAL
REPORT
HARNESSING THE
POTENTIAL OF
FLEXIBLE DEMAND
RESPONSE IN
EMERGING MARKETS
Lessons Learned and
International Best Practices
UNLOCKING THE ENERGY TRANSITION 1
TECHNICAL
REPORT
HARNESSING THE
POTENTIAL OF
FLEXIBLE DEMAND
RESPONSE IN
EMERGING MARKETS
Lessons Learned and
International Best Practices
ABOUT ESMAP
The Energy Sector Management Assistance Program (ESMAP) is a partnership between the World Bank
and over 20 partners to help low- and middle-income countries reduce poverty and boost growth through
sustainable energy solutions. ESMAP’s analytical and advisory services are fully integrated within the World
Bank’s country financing and policy dialogue in the energy sector. Through the World Bank, ESMAP works
to accelerate the energy transition required to achieve Sustainable Development Goal 7 (SDG7), which
ensures access to affordable, reliable, sustainable, and modern energy for all. It helps shape World Bank
strategies and programs to achieve the WB Climate Change Action Plan targets.
Learn more at: https://www.esmap.org.
© November 2024 | International Bank for Reconstruction and Development / The World Bank
1818 H Street NW, Washington, DC 20433
Telephone: 202-473-1000; Internet: www.worldbank.org
This work is a product of the World Bank Group, with contributions given by the staff, consultants, and experts
listed in the acknowledgments. The findings, interpretations, and conclusions expressed in this work do
not necessarily reflect the views of the World Bank, its Board of Executive Directors, or the governments
they represent. The World Bank does not guarantee the accuracy of the data included in this work. The
boundaries, colors, denominations, and other information shown on any map in this work do not imply any
judgment on the part of the World Bank concerning the legal status of any territory or the endorsement or
acceptance of such boundaries. Nothing herein shall constitute or be considered to be a limitation upon or
waiver of the privileges and immunities of The World Bank, all of which are specifically reserved.
Rights and Permissions
This work is available under the Creative Commons Attribution 3.0 IGO license (CC BY 3.0 IGO)
http://creativecommons.org/licenses/by/3.0/igo. Under the Creative Commons Attribution license, you
are free to copy, distribute, transmit, and adapt this work, including for commercial purposes, under the
following conditions:
Attribution—Please cite the work as follows: Energy Sector Management Assistance Program (ESMAP).
2024. Harnessing the Potential of Flexible Demand Response in Emerging Markets: Lessons Learned and
International Best Practices. ESMAP Technical Report. Washington, DC: The World Bank. License: Creative
Commons Attribution CC BY 3.0 IGO
Third-Party Content—The World Bank does not necessarily own each component of the content contained
within the work. The World Bank therefore does not warrant that the use of any third party-owned individual
component or part contained in the work will not infringe on the rights of those third parties. The risk of
claims resulting from such infringement rests solely with you. If you wish to re-use a component of the
work, it is your responsibility to determine whether permission is needed for that re-use and to obtain
permission from the copyright owner. Examples of components can include, but are not limited to, Tables,
Figures, or images.
All queries on rights and licenses should be addressed to World Bank Publications, The World Bank Group,
1818 H Street NW, Washington, DC 20433, USA; e-mail: pubrights@worldbank.org.
Production Credits
Production Editor | Heather Austin
Designer | Circle Graphics, Inc.
Cover Image | © World Bank
All images remain the sole property of their source and may not be used for any purpose without written
permission from the source.
Contents
Abbreviations vii
Acknowledgments viii
Executive Summary ix
1. Demand Response And Its Role in the Energy Transition xii
The Need for Demand Response 1
Defining Demand Response 2
Demand-Response Applications 3
Demand Response as a Climate Asset 4
2. Classifying Demand-Response Instruments 8
Price-Based Demand-Response Instruments 11
Quantity-Based Demand-Response Instruments 19
3. Increasing the Uptake of Demand Response: Enablers and Barriers Enablers 26
Barriers and Challenges 27
34
4. Integrating Demand Response into Power Systems
Contracting Framework: Price-Based Demand Response 38
Contracting Framework: Demand-Side Offers for Wholesale Energy Provision 39
Contracting Framework: Demand-Side Offers for Capacity 40
Contracting Framework: Demand-Side Offers for Ancillary Services 42
45
5. Lessons Learned for Demand Response in Developing Countries
Demand Response in India 48
Demand Response in People's Republic of China 49
Demand Response in South Africa 52
Demand Response in Brazil 55
Demand Response in Viet Nam 57
Demand Response on Small Islands 60
Demand Response in the United States: Best Practices from Advanced Frameworks 62
Comparative Analysis of Demand-Response Programs 66
71
6. Roadmap for Implementing Demand Response in Developing Countries
Summary of the Roadmap 74
System Diagnostic 75
Mechanism Assessment 76
Implementation Design 77
82
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS iii
7. Conclusion 88
Bibliography 91
Appendix A: Definitions of Key Ancillary Services 103
Appendix B: Additional Information on the Case Studies 105
Spanish Time-of-Use Residential Tariff 105
Demand Response in India: Additional Information on Pilot Programs 108
Demand Response in South Africa: Additional Information on ESKOM Programs 109
Demand Response in the United States: Additional Information 112
Appendix C: Possibilities for Load Control 114
iv CONTENTS
List of Figures, Tables,
and Boxes
Figures
Figure 1.1: Electricity System Flexibility by Source 1
Figure 1.2: Objectives of Demand Response 4
Figure 2.1: Types of Demand-Response Mechanisms According to the
Nature of Incentives 9
Figure 2.2: Types of Price-Based Instruments 11
Figure 2.3: Share of Countries with Time-of-Use Rates, by Customer Class 13
Figure 2.4: Average Peak Reduction from Time-Differentiated Rate Pilots 19
Figure 2.5: Elia’s Same-Day Adjustment Process 25
Figure 3.1: Price Responsiveness With and Without Emerging Technology 29
Figure 4.1: Contracting Framework for Price-Based Demand-Response Providers 40
Figure 4.2: Contracting Framework for Demand-Side Energy Offers 42
Figure 4.3: Contracting Framework for Demand-Side Capacity Market Offers 43
Figure 4.4: Contracting Framework for Demand-Side Offers to Ancillary Service Markets 46
Figure 5.1: Time-of-Use Tariffs Available in South Africa 55
Figure 5.2: Peak and Off-Peak Comparison of Large, High-Voltage Clients 58
Figure 5.3 Curtailment Events on Jeju Island 64
Figure 5.4: Eurelectric’s Millener Project: Two Setups 65
Figure 5.5: Evolution of Demand-Response Mechanisms in the United States 67
Figure 6.1: Summary of Roadmap for Demand-Response Implementation within a Country 75
Figure 6.2: Guidance Questions for the Diagnostic of the Power System 76
Figure 6.3: Drivers of the Need for Demand Response 77
Figure 6.4: Approach to Mechanism Identification and Assessment 78
Figure 6.5: Matrix of Demand-Response Challenges and Some Mechanisms to
Meet Them 80
Figure 6.6: Implementing a Demand-Response Program in Four Steps 82
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS v
Tables
10
14
21
50
54
56
59
61
61
79
79
83
84
85
85
Boxes
Box 1.1: Data Centers and Demand Response 5
Box 2.1: Time-of-Use Tariffs for Commercial and Industrial Consumers in Costa Rica 12
Box 2.2: South Africa’s Experience with Critical Peak Pricing 15
Box 2.3: SmartHours Variable Peak Pricing, 2023 16
Box 2.4: Hybrid Price-Based Instruments in Spain 17
Box 2.5: India: Load Control for Large and Medium Customers 21
Box 3.1: Control of Hot Water in Botswana 28
Box 3.2: Electric Vehicles and Vehicle-to-Grid 30
Box 4.1: Business Model: OCTOPUS ENERGY Tailored Tariff Provider 41
Box 4.2: Business Model Example: Virtual Power Plants 43
Box 5.1: Pennsylvania-New Jersey-Maryland Interconnection: A Frontrunner in Demand-
Response Deployment 68
vi CONTENTS
Abbreviations
Acronym Definition
CBL customer baseline load
CEMIG Companhia Energética de Minas Gerais
CPP critical peak pricing
CSP curtailment service provider
DSB demand-side bidding
DSO distribution system operator
EV electric vehicle
EVN Viet Nam Electricity
IEA International Energy Agency
PJM Pennsylvania-New Jersey-Maryland Interconnection
PV photovoltaic
RPM Reliability Pricing Model, PJM’s capacity market
RTO regional transmission organizations
RTP real-time pricing
RVD Redução voluntária da demanda (Brazil)
ToU time-of-use
TSO transmission system operator
VRE variable renewable energy
V2G vehicle-to-grid
VPP variable peak pricing
VRE variable renewable energy
All dollar amounts ($ OR USD) are US dollars unless otherwise indicated. The word “cents”
refers to US cents, unless otherwise indicated.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS vii
Acknowledgments
This report is a product of the World Bank’s Energy Sector Management Assistance
Program (ESMAP). Supervising its preparation were Demetrios Papathanasiou, the World
Bank’s Global Director for Energy and Extractives, and Chandra Govindarajalu and Fanny
Missfeldt-Ringius, ESMAP Practice Managers.
The World Bank’s authorial team comprised Gabriela Elizondo-Azuela, Luiz Maurer,
and Kabir Malik.
Internal peer reviews were conducted at various stages of the report’s preparation.
The team is grateful to the following peer reviewers: Ashok Sarkar, Jas Singh, Tatyana
Kramskaya, Christophe de Gouvello, and Mitsunori Motohashi.
The report benefited from discussions with representatives of private firms, power utilities,
and governments. Those representatives include: Praveer Sinha (TATA Power); Shirley
Salvoldi, Hitesh Umley, Lawrence Padachi, Ferdie Bekker, and Dharmesh Bhana (Eskom,
South Africa); Ahmad Faruqui (consultant on demand response, formerly with Brattle
Group); Sandoval Feitosa and Rafael Ribeiro (ANEEL, Brazil); Glaysson Muller (EPE, Brazil);
and Carlos Battle Lopez, Paolo Mastropietro, and Paulo Rodilla (Universidad Pontificia
Comillas (Spain).
The team appreciates the initial reviews of the paper’s structure received from Marina
Azevedo and Pravesh Raghoo.
The team also appreciates the thorough review and other inputs provided by David
Williams, Andrew Tipping, and Thalia Goode from Economic Consulting Associates (UK).
Steven Kennedy edited the report.
All views expressed and any shortcomings are those of the authors. They do not reflect the
views of the World Bank, its Board of Directors, the peer reviewers, or contributing experts.
viii ACKNOWLEDGMENTS
Executive Summary
Demand response is a “short-term, voluntary decrease in electrical consumption
by end-use customers that is generally triggered by compromised grid reliability or high
wholesale market prices.”1 In demand-response schemes, customers are remunerated for
curtailing their loads (Federal Energy Management Program. n.d.). Curtailment can help
balance energy supply and demand, ensure adequate capacity during times of stress on
the system, and provide ancillary services such as frequency support and management of
network congestion.
The role of demand response in power markets is expected to grow significantly in
coming years. In its “Net Zero by 2050 Roadmap for the Global Energy Sector,” the
International Energy Agency (IEA) foresees a need for “flexibility,” the components of which
are demand response, batteries, and low-carbon flexible power plants to help balance wind
and solar with evolving demand patterns (IEA 2021b). In the roadmap, the use of demand
response in emerging markets and developing economies grows from a negligible base in
2020 to almost 20 percent of all flexibility by 2050. These global expectations are mirrored
in decarbonization pathways for individual countries. For example, the United Kingdom’s
scenarios for achieving net zero energy by 2050 foresee demand response (including vehicle-
to-grid technology) expanding six to eleven times its 2021 levels in terms of capacity, while
IEA (2021c) foresees a “massive increase in power system flexibility” being required for India,
including between 140 and 200 gigawatts of batteries, supplemented by demand response.
Three factors are contributing to this growth. First, some countries have long used
demand response to manage demand fluctuations and support grid security. Continued
growth in the deployment of variable renewable energy reinforces those uses by making
periods of system stress less predictable and imposing further technical challenges for
system operators regarding frequency and voltage control. Second, the electrification of
transport and heating and a rapid growth in demand from data centers are changing load
profiles across various markets. Demand response offers a cost-effective way to balance
supply and demand and manage these developments. Third, smart grid and appliance
technology, coupled with internet-based devices that communicate instantaneously,
extends the range of potential demand-response providers—right down to individual
households.
Maximizing the potential of demand response requires a conducive policy and
regulatory framework. Market design and other regulatory barriers can restrict the
participation of demand-response providers that would otherwise be cost-effective,
increasing the overall cost of electricity supply and slowing the energy transition.
Understanding the options available, international lessons learned, and how these can be
best applied is therefore valuable for policy makers and regulators seeking to understand
how best to leverage the potential of demand response in their power systems.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS ix
Demand-response programs can be broadly categorized as indirect price-based
approaches or direct quantity-based approaches. Price-based (also known as “implicit”)
demand-response mechanisms use time-differentiated electricity prices designed to
incentivize customers to shift consumption away from peak hours or when the power
system is under stress from high demand or low supply. Quantity-based (or “explicit”)
demand response includes a broad set of solutions designed to shape customer
consumption directly through incentive payments. Under quantity-based schemes, the
utility, service provider, or grid operator directly contracts for a given quantity of load
reduction, typically in exchange for an explicit payment. Certain loads, such as water
heaters, air conditioners, lighting, pool pumping, and electric vehicle charging can be
shifted with minor inconvenience to end users. With the increasing deployment of smart
meters and remote monitoring and control tools, such load control can also be automated
and thereby more closely controlled by the contracting entity.
Demand-response programs can be implemented at the retail or wholesale level to
support energy and network services. Electricity market retailers2 may offer time-of-use
tariff options for all customer types, provided their metering equipment can support
implementation. Distribution system operators (which may be bundled with the local
retailer) can also enlist demand-response providers to support local system services such
as management of congestion in the distribution network. At the wholesale level, the grid
operator can contribute to load control and encourage the use of demand response to
achieve wholesale energy balancing, capacity adequacy, and ancillary services. It does so
through various market mechanisms and bilateral contracts. Large consumers will often be
able to access these demand-response markets directly, while smaller consumers may be
required—or prefer for reasons of transaction cost and simplicity—to work via demand-
response aggregators.3 Retail and wholesale mechanisms are complementary.
Policy makers and regulators can encourage innovation and new business models by
improving market design and widening access to demand-response mechanisms.
Emerging demand-response business models are being designed and piloted to aggregate
multiple, fragmented loads and allow the utility or grid operator to control them directly.
Demand-response aggregators can build multiple distributed energy resources—including small,
fragmented demand-response opportunities—to a scale that can deliver substantial grid services
to utilities and system operators. These new companies are equivalent to the energy service
companies operating in the energy efficiency space. Such innovation is greatly supported by
conducive market design and the adoption and installation of new smart technologies.
While the absolute level of demand response remains low, demand-response
instruments have already been widely deployed and tested in many emerging
markets, from the large power systems of India, People's Republic of China, and Brazil down
to those of small island nations. Assessing global lessons learned and barriers encountered
can help other emerging markets understand how best to design and implement their own
instruments in line with the local context and with maximum chance of success. Static time-
of-use tariffs have been adopted in most of the countries reviewed in the case studies of this
report. Still, dynamic pricing (prices that vary in real time in response to prevailing supply-
and-demand conditions) has not been mainstreamed, except for some pilots in South Africa
and exposures
x EXECUTIVE SUMMARY
to spot-prices by large customers in Brazil. Quantity-based demand-response schemes exist
in People's Republic of China and are relatively well developed in South Africa, which has
implemented sophisticated load control systems and business models to manage
nonessential loads to mitigate capacity shortages. However, a wide gap in demand-response
deployment remains, and performance in those countries is compared in this report with the
best demand-response practices in the United States, a global forerunner.
The case studies presented here provide evidence that demand-response programs
should be grounded in a country’s context, targeting identified power system
constraints. Demand-response mechanisms can be well targeted to specific objectives but
are likely to be less effective if applied too broadly. As a first step in designing a demand-
response program, a system diagnostic will identify the drivers of the need for demand
response, which often include variable generation, network constraints, customer demand,
and circumstances external to the system, such as a mandate for decarbonization. The
driver’s location is also essential—that is, how it is dispersed geographically or over time,
how immediate it is, and how it affects different customers. Finally, the structure of a
country’s power system and its experience with demand response will determine possible
mechanisms and how they can be adopted.
The benefits of demand response should be identified and evaluated through a cost-
benefit analysis. The analysis should be based on a country’s enabling conditions, market
structure, and available (or soon-to-be-available) technologies. It should plot all demand-
response options (including “do nothing”) on a “supply curve” based on their net and distributed
benefits, allowing them to be ranked in order of preference. After designing the mechanisms,
enhancements to pertinent policy, legal, and regulatory frameworks should be accompanied by
market education on their potential. Finally, the program should be implemented in accordance
with a detailed roadmap and timeline, supported by pilot projects, quick wins, technical
assistance for implementing institutions, and systematic monitoring and reviews.
Endnotes
1. “Wholesale market prices” here are driven by the high marginal cost of generation,
which would apply equally to a jurisdiction with no active market operating under
least-cost dispatch regulations
2. This report uses “retailer” to refer to entities licensed to supply electricity procured
at the wholesale level (through wholesale markets, directly from generators, or from
a designated single buyer) to end consumers. In many jurisdictions the retailer will
remain bundled legally or in accounting terms with the distribution system operator
or distribution network owner, in which case it will be referred to as a “local utility” or
“distribution company” operating under monopoly price regulation.
3. “Demand-response aggregator” is used as a general term to refer to an entity contracting
several demand-response providers. A retailer may also operate as an aggregator—for
example, when bidding on demand response in wholesale markets.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS xi
ONE
DEMAND RESPONSE
AND ITS ROLE IN THE
ENERGY TRANSITION
The Need for Demand Response
Supported by new and innovative technologies, demand response is rapidly gaining
prominence in power system planning and has the potential to help facilitate the energy
transition. The International Energy Agency (IEA) foresees the need for sources of flexibility—
defined as the ability to adjust demand and supply when faced with an imbalance in the
system—to quadruple by 2050 under a global net zero scenario despite the loss of dispatchable
fossil fuel generation (IEA 2021b). Demand response is identified as a critical and significant
component for enabling this transition in both developed and emerging markets (IEA 2021a).
It should represent about 30 percent of the electricity system’s flexibility in advanced economies
and almost 20 percent in emerging markets and developing economies, as shown in Figure 1.1.
These international projections are replicated in national plans. For example, the United
Kingdom’s scenarios for achieving net zero energy by 2050 foresee demand response (including
vehicle-to-grid technology) to expand to six to eleven times their 2021 levels in terms of capacity,
while IEA (2021c) foresees a “massive increase in power system flexibility” being required for
India, including between 140 and 200 gigawatts (GW) of batteries, supplemented by demand
response (National Grid ESO 2024). Thus, the growing need for flexibility should motivate policy
makers and regulators to embrace demand response aggressively.
Flexibility is crucial to increasing the penetration of variable renewable energy in
the power system. Flexible resources, such as synchronous generation, storage, and
demand response, will become critical elements to balance the system and provide
ancillary services.
FIGURE 1.1
Electricity System Flexibility by Source
economies
Advanced
2020
Coal
2050 Natural gas
Oil
Hydrogen-based
developing economies
Emerging market and
Nuclear
Hydro
2020
Other renewables
Batteries
2050
Demand response
20% 40% 60% 80% 100%
Source: IEA 2021b.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 1
There are three key drivers behind this acceleration in the deployment of demand response.
• Very high levels of variable renewable energy (VRE) are critical to meet global net
zero and decarbonization goals. In their 2050 global net zero pathway, IEA estimates
solar photovoltaic and wind capacity must grow by around four times between 2020
and 2030 and reach almost 70 percent of all electricity generation by 2050 (IEA 2021b).
This rapid expansion will entail significant disaggregation of energy supply, with a much
larger share of generation capacity connected at the distribution level both upstream of
the meter and behind it. These developments are creating new generation patterns and
challenges for achieving supply-demand balance on the grid and for system operators
that must manage additional technical challenges in ensuring reliable supply.1 Energy
system flexibility is thus an increasingly essential part of the power system.
• Structural changes to electricity demand patterns. Climate change, economic growth
in emerging economies, and electrification of end uses (especially transport systems)
are changing the intensity and load shape of electricity demand. Cooling represents
10 percent of global electricity demand, and that share is increasing fast, driven by rising
temperatures and the adoption of air-conditioning and refrigeration, especially in
emerging and developing economies (IEA 2023). Electrification of the transport sector, led
by electric vehicles (EVs),2 is also changing load profiles, as is the rapidly expanding need
for energy-intensive data centers. These demand-side factors and associated changes
in load profiles are increasing the strain on the grid and the complexity of maintaining
reliable operations. Thus, there is an increasing need for flexible resources, including
demand response, to manage structural changes in load shape.
• New technologies are enabling the emergence of fast, cost-effective demand-
response schemes. Digital tools, innovative grid technologies, smart appliances, and other
Internet-of-Things devices enable a suite of options to deploy demand response far beyond
its historical base of large consumers—even to individual households—thus increasing the
ability of system operators to manage loads flexibly and cost-effectively in response to grid
conditions. Smart meters, a foundational smart grid technology, are crucial to optimizing
price- and quantity-based demand response because they provide granular data on prices
and consumption to utilities and customers. Increasing digitalization has allowed more
cost-effective coordination and management of distributed resources and their aggregation
into virtual power plants that can deliver flexibility to grid operators.
Flexibility, and particularly demand response, thus provides substantial value to power
systems by displacing expensive thermal generators and deferring investments in new
generation and grid expansion.
Defining Demand Response
Demand response lies within the broader concept of demand-side management. Although
the terms are sometimes used interchangeably, demand-side management typically
refers to measures designed to shape load profiles in the medium to long term.
2 Demand Response And Its Role in the Energy Transition
In contrast, demand-response programs are focused on changing demand in response
to events of shorter duration—and even in real time. Both stand in contrast to energy
efficiency, which signifies permanent change in energy consumption, generally with no
decrease in service level.
For this report, demand response is defined as encompassing “changes in the electric
usage by demand-side resources from their normal consumption patterns in response to
changes in the price of electricity over time, or to incentive payments designed to
induce. . . lower electricity use at times of high wholesale market prices or when the
system reliability is jeopardized” (FERC 2018). This report also considers behind-the-meter
forms of energy storage
(domestic battery systems and vehicle-to-grid systems) owing to their similar purpose
and their potential for being bundled with providers’ demand-response measures.
Demand-Response Applications
Demand-response measures can increase affordability and reliability. They can also aid in
decarbonation.
• Demand-response measures can lower generation costs by shrinking the need for
expensive thermal plants during peak hours or hours of low VRE output and optimizing
the use of the existing generation assets.
• They can make electricity supply more reliable by minimizing the risk of curtailment and
outages and helping to alleviate network congestion.
• They can reduce carbon emissions by enabling consumers to shift load away from peak
periods (or periods of low VRE output) in situations where the marginal plant is typically
carbon-intensive thermal generation.
The objectives are summarized in Figure 1.2.
Many stakeholders can benefit from demand-response programs. Electricity customers
who participate in demand-response programs obtain direct financial savings from more
efficient usage and may receive compensation for services provided to the grid.
Nonparticipating customers will benefit from reduced peak demand, reduced
infrastructure costs, and increased grid reliability. The cost of non-served energy is very
high, and a more reliable grid mainly benefits people with low incomes who cannot afford
backup generators.
Distributional concerns must be considered in the design of demand-response programs.
For example, when utilities shift from flat to time-varying tariffs (a practice termed “price-
based” demand response in this report), customers who are not adequately informed or
able to adjust their consumption profiles may end up paying larger electricity bills, which
is especially problematic if they are low-income customers.3 There are several ways to
address this distributional challenge. First, in the design phase of demand-response
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 3
FIGURE 1.2
Objectives of Demand Response
Energy Least-cost Maximizing consumer utility through least-cost
balance energy matching of supply and cost-reflective demand
provision Affordability
Capacity Capacity Ensuring adequate de-rated capacity available on Meeting system needs
provision adequacy system for managing system stress events within at lower cost
LOLP limits
Ancillary Reserve Arresting and restoring system frequency Reliability
services provision following loss of load
Reducing system
Frequency Managing continuous fluctuations in frequency outages
regulation caused by fluctuations in demand and in the
supply of renewable energy
Decarbonization
Voltage Managing voltage deviations to retain power
control quality Facilitating greater
penetration of
Constraint Managing network congestion and constraints
renewable energy
management that otherwise curtail generation in substitution
of upgrading or reinforcing the network directly
(“non-wire alternative”)
Source: Author’s analysis.
Note: Non-wire alternatives address congestion management without expansion of the grid. LOLP = loss of
load probability; RES = renewable energy.
programs, a cost-benefit analysis should be carried out to assess the overall impact on
rates across all customer categories. Second, tariff design can be fine-tuned so that no one
loses at the outset (although this may necessitate diluting the incentive). Third, safety nets
can be established for low-income customers. Fourth, the demand-response tariff system
can be made optional (on an opt-in or opt-out basis).
Demand Response as a Climate Asset
Demand response’s primary focus is to enhance the power system’s stability, to address
the challenges of meeting peak demands or low VRE output, and to help the system
operator manage contingencies. This goal is also significantly aligned with climate-related
benefits.
Demand-response programs can help reduce greenhouse emissions in the electricity
sector by shifting consumption to off-peak periods or periods of higher renewable energy
output. An analysis conducted for several US states in 2017 revealed that, in California,
there was a 96 percent difference in emissions between the most and least carbon-
intensive hours of the year. The variation in the New York grid was around 25 percent, and
4 Demand Response And Its Role in the Energy Transition
the variation in the Midwest was close to 11 percent (Zhou and Trieu 2021). The fact that
peak demand is typically met with fossil-fuel-based energy4 highlights the substantial
climate benefits to be had by shifting load away from peak periods toward times when
renewables are generating power.
Demand response provides additional flexibility to displace carbon-intensive thermal
power plants that provide ancillary services. Fast-response resources can handle sudden
variations in solar production or in the ramp-up of EV charging (Conteh et al. 2020). This is
important because inverter-based renewables that are not synchronously connected to the
grid network reduce the level of rotating inertia in the system, which can cause the grid’s
frequency to react more rapidly in the event of a supply or demand shock. Fast-responding
reserves, such as load control (and batteries), help manage such fluctuations and minimize
the need to retain fossil fuel generators (and their associated inertia) on the network.
Therefore, demand response enables continued expansion of the VRE base, displacing
expensive thermal peaking plants. Box 1.1 offers an example of close alignment of demand
response with climate goals.
BOX 1.1
DATA CENTERS AND DEMAND RESPONSE
Driven by the explosion of cloud computing, cryptocurrency mining, and artificial
intelligence, power demand from data centers is rising dramatically around the
world. The scale and speed of growth in demand is very likely to complicate
utilities’ efforts to assure grid reliability and decarbonization. In recognition of the
challenge, data center companies are increasingly focusing on energy efficiency
(largely in relation to processing and cooling) and demand flexibility.
For example, Google’s carbon-intelligent computing platform allows Google to
shift less urgent computing tasks to times and places where renewable energy
is most available, thereby helping lower the company’s carbon footprint.
Google classifies tasks by their urgency to assess which can be shifted without
a substantive impact on user experience or performance (for example,
YouTube video processing). The company is applying the intelligent computing
system to some of its most prominent super-scale data centers worldwide
(Bonifacic 2020).
(continues)
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 5
BOX 1.1 (Continued)
Google has also used the carbon-intelligent computing platform to work with
utilities to support grid reliability. During the energy crises in Europe in the winter
of 2022–23, the company used the platform to reduce electricity demand during
the evening peak period. In the United States, the company reduced data center
power consumption in response to extreme weather events to maintain local
grid reliability.
Other technology companies and data centers are also testing demand-response
initiatives by optimizing their operations. In the face of rapidly growing energy
demand, these efforts could yield a very significant demand-side benefit.
Source: Mehra and Hasegawa 2023.
Lastly, demand response has the potential to lower overall electricity use. It is often assumed
that demand response is “energy neutral”—that it simply shifts consumption to another
period, with no energy savings, but empirical evidence has shown otherwise. Demand-
response programs have brought energy savings of 3–6 percent of total energy consumption.
The savings are achieved because not all the energy deferred (or not consumed) during peak
hours is shifted to other periods. Furthermore, the technologies that enable demand
response, such as energy management systems for buildings, can help customers reduce
their overall energy consumption (Nemtzow, Delurey, and King 2007).
Demand-response programs have emerged as a critical, flexible tool in the energy
transition, offering a range of benefits for grid operators and energy consumers. As the
world continues to shift toward a cleaner energy future, demand response will play an
increasingly important role in maintaining grid stability, reducing greenhouse gas
emissions, and supporting a more sustainable energy system.
Endnotes
1. Specific issues include high ramp rate requirements caused by falling solar photovoltaic
output concurrent with the evening peak, a low system inertia increasing the rate of
change of frequency in the system following an outage event, and grid system congestion.
2. EVs accounted for 14 percent of all car sales in 2023 in People's Republic of China,
Europe, and the United States, up from 5 percent in 2020 (IEA 2023).
6 Demand Response And Its Role in the Energy Transition
3. After hourly pricing for residential customers was introduced in Illinois, a study was
conducted to investigate the allocative efficiency and distributional impact of the
tariff change (Environmental Defense Fund 2021). Annual savings from the program
were $29.8 million. Most of the 344,000 customers reduced their electricity bill, for an
average savings of $86.6 per year, but 5,800 experienced increases in their electricity
bills of an average of $11 per year.
4. As countries advance on the pathway to net zero and reach very high VRE penetration
levels, peak demand will increasingly be met by stored energy. However, this stage is
not imminent for the large majority of power systems, large or small.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 7
TWO
CLASSIFYING
DEMAND-RESPONSE
INSTRUMENTS
A variety of demand-response program designs with differing mechanisms and incentive
structures have been implemented around the world. Demand-response interventions can
broadly be classified as either indirect price-based mechanisms or direct quantity-based
mechanisms. Price-based mechanisms rely on time-differentiated rates to induce customers
to shift consumption away from peak hours or during system contingencies. Several types
of time differentiation are possible; they vary in how precisely they reflect real-time grid
conditions. Various quantity-based mechanisms are designed to shape user consumption
in exchange for an incentive payment. Price- and quantity-based demand response is also
known as implicit and explicit demand response, respectively.1
Each mechanism entails different incentives, business models, and response speeds
(seconds, hours, day-ahead) and requires a different level of coordination between the grid
operator, utility company, and the demand-response providers (customers or aggregators)
that will interact with the market design in which the mechanism is deployed. Figure 2.1
FIGURE 2.1
Types of Demand-Response Mechanisms According to the Nature of Incentives
Price-based Time-of-use Fixed ToU retail tariffs eliciting a demand-response in relation
(indirect/implicit) (ToU) to predictable patterns of system stress (demand peaks and
solar peaks)
Narrow and high peak time price incentives to reduce demand
Consumer responds Critical peak
reflecting both predictable periods of system stress and
tomarket prices by pricing (CPP)
network investment drivers
reducing consumption
Variable peak As per CPP but peak price level is variable dependent on level
pricing (VPP) system stress (may be linked to wholesale market prices)
Real-time Fully flexible pass-through (may be within a hedge) of dynamic
pricing (RTP) pricing to retail tariffs
Peak-time Payment for reduced consumption relative to baseline at
rebate (PTR) peak periods
Direct bidding/contracting in wholesale energy markets
Quantity-based Demand-side
(day-ahead, intraday, balancing; where no market exists via
(direct/explicit) energy offers
bilateral contract and dispatch rules)
Demand-side Direct bidding/contracting in capacity markets
capacity offers
Consumer receives a
direct payment for
reducing load Auto load control/ Remote management of devices (tune or turn off/on) capable
interruptible load of supporting network and system operation services
Manual load control/ Manual management of devices (tune or turn off/on) capable
interruptible load of supporting network and system operation services
Source: Adapted from Morales-Espana, Martinez-Gordón, and Sijm (2022) and NERC (2013).
Note: Some quantity-based instruments entail the offer of both price and quantity pairs to the system operator
or market operator from which least-cost dispatch can be selected (sometimes involving co-optimization
between wholesale and ancillary service markets). Here, these are termed quantity-based to differentiate
them from pure price-based approaches where there is no explicit offer of a given quantity.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 9
describes the different types of demand-response mechanisms, grouped into price- and
quantity-based demand-response mechanisms.
The various demand-response mechanisms are not mutually exclusive. Most can be
implemented in conjunction, targeting various customer segments and markets. For
example, some countries combine time-of-use (ToU) rates for smaller customers with
real-time pricing (RTP) for large customers. Others allow consumers to participate in either
mechanism, depending on their preferences. The existence of, and access to, competitive
wholesale and retail markets may drive choices and decisions. Small retail customers may
have a form of time-differentiated rates and still participate in load control. This option will
likely gain momentum with electric vehicles (EVs), because charging patterns are a crucial
target for incentivizing demand shifting, while their batteries offer a potentially valuable
storage resource. Small customers may be subject to ToU rates and still participate—via
aggregators—in several demand-response programs at the wholesale level. Alternatively,
large customers can often participate directly in the wholesale market (via market pools or
bilateral contracts).
The alignment of each of the above demand-response instruments with the critical
objectives of meeting the least-cost supply-demand balance, ensuring capacity adequacy,
and delivering the suite of ancillary services to the system operator, is illustrated in
Table 2.1.2
TABLE 2.1
Spectrum of Demand-Response Interventions
TOOLS PRICE BASED DIRECT PAYMENT BASED
TOU CPP VPP RTP PTR DR ENERGY DR AUTO MANUAL
OFFERS CAPACITY LOAD LOAD
OFFERS CONTROL CONTROL
Least-cost Energy Yes Yes Yes Yes Yes Yes
provision
Capacity Adequacy Yes
Reserve Provision Yes Yes
Frequency Regulation Yes
Voltage Control Yes
Constraint Management Yes Yes Yes Yes Yes Yes
Source: Author’s analysis.
Note: CPP = critical peak pricing; DR = demand response; PTR = peak time rebate; RTP = real-time pricing;
ToU = time-of-use; VPP = variable peak pricing.
10 Classifying Demand-Response Instruments
Price-Based Demand-Response Instruments
Price-based mechanisms (or implicit demand response) establish time-differentiated retail
electricity rates. Unlike uniform pricing, rates are typically differentiated during peak,
valley, and off-peak hours, or with even more granularity.
The four basic types of price-based instruments vary based on how closely they reflect
grid and market conditions. Static ToU prices are fixed at predefined intervals typically
differentiated during peak, valley, and off-peak hours. At the other end of the spectrum
is RTP, which can vary in real time based on grid conditions. Other types of price-based
instruments, such as critical peak pricing (CPP) and variable peak pricing (VPP), are
somewhere in between (Figure 2.2).
FIGURE 2.2
Types of Price-Based Instruments
Static ToU pricing Critical peak pricing Variable peak pricing Real time pricing
Market
€/kWh linked peak
pricing
Time Time Time Time
Source: IRENA 2019a.
Note: kWh = kilowatt-hour; ToU = time-of-use.
Price-based instruments can reflect time-differentiated costs associated with wholesale
electricity generation and electricity network infrastructure. The former is driven by
variances in the marginal cost of production as more expensive forms of generation are
called upon to meet peak demand or periods of low variable renewable energy output.
These variances are best priced in volumetric terms (US dollar per kilowatt-hour, $/kWh),
reflecting the variable cost of peak generation plants. Network costs, by contrast, are
overwhelmingly fixed and driven by the cost of meeting peak demand ($/megawatt [MW]).
Cost-reflective tariffs, therefore, seek to reflect this pricing structure with volumetric
elements that vary with usage and fixed elements represented by standing charges (a flat
monthly cost regardless of use) and demand charges priced to reflect contracted capacity
or maximum demand. However, fixed costs are frequently recovered in part or in whole
through volumetric pricing. Countries may, therefore, use price-based instruments to
reflect network costs through a CPP approach or by having customers pay (higher) demand
charges ($/MW) focused on usage during peak periods.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 11
Fixed Time-of-Use Tariffs
ToU is a static pricing mechanism, with the price of electricity varying according to
preestablished time intervals, which may be daily or, in some cases, seasonal. ToU tariffs
may be implemented to reflect the underlying costs of generation and the grid at different
times. If the challenge is to meet peak loads, the most common form of ToU is that in
which tariffs differ for energy consumed during peak and off-peak hours (volumetric
charges), reflecting the difference in generation costs.3
ToU tariffs are the most commonly used price-based demand-response instrument
and are more frequently applied to commercial and industrial customers (Box 2.1).
survey conducted in 2019 by the World Bank in 65 countries (mostly middle- and low-
income countries) showed that large industrial and commercial customers in about
40 percent of the countries surveyed are subject to ToU rates. Residential customers,
by contrast, were subject to ToU tariffs in less than 10 percent of the countries
surveyed (Figure 2.3).
BOX 2.1
TIME-OF-USE TARIFFS FOR COMMERCIAL
AND INDUSTRIAL CONSUMERS IN COSTA RICA
Costa Rica has used time-of-use rates, which are targeted to large industrial and
commercial customers, with time differentiation for both the energy tariff (in
$/MWh) and for the demand charge (in $/MW)—the so-called “two-part tariff.”
The day is divided into three periods: peak (5 hours per day), valley (9 hours per
day), and night (10 hours per day). The ratio of peak to off-peak rates is high by
international standards. Demand charges during peak hours are 2.24 times
higher than during night hours, and energy charges are 4.4 times higher. This
price differentiation strongly incentivizes customers to shift nonessential loads
and production schedules from peak to off-peak hours. Despite the lower ratio
for demand charges, customers have an extra incentive to control load and avoid
peaks because demand is measured over 15-minute intervals; the maximum
monthly read-out serves as he basis for the payment of demand charges for the
entire month.
12 Classifying Demand-Response Instruments
FIGURE 2.3
Share of Countries with Time-of-Use Rates, by Customer Class
100%
Share of countries with ToU rates (out of 65 countries) 90%
80%
70%
60%
50%
40%
30%
20%
10%
0%
Residential Commercial Industrial Agricultural Public
Source: Foster and Witte 2020.
Note: Based on a database contained in the World Bank Regulatory Indicators for Sustainable Energy,
which includes tariff schedules of 65 developed and developing countries: 7 high-income countries,
35 middle-income countries, and 23 low-income countries. ToU = time-of-use.
A 2013 study of US and European utilities showed that ToU is the most prevalent type of
price-based demand response. Table 2.2 presents price-based instruments in reviewed
European countries and US states. These instruments can be implemented as mandatory,
opt-out, or opt-in. The choice of modality will affect adoption rates. Opt-in modalities,
which rely on customers choosing to sign up for the program, tend to result in much
lower subscription rates than opt-out modalities. Mandatory rates are the most effective
for subscriptions. In jurisdictions with competitive retail markets, it is typically the
prerogative of the individual retailers to design tariff structures and options that they
consider most attractive to their customers, although some regulatory requirements
may still apply.
ToU tariffs are generally simple and inexpensive to implement. However, static ToU
tariffs do not reflect the system’s condition at any given moment, and the flexibility of
ToU rates is limited. This may become an increasingly important limitation as the share
of VRE grows and periods of system stress become less predictable owing to low wind
or high levels of cloud cover, threatening generation shortages outside of peak demand
periods.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 13
TABLE 2.2
Time-Differentiated Modalities in Various Energy Markets
ENERGY MARKET TYPES OF PRICE-BASED MODALITY ACCEPTANCE (%)
INSTRUMENTa
Colorado (Fort Collins) ToU Mandatory 100
California (Pacific Gas and Electric, Southern California ToU Opt out (in 2020) 75–90
Edison, San Diego Gas and Electric)
California (Sacramento Municipal Utility District) ToU Opt out 75–90
Michigan (Consumers Energy) ToU Opt out (in 2020) 75–90
Spain Real-time pricing Opt out 40b
Arizona (Arizona Public Service) ToU Opt in 57
France ToU Opt in 50
Arizona (Salt River Project) ToU Opt in 36
Oklahoma (Oklahoma Gas and Electric) VPP Opt in 20
Great Britain ToU Opt in 13
Source: Adapted from Faruqui and Sergici (2013).
Note: ToU = time-of-use; VPP = variable peak pricing.
a
These are typically rates applicable to the megawatt-hour (MWh) component (volumetric).
b
The adoption is low despite opt-out. This is likely because customers were wary of real-time pricing due to
the lack of understanding and uncertain impact on the electricity bill.
Critical Peak Pricing Tariffs
CPP involves a dynamic rate wherein prices rise significantly on critical days when there is
a risk of low reserves or even blackouts. The energy price during those days may increase
several-fold, reflecting the underlying generation costs and providing a significant incentive
for customers to adjust their load profiles. Under a CPP structure, the utility notifies
customers a day in advance and sometimes even on the day of the event. This tariff applies
for a maximum number of critical days per year. Box 2.2 illustrates the case of South
Africa, which has a CPP tariff for large clients.
CPP has been applied to a smaller extent in France, Lithuania, Portugal, and Romania
(ACER 2016). France has been using the so-called “Tempo” tariff for many years. Under this
scheme, each year is divided into three color-coded tariff categories: a maximum of
22 days is considered critical (and the highest tariff applies), 43 days are categorized
as requiring attention (high tariff), and at least 300 days are considered standard
(regular tariff).
Some US states, such as California, have introduced a type of CPP rate called CPP-variable
for which the number and duration of peaks are not set ahead of time. The current
14 Classifying Demand-Response Instruments
BOX 2.2
SOUTH AFRICA’S EXPERIENCE WITH CRITICAL
PEAK PRICING
In South Africa, where electric heaters and stoves drive winter peak consumption,
the government implemented the Nightsave program in 2020 to reduce peak loads.
On critical days, the energy cost would jump from US¢3.2/kWh to US¢18.2/kWh.
The peak–to–off-peak ratio (5.7) is thought to be high enough to motivate
customers to change their consumption profiles. South Africa has been working
to improve the program’s pricing methodology. A new critical peak pricing (CPP)
scheme was recently piloted by the state utility Eskom, offering two options for
customers: 16 hours per day (06:00−22:00) for 25 critical peak days, or 8 hours
per day (06:00−14:00) for 50 critical peak days, up to a maximum of 400 hours
per year. The extended duration of the critical periods reflects the constraints
that the power sector faces in South Africa. Both programs had excellent uptake,
and the business processes for a dynamic tariff have been tested, with critical
lessons learned for large-scale implementation of CPP. The 8 hours per day scheme
had better uptake. According to initial surveys, customers were able to manage
load reductions. But when the number of critical peak days in a week or month
was too high, customer fatigue prevented significant load reduction. Therefore,
dispatching the 50 days throughout the year requires careful implementation by
the system operator. The benefits of CPP depended on Eskom avoiding higher
open cycle gas turbine generation costs and reducing the risk of unexpected
blackouts.
CPP-variable program for Southern California Edison, for example, provides four months
of summer-season bill credits in exchange for customers paying higher electricity prices
during 12 to 15 annual CPP events. The substantial difference between peak and off-peak
rates is a powerful incentive for customers to react during critical periods (between 4 p.m.
and 9 p.m.). Customers are notified a day before a CPP event occurs.4
Variable Peak Pricing Tariffs
VPP is a hybrid of ToU and RTP (discussed below). With VPP, as with ToU pricing, peak and
off-peak intervals are predetermined, but unlike with ToU, during the peak period VPP
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 15
customers are charged a rate that varies according to the utility/retailer and usually
reflects the wholesale price of electricity. Because peak prices emulate market prices for
electricity, VPP rate designs more accurately match the cost of producing and distributing
electricity. The risk of high power prices is shifted during peak periods to customers, who
can respond by reducing consumption. Box 2.3 describes Oklahoma’s VPP scheme.
BOX 2.3
SmartHours VARIABLE PEAK PRICING, 2023
With SmartHours VPP, the daily peak price varies between low, standard, high,
and critical rates. Customers can check pricing information or receive day-ahead
price notifications via email, phone, or text from Oklahoma Gas & Electric. On
occasion, high energy demand may cause a critical event. Although these events
are rare, they can occur at any time of day and any time of year and last up to
eight hours. Oklahoma Gas & Electric sends critical event notices at least two
hours ahead to enable customers to prepare.
In the case of time-of-use (ToU) pricing, peak hours are charged at US¢26. With
VPP, off-peak hours are US¢8; standard hours, US¢13 (half of the average ToU);
and peak hours, US¢48.
Real-Time Pricing Tariffs
Under RTP, the energy (MWh) component of tariffs to the end users reflects spot-price
variations (typically day-ahead) in the wholesale market.5
Some countries and regions have been trying a variety of dynamic pricing mechanisms to
provide a better linkage between prices at the wholesale and retail levels (Faruqui 2005).6
For example, in several European markets and New Zealand, energy retailers in competitive
markets have offered dynamic RTP tariff options to various customer categories. These
offers stand alongside more traditional fixed and static ToU tariff options and are crafted
to appeal particularly to owners of solar photovoltaic systems and EVs.
The United States has been trying to implement RTP for a while now, but customer uptake
has been slow. Reasons for the slow uptake are several and vary by state. Customers’ fear
that savings are uncertain is one reason. Their apprehension makes them hesitant to take
16 Classifying Demand-Response Instruments
on the risk of exposure to volatile spot prices. Another reason is that, in some states where
retail competition is more active, some commercial and industrial customers are no longer
interested in the regulated RTP rates offered by the utility because they can procure energy
from alternative suppliers in the nonregulated market. Those customers may still be
exposed to—and respond to—spot price variations (Barbose and others 2005).
The essential question is who bears the price risk, the retailer or the consumer? In a
competitive retail market with various options available, consumers will self-select the
degree of risk they are willing to take in return for the potential to lower their bills through
demand response.
Spain has implemented elaborate price-based instruments with a hybrid of RTP and ToU.
Generation costs are priced in real time, considering variations in the wholesale energy
cost. Grid costs are priced in three differentiated time intervals. A summary of the current
system is presented in Box 2.4, with additional information provided in Appendix B.
BOX 2.4
HYBRID PRICE-BASED INSTRUMENTS IN SPAIN
Spain has implemented sophisticated time-differentiated rates for residential
customers to better reflect underlying generation and grid costs and
encourage customers to change their consumption patterns accordingly.
Since 2021, electricity bills have had the following components:
• System access fees to recover regulated costs (network costs and a group of
energy- and policy-related costs, such as subsidies for renewables and system
operation) are standard for all residential customers.
• A capacity charge is based on the capacity that each customer contracts; it is
enforced via automatic disconnection. This component is offered in two
options: one is fixed; the other is provided in peak and off-peak intervals keyed
to EV evening charging.
• A network access energy component (€/kWh) has three prices—peak, low, and
average. The average charge is about 10 times the off-peak price, conveying a
powerful signal for load shifting.
• A dynamic hourly price (€/kWh) results from the direct pass-through of the
published day-ahead and intraday market prices, plus the cost of ancillary
services for the day after.
(continues)
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 17
BOX 2.4 (Continued)
The scheme should lead to savings of around 3.4 percent for 19 million households
billed using time-differentiated rates (CNMC 2022). It poses an additional expense
of €2 a month for another 8 million households. Changing habits could result in
annual savings of €200 to €300 (CNMC 2022). According to consumer organizations,
the average customer could save up to €574 per year if appliances were used half
the time during the cheapest periods. However, there has been concern that the
new methodology will penalize smaller consumers and benefit larger ones.
Source: CNMC 2022.
Effectiveness of Price-Based Demand Response
Price-based instruments (optional or mandatory) are essential to foster implicit demand
response in the retail market. In 2017, a detailed study (Faruqui, Sergici, and Warner 2017)
was conducted to analyze the impact of such mechanisms on electricity consumption. The
study included 334 case studies from various parts of the world categorized into six basic
pricing categories: ToU, peak-time rebates,7 and CPP, each with and without the use of
technologies to manage consumption during peak hours (Figure 2.4).
The analysis revealed significant variation in performance across the six pricing categories
and within each one. The study found that enabling technologies such as smart
thermostats can significantly boost the effectiveness of pricing schemes. For instance, in
the case of traditional ToU pricing, peak reductions across the sample of projects ranged
from virtually zero to about 25 percent, with a median of approximately 7 percent. Still,
with the application of technology, the peak reduction ranged from 5 percent to almost
40 percent, with a median of about 16 percent. More sophisticated tariff mechanisms, such
as CPP, resulted in peak savings ranging from 2 percent to 60 percent, with an average of
18 percent. When technology is applied, the peak reduction stays within the same range,
but the median increases to more than 30 percent.
Several factors explain these significant differences in performance. The main one is the
design of the tariff mechanism. Tariff design elements include the price ratio between
peak and off-peak periods, the length of the peak period, and the number of pricing
periods.
18 Classifying Demand-Response Instruments
FIGURE 2.4
Average Peak Reduction from Time-Differentiated Rate Pilots
70%
60%
Time-of- Peak Peak Time Critical Peak
Critical Peak
Time-of-Use Use w/ Time Rebate w/ Pricing w/
Pricing
Tech Rebate Tech Tech
50%
Peak Reduction
40%
30%
20%
10%
0%
Pricing Treatment
Source: Faruqui, Sergici, and Warner 2017.
Note: “w/ Tech” signifies the application of smart technology.
Quantity-Based Demand-Response
Instruments
Several quantity-based instruments exist, encompassing interruptible contracts, direct
or automated load control, and demand-side bidding in wholesale energy and capacity
markets. These mechanisms may overlap depending on the structure of a particular power
sector or market. All such mechanisms explicitly offer a combination of a given quantity
level (lower demand relative to baseline) for a given price.
Interruptible Contracts
An interruptible contract is an arrangement in which a transmission or distribution system
operator makes a payment to a customer or lowers tariffs in exchange for the ability to
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 19
reduce or interrupt the customer’s electrical service during system contingencies.8 This
enables some level of “dispatchability” (defined as the possibility of load response by the
system operator) under the conditions spelled out in the agreement, such as notice period,
duration of curtailment, and frequency.
Interruptible contracts are very effective. For example, they proved extremely important
for the system operator in managing the 2001 power crisis in California. The rolling
blackouts would have been more frequent and widespread without them (Sweeney 2013).
South Africa has a negotiated pricing agreement with an aluminum smelter that provides
two hours of interruptibility for an agreed maximum number of events per year. In several
developing countries, the ability to interrupt large users is an additional balancing resource
for the grid operator. Still, it happens ad hoc, without standard protocols for interruptibility
or compensation. Interruptible contracts will remain important in future demand-response
plans, particularly when operations can be automated, enhancing the speed and value of
demand response (Hledik and others 2019).
Load Control
Distribution utilities and power system operators can selectively control loads remotely
through direct load control. Load control may be manual (requiring an end-user response)
or automated. In the latter case, switches are activated and deactivated without human
intervention. Typically, load controls are opt-in mechanisms, and the customer subscribing to
the service receives some form of monetary compensation for allowing the utility to control
load during critical period “events.” Unlike with interruptible tariffs, direct load control does
not fit into a particular customer tariff group. Special incentives are set for those users,
across customer categories, who subscribe to the utility’s load control programs.
In all customer categories several pieces of equipment may be controlled. At the residential
level, EVs are strong candidates, alongside cycling, air-conditioning, or heat pumps
(controlled via switches or smart thermostats). At the commercial level, air-conditioning
and refrigeration can be effectively controlled. Some simple but fast-responding technologies,
such as water heaters, can autonomously respond to control signals. At the industrial level,
the load control becomes process- and end-use-specific in situations in which loads can be
interrupted temporarily or where there is some form of storage in the system (electric or
thermal). Load control makes sense because it benefits the system and causes only modest
disruption to users. Table 2.3 lists some kinds of loads that can be controlled and classified
according to customer segment. One of India’s early pilots of load control for medium-
sized and large customers is profiled in Box 2.5.
Demand response is a beneficial, cost-effective way to provide many ancillary services,
including voltage and frequency regulation, load following, and reserves, compensating for
a lack of rotating inertia that most inverter-based renewables cannot provide.
Future demand-response resources should be able to provide an even faster response than
traditional demand response (down to milliseconds), with the speed of response ranging
20 Classifying Demand-Response Instruments
TABLE 2.3
Loads that can be Controlled According to Customer Segment
END-USE TECHNOLOGY SECTOR
RESIDENTIAL COMMERCIAL INDUSTRIAL
Electric vehicles Yes Yes
Plug-in hybrids Yes Yes
Air-conditioning and heating Yes Yes
Water heaters Yes Yes
Pool pumps Yes Yes
Lighting Yes
Refrigeration Yes Yes
Process industry and large facilities Yes
Agricultural pumping Yes
Data centers Yes
Wastewater treatment Yes
BOX 2.5
INDIA: LOAD CONTROL FOR LARGE
AND MEDIUM CUSTOMERS
India has been implementing several pilots on load control. The first was designed
by Tata Power in Mumbai. It was implemented in 2012 to reduce load during peak
shortages, transmission line tripping, and generator set tripping, as well as when
power purchase costs were high. Demand-response events were limited to a
maximum of 2 hours, totaling 100 hours a year. Participation was optional, and
customers could opt out at any time. Participating customers included malls, hospitals,
municipal sewage plants, IT parks, and airports. The figure below illustrates the load
reduction during a demand-response day. The peak load was reduced from about
55 GW to 42 GW, representing a significant contribution of demand response when the
actual load is compared to the estimated load if no demand response had occurred.
(continues)
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 21
BOX 2.5 (Continued)
Load Control Pilot Program by Tata Power in Mumbai on a Demand-Response
Event Day
60
50
Half hourly load (MW)
40
30
20
10
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
Curtailed Load (MW) Estimated load curve without DR Load curve with DR
Source: Tata Power 2012.
Note: DR = demand response; MW = megawatt.
from a few days in advance (in anticipation of heat waves or forecasted decline in wind
production) to 30 minutes or less (e.g., to address system contingencies, sudden variation
in PV production, or an increase in EV charging).
Demand-Side Energy Offers
Demand response can also be part of wholesale energy provision. This may include
demand-side bidding (DSB) in organized energy pools at day-ahead, intraday, or real-time
(balancing) markets to compete with supply-side offers. DSB, in this way, requires a
functional wholesale pool market with central dispatch. Where such a market does not
exist, demand response must contract via generators or suppliers trading on a power
exchange. Should there be no wholesale market of either form, then a bilateral contract
with the system operator, dispatching the demand response according to their defined
dispatch rules, would be necessary. DSB offers under a merchant market approach
(i.e., without long-term contracts committing availability) are essentially voluntary, the only
penalty for failure to deliver being the prevailing electricity price. This has led to some
22 Classifying Demand-Response Instruments
system operators not perceiving the mechanism as a flexible, dependable demand-
response option to help mitigate expected supply shortages (for example, reductions in
generation) or demand increases (for example, during heat waves) (Cappers, MacDonald,
and Goldman 2013).
Demand-Side Capacity Offers
Consumers can also participate in demand response in some capacity markets by bidding
on contracts for demand reduction. The grid operator instructs winning bidders to reduce
load when required. Unlike the DSB mechanism, bidders receive a fixed payment for
availability, as grid operators need a “dependable” resource to be available for dispatch
when needed. Demand-response mechanisms should be designed to compete on a level
playing field with generators, which makes demand response equivalent to a supply
resource.
System operators must count on dependable and reliable resources when dispatching
power. Some demand-response mechanisms may be beneficial in theory but not
considered strong enough to give system operators the assurance that the resources
will be available for dispatch when needed. There is a perception that demand-response
programs are not as dependable as thermal generation plants, which can presumably
be dispatched on demand when necessary.
There are indeed cases when demand-response resources were overestimated and
unavailable on request. For example, in 2001 in California, about a third of the resource
adequacy requirements met by demand response was not available or directly accessible
to the system operator in peak net load hours on days where “Flex Alerts” or system
warnings were issued during heat waves (California ISO 2022). The Pennsylvania–
New Jersey–Maryland interconnection has encountered similar situations in the winter
(FERC 2023). There is a concern that demand-response programs used to meet resource-
adequacy requirements are significantly overcounted compared with the actual availability
of these resources, particularly during peak net load hours.
The key message is that availability mechanisms should be well designed, with frequent
testing for demand response and generation assets and with penalties for nonavailability.
Measurement and Verification
In an explicit demand response scheme, electricity customers can offer to reduce
their consumption, individually or through aggregation by an intermediary. However,
measuring such a reduction requires the identification of a customer baseline load (CBL).
A counterfactual must be established, as only actual consumption can be observed.
Such a counterfactual is necessary to measure a demand resource’s effective performance
and adequately compensate the provider.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 23
Determining a CBL is challenging and will inevitably involve uncertainty and error. End
users’ electricity consumption is variable for several reasons unrelated to demand-
response interventions. Weather conditions, production schedules, seasonal variations in
firms and household needs, holidays, and other factors strongly affect the amount of
electricity a customer consumes, independent of any price variation or incentive payment.
Attempts to account for these aspects may help reduce (but not eliminate) the related
uncertainty and increase efficiency to the extent they do not make the mechanism’s design
too complex. Conversely, the less predictable consumption is, the greater the scope for
error and manipulation in deriving a CBL, thereby reducing the scheme’s efficiency.
There are several methodologies for estimating CBL. Their suitability depends on the
nature of the demand-response program and local measurement and verification
standards. CBL estimations should balance various desirable criteria, including accuracy,
simplicity, and integrity. The choice of the best methodology depends on factors such as
the function the relevant demand-response mechanism fills in the system, the broader
regulatory framework for demand-response participation in wholesale markets, and
the characteristics of the demand-response providers.9 The most common approach is
using historical metered data, possibly amended by season and weather, with statistical
processing to improve its use as a predictor of future use. The result is an assumed load
curve representing the CBL of the consumer, or aggregated set of consumers. An example
from Elia, Belgium’s transmission system operator, follows.
Elia uses a set of different baseline methodologies to fit the various demand-response
applications it deploys, all developed in conjunction with stakeholders to ensure they
match the specific delivery characteristics of the providers and applications in question
(Elia 2021). For manually instigated reserves, a combination of the last 15-minute interval
plus a “high X of Y” approach is used, whereby the latter refers to taking the average of the
highest X number of days among a set of Y for the equivalent hours in the preceding Y days.
These results are adjusted under a “same-day adjustment” process. This adjustment seeks
to account for weather-related impacts, as illustrated in Figure 2.5.
Elia considers its baseline approaches (including a declarative approach for automated
frequency response services) broadly in line with best practices internationally. However,
considering the diurnal volatility in output and consumption profiles, concerns persist
about their future suitability for harnessing residential flexibility.
Endnotes
1. Alternative classifications are often used. For example, behavioral demand response
relies on behavioral nudges and incentives to induce customers to shift or shed
load. These are cost-effective, as they do not depend on any specific technology
investments. Over time, however, they may cause customer fatigue, rendering them
less effective.
2. See appendix A for details on ancillary services.
24 Classifying Demand-Response Instruments
FIGURE 2.5
Elia’s Same-Day Adjustment Process
800
Load drop with adjustment
700
600
Electricity Demand (kW)
500
400
Load drop w/out adjustment
300
3-in-10 Baseline DR Event
200
Adjusted 3-in-10
Baseline
100
Actual Usage Morning-of Adjustment Window
0
AM
AM
AM
AM
AM
AM
AM
AM
AM
AM
AM
PM
PM
PM
PM
PM
PM
PM
PM
PM
PM
PM
n
t
gh
oo
ni
1
2
3
4
5
6
7
8
9
10
11
1
2
3
4
5
6
7
8
9
10
11
N
id
M
Source: Elia (2021), citing California Independent System Operator.
Note: DR = demand response; kW = kilowatt.
3. These are typically either two-tier or three-tier. Interest in recent years has shifted
from simple daytime and night-time splits to a focus on separate pricing for peak
demand periods (typically evening).
4. Southern California Edison, https://www.sce.com/business/rates/cpp.
5. Spot prices may be calculated with different levels of granularity, ranging from five
minutes to one day depending on the existing settlement rules in the wholesale
market. In bilateral contracting markets using self-dispatch, a liquid power exchange
may be referenced. Where no wholesale market exists, a shadow price may be derived
from the system’s short-run marginal cost.
6. In competitive retail markets these tariff options may be offered by individual retailers.
7. Peak-time rebates are monetary incentives offered to customers who reduce their
electricity consumption during periods of high-cost electricity. Those customers
who do not reduce usage during peak periods are simply charged the rate normally
applicable during those periods. The rebates may be considered similar to a static
ToU scheme but with a sharp, differentiated peak period, the reward being an explicit
payment as opposed to implicit saving.
8. The operator, particularly at the distribution system level, may be bundled with the
retail entity, in which case the network tariff is aggregated with energy costs.
9. A detailed analysis of CBL methodologies exceeds the scope of this report.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 25
THREE
INCREASING THE
UPTAKE OF DEMAND
RESPONSE: ENABLERS
AND BARRIERS
The potential for demand response is far more significant than current levels of deployment
and utilization suggest. This chapter describes critical enablers and barriers to implementation
and uptake. Some enablers are recent, such as new technologies affecting loads that offer
demand response. Others refer to remote management, and specifically the communications
and data infrastructure for monitoring, processing, and relaying information in real time.
The perceptions of system operators, other energy sector entities, and consumers also
play a role. Incentives must be aligned for regulated entities, while customer engagement
can help overcome consumer fears and produce greater understanding and buy-in.
Enablers
Technology and Data
Technology is a critical enabler of mainstreaming price- and quantity-based demand-
response programs. Many technologies already support demand-response programs. New
technologies will emerge to support more granular, faster demand response to meet the
needs of future power systems amid the expanding penetration of variable renewable
energy. Existing technologies include bidirectional communication, smart metering,1 smart
controls, smart thermostats, and energy management systems affecting the end user.
Digitalization will strengthen the opportunities for demand response and behind-the-meter
generation with smart inverters.
Smart meters deployed in most countries have helped lower losses and operating costs,
but they can also support the mainstreaming of demand-response programs. Advanced
metering infrastructure tracks the consumption of individual consumers. Smart meters
record consumption on hourly, half-hourly, or quarter-hourly bases, allowing retailers to
refine tariff structures for energy production and supply costs. They are necessary for
market-based pricing schemes. Advanced metering infrastructure enables two-way
communication between suppliers and customers. It integrates additional technologies,
such as web-based portals, allowing customers to analyze their hourly electricity use,
compare their use with that of other local consumers, and gather information about
options to manage their electricity consumption better (IRENA 2019b). Smart meters
can also help implement remote load control if equipped with relays (on-off switches).
A broad range of technology options—including automation equipment, smart devices,
storage, smart inverters, and communication infrastructure—are available to support
quantity-based demand-response programs with various goals and degrees of
complexity.
Ripple control—a load control technology in which a higher-frequency signal (carrier)
is superimposed onto the standard 50–60 Hz of the main power signal—enables some
equipment to be switched on and off remotely, reducing consumption during peak hours
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 27
or contingencies on a systemwide or local level. The relatively simple devices required do
not require smart metering infrastructure. Protocols and incentives should be agreed
upon beforehand with the customer. It is a mature technology; the Czechia and New
Zealand started using it in the 1950s. In emerging markets, it has been used in southern
Africa for many years.2 Box 3.1 illustrates the example of water heaters in Botswana. Simple
load control switches use one-way communication that cannot be verified. Modern control
systems can be web based and benefit from switching on/off features embedded in
smart meters.
BOX 3.1
CONTROL OF HOT WATER IN BOTSWANA
Like South Africa (see section 5, under “Demand Response in South Africa,” for a
detailed discussion), Botswana implemented a hot water (“geyser”) load control
program in 2010. Prepaid smart meters were installed in key areas, enabling
remote load control by the utility, the Botswana Power Corporation. This program
reportedly achieved a peak load capacity reduction of 20 MW with an estimated
potential of 40 MW.
In 2015, Botswana experienced a supply shortfall caused by deficits in the
Morupule B generator and constrained imports from Eskom. In response, the
Botswana Power Corporation implemented a load management program to
reduce load-shedding events. The program was tasked mainly with lowering
demand during the four-hour peak periods of 6–10 a.m. and 6–10 p.m.; the
schedule was posted on the Botswana Power website. If the program were
called upon, household and small business consumers were required to switch
off all nonessential appliances to limit their load to less than 10 amperes.
Consumption exceeding 10 amperes would be curtailed by the utility, which
would notify the customer via an alarm. The customer would have three
minutes to reduce their load below 10 amperes; if the reduction were not
achieved, their load would be interrupted for 60 minutes. The 10-ampere limit
was set to allow usage of typical 10 CFL lights, refrigerators, televisions, and
decoders (BPC n.d.).
28 Increasing the Uptake of Demand Response: Enablers and Barriers
Remotely controlled thermostats evolved from traditional load controls. A new generation
of smart thermostats (for example, Nest by Google) can be programmed and controlled
online for heating, ventilation, and air-conditioning applications. They rely on two-way
communication and can be used for variable load shedding, which allows for more refined
control, possibly precooling by the grid operator, in anticipation of a heat wave. Home
energy systems enable a comprehensive range of loads to be switched sequentially.
The required process entails two-way communication, leveraging behind-the-meter
automation.
Smart appliances can work in two ways; their use may be deferred in response to price
signals or provide frequency services to the grid. Autonomous responses are critical for
reliability and include refrigeration appliances that offer frequency regulation to the
system operator. Both approaches are forms of demand response. The incremental cost
of building the hardware and software to increase those capabilities is modest (around
$20 per refrigerator).3 More information about load control possibilities is provided in
appendix C and Box 3.2 on electric vehicles (EVs) and vehicle-to-grid (V2G).
The use of enabling technologies boosts the impact of tariff mechanisms. Figure 3.1
illustrates the impact of tariff design and enabling technology on peak demand reduction.
The graph has two curves. The lower one represents the impact of sending price signals to
customers without smart technologies. The vertical axis shows the percentage reduction in
peak demand induced by various peak and off-peak prices. Customers respond to higher
peak to off-peak ratios by lowering peak demand but at a diminishing rate. For instance,
at a price ratio of 2:1, the drop in peak demand is a little more than 5 percent; at 4:1, the
reduction is 10 percent; and at 6:1, it is a bit more than 12.5 percent.4 The upper curve
shows the impact on peak response when technology is deployed.
FIGURE 3.1
Price Responsiveness With and Without Emerging Technology
60%
Peak Impact
40%
20%
0%
2 4 6 8 10
Peak to Off-Peak Price Ratio
Rate Design Enabling Tech Price Only
Source: Faruqui, Hledik, and Sergici 2019.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 29
BOX 3.2
ELECTRIC VEHICLES AND VEHICLE-TO-GRID
Electric vehicles (EVs) are a uniquely exciting technology and present a potentially
significant source of demand response for power systems. They can make two
possible contributions as a resource. The first uses demand response to manage
the load by charging batteries off-peak and discharging during peak hours. The
second represents an emerging business model that uses EV storage capacity to
provide services to the grid.
The EV is typically the most significant individual load among households with an
EV. A level-2 charging station, used in most households with EVs, draws 7 kW from
the grid. Charging EVs is a sizeable increment in peak consumption, particularly
because so many EVs are plugged in as drivers return home in the early evening
(peak coincident). Utilities are concerned about the additional investment in
peaking plants and grid expansion needed to accommodate this incremental
load. In a scenario of growing EV penetration, demand response becomes
necessary. The grid will be overloaded if all electric car owners decide to charge
simultaneously. The good news is that EV charging has a low-capacity factor (for
example, two hours per day), so this charging period can easily be shifted to
off-peak hours. Utilities are designing special programs and encouraging
customers to participate in demand-response programs. Special price-based
instruments, including critical peak pricing and real-time pricing, are being
offered, and in some cases, EVs are being treated as separate loads requiring
submetering (Southern California Edison 2023). The demand-response process
can be automated because most EVs have features that enable them to be
programmed to be charged off-peak, making the demand-response program
more effective.
EVs as an energy source is a breakthrough in terms of demand-response
provision.a Vehicle-to-grid (V2G) takes advantage of EVs’ large electric battery
capacity, which sits idle most of the time. This storage capacity will far exceed
the storage capacity in stationary batteries. V2G will increase flexibility and grid
resilience (smart grid). This business model entails using parked EVs as a power
source for the electric grid during periods of high demand while returning that
power to the EVs during times of low demand. Studies have indicated that
vehicles are not used for active transportation up to 95 percent of the time
(continues)
30 Increasing the Uptake of Demand Response: Enablers and Barriers
BOX 3.2 (Continued)
(Letendre, Denholm and Lilienthal 2006) and can be used to help the power
system. For end users, V2G technology can provide an additional revenue stream.
EV owners can discharge the stored energy in their vehicle’s battery to the grid
during peak demand periods, earning credits or cash payments. Unlike one-way
smart charging, bidirectional charging should offer daily flexibility to the grid.
V2G offers deep load-shifting benefits and an effective real-time mechanism
for mitigating dynamic network constraints. The flexibility generated from
V2G-enabled devices has the potential to accelerate the uptake of renewables
(Horschig, Özgün, and Jones 2023).
V2G is still under development and testing. Despite its alleged benefits, there are
challenges and concerns.5 V2G has been criticized because its economics may not
favor EVs feeding into the grid. Manufacturers and EV owners have expressed
concerns that using the battery to provide price arbitrage and peaking power
(deep cycles) will reduce the battery life cycle. Using a battery energy storage
system to offer other ancillary services to the grid can be explored. Aggregators
must consolidate multiple loads and offer them jointly to the utility or the system
operator. Some manufacturers are mainstreaming V2G capabilities, and regulators
are encouraging behind-the-meter batteries to provide ancillary services to the grid.
a
Other business models include vehicle-to-home.
Supportive Policy and Regulatory Frameworks
Policies and regulations are essential to enable the adoption of demand-response
mechanisms. Developing countries are at various stages of developing price- and quantity-
based demand-response mechanisms. Some are still piloting time-of-use (ToU) rates,
whereas others are advancing toward dispatchable demand-response mechanisms in the
energy market. The nature and intensity of demand-response interventions will depend
on various enabling factors, including sector structure, enabling technology, and policy
direction. Regulators can support the evolution of appropriate, cost-effective demand-
response mechanisms in several ways:
• Propose new pricing mechanisms to foster customer engagement and demand response.
Propose new tariff methodologies, starting with ToU rates and evolving toward dynamic
tariffs such as critical peak pricing and variable peak pricing. Review barriers and other
incentives for customer engagement (for example, opt-in, opt-out, mandatory).
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 31
• Define the pace of implementation of new tariff regimes, proposing the development of
specific pilots, if necessary. Conduct cost-benefit analysis along the way. Regulations
must ensure that, at the institutional level, all roles and responsibilities are adequately
defined and assigned to each actor involved.
• Support system and market operators to develop and implement quantity-based
demand-response mechanisms, ranging from simple load control to more sophisticated
and flexible demand-response product participation in energy markets.
• Identify winners and losers resulting from the implementation of demand-response
programs and, if possible, eliminate conflicts of interest and perverse incentives.
• Agree on how enabling technology should be deployed and how customers should pay
for it.
• Enable the emergence of new players and innovative business models.
New forms of demand response are being adopted in distribution. Historically, the owners of
distribution networks have been largely passive, involved in connections and infrastructure
delivery but not active as “system operators” instructing connected entities on load and
generation management. This pattern is changing, however, and distribution system
operators (DSOs) are discussed throughout this report. As noted above, smart technology
can give the DSO and service providers greater real-time network status visibility. The
combination of battery storage, EVs with V2G potential, distributed generation such as
rooftop solar, and smart meters all enable a DSO to optimize electricity, and potentially
energy as a whole, in a future, integrated system. These changes may help bring forward
technologies to support decarbonization at the least cost on the one hand. On the other,
these technologies face regulatory and market hurdles because the framework has been
developed to accommodate traditional fossil fuel power generation. Such barriers can
include a default preference by DSOs for wire (rather than non-wire) solutions for managing
system constraints because the former contributes to their regulated asset base on which
they earn a return.
The European Union is working to ensure its member states address such concerns
(European Union 2019). Directive 2019/944 relates to incentives for procuring flexibility
services from providers of “distributed generation, demand response or energy storage
and promote the uptake of energy efficiency measures.” Product categories for such
procurement should ensure effective and nondiscriminatory provision. DSOs must also
publish a network development plan every two years. This requirement provides
transparency both on the flexibility provision in the coming decade and on requirements
to integrate EVs in a way that allows for nonnetwork solutions. The development plan
architects will need to consult system users for input.
A 2022 report by the Centre on Regulation in Europe, “The Active Distribution System
Operator,” describes three phases in this evolution of the DSO’s role (CERRE 2022):
• Phase 1: Unbundling, efficiency-inducing regulation, including better supply continuity
and quality.
• Phase 2: DSOs incorporating distributed energy resources (smart meters, EVs,
photovoltaics, etc.) into their processes.
32 Increasing the Uptake of Demand Response: Enablers and Barriers
• Phase 3: Active management of these distributed assets to optimize their use and
maximize decarbonization.
Demand response is critical in Phases 2 and 3. The right incentives must be in place,
however, for the DSO to make this journey. Mechanisms used by regulators have included:
• Mechanisms that allow a portion of capital expenditure to be passed through to
consumer bills for the highly uncertain expenditure related to the energy transition (for
example, unit costs may be fixed, but volumes passed through—this approach has been
used for smart meter rollout).
• Incentivizing better coordination between transmission system operators and DSOs
where these are disaggregated data exchange platforms paired with increased
digitalization/automation. Such initiatives may allow for better utilization and
penetration of distributed generation (often from renewable sources), more system
flexibility, and potential deferral or avoidance of network investments.
• Innovation funds and regulatory “sandboxing” (limiting derogations from usual
regulatory constraints) to allow pilot processes to be trailed.
• A premium weighted average cost of capital for the allowed returns from certain
investments considered high risk.
There is more in chapter 6 on the steps regulators and policy makers could take in
accelerating the deployment of demand response, with a focus on developing countries.
Customer Engagement
The willingness of customers to take part in demand-response programs depends on a
number of factors driven mostly by incentives and behavioral responses. The perceived
net benefits of participation, perceptions of risk or uncertainty, ease of understanding
the program, and trust in the program provider all play a role in voluntary participation.
One crucial factor is overcoming customers’ fear of new demand-response schemes. Complex
combinations of tariff types and load control programs can seem opaque and prevent
consumers from choosing more dynamic pricing-based instruments or subscribing to load
management programs. Customers are rarely familiar with their load profiles, how their
actions may influence consumption, and the impact of all those factors on their electricity bill.
Given this uncertainty, people often prefer a risk-averse approach, staying with flat tariffs or,
if mandatory, accepting ToU tariffs. Many factors drive customer participation. The first is
pricing structures. High peak to off-peak energy price ratios (that is, >4:1) are more effective at
changing consumer behavior than low ones (that is, 2:1).6 Several demand-response programs
at the residential and small commercial levels allow customers to opt in or out.
Greater customer engagement addresses these concerns. The first step is ascertaining that
proposed demand-response schemes are easy to understand. Information can change
behavior because it facilitates the user experience and provides data about consumption
patterns and the impact of load controls on customer discomfort. Energy suppliers,
distribution companies and system operators, regulators, and trade associations should
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 33
play a key role in communication. Communication strengthens customer understanding
about how bills may change because of new tariff designs. Customers should be given
simulation tools to see how their electricity bills change with different consumption
profiles. Before the new tariff systems are mainstreamed, well-designed pilot programs
should be run to test customer response.
Dynamic tariffs offered to the regulated market must provide a hedge against the most
catastrophic price scenarios. Customers may fear events like the 2021 Texas power crisis,
caused by intense cold weather, when spot prices skyrocketed to about $10,000/MWh.
Customers who had shifted to variable-rate plans were suddenly exposed to spot price
volatility and monthly bills exceeding $3,000.7 Extreme cases amplify customer fear of
accepting different tariff regimes, so additional incentives are needed to change the load
profile in response to market signals.
As people accept and grow familiar with digitalization and automation in their daily lives,
their attitude toward dynamic electricity tariffs may change. In the Republic of Korea, more
than 60 percent of survey respondents preferred real-time pricing over less dynamic tariffs
such as ToU (Clean Energy Ministerial 2014). The wider prevalence of distributed energy
resources, including EVs, engages consumers with their energy use. Therefore, perhaps the
biggest hurdle is reaching critical mass and then explaining why dynamic electricity tariffs
work (Faruqui, Hledik, and Palmer 2012).
Barriers and Challenges
There are several barriers to the implementation of demand-response programs. Some
are the same as those to mainstreaming energy efficiency programs, such as load
fragmentation, customer resistance to changing consumption habits, inadequate
incentives, and undermotivated policy makers and regulators facing thousands of
customers and the need to achieve demand reduction while deploying enabling
technologies. This section details the key barriers to demand-response deployment.
Supply-Side Bias
Utilities generally focus on the supply side (for example, building generation or transmission
facilities) to meet customers’ needs. Utilities are capital-intensive, and customers often have no
choice but to buy energy from the local utility. Several other factors compound the supply-side
orientation. First, interacting with a few existing supply-side participants seems easier and
potentially more cost-effective to the electric power industry than creating new strategies to
include the emerging demand-side resource (FERC 2013). Second, traditionally the utility earns
a guaranteed return on the asset base, while the profit-sharing mechanism when the utility
invests in demand response or energy efficiency is uncertain. There may also be a perception
that having the customer help bridge the supply-demand gap will be interpreted as a failure in
34 Increasing the Uptake of Demand Response: Enablers and Barriers
planning and execution. Consequently, demand-side resources (for example, large-scale
deployment of demand response and load management systems) are frequently overlooked.
Customer Fatigue
An important factor in demand-response programs is customer fatigue. Customers may
tire of making manual load adjustments or tracking prices over time, making the demand-
response program less effective. This is true even for customers who have subscribed to
demand-response programs (price or quantity based).
Technology is a crucial element maintaining effective demand-response programs over time.
Remote consumption control can mitigate fatigue, make manual adjustments, and make
daily decisions about when to activate which appliances. In the residential market, it is
expected that management of distributed energy resources (demand response, self-
generation, storage, EV charging) for demand response will increasingly be automatic
without affecting homeowners. Utilities could be empowered to dispatch equipment
remotely or be programmed to automatically react to a critical peak price signal (auto-
demand response). In return, the homeowner would receive larger payments.
Fatigue occurs with all types of demand-response mechanisms that require human
interference. Pricing mechanisms such as critical peak pricing should be applied to a maximum
number of days per year. Otherwise, customers may not respond, and the mechanisms may
lose their effectiveness. Likewise, the utility’s load controls of smart thermostats are limited to
several events per year. Otherwise, customers may override controls.
Behavioral science and “nudges” (subtle changes in the way choices are presented to
influence behavior) have become increasingly relevant in the implementation of DR
programs in the electricity sector. These can make DR programs more effective by helping
utilities engage with consumers by applying principles like social norms, loss aversion,
and the framing of incentives. By engaging consumer in a way that is both motivating
and convenient, utilities can drive behavior changes that not only reduce electricity
demand during peak periods but also promote a more sustainable energy future. With a
combination of behavioral insights and technology, DR programs can be more efficient,
scalable, and effective at increasing consumer participation.
Privacy and Cybersecurity Concerns
As smart meters multiply, there are concerns about collecting and sharing data, particularly
disaggregated consumption data, with utilities and providers of demand-response services.
Advanced metering infrastructure has the potential to enable demand response while
improving both efficiency and grid reliability. The new generation of fast demand response
requires time-disaggregated (nearly real time) data. Still, many are concerned about
sharing these data because they reveal consumer habits. Some software (for example,
Zigbee)8 can examine time-disaggregated data (load profile) and determine which equipment
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 35
was in use, when it was turned on and off, and other information that customers may want
to keep private. It is perceived as a window into people’s lives.
Cybersecurity is a second concern. Smart meter usage data are transmitted over great
distances using communication networks that serve the smart grid. Customers fear that
electronic equipment can be hacked, and that unauthorized parties or hackers could
intercept this information. A smart house will contain several web-based sensors, equipment
(for example, smart thermostats), and energy management systems to interface with the
external world. There is also a concern that hackers will target smart meters to attack the
grid. If hackers can take control of multiple smart meters, they can cause the load to vary
in a regular pattern (oscillation attack), potentially compromising the grid (Latief 2023).
Big Data Management
Smart meters provide vast amounts of data. Compared with a monthly kilowatt-hour bill,
the volume and variety of data may be overwhelming. For a typical residential customer,
electricity may be metered at 15-minute intervals, including energy and peak consumption.
If there is submetering for some special loads, such as EVs, the information doubles. If the
customer is under any time-differentiated tariff (price-based demand response), consumption
must be valued at each corresponding interval pricing. This meets the classical definition of
big data, encompassing information that grows at an increasing rate, the pace at which it is
created, and the variety of data points collected. Sorting out these data, extracting relevant
information about consumption habits, translating it into useful information, forecasting
customer behavior, proposing effective tariff structures, and testing the results are challenges
that utilities face when implementing sophisticated demand-response programs.
System Operators and Perception of Dependability
of Demand-Response Schemes
System operators are generally hesitant about the effectiveness and dependability of
demand-response schemes. Empirical evidence in some US power pools, particularly that of
the California Independent System Operator (CAISO), suggests that the potential of demand
response can be overestimated, which can harm their reputation with system operators.
Demand-response resources were not available when called by the system operator during
emergencies, which raises issues about their dependability in the eyes of the system
operator. Incentives and penalties should be carefully assessed in the design of demand-
response mechanisms; actual demand-response potential should be reassessed periodically.
These concerns are particularly strong in relation to typical demand-side bidding (DSB)
instruments when customers bid to reduce load (usually the day ahead). Customers may
often withdraw their offers in real time without penalties other than the prevailing electricity
cost; if removed in real time, those load reductions will not be available for the system
operators to control (or dispatch). Furthermore, because customer decisions to bid are
based on price, load reductions may take place too slowly to balance the system in the case
36 Increasing the Uptake of Demand Response: Enablers and Barriers
of a contingency. Therefore, it is argued that DSB mechanisms are not as fast and reliable as
a power plant, which may commit the day ahead and be penalized if it withdraws its offers in
real time. These drawbacks, which have affected perceptions about the dependability of DSB
in particular, could be addressed through automation and/or enforceable penalty provisions.
Load control may also be procured in capacity and ancillary service markets, allowing
demand response to behave like a power plant and therefore be considered a reliable
source by the system operator as a reliable resource.
Endnotes
1. Smart meters have been deployed rapidly around the world. Even though their main
aim is to reduce loss and operation costs, most have built-in features to control load
and convey granular ToU pricing, supporting more frequent billing.
2. In early applications, the control signal was transmitted via the grid using transmitters
with a range of up to hundreds of kilometers. The number of receivers per transmitter
is unlimited. Basic load control switches use one-way communication to control load,
with no verification.
3. Personal conversation with Edu Chaves, formerly of Whirlpool.
4. While this result is clear in illustrating the increased responsiveness of demand to
higher ratios, economic theory would suggest an optimum response is gained through
cost-reflective pricing that corresponds to the variation in the marginal cost of supply.
5. Most major original equipment manufacturers have committed to deliver V2G and
vehicle-to-everything compatible EVs in this decade, and charge point manufacturers
have made a similar commitment. For example, the Renault 5 electric is the first in
a long series of cars to come equipped with a bidirectional charger. The innovative
architecture integrates hardware such as natively reversible electrotechnical
components and electrical-current management software that provides ongoing
access to the V2G service while preserving battery capacity. A smart platform, Mobilize
PowerBox, communicates with the car and the cloud to determine whether it should
recharge the battery or send power back to the grid depending on battery charging
needs, domestic needs, and incentives from the energy market and power grid. On
the other side of the Atlantic, Ford F-150® Lightning® truck provides backup power to
homes (vehicle-to-home) using an 80-ampere charging station. With extended-range
batteries, the truck should be able to power a single home for three days.
6. This does not necessarily mean a higher ratio is always desirable; the objective should
be cost-reflective pricing to encourage rational consumption decisions.
7. Some electricity providers offer variable-rate plans at the end of a fixed-rate contract,
fully exposing customers to wholesale market prices.
8. A standard-based wireless technology developed to provide a low-cost, low-power,
wireless machine-to-machine and Internet-of-Things network. Utility companies
can use Zigbee on their smart meters to monitor, control, and automate delivery
of energy.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 37
FOUR
INTEGRATING DEMAND
RESPONSE INTO
POWER SYSTEMS
In theory, all market structures, including vertically integrated monopolies, can provide a
route for demand response to participate in the electricity sector. However, the prevailing
structures will influence the cost efficiency, scope for innovation, and available options. As
markets evolve, there are opportunities to integrate demand-response mechanisms into
the regulatory framework and the market design for energy provision, capacity, and
ancillary services. While functional energy markets exist in many countries, demand is
sometimes disregarded due to regulatory constraints, such as specific technical criteria
designed around traditional forms of power generation.
This chapter introduces the contracting framework for the providers of price- and quantity-
based demand-response mechanisms within the many power sector structures. It shows
ways for a demand-response provider to participate in a liquid market (such as a power pool)
or through bilateral contracting with a monopoly utility. The various demand-response
instruments are introduced and illustrated alongside sample business models. These models
may involve “revenue stacking,” which includes service provision and revenue generation
from multiple applications. For example, vehicle-to-grid (V2G) and virtual power plants may
offer both wholesale and ancillary services and trade in multiple markets.
Contracting Framework: Price-Based
Demand Response
An indirect price-based demand response elicits a reduction or shift in demand from
consumers in response to an electricity retailer’s time-differentiated tariffs (price-based
demand response). The retailer may set appropriate rates considering the time-dependent
variable cost of wholesale electricity generation and purchases, as well as network costs,
which are driven by peak demand. (See Box 4.1 for the example of the Octopus tailored tariff
provider.) Large consumers able to procure directly from any wholesale energy market may
face disaggregated energy and network tariffs, each including time-differentiated elements.
Distribution system operators (DSOs) and transmission system operators (TSOs) bear
network costs at the distribution level and transmission level, respectively. How these costs
are converted to network charges and levied on consumers relies heavily on the market/
system. Where the DSO and the TSO are independent entities, they may levy network
charges directly on system users (including generators, retailers, and demand-response
aggregators). Alternatively, in a “gross pool” market (that is, one where a single market
operator clears all generation and demand bids and offers with centralized dispatch),
transmission network charges are often bundled with the wholesale energy price, possibly
by location. In contrast, distribution network charges may be bundled with retail costs and
may or may not be separately itemized on consumers’ bills. Power systems that have yet to
develop active markets may bundle all costs within one vertically integrated utility or a
single buyer (also when purchasing from independent power producers).
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 39
Even when entities are fully bundled, tariffs will be developed based on cost-of-service
assessments and seek to be cost-reflective—meaning pricing principles seek to reflect an efficient
market outcome. This means the principles undergirding a price-based demand-response
instrument (that is, to reflect energy and/or network costs) can be developed in all cases.
Figure 4.1 illustrates the contracting framework for the provision of a price-based demand
response. The arrows run from relevant entities (on the right), which set the applicable
tariffs, to consumers (on the left), who respond.
FIGURE 4.1
Contracting Framework for Price-Based Demand-Response Providers
*Arrows for charts indicate direction of offer Transmission
and distribution
charges
charged to
retailers can be TSO / DSO
Consumers respond
to time-differentiated set on time
prices by reducing differentiated
consumption–these price basis
Retailer (may be
Customers rates may reflect
a bundled utility)
time-differential costs
corresponding to Wholesale
both energy and energy markets
network aspects will time Wholesale
differentiate energy
based on counterparty
settlement
periods
Large Some large
consumers consumers
(direct trading) will trade
directly Network charges and wholesale energy
costs can be bundled by an integrated 4
utility, single buyer (of generation), or
market operator
Source: Author’s analysis.
Note: DSO = distribution system operator; TSO = transmission system operator.
Contracting Framework: Demand-Side Offers
for Wholesale Energy Provision
Demand-response aggregators and large consumers may be regarded as equal to generators
in wholesale energy markets (or through bilateral contracting and within dispatch rules where
no such market exists). The United States provides an example of demand-response
involvement in wholesale energy markets where demand-side bidding (DSB) rules enable
customers (typically large consumers in the day-ahead market) to participate. DSB can be
an efficient demand-response mechanism, especially if automated, by helping to clear the
market at a lower price point as a cheaper demand-response substitutes for the marginal
generator (that is, the most expensive generation unit required to meet demand). Integration
40 Integrating Demand Response into Power Systems
BOX 4.1
BUSINESS MODEL: OCTOPUS ENERGY TAILORED
TARIFF PROVIDER
Retailers use tailored time-of-use (ToU) tariffs to design a range of packages for
their customers. They create an incentive to shift the load to meet the system
owner’s needs and are based on customer requirements.
Emerging companies perform various roles in the price-based demand-response
space. Octopus Energy Group1 is one retailer that offers a range of innovative,
tailored tariffs, including dynamic pricing, leveraging the national introduction of
smart meters. The design is tailored to mass-market residential customers.
Octopus started developing a half-hourly ToU tariff that is tied to wholesale prices and
updated daily (Agile Octopus); a bespoke tariff for indoor vertical farms; a tariff that
falls as wind speeds rise (Fan Club); and has recently introduced the United Kingdom’s
first vehicle-to-grid tariff (Octopus Power Pack). Recently, Octopus has been offering
savings to electric vehicle owners during “plunge pricing” events, when there is excess
renewable energy on the grid and wholesale market prices fall. During these periods,
discounts of 15–45 percent per kilowatt hour are available on the tariffs for charging
electric vehicles, and vehicle owners might even receive payments for charging if
wholesale market prices become negative at times. Customers are typically notified of
a “plunge pricing” event 24 hours in advance via an app.
Sources: Jackson 2018; Octopus Energy 2023.
in the market can be complex, however, and weak penalties for failure to deliver can result in
perceptions that a provider is not dependable.
In this form, DSB requires a market that is designed around a power pool with central
dispatch, whereby a system operator can select and dispatch the least-cost mix of generation
and demand response required to satisfy forecasted demand. Where no such pool with
central dispatch exists, the demand response must contract through other market players,
typically, retailers, to access the wholesale markets.
The specific market or power system structure will define the trade counterparty for a price-
based demand response. In a gross pool framework, this would be the market operator.
In other systems, it could be the bilateral trade counterparty, relevant power exchange, or a
single buyer/integrated utility. A generalized framework is illustrated in Figure 4.2.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 41
FIGURE 4.2
Contracting Framework for Demand-Side Energy Offers
Some markets do not enable
direct demand-response bids Balance
requiring bilateral contract Responsible Party
with BRP (e.g. Retailer)
Residential and
commercial demand Wholesale
Customers response will Aggregator energy
typically require counterparty*
aggregation
Large Large consumers may bid on a
consumers demand response directly within
(direct trading) a pool or enter into a bilateral
contract with a single buyer
* Counterparty will depend the on market/system structure.
Source: Author’s analysis.
Note: BRP = balance responsible party (an entity responsible to the system operator for imbalances);
DR = demand response.
Contracting Framework: Demand-Side Offers
for Capacity
Capacity markets seek to instill confidence in system operators and enable them to secure
adequate long-term capacity. There are concerns that so-called energy-only market
designs—whereby the fixed costs associated with generation facilities are recovered
through general volumetric wholesale electricity prices—may not provide sufficient
certainty to attract the needed investment in new facilities. Capacity markets seek to
address this concern by providing a reliable revenue stream so that generators can
recover fixed costs. Such markets either do not exist or are new in many countries, but
have become more prevalent, with concerns mounting about insufficient investment in
new capacity, especially flexible capacity, to support the energy transition. Demand
response in capacity markets offers good opportunities for customers to reduce their
electricity costs (through payments) in exchange for a few demand-response events.
It benefits the system by providing adequate capacity at a lower cost. Demand-response
aggregators, including virtual power plants, can manage participation as shown in
Figure 4.3 and discussed in Box 4.2. The financial benefit of the capacity payment
(e.g., $150/MW per day) and substantial penalties for failure to pay explain why the
participation of demand response in the capacity market can be robust (PJM 2024).
If demand-response resources are reliable, they are comparable with generation
resources. (See Box 2.6 for a discussion on the California ISO case.)
42 Integrating Demand Response into Power Systems
FIGURE 4.3
Contracting Framework for Demand-Side Capacity Market Offers
Residential and
Demand-
commercial demand Capacity market
Customer response
response will typically counterparty*
aggregator
require aggregation
Large Large consumers
consumers may bid on a demand
(direct trading) response directly
* As for the wholesale energy market, different market structures may define a different entity as the
capacity market counterparty.
Source: Author’s analysis.
BOX 4.2
BUSINESS MODEL EXAMPLE: VIRTUAL POWER PLANTS
In this business model, demand-response capabilities are combined with other behind-
the-meter resources to emulate the behavior of conventional generators in the grid.
Virtual power plants can offer various products to the grid, including trading in the spot
(wholesale) market, network support to alleviate congestion, ancillary services, peak
management, and capacity reliability. The existence of new distributed resources
(e.g., photovoltaics and storage), combined with a portfolio of intermittent and
dispatchable generators integrated with demand-response programs, enables
aggregators to offer a broader range of sophisticated products and services in the
market, operating like a virtual power plant. For example, slight, infrequent adjustments
to the temperature settings of a smart thermostat (two degrees) can provide hundreds
or even thousands of megawatts (MW) of peak demand reduction if aggregated across
enough participants within a power market (Hledik, Viswanathan, and Peters 2023).
Virtual power plants aggregate small, distributed systems into single, controllable
resources. Sunrun’s virtual power plant in New England, United States, which
(continues)
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 43
BOX 4.2 (Continued)
encompasses thousands of homes with solar and battery systems, provided
1.8 gigawatt hours of energy over three months of the summer of 2022. AGL
Energy’s aggregation of battery energy storage systems in 1,000 residential and
business properties in Adelaide, Australia, provided 5 MW of capacity between 2016
and 2019. The increased capacity could reduce customers’ bills, lower the peak
demand on the network, provide a wholesale market arbitrage, and supply frequency
control ancillary services and voltage support.
The nature of the services and products can be tailored to meet the sophistication
of individual power markets (Burger and others 2016). The figure below illustrates
how a virtual power plant and the various behind-the-meter assets can be
combined to provide grid services.
Behind-the-Meter Asset Classes in a Virtual Power Plant
Storage
Power markets
Renewables
Virtual
Power Plant
Grid Stability
Industrial and
commercial consumers
Asset
Optimization
Distributed
Generation
Other resources with
flexible capacity
Source: Adapted from NEXT (2023).
(continues)
44 Integrating Demand Response into Power Systems
BOX 4.2 (Continued)
The Brattle Group conducted a study to quantify the benefits of virtual power plants.
It concluded that virtual power plants can be significantly beneficial to power systems
(Brattle Group and SEPA 2019). A simulation was conducted to compare the cost and
reliability of generating 400 MW of power through virtual power plants containing
smart thermostats, electric vehicle chargers, smart water heaters, and behind-the-
meter batteries. The study explored whether virtual power plants can provide
resource adequacy with the same level of reliability as gas plants and utility-scale
batteries in a power system with 50 percent renewables. The study suggests that the
net cost for a utility to provide resource adequacy from a virtual power plant is about
40–60 percent less than from natural gas peaking plants (peakers) and utility-scale
batteries. Deploying 60 gigawatts of virtual power plants could meet future resource
adequacy needs of the United States for $15–$35 billion less than the cost of
alternative options over the next decade.
Contracting Framework: Demand-Side Offers
for Ancillary Services
Ancillary services include, but are not limited to, reserves that are able to respond to any outage
event and arrest the associated frequency deviations; ongoing frequency regulation to oversee
fluctuations in demand and variable renewable energy input; voltage support services to
administer localized deviations; and network constraint services to manage congestion.2 Large
consumers that can reduce their demand quickly—sometimes much faster than any generator
can ramp up—should be (and often already are) paid to be ready to do so (an availability-
payment equivalent). Provision has little impact on comfort or convenience because events are
brief (for example, 30 minutes or less). They are paid in return for being prepared to quickly
modify the supply-demand balance of a system. (See appendix A for the main ancillary services.)
More European countries are participating in demand-response ancillary services through
various asset classes, such as home batteries and electric boilers (Belgium), electric heating
and boilers (France), and remote control of appliances and electric vehicles (Norway). In some
cases, households can participate in an explicit demand response for ancillary services based
on the products the TSO defines through aggregation (Slovenia). In other cases, quantity-based
demand-response offers are limited to consumers with the technology to record consumption
in half-hourly increments.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 45
While both TSOs and DSOs can manage ancillary services, for DSOs, they are typically referred
to as “system services” instead. System frequency regulation and reserve procurement and
dispatch are generally a TSO’s sole responsibilities, since those are systemwide attributes.
At the same time, both TSOs and DSOs may contract providers for voltage support and
congestion support for their respective networks. A separate market operator can, in some
cases, undertake the procurement of some ancillary services at the transmission level to
meet stipulated volume requirements. Offers can be co-optimized with energy bids and
offered in a pool market to meet the least-cost dispatch.
This contracting framework is summarized in Figure 4.4.
FIGURE 4.4
Contracting Framework for Demand-Side Offers to Ancillary Service Markets
Direct procurement most common for
interruptible contracts for large
consumers or bundled utility load
control
DSO system
services
The demand response
aggregator may install
Demand-response
and invest in demand
Customer aggregator/
response at consumers’
service provider
premises or subcontract
a third party to do this
TSO ancillary
services
Depending on the market
structure the DSO and/or
TSO may form part of a
bundled “utility” entity
Source: Author’s analysis.
Note: DR = demand response; DSO = distribution system operator; TSO = transmission system operator.
Endnotes
1. Octopus is a UK-based company selling electricity services to 5.4 million customers
globally through its retail arm. It has also licensed its advanced data and machine
learning platform, Kraken, to support millions of customers worldwide.
2. The nomenclature and product definitions associated with these services are market
specific but will typically cover these functions. Reserves are typically subcategorized
into those that are able to respond quickly (often referred to as “spinning reserves”
because, as traditional plants, they are synchronized with the grid) and those with a
slower response time, which are called on to free up the faster reserves.
46 Integrating Demand Response into Power Systems
FIVE
LESSONS LEARNED FOR
DEMAND RESPONSE
IN DEVELOPING
COUNTRIES
To assess the status of demand-response implementation in developing countries, this
chapter looks at programs in five mid-size and large emerging economies—Brazil, People's
Republic of China, India, South Africa, and Viet Nam. It also examines the unique
challenges faced by small island states or territories, using as a benchmark the United
States, one of the most highly developed demand-response markets in the world. In the
United States, several states in the mid-Atlantic region and parts of the Midwest are served
by PJM Power Pool, one of the most aggressive demand-response markets in the country.
When comparing the five countries (Brazil, People's Republic of China, India, South Africa,
and Viet Nam), the study considered the existence and nature of ToU (time-of-use) rates
combined with load control and demand response.
Demand Response in India
Historically, demand management received insufficient attention from Indian policy makers as a
resource for planning short- and long-term electricity systems. System operators and distribution
utilities have responded to changes in customer demand by adjusting generation dispatch,
planning for adequate supply reserves, or managing the load through supply interruptions.
India is, in other words, in the early stages of designing and implementing price- and
quantity-based demand-response initiatives. The country now recognizes the role played
by demand-response mechanisms in meeting power system needs at the national and
subnational levels—challenging given the rapid growth in demand for reliable, affordable,
24/7 power supply. Demand-response mechanisms are seen as an efficient means of
limiting supply disruptions and contributing to the shift toward cleaner energy sources. Air
conditioning is the leading driver of peak power demand in India as per capita income and
urbanization cause the use of air conditioning to surge.
In the 2021 grid code, India’s Central Electricity Regulatory Commission recognized the
value of demand response in lowering electricity supply costs and supporting reliable grid
operations with high levels of variable renewable energy (VRE) (Sasidharan and others
2021). But implementation rests with the distribution companies, which because of
technical, economic, and financial barriers have yet to undertake large-scale programs.
Over the past several years most states in India have implemented various ToU tariff designs,
focusing on industrial and commercial customers connected to high- and medium-voltage.
These tariffs are typically calculated via a surcharge and a rebate applied only to the energy
component of the conventional tariff (Table 5.1). Seasonally adjusted implicit peak-to-off-
peak ratios vary between 1.3 and 2.0 (with an average of 1.5). These are relatively modest
ratios: in Brazil and South Africa ratios may reach 6.0; the ratio in Beijing (People's Republic
of China) is 4.3 (Ferriera and others 2013). The impact on the load profile is more significant
in states with larger off-peak ratios. Still, unmonitored industrial facilities will hamper efforts
to assess the effectiveness of the ToU tariff programs (Chunekar, Kelkar, and Dixit 2014; PwC
2010).
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 49
TABLE 5.1
Time-Differentiated Tariffs in India, by State
STATE TARIFF
Bihar ToD tariff applies to all high-tension consumers. Surcharge of 20 percent or rebate of 15 percent applicable to
energy charges during peak and off-peak periods.
Chhattisgarh ToD tariff applies to select high-tension consumers. Surcharge of 20 percent or rebate of 15 percent applicable
to energy charges during peak and off-peak periods.
Delhi ToD tariff applies to all consumers (except households) whose sanctioned load or maximum demand indicator
(whichever is greater) is 10kW/11kVA or above. ToD optional for household consumers. 20 percent surcharge or
rebate is applicable to energy charges.
Gujurat ToD tariff applies to some high-tension consumers. Surcharge of 10–20 percent applied to energy charges
during peak hours. Nighttime concession available to consumers opting to use electricity only at night.
Haryana Optional ToD tariff applies to high-tension industrial customers from October through March. 19 percent
surcharge and 15 percent rebate applicable to energy charges.
Jharkhand ToD tariff applies to high-tension consumers. 20 percent surcharge and 15 percent rebate applicable to energy
charges.
Punjab Additional charge of Rs. 2/kVAh during peak hours and rebate of Rs. 1.25/kVAh during off-peak hours apply to
medium-sized and large industries, nonresidential, and bulk supply customers. Peak tariff applies only June
through September; off-peak tariff applies for rest of year.
Kerala ToD tariff applies to extra-high-tension, high-tension, and low-tension industrial consumers with connected load
of at least 20kW. Surcharge of 50 percent and rebate of 25 percent applicable to energy charges during peak
and off-peak hours.
Source: Adapted from CERC (2019).
Note: ToD = time of day.
Since 2012, distribution companies in India have initiated commercial and industrial load
control pilots. For instance, Tata Power, which controls several distribution companies, has
designed and implemented demand-response pilots in Mumbai and New Delhi. One of
India’s most innovative utilities, Tata is also rolling out India’s first smart meter–based pilot
program for peak demand and grid stress using automated demand-response manage-
ment. The pilot relies on real-time communication to share information on the load to the
utility and consumers, improving transparency. The utility also conducted a pilot of behav-
ioral demand response that targeted residential customers, demonstrating the potential
savings to the utility and customers. With peak demand four times greater than off-peak
demand, the targeting shaved off-peak demand.
Similarly, BSES Yamuna Power Limited, a joint venture of the government of Delhi and
Reliance Infrastructure Limited, carried out manual demand-response pilots between 2017
and 2019 to assess the benefits of lower peak load costs and flexible load management
(Bureau of Energy Efficiency 2023, accessed 2024). The first pilot saw 30 industrial and
50 Lessons Learned for Demand Response in Developing Countries
commercial customers participate with an avoided peak of 17 MW, while the second pilot
saw participating entities rise to 60, with 32.5 MW avoided, respectively. Requests were
conveyed and confirmation sought through instant messaging services on smartphones
to curtail noncritical loads in return for incentive payments.
These pilots confirmed that customers can lower their consumption during peak hours,
improving grid reliability (Poojary and others 2023). If load control projects like those
described above were deployed in the commercial and industrial sectors across India, the
country’s peak electricity demand could drop by around 7.5 GW, or 5 percent of total peak
demand (Poojary and others 2023). Appendix B contains more information on pilot
programs by distribution companies.
Many different sources can provide demand response. For example, direct load control
and interruptible pumping programs are possible because feeders supplying energy to
agricultural pumping are often segregated from other loads. Nighttime watering occurs
during demand-response events. Utility or demand-response providers could curtail
reliance on switches or timers or manual demand response by the farmers.
Similarly, as 70 percent of the growth in air conditioning is expected to come from the
residential sector alone, it is critical to design demand-response programs targeting this
segment. With respect to the uptake of air conditioning, there are three possible demand-
response scenarios: (1) switching air conditioners on and off, (2) adjusting thermostat
settings, and (3) load cycling. Wifi-enabled smart air conditioners can manage loads by
building in capability for demand response. Others could convert to demand response
with wifi-enabled plugs (Sasidharan and others 2021).
Despite the success of India’s pilot programs and the potential for scale-up of demand
response, barriers remain in policy, regulatory, technical, and economic areas. These
barriers affect the willingness and ability of distribution companies—the companies
responsible for demand-response implementation—to undertake large-scale programs.
A key policy barrier is the subsidization of electricity, which pushes sales below cost-
recovery rates. This can prevent the benefits of avoiding or shifting consumption
from accruing to financial benefit of the provider. The lack of cost-reflective tariffs is
compounded by the small differential between peak and off-peak tariff rates under
ToU pricing.
Regulatory barriers are also constraining. In the 2021 grid code, India’s Central Electricity
Regulatory Commission recognized the value of demand response in reducing electricity
supply costs and supporting reliable grid operations with high levels of VRE (Sasidharan
and others 2021). Nevertheless, regulations often fail to define the term and focus instead
on peak shaving in response to previous supply gaps rather than flexibility for integration
of renewables.
Cost-plus1 regulatory frameworks do not provide incentives to distribution companies to
reduce costs. Instead, they link distribution company revenues directly to the volume of
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 51
energy sold, disincentivizing the adoption of programs to reduce consumption. Taking
more of a revenue approach, supported by performance-based incentives that encourage
demand-response uptake, could ameliorate this conflicting incentive. The regulatory
framework for demand response could be treated on an equal basis from an economic
perspective by taking the following steps (Sasidharan and others 2021):
• Ensuring that distribution utilities are incentivized to consider non-wire solutions to
network congestion or constraints on a basis equivalent to network upgrades
• Allowing greater flexibility in the design of tariff options for consumers
• Ensuring that demand-response aggregators have access to ancillary service markets
• Integrating the cost of emissions in least-cost dispatch calculations.
KEY TAKEAWAYS
• India recognizes the need for demand response to cope with the expected
growth of its economy and power sector, including surging demand for air
conditioning.
• Government and some utilities have been developing load control pilots for
various customer segments. Customer reaction and pilots have been
encouraging.
• Projects offering price-based demand-response can demonstrate potential
and build capacity, but to yield full benefits, they must be supported by a
follow-up strategy, which requires changes in the regulatory framework.
• The incentives within the regulatory framework for controlling revenues and
tariffs will be instrumental in gaining buy-in from distribution utilities to
implement demand-response programs.
• Monitoring pilots and programs is essential to ensure that necessary findings
can be evaluated and adjustments made to improve results.
Demand Response in People's Republic of
China
As a large industrial nation with a forecast peak load of about 1 TW by 2020, China is a large
potential market for demand response. Roughly 5 percent of China’s peak electric load,
about 60 GW, is met by generators, demonstrating a solid case for demand-side
interventions that can lower loads in homes, businesses, factories, and government facilities
via price signals or market payments (Liu 2016). Harnessing the full potential of demand
response would require the creation of markets to unite customers and grid operators.
52 Lessons Learned for Demand Response in Developing Countries
Over the past three decades, People's Republic of China has developed a number of
demand-response programs. In the early 1990s several programs fostered energy
efficiency, reduced consumption, and higher productivity. Some mechanisms were
integrated into demand-side management policies, which have become more
sophisticated, with a national law on ToU tariffs for industrial customers and additional
large-scale demand-response programs.
Since the late 1990s, People's Republic of China has rolled out ToU prices to balance daily
electricity use. The initial objective was to encourage industrial users to shift to off-peak
periods. Most Chinese jurisdictions have introduced ToU tariffs for industrial consumers in
a pilot project or as a permanent solution (Tahir and others 2020), with peak and off-peak
rates that rose over time. Time-differentiated tariffs (price-based demand response) are
sometimes available to residential customers. People's Republic of China does not have a
national ToU policy applicable to all utilities and customer segments. Each city and
province defines the type of ToU rates and the ratio of peak to off-peak tariff.
Shanghai was the first large city in People's Republic of China to widely adopt ToU rates,
although such rates had already been applied to large industrial customers in other
jurisdictions. Beginning in May 2001, the cost of electricity in Shanghai became 50 percent
lower at night than during the day. Four hundred thousand households were eligible to opt
into time-differentiated tariffs. Estimates show that they could cut their off-peak
consumption by 25 percent, reducing their power bills by 12.7 percent. Because so few
consumers signed up for these programs, however, the impacts were limited.
Beijing began engaging in ToU primarily for load management in response to escalating
peak demand and a sinking load factor. In 1996, after determining which industrial users
consumed at least half of their electricity during peak periods, the Beijing electric utility
implemented ToU pricing and other demand incentives to shift consumption to off-peak
periods. The peak to off-peak ratio is now 4.3:1, enough to direct the load away from peak
periods for many.
In 2002, Jiangsu became the first Chinese province to issue demand-side management
regulations and activate a pilot demand-response project that combined ToU rate
structures, interruptible tariffs, voluntary load shifting, and deployment of storage
devices (particularly thermal storage) to facilitate load curtailment and better manage the
supply-demand mismatch. Large industrial customers were instructed to undertake
administratively rationed, uncompensated load reductions to reduce peak demand. In
2012, the Tianjin Economic-Technological Development Area implemented China’s first
automated demand-response project (Navigant 2013; Samad and others 2016). The project
comprised 33 commercial and public buildings and 31 steel, chemical, and automotive
industries as customers with 100 MW of combined load-shedding capacity. The central
government paid each participant two renminbi per kilowatt-hour of reduction, and the
local Shanghai government spent an additional two renminbi per kilowatt-hour, which
adds up to roughly 36 US cents per kilowatt-hour, or four times the local retail price
of electricity.
The main features of these pilot programs are summarized in Table 5.2.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 53
TABLE 5.2
Demand-Response Pilot Projects in China
SUZHOU, JIANGSU BEIJING FOSHAN, GUANGDONG TANGSHAN, HEBEI
Programs offered Interruptible load Interruptible load and Cooling storage pricing Interruptible load
programs (real-time peak load pricing pricing
and contract demand
response)
Load curtailment 1,000 800 450 400
target, MW (2013–15)
Targeted consumers Industrial and Industrial, commercial, Industrial and municipal Industrial facilities
municipal facilities and municipal facilities facilities
Types of projects Nearly 400 facilities 131 projects targeting 80 energy efficiency 35 energy efficiency
connected to a 45 enterprises for projects for industry projects for power
demand-side dynamic pricing and 30 projects for peak plants
management service demand shaving
platform for peak
management
Actual response 2,716 customers, a total 74 customers, 71 MW 129 customers, 176 MW Not available
in 2015 of 2,037 MW across
Jiangsu Province
Source: Stern 2015.
KEY TAKEAWAYS
• Policy makers in People's Republic of China acknowledge the importance of
interventions on the demand side to avoid load shedding or payment for
expensive thermal capacity.
• Proactive identification of energy-intensive consumers with significant
potential for time-shifting load can yield benefits, particularly when combined
with high ratios of peak to off-peak pricing (although care must be taken that
offers not discriminate unnecessarily).
• ToU rates for residential customers have been tried in some jurisdictions, but
as in other opt-in programs in the world, customer buy-in is modest.
• Industry represents a good entry point for implementing demand-response
programs, owing to greater incentives and capacity for engagement.
54 Lessons Learned for Demand Response in Developing Countries
Demand Response in South Africa
South Africa has one of the most comprehensive frameworks for demand response in an
emerging market. It combines ToU rates and load control techniques that are applicable
nationally. Over many years, the vertically integrated electricity utility, Eskom, has implemented
various implicit and explicit (price- and quantity-based) demand-response schemes.
Eskom offers various ToU tariff schemes to its large industrial consumers. Since the early 1990s,
about 80 percent of electricity sales have been time-differentiated (Eskom 2020). Figure 5.1.
presents the various ToU tariffs available in 2021 and the periods they are based on.
FIGURE 5.1
Time-of-Use Tariffs Available in South Africa
Nightsave Urban Large, WEPS, Megaflex, Miniflex, Megaflex Gen, Ruraflex Gen and Ruraflex
Nightsave Urban Small and
Nightsave Rural Low demand season High demand season
23 24 1 23 24 1 23 24 1
22 2 22 2 22 2
21 Weekdays 3 21 Weekdays 3 21 Weekdays 3
20 4 20 4 20 4
Saturday Saturday
19 5 19 5 19 5
Saturday
18 and 6 18 Sunday 6 18 Sunday 6
Sunday
17 7 17 7 17 7
16 8 16 8 16 8
15 9 15 9 15 9
14 10 14 10 14 10
13 12 11 13 12 11 13 12 11
Peak Standard Off-peak
Source: Eskom 2020.
Note: Captions in Figure use Eskom names for ToU programs.
The ToU tariff system constantly evolves to reflect changes in load patterns and system
characteristics. Demand patterns change over time. With the deployment of solar PV, net demand
became more volatile. As a result, the underlying system costs changed, prompting the operator
to propose an overall reform of the ToU tariffs. The peak periods, and the ratios between
peak and off-peak prices, were adjusted to reflect the relative difference in system costs.
In addition to standard ToU rates, South Africa has introduced CPP as a pilot, as shown in the
tariff schedule (Table 5.3). The CPP system allocates 17 days a year that may be designated as
“critical” and have CPP applied, offering a strong price signal. On such critical days, the energy
cost may jump from 3.2 US cents/kWh to 18.2 cents/kWh, a significant incentive for customers
to adjust their load profile.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 55
TABLE 5.3
Innovative Time-Differentiated Designs in South Africa
NON-CRITICAL PEAK DAYS CRITICAL DEMAND CHARGE
(348) (US CENTS/kWh) DAYS (17) (US CENTS/kVA-MONTH
Nightsave urban small High season 3.2 18.2 8.1
critical peak day tariff (June–August)
Low season 2.5 17.5 1.0
(September–May)
Nightsave urban small High season Peak Standard Off-peak
critical peak day ratios (June–August) 1.3 7.3 7.8
Low season Peak Standard Off-peak
(September–May) 1.0 1.0 1.0
Maximum peak-off-peak High season 5.7
ratio
Low season 1.0
Source: Eskom 2020.
Note: Captions in Figure use Eskom names for ToU programs.
South Africa has for many years made extensive use of automated load control and
considers it successful. The target appliances for load control are domestic water heaters
that can be switched on and off remotely during maximum demand periods. Around
400,000 homes are enrolled, representing 200 MW of capacity. Utilities use specially designed
algorithms to control the hot water load to achieve optimum control with minimal interference
and discomfort to the customer. In addition to water heaters, these control mechanisms may
be applied to nonessential intermittent loads such as pool pumps.
South Africa is experimenting with control systems featuring two-way communication.
The greatest challenge is obtaining demand-response technology that is cost-effective with
respect to capital and operating costs. A helpful technology should provide beneficial information
such as tampering notifications, turning water heaters on and off without verification, the
hot water heater temperature, and the energy consumption of the hot water heater.
In 2011, Eskom partnered with demand-response provider Comverge to run a pilot.
Comverge managed the first open demand response in South Africa, which involved 500
MW of commercial and industrial loads, which could respond if requested. The program
was called the Eskom Demand Response Aggregation Pilot Program (Smart Energy
International 2012). Comverge procured and managed nearly 300 MW of demand response
from municipalities with so-called ripple control technologies. (Ripple refers to residual
periodic variation.) The pilot program with Eskom delivered almost 15 GWh of load
reduction over seven months, covering 550 event hours.2
Other mechanisms designed by Eskom for large customers are described in appendix B.
Despite considerable efforts on demand response, the country faces recurrent power
shortages. The rapidly advancing obsolescence of existing coal plants, poor planning for
56 Lessons Learned for Demand Response in Developing Countries
new capacity, and failure to achieve commercial operation with newly built plants are the
main drivers of the current power crisis, so investments in generation have lagged market
requirements.
South Africa uses rotational load shedding, under which the number of hours and frequency
of interruptions increase depending on the criticality of the system. South Africa has a
well-structured load-shedding program; in 2022 load shedding over nearly 3,800 hours of the
year (more than 40 percent) dropped over 20 percent of peak load.3 One interesting aspect
of the scheme is that large industrial customers participating in demand-response programs
are exempt from the early stages of curtailment.
KEY TAKEAWAYS
• South Africa has a long-established tradition of demand side interventions,
including both price- and quantity-based mechanisms.
• South Africa has implemented traditional ToU pricing, refining it to retain cost
reflectivity as the generation mix changes, particularly with greater solar PV
penetration. South Africa has also effected critical peak pricing in its tariffs.
• On the quantity-based front, control of domestic hot water systems has been
operational for many years, offering a strong entry point for automated load
control without requiring sophisticated new technology.
• Despite considerable efforts on demand response, the country faces recurrent
power shortages due to the rapid obsolescence of its existing coal plants.
• Adopting a mix of demand-response mechanisms tailored to the country’s
challenges and consumer profile can strengthen engagement and effectiveness.
Demand Response in Brazil
Brazil implemented ToU tariffs as early as the mid-1980s. The model was based on the
French tariff structure (Ferreira and others 2013). Initially, ToU electricity tariffs were
differentiated according to peak, off-peak, dry, and wet periods; only large and medium-
sized customers participated on a mandatory basis. Over the past few years, smaller
customers have been allowed to participate, but as with other opt-in schemes, uptake has
been modest, with only 65,000 customers as of September 2021.
Large customers are typically subject to two-part tariffs for energy and demand. Time-of-
day differentiation applies to both energy and demand for large customers (blue tariffs)
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 57
In contrast, for medium-sized customers, the time-of-day tariffs apply only to the energy
component (green tariffs).
The peak to off-peak ratio varies according to the customer group and utility. Figure 5.2
compares this ratio for Enel (the largest utility in the state of São Paulo) and CEMIG
(the utility for Minas Gerais). CEMIG leverages ToU rates to reduce or shift peak load
consumption (CEMIG 2021).
FIGURE 5.2
Peak and Off-Peak Comparison of Large, High-Voltage Clients
Electropaulo/ENEL
(State of Sao Paulo) CEMIG (State of Minas Garais)
5.00
4.50
Peak-off and Peak Ratio Comparison
4.00
3.50
3.00
2.50
2.00
1.50
Demand (Blue Tariff)
1.00
Energy (Blue Tariff)
0.50 Demand (Green Tariff)
0.00 Energy (Green Tariff)
Source: Enel and CEMIG.
The ToU methodology achieved the initial objective of shifting loads to off-peak periods. It
is estimated that ToU tariffs at CEMIG resulted (and still result) in a peak load reduction of
500 MW for high-voltage customers and 700 MW for medium-voltage4 (about 10 percent in
total), for an estimated investment saving of $600 million.5
ToU tariffs in Brazil are no longer differentiated seasonally (dry and wet periods). In 2017,
Brazil revised its tariff methodology, introducing a tariff surcharge (called “flag tariff”) based
on the criticality of the power system, ranging from about $5.8/MWh to $19.2/MWh (Enel
2023). The surcharges were designed to help defray some thermal generation costs, such
as the fuel required to operate flexible thermal plants during critical supply conditions
driven by low rainfall and reservoir levels (ANEEL 2017a).
The flag tariff system was designed to ensure cost-recovery rather than elicit a demand
response. Still, it may have a modest impact in terms of energy and peak savings.
Considering a peak energy price of about R$392/MWh and a maximum surcharge of
R$98/MWh, the tariff increase would be 25 percent. Assuming a price elasticity of
58 Lessons Learned for Demand Response in Developing Countries
−10 percent, the expected total reduction in energy consumption and peak demand would
be about 2.5 percent.
ToU tariffs were (and still are) important in Brazil—but they are a static mechanism. Tariffs
are based on predetermined time intervals and rates, which do not vary to respond to the
criticality of the system. ToU was applicable when the load profile was predictable, but VRE
production and changes in the consumption profile now influence net load. For example,
air conditioning loads, which drive the summer peak, tend to occur between 14:00 and
15:00 but may be partially offset by solar production.6 Other TDR methodologies, such as
CPP and RTP, could be helpful but have not been considered, even though customers
operating in the free market may be exposed to hourly prices in the wholesale market,
enabling some response to prices, which track changes in load profile and VRE volatility
more accurately than traditional ToU tariffs (Burger and others 2016).
Brazil has no load control mechanisms of any kind and no plans to introduce them. In
2021, however, the country introduced a simplified form of DSB called redução voluntária
da demanda (RVD).7 At the time, the national grid faced a dual energy and capacity
constraint. RVD was designed expeditiously, building on lessons learned from two projects
piloted by the Agência Nacional de Energia Elétrica a few years earlier. Participation was
limited to large industrial companies (>1 MW) or aggregators that could make offers a
week or day ahead to shift load from peak to non-peak periods (typically for a four-hour
reduction). If accepted for dispatch, qualified bids would be remunerated based on
estimated load reduction (actual vs. deemed or baseline consumption).
The mechanism was available for three consecutive months when the risk of load shedding
during peak hours was high. The total size of the qualified bids ranged between 2 and 3 GW.
The number of bidders, qualified bids, and average price are presented in Table 5.4. RVD was
discontinued in late 2021, when the loss of load probability returned to normal levels.
TABLE 5.4
Redução Voluntária da Demanda (RVD) Bids and Prices
MONTH NUMBER OF BIDDERS QUALIFIED BIDS (MW) WEIGHTED AVERAGE PRICE ($/MWh)
November 2021 52 2,269 179
October 2021 50 3,600 256
September 2021 31 2,323 279
Source: ONS 2022.
RVD proved cost-effective. It saved 28.8 GWh for the months of September and October
2021. It enhanced operating reserves and reduced the risk of load shedding by displacing
the need to dispatch very expensive out-of-merit thermal generators at a cost of about
$460/MWh. The cost of the program ($280/MWh) included: (i) a direct payment of $5 million
for accepted bids (about $180/MWh), recovered via a system service charge; and
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 59
(ii) $3.1 million representing the energy saved by the accepted bidders, which was settled
in the wholesale market (about $100/MWh). The cost-benefit ratio of the program was 1.65
(or 2.65 if only direct payments to bidders are included). New forms of demand-response
participation in the capacity and energy markets are being studied.8
KEY TAKEAWAYS
• Brazil has been using ToU pricing for large and medium-sized customers for
decades with significant benefits.
• ToUs are increasingly unsuitable where VRE penetration is high and dynamic
mechanisms such as CPP or RTP may be more beneficial. None of those tariff
schemes have been considered.
• DSB for wholesale energy markets was tried in 2021, when the power system
was capacity constrained. Results were encouraging, with strong buy-in and
cost-effective response. DSB with demand response in the capacity market are
being studied as a cost-effective option.
• Brazil has not considered certain other quantity-based mechanisms such as
load control and interruptible contracts. These should be further explored.
• Adopting a mix of demand-response mechanisms tailored to the country’s
consumer profile could improve engagement and effectiveness.
Demand Response in Viet Nam
Despite its limited historical experience, Viet Nam has for at least a decade been
interested in the potential for demand response to support its power system. Initial pilots
and programs concentrated on the potential for industrial demand response to reduce the
mid-afternoon peak, alleviating system stress and improving cost efficiency. More recently,
concern has been expressed about a shift to an evening peak due to solar PV expansion
creating a “duck curve” profile to demand. This has led to interest in further demand-
response instruments, including CPP.
In its initial programs focused on industrial load, Viet Nam had a target of 90 MW of
demand response by 2020 and future targets to achieve a demand-response potential
of 300 MW by 2025 and 600 MW by 2030, corresponding to 30 percent of peak load.9
Currently, demand response cannot participate directly in the electricity market (GIZ
2021). However, Viet Nam has implemented the demand-response programs shown in
Table 5.5.
60 Lessons Learned for Demand Response in Developing Countries
TABLE 5.5
Implemented Demand-Response Programs in Viet Nam
PROGRAM YEAR IMPLEMENTED SCOPE PROGRAM RESULT
Pilot load adjustment 2015 Ho Chi Minh City • The registered demand-response capacity reached
program 5,847 kW.
• Four successful demand-response events were
completed.
Voluntary load 2018 Nationwide • Voluntary (that is, no monetary compensation)
adjustment program demand-response implementation agreements
were signed with 2,471 customers, totaling a
potential demand response of 963 MW.
• Ten successful demand-response events were
completed in 2019.
• Maximum capacity reduction was 514 MW.
Source: GIZ GmbH 2021.
Six demand-response programs were identified in a Ministry of Industry and Trade
Circular 23 report in November 2017, which prescribed the processes for implementing
load adjustment programs. The programs fall into three categories, as shown in Table 5.6.
But despite their demand-response targets and a framework to support these programs,
implementation has faced legal constraints that have made voluntary programs largely
ineffective.
TABLE 5.6
Defined Demand-Response Programs
CATEGORY PROGRAM DESCRIPTION
Quantity-based Curtailable load program • 24-hour notice period, response up to 3-hours
duration, compensation paid for usage reduction
compared to baseline
• Targeted industrial and commercial customers
Emergency demand-response program • 2-hour notice period, response up to 3-hours
duration, compensation paid for usage reduction
compared to baseline
• Targeted industrial and commercial customers
Price-based 2-component tariff • Separate energy and charges targeted at customers
Critical peak pricing who have already been on the ToU tariff
• Tariff with peak event prices notified on a case-by-
case basis, targeted at commercial and industrial
customers
Voluntary Voluntary program with commercial • 30-minute notice provided; 10–30 percent
incentives reduction required
Voluntary load adjustment program • Request for reduction with no incentives
Source: GIZ GmbH 2021.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 61
The payment for the quantity-based response programs is based on the reduction in
energy compared to the customer baselines, calculated as the average usage in the same
period in the five days before the demand-response event.
Three key legal constraints limited the implementation of demand response in Viet Nam are:
• The utility, EVN, does not classify demand-response incentive payments as eligible for
cost recovery. As a result, EVN is unable to offer adequate compensation to incentivize
participation in the Curtailable load program and emergency demand-response program.
• Changes to the structure of electricity tariffs require amendments to the prime
minister’s decision. The current tariff framework does not enable two-part tariffs that
separate energy and capacity charges. A ToU tariff has been included in the current
structure; the difference between peak and off-peak prices, however, has been
insufficient to promote demand response (CPCS Transcom Limited 2021). Until such
amendments are made, the use of alternative tariffs cannot be implemented.
• EVN faces restrictions on transferring assets to customers, which means, for example, that
EVN cannot fund demand-response investments and will have to pay customers over time.
These constraints mean that only voluntary programs can be implemented.
KEY TAKEAWAYS
• Voluntary mechanisms have limited reach given the lack of monetary incentives.
• Care needs to be taken to ensure that the legal and regulatory frameworks
governing demand-response procurement incentivize the utility to place its
provision, particularly cost recovery, on a level playing field with supply-side
resources.
• Demand response may be supported by the utility’s direct investments in
demand-response technology on the customer’s premises. Where the
regulatory framework does not permit this, alternative contracting structures
must be possible (for example, competitively sourced third-party provision).
Demand Response on Small Islands
Power systems on small islands face particular challenges around changeable energy supply
and demand. As renewable energy’s penetration grows, a lack of rotational inertia (usually
provided by hydro or thermal generators) in the system can rapidly become problematic for
ensuring stable supply. Small system size can hinder both the implementation of competitive
62 Lessons Learned for Demand Response in Developing Countries
markets and the capacity of stakeholders to invest in and test new technologies.
Nevertheless, much can be done, and the size of the markets and systems can be used to
advantage because of their design simplicity and ability to engage a range of customers.
This section includes case studies of demand-response initiatives.
Hawaii
TOU TARIFFS FOR RESIDENTIAL CONSUMERS
In 2022, solar power provided about 17 percent of Hawaii’s total electricity, primarily from
small-scale, customer-sited solar power generation under a net-metering scheme. The
rapid adoption of solar power is due to the high electricity costs and good solar resources.
Hawaii has some of the highest power prices in the United States.
A ToU pilot for residential customers was rolled out in February 2024. The pilot tested
customer response to opt-in and opt-out options. A total of 15,000 customers were included
in the pilot automatically, and 500 of those opted out. A much larger group of 150,000
customers were not incorporated into the pilot—and just 500 of those have opted in. Solar
customers were given two options within the pilot: to opt-out or to pay/receive the ToU rates
in both directions. The overwhelming majority of those who opted out are solar customers.
It was initially expected that after a year of the pilot, the program’s rates would become the
default, but this deadline is no longer anticipated. Before universal rollout, Hawaii might see
new discussions about time periods and price ratios. No CPP has been considered thus far.
Contrary to most ToUs in the continental United States, the lowest tariffs occur between 9 a.m.
and 5 p.m. (and not overnight), reflecting the availability of solar energy. This is an interesting
example of time-differentiated rates supporting the integration of renewable resources.
DIRECT LOAD CONTROL PROGRAMS
Hawaii has an extensive range of direct load control programs. Events can be dispatched
by local system operators or automatically triggered in case of underfrequency. Dispatch
events are at least one hour, and underfrequency event duration is typically a few minutes.
Ongoing programs include direct load control programs available only for customers on
Oahu, and a frequency control demand-response program available on Oahu and Maui
(Hawaiian Electric Company 2024a). The peak loads for Oahu and Maui in 2022 were
1,102 MW and 195.1 MW, respectively.
• Residential direct load control on water heaters (referenced by the Hawaii Electric
Company as 15 MW controllable from 34,000 customers across both islands, or
1.2 percent of Maui and Oahu 2022 peak load) and air conditioning (25 MW controllable
from 4,000 customers, 1.9 percent of Maui and Oahu 2022 peak load).
• Direct load control for large commercial and industrial customers (18.2 MW from
43 customers, 1.4 percent of Maui and Oahu peak load in 2022) and small businesses
(1 MW from 161 customers, 0.08 percent of Maui and Oahu peak load in 2022).
• Automatic or semiautomatic demand response for frequency control. Customers can opt out
of an event at any time. Participating customers are paid for energy and demand reductions
and receive $3,000 toward a meter upgrade (Hawaiian Electric Company 2024b).
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 63
Jeju Island
In 2012, Korea’s Jeju Island declared its aim of achieving net zero emissions by 2030.
Renewable energy penetration soared on the island, growing from 9.3 percent in 2015 to
18 percent in 2021. This rise led to excess renewable generation during periods of high
renewable resource availability. The excess renewable generation has required significant
curtailment to maintain system balance, even with the 400 MW interconnector to the
mainland. The figure below shows the evolution of curtailment events and volumes on
Jeju from 2015 to 2022.
FIGURE 5.3
Curtailment Events on Jeju Island
100 25,000
Curtailed amount MWh
80 20,000
No. of Curtailments
60 15,000
40 10,000
20 5,000
0 –
2015 2016 2017 2018 2019 2020 2021 2022
No. of curtailments Curtailed amount 1st half (est.)
Source: Kim, Han, and Moon 2022.
To reduce curtailment of power from renewable sources, the government introduced “Plus
Demand Response” in 2020, through which consumers are paid to charge their EVs during
periods of excess generation. This demand-response program is now operating only on
Jeju; it is intended to operate nationwide, however, as renewable energy generation on the
mainland creates a need for this response.
In the 2021 pilot project, 521 public chargers were registered as participating in Plus Demand
Response, totaling 15 MW of capacity. In 2022, 73.5 MW of capacity was registered. Users
could join the project through an app that notifies users of Plus Demand Response time
periods and awards points to consumers who utilize the chargers during those periods
(Korea Bizwire 2021). The points can be converted to the local currency (Invest Korea 2021)
while the EV charger load is aggregated by GridWiz, an aggregator and operator of
distributed energy resources for sale into electricity markets.
One study conducted in 2018 forecast the 2030 EV-charging peak load on Jeju Island to be
94,766 kW.
64 Lessons Learned for Demand Response in Developing Countries
Corsica, Guadeloupe, and La Réunion
Demand response has been applied to water heating and air conditioning loads in La
Réunion, Guadeloupe, and Corsica through the Millener demonstration project of French
national utility EDF. The project was conducted between 2011 and 2015 (EDF 2014; Smart
Grids 2020).
Through the program, EDF offered to install appliance-connected energy controls to enable
it to control water heaters and air conditioners remotely for purposes of demand response
and grid stabilization. The aim was to reduce end users’ consumption, boost renewable
energy penetration in the island systems, and provide balancing services.
The scheme is voluntary. A research paper found that the project installed 1,050 demand-
side management units for households across the three islands, alongside 500 home PV
and battery storage systems. Participants can access the energy controls and control their
consumption at any time (Santi 2013).
Another component of the Millener project was that EDF coordinated solar generation
resources and battery storage systems. The utility aimed to achieve energy savings of
500 MWh per year through the solar and battery systems installed through the
Millener project.
FIGURE 5.4
Eurelectric’s Millener Project: Two Setups
PV panels and electricity storage
configuration
An electricity storage system: PV
• Allows to shape the electricity produced by
Switch
the PV panels Water
heater
• Contributes in maintaining the balance of IT Monitoring
the network (frequency, power at period of Energy gateway & control
peak demand..)
• Enables the client to consume its own
Control center
electricity and have access to energy Meter
Heater AC unit
during a cut-off and grid black out
500 installations to be rolled out in 3 islands
PV
Energy gateway configuration
Switch
An energy gateway monitors the client’s Water
Monitoring heater
equipment (electric heaters, air conditioning IT & control
units, water heaters...) to: Energy gateway
• Reduce its energy demand
• Help to maintain the balance of the network
Control center Meter
(frequency, power at period of peak Heater AC unit
demand..)
1000 installations to be rolled out in 3 islands
11 – EDF june 2012
Source: Pons 2012.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 65
KEY TAKEAWAYS
• Ever-expanding renewable energy penetration can rapidly overwhelm small
island grids and result in substantial curtailment. Demand response can help
mitigate this problem.
• A dearth of large industry has led utilities to focus on controlling residential
water heaters and air conditioners as key technologies for load control.
Demand Response in the United States:
Best Practices from Advanced Frameworks
The United States has one of the most extensive and dynamic demand-response frameworks
in the world. Demand response occurs at utility (retail) and wholesale (power pool) market
levels. The programs complement each other but target different markets and move at
different speeds. The major driving factor behind the introduction of demand response was
the spread of air conditioning.
Vertically integrated utilities recognized that demand response should be part of their
integrated resource plans well before the Federal Energy Regulatory Commission mandated
that demand and supply resources be treated similarly in market design.10 Demand-response
programs have undergone major transformations over the past 20 years. The growth in
incentive-based demand-response resources has occurred mostly in organized wholesale
markets administered by system operators—either independent system operators (ISOs) or
regional transmission organizations (RTOs) in the US market structure. Since 2001, the
Federal Energy Regulatory Commission has required ISO/RTOs to file annual program
evaluations or describe their demand-response program enrollment and performance in
annual state-of-the-market reports.
The three development stages for demand-response mechanisms and corresponding
timeframes are shown in Figure 5.5. At stage 1, demand-response programs rely on manual
load controls and interruptible tariffs, offering capacity planning and emergency response.
Stage 2 has demand-response programs in the power markets, smarter technologies, and
the ability to respond in real time. At stage 3, technologies for ToU metering and load control
are deployed, and new business models and products are created. The disruptive factor was
distributed energy resources (DERs)—for example, behind-the-meter generation and storage
that extended the menu of resources that could “respond” on both the supply (distributed
generation) and demand sides (demand response per se).
66 Lessons Learned for Demand Response in Developing Countries
FIGURE 5.5
Evolution of Demand-Response Mechanisms in the United States
Pre-2000s 2000 2005 2010 2015 2020 2025 & beyond
• Largely manual control • Introduced to Wholesale • Provide Multiple Grid
• Interruptible Tariffs for Markets Services
Large Commercial &
• Increased Automation • Respond to Controls
Industrial
and/or Price Signals
• Increased Precision
• 1-way Direct Load Control for
Residential • Distribution &
• Eventually Ancillary Services
Transmission Relief
• Used for Capacity Planning • Behavioral /Voluntary Options
& Emergencies • Introduction of Storage
• Smarter Equipment
• Migration to Distributed
• 2-way Communications Energy Resources
• Some Near Real-Time Visibility
Stage 1 Stage 2 Stage 3
Source: FERC 2019.
Demand response in the United States is now entering stage 3, which entails integrating
demand response into DERs (solar PV, battery storage) to provide a variety of grid
services.
The potential for demand response is estimated at 29.2 GW at the retail level and 32.1 GW
at wholesale, totaling 61.3 GW (FERC 2023)—or about 6 percent of peak consumption in the
United States. Some practitioners contend that the United States needs more demand
response. The potential of 6 percent demand response has been sufficient to balance the
system so far, but more may be required with the installation of more intermittent
generation and the growth of the EV market.
Technology will allow more extensive customer participation in demand-response programs.
More than 64.7 million electronic meters were in operation in 2015, but only 9.8 million
customers were participating in a demand-response program (CPOWER 2020). Advanced
(smart) meters account for nearly 72 percent of all installed and operational meters in the
United States; 119 million advanced meters were active nationwide in 2022. Residential
customers accounted for about 88 percent of total advanced meter installations, and about
73 percent of residential electric meters were advanced meters (EIA 2023).
New ToU rates are being designed, for example, in Arizona, where a “reverse demand-
response program” has been developed to use excess solar energy in non-summer months
and avoid curtailment. Still, because the peak/off-peak price arbitrage cannot be predicted
(because of the intermittent nature of renewables), the Arizona Public Service program will
be specific to dispatchable nonessential loads. For example, EVs with smart charging and
smart appliances (for example, dishwashers, washing machines, dryers) could offtake free or
negatively priced energy when reverse demand response is activated (Cadmus Group 2018).
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 67
Although rollout of ToU rates was a complex endeavor, implementation of load control has
been easier, easing development of demand-response products and markets so customers
have incentives to curtail loads reducible on short notice. Florida has one of the world’s largest
residential load control systems, managed by Florida Power & Light. The system uses 800,000
remote switches to control 1,000 MW of electrical power (2,000 MW during an emergency).
Florida Power & Light’s load management programs have been successful enough to permit the
construction of new plants to be deferred. Baltimore Gas & Electric has another successful
utility-sponsored load management program, a peak-time rebate, to which 75 percent of
customers (1.1 million) subscribe (Faruqui, Sergici, and Warner 2017). Participating customers
have interval metering. The peak-time rebate program kicks in when high demand is
expected—typically when air conditioning loads soar during heat waves. Customers were
notified up to ten times yearly that the program would kick in. Customers are paid $1.25 for
every kWh saved (compared with a historic baseline) during critical periods (BGE 2019). The
program has sustained peak savings of more than 300 MW over the past few years.
About half the demand-response potential in the United States is harnessed via electricity
markets. Some major features of the demand-response programs in the PJM power pool
are summarized in Box 5.1.
BOX 5.1
PENNSYLVANIA-NEW JERSEY-MARYLAND
INTERCONNECTION: A FRONTRUNNER
IN DEMAND-RESPONSE DEPLOYMENT
Pennsylvania-New Jersey-Maryland Interconnection (PJM) has one of the most active
demand-response programs in the United States. It is the second-largest power pool
for demand-response resources, with about 10.6 GW, or 7.3 percent of peak demand
(for a total of 32.9 GW of demand response in the seven wholesale markets in the
United States, representing 6.5 percent of peak demand in those markets).
The PJM service area includes 13 states and the District of Columbia. The annual
cost of all demand-response programs is about $600 million per year. PJM fosters
the emergence of aggregators. Demand-response aggregators, or curtailment
service providers (CSPs), consolidate multiple loads and offer demand reductions
in the capacity market. Customers may receive a payment from a CSP to reduce
load when the wholesale price is high.
(continues)
68 Lessons Learned for Demand Response in Developing Countries
BOX 5.1 (Continued)
Demand response providers can compete in the PJM capacity market. The products
traded in this market ensure long-term grid reliability by securing the appropriate
power supply resources needed to meet predicted energy demand. Auctions are
conducted on a rolling basis. In the 2018 auction, PJM procured about 163.6 GW of
resources from June 1, 2021, to May 31, 2022. Demand response accounted for
11 GW of capacity in this auction. There has been an 18 percent growth in megawatts
cleared in the PJM market. EnerNOC, a global leader in demand-side flexibility
services, was awarded more than $180 million in capacity payments for demand-
response resources. The capacity market is by far the largest source of revenue.
Contributions from several energy and ancillary services products are much smaller.
The figure below illustrates the revenue earned by PJM’s demand-response products.
Demand-Response Revenues
Demand-Response Revenues
900
load management Dr markets ($ millions)
PJM estimated revenue for economic and
800
700
600
500
400
300
200
100
0 2002** 2003**2004**2005**2006**2007** 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022
Capacity* Emergency Energy Capacity Bonus Payment Economic Energy
Ancilary Services Economic Energy Incentives Price Responsive Demand Credits
Source: PJM 2023.
Note: Capacity bonus payments include payments for load management, economic (including
ancillary services), and price-responsive demand registrations.
* Capacity net revenue inclusive of capacity credits and charges.
** PJM assumes capacity value at $50 MW Day (PJM does not know the pre-RPM value of
capacity credits in the forward market; only a portion of capacity was purchased through the
daily capacity market at the time).
(continues)
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 69
BOX 5.1 (Continued)
PJM is particularly interested in expanding the ancillary service market. Despite
the relative insignificance of the ancillary service market, PJM acknowledges its
ability to respond rapidly when needed, which is important in an environment
of high intermittency from renewables. In 2015, 294 locations were certified to
provide an average of 16 MW annually. Storage will likely have an important role
to play. PJM would like to have easy access to the distributed generation behind-
the-meter to help balance the system. Grid codes, regulations, contractual
arrangements, and coordination between PJM and the distribution company will
need to be revisited to address the emergence of new types of demand response
(Cappers, MacDonald, and Goldman 2013).
One barrier to expanding demand response in the United States is that grid
operators do not understand distribution systems well enough to handle
distributed energy resources with the same ease that they run large, centralized
power plants (FERC 2018). Based on the experience of demand-response programs
in the wholesale markets, aggregators—such as CSPs—have been an effective
intermediary between the fragmented customer base and the wholesale market.
PJM advocates for a competitive CSP-based model with a track record of harnessing
demand-response capability in various retail markets. CSPs have worked with
customers to help them unleash their demand-response capabilities.
KEY TAKEAWAYS
• The United States has the largest, most sophisticated demand-response
market in the world.
• Demand response should be a regulatory requirement for resource planning,
with annual progress evaluations. These actions would raise the profile for
demand response vis-à-vis supply-side solutions.
• Technology that assists automation and expands relevant DERs can kickstart
new business models aggregating smaller consumers.
(continues)
70 Lessons Learned for Demand Response in Developing Countries
• Implementation of ToUs can be a complex undertaking, as is true of most
price-based demand-response instruments. These take time to establish and
to attract consumers, particularly residential consumers.
• The most promising application for demand response in a given case may
depend on the structure of a specific market or sector. Still, demand response
is well-suited to short-term contingency or stress events where the system
operator directly controls the load.
Comparative Analysis of Demand-Response
Programs
Demand response remains an untapped resource in emerging economies. The case
studies presented here highlight good practices for other developing countries and offer
a long-term vision for the scope of demand response in power systems. The analysis
highlights some key accomplishments in terms of price- and quantity-based demand
response, focusing on large developing countries (the so-called BICS group comprising
Brazil, India, People's Republic of China, and South Africa).
Participation in Price-Based Programs
ToU tariffs have been applied across all the markets reviewed—Brazil having the lengthiest
experience. But existing tariff models have not evolved, while uptake in many jurisdictions
has been sporadic and dependent on scheme design, particularly at the residential level, where
uptake has often been disappointing. Many ToU offers are “static,” meaning the peak and
off-peak periods and ratios are pre-set, irrespective of system conditions. Although the
prevalence of ToU tariffs is growing, dynamic pricing, which more closely reflects the
power system conditions, has not been mainstreamed for regulated customers in any of
the BICS countries.
CPP, VPP, and RTP approaches are uncommon, more sporadic, in any event, than static
ToU pricing. Of the assessed countries, only South Africa has experimented, with some
success, with CPP rates. RTP, meanwhile, has yet to be introduced in many developing
countries. Large consumers that trade directly in the wholesale market in Brazil are
exposed to real-time pricing and can react accordingly on an hourly basis. Still, there are
no RTP mechanisms for regulated customers.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 71
Participation in and Existence of Quantity-Based Programs
Quantity-based demand response has been implemented with notable success in some
markets. Interruptible contracts and load controls have been carried out to some degree
in all the countries reviewed in this chapter. It is an area of focus in India, where several
successful projects have been developed. There is a concerted government effort to
mainstream those initiatives. People's Republic of China has established load control
protocols with many customers to avoid load shedding. South Africa, meanwhile, has
implemented sophisticated load control systems and business models to manage
nonessential loads and mitigate capacity shortages. Several small island countries have also
begun exploring load control of energy-intensive devices at the residential level as a priority,
given the vulnerability of small power systems to high penetration of renewables and the
lack of large industrial consumers.
Demand-response participation in wholesale markets is less advanced. This is understandable,
first, because of the complications involved in integrating demand response into functioning
markets as a dependable resource, and, second, because of the absence of active wholesale
markets in many emerging countries. Brazil developed a DSB-like program in 2021, which
attracted 3 GW of demand response. Still, from the countries reviewed here, significant
involvement is apparent only in the more advanced markets such as the PJM Interconnection
in the United States, where demand-response participation is intense in both the energy
and capacity markets. Unlike in Brazil and South Africa, demand-response bids compete
with supply sources in the PJM system and interact to determine spot prices. The role of
demand-response aggregators as curtailment service providers is also well established.
Endnotes
1. “Cost plus” is a form of regulation whereby regulated utilities can recover all costs
incurred plus a regulated profit margin. This differs from a revenue cap approach
that sets the allowed revenues ahead of each regulatory period (with some
allowances for pass-through of uncontrollable costs).
2. Some of these loads may not have been dispatched. Some of the contracted load
providers for existing supplemental demand-response programs were moved to the
Comverge aggregator to test the system.
3. NRS-048-9 is an emergency load reduction program that the system operator and
distribution control rooms implement to prevent national, regional, or local blackouts
when system conditions are such that available power system capacity cannot meet
demand or when adequate reserves required to manage power system security
cannot be maintained without a reduction in load. Technically, this load reduction
program does not fit the definition of demand response used in this report because
of its mandatory (as opposed to voluntary) nature.
4. Personal conversation with Neusa Antunes, Escher Consultoria e Engenharia (2023).
5. CEMIG accounts for about 10 percent of the Brazilian market. If this Figure 5.2 could
be extrapolated to the entire country, total savings would be $6 billion.
72 Lessons Learned for Demand Response in Developing Countries
6. Increasing penetration of air conditioning in Brazil has changed the load profile. This
phenomenon was analyzed for the years between 2000 and 2010. In 2000, the peak load
was observed in early winter (June), moving to April in 2005 and to February in 2010. Air
conditioning loads have been responsible for this seasonal change (Poole 2011).
7. The RVD program was established on August 23, 2021. The main difference from DSB
was that RVD bids did not affect the spot price.
8. ANEEL Normative Resolution n° 1.030 of 2022, established conditions for the new
demand-response program. It authorizes the national system operator to conduct
studies to attract demand-reduction bidders in the capacity market in return for an
availability payment under a one-year contract.
9. Decision 175/QD-BCT dated 28 January 2019 on Approving the Implementation Plan
and Roadmap for the DR Program (GIZ GmbH 2021).
10. The Energy Policy Act (EPACT) of 2005 codified that a key objective of US national
energy policy was to eliminate unnecessary barriers to participation in demand
response by wholesale customers and load aggregators in the energy, capacity, and
ancillary service markets. EPACT directed the Federal Energy Regulatory Commission
(FERC) to develop a comprehensive national assessment of the size and scope of
demand-response resources.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 73
SIX
ROADMAP FOR
IMPLEMENTING
DEMAND RESPONSE
IN DEVELOPING
COUNTRIES
This chapter provides a roadmap for developing countries1 seeking to design demand-
response programs. The roadmap is necessarily high-level to ensure its applicability;
designers may require more details, which can be provided through technical assistance
as needed.
Summary of the Roadmap
The roadmap has three phases, as shown in Figure 6.1. The first phase is a diagnostic
that identifies and clarifies both the operating context and the problem that the demand
response is intended to address. The second assesses the options for demand-response
program mechanisms, which rest on understanding the enabling conditions (such as
the power system’s structure and enabling technologies) and undertaking a cost-benefit
analysis to select the optimal mechanism. The cost-benefit analysis could incorporate
multiple mechanisms, phasing in their introduction, different customer groups, and
a range of other scenarios and sensitivities, as appropriate, and identify each scenario’s
distributional effects. The third phase details the mechanism design and considers
policy and regulatory enhancements, while accompanying measures are fleshed
out and implemented. At this point, a time-based work plan allocates tasks, responsibilities,
and milestones leading to the implementation of demand-response mechanisms.
FIGURE 6.1
Summary of Roadmap for Demand-Response Implementation within a Country
1. Diagnostic 2. Mechanism 3. Implementation
assessment
• Characterize the • Assessing • Detailed
driving need for pre-conditions/ mechanism
demand enabling design
response conditions • Policy/legal/
• How the issue is • Selecting the regulatory
dispersed mechanism • Accompanying
across the • Cost-benefit measures
network and assessment • Work plan (e.g.,
over time demonstration
• How the power pilots, regulatory
system is set up change, scale-up)
• What is known
about demand
response
What is the
What What steps are
problem demand
mechanisms are needed for
response is to 8
appropriate? implementation?
address?
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 75
System Diagnostic
The system diagnostic is critical for establishing the context in which demand response
is being implemented. Figure 6.2 provides guidance questions for each diagnostic topic.
FIGURE 6.2
Guidance Questions for the Diagnostic of the Power System
1.1 The driving need 1.4 What we know
1.2 How the problem 1.3 How the power
for demand about demand
is dispersed system is set up
response response
• How variable is the • Is the problem • Is there private sector • Have we tried it before?
generation mix? dispersed spatially? participation? • Was/is it successful?
• How variable is Locally, regionally, • Is there a competitive Why/why not?
customer load? nationally? market (wholesale or • Is anything specified in
• When does it occur, ie, retail)? policy or legislation?
• Have there been
certain time of day, day • What can we learn from
changes in customer • Are there weaknesses
of the week, or season? international/peer
demand profiles? in the network
experience on certain
• Are there constraints • How urgent is the (transmission or
mechanisms?
in the network? demand-response distribution)?
• Are there any other
requirement when • Do customers have
• Is demand response barriers or enablers?
it occurs, ie, within smart meters?
incorporated into a
seconds/minutes/
least-cost plan? • Who would be the
hours/days? contract counterparty
• Are there
• Which customer groups for a demand
decarbonization or
are/could be affected? response provider?
other targets that
demand response can • What is the capacity of
support? different customer
groups to understand
the issue (and
potentially adopt a
demand-response
mechanism)?
The first two columns (1.1 and 1.2) of Figure 6.2 focus, respectively, on identifying and
characterizing the need that demand response is to address. At this stage, answers need
not be quantitative or highly detailed but should be referenced in later activities. With
regard to the first column, this report summarizes the driving needs for and roles of
demand response across segments of the power system. The questions in the columns
can be read in conjunction with the summary diagram, presented again in Figure 6.3.
The final issue—regarding incorporation of demand response in a least-cost plan—
will have identified demand response and incorporated its impacts into system
planning.2
Answers to the questions in column 1.2 of Figure 6.2 about the issue’s dispersal will provide
further guidance on targeting the demand-response mechanisms, ensuring that they engage
and benefit the right power system participants and wider beneficiaries in the proper way.
Depending on the power system, some of this information may be difficult to obtain; any
information related to these questions will be beneficial.
76 Roadmap for Implementing Demand Response in Developing Countries
FIGURE 6.3
Drivers of the Need for Demand Response
Energy Least-cost Maximizing consumer utility through least-cost
balance energy matching of supply and cost-reflective demand
provision Affordability
Capacity Capacity Ensuring adequate de-rated capacity available on Meeting system needs
provision adequacy system for managing system stress events within at lower cost
LOLP limits
Ancillary Reserve Arresting and restoring system frequency Reliability
services provision following loss of load
Reducing system
Frequency Managing continuous fluctuations in frequency outages
regulation caused by fluctuations in demand and in the
supply of renewable energy
Decarbonization
Voltage Managing voltage deviations to retain power
control quality Facilitating greater
penetration of
Constraint Managing network congestion and constraints
renewable energy
management that otherwise curtail generation in substitution
of upgrading or reinforcing the network directly
(“non-wire alternative”)
Source: Author’s analysis.
Note: “Non-wire alternatives” address congestion management without expansion of the grid. LOLP = loss
of load probability; RES = renewable energy.
The third column of Figure 6.2 assesses the power system’s capacity to support the
implementation of different demand-response mechanisms. Some mechanisms will not be
implementable within certain power system structures—for example, monopolistic generation
supply will not allow energy and capacity offers. Similarly, price-based demand-response
mechanisms require appropriate metering, including smart meters, for more refined
approaches that measure energy consumption over time. The power system review should
consider other weaknesses that may be detracting from demand-response mechanisms.
The fourth column reviews the power system’s experience with demand response. Some
countries may have tried demand-response mechanisms in one part of a country or for
one customer group. They could build on that experience (positive and negative) when
implementing further mechanisms. For those countries that have not tried demand
response, the international experiences described in chapter 5 will be particularly relevant.
Mechanism Assessment
The need for demand response is driven by the context—enabling conditions, a power
system’s structure and characteristics, and the country’s experience with demand
response. With context understood, developers should select the best mechanism(s)
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 77
laid out in this report. Mechanisms can be assessed for effectiveness through a cost-benefit
analysis that ensures positive net benefits for the power system and distributed benefits
to targeted participants without detracting from other contributors. Figure 6.4 provides
guidance on some demand-response mechanisms and their net and distributional impacts,
which are discussed in the following sections.
FIGURE 6.4
Approach to Mechanism Identification and Assessment
2.2 Cost-benefit
2.1 Identifying
assessment of the
potential mechanisms
mechanisms
• Identify and clarify the • If list is too long,
enabling conditions perform quick cost-
• Use the mechanisms benefit analysis to
matrix develop short list
• Start with a long list • Assess different
mechanisms, including
• More than one
combinations
mechanism may be
appropriate • Identify customer
Refer to the group(s)
lessons learned in • Consider market
structure and • Run scenarios
chapter 5
technology availability • Assess distributional/
equity impacts
Identifying Potential Mechanisms
The report earlier provides a matrix of the potential demand-response mechanisms and
the needs each is trying to address. As a guide to identifying appropriate mechanisms, the
matrix is provided again as Table 6.1.
The first question acknowledges the need to address more than one challenge, identifying the
potential mechanisms along the rows. Narrowing the selection will be easy for those three rows
where only one mechanism is suggested. For the others, further analysis will be appropriate.
Table 6.1 lists points to consider for each mechanism. Readers should also refer to the lessons
learned in chapter 5, summarized after each case study.
Not all mechanisms will be possible in every country. For example, broad, price-based
mechanisms grow more sophisticated as one moves across the matrix (see Figure 6.5),
requiring more complex pricing approaches in markets with, for example, dynamic pricing.
This would make sense only if there were a functioning wholesale market with market
prices. The price-based mechanisms also require metering of customer loads at different
times of the day. While such metering may be scarce or absent in some countries, it is
an easier hurdle to overcome than (re)designing an energy or capacity market. Energy
and capacity offers require a market that accepts offers. While such a market may be
on the horizon for some countries, it may not be feasible to implement one as part of a
78 Roadmap for Implementing Demand Response in Developing Countries
TABLE 6.1
Matrix of Demand-Response Challenges and Some Mechanisms to Meet Them
PRICE BASED DIRECT PAYMENT BASED
1. What is the challenge we need to address?
TOOLS TOU CPP VPP RTP PTR DR ENERGY DR CAPACITY AUTO LOAD MANUAL
OFFERS OFFERS CONTROL LOAD
CONTROL
Least-cost Energy Yes Yes Yes Yes Yes Yes
Provision
Capacity Adequacy Yes
Reserve Provision Yes Yes
Frequency Yes
Regulation
Voltage Control Yes
Constraint Yes Yes Yes Yes Yes Yes
Management
Source: Author’s analysis.
Note: ToU = time of use; CPP = critical peak pricing; VPP = variable peak pricing; RTP = real-time pricing;
PTR = peak time rebate; DR = demand response.
TABLE 6.2
Summary of Key Selection Guidance of Different Demand-Response Mechanisms
MECHANISM SELECTION GUIDANCE
Interruptible contract/load control Relatively easy to implement for large customers, to provide reserves and capacity adequacy
for peak periods
Requires certain technologies for metering and remote control
Has been successfully implemented for households in vertically integrated markets, but
aggregators can help provide this with relative ease once the system operator is happy with
reliability and regulations made to match
Static ToU tariffs Relatively easy to adopt for larger customers, with appropriate metering and accurate tariff
design
It is more challenging for smaller customers (especially households) to understand
May be less effective in the age of VRE and as more dynamic tariffs take effect
Mandatory uptake is challenging politically, while opt-in tariffs typically have low response
rates, suggesting market education is critical and automated demand-response resources
highly beneficial
Other price-based instruments Require more dynamic electricity pricing, applicable as markets become more competitive
(particularly in wholesale markets)
Can be marketed with a focus on more engaged consumers with the right technology and
suitable loads (EVs, air conditioning with smart controls, and so forth)
Need to be careful of distributional effects
Energy and capacity offers Most complex of the mechanisms, dependent on market design, and not well established
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 79
FIGURE 6.5
Matrix of Demand-Response Challenges and Some Mechanisms to Meet Them
2. Which mechanisms are appropriate to meet this?
PRICE-BASED DIRECT PAYMENT-BASED
1. What is the challenge we need to address?
Tools ToU CPP VPP RTP PTR DR DR Auto Manual
energy capacity load load
offers offers control control
Least cost
energy
provision
Capacity
adequacy
Reserve
provision
Frequency
regulation
Voltage
control
Constraint
management
demand-response program. At the same time, as markets evolve, mechanism redundancy
can intensify, notably for static ToU tariffs, as market penetration of VRE expands.
Not all mechanisms are available for every customer group; technology options within a
country and among customer groups can dictate mechanism potential. Smart meters with
time-based load measurements are often introduced, first, to large customers—for example,
industrial customers with the most significant load and revenue protection benefits per
customer. These smart meters are then introduced to progressively smaller customers, first
to large businesses then small and then eventually to households. More obviously, remote-
control mechanisms for water boilers/geysers require customers to have effective, cheap,
easily installed switches—for example, ripple control and smart plugs.
The choice does not need to be for a single mechanism. Indeed, options could include
more than one mechanism or combinations of mechanisms; within a single mechanism
there can be multiple applications.3 As noted earlier in the report:
• Small retail customers may have time-differentiated tariffs and still participate in load
control. This option will likely gain momentum for EVs, where charging patterns are a
crucial target for incentivizing demand shifting, while the batteries offer a potentially
valuable grid resource.
• Small customers may be subject to ToU rates and still participate, via aggregators, in
several demand-response programs at the wholesale level.
• Price-based instruments can also be applied to energy and demand, with customers
paying different demand charges ($/MW) depending on their metered peak
consumption during peak and off-peak periods.
80 Roadmap for Implementing Demand Response in Developing Countries
Cost-Benefit Analysis of the Mechanisms
A cost-benefit analysis (CBA) can reveal rate impacts across all customer categories. CBAs
help to determine both an equivalent supply curve of demand-response opportunities
(including combinations of mechanisms) and desired levels of DR deployment, considering
the cost of supply options.
The cost-benefit analysis4 is critical for understanding two aspects of the potential demand-
response mechanisms:
• The net benefit of the application, and
• The distributional impact: which sector participants benefit the most and which
benefit the least?
Costs include smart meters or remote controls, incentive payments for participating
customers, tariff design to determine ToU charges, regulatory changes, and customer
marketing. Benefits include lower peak generation costs, reduced ancillary services (for
example, spinning reserve costs), and the avoided cost of any lost load. Most of these costs
are discrete, one-off items incurred during program implementation. Some benefits—for
example, unused generation and spinning reserve—may require power system modeling.
Similarly, calculating incentive payments will require an understanding of the value obtained
by the utility (which gives an upper bound of what could be shared with customers). In
addition, the calculation discloses the opportunity cost of investing in demand response
vis-à-vis generation assets and reveals both customer willingness and price points that alter
their behavior, which provides the lower bound of the price. Values for each cost and benefit
should be quantified where possible. They should also be detailed as values paid or received
over time before being discounted to a rate that produces a present value. Modeling should
consider the participants’ perspectives—for example, the utility, participating and
nonparticipating customers, the implementing entity—to capture costs and benefits and
identify negative distributional effects for mitigation.
Given the mechanisms for addressing a problem within a power system, it may be that
intervention has a negative net benefit (that is, a net cost) but that this is less negative than
if nothing were done. It is therefore necessary to consider the appropriate counterfactual,
often a do-nothing scenario, against which the net benefit can be compared. Caution is
advised so that the cost of a do-nothing scenario is not considered a benefit (by being
avoided) of a demand-response scenario if the two scenarios are then netted against each
other—that is, double counted.
If the net benefit of the mechanisms is positive, with a positive difference as compared
with the counterfactual, it suggests the mechanism to pursue. If multiple mechanisms or
combinations of mechanisms are considered, rank them, and pursue the option with the
greatest net present value of benefits.5
Where distributional impacts see one customer or stakeholder group—for example,
nonparticipating customers—worse off, lower-ranked options that show less distributional
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 81
inequity may need to be considered. Alternatively, the assessment should ponder
compensation measures, including redistribution of benefits, to the extent possible.
Additional factors to consider when developing the cost-benefit analysis include:
• When evaluating combinations of mechanisms, any interactions should be
considered—that is, where one mechanism may affect customer behavior in relation
to another.
• Mechanisms can be introduced, altered, or withdrawn over time. Therefore, the cost-
benefit analysis must incorporate a time-based component.
• Consider multiple scenarios and sensitivities of aspects like customer growth, rates of
uptake, opt-in/opt-out/compulsory uptake,6 business models, and routes to market.
• Assess the likely uptake through different price signals.
• Would alternative financing arrangements (for example, concessional loans or grant
funding) provide additional value and customer incentives?
Implementation Design
The third and final step in implementing demand response is developing the implementation
process. The four sets of activities of the implementation process are discussed further in
the next four sections.
FIGURE 6.6
Implementing a Demand-Response Program in Four Steps
3.1 Design the 3.2 Ensure supportive 3.3 Adopt 3.4 Establish a
mechanism policy, legal, and accompanying work plan for
regulatory frameworks measures implementation
• Target customers • Guiding policy • Market education/ • Timeline (phasing,
• Price-setting policy • Enabling law communication responsibilities,
• Tax exemptions for milestones,
• Penalties for non- • Supportive regulations
equipment, if interlinkages,
delivery (including technical)
applicable dependencies)
• Provisions for change • Application process
• Social measures • Identification of
in mechanism over • Tariff (re)design, if possible quick wins
time applicable
• Provisions for pilot
• Engagement with • Market (re)design, if project, if applicable
direct stakeholders applicable
• Identification of
• Formal routes to
champions
market and business
(organisation and
models
individual)
• Institutional support
and capacity building
• Monitoring and
evaluation
• Program review
82 Roadmap for Implementing Demand Response in Developing Countries
Design the Mechanism
While the cost-benefit analysis will have developed and tested the concept of a mechanism
or combination of mechanisms, the design phase will work that concept out in practice.
Some points to consider in the design phase are summarized in Table 6.2.
TABLE 6.3
Factors to Consider in the Mechanism Design Phase
FACTOR CONSIDERATIONS
Target customer identification With how much granularity can participating customer group(s) be identified?
Will the groups change over time?
Will customers be added/removed?
Baseline determination and verification How can the objective that demand response is intended to achieve be quantified?
process
What is the baseline (current situation)?
What are the targets?
What is the process for measuring performance against those targets?
Price setting and penalties for non- What is the value of demand response to the power system? What forms of demand
delivery response can be monetized? How much of this value should be passed on to customers
through the price before incentives are reduced?
What is the cost of non-delivery by participating entities? How much of this cost can
and should be recovered through penalties for nonperformance?
Change in the mechanism over time Are changes in the mechanism expected over time?
Is there flexibility in the contractual arrangements to incorporate changes?
Participants in the design phase Which stakeholders are required to participate in the design phase—for example, the
regulator (via regulatory sandboxes), utility, customer representative bodies?
Which others can add value to it—for example, energy service providers, customers
with international experience in other countries/markets, technology designers?
Routes to market and business models Has agreement been reached on how enabling technology should be deployed and
how to split the cost between the utility, the distribution company, and the energy
service provider?
What role is there for supportive financing mechanisms—for example, commercial
loans, concessional loans, and grants, if the CBA agrees?
How can the emergence of new players and innovative business models be enabled
(within supportive policy, legal, and regulatory frameworks)
Provide a Supportive Policy, Legal,
and Regulatory Framework
As some demand-response mechanisms involve technological innovations and changes to
status quo operations, it may be that their implementation is hindered by unsupportive
policy, legal, and regulatory frameworks or could be accelerated through more clearly
supportive frameworks. Elements of such frameworks that should be considered in the
implementation phase are summarized in Table 6.3.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 83
TABLE 6.4
Elements of Supportive Policy, Legal, and Regulatory Frameworks for the Implementation of
the Chosen Demand-Response Mechanism
ELEMENT CONSIDERATIONS
Enabling law Clauses of laws may prohibit certain demand-response activities, technologies, and potential customer
discrimination (participant vs. nonparticipant) and should be addressed.
Changes to laws can be lengthy processes, so such elements should be identified and addressed at the
earliest opportunity.
Supportive regulations As with the enabling law, regulations may require updating or drafting to support the mechanism.
For price-based mechanisms, changes may be required in applicable tariff methodologies. Similarly,
price-based mechanisms may require a cost-of-service and tariff design study to set prices accurately.
Frameworks should accommodate new actors where required or anticipated (for example, aggregators).
Ensure that the legal and regulatory framework governing demand-response procurement incentivizes
the utility to treat it on a level playing field with supply-side resources, in particular cost recovery (see the
case of Viet Nam).
Demand response may be supported by the utility through direct investments in at the customer’s
premises. Where the regulatory framework does not permit this, alternative contracting structures can
be explored (for example, third-party provision, competitively sourced).
Application process Who will initiate an application to participate—for example, customers, installers, or utility?
Are the opt-in, opt-out, and mandatory participation processes clearly defined and well understood?
Who will receive and process applications?
Can timelines and milestones for application processing be formalized, with appropriate penalties?
Market (re)design To what extent is the mechanism dependent on market (re)design?
How will this be implemented, by whom, and on what timeline?
Accompanying Measures
Direct stakeholders will take on many of the activities needed to implement demand-
response measures. But other measures not directly associated with demand response
may be initiated by indirect stakeholders, some outside the energy sector. The relevant
energy ministry and regulators may wish to initiate discussions over such measures.
Table 6.4 provides suggestions for possible non–demand response measures.
Work Plan for Implementation
The final stage of the implementation preparation phase is to design a detailed roadmap
for implementation, which, when followed, will lead to full program implementation.
The roadmap will become the key implementation document for primary stakeholders
(identified in the roadmap), combining all elements of the implementation phase. The
roadmap could cover all activities laid out in this chapter. Key aspects of the roadmap
are presented in Table 6.5.
84 Roadmap for Implementing Demand Response in Developing Countries
TABLE 6.5
Non–Demand Response Measures in Support of Implementation
MEASURES CONSIDERATIONS
Market education/ New technologies and pricing schemes will typically require customer (and other stakeholder) education and
communication clear communication around costs, benefits, misconceptions, timelines, processes, and other elements of
the program.
Tax exemptions on The removal or zero-rating of taxes on equipment—for example, import duties, and value-added taxes—can
equipment lower costs for end customers, improving likely uptake.
Tax exemptions should be included in the cost-benefit analysis. If possible, the relevant tax authority should
be included as a stakeholder to show the net impact of the exemption on its revenues and any offset of
these through increases resulting from greater activity in other areas (for example, income taxes).
Social measures Special safety nets or direct benefit transfer schemes can be established for low-income customers and to
address inequity.
TABLE 6.6
Key Aspects of the Demand-Response Implementation Roadmap
MEASURES CONSIDERATIONS
Timeline The core of the roadmap is a timeline of all activities supported by a visual representation,
such as a Gantt chart.
The timeline should consider the nature of the demand-response program(s), activity phasing,
and the points at which where go/no-go decisions can be made at critical moments.
Activities in the timeline should identify the entities and, if possible, the people responsible,
with appropriate dates and milestones.
Some timeline activities may involve interlinkages and dependencies (activities that must be
completed before another can start).
Pilot projects and quick wins Pilot projects and quick wins should be implemented in regulatory sandBoxes to test the
efficacy of different mechanisms and incentive structures. Their results should be evaluated,
identifying pressing changes to existing regulations.
Examples may include working with existing technologies and customers known to utilities
who are keen to participate in a demand-response program without incentive or further
education.
Champion identification Innovative programs typically require a “champion” organization and, ideally, an individual
within that organization to ensure actions are completed according to the timeline.
A lack of a clear champion can hinder momentum, typically if a program is initiated by one
entity that then delegates to another (without full cooperation).
Institutional support and capacity Where program activities are new, which is likely in most cases, institutions may require
building support for their role in implementation.
Such support may be provided through targeted capacity building.
Monitoring and evaluation Program effectiveness and impact should be monitored and evaluated periodically to ensure
progress against predetermined targets, with adjustments made where progress is falling
short.
Considerations may include assessments of overall outcome impact, uptake, and price
changes. Adjustments may include changing incentives, prices, and requirements for opting in
or out.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 85
Endnotes
1. For simplicity, this chapter refers to countries but acknowledges that unbundled
entities may adopt demand response apart from others within the same country.
2. While this point addresses existing least-cost plans, the implementation process for
demand response should be incorporated within the least-cost planning process.
3. The South Africa case study (discussed in chapter 5.3) shows that the country has
multiple mechanisms operating in a single market.
4. This section provides no formula for conducting CBA, which other resources can
supply, but rather guidance on factors to incorporate in the assessment. In its simplest
form, the net present value of all benefits is divided by the net present value of all
costs to find a cost-benefit ratio. Subtracting the net present value of costs from the
net present value of benefits will give a net present value of benefits.
5. This is preferred over the greatest cost-benefit ratio as it provides a greater aggregate
benefit to the power system, rather than the highest return (which may not have such
a high aggregate benefit).
6. As noted earlier in the report, customers tend to stay with the status quo. Therefore,
opt-in modalities, which rely on customers choosing to sign up for the program, tend
to result in much lower subscription rates than opt-out modalities.
86 Roadmap for Implementing Demand Response in Developing Countries
SEVEN
CONCLUSION
Demand response is a short-term, voluntary reduction in electricity consumption by
end users. Such cutbacks are generally triggered by signals of compromised grid reliability
or high wholesale market prices. But their appeal is growing as power system participants
come to appreciate the utility of demand response as part of a decarbonization strategy.
Demand response can be a highly cost-effective mechanism that maintains the supply-
demand balance during peak hours or when the power system is stressed.
The potential of demand response in developing countries remains largely untapped.
While system operators in developed countries have leveraged demand response for many
years, it has been used sparingly in developing countries. The energy transition and the need
for decarbonization are expected to increase the need for demand response as systems
incorporating mounting levels of renewable generation require greater operational flexibility
to accommodate production and load variations. Policy makers are increasingly aware of the
cost-effectiveness of demand response as a source of flexibility and an essential support to
the energy transition.
Demand-response programs can be classified into two broad categories, depending
on the nature of the incentives adopted. Price-based (or implicit) mechanisms are based
on time-differentiated rates designed to encourage customers to shift consumption from
peak hours or to reduce consumption when the system is stressed. Quantity-based (or
explicit) mechanisms offer customers a direct payment for reducing their load and include
a broad set of solutions designed to shape customer consumption directly, such as
interruptible contracts and direct load control.
Developing countries can benefit from deploying price and quantity-based demand
response, depending on their objectives and market structure. Static ToU rates, which
have been essential in shifting consumption away from peak hours, are the most common
form of participation in demand response. Prices and time intervals do not change,
however, according to the power system’s criticality and need for more flexibility. Price-
based demand-response instruments can be enhanced to reflect system criticality through
more dynamic tariffs (such as CPP). Developing countries may also benefit from quantity-
based load control and interruptible contracts, commonly seen already through control
of hot water heaters, payments for large consumers to reduce short-term demand, and
similar mechanisms. As power sectors evolve, more sophisticated mechanisms, such as
the use of demand response in the energy and capacity markets, can also be explored.
Demand response can be especially cost-effective in small island developing states having
limited resources to accommodate the integration of VRE and manage peak demand.
The most valuable demand-response mechanisms offer maximum flexibility to
the power system operator. This entails eliciting demand response from dependable
resources that can provide a variety of services to the system operator across a range of
timeframes and that can be controlled directly and automated.
Technology, supportive frameworks, and customer engagement will be critical
for mainstreaming demand response. Smart meters, already deployed extensively
worldwide, can support the introduction of dynamic rates and load controls. Other
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 89
technologies, such as smart appliances, smartphones, and V2G (vehicle-to-grid), are
expected to boost the role and functionality of demand-response mechanisms. The right
policies and regulations can enable new business models and new service providers to play
key roles in aggregating and controlling multiple smaller loads that are part of the demand-
response spectrum. Shifting customers to new tariff arrangements and innovative
technologies must be communicated and managed carefully, while allowing customers
to opt out should they wish to do so.
Growing experience with the implementation of demand-response programs,
including pilots and early adoption, is instructive for developing countries. In addition
to the considerable experience in developed countries, developing countries can draw on
the know-how gained by other developing and emerging economies where the operating
context may be more relatable. A greater understanding of the lessons learned from
successes and failures can help them design leapfrog approaches to improving system
reliability and grid flexibility.
Countries should develop demand-response programs designed for the national
context, with appropriate mechanisms and detailed work plans for implementation.
Countries with constraints that could be alleviated through demand response should take
pains to understand their national context so as to ensure that interventions are well-
targeted. Selections should be made from the expansive menu of demand-response
mechanisms, with potential and distributional impacts assessed through a robust cost-
benefit analysis. The identified mechanisms should be implemented through a work plan
accompanied by pilot projects and quick wins; appropriate policy, legal and regulatory
interventions; response measures not directly connected to demand response (particularly
market education); allocation of implementation tasks among clearly identified parties and
along clear milestones; and effective monitoring and evaluation to assess program success.
90 Conclusion
Bibliography
ACER/CEER (Agency for the Cooperation of Energy Regulators/ Council of European Energy
Regulators). 2021. Annual Report on Monitoring the Internal Energy Retail Markets and
Customer Protection. Slovenia: ACER and Brussels: CEER.
Alstone, Peter, Jennifer Potter, Mary Ann Piette, Peter Schwartz, Michael A. Berger, Laurel N.
Dunn, Sarah J. Smith, Michael D. Sohn, Arian Aghajanzadeh, Sofia Stensson, Julia Szinai,
Travis Walter, Lucy McKenzie, Luke Lavin, Brendan Schneiderman, Ana Mileva, Eric
Cutter, Arne Olson, Josh Bode, Adriana Ciccone, and Ankit Jain. 2017. 2025 California
Demand Response Potential Study—Charting California’s Demand Response Future: Final
Report on Phase 2 Results. Berkeley, CA: Lawrence Berkeley National Lab. https://
eta-publications.lbl.gov/sites/default/files/lbnl-2001113.pdf.
ANEEL (Agência Nacional de Energia Elétrica). 2017a. “Nota Técnica No. 100/2017-SRG/
ANEEL.” ANEEL, Brasília, Brazil.
ANEEL. 2017b. “Resolução Normativa No. 792/2017.” ANEEL, Brasília, Brazil.
Balaraman, K. 2022. “PG&E, Tesla Virtual Power Plant Delivers 16.5 MW to California
Grid amid Calls for Energy Conservation.” Utility Drive, August 23, 2022. https://
www.utilitydive.com/news/pge-tesla-virtual-power-plant/630310/?utm_source=
Sailthru&utm_medium=email&utm_campaign=Issue:%202022-08-23%20Utility
%20Dive%20Storage%20%5Bissue:44037%5D&utm_term=Utility%20Dive:%20Storage.
Baldwin, S. and others 2019. Making the Grid Smarter: Primer on Adopting the New IEEE
Standard 1547–2018 for Distributed Energy Resources. Albany, NY: Interstate Renewable
Energy Council (IREC).
Barbose, G., C. Goldman, R. Bharvirkar, N. Hopper, and M. Ting. 2005. Real Time Pricing as a
Default or Optional Service for C&I Customers: A Comparative Analysis of Eight Case Studies.
Berkley, CA: Lawrence Berkeley National Laboratory.
Bertoli, E. and others 2021. Demand Response—More Efforts Needed. Tracking Report,
November 2021. Paris: International Energy Agency (IEA). https://www.iea.org/
contributors/emi-bertoli.
Bonifacic, I. 2020. “Google’s Data Centers Will Process Some Tasks When Clean Energy
Is Abundant.” The Engadget Podcast, April 23, 2020. https://www.engadget.com/
google-carbon-intelligent-computing-platform-200316070.html.
BPC (Botswana Power Corporation). N.d. “Frequently Asked Questions.” https://www.bpc
.bw/about-bpc/faq.
Brattle Group and SEPA (Smart Electric Power Alliance). 2018. Utility Demand Response
Market Snapshot. Washington, DC: SEPA.
Brattle Group and SEPA. 2019. Utility Demand Response Market Snapshot. Washington, DC:
SEPA. https://www.ashb.com/wp-content/uploads/2020/04/IS-2019-137.pdf.
Bureau of Energy Efficiency. 2023. Compendium on DSM Measures by Electricity Distribution
Companies (DISCOMs). New Delhi: Ministry of Power, Government of India.
Burger, S., J. Chaves-Ávila, C. Batlle, and J. Pérez-Arriaga. 2016. “The Value of Aggregators
in Electricity Systems.” CEEPR Working Paper WP 2016-001, MIT Center for Energy and
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 91
Environmental Policy Research, Massachusetts Institute of Technology, Cambridge, MA.
https://energy.mit.edu/wp-content/uploads/2016/01/CEEPR_WP_2016-001.pdf.
BYPL (BSES Yamuna Power Limited). 2018. “Value Proposition from Demand Response
Program.” BYPL, New Delhi.
Cadmus Group. 2018. Assessment of Barriers to Demand Response in the Northwest’s Public
Power Sector. Report for the Bonneville Power Administration. Waltham, MA: Cadmus
Group.
California ISO (Independent System Operator Corporation). 2022. Demand Response Issues
and Performance 2021. Folsom, CA: Department of Market Monitoring, California ISO.
Cappers, P., C. Goldman, and D. Kathan. 2010. “Demand Response in U.S. Electricity
Markets: Empirical Evidence.” Energy 35 (4): 1526–35. https://www.sciencedirect.com/
science/article/abs/pii/S0360544209002461.
Cappers, P., J. S. MacDonald, and C. A. Goldman. 2013. Market and Policy Barriers for
Demand Response Providing Ancillary Services in U.S. Markets. Report LBNL-6155E.
Berkeley, CA: Lawrence Berkeley National Lab.
CCEE (Câmara de Comercialização de Energia Elétrica) and ONS (Operador Nacional do
Sistema). 2018. 1 Relatório de Análise do Programa Piloto de Resposta da Demanda.
São Paulo, Brazil: CCEE and ONS.
CCEE and ONS. 2019. 2 Relatório de Análise do Programa Piloto de Resposta da Demanda.
São Paulo, Brazil: CCEE and ONS.
CEMIG (Companhia Energética de Minas Gerais). 2021. “Valores de Tarifas e Serviços.”
https://www.cemig.com.br/atendimento/valores-de-tarifas-e-servicos/.
CER (Centre for Energy Regulation). 2019. “Tariff Principles and Design with a Focus on
ToD Tariff and Market Based Dynamic ToD.” Presentation at the12th Capacity Building
Workshop for Officers of ERCs at Indian Institute of Technology Kanpur, February.
CERRE (Centre on Regulation in Europe). 2022. The Active Distribution System Operator (DSO):
An International Study. Brussels, Belgium: CERRE. https://cerre.eu/wp-content/
uploads/2022/09/CERRE_DSO_FinalReport_2709.pdf.
Chen, Y., L. Zhang, P. Xu, and A. Di Gangi. 2021. “Electricity Demand Response Schemes in
China: Pilot Study and Future Outlook.” Energy 224 (June): 120042.
Cho, G., and G. Son. 2022. “The Unlevel Playing Field: How the Power Market Structure
Discriminates against Demand Response to Favor Gas Power Generation.” Solutions for
Our Climate (SFOC). https://forourclimate.org/hubfs/%5BENG%5D%20The%20Unlevel
%20Playing%20Field_How%20the%20Power%20Market%20Structure%20Discriminates
%20Against%20Demand%20Response%20to%20Favor%20Gas%20Power%20
Generation.pdf.
Chunekar, A., M. Kelkar, and S. Dixit. 2014. Demand Side Management in India: An Overview of
State Level Initiatives. Pune, India: Prayas Energy Group.
Clean Energy Ministerial. 2014. “Energy Efficient Cooling and Demand Response: Pre-Read
for Public-Private Roundtable.” Clean Energy Ministerial, Seoul, Republic of Korea,
May 12, 2014. https://www.cleanenergyministerial.org/content/uploads/2022/03/
cem5-rt-coolinganddr-pres.pdf.
CNMC (Comisión Nacional de los Mercados y la Competencia). 2022. “Mercado eléctrico.”
https://www.cnmc.es/ambitos-de-actuacion/energia/mercado-electrico.
92 BIBLIOGRAPHY
Conteh, A., M. Lotfy, O. Adewuyi, P. Mandal, H. Takahashi, and T. Senjyu. 2020. “Demand
Response Economic Assessment with the Integration of Renewable Energy for
Developing Electricity Markets.” Sustainability 12 (7): 2653.
CPower. 2020. State of Demand Side Energy Management in North America. Baltimore, MD:
CPower. http://info.cpowerenergymanagement.com/rs/377-RWN-330/images/CPower
%202020%20State%20of%20Demand%20Side%20Energy%20Management%20in%20
North%20America.pdf.
CPower. 2023. 2021–2022 State of Demand Side Energy Management in North America: Volume III.
Baltimore, MD: CPower.
Davis, L. 2022. “How High Did California’s Electricity Prices Get?” Energy Institute Blog,
UC Berkeley, September 12, 2022. https://energyathaas.wordpress.com/2022/09/12/
how-high-did-californias-electricity-prices-get/.
Deshmukh, R., R. Yin, G. Ghatikar, and G. Saha. 2023. “Estimation of Potential and Value
of Demand Response for Industrial and Commercial Consumers in Delhi.” Tata Power
Delhi Distribution Limited, New Delhi, India. Unpublished title.
Dominion Energy. 2021. “Smart Cooling Rewards Program: Making a Decision to Make a
Difference.” https://www.dominionenergy.com/virginia/save-energy/smart-cooling-
reward.
EDF (Électricité de France). 2014. “Technology Unveiled.” https://www.edf.fr/sites/default/
files/Lot%203/CHERCHEURS/Publications/advanced-management-distribution-grids.pdf.
EIA (US Energy Information Administration). 2023. “Independent Statistics and Analysis.”
https://www.eia.gov/tools/faqs/faq.php?id=108&t=3#:~:text=How%20many%20
smart%20meters%20are,of%20total%20electric%20meters%20installations.
Elia. 2021. “Baseline Methodology Assessment.” Public Consultation. https://www.google.com/
url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwiAh73zo7
GGAxVgVEEAHRfOCCMQFnoECBIQAQ&url=https%3A%2F%2Fwww.elia.be%2F-%2Fmedia
%2Fproject%2Felia%2Felia-site%2Fpublic-consultations%2F2021%2F20210927_study_
baseline_methodologies_draft_clean_en.pdf&usg=AOvVaw2ucuT7i0LiB4yt-yg23rHf&
opi=89978449.
Environmental Defense Fund. 2021. “The Costs and Benefits of Real Time Pricing in Illinois—
An Empirical Investigation into Consumer Bills Using Hourly Energy Data and Prices.”
CPUC Workshop on Advanced DER and Demand Flexibility Management.
Enel. 2023. “Tabela de tarifas.” https://www.enel.com.br/pt/Corporativo_e_Governo/
tabela-de-tarifas.html.
EPE (Empresa de Pesquisa Energética). 2018. Plano Decenal de Expansão de Energia 2027.
Brasília, Brazil: EPE, Ministério de Minas e Energia. https://www.epe.gov.br/pt/
publicacoes-dados-abertos/publicacoes/plano-decenal-de-expansao-de-energia-2027.
Eskom. 2020. Tariffs & Charges Booklet 2020/2021. Sunninghill, South Africa: Eskom.
Eskom. 2021. “Transmission Business System Operator (SO) Overview.” Presentation,
Eskom, Sunninghill, South Africa.
ESMAP (Energy Sector Management Assistance Program). 2011. “Best Practices for Market-
Based Power Rationing: Implications for South Africa.” ESMAP Briefing Note 008/11,
World Bank, Washington, DC. https://www.esmap.org/sites/default/files/esmap-files/
ESMAP_PowerRationing_008-11.pdf.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 93
European Union. 2019. “Directive (EU) 2019/944 of the European Parliament and of the
Council of 5 June 2019 on Common Rules for the Internal Market for Electricity and
Amending Directive 2012/27/EU (recast).” Official Journal of the European Union.
https://eur-lex.europa.eu/eli/dir/2019/944/oj.
Faruqui, A. 2005. Applications of Dynamic Pricing in Developing and Emerging Economies.
Washington, DC: World Bank.
Faruqui, A., R. Hledik, and S. Sergici. 2019. “A Survey of Residential Time-of-Use (TOU) Rates.”
The Brattle Group. https://www.brattle.com/wp-content/uploads/2021/05/17904_a_
survey_of_residential_time-of-use_tou_rates.pdf
Faruqui, A., R. Hledik, and J. Palmer. 2012. Time-Varying and Dynamic Rate Design. Regulatory
Assistance Project (RAP) and the Brattle Group. https://www.raponline.org/wp-content/
uploads/2023/09/rap-faruquihledikpalmer-timevaryingdynamicratedesign-2012-jul-23.pdf.
Faruqui, A., S. Sergici, and C. Warner. 2017. “Arcturus 2.0: A Meta-Analysis of Time-Varying
Rates for Electricity.” The Electricity Journal 30 (10): 64–72.
Faruqui, A., N. Lessem, S. Sergic, D. Mountain, F. Denton, B. Spencer, and C. King. 2016.
Analysis of Ontario’s Full-Scale Roll-Out of ToU Rates—Final Study. Prepared for
Independent Electric System Operator (IESO). The Brattle Group. https://www.brattle.
com/wp-content/uploads/2017/10/7289_analysis_of_ontarios_full_scale_roll-out_of_tou_
rates_-_final_study-3.pdf.
Federal Energy Management Program. N.d. “Demand Response and Time-Variable Pricing
Programs.” https://www.energy.gov/femp/demand-response-and-time-variable-pricing
-programs.
FERC (Federal Energy Regulatory Commission). 2008. Wholesale Competition in Regions with
Organized Markets. Order 719. Final Rule. Washington, DC: FERC. https://ferc.gov/media/
order-no-719.
FERC. 2018. Electric Storage Participation in Markets Operated by Regional Transmission
Organizations and Independent System Operators. Order 841. Final Rule. Washington, DC:
FERC. https://www.ferc.gov/media/order-no-841.
FERC. 2019. Assessment of Demand Response and Advanced Metering. Staff Report, December
2019. Washington, DC: FERC.
FERC. 2023. Assessment of Demand Response and Advanced Metering. Staff Report, December
2023. Washington, DC: FERC.
Ferreira, R., L. Barroso, P. Lino, P. Valenzuela, and M. Carvalho. 2013. “Time-of-Use Tariffs
in Brazil: Design and Implementation Issues.” Presentation at the 2013 IEEE PES
Conference on Innovative Smart Grid Technologies (ISGT Latin America), April 15–17,
2023. https://ieeexplore.ieee.org/document/6554486/authors#authors.
Foster, V., and S. H. Witte. “Falling Short: A Global Survey of Electricity Tariff Design.” Policy
Research Working Paper 9174, World Bank, Washington, DC. https://ssrn.com/abstract=
3549004.
Future Market Insights. 2022. Demand Response Market: Demand Response Market by Type,
Vertical and Region. Forecast 2022–2032. Newark, Delaware: Future Market Insights. https://
www.futuremarketinsights.com/reports/demand-response-market#:~:text=%5B323
%20Pages%20Report%5D%20The%20global,Million%20by%20the%20year%202032.
García-Villalobos, J., I. Zamora, J. I. San Martín, F. J. Asensio, and V. Aperribay. 2014.
“Plug-In Electric Vehicles in Electric Distribution Networks: A Review of Smart Charging
Approaches.” Renewable and Sustainable Energy Reviews 38 (October): 717–31.
94 BIBLIOGRAPHY
Gerke, Brian, Giulia Gallo, Jingjing Liu, Mary Ann Piette, Peter Schwartz, and Peter Alstone.
2018. “Shift Demand Response: A Primer.” Lawrence Berkeley National Laboratory,
Berkeley, CA. https://gridworks.org/wp-content/uploads/2018/02/Shift-Demand-
Response-Primer_Final_180227.pdf.
Gerke, Brian, Giulia Gallo, Sarah J. Smith, Jingjing Liu, Peter Alstone, Shuba Raghavan,
Peter Schwartz, Mary Ann Piette, Rongxin Yin, and Sofia Stensson. 2020. The California
Demand Response Potential Study, Phase 3: Final Report on the Shift Resource through 2030.
Berkeley, CA: Lawrence Berkeley National Laboratory. https://eta-publications.lbl.gov/
sites/default/files/ca_dr_potential_study_-_phase_3_-_shift_-_final_report.pdf.
Giraldez, Julieta, Adarsh Nagarajan, Peter Gotseff, Venkat Krishnan, Andy Hoke, Reid Ueda,
Jon Shindo, Marc Asano, and Earle Ifuku. 2017. Simulation of Hawaiian Electric Companies
Feeder Operations with Advanced Inverters and Analysis of Annual Photovoltaic Energy
Curtailment. NREL Technical Report NREL/TP-5D00-68681. Golden, CO: National
Renewable Energy Laboratory. https://www.nrel.gov/docs/fy17osti/68681.pdf.
Girouard, Coley. 2019. “Behavioral Demand Response Gives Baltimore Gas and Electric a
Business Reason to Reduce Peak Usage.” https://www.utilitydive.com/news/behavioral-
demand-response-gives-baltimore-gas-and-electric-a-business-reas/546895/.
GIZ (Deutsche Gesellschaft für Internationale Zusammenarbeit) GmbH. 2021. SGREEE-AA1:
Promoting Implementation of Demand Response Programs in Vietnam. Bonn and Eschborn,
Germany: Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ). https://
energypedia.info/images/b/b4/011_SGREEE-AA1_Promoting_Implementation_of_
Demand_Response_Programs_in_Vietnam.pdf.
Going Solar. 2013. “I Have the Power! (Ripple Control, and Load Shedding in Power Systems).”
https://goingsolar.co.za/2013/06/19/i-have-the-power-ripple-control-and-load-shedding
-in-power-systems/.
Goldman, C., G. L. Barbose, and J. H. Eto. 2002. “California Load Reductions during the
Electricity Crisis: Did They Help to Keep the Lights On?” Lawrence Berkeley National
Laboratory, Berkeley, CA. https://escholarship.org/uc/item/8rq0j1c2.
Government of Botswana. 2019. Botswana Energy Accounting Report: 2015/16. Gaborone:
Government of Botswana. https://www.wavespartnership.org/sites/waves/files/
kc/201516%20Botswana%20Energy%20Accounts%20report_fv%20.pdf.
Government of South Africa. 2008. “National Response to South Africa’s Electricity
Shortage.” Inter-Ministerial Briefing. https://www.gov.za/documents/other/
national-response-south-africas-electricity-shortage-25-jan-2008.
Hale, E., L. Bird, R. Padmanabhan, and C. Volpi. 2018. Potential Roles for Demand Response in
High-Growth Electric Systems with Increasing Shares of Renewable Generation. NREL Technical
Report NREL/TP-6A20-70630. Golden, CO: National Renewable Energy Laboratory. https://
www.nrel.gov/docs/fy19osti/70630.pdf.
Han, H-S., E. Oh, and S-Y. Son. 2018. “Study on EV Charging Peak Reduction with V2G Utilizing
Idle Charging Stations: The Jeju Island Case.” Energies 11 (7): 1651. https://doi.org/10.3390/
en11071651.
Hawaiian Electric Company. 2024a. “Fast Demand Response.” https://www.hawaiianelectric.
com/products-and-services/customer-incentive-programs/fast-demand-response.
Hawaiian Electric Company. 2024b. “Frequently Asked Questions: What Is the Fast DR Pilot
Program.” https://www.hawaiianelectric.com/customer-service/frequently-asked-questions?
FAQFilter1=845&page=1.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 95
Heffner, Grayson, Charles Goldman, Brendan Kirby, and Michael Kintner-Meyer. 2007.
Loads Providing Ancillary Services: Review of International Experience. Berkeley, CA: Ernest
Orlando Lawrence Berkeley National Laboratory.
Heffner, G., L. Maurer, A. Sarkar, and X. Wang. 2009. “Minding the Gap: World Bank’s Assistance
to Power Shortage Mitigation in the Developing World.” Energy 35 (4): 1584–91.
Hledik, R., and A. Faruqui. 2015. Valuing Demand Response: International Best Practices, Case
Studies, and Applications. Report prepared for EnerNOC. Boston, MA: The Brattle Group.
Hledik, R., and K. Peters. 2023. Real Reliability: The Value of Virtual Power. Volume I: Summary
Report. Boston, MA: Brattle Group. https://www.brattle.com/wp-content/uploads/
2023/04/Real-Reliability-The-Value-of-Virtual-Power_5.3.2023.pdf.
Hledik, R., K. Viswanathan, and K. Peters. 2023. “Virtual Power Plants: Resource Adequacy
without Interconnection Delays.” Utility Dive, August 17, 2023. https://www.utilitydive
.com/news/virtual-power-plants-vpp-distributed-energy-resource-adequacy-der-distributed
-energy/691135/.
Hledik, R., A. Faruqui, T. Lee, and J. Higham. 2019. The National Potential for Load Flexibility:
Value and Market Potential Through 2030. Boston, MA: The Brattle Group.
Horschig, T., O. Özgün, and F. Jones. 2023. “How Vehicle-to-Everything (V2X) Can Turbocharge
the Energy Transition.” DNV. https://www.dnv.com/energy-transition-outlook/v2x.html.
Hurley, Doug, Paul Peterson, and Melissa Whited. 2013. Demand Response as a Power
System Resource: Program Designs, Performance, and Lessons Learned in the United States.
Regulatory Assistance Project. http://www.synapse-energy.com/sites/default/files/
SynapseReport.2013-03.RAP_.US-Demand-Response.12-080.pdf.
IEA (International Energy Agency). 2018. The Future of Cooling: Opportunities for Energy-
Efficient Air Conditioning. Paris: IEA. https://www.iea.org/reports/the-future-of-cooling.
IEA. 2020. Demand Response: Tracking Report. Online report, June 2020. https://www.iea.org/
reports/demand-response.
IEA. 2021a. Demand Response: More Is needed. Tracking Report, October. Paris: IEA. https://
www/iea.org/reports/demand-response.
IEA. 2021b. Net Zero by 2050: A Roadmap for the Global Energy Sector. Tracking Report,
November 2021. Paris: IEA. https://iea.blob.core.windows.net/assets/deebef5d-0c34-4539
-9d0c-10b13d840027/NetZeroby2050-ARoadmapfortheGlobalEnergySector_CORR.pdf.
IEA. 2021c. India Energy Outlook 2021. Paris: IEA. https://www.iea.org/reports/india-energy
-outlook-2021#overview.
IEA. 2023. “Keeping Cool in a Hotter World Is Using More Energy, Making Efficiency More
Important than Ever.” https://www.iea.org/commentaries/keeping-cool-in-a-hotter
-world-is-using-more-energy-making-efficiency-more-important-than-ever.
Invest Korea. 2021. “Use Renewable Energy as Much as You Want, and Get Local Currency
as a Bonus.” Jeju Industry News, November 8, 2021. https://www.investkorea.org/jj-en/
bbs/i-2584/detail.do?ntt_sn=8.
IRENA (International Renewable Energy Agency). 2017. Adapting Market Design to High Shares
of Variable Renewable Energy. Abu Dhabi: IRENA. https://www.irena.org/publications/2017/
May/Adapting-Market-Design-to-High-Shares-of-Variable-Renewable-Energy.
IRENA. 2019a. “Time of Use Tariffs: Innovation Landscape Brief.” IRENA, Abu Dhabi. https://
www.irena.org/-/media/Files/IRENA/Agency/Publication/2019/Feb/IRENA_Innovation_
ToU_tariffs_2019.pdf.
96 BIBLIOGRAPHY
IRENA. 2019b. “Demand-Side Flexibility for Power Sector Transformation.” Analytical brief,
IRENA, Abu Dhabi. https://www.irena.org/-/media/Files/IRENA/Agency/Publication/2019/
Dec/IRENA_Demand-side_flexibility_2019.pdf.
Jackson, G. 2018. Agile Octopus: A Consumer-Led Shift to a Low Carbon Future. London: Octopus
Energy. https://octopus.energy/blog/agile-report/.
Jones, J. 2022. “Demand Response Proves More Profitable than Bitcoin Mining in Texas.” Smart
Energy International, August 23, 2022. https://www.smart-energy.com/industry-sectors/
energy-grid-management/demand-response-proves-more-profiTable-than-bitcoin-mining
-in-texas/.
Kalsi, K., M. Elizondo, J. Lian, W. Zhang, L. D. Marinovici, and C. Moy. 2014. Loads as a Resource:
Frequency Responsive Demand. Richland, WA: Pacific Northwest National Laboratory. https://
www.pnnl.gov/main/publications/external/technical_reports/PNNL-23764.pdf.
Kim, J., G. Han, and H. Moon. 2022. “Jeju’s 2030 Carbon-Free Vision Begins with Renewable
Curtailment Freedom: Cost Analysis of Solutions to Renewable Curtailment on Jeju
Island, South Korea.” Solutions for Our Climate (SFOC) and NEXT Group. https://
forourclimate.org/hubfs/Jejus%202030%20Carbon-Free%20Vision%20Begins%20
With%20Renewable%20Curtailment%20Freedom_ENG-1.pdf.
King, C. 2004. “Integrating Residential Dynamic Pricing and Load Control: The Literature.”
Energy Central, December 14, 2004. https://energycentral.com/c/iu/integrating-residential
-dynamic-pricing-and-load-control-literature.
King, Robert, Jaden Crawford, Barry Huddleston, and Steve Isser. 2015. “The Debate about
Demand Response and Wholesale Electricity Markets.” South-Central Partnership for
Energy Efficiency as a Resource (SPEER), Austin, TX. https://eepartnership.org/
wp-content/uploads/2015/10/The-Debate-About-Demand-Response-Final-10.28.15.pdf.
Korea Bizwire. 2021. “Jeju Eyes Compensation System to Encourage Use of Residual Electricity
for EV Charging.” Korea Bizwire, October 1, 2021. http://koreabizwire.com/jeju-eyes
-compensation-system-to-encourage-use-of-residual-electricity-for-ev-charging/200579.
Kornfield, M., and P. Firozi. 2021. “Some Texans Received Electric Bills up to $17,000 after
the Storm. How Does That Happen?” Washington Post, February 21, 2021. https://
www.washingtonpost.com/nation/2021/02/21/texas-high-electric-bills/.
Latief, Y. 2023. “How Hackers Target Smart Meters to Attack the Grid.” Smart Energy
International, May 04, 2023. https://www.smart-energy.com/industry-sectors/
cybersecurity/how-hackers-target-smart-meters-to-attack-the-grid/#:~:text=According
%20to%20the%20study%2C%20with,or%20inserting%20false%20control%20data.
Letendre, S., P. Denholm and P. Lilienthal. 2006. “New Load or New Resource.” Public Utilities
Fortnightly 144 (12): 28–37.
Liu, Y. 2016. Frequency Response Assessment and Improvement of Three Major North American
Interconnections Due to High Penetrations of Photovoltaic Generation. Oak Ridge, TN: Office of
Scientific and Technical Information (OSTI), US Department of Energy. doi:10.2172/1569649.
Lynch, M., S. Nolan, M. T. Devine, and M. O’Malley. 2019. “The Impacts of Demand
Response Participation in Capacity Markets.” Applied Energy 250 (September): 444–51.
https://www.sciencedirect.com/science/article/abs/pii/S0306261919309195.
Marcos, P., and L. Maurer. 2020. Designing Auctions to Increase Competition in Non-
Conventional Renewable Energy: The Colombian Case. World Bank, Washington DC.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 97
Matsuda-Dunn, R., K. McKenna, J. Desai, and P. Mukoma. 2024. Determining and Unlocking
Untapped Demand-Side Management Potential in South Africa: Demand Response at the
Grid Edge. NREL Technical Report NREL/TP-6A40-88042. Golden, CO: National Renewable
Energy Laboratory. https://www.nrel.gov/docs/fy24osti/88042.pdf.
Maurer, L. 2011. “Best Practices for Market-Based Power Rationing: Implications for South
Africa.” ESMAP Briefing Note 008/11, World Bank, Washington, DC. https://www.esmap
.org/node/1259.
Maurer, L. 2019. “Tarifas que incentivam a resposta da demanda.” ANEEL, International
Seminar on Electricity Tariff Structure, Brasilia, Brazil.
Maurer, L. 2022. “Crise da Califórnia e fim dos veículos elétricos—Projeções, fatos e opiniões.”
iNFRADebate, iNFRA, September 14, 2022. https://www.agenciainfra.com/blog/
infradebate-crise-da-california-e-fim-dos-veiculos-eletricos-projecoes-fatos-e-opinioes/.
Maurer, L., M. Pereira, and J. Rosenblatt. 2005. Implementing Power Rationing in a Sensible
Way: Lessons Learned and International Best Practices. Energy Sector Management
Assistance Program (ESMAP) Formal Report 305/05. Washington, DC: World Bank.
Mehra, Varun, and Raiden Hasegawa. 2023. “Supporting Power Grids with Demand
Response at Google Data Centers.” Google Cloud blog post, October 4, 2023. https://
cloud.google.com/blog/products/infrastructure/using-demand-response-to-reduce-data
-center-power-consumption.
Meier, A. 2005. Saving Electricity in a Hurry: Dealing with Temporary Shortfalls in Electricity
Supplies. Paris: IEA.
Ministry of Industry and Trade (Vietnam). 2017. “Circular no. 23/2017/TT-BCT . . . on
Prescribing Contents and Processes for Implementation of Load Adjustment Programs.”
November 16. https://english.luatvietnam.vn/circular-no-23-2017-tt-bct-dated-november
-16-2017-of-the-ministry-of-industry-and-trade-on-prescribing-contents-and-processes
-for-implementation-of-118496-doc1.html.
Modi, V. 2007. A Power Crunch in Sub-Saharan Africa (draft). Department of Mechanical
Engineering. Earth Institute Columbia University, January 2007. Unpublished title.
Morales-España, G., R. Martínez-Gordón, and J. Sijm. 2022. “Classifying and Modelling Demand
Response in Power Systems.” Energy 242 (March): 122544. https://www.sciencedirect.com/
science/article/pii/S0360544221027936.
Muller, G. 2016. “Impacto de Novas Tecnologias e Smart Grids na Demanda de Longo Prazo
do Sistema Elétrico Brasileiro.” Ph.D. Thesis, Universidade Federal do Rio de Janeiro,
Rio de Janeiro.
Muller, G. 2020. Impact of Demand Response on Generation Expansion Planning in the
Brazilian Interconnected Power System. Technical Report. Rio de Janeiro, Brazil: Empresa
de Pesquisa Energética (EPE).
National Grid ESO. 2024. “Future Energy Scenarios (FES).” https://www.nationalgrideso.com/
future-energy/future-energy-scenarios-fes.
Natural Resources Defense Council. 2003. Demand-Side Management in China: Benefits,
Barriers, and Policy Recommendations. NRDC. https://www.nrdc.org/sites/default/files/
dsm.pdf.
Navigant Research. 2016. “Market Data: Demand Response. Executive Summary.” Document
no longer available online.
98 BIBLIOGRAPHY
Nemtzow, David, Dan Delurey, and Chris King. 2007. “Demand Response: The Green Effect.”
Demand Response Special Edition, Public Utilities Fortnightly, March 2007.
NERC (North American Electric Reliability Corporation). 2013. 2011 Demand Response
Availability Report. Atlanta, GA: NERC. https://nercstg.nerc.com/comm/PC/DADSWG
RelatedFilesDL/DADS%202011%20Demand%20Response%20Availability%20Report%20
(March%202013).pdf.
NERC. 2019. 2019 State of Reliability. Atlanta, GA: NERC. https://www.nerc.com/pa/RAPA/PA/
Performance%20Analysis%20DL/NERC_SOR_2019.pdf.
Next. 2023. “Virtual Power Plant: How to Network Distributed Energy Resources.” https://
www.next-kraftwerke.com/vpp/virtual-power-plant.
Nissan. 2016. “Nissan, Enel and Nuvve Operate World’s First Fully Commercial Vehicle-to
-Grid Hub in Denmark.” Enel, August 29, 2016. https://www.enel.com/media/explore/
search-press-releases/press/2016/08/nissan-enel-and-nuvve-operate-worlds-first-fully
-commercial-vehicle-to-grid-hub-in-denmark.
NREL (National Renewable Energy Laboratory). 2018. Potential Roles for Demand Response
in High-Growth Electric Systems with Increasing Shares of Renewable Generation. Report
elaborated by E. Hale, L. Bird, R. Padmanabhan, and C. Volpi. NREL Technical Report
NREL/TP-6A20-70630. Golden, CO: NREL.
Octopus Energy. 2023. “Octopus Energy Group Boosts Kraken’s Capabilities with Innovative
‘EnTech’ Software.” Octopus Energy, March 9, 2020. https://octopus.energy/press/
octopus-energy-group-boosts-krakens-capabilities-with-innovative-entech-software/.
NOS (Operador Nacional do Sistema). 2022. Avaliação do Mecanismo de Resposta Voluntária
da Demanda—Portaria 22/2021. RT-NOS DTA 0091/2022.
Orvis, R., and S. Aggarwal. 2017. A Roadmap for Finding Flexibility in Wholesale Markets: Best
Practices for Market Design and Operations in a High Renewables Future. San Francisco, CA:
Energy Innovation Policy and Technology. https://energyinnovation.org/wp-content/
uploads/2017/10/A-Roadmap-For-Finding-Flexibility-In-Wholesale-Power-Markets_FINAL.pdf.
Park, G., K. Y. Lee, and N. H. Song. 2022. “Research and Demonstration on Mitigation of
Renewable Energy Curtailment Using Electric Vehicle Batteries.” CIRED Porto Workshop
2022: E-Mobility and Power Distribution Systems: 274–78. Presented at the Hybrid
Conference, Porto, Portugal. https://ieeexplore.ieee.org/document/9841662.
Park, K., D. Jang, S. Kim, Y. Lim, and J. Lee. 2023. “A Grid-Friendly Electric Vehicle
Infrastructure: The Korean Approach.” IEEE Power and Energy Magazine, Vol. 21, no. 6,
pp. 28–37. https://ieeexplore.ieee.org/document/10287713.
Pasquier, S. 2011. Saving Electricity in a Hurry: Update 2011. Paris: IEA. https://iea.blob.core
.windows.net/assets/dd89096f-3676-4859-a34a-ef21331f18a1/Saving_Electricity.pdf.
PJM (Pennsylvania–New Jersey–Maryland Interconnection). 2017a. 2022 Demand Response
Operations Markets Activity Report: July 2023. Norristown, PA: PJM. https://www.pjm.com/
-/media/markets-ops/dsr/2022-demand-response-activity-report.ashx.
PJM. 2017b. Demand Response Strategy. Norristown, PA: PJM. https://www.pjm.com/~/
media/library/reports-notices/demand-response/20170628-pjm-demand-response
-strategy.ashx.
PJM. 2024. “Demand Response Fact Sheet.” https://learn.pjm.com/-/media/about-pjm/
newsroom/fact-sheets/demand-response-fact-sheet.ashx.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 99
PLMA (Peal Load Management Alliance). 2017. “Evolution of Demand Response in the
United States Electricity Industry.” Whitepaper, May 25, 2017. Evolution-of-DR-White
-Paper.pdf (peakload.org).
Pons, Thiery. 2012. “How to Manage Intermittency on Islands? EDF Case Studies.” https://
www3.eurelectric.org/media/50462/PONSnew.pdf.
Poojary, R., G. Das, Ghatikar, and S. Saha. 2023. “Open Automated Demand Response:
Industry Value to Indian Utilities and Knowledge from the Deployment.” Unpublished.
Poole, A. 2011. “Implicações do Novo Perfil de Carga do Sistema Elétrico Brasileiro para
a Nova Estrutura Tarifária, Gestão da Demanda e Eficiência Energética.” https://
www.scribd.com/document/56289031/Novo-perfil-de-demanda-e-re-estruturacao
-das-tarifas-eletricas.
PSR. 2020. Demand Response: The Perfect Match to Supply? Energy Report, August 2020.
Rio de Janeiro, Brazil.
PwC (PricewaterhouseCoopers). 2010. Assignment on Implementation & Impact Analysis of
Time of Day (TOD) Tariff in India: Final Report. PricewaterhouseCoopers India Private
Limited. https://forumofregulators.gov.in/data/study/Implementation_Impact_Analysis_
of_Time_of_Day_TOD_tariff_in_India.pdf.
Radovanovic, A. 2020. “Our Data Centers Now Work Harder When the Sun Shines and
Wind Blows.” Google blog post, April 22, 2020. https://blog.google/inside-google/
infrastructure/data-centers-work-harder-sun-shines-wind-blows/.
Rastegar, M., M. Fotuhi-Firuzabad, and H. Zareipour. 2016. “Home Energy Management
Incorporating Operational Priority of Appliances.” International Journal of Electrical Power
& Energy Systems 74 (January): 286–92.
Republic of South Africa. 2008. “Government Gazette. N0 8928. Vol. 517. (31250).” Pretoria,
July 18, 2008. http://www.energy.gov.za/files/policies/Electricity%20Regulations%20on
%20Compulsory%20Norms%20and%20standards%20for%20reticulation%20services
%2018Jul2008.pdf.
Ruff, L. 2002. Economic Principles of Demand Response in Electricity. Washington, DC:
Edison Electric Institute. https://hepg.hks.harvard.edu/files/hepg/files/ruff_economic_
principles_demand_response_eei_10-02.pdf.
Samad, T. E. Koch, and P. Stluka. 2016. Automated Demand Response for Smart Buildings and
Microgrids: The State of the Practice and Research Challenges. Proceedings of the IEEE,
• February. https://www.researchgate.net/publication/295833519_Automated_Demand_
Response_for_Smart_Buildings_and_Microgrids_The_State_of_the_Practice_and_
Research_Challenges.
Santi, Marilyne. 2013. “Who Wants to Join the EDF-Corse ‘Millener’ Project?” https://www
.corsenetinfos.corsica/Qui-veut-adherer-au-projet-Millener-d-EDF-Corse_a1729.html.
Sarkar, Ashok, and Amit Kumar. “Grid Integrated Vehicles (GIV) and Standards for GIVs.”
India Smart Utility Week 2021.
Sasidharan, C., I. Bhand, V. B. Rajah, V. Ganti, S. Sachar, and S. Kumar. 2021. “Roadmap
for Demand Flexibility in India.” White paper, Alliance for an Energy Efficient Economy
(AEEE), New Delhi, June 2022. https://aeee.in/wp-content/uploads/2022/06/Roadmap
-for-Demand-Flexibility-in-India.pdf.
100 BIBLIOGRAPHY
Shakti Sustainable Energy Foundation. 2016. Demand Response in India: Technology
Assessment, M&V Approach, and Framework for DR Implementation. Technical report.
Delhi, India: Shakti Sustainable Energy Foundation.
Sioshansi, F. 2018. “The Sorry State of Demand Response in the U.S.” Energy.post.eu,
January 23, 2018. https://energypost.eu/the-sorry-state-of-demand-response-in-the-u-s/.
Smart Energy International. 2012. “Eskom Sources Demand Response Services from
Comverge.” Smart Energy International, January 3, 2012. https://www.smart-energy
.com/regional-news/africa-middle-east/eskom-sources-demand-response-services
-from-comverge/.
Smart Energy International. 2021. “Tata Power Delhi Launches Residential Demand Response
Programme.” Press release, July 9, 2021. https://www.smart-energy.com/industry-sectors/
energy-grid-management/tata-power-delhi-launches-residential-demand-response
-programme/.
Smart Grids. 2020. “Improving the Insertion of Intermittent Renewable Energies in Island
Areas.” https://www.smartgrids-cre.fr/projets/millener.
Southern California Edison. 2023. “Rate Options for Clean Energy Technology—Go Electric
with a Great Plan.” https://www.sce.com/residential/rates/electric-vehicle-plans.
Steele, P. 2019. “Agile Pricing Explained.” Octopus Energy, February 13, 2019. https://
octopus.energy/blog/agile-pricing-explained/.
Stern, F. 2015. Demand Response in China: The Market & Strategic Positioning of Active Players.
Technical report.
Sweeney, J. 2013. The California Electricity Crisis. Stanford, CA: Hoover Press. https://www
.google.co.in/books/edition/_/3kVmZDEe2_wC?hl=en&sa=X&ved=2ahUKEwi38vGHs_
uGAxVL3zgGHZ8QDkgQre8FegQIBxAE.
Tahir, M. F., H. Chen, A. Khan, M. Sufyan Javed, K. M., Cheema, N. A. Laraik. 2020. “Significance
of Demand Response in Light of Current Pilot Projects in China and Devising a Problem
Solution for Future Advancements.” Technology in Society 63 (November): 101374.
Tata Power. 2012. “Tata Power Demand Response Program.” The Tata Power Company Ltd.
https://bridgetoindia.com/backend/wp-content/uploads/2020/03/Shekhar-Khadilkar
-Tata-Power-20032020.pdf.
US Department of Energy. 2006. Benefits of Demand Response in Electricity Markets and
Recommendations for Achieving Them. Washington, DC: US Department of Energy.
https://eta.lbl.gov/publications/benefits-demand-response-electricity.
Walton, R. 2022. “Texas Expands Industrial Demand Response Program as Grid Goes to the
Brink of Rolling Outages.” Utility Dive, July 15, 2022. https://www.utilitydive.com/news/
texas-expands-industrial-demand-response-program-as-grid-goes-to-the-brink/627358/.
Wang, J., C. N. Bloyd, Z. Hu, and Z. Tan. 2010. “Demand Response in China.” Energy 35 (4):
1592–97.
Wines, M. 2007. “Tolling in the Dark: Africa’s Power Crisis.” The New York Times, July 29, 2007.
https://www.nytimes.com/2007/07/29/world/africa/29power.html#:~:text=In%20normal%20
times%2C%20South%20Africa’s,power%20failures%20have%20persisted%20since.
World Bank. 2004. Demand-Side Management in China’s Restructured Power Industry: How
Regulation and Policy Can Deliver Demand-Side Management Benefits to a Growing Economy
and a Changing Power System. Draft ESMAP Report. Washington, DC: World Bank.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 101
Yin, R., R. Deshmukh, G. Ghatikar, and A. Yin. 2023. “Findings from an Advanced Demand
Response Smart Grid Project to Improve Electricity Reliability in India.” Lawrence
Berkeley National Laboratory. Unpublished title.
Zhou, E., and M. Trieu. 2021. “Electrification Futures Study: Operational Analysis of U.S. Power
Systems with Increased Electrification and Demand-Side Flexibility.” NREL Technical Report
NREL/TP-6A20-79094. Golden, CO: National Renewable Energy Laboratory. https://
www.nrel.gov/docs/fy21osti/79094.pdf.
ZPA Smart Energy. 2016. “Going the Smarter Way.” Ripple Control Systems. http://www.zpa
.cz/en/products-solutions/RCR:c2/.
102 BIBLIOGRAPHY
APPENDIX A
Definitions of Key Ancillary Services
Reserves are backup generation that can be called upon within a certain time frame in
the event the power grid comes under stress (contingency), for example, an anticipated
demand surge or the unexpected loss of a generator or transmission line during a heat
wave.1 Categorization of reserves varies by jurisdiction; response time and duration make
the key difference. Historically, fast-responding reserves have often been called “spinning
reserves,” reflecting traditional hydro or thermal generation synchronized (“spinning”) with
the grid ready to input energy and arrest a frequency drop within minutes. Given their role
in arresting frequency drops, these reserves are also known as “frequency containment
reserves.”
With increasing renewables penetration, dedicated products targeting exceptionally fast
responses (less than a second) and inertia support (to counteract the lower inertia on
grid networks with inverter-based equipment) have been developed in markets like the
Republic of Ireland. These products cause frequency to drop more steeply during an
outage. Batteries have become a popular choice for such services due to their technical
capabilities.
Slower responding reserves (“non-spinning” or “restoration reserves”), which come online
30 minutes or more after instruction, are then used to help free up the fast-responding
units so that they may be ready for the next event.
Other, more bespoke products have also been developed to help support ramp rate concerns,
particularly in relation to a decline in solar photovoltaic output and a rising demand for
evening peak. Grid-forming inverters are one of the solutions being tested. A combination of
synchronous condensers and batteries can help control voltage and frequency by boosting
system capacity.
Frequency regulation, or regulation, refers to generation that can respond automatically
to detected deviations from the frequency at which all generators in a synchronous air-
conditioning system are rotating (In the United States this frequency is 60 hertz, whereas
some other countries use 50 hertz). Regulation is sometimes called “automatic generation
control” because the response is typically too fast for a human being to initiate. Frequency
regulation as an ancillary service corrects for frequency deviations by increasing or
decreasing the output of specific generators, usually by small amounts. The response times
for generators providing regulation are typically in the order of seconds. Frequency
regulation is provided for system operators to ride through unexpected fluctuations in
variable renewables output. Frequency is a systemwide feature.
Voltage support service is essential to stabilize the grid and prevent collapse and cascading
blackouts. Voltage levels should be maintained by balancing active (megawatt-hour, MWh)
and reactive (megavolt-amperes reactive, MVAr) power. Capacitors embedded throughout
the grid provide static reactive power support, and generators, synchronous condensers, or
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 103
dynamic transmission devices provide dynamic reactive support. Reactive supply is typically
not procured through competitive markets. Voltage is a local issue managed at the nodal
level by changing the amount of reactive power.
System operators use network congestion/constraint services to control flows on the
network and ensure they remain within technical limits. Managing flows on particular lines
typically involves a pair of actions: requesting one service provider (generator, storage, or
demand response) to constrain the amount of electricity it is producing or consuming, and a
corresponding request to another provider, in a different locale, to take the opposite action.
Black start capability is necessary for a power system operator to restart the system in
the event of a massive blackout. Black start services must originate from generators that
can start independently and have sufficient real and reactive capability to energize a grid
and restart additional generators.
Endnote
1. It is not clear whether the unanticipated loss of a variable renewable energy for
resource reasons (for example, wind stops blowing) is considered a contingency.
104 APPENDIX A
APPENDIX B
Additional Information on the Case Studies
This appendix contains additional information related to the case studies and examples
presented throughout this report. The focus is on the case studies and examples
considered of greatest potential interest to readers.
Spanish Time-of-Use Residential Tariff
The National Regulatory Authority (Comisión Nacional de los Mercados y la Competencia)
designs tariffs in Spain. The current tariff design for residential customers was introduced
in 2021 and is considered an advanced tariff system built on experience acquired with
time-differentiated tariffs implemented in 2014 for residential customers in Spain.
The 2021 tariff methodology is time differentiated for the capacity (demand) and energy
components and applies to customers with less than 15 kilowatt (kW) installed capacity.
The capacity (demand) charges are two-tiered. The peak capacity charge is from 08:00 to
12:00 on weekdays (Figure B.1). The off-peak charge is more than 95 percent lower than
the previous single charge.
FIGURE B.1
Two-Tiered Capacity (Demand) Charges for Residential Customers
10 11 12 13
9 14
8 15
7 16
6 17
Monday to Friday
5 Working days 18
4 19
3 20
2 21
1 0 23 22 Peak
Valley
Source: CNMC 2022.
Consumption has three tariff brackets (Figure B.2).
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 105
FIGURE B.2
Three-Tiered Consumption Charges for Residential Customers
10 11 12 13
9 14
8 15
7 16
6 17
Monday to Friday
5 Working days 18
4 19
3 20
2 21 Peak
1 0 23 22 Mid-load
Valley
Source: CNMC 2022.
Price differentiation is mandatory for all residential customers. The objectives are to
incentivize all customers to shift load to off-peak periods, enable electric vehicle owners to
charge at lower rates, and help decarbonize electricity consumption.
Customers can review their contracted demands to benefit from the new tariff structure. A
capacity component, which has a fixed value defined on a €/kW per year basis, has been a
part of electricity rates in Spain since the early 1960s. Initially, all households had a fuse,
which tripped when consumption exceeded the contracted value. The electronic meter
performs this function; it trips when consumption remains above the contracted capacity
threshold for longer than 10 minutes.
The final tariff for end users has multiple components (Table B.1). The energy component
is linked to the price variation in the wholesale market. This incorporates an element of
real-time pricing into the tariff structure.
Figure B.3 shows the relative importance of the energy and grid components.
The energy component includes energy itself (traded on the wholesale market in real time),
in addition to capacity fees and all other fees necessary to support the grid and market
operators. The cost of energy in the wholesale market accounts for 85.4 percent.
The grid component has four critical elements. The most important element is the cross-subsidy
to renewable energy, which were contracted at above-market prices (37.4 percent). The second
element is the distribution cost (31.4 percent). The third element is a cost component covering
past revenue shortfalls (16.4 percent). The fourth element is transport costs (10.1 percent).
The impact of the new time-differentiated tariff implemented in 2021 in Spain has not been
evaluated. Doubts about the new tariff design have helped increase consumer awareness that
electricity costs are far from constant. There is some optimism, that with proper information and
motivation, consumption patterns can be modified in response to the new pricing structure.
106 APPENDIX B
TABLE B.1
Cost Components of the Residential Tariff
TRADING MARGIN
Energy costs • Cost of energy in the real-time market
• Cost of ancillary services
• Capacity payments
• System operator
• Market operator
Use of transmission and distribution fees • Transmission costs
• Distribution costs
Fees • Incentive for renewable energy
• Extra costs to serve non-peninsular customers
• Annual shortfalls
• National markets and competition commission
Metering equipment rental
Source: CNMC 2022.
FIGURE B.3
Cost Components of the Residential Tariff
Cost of energy Cost of ancillary
in the real-time services
market 7%
85.4%
Capacity
Payments
Market 7%
System
Operator Operator
0% 1%
Extra costs to serve
non-peninsular
Incentive for customers
renewable energy 4.6%
37.4%
Annual Shortfalls
16.4%
Transport
National 10.1%
Markets and
Competition Others
Commission 0.0%
0.1% Distribution
31.4%
Source: CNMC 2022.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 107
Demand Response in India: Additional
Information on Pilot Programs
Distribution companies have conducted several demand-response pilots; most of them
have been to manage peak demand by shedding or shifting load. Sasidharan and others
(2021) provide an overview of some of the demand-response pilots implemented by
Indian utilities. Table B.2 highlights seven major demand-response pilots that have been
conducted in the last 12 years. Although all the demand-response programs were able
to reduce demand, there are many barriers to scaling up.
TABLE B.2
Seven Major Demand-Response Pilot Projects in India
STATE ELECTRICITY YEAR RATIONALE TYPE OF STRATEGY OF CONSUMER
UTILITY DEMAND DEMAND SEGMENTS
RESPONSE RESPONSE
Maharashtra Tata Power 2012 Peak demand Shed Aggregator-based and Commercial and
Company Ltd— automated demand industrial
Mumbai response
Delhi Tata Power Delhi 2014 Peak demand, Shed Automated demand Commercial and
Distribution Limited grid stress response industrial
Rajasthan Jaipur Vidyut Vitaran 2013–14 Deviation from Shed Manual demand Commercial and
Nigam Ltd schedule response with energy industrial
market integration
Delhi BSES Yamuna Power 2017 Deviation from Shed, shift Behavioral demand Residential
Limited schedule response
Delhi BSES Rajdhani Power 2018–19 Peak demand Shed, shift Behavioral demand Residential
Limited response
Uttar Uttar Pradesh Power 2019 Peak demand Shed Manual demand Commercial and
Pradesh Corporation Limited response industrial
Delhi BSES Yamuna Power 2020 Peak demand Shed Automated demand Residential and
Limited response commercial
Delhi Tata Power Delhi 2021 Peak demand Shed, shift Behavioral demand Residential
Distribution Ltd response
Source: Sasidharan and others 2021.
Some salient features of the demand-response initiatives undertaken by distribution
companies in different states are described below.
BSES Yamuna Power Limited implemented a manual demand response pilot program and an
automated demand response pilot program in 2017 and 2020, respectively. Commercial and
industrial customers participated in the manual demand-response pilot program. Their load
was curtailed for one hour during the summer months. The manual demand-response pilot
108 APPENDIX B
program compensated the customers and successfully achieved a cumulative load reduction of
17 megawatts (MW) from the 19 participating customers. The automated demand-response pilot
demonstrated the feasibility of controlling the air-conditioning load during summer months
through remote control (cloud based). It achieved a 30 percent reduction in consumption.
Tata Power Delhi Distribution Limited has also implemented pilot demand-response
programs. It implemented India’s first smart meter–based pilot program for peak demand
and grid stress management using an automated demand response. The pilot relied on
real-time communication to provide information on the load to the utility and consumers,
increasing transparency. The utility also conducted a pilot of behavioral demand response.
It was targeted at residential customers and was meant to demonstrate the potential
savings to the utility and to customers.
With peak demand being four times as much as off-peak demand, the focus was to shave
peak demand. Schemes planned to be implemented in FY 2021–22 included energy audits
on cold storage for industrial consumers, the replacement of nonefficient commercial
chillers, and the replacement of hot water geysers.
BSES Rajdhani Power Limited (BRPL) designed an automated demand-response program,
which included onboarding customers via a customizable web-based platform. The
distribution utility used the platform to publish events and send notifications to the
participating consumers. If customers approved it, the utility could operate the participating
load as per the scheduled events. The program’s first phase included peak shaving (turning
air-conditioning loads on/off, for example). The second phase (load shifting) tried to shift
water heating loads from peak to off-peak hours. The load reduction was about 1 MW.
BRPL has plans to implement a large-scale demand-response program.
At Jaipur Vidyut Vitran Nigam Limited, a demand-response pilot was designed to reduce the
deviation penalty incurred by the utility when procuring energy from the national grid. The
pilot was designed as a demand-bidding-to-the-market mechanism. Industrial consumers
could submit demand bids directly to the energy market. Seventeen commercial consumers
who had enrolled voluntarily participated in the pilot program. Customers were notified four
hours before a demand-response event. After the event, the software was used to measure
and verify the participants’ load curtailment. Four events resulted in an average of 22 MW in
demand response, proving a cost-effective solution. The scale-up of such projects will require
dynamic pricing programs and advanced metering infrastructure.
Demand Response in South Africa: Additional
Information on ESKOM Programs
Besides the load control programs described above, Eskom has designed some
mechanisms for large customers, as described below.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 109
Supplemental demand-response programs target industrial consumers with a minimum
load entry level of 500 kW or 15 percent of the average load. The load provider should
guarantee up to 300 reduction hours per year, with at least one hour per event per day.
The reduction is requested 30 minutes in advance. The load provider receives a capacity
payment and an energy payment for energy not consumed during the reduction period.
Figure B.4 shows the maximum certified capacity and the average monthly reduction
provided last year.
FIGURE B.4
Performance of Demand Enrolled in the Supplemental Demand-Response Program
600
500
400
MW
300
200
100
0
Aug-20 Sep-20 Oct-20 Nov-20 Dec-20 Jan-21 Feb-21 Mar-21 Apr-21 May-21 Jun-21 Jul-21
Maximum Certified Capacity (MW) Average Dipatched by National Control (MW)
Average Provided (MW)
Source: Eskom 2021.
Note: MW = megawatt.
Instantaneous demand-response programs target industrial consumers with fast-response
capabilities and a minimum load entry level of 10 MW. The instantaneous demand-response
load provider should guarantee up to 200 fewer event hours per year and at least three
10-minute reductions per event per day. DR must respond within six seconds after the shed
signal. The load provider receives a capacity payment based on the median performance for
a month.
Non-dispatchable demand-response programs target industrial consumers that cannot
reduce their load on short notice. The minimum load entry level for these consumers is
2 MW or 15 percent of the average plant load. Consumers can opt to participate with a
110 APPENDIX B
one- to four-hour demand reduction, which will always be requested between 16:00 and
20:00. The reduction request is made two to four weeks in advance, and the consumer
receives an energy payment that varies inversely as the reduction request time that it has
selected. This program was discontinued.
A power alert initiative is meant to engage customers in managing imminent power crises,
but it has no associated remuneration. Power alerts inform the population in general
about the status of the power sector and are triggered when an impending supply-demand
imbalance may lead to load shedding. The status level is indicated by a color code: green
indicates a stable situation, yellow significant strain on the network, red a severe stress
event, and black a critical condition already causing load shedding. Figure B.5 shows the
average demand reduction obtained through red and black power alerts in 2020–21.
FIGURE B.5
Demand Reductions Achieved through Power Alerts (red and black status levels),
2020–21
500
450
400
350
MW reduction
300
250
200
150
100
50
0
ay
ne
y
t
r
er
r
r
y
y
ch
us
be
be
be
l
ar
r
Ju
ua
ob
M
ar
Ju
g
nu
em
em
em
Au
M
br
ct
Ja
O
pt
ov
ec
Fe
Se
D
N
Source: Eskom 2021.
Note: MW = megawatt.
The average megawatt reduction achieved through power alerts is comparable to the
certified and dispatched capacity under the supplemental demand-response program.
Power alerts reach a much larger customer base and require a concentrated effort by
customers to respond under the imminence of a blackout. Therefore, since no incentives
are provided, power alerts are meant to be used only during emergencies. If used too
often, they will likely lead to customer fatigue, making the program less effective.
Besides traditional price- and quantity-based demand-response mechanisms, South Africa
used efficient market-based demand-response business models to manage power
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 111
shortages during the 2008 dual energy and capacity crisis. A 1,500 MW peak demand
reduction was achieved through programs and interventions involving mainly industrial
customers, particularly mines and smelters (ESMAP 2011).1 Eskom was instrumental in
designing and implementing these programs.
Demand Response in the United States:
Additional Information
TABLE B.3
Results of Peak-Time Rebate in Maryland
YEAR ENERGY ELIGIBLE AVERAGE CREDIT ON PEAK DEMAND PARTICIPATION
SAVINGS DAYS CUSTOMERS ELECTRICITY BILL REDUCTION (%)
(N) (N) ($) MW ($, millions)
2013 4 315,000 9.03 96 7.0 82
2014 2 860,000 6.55 209 5.6 76
2015 4 1,020,000 6.67 309 15.5 81
2016 3 1,074,000 6.73 336 11.0 71
2017 2 1,095,000 6.13 330 6.1 74
Source: Studies by EPE (2018).
Note: MW = megawatt.
Figure B.6 presents the 2015 demand-response capacity available to the three largest
utilities in California: Pacific Gas and Electric, Southern California Edison, and San Diego
Gas and Electric, totaling about 2.1 GW.
The most significant contribution came from interruptible tariffs (~1 GW), followed by
residential-level air-conditioning load control, with a 0.4 GW contribution. Non-residential-
level aggregator programs made a 0.35 GW contribution, followed by demand-side bidding
with a contribution of about 0.14 GW.
112 APPENDIX B
FIGURE B.6
Demand-Side Resources in California
2,400
2,200 2147 MW
2,000
2015 1-in-2 Load Reduction Capacity (MW)
1,800
1,600
1457 MW
1,400
1,200
Residential - Air conditioner load control
1,000
Residential - Critical Peak Pricing
800 Residential - Peak time rebates
609 MW Non-Res - Aggregator programs
600 Non-Res - Agricultural pump control
Non-Res - Air conditioner load control
400 Non-Res - Critical Peak Pricing
Non-Res - Demand bidding
200 84 MW Non-Res - Interruptible rates
Non-Res - Real Time Pricing
0
PG&E SCE SD G&E Total
Source: Monthly reports from utilities on interruptible load and demand-response programs filed with
California Public Utilities Commission (A.11-03-001).
Note: MW = megawatt; PG&E = Pacific Gas and Electric; SCE = Southern California Edison; SDG&E = San Diego
Gas & Electric.
Endnote
1. This was predominantly a power buy-back scheme under which large industrial
customers with furnace loads were paid to switch off their plants.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 113
APPENDIX C
Possibilities for Load Control
Some typical loads can be controlled to benefit the power system.
Utilities may control air-conditioning devices during critical periods when there is risk of an
outage. In most places, this typically occurs on hot summer days, when demand strains the
power system. The utility remotely cycles air-conditioning devices in a specific region or
neighborhood—preventing a large number of units operating simultaneously; the load
factor is in turn reduced, lessening coincident peak demand in distribution feeders and
creating virtual capacity for the system. Some load control programs in the United States
compensate customers with a fixed payment (rebate), $40–$120 per year,1 depending on
the region and the utility’s level of control.
At the commercial level, there are other opportunities to explore for more efficient and
more responsive cooling: storing ice in large offices and commercial buildings; heating,
ventilation, and air-conditioning may be controlled automatically, providing a fast demand
response; ice could be stored during off-peak hours and used during peak hours to feed
the heating, ventilation, and air-conditioning systems. Another possibility is district cooling,
chilling water at a centralized location, to distribute to nearby buildings. Chilled water
(or ice) can be stored to support consumption for several hours. District cooling is an
interesting alternative in some niche markets. Its adoption has been modest, but it has
gained momentum.2
Water heaters can also be cycled. They do not play an essential role in driving the peak
demand in the United States, except in the winter in colder climates. Still, load control for
water heaters (with storage capacity) is common in other countries. A demand response
can be initiated rapidly, providing ancillary services if needed.
Crypto miners are a large and growing group of energy users worldwide. For example, the
value of Bitcoin and equivalent currencies is increasing demand in several parts of the
world. Places with cheap energy are attracting crypto miners. This demand growth has
required some US utilities to review their energy and demand pricing strategies to recover
incremental energy and investment costs.
Crypto mining is potentially a reliable demand-response resource, provided the mining
activities shed load when instructed. However, the model has failed in the United States
because most crypto miners do not shed loads during peak hours, mainly when the
cryptocurrency’s value is high. In some states, such as Georgia (with the largest crypto
mining operation in the United States), miners operate almost like a giant baseload. Better
regulatory and demand-response mechanisms must be devised to manage the challenges
posed by crypto mining and the opportunities that it provides.
Data centers are becoming large energy users, and they could potentially participate in
demand-response programs, although some barriers exist. Reliability is crucial; data
114 APPENDIX C
centers require continuous power and invest in batteries and emergency generation to
ensure service continuity. Technically speaking, data centers could operate as islands,
reducing consumption from the grid, but there are practical considerations. Switching
to backup entails some technical risks. Expanding a battery energy storage system is
also costly. The processing load of data centers could be distributed across several
geographically dispersed centers. Still, cloud migration is not easy, and the benefits of
demand participation are not attractive enough to compensate for the complexity and
possible disruption in processing capacity.
Despite those difficulties, leading-edge data centers like Google’s participate in demand-
response programs. When a system operator informs Google of a forecasted grid event
that can cause a supply constraint, this information is conveyed to a global computing
planning system, which generates hour-by-hour instructions to limit nonurgent tasks for
that event’s duration. Strong interconnectivity among computer centers worldwide enables
companies like Google to shift processing tasks and reduce local consumption. When
feasible, some tasks may be rerouted to other data centers.
Endnotes
1. Cycling may cause some minor discomfort, which can be ameliorated by reducing
the temperature set point by 1–2 degrees. Some more sophisticated utility protocols
precool and change the temperature setting automatically via smart thermostats.
2. District cooling is standard in Europe, where the district heating infrastructure is
sometimes shared. This is also under trial in India. It is seldom used in Latin America.
A survey developed by the World Bank in 2017 revealed three (public) district cooling
facilities in Latin America and the Caribbean: Panama City, Rio de Janeiro, and
Medellin. The first facility was designed to provide air-conditioning for the Panama
Canal administrative buildings. A utility-owned energy service company developed
the second facility, and EPM, a local utility company, created the third facility to serve
a cluster of adjoining public buildings. The Lusail City district cooling system that
Marafeq, a utility company in Qatar, developed will use electric chillers and thermal
energy storage to supply chilled water. It is estimated that storage will enable about
1,000 gigawatt-hours in savings annually and avoid the need for some 35 MW of
capacity when compared with conventional air-conditioning systems using air-cooled
chillers or split air-conditioning.
HARNESSING THE POTENTIAL OF FLEXIBLE DEMAND RESPONSE IN EMERGING MARKETS 115
Photo Credits
page xii | © SolStock / E+ / Getty Images
page 8 | © AvailableLight / iStock / Getty Images Plus
page 26 | © PonyWang / E+ / Getty Images
page 38 | © Tunvarat Pruksachat / Moment / Getty Images
page 47 | © cdwheatley / E+ / Getty Images
page 48 | © sinology / Moment / Getty Images
page 74 | © pipat wongsawang / Moment / Getty Images
page 87 | © Simonkr / E+ / Getty Images
page 88 | © Richard Newstead / Moment / Getty Images