IEA TASK XIII: Demand Response Resources 1st Meeting Economic Working Group Meeting

1
IEA TASK XIIIDemand Response Resources1st
Meeting -- Economic Working Group Meeting

• Host
• Mikael TogebyElkraft System, Denmark
• Daniel Violette, Ph.D. Mr. Pete ScarpelliSummit
  Blue Consulting RETX, Inc.Boulder,
  Colorado Chicago, Illinois
• Ph. 720-564-1130 Ph 312-559-0756
• E-mail dviolette_at_summitblue.com pscarpelli_at_retx.
  com
• August 17-18, 2004

2
AgendaDay 1

• --MORNING--
• 1. Introductions
• 2. Background on overall IEA Annex
• 3. Review goals and objectives for economic
  working group
• 4. Defining Benefits and Costs
• 5. DR benefit-cost analysis frameworks
• 6. DRR in forward-looking planning analyses
• -- AFTERNOON --
• 7. Critical success factors for DR assessments
• 8. Discussion (Methods, applications, models,
  estimates, institutions)

3
AgendaDay 2

• 1. Summarize results from day 1
• 2. DRR cost-benefit and planning methods
• Questions that need to be addressed
• Gap analysis, i.e., where does new work need to
  be done
• 3. Develop research and action agenda
• 4. Outline of Final Report
• 5. Action Agenda

4
Demand Response Definition
Demand response is the ability of electricity
demand to respond to variations in electricity
prices in 'markets' or in 'real' time.
Inclusion of DRR in energy markets can take the
form of reduced energy costs, direct payments for
energy not consumed, and/or a reservation
payment for being available to reduce consumption
upon request.
5
IEA TASK XIII DRR

• Define and build turnkey DRR infrastructure model
  including Business Model, Business Rules,
  Enabling Technology, Standards and Implementation
  Plan.
• Deliver DRR into Any emerging or existing
  liberalized electricity market.

6
Project Participants

• Australia
• Canada
• Denmark
• Finland
• Italy
• Japan

• Korea
• Netherlands
• Norway
• Spain
• Sweden
• USA

Pending Mexico, India, New Zealand
7
IEA Task XIII Chronology

• February 2003 PLMA invited to present at IEA
  Demand Response workshop
• March 2003 IEA Secretariat attends Spring PLMA
  Meeting in Washington D.C.
• April 2003 DOE agrees to have USA sponsor DRR
  project if approved by IEA DSM ExCo
• April 2003 IEA DSM ExCo approves DRR Project in
  Concept
• May 2003 DRR Project Working Group formed

8
Task XIII Chronology (cont)

• June 2003 IEA/PLMA Press Releases announcing
  project and September meetings
• July 2003 PLMA asked to participate in DOE
  Planning for demand response
• September 2003 IEA/PLMA International
  Seminar/Experts Workshop
• April 2004 DRR Project Approved by ExCo
• May 2004 DRR Project work begins

9
Activities Timeline
10
Task 2 Market Comparisons

• Goals
• Identify how DR is being used
• Highlight common utilizations and unique efforts
• Report on State of the Practice
• Methodology
• Marketplace Overview
• Independent Research
• Data Request of Experts (released soon)
• Establish State of Practice Work Group
  (technology usage customer marketing efforts)
• Work Products
• Country Comparison Charts
• Program Benchmarking Charts
• State of Practice Case Studies

11
Additional Efforts

• Task XIII Project Portal
• www.demandresponseresources.com
• Project Information
• Work group workplace
• DR Research Library
• Operational Issues Work Group
• Initiated at September Experts Meeting
• Focused on DR market design, data management,
  operations, etc.
• Project Communication
• Articles
• Conferences
• Canada Power Conference, Toronto, Canada
• Metering Europe, Berlin, Germany
• Critical Infrastructure, Grenoble, France
• NARUC, United States

12
Economic Working Group

• The initial draft memo indicated four areas were
  targeted for work by the Economics Working Group
• 1. Appropriate Benefit/Cost Frameworks for DRR.
• 2. Valuation of DRR in Markets and Incorporating
  DRR in Resource Planning.
• 3. Conducting DRR Economic Potential Studies
• 4. Ex-post Evaluation of DRR Programs and/or
  Policies.
• Recognize that we are near the beginning -- your
  input now influences the entire project.
• OBJECTIVES FOR THIS MEETING
• 1. ARE WE ASKING THE RIGHT QUESTIONS?
• 2. ARE WE FOCUSED ON THE RIGHT END-PRODUCTS?

13
Work Area 1 -- B/C Analyses

• Work Area 1 Benefit/Cost Framework -- Develop
  the Benefit-Cost Framework that appropriately
  supports the economic case for DRR as part of a
  resource plan.
• Product This work effort will produce
• A listing of benefits that appropriately credit
  DRR for the value it provides, and a list of
  costs that fully account for the costs of DRR
  programs and
• For each identified benefit and cost, methods and
  guidelines (as practicable) will be identified
  that can be used to estimate each that benefit
  and cost.
• Note The view is that, by doing this
  Benefit/Cost effort first, we are providing a
  framework for the other work areas.

14
Work Area 2 DRR in Planning

• Work Area 2 Valuation and Planning -- Develop
  "approaches" (not specific models) to incorporate
  DRR in forward planning short-term (1 to 2
  years), medium-term (3 to 6 years), and long-term
  plans (10 years).
• Goal Develop planning structures and approaches
  that identify and explain the tools for
  incorporating DRR in longer-term market and
  resource planning.
• Product Report on approaches and guidelines for
  assessing DRR in resource planning.
• Identify existing tools and models that are
  available in member countries (e.g., capacity
  expansion planning models, resource planning
  models, and other useful tools/models, e.g.,
  distributed generation planning tools).
• Develop guidelines for DRR valuation and
  planning.
• Develop examples and case-studies as practicable.

15
Work Area 3 -- DRR Potential

• Work Area 3 DRR Potential -- estimation of DRR
  potential in a market.
• Focus on structuring approaches and developing
  appropriate guidelines for their application.
• Should we begin with work on economic potential
  of energy efficiency programs and for estimating
  the economic potential of distributed generation
  will serve as useful starting points.
• This effort will be linked with the DRR program
  design efforts and "lessons learned" from Task 2.
• Product Produce templates for assessing the
  potential contribution of DRR in different
  markets and for different types of DRR programs.

16
Work Area 4 -- Evaluation of DRR

• Work Area 4 Ex-Post Evaluation of DRR Discuss
  and develop approaches and requirements for
  evaluating and verifying the benefits and costs
  of DRR for specific programs that are in place.
• Focus on assessing the economic value attributed
  to DRR as part of these ex post evaluations
• Do we need to take into account the longer-term
  impacts from DRR that might represent important
  components of the total benefits (i.e., develop
  an annual average).
• Product Produce guidelines to approaches for
  the ex-post evaluation of DRR for different types
  of DRR and different market circumstances.

17
Overview Issues with DR Valuation

• 1. Appropriately capturing all the values
  associated with a DRR.
• Many values associated with DRR are difficult to
  quantify,
• But, they are growing in importance as
  supply-side resources become more constrained
  (e.g., transmission congestion, and natural gas
  prices)
• 2. Need to dimension uncertainty around future
  outcomes.
• Simple planning paradigms such as 1 in 10 year
  events are not very useful in assessing option
  and hedge values as they only represent one
  point.
• Different approaches are needed for dimensioning
  uncertainty if new tools are to be useful.
• 3. Categorization -- There are many types of DRR
  programs -- how to classify them such that the
  guidelines are appropriate.

18

• -- Section 1 --
• Identification of Benefits and Costs
• What should be addressed?

19
DRR -- Market Benefits?

• 1. Reliability -- Increased system reliability
  through investments at load centers, i.e., the
  locational value of the resource.
• 2. Market Power -- Demand reductions curb market
  power and supply-side reliance
• 3. Market Price Reductions -- Reduced regional
  prices (part private and/or market?)
• 4. Efficient markets -- Better pricing and the
  interaction of demand and supply can produce
  technology and overall productivity gains (e.g.,
  1 per year).
• 5. Insurance Value -- Creates the ability to
  lower/minimize costs of low probability high
  consequence events given current infrastructure
  (looking 1 to 2 years out).
• 6. Option Value -- Creates more future planning
  options, e.g., lower demand growth allows for
  more time to assess new infrastructure options
  and adapt to new or changed circumstances (makes
  gradual changes more economic).

20
DRR -- Market Benefits?

• 7. Reduced hedging costs -- Lowered average
  prices and price volatility creates a forward
  price curve that lowers the costs of hedging.
• 8. Risk management benefits -- By allowing
  customers to manage part of the price and
  commodity risks according to their risk
  preferences.
• 9. Portfolio benefits -- DRR provides for
  increased diversity resources over time (are we
  double counting with other benefits?)
• 10. Environmental benefits -- By promoting
  efficient use of resources.
• 11. Customer services -- Through increased
  comfort, customer choice and reward for energy
  management -- Non-Energy Benefits.
• 12. Other -- What is being missed? Where does
  the avoided costs of a combustion turbine fit in?
  Assumes competitive market in generation so no
  need to specify avoid turbine costs?

21
DRR -- Private Entity Benefits

• 1. Specialty DRR providers (in the U.S., they are
  called "aggregators" or "curtailment service
  providers")
• Payments received for providing DRR.
• 2. Distribution Companies
• Lowered distribution OM.
• Lowered capital costs for distribution.
• Payments from others for implementing DRR
• 3. Transmission Company Benefits
• Lowered TD OM costs
• Deferred capital costs

22
DRR -- Private Entity Benefits

• 4. Commodity providers (i.e., electricity
  suppliers to retail customers or retail
  aggregators)
• Lowered costs of purchasing wholesale electricity
  (but what is the impact on margins -- i.e., do
  they really benefit?)
• 5. Reliability Entities (i.e., ISOs or power
  pools)
• They are non-profit, so how do they benefit --
  are they just facilitators?
• 6. End-use Customers
• Lower retail prices for electricity
• Increased reliability
• Payments for providing DRR
• 7. What private benefits are being missed?

23
DRR -- Dimensioning the Costs

• Private Costs (for anyone who runs a DRR
  program)
• Costs of DRR set-up (one-time expenditures)
• Marketing and program design
• Equipment and software
• On-going operating costs
• Payments to participants (capacity and/or energy)
• Overhead management
• Market Costs
• Are there any?
• Miss-allocation of resources?
• Lost profits to generators? (We don't make up
  lost profits to firms when a more efficient
  option comes along so, why do it here?)

24

• -- Section 2 --
• Examples of Benefit Cost Studies

25
Side Line -- Dimensioning Uncertainty in DR
Valuation

• Expressing and dimensioning uncertainty for use
  in analyses.
• Uncertainty is what makes hedges and options
  valuable.
• If we could use point estimates and were certain
  about their values, there is no need for options
  or hedges since the optimal solution would simply
  be picked.
• Industry has used few tools to express
  uncertainty
• Key problem -- How to dimension uncertainty for
  use in planning analyses (simplest to more
  complex)
• 1. Scenario analyses
• 2. Range estimates -- construct confidence
  intervals based on key inputs.
• 3. Range estimates with the range filled in with
  likelihood estimates to provide a rough-cut
  probability distribution.

26
Scenarios Versus Distributions

• An assessment about likelihoods of the different
  scenarios can provide additional, useful
  information.
• Individuals familiar with the market can supply
  the best available information on
  probabilities.
• Derived from judgment, expert opinion and
  augmented by secondary research.
• End-result A distribution is a better
  representation of the scenarios being assessed

27
DRR Benefit/Costs based on Energy Efficiency (EE)
Frameworks

• A common US DSM framework is from the California
  (CA) Standard Practice Manual for Economic
  Analysis of DSM or EE Programs.
• The CA approach defines five stakeholder
  benefit-cost tests
• 1. Participant test,
• 2. Utility test
• 3. Rate impact (or non-participant) test,
• 4. Total Resource Cost (TRC) which is most
  commonly used and
• 5. Societal test (includes externalities).
• BUT, and approach suitable for EE may not be
  appropriate for DRR.
• However, it is widely used in the US in states
  with active DSM programs and that also have DRR
  programs.
• Possible Justification -- Could this set a lower
  bound? If DRR passes the TRC test then, it
  would certainly pass a more appropriate market
  test designed for DRR?

28
CA Standard Practice Manual

• Widely adopted in the US for DSM benefit cost
  analysis.
• First published in 1983, revised in 1988, and
  again in 2001.
• Manual covers conservation, load management, fuel
  switching, and load building programs.
• But all are permanent reductions programs and are
  not dispatchable in response to market factors or
  events.
• QUESTION -- Can we really compare benefits of
  dispatchable programs in this static benefit-cost
  framework? Initial answer is no.
• All stakeholder tests focus on net present
  values (NPVs) over the lifetime of DSM measures.

29
Participant Test

• Evaluates whether a DSM program/measure is cost
  effective to program participants.
• Compares the DSM measure cost after utility
  rebates to the net present value benefits of the
  energy savings over the life of the measure.
• This is one of the most important tests in the
  US, since regulators would not want to encourage
  customers to install DSM measures that are not
  cost effective to them.
• Test results often have B/C ratios gt 2.

30
Utility Test

• Evaluates whether a DSM measure/program is cost
  effective to the utility.
• This test compares the present value of the
  avoided cost benefits over the DSM measures
  lifetimes to the total DSM program costs.
• This test result is usually very positive, with
  B/C ratios gt 3, and often above 10.

31
Rate Impact Test

• Measures programs impacts on electric or gas
  rates. Used to be called the non-participant
  test.
• It compares the avoided cost benefits from a DSM
  program to the sum of the program costs and lost
  revenues.
• This is similar to economics Pareto efficiency
  test a policy which makes everyone better off.
• This test is often slightly negative for
  conservation programs, with B/C ratios of
  0.8-0.9,
• And slightly positive for DR programs, with B/C
  ratios of 1-2.

32
Total Resource Cost (TRC) Test

• Evaluates whether the benefits from the DSM
  resource are greater than the costs of such the
  sum of the program costs and the DSM measure
  costs.
• Also evaluates whether a program is cost
  effective to the utilitys ratepayers as a total
  group.
• This test varies depending on the costs and
  benefits of the programs.
• Usually positive for conservation programs, with
  B/C ratios over 1 even taking into account net
  benefits (i.e., net of free-riders)
• For DR programs, with B/C ratios depend on the
  avoided supply costs and can be less than 1 or
  greater than 1 accordingly.
• Key is how the avoided costs for meeting peak
  demand from other alternatives (usually a
  combustion turbine) are calculated (Is there an
  opportunity to develop guidelines for calculating
  avoided costs?)

33
Societal Test

• Very similar to TRC test.
• Main difference is adding avoided environmental
  externalities (avoided pollution costs) as a
  program benefit.
• CA also uses "other" DSM benefits in the societal
  test such as non-energy benefits, reliability
  benefits, fuel diversity.
• This is similar to the Kaldor-Hicks compensation
  principal in economics
• Winners from a DSM set of programs could
  compensate the losers with some of their benefits
  so that everyone would be better off.
• In the US, the societal test results are often
  very similar to the TRC test results but it all
  depends upon the externality values that are
  added to the TRC test.

34
Case Study 1 -- Xcel Energy's use of B/C Tests
(Utility in Minnesota, U.S.)

• Mass-Market Program Xcel Energy cycles
  participant customers central air conditioners
  and electric water heaters
• 15 minutes on/off during peak periods.
• Uses radio signals and pagers for cycling.
• Offers customers a 15 summer rate discount for
  participating in program.
• Program has been operating for 14 years, and over
  20 of residential customers participate.
  Program impacts are about 0.7 kW per residential
  customer.

35
Xcel ProgramBenefit-Cost Results

• Analysis done for 15 year periodlifetime of a
  cycling device.
• Key benefits and costs are
• Total Program Costs -- 5,453,902
• Number of Participants -- 35,100
• Energy Savings at the Generator -- 510,060 kWh
• Demand Savings at the Generator -- 25,105 kW
• Avoided costs
• Per KW 217.20
• Per kWh 0.713

36
Benefit-Cost Results (cont.)

• Participant test B/C ratio is infinite
• customers have no direct costs to participate in
  program.
• Customer surveys show they experience some minor
  discomfort during control periods, but costs for
  that are not estimated.
• Utility B/C ratio is 5.5avoided generation, TD,
  and energy costs much larger than program costs.
  Rate discounts not included in this B/C ratio.
• Rate impact B/C ratio is 0.86rate discounts
  about equal avoided costs, and program costs push
  ratio below 1.
• TRC and societal B/C ratios are also 5.5, i.e.,
  the same as the utility test ratios, since there
  are no DSM measure costs that customers have to
  pay directly.

37
Uncertainties in Xcel Energys DSM Benefit Cost
Analysis

• Retrospective evaluation where actual program
  impacts are compared to estimated program
  impacts.
• Total energy and demand savings per program.
• Load shapes of actual savings.
• Free ridership/free drivership
• Avoided costs and lost revenues 20 year
  forecasts are inherently uncertain.
• These forecasts do not include low probability
  but high cost events.
• Environmental externalities and other non-energy
  benefits. Methods are not well agreed upon.

38
Benefit-Cost based on LOLP

• LOLP stands for loss of load probability and is
  the focus of most reliability-based
  organizations.
• This approach has been used by the NY ISO, and
  the ISO NE.
• Looks at how DRR has affected the probability of
  an outage.
• To date, these evaluations have been
  retrospective and focused on whether benefits
  attained to date exceed the program costs to
  date.

39
CASE STUDY 2 -- NY ISO Emergency Demand Response
Program (EDRP)

• Assess the benefits of EDRP by looking at how an
  increase in reserves would reduce the loss of
  load probability (LOLP)
• A measure of the benefits of EDRP can be defined
  by the change in the Value of Expected Un-served
  Energy (VEUE) as follows
• ?VEUE (Change in LOLP) (Outage Cost/MW)
  (Un-Served Load in MW)
• By calling EDRP, the load reduction works to
  restore reserve margins
• The extent to which reserve margins are
  completely restored is a function of the amount
  of load reduction of onsite generation provided
  by EDRP participants

40
NY ISO Emergency Demand Response Program (EDRP)
(cont.)

• If deployment of EDRP resources results in a
  positive change in VEUE, then that benefit
  qualifies as a contribution to system security
• Under the most conservative assumptions -- Outage
  costs at 1000/MWh and a reduction in LOLP of
  0.05
• THEN, only 3.6 of the load would have had to be
  at risk in order for the benefits in terms of
  VEUE to exceed program costs as implemented by
  the NY ISO.
• At the other extreme (5000/MWh and 0.50 LOLP),
  only 0.1 of the load would have to be at risk
  for the program benefits to equal program costs
• This B/C approach sets boundary conditions which
  must be true for the program to be cost
  effective, even though those conditions are very
  uncertain.

41

• -- Section 3 --Retrospective Evaluations versus
  Forward-looking Planning Studies
• Case studies 1 and 2 were retrospective
  evaluations designed to determine if a DRR
  program's benefits exceeded its costs for the
  past year.
• Economic planning analyses are needed to
  determine future investment in DRR versus
  supply-side investments.
• How should we do future planning?

42
DISCO Utility Assessment of DRR

• Reference -- Report filed with Massachusetts
  State Regulators.
• Two DISCO utility programs with both focused on
  reliability.
• 1. A large customer call-option program for 160
  MW (in year 1)
• 2. A mass market AC load control program that
  provides 10 MW increment per year for five years
  resulting in a total of 50 MW in year 5.
• Reliability was the focus since the DISCO wanted
  to see if these programs could be justified based
  on DISCO cost reductions.
• To defer OM and capital expenses, the utility
  believed it would have to make reductions
  mandatory.
• As a result, pricing options were not considered.

43
Conceptual Issue Overview

• Should the DISCO take a more proactive approach
  to DR in the future and seek to acquire 170 to
  210 MW of DR?
• ISSUE There is a belief that DR market benefits
  are large, but when DISCO benefits are examined
  they are found to be much smaller than the
  perceived market benefits.
• The Discussion Scenario -- An NSTAR DR investment
  producing 210 MW from a CI call program and a
  mass-market DLC program
• Initial estimate of 80M in market-wide benefits
  over 5-year horizon with market-wide B/C ratio of
  3.4.
• But, from an DISCO perspective, this investment
  provides 7.7M in benefits and 23.8M in costs
  over 5-year horizon with NSTAR B/C ratio of only
  0.3.
• DISCO customer perspective -- Customers receive
  at least 20 of the market benefits or 16.4M
  plus incentive payments of 16M from program
  participation NPV discounted over five years for
  a customer B/C of 1.4.

44
Setting the Stage Views on Market Benefits of
DR

• Overall, there is a basic belief among many
  organizations that market benefits from DR are
  sizeable
• ISO-NE Regional Transmission Expansion Plan
  states that DR can have significant benefits in
  terms of reliability and savings in congestion
  costs.
• New England Demand Response Initiatives Final
  Draft Report states that a small amount of DR can
  enhance system reliability and substantially
  reduce market-clearing prices, producing
  significant benefits to consumers.
• ISO-NE 2002 DR Program Evaluation states that
  magnitudes of DR sufficient to clear the market
  at lower bid prices (ECP), will reduce the price
  of energy for all purchasers in the spot market.
• The NYISO states that it has had a successful DR
  program in operation through two summers which
  has delivered benefits to the grid in terms of
  reduced market price and improved system
  reliability.

45
Setting the Stage (cont.)

• FERC SMD and White Paper
• Demand response is essential in competitive
  markets to assure the efficient interaction of
  supply and demand.
• Demand response options should be available so
  that end users can respond to price signals.
• California PUC -- Demand Response is a vital
  resource to enhance electric system reliability,
  reduce power purchase cost and individual
  consumer costs R. 02-06-001, Order Instituting
  Rule making, June 6, 2002.
• California Energy Commission 2002 2012
  Electricity Outlook Report estimates that an
  increased level of DR could have saved California
  2.5 billion in year 2000.
• FINALLY -- California Energy Commission Order
  Instituting Rulemaking (June 17, 2003) states
  that the CEC will consider the acquisition of
  2,500 MW of DR (approx. 5 of peak demand) to
  moderate price increases and improve system
  reliability.

46
Setting the Stage (cont.)
August 14 2002 Hour 15 Supply Curve (ISO-NE)
Peak load reduced by 5

• A 5 reduction in peak load can generate a 50
  reduction in ECP.
• DR can reduce the price of energy for all
  purchasers in the spot market.
• DR can play a significant role in price spike
  mitigation in lieu of price caps.
• Source ISO-NE 2002 DR Program Evaluation.

47
DISCO Benefits

• Defer or eliminate TD capital expenditures
• Provide (n-1) reserve
• Standby generation
• Curtailable load
• Emergency Response Resource
• Hedge against questionable forecasts
• Extreme weather events
• Unforeseen load growth
• Rapid development
• Assist during contingency/emergency events
• Revenues from ISO-NE DRR capacity and energy
  payments to any DRR provider.

48
NSTAR DR Program Costs

• Proposed Program Cost Categories
• Staff time (development, marketing,
  implementation)
• Software systems
• Physical infrastructure
• Incentive payments
• Financial assistance to participants
• Revenue differentials due to rate impacts

49
DISCO Benefits Costs

• An interactive process involving NSTAR staff and
  enabling technology manufacturers and vendors was
  used to calculate the benefits and costs
• Distribution System Planning
• Regulatory Policy and Rates
• ISO-NE Liaison
• The calculations were based upon estimates of
  achievable DR by key rate classes
• Geographic and demographic analyses of customer
  characteristics in relation to key distribution
  system nodes
• Rate analyses were also performed to estimate DR
  impacts on billing determinants

50
DRs Impact on DISCO Revenues

• Monthly peak demand determines billing kW.
• Analyses assume a 10 reduction in billing
  determinants.
• For programs examined, revenue differentials
  minimal (under .1) as only participants are
  impacted and other high consumption hours make up
  for the reduced demand charge for these
  participants in hours with load control.

51
NSTAR BCR
NSTAR BCR 0.32

Deferred TD Expenditures Revenue from ISO-NE
NSTAR Benefits 7.7M
Market Benefits 80.1M
Benefits
Costs

• NSTAR Expanded DR
• Program costs
• Lost revenues

NSTAR Investment 23.8M

• - Calculations assume 160 MW enrolled in CI call
  option program and 10 MW enrolled in mass-mkt
  program in 2004 increasing by 10 MW per year
  until 50 MW are enrolled in 2008.
• Values are NPV over 5 years at 7.72 discount
  rate.

52
Market-Wide Benefits

• A pivot factor in assessing NSTAR options
• Simple analyses used to justify assumed large
  benefits in terms of
• 1. Hedges against price spikes in spot markets,
• 2. Reduced price volatility influencing all MW
  transactions through lowered forward price curve,
  and
• 3. Increased reliability (TD system and
  generation adequacy)
• 4. Portfolio value due to resource diversity --
  DR supply costs not correlated with fuel prices,
  plant outages, and transmission congestion.
• 5. Reduced market power -- Mkt power exists on
  peak days when transmission constraints do not
  all for the import of power into zones.
• Methods for estimating benefits have not been
  standardized and are still being developed.
• Existing studies are retrospective and show low
  price effects from DR -- BUT what might occur in
  the next five years?

53
Market Extended Market Benefits

• Market Benefits
• Transfer/Collateral Benefits
• Reduction in long-term hedging costs
• System reliability improvement
• Extended Market Benefits
• Transmission Real Options Value
• Increased incentive for innovation
• Increased resource portfolio diversity
• Reduction in market power
• Operating reserves demand curve

54
Market Benefits Definitions

• Collateral Savings Reduction in market-clearing
  prices in spot markets impacted by DR. Also
  referred to as Benefits to Non-participant
  Buyers.
• Hedging Benefits Savings due to reduced average
  prices and price variability in the market.
  Results lowered forward price curve for all MW
  transactions.
• Reliability Benefits Enhanced grid reliability
  and reduced probability of customer outages
  (reduced ISO-NE calculated LOLP).
• Transmission Real Options Value Hedge against
  low probability, high consequence events Only
  counted in Extended Market Benefits Analysis.

55
Rationale for including Market Benefits

• DR has positive impacts beyond those captured by
  traditional benefit/cost tests.
• DR has potential to ameliorate generation,
  transmission, and distribution issues in specific
  situations.
• FERC estimated that a 5 reduction in peak demand
  could have reduced recent California price spikes
  by 50.
• NSTAR has received societal benefits funds to
  assess a small-scale demand response program for
  commercial customers.

56
Market BCR for DISCO Program
Market BCR 3.36 Extended Market BCR 5.34Based
on an Options Value of 25 million (best guess)

• Generation Reliability
• - Reserves
• - Outage costs
• Reduced Energy Costs
• - MWh prices
• - Hedging costs

Market Benefits 80.1M
Benefits
Costs

• NSTAR Expanded DR
• Program costs
• Resource costs

NSTAR Investment 23.8M

• Best practices used to estimate market benefits.
• Much debate about the magnitude of the benefits.
• Calculations assume 160 MW enrolled in CI call
  option program and 10 MW enrolled in mass-mkt
  program in 2004 increasing by 10 MW per year
  until 50 MW are enrolled in 2008.
• Values are NPV over 5 years at 7.72 discount
  rate.

57
Go/No Go Factors for DRR

• Conditions favoring DR initiatives
• Escalating or volatile energy prices
• Plant outages and reduced generation availability
• Uncertainty in capital markets
• Unexpected growth in electric demand
• Wide availability of advanced metering technology
• Increased risk management costs
• Conditions hindering DR initiatives
• Bifurcation of incentives
• Supply/generation technology breakthroughs
• Low cost of capital for plant additions
• Stable or low growth in electric demand
• Status quo regulatory posture (e.g. no
  innovative rates, etc.)

58
Regulatory and Market Unknowns

• Regulatory environment
• Will regulators encourage distribution companies
  to promote DR that serves dual purposes (DISCO
  benefits and MARKET benefits) via rates or
  incentives?
• Depends upon perceived market-wide benefits.
• Market environment
• Will suppliers (load serving entities -- LSEs)
  pay the DISCO for commodity price risk protection
  via DR?
• Will there be competitive curtailment service
  providers (CSPs) that serve the market function?
• Will outsourced DRR become an industry convention
  (e.g., the purchase of a Texas Utility's DRR
  program by a technology company, i.e., Comverge,
  Inc.)?

59
KEMA-XENERGY ExampleMethod 1 Black Scholes

• Value of options to buy or sell at fixed prices
• Call option to buy at X if market goes higher
• Put option to sell at X if market goes lower
• Other Inputs
• r risk-free rate of return
• s volatility
• DR as call option that reduces risk
• Avoided expected price base value
• Call option says I can buy at a fixed price
• Value of the call option is the value of avoided
  exposure to prices above base

60
KEMA-XENERGY Method 1 Black Scholes - Problems

• If market drops, could have paid too much for DR
• To protect against overpayment risk, could buy a
  put option
• Strike price cost of DR (/ MWh)
• Cost of overpayment risk value of the put
  option, whether or not we actually buy it.
• NOTE Black Scholes not typically assumed to
  apply to electricity markets, but it is easily
  applied as and example.

61
KEMA-XENERGY Method 2 Portfolio Optimization

• Monte Carlo simulation compares cost and risk of
  alternative portfolios (used instead of
  closed-form Black-Scholes)
• Can construct a supply portfolio to satisfy
  reliability requirement
• Balancing cost and risk
• Define cost-risk trade-off explicitly or
  implicitly
• In simple illustration,
• Price volatility is only source of risk
• Model interaction between native generation and
  market

62
KEMA-XENERGY Method 2 Portfolio Comparison
Generation Capacity

Demand and Market Prices
Market Price Volatility 28
63
KEMA-XENERGY Method 2 Portfolio Comparison

NOTE This is only one piece of the puzzle, but
it is indicative of new thinking in approaching
these problems. From AESP/EPRI
Pricing Conference, May 18-19, 2004
64
Alternative Portfolio Approach

• 1. Information needs for DRR planning
• Resource characterization and value analyses.
• Need to dimension uncertainty around key factors.
• 2. What is needed from the planning tools
• Ability to work with distributions as inputs.
• Address the value of information as uncertainty
  is reduced over time.
• Time steps are required in the analyses.
• 3. A simplified example
• 4. Conclusions

65
Application of Portfolio Analyses

• Today's planning environment requires analyses
  that
• Uncertainty be incorporated in the analyses.
• Risk mitigation options must be identified and
  valued.
• Appropriately credit Demand Response (DR) and
  Energy Efficiency (EE) for risk management and
  other values.
• Hedging values as expressed in reduced mean peak
  period prices and price volatility -- both
  influence forward price curves.
• Direct price impacts in spot market transactions
  (gas and electric).
• Other values (market power, innovation, customer
  values).
• Address the value of information and learning
  over time.
• Assess the value of flexibility, i.e., creation
  of real options to address future contingencies
  (some may not yet be known).
• Continue to appropriately analyze supply-side
  economics.

66
Application of New Tools

• Need to dimension uncertainty.
• Assess Value at Risk from different options.
• Fully address the portfolio of demand-side and
  supply-side options.
• Need to work with distributions of outcomes
• Closed form solutions and analytics.
• Monte Carlo methods.
• Decision-tree variants.
• Must incorporate time steps to address
  flexibility.
• New models such as _at_RISK and Crystal Ball allow
  for analyses based on representations of market
  uncertainties.
• Supply-side models also incorporate Monte Carlo
  solutions, e.g., General Electric's MAPS model
  and Global Energy Decision's MIDAS model.
  Adaptations of these models may be useful.

67
Information Needs for Portfolio Analyses

• 1. Appropriately capturing all the value
  associated with a resource option.
• Many values associated with demand-side options
  are difficult to quantify, but are growing in
  importance as supply-side resources become more
  constrained (e.g., transmission congestion, and
  natural gas availability and prices)
• 2. Need to dimension uncertainty around future
  outcomes.
• Simple planning paradigms such as 1 in 10 year
  events are not very useful in assessing option
  and hedge values as they only represent one
  point.
• Different approaches are needed for dimensioning
  uncertainty if new tools are to be useful.

68
2. Dimensioning Uncertainty

• Expressing and dimensioning uncertainty for use
  in analyses.
• Uncertainty is what makes hedges and options
  valuable.
• If we could use point estimates and were certain
  about their values, there is no need for options
  or hedges since the optimal solution would simply
  be picked.
• Industry has used few tools to express
  uncertainty
• Key problem -- How to dimension uncertainty for
  use in planning analyses (simplest to more
  complex)
• 1. Scenario analyses
• 2. Range estimates -- construct confidence
  intervals based on key inputs.
• 3. Range estimates with the range filled in with
  likelihood estimates to provide a rough cut
  probability distribution.

69
Scenarios Versus Distributions
70
Application of New Tools

• Need to dimension uncertainty.
• Assess Value at Risk from different options.
• Fully address the portfolio of demand-side and
  supply-side options.
• Need to work with distributions of outcomes
• Closed form solutions and analytics.
• Monte Carlo methods
• Decisions tree variants
• Must incorporate time steps to address
  flexibility.
• New models such as _at_RISK and Crystal Ball allow
  for analyses based on representations of market
  uncertainties.

71
Simplified Example -- Decision Tree
Time Period T 1
Objective MinimizeRevenue Requirementsover 10
years. Time Step One-year steps overa 10-year
period. Proxy Example Real applicationwould
includedistributions insteadof single
probabilitynodes.
SupplyPortfolio2
SupplyPortfolio3
Other Time Steps
SupplyPortfolio1
GasPrices
SeasonalEnergyDemandMetrics
PeakDemandMetrics



High .6
Low .4
High .7



High .5
Low .3



High .5
Low .5



High .5
Low .5



Low .5
High .5



High .4



Low .5
Low .6



NPVVAR
NPVVAR
NPVVAR
72
Example DistributionStochastic Price Forecasts
73
Example DistributionForecasted Bill Changes
74
Conclusions (Agree/Disagree?)

• The tools exist to assess portfolio of
  supply-side and demand-side options.
• This requires
• 1. Appropriate resource characterization.
• 2. Representations of the uncertainty around key
  factors in the analysis.
• The challenge is to change perspectives and to
  get planners to move out of their comfort zone to
  develop better (i.e., more accurate)
  representations of uncertainty.
• Representing uncertainty and the value of
  information over time is the key challenge as
  both contribute to the value of options and
  hedges.
• This is new to planners, but it is necessary --
  the good news is we have the tools and processes
  that will allow these analyses.

75
Conclusions -- Unique DRR Valuation Problems

• Values accrue to different entities
• Distribution companies in terms of deferred
  maintenance and new facilities, plus contingency
  avoidance.
• Transmission owners through reduced capacity and
  maintenance.
• Reliability managers through lowered costs of
  better Loss of Load Probabilities (LOLPs).
• Customers who now are able to receive payment for
  their ability to use electricity flexibility.
• A key attribute of consumption is now given a
  value.
• SO -- Values accrue to many and to the market at
  large, but costs are concentrated at the program
  level.
• Value is segmented with no one group is willing
  to provide full value for DR, but generators can
  be consolidated opponents.

76
Day 2 Topics

• Suggestions for additional topics
• 1. Final report outline just to see if one view
  of the project deliverable fits with the needs of
  most IEA participants.
• 2. What tools might we want to include in our
  review and try to build on
• Supply-side production cost models (probabilistic
  and non-probabilistic)
• Load shaping tools
• Monte Carlo and decision analysis tools (_at_RISK
  and Crystal Ball)
• Others?
• 3. Can we use information used to assess
  distributed generation to help address
  curtailable load options?
• 4. Other topics

77

• Meeting Moderators
• Dan Violette Pete ScarpelliSummit Blue
  Consulting RETX, Inc.1722 14th Street,
  230Boulder, Colorado 80302 Chicago,
  IllinoisPh 720-564-1130 Ph312-953-4642E-Mai
  l dviolette_at_summitblue.com E-Mail
  pscarpelli_at_retx.com
• Meeting Host
• Mikael TogebyElkraft SystemLautruphøj 72750
  BallerupPh 44 87 36 16www.elkraft-system.dk