Does Real-time Pricing Deliver Demand Response?

1
Does Real-time Pricing Deliver Demand Response?
Charles Goldman Lawrence Berkeley National
Laboratory CAGoldman_at_lbl.gov New England
Restructuring Roundtable October 28, 2005
2
Policy Questions

• What evidence is there that RTP delivers demand
  response (DR)?
• What factors determine how much DR you get?
• Customer enrollment (amount of load exposed to
  hourly varying prices)
• Price response of enrolled customers
• Relative roles of RTP and Emergency DR programs
• How does default service RTP impact retail market
  development?

3
Recent Projects Examining Large Customer RTP
Experience

• Optional RTP Tariffs Utility Experience
• Summarized 43 RTP programs offered by vertically
  integrated utilities in 2003
• Analyzed trends in program participation
  participant price response
• RTP as Default Service
• Case studies of eight states where RTP is
  considered for large C/I customers in context of
  competition for retail load
• Comparative analysis of market and regulatory
  context, RTP tariff design, and customer choices
  and response
• Case Study of Niagara Mohawk RTP
• In-depth study of customer choices and response
  to day-ahead hourly RTP tariff as default-service
  in a competitive retail environment
• Interviewed and surveyed customers and estimated
  price response

4
Optional RTP Tariffs Overview

• RTP offered by
• Most investor-owned utilities (IOU) in Southeast
  and TVA
• All IOU in Illinois and NY (statutory/ regulatory
  requirement)
• Many mid-west utilities (First Energy, Cinergy,
  Xcel, KCPL)
• All CA IOUs in 2003, but two programs since
  cancelled
• RTP not offered by many utilities in
• The West
• New England

Number of Utilities in Each State withRTP as
Optional Tariff (2003)
5
RTP as Default Service Regulatory Market
Context

• Default RTP driven primarily by retail market
  restructuring goals not DR
• In New York, PSC first decided against and then
  recently in favor of statewide default service
  RTP for large CI

6
Case Study of Niagara Mohawk RTP Choices
Available to SC-3A Customers
7
RTP Enrollment A Snapshot

• Enrollment in Default RTP
• 3-15 in PJM states
• 25-34 in New York
• Why the differences?
• Alternative, fixed-price utility supply option
  (e.g., Duquesne)
• Tariff design day-ahead vs. after-the-fact price
  notice
• Retail market development?
• Enrollment in Optional RTP
• Anemic in all cases except Ga. Power
• Offer inadequate savings opportunities or too
  risky
• Lack of aggressive marketing by utility

8
Niagara Mohawk RTP Customer Migration Patterns

• You can build it, but they may not (or may)
  stick around
• 17 of customers left NMPC for ESCO and never
  returned
• 18 went back and forth
• 37 switched later to ESCO and never returned
• 28 of customers stayed on NMPC RTP
• Load served by ESCO 30 (2000) increased to 63
  by 2004
• Surprise Load facing hourly prices (45-60 in
  2004)

9
Niagara Mohawk RTP What Customers Told Us
Barriers to Price Response (N76) Frequency
No barriers encountered 9
Organization/ Business Practices
Insufficient time to pay attention to prices 39
Institutional barriers 23
Inflexible labor schedule 16
Inadequate incentives
Electricity is not a priority 17
Cost/inconvenience outweighs savings 17
Risk averse/ hedged
Management views price response as too risky 10
Flat rate or time-of-use contract makes responding unimportant 9

• 30 of customers say they are unable to curtail
  load
• 70 can either forego or shift load or utilize
  onsite generation
• Most customers report multiple barriers to price
  response15 respond without obstacles

10
Niagara Mohawk RTP What customers actually did?
Business Category N Average Elasticity
Government/Education 34 0.10
Public Works 17 0.02
Commercial/Retail 16 0.06
Healthcare 8 0.04
Manufacturing 44 0.16

• Relative price responsiveness varies
  substantially across and within business sectors
• Key Findings
• 18 of customers account for 75-80 of aggregate
  DR
• 119 customers reduced their peak demand (500 MW)
  by 10 (50 MW)

11
Optional RTP Tariffs Maximum Load Reductions

• Aggregate load reductions are modest for nearly
  all RTP programs (lt1 of utility peak)
• Only two utilities (Duke Georgia Power)
  reported load reductions greater than 100 MW

Public Service of Oklahoma
40 MW
Duke Power
200 MW
Com Ed (Rate RHEP)
Jersey Central Power Light
60 MW
Florida Power Light
Kansas CityPower Light
Otter Tail Power(Option 1)
Pacific Gas Electric
Georgia Power
750 MW
23 MW
Gulf Power
0
1
2
3
4
5
6
Maximum Load Reduction ( of Utility's Peak)
12
RTP as Default Service Customers Exposed to Spot
Market Prices

• Potential market impact
• Niagara Mohawk curtailments equivalent to about
  0.6 of system peak load
• New Jersey and Maryland unknown

13
ISO/Utility DR Program Performance
DR Program Maximum Load Reductions Percent of
Total System (State or Utility) Peak

• Actual performance tends to vary with program
  type
• Emergency DR programs yield large reductions when
  high payments offered (e.g., 0.50/kWh)
• Call option programs yield close to participants
  nominated amount
• Emergency DR programs have thus far
  demonstrated larger load reductions than RTP
  (except for GA)
• but not a direct substitute for RTP

14
Conclusions RTP as a Demand Response Strategy

• You can build it but they may not come
• Low enrollment in most optional RTP programs
• Participation doesnt guarantee price response
• Only 10 of 42 Optional RTP programs report load
  reductions
• 18 of NMPC customers account for 75-80 of DR
• Even if they come, they may not stick around
• Expect that most customers will switch from
  Default RTP
• But Default RTP can yield significant indirect
  market benefits
• More retail choice customers willing to face
  hourly prices but will they respond?
• Program design, supporting infrastructure and
  utility incentives are keys to success
• Default Service RTP Day-ahead, hourly pricing
  balances retail market development and DR
• Optional RTP Georgia Powers secrets to success
  (corporate commitment, aggressive marketing
  customers can hedge and CBL rules allow
  customers to generate bill savings)
• Policymakers must make long-term commitment to
  build DR infrastructure (e.g.,customer info,
  tech. assistance, codes/standards, mkt.
  assessment)

15
Implications RTP as a Demand Response Strategy

• Retail choice states
• Will state PUCs have the political will to
  establish RTP as default service and for which
  groups of customers?
• Many customers willing to face hourly prices for
  some load under current market conditions
  (moderate price volatility, reasonably
  competitive retail market)
• NMPC experience suggests moderate levels of
  demand response at high prices (10 50 MW)
• Policymakers need more information on retailer
  contract types and response
• Wholesale market design
• Customers will enroll and respond to emergency DR
  programs (1-3 of system peak)
• These DR programs are complementary with RTP
• Linking Price Response to ISO Spot Markets

16
LBNL Reports on RTP Experience

• A Survey of Utility Experience with Real Time
  Pricing
• G. Barbose, C. Goldman and B. Neenan. LBNL-54238,
  December 2004.
• Real Time Pricing as Default or Optional Service
  for CI Customers A Comparative Analysis of
  Eight Case Studies
• G. Barbose, C. Goldman, R. Bharvirkar, N. Hopper,
  M. Ting and B. Neenan. LBNL-57661, August 2005.
• Customer Strategies for Responding to Day-Ahead
  Market Hourly Electricity Pricing
• C. Goldman, N. Hopper R. Bharvirkar, B. Neenan,
  R. Boisvert, P. Cappers, D. Pratt, and K.
  Butkins. LBNL-57128. August 2005.
• Reports available at
  http//eetd.lbl.gov/ea/EMS/drlm-pubs.html

17
Background slides
18
Two-part Real-Time Pricing Tariff How It Works

• Customer sees hourly prices for their marginal
  usage
• Customer baseline (historic) usage (CBL)
  partially hedges customer against hourly price
  volatility

19
Block and Index Pricing
20
RTP as Default Service Customer Enrollment
21
Niagara Mohawk RTP Case Study Most Customers
Dont Check Hourly Prices

• 70 of NMPC customers report never or rarely
  checking day-ahead hourly prices
• 13 check when other signals NYISO events or
  hot weather suggest they are high
• 17 consult prices routinely

22
Response to Day-Ahead Prices NMPC Aggregate
Price Response Curve

• 119 SC-3A customers would reduce their load by
  about 50 MW, or 11 of their peak demand (500
  MW), at high prices

23
Price Response of Customers on Utility RTP Tariffs

• Georgia Power Company (Summer 1999 estimates)
• 750 MW load reduction on an exceptionally high
  priced day
• Load reduction 15 of participants combined
  billing demand
• Niagara Mohawk Power Company (2004)
• 50 MW reduction when peak prices 5x off-peak
  prices
• Load reduction 10 of participants combined
  billing demand

24
Trends in Day-Ahead Market Prices Summer,
Eastern New York
On-Peak defined as 2pm-5pm on weekdays

• Less price volatility since 2002 compared to
  summers of 2000 and 2001
• Average hourly prices for summer period are
  relatively stable over 5 years

25
Georgia Powers Secrets of Success

• Unique Georgia retail market underlies Georgia
  Powers success with RTP
• New CI customers have a one-time choice of
  supplier and GPC is allowed to compete
• High-level corporate commitment to RTP as a tool
  to compete for new load
• Key tariff design and implementation details
• Aggressive marketing for gt10 yrs
• High degree of ongoing customer support and
  (re-)training
• Attractive hedging options
• Two-part tariff design with CBL
• Supplemental financial hedging products (caps,
  collars, contracts for differences, adjustable
  CBLs)
• CBL rules have enabled many participants to
  obtain substantial bill savings, regardless of
  load response

26
Georgia Power RTP CBL Rules Enable Bill Savings

• Key fact On average, each customer on Georgia
  Powers RTP rate has a CBL equal to 60 of its
  actual load
• Across a sample of 85 accounts, the CBLs ranged
  from 0-80 of the customers total load
• How can this be?
• Customers previously on Georgia Powers
  Supplemental Energy rate (a curtailable rate)
  could receive an initial CBL equal to their Firm
  Load Level
• New customers can receive a CBL below their
  projected load
• All customers can expand their facilities or add
  load without adjusting their CBL upward
• Hourly RTP prices for load above the CBL have
  historically averaged less than standard tariff
  rates
• Marginal vs. Embedded Costs

27
Utility and ISO/RTO DR Program Enrollment (2004)

• Most case study states have call option (or
  interruptible) and voluntary load reduction
  programs for large CI