QUANTIFYING THE BENEFITS OF DYNAMIC PRICING

1
QUANTIFYING THE BENEFITS OF DYNAMIC PRICING

• Ahmad Faruqui and John Tsoukalis
• (925) 408- 0149
• An e-forum of the
• New Mexico State University
• Edison Electric Institute
• April 17, 2008

2
There Is Some Good News

• There is general agreement that appropriate
  tariffs are essential to any rapid development of
  electricity supply

3
And Some Bad News

• and there is complete disagreement as to what
  constitutes an appropriate tariff.
• D. J. Dalton
• London, March 1938

4
70 Years Later, The Debate Continues On This Side
Of The Pond

• The new rate design is dynamic pricing
• It takes time-of-use rates to its logical
  conclusion
• The highest prices occur in near-real time or
  real time conditions
• It is designed to lower system costs and benefit
  all customers
• It is still not economical to store electricity
• With the continued penetration of central air
  conditioning systems, the top 60 hours of the
  year now account for 10-15 percent of the annual
  system peak

5
5 Peak Reduction Worth 31 billion (over a
20-year period, just based on avoided costs)

NPV of Avoided Costs 31 billion
6
Assumptions

• 5 demand reduction in 757 GW
• 52/kW-year capacity price
• 20 year horizon
• 15 discount rate
• 2 peak growth rate
• Avoided cost of energy is 36 of avoided cost of
  capacity
• Value of wholesale price reduction is 278 of
  avoided cost of capacity
• Derived from a study on the value of DR in PJM
  
• The Brattle Group, 2007, Quantifying Demand
  Response Benefits in PJM, Prepared for PJM and
  MADRI

7
Dynamic Pricing Comes In Many Flavors

• The list includes
• Critical peak pricing (CPP)
• Peak time rebates (PTR)
• Variable peak pricing (VPP)
• Real time pricing (RTP)
• Any of these options can be combined with
  inclining block rate designs and/or standard TOU
  rates
• They can be combined with enabling technologies
  such as in-home displays and/or smart thermostats
• Finally, the rates can be offered on a day-ahead,
  day-of or hour-ahead basis

8
Critical Peak Pricing /Time of Use Pricing
Higher prices during Critical Peak Events
50-150 hours/year
Discounted price during off peak hours 7,700
hours/year
Higher TOU prices during peak hours 1,000
hours/year
9
Peak Time Rebate
Credit that mirrors the CPP surcharge
10
Each Rate Design Reflects A Risk-reward Trade-off
11
Dynamic Pricing Enables Customer Choice
12
Do Customers Respond?

• Not all do
• Some respond marginally
• Some respond a lot
• In the aggregate, they do respond quite a bit and
  their responses persist over time
• Three decades of experimentation attest to these
  findings

13
Newest Pricing Pilots Australia, Canada,
France, US
14
Pilots Reveal Compelling Evidence Of Demand
Response, Even To Static TOU Rates
15
Higher Impacts Are Observed For Dynamic Pricing
Than For Static Time-of-use Rates
16
Customer Response Varies By Price And Market
Segment
17
How Do We Quantify The Benefits Of Dynamic
Pricing?

• Estimate individual customer demand response
• Do your own pilot to get custom-tailored results
• Or use the PRISM software (see next slide)
• Estimate the number of program participants
• Decide whether dynamic pricing will be the
  default rate or an optional rate
• Estimate aggregate demand response
• Put a value on this demand response

18
What is PRISM?

• The PRISM model is used to estimate individual
  customer-level demand response
• PRISM needs several types of inputs
• Weather data
• Existing and dynamic rate designs
• Customer load shapes
• Central Air Conditioning saturation
• PRISM generates the change in the average
  customers load shape
• This is used to estimate short run customer bill
  impacts and long run avoided costs

19
How PRISM Works
20
Avoided Generation Capacity Cost Typically
Largest Portion Of Savings
21
The Key To PRISM Are The Elasticities Of Demand

• Elasticities estimated using experimental data
  from CAs 3 IOUs
• PRISM uses two elasticities
• Substitution elasticity estimates the shift
  from peak to off peak periods
• Daily price elasticity estimates the change in
  daily energy consumption
• Sample elasticities

22
Price Responsiveness Varies By Customer
Characteristics
23
Elasticities Applied To Difference Between The
Existing And The New Dynamic Rate
24
Problem Bills Initially Rise For 50 Of The
Dynamic Pricing Customers
How do we make dynamic pricing attractive for
these customers?
25
Flat Rates Embody Implicit But Very Real Price
Volatility Risk Premium
26
With Risk Premium Credit, Dynamic Pricing Rates
Attractive For 70
27
With Demand Response, Dynamic Pricing Attractive
To Over 95
28
So Why Is There So Little Dynamic Pricing?

• It strikes many people as a radical new idea
• As Arthur C. Clarke noted, when confronted with a
  new idea
• The first reaction is, It is completely
  impossible
• The delayed second reaction is, Its possible
  but not worth doing
• Ultimately, the reaction is, I said it was a
  good idea all along

29
References

• The Brattle Group, Quantifying the benefits of
  dynamic pricing, Edison Electric Institute,
  January 2008 (Downloadable from www.eei.org/ami)
• Ahmad Faruqui and Sanem Sergeci, The power of
  experimentation, Discussion Paper, The Brattle
  Group, April 4, 2008
• Plexus Research, Inc., Deciding on Smart
  Meters, Edison Electric Institute, September
  2006 (Downloadable from www.eei.org/ami)
• Federal Energy Regulatory Commission, Demand
  Response and Advanced Metering, Staff Report,
  August 2006
• Robert Earle and Ahmad Faruqui, Toward a new
  paradigm for valuing demand response, The
  Electricity Journal, May 2006
• US Department of Energy, Benefits of Demand
  Response in Electricity Markets, February 2006
• Join the AMI/Smart Grid list serve send name
  and affiliation to moldak_at_eei.org