C HARLES R IVER A SSOCIATES The Economics of Critical-Peak Pricing (CPP) for Mass Markets September 23, 2004 Electricity Pricing for Rate Cases An EUCI.

CHARLES RIVER ASSOCIATES

The Economics of Critical-Peak Pricing (CPP) for Mass Markets

September 23, 2004Electricity Pricing for Rate Cases

An EUCI ConferenceDenver, Colorado

Ahmad Faruqui and Steve George

2EUCI Pricing Conference

Presentation Objectives

Define critical-peak pricing (CPP)

Discuss where it has been implemented

Specify a cost-benefit analysis framework

Review impact evaluation methodologies

Provide impact estimates from California’s statewide pricing pilot and associated demand curves

Discuss barriers to implementation


What Is Critical-Peak Pricing (CPP)?


CPP is a sub-set of dynamic pricing

Dynamic pricing is an advancement over traditional TOU pricing, which allows prices to vary across the day and perhaps seasonally, but where both prices and time periods are known a priori and are fixed for some duration

Dynamic pricing includes all electricity tariffs that recognize the inherent uncertainty in supply costs

• Option 1: The timing for which known prices are in effect is uncertain

• Option 2: Both timing and price levels are uncertain

• Option 3: Timing, price levels and time blocks are all uncertain


Examples of dynamic pricing

Critical Peak Pricing (CPP): Layers a much higher critical peak price on top of TOU rates. The CPP is only used on a maximum number of days each year, the timing of which is unknown until a day ahead or perhaps even the day of a critical pricing day

Extreme Day Pricing (EDP): Similar to CPP, except that the higher price is in effect for all 24 hours for a maximum number of critical days, the timing of which is unknown until a day ahead

Real Time Pricing (RTP): Prices that vary hourly or sub-hourly all year long, for some or all of a customer’s load


Why focus on the mass market?

The largest number of customers are in the residential and small commercial and industrial market, often called the mass market

They account for a third or more of the energy consumption in many utilities and a bigger fraction of the peak demand

These customers often do not make attractive candidates for retailers and are often left out of the picture

Letting the incumbent utilities give them a choice of pricing products brings them back into the game


Who Is Doing What?


The French are the global leaders in dynamic pricing for mass markets

About 10 million residential customers are on a two-part TOU tariff in France, out of a population of 31 million customers

Several hundred thousand customers are on a special rate, called tempo, which was initiated in 1996

Prices on the tempo TOU rate vary by day type, being the highest on 22 red days, followed by 43 white days and 300 blue days

The days are called at 4:30 pm the day before, through the Internet, Minitel and a vocal system; they are posted at 8:30 pm on a compteur electronique


EDF’s offers TOU and tempo rates

5.44

10.5

45.71

6.324.39

8.87

16.39

10.32

05

101520253035404550

Regular TOU tempo BlueDay

tempo WhiteDay

tempo RedDay

Eu

ro c

ents

per

kW

h

Peak Off-Peak


Gulf Power

The Good Cents Select program combines TOU and CPP pricing elements with sophisticated energy management technology for residential consumers

The voluntary program has been in place for roughly two years and has about 3,000 customers enrolled

Each customer pays roughly $5/month to participate and still saves 15% on their electricity bill

Average coincident peak demand reductions exceed 2 kW

40 percent reduction in peak period energy usage

96% customer satisfaction rating


The first PG&E experiment

In 1992, PG&E implemented a CPP rate that used two-way communication with automated thermostat control of air conditioners

• Peak/off-peak price ratio of roughly 4.5/1 and a CPP/off-peak price ratio of nearly 8/1

Usage during the CPP period dropped by roughly half for low, medium and high usage customers

Usage reductions on the hottest CPP days were slightly greater than on other CPP days

• That is, customers did not override their thermostat setbacks even on the very hottest days


The GPU experiment

In 1997, GPU implemented a CPP pilot program similar to the PG&E pilot, and got similar results

3-Part TOU rate with occasional use of critical peak price for limited number of high-cost hours

Peak/off-peak price ratios of 4/1 and 5/1 for 3-part rate with critical price ratio closer to 8/1 (50¢/kWh)

Energy reductions

Peak Period 26%

Critical Peak 50%


California’s statewide pricing pilot (SPP)

The SPP is an outgrowth of a regulatory proceeding initiated by the CPUC (R.02-06-001) in June 2002 on advanced metering and dynamic pricing

• The outcome of a working group process involving multiple stakeholders

• The first large-scale scientific experiment focused on dynamic pricing for mass-market consumers

It’s being carried out to narrow the range of uncertainty in assessing the benefits of deploying advanced metering infrastructure in the state


Are Dynamic Prices Cost Effective?


Who do we need to carry out a cost-effectiveness analysis?

Implementing dynamic rates typically requires a significant investment in advanced metering infrastructure (AMI), which can be a billion dollar capital investment for some companies

• Additional incremental costs might include higher data processing costs, changes to billing and customer service systems, and customer education and marketing costs

To determine whether such an investment is warranted, a company must compare these costs with potential offsetting benefits in the form of:

• Avoided meter reading costs and other operational savings

• Avoided G, T and D capacity investments

• Lower energy costs

• Potential reductions in environmental externalities

• Reduced risk for default suppliers, due to lower price volatility and risk sharing


California’s standard practice manual provide a way of determining cost effectiveness

Test Key Question Benefits Costs

Participant Test (PT) Is the participant better off?

Bill Decrease

Customer Incentives

Program Costs (Participant)

Participation Fees

Total Resource Cost Test (TRC)

Is resource efficiency improved?

Avoided supply-side costs

Program Costs (Total)

Ratepayer Impact Test (RIM)

Are rates lowered? Avoided supply-side costs

Participant Fees

Revenue loss

Customer Incentives

Program Costs (Utility)

Utility Cost Test (UC) Are revenue requirements lowered?

Avoided supply-side costs

Participant Fees

Customer Incentives

Program Costs (Utility)


What Are The Implementation Challenges?


There are at least four sets of challenges to implementing dynamic pricing

Utility

• Financial, cultural and operational

Regulatory

• Educational and political

Customer

• Financial and educational

Technology

• Scale, scope and density economies


Early Results From The SPP


The four climate zones of SPP


We were able to estimate a single, statewide demand model

For the CPP-F residential analysis, separate demand models were initially estimated for four climate zones. Ultimately, the data were pooled over the zones and binary variables representing individual zones were interacted with the price term to determine whether price responsiveness varied across zones

• The zonal binary variables were significant in Zones 3 and 4

However, these zonal binaries became insignificant in the regression models once we introduced interaction variables that allowed price responsiveness to vary with weather conditions and the ownership of central air conditioning (CAC) equipment

That is, the variation across zones is due to variation in weather and CAC ownership. Once these two factors are accounted for, there are no significant differences in price response across zones


Residential Elasticities


Elasticities for the CPP-F tariff

The statewide elasticity of substitution, derived from the CES model, is –0.069 for the typical weekday and the statewide daily price elasticity is –0.023.

• The elasticity of substitution varies significantly across climate zones, rising from a low of –0.032 in Zone 1 to –0.111 in Zone 4.

The statewide own-price elasticity of demand for peak period energy use is –0.094. It varies from a low of –0.055 in Zone 1 to a high of –0.139 in Zone 4.

• The statewide cross-price elasticity of demand for peak use is –0.140

The statewide own-price elasticity of demand for off peak use is –0.151. It rises from a low of –0.127 in Zone 1 to a high of –0.172 in Zone 3.

• The statewide cross-price elasticity for off-peak use is +0.01.


The elasticity of substitution varies with weather, daytype and CAC ownership

Average Household Household with no CAC

Household with CAC Climate Zone

CPP day Non-CPP Day

CPP Day Non-CPP Day

CPP Day Non-CPP Day

1 -0.04 -0.03 -0.04 -0.03 -0.07 -0.06 2 -0.06 -0.05 -0.05 -0.04 -0.08 -0.08 3 -0.10 -0.09 -0.08 -0.07 -0.11 -0.10 4 -0.12 -0.11 -0.10 -0.09 -0.13 -0.12

State -0.08 -0.07 -0.06 -0.05 -0.10 -0.09


Key findings for the CPP-V tariff

The elasticity of substitution is considerably larger on CPP days, with a value of –0.204, than on non-CPP days, with a value of –0.012

The elasticity of substitution on CPP days is much higher for the CPP-V rate than for the CPP-F rate

However, it is important to note that these elasticities pertain to single family households with central air conditioning in climate zone 3 who had already volunteered into an earlier pilot program involving smart thermostats. They can be generalized to a similar population but not to the residential population as a whole


Residential Impacts


Impact analysis is based on the weighted average prices for treatment customers in each climate zone

71.5

23.7

7.5

51.7

21.3

11.0

61.0

22.5

9.4

0

10

20

30

40

50

60

70

80

ce

nts

/kW

h

CPP-F High Ratio CPP-F Low Ratio Average

Residential Price For Consumer At Midpoint of Tier 3(Weighted Average, Climate Zone 2)

CPP Period Peak Period Off-Peak Period

Control Group Average Price 13.3 cents/kWh


Impact of CPP-F rate on CPP and non-CPP days

Percent Change In Peak Energy Use For Average SPP Prices

-13.37 -12.5

-3.32-5.59 -4.8

-17.13

-9.61-8.35

-1.91

-6.83

-20

-15

-10

-5

0

Zone 1 Zone 2 Zone 3 Zone 4 All

% C

ha

ng

e In

kW

h

CPP Day Non-CPP Day


Residential Demand Curves


Demand curves for peak, off-peak and daily use have been derived through simulation

Ahmad Faruqui:

Need to update all demand curves

Ahmad Faruqui:

Need to update all demand curves

Peak Period Demand Curve, Statewide

0.00

0.10

0.20

0.30

0.40

0.50

0.60

0.70

0.80

0.90

0.95 1.00 1.05 1.10 1.15 1.20

Peak Period kWh/hour

Pea

k P

rice

($/k

Wh)

0.13

1.16

0.62

0.22

1.04 1.12

Average CPP Price

Average TOU Price

Average Control Price


Demand curves vary by climate zone

Peak Period Demand Curves by Zone

$0.000

$0.100

$0.200

$0.300

$0.400

$0.500

$0.600

$0.700

$0.800

$0.900

0.3 0.6 0.9 1.2 1.5 1.8 2.1

Peak Period kWh'

Pea

k P

rice

($/k

Wh)

Zone 1 Zone 2 Zone 3 Zone 4 Statewide

0.13

0.62

1.16 1.720.840.770.490.46 1.46 1.65 2.021.04


Conclusions

Studies have shown that CPP can produce large economic gains for utilities and their customers

However, utilities and commissions face several barriers before such gains can be harnessed

Many utilities are seriously studying the economics of dynamic pricing and conducting experiments to overcome the barriers

The tools and data for analyzing and overcoming these barriers are available

Ultimately, the decision to roll out CPP as the default rate will partly be a philosophical decision, partly a political decision and partly an economic decision


References


Bibliography

Ahmad Faruqui and Stephen S. George, “Dynamic Pricing for the Mass Market: The California experiment,” Public Utilities Fortnightly, July 1, 2003, pp. 33-35.

Ahmad Faruqui, “Toward post-modern pricing: guest editorial,” The Electricity Journal, July 2003.

Ahmad Faruqui and Stephen S. George, “Demise of PSE’s TOU program imparts lessons,” Electric Light & Power, January 2003, pp.1 and 15.

Ahmad Faruqui and Stephen S. George, “Reforming pricing in retail markets,” Electric Perspectives, September/October 2002, pp. 20-21.


Bibliography (continued)

Ahmad Faruqui and Kelly Eakin (editors), Electricity Pricing in Transition, Kluwer Academic Publishers, 2002.

Ahmad Faruqui and Melanie Mauldin, “The barriers to real-time pricing: separating fact from fiction,” Public Utilities Fortnightly, July 15, 2002, pp. 30-40.

Ahmad Faruqui and Stephen S. George, “The value of dynamic pricing,” The Electricity Journal, July 2002, pp. 45-55.

Ahmad Faruqui and Stephen S. George, “Time to get serious about time-of-use rates,” Electric Light & Power, February 2002, Volume 80, Number 2, pp. 1-8.

Ahmad Faruqui, Hung-po Chao, Vic Niemeyer, Jeremy Platt and Karl Stahlkopf, “Getting out of the dark,” Regulation, Fall 2001, pp. 58-62.


Bibliography (concluded)

Ahmad Faruqui, Hung-po Chao, Vic Niemeyer, Jeremy Platt and Karl Stahlkopf, “Analyzing California’s power crisis,” The Energy Journal, Vol. 22, No. 4, pp. 29-52.

Ahmad Faruqui, Hung-po Chao, Vic Niemeyer, Jeremy Platt and Karl Stahlkopf, “California syndrome,” Power Economics, May 2001, Volume 5, Issue 5, pp. 24-27.

Ahmad Faruqui and Steve Braithwait, “The choice not to buy: energy savings and policy alternatives for demand response,” Public Utilities Fortnightly, March 15, 2001.

Ahmad Faruqui and Kelly Eakin (editors), Pricing Electricity in Competitive Markets, Kluwer Academic Publishing, 2000.

C HARLES R IVER A SSOCIATES The Economics of Critical-Peak Pricing (CPP) for Mass Markets September 23, 2004 Electricity Pricing for Rate Cases An EUCI.

Documents

criticalpeak pricing

electricity pricing

euci pricing conference

cpp pricing elements

euci pricing conference

duration dynamic pricing

traditional tou pricing

choice of pricing products