Using Residential Customer Response to Manage Hydro Risk Frank A. - - PowerPoint PPT Presentation

Using Residential Customer Response to Manage Hydro Risk

Frank A. Wolak Department of Economics Stanford University Stanford, CA 94305-6072 wolak@zia.stanford.edu http://www.stanford.edu/~wolak

Goals of Talk

• Demand can and must participate in wholesale

electricity market to manage hydro risk

– Higher wholesale prices do not cause more rain so demand must be reduced when there is less water available – Real-time pricing of electricity makes this possible without declaring crisis or curtailing firm load

• Critical peak pricing (CPP) with a rebate is a form of

real-time pricing that is popular with consumers

– May become default residential tariff in California

• Critical peak pricing schemes can be implemented in

hydro-dominated systems such as New Zealand

Outline of Talk

• Real time pricing versus time-of-use pricing
• Rationale for and description of Anaheim

experiment

• Critical peak pricing (CPP) with a rebate

experimental design

• Assessing validity of experimental design
• Measuring treatment effect of CPP event
• Applying results to hydro-dominated system

such as New Zealand

Real-time vs. Time-of-use pricing

• Real-time pricing (RTP)

– Retail prices that vary with real-time system conditions – In fossil-based system this requires hourly meters to implement

• Must measure consumption within hour

– In hydro-dominated system real-time pricing possible with monthly meter reading

• Time-of-use pricing (TOU)

– Retail prices that vary with time of day, regardless of system conditions

• Low price from midnight to 12 pm and 6 pm to midnight
• High price from noon to 6 pm

– Does not require hourly meter

• Only meter that records monthly consumption in two time periods

Real-time vs. Time-of-use pricing

• Real-time pricing

– Customers have incentive to reduce demand during periods with high wholesale prices and stressed system conditions

• Reduces wholesale price volatility and increases system reliability
• Limits ability of suppliers to exercise unilateral market power

– Retailers with RTP customers can use them to exercise monopsony power

• Time-of-use pricing

– Customers have no incentive to reduce demand during periods with high wholesale prices and stressed system conditions

• Similar incentive to single fixed price tariff

– Two fixed prices all days as opposed to one fixed price all days

– Inelastic hourly demand for electricity with respect to hourly wholesale price

Anaheim Critical Peak Pricing Experiment

• Why is a Anaheim critical peak pricing experiment necessary?

– Interval metering must be installed to implement real-time pricing tariffs

• A major source of benefits of interval metering is ability to implement real-

time pricing tariffs

– Savings on meter reading costs because meters are remotely readable – Operational benefits because can see exactly where power outages are in real time

• Symmetric treatment of load and generation for just large customers is very

difficult for regulator to implement politically

– All customers with peak demands above 200 kW have had interval meters (paid for by California taxpayers) since summer of 2002 – Default price for these customers is still fixed retail price

• Results of experiment can be used to quantify benefits of implementing

universal interval metering

– Results can also be used to estimate dynamic model of intertemporal behavior to compare treatment effect of CPP day to own-price demand elasticity

• Wolak (2007) “What Does a Treatment Effect Measure: Evidence from Anaheim

Critical Peak Pricing Experiment”

Telephone Telephone Internet Internet Wireless Network Wireless Network

Data Center Utility User Consumer Local power lines Wireless Distribution lines

AMI Communication Networks

Local Area Networks Wide Area Networks

Hourly Consumption of a Northern California Residential Consumer Weekly Consumption Monday to Sunday

Anaheim CPP Experiment

• During the summer of 2005, the City of Anaheim Public

Utilities (APU) ran a Critical Peak Pricing (CPP) experiment

• During late 2004, a random sample of APU residential

customers were selected to participate in experiment

• Customers in this sample were randomly assigned to the

control and treatment groups

– Control customers were not notified of this selection but simply had interval meters installed at their dwelling – Treatment group customers first received a letter notifying them that they had been selected to participate in CPP program and were asked to return a reply form with their phone number and/or e-mail address

• Follow-up phone calls to sign-up those that did not respond to mailing
• Follow-up mailing to recruit those who could not be contacted by phone
• Final result--Very little attrition from randomly selected treatment group
• Process ultimately resulted in 52 control customers and 71

treatment customers, or 123 total customers

Anaheim CPP Experiment

• All customers (treatment and control) paid a fixed

retail price of 6.75 cents/kWh up to their monthly baseline of 240 KWh

– Monthly consumption beyond 240 KWh baseline charged at 11.02 cents/kWh – This is tariff paid by all APU customers

• Customers in treatment group were subject to a

maximum of 12 CPP days for experiment period

– Day-ahead notification of CCP days via telephone or e- mail (depending on customer’s choice on reply card)

Anaheim CPP Experiment

• CCPs days are required to be on weekdays that are

not holidays

• Consumption below reference level during peak

period (noon to 6 pm) of CPP days eligible for refund of 35 cents/KWh

– Consumers receive a rebate if their average consumption during peak periods of CPP days is less than their reference peak period consumption

• Rebate on day d = max(0,(q(ref) – q(peak,d)))*p(rebate)
• Rebate implies that customers guaranteed not to pay

more than they would have under control tariff

– Recruitment letter emphasized “You can’t lose”

Anaheim CPP Experiment

• Rebate mechanism addresses customer concern that

real-time pricing “charges high prices at time when electricity is really needed”

– CPP-R rewards customers that reduce when others need electricity

• Reference peak period consumption is customer’s

“typical” peak period consumption

– Defined as average peak period consumption of that customer during three highest consumption non-CPP days that are eligible to be CCP days during entire experiment period

Dataset Used in Analysis

• Daily Peak and Off-peak period consumption for 123 locations

Peak period—noon to 6 pm

• Peak(i,d) = Peak period consumption for location i on day d

– Off-period—all other hours of the day

• OffPeak(i,d) = Off-Peak period consumption for location i on day d
• Temp(d) = Maximum daily temperature at Fullerton Airport

for day d

• Day(d) = Indicator for whether day d=1,…,136 (all days

during sample period)

• LOC(i) = Indicator for location i, i=1,…,123
• Treat(i) = Indicator for whether location i is in treatment group
• CCP(d) = Indicator for whether day d is a critical peak day

Pre-Treatment Period Comparison

Meters installed for all customers in experiment before June 1, 2005 start date of experiment Consumption recorded at 15-minute intervals throughout the day for customers in both groups Comparison of pre-treatment 15-minute means to assess randomness of selection of customers into experiment and their assignment to treatment and control groups

r ed=cont r ol gr oup bl ue=t r eat m ent gr oup consum t i on

• 0. 09
• 0. 1
• 0. 11
• 0. 12
• 0. 13
• 0. 14
• 0. 15
• 0. 16
• 0. 17
• 0. 18
• 0. 19
• 0. 2
• 0. 21
• 0. 22
• 0. 23
• 0. 24
• 0. 25
• 0. 26

t i m e 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24

Treatment Control

Pre-Treatment Period Comparison

For virtually all 15-minute periods, 95 percent confidence interval on mean difference in pre-experiment period consumption by treatment and controls groups contains zero Conclusion—No evidence of non-random selection into experiment or subsequently into treatment versus control groups

Difference in Pre-Experiment Period Treatment and Control Means

• 0.2
• 0.1

0.1 0.2

1 6 11 16 21 26 31 36 41 46 51 56 61 66 71 76 81 86 91 96

15-Minute Intervals Difference in Means Upper 95% Confidence Bound Lower 95% Confidence Bound

Measuring Price Response

• Two models estimated for peak period
• Average peak period treatment effect

– ln(Peak(i,d)) = αCCP(d)*Treat(i) + λd + δi + εid – δi = location-specific fixed effect (controls for persistent differences in consumption across locations) – λd = day-specific fixed effect (controls for persistent differences in consumption across days in sample)

• εid = observable mean zero stochastic disturbance
• Temperature dependent peak period treatment effect

– ln(Peak(i,d)) = αCCP(d)*Treat(i) + γCPP(d)*Treat(i)*TEMP(d) + νd + μi + ηid

• μi = location specific fixed-effect (controls for persistent differences

in consumption across locations) – νd = day-specific fixed effect (controls for persistent differences in consumption across days in sample) – ηid = observable mean zero stochastic disturbance

Measuring Price Response

• Two models estimated for off-peak period
• Average off-peak period treatment effect

– ln(Off-Peak(i,d)) = αCCP(d)*Treat(i) + λd + δi + εid – δi = location-specific fixed effect (controls for persistent differences in consumption across locations) – λd = day-specific fixed effect (controls for persistent differences in consumption across days in sample)

• εid = observable mean zero stochastic disturbance
• Temperature dependent peak period treatment effect

– ln(Off-Peak(i,d)) = αCCP(d)*Treat(i) + γCPP(d)*Treat(i)*TEMP(d) + νd + μi + ηid

• μi = location specific fixed-effect (controls for persistent differences

in consumption across locations) – νd = day-specific fixed effect (controls for persistent differences in consumption across days in sample) – ηid = observable mean zero stochastic disturbance

Estimation Results

Table 2: CPP Treatment Effect Estimates Under Different Assumptions Natural Log of Peak Period Consumption in KWh Natural Log of Off-Peak Period Consumption in KWh Coefficient Estimate Standard Error Coefficient Estimate Standard Error Variable Name Fixed Effects* Treat(i)*CPP(d)

• 0.1215

0.0579 0.0138 0.0316 Random Effects Treat(i)*CPP(d)

• 0.1207

0.0308 0.0144 0.0189 Feasible Generalized Least Squares** Treat(i)*CPP(d)

• 0.1462

0.0505 0.0118 0.0220 *Arrellano (1987) covariance matrix used, **Estimates computed using Cochrane-Orcutt procedure assuming AR(2) errors. All regressions include 135 day-of-sample fixed effects.

Estimation Results

Table 3: Temperature Dependent CPP Treatment Effects Under Different Assumptions Variable Coefficient Standard Error Fixed Effects* Treat(i)*CPP(d) 2.9869 1.7184 Treat(i)*CPP(d)*Ln(Temp(d))

• 0.6894

0.3878 Random Effects Treat(i)*CPP(d) 2.9927 1.6510 Treat(i)*CPP(d)*Ln(Temp(d))

• 0.6905

0.3677 Feasible Generalized Least Squares** Treat(i)*CPP(d) 2.1326 1.4345 Treat(i)*CPP(d)*Ln(Temp(d))

• 0.5058

0.3233 *Arrellano (1987) covariance matrix used, **Estimates computed using Cochrane-Orcutt procedure assuming AR(2) errors. All regressions include 135 day-of-sample fixed effects

Temperature Dependent Treatment Effects

Dynamics of Price Response

• Examine if substitution across days occurred as a result of CCP days
• Include lagged value of CPP(d)*Treat(i)

– ln(Peak(i,d)) = α1CCP(d)*Treat(i) + α1CCP(d-1)*Treat(i) + λd + δi + εid – δi = location-specific fixed effect (controls for persistent differences in consumption across locations) – λd = day-specific fixed effect (controls for persistent differences in consumption across days in sample)

• εid = observable mean zero stochastic disturbance
• Include lagged value of CPP(d)*Treat(i)

– ln(Off-Peak(i,d)) = αCCP(d)*Treat(i) + γCPP(d-1)*Treat(i) + νd + μi + ηid

• μi = location specific fixed-effect (controls for persistent differences in

consumption across locations) – νd = day-specific fixed effect (controls for persistent differences in consumption across days in sample) – ηid = observable mean zero stochastic disturbance

• Same regression with lead value of CPP(d)*Treat(i)

Summary of Results

• Load-reduction effect--Peak period consumption of treated

group approximately 13% lower than consumption of control group during CCP days

– Controlling for all fixed differences across locations, and fixed differences across days

• Load-reduction effect—Evidence of larger consumption

reduction in higher temperature days

– Five degree temperature increase implies 4 percentage point increase in the consumption reduction of treated group versus control group

• Little evidence of load shifting to off-peak periods

– No statistically significant difference in treatment versus control group mean consumption during off-peak periods on CPP days – No statistically significant difference in treatment versus control group mean consumption during peak and off-peak periods in day before or day after CPP day

Designing CPP-R Programs

• Design parameters of CPP-R programs

– Rebate price – Level of fixed price(s) – Maximum times an event can be called – Mechanism used to set reference level relative to which rebates are issued

• For peak-shifting, level of rebate price very important

– Need to get attention of customer to reduce during hour needed

• Set level of fixed price and mechanism used to set reference

level to ensure revenue adequacy of retailer

– Customer could be offered menu of fixed prices and reference levels

• Set maximum time CCP events called to balance

– Credibility that events will actually occur—interruptible customers – Not too frequently to cause CPP events to be ignored

Designing CPP-R Program for NZ

• If problem is managing hydro risk

– Program should be designed on an annual basis

• Customers sign one year or longer contracts to remain on tariff

– With monthly meter reading CPP events should last one month – Rebate price does not need to be significantly higher than fixed price because rebate will be paid for total monthly reduction relative to monthly reference level – CPP called based on publicly available and verifiable hydro information

• For instance, water levels in major hydro basins from COMIT database

– Reference level based on average consumption during this month for past N years

• Aggregate reference level information should be submitted to Electricity

Commission so that it can determine water levels that trigger CPP event

• Electricity Commission may suggest that all retailers offer a

CPP-R rate

Designing CPP-R Program for NZ

• If problem is managing intra-day price risk

– Hourly metering must be installed

• CPP events last portions of the day

– Rebate price should be significantly higher than fixed price because rebate must overcome inertia of customer to respond – CPP events will be called based on publicly available and verifiable information such as prospective wholesale price

• M-co produces several rounds prospective prices as part of market
• peration

– Reference level mechanism should be designed to reduce peak demand

• Provide opportunity for customer to earn sizeable rebate for significant

reduction (overcome inertia to respond)

• Avoid paying for reductions that would occur anyway

– Wolak (2006) “Residential Customer Response to Real-Time Pricing: The Anaheim Critical-Peak Pricing Experiment”

Conclusions

• No need for expensive capacity payment mechanisms to

manage hydro risk

– Demand-side participation works in markets for other products and it can work in the market for electricity

• Anaheim experience demonstrates significant demand

reductions possible with modest rebate payments

• Program can be designed to achieve and/or annual energy

and peak energy reduction goals

• Hydro-based system can implement an annual CPP-R

program without hourly meters to reduce annual electricity demand

• Peak reduction programs requires hourly meters to

implement which may be more complicated for New Zealand given current meter ownership and control situation

Questions/Comments For more information: http://wolak.stanford.edu/~wolak