Smart Grid Increasing the IQ of the Smart Grid Unclear exactly - - PowerPoint PPT Presentation

1

Increasing the IQ of the Smart Grid Through Active Customer Participation in Wholesale Electricity Markets

Frank A. Wolak Director, Program on Energy Sustainable Development Professor, Department of Economics Stanford University wolak@stanford.edu http://www.stanford.edu/~wolak (Joint work with Matthew Kahn)

Smart Grid

• Unclear exactly what the smart grid is, but

it does include

– Interval meters – Storage and load flexibility technology – Automated response technology

• Technology without information or an

incentive for the consumer to use it is unlikely yields economic benefits

• Two important necessary conditions

ignored in discussions of smart grid

– Information provision – Dynamic Pricing

2

Outline of Talk

• Why active participation of consumers is essential
• Managing intermittency
• Managing unilateral market power
• Three necessary conditions for active participation
• Interval meters, adequate information, dynamic pricing
• The role of information in active participation
• Information experiment
• The role of dynamic pricing in active participation
• Dynamic Pricing versus Time-of-Use Pricing
• Symmetric treatment of load and generation
• Dynamic Pricing Experiment
• Hourly Pricing (HP)
• Critical Peak Pricing (CPP)
• Critical Peak Pricing with Rebate (CPP-R)
• Day-ahead versus real-time dynamic pricing programs
• Automated dynamic demand response
• The role of symmetric treatment of load and generation

3

Why Active Participation is Essential

• Many jurisdictions have ambitious renewable energy goals

– California has 33 percent renewable energy share goal by 2020

• Renewables are often unavailable during peak periods

– During July 2006 heat storm, July 24 demand in California ISO control area hit a 1 in 50 year peak of 50,200 MW

• Less than 5 percent of installed wind capacity was operating at the time

– In California, wind energy comes primarily during night and solar energy can only come during the day

• Cloud cover can significantly reduce solar PV output

– Wind and solar output are highly positively correlated across locations in California

• If there is no wind at one location, there is likely to be no wind at others
• Major factor driving need for dynamic pricing—High wholesale prices

do not cause more wind or solar energy to be produced

– As share of renewable energy grows final consumers must supply more “dispatchable negawatts” to maintain system balance

• Load-shifting or investments in energy storage technologies

4

2

10000 20000 30000 40000 50000 60000 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Demand in Megawatt-Hours Hours of the Day

Hourly Demand July 24, 2006

Actual System Load Scheduled Load Hour Ahead Forecast 2-Day Ahead Forecast

Daily Load Shape in California

5 6

Solar Output During Daylight Hours in California ISO Control Area

7

Economics of Energy Efficiency

• Variation in electricity demand throughout day and year

– On 7/24/06 demand ranged from 28,300 MW to 50,200 MW – 50,200 MW is still historic peak for California ISO system

• Average MW consumption per hour during 2006

– Approximately 27,000 MW – Peak demand for 2006 is 50,200 MW

• Reducing peak demand through active participation

– Eliminate need to construct new generation capacity – Can retire old inefficient units located close to large cities

• Significant fraction of generation capacity used very

infrequently

– In California approximately 5,000 MW (10 percent of peak demand) used less than 2 percent of hours of the year – With climate change larger fraction is likely to be used even less frequently

8

3

California Summer Loads Conditions: 1998 to 2009 (Peaks are More Variable than Total Demand)

9

California ISO Control Area

Annual

10

Barriers to Active Participation

• Substantial state-level regulatory barriers to

active participation

– “Consumers must be protected from short-term price risk” – “Electricity is a right, not a commodity”

• Wolak, Frank (2007) “Managing Demand-Side Economic and Political Constraints on Electricity

Industry Re-structuring Processes,” on web-site.

• Stakeholders in regulatory process realize few,

if any, benefits from active participation

– Most lose--Regulatory staff, Generation unit owners, Distribution utilities

• Only consumers realize benefits

– Wolak, Frank (2010) “An Experimental Comparison of Critical Peak and Hourly Pricing: The PowerCentsDC Program,” on web-site – Wolak, Frank (2006) “Residential Customer Response to Real-Time Pricing: The Anaheim Critical-Peak Pricing Experiment,” on web-site

11

Necessary Conditions for Active Participation

• Lack of hourly metering of final demand makes it

impossible to set hourly retail prices that pass-through hourly wholesale price

– Customer reduces monthly bill by same amount by reducing consumption by 1 KWh during hour when wholesale price is $5000/MWh as he does when price is $0/MWh

• Economics of hourly meters is rapidly changing because
• f technological change

– Modern hourly meters are read remotely by wireless or wireline technology – Interval metering investment in California justified primarily using labor cost saving and increased outage monitoring quality

• All California investor-owned utilities should have

interval meters in place for all customers by end of 2012

– Need retail prices and information provision that maximize the benefits consumers realize from these meters

12

4

Necessary Conditions for Active Participation

• Consumers need to understand how their energy-

consuming actions translate into dollars on their monthly electricity bill

– Do not directly consume electricity – Electricity is a derived demand from the consumption of services from electricity-consuming durable goods

• Watching television, washing clothes or dishes, using computer
• Consumer needs to have information on costs of its

energy-consuming actions to make informed choices

• Most electricity utilities in California charge according to

nonlinear price schedule which complicates this process

– Information provision experiment described below attempts to assist in this process

13

Nonlinear Price Schedules

(SCE = Southern California Edison) (SMUD = Sacramento Municipal Utility District)

14

The Role of Retail Pricing

• Dynamic retail prices are source of the economic benefits to

consumers that alter their consumption to maintain system balance

– Manually in day-ahead time frame – Automatically (using technology) in real-time – Requires hourly meters to implement

• Must measure consumption on hourly basis to charge hourly prices

– Consumers must understand how their actions translate into dollars

• Dynamic retail prices also provide business case for consumers to

invest in energy efficiency and storage technologies

– Value of storage is ability to buy energy at low price and sell or displace purchase

• f energy at high prices

– Requires hourly meter to implement – Consumers must understand how actions translate into dollar savings

• Conclusion--Combination of hourly meters, information

provision, and dynamic retail pricing are necessary for consumers to benefit from active participation

15

Information Experiment

• PESD researchers partnered with two California electric utilities to

address the question of the role of information provision on a customer’s ability to participate in wholesale market

• Designed an “information treatment” to investigate two questions

– How nonlinear price schedule impacts a customer’s monthly electricity bill – How customer’s electricity-using actions translate into monthly bill

• On-line “information treatment” (roughly 30 minutes)

– 1) Shows customer nonlinear price schedule they face and where they were on that schedule during most recent months – 2) Takes an inventory of customer’s energy consuming durable goods and utilization of these durable goods – 3) Suggested energy savings that customer could undertake based on inventory and utilization and showed expected impact on monthly electricity bill

• Interactive and allowed customer to consider many options and then “commit” to actions

– 4) Customer was sent follow-up .pdf file of results to remind them of their “commitments”

16

5

Information Experiment

• For each utility, a random sample of customers was drawn and

divided into treatment and control groups

– Treatment groups were solicited by mail (in one case) and e-mail (in the other) to take information treatment and were offered Amazon gift cards if they did

• For both treatment and control groups, monthly consumption (in one

case) and hourly consumption (in the other case) were collected for six months before and six months after first solicitation was sent

• Effect of “information treatment” estimated

– Compare difference in consumption between treatment and control groups before versus after treatment occurs – Difference-in-difference estimate of “impact” of intervention

• Important factor determining impact of information on consumption

– Where consumer was on schedule before treatment impacts what actions they will take given this information and other information provided in “treatment” – Consumers that learn they are on lower tier(s) face a lower marginal price than customers on higher tier(s) 17

Nonlinear Price Schedule

(SMUD = Sacramento Municipal Utility District)

18 19

Information and Consumption Information Experiment

• Consumers on lower tier consume more in response to

information

• Consumers on higher tier consume less in response to

information

• Net overall effect of “information treatment” is a predicted

reduction in aggregate consumption from information because those on higher tier consume more and there are slightly more of them in population

• Does this result hold for utility that has five price steps?

– Does information about where consumer is on nonlinear price schedule in months leading up to intervention lead to different response from consumers?

20

6

Nonlinear Price Schedule

(SCE = Southern California Edison)

21

Inland Price Schedule has same levels for price steps but larger quantity on each step

22

Information and Consumption Information Experiment

• Consistent with results for two tiers

– Larger in absolute magnitude reduction for higher tiers – Larger in absolute magnitude increases for lower tiers

• Results demonstrate that information provision is

important factor in allowing consumers to participate actively in wholesale market

• Need to provide price signal and information on how

customer’s actions impact monthly bill

– Price signals without information are unlikely to be as effective as price signals with information

23

Dynamic Pricing Experiment

• Before describing dynamic pricing experiments

need to provide some background

– Distinguish between dynamic pricing and time-of-use pricing – Describe symmetric treatment of consumers and producers that characterizes all products sold through market mechanisms – Describe history of regulated retail prices that created initial conditions that provide consumers with “free hedge” of retail price and quantity risk

• Asymmetric treatment of consumers and producers

24

7

Dynamic vs. Time-of-use pricing

• Dynamic pricing

– Retail prices that vary with real-time system conditions – Requires hourly meters to implement

• Must measure consumption on hourly basis to charge hourly

prices

• 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 during day

25

Dynamic vs. Time-of-use pricing

• Dynamic 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

– Downward sloping aggregate hourly demand for electricity with respect to hourly wholesale price

• 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

– Perfectly inelastic hourly demand for electricity with respect to hourly wholesale price

• Does not limit ability of suppliers to exercise unilateral market power any more

than a single fixed price of electricity

26

Symmetric Treatment of Consumers and Producers

• In all markets, default price all consumers must

pay and producers must receive is real-time price

– Without symmetric treatment, maximum amount of active demand-side participation that benefits market efficiency is unlikely to develop – Neither consumers or producers are required to pay or receive this price

• To avoid it, customer must sign a hedging arrangement
• Example from airline industry

– Customers always have option to show up at airport and purchase ticket for flight they would like to travel on at real-time price – To hedge risk, consumer purchases ticket in advance (fixed-price forward contract)

• Electricity consumers must face same default price as

consumers of all other products to benefit from active participation

27

Historical Asymmetric Treatment

• f Consumers and Producers
• Because of legacy of vertically integrated-monopoly

market structure, in many jurisdictions customers have “free hedge” against real-time price for unlimited quantity

– In vertically-integrated monopoly regime, utility provided electricity price and quantity insurance to customer

• Customer paid firm’s average cost for each KWh consumed and

utility ensured supply was always available

• In wholesale market regime it is very difficult to set a fixed

retail price for unlimited quantity that is guaranteed to always cover wholesale energy costs

– Recall “California Electricity Crisis”

• Hedge for fixed quantity of energy is standardized product

that can be sold in secondary market

– How electricity is bought and sold in wholesale market, which implies that electricity in retail market must be bought and sold this way too (more on this point later)

28

8

• Fixed-retail price pays real-time hourly wholesale price

“on average” for hourly demand at fixed-retail price

– Retailers will go bankrupt if retail price does not satisfy equation given below on an annual basis

• P(retail) ≥ P(wholesale) + P(transmission) + P(distribution)
• Conclusion—Cannot “protect customers from volatile

wholesale prices”

– Can only prevent them from taking actions to limit wholesale price volatility and reduce their monthly bill – Investments in energy storage and demand flexibility can only be profitable with symmetric treatment of load and generation

• If pay 10 cents/KWh for all KWH, how you do make storage and load-shifting

investments pay?

Important Point

29

Politically Acceptable Dynamic Pricing

• Major complaint with dynamic retail pricing is that

customers cannot respond to hourly wholesale prices

– Difficult to determine when is best time to take action

• If taking action is costly and price increase is one hour

in duration, a very large price spike is needed to cause most customers to respond

– For residential customer with (2.5 KW) flat load shape, a large price spike is needed to overcome $5 cost of taking action to reduce demand by 20 percent

• $10,000/MWh for a 0.5 KWh demand reduction for 1 hour

– Longer duration of high prices requires smaller increase in prices

• $5,000/MWh average price for 0.5 KWh demand reduction for 2

hours

• Mechanisms that address cost-of-taking-action problem

can increase willingness of consumers to participate

– Critical Peak Pricing (CPP) is a popular way to do this

Politically Acceptable Real-Time Pricing

• Critical Peak Pricing—Customer purchases according

to usual fixed-price tariff or nonlinear price tariff during all hours of each day

• Customers face risk of Critical Peak Pricing (CPP) day

– Retailer commits to no more than N (N ≈ 10) CPP days in a pre-specified time interval – During peak-period of a CPP day, customer pays a much higher price for electricity

• Strong incentive reduce demand during this time period
• Peak period is typically 4 to 6 hours during day
• Overcomes cost of taking action problem by committing to a

sustained period of high prices

• Potential “moral hazard” problem for retailer

– Can declare CPP day to manage short-term wholesale energy purchase costs due to inadequate forward market procurement – Retailer has incentive to use all available CPP days because these are high profit days for retailer

• CPP price much higher than average retail price

Politically Acceptable Real-Time Pricing

• CPP with rebate is more popular with

consumers because it addresses this moral hazard problem

– Consumption during peak hours of CPP days receives a rebate relative to household’s reference consumption, if its actual consumption is less than reference consumption – Retailer faces risk that total rebates paid will be more than wholesale energy procurement cost savings

• If CPP period wholesale price is $300/MWh (implicit in retail

price), then if wholesale price is below $300/MWh, retailer loses money paying for rebate

• Retailer only wants to declare CPP days when rebates paid

are less than wholesale cost savings

9

Politically Acceptable Real-Time Pricing

• CPP with rebate (CPP-R) implies that

customers guaranteed not to pay more than they would have under baseline tariff

• “You can’t lose from rebate mechanism”
• Customers have the option to quit with no cost

implications if it is too difficult to reduce their consumption

– Pay for consumption above reference level during CPP period at fixed retail price

• Under CPP-R, marginal price of fixed retail

price plus rebate is only relevant if consumption is less than reference level

– Only carrot of rebate is used under CPP-R

• Under CPP, both carrot and stick used

– Higher price for all consumption during CPP period

Option to Quit and Average Treatment Effect

PC PN DH(p) DL(p) QHR = QHCon QHC QLCon QRef QLR = QLC ATE(CCP) = prob(Low)*(QLCon – QLC) + prob(High)*(QHCon – QHC) ATE(CCP-R) = prob(Low)*(QLCon – QLC), because QLR = QLC and QHR = QHCon ATE(CCP) > ATE(CCP-R), if PC ≈ PN + PRebate

Dynamic Pricing Experiment

• Do customers respond to high real-time price warnings

and CPP events?

• Treatment effect of discrete event (price elasticities will come later)
• How do these price responses differ across customer

classes?

• Regular (R) versus all electric (AE) customers
• Low-income (RAD) versus regular (R) customers
• Summer versus winter
• Does “cost of taking action” limit demand response of HP

customers versus CPP customers?

• Does “option to quit” result in CPP response greater than

CRR with rebate (CPP-R) response?

• Do Smart thermostats boost demand response?

PowerCentsDC Program Overview

• Residential pricing pilot

– Interval meters to record real-time consumption – Hourly pricing (HP) – Critical Peak Pricing (CPP) – Critical Peak Pricing with Rebate (CPR)

• Governed by “Smart Meter Pricing Pilot, Inc.” (SMPPI)

– Public Service Commission, DC – DC Office of People’s Counsel – Consumer Utility Board – IBEW

• International Brotherhood of Electrical Workers

– Pepco (contributed $2 million from shareholder funds)

10

Program Participants

Customers are from all eight Wards

Four Customer Codes

R = Not All Electric AE = All Electric RAD = Residential Aid Discount (Low Inc) RAD-AE = RAD All Electric (Low Inc)

Location of Participants

Smart Thermostat

• Offered to customers

with central A/C and who controlled their thermostat

• Approximately 25% of

customers opted for the smart thermostat.

• LED lights up during

CPP or High Price event (depending on pricing plan)

PowerCentsDC Pricing Options

• Critical peak pricing (CPP)

– A maximum of 12 CPP days during summer and 3 during winter

• Between 2 pm and 6 pm during summer (4 hours)
• Between 6 am to 8 am and 6 pm to 8 pm during winter (4 hours)

– Customers pay according to an increasing block schedule during all other hours – Customer charged approximately 75 cents/KWh for energy during CPP period

• Critical peak rebate (CPR)

– Customer earns rebates during critical peak hours by reducing usage below reference level set by SMPPI – Customers pay according to an increasing block schedule during all other hours – Customer receives rebate approximately equal to 63 cents/KWh and 12 cents/KWh is average energy price from standard pricing schedule

• Customer faces approximately same marginal price as CPP customer during

CPP period is rebate is being paid (63 cents/KWh + 12 cents/KWh = 75 cent/KWh)

11

PowerCentsDC Pricing Options

• Hourly pricing plan

– Hourly energy prices based day-ahead PJM prices – Hourly pricing curve made more extreme

• High price periods upweighted slightly
• Treatment received by customers

– All types of customers notified day before via automated phone call, email, or text page – CPP and CPR customers notified day before CPP event occurs

• CPP event days during summer called when forecast of high temperature for

day is above 90 degrees

• CPP event days during winter called when forecast of low temperature for

day is below 18 degrees

– Hourly pricing customers receive notification of high price (HP) warning hour

• Hours when day-ahead price for energy is above 23 cents/KWh

(> $230/MWh) during summer months (between 1/3 to 1/4 of CPP price)

• Hours when day-ahead price for energy is above 15 cents/KWh

(> $150/MWh) during winter months

Answers to Research Questions

• Price responsiveness

– Both R (Regular) and AE (All-Electric) customers reduce their consumption in response to CPP and HP hours

• In the range of 15% to 20% reductions

– Effect (% reduction in consumption from CPP or HP event) larger for AE customers relative to R customers in both summer and winter

• Roughly 5 to 10 percentage points larger

– For R customers effect primarily confined to summer periods

• RAD (Low Income) customers

– RAD-R and RAD-AE customers reduce their consumption in response to CPP event

• Treatment effects are larger than CPR treatment effect for R and AE customers
• Roughly double the percentage reduction of R customers on CPR rate
• Difficult to see evidence of “cost-of-taking action” for

hourly pricing

– Hourly pricing effect is between 1/3 to ¼ of size of CPP effect consistent with HP warning being for energy prices that are 1/3 to ¼ the size of CPP energy price – Can be explained by pattern of hourly prices during day

• Strong evidence in favor of option-to-quit effect

– For both R and AE customers CPR effect is ½ to ¼ of CPP effect – For RAD customers not possible to examine this hypothesis because only CPR treatment was applied to RAD-R and RAD-AE

• Smart thermostat significantly enhances

treatment effect

– Almost doubles effect for CPP treatment for AE customers – Also increases treatment effects for for R customers – Increases treatment effect for RAD-R and RAD-AE customers, but results not very precisely estimated

Answers to Research Questions

Conclusions from Dynamic Pricing

• Default hourly-pricing may not be that difficult for consumers to

respond to

– High price periods tend to cluster together, similar to CPP periods – Cost of taking action does not seem substantial

• Default CPR tariff inferior to default CPP tariff

– Loss in price-responsiveness could be large – “Option-to-quit” produces substantially smaller treatment effect – Further argument for default pass-through of hourly price or CPP default

• Smart thermostats significantly enhance price responsiveness of all

customers

– Air-conditioning and electric heating intensive areas may benefit most

• Low-income consumers can achieve significant price

responsiveness

– Almost double treatment effect of RAD-AE customers on CPR relative to R and AE customers on CPR – Low income consumers can achieve significant economic benefits from dynamic pricing

12

Day-Ahead versus Real-Time Dynamic Pricing

• All dynamic pricing plans currently based on day-ahead prices
• Critical peak pricing (CPP), CPP with rebate, Hourly pricing (HP)

plans

• Day-ahead prices are substantially less volatile than real-time

prices

• Consumers adjust day-ahead schedules based on day-ahead

prices

• Day-ahead price-responsiveness of customer assessed in
• Wolak, Frank (2010) “An Experimental Comparison of Critical Peak and Hourly

Pricing: The PowerCentsDC Program,” on web-site

• Wolak, Frank (2006) “Residential Customer Response to Real-Time Pricing: The

Anaheim Critical-Peak Pricing Experiment,” on web-site

• Consumers can use automated technologies to respond to differences

between day-ahead and real-time prices

Day-Ahead versus Real-Time Dynamic Pricing

• All US markets are multi-settlement markets

– Suppliers sell power in in day-ahead forward financial market that clears against real-time price at that location – Hourly pricing in day-ahead market and 5-minute pricing in real-time market

• Example of multi-settlement process

– Generation unit at node A sells 100 MWh in day- ahead market at $70/MWh – In real-time produces 110 MWh, which implies that unit sells 10 MWh at real-time price of $60/MWh – If unit produced 95 MWh, this implies that it must purchase 5 MWh at $60/MWh

46

• Active participation of final demand

– Suppose PDA < PRT – Final schedule of 100 MWh in day-ahead market at PDA – Consume 90 MWh – Consumer earns 10 MWh*(PRT – PDA) for 10 MWh demand reduction relative to day-ahead schedule

• Consumer can also make money by

increasing consumption

– Suppose PDA > PRT – Final schedule of 80 MWh in day-ahead at PDA – Consume 85 MWh – Consumer earns 5 MWh*(PDA – PRT) for 5 MWh purchased in real-time versus day-ahead market

47

Day-Ahead versus Real-Time Dynamic Pricing Day-Ahead versus Real-Time Dynamic Pricing

• Symmetric treatment of consumers and producers
• Default price that producer receives is real-time price
• Only if sells in day-ahead forward market can it be paid

the day-ahead price, but only for quantity sold in day- ahead market and not for actual production

• If default price that all consumers pay is real-time price,

this will foster investment in automated and human intervention-based demand response

• Automated demand-side participation in wholesale market

can help overcome regulatory barriers to symmetric treatment of load and generation

• Customer need not know day-ahead or real-time prices
• Day-ahead behavioral response to retailer signals
• Real-time automated response to retailer signals

13

Monthly Average Off-Peak Period Prices for 2009 (SCE LAP) Monthly Average Peak Period Prices for 2009 (SCE LAP) Quarterly Real-Time Price Duration Curves for 2009 SCE LAP Quarterly Real-Time Price Duration Curves for 2009 (SCE LAP)

14

Day-Ahead versus Real-Time Dynamic Pricing

• Even during a year with a depressed economy and mild weather, there

were a number of periods with very high real-time prices

• With symmetric treatment of load and generation and automated

response technology, shifting demand away from certain periods can yield significant cost savings

• Buy energy at $50/MWh in day-ahead market and sell it back at

$2,000/MWh in real-time market

• Most volatile prices are near major load centers
• California retailers are currently able to buy at Load Aggregation

Point (LAP) prices averaged over large geographic areas covered by three investor-owned utilities

• This is likely to end in the near future
• Wolak, Frank (2010) “Quantifying the Benefits of Spatial versus Temporal

Granularity in Retail Electricity Pricing,” on web-site

• Wolak, Frank (2011) “Measuring the Benefits of Greater Spatial Granularity in

Short-Term Pricing in Wholesale Electricity Markets”

• Supply-side benefits of greater spatial granularity in pricing in California

Managing Short-term Price Risk

• Retail customer purchases analogue to cellular

telephone “calling plan” for electricity consumption

– Fixed price contract for fixed quantity of energy – Examples

• 7x24 for 2 KWh at 10 cent/KWh
• 6x16 for 0.5 KWh at 12 cents/KWh
• 5x4 for 0.5 KWh at 15 cents/KWh

– This yields a load shape that approximates customers actual consumption

• Customer only exposed to real-time price for

deviations from this load shape, upward and downward

54

Load Profile: Purchased and Consumed Weekly Consumption Monday to Sunday

55

Conclusions 1

• Default real-time pricing for all consumers maximizes

benefits of smart technologies

– Makes day-ahead dynamic pricing, storage and automated load shifting technologies financially viable – No customer needs to pay real-time price for any consumption,

• nly face it as a default price, just like in all other markets
• Default fixed price increases average prices to

consumers or increases risk of retailer bankruptcy

– Does not protect consumers from paying volatile wholesale prices

• Allow consumers to purchase fixed load shape at a fixed

prices, not all they want at a fixed price

– Consumers buy and sell deviations from fixed load shapes in day-ahead and real-time markets to minimize bill risk

– Similar to cell phone model

• Purchase total monthly minutes at fixed price in advance
• Real-time price per minute for consumption above total monthly minutes
• Rollover of unused minutes similar to selling unconsumed contract quantity in day-ahead
• r real-time market

15

Conclusions 2

• Whether symmetric treatment is implemented in

most US states is still uncertain

– Significant regulatory barriers in all but one state – Information provision is key to successful implementation

• Ongoing experiments at PESD on information provision
• Texas is test case for potential benefits of

symmetric treatment

– Large renewable energy share

• Wind in west Texas

– Interval meters are currently being installed – Full retail competition allowed – Default wholesale price for all consumers is real-time price

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

58