THE BENEFITS OF DYNAMIC PRICING IN MASS MARKETS Ahmad Faruqui, Ph. D. Carnegie Mellon University Pittsburgh, Pennsylvania March 13, 2007 Précis • Many utilities are considering the system-wide deployment of smart meters ► ► They can improve utility operations and the cost savings can cover a substantial portion of the multi-million dollar investment However, depending on the utility, the “gap” between operational benefits and AMI costs may still be quite large • One way of bridging the gap is to use smart meters as a means of providing “smart” prices to customers that would induce demand response, obviating the need for expensive peaking capacity and energy • As a bonus, smart pricing would eliminate an important inequity in existing rate designs ► Consumers who use relatively less power during expensive peak periods subsidize others 2 The inequity in flat rates may amount to $4 billion dollars for a state with 10 million customers Load Shapes by Customer Type (10 million customers total) Consumption (kWh per Month) 600 500 Peak Rates Peak Peak 400 Flat = $0.10/kWh 300 200 Peak = $0.20/kWh Off-Peak Off-Peak Off-Peak 100 Off-Peak = $0.067/kWh 0 Flat Average Peaky Amount of Cross-Subsidy Per customer = $10/month Total per month = $33.3 million NPV (10 years) = $3.9 billion 3 Smart prices allow for risk sharing between suppliers and consumers Supplier Risk Flat Rate Seasonal Rate TOU CPP-F CPP-V VPP RTP Consumer Risk 4 Critical-peak pricing (CPP) is by far the most popular design • It is essentially a time-of-use (TOU) rate on most days of the year • When the power system encounters critical conditions, the peakperiod price rises to much higher but known levels, either on a dayahead or day-of basis • In variable critical-peak pricing (VPP), the critical-peak price rises to an unknown level that reflects actual market conditions • Both of these rate designs approximate real-time pricing (RTP) rates and are easier for mass market customers to deal with 5 A CPP rate will provide customers with substantial opportunities to save money Current Residential Rate vs. Cost-Based CPP/TOU All-In Rate 1.00 0.90 0.80 $0.90934/kWh Current Rate Critical Peak New Rate Rate ($/kWh) 0.70 0.60 0.50 0.40 0.30 0.20 0.10 $0.14807/kWh $0.14824/kWh $0.11312/kWh 0.00 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Hour of Day 6 For a vast majority of summer hours, the customer will save money Price Duration Curve 1.20 Rate ($/kWh) 1.00 Cost-Based CPP/TOU Rate 0.80 0.60 0.40 Current Rate 0.20 0.00 0 500 1000 1500 2000 Number of Hours in Summer Period 7 2500 Dynamic prices have had a substantial impact in a hot climate such as California’s Central Valley Figure 11 Hourly Load Shape - Complex Daily Share Model - Zone 4 2.5 kW Load 2.0 1.5 1.0 <-Peak Period-> 0.5 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Hour Control 4 Treatment 4 8 They produce an impact even in a mild climate such as San Francisco’s Figure 8 Hourly Load Shape - Complex Daily Share Model - Zone 1 0.9 0.8 0.7 kW Load 0.6 0.5 0.4 0.3 <-Peak Period-> 0.2 0.1 0.0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 Hour Control 1 Treatment 1 9 The higher the price, the greater is the drop in peak usage, with the reduction varying with market segment Residential Customer Response Curves Decrease in Critical Peak Demand 0.0% -5.0% -10.0% -15.0% -20.0% -25.0% -30.0% -35.0% 0.00 CAC w/ Technology CAC Average Non-CAC 0.20 0.40 0.60 Critical Peak Rate ($/kWh) 10 0.80 1.00 Load reductions can be enhanced through enabling technologies Type of technology 60 50 40 30 20 10 0 Percent drop in critical peak load Smart Meter Smart Thermostat 11 Gateway Systems Weighted Average Under traditional ratemaking, 50% of the customers would be worse off under dynamic pricing Distribution of Bill Impacts 20% Electricity Bill Increase (Decrease) 15% 10% 5% 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% -5% -10% Customers with Flatter Consumption Customers with Peakier Consumption -15% Percentile of Customer Base 12 100% That fear may keep customers from even trying out the new rates • And fear of that fear may keep us from even offering dynamic pricing to customers, since we are anxious to “protect the customers from themselves” • How do we break out of this bubble? 13 Enter the risk premium • Flat rates embody an implicit but very real risk premium that insures customers against price volatility • The risk premium is proportional to the volatility of loads, the volatility of spot prices and the correlation between loads and spot prices ► Thus, if load volatility is 0.2, price volatility is 0.6 and price-load correlation is 0.4, the risk premium is about 5% • π = exp( σL. σP. ρL,P ) Where: ► ► ► ► π = Risk Premium σL = Load Volatility σP = Spot Price Volatility ρL,P = Correlation Between Load and Spot Price 14 A Monte Carlo simulations suggest that a 3% risk premium is a conservative estimate Probability Distribution of Risk Premium 8% 3% risk premium 6% 5% 4% 3% 2% 1% 0% 0% 2% 4% 6% 8% 10% 12% 14% 16% 18% 20% 22% 24% 26% 28% 30% 32% 34% 36% 38% 40% 42% 44% 46% 48% 50% 52% 54% 56% 58% 60% 62% 64% 66% 68% 70% Probability of Occuring 7% Risk Premium 15 After crediting for the risk premium, dynamic pricing rates become attractive for 70% of customers Distribution of Bill Impacts 20% Revenue Neutral Electricity Bill Increase (Decrease) 15% Risk Adjusted 10% 5% 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% -5% -10% -15% Customers with Flatter Consumption Customers with Peakier Consumption -20% Percentile of Customer Base 16 100% Enter demand response • There is substantial evidence that dynamic pricing will lower critical peak loads by more than 10% per average household • The bigger the household’s monthly consumption level, the more will be the load drop • Customers in hot climate zones will exhibit the most demand response 17 After crediting customers with the risk premium and demand response, we can attract over 95% of customers Distribution of Bill Impacts 20% Revenue Neutral Electricity Bill Increase (Decrease) 15% Risk Adjusted Load Shifting 10% 5% 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% -5% -10% -15% -20% Customers with Flatter Consumption Customers with Peakier Consumption -25% Percentile of Customer Base 18 100% Aggregate MW impacts and financial benefits depend on the number of participating customers • The participation rate will depend on the deployment scenario and marketing strategy • A mandatory scenario will generate the highest number of participants, followed by an opt-out scenario (around 70-90 percent) and finally by an opt-in scenario (from 10 to 30 percent) • In all cases, the CPP rate needs to generate substantial bill savings for customers 19 Impacts vary by utility size and location Impact on Four Representative U.S. Utilities MW or Millions of 2007 Dollars 2,500 2,000 MW $ 1,500 MW $ 1,000 MW 500 $ MW $ 0 Large Southeastern Large Western Midsize Mid-Atlantic 20 Small Midwestern Additional reading • Brattle Group, The. “Quantifying the benefit of demand response for PJM,” prepared for PJM Interconnection LLC. and MADRI, January 2007 • Faruqui, Ahmad. “Breaking out of the bubble: how dynamic pricing can mitigate rate shock,” Public Utilities Fortnightly, March 2007. • Federal Energy Regulatory Commission (FERC), The US. “Demand Response and Advanced Metering,” Staff Report, August 2006 • North American Electric Reliability Corporation (NERC). “2006 Long-Term Reliability Assessment,” October 16, 2006. • Plexus Research, Inc., “Deciding on Smart Meters,” Edison Electric Institute, September 2006. 21 Contact information Ahmad Faruqui, Ph. D. Principal The Brattle Group 353 Sacramento Street, Suite 1140 San Francisco, CA 94111 Voice: 415.217.1026 Fax: 415.217.1099 Cell: 925.408.0149 Email: Ahmad.Faruqui@Brattle.Com 22