Dynamic pricing

Transcription

Dynamic pricing
Thanks to technology and institutions,
Electricity demand could become active !
Workshop « Smart grids » Chaire développement durable-X,
21 October 2010
Dominique FINON
Directeur de recherche CNRS, Cired
Introduction
Demand activation : a needed complement to smarting
grid&meters for the cost-benefit of smart meters roll-out
results from Faruqui, Harris and Hledik on Europe , published in J. Energy Policy
(2010) Vol.38
• Costs:
Indicative cost of smart meters roll-out in EU: 51 B €
• Benefits
1. Operational savings for DSO: 26 B€ (estimation deduced from
some US experiences)
2. But ….gap of 25 B€:
To be gained by improvement of the surplus of the
consumers, and suppliers (LES)
If improvement of tariffs by default suppliers:
1.
2.
With simple time of use tarif : 14 B€ of benefits
With dynamic pricing: 68 B€ of benefits (better correspondence of
tariffs with time variation of wholesale price
To be gained by savings on capacity adequacy
Introduction
• Smart metering/ grid needs dynamic pricing to be cost
effective
– Improvement of social efficiency of the system by
activation of demand
– by information on price , quantity and expense
– By dynamic pricing i.e. retail price related to wholesale price
– Many coutries with smart meters have no dynamic pricing
– Caution from regulators to be faced to consumers backlash if
dynamic pricing
Content
•
•
•
•
1.The ways of activation of demand
2. Contracts in dynamic pricing
3. Experiment in dynamic pricing
4. Where gains of surplus by demand
activation are located?
1.The ways of activation of demand
1.1. The very specificity of electricity demand function
• Electricity non storable
– Need of semi hourly markets with weak intertemporal relations at the
wholesale level)
• Electricity consumer is blind
– No price information transfer to consumers:
– No pass-through of wholesale price in real time
– No possibility of revelation of WTP concerning reliability
• No information on consumed quantity transmitted to suppliers in real
time
• For a supplier/retailer , selling a contract :
– similar to a call option contract on unlimited quantity
– making risk management quite complex for retailer
• Consumers could consume the same quantity of electricity when the
price on the hourly market could be as much as 1000 times off-peak price
• If price-quantity information transmission to consumer in
a timely way, consumers will react to price signal.
• So two successive & complementary processes of
demand activation:
– Information acquisition by the consumer: effects on his
behaviour
– Demand becomes reactive to a menu of price /quantity
contracts
• Experiments in the US and in Europe
– On information acquisition
– On dynamic pricing
1.2. Importance of the information acquisition
• Les “display devices” : rôle fondamental dans l’”éducation”
du consommateur et sa réponse aux signaux prix.
•
Ils font varier favorablement les élasticités prix.
• Un grand nombre d’objets communicants ont été testés (Orb, Flower
lamp, Aware cord…).
• Certains objets ont fait leur preuve sur le terrain (Orb, thermostat
communicant, …).
• Peu de données.
• L’hétérogénéité des modes expérimentaux ne permet pas de relier un
display en particulier à une réponse “moyenne” du consommateur.
• US Experiments Faruqui (2009) : effects on 6% max on the energy
consumption
• UK (Sarah Sharby) on info feed back : 11% as an effect on consumption
• OFGEM much more skeptical: effect of 1% with the display on smart
meters
1.3. Dynamic pricing
price
Solution 1:
dynamic/real time
pricing
• Smart meter & Activation of
demand: the key of market
normality
–If the demand will be price
responsive,
–price variablity decreases
–Market power decreases
price
2. Dynamic pricing and contractualization
Time of use pricing (TOU ):
– Retail price that vary with time of day
– But regardless of system conditions
– Does not require hourly meter
Dynamic pricing :
• retail price that vary with real time system and market conditions
• Requires hourly meters to be implemented
Load shaving contracts: (outside demand function (price quantity)
The customer delegates to a supplier or another player the management of his
demand during variable peak period (implicit payment by savings)
Possiblity of delegation:
• to aggregator of load shaving services,
• to distributors ( more and more interesting for DSO if intermittent
sources)
Dynamic pricing versus Time of use pricing
Dynamic pricing
•
•
Time of use pricing
No incentive to to reduce demand
during periods with high wholesale
price and stressed system
Incentive to reduce demand during
periods with high wholesale prices
and stressed system
Reduces price volatility and increase
system reliability
•
Perfectly inelastic hourly demand for
electricity
•
This solution improves partly
collective surplus on long term:
Partial reduction of peak demand on
long term adequate capacity
•
•
This solution improves collective
surplus on short term and long term
•
Predictable and easy to understand
Different types of dynamic pricing
• Purest form : Real time pricing (RTP)
–
Elec prices linked to the wholesale price
(day ahead basis to reduce
uncertainty)
–
Only for industrial consumers able to manage their risks
• Critical peak pricing (CPP) to reflect the true cost during a
small period (12 days)
– Very high prices (Times 5)
– Small Discount for the remaining hours
– Incentives to decrease conumption in peak hours or to shift to less expensive
hours
Different types of dynamic pricing
(following)
• Flexible critical peak pricing
– Designation the day before (in sime case few hours before)
– More efficient than fixed CPP
• A variant : Peak time rebate (PTR):
– Cash rebate for each kWh of load below a baseline
– To establish a baseline
– Payment to the wholesale price (or whlesale price minus retail price)
– Eventually remote control on heating/ air-conditoning devices
Economic problems in implementing dynamic pricing
(Chao, 2010; Bushnell, Hobbs, Wolak, 2009))
• Establishing a baseline for each consumer for the CPP and for
the Peak time rebate:
– Assymetry information between the supplier and the customer
– Credibility of demand reduction under the peak price signal which is a
problem for the supplier’s risk management
•
Economic problems in implementing dynamic pricing
(Chao, 2010; Bushnell, Hobbs, Wolak, 2009))
• The solution is demand subscription: « buying your baseline
approach »
– Identical treatment of the supply and the demand
– Baseline is contractualised with an exogenous formula (proper
incentive rathe than the )
– Consumers « may sell energy » on which they have made a firm
commitment to purchase
– One example : a specific load profile contract/Nominal quantity bis
indexed (with a formula to total demand of the class)
3.2. Experimentation on dynamic pricing
Programs on pilots of dynamic pricing in USA, Australia,
Canada
Programs
DOE Utilities 5 Etats, volontaire, obligatoire
DOE Utilities dans 3 Etats, volontaire
PG&E, volontaire
Midwest power system of Iowa, volontaire
The Gulf Power Selec Program, Floride 2000-2001
Puget sound Energy (Seattle) 2001-2002
Ameren UE- (2004, 2006)
California Statewide Pricing Pilot (SPP) 2003-04
volontaire California Public Utilities Comission
Community Energy Cooperative’s energy-Smart Pricing
Plan (ESPP) 2003-2005 – Illinois, volontaire
Anaheim Critical Peak Pricing experiment, 2005
Tariffs
refereneces
TOU
TOU
TOU
TOU
TOU, CPP
TOU
TOU, CPP
TOU
CPP-fixe
CPP var
RTP
Kohler and Michell (1984)
Caves et al. (1984)
Caves et al (1989)
Baladi et Herriges (1998)
Borenstein, et al (2002)
Faruqui et George (2003)
RLW Analytics AmerenUE
Charles River Assoc (2005)
Energy Australia’s Network Tariff Reform 2006, volontaire
Idaho Residential Pilot Program, Idaho Power Cie,
volontaire, 2006
TOU
TOU, CPP
Daniel Violette Summit Blue
(2006)
University of California
Energy Institute, F. Wolak
Harry Colebourn (2006)
Sanem Sergici (2006)
Olympic Peninsula Project, 2005-2007
TOU, CPP,
RTP
Pacific Northwest National
Lab du DOE (2007)
Ontario Pricing Pilot, 2006-2007
CPP
TOU
Ontario Energy Board
16
Main results
• Incentives by contracts and by tariffs Faruqui (2009)
• With variables prices related to the wholesale prices and the
load demand on the system,
• lowering of consumption and load is estimated between 10%
and 15% during peak hours.
• Addition of automatic control on some electrical uses inside
home
Increase of reduction effect by a factor 2 if possibility to
connect advanced metering devices to different appliances
and heating, air-conditioning (with automatic control)
NB Experiments has been made on ToU and on CPP
Results : heterogeneity of Price Elasticity
Exemple of California SPP
CPP days CAC
Electricity Price Elasticity Estimates - Range and Mass Central
Points ( Absolute Values) for 15 Studies
Points are mass center, lines the values range (where appropriate)
0.45
Estimated Elasticity - Absolute Value
Réduction des pointes : 2 à
25%
Own-price
elasticity
(all others
substitution)
0.4
0.35
0.3
0.25
 Large plage de valeurs de la
0.2
réaction
du
consommateur
(mesurée par l’élasticité prix)
0.15
0.1
0.05
0
1
2
3
Household
TOU
4
5
6
7
8
9
10
11
Study Num ber (Table 4)
Business
HH B
CPP
12
13
14
15
17
18
HH Business
RTP
11 et 12 SPP
28 octobre 2010
16
17
Illinois
Entité d'appartenance
 Consommateurs hétérogènes =>
large dispersion des valeurs
18
4. Where gains of surplus by demand activation
are located?
Where possible gains of surplus by demand
activation are located?
• For the consumers:
– Reduction of expenses on energy during peak and eventually during all
the year
• For the suppliers:
– risk management
– Increase of the competitive intensiveness by new opportunities of
offering a wider set of offers (It favors « switching »)
• For the TSO and the DSO:
– economic saving on capacity adequacy and program of reliability
– increasing stakes of dynamic pricing /smarting DSO grids with
deployment of intermittent sources in windpower and PV
• And again for the consumers
– Reduction of the need of capacity adequacy with the help of peak
shaving
– Reduction of market power during peak and extreme peak
One aspect of surplus sharing by market rules:
How to give a value to demand repsonse program via
dynamic pricing?
– Contracts of load shifting with DSO, TSO, aggregators, suppliers
– Integration in load management resources inside capacity mechanism
(East coast
• NB : 10% in New England and PJM , in the next future France
with the obligation of capacity credit, etc.)
• Direct load control. Load serving entity (LSE) initiates load reduction by
remotely cycling end user equipment.
• Firm service level. End user reduces load to a predetermined level based
on notification from LSE.
• Possibility d’intégrer les effets probabilisés de dynamic pricing