Question? Leave a message!




An introduction to electricity markets and derivatives

An introduction to electricity markets and derivatives 38
An introduction to electricity markets and derivatives Lecture 1 Electricity markets University of Oslo Department of Mathematics Ren e A d EDF RD Finance for Energy Market Research Centre Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 1 / 54Agenda 1 Electricity features Nonstorability Transport 2 Electricity markets microstructure Intraday market Dayahead market Forward market 3 Derivatives Risk management Power plants tolling contracts Energy storage swing contracts Other derivatives 4 Conclusion 5 References Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 2 / 54Disclaimer Disclaimer Any views or opinions presented in this presentation are solely those of the author and do not necessarily represent those of the EDF group. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 3 / 54Electricity features Electricity features Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 4 / 54Electricity features Main electricity features A local commodity Electricity is nonstorable. Electricity transport satis es speci c laws. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 5 / 54Electricity features Nonstorability Storability Comments Generally, subscribed capacity consumption exceeds installed capacity. Example in France: 250 GW of subscribed consumption capacity vs 128 GW installed capacity. Present best way to store large volume of power: hydroreservoir. Limited by pumping rate 0.74. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 6 / 54Electricity features Nonstorability Consequences On shortterm (next hours) A too long excess of demand compared to production may rst resolves in a decrease of frequency, ... and if not properly corrected, may lead to dramatic blackouts. July 30th, 2012: India, 670 millions people. August 13th, 2003: Ontario and North America, 50 millions people. November 4th, 2006: UCTE, 15 millions people. ) Minute by minute realtime assessment of the equilibrium between consumption and production. The Transport System Operator (TSO) is responsible for the electric system security and reliability. He manages the uncertainties on demand and production by a serie of operating reserves. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 7 / 54Electricity features Nonstorability Reserves Operating reserves An operating reserve is a generation that can be mobilised with a shortterm noti cation. Operating reserves vary by response time. Three kinds of reserve: primary reserve: response time 20s. Automatic devices.  500 MW in France. secondary reserve: response time 3mn. Automatic.  600 MW in France. tertiary reserve: response time 15mn. Manuel.  1,500 MW in France. The volume of each reserve may vary depending on the nature of the uncertainties on a particular electric system. p They tend to grow in T , where T is the time to mobilisation. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 8 / 54Electricity features Nonstorability Consequences On midterm basis Reliability assessment analysis: is there enough capacity to full ll demand in the next months within a certain default probability Means: changing planned outtage schedule, buy on the market, demandside management policy Extreme way: load shedding. On longterm basis Build new capacity to allow enough excess capacity. Demandside management policy. Sound tari cation. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 9 / 54Electricity features Transport Transport Comments The transport of electricity satis es Kirchho 's laws. The intensity at each node should be zero and the tension in each loop should be also zero. Consequences In a meshed electricity network, power will go from one point to another using all available paths. ) Electricity ow interference. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 10 / 54Electricity features Transport Interference between commercial ows and physical ows Situation A power producer G1 has client in node C whose consumption is 180 MW, while a power producer G2 has also a client in node C whose consumption is 90 MW. Each producer holds enough generation capacity and no production cost advantage. G1 90 180 G2 C 90 Figure : Network physical capacity limits. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 11 / 54Electricity features Transport Interference between commercial ows and physical ows G1 G1 0 60 30 180 120 + 30 G2 C G2 C 90 60 + 60 Figure : Desired commercial ows. Figure : Physical ows. Congestion Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 12 / 54Electricity features Transport Transport consequences on trading Consequences ) Crossborder trading opportunities. Transfert capacities available for trading between countries need some generation hypothesis. In Europe, the available net transfert capacity (NTC) are managed and published by the ENTSO (European Network System Operator) Publicly available in her transparency platform (www.entso.net). Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 13 / 54Electricity markets microstructure Electricity markets microstructure Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 14 / 54Electricity markets microstructure Common market structure for a local commodity Comments Electricity is a local commodity. As many electricity market as they are states: USA: Europe: South America: Asia: Paci c: Market microstructure highly depends on national regulation. Nevertheless, common structure emerges driven by the necessary equilibrium between consumption and production. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 15 / 54Electricity markets microstructure Common market structure A sequence of markets ordered by timehorizon The intraday market and/or balancing mechanism The dayahead market The forward market Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 16 / 54Electricity markets microstructure Intraday market Intraday Commons Ensure the security of the system Balancing Mechanism. Transparent market price for the cost of imbalance Imbalance Settlement Price. Remark May coexist at the same a market for next hours where rms exchange power. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 17 / 54Electricity markets microstructure Intraday market Balancing mechanism Example French TSO adjustment market mechanism as of April, 2013. Balance Responsible Entities (BR) submit bids and o ers to increase or decreases their production (or consumption). TSO selects o ers based on economic precedence. BR are paid as bid. Every power available plant should be declared on the adjustment market Producers declare their price to increase their production System operator uses all these o ers to insure realtime production consumption equilibrium But, some time later, each balance responsible entity receives the bill of her imbalances... Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 18 / 54Electricity markets microstructure Intraday market Balancing mechanism Example: French TSO imbalance price settlement mechanism S represents the dayahead price settled the day before for the hour of interest. d P is the average price of the o ers used by the TSO on the balancing mechanism to decrease the production (or increase the consumption). u P is the average price of the o ers used by the TSO on the balancing mechanism to increase the production (or decrease the consumption). Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 19 / 54Electricity markets microstructure Intraday market Imbalance mechanism Network Network Adjustment Trend Adjustement Trend Positive Negative Actor Imbalance Positive   d P Actor is paid S min S; 1 +k Actor Imbalance Negative u Actor pays max (S;P  (1 +k)) S Lecture Network needs upward adjustment Actor is producing too much Actor d P is paid but not more than S. 1+k Network needs upward adjustment Actor is producing not enough Actor u pays P  (1 + k) and at least S. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 20 / 54Electricity markets microstructure Intraday market Balancing mechanism 450 400 350 300 250 PMPH PMPB 200 150 100 50 0 1 49 97 145 193 241 289 Weigthed Average Upward and Downward adjustment price in French power market from january 4th, to 10th 2010. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 21 / 54Electricity markets microstructure Intraday market Imbalance prices 160 140 120 100 80 60 40 20 0 1 49 97 145 193 241 289 Imbalanced Settlement Prices (Upward and Downward) in French power market from january 4th, to 10th 2010. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 22 / 54Electricity markets microstructure Intraday market Intraday market Note Beside this balancing mechanism, an intraday market for energy delivery for the hours of the day or of the next day exists. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 23 / 54Electricity markets microstructure Intraday market Epex intraday market prices 400 350 300 250 200 150 100 50 0 50 13/12/2011 01/02/2012 22/03/2012 11/05/2012 30/06/2012 19/08/2012 08/10/2012 27/11/2012 16/01/2013 Figure : Epex intraday hourly prices during 2012. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 24 / 54Electricity markets microstructure Dayahead market Dayahead market mechanism Example of Epex spot Based on a xed trading auction. Participants submit bids before a certain time (around 10:00). Bids can concern a particular hour of the next day or a set of hours (order block). bids of market participants for a particular hour form a bid curve because she can submit a list of prices and quantities. Market organizer clears the market: she xed a price for each hour of delivery and determines the seller and the buyers. Market players have then enought time to send production orders to their power plants and send their schedule to the TSO. Note: the clearing process results in a nonconvex optimization problem (block orders), for which de ning a market price requires caution. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 25 / 54Electricity markets microstructure Dayahead market Dayahead market mechanism Transcountries trading In continental Europe, each country has its own electriciy dayahead market cleared by her own rm. Without coordination, resulting quoted prices may provide the wrong signal when compared to transit ow between countries. Example given between France and Germany: ows would not follow spot prices di erence between countries. Since quoted dayahead prices by market organizer have a transparency function, mechanisms have been developped to ensure a consistent relation between crossborder transactions and local dayahead prices. Market coupling: performing implicit auction mechanism. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 26 / 54Electricity markets microstructure Dayahead market Dayahead market mechanism Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 27 / 54Electricity markets microstructure Dayahead market Market coupling Implicit auction mechanism In each country, market participant do not have to care about nding a counterparty in neighboring countries. She has juste to submit her bid in her country (sell or buy). Market organizers perform a clearing process with transport constraints implied by the available transfert capacity (ATC). If there is no binding transit capacity constraints, then there will be a single price for the clearing area. If there is at least one binding transit capacity constraint, two prices will emerge. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 28 / 54Electricity markets microstructure Dayahead market Dayahead market price exhibits seasonnalities Daily and weekly seasonality. Epex hourly spot price, january, 4th to 10th 2010. 120 100 80 60 40 20 0 1 25 49 73 97 121 145 Exhibits also, annual saisonnality. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 29 / 54Electricity markets microstructure Dayahead market and strong dependence with consumption Spot price (eur/MWh) Demand (GWh) 80 60 40 20 Mon Tue Wed Thu Fri Sat Sun Mon Example of Epex spot ong Janaury, 2012. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 30 / 54Electricity markets microstructure Dayahead market Dayahead market price exhibits spikes and negative prices EEX dayahead price. 350 300 250 200 150 100 50 0 50 100 07/07/2005 19/11/2006 02/04/2008 15/08/2009 28/12/2010 11/05/2012 23/09/2013 Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 31 / 54Electricity markets microstructure Dayahead market Relation between intraday and dayahead prices 1200 1000 800 600 400 200 0 −200 −500 0 500 1000 1500 2000 Epex dayahead 2012 Figure : Epex intraday and dayahead hourly prices during 2012. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 32 / 54 Epex intraday 2012Electricity markets microstructure Dayahead market Comments Extreme situations High dayahead price and low intraday: uncertainty resolved. Low dayahead and high intraday: very shortterm uncertainty realisation. Dayahead market prices is refered as the spot price Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 33 / 54t t 0 1 2011 2012 Q1 Q2 Q3 Q1 Q2 Q3 1 2 3 4 5 1112 1 2 3 Zoom 15/12/2010 25/10/2011 Temps October 2011 November 2011 W1 W2 W3 W4 W W1 1 W W2 2 W W3 3 W4 25/10/2011 Electricity markets microstructure Forward market Forward market nested contract structure Temps 2011 2012 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 2010 2013 1 2 3 4 5 6 7 8 9 101112 1 2 3 4 5 6 7 8 9 101112 Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 34 / 54t 1 2012 Q1 Q2 Q3 1112 1 2 3 Zoom 25/10/2011 Temps October 2011 November 2011 W1 W2 W3 W4 W W1 1 W W2 2 W W3 3 W4 25/10/2011 Electricity markets microstructure Forward market Forward market nested contract structure t 0 Temps 2011 2011 2012 Q1 Q1 Q2 Q2 Q3 Q3 Q4 Q1 Q2 Q3 Q4 2010 2013 1 1 2 2 3 3 4 4 5 5 6 7 8 9 101112 1 2 3 4 5 6 7 8 9 101112 15/12/2010 Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 34 / 54Zoom Temps October 2011 November 2011 W1 W2 W3 W4 W W1 1 W W2 2 W W3 3 W4 25/10/2011 Electricity markets microstructure Forward market Forward market nested contract structure t t 0 1 Temps 2011 2011 2012 2012 Q1 Q1 Q2 Q2 Q3 Q3 Q4 Q1 Q1 Q2 Q2 Q3 Q3 Q4 2010 2013 1 1 2 2 3 3 4 4 5 5 6 7 8 9 1011 1112 12 1 1 2 2 3 3 4 5 6 7 8 9 101112 15/12/2010 25/10/2011 Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 34 / 54Electricity markets microstructure Forward market Forward market nested contract structure t t 0 1 Temps 2011 2011 2012 2012 Q1 Q1 Q2 Q2 Q3 Q3 Q4 Q1 Q1 Q2 Q2 Q3 Q3 Q4 2010 2013 1 1 2 2 3 3 4 4 5 5 6 7 8 9 1011 1112 12 1 1 2 2 3 3 4 5 6 7 8 9 101112 Zoom 15/12/2010 25/10/2011 Temps October 2011 November 2011 W1 W2 W3 W4 W W1 1 W W2 2 W W3 3 W4 25/10/2011 Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 34 / 54Electricity markets microstructure Forward market Available forward contracts Example EEX Are available at the same time: 6 calendars 11 quarters 9 months 4 weeks 2 weekends 8 days In three avours: baseload (each hour), peakload (07:0020:00 Monday to Friday) and o peak (complementary to peakload, not available for weeks, weekends and days). Thus, 106 contracts are available Compare with the 525,684 hours in the next six years. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 35 / 54Electricity markets microstructure Forward market Forward contract settlement Settlement Electricity forward contracts implies a delivery during a period of time. Delivery of power every hour of the week, month, quarter or year (base load) or a set of hours frome Montay to Friday (peak load). Possible settlement at maturity or continuously during the delivery period. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 36 / 54Electricity markets microstructure Forward market German baseload forward curve dynamic Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 37 / 54Electricity markets microstructure Forward market German baseload forward curve dynamic Comments Very di erentiated behaviour between spot, month and yearly contracts. Slow motion of yearly contracts. May exhibit report or deport con guration. Strong seasonal pattern of monthly contracts (blue dots). Dissapears Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 38 / 54Electricity markets microstructure Forward market German baseload open interest curve dynamic Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 39 / 54Electricity markets microstructure Forward market German baseload open interest curve dynamic Comments Close maturities catch all liquidity. Linear growth of closest maturity open interest. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 40 / 54Electricity markets microstructure Forward market FrenchGerman spread curve dynamic Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 41 / 54Who was who France is in blue, Germany is in red. Electricity markets microstructure Forward market FrenchGerman spread curve dynamic Comments Strong dependence. Possible signe inversion with time and maturity. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 42 / 54France is in blue, Germany is in red. Electricity markets microstructure Forward market FrenchGerman spread curve dynamic Comments Strong dependence. Possible signe inversion with time and maturity. Who was who Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 42 / 54Electricity markets microstructure Forward market FrenchGerman spread curve dynamic Comments Strong dependence. Possible signe inversion with time and maturity. Who was who France is in blue, Germany is in red. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 42 / 54Derivatives Risk management Derivatives Risk management Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 43 / 54Derivatives Risk management The name of the game Problems An electric utility is exposed to a whole set of risk factors: Electricity prices Fuel prices: coal, crude, gas, Emission prices Currencies Consumption and market shares Outtages and in ows Climate Two kind of problems: local problems and global one. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 44 / 54Derivatives Risk management The name of the game Local problems Simulating realistic spot prices. Pricing and hedging a particular structured product. Assessing the impact a particular factor on the value of an asset. ... Global problem Designing the models for a whole risk management system for a utility exposed to those risk factors. Need for one or several models to allow realistic representation of prices and their dependencies, fast and robust estimation procedures, calibration to market prices, fast computation of simulated forwards, fast computation of derivatives, fast computation of greeks. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 45 / 54Derivatives Risk management Derivatives Problems Like any other commodity market, electricity markets have their options on quoted futures. But, the most challenging problems comes from the pricing, hedging and structuring of exotic tradable products linked to physical assets. Can be called Real Derivatives. Real Derivatives Embedded options in producers or retailers portfolio. Power plant strip of call options on fuel spread. Tolling contracts) structured contract counterpart. Water reservoir strip of call options on calendar spread. Swings) structured contract counterpart. Demandside management strip of puts. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 46 / 54Derivatives Risk management Power plants tolling contracts Power plants as derivatives First approximation of power plants value Strip of calls on the spread between its fuel price and the electricity price. The value of a power plant on a period of time 0; T would then be given by : " Z T + E (P hS ) dt ; t t 0 where P is the price of power, S the price of its fuel and h its heatrate and + x = max (0; x). Margrabe's 1978 closedform formula for exchange options applies and gives value and Greeks. But, there are some problems with this approximation. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 47 / 54Derivatives Risk management Power plants tolling contracts Diculties with power plants as derivatives Numerics With emission price, value reads: " Z T c + E (P hS gS ) dt ; t t t 0 Margrabe's formula does not apply anymore. But, it still neglects operational constraints: startup cost act as nonzero strike price. Rampup time, minimum power... Taking constraints into account stochastic control problem (optimal switching or singular control). Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 48 / 54Derivatives Risk management Power plants tolling contracts Diculties with power plants as derivatives Riskneutral probability If the market was complete, the probability under which the former expectations are computed would not be ambiguous. At time 0, future prices would exist for each hour t of the year and the value of the power plant would be : " Z T e f c + E (F (0; t) hF (0; t) gF (0; t) dt ; 0 e f c where F (0; t), F (0; t), F (0; t) are respectively the futures price for power quoted at time zero for delivery at hour t, the futures price for its fuel and the futures price for the emission permit. But, such information is not available. In realistic cases, the valuation of a power plant cannot avoid the problem of the incompletness of the electricity market. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 49 / 54Derivatives Risk management Energy storage swing contracts Water reservoir and energy storage Valuation Theoretical diculties of power plants valuation apply also for hydro power plants But, the limited ressource of fuel leads to a problem a storage management. Most simple hydrolic storage management problem deals with a single reservoir. The valuation problem writes " Z T sup E q S du + g(S ; X ) ; u u T T q 20;q t t t;x t;a dX = (a q )ds; s s s with S the electricity spot price, X current level of the water reservoir, a t t s random in ows. X is subject to level constraints and should stay within x; x. The function g represents a nal value for having a certain level of water at a nal time. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 50 / 54Derivatives Risk management Energy storage swing contracts Water reservoir and energy storage Valuation Leads to stochastic control problems Intensive use of Dynamic Porgramming Main diculies: dimension state. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 51 / 54Derivatives Risk management Other derivatives Other derivatives Retail contracts Tari cation policy for di erent nal consumers (industrial, small business and households) Financial risk embedded in a client load curve vs portfolio e ect. Market share vs margin. Weather derivatives Producers nancial risk highly depends on weather. Weather derivatives on temperature, rain, wind procure insurance on bad outcomes. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 52 / 54Conclusion Conclusion Complexity Power systems knowledge. Nested market microstructure. Complex products. Incompletness valuation. Numerical methods for stochastic control problems. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 53 / 54References References Monographies Fred E. Benth, Jurate S. Benth, Steen Koekabakker, Stochastic Modeling of Electricity and Related Markets, World Scienti c, 2008. H elyette G eman, Commodities and Commodity Derivatives: Modelling and Pricing for Agriculturals, Metals and Energy, Wiley, 2005. Alexander Eydeland Krzysztof Wolyniec, Energy and Power Risk Management: New Developments in Modeling, Pricing and Hedging, Wiley, 2002. Les Clewlow Chris Strickland, Energy Derivatives: Pricing and Risk Management, Lacima Group, 2000. Ren e A d EDF RD Finance for Energy Market Research Centre Electricity Derivatives 54 / 54
Website URL
Comment