5. Vorlesung Energiewirtschaft II: Risk Management and Electricity Trade

Size: px

Start display at page:

Download "5. Vorlesung Energiewirtschaft II: Risk Management and Electricity Trade"

Lindsey Ford
5 years ago
Views:

1 5. Vorlesung Energiewirtschaft II: Risk Management and Electricity Trade Georg Zachmann V

2 Agenda of Today's Lecture 1) Organizational Issues 2) Summary of Last Weeks Findings 3) Market Efficiency 1) Cross-Border Electricity Trade 2) Cross Commodity Arbitrage 3) Intertemporal Arbitrage 4) Solution of the Test Exam 5) Assignment 3 6) Exercise 5: Cross Border Efficiency - 2 -

3 Organizational Issues Klausur: -Tachenrechner -Normalverteilungstabelle - 3 -

4 Summary of Last Weeks Findings 1) Organizational Issues 2) Summary of Last Weeks Findings 3) Market Efficiency 1) Cross-Border Electricity Trade 2) Cross Commodity Arbitrage 3) Intertemporal Arbitrage 4) Solution of the Test Exam 5) Assignment 3 6) Exercise 5: Cross Border Efficiency - 4 -

5 Forecasting Electricity Prices 1) Model Selection 2) Data Reduction 3) Parameter estimation / Calibration for reduced data 4) Forecast reduced data 5) Re-extent data + 6) Option Pricing - 5 -

6 Cross-Border Electricity Trade 1) Organizational Issues 2) Summary of Last Weeks Findings 3) Market Efficiency 1) Cross-Border Electricity Trade 2) Cross Commodity Arbitrage 3) Intertemporal Arbitrage 4) Solution of the Test Exam 5) Assignment 3 6) Exercise 5: Cross Border Efficiency - 6 -

7 UCTE Map - 7 -

8 Congestion Prevents Price Convergence - 8 -

9 Congestion Management Systems 1) Nodal Pricing 2) Implicit Auctions (Zonal) 1) Market Coupling 2) Market Splitting 3) Explicit Auctions - 9 -

10 Nodal Pricing 1 -Each injection and/or withdrawal point of a network is called a node -The physical grid is owned by one or multiple companies -The grid is managed by an independent system operator -At each node all interested parties might submit offer and demand curves -Then all bids and offers at each node are aggregated M D 50 MW 100 MW 10 MW G

11 Nodal Pricing 2 -An optimization software solves the system with respect to various restrictions: - Line capacity - N-1 criterion - Block bids - -The result specifies: - Production and consumption at each node - Prices at each node - Flows -Investment incentives: - Prices at the nodes might differ significantly, thus stipulating better generation and/or consumer allocation - If flow = capacity (i.e. capacity is binding) the shadow price for the line utilization is paid to the line owner. This partly incentives line investments

12 Implicit Auctioning I -A group of nodes is assembled in zones. -Each node of a zone has the same price. -Again, demand and supply bids of each zone are collected by a zonal authority. -Now a central dispatch center has two possibilities: - Market coupling - Market splitting -Market coupling: - The optimization program couples zones such that they have an equal price as long as this produces no congestion. - There often is no single solution, e.g. because of - A center zone could be either coupled with the north zone or the south zone but not with both - Block bids - Inner zonal flows -Market splitting - The entire market is viewed as one zone with an single price. - If this simple solution is infeasible i.e. it would produce congestion, the market is split in sub-zones until a solution is reached

13 Implicit Auctioning II -To optimize the system, virtual lines between the zones (with estimated capacity constraints: the so called NTC) have to be calculated and taken into account -Result in both cases - Multiple nodes are grouped in zones with equal prices - More liquid markets than in nodal pricing - In perfect markets : Less optimal dispatch than nodal pricing Zone 1 50 MW 90 MW 100 MW 10 MW Zone 2 -Market participants trade Contracts for Difference (CdF)

14 Explicit Auctioning I -The market for electricity and the market for transmission are separated. -First, the market participants bid for transmission capacities in an auction. -Second, the market participants can buy and sell electricity in the respective markets, subject to their national production and consumption as well as their capacity holdings. -Simple example: - Trader A buys the right to transport 500 MWh for the 13th hour of the 29 June 2006 from Germany to the Netherlands ( Jun) - He now sells 500 MWh at the APX ( Jun) - He then buys 500 MWh for the 13th at the EEX ( Jun) -The main challenge for the grid operators is to offer the correct amount of capacity rights for auctioning - Too much rights => lines are congested and the system operator is responsible for potentially expensive rescheduling - Too few rights => the potentially valuable rights remain unused. - Doubtful: which incentive has an integrated company too offer as much capacity as possible?

15 Explicit Auctioning II -Better then othertypical transitional arrangement for rationing congested lines: - First come first serve - Pro rata -Illiquid markets -Capacities remain unused or are even misused -Prices are no reliable signal -The current system has no sustainable investment incentives - Example: 2005 the annual and monthly Polish-German auction attained ~68 Mio.. Building a 100km 380 kv line that would be able to reduce this bottleneck would cost around 90 Mio.. But nobody has an interest to do so

16 System Comparison Allocation Efficency (flows) Economic Pricing Transmission Investment Incentives Location Signals for Generation & Consumption Market Power Concerns Nodal Pricing Implicit Auction Explicit Auction

17 The European Reality

18 Why we do not have nodal pricing? -Deep system changes needed: - Separation of grid ownership and system management (TSO vs. ISO) - A new European System Operator (ESO) has to be founded. - The ESO has to seize the right to dispatch the system from the former TSOs. - Combination of system operation and electricity exchange - The existing electricity exchanges (spot) have to merge or to disappear - The new European power exchange has to go together with the ESO -Many property rights affected -Many national political decisions to coordinate -Much political opposition from the potential losers: - Existing TSOs - Existing Electricity Exchanges - Generators with market power - Consumer in low price countries

19 What is a Dutch-German Arbitrageur doing? (Explicit Auction) 1) He bids for transmission rights in the potentially profitable direction. His bid price should be E(p APX -p EEX )-ε. Sealed bid uniform price auction 500 MW à 10 /MWh 2) He is informed, that he obtained the desired capacity. 3) He sells electricity at the Amsterdam power exchange. His sell price should be E(p EEX )+ε. Sealed bid uniform price auction 500 MW à 80 /MWh 4) He is informed that he obtained the desired capacity. 5) He buys electricity at the EEX. His bid price should be p balancing -ε.* Sealed bid uniform price auction 500 MW à 60 /MWh The profit of the trader is thus 500 x (80 (60+10)) = The maximum loss is V x (p capacity + (p balancing p APX )). This could be significantly reduced by holding a portfolio of financial and physical rights in both countries which could be timely adjusted. *p balancin is the price for having to buy electricity in the Dutch balancing market

20 Timing of auctions Cross Border Auctions Power Exchanges EON- TenneT (EEX-APX) EON- ELTRA (EEX-DKW) APX Elspot (NordPool) EEX Available capacity known to market participants 8:30 9:00 End of bidding 9:00 9:30 10:30 12:00 12:00 Publication of Results 9:30 10:00 11:00 12:00 12:

21 What should be the market outcome? In the case of perfect full information markets there are two cases: -No congestion: p capacity = p APX p EEX = 0 -Congestion: p capacity = p APX p EEX > 0 -Otherwise a fully informed trader could do risk free arbitrage

22 What happens in imperfect markets? Market power: the ability of a company to profitably move prices away from the competitive level. A is a generation monopolist in the North with low cost and high prices. B is a competitive generator in the South with high cost and marginal cost prices. Monopolist Price North c A Competitive Price c B Monopolist Price South Monopolist Trader Things get more complicated if we allow for simultaneous trade in both directions. If there is market power in one country, the owner of market power might for example be interested in withholding cross-border capacities

23 What does reality teaches us? Electricity often flows from high to low price E.on and RWE auction have fairly similar prices EON =>TenneT RWE =>TenneT

24 Law of One Price Time variant coefficient framework with and white noise process

25 Time variant coefficient (α t ) EEX-APX 3h: Smooth convergence EEX-APX 13h: Outlier dominate the picture

26 Proximity indices (γ t ) and convergence indicator (θ) -Convergence towards arbitrage-freeness in off-peak. -No significant convergence in peak due to outliers

27 Cross Commodity Arbitrage 1) Organizational Issues 2) Summary of Last Weeks Findings 3) Market Efficiency 1) Cross-Border Electricity Trade 2) Cross Commodity Arbitrage 3) Intertemporal Arbitrage 4) Solution of the Test Exam 5) Assignment 3 6) Exercise 5: Cross Border Efficiency

28 Cross-Commodity Arbitrage -Hedging fuel price risks for generators - Spread (units: /MWh el - /MWh th / MWh el /MWh th = /MWh el ) - Spark Spread (P el - p Gas / efficency Gas ) - Dark Spread (P el - p Coal / efficency Coal ) -Use correlations to hedge financial portfolios - Sometimes there are no derivates on our underlying available that feature the desired properties. Historical data allow to find other commodities that are strongly correlated (positive or negative) with our underlying. - Now one can go long or short in those commodities (or related derivatives) to hedge the portfolio. -Make arbitrage gains - In imperfect markets it might be possible to buy and/or sell derivates on different commodities such that the portfolio remains mainly hedged but arbitrage gains are attained

29 Hedging fuel price risks for generators There are many forms of contracts that help generators to reduce price risks: -Contract for difference (CfD) is an assurance against interregional price differentials -Dirty (Spark or Dark) Spread is the fuel cost - electricity price difference -Clean (Spark or Dark) Spread is the fuel and emission cost electricity price difference -Spreads exist as weakly standardized contracts or as individual contracts (of e.g. big traders and small generators) -Spread contracts allow generators to stabilize their cash flow. This is important to assure banks of the profitability of new investment projects

30 Use correlations to hedge financial portfolios What assets are correlated with German electricity prices? -Stocks of electricity generators (e.g. RWE AG) -Stocks of electricity consumer (e.g. Thyssen Krupp) -Stocks of substitute provider (e.g. Gazprom) -Stocks of fuel provider (e.g. RAG) -Electricity intense commodities (e.g. aluminum) -Fuels (e.g. gas / coal / uranium / EUAs) -foreign electricity prices (e.g. PNX or APX) -cross border capacity (e.g. Annual Dutch-German auction)

31 Make arbitrage gains Use the aforementioned cross-correlations of assets to construct a portfolio that lies above the return-risk line of capital market products. This is only possible in imperfect/immature markets

32 Intertemporal Arbitrage 1) Organizational Issues 2) Summary of Last Weeks Findings 3) Market Efficiency 1) Cross-Border Electricity Trade 2) Cross Commodity Arbitrage 3) Intertemporal Arbitrage 4) Solution of the Test Exam 5) Assignment 3 6) Exercise 5: Cross Border Efficiency

33 Intertemporal Arbitrage 1) Buy or sell forward/future 2) Sell or buy spot Again, in perfect markets, no arbitrage gains could be attained. The literature showed however, that electricity markets are still imperfect

34 Solution of the Test Exam 1) Organizational Issues 2) Summary of Last Weeks Findings 3) Market Efficiency 1) Cross-Border Electricity Trade 2) Cross Commodity Arbitrage 3) Intertemporal Arbitrage 4) Solution of the Test Exam 5) Assignment 3 6) Exercise 5: Cross Border Efficiency

35 Assignment 3 1) Organizational Issues 2) Summary of Last Weeks Findings 3) Market Efficiency 1) Cross-Border Electricity Trade 2) Cross Commodity Arbitrage 3) Intertemporal Arbitrage 4) Solution of the Test Exam 5) Assignment 3 6) Exercise 5: Cross Border Efficiency

36 Normalized PC Slide 27 in lecture 3: Standardizing i.e. division by std would give all hours the same weight Two possibilities: - Standardize the original data matrix (Data Mean)/STD - Decompose the correlation instead of the covariance matrix (corrcoef() instead of cov()) Result:

37 Spikes

38 Parameter uncertainty

39 Estimate the Option Value With spikes

40 Exercise 5 1) Organizational Issues 2) Summary of Last Weeks Findings 3) Market Efficiency 1) Cross-Border Electricity Trade 2) Cross Commodity Arbitrage 3) Intertemporal Arbitrage 4) Solution of the Test Exam 5) Assignment 3 6) Exercise 5: Cross Border Efficiency

41 Exercise 5 1) Load prices for the German and Dutch electricity market and the daily RWE-TenneT auction results (international.mat) 2) Plot (x-axis: hours, y-axis /MWh) 1) The average price differentials for all hours, 2) The average capacity prices for all hours 3) And the average remaining import and export arbitrage opportunities for all hours 3) Plot remaining arbitrage opportunities over time for the 3rd and 13th hour (x-axis time, y-axis /MWh) 4) Calculate how often the Auction was overpaid and how often it was underpaid by more than 5% (10%)

Wholesale power market challenges:

EU Electricity Market Reform Seminar Dublin 13 March 2013 Wholesale power market challenges: from simplicity and efficiency to complexity and regulation Peter Styles European Federation of Energy Traders