Market monitoring tools David Newbery Southeast Europe Electricity Market Monitoring Workshop Athens 4-5 October 2005 http://www.electricitypolicy.org.uk
The Cambridge-MIT Institute A Review of the Monitoring of Market Power Paul Twomey, Richard Green, Karsten Neuhoff and David Newbery download CMI EP 71from http://www.electricitypolicy.org.uk/pubs/wp.html Part of the research was funded by the Association of European Transmission Operators ETSO. D Newbery Athens 2
Market Power Detection Tools Choose tools suitable for different tasks: Ex-ante versus ex-post analysis Long-term vs. short-term/real time analysis System-level market power vs. local market power vs firm-level market power Horizontal market power vs vertical market power D Newbery Athens 3
Applications of Market Power Detection Tools Long-Term Short-Term Ex-Ante - Merger rulings - Assessing applications for market-based rates - Determining potential must-run generators - requiring contracts - Spot market bid mitigation - Must-run activation & other system operator contracting Ex-Post - Litigation cases (e.g. California refund case) - Changing market design - requiring contracts and VPPs - Short term price recalculations - Penalties for withholding Inspired by a similar table in Helman (2004) 4
Market Power Detection Tools List Behavioral Indices and Analysis Bid-Cost Margins (e.g. Lerner Index) Net Revenue Benchmark Analysis Structural Indices and Analysis Concentration ratios and HHI Residual Supply Index Residual Demand Analysis Simulation Models Competitive Benchmark Analysis Oligopoly Models D Newbery Athens 5
Bid-Cost Margins Lerner Index: Price Marginal Cost LI = Price In a competitive market LI is zero if MC correctly interpreted as scarcity price Cournot oligopoly LI = market share/elasticity Do not require geographic market definitions Is a standard measure of exercise of market power but which MC? Short-run or long-run? D Newbery Athens 6
Market share methods Concentration ratios C1: share of largest firm C3, C4 total share of top 3 or 4 firms Capacity, available capacity, with or without imports (depending how interconnector controlled?) shares of production also revealing C1 > 20% can be a concern but depends on extent of spare capacity D Newbery Athens 7
100 90 80 70 60 50 40 30 20 10 0 Concentration ratios Installed capacity 2001 C1 C1 net of impor capacity C3 Cyprus Malta Albania Latvia Croatia Greece Estonia** Belgium France Ireland Slovakia Slovenia Portugal Czech Italy Lithuania Sweden Spain Denmark Austria* Germany Norway Hungary Finland UK Poland Source: EU 3rd Benchmarking Report D Newbery Athens 8 percent
Herfindahl Hirschman Index: HHI Standard tool for anti-trust (esp. in US) HHI = sum of squared market share %: e.g. 5 firms of 20% each => HHI = 2,000 number of equivalent firms n = 10,000/HHI screens <1000 unconcentrated 1000-1800 moderately concentrated >1800 highly concentrated => serious concern if HHI>1800 and merger raises HHI by more than 100 D Newbery Athens 9
6000 5000 4000 3000 2000 1000 0 HHI by capacity for England and Wales 1990-2002 HHI all fuels and imports HHI for coal capacity Apr-91 Oct-91 Apr-92 Oct-92 Apr-93 Oct-93 Apr-94 Oct-94 Apr-95 Oct-95 Apr-96 Oct-96 Apr-97 Oct-97 Apr-98 Oct-98 Apr-99 Oct-99 Apr-00 Oct-00 Apr-01 Oct-01 D Newbery Athens 10 Apr-90 Oct-90
Market Share and HHI Difficult to determine appropriate geographic region (e.g. SSNIP test, Hub-and-Spoke) Ignores demand side, entry conditions, strategic incentives and often congestion issues Little empirical justification California - under some market definitions, no single supplier in California had a 20% market share during the crises D Newbery Athens 11
Residual Supply Index Measures the extent to which a generator s capacity is necessary to supply demand after taking into account other generators capacity Residual Supply Index continuous variable RSI = Total Capacity Company i's Relevent Capacity Total Demand Sheffrin s screen test: RSI must not be less than 110% for more than 5% of hours per year D Newbery Athens 12
California RSI duration curve June-Sep 2000-2002 all hours Sheffrin (2002) D Newbery Athens 13
130% 125% 120% 115% 110% 105% 100% 95% 90% RSI duration curve GB Winter 1999-2000 0 100 200 300 400 500 600 700 800 900 1000 1100 1200 1300 1400 1500 1600 1700 1800 1900 2000 number of hours RSI lower than D Newbery Athens 14 RSI percent
Significant Correlation between RSI and Price-Cost Markup Sheffrin (2002) D Newbery Athens 15
Spot price vs Residual Supply Index GB Winter 1999-2000 80 70 60 50 Note - marginal cost provides lower bound /MWh 40 30 20 10 0 90% 100% 110% 120% 130% 140% 150% 160% 170% 180% 190% 200% 210% 220% RSI % D Newbery Athens 16
Generation companies have MP within countries... and retain market power due to transmission constraints capacity demand 150% 125% 100% 75% 50% 25% 0% Austria Belgium France Germany Italy Netherlands Switzerland Spain Gen 1 Gen 2 Gen 3 Fringe Imports D Newbery Athens 17 Source: Remaining capacity and availability factor from UCTE Power Balance Forecasts 2002-2004, NTC from ETSO (Winter 2001/2002), National Generation Shares from ICF consulting, Annual reports and presentations
Residual Supply Index Takes account of capacity scarcity Suited to dynamic analysis on an hour-byhour basis and local market power analysis Empirical support of predicting market power Needs access to availability data (from TSO?) Arguably the best tool D Newbery Athens 18
Collective dominance if: Market characteristics conducive to tacit coordination, and Tacit coordination sustainable: firms lack ability and incentive to deviate, given incentives for retaliation, and Buyers, fringe firms, entrants cannot challenge tacit coordination D Newbery Athens 19
Collective dominance criteria Markets concentrated, transparent, mature Low elasticity of demand Homogenous product, similar costs, shares Little excess capacity, barriers to entry Excess pricing, profit little response to cost fall, barriers to switching Electricity as a test case D Newbery Athens 20
/MWh (Jan 2000 prices) 70 60 50 40 30 PG gaming Pool prices since vesting Ofgem price review Nuclear outages reduce plant margin Annual price cap agreed for 2 years French strike & station failure Low availability & Eastern bidding strategy SMP manipulation Capacity withdrawal 20 10 0 Scheduling problems Price falls to meet price cap Plant divestment to Eastern NP & PG manipulation Further plant divestment Apr-90 Apr-91 Apr-92 Apr-93 Apr-94 Apr-95 Apr-96 Apr-97 Apr-98 Apr-99 Apr-00 D Newbery Athens 21
Generation in England and Wales D Newbery Athens 22
40,000 35,000 30,000 25,000 Capacity Ownership of Coal Generation 1990-2004 Offer encourages sale of 6,000 MW NETA live PG & NP trade horizontal for vertical integration ALCAN Innogy National Power British Energy Independent AES MW 20,000 15,000 10,000 Intnl Power Edf TXU/Eastern Powergen SS&E 5,000 AEP Edison 0 Apr-90 Oct-90 Apr-91 Oct-91 Apr-92 Oct-92 Apr-93 Oct-93 Apr-94 Oct-94 Apr-95 Oct-95 Apr-96 Oct-96 Apr-97 Oct-97 Apr-98 Oct-98 Apr-99 Oct-99 Apr-00 Oct-00 Apr-01 Oct-01 Apr-02 Oct-02 Apr-03 Oct-03 Apr-04 Oct-04 Source: NGC Seven Year Statements, various years, and data from J Bower and C Humphries, slide from D Newbery 23
Collective dominance: the GB Electricity Pool Markets concentrated, transparent, mature Low elasticity of demand homogenous product, similar costs, shares little excess capacity, barriers to entry? excess pricing, profit little response to cost fall, barriers to switching?? Need to be able to test for tacit collusion D Newbery Athens 24
Residual Demand Analysis Best response to generator s residual demand Theoretical justification Supply Function Equilibria (locally profit maximising) Requires individual bid data to construct residual demand curves Can detect collusion as well as market power e.g. Wolak, Sweeting, Hortacsu/Puller D Newbery Athens 25
D Newbery Athens 26 Hortacsu and Puller (2003)
Real GB electricity prices and costs 35 Sweeting s periods 7000 30 Restraint Price control Profit maximising Tacit Collusion 6000 25 plant withdrawal 5000 (2001)/MWh 20 15 Electricity NETA 4000 3000 HHI 10 coal cost 2000 5 gas cost Coal HHI 1000 0 0 1990 1991 1992 1993 1994 1995 1996 1997 1998 1999 2000 2001 2002 2003 2004 D Newbery Athens 27
Net Revenue Benchmark Analysis Compares estimated revenues with total costs Assess financial viability and barriers to entry important in presence of price caps Spark and dark spreads useful proxy need to allow for EUA opportunity cost Persistent excess profit suggestive of market power and barriers to entry Persistent failure to cover total costs suggestive of predatory behaviour? D Newbery Athens 28
35 30 25 20 15 10 5 0-5 Spark spread month ahead 50% efficiency NL DE UK EUA cost 1-Sep-02 1-Dec-02 2-Mar-03 1-Jun-03 31-Aug-03 30-Nov-03 29-Feb-04 30-May-04 29-Aug-04 28-Nov-04 27-Feb-05 29-May-05 28-Aug-05 D Newbery Athens 29 Euros/MWh
35 30 25 20 15 10 5 0-5 Spark spread month ahead 50% efficiency NL DE UK EUA cost 1-Sep-02 1-Dec-02 2-Mar-03 1-Jun-03 31-Aug-03 30-Nov-03 29-Feb-04 30-May-04 29-Aug-04 28-Nov-04 27-Feb-05 29-May-05 28-Aug-05 D Newbery Athens 30 Euros/MWh
25 20 15 10 5 0-5 Spark spread net of EUA UK DE NL EUA starts 1-Jun-02 31-Aug-02 30-Nov-02 1-Mar-03 31-May-03 30-Aug-03 29-Nov-03 28-Feb-04 29-May-04 28-Aug-04 27-Nov-04 26-Feb-05 28-May-05 27-Aug-05 D Newbery Athens 31 Euros/MWh
Competitive Benchmark Analysis Simulate the competitive market in order to calculate Lerner Index of actual price over simulated competitive price Increasingly popular tool of analysis Does not identify individual generators exercising market power Difficulties in identifying appropriate costs Subsequent controversy over quantitative results D Newbery Athens 32
Competitive Benchmark Model of German ESI Müsgens (2004) for period 2000 to 2003 Multi-regional approach with dynamics (e.g. hydro) 50 >400,000 variables and equations! Monthly Base EEX 40 Monthly Base SMC 30 20 10 0 1. Period: Average deviation between marginal costs and prices about 2% 2. Period: Average deviation between marginal costs and prices nearly 50% Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec Jan Feb Mar Apr May Jun D Newbery Athens 33 2000 2001 2002 2003
Oligopoly Models Ideally integrates relevant factors (costs, demand, strategic incentives, transmission constraints) Equilibrium problematic, especially with contracts Recent European examples ECN s COMPETES model - Cournot and Conjectured Supply Functions (CSF) model of Netherlands, Belgium, France and Germany with transmission constraints for market integration Frontier s SPARK model a supply function equilibrium model (but there are multiple equilibria) Results influential in Nuon-Reliant Merger Case in the Netherlands D Newbery Athens 34
Acronyms - 1 AMPs: Automatic Mitigation Procedure (very US) ATC: Available Transmission Capacity CEC: Commission of European Communities CEGB: Central Electricity Generation Board ESI: Electricity supply industry EUA: EU allowance (permit to trade 1 tonne CO 2 ) FERC: Federal Energy Regulatory Commission GW: Gigawatt = 1000 Megawatt = 1m kw G: Generation HHI: Herfindahl Hirschman Index ISO: Independent System Operator MC: marginal cost MO: market operator 35
Acronyms - 2 MOU: memorandum of undestanding MM: Market monitoring MP: Market power NETA: New Electricity trading Arrangements NRA: National Regulatory Authority OTC; Over the counter (markets) PUC: Public Utility Commission PX: Power exchange S: Supply SSNIP: small but significant non-transitory increase in price RSI: Residual Supply Index T: Transmission TSO: Transmission System Operator 36