Derivative market design & performance Masterclass for the Restructured Electricity Industry 24-26 August 2005 CEEM, 2005
Participant motivation for trading electricity derivatives: price-risk management Generators may have fixed costs but face variable electricity spot price Retailers may buy electricity at variable spot price & sell at a pre-determined retail price Large end-users may buy electricity at variable spot price but sell their products at market-set price Opposing (complementary) risk profiles: Generator is a natural seller of derivatives Large end-user is a natural buyer of derivatives Retailer potentially both a seller & a buyer 2
Participant risks in derivative trading Market risk: price or volume changes in spot or derivative markets Credit risk: counterparty fails to meet contractual obligations Regulatory risk: impact on derivative value due to regulatory decisions Operational risk: internal decision making, equipment performance, liquidity management, exogenous phenomena 3
Market power issues in spot & derivative markets Spot markets Buyers have little discretion in spot market trading Buyers cannot be sellers Significant barriers to entry Derivative market positions restrict spot market power Auction: market maker must manage counter-party risks Derivative markets Buyers have more discretion in derivative market trading Buyers can also be sellers Few barriers to entry Risk assessment depends on spot market power Greater role for bilateral trade 4
Derivatives:- definitions A derivative (contract) is a paper product: Derivatives are also called financial instruments Its only relationship to the underlying (physical) product is through the underlying spot price: The derivative creates a financial obligation on seller related to a future spot market price outcome Key derivatives for electricity: Two-sided contract for differences (CFD) or swap Call option or cap (a form of one-sided CFD) Put option or floor (a form of one-sided CFD) 5
Derivatives:- key parameters Quantity of spot market units to which the derivative applies (e.g. MWh) Spot market period(s) to which derivative applies Strike price: Interpretation depends on the type of derivative 6
Two-sided CFD or swap:- definitions A CFD or swap for a single future spot market interval is a piece of paper stating: 1. Contract price (strike price) = p c 2. Contract quantity = x c 3. The future spot time at which contract will be reverse traded (or closed out ) at spot price p s Trade in CFDs is only related to trade in the physical commodity by the spot price at which the reverse trade is carried out 7
Trader s view of CFD trading CFD buyer: Buys CFD at contract time; cost = p c x c Sells CFD at spot time; income = p s x c Net cost of CFD to buyer = x c (p c -p s ) CFD seller: Sells CFD at contract time; income = p c x c Buys CFD at spot time; cost = p s x c Net value of CFD to seller = x c (p c -p s ) Market in CFDs required to allow adjustment of x c 8
Effect of CFD close-out Price CFD seller pays CFD buyer x c (p s -p c ) variable spot price uniform contract price CFD buyer pays (consensus view of CFD seller x c (p c -p s ) expected spot price for the duration of the contract) Time 9
CFD & electricity trading generators market in CFDs x c, p c spot market in electrical energy x s, p s retailers or direct end-users 10
Cash flow for CFD & electricity trading #1 +$ retailer CFD settlement: x c (p c -p s ) (cash flow direction depends on sign) +$ generator $ end-user $ spot market settlement: x s, p s $ Retailer recovers cash flow from end-user via a retail tariff 11
Cash flow for CFD & electricity trading #2 Variable price cash flow (p s x c ) retailer Fixed price cash flow (p c x c ) generator $ $ spot market settlement: x s, p s end-user $ Retailer recovers cash flow from end-user via a retail tariff 12
CFD & electricity trading: generator perspective Generator sells: CFD for net income after buyback: x c (p c -p s ) Electricity in spot market for income: p s x s Income to generator (= cost to consumer): = p s x s + x c (p c -p s )[spot + CFD] gross pool = p c x c + p s (x s -x c ) [contract + variation] net pool = p c x s + (p s -p c )(x s -x c ) [projection + price response] p c & x c : market estimates of future p s & x s where is the net sum of CFD quantities traded to date for a future spot market period ( open position ) 13
CFD & electricity trading: generator with 1MW CFD @ 30 $/MWH Spot Mkt Period Spot price ($/MWH) CFD (p c -p s ) ($/MWH) Income if x s =x c ($) Income if x s =0 ($) 1 30 0 30 0 2 32-2 30-2 3 28 +2 30 +2 4 100-70 30-70 5 10,000-9,970 30-9,970 14
CFD & electricity trading: retailer with 1MW CFD @ 30 $/MWH Spot Mkt Period Spot price ($/MWH) CFD (p c -p s ) ($/MWH) Cost of 1MW if x s =x c ($) Cost of 1MW if x c =0 ($) 1 30 0 30 30 2 32-2 30 32 3 28 +2 30 28 4 100-70 30 100 5 10,000-9,970 30 10,000 15
Direct end-user using CFD as a hedge End-user with 15 MW factory buying directly from spot market: Assume that product value = (150-p s ) $/MWh At contract time, end-user buys: 15 MW CFD @ 50 $/MWh = 750 $/h Locked in profit = 15(150-50) = 1500 $/h 15 Quantity MW Profit p s 150 Price $/MWh 16
Direct end-user using CFD as a hedge At spot time, consider two cases: 1. Spot price = 100 $/MWh With consumption of 15 MW, spot + CFD trading cost = p c x c + p s (x s -x c ) =50x15+100(15-15) = 750 $/h (same as cost of CFD, regardless of spot price) 2. Spot price = 200 $/MWh Either consume 15 MW with same result as case 1 Or shut down factory & earn a profit from the CFD transaction alone: profit from CFD = x c (p s -p c ) = 15(200-50) = 2,250 $/h 17
Summary of CFD properties CFD protects against future price risk: Incentive to fully hedge expected spot position Thus CFD market predicts future spot market in both price & volume (hedge volume only) Even when fully hedged, there is still an incentive to respond to spot price: Rewards voluntary price response Generator not protected against outage risk 18
Example: Hedging regional spot price differences in the NEM using spot market settlement residues A hedge against differences between regional spot prices for one direction of flow Settlement residue for regulated interconnectors: Difference in regional reference prices multiplied by interconnector power flow for each spot market interval NEMMCO runs 3-monthly auctions of settlement residue An incomplete hedge: doesn t cover interconnector losses or outages Doesn t cover unregulated interconnectors 19
NEM inter-regional settlement residue hedge ctd Region B NSP $60 - costs SR Auction $60 gen r net income $240 10 MWh $50 Settlement residue $50 $100 $150 $200 A B 15 $/MWh 20 $/MWh constrained lossless link with small spot price difference cust r net cost $300 10 MWh two-region network & regional spot prices Generator sells a 10 MWh hedge contract on region B price to customer at $30 MWh & buys a directed 10 MWh hedge (B-A) from NEMMCO SR auction at cost of 6 $/MWh (expected spot price difference between regions) 20
NEM inter-regional settlement residue hedge ctd Region B NSP $60 - costs SR Auction $60 gen r net income $240 10 MWh Settlement residue $49800 $49800 $200 $50000 A $49700 B 20 $/MWh 5000 $/MWh constrained lossless link with large spot price difference cust r net cost $300 10 MWh two-region network Generator sells a 10 MWh hedge contract on region B price to customer at $30 MWh & buys a directed 10 MWh hedge (B-A) from NEMMCO SR auction at cost of 6 $/MWh (expected spot price difference between regions) 21
Annual outcomes of NEM SRAS for FY03 (SRC report FY03 see www.nemmco.com.au) 60 50 40 30 20 10 Sale price $M Distribution $M 0 SaVi ViSa ViSn SnVi SnNs NsSn NsQld QldNs 22
Annual outcomes of NEM SRAS for FY04 (SRC report FY04 see www.nemmco.com.au) 35 30 25 20 15 10 5 Sale price $M Distribution $M 0 SaVi ViSa ViSn SnVi SnNs NsSn NsQld QldNs 23
Negative weekly residues for SRAS in FY03 (SRC report FY03 see www.nemmco.com.au) 400 350 300 250 200 150 100 50 Ann sum of negative weekly residues $K 0 SaVic SnVic NswQld QldNsw 24
Negative weekly residues for SRAS in FY04 (SRC report FY04 see www.nemmco.com.au) 2500 Ann sum of negative weekly residues $K 2000 1500 1000 500 0 SaVic SnVic NswQld QldNsw 25
NEM unregulated interconnector An unregulated interconnector (MNSP): Submits offers into the NEM spot market & retains spot market income Does not participate in SRA process but can independently sell inter-regional hedges MNSP hedge versus SRA auction: MNSP not restricted on hedge design or duration but faces similar issues (availability,losses) SRA auction competes for hedge volume & sets a benchmark price 26
Call option or cap:- definition & role The seller must compensate the buyer if the spot price is above the strike price Potential call option buyer: consumer with inflexible demand unreliable base load generator Potential call option seller: reliable, high operating cost thermal generator low capacity factor hydro generator 27
Put option or floor:- definition & role The seller must compensate the buyer if the spot price is below the strike price Potential put option buyer: inflexible base load generator Potential put option seller: large electricity consumer 28
Call & put options: illustration (buy a CFD = buy a Call & sell a Put at the same strike price & same option fee) Price call option seller pays call option buyer x c (p s -p c ) variable (possibly high) spot price uniform strike price put option seller pays put option buyer x c (p c -p s ) Time 29
Call & put options: features Options are single-sided versions of CFDs Unlike CFDs, the option buyer must pay a fee to purchase the option: The option fee is based on an estimate of the close-out value of the option at spot time: a call option will have non-zero close-out value if the spot price exceeds the option strike price a put option will have non-zero close-out value if the spot price is lower than the option strike price Can create composite derivatives, eg: A collar combines a call option at a higher strike price with a put option at a lower strike price 30
Final value of a call option call option value strike price value to buyer option fee spot price break-even outcome value to seller 31
Final value of a put option put option value strike price value to seller option fee spot price break-even outcome value to buyer 32
Buy CFD: Sell a put option & buy a call option at same strike price & option fee Option value ($/MWH) Sell put Buy call & sell put = buy CFD Buy call Spot price ($/MWH) 33
Retailer or end-user who buys CFD & spot Net cost ($/MWH) Buy spot Buy (CFD+spot) Spot price ($/MWH) Buy CFD 34
Call option & generator with high operating cost #1 probability density average spot price strike price for generator with high operating cost spot price probability distribution spot price ($/MWh) 35
Call option & generator with high operating cost #2 Generator would like assured operating surplus to earn return on investment But operating cost > expected spot price: Cannot benefit from a CFD contract If generator reliable (both start & operate): Sell call option @ strike price = operating cost Then option fee provides return on investment: Size of fee depends on likelihood of: spot price > strike price 36
Call & spot: generator surplus Assumes spot quantity equals option quantity when spot price > strike price (i.e. fully reliable generator) generator surplus strike price = operating cost spot market surplus generator surplus from spot & option option value spot price 37
Call & spot: cost to retailer or end-user Assumes spot quantity equals option quantity when spot price > strike price (i.e. fully reliable generator) cost strike price spot cost spot cost plus option cost option cost spot price 38
Call option as end-user hedge #1 End-user with 100 MW inflexible load Generator with 60 $/MWh operating cost End-user buys call option from generator: 100 MW, 60 $/MWh, period T, option fee = F Period T scenario 1:- spot price = 20 $/MWh Call option inactive, generator does not operate: Earns fee F End-user buys at spot price: Total cost = spot cost + F 39
Call option as end-user hedge #2 Period T scenario 2: spot price = 200 $/MWh Generator pays end-user the option close out value: 100x(200-60) = 14,000 $/h Generator sells electricity, incurs operating cost: Surplus from spot market = 100x(200-60) = 14,000 $/h Generator surplus from spot + call option = F End-user buys 100 MW electricity at spot price, receives option close out value from generator: End-user cost = 100x200-14,000 + F = 6,000 $/h + F Shielded from spot price > strike price (60 $/MWh) 40
Call option as end-user hedge #3 Setting the option fee F at contract time: Assume: Reliable generator Expected spot consumption = 100MW Spot price probability distribution given in table below Then: Ave. spot price =.35x20+.5x30+.15x200 = 52 $/MWh price 20 30 200 ($/M wh) probability 0.35 0.5 0.15 41
Call option as end-user hedge #4 Expected outcomes for end-user: Cost without option = 52x100 = 5,200 $/h Cost with option = (.35x20+.5x30+.15x60)x100 = 3,100 $/h Benefit from option = 2,100 $/h Expected end-user benefit from option equals expected generator spot income Basis for negotiating F if both are risk averse Both generator & end-user still see an incentive to respond to spot price 42
Buy collar: sell a put option at a low strike price & buy a call option at a high strike price Option value Buy collar Sell put Buy call Spot price 43
Retailer or end-user who buys collar & spot Cost ($/MWH) Buy (collar+spot) Buy spot Spot price ($/MWH) Buy collar 44
Example: Premium ($/MWH) for all NEM spot prices above strike price for year to 4/02 (Reliability Panel, 2002) 60 50 Queensland New South Wales Victoria South Australia Premium ($/MW/h) 40 30 Premium at zero strike price = mean spot price (in hindsight) 20 10 Premium at non-zero strike price = fair option fee for call option at that strike price (in hindsight) 0 0 50 100 150 200 250 300 Strike Price ($/MWh) 45
Example: d-cyhpatrade exchange-traded call options for NSW peak period (www.d-cyphatrade.com.au) 2000 1800 1600 1400 1200 1000 800 600 400 200 0 65 70 75 80 85 90 Strike price $/MWH 05Q1 06Q1 07Q1 46
Example: AFMA OTC 10 MW CFDs (www.afma.com.au) AFMA NEM forward price curve August 2004 70 60 50 40 30 20 SA peak Vic peak NSW peak Qld peak SA off-peak Vic off-peak NSW off-peak Qld off-peak 10 0 2005 2006 2007 2008 Calendar year 47
Example: NSW 3rd Tranche vesting contracts (July 1998 to December 2000) Two types of contract: 1. CfDs covering around 80% of franchise load: NSW govt preferred a strike price of 44.5 $/MWh ACCC proposed to reduce this to 37 $/MWh 2. Cap (buy call) & binary option floor (sell adder rather than put) ACCC proposed cap of 37 $/MWh (NSW 65 $/MWh) Floor price of 14 $/MWh representative of fuel costs This structure was designed to: Allow 95-98% of franchise load to be vested Expose generators to pool prices at the margin 48
NSW 3rd tranche vesting contract, continued. Cost to retailer for spot & type 2 vesting contract (7/98-12/00) Net cost to retailer ($/MWh) 37 Option cap (ACCC version) Spot price Overall cost to retailer: spot + vesting contract 14 Side effect of binary option : NSW generators discouraged from wanting the spot price in this region Binary option fee = $14/MWh adder 14 37 Spot price ($/MWh) 49
Example: NSW Electricity Tariff Equalisation Fund (NSW Electricity Supply Amendment Act, 2000, applying from 2001) A form of CFD to cover default tariff load: Between state-owned retailers & Ministerial Corp n, underwritten by state-owned generators: Strike price based on IPART estimate of LRMC Volume set ex-post at NEM settlement: Uncertain volume deters gens from other contracts Return of ETEF surplus to gen s is discretionary: Reduces their incentive to offer lower than LRMC Encourages them to bid up spot price if ETEF high Poor social, economic & environmental policy: Winners: residential with air-con, interstate gen s Losers: low income & contestable consumers 50
Institutional structure for ETEF (NSW Treasury, Dec 2000) 51
NSW Electricity Tariff Equalisation Fund (NSW Treasury, December 2000) For each retailer, the volume of energy covered is that supplied under regulated retail tariffs. This creates moral hazards for both retailers & customers. The load-shape may worsen & the effective LRMC rise. 52
Calls will be made to state-owned generators when required to top-up the fund (NSW Treasury, Dec 2000) 53
Issues raised by ETEF Regulated tariff will subsidise consumers with expensive habits (e.g air-conditioning) Retailers can pass on these risks via ETEF Underwrites A/C marketing drives (eg EnergyAust) Retailers may try to bias ETEF volumes: Up when spot price high, down when low Discourages NSW generators from selling other forward contracts (as in South Aust): Increases risk to other customers & retailers Generators are effectively uncontracted when the ETEF fund is in surplus & no calls expected 54
EnergyAustralia residential airconditioning marketing campaign, December 2000 EnergyAustralia can provide you with advice on your airconditioning needs. Just call 131364 and talk to our energy experts about 2 years interest free on any airconditioning system (EnergyAustralia bill stuffer, 12/00) Supported by NSW government s Electricity Tariff Equalisation Fund, with volume set ex-post 55
In 2001 NSW load >90% peak for ~5% of time NEM load duration curves, January-March 2001 & 2003 (NECA quarterly Market Statistics) In 2003 NSW load >90% peak for <2% of time 56
Average small consumer load profiles in Sydney for 2000/1 Summer (H Coleburn, Energy Australia, August 2001) Air-conditioning load is very temperature sensitive: A/C load on hot summer day much greater than on average summer day Load for consumers without A/C show much less temperature sensitivity Average customer consumption - Summer 2000/01 Profiles 2.50 2.00 1.50 Air conditioning usage AC, Hot Workday AC, Avg Workday kw 1.00 No AC, Hot Workday 0.50 0.00 2:00 4:00 6:00 8:00 10:00 12:00 14:00 16:00 18:00 20:00 22:00 No AC, Avg Workday 57
Growing temperature sensitivity of demand in South Australia (SA DM Taskforce, 2002) 3000 SA System Load vs Temperaturen 1994-2001 WD 19:00 2900 2800 2700 2600 2500 2000-01 MW 2400 2300 2200 2100 1999-2000 1997-98 2000 1900 1994-95 1800 1700 1600 1500 1400 1300 1200 0 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 32 34 36 38 40 42 Deg C 58
Alternatives to ETEF Careful use of CFDs, caps & collars Swaptions: options over CFDs, caps & collars Flex products: Flexible volume CFDs, caps & collars Weather derivatives: Call option on weather event, such as temperature or rainfall Broker tries to match counter-party interests Physical actions such as improved house design (derivative market assists valuation) 59
Summary, derivative contracts When used well, derivative contracts: Reduce spot market price risk Do not interfere with spot market incentives Can be used as a vesting contract to impose transitional financial obligations Given well designed spot & derivative markets, trade in derivatives can: Predict future spot market conditions Provide flexible risk management facilities Improve control of market power 60
Current use of derivatives in the NEM Risk management framework incomplete: Aggregate volume information not available to support network planning Little end-user participation in derivative trading, so risk chain incomplete Possible remedies: Develop measures of hedge volume by region Restructure retail tariffs to spot & forward form: Consider using profile forward volumes for small end-users 61