1

Colombia Firm Energy Auction

Peter Cramton(joint with Steven Stoft)

18 December 2006

2

Outline• Purpose of market• Why forward procurement?• Key features

– Product– Planning period– Commitment period– Cost of new entry (CONE)– Demand curve– Descending clock auction– Price formation– Performance incentives– Fail-safe mechanism– Secondary market

• Transition• Supplier concerns

3

Purpose of Market

4

Purpose of market

• Induce just enough investment to maintain adequate resources

• Induce efficient mix of resources

• Reduce market risk

• Avoid market power in firm energy market

• Reduce market power in energy market

• Pay no more than necessary

5

Why forward procurement?

6

Why forward procurement?

• New projects compete in advance of entry– Coordinated entry

• Less uncertainty in achieving target• Avoid boom/bust• New resources set price directly• Less reliance on demand curve for price setting

• Long-term commitment for new resources– Reduced investor risk– Better price signal for new investment

7

Key features

8

Key Features

• Product• Planning period• Commitment period• Cost of new entry (CONE)• Demand curve• Descending clock auction• Price formation• Performance incentives• Fail-safe mechanism• Secondary market

9

Product

10

Product

• Firm energy ― availability of energy during scarcity events– Dry year (seasonal scarcity)– Outages (spot scarcity)

• Scarcity event defined by high energy price– Energy price is a transparent trigger– Energy price is a reliable trigger

11

Product

• Physically-backed energy option• Firm energy defined by delivery capability in worst-case

benchmark ― a very dry year– Nameplate capacity (maximum output rate)– Firm energy (average energy output rate in worst-case

benchmark)• Thermal example: 92% of nameplate due to outages• Hydro example: 35% of nameplate due to limited water

• Strike price = Gas index high heat rate + other VC– Gas index = New York Harbor residual fuel oil– High heat rate = 12.482 MBTU/MWh– Other variable cost = $15.20/MWh

• This is less than the heat rate of all existing gas units• New thermal peakers have heat rates of 9-10

12

2010 target

2005 actual

0% 50% 100%

2

4

6

8

10

0

Duration

Load(GW)

Firm Energy Target

Load Duration Curve and Firm Energy Target

growth

Target peak load

13

Single product; single price

• In a hydro-dominated system, there is a single reliability constraint:firm energy to cover dry year

• Single constraint implies single product and single price

• Each new resource is rated for its incremental contribution to the firm energy constraint

14

Product is:• Firm energy + mandatory hedge• Firm energy =

– Expected annual drought-year energy contribution to system (if unit disappeared, how much less energy would the system have—same for hydro and fossil)

• Mandatory hedge = – Obligation follows load in aggregate

• Unit’s daily obligation based on its firm energy• Obligation over day tied to dispatch• Matching obligations with dispatch improves the performance

of the spot energy market– Rewarded by spot price if produce more than

obligation in scarcity hours

15

Obligation

• Purchase by load is translated into an obligation for each supplier– If 98% of target is purchased, then aggregate supplier

obligation is 98% of load in each hour• Load is unhedged for 2% of load that is not purchased

– True-up at end of each month to adjust for deviation between monthly target and actual monthly load

• If actual load is 102% of month’s target, then obligations are scaled down by 100/102. Penalties/rewards are calculated on this basis.

– Risk from unanticipated load growth is born by load, not suppliers

16

Settlement

• Settlement is just like settling a conditional contract for differences in each hour

• If spot price < strike price, then no obligation• If spot price > strike price, then settle

differences: reward or penalty =(Q supplied – Q obligation) (P spot - P strike)Same as supplier buying from spot market to satisfy obligation.

• Same outcome if done on unit basis or portfolio basis

• Supplier optimizes portfolio just as without option

17

Planning period

18

Planning period

• Time between auction date and start of commitment

• 4 years ― long enough for new entry to occur (except large hydro projects)

• Makes firm energy market contestable and allows new entry to set the price– Existing resources would set the wrong price

because of sunk costs and market power

19

Long lead-time projects

• 4-year planning period may be too short for large hydro projects (6-8 years to build)

• Allow large hydro projects to lock in auction price from 4-year ahead auction seven years (or less) ahead

• Large hydro project is price taker• Decides after auction a fraction of its firm energy to lock

in at 4-year ahead auction price• Total quantity of firm energy in years > 4 that load

purchases is limited by a percent of new firm energy required in that year based on planning projections

Years ahead 7 6 5Percent limit 40% 50% 60%

Load's Limit on Forward Purchase

20

Planning period

• First auction Nov 2007 (2011): 3 year

• Second auction Jun 2008 (2012): 3.5 year

• Third auction Nov 2008 (2013): 4 year

• All later auctions (2014…): 4 year

21

Commitment period

22

Commitment period

• New resources ― up to 20 years– Supplier selected at qualification– Long commitment lets new resources lock-in firm

energy price, reducing risk and encouraging investment

– Price is in constant $ (adjusted for inflation)

• Existing resources ― one year– Does not need long commitment, since costs are

already sunk– Short commitment reduces risk (more draws from

price distribution)

23

Demand curve

24

Demand curvePrice offirm energy

Firm energyTarget

CONE

2 CONE

CONE = Cost of New Entry (marginal unit)

0

Curve reflects marginal value of firm energy

Able to withstand scarcity events worse than worst-case benchmark

½ CONE

Price ceiling

Price floor

4%

Load not fully hedged

25

Descendingclock auction

26

Descending clock auction

• Auctioneer announces high starting price

• Suppliers name quantities

• Excess supply is determined

• Auctioneer announces a lower price

• Process continues until supply equals demand

27

Starting price

• Starting price must be set sufficiently high to create significant excess supply

• Setting too high a starting price causes little harm– Competition among potential projects determines clearing price;

high start quickly bid down

• Setting too low a starting price destroys auction– Inadequate supply or insufficient competition

• Price of 2 times Cost of New Entry is recommended• Note that clearing price will exceed CONE in some years

to the extent it is below CONE in other years (of surplus)

28

CONE updates

• If auction in year t is successful,CONE in year t+1 = .7 CONE in year t

+ .3 clearing price in year t

• If auction in year t fails,

CONE in year t+1 = CONE in year t

29

Mechanics

• Clock auction done in discrete rounds• In each round,

– Auctioneer announces• Excess supply at end of prior round• Start of round price (higher price)• End of round price (lower price)

– Each bidder submits a supply curve at all prices between start of round price and end of round price

– Auctioneer determines excess supply at end of round price

• If no excess supply, clearing price determined

30

Price

$7.00

$6.63

$6.17

$6.00

Quantity (MW)

175 300 400

start-of-roundprice

end-of-roundprice

Individual Supply Bid, Round 6

• Activity rule– Bidders can only maintain or reduce quantity as price falls

(upward sloping supply curve)• “Intraround bids”

– More accuracy without too many rounds– Better control of pace of auction– Ties are reduced

31

Price

$12.00 = P0

P1

P2

P3

QuantityDemand

Round 5

Round 4

starting price

clearing price

Aggregate supply curve

P4

P5 $6.17 = P6

Round 3

Round 2

Round 1 excess supply

$6.00 = P6’

Descending clock auction

32

P

Q

D

PC

S

1. P

Q

D

PC

S

2.(a) P

Q

D

PC

S

2.(b)

Unit accepted

Unit rejected

Clearing rule: maximize net value

33

Information policy

• Demand curve and starting price announced before auction

• After every round, auctioneer reports– Aggregate supply by hydro, baseload, peaker– Excess supply at end of round price– End of round price for next round

(determined from extent of excess supply)

34

New projects are all or nothing

• Lumpy investment respected; investor does not fear partial acceptance

• If multiple bidders drop at the clearing price, the group of bids are accepted that minimizes excess supply

35

Price formation

36

Market power

• Addressing market power in firm energy market is essential

• Strong incentive to exercise market power– Existing resources have substantial sunk costs– New resources are only a tiny fraction of total– Market is concentrated

• Any of top-4 suppliers could unilaterally set price

• Long-term price signals are more stable and efficient if determined from competitive forces, rather than market power

37

Market power solution• New resources

– Bids are not mitigated in any way– Assumes competition for new resources

• Existing resources– Resources can opt out of market with either an opt-out bid or retirement bid– Opt-out bid

• Not revealed during auction• Cannot impact the price for existing supply• May be rejected for reliability reasons; gets reliability must run payment if rejected

– Retirement bid ― permanent opt out of firm energy market• Submitted four weeks before start of auction• Accepted retirements excluded from any future firm energy payments• Retirements may be rejected for reliability reasons, but only if the reliability problem cannot be

resolved during the planning period with alternative actions, such as transmission upgrades or new resources

• Retirements are posted as soon as they are accepted• Retirements are replaced with new resources in the auction

(represented as a shift to right in the demand curve for all prices below the retirement bid)

• Repower bids– Replacement of generating unit(s) at existing plant– Replacement unit is bid as new resources– Existing unit is a conditional-retirement

38

Repowering bids

• Easily accommodated in auction

• Two types:– Quick switchovers (down time less than 1

year)• Repower bid is a new entry bid and a conditional

retirement

– Extended down time (more than 1 year)• Retirement followed by new entry bid 1 or more

years later

39

Market power solution

• New resources almost always set the price

• Demand curve sets the price in surplus years in which new entry is not needed

• Retirements occasionally set the price

• Other than retirements, existing resources never impact the price

40

Replacing accepted opt-outs

• Opt-outs are replaced by new resources– March up the supply curve revealed in the

clock auction• But not more than a 30% increase in price• Any additional replacements occur in

reconfiguration

– All new resources receives this higher price

• Existing resources receive the original clearing price

41

Performanceincentives

42

Incentives and hedge

• Performance incentive comes primarily from energy spot price; this is not changed by hedge

• Hedge assures that normal performance will receive normal reward in wet and dry years alike

• Every extra MWh of energy is rewarded the same with or without hedge

43

Energy price motivates performance

• Hedged resources still face the energy spot price at all times– Those that perform better receive more– Those that perform worse receive less

• An additional incentive to perform is impact on firm energy qualified for sale in auctions in future years

• Under performing units are downgraded• Over performing units are upgraded

44

Fail-safe mechanism

45

Inadequate supply

• If, at the starting price, there is insufficient supply of firm energy– New resources are paid starting price– Existing resources are paid 1.1 CONE

• Rule does not discourage new projects

46

Insufficient competition

• Existing resources, less retirements, are less than demand at the starting price,

• At the starting price, the firm energy bid exceeds demand but less than 4% excess, or a supplier’s new resources are pivotal, and

• At qualification, quantity of new projects from small players (those with less than 15% maximum firm-energy market share) > 50% of required new firm energy

• Otherwise insufficient competition:– Auction held– New entry paid clearing price– Existing capacity paid 1.1 CONE (or clearing price if less)

• Rule does not discourage new projects

47

Secondary market

48

Reconfiguration auction

• Takes place at same time as primary auction– Primary: 4 years ahead– Reconfiguration: 3, 2, 1, 0 years ahead

• Reconfiguration includes– Adjustment of firm energy target for current

forecast– Supplier’s buy/sell to balance position

(including dispatchable load)

49

Reconfiguration auction

• Standard sealed-bid clearing-price auction

• Same demand curve as in primary auction, netting out resources already purchased

• No bid mitigation

50

Monthly spot exchange

• Monthly simultaneous clearing– Standard sealed-bid clearing-price auction– Suppliers buy/sell to balance positions– Demand curve same as in primary auction

51

Transition

52

Transition and timing

• Transition period 2007-2010– Administrative firm energy price includes a premium

for cost of energy option– Existing contracts are modified to subtract the same

premium since prices above the strike price are now covered by the firm energy payment, not the contract

2011 Primary auction

… 2030 … 2031

Planning period

2012 Primary auction

New resource commitment period

Existing resource commitment period

2009

Existing resource commitment period

New resource commitment period

Planning period

Administrative Pricing for New and Existing Based on CONE

2007 2008 2010 2011 2012 2013

53

Reducing risk in early years• Early years of auction

– Ceiling and floor on firm energy payment to existing suppliers

– Spread between ceiling and floor expands after each competitive auction

– Spread starts at 0 (transition years)– Increases to

• Ceiling = 2 CONE• Floor = .5 CONE

Transition 0 1 2 3 4Ceiling = CONE 1 1.2 1.4 1.6 1.8 2

Floor = CONE 1 0.9 0.8 0.7 0.6 0.5

Number of competitive auctionsCeiling and Floor for Existing Suppliers in Early Years

54

MWDemand

Transition

Auction date

Firm energy to be auctioned

20072008

20092010

20112012

20132014

2015

Existing supply

New supply

Nov Jun Nov Nov Nov2007 2007 2008 2009 2010

55

Supplier concerns

56

Supplier concerns

• Please note that these concerns were as expressed to CREG on Monday, 5 June 2006, in response to the original CREG proposal. These concerns may not apply to proposal presented here, which is significantly different from the original proposal.

57

1. Approach not used elsewhere

• Similar approach was adopted in New England (32 GW peak load) in March 2006– Approach currently being implemented– Approach endorsed by FERC

• All elements of proposal are commonly used in markets around the world– Options are used everywhere– Clock auctions used for many years in US, UK,

France, Germany, Belgium, Hungary, Denmark, …

58

2. Call option increases risk

Price Coverage

Forward Energy Contract

US$100

US$250

US$0 US$0

US$250

Forward Energy Contract

Energy Option

59

3. Consumers will pay more

• Minimizing risk while addressing market power and performance incentives means that consumers will pay less

60

4. Price not a reliable measure of scarcity

• High prices come from two sources– Scarcity– Market power

• Option fully addresses market power in spot market; thus, high prices can only come from scarcity

• If spot prices are still unreliable, then market must have another flaw. Fix it!

61

5. Strike price is too high

• Why not have a very high strike price?(US$250 or more)– Benefits of call option are largely lost

• Load hedge• Mitigation of market power in spot energy market• Mitigation of market power in contract market• Risk reduction

– No reason to set strike price higher than marginal cost of an expensive thermal unit

62

6. Approach is complex

• Effective firm energy markets are necessarily complex because the economic challenges are great, especially market power

• Proposal uses clear and simple market-based methods

63

Questions