Transcript No Slide Title

The Value Paradigm
and Future of ISOs
Richard O’Neill
Chief Economic Advisor
Federal Energy Regulation Commission
[email protected]
APEX Conference
Seoul, Korea
October 30, 2006
Views expressed are not necessarily those of
the Commission
1
What changed?
Technology has eliminated some market
failures
Generator size small compared to market
vertical integration not required for
communication
What is at stake?
consumption generation
Tkwh
GW
US
world
3.71
15.80
revenues price
billions/yr $/kwh
950.00 270.00
3625.00 2100.00
0.07
2
0.14
Growing pains
Markets have become larger than the
traditional political/regulatory boundaries
Recognize the efficiency benefits
Problems with sovereignty issues
Growing pains abound (U.S., Europe, Asia,
Australia, South America)
State v federal issues
East v west issues
Change takes time
3
History of merchant generation
and transmission
merchant generation is the U.S.
Pre-1978: almost none
PURPA: mandatory purchase
Opposition: threaten reliability
Wrong
Merchant transmission
Pre-2000: none
Crosslink (AU)
Opposition: threaten reliability
Wrong
4
cultural issues
2000 year debate What is the “just” price?
Is electricity a public good?
Enron/California backlash
Should the government hedge for citizens?
How important is administrative simplicity?
Competition
does not promise lower prices
Does produce lower prices than cost-based
regulation
5
Electricity markets
 Electricity markets have become highly interconnected.
 The policy question is how should the markets be organized?
 Externalities abound including
 lack of demand response (the central problem), reliability rules
 the laws of physics, business stealing, pollution
 free riding, market power and blackouts
 The vertically integrated utility (franchised monopoly)
 internalizes externalities
 creates efficient incentive problems and
 usually introduces rate regulation.
 a relatively unregulated market is politically unacceptable
 an exchange with rules and regulations that prevent
market power and promote efficiency
6
Value of competition
cost-based regulation
At best weak efficiency incentives
Weak innovation incentives
Naïve market power mitigation
Price = cost
Return on equity and imprudence?
Competition (market-based) regulation
strong efficiency incentives
strong innovation incentives (may take time)
7
market power mitigation (bidding rules)
information technology upsets the
vertically integrated utility paradigm
Computer technology improvement since
1970
Hardware: 109
Software: 106
Telecom:
fiber optics
Wireless
Internet
Expands the breath and depth of
competition possibilities
8
electronic
trading
 FERC requires
 1988 electronic access to pipeline information.
 1992 pipelines to conduct capacity auctions on EBBs.
 1994 electric utilities to transact business on EBBs
 1995 EBay: online auction site founded
 1997 Continental Power Exchange the predecessor to ICE
 1999 Enron Online
 bought and sold $880bn just two years.
 2002 sold to UBS and dies
 Floor/desk v. screen/computer
 NASDAQ: screen based
 NYMEX: dual system
 NYSE: 2006 starts dual system
9
Transitioning from
planning and costbased prices to
competition, auctions
and market power
mitigation models
10
Analogies and their problems
analogy
owners
Commodity
/conduit
Displace
ment
network
other
issues
pricing
highways
public
unbundled
No
congestion
Gas tax and toll
roads
water
public
bundled
yes
other uses
usage
Natural
gas
private
unbundled
yes
storage
and valves
Price caps and no
withholding
Air traffic
public
unbundled
no
No pricing
Ticket tax
parks
public
unbundled
no
congestion
Income tax
telecom
private
mixed
No
Busy signal
Price caps
railroads
private
bundled
no
congestion
Loose price caps
2015년 7월 21일
11
Market operator for
commodity
capacity
tx rights
ISO
LSEs/discos
system operator
security coordinator
tariff administrator
Transco/
Gridco
gencos
What did we do before we went
back to the drawing board?
ENRON creates ether money;
wall street believes!
iou
12
There is no perfect
market design
there is no efficient market design that
satisfies
Individual rationality (value maximization),
incentive compatibility (efficient outcomes), and
Revenue adequacy (no outside money)
There is no perfect market design (costbased or market–based)
All markets are regulated
The policy question is how to design a ‘just’
market
13
A decade of expensive
failed market designs
Zonal markets (Cal, PJM, NE, UK)
Contract path
Sequential markets for energy and anc services
One settlement systems
Infeasible markets (Cal PX and UK)
Ignore non-convexities (start-up and no-load)
Ignore market power
As-bid/first-price pricing
all ended in administrative intervention
"Everything should be made as simple
as possible ... but not simpler." Einstein
14
Pre-day-ahead markets
for transmission rights: CRT/TCC/TRCs/FGRs
for generation capacity/reserves (ICAP)
market power mitigation via options contracts
day-ahead market for reliability(valium substitute)
simultaneous nodal market-clearing auctions for energy,
ancillary services and congestion
allow multi-part bidding
higher of market or bid cost recovery
allow self scheduling
allow price limit bids on ancillary and congestion
Real-time balancing myopic market
15
markets are nodal-based LMP with fish protection
similar
market
designs
 US: PJM, NY, NE, MISO, CA, SPP, ERCOT
 Asia/Pacific: New Zealand, western China, eastern
Australia, Philippines, ...
 Europe: Ireland
 Russia
 South America: Chile, Argentina, Columbia, Brazil, …
16
What’s next?
Complete the market design
Choice: don’t eliminate choice because
market participants don’t want it!!!!
More active demand side
Price all scarce resources
Mitigate market power
17
End-use markets
 Lacking demand response
 capacity markets
 build rate base generation and transmission
 energy-only market may not clear
 with demand response
 capacity markets become hedging markets
 energy-only market clears
First: Generation in drag
 Simple policy
 all replacement meters are interval
 Large customers put on real time prices
 Intelligent remote circuit breakers
 Offer hedges against real-time market
18
Rethinking reliability
Unbundling and Paying for reliability
 Old way: Voluntary “gentlemen’s agreements”
 Little or no liability; No Demand-side role
 “absolute” limits
 Primary cause of cascading blackouts: Vegetation and
Operator error
 Many models do not differentiate between
 Simple controlled curtailment
 Cascading uncontrolled blackout
 Some simplifications can cause anomalous results
 One event criteria
 Amount curtailed criteria
 Improving reliability
 Constraints are soft not hard: use pricing
 Some liability creates good incentives
 Demand curve for reserves
19
 reliability markets: pay for generation and transmission reserves
electricity market design
electricity market models and regulation have
generally been circumscribed by hardware and
software capabilities.
Modeling simplifications include
energy only,
DC only,
oversimplified topologies (contract path)
large time steps.
Simplifications create ‘missing commodities’
including
reactive power,
reserves (real and reactive) and
other ancillary services.
20
Computational considerations
“perennial gale of creative destruction”
Schumpeter
 Pre-1996: patched pool software
 1996: LMP in NZ
 300 nodes
 transmission constraints are manual
 1990s: linear programs improved by 106
 103 in hardware
 103 in software
 2000s: mixed integer programs improved by 103
 Hardware: parallel processors and 64 bit FP
 Software: cplex
 2006: 30000 nodes
 10000+ transmission constraints
 1000 generators with n-part bids
 MIP introduced in DAM
21
Benefits of better modeling
 A good portion is traded or dispatched using sophisticated
mathematical software
 Minor gains in auction efficiency are measured in billions
 Benefit/cost for better modeling can be over 250
 Where is the innovation occurring?
ISOs
exchanges
 Vertically integrated utility
 Examples
 MIP
 CCCT modeling
 Topology estimators
 Reactive power modeling
22
value of modeling
improvements
low
better optimizers
better ACOPF
better real time pricing
better reliability
better asset utilization
ex ante mitigation
better capital investment
total
1
1
5
1
1
3
1
13
billions/year
middle high
3
10
3
10
15
30
4
30
2
30
6
30
5
10
38
150
23
a ideal complete market design
(a target rich environment)
 all scarce commodities are
 simultaneously priced and
 cleared over both space and time
 using a full Mixed Integer ACOPF
 all assets as active market participants
 Market power mitigation.
 The market design has
 combinatorial marginal cost bidding with budget constraints,
 Non-confiscatory settlements and
 simultaneously clearing of all products.
 If different models are used for different time horizons
for planning, compatibility is important
 No Lagrange multipliers left behind
24
λ λ λ λ ...
Value of an ISO
Owns no electric assets
Promote competition
Integrates dispatch, competition and reliability
Revenue-neutral non-confiscatory auctions
Cutting edge of software development
Better information
Planning for competition
Sounds unusual but all markets that involve land use are
regulated and planned
Entry sites: good electric locations
25
Joint projects
Wholestic Market Design AGORAPHOBIA
You don’t always
get it right the
first time.
Now you have
experience
Are you a
Copernican or
a Ptolemain?
All power corrupts, but we need the electricity
26
“A new scientific truth does not triumph by
convincing its opponents and making them see
the light, but rather because its opponents
eventually die, and a new generation grows up
that is familiar with it.”
Max Planck, “Scientific Autobiography and
Other Papers”