k*k Concurso Nacional de Proyectos FONDECYT

Comentarios

Transcripción

k*k Concurso Nacional de Proyectos FONDECYT
k*k Concurso Nacional de Proyectos FONDECYT 2007
COMPROBANTE DE RECEPCION DE INFORME FINAL
NUMERO
RESPONSABLE :ENZO SAUMA SANTIS
RUT
:8827206-4
TIPO
:INCENTIVO A LA ETAPA :2007
: 7070239
DURACION : 1 A-no(s)
TITULO : ASSESSMENT OF THE ECONOMIC IMPACT OF TPANSMISSION INVESTMENT
S, AND ITS USE IN STUDYING TI-lE ECONOMIC-INCENTIVE STRUCTURES
DISCIPLINA :ESTRUCTURA DE MERCADO (AN (CS. ECONOM/ADMI)
FECHA RECEPCION :14/03/2008
TIMBRE RECEPCION
*** Concurso Nacional de Proyectos FONDECYT 2007
COMPROBANTE DE RECEPCION DE INFORME FINAL
NUMERO
RESPONSABLE :ENZO SAUMA SANTIS
:8827206-4
RUT
:INCENTIVO A LA ETAPA :2007
TIPO
: 7070239
DURACION : 1 Ano(s)
TITULO : ASSESSMENT OF THE ECONOMIC IMPACT OF TRANSMISSION INVESTMENT
5, AND ITS USE IN STUDYING THE ECONOMIC-INCENTIVE STRUCTURES
DISCIPLINA :ESTRUCTURA DE MERCADO (AN (CS. ECONOM/ADMI)
FECHA RECEPCION :14/03/2008
TIMBRE RECEPCION
(ONICYT - RECIBID
OFICINA DE PARTES
MAR 2008
)OPFJON
VI.- INFORME DE PROYECTO DE INCENTIVO A
INTERNACIONAL
Proyecto N2 7070239
Proyecto N 11060347
NÚMERO DE PROYECTO DE INCENTIVO A LA COOPERACION
INTERNACIONAL
NÚMERO DE PROYECTO FONDECYT REGULAR
13
Enzo Enrique Sauma Sanas
F nW
INVESTIGADOR(A) RESPONSABLE
PERÍODO QUE SE INFORMA
01
06
15
03
2008
HASTA
Shmuel Shimon Oren
University of California, Berkeley
NOMBRE COLABORADOR(A) EXTRANJERO(A)
AFILIACIÓN INSTITUCIONAL ACTUAL
10
12
DESDE
2008
FECHA PRESENTACIÓN
DESDE
FECHAS DE ESTADÍA
03
1
2007 ]
26
12
2007
HASTA
Describa las actividades realizadas y resultados obtenidos. Destaque su contribución al logro de los objetivos del proyecto
Regular. Si es pertinente, indique las publicaciones conjuntas generadas, haciendo referencia a lo informado en el punto 111
del informe de avance! final. Agregue los anexos necesarios.
El trabajo desarrollado con esta colaboración internacional fortaleció el proyecto de Iniciación en
Investigación N. 11060347, particularmente en los objetivos específicos 3.1 (Analizar las
propiedades de los dos sistemas de transmisión considerados que puedan inducir desincentivo
para realizar inversiones de largo plazo en transmisión que sean eficientes desde una
perspectiva social), 3.2 (Analizar los incentivos/desincentivas que tienen los propietarios de los
derechos de transmisión para realizar inversiones de largo plazo en transmisión que sean
eficientes desde una perspectiva social, en el contexto del sistema de transmisión actualmente
vigente en los Estados Unidos), 3.3 (Analizar los incentivos/desincentivas que tienen
inversionistas privados para realizar inversiones de largo plazo en transmisión que sean
eficientes desde una perspectiva social, en el contexto del sistema de transmisión actualmente
vigente en los Estados Unidos), 3.4 (Analizar los incentivos/desincentivos que tienen las
empresas generadoras de electricidad para realizar inversiones de largo plazo en transmisión
que sean eficientes desde una perspectiva social, en el contexto del sistema de transmisión
actualmente vigente en los Estados Unidos), 3.5 (Caracterizar las implicancias de la asignación
de derechos de transmisión en los incentivos/desincentivos que tienen las empresas
generadoras de electricidad para realizar algunas inversiones de largo plazo en transmisión que
sean eficientes desde una perspectiva social, en el contexto del sistema de transmisión
actualmente vigente en los Estados Unidos) y 3.7 (Analizar los incentivos/desincentivas que
tienen las empresas generadoras de electricidad para realizar inversiones de largo plazo en
transmisión que sean eficientes desde una perspectiva social, en el contexto del sistema de
transmisión actualmente vigente en Chile).
Para lograr los objetivos planteados, Enzo Sauma y Shmuel Oren realizaron conjuntamente las
siguientes actividades de investigación:
1.- Se estudió el sistema de transmisión eléctrica actualmente vigente en los Estados Unidos,
identificando las propiedades que puedan inducir desincentivo para realizar inversiones de largo
plazo en transmisión que sean eficientes desde una perspectiva social.
2.- Se estudiaron los incentivos/desincentivos que tienen los propietarios de los derechos de
transmisión para realizar inversiones de largo plazo en transmisión que sean eficientes desde
una perspectiva social, en el contexto del sistema de transmisión actualmente vigente en los
Estados Unidos.
3.- Se estudiaron los incentivos/desincentivos que tienen las empresas generadoras de
electricidad para realizar inversiones de largo plazo en transmisión que sean eficientes desde
una perspectiva social, en el contexto del sistema de transmisión actualmente vigente en los
Estados Unidos, y que permita la caracterización de las implicancias de la asignación de
derechos de transmisión en los incentivos/desincentivos que tienen las empresas generadoras
de electricidad para realizar algunas inversiones.
4.- Se evaluaron las reformas en el procedimiento de inversión en transmisión propuestas por la
Federal Energy Regulatory Commission (FERC) en la "Notice of Proponed Rulemaking on
Promoting Transmission Investment through Pricing Reform (RM06-4-000)" desde la perspectiva
de la eficiencia social de las inversiones de largo plazo en transmisión. Y se propusieron otras
variantes a la estructura de incentivos de inversión en transmisión usada en los Estados Unidos
con el fin de mejorar la eficiencia social.
5.- Se estudió los incentivos/desincentivas que tienen las empresas generadoras de electricidad
y otros inversionistas para realizar inversiones de largo plazo en transmisión que sean eficientes
desde una perspectiva social, en el contexto del sistema de transmisión actualmente vigente en
Chile. Para ello, se sostuvieron reuniones bilaterales con las siguientes empresas del sector
eléctrico Chileno: AES Gener (11-12-2007), Transelec (14-12-2007) y CDEC-SIC (19-12-2007),
además de una reunión con el ministro de energía, señor Marcelo Tockman (20-12-2007).
Además, se realizaron reuniones con profesores de la Pontificia Universidad Católica de Chile
(Hugh Rudnick y Sebastián Ríos entre otros) y de la Universidad de Chile (Alejandro Jofré y
Andrés Weintraub entre otros), para discutir sobre el tema. También se realizó una reunión con
el señor Bruno Phillipi para conversar sobre el tema y sus implicancias en las políticas públicas.
Adicionalmente, las cinco actividades de investigación mencionadas antes fueron reforzadas por
medio de la organización y realización de dos seminarios: uno de difusión general al sector
eléctrico (realizado el 12 de Diciembre de 2007, en el hotel Santiago Park Plaza - se adjunta el
programa y el material de dicho seminario) y otro de difusión interna entre los académicos
(realizado el 19 de Diciembre de 2007, en las dependencias del Centro de Modelamiento
Matemático de la Universidad de Chile). Mientras el seminario realizado el 12 de Diciembre de
2007 fue completamente organizado por mi, el seminario realizado el 19 de Diciembre de 2007
fue organizado conjuntamente con los académicos de la Universidad de Chile. Por otra parte,
estos seminarios ayudaron también a promover los resultados de nuestra investigación conjunta
a nivel nacional.
Como resultado de las actividades realizadas, se trabajó en la elaboración de una publicación
conjunta que contiene los temas señalados anteriormente. Un título posible para esta
publicación (obviamente sujeto a modificación) es: "Do Generation Firms in Restructured
Electricity Markets Have Incentives to Support Social ly-Efficient Transmission Investments?". Se
espera seguir trabajando en esta publicación durante el presente año.
Seminario Internacional
"Riesgo Energético:
Una Mirada Científica"
12 de diciembre de 2007
Hotel Santiago Park Plaza
Auspician
(4 Gener
naençsaAES
FOPtCVT
L
Seminario Internacional
"Riesgo Energético: Una Mirada Científica"
Miércoles 12 de diciembre de 2007.
Hotel Santiago Park Plaza
(Av. Ricardo Lyon 207, Providencia).
Programa:
8.00 - 8.30:
Acreditación
8.30 - 8.45:
Bienvenida y Conferencia Inaugural
Enzo Sauma,
Profesor del Departamento de Ingeniería Industrial y de
Sistemas,
Pontificia Universidad Católica de Chile
I
8.45 - 10.00:
"Diferentes Aproximaciones para Garantizar el Abastecimiento
Energético"
Shmuel Oren,
Earl J Isaac Professor del Departamento de Ingeniería
c
Industrial e Investigación de Operaciones,
Universidad de California, Berkeley
10.00 - 10.20:
Coifee Break
10.20 - 10.45:
"Seguridad en el Abastecimiento Eléctrico: El Caso Chileno"
'
10.45 - 12.00:
Juan Ricardo Inostroza,
Gerente de Regulación y Negocios,
AES Gener S.A.
"Gestionando el Riesgo Energético"
Shmuel Oren,
Earl 1 Isaac Professor del Departamento de Ingeniería
Industrial e Investigación de Operaciones,
Universidad de California, Berkeley
12.00 - 12.30:
Auspician
Panel de preguntas
'
Gener
sAE5
L DICT1
Departamento de Ingeniena Industrail y de Sistemas. Escuela de Ingeniena, Pordificia Universidad Gatolica de Ch ¡le, Tel. 56-2-3544272, 3544081 Fax: 56-2-5521608
Generation Adequacy: Theory,
Alternatives and International
Experience
Shmuel Oren,
University of California at Berkeley
Presented at
Santiago, Chile
December 12, 2007
Shmuel Oren, December 12. 2007
Economic Theory of
Generation Investment:
The Energy-OnIy Market "Goid
Standard"
Shmuel Oren, December 12 2007
Efficient lnvestment Paradigm in
Competitive Energy Only Markets
• Energy is priced at marginal cost with demand side
setting the price during scarcity hours.
• Competitive forces drive generation capacity, technology
mix, and prices toward a long-term equilibrium, where
the total amount and technology mix of generation
capacity is optimized with respect to supply and demand
preferences for reliability and cost.
• Fixed costs of generation capacity at long run equilibrium
are exactly covered by inframarginal costs and scarcity
rents.
• Forward markets and hedging instruments enable
parties to manage their risk exposure.
Shmuel Oren, December 12. 2007
Capacity Planning and Cost
Recovery with Marginal Cost Pricing
MI
.
fl.4AUD REDIfTIOI1
- CURTAJIEÜ 10A1
V1T'\ALL
.
GW
\TIME
1IE
I ---
COST
ÍIIHC
—4:T
re - 'l 14— T2
-
:
1
MM
.MIrJ
M?SC?1L COST FUNC
1 FPRICE DURATION c.UR'l
OLL
ENERGYREvl4IJE
C272
¿FI
w,r,
L \MG
2
Collecting Capacity Payments
through Scarcity Prices
I
ON
GNI
\TI4E
Tú
4- Ti -:4--- 12 — 12 4—
Tú
4- Ti *4— T2
:
COST
FUNC
- TJ 4—
- 3-
MARGINAL COST FUIJC
ii
10N (PRICE DUP.ATION CURVE)
YOIL
YE
C71`2
C31D
rppirY Pi4EUT (FI VIL'TD)
Illustration of Inframarginal Profits
and Scarcity Rents
DndI700-000pr,
$/MIh)
Pric e
0700-800 p.n
Scarcity rent
Inframargiria) Profits
Dend
900- l00ofl
—
1000—
Response
1-1
Demand al
200 - 300
ent
:^s,_
GEN CENO
al
GEN GEN 02
GEN
Optinlal
Capacity
MW
Shmuel Oren, December 12, 2007
3
Dynamic Adj ustment of Capacity
Mix Toward Equilibrium: Example
Generatortype
No of units
Unit capacity
Fixed cost
Marginal cost
Gi
50
80 MW
S926,400/Month
$15/MWh
G2
100
60 MW
$288, 000fMonth
$25IMWh
The demand is characterized by two demand functions
for peak and off-peak hours (PPrice, Q=Quantity)
Off-Peak: 420 Hours/Month
300 Hours/Month
Peak:
Q/1000
P30-Q/1 000
P=50-
Shmuel Oren, December 12, 2007
Disequilibrium
MC O'MWh
Peak
Off-Peak
'HW
0
.
HHWH
1
.
4000 000
0,000 MW
Peak price = $40/Mwh, Off-Peak price = $25/MWh
GI: 80*[(4015)*300+(25 .15)*420)] - 926,400 = $9600/Month (excess profit)
G2: 60*(4025)*300 - 288,000 = ($18,000/Month) (deficit)
Shmuel Oren, Decerober 12, 2007
4
Long Run Equilibrium
MC
50
7h
Peak
30
.Off-Peak
'
5 LW.HWWWH.
0
4000 5000 0000
0000 10,000 MW
Entry: 2000 MW GI, Exit: 3000 MW 02
Peak price = $41 /Mwh, Off-Peak price = $24/MWh
GI: 80*[(4115)*300+(24.15)*420] - 926,400 = O (Break-even)
02: 60*(4125)*420 - 288,000 = O (Break even)
Shmuel Oren, December 12, 2007
Reality Check
• Fuel prices and reservoir content can change quickly
affecting short term electricity prices
• Capacity mix take a long time to adj ust and such
adjustment occurs through a series of "boom and bust"
cycles
• While the capacity mix is out of equilibrium sorne
generators can make windfall profits at the expense of
consumers while other generators may not be able to
cover their investment cost
• Forward contracting can mitigate and smooth out the
effect of volatility in fuel prices and adverse weather
conditions.
• Forward contracting also enable risk sha ring between
different technologies and between generators and
con su me rs.
• Forward contracting suppresses the capability and the incentives of generators to exercise market power.
Shmuel Oren, Decerober 12, 2007
5
Impediments to Energy OnIy
Markets
• Inelastic demand and limited demand response
• Suppression of scarcity prices by r&iability motivated
operator actions
- Reliability unit commitments
- Out of market dispatch of must run units
- Deployment of reserves triggered by emergency conditions
• Suppression of scarcity prices by price caps and market
mitigation to prevent economic and physical withholding
(difficult to distinguish between legitimate scarcity prices
and market power abuse)
The Missing Money Problem
Shmuel Oren, December 12, 2007
Suppression of Energy Prices due
to Spinning Reserve Deployment
Shmuel Oren. Decernber 12. 2007
2
Suppression of Energy Prices due
to Out-of Market Unit Commitment
Markel pi,c, v/thOUt 00
n0 Maoy r,o Ur O E0I Pr.Ce
7c'3
j
/
000
a
€c2
720
14O0Po
Ix
o
•Oox
:o:
o
—Sefc.no M.l2
*,Goo
3,)W3
Sw3
2WA
25)w
X0MCfn'MA,1 —S05
—O-er'd
Source. ¡S0 NE
Scarcity Pricing Through
Administrative Demand Function for
Operating Reserves
et .tearin
Scarcity "Enery OnI' Mar
Normal Energy Ordy Mark Clearin
O
p I3WIl
$30.002
G*fl2{21130
2U$P11
/
Wfl
IZO 000
G*f101OIfl
\
1O000
$7.000
2l00$
— ---$30 ______
$30
-
QMW)
OjMW
nr.r
o I* r0p2c0) 2
rene.t$ nr rOl'nM rar;.s ZA 0w pICO
wlor W~2
h7 310 0OCC1 crnro opm.
r,ne 11,17m.
O
Shmuel Oreo, December 12, 2007
7
Further Challenges to Energy OnIy
Markets
• Steep supply function and uncertainties make
scarcity rents highly volatile and sensitive fo
market error in determining the optimal capacity
• It is practically impossible fo differentiate
legitimate scarcity rents from inflated prices due
to exercise of market power.
• Demand response is limited by technological
barriers and operational practices
Shmuel Oren, December 12, 2007
Challenges to Energy OnIy Markets
(Cont'd)
• Very high scarcity rents even it they are
legitimate are politically unacceptable (reason
for price caps)
• Low leveis of reserves foster collusive behavior
and market power abuse
• Capacity shortages cannot be resolved
overnight and while the entry occurs the
persistent scarcity rents result in wealth
transfers from consumers to producers.
• Exposures in the electricity supply chain are not
properly allocated to insure vountary, socially
efficient risk management practices by the
market participants (free riders)
Shmuel Oren, December 12, 2007
8
Alternative Approaches to
Generation Adequacy Provision
Variations on energy only markets with limited mitigation
(e.g. high offer cap) that rely on energy remuneration
and scarcity pricing to guide investment
Adequacy mechanisms based on capacity products
which takes two forms::
- Capacity payments to installed or operational capacity
- Capacity or hedging obligations imposed on LSEs which can be
met in severa¡ ways:
• Centralized capacity market or market for standardized hedges
• Bilateral contracting with regulatory verification
• Combinations of bilateral contracting with bulletin board trading of
standardized contracts or a central capacity market
Central resource procurement that can take the form of:
- Competitive tendering through an RFO process or bilateral
negotiation
- Strategic reserve contracts between the ¡SO and critica¡
resources
Shmuel Oren, December 12, 2007
Variations on Energy—OnIy Markets
• Energy-only markets have been operating in Australia,
Alberta, New Zealand and more recently in Texas and MISO
• In the Australian (the model for energy-only markets) market
spot prices can rise to $8,000/MWh, scarcity rents result from
"hockey stick" bidding by generators, consumers are
protected through voluntary hedging and bilateral contracting
while suppliers protect themselves against performance risk
by maintaining a reserve margin (no market for reserves)
• Spot prices in the Australian market are capped temporarily
(at $80/MWh) it average energy prices over a week period
exceed a threshold (never happened more than once in an
annual cycle).
• Suppliers are bound by "fair play" rules that allow them to
capture high scarcity rents but prevents them form "ambush"
strateg ies
• Competition is facilitated by disclosure policies that informs
generators of investment opportunities arid exposes bidding
behavior
Shmuel Oren, December 12, 2007
9
Price Spike in Australia June 13, 2007
wn,ts w*tAajIo
4-
11% ) '
.lawa,
TL-
1
444
FiB9.1S
C2
jí 148.60
2206M.
.... -.4.
3_?
1336712
7763
N
r1 -999.7ij
1551 MW 1
PUCT Market Power and
Resource Adequacy Rule
• Rule developed through a series of PUCT hearings and
stakeholders comments over a two year period. Rule finalized
in August 23, 2006
• Rule influenced by two major forces:
- Generators dissatisfaction with existing market mitigation
(including price cap and hockey stick mitigation
proced u res).
- Strong objection by retail energy providers (REPs) to any
type of capacitypayment that will reduce their "headroom"
relative to the 'price to beat".
• REPs see any capacity mechanism as interfering in
their business activity of managing wholesale price risk.
• Rule attempts to replicate the success of the Australian model and its variants that have been implemented in New Zealand
Shmuel Oren, December 12, 2007
and Alberta
10
Elements of the Texas
Generation Adequacy Rule
Simple definition of market power
- Safe-harbor (Less than 5% of installed capacity - "small
fish swim free")
- Voluntary mitigation plan for larger players
- Market power test is not proposed and is left to each
enforcement case
Energy-only mechanism is maintained (no capacity
payments)
Offer cap is proposed to gradually increase to $3,000/MWh
Wealth transfer backstop mechanisms
- If profits of generic peaker exceed $1 75,000/MW in an
annual cycle offer cap reduced to $500 for remaining of
year.
Shmuel Oren, Decernber 12, 2007
Elements of the Texas Generation
Adequacy Rule (cont'd)
• Removal of system-wide mitigation mechanism: Modified
Competitive Solution Method (MCSM), and $300 "Shame
Cap" (a 110w "hockey stick" bidding to produce scarcity prices
• Aggressive price transparency measures:
- Disclosure of price setter's ¡dentity within 48 hours if price
aboye a threshold
- Disclosure of individual offer curves 60 days after market
close
- Other inputs to be published no later than 90 days
Aggressive development of demand response:
- Advanced Metering (Project No. 31418)
- Demand Side Participation (Project No. 32853)
- Emergency Interruptible Load Service (EILS) for 1000 MW
- Long term alternatives for scarcity pricing under
development Shmuel Oren, December 12, 2007
11
MISO Energy-OnIy Market
• No payment for capacity and no "must-offer"
requirement
• $1000/MWh offer cap for energy + Scarcity rents for
operating reserves that are capped at $2500/MWh
• Administrative demand function for operating reserves
used to set scarcity rents for reserves during reserve
shortage. Scarcity rents for reserves paid to ah spinning
reserves and added to energy price
• State commissions wilh enforce mandatory load hedging
obligations
Shmuel Oren, December 12. 2007
Markets with Capacity Payments
• Pre NETA UK
- Hourly capacity payments to available capacity base on VOLL times
hourly LOLP. (pervasive gaming)
• Brazil
- Hourly ca pacity adder based on clearing price set by a fictitious slack
generator' representing capacity shortfall and an administrative price
curve representing shortage cost. (replaced by LSE forward contracts and
call options obligations for firm energy procured through central auctions.)
• Spain
- capacity payments for instaHed capacity commingled with stranded cost
payments based on technology mix
• Chile, Peru and Colombia
- Capacity payments to capacity in the merit order for energy and reserve
needs (based on mandatory cost-based offers). Payment are based on
annual LOLP calculations base on "effective capacity" and hydrological
forecasts, using computer simulation. (Chilean system was reformed and
replaced by contracting obligation and central auction of forward and cal¡
option contracts, Colombian system is being reformed implementing an
auction for multi-year firm energy options)
Shmuel Oren, December 12, 2007
12
Markets with Capacity Payments (cont'd)
Argentina
- Two types of capacity payments
• Hourly payment to capacity committed in the day ahead tor energy or
reserves
• Annual payment (determined annually) for long term thermal backup
capacity for dry years
• Different rules apply to thermal and hydro capacity
- Price cufting in energy in order to collect capacity payments resulted in
dispatch inefficiency
Italy
- Temporary capacity payments established in 2004 but it became semipermanent. It remunerates capacity dispatchable in the day ahead and
available on "critica¡ days". Not available to units engaged in physical
bilateral contracts or receiving other incentives (e.g. renewabies)
South Korea
- lnitially had separate energy markets and different capacity payments
for peaking capacity and baseload capacity. The two markets were
merged and a single capacity payment was estabished that vares with
the amount of installed capacity
Shmuel Oren, Decembeí 12, 2007
Capacity Payment Curve in Korea
Price
(Wc kW-h)
S
D
1.12z1,.
Capaoty
C&T - - - - - - - 1
<Figure
1 c'
4
r
Instafled
Capaczty
Demand and Sppty Curves of Caradty>
Shmuel Oren, December 12, 2007
13
EARLY ¡CAP MARKETS in the
US and RECENT REFORMS:
PJM, NYISO, ¡SO-NE
Sbmuel Oren, December 12, 2007
Problems with Short Term ICAP
Markerts
• "Bipolar prices due to ¡nelastic supply and
inelastic demand (price either zero or equal to
the deficiency penalty, eventually resulted in
price collapse)
• Deliverability problems (capacity procured in the
wrong location and system shows enough
capacity when there are locational shortages)
• Leakage (poor incentive structure and
insufficient performance penalties result in
opportunistic delisting of capacity)
• No opportunity for participation by new entrants
due to short product duration
Shrnuel Oren, December 12, 2007
14
Binary Pricing
[)eri anc
Prico
Deficiency penalty
True upply
--
Conipetitive
QLianlity
Shmuel Oren, Decernber 12, 2007
Monthly Average of PJM Daily Capacity
Credit Market Clearing Prices
Figure 1, Monthh' Average of FJM Di1v CApacin Credit Market Cleariug Priei (5/MW per doy) aiid
\o1urne (V(
A
W co rí
1
i;
JI
1
C-.°recPucc
15
¡CAP Market Daily and Monthly
Price Trends at PJM
Capaciy M3rkçd WIUhd-Aera PrIc
51.-0 Gú-
5,
W0 00
S2-0 00
Veo 00
$so 00
C1,I •
4.
-Iio.V/ CCM 1
PJM Capacity Market
¡SO NE ¡CAP Market Clearing
Prices
LC
1
,Lc
,-O3Ot&
IUi III
L.
M1 fil
.11 Axa1-O' c
E
,ai-Q L-?1-&b
Shmuel Oren, December 12, 2007
Prices in the ¡SO-NE ¡CAP Market
17
Average Net Revenue Shortage of
Combined -Cycle and Combustion Turbine
Generators in the ¡SO-NE
Net Revenue and Capacity Coste 1999 -2003
1200
---
-
10 00
500
088
oso
(yCooe
4 48
-
20O
0 00
541
o
FrnnE,.y
2
Ul- ^- -- -LowC
HigO Ce
cc
CC
Soe btwloa - Yang and Zheng 2004
LowC
CT
~C—
Cr
Key Elements of the NYISO ¡CAP Market
• Annual ¡CAP requirement set administratively relative to
forecasted annual peak demand (currently 18% reserve
ma rg 1 n)
• Locational requirements for NYC and Long Island and rest
of state (in NYC 80% of demand must be met with lfl City
generation)
• [SE ¡CAP obligation can be met with bilateral contracts
• Centralized capacity market conducts auctions for three
prod u cts
- Seasonal strip auction (semiannual)
- Monthly auction for adjustments to seasonal strip
- Daily spot auction (clearing residual needs for LSE) using an
administrative downward sioping demand curve
• Pena Ities levied by SO on LSE for noncompliance via a
deficiency payment mechanism
• import capacity qualifies for ¡CAP only if backed by
transmission rights
Shmuel Oren, December 12. 2007
18
NYISO Summer 2007 Locational
Demand Curves for UCAP
Suínlrt, 2007 D,niaird C.wvc,
o JcArraqUIn.,
NYISO Locational Capacity Auction
Resu Its
120.00
•LI
—. 100.00
prc
1
a ROS
80.00
2
60 00
40.00
1
20.00
1
1
1
iSs
1
19
Stabilizing Effect of VRR on the
NYISO ¡CAP Market
- nnc
•e
•
1
s
1
\
1
1
•-•-.
IW1
Ll
lLll
}11
l91I ¶
Shmuel Oren, December 12, 2007
Does the NYISO Approach Work?
There is no evidence that the demand function approach
has achieved its primary objective of attracting new
generation resources.
The New York Public Service Commission (NYPSC) has
been exploring approaches such as imposing load
hedging obligations on LSEs and resource procurement
options.
In a recent press release dated April 18. 2007 the
NYPSC stated that the existing wholesale etectricity
market structure has not led to much merchant-driven
supply nor shown much promise for new merchantdriven entry, both of which are needed in order to meet
reliability needs.
Shmuel Oren, December 12. 2007
20
Key Features of the PJM Reliability
Pricing Model RPM
• Locational capacity requirements in 23 capacity
zones established by PJM 3 years forward. Procured
resources must satisfy deliverability conditions to the
designated zones.
• LSEs assigned local capacity obligations based on
load share in each zone and must procure the
capacity three year forward for a one year minimum.
• LSEs allowed to meet their capacity obligation with
generation, existing or planned demand resources
and plan ned tía nsmission upgrades that increase
import capability into constrained areas.
Shniuel Oren. December 12, 2007
Key Features of RPM (contd)
• PJM conducts an annual three year forward procurement
sealed bid auction for local capacity on behalf of LSEs
that are short by the deadline
• Variable resource requirement (VRR) based on a down
sioping demand curve used to adjust procurement
quantity according to the auction clearing price.
• Explicit market power mitigation rules addressing market
structure concerns in the capacity markets
• LSEs can select an "Opt-out alternative" by submitting a
four year resource plan, including the RPM delivery year
• Capacity payments reduced by an energy rents
adjustment based on a six years average enegy
revenues for a generic peaking unit.
Shmuel Oren. December 12. 2007
21
RPM Demand Function
(per August 2006 Settlement)
Price t
CONE=$65.000/MW/YR
1 .5xCONE
. CONE - Cost of new entry
\ TC - Target Capacity
0.2xCONE 1.................
Capacity
.98xTC
1 .O5xTC
Shmuel Oren, December 12, 2007
Proposed LICAP Demand Function
Price
Locational 1CAP
OC
2'FC ....
1 U4
cm.ax
M,r'
Capccity
histoncal ..eage cacflv)
The LICAP proposal "killed" by strong opposition from the states
served by ¡SO-NE Shmuel Oren, December 12, 2007
22
Key Elements of the NE Forward
Capacity Market (FCM)
• Capacity resources procured 3 years forward for one year
commitment (new generators can get 1-5 year contracts
with payment based on first year clearing price, indexed for
inflation) with payments at delivery time.
• Zonal procurement based on 3-5 year "binding" load
forecast (by ¡SO), local reliability needs and transmission
i m its
• Fixed zonal target quantities (no demand curve)
• Procurement through descending clock auction for existing
and new qualified capacity resources, starting with
2xCONE
• Uniform clearing price for each vintage of procured
capacity in a zone (except for capacity of new generators
procured in previous years).
• Transition mechan sm with fixed capacity payments to
existing resources starting Dec. 1, 2006 ($3.05-4.10/kWShmuel Oren, December 12, 2007
Month)
Key Elements of FCM (2)
Annual reconfiguration auctions (static double auctions)
allow balancing of obligations in response to changing
supply and demand conditions through incremental
procurement, capacity release, commitment trading among
resources and de-listing by committed resources
- Auction heid at 28, 16 and 4 months ahead of delivery
- De-listing is annual or permanent
- Tender is annual commitment
Monthly and seasonal reconfiguration auction allows
adjustments of annual commitment during commitment
period
Bilaterally contracted load can self-provide its capacity
obligation by offering it as price taker into the auction.
Capacity payments ar reduced by peak energy rents (PER)
based on energy profits of a generic generator with 22,000
heat rate and indexed fuel cost. PER computed for each
month based on the last 12 month average.
Shmuel Oren, December 12, 2007
23
Key Elements of FCM (3)
• PER deductions in any month is capped to two month
capacity payment and for the year to the annual capacity
payment
• Capacity must be available at predesignated critica¡
hours. Unavailability during these hours is penalized
through proportional red uction of net capacity payment
(net of PER)
• Demand and intermittent resources can participate in
capacity auction.
- Demand and intermittent resources are not subject to peak
hours availability but are subject to average performance criteria.
- Demand and intermittent resources capacity is derated based on
historical performance
- Demand capacity is credited for Iosses and avoided reserves
Shrnuel Oren, December 12, 2007
Key Elements of FCM (4)
• Al¡units have a must offer obhigation unless de-usted
• Units can submit de-Iist bids for one year or permanent
de-Iist, subject to reliability review by system operator)
and market power review (by market monitor). De list
price aboye 0.75xCONE must be justified by cost
documentation.
• New generation units must either bid abo ye 0.75xCONE
or enter as price taker (to avoid price suppression by
large LSE procuring new generators through bilateral
contracts).
• If there iS fol enou9h competitive new capacity to set
price, auction clearing price is set by an alternative
(administrative) price rule.
• Al¡ settlements to generators and billing to load (on a
peak monthly load prorata allocation) occur at
performance timeShmuel Oren, December 12, 2007
24
The FCM Descending Clock Auction
Aggregate supply curve
Pre
staIng price
po
------ ------------------------ v
jRound 1
excess supply
fRound2
P2
3
P3
4
'5
nd
P6 -
ciearng pvce
Qcertily -
Target Capacity
Market Power Suppression and the
Connecticut Strategy
• The FCM procurement auction is designed to mitigate market
power Qn supply side by aliowing participation of new entrants
• Price set by new entrants ensure that capacity payments
solves missing money problem
• Energy prices still capped at $1 00011VIWh for de-usted and
import resources (Cap plus PER aimed to mitigate market
power in energy)
• The Connecticut Public Service Commission has issued an
RFP for procurement of new generation capacity with the
intention of offering that capacity at reduced prices into the
FCM auction
• Such a strategy constitutes exercise of monopsony power
which can suppress the auction clearing price and reduced
payment to incumbent generators.
• FERC ordered that new capacity procured bilaterally out of
the FCM auction must be offered as price taker so that it
counts toward activation of the alternative pricing rule
designed to protect incumbent generators.
Shmuel Oren. December 12, 2007
25
Relaxing the Energy Price CAP on
Uncontracted Capacity
Capacity value defined as a cali option on unmitigated spot prices
equilibrates to either CONE (if new entry sets the price) orto the
expected energy rents given available capacity
U ncontracted
Capacity
Price
New Capacity Offered
2xCONE
Contracted
CONE Capacity \
expected profits from unmitigated spot energy market
Clearing
Price
Capacity
Target Capacity
Shmuei Oren, December 12, 2007
Implied Demand Curve for Capacity
Based on Cali Option Value
cnn,,,
Cali Option
Value
Cali Option Pnce
Per Ilour
ap -Stnke
loo
Pece Duration Curve
Time
Capacity
implied Demand Curve for Capacity
Al optimal capacity Q Ihe cal¡ oplion price
= Average Hourly {CT fixed cosI - CT energy profit wilh price capped al Slrike} Shmuel Oren, December 12, 2007
26
Capacity Prices in the Northeastern USA
-
-
j
.xc
•x 2-r
L
:-?-:co;
,•':cl-
so oo
410 00
cio
PJM
PJM
EMPAC
PJM
SWMkAC
Market
PJM - Auc
Avnre t
PJI - RPM AJctor
PJM- RPM 4ction
PJ1. RPM &jtJo
ISOE - Auci Average Est
IÜE - Trar rirc.'l anr0
rIriso - ALcrj Average SaL
N viSO - Aactran Average Ea
rIYISO- ructjy Average a1.
ISOl€
Region
wc
200&'2007
1€
ROS
20O72008
IL
RTO
EMAAC
SV/AAC
LI
N'C
ROS
LI
1330
$74r
10945
2761
50
The California Resource Adequacy
Requirement (RAR) Program:
Putting Humpty Dumpty Back Together
Fftj
wt.L,
ttj st o
ktj
grt fLL;
H-vwpt
ALL tke dg iorse iic oLL ti K
oLc19 pt uci&ptj toger
it
co
Shmuel Oren, December 12, 2007
27
Development of the California (RAR)
Program
• lniplemented by CPUC in coordination with
CEC and CAISO.
• Ruling developed through a proceeding involving
preliminary orders and stakeholders comments
over a three year period.
• RAR program amounts to "virtual" re-bundling of
generation and distribution.
• Applies only to CPUC - Jurisdictional LSEs.
• Designed to assure resource availability, protect
customers and reduce dependency on RMR for
reliability assurance.
Shmuel Oren, December 12, 2007
Elements of the RAR Program
Short Term:
• CPUC "Phase 1" Ruling in June 2006
established Local Capacity Requirements for
2007 and defined a tradable capacity product.
• "Phase 2" began September 2006 to develop
Capacity Market or alternative design for
implementation in 2008-9 time frame (twa majar
camps arguing for a North Eastern style
centralized capacity market vs. continuation of
the RAR program)
Long Term
• Empowers lOUs to enter into contracts with new
generation capacity on behalf of LSE in theír
territory
Shrnuel Oren, December 12, 2007
28
Elements of the Short Term RAR
Program "Phase 1"
• California Energy Commission (CEO) provides annual load
forecast.
• CAISO calculates year ahead local reliability requirements for
generation capacity in fine local reliability zones considering
generation expansion plans, transmission limits and operational
response possibilities.
• CPUC-Jurisdictional LSEs are assigned (by CPUC) bilateral
contracting obligations to cover 115-117% of the annual peak
load, based on load share in local reliability area.
• Procurements by lOUs, monitored by independent reviewer, are
deemed to meet prudency criteria (cost can be passed to
consumer).
• Seven local reliability areas in PG&E territory combined into two
areas to increase liquidity.
Shmuel Oren, December 12, 2007
Elements of the Short Term RAR
Program "Phase 1" (2)
• lmports allocated on a load share basis to reduce LSE
obligation.
• LSEs procure resource-specific bilateral capacity (and
energy) contracts through direct negotiation with
generators and via bulletin board.
• CPUC Energy Division monitors with RAR of the
contracted resource portfolio for coming year.
• Exemptions if capacity resources are not available or
are not offered below $40/Kw-Yr. (max offered by
CAISO for RMR)
• CAISO Checks procured resources against local
reliability needs and procures supplemental backstop
RMR resources as needed (cost uplifted).
Shmuel Oren, December 12, 2007
29
Long Term Procurement Program
• Recent July 2006 decision recognizes that short term RAR
cannot address new investment needs.
• Decision is transitional
• lOUs are designated as entities responsible for new generation.
• bU not responsible for managing the energy produced by new
generation capacity.
• Value of energy to be revealed through a competitive auction for
tolling rights to the energy from new generators.
• The cost of new capacity given by the contract cost Iess the
energy value.
• Cost of new capacity over up to ten years can be allocated to
LSEs in the bU territory
• LSEs can opt out from the capacity cost albocation provision if
they are resource adequate of a sufficiently long time period.
Shmuel Oren, December 12, 2007
RESOURCE ADEQUACY IN
EUROPE
Shmuel Oren, December 12, 2007
O UCTE
Nordel
\
ME GB
O
Ireland + N. Ireland
y
• Turkey
O IPS/LJPS + Baltic Statos
O red Curerit
n:erco1nedIion Irie
UCTE lii 2004:
- peak domand 400 G • annual load 2500 TWh
- onnrgy oChaned 250 TWh
-450 M ¡nhabitants
25 countrlo
-40 TSOv
0'
Background
• Genera) frame'ork European Drectves tTPA and vhoIesa!e
competition n 1999. extended to reaiI compettion ¡o 2007)
• 13 of European countries have aiready mplemented fuH retail
campetition, Me otber part will do it in 2007
• 1 reulator for each country
• 1 or more) Transmission System Operator for each country
• Transmission charges defined at country leve) i l mainly postage stamp)
• Markets are interconnected, but freqjent congestions occur at the
national borders
• Particu)arty strong interaction
2005)
.vih 002 market operat ono) snce
Shrnuel Oren, December 12, 2007
31
Dominant Market Design in Europe
Forward and Futures
Markets
Day ahead ntra-day Balancing
Market
Market
Market
OTC and PX
cI
anPX
(andPX)
TSO
Time
Ye&s. months. weeks
before delivery
Day
befare dolivery
Time
of delivery
OTC : Ove- the Counter . bilateral transactior,s
PX : Voluntary pavor exchango, tradg s tandard praducts
TSO TrarsmIssion System C)perator
Shmuel Oren, December 12, 2007
Investment History in Europe and
Resulting Generation Mix
rw
wO4 uS.d by tt,c
IhonnJI gone,M,',ç MIII. CMnII*s aliad Ii 9 IiI!OpSarI Mo j arla. duda9 rla Ial doMada
IAU, BE, IT, FR, GE, BE. SP 8W, UR)
snumo • iaoch Iru*iofry MosPy
Mix of powei generated in sorne EU countries (compared to the US one(
¡ata 2OO4
100%
100% j_ i
i4i1-i4
60%
• natural ga 60% cl
40%
40%
40%
•nuclear
20%
20% ........................................
20%
n- hydro
Europe-15
USA
DF
32
Assessment of Generation
Adequacy ¡n the UCTE
Soenario includin g future potenfial investrnents
Conservative
2006 2001 2008 2000 2010 201
1
2012 2033 2014 2015
2o& U TE 2005
Sources of Generation Investment
Risk
• Base-load and mid-rnerit units
- Long-term cash fle; rísk
- Regulatory risk
- Business model risk
• Peaking units
- Revenue risk
- Regulatory risk
• Regional exposures
Balancing ongoing fuel portfolio
Shmuel Oren December 12, 2007
33
European Commission Framework
for Generation Adequacy Ruling
A
ter >IepPrarFueIs
Long Terrn
Sho1 Term
*
System
AdequacMarket
y Adequacy
opeJionai
Securtv
Gene ratiori
Network
Ade(.juacy Adequacy
Shmuel Oren, December 12, 2007
January 2006: a New European
Directive Qn Security of Supply
• Triggered by the outages occurnng in 2003
• Addresses generation and network issues
• Main features affecting Generation Adequacy
- each Member State is responsible for the Secur ty of Suppy withiri
its borders
- Obligation of monitoring and reporng on generation adequacy
- To attract sufficient investment : make use of rnarket-based
rnechansms or (last resort) l e rijer-c for new capacity
Shmuel Oren, December 12, 2007
34
Foundation for the European
Energy PoIicy
"
ç
V
o
II
.
o
o
1
Environment
Noture
preservation
Shareof
renewables
Kyoto
8 post Kyoto
Shmuel Oren, December 12, 2007
Resource Adequacy Mechan isms
in Europe
Competltive tendering
Capacity payment
Stratoglc reserves
Peak unit procuroment by TSO
Capacity market with TSO as single buyer
4. *
eDF
Shrnuel Oren, Decernber 12, 2007
35
Resource Adequacy Mechanisms in Europe (2)
• Competitive Tendering: (France, Germany, Portugal)
- A market friendly extension of the traditional central planning paradigm in the
monopolistic context. Used by governments or regulator lo correct forecasted
supply and demand imbalances
• Strategic Reserves: (Sweden and Norway 2001-2002, UK, Nederland)
- TSO enters into medium term contracts with producers paying them for keeping
capacity in reserve wtiich is dispatched at the discretion of the TSO. Energy pnce
set in the contract
- Quantity approved by regulator (in Sweeden 2GW out of 27GW peak load)
- Cost of contracted uplifled through transmission tariff
• Peak Unit Procurement: (Sweden, Finland)
- TSO owns and operates a limited amount of peaking capacity
- Units dispatched by TSO for real time balancing and quick response reserves.
Should fol be used lo often in arder lo minimize distortion of price signals.
• Capacity Market with TSO as single buyer. (Norway)
- Implemented since the shortage of 2001-2002
- Optional monthly market procuring capacity from generators and demand
reduction
- Procured capacity cannot seli energy in day ahead but can selI in real time at
market prices
- In Norway about 2GW out of 23GW peak load are procured.
- Cost uplifted lo transmission tariff
Shmuel Oren, December 12. 2007
Questions?
a,
Shmuel Oren, December 12, 2007
36
Risk Management and
Financia¡ Instruments in
Competitive Ekctricity Markets
Shmuel Oren,
University of California at Berkeley
Presented at
Santiago, Chile
December 12, 2007
'reo [01-ocr lo, 1 .21,7
Typical Electricity Supply and Demand
Fu n cti o n s
$100 Morgr ecl 0001
lS/MOrO 1
090
Marginal Cont
SSO
Nra. pl.rt.
07(1
$60
$50
$40
1
$30
$20
0/0
$0 0
20000
40000
60000
De,nnd IMWI
80000
100000
'ron Deçrrnl,er 12, 2,Hl"
1
Electricity On-Peak Spot Price
$160.0
$8000
Spot Price
$1400 -
ERCOT on-peak
- COB on-peak
$70 00
5120.0
$6000
$1000
$50,00
580.0
$40.00
$600
1
:::
JL4
520 W
p"'4
-
$00
12111199 31101199 6/16/199 91261199 11411997 41141199 71231199 10131119 21811998 5/191199
5
6
6
6
7
7
97
8
$000
T/me
}rcr'I)t'(,I,Içi 2
3 I'
PJM Real-Time Market Prices
500
. 450
400
• 350
300
1-
250
' 200
150
100
50
O
Apr-98 Sep-98 Mar-99 Sep99 Mar-00 Sep-00 Mar-01 Sep-01
.3
Volatility can Result from Market Power
(e.g., Economic Withholding by Pivotal Supplier)
$1.000
$800
HiJhest
Ofler
Price
$600
$400
$200
$0
0
200
100
300
400
Cumulative MW Ofíered
2.
Volatility can Result From Rigid
Rules + "Hockey Stick" Bidding
$1,000
.
.... .....
$800
Highesl
Offer
Prce
$600
:
0
100
200
300
Cumulalive MW OlTered
)rt
400
L)ccuihc 2.
3
Examples of Extreme Price Events
The wholesale eiectricity price in the US Midwest in the summer of
1998 rose to $7000/MWh, while a normal range was
$30/MWh$60/MWh.
South Australia and Victoria, Jarivary 16, 2007: persistent high
temperatures led to record consumption of electricity resulting in
wholesale spot prices abo ye $5000/MWh.
The Texas Ice Storm 2125-2712003
..i ERCOT had to procure 100% of the balancing offer stack during severa¡
hou rs
j Market clearing price was set by 1MW offered on a regular basis at $990
by a OSE that offered the rest of ¡ts balancing energy at $200/MW or
less.
$17 million changed hands in few hours
One reta¡¡ energy provider (TCE) that was over-exposed to real time real
time spot prices declared bankruptcy.
ren Drrernl.rr, 12,
Price-Setting Offer Curve at ERCOT, 2/2512003
$1000
$990
2 $800
$600$400
$200
$149
$0
0
100
50
150
MW
(rrn' l)
'rrrIrr 12, 2Ufl7
4
ERCOT MCPE by Interval, 2/25/2003
$1,000
$800
--
$600
$400 -
______
-
I-V7 A
$200-------
-
-
$0
0:15
3:15
6:15
9:15
- D
12:15
2. 1 , ú7
15:15
18:15
21:15
9
5
Forward Contracts Smooth out Price
Volatility
11 f ( onrnodÍtv (boris
P.JM \\e'tez'ii
Eleclaicits' (J-NI, N\MEX)
\r1l Piioi (kto,
104'2305 C-8931
087 88 k84.98 L87 58 MOv A,9 3 Iin.0
Z0.0c
-11000
-10o_0c
1-nono
-80 00
1*00
1
•44 00
-30_00
ROl 8',84S 2000
Voluin. Op.r Inl.r.st 1100
-
--•
7.
t-
u
-
II
2.
Both Price and Quantity are Volatile (PJM Market Price-Load Pattern)
1200 -
Period: 4/1998-12/2001
800Ql
1.)
600-
1..
Ql
E
P 400
0)
•
.4..
200r
0
10000
20000
30000
40000
50000
60000
Actual load (MW)
n.
Price and Demand Correlation
Electnclty Demand and PilcC in California
16000
16000
160
160
p
1
140
=
000
100 z
so 2
6000
60
4000
40
2000
20
9. July 16. July 19
Correlation coefficients:
0.539 for hourly price and load from 4/1998 to 312000 at Cal PX
0.7, 0.58, 0.53 for normalized average weekday price and load in Spain, Britain, and
Scandinavia, respectively
Drrubcr 12. 2011'
13
Electricity Supply Chan
co-;:
ÇJQ
Wholesale
electricity market
(spot market)
Generators
Customers
(end users served
at fixed regulated
rate)
LSE
load servin
entity
A similar situation rs faced by a trader with a fixed price load following obligation
(such contracts were auctioned off in New Jersey to cover default service needs)
l)rrrul,r 12,201'
4
7
Volumetric Risk for Load-Serving Entities
Properties of electricity demand (load)
Uncertain and unpredictable
Weather-driven - volatile
Sources of LSEs exposure
Highly volatile spot price
Fiat (regulated) reta¡¡ rates & limited demand response
Electricity is non-storable (no nventory)
EIectricity demand has tobe ser-ved (no "busy signal")
Adversely correlated wholesale price and demand
Covering expected load with forward contracts will result in a
contract deficit when prices are high and contract excess when
prices are Iow so the LSE faces net revenue exposure due to
demand fluctuations
- I),nI,e, 2 31)11
15
Generators' Risk Exposure
Fuel cost risk
Quantity fluctuation
Plant outages
Wholesale price volatility
Transmission outages and congestion
Regulatory risk (market mitigation, operator
intervention, suppression of scarcity rents)
Ability to recover fixed costs.
12.
1-
201
8
Risk Management is a Necessary Condition for
Efficiency and Reliability of Competitive Power Markets
Generators, load serving entities and power marketers can manage
their risk through the use of long term contracts and financia¡
¡nstruments
Efficient risk management requires tools:
Financia¡ instruments and contracts compatible with the industry
risk profiles
Pricing methodologies
Hedging strategies and portfolio optimization techniques
Availability of such toois will increase the industry comfort level with
the use of financial instruments and facilitate the iiquidity and
efficiency of markets for risk.
Market design reiles on the ability of participants to reallocate risk
efficiently
Even in absence of liquidity, understanding financia¡ instruments can
help in structuring and pricing bilateral contracts and market
mechanisms (e.g. transmission rights, capacity mechanisms)
- Der.I,er 12. 2XI7
17
Financial Instruments for Electricity Risk
Management
Electricity contracts and derivativos
Forwards or futures, contracts for difference (CFD)
Plain-vanilla options (puts, calis, one sided CFD)
Swing options (options with flexible exercise rate)
Spark spread options (option on the difference between the electricity
and heat-rate-adjusted fuel spot prices)
Locational spread contracts and options ( contracts and options on the
difference between two locational spot prices)
Structured transactions (tolling contracts, default service contracts)
Temperature-based weather derivatives
Heating Degree Days (HDD), Coohng Degree Days (CDD)
Power-weather Cross Commodity derivatives
1 Payouts when two conditions are met (e.g. both high temperature &
high spot price)
Composite Iristruments
j Callable forwards (a forward contract bundied with a cal¡ option)
Callable forward with multiple exercise points and strike prices
Putable forward (a forward contract bundied with a put option)
1
1$
Electricity Forwards
Payoff per MWh = S, -
I
(spot prce at matunty - forward price
CDF is a strip of forward contracts
Most common forni of bilateral contracts
Contract written for: On-Peak (06:00-22:00), Off-Peak, 24 Hours.
Spot settlement price for the day averaged over period.
Generators such as IPP are natural sellers, LSE are natural buyers
Day ahead and hour ahead forwards traded through organized
exchanges, week ahead and longer traded OTC, through brokers or
directly by parties.
One hour ahead forwards are physical but Day ahead and longer
are financia¡ in nature (can be settled financially)
run Drçtnihcr 12, 2(M}7
lO
Electricity Futures
Firsttraded Qn NYMEX in March 1996, same payoff as
forwards but highly standardized
Smaller quantities than forwards. For example MidColombia electricity futu res traded Qn NYMEX specified
432MW1-1 of firm electricity, delivered to the MidColombia hub at 1 MW per hr., 16 on-peak hour per day
(corresponding forward has delivery rate of 25 MW per
hr. for the same delivery period in a month.
Exclusively traded in organized exchanges (the
exchange is the counter-party)
Settled financially - lower transaction cost
Strict margin requirements - lower credit risk.
12,20117
10
Electricity Plain
Cali and
Put Options
Payoff of Cali per MWh = Lla.v(S, - A.
K is the strike price
a.v( K .S
Payoff of Put per MWh =
Cali options can be used by an LSE to insure supply at a fixed price
K without the obligation of taking deiivery (can be executed by
buying from the spot balancing market and exercising the cal¡ option
to offset extra cost if the spot price exceeds K)
Put option can be used by IPP to lock in a fixed sales price (can be
executed by selling the power into the spot baiancing market and if
the spot price falis below K exercising the put option to offset the
revenue shortfall.
Cali options obligations with physical cover (i.e., the seller must have
the physical resources to cover the option) have been proposed as a
generation adequacy mechanisms.
( rn .
12. 2lX1
21
Cali Option Value in a PriceCapped Energy Market
Energy Pricet $/MWh
Cap
Annualized
Cali option value
$/MW
Price duration curve
—
Strike
8760 Hours
Or
-
n11.r 12. 2111P
22
11
Cali Option Premium as Function of
Generation Capacity Based on
Opportunity Cost of Selling at Spot Prices
Arinualized Cali Option
Value
Energy Price
Cvii Option Price
Per Hour
Cap
Strike
Cap -StrikE
8760
Price Durahon Curve
Hours
Capacity
irnplied Demarid Curve for Capacity
Al optinial capacity Q ihe cal¡ optiori price
= Average Hourly (CT fixed cosi - CT eriergy prof it wiih price capped al Sthke
iren - Dc crrilc 12.
!iaL
23
Swing Options
Settled like plain options but the exercise time is flexible.
The holder buys a certain number of MM and can
exercise the options within a certain time period subject
to restriction (upper and lower bou nds on the rate of
exercise, minimum and maximum total quantity)
Swing Cali options useful for a load serving entity that
must serve its load at fixed prices to offset risk due to
load fluctuations
Swing Cali options can be used as a model to price
interruptible service contracts with bounds on total
number of interruptions and interruption frequency
Swing Put options can be used by a load serving entity
to offset losses due to underutilization of a "take or pay"
procurement contract.
(irrn - Decrnrber
22, 21 M) -
2
12
Spark Spread Options
)
is the strike heat cate)
(
•Payoff per MWh = . t Iav( , -
• Represents the financia¡ arialog of generation capacity
• Useful in pricing capacity value
-Can be used to hedge investment risk in generation capacity
-Can be used instead of cal] options in generation adequacy mechanism
(cheaper due to stripping of the fuel cost risk)
8000 heat
LFhi
iIÍL U
iír.Sg
Se
\r2
r'
-.
-
rce
Spark Sprmd
S(MWh)
Time 1
Drmliei ¡12(11-,
Locational Spread Contracts and Options
Useful as financia¡ transmission rights (FTR) or for hedging
congestion charge risk of bilateral energy contracts in markets with
locational marginal pricing (LMP).
FTR Obligation settlement per MWh from A to B =
FTR Option settlement per MWh from Ato B= lJv(
-S;'. 0)
.II
[Ectj4
o:
Power
at
r;:::ri
electrtclty
Transmissio
[OSt?j
-
J
Sign bilateral supply
NYMEX COB/PV Futures Price Series
$60,0 -
565.0
550.0
540.0
NYMEX COB and PV Futures Price $ 55.0
$45,0
J
030.0
li,
ii
$35.0
020.0
- $20,0
PV 7/98
- - PV9/98
- COB 7/98
COB 9/98
$10,0
$15,0
00.0
55.0
430197 6.19/97 8/0/97 9/27/07 11/16/97 1/5/08 2/24/98 4/15/08 6/4/98 7/24/98 9/12)98 11/1/98
/ mi Dctcmmrhçr 2, 21Km'
Structured Transactions
Contracts thai are tailored to ihe specific risk
management needs of the seller or the buyers and are
either covered by physical assets or through hedging
u Tolling contracts (typically covered by physical capacity or
future investment in capacity): similar to forward contract
but bundled with real options. For example the buyer pays
an upfront fixed premium in exchange for the option to
control the scheduling of the plant with the ¡SO or take
delivery at specific times. Also may be restricted by
operating constraints, number of starts etc.
u Default service contracts: fixes sales price but aliows the
buyer to take variable quantity (LSE implicit contract with
regulated retail customers)
/ /mr,m . D'r,mImKr 12, 2(9/7
14
Weather Derivatives
Most commonly traded weather index
HDD and CDD derivatives
Power demand is correlated with HDD and CDD
Transparency
i Measured by independent source
Forecast and data are widely available
Hedging and diversification of weather risk
Weather derivatives provide hedging capability to many
industries whose performance is weather sensitive
(energy, agriculture, tourism)
Weather derivatives enables risk sharing and
diversification of weather risk among industries
u HDD (Heating degree days), CDD (Cooling degree days)
CDD=J T, -65I where Td
JIDD=65-T,
=
CDD(HDD,)
Monthly orseasonal CDD(HDD):
Six attributes of WDs: type, period, underlying index, City,
strike, and tick size.
29
I),hr 12, 2007
HDD vs. Power Demand
Miit0Iv UK NatoroI Eec1ricity dErrr
tor Londc.n H2t0ro helw902 AuQU''
0
50
130
2c)
0
250
HOPo frr Lonlon Henthrcw
Source: Weather Risk Advisory
Ciiis1 HDL
june
300
350
400
2, 2007
15
How the HDD Derivatives work?
Winter HDD Put
Profit $
Index: HDDs
Location:
Unhedged profit
______,/" \\
,
6
Z
Airport
Period: Dec, 2004
no::
HDDs
HDD
ff
Tick: $10000 per
degree day
Limit: $2000000
Buy a Put option with strike 300: if HDD<300, payout (300HDD)*$tick
Generator can offset profit shortfall due to coid weather with weather derivatives
12
The Market for Weather Derivatives
Early WDs were traded over-the-counter (OTO). First contract
with a weather based contingency was written in 1996
To increase Iiquidity and remove credit risk, the Chicago
Mercantile Exchange (CME) started an electronic market
place for WDs in Sep. 1999
CDD and HDD futures and options are traded for 19 cities of
the US; 9 in Europe; 6 in Canada; and 2 in Japan.
Notional value of CME weather products in 2004 was $2.2
billion, and grew ten-fold to $22 billion through September
2005, with volume surpassing 630,000 contracts traded. In
2006 the value of weather derivatives traded grew to $45
billion.
1 ,,,, -
j)eI,,
12. 21(07
Composite Instruments
Callable forward - Rephcates a curtaiiabie service contract
(Cali premiumRate discount)
ects Struke
^S^ely
K
:^^: ^K
Exercises Option
Putable forward - Rephcates a dispatchabie IPP contract
(Put premium = Standby payment)
Buvs 1 Put
—:::El ^.n
Selis 1 Put
rn Dbr 12
21ff
Volumetric Hedging Model Setup
One-period model
At time O: construct a portfoiio with payoff x(p) j At time 1: hedged profit Y(p,q,x(p)) = (r-p)q+x(p)
Load
LSE
marke
lx))
POIO
for a delivery at time 1
ObjeCtive
Find a zero cost portfoiio with exotic payoff which maximizes
expected utility of hedged profit under no credit restrictions.
- D.-hr 12. 21
14
17
Mathematical Formulation
Objective function
Utiity function over proflt
max E[U[(r - p)q + x(p)]]
Joint distribution of p and q
1
max f fu[(r - p)q + x(p)]f(p,q)dqdp
,(p)
Constraint: zero-cost constraint
1B
E[x(p)]
,
!
A contract is priced asan expected
discounted payoff under risk-neutral measure
2 risk-neutral probability measure
\B: price of a bond paying $1 at time 1
2. 2II
Optimality Condition
The Lagrange multiplier is determined so that the constraint is satisfied
E[U?((r_p)q+x*(p))p]=2* g(p)
fi, (p)
i Mean-variance utility function:
E[(" (Y)] = E[Y] - ± (IJ(/J(Y)
g(p)
x*(p)=iíl__.t(p) g(p)
+E[E[)'(p,q)pJ]_. fp(p) -E[y(p,q)p]
)
a[ Elfp(p)],
LfJ
1 1,,',
l),—,b,, 2. 21ki
18
Illustrations of Optimal Exotic Payoffs Under
Mean-Var Criterion
Bivarate lognormal distribution:
7-- — .- po -------------pO.3
6-- - --pO.S - - - -, - - - - -.
p=07
5-- ______------ -----
(logp,logq) - A'(4. 0.72, 5.69. 0.2 2 .p) utidcrP&Q
(E[p] =$7OIAIWh, 0(p)=$56/MJI71
E[q] = 300MWh. cr(q) = 60Mtii
r =$120/MWh (fiat retail rate)
LJ
'p"y (r-p)q o o(p)
Optimal exotic payoff
:01."
Note: For the mean-variance utility,
the optimal payoff is linear in p
when correlation is O,
( )n 1)crrrIr, 1 2. 2OtF
Replication of Exotic Payoffs
Strike E
X
Strike> E
(P) = x(F) 1+x,(F)(p—I-)+ x(K)(K - pdK + x(K)(p_K)dK
Bond
payoff
Forward
pa off
Payoff
Cal! option
pa off
Rut option
pa off
Payoff
Payoff
Forward price
Strike price
Strke price
Exact replication can be obtalned from
a long cash position of size x(F) j a long forward position of size x(F)
long positions of size x"(K) in puts struck at K, for a continuum of K
which is less than F (Le., out-of-money puts)
long positions of size x(K) in calls struck at K, for a continuum of K
which is larger than F (i.e. out-of-money calis)
7 7 0= . 1k ffi, , 17
38
19
y'
•
• •
•-
TI2
-j1111n1 4
8 10P2'{l.4L%
rs nl.,, ,n 1 ....Jç.tI *4nn
-. vakrlra.-,K
- - -; ---
1 1*; l II .b;
II2Ii..nl;; ;,r,;;',
/
c4,,rAk-rr-
4
I
..
t
lt2fl
l,<.(tí4;
*,rl, u';, ;r,nIu;.' i,iIi,;' tijiiiiiiuiziiiL ;ri,u;';'
'r*;• ;,iul *d E' 1) wui,z 1 j vuu,t. I,;;'u;i iit*J ii;;
iii 1; g
4
iuh
li,IllI
jn
r
1
Optimal Timing of Statíc Hedge
Assuming GBM dynamics for price and quantity when is the best time
to buy optimal hedge and what is the best hedging portfolio at that
time?
Price and quantity dynamics:
t iT ± •: 7i E i;.!B )
di1t = q,
lt (fB + ( ri: .
tul' +
Optimal hedging portfolio if entered at time T
inax
t.
— i'7-Ii-J- +
111 1'1
=trnia.x.E,[Uiir —-!-h ?T — J'E:.'r.IJ
ripT
For mean - variance utility maximizing expected utility of optimally
hedged profit with respect to time is equivalent to minimizing
11(r) - w'iJ
JPJ-))
+ 2ea'ir —pr:qr.(;'-() + lw'Ur —pT:IqT
2. 2*l17
lo
20
Example:
Price and quantity dynamics
=
LaB1 +
=
Th fcirwíird riri uiid 1, nid etin tal e fi 'r a ini inth ubv v vtr latu 1 is
sitmed lo i 2L1E1w.1\I\\ii and 1000 J\V1 Th f111u.whig nilile uiniariz
the base values et the parameters:
Fiaiiir. ir T r
p
q
\uluu.
1 1 40 1 20 1 111011
1112
1
(Tj
Ø
07
0.1
0.7
- Deurruhcr 2, 2011
1
u
j
-
?tçr rt
:ur1s
UI,
Wtu;,nuf n ut,uf
D..tc-I c',U rciu
21v.c,uu 201rcn Ç
5T1u1IlulrI Lli -viaji) í E J:ll 1rtit Vtuii, 1-Ii líd
((ro
-a
12 201r
Tiniu.
42
21
Distribution of Hedged Profits as Function of
Hedging Time
O
r=
2
Optimal
S.
(7t2
1
7
1
'Ç
1.
Ne ged Prófit for peiiod T=i TS
05
3
- L2rrr,I,, 2, 2U(í
13
Optimal Hedging Function and Portfolio
based on info at time 0*
[,j,f,7 /í :c1
--------------------------
1
T''I1JjI
'tT í!iJ
.1, Tu,
1111 'Ii-'
ii!''i
.nu
*In reahty hedging portfoho will be determined at hedging time based
on realized quantities and prices at that time
2.
-II
22
Extensíons: VaR constrained hedging
VaR is the value such that
Pr(x VaR) - a for sorne a (typically 5%)
The VaR Constrained optimal hedging problem: VaR
Max X( P ) E [ Y ( x)] . s.t E[x(p)] = O. VaR(Y(x)) > V
Proposition: If the VAR is a function of the mean and variance of a
distribution and is monotonicafly decreasing in the variance and
ncreasing in the mean of the distribution, then a VaR constrained
optimal portfolio is on the efficient mean-variance frontier.
We prove that for the bivariate (ln(p), In(q)) distribution the hedged
profit under a Mean-Variance criteria satisfies the abo y e proposition
and hence an optimal MV hedge is equivalent to a corresponding
optimalVaRconstrainedhedge.
('ren DçLtmter 1
2, 2107
Equilibrium Model for Pricing Weather Derivatives
W
w_
Rta.a
-
a
F it
'
rri.rr.
M*S
1
-,rrn
23
Conclusion
Risk management is an essential eement of competitive
electricity markets.
The study and development of financia¡ instruments can
facilitate structuring and pricing of contracts.
Better tools for pricing financia¡ instruments and development
of hedging strategy will increase the Iiquidity and efficiency of
risk markets and enable replication of contracts through
standardized and easiiy tradable instruments
Financia¡ instruments can facilitate market design objectives
such as mitigating risk exposure created by functional
unbundling, containing market power, promoting demand
response and ensuring generation adequacy.
ren - I)ecrn,I
it, 12. 2011T
.1
24

Documentos relacionados