University of Southern California Dissertations and Theses

Electric system restructuring and system reliability

(USC Thesis Other)

Electric system restructuring and system reliability

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content ELECTRIC SYSTEM RESTRUCTURING
AND SYSTEM RELIABILITY
Copyright 2001
by
Catherine Miller Horiuchi
A dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
in Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PUBLIC ADMINISTRATION
(PUBLIC ADMINISTRATION)
May 2001
Catherine Miller Horiuchi
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UMI Number: 3027725
Copyright 2001 by
Horiuchi, Catherine Miller
All rights reserved.
®
UMI
UMI Microform 3027725
Copyright 2001 by Bell & Howell Information and Learning Company.
All rights reserved. This microform edition is protected against
unauthorized copying under Title 17, United States Code.
Bell & Howell Information and Learning Company
300 North Zeeb Road
P.O. Box 1346
Ann Arbor, Ml 48106-1346
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
UNIVERSITY OF SOUTHERN CALIFORNIA
SCH O O L O F PO LIC Y, PLANNING, AND DEVELOPM ENT
UNIVERSITY PA R K
LOS ANGELES, C ALIFO R NIA 90089
This dissertation, written by
CATHERINE MILLER HORIUCHI
under the direction of h.&z... Dissertation
Committee, and approved by all its
members, has been presented to and
accepted by the Faculty of the School of
Policy, Planning, and Development, in
partial fulfillment of requirements for the
degree of
DOCTOR OF PUBLIC ADMINISTRATION
Dean
DISSERTATION COMMITTEE ~
nrperson
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Dedication
In memory of Sarah Louise Marks Miller, for her undying love.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
iii
Acknowledgements
This dissertation required the cooperation of many in the electric industry.
Numerous people shared their expertise, made timely responses to inquiries, and
located publicly accessible information. I also would like to acknowledge the
policy makers who participated in the survey; some had to reschedule more than
once due to the pressing nature of the market problems.
The quantitative assessment included non-linear regression models of
panel data. I received assistance in understanding texts and study material and in
interpreting statistical output from Suzanne O’Keefe, Henry and Corinne Miller,
and Lorraine McCall.
I am grateful to my dissertation committee for their insights and helpful
comments. My deepest thanks go to my dissertation advisor, Elizabeth Graddy,
who has provided steady direction throughout the process. I thank Charles
Cicchetti for his interest and support, and for contributing his experience and
knowledge of electric system regulation. I thank Ross Clayton for suggesting this
topic several years ago, as it has been endlessly fascinating.
No acknowledgement would be complete without emphasizing the
fundamental importance of the enduring support of my devoted husband, Wayne.
Lastly, I thank our two children, Walt and Angela, for their youthful confidence.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
iv
Contents
Dedication ii
Acknowledgements iii
Tables vii
Figures viii
Abstract ix
Chapter 1: Introduction 1
Introduction 1
Research Question 7
Outline of Dissertation 12
Chapter 2: Electric System Restructuring in the U.S. and California 16
Introduction to the U.S. Electric Industry 16
The Components of an Electric System 17
Generation 18
Transmission 19
Distribution 20
The Electric Product 21
Reducing Regulatory Controls on Electric System Components 22
Regulatory Agency Acronyms 24
Partial Electric System Restructuring 26
Origins 28
Munis and IOUs 29
California’s Electric System and its Major Service Companies 31
Challenges in the Restructuring Process 35
Technological Challenges 36
Political Challenges 38
Regulatory Challenges 40
Legislation Altering the Comprehensively Regulated System 41
National Legislation 42
State Legislation 43
Trends in Electric Transport: Retail and Wholesale Wheeling 47
Operational Context of Regulatory Activities 50
The Role of Energy Efficiency and Demand Side Management 54
Chapter 3: Literature Review 58
Introduction 58
Market Rationale for Restructuring the Electric System 59
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Successful Antecedents 60
Anticipated Gains 64
New Entrants 66
The Feasibility of Electric Industry Disaggregation 69
Natural Monopoly 69
Component Systems 70
Reliability Constraints 74
Complexity Theory and Control of Disaggregated Components 76
Protecting Stakeholders and Stockholders 80
Stranded Costs Recovery 80
Risk o f Cooptation 83
Managing the Regulatory Remainder 84
Implementation and Evaluation 85
Transition Effects 87
Externalities 89
Public Administration Perspectives 92
Public Good 93
Market Power 95
Use of Analogies 98
“Big Questions” Issues 104
Chapter 4: Research Design 106
Introduction 106
Goal of the Research and Resultant Design 107
Incomplete Knowledge and Policy Making 110
Description of Approach 113
Restructuring 116
Selection of Subjects and Instrumentation 117
Testable Hypotheses Derived from the Institutional Arrangements 119
Field and Analytical Procedures 120
Data Collection and Recording 123
Data Sources 124
Material Directly Supplied by Utilities 124
Weather Information 125
CPUC Required Outage Reports 128
NERC DAWG Database 129
Legislative Hearing Documents 129
ISO Hearing Documents 130
Stakeholder Interviews and Surveys 130
Analytic Design 130
Methodological Assumptions 133
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
vi
Chapter 5: Findings 135
Introduction 135
The Very Public Problem 136
Quantitative Analysis 138
Logit Regression Results 139
Discussion of the Error Term, Tests and Transformations 161
Negative Binomial Regression Results 164
Outage Information 171
Outages and Emergencies 171
NERC Disturbance Data 174
Findings Regarding Institutional Changes and Comparisons 179
New Institutions and Changes in the Traditional Utility Companies 179
Owned and Contracted Power 186
Energy Efficiency and Load Management 195
Direct Access 199
Survey Results 201
Relation of Findings to Hypotheses 211
Chapter 6: Conclusions and Further Developments 235
Introduction 235
Conclusions 236
Limitations of the Analysis 241
Uncertainty 241
Counterfactuals 242
Alterations to the Restructuring Model 243
A U.S. Framework for Electric System Restructuring 245
Reviewing the California Situation 248
Fit of California’s Restructuring to Market Models 248
The 2001 California Energy Crisis 251
Ideas for Further Study 255
Unresolved Issues as Closing Observations 258
Bibliography 263
Appendix A: Interview Protocol 272
Appendix B: Contacts 274
Appendix C: Data Sources 277
Appendix D: Glossary of Terms 280
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
vii
Tables
Table 1: Institutional Changes Due to Partial Electric System Restructuring 14
Table 2: Disaggregated Components of the Vertically Integrated Utility.............22
Table 3: Energy Regulatory Bodies..........................................................................25
Table 4: Majone’s Model for Process Evaluation..................................................I l l
Table 5: Data Applicable to the Research Hypotheses......................................... 122
Table Set 6; NERC Summer and Winter Margin Estimates.................................137
Table 7: Logit Regression Independent Variables.................................................140
Table Set 8: SMUD Data, 1994-1999, Logit Regression ......................... 142
Table Set 9: LADWP Data, 1990-1999, Logit Regression................................... 145
Table Set 10: PG&E Data, January 1993- October 2000, Logit Regression.....149
Table Set 11: SCE Data, January 1993-September 2000, Logit Regression..... 151
Table Set 12: SDG&E Data, 1991-1999, Logit Regression............................... 154
Table Set 13: Multiple Utility Data, 1990-1999, Logit Regression................... 157
Table 14: PG&E Negative Binomial Regression, SAIDI Annual Data as
Dependent Variable................................................................................ 166
Table 15: SCE Negative Binomial Regression, SAIDI Monthly Data as
Dependent Variable................................................................................ 167
Table 16: SDG&E Negative Binomial Regression, SAIDI Monthly Data
Dependent Variable............................ 168
Table 17: SMUD Negative Binomial Regression, SAIDI Annual Data as
Dependent V ariable ......................................................................169
Table 18: LADWP Negative Binomial Regression, SAIDI Annual Data
Dependent Variable................................................................................170
Table 19: California ISO Emergencies Annual Totals 1998-2000..................... 182
Table 20: Generation Capacity Shifts Since 1993, in Megawatts....................... 187
Table 21: SDG&E’s Cost of Purchased Power, 1991-1999.................... 194
Table 22: Cumulative Direct Access Load by Customer Class, July 1998 vs.
August 2000.............................................. 201
Table 23: Significant Independent Variables, Logit Regression......................... 237
Table 24: Significant Independent Variables, Negative Binomial Regression... 239
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
viii
Figures
Figure 1: Multiple Utility Data, 1990-1999, SAIDI Comparison.........................172
Figure 2: California ISO Emergencies by Month, 1998-2000......... 183
Figure 3: Individual Utilities’ Change in Generation Capacity Over Time.........187
Figure 4: Summation ofUtilities’ Contribution to Generation Capacity............188
Figure 5: Growth in Peak Demand, by Utility....................................................... 188
Figure 6: SMUD’s Use of DSM Dispatchable Load Shedding, 1994-2000..... 197
Figure 7: Comparison of Discrete Incidents, SMUD Load Shedding, 1994-2000
...... 198
Figure 8: Direct Access Customer Shifts, by Investor-owned Utility.................200
Figure 9: A Cyclical Framework to Fit Deregulatory Reform Intentions...........248
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
Abstract
In 1996 the California legislature passed AB 1890, explicitly defining
economic benefits and detailing specific mechanisms for initiating a partial
restructuring the state’s electric system. Critics have since sought re-regulation
and proponents have asked for patience as the new institutions and markets take
shape. Other states’ electric system restructuring activities have been tempered
by real and perceived problems in the California model. This study examines the
reduced regulatory controls and new constraints introduced in California’s limited
restructuring model using utility and regulatory agency records from the 1990s to
investigate effects of new institutions and practices on system reliability for the
state’s five largest public and private utilities.
Logit and negative binomial regressions indicate some negative impact
from the California model of restructuring on system reliability as measured by
customer interruptions. Time series analysis of outage data could not predict the
wholesale power market collapse and the subsequent rolling blackouts in early
2001; inclusion of near-outage reliability disturbances—load shedding and energy
emergencies—provided a measure of forewarning. Analysis of system
disruptions, generation capacity and demand, and the role of purchased power
challenge conventional wisdom on the causality of California’s power problems.
The quantitative analysis was supplemented by a targeted survey of electric
system restructuring participants.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
X
Findings suggest each utility and the organization controlling the state’s
electric grid provided protection from power outages comparable to pre
restructuring operations through 2000; however, this reliability has come at an
inflated cost, resulting in reduced system purchases and decreased marginal
protection. The historic margin of operating safety has fully eroded, increasing
mandatory load shedding and emergency declarations for voluntary and
mandatory conservation. Proposed remedies focused on state-funded contracts
and government-managed power authorities may not help, as the findings suggest
pricing models, market uncertainty, interjurisdictional conflict and an inability to
respond to market perturbations are more significant contributors to reduced
regional generation availability than the particular contract mechanisms and
funding sources used for power purchases.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
1
Chapter 1
Introduction
Introduction
In the late 1980s, following the lead of other regulated industries toward
broader use of market forces, electric system regulators and the federal
government issued a set of orders opening more of the century-old monopoly to
market participation. The 1990s saw numerous follow-on actions at the Federal
level and in most states. As the twenty-first century unfolds, and the results of
these actions are dissimilar between states, the feasibility of beneficially
introducing market forces into the electric industry has been questioned. Many
additional changes are proposed, some continuing forward in the market model,
others restricting participation, operations, and limiting competition. In
California’s case, these proposals include taking giant leaps backward from
market designs.
This dissertation deals with a small part of the regulatory environment: the
question of changes in end-to-end reliability since California passed legislation in
1996 partially restructuring the electric system. These changes are investigated in
the context of electric system reliability in the 1990s, focusing on reliability
disturbances pre- and post-restructuring in California.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
2
The North American Electric Reliability Council (NERC) defines electric
system reliability in terms of the adequacy and security of the system: adequacy
to supply electric demand at all times despite planned and unplanned outages of
system facilities, and security to withstand sudden disturbances. (North American
Electric Reliability Council 1998)
This study looks further into the adequacy of reliability, beyond the failure
of supply evidenced in outages to the threat of failure shown in reliability-related
events such as customer load shedding and energy margin emergencies. The
study period is selected in the wake of the Federal Energy Regulatory
Commission (FERC) Orders 888 and 889, the Energy Policy Act of 1992, and
California Assembly Bill 1890 (hereafter referred to as AB 1890).
To assess reliability, a quantitative analysis was performed of reliability
events in the first years of the partial electric system restructuring. The analysis
reviewed the partial restructuring’s effects on the ability ofthe combined efforts
of altered generation, transmission, and distribution institutions to reliably deliver
sufficient generation to the users of electricity. A corollary of this reliability
question is the capacity of the electric industry under this partial restructuring to
deliver the promise of deregulation-increased competition and reduced cost
through efficiencies and market forces. While rate reductions and competitive
pricing strategies are not the objects of quantitative analysis in this dissertation,
institutional structures and actions designed to meet these goals have altered
Reproduced with permission o fth e copyright owner. Further reproduction prohibited without permission.
3
industry practices and regulatory decisions. Where these changes are postulated
to have contributed to reliability problems, they are discussed.
Expression of concern is widespread. On the front page of the Wall Street
Journal, May 11, 2000, the lead story states "Doom and Gloom: New Rules,
Demands Put Dangerous Strain On Electricity Supply" with the subhead, "Partial
Deregulation Breeds Confusion in Industry; Summer Shortage Feared: Oracle
Builds Its Own Bunker". It describes conditions of fear, uncertainty and doubt
regarding sufficient and deliverable generation to meet anticipated needs. Large
companies have altered their risk management profiles anticipating electric
system problems; some are disengaging entirely from their historic providers,
setting up their own generators and selling any excess to the market. The federal
government completed its own Power Outage Study Team (POST) report, and
found numerous historical designs and current practices interacted to cause the
East Coast outages of 1998 and 1999.
This dissertation uses the example of partial deregulation in California as
its source for data to explore how well actual events fit the theoretical
expectations of restructuring. California is a significant bellwether, and has also
been an early adopter of restructured operations. California is the sole U.S. state
to experience what many view a comprehensive failure of its restructured
environment, challenging market model assumptions of robustness and
responsiveness in a market-based environment.
Reproduced with permission o fth e copyright owner. Further reproduction prohibited without permission.
4
In addition to quantitative analysis to determine contributions from
possible causal conditions to reliability disturbances, this dissertation includes a
qualitative assessment from the managers and public officials responsible for
designing, legislating, and implementing the California model of partial electric
restructuring. These participants are surveyed using an instrument derived from
the federal POST questions and findings, with emphasis on California activities.
For the past century, electric system management has been an area of
technical arcana. Advocates for reduced regulatory controls and open markets
bring discussion of the nature and management of electricity and regulation into
wider political and public discourse. Participation in this discussion requires
learning a new set of acronyms; cognitive dissonance may explain some of the
difficulties in the implementation of restructuring and the problematic nature of
certain political solutions.
Electric service in the U.S. originally developed in many separate places
and followed separate designs. Each island of electricity was managed
independently. Before electricity was deemed a natural monopoly, multiple
electric companies vied for customers in the same area, each setting up poles and
stringing wires. Over time, this practice declined with establishment of exclusive
electric franchises. As electric grids grew, and potential for transit agreements
developed between neighboring non-competing electric companies, interconnects
were developed bridging each company's territory.
Reproduced with permission o fth e copyright owner. Further reproduction prohibited without permission.
5
Electricity has been, and continues to be, regulated on multiple levels-
Federal, state, and local. It is a commodity provided partially by municipal
utilities but mainly by highly regulated investor-owned utilities. Restructuring
activities in the 1990s explored the discrepancies in electricity pricing between
states, and started efforts to apply the values of deregulation found in other
industries to one of the few remaining monopolies. The United States trails many
countries in its electric restructuring effort. Utilities and regulators have used this
opportunity to study energy system restructuring in South America, England,
Europe and elsewhere, to understand how the restructured organizations could
best be constructed in particular jurisdictions. The U.S. situation is complicated
by its constitutional design. In large part the 50 states regulate power
individually, and only a subset of states has experienced the high power costs,
externalities, and other technical and political factors resulting in drives for
restructured operations.
Much attention in the late 1990s was devoted to the establishment of new
models for transmission. This developed from evidence and a belief that the
transmission system was not a suitable candidate for deregulation. Additionally,
the separation of generation resources from transmission operations in the
formerly vertical monopolies required development of monitoring points at each
generator, establishing for an open operator the quantity of power entering the
grid. The nature and mechanics of these edge systems were developed through a
Reproduced with permission o fth e copyright owner. Further reproduction prohibited without permission.
6
collaborative committee process, designed in response to FERC Orders 888 and
889. These and other activities associated with development of the restructured
electric operations are discussed in detail in Chapter 2.
FERC Order 2000 requires transmission system designs. Two basic
operating models for transmission systems have been developed: the command
and control model found in California, and the market model proposed in other
areas. This dissertation peripherally addresses this difference wherever the design
choice for California has impacted the availability of sufficient generation
capacity.
Implicit in the general attention on designing new transmission systems is
the assumption of adequate generation. The early problem of sufficient
generation capacity was resolved by each utility; it was not anticipated that this
would need to be extensively revised. Generation is the most fully competitive
resource, and independent power producers have been part of the overall electric
system operating model for decades due to PURPA legislation following the
1970s energy crisis. The assumption of adequate generation has been sorely
tested since 1996. This close study of the period preceding the 1996 enactment of
AB 1890 through 2000 focuses on issues of generation management and delivery.
It evaluates whether changes in the comprehensively regulated energy
management and pricing systems resulted in insufficient generation resources (or
Reproduced with permission o fth e copyright owner. Further reproduction prohibited without permission.
7
local generation substitutes) to meet demands. It suggests causes for the perceived
insufficient and unreliable supply of electricity to utility customers.
Research Question
The study addresses the following research question: How have
regulatory decisions and restructuring legislation in California affected electric
system reliability, particularly the availability o f sufficient generation to meet
localized demand?
The California regulatory decisions and restructuring legislation discussed
in this research refer to the limited and specific actions taken in the development
and response to AB 1890.
Aspects of restructuring superficially peripheral to generation are
examined for their impact on generation management. That is, utility managers
charged with system reliability can take certain actions to substitute one
generation source for another, use transmission to mitigate generation problems,
or alter customer behavior to limit demand for generation. Problems are
suggested through the frequency and severity of reliability disruptions, which
interrupt normal electric supply and demand. Due to the interlocking nature of
the electric system components, the reliability disruptions are measured in terms
of their impact on end users, e.g., a known interruption of expected reliable
power.
Reproduced with permission o fth e copyright owner. Further reproduction prohibited without permission.
8
The hypotheses related to the research question are provided below.
HI: The rate freeze imposed upon investor-owned utilities has reduced system
reliability.
This hypothesis is examined by comparing any changes in general system
reliability involving distribution, transmission, and generation disturbances
with specific reliability data from the distribution companies subject to the
rate freeze. It also contrasts reliability disturbances experienced by investor-
owned utilities with incidents experienced by municipal utilities; municipals
were not directly required to freeze or reduce rates, but are subject to political
expectations based on rate comparisons with the IOUs.
H2: Industry investment has shifted in response to restructuring, resulting in less
investment in reliability-related systems.
Each utility investigated has a capital investment segment of their budget.
While much of the capital is used directly to support the core functions of
generation, transmission, and distribution, a portion is dedicated to non-core
investments. These range from energy marketing to call center software to
research on renewable technologies. This hypothesis compares non-core
investment in each utility to total investments, and to investments made by
new participants seeking open markets in the non-core areas. Discussion
includes changes to investment in generation, following the separation of
generation from the emerging distribution companies.
Reproduced with permission o fth e copyright owner. Further reproduction prohibited without permission.
9
H3: Increased power purchasing has reduced system reliability.
This hypothesis compares over time the percentage of total power supply that
is purchased to each utility’s reliability profile. Each utility maintains outage
records that constitute a profile of the degree of expected reliability provided
its customers. Historically, each utility managed its overall portfolio owned
generation, long term contracts, and limited supplemental (spot) purchases to
be available at short notice to meet energy demands. The separation of
generation, and the introduction and use of new generation holding
companies is part of this hypothesis.
H4: New intermediary structures (the ISO and power exchange) have reduced
system reliability.
This hypothesis contrasts the number and severity of reliability disturbances
before and after the establishment of the California ISO. This reflects the
shift from transmission control by individual utilities to a separate, statewide
agency tasked with maintaining system reliability and facilitating open access
to the electric grid.
H5: The introduction o f direct access customers has decreased system reliability.
This hypothesis compares the number and load of direct access customers to
the reliability profile of each utility. Sources of generation and transport
management for these customers are investigated.
Reproduced with permission o fth e copyright owner. Further reproduction prohibited without permission.
10
H6: The reduction o f DSMpractices does not affect system reliability.
This hypothesis compares over time the amount of load subject to demand
side management (DSM) to the peak load associated with reliability
disturbances. Changes in DSM's role as a substitute for generation is
investigated.
H7: The separation o f generation from distribution does not affect system
reliability.
This hypothesis compares the reliability profile of each utility against its self-
generation capacity over time. Additionally, it looks for differences between
the IOUs who have divested a proportion of their generation and whose
transmission the ISO controls from the municipal utilities who have
maintained a more vertically-integrated business model.
The above hypotheses focus on measurable electric system operations.
Other aspects of electric system restructuring are less easily captured.
Externalities such as environmental issues of water usage in the western
U.S. or acid rain in the East may be irrelevant to the establishment of a
competitive market environment but remain important for the public interest. For
instance, in 1999 the California Public Utilities Commission (CPUC) assigned
over 100 staff members the task of preparing a draft environmental impact review
(.EIR) relevant to a hydroelectric generation divestiture application made by
Pacific Gas and Electric (PG&E). This divestiture would move much of the
Reproduced with permission o fth e copyright owner. Further reproduction prohibited without permission.
n
Sierra Nevada watershed from current regulatory oversight. FERC operational
guidelines do not include the many formal and informal agreements in place
between PG&E and watershed users; the property involved includes both FERC
and non-FERC acreage as well. The Draft EIR, issued in November 2000, runs
over 1000 pages.
These externalities command political attention and must be considered as
relevant factors in policy decisions which may affect the ability of restructured
assets to be operated effectively in the emerging energy environment. Generation
disruptions occur because of these externalities, as occurred in Southern
California in December 2000; thousands of megawatts were taken offline when
power plants that had operated at higher-than-average load throughout the
summer exceeded their annual permitted emissions.
The electric system faces certain physical and technical constraints. These
constraints include line loss, the resultant role of some measure of nearby
generation, the non-storability of electricity, and an assured maintenance of
specific current. Line loss describes the reduction of current as it passes through a
wire over a distance; this line loss can be affected by weather or air pollution
(such as smoke from brush fires) which can draw additional current from the line.
Electricity cannot be stored like water behind a dam; it must be generated minute-
by-minute, to meet the draw from individual sites. Appliances and equipment
expect a particular voltage to be delivered, so the appropriate voltage must be
Reproduced with permission o fth e copyright owner. Further reproduction prohibited without permission.
12
maintained on the line at all times. Altering equipment, by adding line
conditioners and uninterruptible power, is an expensive if seemingly increasingly
necessary alternative to a steady voltage. The transmission system must also
adhere to institutional practices required for high reliability systems. These
practices have been revised by the state’s restructuring model.
The study of electric system regulatory controls and the application of
partial restructuring is a life’s work, even if the effort is restricted to activities in
the United States or in California. In order to focus clearly on one restructuring
effect, many sub-areas must be glossed over or set aside in this research. The
research scope is limited to issues associated with adequate generation and
reliable delivery in the pre- and post-AB 1890 timeframe, through an examination
of reliability events.
Outline of Dissertation
This chapter introduces the research question and introduces the topic
terminology to assist in interdisciplinary discourse. It also contains the
dissertation outline.
Chapter 2 discusses the electric industry and the changes it has undergone.
This chapter provides a detailed history of early decisions in energy deregulation.
It includes the denotations and definitions used by the utilities industry to describe
and explain its activities. It also describes technological, political, and regulatory
Reproduced with permission o fth e copyright owner. Further reproduction prohibited without permission.
13
challenges in the restructuring process relevant to a discussion of changes to
generation requirements and resources.
The literature review in Chapter 3 investigates the academic and
professional literature on electric system restructuring, its relationship to other
restructuring efforts, and the international move toward reduced regulation. This
literature describes special concessions granted for electric system restructuring
due to its technological challenges (nuclear power, for instance) or its lengthy
history as a non-competitive, highly regulated industry (as in the historic practice
of "fair rate of return", and assessments to capture revenue for stranded costs).
This review provides the theoretical context for the dissertation while the research
and analysis itself is very much focused on the actions and consequences of
decisions to move toward market mechanisms and away from regulation. The
review will further ground this research in the context of other work in public
administration, including literature on implementation and evaluation. It will also
develop the research question’s relevance in both a professional and theoretical
context.
Chapter 4 describes in detail the research design. The categories of
reliability disruptions are established as the dependent variables for analysis. The
chapter begins with discussion of the institutional responses to the California
restructuring model, as noted in Table 1.
Reproduced with permission o fth e copyright owner. Further reproduction prohibited without permission.
Power purchasing
ancillary markets
14
Mixture of contract and build
options for standby power
Sell-off of generation; assumption of power
availability
Own generation; end-to-end control
of power availability
Reduction o f demand side management
{DSM) practices and energy efficiency (EE)
promotion
Use of DSM techniques to equalize
load demand and maximize
generation assets
Independent System Operator (ISO) or power
pool manages transmission system; many
contributors to grid via bid process at the
Power Exchange (PX).
End-to-end control of power
availability-coordinated, non
competitive contributions to grid
energy
Large capital investments in non-core
operations such as customer information
systems (CIS), automated meter reading
(AMR), and unregulated affiliates
Large capital investments in core
operations (generation,
transmission, distribution)
Table 1: Institutional Changes Due to Partial Electric System Restructuring
Chapter 4 also describes the research methodology adopted, the analytical
approach, and data sets in detail. The hypotheses investigated are stated, and the
data necessary to answer them are explored.
In Chapter 5, findings developed from the analysis of data are reported.
Data supporting or refuting effects of institutional changes are presented;
shortcomings of the data are also discussed here. The qualitative material from
interviews and direct observation of restructuring participants is also presented.
The implications of the findings for the hypotheses is discussed.
In Chapter 6, conclusions drawn from analysis of the findings are
presented and summarized. Limitations of the research and applications of the
research are detailed. The fit of California’s restructuring to existing theory on
reduced regulatory controls is described. The energy crisis that developed in late
fall of 2000 and the activities of early 2001 are reviewed, especially in light of the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
findings and data that were available at the time the crisis was developing.
Promising areas for further development are specifically outlined, both where
additional investigation of data at a more precise unit of analysis is indicated, and
where interesting corollary questions would require a different data set. Finally,
many interesting questions derived from examination of the evidence included
herein are reviewed in this final chapter. Three unresolved issues that have
received limited discussion in California’s public policy discourse close the
chapter.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
16
Chapter 2
Electric System Restructuring in the U.S. and California
Introduction to the U.S. Electric Industry
To understand the nature of electric system restructuring, the industry
needs to be viewed in context with a sense of its history. The U.S. case differs
from foreign examples and other U.S. industries deregulated in the latter half of
the twentieth century in its development of the technology and in the regulation of
electric service. The economist Douglass North develops the theory of path
dependence to explain the differing effects of similar decisions and actions on the
Spanish and English empires (North 1990,112-17). Similarly, the path of electric
service has relevance in the method and degree of success in electric system
restructuring.
This chapter reviews the electric industry in general and the specific
history of electricity in California. It describes the initial actions and decisions
that defined the reduced regulatory controls. It details new constraints that
comprise California’s limited restructuring of the electric system. In this process,
the chapter describes many relevant terms and players that populate the energy
landscape.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
17
The Components of an Electric System
The traditional vertically integrated utility handled all component
activities, from the generation of electricity to its transmission along high voltage
wires through distribution substations, transformers, meters, and into its
customers’ facilities. Until the advent of restructuring, the utility has been
guaranteed an opportunity to earn a fair rate o f return, the cost of prudent service
plus a reasonable percentage for profit.1 In exchange, the government regulated
the monopoly operation to protect the public interest.
The first phase of restructuring involved analysis of the regulated
monopoly power companies, with policy makers seeking natural division points
and considering likely opportunities for competitive economic forces. The effort
to create a competitive market for electricity also considered pricing models other
than the historic fair rate of return or cost-plus model. For example, proponents
of marginal cost models argued marginal cost design would foster innovation by
increasing or altering electric system participants and overall load.
Legislators and public officials with general oversight are challenged by
the complexity of restructuring electric service. Public sector leaders must
provide regulation, oversight and legislation to multiple areas, each defined and
limited by specific, detailed technical knowledge; electric system management is
no different. Terminology and the defined components of electric system
1 California Energy Commission, “Electricity Report 1996,” November 1997, 5.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
18
operation and distribution services comprise a language of electric system
restructuring that must be shared among regulatory parties and stakeholders to the
process. This knowledge is gained slowly.
The three structural components that commonly describe the vertically
integrated utility are individually discussed below. Electricity is an essential
component of everyday life, and this essential nature is inherent in the highly
reliable system structure. Electric system restructuring seeks to alter this
traditional structure through regulatory change with no negative effect on the
reliable creation and delivery of electricity to customers.
Generation
The generation component creates electricity. Due to the non-storable
nature of electricity, adequate generation capacity must be available twenty four
hours a day. Electricity is most frequently generated by steam turning a turbine;
the steam is created from burning fossil fuels, from nuclear fission or by falling
water. Advanced generating technologies include biomass, wind turbines and
solar cells; these advanced technologies are more expensive per unit of energy
produced (and comprise less than 5% of overall generation in California).
Flexibility in fuel source is receiving renewed attention in 2000 in energy
policy discourse, as buyers of generation seek substitution strategies in response
to high natural gas prices. Additional coal, oil or other generation can be located
in less densely populous western states such as Nevada and Utah without
exceeding Federal emissions standards, but all new generation plants in California
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
19
must seek pollution credits that exceed its projected emissions as part of overall
efforts to comply with EPA emission standards.
Transmission
The transmission component moves electricity along high-voltage lines
from the points of generation through the electric grid into a distribution system.
This part of the business, the wholesale trade of electricity among utilities, has
been regulated by FERC since its transformation from the Federal Power
Commission with the creation of the Department of Energy in 1973. A unified
electric network of five grids operates in the U.S., most of Canada, and part of
Mexico. Three of these power grids operate in the United States, the Western
Interconnected System, the Texas Interconnected System, and the Eastern
Connected System. California is part of the Western Interconnect. As noted
elsewhere, electricity does not move along an otherwise empty wire. Generation
supplies current, which is drawn by the consuming equipment at the far end. A
complex system manages this process, which stabilizes frequency, maintains
power quality, prevents overload, and efficiently integrates generation resources.3
U.S. and European systems of transmission differ, as do U.S. and
European distribution systems. Investments in Europe toward high-speed data
links using the transmission system, combined with technical differences in the
design of the systems, have led to different capacity between these industrialized
2 Department of Energy, Energy Information Administration. The Changing Structure of the
Electric Power Industry: Selected Issues, 1998, DOE/EIA-0620, July 1998,7-10.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
20
sectors for new lines of business. The government-managed nature of the electric
service is another difference between U.S. and European systems.
The U.S. restructuring model alters the operations of existing power
producing corporations, and attempts to structure opportunities for new entrants in
meeting specific cost, availability, and reliability targets for utility customers.
Transmission entities remain regulated monopolies, but the operating rules and
oversight bodies are non-identical across the U.S..
Distribution
The distribution component of electric service accepts high voltage
electricity from transmission systems, transforms it to lower voltages suitable for
commercial or residential customers, and delivers it to these customers over
company owned overhead or underground lines. Distribution companies also
own the meters attached to buildings, servicing and reading these meters. Lastly,
distribution companies manage customer service and billing.
In conjunction with traditional services, distribution companies have been
responsible for rule-setting and requirements necessary for customers to arrange
for electricity to be supplied by new entrants, power marketers who resell power
purchased from third parties or qualifying facilities (QFs), independent power
producers. These requirements vary by state; in California, they have been
substantial. As a result, few Californians are participating in direct access
programs which allow them to bypass their traditional energy provider.
3 Ibid., 139.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
21
Other work handled by distribution companies includes specialized
customer service and energy consulting. These energy services are also provided
by firms other than vertically integrated utilities, energy services providers
(ESPs). California’s IOUs continue to offer some energy services for existing
customers within the utilities.
The Electric Product
Electricity is entirely homogeneous; electrons emerging from a wire do
not differ based on the energy production source. High sulfur coal produces more
residual emissions than hydroelectric power, but turbines create identical current
regardless of fuel source. Each source has its side effects, even so-called green
energy. The electric product that is purchased-the kilowatt or megawatt hour-is
not homogeneous, as evidenced by rate classes, rebates, and incentives to
participate in load shedding programs.
Producing electricity historically has meant large capital investment and
proximity to fuel sources coupled with a right of way for wires transporting the
product to the site of consumption. In energy wheeling a buyer contracts from a
faraway producer, but that particular producer’s electrons do not travel through
the electric wires to the customer’s location. Rather, the producer’s electricity
enters a common carrier transportation network, from which the buyer pulls off
their contracted amount. Originally wheeling occurred only between utilities,
then between utilities and independent power producers. With open access
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
22
offered through restructuring, first commercial and then retail customers were
provided the opportunity to obtain a more competitive generation price.
Much energy management revolves around keeping the current in the wire
at a constant voltage. Producers and users independently, and with limited
coordination, put electrons on and take electrons off through creating and drawing
electric current. Sophisticated technology monitors these activities, correlating
the physical behavior of the electric product with the financial contracts that
estimate and price the intended activity. Some of this management is done by the
ISO, some by the distribution companies.
Reducing Regulatory Controls on Electric System Components
Table 2 describes the three traditional components and the theorized areas
of regulatory reduction and market introduction.
Component Role Initial Restructuring Intention
in California and the U.S.
Generation Creates electricity Open to competition
Transmission High voltage transport Remains monopolistic and
highly controlled
Distribution Low voltage transport Transport remains monopolistic;
other aspects of distribution
company operations remain
monopolistic over short term
Table 2: Disaggregated Components of the Vertically Integrated Utility
Parts of the electric system appear less suited to the introduction of
competition and markets. For instance, deregulating transmission would
necessitate large capital expenditures by new entrants and the creation of
additional physical infrastructures (towers, trenches, wires, rights of way) for
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
23
competitors. As a result, the term restructuring is preferred over the more
comprehensive term deregulation. Existing regulatory structure remains, and
new oversight bodies, such as those overseeing the transmission organizations, are
being established.
The new transmission management bodies, independent system operators
(ISOs) or regional transmission operators (RTOs), are designed on different
assumptions and models. These new entities operate the transmission grids to
manage and assure a high-reliability transport system. The RTO or ISO manages
the overall schedule of power deliveries and movement along the transmission
grid in its control area. California established its ISO as a special instance RTO; it
follows a command-and-control based model, while east coast RTOs such as the
NYMEX are market based exchanges. (The market strategies pursued by sellers
and buyers in the second half of 2000 greatly increased the California ISO’s
market role, but did not alter its sense of mission and style.)
Focusing on the three components in delivering electricity excludes an
area of additional services of much interest to restructuring proponents. Energy
services has been an umbrella term describing a wide range of energy
management activities outside the traditional regulatory constraints. In the 1980s,
energy services departments within utilities performed energy audits and energy
conservation projects. Independent energy services consultants provided similar
analyses for both utilities and their largest customers. New energy service
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
24
providers (ESPs) have entered restructured electric markets offering generation.
Traditional distribution company processes can reorganize into regulated and
unregulated units; establishing an unregulated subsidiary of the parent utility to
provide energy services is a common strategy. This provides utilities with
flexibility in responding to market conditions in developing and marketing value-
added services in the unregulated areas, thus gaining or preserving market share
in these new economic spaces. Energy services may also include analyses of
operations to reduce consumption, specialty services such as coordinating with
multiple suppliers for bill aggregation and rate negotiation, and premium power
services providing quality guarantees of uninterruptible service or more constant
voltage.
Regulatory Agency Acronyms
Many terms and acronyms are found in discussions of electric industry
restructuring and will be frequently referenced herein. The industry and its
regulatory history are complex; mastery of terminology is a sine qua non for
following disputes on the details of the restructured environment. The commonly
used acronyms for the regulatory bodies associated with the electric system
operations in California are listed and described briefly in Table 3, below.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
25
AGENCY ROLE
CEC The California Energy Commission oversees municipal power in California. It
also evaluates generation siting in coordination with the California
Environmental Protection Agency (Cal EPA).
DOE The Department of Energy originally derived from the Atomic Energy
Commission tasked after World War II with the peacetime use of atomic energy.
The DOE regularly finds itself a target for abolition by members of Congress.
Brown (1997) states the energy savings from three DOE initiatives -- heat-
reflecting windows, electronic ballasts for fluorescent lighting, and variable
speed refrigeration systems - more than defray its research and development
budget.
FERC The Federal Energy Regulatory Commission issued FERC Order 888 and 889 in
1996 establishing guidelines for deregulating transmission. FERC’s role is the
oversight of the wholesale sale of energy, and its interstate movement through
the national transmission system.
NERC The North American Energy Reliability Council was formed in 1968 in response
to the 1965 blackout in the Northeast. Ten regional councils assure reliability of
the electric grid, the transmission infrastructure crucial to continued reliability.
Technical changes to the grid are required as reduced regulatory controls and
open access alter power flows; transaction accounts must be balanced between
participants and non-participants of different power pools.
NRG The Nuclear Regulatory Commission is the Federal agency responsible for
oversight of the nuclear power industry’ s licensing, construction, and
maintenance. High capital requirements for building, operating, and
decommissioning nuclear plants have been a significant driver for stranded costs
recovery.
PUC/PSC Public Utility Commissions-called Public Service Commissions in some states-
hold responsibility within states to regulate electric companies. Traditionally
their role has been the review and approval of rate cases for privately-owned
utilities (IOUs). The PUCs have been actively involved in determining the
degree and form of electric restructuring that will take place in their jurisdiction.
Their decisions are state-specific with minor overall coordination. Not all PUCs
discern benefit for their jurisdiction from the introduction of electric system
markets. Private interests desiring uniform structures across the nation to
facilitate restructuring are seeking federal judicial opinion to override or abolish
state level controls. California has a PUC, often abbreviated CPUC.
wscc The Western Systems Coordinating Council is the geographic sub-unit of NERC
that includes California.
Table 3: Energy Regulatory Bodies
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
26
Partial Electric System Restructuring
Reliable and affordable electric power has until recently been purchased
by virtually all U.S. households and businesses from monopoly providers. These
providers held end-to-end responsibility for electricity, which meant they
produced electrical current, transmitted at high-voltage, transformed the current
into appropriate voltage for commercial or residential use, delivered the current to
the consumer, metered the utilization of electricity, managed customer concerns,
and billed and collected for the services provided.
Production, transmission, and distribution of electricity has been a
comprehensively regulated industry for a century, with demarcations of
responsibilities established between federal, state, and local jurisdictions. The
federal government made rules on interstate matters involving the transmission
grids and the NRC monitors nuclear power.4 The states regulated the power
companies holding franchises in their area. Generally, local political authorities
or municipalities managed the electric companies in areas where municipal
utilities were established. There is much inter-jurisdictional overlap and dispute.
The early drive for public oversight of electricity late in the 19th century
came as part of Progressive Era reforms in response to what Samuel Insull termed
“debilitating competition”. Over three thousand individual systems developed in
the first twenty years of electric system production between 1882 and 1902; a
4 CEC, “Electricity Report 1996,” 5.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
27
quarter of these systems were municipal operations. Insull, who had been
Thomas Edison’s secretary prior to becoming president of Commonwealth
Edison, also saw in regulation a mechanism to foster confidence in electric
company bond issuance; this confidence was essential to raise the capital
necessary to create the electric system infrastructure. In California, the period
from 1850 to 1930 resulted in over 500 gas, water, and electric companies being
created and then being consolidated into Pacific Gas and Electric. Governmental
oversight heightened after depression-era investigation of utility stock collapses
by the Federal Trade Commission (FTC) and the Securities and Exchange
Commission (SEC).
Electric system restructuring at the turn of the 21st century follows
considerable success in deregulating other industries, specifically transportation
(rail services and trucking), airlines, telecommunications, and natural gas
generation and pipelines. Restructuring the electric system has begun with a
disaggregation of functional components and products, and with the development
of market mechanisms for generation. The interest in restructuring is not uniform
across states although some degree of restructuring is underway in 42 states. For
a variety of reasons, some states have a much higher average system cost than
others.
States with prices higher than average have shown the most interest in
electric system restructuring as a mechanism to introduce competition and lower
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
28
costs.5 Low cost states have argued that no imminent value to their citizens is
apparent, but proponents of reduced regulatory controls consider management of
electricity an interstate commerce matter that should be regulated not state by
state but nationally. The Department of Energy estimates that some value will
come to all states from retail competition and customer choice. DOE predicts
greater value for high cost states and reduced overall variation in prices between
states, with remaining regional variation due to differing fuel prices, operating
costs, transmission costs and constraints, distribution costs, and the recovery of
stranded costs.6 New entrants to the business of electrification and older vertical
utilities seeking new markets in the territories previously not accessible to their
business have sought Federal support for restructuring across all states equally.
Origins
California’s natural interest in leading the country in energy restructuring
were heightened by an economic slowdown in the late 1980s and early 1990s. The
state faced a problem attracting new industry in the post-Cold War era with its
non-competitive price of energy when compared to neighboring states. Its
economy ranks sixth largest among the world’s nations, so restructuring decisions
in California would have regional if not national effects. In 1996, when AB 1890
was passed, California's electricity cost approximately 50% more than the
national average, and reducing the cost of electricity for all rate classes was a
5 White, Matthew W., “Power Struggles: Explaining Deregulatory Reforms in Electricity
Markets,” 221.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
29
principal driver for the legislation, which added section 330 et seq. to the PUC
code. Explicit in the legislation is a 10% immediate reduction, and a promise for
an overall 20% reduction in rates by 2002.7 Even with the promised AB 1890
reduction, the state would still have higher base prices, but restructuring also
offered the hope of many new electricity products and providers, allowing more
targeted options for consumers and businesses.
Early activities to introduce competition in the wake of PURPA moved
forward in California and other states. Independent power producers could
depend upon regulatory requirements to assure their product would be part of the
energy mix. The development of “standard offer” contracts based on short term
avoided costs, and subsequent development of the S04 contracts resulted in
o
favorable terms and a high response from QFs. Both public and private power
companies participated in the planning committees around the FERC open access
legislation.
Munis and IOUs
Traditional vertical utilities developed either as privately-held or public
stock ownership companies, or as public sector operations. Monopoly franchises
limited direct business competition between governmental institutions and
investor-owned utilities. With restructuring, one issue that has emerged is the
6 Department of Energy, Office of Economic, Electricity, and Natural Gas Analysis,
“Comprehensive Electricity Competition Act: Supporting Analysis,” July 1998, 4
7 The promise is foremost of the legislative findings and declarations: “It is the intent of the
Legislature that a cumulative rate reduction of at least 20 percent be achieved not later than April
1, 2002.”Public Resources Code Section 25000 et seq., Section 330(a).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
30
unusual situation of public and private companies competing with each other for
power supply and customers.
Investor-owned utilities (IOUs) often have names that include the term
public, but they are not government owned companies. They are termed public
companies because they are owned by shareholders of publicly traded stock.
These private sector companies hold exclusive franchises for service, with
regulated rates based on fair rates of return for their shareholders.
Publicly owned power companies, also called municipals or munis, often
developed in cities adjacent to rivers. In these locations water management and
electric services developed in concert early in the twentieth century. A municipal
utility operates similarly to an IOU with the exception that any amount over cost
goes into general funds of the public enterprise. Publicly owned enterprises are
governmental agencies which may be self-governing, or departments in city
governments. They range in size from the Los Angeles Department of Water and
Power (LADWP) with two billion dollars in annual electric revenue and 1.4
million customers to small town agencies. The Federal government also has
ownership of some power resources; these often resulted from the construction of
major dams in the arid West. Coordination between municipal power agencies
occurs through the American Public Power Association (APPA). California has
three additional municipal power associations: the Northern California Power
8 Hirsh, Power Loss, 96-97.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
31
Agency (NCPA), the Southern California Public Power Authority (SCPPA), and
the California Municipal Utility Association (CMUA).
Two other utility structures exist, but are not under investigation in this
study on effects of the partial restructuring. Ten federally owned utilities produce
and provide mainly wholesale power, preferentially sold to nonprofit entities such
as municipal utilities; the Tennessee Valley Authority is the largest producer in
this category. There are also over 900 cooperatively owned utilities; most were
created as part of the Rural Electrification Administration and owned by rural
farmers and communities.9
California’s Electric System and its Major Service Companies
This study examines the five largest public and private utilities in
California. California’s utilities and independent power providers can generate
55,000 megawatts; in 2000 demand peaked at about 44,000 megawatts.
California does not generate all the power it uses. On any given day, some plants
are offline for planned or unplanned shutdowns. California utilities also have
ownership or long term contract interests with power generated in other states.
The five utilities represent the large majority of California’s electric system and
customers. These five provide the opportunity to compare utilities that are
directly subject to the restructuring rules of AB 1890 (the three IOU’s) with
9 Department of Energy, Energy Information Administration. The Changing Structure of the
Electric Power Industry 1999: Mergers and Other Corporate Combinations, DOE/EIA-0562(99),
December 1999, 12.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
32
utilities that are only indirectly affected by the new rules (the two large municipal
utility agencies).
Pacific Gas and Electric (PG&E) began as the San Francisco Gas
Company, incorporated in 1852, and its first gas plant was completed in 1854.
Many other companies also sought to develop the San Francisco utility
marketplace, resulting in a price war that by reduced prices from the initial rate of
$15 per thousand cubic feet to $1.60 per million cubic feet in 1870. Numerous
buyouts and consolidations followed, in a regular pattern that resulted in PG&E's
incorporation in 1905. PG&E's last major merger was arranged in 1930, as it took
over Great Western Power and San Joaquin Light and Power, capping the
consolidations, with monopoly coverage of most of northern California. For
1999, PG&E's electric operating revenues were $9.23 billion for a service area
covering 70,000 square miles and 13 million people. In 1998 and 1999 PG&E
sold its fossil fuel and geothermal generation; the fossil fuel units sold for $1.3
billion, more than double their depreciated book value of $602 million. The
geothermal generation was sold for $213 million, below its book value of $244
million. The three plants sold in 1998 had a generating capacity of 2,645 MW,
the 1999 sale divested an additional 4,289 MW.1 0 As of 2000, only its nuclear
plant at Diablo Canyon and the Sierra watershed hydroelectric generation remain
as PG&E units. PG&E also established an unregulated retail energy service
1 0 SEC 10-K filing for Pacific Gas & Electric Company, SEC File Number 001-02348, 9-10
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
33
company; by 1999, this unregulated line of business had over $2 billion in
contracts for large-scale energy projects and commodities.1 1
San Diego Gas and Electric (SDG&E) is the principal subsidiary of Enova
Corporation Sempra Energy, which was formed in 1998 as a holding company for
SDG&E's then parent company Enova and Pacific Enterprises. SGE&E
recovered costs of stranded assets early, through the sale of its generation assets
and the difference between its regulatory rate of return and the frozen rates paid
by customers in 1998 and 1999.1 2 This resulted in SGE&E receiving approval to
shift from regulatory based rates, as part of its overall shift to performance based
regulation (PER). In 1999 SDGE stopped using fixed rates and begin using a
pass-through charge for electricity to its customers; July 2000 was the first billing
cycle where this impact was noticed when ratepayers received bills reflecting as
much as a three-fold increase in their energy costs.
Southern California Edison (SCE) is the second largest investor-owned
utility in the United States, and a subsidiary of Edison International. Its service
area consists of 50,000 square miles in central and southern California. SCE is the
distribution utility for 4.3 million customers; the total population of its service
area is approximately 11 million Californians.1 3 While 85% of SCE’s fossil
generation was sold as a result of AB 1890, SCE continues to operate San Onofre
1 1 PG&E Corporation, “Annual Report 1998,” 3.
1 2 As stated by management, “San Diego Gas and Electric succeeded in recovering the majority of
its stranded costs-and in lifting its rate cap-more than two years ahead of schedule .. .” Sempra
Energy, “Annual Report 1999,” 5.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
34
Nuclear Generation Station with two operational and one decommissioned
generation unit.
The Sacramento Municipal Utility District (SMUD) has provided electric
service to the city of Sacramento, Sacramento County, and a small portion of
Placer County since December 31, 1946. A municipal utility district formed
under California’s MUD Act, it is governed by a seven member board serving
staggered terms and responsible for appointing the general manager. In 1989 an
advisory ballot measure in Sacramento passed, requesting SMUD to close Rancho
Seco, its sole nuclear power plant. The plant had endured many disruptions and
rates had been raised several years in a row. A string of general managers was
replaced after the closure vote with S. David Freeman, the former head of the
Tennessee Valley Authority and a Carter administration advisor. After four years
at SMUD and a term at the New York Power Authority, Freeman would
eventually return to California to set up the ISO and PX. He then moved to the
City of Los Angeles Department of Water and Power, replaced by SMUD’s
general counsel Jan Schori.1 4 SMUD has chosen not to join the ISO. Its decade
of power contract buying, retained hydropower generation, and construction of
380 megawatts in cogeneration capacity positioned it well early in restructuring.
Persistent high generation prices the latter part of 2000 combined with a delay in
rains needed to replenish its hydroelectric system. These factors exposed SMUD
1 3 Southern California Edison Company, “Annual Report 1998,” 2.
1 4 "L.A.’s Latest Luminary,” Los Angeles Times, December 23, 2000.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
35
to higher-than-expected short-term power costs and depleted its “rate
stabilization” fond, so SMUD has begun a rate review process about six months
earlier than previously anticipated.
Los Angeles Department o f Water and Power (LADWP) was established
in 1925 as a department within the overall City of Los Angeles bureaucracy. It is
the largest public electric service in the United States with over 1 million
customers and $2 billion in annual revenues. Traditionally the head of LADWP
was a political appointee of the mayor with little experience in electric system
management. With the onset of California’s partial restructuring, the City
Council decided that the department would need a new vision and hired S. David
Freeman. LADWP has pursued a course similar to other utilities in downsizing
its workforce by over 20% and retiring debt. As a municipal utility LADWP is
not subject to AB 1890’s requirements and has not divested its generation or
joined the ISO. In 2000, LADWP sold excess generation into the Power
Exchange on many days, resulting in a large reduction of its debt load.1 5 Also in
2000, LADWP developed a ten-year integrated resource plan (IRP) to add
substantial generation within its service area.
Challenges in the Restructuring Process
The separation of generation resources from transmission operations in the
formerly vertical monopolies has required at each generating site monitoring
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
36
points to validate for open access grid operators the quantity of power entering the
grid. The nature and mechanics of these edge systems were developed through a
collaborative committee process designed in response to FERC Orders 888 and
889. In addition to systems developed to manage transmission, new monitoring
and information systems have been required to introduce direct access. A direct
access customer participates in a form of retail wheeling where a non-local energy
provider coordinates load and customer usage information with the monopoly
provider of distribution services. California’s direct access programs have been
problematic. In other states that have deregulated, a larger percentage of
residential customers have changed power providers. Though most areas have
some providers for commercial customers, few have found it possible to compete
for residential customers. For example, by mid-2000 in San Diego there were no
energy providers offering competition for San Diego Gas & Electric customers.
Technological Challenges
Several technological challenges result from the new competitive
structures. They include the information systems at generation and edges,
metering and load shedding to better manage demand, and distributed generation
as an option to increase generation availability without incurring large capital
requirements for expanded transmission systems.
1 5 LADWP issued press releases nearly daily, stating its expected peak and power available for
sale. See, for example, “Los Angeles Has Adequate Power Supplies,” December 12, 2000.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
37
In the former vertically integrated utilities, all operating departments were
aware of conditions in other operating units. The divestiture of generation by the
IOUs and the development of direct access customer relations beyond the former
verticals’ geographic boundaries have created the need for new information
systems. One set of systems has been put in place at generation sites which
provide the ISO realtime information on generation into the grid. For non
participating municipal utilities new systems at their boundaries provide edge
reporting to the ISO.
Although the new utility distribution companies (UDCs) remain
monopolistic enterprises, there is considerable pressure and some interest in
further reducing the monopolistic scope in the delivery of customer billing
systems and metering. Specific metering solutions have been proposed by the
IOU and by the legislature to provide more customers with time-of-use (TOU)
information that could be used in peak periods to establish special peak rates.
With the new metering technologies customers would have the ability to check
the cost of energy and adjust their load according to their tolerance for cost
variability. Additional load shedding equipment such as radio transmitters to turn
off air conditioning remotely is being considered more seriously since the summer
of 2000 load seemed resistant to normal price variability. The average price for
August 2000 around the clock was $150 per megawatt hour; the price ranged
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
38
between $250 and $400 in December 2000, despite a 50% reduction in load from
summer peak demand.
Distributed generation (DG) can provide local generation in smaller
amounts-10 to 50 megawatts. It offers the potential to increase grid reliability by
reducing the draw into excessively energy-hungry and congested electric routes
such as Silicon Valley. However, these solutions are not yet as economic or
technically feasible as traditional energy efficiency programs (California Energy
Commission 2000). They have also faced substantial opposition from existing
utilities and the ISO. The statewide financial distress brought on by the pricing
problems has renewed interest in reducing obstacles to distributed generation.
Political Challenges
AB 1890 expectations of and promises to ratepayers and stakeholders
alike have created considerable political challenges. A strong and negative public
reaction to the apparently unforeseen problems in San Diego rates in the summer
of 2000 threatened one state Senator’s seat; prompt action by the legislature to
roll back residential rates dampened the public outcry until after the election.
Even in the benign energy environment of 1998, utilities, legislators, and
regulators had to fight off a grassroots initiative, Proposition 9. The proposition
developed in response to the steep discounts large customers received which
contrasted with the minor residential rate reduction paid for through revenue
bonds that resulted in long-term surcharges on their bills.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
39
A political propensity for immediate remedies, particularly in election
season, limits legislative response to fast or simplistic approaches. It is not
unreasonable for politicians to behave this way, since the ability to author
substantive legislative action requires one to hold elected office. However, issues
in managing market introductions are unlikely to be meaningfully addressed
without extended discussion, investigation, and process modification.
California’s ISO design created political challenges. The California ISO
began operations on March 31, 1998, and controls 124,000 square miles of
transmission lines, the second largest control area in the U.S. and the fifth largest
worldwide. Despite its size, the ISO is state-specific, unlike others which cross
state boundaries as did their power pools from which the RTO groupings
originated. This is a significant distinction. The California ISO's board of
governors was designed with 25 members and 4 advisory members to represent a
depth of technical background and provide a seat for each electric system
operational interest. As one part of the legislative response to the restructuring
problems in California, this expert board was replaced at the end of 2000 by a
governor-appointed board with 5 members. The replacement board has one
member from the prior board, and one volunteer from the governor’s Electricity
Oversight Board (EOB); the other three appointees have no background in energy
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
40
policy or electric operations. FERC Commissioner Hebert has expressed concern
at the politicized nature of the revised board structure.1 6
Regulatory Challenges
Several regulatory challenges to electric system restructuring exist. For
one, there is substantial overlap in jurisdiction, to the degree that utilities struggle
to operate in accordance with often conflicting directives. First, there is the direct
regulation of the California Energy Commission, the California PUC, FERC, and
in the case of SCE, PG&E, and SMUD, the Nuclear Regulatory Commission.
While the regulatory roles are supposedly distributed and demarcated, in practice
the most difficult issues often get shoved from one agency to the next; in the
summer of 2000 the CPUC asked FERC to declare the generation market non
competitive and to set lower wholesale prices.
Secondly, there is a problem when the regulatory agencies have multiple-
area conflicts. FERC has supported restructuring initiatives for many years, and
so the request to FERC from California agencies to move back to cost-based
pricing does not have strong support in Washington. These problems are
worsened in the partial restructuring design of California, as it tries to move part
of the industry into a market design while still experiencing problems with aspects
of the regulatory contract.1 7 The state of regulation post-AB 1890 in California
1 6 Hall, Carl T., “Critics Say ISO Rookies Will Hinder Solving Crisis”, San Francisco Chronicle
http://www.sfgate.com/cgi-bin/article.cgi7fileNchronicle/archive/2001/02/20/MN133473 .DTL
02/02/2001
1 7 “The three components of the regulatory contract are entry controls, rate regulation, and utility
service obligations. The state commission controls the entry of the utility's competitors and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
41
resembles neither an open market nor a regulatory contract. Rather,
circumstances favor large players on both the supply and demand side.
While any new company potentially could bid to supply power to
customers, utilities and not the PUC determine the rules and requirements for
participation in their service areas. By the beginning of 2000 only 156,397
residential and 53,355 commercial customers statewide, representing 2.1% of the
residential and 13.8% of the total load, have exercised their direct access option
and switched their power suppliers. In Sacramento, only one company offered
residential customers power and it withdrew from this line of business in July
2000. Potential participants have stated that the rules SMUD developed for
participation were too restrictive and serving small customers could not be done
profitably (Peyton 2000). Conversely, 31.7% of commercial customers with
loads over 500 kW moved to alternative suppliers (California Energy Commission
Supplemental Direct Access Implementation Activities Report: Statewide
Summary 2000).
Legislation Altering the Comprehensively Regulated System
Legislative measures at Federal and state levels have moved the
restructuring process for electric systems forward. These are discussed below.
authorizes rates that give the utility's investors the opportunity to earn a 'fair' rate of return on their
investment.” (Sidak and Spulber, 113)
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
42
National Legislation
Early federal level deregulatory legislation, PURPA in 1978 and the
Energy Policy Act of 1992, required utilities to buy energy produced by
independent producers and allowed customers to develop their own energy
supplies. These measures resulted in the emergence of numerous small
generating firms and advanced the technology of generation by assuring small
producers a guaranteed purchaser and price for the energy they generate.
In April 1996 FERC issued Order 888, shortly followed by Order 889.
These Orders described the means to implement the 1992 Energy Policy Act
spurring the widespread investigation of energy restructuring by the states. They
affirmed the requirement that utilities open their transmission systems to allow
transport of generation products between external parties over their formerly
private networks. The FERC Orders also initiated the separation of utilities’
generation and transmission functions.
FERC has since issued its Order 2000 which requires RTOs and the
California ISO to file their transmission system designs. The Board of Governors
has held discussions at the ISO on the apparent difficulties that some hold to be
due to its present design; others feel the ISO is hampered by its California-
specific scope. In response, the ISO has requested a FERC inquiry and findings
on the market structure in California and would like FERC to offer a return to
cost-based generation pricing. FERC found the California summer 2000 market
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
43
“unjust and unreasonable” but chose not to intervene, requesting the CPUC to
take corrective action.
State Legislation
California’s AB 1890 was developed by the legislature in response to the
efforts by the three private utilities in California to restructure the electric system
on their own terms through the California PUC. This bill froze residential rates
from 1998 through 2002, authorized the collection of a competition transition
charge (CTC) to compensate utilities for stranded investments, established a fixed
percentage for public goods and low income programs, and authorized the
floating of bonds to allow utilities to offer an immediate 10% reduction in their
residential rates. It also promised general ratepayers at least an additional 10%
reduction from their 1998 rates in 2002, following four years of CTC charges.
In 1998, a coalition of consumer advocates led by Harvey Rosenfield put
an initiative on the November ballot to undo AB 1890 and reassess the
restructuring scenario to be more favorable to residential ratepayers. The
consensus of these groups revolved around the immediate strong rate reductions
for large commercial customers, and the availability of alternate energy providers
for these same large commercial accounts. While many of the largest customers
had negotiated steep reductions as soon as AB 1890 was enacted, the only
reduction small commercial and residential customers received was a 10%
reduction, funded not by improved utility operations but by the issuance of special
bonds supported by a 10-year residential rate surcharge.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
44
Proposition 9 failed, in the minds of restructuring supporters, from its ill-
conceived and erroneous nature, and from a general perception by the public that
the California restructuring model was functioning as promised. Conversely, the
utilities and other AB 1890 supporters spent $40 million on the “No On 9”
campaign; Proposition 9’s grassroots supporters were certain this provided the
winning edge. Regardless, the California voters soundly defeated Proposition 9,
and so the only challenge to change AB 1890 in the first three years of
restructuring ended abruptly. Then gubernatorial candidate Gray Davis opposed
Proposition 9; some political strategists feel the multi-billion dollar unpaid bill for
summer of 2000 market-based electricity may cause Governor Davis problems in
his bid for re-election in 2002 (Morain, 2000).
Other small bills submitted and passed by the California legislature in the
early years of restructuring (1996-2000) were unremarkable. AB 1890 was held
to be a very successful bill as augured by the selling prices of generation sold off
by the IOUs in anticipation of wide participation in new and unregulated energy
services.
The congratulatory mood changed in the summer of 2000, when San
Diego Gas and Electric began to pass through to the customer the prices it was
paying for electricity in the peak summer season. Unlike two or three years
earlier when energy costs of one to two cents were common, the generation price
passed through to unsuspecting SDG&E customers averaged seventeen cents in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
45
August, a ten-fold increase. Political maneuvering by Governor Gray Davis led to
the state Senate drafting and passing in a single day a bill designed to roll back
and hold SDG&E customers rates constant pending a re-tooling of the energy
marketplace. At this point the Republicans in the legislature requested a special
session on electric restructuring. Governor Davis declined to do so at that early
juncture.
Like all regulatory commissions, the California PUC sometimes risks
providing more benefits to the companies they are entrusted to regulate than to the
ratepayers. The state legislature codified an example of this in the AB 1890
restructuring process by establishing the somewhat arbitrary 10% rate reductions
funded by special revenue bonds received through a rate surcharge. Freezing
retail rates while letting wholesale generation prices float allowed utilities several
years to recover uneconomical investments through the special CTC and through
additional revenue due to the difference between the wholesale and retail price of
generation. This also pushed direct effects on ratepayers from less predictable,
more dynamic market forces several years into the future. Divestitures of
generation improved the cash position of the utilities as well; the plants were sold
at much higher than book value. San Diego Gas & Electric recovered their
stranded investments over a year early and sought to bring market pricing to their
customers much earlier than the 2002 original timeline. This is why SDG&E
customers were the first to experience direct pass-through pricing.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
4 6
Few of the two dozen bills proposed in the 2000 legislative season were
passed into law. Among them were last minute joint resolutions to freeze and roll
back rates in San Diego and AB 970 to expedite new power plants in California.
In 2001, an extraordinary session of the legislature was called to address
what had become an energy crises of colossal proportion: the upside-down nature
of the wholesale/retail price gap had gone unchecked for several months, resulting
in financial instability for two of the UDCs. Power providers stopped selling
through the PX or the ISO to the two large UDCs, since they were no longer
getting paid. A flurry of legislation put California in the power buying business.
AB7 authorized the Department of Water Resources to purchase on behalf of the
UDCs. ABlx 1 authorized a ten-billion dollar bond issuance for long term power
contracts, $500 million in short term money, and a mechanism to allow DWR to
spend as much additional money as it deemed necessary in the interim, simply by
notifying the legislature ten days in advance. The early legislation in this session
that was passed into law does not focus on the market or institutional issues most
fundamentally linked to the fiscal problem, merely provided a new monetary
source allowing the market behavior to continue unabated. In order to assure a
market for the new state-managed power contracts, ABlx 1 suspended direct
access; existing DA customers may find their pricing contracts restored in later
legislation.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
47
Trends in Electric Transport: Retail and Wholesale Wheeling
To a greater degree than other deregulated services electric system
restructuring faces strict limits, especially in the narrow range of acceptable
performance allowable in the transmission and distribution grids. The model of
limited restructuring design to produce competitive generation markets may cause
non-trivial disruptions of the transport system; such claims were made during
blackouts and brownouts in the summer of 1998 and 1999, and resulted in the
formation of a Department of Energy Power Outage Study Team (POST).
Transmission or wheeling of energy continues to be highly regulated.
Restructuring describes the process of deregulating one or more components of
electric energy production and distribution. It includes determination of which
parts of the electric industry constitute natural monopolies requiring continued
state and federal regulation, what can be opened for market processes, and the
degree to which an unregulated market will be allowed. The task of moving
electricity between the source and the consumer, wheeling the power involve two
separate applications of restructuring: wholesale wheeling, between producers
and bulk purchasers (generally, the electric companies) and retail wheeling,
between producers and the end consumers. Generation and power purchase can
involve any two players, but the transport through the grid must remain highly
controlled to assure grid reliability.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
48
Retail wheeling permits a company to wheel-offer to contract and arrange
transport- energy from their generation source to an individual customer, using
the shared transmission system and in turn an existing utility’s distribution
system. The accounting which allows for this type of transaction includes both
the system scheduling and the customer billing which must be coordinated
between the third party supplier and the distribution utility. It differs from
wholesale wheeling in the ability of individual customers to structure contracts
with power providers, and requires their local power company to accept the
electricity from the third party for the customer’s consumption; the customers pay
a wires charge for this service.
Disruptions are considered a national security risk, and a sense of electric
system reliability impacts our economic well-being. The development of
premium power programs creates the potential for a new class bifurcation: the
haves and have-nots of reliable power supply. National security concerns about
the electric system have been voiced by the Critical Infrastructure Assurance
Office established in 1998 following a report on the overall state of the national
infrastructure.
Deregulating generation and establishing separate organizations to manage
the transmission system carry a cost, the cost of the new organization plus the
creation of systems to monitor power moved onto, off of, and across different
companies’ formerly private systems.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
49
In states where restructuring is altering the management of the electric
system the cost of correcting system problems and maintaining reliability could
be disproportionately allocated to small customers and retail consumers. Large
commercial and government users can exercise newly-created options to shift to
other than their historic suppliers or use their large customer status to obtain
advantageous pricing from the utility in exchange for fixed-length, frxed-quantity
commitments.
Utilities, large customers, ratepayer advocates, legislators, regulators,
policy makers, and to a lesser degree the general public have discussed the issue
of understanding how costs are allocated. Responses to a desire for more
complete information on the cost and charges for electricity include bill
disaggregation. This process separates the formerly unified billing into numerous
charges subject to different rules, regulations, and limits. Utilities have sought,
and the PUC/PSCs have approved, exit fees paid when a customer chooses to
contract with energy from another provider. This alters the overall electric system
funding and expenditures, effectively introducing new inefficiencies and possibly
raising overall costs.
Conflicting opinions are held on the manner and value of deregulating the
electric industry. There are several issues under discussion but not yet fully
developed: the future role of publicly owned power, the implications and
consequences of electric restructuring on corollary resources such as scarce
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
50
western water supply and management, and the regulation of costs and rates for
the mixed deregulated and regulated components. These need further explication
for public decision makers who must determine what course of action provides
the greatest public value. The small and decreasing number of investor-owned
electric utilities offers opportunity for market power and political pressure-based
strategies (rent-seeking behavior). Protecting public value under these
circumstances requires much study and analysis, with results plainly stated.
Information on a macro level is spotty statewide, though the ISO has provided
access to a great deal of its transaction history. Lack of sufficient information to
determine appropriate responses to signs of market instability has been repeatedly
voiced throughout 2000 and into 2001. The lack of plain language, open
discourse, and an arbitration structure for representatives of different
constituencies may have contributed to inaction when the market perturbation in
California had not caused such major fiscal damage.
Operational Context of Regulatory Activities
Regulatory theory, market theory, regulation and legislation do not
establish a deregulated industry. Action issues dominant in reducing regulatory
constraints and introducing competition involve the establishment of institutions
and organizational models to carry the process forward.
Incentives and constraints are also significant. Attempts have been made
to protect small residential and low income customers through mandatory rate
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
51
cuts and fixed amounts to be spent on public good initiatives. Mandatory
divestitures of generation have been required as part of mergers. Both of these
tactics are problematic. The public good portion of rates that would have expired
in 2002 has been discussed in a number of legislative proposals. The principal
developers of these proposals have been supporters of advanced and renewable
technologies, such as solar power and biomass (garbage, usually). The allocated
charge is not designated specifically as a reduction in rates for low income and
certainly not as a rate stabilization fund to protect all residential customers from
fluctuations in market pricing. The mandatory divestitures of generation have
gone more smoothly as the generation facilities have been sold at above book and
estimates. This fact has allowed for more rapid payoff of the stranded assets. San
Diego Gas and Electric speeded up their payoff and ended AB 1890-controlled
rates to residential customers in 1999. The fluctuating costs of peak power were
considered no longer SDG&E's problem-the IOU as of 2000 can pass these
increases directly to the consumer. Since the retail consumer has few to no
alternatives for power, this caused considerable uproar in the summer of 2000
when the cost of power rose from under three cents to over ten cents.
Managing the spot market is a significant concern which has been studied
more completely overseas where the deregulatory reforms have been in place
longer and the electric infrastructure is more unified. The California-specific
Power Exchange (PX) has come under scrutiny with the pricing problems of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
52
2000; multiple small and large alterations have been suggested. These
suggestions derive in part from presumptions regarding the absence or presence of
market power. Since no consensus regarding market power has developed, these
solutions tend toward contradictions.
Discussion and legislative/regulatory responses to restructuring proposals
are ongoing in state legislatures, at the Federal level, and around the world.
Economic theorists concur on the value of markets; but the technical constraints
of existing technology for electric generation, transmission, and distribution
create interconnections that limit the shift to a completely open market.
Additionally, legislators and regulators must determine the extent of rate
protection for small consumers and public institutions, and decide on subsidies for
research and implementation of new, more efficient, or less polluting electric
services.
Mergers and acquisitions move the management of companies with
generation, transmission, and/or distribution responsibilities out of local
jurisdictions and even to other countries. In 1997, PG&E arranged to sell its
fossil and geothermal plants in the San Francisco Bay Area. Southern Energy,
Inc. bought the Delta and Potrero power plants, 2,700 megawatts. FPL Energy,
Inc., an unregulated subsidiary spun off from the traditional utility Florida Power
and Light, bought the Geysers geothermal facility as part of its acquisition of 208
megawatts of California’s renewable energy generation. Calpine has
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
53
approximately 900 megawatts of energy formerly owned by SDG&E and PG&E
as part of its worldwide portfolio. Dynergy now owns 2,000 megawatts of
SDG&E generation. Enron acquired 145 megawatts of wind power for use in its
green energy program. Reliant Energy bought 3,500 megawatts of SCE’s
generation. While the California PUC does not regulate production, the
California Energy Commission still maintains its database of all generation and
ownership (California Energy Commission 2000).
Could out-of-state control of in-state resources be a contributor to current
in-state reliability problems? The locus of responsibility for brownouts and high
peak prices has not been established, so in-state generators are not directly faulted
for selling their product to the highest bidder, in-state or out-of-state. FPL, Inc.
has the fiduciary responsibility to achieve the return on investment to its
stockholders. PG&E had a historical commitment to offer reasonable electric
prices and high quality of electric service to Bay Area customers, protected by a
guaranteed rate of return. Restructuring allowed all three California IOUs to
separate into UDCs, unregulated generation companies, holding companies, and
energy services firms. When PG&E asked the CPUC for a 39% rate increase in
December 2000, an audit revealed the utility had moved much of the money over
collected since 1996 into the parent corporation, as did SCE. While the public
reaction was not favorable, these CTC funds were provided for stranded cost
recovery. It was not the utilities’ obligation to keep the funds in the UDC for
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
54
future investment in the regulated segment. California-based utilities find their
priorities altered by restructuring, and must operate in similar fashion to all other
energy-related firms, as industry mergers, acquisitions, bankruptcies and
consolidations affect all players in a competitive market.
The Role of Energy Efficiency and Demand Side Management
Energy efficiency denotes programs, research and activities a utility uses
to reduce demand for energy. Using neo-classical economic terminology, utilities
employ energy efficient technologies to alter for their advantage the demand side
of the supply and demand equation, thus the term demand side management, or
DSM. Energy conservation came into widespread use in the mid-1970s when
production changes by the OPEC oil cartel resulted in nationwide shortages. In
the 1980s and early 1990s, energy efficiency programs were developed to manage
the load curve. The 24 hour consumption of electricity undulates in a wave form,
peaking in the mid- to late- afternoon, bottoming out in the pre-dawn hours.
Flattening the curve reduces the amount of generating capacity that is needed to
meet electrical demand.
While a number of initiatives ended with the 1970s crisis, utilities found
the programs had beneficial effects in flattening the load curve, particularly in
reducing summer peak load. There was a dual value in this reduction. First, since
consumers paid the same rate at 10 a.m. and 6 p.m., the utilities wanted to make
sure they did not pay overly high prices for 6 p.m. generation should a heat storm
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
55
or other circumstance result in them obtaining energy in the spot market. Second
and more importantly, additional load at peak could trigger the need to build a
new power plant. Utility-sized plants are expensive, $100 million or more, and
the super-peak demand (highest demand) occurred only about 200 hours per year.
The value of using energy efficiency measures greatly exceeded the cost of new
generation; avoiding the generation also avoided the externalities, such as
greenhouse gas emissions or birds of prey killed in collisions with windmills. The
policy of seeking efficiency first before constructing new generation was popular
in California for over a decade. This practice ended with divestiture and the PX
market for power, but was revitalized late in 2000 with the increase in ISO low-
margin emergencies.
Energy efficiency works best with a range of efficiency measures. Load
shedding or curtailable load has been a staple of energy system management for
over a quarter century since the energy crisis of the mid-1970s. Many large
companies have load shedding clauses in their energy contracts. These allow the
utility to request the customer to reduce its load at short notice in exchange for a
more favorable rate for the business. This same approach could also be used for
residential and small commercial customers. In the Sacramento area alone,
SMUD has over 100 megawatts of load shed associated with curtailment
agreements with over 90,000 customers who have agreed to limited radio-signal-
initiated cutoff of air conditioning at peak load.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
56
Energy efficient appliances have also been a significant part of energy
management planning. New furnaces, air conditioners, and refrigerators have
greatly reduced energy requirements. Compact florescent light bulbs designed to
fit standard sockets reduce lighting costs and save the energy wasted in the
generation of heat associated with incandescent illumination.
AB 1890 and the required divestiture of generation caused disruption and
reduction in energy efficiency programs. New and old generation owners alike
found a lucrative selling opportunity at peak, selling into the hour-by-hour spot
market. Managing to reduce peak load was no longer an advantageous strategy,
even for the municipal utilities and for the portion of the lOU’s generation that
was not immediately divested. Generators who sold at the peak could create a
per-hour revenue stream an order of magnitude higher than the same hour before
the passage of AB 1890. The rise in peak hour spot market is a common
phenomenon throughout the U.S.; under most circumstances, the price falls back
to standard levels within a few hours. The rise of pricing at the end of 2000 in
California did not follow the peak spot market pattern, as the prices were high
around the clock, and for many months beyond the summer’s peak demand
period.
SMUD data on load reduction exemplifies this disruption of energy
efficiency programs. Prior to AB 1890’s passage, SMUD’s dispatchable
reduction program had several incentive levels, and was invoked 25 times
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
57
between 1994 and 1996. A median reduction of 55 megawatts load reduction was
only exercised once in 1997, twice in 1998 and not at all in 1999. For the four
summer months of 2000, SMUD invoked its peak emergency cycling six times in
voluntary association with ISO statewide emergencies for a median reduction of
42 megawatts. Similarly, residential incentives to participate in load reduction
were reduced from $20 per month to $10 for the entire summer.
The role of energy efficiency has grown with debates on the fiscal and
logistic constraints of adding new generation and transmission to provide
adequate supply for the projected market. AB970’s passage established the
mandate to obtain as much as possible of the 1,000 additional megawatts expected
to be needed by the summer of 2001 through energy efficiencies. The
supply/demand mismatch at the end of the year 2000 resulted in the governor
establishing a new energy efficiency target of 15% of summer load, a reduction of
6,000 megawatts.
The debates on these many issues invoke the theoretical values and
empirical unfolding of deregulation efforts carried out in other industries or in
other places throughout the last half of the twentieth century . The literature
reviewed in the following chapter explains the expectations and validation from
which AB 1890 arose to justify subsequent utility, legislative, and regulatory
actions.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
58
Chapter 3
Literature Review
Introduction
This dissertation’s literature review summarizes the theoretical basis for
restructuring and the regulatory context for early activities in developing
regulatory models for restructured electric systems. The research question, “How
have regulatory decisions and restructuring legislation in California affected
electric system reliability, particularly the availability o f sufficient generation to
meet localized demand?” rests upon an economic rationale discussed in the
literature. This literature review further relates the research to the context of other
work in public administration, particularly the development of implementation
and evaluation, decision making, and complex systems theory. The review
develops the research question’s relevance in both an action and theoretical
context.
This dissertation derives from the theoretical basis for deregulation in
general, for network systems and public/private infrastructure restructuring in
specific, as well as a preliminary theoretical work in electric system restructuring.
It is hypothesized in this research that the electric system’s need for reliability
may negatively affect the predicted results of energy restructuring. Vice versa,
introductions of markets in lieu of regulatory structures may negatively affect
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
59
reliability. Discussion of the risks of deregulation or partial restructuring is found
in literature cited here.
The literature presented here can be divided into five streams: the validity
of the market model for network industries in general and the electric system in
particular; the feasibility of disaggregating the electric system; the means to
manage the transmission and distribution systems deemed not well-suited for
market mechanisms; the necessity to protect existing stakeholders from unfair
losses due to the introduction of market forces; and the public administration
perspective of this restructuring environment. Particular issues within these
streams have received heightened attention.
Market Rationale for Restructuring the Electric System
The market model is believed appropriate for portions, if not all, of the
electric system. Several general analyses of electric restructuring potential are
available (Joskow (1997), Baumol (1983), White (1996), Winston (1998)). The
fundamental argument in favor of electric industry restructuring centers on the
theoretical value of market mechanisms as being more responsive than any
regulatory approach.
Under monopoly franchises with a captured market and guaranteed rates
of return, few incentives have existed for improved operations. The state-by-state
rate setting for IOU’s and local governance for municipals meant that each utility
would make its case to its regulators or elected officials on the appropriate rate for
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
60
reliable electric service. Depending on the skills of utility officials and regulators
to establish effective operations, these rates varied substantially. Rates were 50
percent higher for some companies; in some areas, neighbors on differing sides of
a street were allocated to companies with dissimilar rates and service structures,
with no opportunity to be supplied by their neighbors’ provider. Rates also varied
for industrial and commercial customers.
Technological improvements also reduced the need to build very large
electric generation plants. More companies could consider onsite power
production, though Joskow (1997) notes that technology does not fundamentally
alter the nature of generation where competing sources have been putting
identical product onto the transmission system for many years. The reduced cost
of entry requires extra coordination at a system wide level to manage generation
capacity and production, and to assure available power to send into the grid
whether prices are at their highest or lowest.
The above activities were the beginning of envisioning power companies
not as vertically integrated suppliers of energy but rather as a collection of
generation, transmission, and distribution functions which could be redesigned as
separate units, operating competitively.
Successful Antecedents
During the latter part of the twentieth century, the economic value of
deregulation has been demonstrated in multiple industries in the U.S. and abroad.
Some industries, such as railroading, serve principally large commercial and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
61
industrial customers, while other deregulated industries have greater presence in
the general public marketplace. The electric industry has both these customer
bases.
For many Americans, the dramatic breakup of AT&T into the baby Bells
in the early 1980s exemplifies deregulation. Since that time, an astonishing array
of new technologies and products based on telephonic communication have
emerged, such as wireless phones, ubiquitous faxes and pagers, and value added
services from caller ID to fast Internet connectivity via digital subscriber line
(DSL). Many former services included in the bundled telephone charge are now
billed separately. These include formerly uncharged 411 directory assistance and
in home telephone repairs.
An additional rationale for electric system restructuring came from the
successful introduction of independent producers of electric power. The industry
experienced the beginning of restructuring post-PURPA in the substantial number
of independent power producers involved in qualifying facilities contracts.
Extension of generation competition and the lifting of remaining regulatory
barriers through establishing a market for generation seemed a reasonable
approach to initiating electric system restructuring.
Successful restructuring in England, Europe, and South America preceded
U.S. restructuring activities. Activities in other countries exemplified the process,
supplying impetus to reduce regulatory constraints on the electric industry in the
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
62
U.S. by demonstrating regulatory constructs and technical mechanisms required
for a restructured environment. Many U.S. companies are marketing services,
computer hardware and software, and expertise developed elsewhere. Most
adoptions from overseas restructuring come from Scandinavia and from the
United Kingdom. England has been especially aggressive in selling off
government owned power as part of its privatization movement while leaving its
Regional Electric Companies (Rocs) intact. American companies increase their
understanding of the deregulated energy marketplace through buying and
operating generation in South America and around the world. Northern Telecom
has developed a power line carrier technology based on the European style
transmission and distribution system for carrying data on power lines. This
technology is similar to the cable industry’s introduction of cable modems, and
offers potential for utilities to market data services over electric wires to their
existing electric service customers. Not all practices and technologies work
similarly in the U.S., but examining and applying restructuring models that have
been successful abroad allows for experimentation, innovation, and testing trials
while also heightening demand for the end of monopolistic power structures here.
Restructuring electric regulation in England and other countries was
simplest where formerly state-owned and operated utilities were privatized. The
i o
United States by contrast has comparatively little publicly owned power .
1 8 Comparing investor-ownership with the various government ownership models, the
approximate capacity balance in 1989 was 76% investor-owned, 10% publicly owned, 4%
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
63
Analysis of energy restructuring overseas is useful as other countries have moved
more aggressively than the United States to deregulate. For example, 30% of the
power in California is municipal power (CMUA 1998, in California State
Assembly Briefing 1999). By contrast, Britain fully nationalized its electric
system in 1948, and provided electricity as a state service for over forty years
before privatizing electricity in 1990 (Newberry and Green in Gilbert and Kahn
1996).
Managing under the new market model has been discussed extensively,
dating back to Bauman (1982) and his definition of the fundamental principles of
contestable markets, continuing in Newbery (1998) and Wolfram (1998). Baumol
states a concern that regulators may try to prevent exit of companies from
markets. Theoretical models are also strained by actual events when generating
companies, system operators and distribution company power managers must
determine how to deal with spot market spikes, decide when participants must put
power on the grid, and manage through unexpected price jumps. Less common
are unexpected price drops though rare occasions are occurring when the
generation price per kilowatt hovers near zero. This extremely low price can
occur, for instance, when water policy dictates a flow quantity to protect fish
populations when power conditions would not produce economic generation.
cooperative, and 10% federal. See Richard J. Gilbert and Edward P. Kahn, International
Comparisons of Electricity Regulation (Cambridge: Cambridge University Press, 1996), 179-184.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
64
Gilbert and Kahn (1996) edited a comparative study of regulation and
restructuring with chapters on seven individual countries and three regional
clusters. These authors attempt to map the commonalities of regulatory processes
to economic and political theory. Through their discussions on information
asymmetry and institutional analysis, they reiterate the finding that the ability to
capture value from restructuring depends on building elastic demand. As noted
earlier, the nature of electric systems tends toward short term inelasticity with
long term high elasticity.
Anticipated Gains
The United States began its approach to electric system restructuring later
than some other parts of the world. For instance, in California, the investor-
owned utilities sponsored a trip for the California PUC to South America to
review progress in that part of the world; these early adopters of electric system
restructuring were also discussed frequently in business newspapers such as the
Wall Street Journal.
The reasonable expectation from a good restructuring design is increased
transaction efficiency leading to reduced cost and increased innovation. As noted
in chapter two, U.S. states have historically operated independently and have
divergent energy pricing and consumption profiles. The desire for reduced cost
has driven restructuring in the U.S.; and so there is uneven interest in
restructuring. Low cost states may agree that restructuring can bring innovative
services, but they feel less compelled to move forward aggressively. Political and
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
65
social pressure in California were factors in the early implementation of
restructuring with its electric rates 50% higher than average.
This anticipated gain may be dependent on both the initial state of energy
regulation and also on the institutional constructs used for restructuring. Ruff
(1998) notes the distinction of prior public ownership in countries where
restructuring has been most successful. This distinction allows the government to
control the details of restructuring options and outcomes. Also, public managers
do not have the fiduciary responsibilities to their shareholders that may affect the
judgment and decisions private utility executives make. Ruff also feels that
California’s aversion to a non-governmental trading process, opting instead for a
quasi-govemmental institution through its construction of the ISO, bodes ill for
the partial restructuring model adopted by the state.
Significant dimensions of industry analysis on the issue of monopoly and
establishing a market are concentration and barriers to entry. The focal problem
is the price-cost margin (Williamson 1993). This analysis is borne out in
complaints of new entrants to the electric industry; they argue the incumbent
organizations are able to set rules making it extremely difficult for new companies
to participate as energy suppliers and generate profit. One mechanism limiting
these new entrants is the approval in AB 1890 of stranded costs recovery through
a non-bypassable charge per kWh on the typical disaggregated bill. This charge
reduces the advantage for retail customers interested in selecting new suppliers.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
66
Findings of market power (Borenstein, Bushnell, and Wolak, 2000; Joskow and
Kahn, 2000) support this complaint. Joskow and Kahn state that their analysis
complements the FERC staff report by quantifying factors identified by FERC,
though FERC’s November 2000 summary of the California market review
focuses more on market rules than on market power.
New Entrants
Restructuring has created hundreds of new energy firms participating in
the modified industry. Many new products and services are being created, such as
weather hedging contracts. Though some firms are managed by new entrants,
many of these firms are led by managers who have left behind their former utility
roles. Though executives of the private IOUs have not experienced radical
turnover, their background and experience is another important area to consider as
the market model for electricity is in sharp contrast with the vertically integrated,
highly regulated institutions where most executives have spent their entire
careers.
Management of utility companies is experiencing radical change.
Mergers, acquisitions, the development of parent companies similar to the holding
companies of a century ago, these create new and different organizations and alter
management requirements. Geddes (1997) examines the connections between
turnover in management of IOU’s and the wealth of owners and customers.
Geddes suggests this empirical research be repeated annually as utility stock
prices have vacillated substantially through the early years of restructuring
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
67
activity. Also, utility executives have sought protection from the effects on their
stock prices caused by competition through specialty legislation allowing them to
pass-through transition costs to ratepayers. Thirdly, the number of utility mergers
and spin-offs will have impacts on management turnover and corporate value.
Downward pressure on CEO salaries at utilities was also studied by Joskow.
Management teams at traditional utilities have sought to prepare their utilities for
the new competitive market. They have taken notice of products and offerings of
new entrants, and adapted their existing practices to re-market their activities for
the new environment.
Despite electricity’s inherently non-differentiated nature, one electron
being identical to the next, utilities and new entrants are developing brands and
specialty products typical of free-market maneuvering for market share.
Predominant among these activities for non-time-dependent products are green
power and premium power. Green power represents a supplier’s commitment that
only renewable resources will be used to generate a customer’s electricity; in
return, the consumer pays an extra cost, perhaps a penny a kilowatt hour (kWh).
Premium power represents a level of heightened customer service; the customer is
provided with an assigned account representative to work internally with utility
organizations to expedite service and other requests. Specialized products may be
separate value-added items available from the historic UDC or a new energy
supplier. A large customer desiring more comprehensive energy services can
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
68
make a request to their traditional provider. The products and services may be
bundled with a power contract offer by a new energy services company or by the
historic utility concerned about its large customers defecting to alternative energy
suppliers. These activities position the commodity energy product as a
differentiated product set; early indications demonstrate the promise of these
offerings for the providers of value-added differentiators. The veracity of claims
for green-ness in particular is not provable on delivery of product. Because a
segment of the customer base will pay more for electricity marketed as
environmentally friendly, monitoring and other regulatory controls will be
required to establish what constitutes green power and whether suppliers are
providing energy from these sources. Transmission system managers may find
claimed determination of green-ness to be another feature added to their perimeter
accounting mechanisms.
Neither of these products capitalizes on the most salient distinction in
value between electrons-its time of use. As noted by Joskow and Schmalensee
(1983), the peaking electrons late on a hot California summer afternoon are much
more valuable to utilities and consumers than the electrons available winter
evenings when most are asleep. The peak prices on wholesale markets since 1998
clearly demonstrate this. Seasonal and diurnal load cycles have long been objects
of transmission and distribution energy management planning. Marketing them
separately generally requires time-of-use (TOU) metering which is currently in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
69
limited use, mainly in commercial environments. Legislation requiring meter
replacement or augmentation to support TOU pricing was among the 26 separate
pieces of legislation introduced in the California Senate and Assembly during
their 2000 session. Metering requirements are controversial in their precise
details, however, as they require high capital investment and may eliminate a
potential realm for competitive participation by introducing a new stranded costs
component.
The Feasibility of Electric Industry Disaggregation
This section discusses literature on vertical industries and their
disaggregation. The questions here include re-examination of natural monopoly
arguments, possible components in a disaggregated electric service, reliability
constraints on disaggregration, and the complexity of the disaggregated system.
Natural Monopoly
The concept of natural monopoly is less compelling than before, largely
resulting from widespread restructuring in other industries. Early in its
development, the telephone industry was considered as much a natural monopoly
as the electric utility. The two industries shared high capital and universal service
requirements as well as the physical poles and wires needed to carry its product:
the voice signal for the telephone company, the electron stream for electric
utilities. Other monopolistic industries also underwent mostly positive
restructuring. The restructured airline industry provides many more flights in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
70
major markets and sophisticated, almost individualist pricing. Of concern to retail
ratepayers and rural counties facing electric restructuring, some smaller markets
have seen less satisfactory effects from airline deregulation:
Congressman Conyers from Detroit, Michigan, ignored estimates of the
large nationwide benefits of airline deregulation and questioned the
wisdom of the entire policy because he felt fares at Detroit’s Wayne
airport were “too high.”... To be sure, consumers have not shared
equally in the gains from deregulation. Long-distance air travelers on
high density routes have benefited more than short-distance travelers on
low density routes. (Winston 1998, 90-91, 102)
Joskow and Schmalensee (1983) describe a natural monopoly as a
situation where “a single firm can provide all of the output of all of the products at
lower total cost than could be achieved by more than one firm” (p. 29). These
authors continue that economies of scale of the range of demand is a sufficient but
not necessary condition for these natural monopolies; the classic example is
water supply. Hirsh (1999) notes that the monopoly status granted to utilities a
century ago came with both the obligation to serve and regulatory oversight,
supported by a guaranteed rate of return. While the technical conditions for a
monopoly status may no longer exists, as Hirsh argues, the technical and
institutional constructs that support the verticals also allow for the cooptation of
the PUC, as demonstrated by Selznick (1949) in the case of the Tennessee Valley
Authority (TVA).
Component Systems
Opinions differ on whether it is economically efficient to disaggregate the
electric system to achieve restructuring effects in parts. Some argue that
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
71
restructuring of other components has little bearing on issues related to the
transmission system. Legislative proposals are brought forward related to pricing
structures, generation divestiture, advanced technology supports and similar
measures. Rarely does the discussion on each measure include its effect on
existing systems, proposed systems, or on the parts of the system thought to be
distinct from the effects of these decisions (Barker et al. 1997).
If parts of California restructuring appear volatile or persistently
inefficient, and bill reductions unlikely, then other areas are more promising. The
conventional disaggregation of a vertical utility into generation, transmission, and
distribution was always considered an artificial if convenient construct.
Additional consideration of new technologies and ideas on restructuring
opportunities have further disaggregated the distribution company concept,
suggesting more areas for reduced regulatory controls and increased competitive
efficiencies. Services such as customer billing, metering, and similar business
operations appear generic and easily available as competitive component
operations.
While UDCs strenuously defend their customer service operations, a
deregulated meter similar to the deregulated telephone handset raises more
immediate issues. Discussion on the disaggregation of metering services has been
ongoing for years. It has grown more intense since the ISO included new
metering solutions in its set of reforms for “immediate implementation” in August
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
72
2000.1 9 Related to the development of automated meter reading (AMR)
technology which requires substantial capital investment to implement, there is
greater interest in using the new meters to institute time of use (TOU) rates and
encourage voluntary curtailment through the market mechanisms governing
pricing, supply, and demand.
A utility distribution company or third party investing in AMR or TOU
meters for an entire distribution company must invest approximately $100 per
meter times the number of customers, plus costs for network lines, central and
hub computers, and software to collect and process the meter reads. If legislation
for such a substantial capital investment is passed, it could result in the metering
supplier requesting exclusivity at the meter thus delaying any competition and its
concomitant efficiencies by many years. Alternatively, UDCs implementing
mandated meter enhancements could request that the PUC include these capital
costs as new stranded investments and allow for continued CTC collection on the
utility bill.
Capital requirements and structure have a substantial impact on regulated
firms (Spiegel and Spulber, 1994). The role of metering in a restructured
environment underscores a potential new arena for competition that might instead
result in continued regulatory protection of capital investment. Proponents of
1 9 From the proposed price cap resolution considered at the July 6th, 2000 Board of Governors
meeting. This special meeting was held specifically to rescind the Motion on Price Caps adopted
the week before, to leave the price cap unchanged at $500. The Board was under intense political
pressure to reduce the price cap to $250 (California ISO, July 6,2000).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
73
electric system restructuring have sought the creation of distinct components and
new market segments free from massive capital requirements and the regulatory
controls that accompany cost-based rates. Advocates of re-regulation have
requested protection from charges for new infrastructure related to specialty
services, but continued regulatory support for traditional public goods
expenditures that will also be paid by those who select other than traditional
energy providers and services.
A disaggregated market complicates pricing of certain components.
Demand-side management evaluation at Ontario Hydro in Canada is typical in
comparing the avoided or marginal costs with the average cost of the DSM
measure (Waverman and Yatchew, in Gilbert and Kahn 1996). These authors
note that economic efficiency is limited to the difference between avoided costs
and average cost, whereas often DSM programs are funded to the level of the
avoided costs. In California, the very low fuel costs of the mid- to late- 1990s
reduced interest in DSM programs; interest is resurgent both due to current supply
shortages and the high price of additional energy. If the cost of a new megawatt
of power is 120 dollars, DSM at five to ten cents per kilowatt hour has high value
in this marginal cost market. This example excludes any new costs for the
transmission to move the new megawatt and any externalities such as noise or air
quality degradation. Demand side solutions also eliminate the transaction costs
associated with obtaining generation siting approval by the CEC.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
74
Reliability Constraints
Reliability is an important differentiator between electric system
restructuring and other industries that have been deregulated. The electric grid
requires persistent, constant, invariant voltage. The electric grid is generally
considered the high voltage transmission system from the generator into the
stepped-down distribution system; the California ISO's management area is this
high voltage transportation system. But the distribution system is itself subject to
the same voltage management constraints, and the line between transmission and
distribution is not immutable. Both the transmission system and the distribution
system have remained regulated monopolies. The management of the
transmission system has been substantially altered in California and elsewhere
through separation of transmission management from the former vertically
integrated utilities.
The demand for reliability alters the ability to apply general economic
regulatory and market theory. High reliability systems have been discussed in the
literature; scholars show some differentiation in their analysis. The literature only
peripherally addresses the California ISO articulated ideas of command-and-
control and reliability at any cost, or limited reliability in limited regions at
negotiated lower costs.
Perrow (1984) analyzes the sources of so-called accidents in complex
systems such as nuclear power, petrochemical, aircraft and airways, marine
shipping, dams, and weapon systems. Perrow finds an irreducible number of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
75
“normal” or system accidents that augment better understood component failures.
Perrow uses a mnemonic, DEPOSE, to describe component failures: design,
equipment, procedures, operators, supplies and materials, environment.
Increasing risk derives from tightly coupled systems. Centralized management
and control in tightly coupled systems reduces many component failures, but it
cannot prevent system accidents. Perrow argues that many accidents are wrongly
attributed to “operator error,” based on unreasonable expectations of behavior,
prior training, and judgments under ambiguous circumstances. Managing for
reliability in these industries revolves around appropriate control, loosening of
coupling, and reduction of complexity. In cases where no potential reductions
exists, but catastrophic risks remains, Perrow closes his book with a
recommendation that substitute provisions be made and the high-risk, tightly
coupled system be abandoned.
Perrow's somewhat gloomy assessment contrasts with the La Porte model
of high-reliability organizations, although La Porte’s 1996 article notes no case
where a sloppy organization has later become quality conscious and successfully
implemented a high reliability system (La Porte and Keller, 1996). La Porte’s
work on institutional constancy and trustworthiness is particularly relevant for the
California ISO which began operations March 31st, 1998 and is establishing itself
as a new institution, largely self-defined. Prior to AB 1890, blackouts and
brownouts were rare events owing in part to the activities of the North American
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
76
Electrical Reliability Commission (NERC), established to focus on overall
reliability following the major urban blackouts in the 1960s.
The California ISO’s performance has mixed reviews. Stage I and II
emergencies seeking voluntary and required curtailment of consumption have
been issued with increasing frequency in 1998, 1999, and 2000. The ISO has also
been criticized for expanding its role from grid operator to market maker. In the
fall of 1999 the ISO Board of Governors established a peak price of 750 dollars
per megawatt hour. During the summer of 2000, it was forced by political
pressures to reduce the price three times. The public may wonder about the
integrity and trustworthiness of this new entity. La Porte’s question as to whether
our current political and legal system can sustain the required level of reliability
remains unanswered though certain interested parties would answer with a
resounding “no”. They request federal usurpation of long-term states’ rights and
responsibilities for energy management as a necessary condition for the provision
of the economic benefits of electric system restructuring.
Complexity Theory and Control of Disaggregated Components
In a vertically integrated utility, end-to-end tight coupling and command
and control structures resulted in high reliability. The pre-AB 1890 sale of
generation into the verticals was totally controlled by the utilities. Market
controls are polar opposites of command and control mechanisms; high prices
related to high demand encourages suppliers, while low prices related to low
demand may discourage production. The component systems of the electric
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
77
system were not designed for market driven operations. The adoption of the
market models thus compelled creation of the ISO and the PX as mediating
command and control structures. The requirement that all power be bid into the
PX for sale necessitated a new, if different, command and control structure. The
requirement for uniformly energized system operation combined with the PX
market design lacks the fluidity market systems usually need to thrive. FERC
would remedy this market weakness in its preliminary analysis of the California
market problem by allowing UDCs to buy power from elsewhere. Other solutions
to reduce coupling would include competitive markets. An independent, separate
market was proposed, but regulators and legislators chose to implement a single
market at the PX. This decision derived from an assumption that the best
mechanism for high reliability would be a continuation of the tight coupling in
new operating and market structures.
In theory, restructuring and divestiture of generation will affect the
transmission system, possibly resulting in a need for re-regulation or adaptation in
marginal cost pricing. Marginal pricing allows distribution utilities to pass
through the risk of generation price fluctuation to their customers, eliminating the
utility’s risk. This and other incentives to develop competitive generation, and
increase demand side (DSM) or load management programs, have been tested in
various overseas markets (Cicchetti with Sepetys 1995; Cook 1999).
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
78
Perrow’s work on complex organizations and complex systems stresses
the importance of understanding whether systems are tightly coupled or loosely
coupled (Perrow 1984 and Perrow 1986). In Perrow’s analysis, loose coupling is
essential for complex system robustness and reliability in the absence of strong
command and control circumstances. Perrow illustrates this through analysis of
the space program where many critical systems were tightly coupled. In the
absence of loose coupling, mission control’s authority outweighed even the direct
observation of the astronauts, a problematic management structure, especially
when earth-bound instrumentation produced unfamiliar readings.
Inelasticity of electric demand over the short term and the limited
tolerance of the general population for the negative incentives high prices provide
have sorely tested the reduced regulatory picture. A legislative call for a much
more heavy-handed regulatory response from Sacramento results from this
dilemma.
Although the transmission system is not considered an arena for
restructuring and competition, it will not simply continue as before. The
transmission system, its regulatory controls, and its management face mandatory
and substantive shifts which must be accomplished without disruption to normal
operations. These result from changes occurring both in the deregulated
generation component and the still-regulated monopoly distribution systems.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
79
Despite the potential for competition to reduce inefficiencies, central
dispatch and network operations offer efficiencies that cannot easily be replicated
(Joskow 1997). This explains the significance of the California ISO. Establishing
its perimeter and management systems cost hundreds of millions of dollars, and
management of them introduces ongoing budget requirements. In addition to
these transition startup and ongoing maintenance costs, the separation of
generation from other utility operations has altered the load profile. Generators
get little or no value from energy efficiency and conservation activities triggered
by and methodically put into place since the 1970s OPEC oil embargo. Their
greatest revenue has been seen during summer peaks. The alteration in summer
peak demand has stressed the transmission system; costs for expansion of this
system will be directly borne by ratepayers, as they are fully regulated and the
rate of return on transmission is set by the PUC as part of normal rate cases.
Joskow examines political rent-seeking in the allocation of emission
permits to electric utilities (Joskow 1998). Joskow considers the difference
between the models of emission trading and the much more complex actual
legislative process. In a similar fashion, he examines early load forecasting
models employed by utilities to assess their operating requirements on the day
ahead and hour ahead timeframes; they do not reflect the complexities of the real
day ahead and hour ahead markets. There are several competing actual processes.
California’s system is made more complex due to its properties as a high cost
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
80
state while interconnected to neighboring low cost states. These state-specific
differences extend to peak price caps, power supplies rules and the relative
interest or disinterest in major restructuring activities underway (American Public
Power Association 1998). FERC (2000) describes this as a “balkanization” of
transmission and a contributor to California’s “unjust and unreasonable”
wholesale market. Issues here include the potential for gaming in the nascent
market structures, market power in generation, and unknown changes in ratepayer
behavior once affected by substantially higher bills.
Protecting Stakeholders and Stockholders
The regulated utility industry engaged in many practices based on long
term depreciation, regulatory requirements, and fixed rate of return. Altering this
model to allow competition and new products without unfair damaging existing
providers and their customers became a critical issue for the success of a
restructured electric industry.
Stranded Costs Recovery
Stranded costs recovery was the most strenuously argued issue in the early
framing of electric system restructuring; it is the issue most represented in
academic literature. This well-defined body of literature addresses the negotiated
requirement for restructuring to proceed only with protection of past capital
expenditures by existing monopoly providers. This subject became the most
widely discussed, analyzed, and hotly contested issue of early stage electric
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
81
system restructuring. Its implementation relevance became evident in California
when San Diego Gas and Electric finished its stranded costs capture and was
authorized to pass through the rising price of generation in California to its retail
and commercial customers. The stranded costs argument returned with the 4 to 5
billion dollars in uncollected fuel costs from the summer of 2000, although all
prior discussion was specific to capital, not fuel, expenditures.
In simplest terms, stranded costs have been defined as capital
expenditures made under the assumption that all future costs of repaying these
long-term debts would be captured through standard rate cases. The competitive
operating model sets prices in the industry via free market mechanisms. In a
market scenario early costs incurred under different assumptions and directions of
the CPUC are stranded and would be borne directly by the utility company, its
customers and its shareholders. Premier among these are costs associated with
nuclear power.
Nuclear plants provide the electricity industry with a unique distinction
among deregulating entities given the extremely high costs associated with
initially constructing, currently operating, and ultimately decommissioning these
plants. These costs have been much higher than anticipated, and many of these
plants were built under government encouragement to lessen U.S. dependency on
foreign oil and as environmentally more benign than coal. The negative
implications of paying these costs back directly and expediently could result both
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
82
in the value of utility shares plummeting and in substantial rate increases; this
would be an unpleasant outcome for stockholders and ratepayers alike. Because
of this, economic analysts and regulators jointly developed and recommended rate
mechanisms to recover the bulk of the stranded investments through rate
surcharges that cannot be bypassed by consumers changing energy suppliers.
A corollary to the stranded costs question is an understanding of the
regulatory compact and the agreements inherent between regulators and the
regulated industry. Investor-owned utilities have been guaranteed a reasonable
profit in exchange for provision of service as decreed by the CPUC. Baumol and
Sidak (1995) discuss the need to understand the effects of takings, regulatory risk,
investor goals, consumer interest, and cost mitigation duties in a recovery method.
Sidak and Spulber (1997) continue the discussion by considering the
constitutional limits on the government’s ability to redefine or otherwise
determine how utilities use their own property, and whether this constitutes a
regulatory taking justifying compensation through recovery of stranded costs.
The high emotion and discomfort expressed regarding PG&E’s attempts to
divest its hydroelectric resources exemplify these takings and their economic
implications. Representatives of the CPUC traversed California in 2000 listening
to Californians describe their multiple uses of the watershed involved in the
hydroelectric production. Divestiture would result in the CPUC’s loss of
jurisdiction over this territory, and citizens’ longstanding if often non-contractual
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
83
agreements with PG&E for these non-generating uses of the land and water would
be nullified. The draft EIR produced through the hearing process runs 2000 pages
and examines 19 scenarios, concluding only one-maintenance of the status quo-
truly and fully protects the public value of the watershed involved in PG&E’s
Sierra hydroelectric system.
In practice, a large amount of the unbundled electric bill is not avoidable.
The bill includes the stranded costs charge, transmission expenses and the utility’s
self-determined portion of costs for distribution and metering services. Only a
fraction of the price-about 10%— is allocated to the cost of energy for which
alternative providers can offer discounts. This reality has resulted in less than 1%
of California’s residential electric customers changing their energy provider. This
lack of interest counters the touted hope (Joskow 1997) for electric system
restructuring: reduced electricity costs resulting from a competitive generation
market extend beyond the largest commercial and industrial customers. If
anything, the creation and layering of many non-bypassable surcharges such as
required in California legislation AB 995 and SB 1194, suggests bills will continue
to rise. Each unbundled component has its own supporters continuing to lobby
for its inclusion and increases.
Risk of Cooptation
Selznick’s study of the TV A provides an early and quintessential example
of cooptation, which he defined as “... the process of absorbing new elements
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
84
into the leadership or policy-determining structure of an organization as a means
of averting threats to its stability or existence” (Selznick 1948).
Cooptation may also explain why the California ISO Board of Governors
was established with 29 members. Since the term cooptation has a slightly
negative connotation, this board composition is called instead a stakeholder
process. As one ISO governor has stated, this provides each governor a voice to
shape restructuring to fit his or her organization, and to argue for or against
specific additional charges being added to the transmission tariff.
As stated in the case of the ISO board, cooptation is built into the board
design. The fragmented nature has made the board fractious at times. The ISO
board structure suggests the difficulty of overcoming this driver of restructuring,
the inherent and seemingly inevitable propensity of regulatory bodies such as the
California PUC to be co-opted by the organizations they regulate. The natural
monopoly based on 1910 technologies may be solely an object of political
structure fifty or 100 years later as Hirsh (1999) points out.
Managing the Regulatory Remainder
Once disaggregation begins, many issues develop in regards to the
remainder. The restructured and new functions with reduced or no regulatory
oversight are connected to highly regulated monopoly segments in a strenuous
technical environment requiring maintenance of current at precise levels.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
85
Implementation and Evaluation
The role of implementation research has expanded to give more attention
to improving as well as documenting outcomes (O’Toole, 2000). This focus may
increase the value of policy analysis in the midst of implementation activities.
Certainly, electric system restructuring in California would benefit from this shift.
Ongoing monitoring and oversight are another area of concern for
implementation research. Inadequate controls are in place, as voiced by the heads
of the CPUC and the EOB in an August 2000 letter to Governor Grey Davis. At a
workshop on 2001 Summer Reliability, CPUC Commissioner stated his
frustration over a hundred unanswered subpoenas for information to participants
in the industry. There is a specific cause for this information problem. When the
vertical utilities broke into pieces, information that was part of the public record
on utility operations narrowed and widened simultaneously. Generation
companies are not subject to CPUC oversight, but some generation information is
still required and collected by the CEC and by the NRC for nuclear assets. The
ISO provides online access to much of its hourly pricing and purchasing data.
Privately held distribution companies continue to be regulated, but it is not yet
apparent what information will provide the Commission sufficient understanding
of the operating changes relevant to reviewing the rate cases these companies
propose and the special assessments they request. Public utilities continue to run
as before, but they have requested to conduct more business privately, particularly
negotiated prices and services for large customers. After its initial
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
86
implementation, the California ISO sought release from open meeting laws and
other constraints faced by state agencies. O’Toole’s broad body of work on
implementation and implementation research includes reminders that
implementation reflects the interaction of many actors (O’Toole 1996).
Pressman and Wildavsky’s Implementation (1973) is the seminal text on
implementation and bureaucratic effects in California. Their evaluation of the
results of a federal grant from the Economic Development Administration to the
city of Oakland parallels on a small scale the initiation of restructuring in the
state. The project with its vision of reducing poverty and unemployment through
providing capital to businesses failed to deliver these results. Similarly, beyond
the short term energy emergencies, the medium range expectations are downbeat,
centering on continued conflict between communities, state air quality
management districts, and the utility structures. The prognosis is for increased
bills statewide in stark contrast to the legislated reductions.
The discussion above on how to implement restructuring included the
issues of the ISO’s establishment and support for the stranded costs recovery.
FERC 2000 requires review and submission of regional transmission
organizations; in November, 2000, FERC restated that the western grid’s
balkanization and lack of a true ISO contributed to the previous summer’s
problems. The continuing monopoly of the transmission system-albeit a new
institutional player-brings up the question of whether it is inappropriate for this
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
87
monopoly to unilaterally buy power to meet the mandated power reserves when
they will pass the cost directly on to all scheduling coordinators. The
transmission system has already changed radically and will continue to do so in a
geographic manner that crosses state boundaries. However, the electric industry
as a whole is not a strictly federalized structure.
Transition Effects
Transition problems and effects suggest continued information availability
is critical to maintaining or enhancing reliability. Taylor (1999), and Grudinin
and Roytelman (1997) describe 1996’s large scale outages, and the lessons
learned and actions taken to reduce risk of re-occurrence. The transmission and
distribution systems have been radically altered, they note, both in their
institutional supports and in the technologies applied to manage and maintain
reliable power in the wake of widespread regulatory changes. These authors’
diagnostics are technical, but they mention the institutional support required to
create these systems. Taylor also provides a sidebar on the NERC response which
requires additional investments introducing new technical complexities. These
NERC actions are taken in relative isolation from local political processes, and
they are separate from activities initiated at the state legislature or at the CPUC.
Restructuring an industry is a gradual process, whether the end result is
total deregulation or partial reduction in regulatory oversight. Many effects are
sure to be transitory. Winston (1998) describes the transition effects of airline
deregulation as continuing twenty years into the process. He acknowledges his
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
88
analysis of empirical restructuring effects fails to address the counterfactual of
changes that would have happened regardless of deregulation, but he argues that
the simplified analysis of price changes over time is drawn from other studies of
deregulation. A complete counterfactual would estimate price, costs, and changes
that would not have occurred outside restructuring. This omission, plus his
expectation that deregulation could work for all components of electric services,
weakens his argument for the economic validity of the electric system
deregulatory reforms as initially operationalized. Legislation extending the
collection of surcharges for renewables until 2012, or the proposal from the IOUs
that the extra fuel costs from the summer of 2000 be collected as a five-year bill
surcharge, exemplify problems inherent in an only slightly deregulated
restructuring model. Technological distinctions in electricity and the continuing
regulation of systems tightly coupled with the restructured segments make
electricity restructuring dissimilar to other examples.
Municipal power has operated with different monetary objectives than
IOU’s. Public sector utilities are structured to cover the cost of service, not to
meet shareholder expectations of return on investment. This difference and the
ability of government owned agencies to issue tax-exempt bonds have historically
resulted in an approximate 15% price advantage for consumers. Any amount
collected over the cost of service may be set aside to hedge potential rate
increases, mitigate contract cost fluctuations, or, in the case of integrated public
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
89
utilities, may be sent into a city or county’s general fund. These distinctions have
contributed to the persistence of publicly owned power agencies.
In October 2000, CEC Commissioner Art Rosenfeld discussed ISO daily
demand from October 1999 through September 2000, and noted an immediate
need for approximately 2,000 megawatts to meet summer 2001 demand (CEC,
2000b). In his opinion, the largest rapidly available source for this quantity of
energy comes from a 4-degree reduction in indoor temperatures (“no swear, no
sweat”), accompanied by a 30% diminution of lighting. Commercial and
residential air conditioning, and commercial lighting account for 41% of total
load; this solution would not require costly augmentations to the transmission and
distribution of system that additional central generation necessitates. In other
words, while electric demand over the short term is inelastic, over the longer term
it is “perfectly elastic.” (CEC, 2000b.) Distributed generation shares the limited
transmission system augmentation value, but UDCs are reluctant to support this
technical approach.
Externalities
Economic theory defines externalities as costs or benefits applied outside
the organization which initiates these impacts (Weimer and Vining 1992). In the
case of electric system restructuring, several negative externalities are evident.
These externalities include environmental impacts of increased generation and
environmental impacts and loss of property rights for corollary increases in
transmission systems. Externalities from increased generation can be avoided by
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
90
siting generation directly adjacent to underpowered distribution systems, but
historically there has been substantial local and regulatory opposition to this
approach. These siting decisions also add to the perceptions of certain classes of
customers that they are unequally and unfairly allocated their distribution of the
electric restructuring costs. Positive externalities potential also exists, such as
additional steam availability for industrial processes as more medium-sized
cogeneration facilities are developed and new products and services are
introduced.
Externalities can be a source of market failures, and of “nonmarket” or
government failures (Wolf 1979). Combining the ideas of market failure and the
nonmarket or government failure can provide a better explanation of the sources
and solutions to the market failures possibly associated with electric system
restructuring. For example, maintaining adequate voltage in the grid-reliability
at all times and all costs-creates an ISO-created charge that is not paid by the
ISO. This charge is billed directly to the schedule coordinators. Non-bypassable
surcharges used for measuring and managing the grid are also dissimilarly
allocated to grid users and end customers as some of the state’s largest electric
users have contracted or regulated fixed rate power.
In a restructured energy system, the management of externalities falls
upon the regulatory agencies. In California’s case this would be the CPUC,
although the commission may attempt to pass politically charged issues onward to
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
91
the Federal regulatory agency, FERC. An example would be the difficult issue of
who was to pay for the high cost of energy in the summer of 2000. FERC
indicated that their role did not include resolving this issue.
Traditional externalities theory expands beyond immediate changes to
corollary effects. Mack et al. (1998) note the EPA is investigating coal as it
considers the externalities deriving from the likelihood of much broader use of
this fuel due to electric system restructuring. In many regions, coal is cheaply
available and utilities have sufficient emission credits; this makes increased acid
rain in the East and particulate emissions in the West unintended but recognized
consequences of restructured electric systems. A positive externality potential
also exists. Diversification of fuel sources again rises to strategic interest, and
California’s strict environmental rules may spur new advances in the technologies
of coal plant generation. The California Energy Commission lists 1,012
generation plants in the state, only 15 use coal for fuel. If emission constraints
make coal use in California unfeasible, generation based on this inexpensive fuel
will be sited outside the state to sell into the California market, similar to the
present generation in the Four Comers area. The necessity to enhance the
interstate transmission system, combined with limited California regulatory or
legislative control over out-of-state plants, would introduce further difficulties
with inter-agency management.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
92
Public Administration Perspectives
Joskow and Schmalensee (1983) explain that the economic efficiency of
restructuring takes precedence in the discussion over measurability and agreed
upon value. However, they acknowledge that this is a simplification. Looking at
the issue from a public administration perspective provides a degree of insight
into why the first few years have turned up some unimaginably severe
unanticipated effects. While there are always unintended side effects from public
policy actions, the severity and rapidity of the summer of 2000 problems caused
general shock and a widespread call for meaningful alterations in the restructuring
plan.
Public administrators and policy makers have a specific perspective.
Although not economists or engineers, they provide guidance and management
for agencies and communities that are greatly affected by this latest regulatory
change. Administrators and policy makers must rely on specialists and experts,
while having sufficient understanding and judgment to move beyond facts and
findings to evaluate whether these facts warrant proposed actions. Additionally,
public administrators must balance their administrative and political
responsibilities, and temper them with their commitment to the responsibilities
and obligations of public service. Public administrators also are called upon for
assistance in establishing, revising, and monitoring new institutions. For these
actors, an abbreviated analysis provides information to apply to their operations.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
93
While questions, opinions, and preliminary rulings abound, a small set of
core issues are presented to public administrators and policy makers. These
include protection of public good initiatives and associated energy efficiency
expenditures; investigation of and protection against market power; caution in the
use of analogies to restructuring predecessors, and selecting actions with
reference to the big questions of public administration.
Public Good
Friedman (1984) discusses the effect of a public good requirement on
markets in an imperfect world. In his discussion he considers the achievement of
a Pareto-optimal solution unlikely; one or more of the necessary conditions will
usually be unmet. As a result he focuses on second-best solutions and notes that,
with policy intervention, the second-best solution is also not found. The case of
mandatory rate reduction exemplifies his analysis. First, the efficient market is
not available so a proxy is constructed in the 10% general reduction. IOUs
convince legislators that the utilities cannot meet this reduction through altered
business practices or other conventional measures. The legislature authorizes the
utilities through AB 1890 to fund this reduction by floating bonds, which must be
repaid with interest by those same ratepayers through a non-bypassable surcharge.
Altering the structures for the creation, transmission, and distribution of
electric power creates issues. All parties involved in AB 1890-which passed the
California State Assembly 80 to 0-expected electric system restructuring to offer
benefits. But the timing and distribution of benefits were anticipated to spread
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
94
dissimilarly among customer classes. This brought out the issue of achieving
equity. Young (1994) systematically describes the strengths and weaknesses of
assorted methods for dividing and sharing goods. These go beyond the primacy
of efficiency as a goal of reduced regulation to the issues of equality, priority, and
proportionality. Lacking internal or public discourse on Nash equilibrium or
Pareto optimal solutions, the California model rests on stakeholder boards at the
ISO and PX. FERC has suggested abolishing these boards, ostensibly designed so
all participants would have a say, because of the difficulty they have had in
reaching decisions (Federal Energy Regulatory Commission, 2000).
Section 30 of the Warren-Alquist Act (California Energy Commission,
2000)-the section added by AB 1890-states baldly that the ratepayer will receive
a 10% immediate benefit through a reduction in the rates, and a 20% reduction
after 4 years of collecting the competitive transition charge (CTC). This effort to
explicitly define and capture the public economic value, in retrospect, seems ill-
advised. Equity in rate reductions does not fit a classic definition of a public
good; nevertheless, elected officials and regulators consider part of their decision
making to be driven by attempts to achieve the best result for the public at large.
The transmission grid, properly energized, might be considered a public good as
well.
The 20% rate decrease by 2002 seems unlikely as the first utility to recoup
the CTC tripled its peak rates in the summer of 2000, based on the cost of
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
95
generation. Political response to public outcry resulted in restoring the bundled
rate in San Diego. The difference between the residential rate structure and the
purchased price of power caused a financial crisis for the IOUs, and the money
spent first by the IOUs and later by the state of California requires repayment.
Ideas in early 2001 for recouping the uncollected cost of generation include a 10-
year additional surcharge, a multi-billion contribution from the state, or major rate
hikes. There does not appear to be a method to return to the previous status quo;
any newly-restructured rate will be substantially higher than the 1996 rate in
effect prior to the enactment of AB 1890.
M arket Power
Preventing the establishment of market power was a major concern in the
design of electric system restructuring in California. Market power refers to the
ability of a supplier, or set of suppliers, to control the quantity sold and establish
terms of trade most beneficial to themselves. Empirical analysis has suggested
that market power could be a problem in California for its super-peak hours, the
few daily peak hours during the summer months, about 200 hours per year overall
(Borenstein and Bushnell 1998; and Borenstein, Bushnell and Wolak 2000),
though energy prices remained much higher post-summer 2000 than past years,
persisting well beyond the traditional summer peaks. FERC in its November 1,
2000 finding did not agree with this assessment of the existence of market
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
96
-1A
power. The question of market power in the California design of partial
restructuring is not settled.
Disinterest in switching energy suppliers provides one indicator of market
power early in the restructuring timeframe. In development of a mechanism to
allow for the collection of stranded investments the California PUC has set up a
structure that gives retail customers virtually no savings in switching to a cheaper
supplier of electricity. This problem results from high total electric bills, and only
a minor portion allocated to supplied generation. New entrants in retail energy
supply, conversely, have found it difficult if not impossible to buy and resell
energy profitably into the retail market. The high cost of generation often has
exceeded the price offered to retail customers. Most of these businesses have left
California’s restructured electric market.
Evidence of market power is found in the new disaggregated bills.
California’s IOU customers contain separate charges for transmission,
distribution, the energy itself, and a competitive transition charge (CTC) based on
the amount of money to be collected against sunk costs between 1997 and 2002.
Residential customers also pay an additional charge to service rate-reduction
bonds floated in 1997 to grant them the mandated 10 percent rate reduction.
Several other surcharges have been proposed suggesting persistent efforts to
2 0 Chairman James J. Hoecker instead called the market structure “seriously flawed” and sought
among remedies a return to 95% of transactions in the day-ahead market to reduce chronic
underscheduling. Instead, ever increasing amounts were still being purchased at the ISO in
January 2001. See FERC, “Seriously Flawed”, Docket EL00-95-000, et al., November 1, 2000.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
97
manipulate the state legislature in hopes of protecting vested special interests.
State Senator Debra Bowen, chair of the Senate Committee on Utilities, has
acknowledged this problem, declaring that restructuring decisions should be
managed within the CPUC framework (Nemec, 2000). The energy charge itself
in 1998 was allocated at only about two cents per kilowatt-hour. New suppliers
could offer a 50 percent reduction in energy costs, translating to an approximate
5 percent reduction in an average residential bill. The actual offerings by power
marketers ranged from 0.5 percent to 2.5 percent (Wiser et al. 1998). For
suppliers, the minimal amount of deregulated space in the rate leaves little room
for profit; most are no longer signing up retail customers. All this preceded the
summer of 2000 wholesale price problems which resulted in 4 to 5 billion dollars
in energy charges not collected from ratepayers.
Market power can possibly develop if there are insufficient sellers and
product in the market. High energy costs first appeared during the 1998 summer
peaks in the Midwest when power costs spiked at $7,000 per megawatt hour
(Kranhold 1998). California’s power spikes resulted in the ISO capping prices
first at $750 per megawatt hour, then downward to $500, $250, and even as low
as $60 per megawatt hour though this latter price was set aside by FERC.
Borenstein and Bushnell (1998) published economic models and evidence of
market power potential, they also predicted the lowering of prices paid to
qualifying facilities (QFs). The use of transmission line charges to limit new
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
98
entrants has been noted as well (American Public Power Association 1999) and
has been called “substantial” (Borenstein, private correspondence). These
charges extend to qualifying facilities in contract renewal which has caused some
to consider ending their participation in the marketplace.2 1
Use of Analogies
Analogies provide a convenient method to simplify and accelerate
discussion on complex issues, through comparison with other issues with which
discussants are more familiar.
Considerable disagreement exists about the degree of similarity the
electric system has with other industries that have undergone deregulation or
restructuring. The differences include issues of state vs. federal jurisdiction, the
role of externalities, the possibility of market power within deregulated segments,
and the validity of policy making by analogies to other industries.
Using analogies improperly can lead to erroneous estimates of effects, and
to suggestions of parallels that may not exist. As Neustadt and May (1986)
discuss, reasoning from analogous thinking can be useful or dangerous or both at
once. There are a few similarities, but many differences between airline
deregulation and energy restructuring, for instance. Savage (1997) describes the
effect of deregulation on reliability in the airline industry. He reviews forty years
of safety performance (accidents, fatalities, near misses and reportable incidents),
2 1 One survey participant acknowledged these price changes resulted in his facility’s non-renewal
of the QF contract.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
99
mapped against total departures and passenger miles by carrier type. Savage
estimates fatalities to be flat while passenger miles have steadily increased; total
customers have also increased. Savage calculates that the few additional fatalities
will have offset hundreds of auto fatalities, had people driven rather than flown.
Comparatively, reliability in energy restructuring is equally significant and may
follow a similarly smooth course. In contrast, demand for flying went up post
deregulation, as ticket prices fell; this differs from the early restructuring behavior
of energy prices, which increased as demand for energy increased. On commuter
routes, the increased traffic resulted in up-sizing of aircraft, which benefited
reliability since the larger craft have better safety profiles. In restructuring, much
discussion centers on down-sizing power plants, and localizing them in areas most
prone to reliability problems. Savage also cites Alfred Kahn, who argued that
quality is overprovided or produced inefficiently in regulated markets;
restructuring electricity may differ in that the demand for power quality has never
been higher with the increasing dependence of sensitive electronic equipment.
The likelihood of misguided insight from analogies with other industries
was discussed by Joskow and Schmalensee in 1983. The technological, political,
and physical specifics in electric systems suggested that attempts to draw insights
by means of analogy would be weakened by the likelihood of erroneous aspects
of the comparison. For instance, electric system capital is “almost perfectly
immobile” (p. 8), unlike airlines which have extreme capital mobility and are
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
100
limited only by their ability to contract gate space at airports. This one mobility
limitation, it might be noted, has resulted in the large increase of utilization of
slightly off-market airport locations, e.g., Colorado Springs for Denver, Baltimore
for Washington D.C. These are economic substitutions, not perfect, but adequate
for market entry.
Analogy to telecommunications deregulation is common. The image of
energy telemarketers calling during dinner to ask if the consumer would like to
change their electric service is borrowed directly from telecomm telemarketing,
and draws nervous laughter-no one is certain whether or not that image has any
validity for this market.
Equipment comparisons between the meter and the telephone are
common. With telephone deregulation, customers were offered the opportunity to
buy their Bell equipment, continue to lease it from the phone company, or replace
it with a new phone purchased through standard retail channels. Strict analogy
suggests having the consumer determine whether to incur costs for a new meter
and the selection of features beyond basic metering. The risks inherent to analogy
suggest further exploration of the supposed similarities between these equipment
types that bridge providers and consumers. The conscious and subconscious
assumptions and suppositions of regulators affect the decisions they make. These
assumptions need to be understood and discussed publicly before these regulatory
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
101
bodies decide to grant 10 or 20 years of exclusive rights to energy customers’
meters as part of restructuring activities.
Regional transmission organizations have also been called power pools,
since they give utilities an opportunity to share costs and resources that contribute
to reliability more efficiently than if each managed their own standby systems.
The notion of a pool can be misleading, however. Kuhn et al. (1996) suggest the
analogy of a lake with many inlets and outlets for the electric grid, but this is
overly simplistic. Kuhn et al. argue that transmission accounting unnecessarily
complicates energy management and raises the costs of wholesale and retail
wheeling; they suggest the grid be considered a pool resource more efficiently
managed as an institutional common as developed in Hall 1998. These authors’
argument denies the right of transmission ownership; this violates their property
rights. Buchanan (1993) calls these rights the basic guarantor of liberty. The
analogy also fails based on the actual, physical nature of the electric grid. Each
part must be energized to a certain level, no higher or lower; transmission loss
over distance is real and non-trivial. The existing generation system cannot
simply pour all the energy from western Canadian hydropower, for instance,
through the grid and instantly provide stable power everywhere in the United
States. Some transmission experts consider over-dependence upon and over
subscription to hydroelectric power from British Columbia among contributing
factors to the August 10,1996 blackout.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
102
There are similarities with the automotive industry. Meyers et al. (1997)
in making a case for recovery of stranded costs, compares the electric industry
with this industry. Unfortunately, their conclusion that import controls were
useful in redistributing the costs of transforming the automotive industry contrasts
with the academic analysis of Gomez-Ibanez et al.(1983), who argue that the
import controls caused long term loss to all principal interests, a virtual reverse-
Pareto efficiency where everyone is made worse off.
Subsidies and the telecommunication industries are the topic in Britman et
al. (1998). This study analyzes targeted and untargeted subsidies for their
effectiveness in achieving the normative policy goals of deregulating the
telecommunication infrastructure, in particular the promotion of universal
telephone service. These authors note negative effects of particular financing
mechanisms, and the higher rate of effectiveness from targeted subsidies over
untargeted. Any universal access service runs counter to the normative economic
theory that prices should reflect the economic cost of providing the service.
However, serving lower income customers is an assumed requirement to the
holders of telecom service licenses. Electricity is also a utility with a universal
access value. As California’s use of the competitive transition charge (CTC)
comes to its originally legislated end date, and as much of the loosely-structured
public goods portion of the rate goes to contract programs, the obligation and cost
for universal service is likely to be re-addressed. Potentially, this could result in
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
103
another increase in rates. Indeed, Winston states that “policies to preserve
universal service have a tendency to undermine the competitive forces that can
enable consumers to obtain reasonable prices and services” (Winston 1998,107).
Applying simple analogies and developing national policy based on the
attractiveness of the analogy is hazardous. Schlager and Blomquist (1996)
discuss the more complex theoretical models of policy making currently in use:
Sabatier’s advocacy coalitions framework, institutional rational choice, and Moe's
politics of structural choice. Examining California’s electric restructuring
decision making in this theoretical context provides more explanatory power and
better predicts the constraints and factors necessary to introduce more economic
efficiency into the industry.
Kessler (1998) and Schlee (1996) provide a disquieting analysis of the
strategic value to policy makers in not being well-informed and knowledgeable.
This theory explains the reliance on and usefulness of Assemblyman Steve
Peace’s advocacy for the particular design of AB 1890. The voting public
directly suffered from the summer 2000 generation costs passed on in San Diego,
so several California legislators faced substantial opposition during the election
cycle. This resulted in the rapidly constructed late session legislation to again
freeze San Diego’s rates; the threatened incumbents retained their seats in the
state legislature. Senator Peace’s own political future is clouded, though he now
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
104
denies as a “myth of deregulation”2 2 the widely held perception that he was the
architect of AB 1890.
“Big Questions” Issues
Kirlin (1996) develops seven big questions for public administration in a
democracy, separate from issues in public management. These involve tools of
collective action, tradeoffs between designs based on function or geography,
national versus local political arenas, the potential and timing to isolate decisions
from political processes, balancing competence with representation and
leadership, and societal learning. Some issues and discussions in electric system
deregulation fall into these categories.
This issue of place, and the inteij urisdictional nature of electric
restructuring, is relevant. In the late 1980s and early 1990s, electric rates became
one element of the bargaining packages communities would assemble as part of
enticements for large companies siting new facilities. Certain commercial
ventures would consider relocation from high electric rate areas to low rate areas;
White (1996) notes that states with low energy prices have a much reduced desire
to see restructuring legislation passed and policy set on a federal basis. This
distinction directly affects the introduction of an energy market in California. The
western states grid is interconnected, but the ISO remains state-specific unlike the
New England RTO. State PUCs and PSCs direct the precise restructuring models
2 2 Senator Peace placed a six-minute video explanation of his role in the development of AB 1890
on his California State Senate web page. See at http://www.ca.senate.gov, January 31, 2001.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
105
which differ state to state. These efforts intend to protect the low cost states but
limit the California energy market thus increasing the potential for market
weakness.
Evaluating restructuring from a different rational than political or
economic situations is especially important for city utility departments. The
ability of electric power service to enhance the general funds has been a boon for
city or county governments with electric utility departments. This potential
revenue stream has in cases resulted in routine municipal dependence on the
utility to enhance general funds not directly tied to electric system costs (Hildreth
and McMillan 1998). An electric utility department within an existing municipal
structure represents a local, unrestricted revenue source now destabilized by
restructuring. Unlike the pricing example of airlines cited by Winston, utilities in
California have had limited price offers. PX rules in place since AB 1890 set all
loads at the highest accepted market clearing bid. There have been modifications
to this rule, but prices remain high. This can and has benefited municipal
utilities able to sell excess generation, but few have this capacity. Instead, the
entire municipal operation must be adjusted to the new electric marketplace.
Concluding this chapter with Kirlin’s big questions leads to an
investigation on electric system restructuring from the public administrative
perspective. The next chapter presents the research design to collect empirical
evidence on restructuring.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
106
Chapter 4
Research Design
Introduction
This chapter describes the research design, methodology, analytical
approach, and data sets in detail. It explains the types of reliability measures that
are available and the relevance of categorizing reliability events. The chapter
reviews the institutional changes and testable hypotheses and maps them to the
data collection.
The research design includes both a quantitative and a qualitative
component. There are inherent difficulties in interpreting a meta-data collection
and analysis. The qualitative assessment is designed to assist in interpreting
results, add depth, and possibly increase the predictive value of the findings.
Electric system restructuring is a multi-disciplinary process, involving technical
specialists, public and private utility managers, regulators, politicians, economists,
and the general public. The research connects the quantitative data analysis with
the experiences and interpretation of events by participants in the process. The
method by which this is attempted is discussed in this chapter.
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
107
Goal of the Research and Resultant Design
The research evaluates restructuring results while acknowledging
difficulty in separating incipient long term effects from transitory effects
associated with transition activities. Political and fiscal disarray in the summer of
2000 associated with wholesale generation and SDG&E’s end customer pricing
problems added complexity to the data collection and to determining transitory
effects. Even in the best of circumstances determining which effects are transitory
is not minor. Using participant surveying to supplement technical reports and
empirical evidence has assisted in deciding if a particular effect is largely an
irrelevant transition impact. However, changes to the restructuring plans that
have already been instituted and more radical shifts under discussion also
introduce new effects and new transitions. Additionally, it is widely understood
that the 2000 market pricing problem will have medium to long term
consequences, both in changes to the market design and in mandatory residential
bill increases to pay for fuel purchased in the summer and fall of 2000.
Regardless of the transition issue, this research retains relevance in its
compilation of empirical material and analysis thereof. The application of
restructuring processes to California’s electric industry has challenged pure
market theorists and lawmakers alike. The research may show the naivete
associated with legislating a market design and decreeing the precise alteration in
consumer bills. It also suggests that evaluating excess capacity under regulatory
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
108
models does not ensure adequate capacity for market function. Particularly,
comparison between municipal utilities and IOUs may shed light on the feasibility
of government ownership ("municipalization") to resolve consumer anger over
the current situation, an approach being investigated by such northern California
cities as Berkeley and Davis.
Despite pre-restructuring plans, many unexpected results have occurred.
New participants in California energy markets purchased hydroelectric and
thermal plants at valuations initially underestimated by 50% or more. Financial
assessors attempting to assist producers in establishing values for these resources
had minimal experience in such assessments, causing loss of confidence in the
valuations. These selling prices had a ripple effect as participants attempted to
adjust the AB 1890 formula to compensate for this unanticipated increase in
value. The inaccurate estimate of the stranded costs associated with these
facilities accelerated the collection of the legislated competitive transition charge.
This amount was over-collected and rapidly captured in San Diego,
foreshortening the four-year transition period by half, while the new institutional
structures, the ISO and PX, lacked the longer term operating history to manage
their roles without major disruptions to the new energy market model.
The especially high valuations of hydro facilities may have reflected
bidders’ detailed comprehension of the unique properties of hydroelectric
generation or their estimate of the anticipated demand for generation after
Reproduced with permission of the copyright owner. Further reproduction prohibited without permission.
109
divestiture. This anticipation seems well-placed in reviewing production and
profits for 2000; Enron’s revenues for the fourth quarter rose four-fold, from $10
billion in 1999 to $40 billion in 2000. Balancing generation with demand
displays the different properties in market based operations over traditional
vertical integration. Energy analysts and managers focused on providing the
precisely correct quantity of electricity in a multi-producer market describe the
rationale for these valuations. Hydroelectric resources contribute only 15% of
California’s energy, but comprise half of California’s ancillary or standby
capacity for the summer peak.
Misunderstanding the underlying rationale for the increased values, one
state senator suggested in an early 1998 hearing that the rising bids demonstrated
the so-called greater fool theory of buying and selling. Among other arguably far
fetched ideas this same state senator has since suggested are state takeover of all
generation-state municipalization-and consumer refusal to pay the utility bills.
Meanwhile, a thousand-page draft environmental impact report on PG&E’s
hydroelectric divestiture recommends continued status quo operation, warning of
several environmental externalities should the divestiture remove the watershed
from regulatory oversight.
The empirical evidence collected, analyzed and evaluated in this
dissertation includes statistics and reporting on outages and electrical emergencies
through the 1990s; load profiles pre- and post- AB 1890 restructuring, and
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
110
discussions on the new legislation for California’s public and private utilities.
The data sources are catalogued in Appendix C. A summary of the data’s
relationship to the hypotheses is shown in Table 5 found in the hypotheses section
below.
Incomplete Knowledge and Policy Making
Collecting available data and analyzing it is important. Publicly available
information on electric system restructuring is limited, as noted by Joskow and
Kahn in their study of possible application of market power in the summer of
2000:
Only FERC can gain access to the kinds of confidential data necessary to
fully understand the behavior and performance of wholesale markets in
California, to quantify the magnitude of departures from competitive
pricing, and to identify the behavior leading to increased prices. (Joskow
and Kahn, 2000, p.3)
Acknowledged, the complete knowledge base for this research question is
not available; the material herein represents many months of inquiry, however,
and can provide insight into an empirical evaluation of the hypotheses below.
Majone (1989) discusses the non-neutrality of policy instruments that have
been applied to manage failures in distributions of effects. He dichotomizes
Ouchi’s two parameters for evaluation, measurability of outcome and knowledge
of the process that generates the outcome, as shown in Table 4, below.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
I l l
Knowledge of Process
complete
incomplete
Measurability
Of
high Evaluation by process or by outcome evaluation by
outcome
Outcome low Evaluation by process evaluation by
input
Table 4: Majone’s Model for Process Evaluation
Reviewing Majone’s model, it can be seen that in the case of electric
restructuring, the preferred evaluation method for system reliability at an early
stage with incomplete knowledge of process is evaluation by outcome. This
method requires collecting information on planned actions and regulation,
processes and implementation, and expected and actual outcomes. Available
information provides material for quantitative assessment. These are the principal
data applied herein.
Some process variables are emerging, suggesting the beginning of more
complete knowledge. The available information base not only differs according to
access to confidential or proprietary information, it also differs between the
private IOUs and the public municipal firms, as they operate under alternative
institutional rules. Also, even where required, the information provided by firms
has not been uniformly supplied to regulatory commissions. These distinctions
are stated where they exist. The processes are also shifting in response to the
significant problems with the initial design; radical alteration of the ISO board of
governors and a FERC recommendation that the Electricity Oversight Board be
disbanded exemplify these process changes.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
1 1 2
Sometimes information appears in isolation, peripheral from direct
reporting requirement or requests. For instance, in late 1999 the California Senate
Committee met to discuss AB 811, which sought support to add a charge to a
customer class as part of a rate case review. Representatives of two of
California’s major IOU’s asserted that between one-quarter and one million
dollars would be sufficient to alter the customer billing system to support direct
access, making the assessment charge supported through the bill appear
negligible. Not discussed in committee was the typical cost of utility customer
billing systems, which run in the tens of millions. Over $100 million was spent
by PG&E before a proposed IBM system was abandoned (Moad 1997). SMUD
has spent over $60 million on an SAP implementation that included new customer
service software. Changing computer billing systems to support market flexibility
is not a trivial matter. The AB 811 request for funding to change computer utility
billing systems to support restructured operations is one example outcome to
evaluate; this contribution can be included even though the restructuring process
itself and the knowledge of that process is incomplete. Given the difficulty of
direct information collection, these peripheral information sources increase in
importance.
As covered in chapter three, several literature topic areas related to
restructuring provide the grounding for this particular study. Work has been done
in the United States predicting and reviewing restructuring efforts in several
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
113
industries. This research effort continues to compare U.S. restructuring to
research on Britain and elsewhere. This analysis of the transmission system also
satisfies La Porte’s call for case studies of high reliability systems. Particular
aspects of interest in the U.S., stranded investments for instance, have been
discussed at length. This discussion has shaped California’s electric system
restructuring design, wherein the CPUC and the Legislature have allowed the
recovery of these so-called stranded costs. Postulated effects of electric system
restructuring on transmission reliability are being examined. Government and
regulatory circles discuss analyses of proposed rate increases and outages,
generation imbalance and other shortfalls. This study builds upon these economic
and public management theories and analyses.
Description of Approach
In addition to theory building, a practical value of an early implementation
analysis derives from the ability to use findings to improve the quality of the
implementation. The difficulty of an early implementation analysis derives from
limited empirical results. To capture the value of this study while minimizing the
difficulty associated with a preliminary assessment, the approach taken involves
principally quantitative assessment, supplemented by qualitative information.
This is a meta-data, non-experimental design, using standard econometric
statistical methodology for the quantitative portion of the analysis. The
quantitative data on energy consumption are event-driven; there is no pure control
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
114
group of non-participants in the process, especially given the interconnected grid
structure. In lieu of a strict control, parts of the restructuring process apply only
to the IOUs and in these areas, public utilities serve as a non-participating control.
Additionally, SDG&E accelerated its capture of the competitive transition charge
and changed its rate structure, providing another distinction.
For the years preceding and following the passage of AB 1890, baseline
data exist on power contracts, power consumption, as well as outages and load
disturbance incidents. Major and minor outages, as well as load disturbance and
declared emergencies comprise the category of reliability events studied.
Weather throughout the decade is also considered as it is often a factor in
reliability events. Interviews with stakeholders and policymakers provided
supplemental qualitative information.
Relevant material includes published research on aspects of electric
restructuring, theoretical analyses of projected economic value to be derived from
electric system restructuring, and impact analyses completed on implementations
to date. In addition to the academic literature that provides the research dialog,
material produced by the California PUC, the CEC, state and federal agencies has
been considered. The U.S. Department of Energy’s Power Outage Study Team
(POST) investigated similar eastern U.S. reliability disturbances; this study of
western events references their categorization of issues, particularly in the
qualitative survey.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
115
The data on outages that each utility makes public have been brought
together for this investigation, and evaluated with a monthly unit of measure for
as many of the years in the 1990s as available data have permitted. The research
design assesses outages for several years prior and post passage of AB 1890 in
1996 and the operational start of the ISO in 1998. This extended period was
necessary because utilities anticipated restructuring and each developed its unique
operational and strategic response. Logit analysis was performed to determine
which fundamental environmental conditions or restructuring actions and
responses are most associated with the reliability events. Based on the findings,
the analysis suggests a conceptual framework to best fit restructuring intentions
and actions; this is presented following the findings in chapter 6. Through
investigation of publicly-stated positions held by parties involved in the politics,
policy making, operational, and other aspects of the industry, the findings and the
framework provide implementation evidence relevant for further alterations in the
restructuring structure.
The first chapter describes five changes in institutional arrangements
supporting the new regulatory processes:
• Power purchasing through PX daily and ISO ancillary markets
• Divestiture of generation; assumption of power availability
• Alteration of demand side management (DSM) practices
• ISO manages transmission system; many contributors to grid via PX bid
process.
• Large capital investments in non-core operations such as customer
information systems (CIS) and automated meter reading (AMR)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
116
These are postulated to have collateral effects increasing disruptions. The
investigation of the research question and specifically hypotheses HI, H2, H3,
and H4 examine data that show evidence of system reliability impacts from these
changes.
Restructuring
In searching for an answer to the research question, three categories of actions and
effects have been investigated:
• cost and profit shifting between regulated and non-regulated electric
system components
• added complexity in the end-to-end production and delivery of electricity
• altered institutions and regulations
Each of these categorical changes could hamper or help deliver the stated goal of
electric system restructuring: reduced cost of electricity to its consumers. Each
may have negative unintended consequences.
Many issues faced in a large democracy involve complex solutions and
complex organizations. Public policy makers have a knowledge base available to
aid in the understanding of their own organizations and of the processes of
developing solutions that are acceptable to their constituencies (Perrow (1986),
Yankelovich (1991)).
Unintended consequences follow from most actions, so the presence of
such consequences is neither surprising nor especially interesting. Predicting
consequences and providing institutional or regulatory support to enhance or
mitigate such consequence as need be are matters of considerable interest
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
117
especially in cases such as this, where the inputs for reliable grid operation have
many publicly significant externalities. These predictions and responses may also
be useful for refining underlying economic theory and models for restructuring
electric service delivery.
It is important to note that the fundamental economic intention of
regulatory reform is to reduce the bundled cost of electricity to consumers. This
intention served as the foundation for large customers of the monopolistic
environment to request electric system restructuring, in an effort to stimulate
better cost management by producers and to provide competitive price options for
consumers. An electric bill that in its totality remains flat or rises does not
indicate a drop in either production costs or consumer prices. This point is often
obscured in the creation of unbundled billing. Unbundling has initiated a half-
dozen separately labeled cost components. New agents are assigned rights to
assess component charges. Limiting restructuring to the introduction of a
competitive market for bulk generation, and seeing a bill component with that
label drop in price, neither validates the existence of competitive electric services
nor assures a reduction in bundled costs.
Selection of Subjects and Instrumentation
The quantitative study is based on public information disclosed in the
process of legislation, regulation, and public disclosure rules. These data are
cataloged in Appendix C.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
118
In addition to the reliability events data, this data set includes some
information on the structure and costs associated with the establishment of the
California Independent System Operator, responsible for the management of the
state’s transmission system, and of the Power Exchange, responsible for the
selling and buying of electricity on the day-ahead market.
Direct evidence of changes in operations and in energy demands has been
evaluated with additional interpretation from management, especially in regards
to determining the counterfactual case, in light of pre- and post- AB 1890
predictions. Inclusion of SMUD and LADWP provides contrast because as public
utilities like the District, and the City of Los Angeles’ Department of Water and
Power are not bound directly to AB 1890. Additionally, SMUD shut down its
nuclear reactor at Rancho Seco in 1989 after many high-operating cost years and
a non-binding vote of the public to close the facility. This meant that SMUD was
compelled to replace 900 megawatts of power through power contracts, thus
participating in a precursor of a restructured generation market known as the
portfolio manager model, a framework envisioned and intended by both PURPA
and the Energy Policy Act of 1992 (Joskow 1997). This portfolio management
model has been set aside in many countries and in California to pursue the retail
wheeling or customer choice model. While SMUD is not compelled either to join
the ISO or to adhere to AB 1890, the District began early in the process to
implement the customer choice approach.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
119
The subjects for the study comprise the large, traditional power agencies
in California, the new agencies for managing the transmission system, a large
public utility, and the legislative and regulatory institutions that govern the
electric industry.
Testable Hypotheses Derived from the Institutional Arrangements
HI: The rate freeze imposed upon investor-owned utilities has reduced system
reliability.
This hypothesis contrasted results from the distribution companies with more
generalized reliability metrics.
H2: Industry investment has shifted in response to restructuring, resulting in less
investment in reliability-related systems.
Each utility investigated has a capital investment segment of their budget.
This hypothesis compares non-core investment in each utility to total
investments, and to investments made by new participants seeking open
markets in the non-core areas.
H3: Increased contracting for power has reduced system reliability.
This hypothesis compares over time the percentage of total power supply that
is provided through purchased power to each utility’s reliability profile.
H4: The new intermediary structures (ISO and power exchange) have reduced
system reliability.
This hypothesis will contrast the number and severity of reliability
disturbances before and after the establishment of the California ISO.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
1 2 0
H5: The introduction o f direct access customers has decreased system reliability.
This hypothesis will compare the number and load of direct access customers
to the reliability profile of each utility.
H6: The reduction o f DSMpractices does not affect system reliability.
This hypothesis will compare over time the amount of load subject to demand
side management (DSM) to the peak load associated with reliability
disturbances
H7: The separation o f generation from distribution does not affect system
reliability.
This hypothesis will compare the reliability profile of each utility against its
self generation capacity over time.
The survey instrument is based on a Department of Energy study of East
Coast power outages in 1998 and 1999; it also references proposed or
implemented legislation introduced in 2000 to mitigate unintended effects of AB
1890. The survey provides qualitative supplemental information on the causes
and cures for recent system problems, and the industry perception of the ISO
model. The questions derive from the five areas of possible causality discussed in
the Department of Energy’s Power Outage Study Team’s Final Report
(Department of Energy, 2000).
Field and Analytical Procedures
Several sets of data are used for this research. Some sets are measured,
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
121
such as weather data (temperature, precipitation). Others are collected
automatically at creation, such as current or peak electrical load. Certain data fall
within a derivative category, such as average system cost. There are missing data
points which have been extrapolated or reconstructed where possible; constructed
data are noted in chapter 5. The data have been aggregated into monthly units to
the degree possible and meaningful. While there exist considerable projected
data, such as load predictions for future years, this has not been used, particularly
since post-implementation projections have had mixed accuracy. These data have
been collected and applied against the hypotheses, as noted in Table 5 below.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
1 2 2
Evidence LADWP SMUD PG&E SCE SDG&E State/
region
Hypotheses
Weather HI et al.
Temperature
CDD/
HDD
Precipitation rain
Baseline data HI etal.
Generation owned MW MW MW MW MW
Peak load MW MW MW MW MW H6,H7
Purchased power MW MW MW MW MW H3
Direct access
customers #'s #'s #'s H5
DSM/load
management # MW # MW # MW # MW H6
Reliability HI et al.
NERC major outages #'s #'s #'s #'s #'s iflarge HI
SAIDI reports
monthly annual
annual/
specific
annual/
specific
annual/
specific
HI
Stage 3 emergencies ISO HI
Stage 1 emergencies ISO H6
Stage 2 emergencies ISO H6
Load shedding events day MW day MW H6
Capital Investment H2
Generation sales $/MW $/MW $/MW
Non-core investment $ $ $
New participant
investments $/MW $/MW $/MW
Purchased Power and Contracts H3
purchased vs. owned %total %total %total %total %total
purchasing via markets margin
ISO and PX Creation H4
ISO/PX participation no no yes yes yes H7
Direct access H5
customers, residential # # # % load
customers, commercial # # # %load
Demand Side Mgmt H6
Load reduction annual monthly annual annual annual H7
Energy efficiency annual monthly annual annual annual
Time-specific Measures HI, H4
ISO startup statewide transmission for the IOUs, end of March 1998
CTC payoff complete SDG&E only, CPUC decision July 1999
Table 5: Data Applicable to the Research Hypotheses
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
123
Face-to-face and teleconference interviews were scheduled with managers
at utilities, and also with other participants in the legislative and policy making
processes. The selection of these persons began with an invitation to all ISO
Board members, as well as seeking out utility-specific staff associated with
specific research data. Each was asked who else might be interested in
participation, either as a formal survey participant or for general interview
questions. Twelve formal survey interviews were completed; more than 20
additional persons provided specialized or limited information. These persons are
noted in Appendix B as informal contacts; supplementing these research contacts
are the public comments made by many additional restructuring stakeholders at
legislative hearings and in published accounts.
Public hearings attended included Capitol informational hearings, agency
sponsored workshops, and Senate and Assembly committee meetings on
legislation in process. These forums serve as an official setting for observation of
policy makers, legislators, and interested parties. Videotaped hearings document
evidence not found in print elsewhere in the publicly available records.
Data Collection and Recording
Large if dissimilar raw data sets document the electric system and its
restructuring. These data may be found in various configurations, according the
role of the publishing organization. Each organization manages data relevant to
its role in the industry and the regulatory requirements of its role. Data come
from NERC; through the California PUC for required IOU submissions; through
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
124
the CEC for data collected on municipal utilities. The California ISO also
manages a large data store. Public information is available to all, including
utilities or their subsidiaries now with competing interests. Utilities also manage
proprietary data collections. Several data sets are standard in their reporting and
part of submissions to state and federal regulators. Additional data are available
within reports on special investigations of especially large outages, obscured
within SEC filings and financial statements, included in bond issuance
prospectuses, and various other operations documents. Limited material is also
included in annual reports, seemingly based on utility discretion.
Electric system reliability is very high. There are large operational and
collateral costs for bad reliability. Thus, a matrix showing daily reliability would
be mostly zeros. One important decision that had to be made for this dissertation
was the unit of measure of time. Much of the available and published data points
are annual, while reliability disturbances usually span minutes or hours. In order
to compare across these many measures, the data collection was aggregated or
disaggregated into month time units to the fullest degree possible.
Data Sources
Material Directly Supplied by Utilities
The initial data source is material provided by the five large public and
private utilities studied. These include SEC filings, material in their annual
reports, and some additional material provided on request. The municipal and
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
125
investor-owned electric utilities have differing reporting requirements and also
diverge in their perspective on the proprietary nature of their information. This
led to some missing data in the data set. Where possible, the missing data were
reconstructed through review of other materials. However, some data were not
available from all five utilities. In these cases, the comparison across utilities
necessarily excludes consideration of one in the set.
The data collection problems were consistent and not unique to this
researcher. Many calls and emails were not returned. In August 2000, a special
task force appointed by Governor Davis to study the electric system issues
complained in their first report’s cover letter that much of the information they
needed would not be provided by the utilities. In October 2000, one CPUC
commissioner complained in a workshop on summer reliability that despite over
100 subpoenas, the utilities were not providing the CPUC with information
requested as part of their Order to Investigate. Thus, the data that have been
collected must suffice for the present time.
Weather Information
In determining how much weather data would be relevant, several sources
were reviewed, and the data discussed in coordination with utility and regulatory
staff who work with weather data.2 3 These experts did not feel that weather was a
2 3 Weather assessment was aided by Tom Gorin, California Energy Commission, and Vito Blomo,
Sacramento Municipal Utility District. Supporting weather information also came from the
Climate Prediction Center at NOAA. M r. Gorin additionally noted that the impact of local vs.
regional weather on electricity reliability could be directly ascertained through reviewing energy
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
126
relevant factor in the increase in electricity shortfall. SAIDI reports and other
outage explanations often cite the weather as the single most salient variable,
especially for large outages. For this reason, weather data were collected. While
there is some variability of weather in the state, on any given day the general
conditions are similar enough to use one set of data as the baseline. The weather
data represent meteorological conditions in the central valley, specifically
Sacramento.
The precise data set includes high, low, average temperature, heating
degree days and cooling degree days for Sacramento, as a proxy for weather
statewide. According to the agencies’ weather specialists, “When it’s hot in
Sacramento, “... it’s hot elsewhere, basically” (Tom Gorin, Demand Analysis
Office, California Energy Commission; personal communication). In fact,
understanding weather’s impact on demand supercedes the statewide unit.
Weather patterns regionally-the weather in California and adjacent states-affect
the overall load of the western U.S. grid, including the capacity available for
import into California. The state is a net importer due to historic availability of
inexpensive hydropower from the Northwest coupled with California’s strict air
quality and environmental standards which have resulted in more use of
generation sited beyond state boundaries.
exports and pricing hourly on the ISO’s website. The immediacy and completeness of ISO posted
data has since been implicated in the market power debate, reducing its public availability.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
127
The concept of a heating degree day was developed by heating engineers
as a mechanism to relate each day’s temperature to the demand for fuel.
Similarly, the concept of a cooling degree day relates the temperature to the need
for energy demands for air conditioning. Heating and cooling degree days are
based on deviations from an average daily temperature of 65°. The daily average
temperature is used in calculating a day’s HDD or CDD; the high and low
temperatures are added, and divided by two to produce a day’s average
temperature. The difference between 65° and the average temperature is that
date’s HDD or CDD. If a summer day’s high is 100° and the low is 70°, the
average is 85° and the day’s CDD value is 20, representing the amount of air
conditioning load that is likely for that day. The HDD and CDD are used by
energy-oriented weather researchers as daily or aggregate values. These can be
summary or averages, and can be used to compare energy impact of different
locations, or to compare energy loads from weather in the same area at different
times of year. They can also be used longitudinally to compare weather changes
over time.
Because the electric grid is a high reliability system, information on
outages was aggregated by month, so weather data were also aggregated that way.
When looking at averages for the month, the daily weather dynamics are less
relevant. “... the rainfall value on any particular day could possibly explain that
day’s load, but in the overall scheme of a month’s energy it doesn’t provide much
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
128
explanatory information.” (Vito Blomo, Principal Resource Planner, SMUD;
personal correspondence).
CPUC Required Outage Reports
The California Public Utility Commission requires specialized outage
reporting from the investor-owned utilities. These numbers include transmission,
substation, and distribution outages over 5 minutes in duration; they exclude
planned outages. The SAIDI, SAIFI and MAIFI indexes also exclude events
caused “. .. by earthquake, fire, or storms of sufficient intensity to give rise to a
state of emergency being declared by the government. . . or any other disaster
that affects more than 15% of the system facilities or 10% of the utility’s
customers, whichever is less” (CPUC, 2000b). CPUC Decision 96-09-045 thus
serves as a reasonable measure of utility performance, unbiased by circumstances
beyond the utility’s control, as imposed additional reliability recording,
calculation, and reporting requirements. The SAIDI and SAIFI information
covers the period of 1990 through 1999 under study; the MAIFI index was not
added until 1994. Additionally, as PG&E notes in their 1999 Reliability Report,
advanced technology found in new monitoring equipment (“electric outage
notification devices”) caused a large increase in their MAIFI data between 1995
and 1996, nullifying comparisons of years 1990-1995 with 1996-1999. PG&E
also did not submit specific SAIDI incident information for years 1997-1999;
inquiries to both CPUC and PG&E were made, but these data have not become
available.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
129
Again, summarizing these metrics, the units for the reliability measures are:
• SAIDI - minutes of sustained outages per customer per year
• SAIFI - number of sustained outages or more per customer per year
• MAIFI - number of momentary outages per customer per year.
In addition to annual SAIDI numbers, utilities provide some additions to the
overall SAIDI number, such as contributions of specific outages or other monthly
equivalent SAIDI data. The SAIDI data are the most complete reliability
statistics publicly available and are used in the cross-utility analysis. These data
focus on the distribution companies, not on the generation or transmission portion
of electric utility service.
NERC DAWG Database
The Disturbance Analysis Working Group of the North American Electric
Reliability Council (NERC) provides summary information on major disturbances
throughout the national transmission grid (“bulk electric systems of the electric
utilities in North America”, NERC 2000) from utilities’ reports to the Department
of Energy. This is the DAWG database. For this research data from 1990-1999
has been reviewed.
Legislative Hearing Documents
The California Legislature has held several hearings on electric system
restructuring, where numerous stakeholders have contributed company
documentation supporting their arguments. The Legislative Analyst also provides
information for the use of the legislators and the public at these meetings.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
130
ISO Hearing Documents
The Independent System Operation meets regularly and also holds special
meetings to evaluate the market and supply operations. These have provided
invaluable information useful in understanding the development of these
institutional structures. Their meeting documents also include operations
statistics and reports for the various governance subcommittees.
Stakeholder Interviews and Surveys
A number of principal participants and elected officials have discussed the
various issues one-on-one. Due to the extreme political sensitivity of the matters
discussed in the document, this information is presented in the aggregate or as
unattributed comments exemplifying survey responses. Many of these individuals
regularly speak out in the public media or in public meetings, and attributed
remarks are limited to those made in these public contexts or with specific
approval from the interviewed stakeholders.
Analytic Design
The empirical data used in the quantitative assessment include:
• energy-related weather data
o rainfall,
o cooling degree days, and
o heating degree days
• outage-related data
o SAIDI
o top SAIDI events
o NERC-reportable events
• system capacity and load management data
o ISO stage 1,2 and 3 declared emergencies
o utility-specific load shedding
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
131
• restructuring process milestones
o implementation of the ISO
o completion of collection of the CTC
• energy load offset by energy efficiency measures
• changes in the amount of purchased and contracted power
• changes in direct access customers
These data were compiled from multiple sources, as shown in Appendix
C; they were then analyzed using conventional spreadsheet and statistical
software. For the regression analysis, a data matrix was constructed for each
utility and for the five utility comparison. The logit was performed by
constructing a logical dependent variable representing each month whether or not
an outage or near-outage (system wide or local capacity) incident occurred.
Negative binomial regression was also used, verifying the logit regression and
allowing calculations based on one of the continuous variables (SAIDI). The two
dependant variables for the two regressions captured different aspects of the
California model of partial restructuring.
The regression analysis uses monthly data. Summation of data into
months was done due to the historic high reliability resulting in many zero values.
The raw data have been modified as follows:
• energy-related weather data
o rainfall: daily rainfall summed into monthly total
o cooling degree days: mean daily CDD for the month (daily
values for the month, summed and divided by number of
days in month)
o heating degree days: mean daily HDD for the month
• outage-related data
o SAIDI: annual value, constant for all months in any year
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
132
o top SAIDI events: value of any top incidents in any month
summed
o NERC-reportable events: value of any NERC events in
any month summed
• system capacity and load management data
o ISO stage 1,2 and 3 declared emergencies: total number of
emergencies in the month
o utility-specific load shedding: total number of days of load
shedding in the month
• energy load offset by energy efficiency measures: monthly
number
• changes in the amount of purchased and contracted power: annual
number
• changes in direct access customers: monthly number
For use in the matrix analysis, dummy variables separated data for the five
utilities to be pooled; another dummy variable separated municipal utilities from
IOUs, in case the significant distinction between these two classes was not
captured by the selected independent variables.
The reliability disturbance logical dependent variable constructed for the
regression analyses, djlsturb, was assigned the value 1 if the month experienced
load shedding, had a top SAIDI incident, had a NERC-reported major incident, or
if reserves fell below the 7% margin resulting in an ISO Stage One Emergency.2 4
The use of a constructed variable combining these utility and operating entity data
allows historic load shedding data to approximate the required responses under
the ISO emergencies.
2 4 Stage O ne Emergency declared at 7% of operating reserves; Stage Two at 5%, and Stage 3 at
1.5% . News releases for emergencies contain this information, and are available online at
www.caiso.com.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
133
Using two dependant variables for the two regressions allows for a
comparison between the effects of partial restructuring on distribution companies
(captured with the SAIDI data), and a more systemic impact that includes
institutions and systems directly established or altered by AB 1890 (captured with
the constructed dummy variable). This can investigate whether the rate freeze
under which the distribution companies operated may have led to operating
pressures that reduced their capability to provide reliable service.
The qualitative component of the research involved a survey related to the
Department of Energy’s Power Outage Study Team findings, with attention to the
California model of partial restructuring and focus on salient organizational and
market issues from the summer of 2000. The survey findings do not comprise a
test of the research hypotheses; instead, they allow participants’ ideas in the
general area of inquiry to be included in the overall assessment of the California
model.
Methodological Assumptions
Logit regression is a data technique with limited tolerance for missing
values. As an alternative to discarding all rows with no specific data, missing
data have been reconstructed to the degree possible from other aggregate
numbers. In other cases, where this is not possible, the logit regression for the full
set of utilities excludes data categories or years where substantial data are not
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
134
available. In these cases, utility-specific regression will include variables that are
not available across the full set.
It is assumed that the non-random nature of the participants sampled for
qualitative input do not adversely alter the findings or conclusions which are
based on an empirical analysis of outage and utility performance data. Although
not random, parties from recognized participants and special interests have been
interviewed as noted above. These interested parties are specifically identified
through their involvement in hearings, legislative analysis, and other public
documents, thus constituting a purposive sample. While some of the parties
contacted declined to be interviewed, opinions from all parts of the spectrum on
electric system restructuring are available and have been made known through
published statements, votes as members of elected or appointed bodies, and other
public actions.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
135
Chapter 5
Findings
Introduction
This chapter reports the findings developed from the analysis of data.
First discussed are the regression results from the quantitative body of data for the
five investigated utilities. The logit regression results are presented, followed by a
discussion of the error term issues associated with the collected panel data.
Negative binomial regressions results are reported next; these regressions were
performed to address questions raised by the logit regressions. The outage data
are separately described and analyzed in the next section.
The first chapter described institutional practices and arrangements altered
by AB 1890. Findings related to these five changes are discussed in the four sub
sections on institutional changes and comparisons. There are four sections, not
five, due to some overlapping between changes in owned power and the roles
taken on by the ISO.
The final sections in this chapter begin by describing the qualitative
material from interviews and the direct observation of restructuring participants
are presented. The fit of the findings to the hypotheses concludes the chapter.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
136
The Very Public Problem
In 2000 California saw a shift in the public perception of the electric
industry. Little considered previously unless the lights went out during a storm
or following an accident, California’s electricity industry at the beginning of 2001
seems a high profile, dysfunctional, and expensive mess. No simple solution to
higher customer bills is evident despite the September 2000 legislation rolling
back San Diego’s retail and small commercial rates and quieting public concerns
until after the November election.
Only one aspect of this very complex public problem has been
investigated here: How has electric system reliability been affected by AB 1890’s
implementation? Has the incident rate of reliability disruptions been altered by
restructuring and the change in generation ownership for IOUs?
Much has been said about the lack of generation capacity built in the
1990s; a look at the NERC summer and winter load forecast for the WSCC region
that includes California suggests a trend of gradually declining margins in winter;
the trend is less gradual in summer, yet still sufficient for safe and reliable grid
operations in the western United States, far higher than the 7% margin that
triggers an ISO Stage 1 emergency.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
137
Available
Net
Internal Available
Mon-Yr Resources Demand Margin
Dec-96 58,621 39,144 33.2%
Jan-97 59,376 38,102 35.8%
Feb-97 59,344 37,585 36.7%
Dec-97 56,172 38,834 30.9%
Jan-98 56,926 37,983 33.3%
Feb-98 56,045 36,772 34.4%
Dec-98 54,203 37,207 31.4%
Jan-99 53,875 36,462 32.3%
Feb-99 53,437 35,381 33.8%
Available
Net
Internal Available
Mon-Yr Resources Demand Margin
Jun-96 63,229 43,494 31.2%
Jul-96 63,449 47,722 24.8%
Aug-96 63,309 48,517 23.4%
Sep-96 62,825 45,071 28.3%
Jun-97 60,622 44,529 26.5%
Jul-97 61,886 47,328 23.5%
Aug-97 61,439 48,449 21.1%
Sep-97 59,538 46,346 22.2%
Jun-98 56,245 41,847 25.6%
Jul-98 57,678 44,581 22.7%
Aug-98 57,458 45,750 20.4%
Sep-98 55,610 44,136 20.6%
Jun-99 56,131 46,003 18.0%
Jul-99 56,355 48,255 14.4%
Aug-99 58,552 50,996 12.9%
Sep-99 55,401 48,008 13.3%
Jun-00 56,093 44,673 20.4%
Jul-00 58,001 49,273 15.0%
Aug-00 57,192 48,703 14.8%
Sep-00 56,042 45,194 19.4%
Table Set 6: NERC Summer and Winter Margin Estimates
WSCC changed the regional composition over this period. The 1996-1997
Assessments covered the California-Southern Nevada Region. 1998 and later
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
138
Assessments covered the Califomia-Mexico (U.S.) region. For 1999/2000 winter
the WSCC regions were aggregated U.S., Canada, and Mexico only.
The increasing frequency and severity of ISO emergencies suggest that
other factors than regional generation capacity are in play. No ISO Emergency is
declared prior to a 7% margin; a Stage 2 Emergency is declared for a 5% margin;
and Stage 3 is only invoked at the 1.5% margin. Given the forecast, one might
expect no emergencies to have been declared at all, rather than separate
emergencies on 36 days through November 2000, up from only 3 in 1999 and 12
in the ISO’s first year of operation.
This research sought to characterize these factors among other empirical
measures of reliability disturbances. The research results are collected and
presented in the following sections.
Q uantitative Analysis
The quantitative analysis section includes regressions results performed on
data matrices constructed from multiple data sources, as noted in Appendix C.
These matrices were analyzed as non-linear models, thus freeing them from
unreasonable assumptions of linearity (Aldrich and Nelson 1984). This prevents
assumptions such as the idea that the marginal effect of the independent variables
is constant, and is a requirement in light of the discrete dichotomous qualitative
dependent variable, djisturb, constructed for the logit regression.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
139
Logit Regression Results
The constructed dummy variable allows an examination of the effects of
the California design which deregulated generation and froze distribution. The
construction includes information relevant to the generation, transmission, and
distribution of electricity.
The relevant model for the logit regression is
DVit = aXt + pZit
where DVit represents the reliability disturbance measure; Xt represents
the independent variables constant across utilities such as the date the ISO started
and Zit represents independent variables that vary across time and utility, such as
generation capacity or changes in purchased power contracts.
A more precise description of the variables is given in Table 8. This table
describes each independent variable used in the calculations. It notes whether the
variable can be classified as an X- or Z- type variable, and whether its value is
constant within any given utility (e.g., all occurrences of the djnuni value within
SMUD’s data set will be assigned the value 1) so it has value only in cross-utility
analysis.
Each utility’s published data were sought in these categories for as many
of the past 10 years as was available. Each utility’s data were run individually to
maximize the years of data covered and categories of data included. As noted in
chapter 4 the regression analyses use months as the unifying unit of measure for
aggregating data, though some collected data are annual only and have been noted
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
140
as such. Dummy variables were constructed to represent the introduction of the
ISO, the end of the CTC in San Diego, to separate data for the five utilities within
the cross-utility data collection, and to separate municipal from IOU data for
cross-utility comparisons. The summary dependent logical variable representing
reliability disturbance occurrence, djisturb, received the value 1 if the month
experienced load shedding, had a top SAIDI incident, had a NERC-reported
major incident, or if reserves fell below the 7.5% margin resulting in an ISO
emergency. A dummy variable for earthquakes was included in the initial
analysis, but ultimately discarded as not providing any value; in all cases, major
earthquakes result in power outages. Cataloging and quantifying all California
temblors for the past decade and correlating them to outages to determine the
relevance of quake size and proximity goes beyond the scope of this research; this
would be an interesting area for additional research.
X orZ
variable
description comment
d_muni X dummy - muni or IOU
w_rain Z total rain for the month similar, statewide weather considered
for this model
w_cddavg z average ODD value for the month "ODD" and "HDD'' are energy planning
metrics related to heating and cooling
loads
w jid d avg z average HDD value for the month heating degree days
d js o u p X dummy - month since ISO startup demarks beginning of restructured ops
purchpwr z purchased and contracted power percentage utilization of
contracted/purchased generation
dsm ee z load affected by DSM measures
only available for SMUD
da_delta z customers switching suppliers only available for lOUs
d_endctc X dummy - passthrough pricing SDG&E completed collection of CTC;
but this ability was immediately
rescinded
Table 7: Logit Regression Independent Variables
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
141
This results in an equation of the form:
ddsturb = aidjtnuni + PiW_raintot + (3 2 w_cddavg + Pswhddavg +
a2 d_ISOup + p4 purchpwr + p5 dsm_ee + p6 da_delta +
ajdendctc
The upper and lower bounds of data collections were designed to
m axim ize relevant factor collection opportunity. None of the utilities provided
comprehensive data for analysis. Available data from five utilities produced the
following logit regression results. Results are presented first from the two
municipal utilities, then the three investor-owned utilities, and finally, from the
cross-utility analysis. The range of data reflects months for which data were
available relevant to the research question and hypotheses.
The regression extends to available SAIDI outage data for the municipal
utilities, and available direct access data for the IOUs. While this limits the
regression results somewhat, additional categorization and analysis of data occurs
in the other sections of this chapter, in direct reference to the hypothesized
institutional changes.
Data from Municipal Utilities
SMUD
The first utility-specific regression was performed for the Sacramento
Municipal Utility District (SMUD). This municipal utility has a large energy
efficiency department and provided detailed data on use of demand side
management and energy efficiency. This municipal utility provided less
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
142
comprehensive SAIDI information than the required CPUC reporting of the IOUs.
The data available from January 1994 through December 1999 were used for the
logit regression.
Variable Obs Mean Std. Dev. Min Max
w rain 72 1.201389 2 . 014 0 9 . 7
w cddavg 72 3 .2625 4.222256 0 13 . 9
w hddavg 72 6.669444 6.794377 0 22
d isoup 72 0 1
purchpwr 72 .4993333 .1063102 .366 .632
dsm_ee 72 1.43425 1. 078016 - .403 4 . 572
Case Processing Summary
Unweighted Cases N Percent
Selected Cases Included in Analysis 72 100.0
Missing Cases 0 .0
Total 72 100.0
Unselected Cases 0 .0
Total 72 100.0
Omnibus Tests of Model Coefficients
Chi-sauare df Sid.
Block 31.826 6 .000
Model Summary
-2 Log
likelihood
Cox & Snell
R Sauare
Nagelkerke
R Sauare
44.452 .357 .547
Classification Table?
Observed
Predicted
D DSTURB
Percentage
Correct 0 1
D_DSTURB 0 52 4 92.9
1 6 10 62.5
Overall Percentage 86.1
a. The cut value is .500
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
W_RAIN -.412 .530 .605 1 .437 .662
W_CDDAVG .473 .159 8.825 1 .003 1.605
W_HDDAVG .103 .103 1.001 1 .317 1.108
PURCHPWR -7.398 9.802 .570 1 .450 .001
DJSOUP .150 1.975 .006 1 .940 1.161
DSM_EE .053 .452 .014 1 .906 1.055
Constant -.247 5.581 .002 1 .965 .781
Table Set 8: SMUD Data, 1994-1999, Logit Regression
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
143
The sole significant variable in SMUD’s reliability disturbance profile is
cooling degree days, w_cddavg. Reliability disturbances in hot weather are more
frequent, and heat storms are good predictors of high load problems. This reflects
SMUD’s most frequent challenge, air conditioning loads during the central
valley’s hot summer weather. To manage its load at these peak times, SMUD has
used voluntary load shedding similar to that invoked by the ISO in a State 2
emergency. SMUD made regular use of load curtailment as an alternative to
buying extra power at peak prices; this strategy limited the need for additional
local generation and also assisted in maintaining relatively low rates. After AB
1890’s passage SMUD re-designed its load shed program and reduced its
invocation as is discussed separately in the section on energy efficiency.
LADWP
LADWP provided the least amount of outage and reliability information
of all five utilities. The data available from January 1990 through December 1999
were used for the logit regression.
Purchased and contracted power represents the capacity, not the quantity,
of power purchased. LADWP was not subject to the divestiture requirements of
the IOUs, and is free to contract in whatever manner seems appropriate for
management. The amount of reported contracted and purchased power capacity
was stable from 1994 through 1998 at 2,900 MW, and most of this power is
contracted for a very long time. The IPP contract expires 2027, their Hoover Dam
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
144
contract expires 2017. The importance of this capacity fluctuated with
operational capacity of LADWP’s owned generation.
LADWP has made some contractual changes, such as terminating a
Montana Power Company contract, but that transaction was not completed until
December 21, 1999. In August 2000, the City Council approved LADWP’s 2000
Integrated Resource Plan which calls for 2,900 MW of in-basin power additions
over the next 10 years. How this plan integrates with existing long-term power
purchases is not known, but LADWP has reduced its debt load substantially in
2000 by selling excess power into California’s expensive spot market.
Two purchased power numbers may be reported by utilities, the
megawatts of energy made available for use or the number of energy units bought
and sold. The former capacity is measured in megawatts; the latter is measured in
megawatt or kilowatt hours by the thousand, million, or billion. For LADWP, no
variance was reported in capacity, but actual purchases did vary, from 10.8 billion
kWh in 1995 to 13.3 billion kWh in 1998.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
145
Variable Obs Mean Std. Dev. Min Max
w rain 120 1.101667 1.958175 0 9 . 7
w cddavg 120 3.3875 4.257698 0 13.9
w hddavg 120 6.783333 7.128984 0 23 .9
purchpwr 120 .196 .0789681 . 12 .35
d_isoup 120 0 1
Case Processing Summary
U nweiqhted C a se s N P ercent
S elected C a se s Included in A nalysis 120 100.0
M issing C a se s 0 .0
Total 120 100.0
U n selected C a se s 0 .0
Total 120 100.0
Omnibus Tests of Model Coefficients
C hi-square df Sig.
Block 2 3 .8 6 0 5 .000
Model Summary
-2 Log
likelihood
C ox & Snell
R Square
N agelkerke
R Square
6 6 .5 6 5 .180 .341
Classification Tabl#
Observed
Predicted
D DSTURB
Percentage
Correct 0 1
D DSTURB 0 103 2 98.1
1 8 7 46.7
Overall Percentage 91.7
a. The cut value is .500
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
W_RAIN -1.361 .776 3.077 1 ,079 .256
W CDDAVG .181 .112 2.595 1 .107 1.199
W_HDDAVG .078 .071 1.212 1 .271 1.081
PURCHPWR 34.819 14.067 6.126 1 .013 1.3E+15
DJSOUP -4.117 2.491 2.732 1 .098 .016
Constant -9.063 2.638 11.804 1 .001 .000
Table Set 9: LADWP Data, 1990-1999, Logit Regression
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
146
The independent variable w_rain is significant at the .10 level. This is less
compelling than the .05 level significance of the purchased power, but suggests
that rain is not positively correlated with reliability disturbances. Other than this
weak effect, the LADWP results show no statistical significance of weather-
related variables on electric system reliability. Rather than being truly reflective
of the role of weather in system disruptions, this finding likely reflects a lack of
incident-specific SAIDI data available.
The purchased and contracted power independent variable purchpwr is
significant at the .05 level. LADWP has kept its long-term contracting capacity
constant at 2,900 MW from 1995 through 1998; over the same period its
purchased kilowatt hours increased and fell, matching production. The quantity of
purchased and contracted power often exceeds the requirements of LADWP,
assuring availability of sufficient generation for local demand; the utility has sold
the excess power into the market, reducing its debt load. Long-term contracting
for power differs from the day-ahead purchasing model carried out at the PX and
required of the IOUs. The expected result would be that LADWP’s greater
flexibility in contracting for power would be negatively correlated with reliability
disturbances, but the coefficient here is positive. The logit analysis included
effects from outside the distribution company, that is, NERC reliability
disturbances and declared ISO emergencies. The negative correlation between
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
147
purchase power and disturbances suggests that these external forces interfere with
LADWP’s strategic internal use of purchased power and contracted power.
The independent variable dJSOup is significant at the .10 level. This is
less compelling than the .05 level significance of the purchased power. This does
not necessarily indicate that the ISO reduced disturbances. The dummy time
variable separates disturbance data from before and after the ISO began its grid
controller role. This point in time is consistent with other factors that could be
relevant for LADWP, such as strong economic growth in California and the
reduction of generation margin in the region. It could also be suggested that it
represents the arrival of S. David Freeman as general manager. Freeman headed
up the group that set up the ISO, then aggressively worked at LADWP to improve
its position in the partially deregulated environment. This regression result means
the hypothesis that the ISO had a positive effect on system reliability cannot be
rejected. This could also be interpreted to suggest the statewide systematic
changes that indirectly affect LADWP and other municipal utilities exempt from
AB 1890.
Data from Investor-owned Utilities
PG&E
Missing or calculated values were handled as follows. For years 1993-
1997, purchpwr (purchased and contracted power percentage) is constructed
from the annual figures for owned megawatts divided by that year’s peak demand.
For 1998-1999, peak demand for PG&E could not be located, so the 1997 peak
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
148
demand has been used as a substitute value on the assumption that the peak was at
approximately this level or higher. Similarly, no updated information on the
percentage of purchased power is available for 2000. The 1999 purchased power
value has been extended into 2000 as the only major generation divesture pending
is the sale of hydro-electric facilities; this divestiture is still under review and
possible legislative constraint.
The logit regression with values constructed as noted above includes the
period from January 1993 through October 2000.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
149
Variable Obs Mean Std. Dev. Min Max
w rain 94 1.323404 2 .219243 0 9.7
w cddavg 94 3 .260638 4 . 099311 0 13 . 9
w hddavg 94 6.575532 6.818581 0 22
d_quake 94 0 1
purchpwr 94 .4303191 .1739574 .22 . 7
d_isoup 94 0 1
da_delta 94 1249.372 2934.023 -13831 12916
Case Processing Summary
Unweighted C ases N Percent
Selected C a ses Included in Analysis 94 100.0
Missing C ases 0 .0
Total 94 100.0
U nselected C a ses
0 .0
Total 94 100.0
Omnibus Tests of Model Coefficients
Chi-square df Sig.
Step 6.530 6 .366
Model Summary
-2 Log
likelihood
Cox & Snell
R Square
Nagelkerke
R Square
123.781 .067 .089
Classification Table a
Observed
Predicted
D DSTURB
Percentage
Correct 0 1
D_DSTURB 0 30 17 63.8
1 20 27 57.4
Overall Percentage
60.6
a. The cut value is .500
Variables in the Equation
B S.E. Wald df Sis. Exp(B)
W_RAIN -.026 .120 .048 1 .826 .974
W.CDDAVG -.037 .077 .233 1 .629 .963
W_HDDAVG -.011 .050 .045 1 .832 .990
PURCHPWR -2.625 2.878 .832 1 .362 .072
DJSOUP 2.048 1.124 3.320 1 .068 7.749
DA_DELTA .000 .000 .977 1 .323 1.000
Constant .809 1.036 .610 1 .435 2.247
Table Set 10: PG&E Data, January 1993- October 2000, Logit Regression
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
150
The PG&E logit model fails to predict or explain the utility’s reliability
disturbance profile, as evidenced by the failure of the Chi-square test.
There are at least two possible reasons for this. First, the data may be
insignificant themselves, either in the number of data points or in the accuracy of
the data, and the assumptions used in constructing missing values may be
incorrect. Second, the elements studied for constructing the model may not have
much effect on the overall reliability of PG&E’s system. It seems possible that
the extrapolated and missing values create enough variation to cause this effect
for this range of data. It is equally arguable that the constructed binary variable
that represents the utility’s reliability disturbance profile fails to truly capture the
range of system reliability disturbances necessary, or otherwise oversimplifies
PG&E’s reliability profile. To check against the latter possibility and look more
closely at the components of the constructed dependent variable, negative
binomial regression was done to determine if the dependent logical variable was a
specific cause of model failure by using the SAIDI statistic as the dependent
variable. The negative binomial analysis is discussed in a later section of this
chapter.
SCE
The SCE data set runs from January 1993 through September 2000. The
1999 value for the percentage of purchased and contracted power, purchpwr, has
been extrapolated to September 2000.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
151
Variable Obs Mean Std. Dev. Min Max
w rain 93 1.32043 2.231083 0 9.7
w cddavg 93 3 .284946 4 .114712 0 13 . 9
w hddavg 93 6.602151 6.850626 0 22
d isoup 93 .3225806 .4699975 0 1
da delt 93 1270.312 3106.796 -5954 14519
purchpwr 93 .286129 .3596708 - .05 . 77
Case Processing Summary
U nw eighted C a se s N Percent
S elected C a se s Included in A nalysis 93 100.0
M issing C a se s 0 .0
Total 93 100.0
U n selected C a se s 0 .0
Total 93 100.0
Omnibus Tests of Model Coefficients
C hi-square df Siq.
Block 9 .668 6 .139
Model Summary
-2 Log
likelihood
C ox & Snell
R Square
N agelkerke
R Square
118.988 .099 .132
Classification Table a
Observed
Predicted
D DSTURB
Percentage
Correct 0 1
D DSTURB 0 30 14 68.2
1 14 35 71.4
Overall Percentage 69.9
a - The cut value is .500
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
W_RAIN -.113 .130 .752 1 .386 .893
W_CDDAVG .136 .080 2.907 1 .088 1.146
W_HDDAVG .150 .058 6.567 1 .010 1.161
D JSO U P
1
' o
CO
- -3 .
1.395 .245 1 .621 .501
DA_DELT .000 .000 1.194 1 .274 1.000
PURCHPWR 1.712 1.954 .768 1 .381 5.537
Constant -1.295 .622 4.336 1 .037 .274
Table Set 11: SCE Data, January 1993-September 2000, Logit Regression
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
152
The SCE logit model fails to predict or explain the utility’s reliability
disturbance profile, as the data model fails the Chi-square test. As noted for
PG&E, there are two possible reasons for this. The data may be insufficient or
inaccurate. Secondly, the elements of the model may not capture significant
specific sources of the incident effect on the overall reliability of SCE’s system.
Unlike PG&E, there were not many missing data elements, and the 93 months of
data should have provided sufficient material. It also may be that the constructed
binary variable that represents the utility’s reliability disturbance profile fails to
capture the range and weight of system reliability disturbances necessary to create
a model with predictive value. Again, negative binomial regression analysis was
done to determine if this latter point was indeed a contribution to model failure.
Both PG&E and SCE underwent major corporate restructuring process
over the study period, beyond the institutional changes required by the legislation
and regulation. For example, as will be noted later, the changes in power
contracting have been much more complex than simply the divestiture of fossil
fuel plants. All three IOUs established parent corporations; other internal changes
may not be part of the public record. These two utilities require further study to
determine what underlies their performance. Major audits by the CPUC may
provide more publicly available data for analysis.
SDG&E
The SDG&E data run from January 1991 through October 2000. The
1999 value for the percentage of purchased power, purchpwr, has been
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
extrapolated through October 2000. SDG&E data offer two separate potential
constructions for purchased power. One involves the owned megawatts
contrasted to peak demand, similar to PG&E. Another purchpwr construction
reflects the annually reported percentage of contracted kilowatt hours. E.g., this is
a contrast between capacity-the ability to supply power on demand-and actual
summary usage. The two concepts are similar but not identical. Neither set is
complete. For this analysis the usage-based measure was selected as it limited the
missing data cases and included the period since SDG&E completed collection of
the CTC, allowing that independent variable to be included in the analysis.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
154
Variable Obs Mean Std. Dev. Min Max
w rain 118 1.249153 2.136646 0 9.7
w cddavg 118 3.35 4.144624 0 13. 9
w hddavg 118 6,609322 6.914496 0 22
d isoup 118 0 1
da delta 118 428.4068 1385.751 -6785 8954
purchpwr 118 .6310339 .1061983 .497 .809
d endctc 118 0 1
Case Processing Summary
Unweighted C ases N Percent
Selected C ases Included in Analysis 118 100.0
Missing C ases 0 .0
Total 118 100.0
Unselected C ases 0 .0
Total 118 100.0
Omnibus Tests of Model Coefficients
Chi-square df Sig.
Block 17.414 7 .015
Model Summary
-2 Log
likelihood
Cox & Snell
R Square
Nagelkerke
R Square
145.863 .137 .183
Classification Table a
Predicted
D DSTURB
Percentage
Correct Observed 0 1
5 bS'tU R B 0 35 21 62.5
1 20 42 67.7
Overall Percentage 65.3
a. The cut value is .500
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
W_RAIN -.082 .116 .500 1 .480 .921
W_CDDAVG .239 .075 10.008 1 .002 1.270
W_HDDAVG .180 .053 11.621 1 .001 1.198
DJSOUP -.181 .941 .037 1 .847 .834
DA DELTA .000 .000 .055 1 .815 1.000
PURCHPWR -2.402 3.711 .419 1 .518 .091
D ENDCTC 1.074 .831 1.670 1 .196 2.928
Constant -.328 2.187 .022 1 .881 .720
Table Set 12: SDG&E Data, 1991-1999, Logit Regression
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
155
The SG&E logit model shows w cddavg and w hddavg are significant at the .005
level.
SDG&E retired its competitive transition charge, and was authorized by
the CPUC to begin passing through its cost of power directly to the ratepayer in
the fall of 1999. Low wholesale prices through the winter and spring of 2000
resulted in little shift in consumer prices until summer. After weeks of ratepayer
outrage and election year concerns regarding San Diego area Assembly
representatives who had voted for AB 1890, the legislature rescinded this
passthrough pricing ability in September and included a rate rollback to the prior
June while the price of generation continued its steady climb. Thus it is not
surprising to see no effect from the few months’ application of the passthrough
generation cost on reliability in San Diego.
San Diego’s case might have provided an interesting comparison with the
other utilities who faced a rate freeze on the generation portion of the bill. The
two IOUs had divested generation and were more exposed than the municipal
utilities to variable generation cost, but had no ability to pass that cost on to the
customer.
Based on these logit regression results, neither the introduction of the ISO
as an intermediary structure nor the shift to purchased power had an effect on
system reliability during the early years after the passage of AB 1890 in San
Diego. Only heating and cooling days are significant at the .005 level.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
156
Combined Analysis for the Five Utilities
In addition to utility-specific analysis, a cross-utility logit regression was
run. This allows comparison between the reliability disturbance profiles of the
IOUs subject to AB 1890 and municipal utilities not bound by most aspects of AB
1890.
As was done for the individual investor-owned utilities in the utility-
specific analysis, the percentage of power contracted in 1999 was extrapolated to
2000 allowing the inclusion of much of 2000 in this analysis. The IOU’s fossil
plants had been sold in 1998 and 1999 as required, and the large PG&E
hydroelectric divestiture had been delayed indefinitely. The municipal utilities
had no mandate to divest. Preliminary load growth and consumption figures for
2000 indicate only a moderate increase in demand. This assumption on purchpwr
allows inclusion of first half of the third year of ISO/PX operations, a critical
period.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
157
Variable Obs Mean Std. Dev. Min Max
d muni 566 0 1
w rain 566 1.242049 2.120836 0 9.7
w cddavg 566 3 .345406 4.146128 0 13 . 9
w hddavg 566 6.616784 6.914744 0 23 .9
d isoup 566 0 1
d endctc 566 0 1
purchpwr 566 .4213428 .2413301 - . 05 .809
Case Processing Summary
Unweighted Cases N Percent
Selected C ases Included in Analysis 566 100.0
Missing Cases 0 .0
Total 566 100.0
Unselected C ases 0 .0
Total 566 100.0
Omnibus Tests of Model Coefficients
Chi-square df Sig.
7 1 .5 5 8 7 .000
Model Summary
-2 Log
likelihood
C ox & Snell
R Square
N agelkerke
R Square
5 4 5 .6 3 9 .119 .179
Classification Table3
Observed
Predicted
D DSTURB
Percentage
Correct 0 1
D_DSTURB 0 414 19 95.6
1 109 24 18.0
Overall Percentage 77.4
a- The cut value is .500
Variables in the Equation
B S.E. Wald df Sig. Exp(B)
D_MUNI -1.323 .241 30.214 1 .000 .266
W_RAIN -.061 .062 .983 1 .321 .941
W_CDDAVG .189 .038 24.368 1 .000 1.208
W_HDDAVG .081 .024 11.114 1 .001 1.084
DJSOUP .666 .269 6.123 1 .013 1.947
D_ENDCTC .416 .562 .548 1 .459 1.517
PURCHPWR -.158 .508 .096 1 .756 .854
Constant -1.997 .372 28.792 .000 .136
Table Set 13: Multiple Utility Data, 1990-1999, Logit Regression
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
158
This cross-utility logit analysis suggests that the independent variables
allow for the correct reliability state to be predicted 77% of the time, based on this
model. The statistically significant independent variables in predicting electric
system reliability are two weather-related variables, wjiddavg and w cddavg, the
institutional change variable djsoup, and municipal utility status, djnuni. As the
prior utility-specific data show, weather variables are the most uniformly
significant across utilities; increases in these measures predict increase in
reliability disturbances. Throughout the 1990s, electric system reliability
remained a weather-related story; these results show this in the individual utility
analysis and in the cross utility regression.
The cross-utility regression finds dJSOup significant at the .05 level.
This does not indicate that the ISO caused these disturbances. This dummy time
variable separates data before and after the ISO began its grid controller role.
This point in time is consistent with other factors not included in the analysis,
such as strong economic growth in California and the reduction of generation
margin in the region. This regression result means the hypothesis that the ISO
had a negative effect on system reliability cannot be rejected.
The large increase in reliability disturbances beginning in the summer of
2000 is not reflected in the municipal utilities’ logit regressions, since incomplete
reliability outage data limited those analyses to the period ending December 1999.
Two of the three IOUs logit regressions failed the Chi-square test and have no
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
159
explanatory power. For the third IOU, SDG&E, the variable d ISOup is
significant; its negative coefficient suggests that in its case, the restructuring and
other factors at the time the ISO began slightly decreased system reliability
problems. None of the regression analyses could include the late-2000 increase in
ISO declared emergencies. The ISO emergencies through the end of 2000 will be
discussed within the section on the ISO below.
The investigated independent variables associated with operating changes
introduced by the California restructuring model do not show statistically
significant impacts on electric system reliability for the period under
consideration, which excludes the fall and winter of 2000 and 2001 since utility
operating statistics have not yet been published. The constraints that the
California state legislature immediately implemented when SDG&E passed on
their cost of power removed any CTC completion-related effects for good or ill.
The ISO itself has neither improved upon nor worsened traditionally high electric
system reliability over the time period reviewed here. This analysis suggests the
ISO initially met their organizational goal of maintaining the high reliability of
the statewide electric grid.
djnuni is significant at the .005 level. The negative coefficient suggests
that being a customer of a municipal utility decreases the likelihood of
experiencing a reliability disturbance or outage. California’s municipal utilities
continue as vertically integrated businesses, and have not been restricted in their
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
160
operating practices, unlike the IOU’s. They have not had to divest their
generation or join the ISO.
In developing the research design, it was thought the differences between
the munis and the IOUs could be captured by looking at specific differences in the
divesture of generation capacity and concomitant increase in contracted energy,
and the end of the CTC in San Diego. But these broad measures do not seem to
capture the statistically significant distinction in system reliability based on the
data available.
Another theoretical difference between IOUs and municipal utilities is not
captured in these data. At the time legislators were drafting AB 1890,
conventional wisdom held that the IOUs over-invested in generation because of
their ability to recapture all costs plus the regulatory rate of return. The municipal
utilities were thought also to do this, but to a lesser degree due to political
constraints associated with their governance. This tendency resulted in over
engineered, highly reliable IOUs with excessive generation.
Other operating differences between municipal utilities and the IOUs
might be worth examining if suitable data measurements can be collected. An
example is any difference is the utilities’ required response to separate ISO
emergencies, as the data collected for this research do not specifically detail how
this difference is enacted. Purchased power might also be further broken down
and investigated, including changes in long and short term contracting as well as
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
161
the role of qualifying facilities (QFs) and the impact of surcharges. In all cases
however, the evolving new company structures complicate comparisons across
the pre- and post-restructuring timeline. These issues will be further developed in
the final chapter.
While the number of ISO Stage One, Two, and Three Emergencies
through the summer of 2000 caused much consternation, they did not result in any
disruption of firm load for the period with complete data. Nevertheless, these
emergencies underscored the problem of deteriorating regional generation
margins. Once the margin reaches zero, then negative system effects such as
power disruptions must occur. These ISO energy emergencies representing low
margins for power available and contracted to California have increased and
worsened, as the firm load power shedding in January 2001 demonstrated.
As will be discussed at length below, these quantitative results are
consistent with the survey results. Electricity policy makers and experts feel that
the early years of restructuring were successful. But they are concerned about
how to make the mixed regulatory system functional and efficient over the longer
term.
Discussion of the Error Term, Tests and Transformations
SPSS and STATA were used to generate the regression models. Care has
been taken to eliminate periods with limited outage data, and missing values have
been filled conservatively, possibly increasing the risk of overlooking a
significant correlation. There is a regression assumption that the error term e has
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
162
a common variance, that the data are homoscedastic. The ratio variables were
examined for heteroscedasticity. The dummy variables were assumed
heteroscedastic, but examined for systematic tendencies. The data were generally
examined for outliers in each independent variable.
Two distinct sets of panel data were used for regression analyses. Panel
data have advantages in reducing collinearity and increasing the efficiency of
estimates. But most panel data are not generated from controlled experiment, but
come from everyday economic life; these data are no exception. Panel data have
to be examined for selectivity and heterogeneity bias (Hsaio 1986). The individual
utility data have an error term whose nature includes a time component. The
cross-utility data have an error term with a cross-utility component as well as a
time series component. Tests for randomness, constancy of variance, outliers and
normality were used to examine the data (Neter et al. 1985).
To test for heteroscedasticity in the data, residuals of the significant
variables were examined. If heteroscedasticity was found, the data were
transformed using logs and deflators; e.g. logy, My, and Vy. Transformations of
the residuals reduced but did not completely eliminate heteroscedasticity in
weather data. The weather-related variables remain heteroscedastic despite
transformation. The fact that weather remains unpredictable does not completely
negate the predictive value of the results. If the weather is fine, reliability
disruptions based on weather will be avoided; but it cannot be predicted with
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
163
certainty that summer 2001 will be a hotter or colder than average summer. The
heteroscedasticity problem biases the estimated variances; the estimators obtained
by ordinary least squares procedures are unbiased and consistent, but inefficient
(Neter, Wasserman, and Cuter 1985). For significant independent variables, a
smaller standard deviation reduces the problem. This is reflected in the
confidence levels.
The purchased power variable was plotted against annual SAIDI numbers
in the cross-utility matrix. This graph showed little heteroscedasticity.
Interpretation of the results would suggest predicting a one unit change in
the independent variable correlates or predicts a p percent change in the
dependent variable. The error analysis suggest that the direction of the change is
valid; the moderate size of the individual utility sample (ranging from 74 months
for SMUD to 120 for LADWP) and the combined utility reduces the estimation
error. The model remains a useful analytic tool for describing individual and
cross utility effects.
Alternative statistical methods were used on the data. The logit
regressions for two of the IOUs failed goodness of fit, preventing those models
from producing descriptive value and limiting the ability of the models to produce
predictive value. To derive greater understanding from the data, the negative
binomial regression served to double-check the logit results on the constructed
variable, djisturb, and to examine the continuous variable SAIDI data as the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
164
dependent variable to search for correlations overlooked by the constructed
dependent variable. Simple time series analyses of the data provide additional
information.
Negative Binomial Regression Results
The second dependent variables used for the negative binomial regression
focuses on the effects of partial restructuring on distribution company, as reported
with the SAIDI data. The rates were frozen for the three IOUs. While municipal
utilities have more flexibility in altering their rates, they face both political
constraints and expectations to keep their rates well below the IOU rates. The
two municipal utilities are by extension exposed to the problem of the IOUs,
keeping rates frozen while responding to market pricing for generation. These
constraints on rates may have resulted in distribution practices reducing reliable
service.
Binary logit regression requires a binary dependent variable. A value of 1
for the constructed variable d isturb represented the occurrence of at least one
reliability disturbance in the month. The reliability events used for the logit
regression consisted of ISO emergencies, load shedding incidents, major outages
reported to NERC, or an above-average SAIDI value for the utility that month.
Electric system reliability is very high and generation margins are also substantial,
resulting in many months with no incidents, a zero value for d isturb.
The use of negative binomial regression allowed verification of the results
using the constructed variable d dsturb, but more importantly investigates the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
165
role of the restructuring independent variables using the SAIDI outage value as
the dependent variable. SAIDI is a continuous ratio variable, unlike the dummy
binary variable used for the logit regression. It directly represents the most
extreme type of reliability disturbance, power outage, isolated to the distribution
company portion of system reliability.
SAIDI reports included the annual number of minutes of outage per
customer for the year primarily; IOUs also reported principal SAIDI events and
their contribution toward the annual number. This allowed calculation of an
estimated contribution each month, or a SAIDI monthly number. The munis
provided similar but non-identical SAIDI data. The use of monthly or annual
SAIDI is noted below.
The expenditure of hundreds of millions in state of California funds to
increase and maintain a generation supply margin supports the importance of
margin measures in assessing system reliability. The approach here of using
SAIDI variables excludes reliability disturbances related to reduction of available
generation supply margin included and evident in the ISO emergencies or
industry-standard load shedding practices. The logit regression did not
successfully construct a model with predictive power for the largest two IOUs that
included these measures, so an exploration of the data without marginal incidents
and focused at distribution company specific information was completed. The
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
166
effects of the marginal incidents on SAIDI are explored in the negative binomial
regression by including the data as independent variables.
Negative binary regression results for the IOUs are reported first, then the
results for the municipal utilities.
The PG&E data produced the following results, using SAIDI annual data.
Using SAIDI monthly measures instead of the annual data excluded the post
restructuring period. Several independent variables such as purchased power do
not become relevant until later than that time; PG&E’s SAIDI submission to the
CPUC has been incomplete since 1997. This restricted the use and value of the
monthly SAIDI data, which if available, would have provided a more detailed
outage assessment.
PG&E Negative Binomial Regression Number of obs 84
Model chi2(7) 19.41
Prob > chi2 = 0.0070
Log Likelihood = -300 .0261955 Pseudo R2 = 0.0313
saidi | Coef. Std. Err. z P>|z| [95% Conf Interval]
lnmean |
w rain | .0068785 .004945 1.391 0.164 -.0028135 .0165705
w cddavg | -.0012891 .0030851 -0.418 0. 676 -.0073359 .0047576
w hddavg | -.0006762 .0019494 -0.347 0.729 -.0044969 .0031445
purchpwr | -.3375427 .1103579 -3.059 0.002 -.5538401 -.1212452
d isoup I .0300368 .0402751 0.746 0.456 -.0489009 .1089744
da delta | 2.67e-06 4.82e-06 0.555 0.579 -6.78e-06 .0000121
isoemerg | .0266292 .0173178 1.538 0.124 -.0073131 .0605715
cons | -4.292038 .0395951 108.398 0.000 -4.369643 -4.214432
lnalpha |
cons I -17.33107 .6434947 -26.933 0.000 -18.59229 -16.06984
alpha 2.97e-08 [_lnalpha] cons = In(alpha)
(LR test against Poisson, chi2(1) = -.005516 P . )
Table 14: PG&E Negative Binomial Regression, SAIDI Annual Data as Dependent Variable
With SAIDI annual data as the dependent variable for PG&E, results show
the percentage of purchased power is statistically significant at the .005 level.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
167
The negative coefficient means that with an increasing ratio of purchased power,
likelihood of power outages decrease. These data do not support the general
hypothesis that post-restructuring divestiture of ownership in generation and
increased contracting result in increased system reliability problems over the
period studied.
SCE regression was possible using annual SAIDI numbers or monthly
SAIDI numbers. The monthly numbers represent a proportional SAIDI
contribution including major outage SAIDI values reported in a particular month
and provide a more complete assessment.
SCE Negative Binomial Regression
Log Likelihood = -171.0896848
Number of obs
Model chi2(8)
Prob > chi2
Pseudo R2
84
42.58
= 0.0000
= 0.1107
saidimo | Coef. Std. Err. z P>|z| [95% Conf Interval]
Inmean I
w rain | .1108697 .0270301 4.102 0.000 .0578917 .1638476
w cddavg | .0136836 .0205099 0.667 0.505 -.0265151 .0538823
w hddavg | .0243572 .0123227 1.977 0.048 .0002051 .0485093
purchpwr | .1625227 .4202524 -0.387 0.699 -.9862024 .6611569
d isoup | .1672659 .3035997 -0.551 0.582 -.7623103 . 4277786
da delt | .0000336 .0000265 -1.268 0.205 -.0000854 .0000183
isoemerg | .1041115 .1290283 0.807 0.420 -.1487794 .3570024
loadshed | .5897324 .325542 1.812 0.070 -.0483182 1.227783
cons | 8.266055 .1514483 -54.580 0.000 -8.562888 -7.969222
_lnalpha |
_cons | 3.770238 1.037021 -3.636 0.000 -5.802761 -1.737714
alpha .0230466 [_lnalpha] cons = In(alpha)
(LR test against Poisson chi2(l) = 1.239576 P = 0.2656)
Table 15: SCE Negative Binomial Regression, SAIDI Monthly Data as Dependent Variable
The negative binomial regression using the monthly SAIDI data for SCE
indicates w_rain and wjiddavg are significant at the .05 level. An increase in
either wjrain and wjiddavg increases the likelihood of power outage. These data
reflect SCE’s reliability vulnerability to bad weather and hot summers, loadshed
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
168
is significant at the .10 level. This is less compelling than the .05 level
significance of the weather variables, but is suggestive of the linkage between
conditions resulting in loadshed and power outages. A loadshed incident means
an increased likelihood of power outage. The conditions that cause a load
shedding decision are the same conditions that might result in power instability.
SDG&E’s negative binomial regression was possible using annual SAIDI
numbers or monthly SAIDI numbers. The monthly numbers provide more
complete information including the proportional total SAIDI contribution plus
major outage SAIDI values reported in a particular month.
SDG&E Negative Binomial
Log Likelihood = -242
Regression
8529162
Number of obs
Model chi2(7)
Prob > chi2
Pseudo R2
108
32.94
= 0.0000
= 0.0635
saidimo | Coef. Std. Err. z P>|z| [95% Conf Interval)
Inmean |
w rain | .0777926 .0197926 3.930 0.000 .0389998 .1165854
w cddavg | .0465772 .013775 3.381 0.001 .0195787 .0735756
w hddavg | .0217847 .0087187 2.499 0.012 .0046965 .038873
purchpwr | -.7348398 .6489024 -1.132 0.257 -2.006665 .5369856
d isoup | -.1000964 .1790826 -0.559 0.576 -.4510919 .2508991
da delta | .0000148 .0000471 0.314 0.754 -.0000776 .0001071
isoemerg | .1062316 .0783979 1.355 0.175 -.0474255 .2598887
_cons I -7.57151 .3918648 -19.322 0.000 -8.339551 -6.803469
lnalpha |
cons | -6.5962 12.92718 -0.510 0.610 -31.93301 18.74061
alpha .0013655
(LR
[_lnalpha]_cons = In(alpha)
test against Poisson, chi2(l) = .0060817 P = 0.9378)
Table 16: SDG&E Negative Binomial Regression, SAIDI Monthly Data Dependent Variable
The SDG&E results again suggest the relevance of weather-related factors
only, as did the logit regression for the utility. Rain, winter heating, and summer
air conditioning are all positively associated with increases in power outages.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
169
Negative binomial regression options for SMUD were limited by the
available outage data. Only annual SAIDI numbers are available.
SMUD Negative Binomial Regression Number of obs 72
Model chi2(9) 7.00
Prob > chi2 = 0.6367
Log Likelihood = -98 .6882526 Pseudo R2 = 0.0343
saidiann I
_ _ _q.
Coef. Std. Err. z P> I z 1 (95% Conf. Interval]
_lnmean 1
w rain | .0779006 .0602808 1.292 0.196 -.0402476 .1960489
w cddavg I -.0354078 .0399329 -0.887 0.375 -.1136749 .0428593
w hddavg | -.0219885 .0224537 -0.979 0.327 -.0659971 .02202
d isoup | -.1645104 .4034458 -0.408 0.683 -.9552497 .6262289
purchpwr | -.4605439 2.255769 -0.204 0.838 -4.881771 3.960683
dsm ee I .0424614 .118217 0.359 0.719 -.1892397 .2741625
dsm load 1 -.019418 .0949316 -0.205 0.838 -.2054804 .1666445
loadshed 1 .128586 .1144344 1.124 0.261 -.0957013 .3528732
isoemerg I .2186298 .1433666 1.525 0.127 -.0623635 .4996232
cons I -8.653476 1.224407 -7.067 0.000 -11.05327 -6.253683
Inalpha I
_cons | -16.43968 1.419879 -11.578 0.000 -19.22259 -13.65676
alpha 7.25e-08 [_lnalpha] cons = In(alpha)
(LR test against Poisson chi2(1) = -2.21e-06 P
.)
Table 17: SMUD Negative Binomial Regression, SAIDI Annual Data as Dependent Variable
The model fails the Chi-square test, and so these results cannot predict or
explain any relationship between the independent variables and SMUD’s reported
SAIDI figures.
For LADWP, negative binomial regression was possible using annual
SAIDI numbers. For the analysis, LADWP data were reduced to months with
SAIDI information, January 1996 through December 1999.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
170
LADWP Negative Binomial
Log Likelihood = -189
Regression
5499773
Number of obs
Model chi2(6)
Prob > chi2
Pseudo R2
48
39.81
= 0.0000
= 0.0950
saidiann I Coef. Std. Err. z
P> 1 z 1 [95% Conf Interval]
Inmean |
w rain | .0390235 .0208504 1.872 0.061 -.0018425 .0798895
w cddavg | -.0033781 .0101245 -0.334 0.739 -.0232218 .0164655
w hddavg | -.0073694 .007219 -1.021 0.307 -.0215184 .0067796
purchpwr | -2.096952 .7124113 -2.943 0.003 -3.493253 -.7006519
d isoup [ -.0352226 .1546399 -0.228 0.820 -.3383113 .267866
isoemerg | .1189877 .0503952 2.361 0.018 .0202149 .2177605
cons I -4.946306 .1279871 -38.647 0.000 -5.197156 -4.695456
lnalpha |
cons | -3.679168 .3551151 -10.360 0.000 -4.37518 -2.983155
alpha .025244 [ lnalpha] cons = In(alpha)
(LR test against Poisson chi2(l) = 24.34145 P = 0.0000)
Table 18: LADWP Negative Binomial Regression, SAIDI Annual Data Dependent
Variable
purchpwr is significant at the .005 level. The LADWP data again suggest
that purchased and contracted power seems relevant to its reliability profile, as
was discussed above in the logit regression analysis, albeit in the opposite
direction. The negative binomial analysis excluded effects from outside the
distribution company, using instead internal customer outages. The positive
correlation between purchased power and outages suggests that within its own
control system, LADWP’s strategic internal use of purchased power and
contracted power improves its system reliability.
The independent variable w rain is significant at the .10 level. This is less
compelling than the .05 level significance of the purchased power, but suggests
that rain is positively correlated with reliability outages within the LADWP
control area. Other than this weak effect, the LADWP results again show no
statistical significance of weather on electric system reliability. As noted earlier,
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
171
this might reflect the lack of incident-specific SAIDI data available rather than
being truly indicative of the role of weather in system disruptions.
This analysis of SAIDI allows the ISO emergencies to be included as an
independent variable. This variable is significant at the .05 level and supports the
concept that ISO emergency actions have effects on municipal utilities. These
actions are allowable for the ISO control area and have been invoked by the ISO
under rules for emergency, low generation margin operations. Municipal utilities
are insulated, but are not islands, and must share in these reliability disruptions.
Note that unlike the logit regression, which finds dJSOup significant, the
negative binomial regression does not. The distinction between outages and
margin reduction is important. An early reliability assessment that focused solely
on outage would miss the degradation effect that is captured by also considering
marginal events such as load shedding and declared emergencies.
Outage Information
The following sections discuss the outage information used in constructing
the dummy variable representing reliability disruptions for the logit analysis, and
graphically compares SAIDI outage data across utilities. NERC outage
information is also detailed.
Outages and Emergencies
Similar but non-identical outage reporting occurs between municipal and
investor-owned utilities. The California PUC requires more comprehensive data
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
172
for the IOUs subject to its requirements; this additional information is included in
the IOUs regression results. All five utilities respond to ISO emergencies;
however, the ISO has limited authority to direct operations of the munis. It
cannot pass on the costs of energy it purchases in emergency operations to non
members. Annual data compared across the five utilities illustrate several points.
SAIDI comparison between utilities
180
120
® §
o
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
LADWP —®— PGE —A—SCE SDGE -* -S M U D
Figure 1: Multiple Utility Data, 1990-1999, SAIDI Comparison
The annual data show large differences in average system interruptions,
measured as minutes per customer per year. All the utilities show an
improvement from 1998 to 1999 in unplanned system interruptions.
The utility with widest variability, SMUD, underwent massive
reorganization and re-engineering in anticipation of electric restructuring; a 1997
management report made public as part of a lawsuit detailed the difficulties of the
time and transition (Peyton 1998). SMUD attributes much of the 1999
improvement to a shift in their tree trimming program. This reduced tree-related
storm outages by 77% and brought SMUD from the 1998 worst quartile SAIDI
ranking to the 1999 best quartile SAIDI ranking among 45 utilities measured by
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
173
Theodore Barry & Associates (TB&A). The operational change came in response
to the California PUC’s General Order 95, Rule 35 on distribution line clearance
and vegetation management.
SMUD’s SAIDI improvement and its relation to the PUC’s rule is
interesting because tree-related outages may be categorized as weather-related,
such as trees thrashed by a winter storm and falling on an electric line. An
argument can be made that many weather-related reliability events are reducible
based on UDC practices and operational priorities. G.O. 95 required an 18-inch
minimum clearance; SMUD increased their target clearance as part of an effort to
assure the 18-inch minimum clearance would be maintained on a three year tree
trimming cycle. This change lead to many trees being radically pruned causing
public outcry and some moderation of the policy.
How relevant was weather in creating conditions for ISO energy
emergencies? 1998 and 1999 were hotter than normal, but 2000 was a cool
summer, according to the Climate Prediction Center at the National Oceanic and
Atmospheric Administration (NOAA). The logit analysis suggests that weather is
a consistent factor, and the NERC DAWG data do not show any increase in
outages in recent years due to weather factors. Despite this, earlier in the year
political leaders tried to blame the weather, not restructuring flaws, for the year
2000 power emergencies. State Senator Steve Peace wrote in an editorial opinion
piece:
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
174
You see, nothing could change the fact that we did not have enough
generating capacity in the state .. . The summer of 1999 was relatively
cool and large hydroelectric supplies cloaked a smoldering crisis ....
(Peace, 2000)
Rather than the weather, the focus in 2000 has been on creating
adequate generation margin. In 1996 Senator Peace and other advocates for
AB 1890 stated that the state had a wasteful excess in generating capacity.
The restructuring goal was to align more precisely generation capacity and
demand through the market mechanisms. Generation margin is viewed both
as an aggregate average power capacity or a fluid hour-by-hour capacity
based on plant or market availability. But the goal of maintaining generation
margin has been overshadowed by price stability due to consistently higher
around the clock prices seen throughout the fall of 2000.
NERC Disturbance Data
The DAWG (Disturbance Analysis Working Group) database of the North
American Electric Reliability Council (NERC) provides summary information
from utilities’ reports to the Department of Energy on major disturbances
throughout the national transmission grid (“bulk electric systems of the electric
utilities in North America”, NERC 2000). Canadian utilities voluntarily report on
electric emergencies. The years from 1990-1999 produced the following profile of
large incidents. For each year, the total number of reports is given, the number of
reports involving the Western Systems Coordinating Council (WSCC), the
umbrella organization for the western states including California, and the number
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
175
and details of incidents for the five California utilities included in this research.
The numbers for 2000 were not available at the time of this analysis. NERC data
involves major outages, the type that are excluded from the CPUC SAIDI data.
“Firm load shedding of any amount is rare.” (Gene Gorzelnik of NERC,
personal correspondence) It is not surprising few reportable incidents have
occurred.
Approximately half these reports involve severe weather or other natural
disasters. Many reports are classified as “unusual occurrences” that reflect
unexpected line equipment behavior that resulted in generation equipment
shutting down automatically. In some but not all cases, a loss of generation
resulted in under-loaded lines, triggering customer outages. In other cases, the
fault was isolated to the generation site or nearby transmission grid with no
customer outage occurring.
In 1990, 34 reports were filed, three involving the Western Systems
Coordinating Council (WSCC), one involving a California utility outside the five
included in this research. This was not an outage but a one hour and twelve
minute load disruption, a frequency variance from 59.871 Hz to 60.0059 Hz. The
cause of this disruption was an earthquake in the Los Angeles area.
In 1991, 31 incidents were reported, four from the WSCC, one involving
176 interruptible customers and California Department of Water Resources
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
176
pumping operations. One of the other three reports involved PG&E, which
dropped 400 MW of interruptible load.
In 1992, 22 reports nationwide included eight from the WSCC, one for
LADWP and one for PG&E. The LADWP load interruption was caused by a tree
on fire, which tripped a line, also resulting in an underfrequency load in Burbank.
The 35-minute PG&E outage involved 530 MW of load interruption due to
unknown causality.
In 1993, 24 reports included eleven WSCC incidents, one for PG&E and
one for LADWP. In the 36-minute LADWP case, a line fault caused an
interconnecting line to block, and 1,680 MW of generation was turned off but no
customer load was lost. The PG&E report involved Diablo Canyon unit #1
tripping 1,077 MW off, and a frequency abnormality lasting 55 minutes resulted
in a loss of 30 MW of electric service to 29,000 customers.
In 1994, 28 incidents were reported including ten WSCC reports. One was
a major multi-agency outage affecting SCE, PG&E, and LADWP, one LADWP,
and two PG&E. The LADWP outage was the result of the 6.6 magnitude
Northridge earthquake. Of the two PG&E outages, one was due to a large brush
fire tripping two large transmission lines moving power from four generating
units; restoration was not complete until nearly 18 hours later. The other PG&E
load reduction affected interruptible load only. The multi-agency outage involved
a series of power line overloads and low voltage culminating in a 1,500 MW
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
177
power surge through Arizona and California toward Oregon. Customer load was
restored “within minutes”.
In 1995, 21 reports include eleven WSCC incidents, one PG&E, one
LADWP. The PG&E interruption was due to a lightening fault at a powerhouse;
only customers of Sierra Pacific in the West Reno, Nevada area were affected by
this fault. The LADWP interruption involved a switchperson’s error in
coordinating operations with the City of Glendale. Only Glendale residents were
affected by this one-hour outage.
In 1996, 28 reports included twelve WSCC incidents, one LADWP, and
two major interconnection disturbances. The LADWP incident resulted from a
“flash” to a palm tree during a maintenance cycle. In the first WSCC
interconnection incident on July 2, an undetermined failure cascaded and created
multiple energy “islands”. The largest such island included California, southern
Nevada, Arizona, New Mexico, and El Paso Texas; the majority of the one
million customers had service restored within 30 minutes, but full restoration took
nearly seven hours. The second major incident occurred on August 10. This was
a hot day throughout the West, and a number of “random multiple transmission
line outages” led to growing voltage oscillations, that resulted in five system
interties going down. The ripple effect resulted in outages to 7.5 million
customers; post outage analysis has led to seven changes in WSCC system
operations to prevent a recurrence of these operating conditions.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
178
In 1997, 29 reports included seven WSCC incidents, two PG&E and one a
combined LADWP and SCE incident. The first incident was an act of sabotage,
shots fired into a substation in response to the Timothy McVeigh verdict. The
second incident appears also have been deliberate sabotage, the manual turning of
39 valves in a substation. The investigation included evaluation of the security
and duplication of substation keys. The LADWP/SCE outage was caused by an
airplane hitting a very high voltage transmission line near Hesperia. While most
customer demand was restored within 20 minutes, the final restoration took over
13 hours.
In 1998, eight reports included one WSCC, which involved a PG&E
interruption. This was the December 8 outage in the San Francisco area resulting
from “human error” in a scheduled maintenance. This incident is also significant
as an early indicator of the coordination needs between the ISO and PG&E;
several procedural and communications failures worsened the outage (ISO
1999c).
In 1999, 17 reported disturbances include three WSCC reports, one from
the Califomia-Oregon intertie and one on a California ISO Stage One Emergency.
In the intertie incident, 743 MW of generation was removed, but no customers
were affected. The hot June 29 Stage One Emergency declaration requested the
public to conserve energy through the afternoon only. No forced outages
occurred.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
179
This summary of the NERC DAWG database reports suggests that early
restructuring activities appear to have not resulted in any significant increase or
decrease in the number of severe bulk electric transmission system incidents. In
other words, significant reliability problems are localized, such as the San
Francisco outage of December 1998 and the Bay Area congestion and generation
weakness in 2000.
Findings Regarding Institutional Changes and Comparisons
The following sections discuss the institutional changes and analyzes data
of particular relevance. These changes include the introduction of the PX and the
ISO, the hypothesized changes in contracting for power, shifts in the use of
energy efficiency and load management, and the ability of utility customers to
contract with alternative power providers.
New Institutions and Changes in the Traditional Utility Companies
This section discusses the creation and impacts of the new structures
designed to provide market and grid management services in response to the
California model of limited restructuring. The new agencies discussed are the PX
and the ISO; the affected entities are the restructured IOUs and municipal utilities,
and the new generation owners.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
180
The Power Exchange (PX)
The relatively low political and public profile of the PX maintained
through 1999 was lost during the year 2000. The displacement of routine daily
purchase from the PX to the ISO’s spot market, and the escalating bid prices led
to scrutiny of the PX’s market clearing price bid process. Differing price caps on
power purchases through the PX, the ISO, and new options for direct purchases
between the utilities and suppliers were considered in the summer of 2000 to
reduce market problems. But the financial insolvency of two of California’s
IOUs caused operating problems at the PX which no longer received the IOU
funds to pay power marketers.
The PX’s task was peripherally relevant to reliability.2 6 Negative effects
from the California restructuring model on power purchases began to be observed
in the changing use of the PX day ahead model. The differing price structure,
persistent underscheduling by schedule coordinators, and the mandatory
reliability purchases for grid balance shifted selling to the ISO. This taxed
resources at the ISO, which anticipated less than 5% of the power would be
bought through their manual process, not the 30% and higher seen in the summer
of 2000; fewer transactions were managed at the PX.
2 5 la a January 1999 Assembly Briefing Book and associated presentation, the PX described itself
as a “competitive marketplace . .. through an electronic auction.” It also describes what it is not:
“... NOT a government agency .. . NOT a utility . . . NOT a generating facility.” See “California
State Assembly, “Electric Restructuring,” section 20.
2 6 Ibid. In 1999, 80% of the purchases were described as PX transactions, with spot transactions at
the ISO responsible for system reliability.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
181
Some wondered whether the PX power auction model contributed to the
increasing price and decreasing availability of sufficient power in California.2 7
Under the PX’s market clearing price (MCP) design, the highest accepted bid was
deemed the MCP and all accepted bidders were paid at that price. This approach
may have raised the average cost per MW as opposed to an alternative model
which would take each accepted bid at the offer price, but the last half of 2000
showed a steady movement of sales from the PX to the ISO. Since the ISO does
not follow the same bidding process, and the prices paid at the ISO have risen
steadily, the auction model does not appear to be a significant factor.
Legislation proposed in the 2001 California extraordinary session on the
problems in the electricity market and restructuring design would establish a state
power authority that will largely displace the financially troubled PX, and move
its operations to the Department of Water Resources (AB1X, Keeley).
Testimony offered at legislative hearings on the bill has stated that the PX’s staff
may be hired by the DWR as part of the state’s response to the financial problems
of the IOUs and resultant generation shortfall.
The California Independent System Operator (ISO)
Data from the ISO reveal margin fluctuations and price effects. In 2000,
net imports dropped from its previous high levels of 7,000 MW on a summer day
to extremely low levels of 300-600 MW. This and similar fluctuations could be
2 7 A proposal to revise PX bidding protocols was among 32 different potential solutions to the
pricing problems presented at the California Legislature’s August 10, 2000 Joint Informational
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
182
verified by any interested party on the ISO’s website by reviewing peak hours
(e.g., 4pm on a hot day) and checking for imports and exports out of the state.
Each hour shows pricing as well, e.g., the day ahead market only $400 per
megawatt, but $750 for the hour ahead market on a summer afternoon.
The logit and negative binomial regression showed mixed results on the
dummy variable djsoup, there is some degree of clarity available by a time-series
analysis of ISO outages, where the period can be extended out further than the
fully populated data collection used for the regression analyses.
1998 1999 2000
Staged Emergencies 7 4 55
Stage 2 Emergencies 5 1 36
Stage 3 Emergencies 0 0 1
Totals 12 5 92
Table 19: California ISO Emergencies Annual Totals 1998-2000
Table 19 compares the declared emergencies, expressing the degree of
generation supply margin weakness experienced by this agency responsible for
grid reliability. The problems emerging in the second half of 2000 were
dissimilar to the price or availability spikes seen in other parts of the United States
over several earlier summers for both persistence and the steady functional
impairment of the new institutions that had worked well for the preceding 27
months. This successful operation is clearly depicted in Figure 2.
Hearing.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
183
California ISO Declared Emergencies
40
Figure 2: California ISO Emergencies by Month, 1998-2000
Vertical integration provided utilities with a set of tools to minimize risk
of high average price or low average margin. A utility might not choose to
routinely pay $750 per MWh if their end customers can only be charged $75 for
consuming it. The ISO by design doesn’t face those fiscal constraints, and will
order power at this price or higher if the alternative is de-energizing a segment of
the grid. The ISO then charges its member for any and all power bought by the
agency. As the fiscal problems with the IOUs began to emerge, generators began
to balk at selling to them; later, they became reluctant to sell to the ISO as well.
This reluctance translated to a reduction in margin, a threat to system reliability.
The problems of the summer of 2000 were not solely financial in nature.
Many plants had been run hard all summer, annual air emission limits for some
plants were reached in early summer, and routine maintenance was reportedly
delayed (Smith, 2000).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
184
The Newly Restructured Operating Entities
Although the names of California IOUs remain the same, they actually
have become very different entities than the vertically integrated systems of 1990.
In addition to divesting generation, the IOUs have spun off unregulated units to
buy generation or offer energy services. They have revised past practices as well.
For instance, the current posited generation shortfall may not have been created
by any short-sightedness, but rather by design. Earlier in the 1990s PG&E
acknowledged the value of using energy efficiency and load management
(demand side management programs, or DSM) in its operations:
Since the mid-1970s, the Company has expended over $1 billion on
DSM programs, allowing the Company to avoid the need for
approximately 1,600 megawatts (MW) of new generating capacity.
(PG&E Securities and Exchange Commission 10K Submittal for 1994,
20)
A vertically integrated company has an incentive to save costs by
efficiency. If the production cost for generating a kilowatt hour are seven cents,
and reducing the same amount of energy demand costs five cents, the effort of
organizing DSM is an efficient strategy for a utility. This is especially true if the
price to the consumer of the kilowatt hour is about six cents; if the cost of either
DSM or generation exceeds the rate paid, the only distinction is the rate at which
the utility loses money. Unlike the vertically integrated utility, a divested
generator cannot improve its revenue and profit picture by better application of
demand side management to reduce consumption and mitigate peaks. The newly
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
185
rediscovered value of energy conservation, evident in the fall and winter of
2000/2001, derives from the funds it protects from peak power purchases and the
fiscal solvency for the UDCs operating under fixed residential rates.
LADWP Operating Changes
The Los Angeles Department of Water and Power changed their operating
procedures in 1999 to benefit from the deregulated market. LADWP increased its
sales into the energy market from 1.3 billion to 4.9 billion kilowatt hours, nearly a
three-fold increase. LADWP’s general manager has made numerous public
statements regarding his intention to use the high market prices and selling
opportunity to pay down LADWP’s debt. In the lobby at LADWP offices in
downtown Los Angeles, a debt “clock” counts down toward the zero level of
indebtedness that is the utility’s target in 2001. LADWP also launched a 10-year
plan to increase in-basin production in August 2000.
The New Owners of Generation Assets
When SCE, PG&E, and SDG&E sold their fossil plants, some legislators
were publicly confused at the above-book values assessed for these plants. Some
of California’s generation dates back nearly a century. Much of the hydro system
is almost that old. Traditional rate of return monopoly operations meant that each
plant was run, maintained, and repaired on regular schedule in coordination with
all other assets. The high cost of building new plants, and California’s higher-
than-average rate structure, encouraged management to cautiously and
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
186
continuously manage each asset, as if it were going to run for another 50 years,
especially since all prudent costs could be fully recouped within the rate case.
The new ownership of generators took a different approach. The goal of a
generation company is to generate and sell power at the best price obtainable, and
to generate at the least cost. The ISO has struggled with the uncoordinated
maintenance schedule of unregulated generation in its operational goal to
maintain reliability.
Owned and Contracted Power
One of the oft-repeated statements in the discussions of adequate
generation is the claim that California has not added any generation the past
decade. To the contrary, PG&E documents show that in the 1990s, 32 separate
cogeneration projects resulted in signed contracts for 1,031 MW of power
(PG&E, 2000). This brought the total for actively delivering contracts to 4,338
MW, approximately 10% of California’s peak summer load. SMUD added 380
MW of cogeneration capacity between 1995 and 1997.
The greatest negative effect demonstrated so far in the limited
restructuring of power generation has been supply and demand imbalance,
resulting in price fluctuations. There has also been a steady erosion of operating
2 8 This least-cost assumption was a part of a critical discussion in AB 1890, and in all subsequent
bills. California’s environmentally less hazardous, alternative generation costs more than
traditional generation. Support for this has been contained within regulatory and legislative
structures for decades, and AB 1890 could have undone that essential support since clean energy
may not be inexpensive.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
187
margin, that is, the percentage of generation availability beyond projected
demand.
LADWP PG&E SCE SDG&E SMUD
1993 14,714 20,606
1994 4,400 14,930 20,615 3,767
1995 4,400 13,588 21,603 2,157 914
1996 4,700 13,583 21,602 2,403 915
1997 4,700 13,583 21,511 2,403 1,032
1998 3,655 10,938 10,546 2,403 1,127
1999 4,161 6,635 10,474 1,098 1.130
Table 20: Generation Capacity Shifts Since 1993, in Megawatts
Table 20 shows the reported generation capacity for the five utilities
studied in this research. The IOUs divested a substantial portion of their
generation capacity in 1998 and 1999. This information is graphed in Figure 3.
Owned Generation Capacity, MW
25000 — ■ —
20000 A--------- O r
15000 n---------
10000 -
— LADWP
—B—PG&E
—A —SCE
—X—SDG&E
—J K —SMUD
1993 1994 1995 1996 1997 1998 1999
Figure 3: Individual Utilities’ Change in Generation Capacity Over Time
The same generation data, stacked to show cumulative capacity,
demonstrate the reduction of utility controlled generation capacity resulting from
AB 1890 divestitures. The divested generation did not leave the state, but was
managed independently. The ongoing inquiry on the development of market
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
188
power hinges on interpreting the operations and bidding behavior of a small
number of the divested generation owners.
Utility Contribution to Statewide Generation, MW
50,000
40,000
30,000 -
20,000
10,000
1993 1994 1995 1996 1997 1998 1999
— 3K-SMUD
— X—SDG&E
—& ~ S C E
— B —PG&E
— LADWP
Figure 4: Summation of Utilities’ Contribution to Generation Capacity
At the same time that investor-owned utilities divested generation, utilities
experienced a rise in peak demand statewide, in contrast to a trend to reduced
demand the first half of the 1990s. Peak demand for 2000 was lower than 1999.
Cumulative Change in Peak Demand, MW
3000
2500
2000
1500
1000
500
0
-500
-1000
-1500
-2000
1990 1991 1992 1993 1994 1995 1996 1997 1998 1999
SMUD
PG&E
LADWP
Figure 5: Growth in Peak Demand, by Utility
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
189
Sixteen of nineteen data points from 1991 through 1995 show a net
reduction in peak demand, while thirteen of fifteen data points from 1996 through
1999 show a net increase. That is, the first half of the decade displayed flat to
negative short term demand growth, while the second half of the decade
experienced steady growth moderating at the end.
Fluctuations have occurred at the traditional time of peak demand-very
hot summer afternoons and evenings. California has only a few hundred such
peak hours each year, no more predictable than the weather itself. That is,
generally predictable over the medium term but unpredictable with precision more
than a few days in advance.
In July 1999, San Diego Gas & Electric received CPUC approval for pass
through pricing of electricity to its residential ratepayers. This removed
SDG&E’s direct risk from price fluctuation; the customer now bore all risk, with
limited information. Residential customers would receive few price signals
related to the variable cost of generation; the customers had no way of knowing at
the time of power consumption the price offered by the utility on their behalf.
SDG&E had no reason to provide individual incentives for energy conservation;
the aggregate behavior of all customers in the class would determine load,
quantity purchased, and price. The individual customer had no signal; and in the
absence of customer choice, the individual customer had no alternative to the
transacted aggregate price.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
190
In July 2000, the implications of SDG&E’s successful recovery of
stranded costs became apparent. Customer bills more than doubled as SDG&E
passed through the rising cost of natural gas generation. Political fallout was
intense, resulting in the single day construction and Senate passage of AB 2290,
enacting a retroactive rate decrease for SDG&E customers and an investigation of
possible price manipulation on the part of generators. The political action was not
a complete reversal, and SDG&E was still able to pass through some of its high
costs for generation purchases to its largest customers.
The other California utilities studied did not experience San Diego’s
problem for two distinct reasons. The municipal utilities had not divested their
generation and generation contracts; they had less exposure to price fluctuation
risk. In addition, they could sell any excess generation to the Power Exchange
and improve the financial portfolio of their utilities; LADWP sold $140 million of
generated power in 1999, and $35 million in June 2000 alone (Shuit and
Mozingo, 2000). The IOUs faced a different situation than either SDG&E or the
municipal utilities. While Southern California Edison and Pacific Gas & Electric
had sold their fossil generation, this process and the collection of the CTC was not
complete. SCE had sold 13 of 15 plants, representing over 85% of their non
nuclear generation. PG&E’s situation was more complex, but it also had not
completed CTC collection and so was still operating under a rate freeze.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
191
PG&E faced difficulties in its hydropower divestiture and also in its
planned generation additions. The IOU was still designing an acceptable
divestiture of its Sierra hydropower system. This network of dams, reservoirs,
and powerhouses produces 5% of California’s electricity (Woolfolk, 2000), and is
extremely important for summer peak loads. It also has multiple environmental
and mixed use interests that would be impacted by a change in ownership. The
Draft Environmental Impact Review for the hydro divestiture ran to over 1,000
pages. PG&E floated a proposal twice to create a deregulated subsidiary under its
corporate parent company; i.e., a method to sell the hydropower generation to an
unregulated entity likely to operate it in a manner friendly to PG&E.
Nuclear power was excluded from the generation divestures. After AB
1890 enactment, SCE and PG&E continue to own and operate California’s two
nuclear power plants, Diablo Canyon and San Onofre respectively. While long
term storage and decommissioning issues remain for the nuclear industries, these
two power plants provide a small hedge against the price of generation in a
market where high demand results in record average prices for the available
supply. The value of these high capacity units can be seen in market price shift
when an unplanned outage takes a nuclear plant offline.
A second change in generation happened related to contracting.
Discussions during the 2001 extraordinary legislative session sought to keep QF
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
192
contributions online. Some have gone offline as their formerly lucrative contracts
have expired. Consider an example QF.
A hypothetical cogeneration facility has a 20 MW capacity, and requires
15 MW for its own business operations. Formerly, the producer established per
PURPA rules a contract with the utility to sell its excess capacity. Once the pre
restructuring contract expires, the producer must establish a new one under
different rules. Under post-restructuring rules, the firm can either generate and
use only the 15 MW or they can sell 20 MW to the PX and buy back 15. In the
second scenario, the producer picks up charges from the PX for the transaction,
from the ISO for the transmission, and from the utility for non-bypassable
surcharges when it consumes the 15 MW. These new costs may affect whether
expiring contracts are renewed, further exacerbating post-AB 1890 generation
shortfalls.
Large shifts in the contribution of QFs were not evident into the fall of
2000. However, 23% of PG&E’s power came from QF providers per its 1999
SEC 10K filing. AB 1890 and ISO/PX operating rules clearly have altered the
rules and rewards for participating QFs. Less-favorable QF contract terms are also
expected. In the midst of the January 2001 State 3 emergencies and large capital
shortfall for fuel purchases, the governor and legislators sought a 50% reduction
in prices paid to the QFs. The ISO power emergencies suggest these providers of
substantial energy could be important to the continuation of a reliable and stable-
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
193
priced energy supply, yet the fiscal rewards for these contributions are being
reduced.
The role of purchased power at SCE became steadily more important for
SCE throughout the 1990s. "Purchased power and other sources" for SCE rose
from 36.6% of total energy requirements to 53.8% between 1994 and 1998
(Southern California Edisonl998 Annual Report, 40).
The State of California’s entry into the daily purchased power market and
their efforts to build a long-term power contract portfolio have been accompanied
by claims that the UDCs should have entered into long term contracts months
earlier, avoiding the financial problem and subsequent shortfall in offered
generation. There is a rational argument why they did not enter into such
contracts.
Prior to late summer, the UDCs were proscribed from entering long term
bilateral contracts. When this rule was revised, the new guideline from the CPUC
allowed them to participate in such contracts subject to a “reasonableness
review”. Requests from SCE and PG&E for clarification and specifics regarding
contract reasonableness did not result in clear guidance from the CPUC. The
utilities had some concern that the contracts they could get would not be
considered reasonable, and would potentially leave them liable for the long term
cost of the contracts.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
194
Defining a reasonable price for a long term contract in 2000 was not a
simple task. Roseville Electric, a municipal utility, signed a five-year 4.9 cent per
kWh contract with Enron. Utility managers considered this a high cost, but
unavoidable given the growth in the area and Roseville Electric’s commitments to
Hewlett Packard, their largest customer and one the utility did not want to lose to
a third-party power provider.
An examination of the price per kWh SDG&E reported paying for each
year in the 1990s provides information on historical prices for contracted power.2 9
Each year in the 1990, SDG&E’s purchased power price was below four cents per
kWh. The prices offered in summer 2000 contracts were 30% higher than the
highest prices paid for the past ten years, and 40% higher than the average prices
over the same period. Conventional wisdom and experience in 1998 and 1999
suggested that competition would bring lower prices, not higher ones, over the
long term. Five year contracts for 5 cents per kWh were considered no bargains
in the summer of 2000.
Year Price (cents/kWh)
1991 3.5
1992 3.8
1993 3.5
1994 3.7
1995 3.3
1996 3.1
1997 2.8
1998 3.5
1999 3.7
Table 21: SDG&E’s Cost of Purchased Power, 1991-1999
2 9 Information from SEC 10-K filings for San Diego Gas and Electric, 1994, 1997 and 2000.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
195
Note that these prices include nearly two years of ISO operation. In 1998
and 1999 SDG&E was buying their power through these new institutions, and had
divested most of its own generation. In the summer of 2000, five year contracts
were offered at 5 or 5.5 cents per kWh. Given past prices and the expectation that
the 2000 price disturbances were temporary, it is unsurprising that these contracts
did not received much favorable review. Utilities faced the added uncertainty of
consequences should the CPUC decide after the fact the contracts were
unnecessarily expensive. The UDCs made a rational but ultimately expensive
decision to avoid these particular bilateral contracts.
Energy Efficiency and Load Management
LADWP
LADWP does not provide load management and energy efficiency data in
a public forum; repeated inquiries soliciting this information went unanswered.
PG&E
This researcher found only limited load management information available
from PG&E which suggests that approximately 5% of its load has some degree of
load-shedding capacity. No comprehensive records of PG&E’s load shedding has
been provided. However, the summer of 2000s ISO Stage 2 and 3 Emergencies
resulted in public discussion on the customers who would be cycled and the
length of any voluntary or involuntary interruptions. Proximity to schools, police
stations, and hospitals results in customers not being subject to load shedding.
PG&E reached its 2001 voluntary load shedding limits in the first weeks of the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
196
year in the protracted low-margin period of Stage 3 Emergencies. At the same
time, regulatory action exempted southern California interruptible customers from
similar shutdowns from concerns of the economic effects of lengthy unscheduled
business interruptions. Dire warnings for the summer of 2001 from the head of
the ISO and others followed these two reductions in load management capacity.
SDG&E
This researcher could not locate SDG&E load management and energy
efficiency data in its SEC disclosures or other public reporting.
SCE
SCE load shed customers found themselves subject to many times the
number of incidents they had experienced in early years. Some customers are re
thinking the value of the arrangement and are removing themselves from this
program, limiting the flexibility or increasing the cost of the program for SCE. A
new incentive structure will likely required for participation.
SMUD
SMUD has provided the most complete set of data on these demand side
approaches to counterbalancing supply side issues. Several observations can be
made from the data relevant to the introduction of the market.
Reviewing the energy efficiency data shows 1998 was a year of change.
SMUD switched from counting savings in megawatts to counting in kilowatts, a
thousand fold reduction. The load shedding program was altered from a multiple
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
197
level, multiple incentive residential load reduction program into their Emergency
Only Peak Corps. The new program operated with a different goal and new
guidelines. It reduced ratepayer incentive to participate from as much as a $20
credit on their electric usage each summer month to a $10 payment for the entire
summer. From 1994-1997 were 26 cycling days; from 1998 through 2000 there
were only eight cycling days, as shown in Figure 6.
SMUD Variation in Load Shed Events 1994-2000
250
700
150
m
9
100
s 50
0
1
♦
_
nip!
■
♦ A
*
♦
♦
t *
y
*
1994 1995 1996 1997 1998 1999 2000
♦ MWredux ■ MWavail
Figure 6: SMUD’s Use of DSM Dispatchable Load Shedding, 1994-2000
SMUD’s preparations for competition took two forms. First, the utility increased
its contractual capacity to shed load by over 50%, from 135 MW to 211 MW.
This gave it more flexibility for hot weather, and protection from high peak
prices, should the utility lack sufficient hydroelectric resources to meet its own
needs. Second, it reduced both the cost of the program, and the frequency of its
utilization. This allowed the utility to maximize its own summer revenue through
increased consumption, building its “Rate Stabilization” fund.
A slightly different presentation of the SMUD load shed data allows
consideration of the slight difference in the amount of load shed per incident. The
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
198
median megawatts shed in an event also dropped from 55 MW to 42 MW, despite
the increased megawatts available for load shedding, as shown in the chart below.
SMUD Load Shedding 1994-2000
250
200
150
♦ MWredux
■ MWavail
S’ 100
♦
1994 1995 1996 1997 1998 '99 2000
Discrete Incidents
Figure 7: Comparison of Discrete Incidents, SMUD Load Shedding, 1994-2000
SMUD was not subject to AB 1890, and neither joined the ISO nor
divested its generation. But its behavior post-AB 1890 has been affected by the
changes statewide. Since generation has been separated from regulatory controls,
market price has risen with new demand peaks. There exists a potential for
generators, including municipal generators, to improve their revenue stream as
overall demand rises with flat supply. Between 1991 and 1997 SMUD peak
demand increased by only ten percent. Between 1997 and 1999 the utility’s peak
demand rose by 215 megawatts, another 15 percentage points. As the figure
shows, SMUD has this amount of load available for shedding and twice in 1998
shed 140 megawatts. But in the summer of 2000 SMUD only shed on average 61
megawatts, when wholesale prices were highest on hot days and in accord with
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
199
two Stage 1 and four Stage 2 ISO emergencies. This indicates one possible
source for peak margin reduction in California.
Direct Access
While all the utilities under study have direct access programs, only the
IOUs are required to report their numbers. Neither SMUD nor Los Angeles
Department of Water and Power provided data on their programs. These findings
represent the information on the three IOUs.
Metering and accounting system changes were required in order to
implement direct access. IOUs and municipal utilities that allowed their
customers to choose alternate suppliers needed to develop mechanisms to provide
one another load information. Institutional shifts in core investment could have
included widespread metering changes, beyond direct access customers. This did
not happen; metering retrofits have been limited to large customers on real-time
pricing.
SMUD was the first utility to offer direct access, first to commercial
customers, and then to firms seeking to aggregate retail customers. As a
municipal utility, it was able to design its direct access rules which many
participating firms found arduous. Ultimately direct access ceased to be offered
to residential customers as alternative providers dropped out of the program.
Direct access reporting by investor-owned utilities started in November
1997. Information provided by each utility includes the number of requests for
direct access received and processed, any backlog, the number of customers
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
200
switching from their UDC to an energy service provider (ESP), the number of
customers switching between ESPs, and the number of customers switching from
an ESP back to the UDC. The cumulative trend in customer shifts is graphed
below.
Participation in Direct Access
a! 20,000 -
in
m
e 10,000-
I o
1 - 10,000 -
O
| -20,000 -
£
t
V .{/•'
V
4
$ V'-'
PG&E - S C E SDG&E
Figure 8: Direct Access Customer Shifts, by Investor-owned Utility
The numbers show initial and ongoing participation in the IOU direct access
programs by residential and commercial customers. The year 2000 saw a number
of customers abandon their direct access provider and return to their UDC. The
single greatest shift back towards a UDC occurred in June 2000 when 13,831
switches back to PG&E occurred mainly among residential and small commercial
customers, counterbalancing PG&E’s biggest customer switch toward third party
provision in April of 1998.
The first six months of direct access resulted in less than 1% of residential
customers but 13.1% of large commercial customers switching to alternative
suppliers. By the summer of 2000, the number of residential customers
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
201
participating had increased to 1.7% of residential customers and 13.2% of large
commercial customers. The total load served by direct access providers increased
from 8.2% to 12.2% over the same two year period. Figure 7 charts this limited
statewide shift.
Residential
Commercial
<20 kW
Commercial
20 - 500 kW
Industrial
> 500 kW Agricultural
Total All
Classes
% Direct Access
Customers 7/31/98
0.7% 1.8% 3.3% 13.1% 0.8% 0.8%
% Direct Access
Customers 8/31/00 1.7% 2.4% 5.5% 13.2% 2.6% 1.9%
% kWh Load 7/31/98
0.8% 2.5% 7.6% 21.0% 1.1% 8.2%
% kWh Load 8/31/00
2.1% 3.9% 12.9% 27.5% 6.0% 12.2%
Table 22: Cumulative Direct Access Load by Customer Class, July 1998 vs. August 2000
The difference between residential and commercial customers is generally
attributed to the minor savings offered to residential customers; early estimates of
2.5% savings on their bill were not compelling to make a supplier change. There
were also a limited number of companies interested in signing up residential
customers. This number has since declined further due to difficulty designing a
retail package both attractive to UDC customers and profitable for the alternative
providers. Commercial accounts, which historically subsidized retail customers
by paying above-market prices, have more readily switched.
Survey Results
“Imagine a drought, plus the heat. Imagine no water and no generation.”
Like this respondent, all survey participants had strong impressions of their years
in electric restructuring policy making, and specific expectations about the near
future. Most continued to voice confidence that a well-functioning market would
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
202
in time provide the benefits of restructuring. They were unsure when and how a
cohesive California leadership would make the changes from the existing market
design that are necessary to assure correct market function. They also focused on
the appropriate role of policy makers and regulators in building public
understanding and support, and limiting damage by providing near-term
corrections to the market problems.
The survey instrument in Appendix A is modeled from the findings of the
Department of Energy’s Power Outage Study Team (Department of Energy,
2000) with reference to California’s restructured design and implementation
activities, especially high-profile actions carried out in the Summer of 2000.
Approximately 40 contacts are reported in Appendix B. Some provided brief
assessments of California’s restructured environment, others are technical
specialists, still others have policy roles within specific sub-areas of the electric
system. Several commented on only a few questions related to their area of
expertise and experience. Information from these partial responses are reflected
in this section. A summary of survey questions and the responses of the twelve
who were formerly interviewed can be found at the end of this chapter.
Approximately half of the policy makers approached to participate in the survey
agreed to be interviewed. A discussion of non-respondents (who are of course not
listed in Appendix B) concludes this section.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
203
Participants have had long experience and expertise in electric systems
policy processes. Most have spent their careers in utilities, government,
regulatory and advisory agencies, and similar environments relevant to the highly
controlled utility industry and dissimilar from competitive deregulated industries.
A few represented consumer or business interests. These participants have had
limited experience in regulatory restructuring outside California’s electric
restructuring process. They are practical and action-oriented, generally optimistic
on the value of markets but without extensive knowledge of technical market
design and function. Most remember lines at gas stations during the 1970s oil
embargo. They are not enthusiastic supporters of price caps because of the
tendency of caps to constrain supply. Reliability and low price are inherently
dependent on abundant supply to meet demand. Several spoke of the negative
effects of politics in the summer of 2000, and repeated pressure on the ISO board
to set particular caps at the governor’s request even though they expected this to
depress supply, decrease ISO operating margin, and negatively impact reliability.
In their organizational roles respondents are generally reactive toward
legislative processes. That is, they avoid getting involved in the drafting or
passage of legislation when not specifically requested to participate unless they
are aware of direct effects on their organization resulting from the legislation.
Where such effects exist respondents become involved in the legislative hearings
to be sure their opinions and views are included in the legislation, to maximize
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 0 4
benefit and minimize problems for their organizations. Respondents did not
support rapidly drafted and enacted legislation which prevents this evaluation by
all affected parties from occurring. Once legislation is approved, the respondents’
organizations alter their business models to match the rules and legislation. AB
1890’s intent was to bring many new players into the California electric market,
but it also resulted in major shifts in the existing firms.
Many respondents participated in the design of the ISO and the PX, and
have been on the ISO Board of Governors, at the PX, or both. The ISO
stakeholder board as established tended towards factions and became susceptible
to political manipulation. Even those at opposite sides of the spectrum in terms of
their restructuring “religion” agreed on the potential of electric system
restructuring to work and deliver benefits to Californians, but the weak,
California-specific institutions ran into difficulties in delivering this and were
unable to establish processes for correction. The role of the ISO shifted in 2000,
adding to its role as the “air traffic control” of the electric grid major power
purchasing. This was not a result of rational rule making, but emergency
purchasing as sellers shifted from day-ahead to more profitable hour-ahead
scenarios.
Like the ISO, opinions on the role and function of the California Power
Exchange (PX) varied. Generators wanted to deal directly with large customers;
IOUs preferred their historic buying processes, but agreed to participate in the PX
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
20 5
in exchange for getting the CTC authorized to collect their stranded investments.
Noting suspicions and economic models suggesting market power and gaming
behavior with the market clearing price concept, respondents felt the PX needed
intervention to correct the problems that occurred with a “confluence of
circumstances” involving better computer models, strategic planning and strategic
marketing within the rules of the restructured system. Because the ISO
performed the last minute buying and declaring the energy emergencies, it
unfairly was blamed for uncorrected problems elsewhere in the design. The ideas
for correction have been in general discussion and include a larger transmission
pool organization, more flexibility in buying power outside the power exchange.
Many of these ideas were part of the original ISO/PX design discussion, discarded
at the time and returning for a second look in 2001.
The NIMBY-ism rampant in California is among other problems that
respondents felt need to be addressed. A summer 2000 emergency peaker unit
that PG&E proposed siting on a barge in San Francisco Bay failed on both a
NIMBY level and on significant and acknowledged environmental concerns in the
siting of the unit. The NIMBY question is symptomatic of a larger issue central
to this research study: what are the appropriate solutions to adequate supply for
demand in each location, given California’s constraints?
While the answer to this question seemed clear under traditional
regulatory controls, the answer is less obvious under restructuring. There are
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 0 6
interlocking issues that are nowhere so clearly highlighted as in Northern
California, in San Francisco and south along the peninsula. There is insufficient
localized generation; the area draws approximately 700 MW but produces only a
fraction of that demand. Transmission capacity into the area is barely adequate,
especially given expected growth. Land use issues demonstrate fundamental
conflicts over housing, transportation corridors, open space, environmental
protection, seismic instability, and limited tolerance of the local population for
changes in land use that negatively impact property values or quality of life
without commensurate compensation.
Respondents raised the question of defining the correct amount of
generation margin under restructuring. An efficient market correctly establishes
the right amounts of demand and supply, but California’s market is not producing
this margin. The California model of limited restructuring intended to end the
IOU’s supposedly wasteful overproduction. This overproduction was
theoretically a side effect of the cost plus rate of return model combined with
operating caution and a monopoly provision of service. Public discussion at the
legislature or elsewhere does not specifically address the appropriate generation
margin. There needs to be more discussion on rulemaking processes and the
principles of restructuring design. This assurance of margin differs from deciding
how much generation to build in-state. Addressing the balance of supply and
demand and the rational operating margin combines the issues of local generation,
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
207
possible transmission changes, energy conservation, time-of-use load shifting,
load shed incentives, and pricing goals.
California is not an island. Respondents shared the impression that rather
than the California specific ISO, a true regional transmission organization (RTO)
is necessary. California buys significant out-of-state power, siting power in
sparsely populated areas is simpler, but there are western transmission constraints.
Restructuring is an opportunity to consider how the western region interacts.
Respondents detailed negative elements in the post-AB 1890 market
operations that require changes. The pragmatic, reactive nature of the stakeholder
process resulted in a restructured industry design that held up well early on, but
unanticipated consequences could not be corrected. New players developed
behaviors advantageous to their part of the restructured industry. The restrictive
market rules limited flexibility, competition, and secondary options in buying and
selling generation load. The ISO was open for municipal utility membership, but
the potential for “private use” tax liability for municipal utilities joining the ISO
precluded them from exercising that option. The traditional processes did not
map to the new industry model creating a the timing mismatch between
hourly/daily markets and annual/biennial rate cases.
As these cracks appeared, the same pragmatic parties wanted to take
corrective action, but no working process for correction was available. Large
differences of opinion existed on whether and how FERC should intervene in
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
208
California to establish rules or settle disputes on required actions suggested by the
CPUC, the ISO, the PX, the legislature, or the governor. Lacking an authoritative
institutional structure, decisions on jurisdictions began to rest on actual incidents,
precluding preventive approaches that could limit damage. For instance, as
general price increases affected the western states wholesale market, FERC was
requested to intervene.
There are questions on responsibility for future investments. Is
transmission the ISO’s problem? Nothing is committed to the transmission
bottleneck at path 15. Is this problem related to changes in imports from the
northwest, or an indicator of an interstate grid weakness that should be addressed
across states? How do independent, separately operated generation facilities
affect statewide inventory and ability to replace faulty equipment?
Direct access, the ability of residential or commercial customers to select
their generation supplier, was a key promise of AB 1890. Even the municipal
utilities offered direct access programs. This competition for power was supposed
to bring lower cost for generation to all customers. But it turned out to be a
complex system of special meters, credits between utilities, customers, and the
power marketers, and modifications of customer billing systems. New meters
supporting real-time pricing and other products could increase availability and
penetration of direct access programs to residential customers. Two respondents
mentioned early efforts by the UDCs to protect their franchise in metering and
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 0 9
customer billing systems, clashing with new entrants hoping to compete for these
services with new technologies. The result from all this is an absence of
competition for 88% of the state’s generation load.
The CPUC seemed slow and unresponsive in establishing rules for
hedging and bilateral contracts to minimize the exposure of the IOUs to the daily
power market. There was a feeling that the CPUC lacked a fundamental
understanding of structural problems, delaying action. Respondents felt the CPUC
is essential in either supporting the status quo, or in making and moving changes
in the restructured design. They also are supposedly assuring the residential
customer is protected, as both the transmission and distribution portion of the
industry are still fully regulated.
Respondents were interested in getting greater flexibility in generating
options, and variety of generation plants. The regulatory process does not have in
place methods to validate the many types of power products that need to be
supported both on wholesale and retail levels. New plant owners run old
generation differently, and think differently about these assets. Restructuring
uncertainty and price caps are disincentives to improve operations, to add new
generation, to design and operate for fuel flexibility.
Optimistic and pragmatic, the respondents felt most of the problems were
man-made and solvable, but were concerned that elapsed time was making
solutions more difficult. They also felt simplistic analyses presented to the public
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 1 0
understated the fundamental issues and masked the serious work that needed to be
done on these structures. Especially important is a reduction in uncertainty and
instability. Survey respondents acknowledged NIMBY-ism, government red tape,
and slow regulatory response to new ideas. But uncertainty was named as the
most significant impediment to the capital investment necessary to create a
flexible and robust supply and demand market.
In addition to the contacts reported in Appendix B, there were a number of
persons who indirectly declined to participate. In very few cases, these persons
referred the request to another party because their involvement with the California
restructuring design and implementation was limited. In other cases, indirect
contact was made through telephone calls and faxes making it simpler to not
participate; it may be possible that staff acting as gatekeepers screened out
requests. In several cases, direct contact was made and agreement to participate in
the survey was reached, but in follow-up with staff to schedule an interview time
no appointment was forthcoming.
In considering the reluctance of some to participate, reference was made to
Dillman’s tailored design method (Dillman, 2000) for clues as to their decision
making process. Four principal causes seem most prominent. First, the electric
system restructuring process that was planned throughout the early 1990s,
legislated in 1996, implemented in 1998, and went well for two years began
wobbling severely in 2000. Most survey participants and non-respondents had
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
211
been involved in this process and were busily involved in working on remedial
action during the survey period. Second, since mid-2000 there has been much
front page reporting on electric restructuring. Many respondents have been
extremely careful to limit their comments in public and private; the nonresponders
are the logical extreme in this caution. Third, the questions facing restructuring in
California are difficult. Some may have little interest in acknowledging the
difficulty of progressing with restructuring in the current structural and political
climate. Finally, despite the many points of view on electric restructuring, a
relatively small number of active organizations and players shaped the California
restructuring plan. Some may feel that others who have participated in the survey
are reasonable representations of their own points of view, and so move on to
other priorities.
Relation of Findings to Hypotheses
This section summarize correlations between the findings and the
hypotheses. The results suggest that the partial restructuring of electric service
has affected system reliability. The logit regression showed limited correlation
between system reliability disturbances and the new institutions and behaviors
that have developed between the passage of AB 1890 in California and the end of
the period for which statistics are fully available. The steadily diminishing
margin foreshadows serious potential for reliability disorder in the near future.
The secretive atmosphere of the industry evidenced in the missing data may mask
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
212
other difficulties as well. While early restructuring went well, several
problematic decisions have created instability. These include the constraints on
contracting generation, the requirement to sell only to the PX, and a daily auction
that gave all the successful bidders the same high price resulting in fiscal distress
and reluctance to sell at all to financially distressed IOUs. Evidence exists that
the simple market rules could be used to gain a higher price than a competitive
market would yield, and mechanisms to correct these weaknesses were not
promptly designed and implemented. The price of electricity does not
immediately, directly affect electric system reliability. But the protracted high
prices resulted in payment defaults by two of California’s IOUs, destabilizing the
supply of generation. Each hypothesis is discussed in turn below.
HI: The rate freeze imposed upon investor-owned utilities has reduced system
reliability.
The logit regression showed a difference in reliability based on utility
status as a municipal utility. The cause seemed more likely long term
contract constraints on the IOUs rather than the rate freeze, as the IOU rate
freeze created political pressure on municipal utilities to retain their low
rates, and the effect was found in the regression that included the general
disturbance data. The negative binomial regression showed an effect on
system reliability for the restructuring variables isoemerg (LADWP) and
loadshed (SCE). The SMUD SAIDI data showed major SAIDI fluctuations
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
213
associated in part to business restracturing associated with retaining rate
parity with PG&E and altered distribution operations. San Diego’s flexibility
in passing through the cost of generation led to severe bill increases and
resulted in legislation rescinding the variable generation portion of the bill;
this action would mask incipient effects of the variable pricing. Thus, while
some evidence is inconclusive, HI is supported.
H2: Industry investment has shifted in response to restructuring, resulting in less
investment in reliability-related systems
Limited data are available on a utility-by-utility basis on their internal
capital investments related to reliability. Data that are available suggest new
CPUC rules such as the tree trimming rule discussed by SMUD officials have
improved reliability. These are separate from restructuring activities, per se.
The hypothesis suggested spending might shift to new systems, such as
automated meter reading (AMR); instead, money collected through the CTC
has been used by the parent companies for other purposes, outside the new
UDCs, such as building up their unregulated generation subsidiary firms or
-2A
utility debt reduction.
3 0 John E. Bryson, President and CEO of Edison International, testified before the Assembly
Subcommittee on Electrical Energy Oversight on February 9, 2000. He stated then that capital
investment at the utility has shrunk since “deregulation” from $10.6 billion in 1995 to $7.1 billion
in 2000, reflecting equally the contraction in SCE’s asset base and EIX’s stock repurchases. He
also noted that T&D capital expenditures increased each year, exceeding the 1995 general rate
case authorization by $500 million.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 1 4
The ISO fills the role of a unified transmission agency directed to
provide reliability. In creating the ISO, the partially restructured California
electric industry invested in a new reliability-related system. One survey
participant stated that the ISO’s reliability mandate has shifted some
responsibility from utilities to the ISO, which then bills back the utilities for
its expenditures.
There have been funding and investment shifts, but it is not clear
whether these constitute less investment related to reliability. There may be
double-spending or under-spending if utilities feel guaranteed adequate
capacity is the ISO’s role; and misallocation could increase with utilities, the
ISO, and the State of California each entering into hedging contracts and
peak load management. The secretive atmosphere of the industry evidenced
in the missing data may mask other difficulties as well.
Audits of PG&E and SCE released in February 2001 cataloged funds
shifted between the parent companies and the UDCs.3 1 Even without specific
details of the investment reduction, the analysis of SAIDI data for the IOUs
provides no evidence of reductions in necessary reliability systems. While
insightful, the observations or testimony of restructuring participants do not
3 1 The CPUC-commissioned audit of PG&E noted that $4.0 billion was transferred from PG&E to
the parent corporation from 1997 to 1999, allocated to $1.5 billion in dividends and $2.5 billion in
stock repurchases. An additional $79 million was transferred from the parent corporation to its
Energy Services subsidiary, and $650 million to its Elm Power Corporation subsidiary in 1999
(VI-5).
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 1 5
provide any test of data for this hypothesis. Limited data precludes drawing
any conclusions about H2.
H3: Increased contracting for power has reduced system reliability
The logit and negative binomial regression show effects of purchased
power on system reliability. The logit analysis included effects from outside
the distribution company, while the negative binomial regression included
only distribution company outage data. The positive correlation between
purchased power and system disturbances and the negative correlation
between purchased power and outages at LADWP suggests that the external
forces interfere with LADWP’s strategic internal use of purchased power and
contracted power. Negative binomial regression also found a negative effect
of the change in purchased power on outages for PG&E.
As noted above, the investigation period from the mid-1990s through
the summer of 2000 shows substantial degradation of marginal capacity and
stops just short of the period of fiscal and operational malfunction of the
restructured environment. The type of power purchasing and contracting
allowed under AB 1890, rather than the mere change of ownership, is a
contributing factor to the marginal capacity and pricing issues.
H3 is supported, and a corollary hypothesis focused on only reduced
long-term contracting and constraints on contracting would be supported.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 1 6
H4: The new intermediary structures (ISO and power exchange) have reduced
system reliability.
The logit regression showed limited correlation between system
reliability disturbances and the new institutions and behaviors that have
developed between the passage of AB 1890 in California and the end of the
period for which statistics are folly available. While the logit regression does
not uniformly implicate these new structures, the relevance of the djnuni
factor highlights the possibility that these structures and the rules by which
they operate may be contributors to the instability experienced at the end of
2000 and beginning of 2001. Additionally, the cross-utility analysis showed
it to be statistically significant and positively associated with increased
reliability disturbances.
The negative binominal regressions found dJSO up was not
significantly associated with SAIDI data for any of the five investigated
utilities. The logit regression results show the variable d ISOup is
significant. This dummy variable separates disturbance data If om before and
after the new institutions were established. This significant variable does not
necessarily indicate that the ISO or the PX increased disturbances. This point
in time is consistent with other factors that could be relevant in the cross
utility analysis, such as strong economic growth in California, or the
reduction of generation margin in the region. The negative association
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
217
between dISO up and system disturbances at LADWP could represent
changes at the municipal utility after S. David Freeman became general
manager. While these other factors may contributed to the statistical
significance of dISO up, the possibility that the new institutions are the
relevant factors in the increase in reliability disruptions cannot be rejected.
Thus, H4 cannot be rejected.
H5: The introduction o f direct access customers has decreased system reliability.
Only the IOUs have published numbers on direct access customers.
The logit regression results for SCE and PG&E were not valid, and the
SDG&E data did not show any significance in the direct access programs in
place. Customer interest in direct access programs as implemented in
California has lessened. It is not known from public information whether
customers withdrew from alternate power suppliers due to price changes,
suppliers exiting the California market, or other causes. Thus, based on
available information H5 is not supported.
H6: The reduction o f DSMpractices does not affect decreased system reliability.
Much has been published about the need for increase demand side
management to ease the generation margin shortfall. The logit regression
results from data analyzed in this dissertation do not indicate that reported
DSM energy efficiency practices have had a significant impact on system
reliability. However, only limited information on these practices was
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
218
available for this research. In SCE’s negative binomial analysis, the sole
IOU reporting its load shedding showed it to be a significant component. The
evidence on H6 is inconclusive.
H7: The separation o f generation from distribution does not affect system
reliability.
The logit regression results are inconclusive for the state’s two largest
utilities. The logit regressions for the cross-utility analysis significantly
indicate municipal utility status is negatively associated with reliability
disturbances. Municipal utilities did not divest their generation. However,
many other differences between the IOUs and municipal utilities exist; it
appears that the range of contracting options for purchased power is more
important than the ownership of generation.
Evidence exists that the simple market rules could be used to gain a
higher price than a competitive market would yield, and mechanisms to
correct these weaknesses were not promptly designed and implemented. The
price of contracted electricity does not immediately, directly affect electric
system reliability. But the protracted high prices resulted in payment defaults
by two of California’s IOUs, destabilizing the supply of generation.
As part of restructuring activity, each IOU established a parent
holding company. Between 1996 and 2000 they shifted large sums collected
through rate surcharges for the rate reduction bonds and the CTC from the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
219
UDCs into the parent firms; this money was then spent in part acquiring new
generation assets in California and other states. For the period included in
the regressions, this resource allocation has not resulted in decreased system
reliability for the UDCs. The decreasing margin of generation availability in
the second half of 2000 did not result in increased outages, and did not appear
directly related to any divestiture of generation. The steadily diminishing
margin foreshadows serious potential for reliability disorder in the near
future. Thus, H7 cannot be rejected.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
220
Chapter Notes
Summary of Survey Responses
Responses follow from the twelve individuals formally interviewed.
These policy and operations leaders in the restructured utility environment present
their views on California restructuring and issues of system reliability. Each
question from the survey is followed by aggregated responses. Some questions
were asked of fewer than all respondents. This resulted from time constraints or
extended responses to other questions based on individual interests and expertise.
1. Please describe your organization’ s participation in the electric industry
and your individual role.
Respondents represented public and private utilities, state and federal
agencies including the California ISO, associations for municipal power and
independent producers, and advocates of large commercial businesses. Their
areas of expertise included generation, transmission, distribution, and energy
services, and the management of utility organizations.
Respondents’ positions in their current and past organizations included
executive and senior management, advocacy, and contract roles. Most have
had experience either as original members on the ISO Board of Governors or
were newly appointed to that board of 27 stakeholders. Many present their
organizations’ perspective regularly in public forums such as testifying before
the California Senate or Assembly in legislative hearings. More than half
were personally involved in the pre-AB 1890 restructuring processes: the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 2 1
CPUC’s early memoranda of understanding (MOUs) with PG&E and SCE, its
December 1995 decision, and the activities of the WEPEX organization, all
precursors of the California legislative action in AB 1890.
Representative responses follow: “We buy and sell power; we
distribute power to our customers and provide customer services, billing and
metering.” “We market the power that's produced.” “We run the real time
market.” “Our members all have interruptible service.” “I was an original
governor at the ISO.” “I've been involved throughout.”
2. What major actions taken by your organization derive from the passage o f
AB 1890?
For the most part, respondents began by describing actions their
organizations took in anticipation of restructuring with the PUC MOUs and
AB 1890. Utility respondents discussed the early importance of collecting
stranded costs and the need to be competitive; the perceived necessity of
establishing a competitive environment extended to respondents with
municipal power interests. “We basically accepted competition as being the
direction of the industry, the outcome that the state was going to see”
Respondents spoke of major organizational restructuring as well.
Fundamental organizational change came to the IOUs. They divested their
fossil generation, joined the ISO, offered direct access to their customers, and
created parent corporations and unregulated affiliated companies. The ISO
and PX were also established by AB 1890. Municipal utilities had the option
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
222
of joining the ISO, but there was a major catch: risk of so-called private use
designation by the IRS. Public bond funds can only be used for public
projects; joining the ISO could carry the re-categorization of certain public
debt incurred for facilities turned over to the ISO as no longer public, thus no
longer tax-exempt.
For association respondents, one shift has been an increased
involvement before a panoply of regulators and rule makers: “.. .representing
[the membership] at FERC, the PUC, the ISO, the legislature, providing the
collective voice.” Other responses included new optional roles undertaken
based on expectations of how restructuring would affect their organizational
operations: “We became scheduling coordinators . . . three months after
restructuring began, to protect ourselves from the costs restructuring would
create.”
3. Does your organization have an official position on SB1388 (introduced
but not passed), or other active California legislation on electric
restructuring? What has been your position on the recent legislation, and
what type o f legislation do you support going forward?
No respondents discussed SB1388, introduced by Steve Peace to
foreshorten review of proposed generation sites but not actively advanced nor
passed into law in 2000. They discussed instead their organization’s
legislative involvement with other more active measures such as AB970, and
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
223
the special legislation protecting San Diego ratepayers from rising generation
charges enacted September 2000.
Several stated their legislative involvement was limited to only those
measures with immediate, direct relevance to their organizations: “On AB970
we provided comments and had a member testify in person.” “We follow the
legislation, but we don’t get directly involved. ..We just try to make sure we
make whatever noise we can make if things affect us.” They stated that
appropriate legislation is important to the success of restructuring, but “Any
legislation . .. put through in a week has a lot of problems.” (This comment
was offered in reference to AB 1890, but it could serve an equal warning for
the legislation offered and enacted in January 2001.)
4. What is your opinion on the floating barge towed to California by PG&E,
or similar peak production facilities ?
To alleviate peak demand supply shortfall during the summer of 2000,
PG&E decided to site a portable generation facility in the Bay Area. The area
has congested transmission, so PG&E hoped to site the unit near SFO airport.
The portable unit was housed on a barge, located in the Gulf of Mexico, and
would be moored in the San Francisco Bay. In transporting the unit through
the Panama Canal, the barge became a visible, publicly discussed symbol of
the power problems. Respondents voiced concern over several interlocking
issues: inadequate amount of generation to meet electric demand on the
Peninsula, inadequate transmission capacity for the same area, low tolerance
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
224
of the local population for the pollution effects of the unit. There was
recognition of increased load interruptions, and so a need for more peaking
power plants: “Using the plant on an occasional basis is not as bad as
everybody turning on their personal generators, and the other costs of the
power going out.” Without discounting the validity of environmental
objections, several pointed out the emissions of the portable plant would be
less than those of personal generators or of the older operating plants in
California.
The question stimulated several comments on meeting peak demand
and the cost of meeting this demand through the ISO power market. One
respondent felt part of the peak demand issue was system gaming. The ISO
could buy power to balance the grid from California producers or from outside
the ISO’s control area, an out-of-market or OOMpurchase. Under the market
structure, a California generator could sell to a third party power marketer out
of state, who could in turn then re-import the power back to California as an
OOM buy at a much higher price. Regardless of the purchase path or price,
the reliability purchase cost was spread among all load schedulers based on
their load rather than targeted to distributors who had not submitted a
balanced load schedule. Similar to this comment, others noted that ISO cost
sharing mechanism was not sensitive to time-of-day issues; that is, there were
not separate allocations for peak and off-peak schedulers. This shifts some of
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
225
the cost for peak usage to non-peak users, another way for schedule
coordinators with peak hour demand to limit their exposure to peak pricing.
Changes in operating practices of existing generation were also
mentioned for the effect on reliability, sufficient generation, and peak pricing:
“One of the plants . . . the new owner decided not to run it because it’s a 50
year old clunker. So we are leasing it for the summer to get another 120
megawatts in the market... even if it doesn’t make money.”
5. Do you feel both reliability and reasonable price can remain constant as
restructuring continues? I f not, what degree o f tolerance is required for
variability in price or reliability?
Respondents noted the timing differences in trying to match an hourly
or daily market to the PUC’s practice and process of periodic rate cases. This
disconnect between market pricing and allowable rates is problematic; the role
of hedging contracts is significant in reducing volatility but until the summer
of 2000 this was proscribed for the IOUs.
Tolerance for variability in price was tested in San Diego during the
summer of 2000, respondents noted. Frustrated ratepayers led to political
action to eliminate variability at the retail level while wholesale transactions
occurred at variable, increasing costs. In regard to the rising price of
generation one respondent noted, “The ISO has statutory requirement to keep
the lights on, there's no statutory requirement for a $250 cap or any other.”
This point has been identified by legislators as well. The 2001 legislative
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 2 6
extraordinary session on electric restructuring has opened with a bill altering
the statutory priorities to stable price foremost, then reliability, and lastly low
price.
Respondents had differing opinions on variability in reliability. One
stated, “Having maximum reliability is going to cost. The question is, how
much margin to you want to carry? We have a degree of reliability we’re
used to.” The regulatory restructuring and introduction of a particular market
design for generation has reduced available margin, as the ISO emergencies
make clear. “Given a reasonable time period, you will get a higher reliability
at lower cost... Short term, we are going to be struggling to find the principles,
the rules.”
Respondents also noted that reliability could be adversely impacted by
any disincentives to add new generation, and with the preponderance of new
generation fueled by natural gas rather than a mix of source fuels. One stated
the question relates directly to the use of price caps, which discourage
generators from building in California. Respondents referred back to natural
gas shortages and price controls in the 1970s and 1980s as examples of how
decisions on fuel preferences and price caps are likely to impact reliability.
6. The proposed price cap resolution discussed at the July 6th, 2000 special
meeting o f the ISO and passed at the August 1st meeting called for
"immediate implementation" o f a set o f reforms. How rapidly do you feel
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
227
legislative and regulatory bodies can create changes in hedging, load
management, and metering?
Respondents addressed each reform separately, as each has its own
implementation profile. The general role of these reforms: “You need
demand response. You have to have elasticity.”
Each reform has its own timeframe. Most stated there was no
substantive impediment to utilities’ immediately implementing hedging
contracts. More load management can be established by next summer, stated
one respondent. Replacing existing meters with smart meters that display
real-time pricing will take time and money.
Respondents also noted that someone has to be looking at the meter to
notice the price at any given moment; if it is mounted outside, the user inside
or at work will not get the price signal and alter behavior, but the meters could
serve other purposes; they could manage air conditioning load shedding and
offer value-added communication services such as Internet connectivity.
7 . Do you feel that traditional methods for supplying power (e.g.,
construction o f new transmission and central generation) are able to keep
up with load growth? I f not, what alterations are needed?
Respondents varied in their opinions on the ability of traditional
methods to supply adequate generation and transmission. Most spoke of
changes already in process, such as increased proximity of plant sites to the
area requiring the load. Others spoke of the 1990s lack of investment for
either generation or transmission due to increased risk related to restructuring-
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 2 8
initiated uncertainty. Joint power authorities (JPAs) have been used to spread
risk successfully in the past decade and can help in the coming years as well.
Several noted the role and potential value of alternative methods such
as distributed generation. “I think that the central station plant with
transmission will become less and less important in the future, especially in
California.” Also important will be increased demand-side management
(DSM) and streamlined siting procedures based on smaller plant sizes and
increased numbers of new plants.
California is a net importer of electricity. Out-of-state generation and
interstate transmission means that FERC can be of major assistance to
California in establishing a successful, economically efficient, deregulated
electric system. Several respondents, ISO governors, noted the value to
California of getting a more regional RTO established. The disconnect
between deregulated generation and monopoly transmission was also noted,
especially in the general perception that the problem is generation: “They’ve
missed the transmission part.”
8. Do you feel incentives for reactive power production are adequate? I f
not, what types o f incentives are needed?
Survey respondents see several different types of electricity involved
in the electric marketplace important for reliably delivering electricity.
Respondents consider the role of reactive power interesting but overshadowed
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 2 9
by the day ahead market shortfall, the unseasonably high price of generation,
the large amount bought on the hour ahead market, evidence of market power,
and how policymakers, regulators and legislators will act to mitigate these
problems. Only when these more urgent issues are under control will
attention be turned to reactive power.
Reactive power is an area where separate incentives could be
established. But there is concern at the ISO and with the generators on the
different types of power and pricing in use now, such as out of market (OOM)
purchases and reliability must run (RMR) generation. This would add yet
another type of electric energy pricing, none directly related to time-of-use
and all for the same raw electron commodity. How to categorize reactive
power is also a question, one that is not likely to be decided for a couple of
years.
Changed decision-making over rules on generation was a persistent
thread through the interviews. For this question, one respondent stated: “If we
paid for VAR, you could run SONGS 1 [SCE’s San Onofre Nuclear
Generating Station Unit 1; operating discontinued November 1992 as part of a
settlement agreement between SCE and the CPUC's ORA] as a big flywheel,
but nobody has an incentive to do it right now.” Similar comments about
different generating facilities were made during the course of the interviews.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 3 0
Most involved generation that has been taken offline since AB 1890 was
implemented, unlike SONGS 1 which has been offline for a number of years.
9. Do you feel mechanisms for sharing information on maintenance best
practices and equipment performance among distribution utilities are
adequate? I f not, what types o f mechanisms would improve information
sharing?
Most respondents felt they did not have the expertise on distribution
systems to answer the question, though they pointed out the ISO has active
technical groups on the maintenance and management of transmission
systems. Those who answered felt the mechanisms were adequate, but not
ideal.
10. In your opinion, have utilities experienced lengthy delays in replacing
failed equipment? I f so, why?
Respondents had not seen this. There was a question regarding the
difficulty of obtaining certain equipment in a timely manner, and also a sense
that proper investment for the future is difficult to determine in the
restructured environment.
11. To your knowledge, have planned distribution system upgrades been
implemented on schedule? I f not, why not?
Respondents have not noticed changes in distribution system upgrades
but discussed transmission system issues. Several noted the extreme difficulty
of siting new transmission. There has been a local aversion to new high-
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
231
voltage lines. Some upgrades, such as to Path 15 responsible for north-south
transmission, can require transmission upgrades on another feeder line into the
point of congestion. In the case of Path 15, an upgrade would create a
transmission weakness in Idaho; no regional version of the ISO exists that
could determine who pays what portion of this type of upgrade.
Adverse comments on system upgrades were related to frequent public
protest of siting or funding infrastructure enhancements, particularly the
addition of new poles, wires, and affiliated equipment in their own
neighborhoods; the NIMBY (“not in my back yard”) factor.
12. How do substation maintenance programs anticipate component
weaknesses? What changes in these programs are suggested by load
changes associated with restructuring?
Apart from restructuring, there is a lot of new information on
manufacturing and maintenance that is relevant for utilities. “I am not
convinced that the 1950s style, old-style utility maintenance is the right
approach.” Utilities perform preemptive maintenance and newer computer
software programs assist them in this task. There has been a shift from time-
based to usage-based equipment maintenance; e.g., a piece that might once
have been on an annual maintenance cycle may now be inspected after a
certain number of hours running.
Restructuring has created a need for inter-organizational coordination.
The December 1998 outage in San Francisco was made worse by incomplete
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 3 2
coordination and communication between PG&E and the ISO. PG&E
eventually paid the ISO a substantial fine for this outage and these failures.
13. Have transmission and distribution maintenance expenditures declined
over time? Are these expenditures adequate?
Those respondents oriented toward IOU operations noted that the
ability to fund maintenance is tied to the amounts authorized by the PUC in
the periodic rate cases. “The difficulty in ratemaking is we project out three
years, and the growth isn’t what you project.” The power flows have been
affected by restructuring and growth, as loads have increased and less
hydroelectric power is available from the Northwest.
This reduction of imports increases in-state movement of power
stressing existing lines. There may be inadequate planned funding for these
lines. A new process to set the wheeling charges for transmission is underway
at FERC. Transmission charges initially in place have made it unattractive for
municipal utilities to join the ISO.
One respondent noted that “the risk of declining maintenance is real,
when people get tight. A company that operates that way will be shaken out,
they can't compete, they aren’t running a viable business.” Other industries
that have undergone restructuring have seen numerous bankruptcies and
mergers. There is no reason to believe California utilities will be exempt from
these competitive pressures and forces. Despite this, several stated that the
answer to these concerns is not re-regulation, but rather structural design to
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
233
help California work through the transition and get to the point where the
benefits of competition are delivered.
Respondents representing distribution companies noted their budgetary
commitments to distribution system operations and maintenance. One
operations manager pointed out that adequate maintenance need not be
directly correlated to spending; their agency had been keeping budget flat
while improving their maintenance practices through reengineering and
reducing “red tape”.
14. What business factors, if any, are most significant or may compromise
reliability performance (e.g., out o f state or foreign ownership, more
power contracted rather than owned generation)?
Respondents did not feel that in-state or out-of-state ownership would
be an issue, nor would owned vs. contracted generation, if the market is set up
correctly: “I would like to believe the right economic incentives will make
everything work right.”
It was clear that the price variability in the shift from cost-based to
market pricing was a strong force, though these respondents didn’t feel this
would necessarily compromise reliability: “These plants are there to make
money and they will operate as best they can to make money.” “I don't know
where we have had a reliability problem because of the changes in ownership,
but obviously the price you pay is different.” These comments came while the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 3 4
difference between wholesale price and rate was gradually building multiple
billion dollar deficits for the IOUs.
The respondents shared a sense that reliability problems would be
man-made and therefore avoidable: “The biggest detriment [to reliability] is
the variety of approaches and instability in regulatory entities. If people see
stability, they will build plants and enter into agreements.” It can be argued
that the rolling blackouts of early January 2001 were related more to fiscal
uncertainty and the lack of regulatory or legislative revision than to limited
capacity in the WSCC.
Operational changes were not anticipated to impact reliability. No one
felt these changes carried a direct and negative impact, despite disaggregation
of functions: “You can't just go across the room and flick the switch any
more.” “[California] is not a third world country. We have a 50 year
reputation for reliability and ubiquity of electric power at reasonable cost.”
But instability could result in uncontrolled factors tipping the balance in the
summer of 2001: “Imagine a drought, plus the heat, imagine no water and no
generation.”
Most respondents continue to voice confidence that a well-functioning
market would in time provide the benefits of restructuring. The role of
regulation should be limited to management of the non-market portions of the
electric system.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 3 5
Chapter 6
Conclusions and Further Developments
Introduction
In 1996 the California legislature passed AB 1890. This legislation
defined economic benefits while creating rules and institutions to restructure the
state’s electric system. The legislature was responding to the CPUC Blue Book of
1995 which in turn was a response to FERC Orders 888 and 889 and the 1992
Energy Policy Act. This study examined the California model of limited electric
system restructuring using utility and regulatory agency records from the 1990s to
investigate system reliability. The research question focused on five changed
practices leading to hypotheses comparing the state’s five largest public and
private utilities. Logit and negative binomial regression analyzed the data. The
quantitative analysis was supplemented by a survey of electric system
restructuring participants.
This chapter summarizes conclusions drawn from the findings and
presents limitations of the analysis. Within the context of the research limits, it
considers application of the findings and conclusions. The chapter then discusses
assumptions and implications from California’s restructuring that suggest an
electric system restructuring framework. It reviews the California situation for its
fit to market models, the early 2001 California energy crisis, and ideas for further
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 3 6
study. The chapter concludes with three unresolved issues that must be addressed
in any future form of restructured electric service in California.
Conclusions
The research results are significant. The results demonstrate a slow but
steady decline of regional generation margin. The divestiture of generation did
not cause an increase in reliability disturbances. The IOUs and municipal utilities
shared some problems, despite the municipal utilities’ exemption from AB1890.
Policy makers show a willingness to modify the design of restructuring and
acknowledge the necessity of this action to re-balance supply and demand. While
admitting problems in the California design, few electric industry representatives
have become disenchanted with reduced regulatory controls and efforts to
establish a more competitive model. These representatives share a conviction that
the vertically integrated regulated system cannot be re-established.
Table 23 summarizes results from the logit regressions which investigated
the effects of California’s partial restructuring on a constructed measure that
included the altered electric system components and reliability disturbances. The
logit analysis created a broader disruption category which included load shedding
and energy emergencies; effects of the generation divestitures and the
introduction of new institutions were among the independent variables.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 3 7
independent
variables significant level direction data
Logit
dv=d_dsturb
djnuni yes 0.005
1
cross-utility
wjain yes 0.10
1
L A D W P
w _cddavg
yes 0.005
t
S D G & E
w _cddavg
yes 0.005
t
S M U D
w _cddavg yes 0.005
t
cross-utility
w _hddavg
yes 0.005
t
S D G & E
w _hddavg yes 0.005
t
cross-utility
djsoup yes 0.05
I
L A D W P
djsoup yes 0.05
T
cross-utility
purchpw r
yes 0.05 L A D W P
dsm_ee no
da_delta no
d_endctc no
Table 23: Significant Independent Variables, Logit Regression
The logit results found municipal utility status, weather variables, the post-
ISO period, and purchased power significant for at least one utility or for the
combined utility data. The logit regressions found no significant effects from
direct access or energy efficiency for the period examined.
The weather-related heating and cooling were the most consistent factors
in increasing reliability disturbances, at the .005 level. The less-compelling .10
level LADWP effect of total monthly rain is not noted in any other utility or in the
cross-utility analysis.
d ISOup was significant at .05 level, associated with increased
disturbances in the combined utility analysis. This was not evident in negative
binomial analysis using reported outage data (system average interruption
duration index, or SAIDI) instead of more general disturbance information. The
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 3 8
significance does not indicate that the California Independent System Operator
(ISO) caused these outages; rather, the hypothesis that the ISO had a negative
effect on system reliability cannot be rejected. This dummy variable separates
reliability disturbances before and after the ISO began operations as grid
controller. The effect is consistent with other changes not explicitly included in
the analysis, such as strong economic growth in California or the reduction of
generation margin for the WSCC area of the Western Interconnection power grid.
The negative correlation between reliability disturbances and LADWP similarly
can suggest either indirect effects of the ISO on the utility, or other activity at
LADWP at the same time, such as the response of general manager S. David
Freeman to the challenges of electric system restructuring.
For the IOUs and the cross-utility analysis, the increased dependence on
purchased power produced no significant results. Purchased power results for
LADWP suggest purchased power practices were associated with increased
system disturbances. The overall purchased power results for LADWP were
mixed, as will be discussed below following the presentation of the negative
binomial results.
Municipal utility status is significant at the .005 level, and negatively
associated with system reliability disturbances, that is, favorably associated with
system reliability. As IOUs first requested large rate increases and then sought
protection from creditors, the general public and policy makers have noticed this
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 3 9
advantage as well. Yet, little specific evidence exists that wholesale government
participation is preferable to improved market design and rules.
Table 24 summarizes results from the negative binomial regressions which
investigated the effects of California’s partial restructuring of the electric system
on utility distribution company (UDC) power outages. Using the distribution
company specific outage data as dependent variables allowed inclusion of
independent variables for ISO emergencies and load shedding. The non-identical
SAIDI outage data did not allow for cross-utility analysis. Directly or indirectly,
all UDCs were subject to fiscal pressure due to frozen rates at the IOUs, which
might have reduced reliability-related operations and resulted in increased outages
following the passage of AB1890.
independent
variables significant lev el direction data
Negative Binom ial
dv= S A ID Iann
w _rain
yes 0.05
t
SC E
or S A ID Im o n
w _rain
yes 0.005
T
S D G & E
w _rain yes 0.10
t
L A D W P
w _cddavg
yes 0.005
t
S D G & E
w _hddavg
yes 0.05 SC E
w _hddavg yes 0.05
t
S D G & E
purchpw r
yes 0.005
1
L A D W P
purchpw r
yes 0.005
1
P G & E
isoem erg yes 0.05
t
L A D W P
loadshed yes 0.10
T
SC E
djsoup no
dsm_ee no
da delta no
d_endctc no
Table 24: Significant Independent Variables, Negative Binomial Regression
R eproduced with perm ission o f the copyright owner. Further reproduction prohibited without perm ission.
2 4 0
The negative binomial results found weather variables, purchased power,
the use of load shedding and ISO emergencies significant for at least one utility.
The negative binomial regressions found no significant effects from direct access,
energy efficiency, the end of the CTC in San Diego, and the introduction of the
ISO for the period examined.
Weather-related variables were again the most consistent factors in
outages. Rainfall was significant for three utilities, heating degree days for two
utilities, and cooling degree days for one utility. Both regressions found weather-
related heating and cooling factors significant. There is a level of uncertainty in
the weather that is irreducible. Adjusting public policy for weather in California
can be a political risk, as weather cycles do not match political seasons.
The purchased power results for LADWP and PG&E show a negative
association between increased purchases for power and actual outages for the
period studied, significant at the .005 level. These results do not confirm the
conventional wisdom that divestiture of generation and increased purchased
power increases risk of system outages.
Other significant negative binomial findings affected only a single utility
each. ISO emergencies were significant at the .05 level for LADWP, and load
shedding was significant for SCE at the less-compelling .10 level.
In addition to regression analysis, time series analysis was performed with
the following results. Time series analysis of generation demand showed peak
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
241
demand in 2000 slightly lower than in 1999. Supply shortfalls at the end of 2000
were not simply the result of inadequate build-up of generation capacity in the
1990s. Generation margins reported by utilities had been high for some time
including throughout the period preceding and immediately following AB 1890’s
passage. The proportion of generation owned by UDCs fell steeply between 1997
and 1998, less steeply between 1998 and 1999. The ISO declared few
emergencies until the summer of 2000; ISO emergencies in the second half of
2000 offered an early warning of a destabilized restructured electric system.
Limitations of the Analysis
Conclusions based on this research are limited by uncertainty, limits to
counterfactuals, and ongoing alterations to California’s electric restructuring due
to political and financial effects in 2000. Each is discussed below. There are also
limitations inherent where irreducible heteroscedasticity in the regression analysis
restricts its predictive value. To mitigate these limitations, data for the entire
decade have been reviewed, covering several years prior to restructuring and
several following the passage of AB 1890.
Uncertainty
Friedman (1984) states that uncertainty results from our lack of
information or the lack of ability to process that information, i.e., our bounded
rationality. Uncertainty in predicting electric system reliability under
restructuring is enhanced by persistent information asymmetry between
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 4 2
regulators, utilities, generators and customers. The public and even regulators are
not fully informed regarding either regulated or unregulated segments. Operating
information has not been made available for components of the formerly
integrated operations; this information may even no longer be collected for
operations now outside the regulatory arena. This uncertainty is exhibited in the
research and findings.
While the quantitative analysis is statistically accurate and descriptive of
the situation for the past decade, incomplete information and the complex set of
reliability disturbance triggers suggest substantial explanatory data are outside the
material covered. Uncertainty can be reduced by data pooling; this research
pooled data from many sources to reduce missing elements. It also combined
qualitative assessments with the quantitative analysis.
One impetus for restructuring is the very unpredictability and uncertainty
of the future. In the absence of market power, market mechanisms are
fundamentally more responsive and impartial than regulatory processes.
Recognizing this regulatory weakness can result in legislation with triggers
precluding anticipated undesired outcomes; no triggers were included in
California’s restructuring design. Establishing triggers can have negative effects,
since they reduce public policy discussion.
Counterfactuals
This research is limited in providing counterfactuals such as the degree to
which economic effects of the early 1990s recession skewed consumption and
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
243
suggested adequate capacity and availability. The 2000 generation supply
problem has been attributed repeatedly and in large part to the limited in-state
generating capacity added in the 1990s. Yet a major recession in California
followed the end of the Cold War, with subsequent dislocations of jobs, revenues,
people and their associated electric load. This was not a period requiring large
localized generation increases, and the hundred-year-old regulated electric system
monopolies had experienced recession, depression, boom, bust cycles many times
before without any disruptions similar to those in the last half of 2000.
Nevertheless, no comprehensive answer has been provided in this research to the
question of any supply/demand imbalance which could have occurred even if
nothing had changed.
Alterations to the Restructuring Model
A related limitation is the difficulty in separating short term transition
effects from longer term problems inherent in the restructured industry design.
The transition from any regulated industry to an open market exceeds the few
years officially legislated in AB 1890. This transition began before passage as
utilities argued for stranded cost recovery and tried to forecast how competition
would alter their operations. These transition effects are complicated and
extended by alternations to the model adopted by California for partial electric
system restructuring. They will persist another decade or two, even longer if the
state becomes a major owner of generation or transmission facilities in response
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 4 4
to the crisis or issues billions in bonds to pay for power purchases and long term
non-bypassable surcharges are enacted to reimburse the state.
Extended transitions create intertemporal imbalances. The AB 1890
structured rate reduction bonds require IOU ratepayers to pay an 8% surcharge for
ten years in exchange for a three year 10% discount. A threatened voter ballot
initiative that may resemble the failed Proposition 9 could add to the extended
transition effects. Resurgence in interest in energy efficiency retrofits also creates
a long term impact.
The California market problems during the second half of 2000 resulted in
billions of additional power costs. Reducing these costs became the paramount
issue for electric restructuring. Allocating the billions in charges, and ending
behaviors and market rules contributing to the high prices resulted in suspending
the AB 1890 formula for electric system restructuring. No replacement formula
has been adopted. This could limit the value of this paper’s research findings,
particularly if the research solely addresses a defunct system. However, the value
of the data compilation and comparison across utilities persists. The assessment of
mismatch between established models for economic regulatory restructuring and
California’s implementation of partial electric system restructuring remains valid
for use in other designs for altered regulatory structures.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 4 5
A U.S. Framework for Electric System Restructuring
The separation of powers among state, federal and local entities is a
defining characteristic of the American form of government and is a source of
stability. However, this separation among agencies, regulatory bodies, legislators,
courts and public processes complicates application of economic theory for the
interlocking systems. The western electric system has been stressed, and the
public has not benefited from the restructuring of California’s electric system.
Even worse are the 20%-50% rate increases imposed in Washington and other
western states which bid into the shared market for power. Citizens of these
states have no voice whatsoever in the politics and legislative processes that have
governed California’s electricity policy. A last reminder of the inter-connected
energy fate of the West is increased air pollution in Colorado for electricity
generated and sold to California. The relationships run parallel if not always
bilaterally at several units of analysis;
CPUC FERC
CA governor western governors
CA grid western interconnect
CA ISO other power pools (potential western RTO)
CA demand western emissions (common)
CA emissions western demand (rare)
CA purchases western prices
Price elasticity appears essential to successful restructuring and an
efficient energy market. Price elasticity is increased through the availability of
substitutes. Direct access with ease of entry and exit creates opportunity for
competition in the price of generation. Substitutes for existing California
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 4 6
generation include distributed generation, creation of negawatts through energy
efficiency, incentives to reduce load through interruptible rates or dispatchable
load reduction, and out-of-state generation with transmission.
Increased localized generation and demand side management can respond
to constrained transmission. Enhancing the transmission system is a more
controlled response, as the transmission system remains a non-competitive
monopoly under regulatory oversight. Adding major generation beyond the
state’s border or far removed from high demand urban corridors resolves the
NIMBY disputes if the additional transmission load can be managed without new
high voltage lines. Significant increase in transmission volume without
increasing infrastructure is unlikely, but mitigations can reduce community
resistance.3 3
Through its discussion of baseline consumption, the California legislature
is acknowledging a universal electric service level. The current baseline is rate
class specific. As long as rates were stable, little public discussion was carried
out about its appropriateness. With unstable prices, the question is again debated
how to establish baseline consumption across all classes of service.
3 2 A term that conceptualizes the value of conserved energy, often associated with Amory Lovins
and his work on energy efficiency. Hirsh, Power Loss, 194.
3 3 Planning new transmission corridors, particularly in the Bay Area of Northern California,
highlights parallel infrastructure problems in the same geographic region such as inadequate
housing and transportation capacity. Competing land use issues were factors in the San Jose City
Council’s rejection of an otherwise-approved generation project in early 2001.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 4 7
The fiscal crisis has forced a reassessment and acknowledgement that
there is a price beyond which reliability is no longer the most important factor.
Governor Davis’s sealed-bid generation contract process may result in
establishing an acceptable price for a portion of the load. Pricing functionally
distinct portions of the load was one reason for separate PX and ISO electric
markets, but the permeability of the boundary shifted selling to the most lucrative
point: peak power and reliability purchases are bought at a premium.
Regulators, new entrants, utilities and policy makers continue to refme
separable components of the electric market. The commodity electron is used in
several different power management applications. Functional differences exist in
grid operations between reliability must run (RMR), standby, and reactive power.
Peaking units, premium power service, and time of use meters are all products
that contribute to a more flexible restructuring model and could be separately
managed, priced, and sold.
Unexpected outcomes and potential for new products suggest a cyclical
framework that can reassess past decisions in light of new developments while
protecting public interests. This framework is illustrated in Figure 9, exemplified
by incidents in the implementation of AB1890.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
248
Step 1: Commitment by public ..
leaders to market mechanisms
f
(Agreement to create and implement \
new legislation, regulation) \
Step 4: Evaluation of institutional
success; investigation of new
technologies and opportunities;
assessment of social equity
Step 2: Clear demarcation of
regulated and market arenas
(Assurance of direct access)
(Dialog on acceptable price and
reliability; on augmentation or reduction
of market) 1
Step 3: Development of /
mechanisms to create or correct /
institutions as needed
-------
(Determine role of ISO/RTO,
functional success defined)
Figure 9: A Cyclical Framework to Fit Deregulatory Reform Intentions
Reviewing the California Situation
This section review the situation in California. It considers the fit of the
partial restructuring to standard market models, the 2001 crisis, and ideas for
further study.
Fit of California’s Restructuring to M arket Models
Economic benefits codified in AB 1890 suggest actual or political naivete
about markets, where prices move in both directions due to supply, demand,
elasticity, and the availability of substitutes. It also indicates a disconnect
between the world of theoretical models and the world of political and social
realities.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 4 9
California’s electric restructuring results initially matched predicted
economic market models, given its constraints due to the limited nature of the
restructuring. Already moderate generation costs dropped below two cents per
kilowatt hour. Stranded costs were largely recovered, nearly thirty percent of the
large commercial load moved to direct access providers, and system reliability
showed no systemic problems.
But the limited market for power destabilized. Several other states with
restructuring activities and regional RTOs have not experienced the systemic,
comprehensive negative effects that destabilized the California restructured
environment the last part of 2000. The California model as initially designed
failed to closely map standard models for market formation. It overly constrained
participants and attempted to control behavior within and without the regulated
portions of the electric system. The California-specific ISO prevented this
transmission organization from developing strategies that transcended state
politics.
Two years after the bulk of the non-nuclear generation assets were
divested by the IOUs, California’s energy market began experiencing unexpected
persistent pricing problems under the new market rules. These pricing problems
were not accompanied by system outages until the utilities ran out of money. In
addition to high wholesale prices at peak demand seen in other regional markets,
the California generation market exhibited two peculiar behaviors. First, ever
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 5 0
larger blocks of power were sold on the ISO spot market rather than the Power
Exchange (PX) day-ahead market. Second, high prices persisted for more hours
and days, while prices did not fall back to typical rates for off hours or even the
off season. The California-specific nature of these two problems suggests the
responsibility rests in the state’s restructuring design, rather than any fundamental
flaw in the shift to a more market-based electric system.
The public and politicians are generally unsupportive of economic market
pricing models based upon better retail price signals to reduce consumption,
because over the near term these tend to increase corporate profits while reducing
public convenience. When retail bills tripled during the summer of 2000 in San
Diego, the mayor was quoted as worried that the next phone call would advise her
of an elderly pensioner dead from heatstroke in an under-cooled apartment. In the
winter of 2000, fear of hypothermia resulted in newspapers and public service
announcements reminding the public not to turn off their furnaces despite high
natural gas prices. These fears have not been grounded in specific occurrences of
tragic outcomes.
California’s public distress over the rising cost of electricity has resulted
in calls for re-regulation, with increase conviction among some public officials
that electricity is not a marketable commodity. Electric system restructuring in
other states has slowed. Proposals to emulate the relative success of municipals
have included formation of new municipal utility districts, ownership of
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
251
generation, ownership of transmission, and a state version of the PX where money
to buy power comes from state-issued bonds instead of the utilities. These
activities have been investigated or undertaken to reduce rolling blackouts and
consumer anger over proposed rate increases and IOU behavior. The strong
executive and legislative reaction in early 2001 toward state ownership in
response to market problems contradicts the stated direction of FERC to reduce
regulatory controls on energy markets, as well as a worldwide trend shifting
formerly regulated industries to competitive models.
Market design and effects in California impact other states in the western
grid. The power California purchases is also marketed to neighboring states.
These interstate effects of in-state operations underscore federal level issues
beyond traditional wholesale wheeling and interstate transmission. The state
PUCs and PSCs will continue to bear responsibility for protecting the public
interest in rate setting for the monopoly UDCs. The CPUC’s continued interest in
cost of service prices and fixed rates contradicts FERC’s general direction to open
market pricing.
The 2001 California Energy Crisis
In responding to federal electric industry restructuring goals, California
leadership expressed optimism and enthusiasm for the benefits promised in
electric system restructuring: lower-priced electricity and new electric services.
But AB 1890 created a partially restructured system with legislation detailing
specific results that were not achieved. As noted by survey participants, the state
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 5 2
had no intrinsic method to adjust the new institutions when unintended
consequences occurred. Existing rules, institutions, and legislation could not
address unsatisfactory market behavior in the second half of 2000. The existence
of market power was investigated, but the question was not settled. Mechanisms
to better match supply and demand were discussed, but not conclusively decided
nor implemented. More than one survey participant noted that the QFs and
nuclear facilities built since the 1978 passage of PURPA ended the last supply
shortfall; this abundance continued through the restructuring design period.
Unfavorable market conditions reduced QF contributions, worsening the supply
and demand imbalance.
With initial restructuring success and limited indication of market
problems before mid-2000, the ISO had little need for inter-organizational
discourse and coordinated action with the PUC or other institutional actors. The
ISO declared Stage 1 and Stage 2 emergencies on seven days in 1998, and only
four days in 1999. By August 2000 the ISO had already declared 13 emergencies
and the rates in San Diego were causing a huge public outcry. The perception
grew that the PUC and ISO were working at cross purposes.
Each institution had its own response, and each called upon others to take
action to resolve the dilemma of rising wholesale prices. The legislature rolled
back residential rates in San Diego, the ISO and PX boards established price caps
under intense political pressure, the CPUC chose a passive response to utilities’
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
25 3
requests for clarification on “reasonableness” reviews for newly-authorized
bilateral contracts. Historic patterns of demand, supply and pricing in fall and
winter suggested fundamental change did not need to be taken immediately. The
assumption that fall and winter 2000 would resemble past years proved to be
incorrect.
These actions did not provide substantive solutions to the lack of
interaction between institutions or the shrinking margin of available generation;
nor did these actions promote end user demand reductions.3 4 Instead, they
resulted in unprecedented reductions in available generation, two IOUs defaulting
on payments with creditors threatening bankruptcy, emergency allocation of state
money for electricity purchases, and Federal intervention in the form of mandated
sales into California.
Much of the capacity shortfall late in 2000 resulted from unplanned
generation maintenance; the causes for this maintenance are disputed.
Approximately 30% of expected load overall state capacity was simultaneously
unavailable in early January 2001. Offline generation represented a major
capacity shortfall. Under vertical operations, there would at minimum have been
significant planning opportunity to coordinate maintenance, avoiding shortages
3 4 Some policy responses worsened the problem. Utilities and the CPUC supported the idea of
“balanced” billing, a method by which consumer distress over high bills is reduced by equalizing
payments, paying for less than total usage in high-use months, more in low-use months.
Bhattacharjee et al. (1993) describe how this approach masks actual costs and decreases
conservation. Immediate, out-of-pocket expenses are more salient losses and have greater impact
on behavior.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 5 4
and the resultant rise in market price. Supply shifts that were well-planned and
coordinated under vertical utilities did not resemble operations in late 2000. A
shortfall due to planned maintenance was not unforeseen; summer 2000 ISO
Board of Governor meetings discussed the value of implementing a unified
maintenance schedule under restructured operations.
January 31, 2001 marked the fifteenth straight day the California ISO
declared a Stage 3 emergency. Rolling blackouts occurred two of the first three
days; the system then settled into a continual low-margin operation with state
purchase and no blackouts for the next month. Governor Davis requested and
received emergency legislation authorizing the Department of Water Resources to
become a spot market purchaser of power for resale by the nearly bankrupt PG&E
and SCE. In less than a week over $100 million of $400 million allocated was
spent at ten times the usual winter peak price. Spending by the end of January
amounted to a half-billion dollars; by mid-February, this amount exceeded two
billion dollars. This exercise in mass spot market purchasing occurred at a low
winter peak of 30,000 megawatts of demand statewide. Peak usage in the winter
of 1999 exceeded 34,000 megawatts without a single ISO emergency stage
invoked. This is an approximate 20% reduction in available supply in only twelve
months, 12% representing the difference between the 34 and 30 megawatts, and
6% representing the minimum difference between normal operations and a Stage
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 5 5
3 emergency. Had this extra 20% in existing generation been available, there
would have been no emergency purchases and no curtailment by the ISO.
The relation between generation margin and reliability is complicated by
the current high cost of generation. The quantitative results suggest there is little
interest in the increase of spending that would be required to reduce outages to
zero for all users. SAIDI outage targets and discussions with survey participants
confirm that utilities and ratepayers have long had a tacit assumption regarding
good-enough reliability at a reasonable cost. Premium power programs focus on
providing higher reliability at an additional cost on a limited basis only for those
customers with stricter operating tolerances than the general public. The cost of
providing generation differs according to the margin desired, so it is important for
policymakers to discuss how much margin is appropriate, and how to manage the
cost for the margin provided.
Ideas for Further Study
This research presented an expansive but not exhaustive analysis of
California electric system issues related to electric system reliability. Further
research could examine variability in system reliability, factors of disruption,
implications of operations policy in determining an acceptable level of system
reliability disruptions, and distinctions between IOUs and munis not captured in
this study.
The relationship between price and reliability under various regulatory
scenarios could be modeled. The fiscal impact of state-supported electricity
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 5 6
purchases questions whether absolute reliability at any cost is acceptable or
reasonable. Beyond actual reliability disturbances, fear of outages could be
considered. The insolvency of two IOUs and the state’s reaction could be
explored for their effects in reducing available generation and the increases in
local shortages and transmission congestion.
California’s historic proclivity for floods, fires, and other emergencies
suggests general benefit from similar analyses based on disturbances and outages.
California temblors could be correlated to outages to determine the relevance of
quake size and proximity. Increased research on best practices in distribution and
transmission maintenance are an important part of electric system reliability, and
the UDCs and transmission grid will remain franchise monopolies with their costs
captured within rates.
The correlation between better tree trimming and reduced SAIDI for
SMUD suggested that historic categorization of outages as weather events may
represent a more subtle factor. The existence of outage metrics suggests an
implicit agreement between utilities, regulators, and customers on the acceptable
level of reliable service. State executives and lawmakers may be unaware of such
assumptions of reasonableness in balancing maintenance, reliability, and electric
cost.
Operating and other differences between municipal utilities and the IOUs
are worth examining if suitable data measurements can be collected. For
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 5 7
example, the utilities might have different required responses to ISO emergencies.
Purchased power can also be further broken down and investigated; this could
include comparing long and short term contracting as well as the role of
qualifying facilities (QFs) contracts and the impact of state-mandated rate
surcharges. Because they remain vertically integrated and are exempt from many
AB 1890 rules, municipal utilities may have more ability to balance supply and
demand within their own regions.
Energy efficiency studies have experimented with levels of lighting and
cooling that reduce consumption without observable discomfort. A similar study
could be done to establish tolerance for reliability disruptions such as load
shedding or scheduled outages. Evidence of tolerance is limited and anecdotal.
Additional research could inform efforts to improve elasticity for demand in a
restructured supply market.
Evolving and new organizational structures complicate comparisons
across the pre- and post-restructuring timeline. For greatest relevance and value
to public policy, electric restructuring research could concentrate on
governmental institutions or utility organizations beginning with their creation.
This approach recommends itself since predictive models are of limited applied
value if the system studied has ceased operations.
The study of electric system restructuring crosses discipline boundaries.
Integrated knowledge from several academic disciplines is needed due to the
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 5 8
public implications of restructured industry designs. Each field researches and
reports using precise language that may not be correctly interpreted without a full
understanding of the terminology. Much of this research therefore may not be
directly accessible to policy leaders, who must consolidate information from
advisors and apply the results in the real world. The largely unforeseen electric
market problem and the immense cost of not seeking an early remedy can serve as
an impetus for developing an improved discourse. This research on the effect of
California’s partial restructuring makes a modest but real contribution to a shared
understanding regarding the delivery of theoretical economic benefits from
restructuring the electric system.
Unresolved Issues as Closing Observations
California’s electricity system restructuring involves tens of billions in
annual and special expenditures intended to resolve the market problem. Several
issues will shape the cost to Californians now and for decades to come Three are
mentioned here in closing as reminders of the ongoing consequences of
government action and intervention, the search for balance between the public and
private sectors, and the need for a common language for discourse.
This first issue involves the monopolistic components of electric service,
as opposed to areas that can be opened to competition and market pricing. This is
a continuation of the natural monopoly discourse. Defining monopolies and the
means to manage them is subject to technological improvements as well as
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 5 9
political and social climate. The current components and rate distinctions are in
ways polite fictions to the consumer. “Customers don’t pay rates; they pay bills,”
S. David Freeman has said. Political rhetoric aside, the customer checks the
number at the bottom of the bill. If the total becomes unacceptable, basic
economic theory predicts the customer will respond, often in innovative and
interesting ways. In this balance between supply and demand, rights and
opportunities, the electric bill will be squeezed. Portions will get more attention
from customers, taxpayers, utility officials, legislators and regulators. In 2000
and early 2001, the steep rise in the price of generation focused attention first on
generation, then on transmission bottlenecks, then on fuel source. In 1996, the
issues were the delay of savings for residential ratepayers and the size of the CTC.
In 1998, the addition of the non-bypassable rate-reduction bond surcharge and
Proposition 9 captured public attention.
The bounds of the billing categories are fungible; over the next twenty years
existing categories such as distribution costs will be further split. Each sub-part
will in turn be debated and cataloged as regulated or competitive in the continuing
fragmentation of the vertical monopoly. This will be accompanied by
consolidation across geo-political boundaries within categories, if other industries
in any way exemplify the pressures of market forces on the traditional model of
electric service provision.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 6 0
Retail wheeling or direct access is the second disputed issue. Legislative
analysis stated competition would reduce the cost of generation, which was
calculated at more than half the total cost of electricity early in the restructuring
discourse. AB 1890 promised retail customers choice of generation suppliers.
After its enactment, dozens if not hundreds of potential suppliers and power
marketers came to California to provide choice. Instead of customers evaluating
these choices and switching over time, a moderate percentage of commercial
customers and tiny percentage of residential customers immediately found a
suitable product; the rest found nothing of interest offered, and the choices
diminished. This is in contrast to the penetration and persistence of choice in
some eastern states. The 2001 extraordinary legislative session passed legislation
authorizing a $10 billion bond issuance for state purchased power; ABlx 1
preempts direct access purchases. Even if later legislation grandfathers in existing
direct access customers, the state’s long term power contract strategy requires a
committed load and can prevent its citizens from gaining two promised goals of
comprehensive restructuring-lower cost and greater choice. Proponents of
reduced regulatory controls, business interests with successful strategies in other
states, and potentially even consumer interest groups will vigorously challenge
the loss of direct access. Customer choice, an important element of the limited
regulatory change in California, has ideological and financial implications.
Without direct access, there does not exist a competitive retail market.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
261
“Fair and reasonable” pricing is the third issue. Californians understand
that the price of gasoline varies from week to week and year to year. Neither its
citizens nor its legislators are yet prepared to treat electricity as the same type of
commodity. Protracted discussion at the state Capitol accompanied the inclusion
of a base quantity of electricity that would come at a price subsidized by users of
larger amounts of electricity. No one has calculated what the target quantity,
130% of baseline, really means in terms of the square footage and ambient
temperature of living space, the efficiency rating of a furnace or air conditioner,
the cubic footage of a refrigerator, the size of a television, or the number of light
bulbs. Suggestions on how many California customers (voters) would avoid
major rate increases by staying under the target varied from 25%-85% in AB 1 x 1
hearings. Alternatively, on the capital investment side, the governor and
legislators seek to increase in-state generation on the assumption that it will
reduce the price of generation. Generation in 1998 and 1999 cost as little as $.02
per kWh; SDG&E’s average cost for purchased power from 1991-1999 ranged
from 2.8 to 3.8 cents per kWh. If new in-state generation costs 5 cents per kWh,
it is reasonable to suppose generators will seek a fair rate of return plus a risk
premium for building in the unstable California market. This can lock in
generation prices at least twice as high as the 1990s. Other surcharges have been
proposed for state-authorized power contracts, extensions of the CTC for the sake
of fiscally strapped IOUs, new transmission lines, specialized energy efficiency
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 6 2
incentives and clean/renewable power initiatives. These endanger the just and
reasonable price postulated under restructuring, one that results in a lower average
bill, all things being equal.
California’s partial restructuring of the electric system failed to reduce the
price of electric service and negatively affected system reliability. Through 2000,
utilities and the ISO provided protection from power outages comparable to pre
restructuring operations, but increased reliability disturbances offered an early
warning on reliability problems. These indicators were missed or misinterpreted
and did not result in appropriate market adjustments or institutional responses.
Assumptions not conclusively supported by this study’s findings have led to long
term state contracts and in-state generation goals. These research findings and
the framework for revision offer empirical information and a mechanism for
policy makers to realign the state’s partial restructuring to continue a transition to
market forces through improved pricing models, reduced market uncertainty, and
the reduction of inteij urisdictional conflict.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 6 3
Bibliography
Alderfer, R. Brent, Monika M. Eldridge and Thomas J. Starrs. Making
Connections: Case Studies o f Barriers to Interconnection o f Distributed
Power. National Renewable Energy Laboratory/SR-200-28053, May 2000.
Aldrich, John H. and Forrest D. Nelson. Linear Probability, Logit, and Probit
Models. Newbury Park: Sage Publications, 1984.
American Public Power Association. “1995 Annual Statistical Report,” Public
Power (January-February, 1995): 75-171.
. “State Legislative/Regulatory Restructuring Summary, July 31,1998,”
http://www.appanet.org/ppeui/chart.html, August 10,1998.
. “Market Power Tales Show Importance of ISOs,” Public Power Daily,
September 14,1999.
Barrington-Wellesley Group Inc. “Review of Pacific Gas and Electric Company
Financial Condition for the California Public Utilities Commission.” January
30, 2001.
Barker, James Jr., Bernard Tenenbaum and Fiona Woolf. “Regulation of Power
Pools and System Operators: An International Comparison,” Energy Law
Journal, 18, no. 2 (1997): 261-332.
Baumol, William J. “Contestable Markets,” American Economic Review 72, no. 1
(March 1982): 1-15.
and J. Gregory Sidak. “Stranded Costs,” Harvard Journal o f Law and
Public Policy 18, no.3 (Summer 1995): 835-849.
Bhattachaqee, Vinayak, Charles J. Cicchetti and William F. Rankin. “Energy
Utilities, Conservation, and Economic Efficiency,” Contemporary Policy
Issues 11, no. 1 (1993): 69-78.
Bilas, Richard A., Kenneth L. Lay, Gordon R. Smith, Michal C. Moore, and
Robert J. Michaels. “Power Markets: Restructuring Electricity,” Contemporary
Economic Policy 17, no. 1 (January 1999): 1-19.
Borenstein, Severin and James Bushnell. “An Empirical Analysis of the Potential
for Market Power in California’s Electricity Industry,” Working Paper 6463,
Cambridge: National Bureau of Economic Research, 1998.
5 and Frank Wolak. “Diagnosing Market Power in California’s
Restructured Wholesale Electricity Market,” Working Paper 7868, Cambridge:
National Bureau of Economic Research, 2000.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 6 4
Boyd, James. “The ‘Regulatory Compact’ and Implicit Contracts: Should
Stranded Costs be Recoverable?” The Energy Journal 9, no. 3 (1998): 69-83.
Brennan, Timothy J., Karen L. Palmer, Raymond J. Kopp, Alan J. Krupnick, Vito
Stagliano and Dallas Burtraw. A Shock to the System: Restructuring America’ s
Electricity Industry. Washington: Resources for the Future, 1996.
Britman, Ross C., David L. Ketterman, and John W. Mayo. “Targeted and
untargeted subsidy schemes: Evidence from postdivestiture efforts to promote
universal telephone service,” Journal o f Law and Economics 41, no. 2 (1998):
477-502.
Brown, Lester R. et al. State o f the World: A Worldwatch Institute Report on
Progress Toward a Sustainable Society. New York: W. W. Norton and
Company, 1997.
Bryson, John E. “Statement of John E. Bryson Before the Assembly
Subcommittee on Electrical Energy Oversight,” February 9, 2001.
Buchanan, James N. Property as a Guarantor o f Liberty. Brookfield, Vermont:
Edward Elgar, 1993.
California Energy Commission. “Database of California Power Plants,”
2000_PLANTS.XLS. Online at http://www.energy.ca.gov/
databases/index.html#powerplants, August 28, 2000.
— — . “Electricity Report 1996,” November 1997.
. “Price Responsive Load Management,” discussed at the CEC/CPUC
Summer Reliability and Load Management Workshop, www.energy.ca.gov,
October 28,2000.
. “SB 735: The Role of Energy Efficiency and Distributed Generation in
Grid Planning,” P300-00-003. April 2000.
. “Warren-Alquist Act: Public Resources Code Section 25000 et seq.,”
P I60-00-001. January 2000.
California ISO. "California ISO System Disturbance Report: December 8, 1998—
San Francisco Area," March 31, 1999.
. “News Release: California ISO Board of Governors Approves Settlement
Agreement with PG&E Following December 1998 Outage,” May 28, 1999.
. “Proposed Price Cap Resolution,” July 6, 2000.
. Terry M. Winter’s “Memorandum re: Release of December 8, 1999 [sfc]
San Francisco System Disturbance Report,” March 19,1999.
California Public Utility Commission. Draft Environmental Impact Report SCH#
2000042110: Pacific Gas & Electric Co. ’ s Application for Authorization to
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
265
Divest Its Hydroelectric Generating Facilities and Related Assets (Application
99-09-053). November, 2000.
--------- . Electric System Reliability Annual Reports. Online at
http://nic.cpuc.ca.gov/divisions/energy/sections/reliability_section/reliability-
reports.htm, October 13, 2000.
California State Assembly. Briefing Book, Assembly Committee on Utilities and
Commerce Informational Hearing: Electric Restructuring — Then and Now,
January 25, 1999.
— . Assembly Bill 811. Time o f Use Credit, 1999.
California State Senate. Senate Bill 1063. Disposition and Future Operation o f
Utility Owned Hydro Facilities, 1999.
Cicchetti, Charles J. with Kristina M. Sepetys. Restructuring Electricity Markets:
A World Perspective. New York: Visions Communications, 1995.
Coleman, Charles. P.G. & E. o f California: The Centennial Story o f Pacific Gas
and Electric Company, 1852-1952. McGraw-Hill, 1952. Excerpted online at
www.pge.com, July 17, 2000.
Committee on Electric Utility Regulation. “Report of the Committee on Electric
Utility Regulation,” Energy Law Journal 18, no. 19 (1997): 197-231.
Conkling, Roger L. “Marginal Cost Pricing for Utilities: A Digest of the
California Experience,” Contemporary Economic Policy 17, no. 1 (January
1999): 20-32.
Cook, Peter. “Privatization and Utility Regulation in Developing Countries: The
Lessons So Far,” Annals o f Public and Cooperative Economics 70, no. 4
(1999): 549-587.
Department of Energy. Interim Findings o f the U.S. Department o f Energy’ s
Power Outage Study Team. Online at http://tis.eh.doe.gov/post January 10,
2000.
—------ . Report o f the U.S. Department o f Energy’ s Power Outage Study Team:
Findings and Recommendations to Enhance Reliability from the Summer o f
1999. Final Report, March 2000.
, Energy Information Administration. The Changing Structure o f the
Electric Power Industry: Selected Issues, 1998. DOE/EIA-0620, July 1998.
, --------- . The Changing Structure o f the Electric Power Industry 1999:
Mergers and Other Corporate Combinations, DOE/EIA-0562(99), December
1999.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 6 6
, Office of Economic, Electricity, and Natural Gas Analysis and Office of
Policy and International Affairs. Comprehensive Electricity Competition Act:
Supporting Analysis. DOE PO-0057, July 1998.
Dillman, Don A. Mail and Internet Surveys: The Tailored Design Method.
Second Edition. New York: John Wiley & Sons, Inc., 2000.
Erikkson, Ross C., David L. Kaserman and John W. Mayo. “Targeted and
Untargeted Subsidy Schemes: Evidence from Postdivestiture Efforts to
Promote Universal Telephone Service,” Journal o f Law and Economics 41
(October 1998): 477-502.
Federal Energy Regulatory Commission. “Commission Proposes to Reshape
California’s ‘Seriously Flawed’ Electricity Markets with Sweeping Changes,”
News Release, November 1, 2000.
Flentje, H. Edward. “The Evolution of Municipal Electric Utilities in Kansas,”
Municipal Finance Journal 19, no. 2 (Summer 1998), 30-39.
Friedman, Lee S. Microeconomic Policy Analysis. New York: McGraw Hill:
1984.
Geddes, R. Richard. “Ownership, regulation, and managerial monitoring in the
electric utility industry,” Journal o f Law and Economics 40, no. 1 (1997): 261-
288.
Gee, Benny Yue. “Electric Utility Returns and Risk: An Analysis.” Unpublished
Master’s Thesis. California: California State University, Sacramento, 1990.
Gilbert, Richard J. and Edward P. Kahn, eds. International Comparisons o f
Electricity Regulation. Cambridge: Cambridge University Press, 1996.
Glazer, Craig A. “One Regulator’s Take on Stranded Costs: Let the market do it!”
Public Utilities Fortnightly (June 15, 1998): 22-23.
Grudinin, Nickolai and Ilya Roytelman. “Heading off emergencies in large
electric grids,” IEEE Spectrum 34, no. 4 (April 1997): 42-47.
Hall, Carl T. “Critics Say ISO Rookies Will Hinder Resolving Crisis,” San
Francisco Chronicle. Online at http://www.sfgate.com/cgi-
bin/article.cgi?file= /chronicle/archive/2001 /02/20/MN133473 .DTL February
20, 2001.
Hall, Charles. “Institutional solutions for governing the global commons: Design
factors and effectiveness,” Journal o f Environment and Development 7, no. 2
(1998): 86-114.
Hartman, Raymond S. and Richard D. Tabors. “The Regulatory Contract and Its
Relevance to Stranded Assets Under Restructuring: A Modest Proposal,” 1996.
Online at http://ee.notes.org/tabors.htm, July 28, 1998.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
267
Hildreth, W. Bartley and Carol McMillan. “Kansas Municipal Electric Utilities in
Perspective,” Municipal Finance Journal 19, no. 2 (Summer 1998): 40-56.
Hirsh, Richard F. Power Loss: The Origins o f Deregulation and Restructuring in
the American Electric Utility System. Cambridge: The MIT Press, 1999.
Hirst, Eric and Brendan Kirby. “Maintaining Reliability In A Restructured
Industry,” Public Power (May-June 1998): 24-27.
Hirst, Eric and Brendan Kirby. “Simulating the Operation of Markets for Bulk-
Power Ancillary Services,” The Energy Journal 19, no. 3 (1998): 49-68.
Hsiao, Cheng. Analysis o f Panel Data. Cambridge: Cambridge University Press,
1986.
Information Assurance Task Force. Electric Power Information Assurance Risk
Assessment Report, Washington, DC: National Security Telecommunications
Advisory Committee, 1997.
Jones, Scott T. and Matthew B. Krepps. “Market Share in Generation: The Impact
of Retail Competition on Investor-Owned Utilities,” Public Utilities
Fortnightly 136, no. 13 (July 1, 1998): 46-49.
Joskow, Paul L. “The Political Economy of Market-based Environmental Policy:
The U.S. Acid Rain Program,” Journal o f Law and Economics 41, no. 1
(1998): 37-83.
. “Restructuring, Competition and Regulatory Reform in the U.S.
Electricity Sector,” The Journal o f Economic Perspectives 11, no. 3 (Summer
1997): 119-138.
and Edward Kahn. “A Quantitative Analysis of Pricing Behavior In
California’s Wholesale Electric Market During Summer 2000,” November
2000. Online at www.edisonnews.com/releasesexe/00-147b.pdf, November
27,2000.
and Richard Schmalensee. Markets for Power: An Analysis o f Electric
Utility Deregulation, Cambridge: The MIT Press, 1983.
Kessler, Anke S. “The Value of Ignorance,” Rand Journal o f Economics 29, no. 2
(Summer 1998): 339-354.
Kirlin, John J. “The Big Questions of Public Administration in a Democracy.”
Public Administration Review 56 (September/October 1996): 416-423.
Komesar, Neil K. Imperfect Alternatives: Choosing Institutions in Law,
Economics, and Public Policy. Chicago: University of Chicago Press, 1994.
Kranhold, Kathryn. “Midwestern Firms Cry Foul on Outages,” Wall Street
Journal, July 10, 1998, A2.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 6 8
Krouse, Clement G., Kenneth L. Danger, Christo Cabolis, Tanja D. Carter, Jon M.
Riddle, and Daniel J. Ryan. “The Bell System Divestiture/Deregulation and the
Efficiency of the Operating Companies,” Journal o f Law and Economics 42,
no. 4 (April 1999):61-87.
Kucewicz, William P. “Short Circuit: California’s Self-Imposed Electricity
Crisis,” GeoInvestor.com August 2,2000.
Kuhn, Thomas R., Pradeep Mehra, Robert L. Ball, Richard C Green, Jr., Donald
E. Felsinger, David L. O’Connor, Michael Shames and John R. Hodowal
(1996). “Electric Utility Deregulation Sparks Controversy,” Harvard Business
Review 74, no. 3 (1996): 150-162.
La Porte, Todd R. and Daniel S. Metlay. “Hazards and Institutional
Trustworthiness: Facing a Deficit of Trust,” Public Administration Review 56,
no. 4 (1996): 341-347.
and Ann Keller. “Assuring institutional constancy: Requisite for
managing long-lived hazards,” Public Administration Review 56, no. 6 (1996):
535-544.
Lebamoff, Craig. “Future electricity restructuring might not lower consumers’
costs,” State Tax Review, Chicago, 59, no. 14 (1998): 3-4.
Loehr, George C. “Ten myths about electric deregulation,” Public Utilities
Fortnightly 136, no. 8 (April 15, 1998): 28-32.
Los Angeles Department of Water and Power. “Los Angeles Has Adequate Power
Supplies: Plans to Sell All Excess Energy Within California,” News Release,
December 12,2000. Online at
http://www.ladwp.eom/whatnew/dwpnews/l 21200.htm, January 15, 2001.
Mack, Joel H., James R. Barrett, and Claudia M. O’Brian. “As Utilities
Deregulate, EPA Concentrates on Coal,” National Law Journal 21, no. 15
(1998): B7-B9.
Majone, Giandomenico. Evidence, Argument, and Persuasion in the Policy
Process, New Haven: Yale University Press: 1989.
Meierhoefer, Cameron and Melissa Kelman. “Rethinking Georgia’s Public Policy
in Electric Power,” Georgia Public Policy Foundation, 1997, Online at
http://www.inteliview.com/pwr.htm, July 27, 1998.
Meyers, Glenn D., Buckner Wallingford II and Horace J. DePodwin. “Stranded
Utilities: How Demographics, Not Management, Caused High Costs and
Rates,” Public Utilities Fortnightly (June 1, 1997): 18-24.
Moad, Jeff. “PG&E Dumps CIS Project,” PC Week, April 28,1997, 1, 16.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 6 9
Monmaney, Terence. “L.A.’s Latest Luminary,” Los Angeles Times, December
23, 2000. Online at
http://www.latimes.com/news/state/updates/lat_dwpOO 1223.htm, December
23, 2000.
Morain, Dan. “Davis Walks a Political Tightrope on Energy Prices,” Los Angeles
Times, September 12, 2000. Online at
http://www.latimes.com/news/front/20000912/t00085969.html, September 12,
2000.
Neter, John, William Wasserman and Michael H. Kutner. Applied Linear
Statistical Models: Regression, Analysis o f Variance, and Experimental
Designs, 2nd ed. Homewood, Illinois: Irwin. 1985
Nemec, Richard. “California Energy: Whose Job Is It, Anyway?” California
Journal (March 2000): 34-39.
North American Electric Reliability Council. “Disturbances Analysis Working
Group: Disturbances, Load Reductions, and Unusual Occurrences,” n.d.
Online at http://www.nerc.com/dawg/welcome.html, July 7, 2000.
Newbery, David M. “Competition, contracts, and entry in the electricity spot
market,” Rand Journal o f Economics 29, no.4 (1998): 726-749.
North, Douglass C. Institutions, Institutional Change and Economic Performance.
Cambridge: Cambridge University Press, 1990.
Okun, Arthur M. Equality and Efficiency: The Big Tradeoff Washington, D.C.:
The Brookings Institution, 1975.
O’Toole, Laurence J. Jr. “Implementing Public Programs,” in James L. Perry, ed.
Handbook o f Public Administration, 2nd ed. (San Francisco: Jossey-Bass,
1996), 150-62.
O’Toole, Laurence J. Jr. “Research on Policy Implementation: Assessment and
Prospects.” Journal o f Public Administration Research and Theory 10, no. 2
(2000): 263-288.
Pacific Gas and Electric Company. “Cogeneration and Small Power Production
Semi-Annual Report to the California Public Utilities Commission.” Pacific
Gas and Electric, 2000.
Peyton, Carrie. “Scathing report finds SMUD not ready for free market,”
Sacramento Bee, April 28,1998, A1, A7.
Peace, Steve. “San Diego Electric Bills Are Soaring; Why Has Deregulation
Started So Badly?” Op-ed, San Diego Union-Tribune, July 16, 2000.
Perrow, Charles. Normal Accidents: Living with High-Risk Technologies. New
York: Basic Books: 1984.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 7 0
. Complex Organizations: A Critical Essay. Third edition. New York:
McGrawHill: 1986.
Pressman, Jeffrey L. and Aaron Wildavsky. Implementation. 1973. Reprint
Berkeley: University of California Press, 1984.
Ruff, Larry E. “An Efficient, Competitive Electricity Industry: Can the Vision
Become Reality?” 1998. Online at http://www.electricity-
online.com/tej/ruffll .html, July 28,1998.
Savage, Ian. “Deregulation and Safety in the United States: The Evidence after
Twenty Years,” in Marc Gaudry and Robert Mayes, eds. Taking Stock o f Air
Liberalization. Boston: Kluwer Academic Publishers: 1997.
Schlager, Edella and William Blomquist. “A comparison of three emerging
theories of the policy process,” Political Research Quarterly 49, no. 3 (1996):
651-672.
Schlee, Edward E. “The value of information about product quality,” Rand
Journal o f Economics 27, no. 4 (Winter 1996): 803-815.
Shuit, Douglas P. and Joe Mozingo. “By Luck or Skill, DWP Gets the Last
Laugh,” Los Angeles Times, August 3,2000. Online at
http://www.latimes.eom/news/state/20000803/t000072713.html, 8/4/2000.
Sidak, J. Gregory and Daniel F. Spulber. “Deregulation and managed competition
in network industries,” Yale Journal on Regulation 15, no. 1 (Winter 1998):
117-147.
Smith, Rebecca. "New Rules, Demands Put Dangerous Strain on Electricity
Supply," Wall Street Journal, May 11, 2000, A l, A8.
Spiegel, Yossef and Daniel F. Spulber. “The capital structure of a regulated firm,”
Rand Journal of Economics 25, no. 3 (Autumn1994): 424-438.
Taylor, Carson W. “Improving grid behavior,” IEEE Spectrum 36, no. 6 (1999):
40-45.
Weimer, David L. and Aidan R. Vining. Policy Analysis: Concepts and Practice,
Englewood Cliffs: Prentice Hall: 1992.
Western System Coordinating Council. “News Regarding Power Outages”, n.d.
Online at http://www.wscc.com/outages.htm, July 27, 1999.
White, Matthew W. “Power Struggles: Explaining Deregulatory Reforms in
Electricity Markets,” in Brookings Papers on Economic Activity:
Microeconomics (1996), 201-250.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
271
Williams, Jack. “Understanding heating, cooling degree days,” USA Today, April
17, 2000. Online at http://167.8.29.16/weather/askjack/waskdays.htm,
10/5/2000.
Williamson, Oliver. “The Economics of Organization: The Transaction Cost
Approach,” American Journal o f Sociology 87, no. 3 (1981): 548-577.
Williamson, Oliver. “Transaction Cost Economics and Organizational Theory,” in
Williamson, Oliver E ., ed. Organization Theory: From Chester Barnard to
the Present and Beyond. Oxford: Oxford University Press: 1995.
Williamson, Oliver E. “The Institutions of Governance,” AEA Papers and
Proceedings: The New Institutional Economics 88, no. 2 (May 1998), 75-79.
Wilson, James Q. Bureaucracy: What Government Agencies Do and Why They
Do It. Scranton, PA: Basic Books: 1989.
Winston, Clifford. “U.S. Industry Adjustment to Economic Deregulation,”
Journal o f Economic Perspectives 12, no. 3 (Summer 1988): 89-110.
Wiser, Ryan, William Golove and Steve Pickle. “California’s Electric Market:
What’s in it for the Customer,” Public Utilities Fortnightly 136, no. 15 (1998):
38-47.
Wolf, Charles Jr. “A Theory of Nonmarket Failure: Framework for
Implementation Analysis.” Journal o f Law & Economics 22, no. 1 (1979):
107-139.
Wolfram, Catherine D. “Strategic bidding in a multiunit auction: an empirical
analysis of bids to supply electricity in England and Wales,” Rand Journal o f
Economics 29, no. 4 (1998): 703-725.
Woolfolk, John. “Offer to sell dams raises hopes, fears,” San Jose Mercury News,
July 18,2000. Online at http://sjmercury.com/local/center/dam0719.html,
7/19/2000.
Yankelovich, Daniel. Coming to Public Judgment: Making Democracy Work in a
Complex World. Syracuse: Syracuse University Press: 1991.
Young, H. Peyton. Equity: In Theory and Practice. Princeton, NJ: Princeton
University Press: 1994.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 7 2
Appendix A
Interview Protocol
Derived from the DOE Power Outage Study Team
1. Please describe your organization’s participation in the electric industry
and your individual role.
2. What major actions taken by your organization derive from the passage of
AB 1890?
3. Does your organization have an official position on SB 1388, or other
active California legislation on electric restructuring?
4. What is your opinion on the floating barge being towed to California by
PG&E?
5. Do you feel both reliability and reasonable price can remain constant as
restructuring continues? (If not, what degree of tolerance is required for
variability in price or reliability)?
6. The proposed price cap resolution discussed at the July 6th,2000 special
meeting of the ISO and passed at the August 1st meeting called for
"immediate implementation" of a set of reforms. How rapidly do you feel
legislative and regulatory bodies can create changes in hedging, load
management, and metering?
7. Do you feel that traditional methods for supplying power (e.g.,
construction of new transmission and central generation) are able to keep
up with load growth? If not, what alterations are needed?
8. Do you feel incentives for reactive power production are adequate? If not,
what types of incentives are needed?
9. Do you feel mechanisms for sharing information on maintenance best
practices and equipment performance among distribution utilities are
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
273
adequate? If not, what types of mechanisms would improve information
sharing?
10. In your opinion, have utilities experienced lengthy delays in replacing
failed equipment? If so, why?
11. To your knowledge, have planned distribution system upgrades been
implemented on schedule? If not, why not?
12. How do substation maintenance programs anticipate component
weaknesses? What changes in these programs are suggested by load
changes associated with restructuring?
13. Have transmission and distribution maintenance expenditures declined
over time? Are these expenditures adequate?
14. What business factors, if any, are most significant or may compromise
reliability performance (e.g., out of state or foreign ownership, more
power contracted rather than owned generation)?
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 7 4
Appendix B
Contacts
In writing this dissertation, information, insight and encouragement has
come from national, state, and local officials as well as participants in the electric
system restructuring process. I am grateful for each opinion and all assistance,
especially with locating supporting materials. These persons are listed below,
along with their position and affiliation at the time of contact. Interview
participants are marked with an asterisk.
A&A Arnold & Associates, Inc., Hialeah, FL
Velazques, Arnold M., President. August 1, 2000.
California Energy Commission, Sacramento CA
Baird, Larry, Policy Analyst. July 19,2000 et seq.
Gorin, Tom, Demand Analysis Office. September 6,2000.
Fukumoto, Dennis. November 20,2000.
California Independent System Operator, Folsom, CA
* Winter, Terry, President and CEO. October 3,2000.
California Large Energy Consumers Association
*Barkovich, Barbara. Also a member of the ISO Board of Governors, October 30,
2000.
California Manufacturers & Technology Association and Proctor and Gamble
*Roscoe, Stacy. Also a member of the ISO Board of Governors. August 4, 2000.
California Public Utilities Commission
Schumacher, Brian D. October 13, 2000.
California State Assembly
Wright, Roderick, Assemblyman. February 20, 2000.
Pescetti, Anthony, Assemblyman. January 22, 2001 et seq.
The Center for Energy and Economic Development
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 7 5
Ross, Terry, Vice President. December 1, 2000 et seq.
The County o f Los Angeles
Muensch, Bob. December 15, 2000.
Duke Energy Field Services
J.R. McPherson, Director of Government Affairs. January 29,2001.
Environmental Energy Technologies Division, Lawrence Berkeley National
Laboratory
Koomey, Jon. February 9,2001 et seq.
Flanagan Consulting Group, Inc.
Flanagan, Dan. January 24, 2001 et seq.
Independent Energy Producers Association
*Smutny-Jones, Jan, also a member of the ISO Board of Governors. August 22,
2000.
Los Angeles Department o f Water and Power
Tharp, Eric. Sep 21, 2000 et seq.
Pacific Gas and Electric
*Hapner, Dede, also a member of the ISO Board of Governors. July 25, 2000.
Timmerman, Mark. November 20,2000 et seq.
Public Financial Management, Newport Beach, CA
Curry, Keith D., Managing Director. August 1, 2000.
Sacramento Municipal Utility District, Sacramento, CA
Blomo, Vito, Principal Resource Planner. August 25,2000 et seq.
Cotton, Isaac, Load Management Supervisor. October 20,2000 et seq.
Frantz, Stephen, Marketing Specialist. December 15, 2000 et seq.
Inbody, Jana, Demand Side Specialist. October 20,2000 et seq.
LaBranche Ed, P.E., Planning Process Coordinator. November 21,2000 et seq.
*Schori, Jan, General Manager. August 21,2000 et seq.
Starkovich, Arthur. Public Information Specialist. September 8, 2000 et seq.
Weedall, Michael, Manager, Energy Services
State o f Nevada, Carson City, NV
Guinn, Kenny, Governor. August 1, 2000.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 7 6
State o f Nevada, Office o f the Attorney General.
*Hay, Tim, Chief Deputy Attorney General. Previously on the ISO Advisory
Board. January 24, 2001 et seq.
State o f California, Department o f Water Resources
* Patel, Viju, Executive Manager, Power Systems. Also a member of the ISO
Board of Governors. October 17,2000.
Southern California Edison
*Fielder, John, CEO, also a member of the ISO Board of Governors. August 11,
2000.
Zamorano, Manuel. October 10, 2000 et seq.
Southern California Public Power Association
* Carnahan, Bill, also a member of the ISO Board of Governors. October 24,
2000.
U.S. Senate, Washington, DC
Murkowski, Frank H., Senator and Chair of the Energy and Natural Resources
Committee. August 2, 2000.
Western Area Power Administration
*Toenyes, Jerry, also a member of the ISO Board of Governors. August 2,2000.
Woods and Daube
* Michael Woods, also a member of the ISO Board of Governors. December 12,
2000.
* Formally interviewed
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 7 7
Appendix C
Data Sources
This dissertation brought together data from many sources including the
California utilities, state and federal regulatory agencies, required filings with the
Securities and Exchange Commission and materials presented by participants at
various public and legislative hearings. These sources are listed below.
Reliability Data
CPUC Decision 96-09-045 requires specific reporting, including: system
indices (SAIDI, SAIFI, and MAIFI), significant outages for the year, and
historical indices for the preceding 10 years. The IOUs are bound to comply with
this reporting requirement; municipal systems such as SMUD and LADWP are
not, though they have provided some of this information.
1997 Reliability Report for Pacific Gas & Electric Company
1998 Reliability Report for Pacific Gas & Electric Company
1999 Reliability Report for Pacific Gas & Electric Company
1997 Reliability Report for Southern California Edison
1998 Reliability Report for Southern California Edison
1999 Reliability Report for Southern California Edison
1997 Reliability Report for San Diego Gas & Electric
1998 Reliability Report for San Diego Gas & Electric
1999 Reliability Report for San Diego Gas & Electric
1996-2000 Reliability Graphs for Los Angeles Power and Light
1999 TB&A Transmission & Distribution Best Practices Questionnaire (1998
LADWP data)
2000 TB&A Transmission & Distribution Best Practices Questionnaire (1999
LADWP data)
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
278
The North American Electric Reliability Council (NERC) includes a Disturbance
Analysis Working Group that produces a database of DOE reported outages. This
paper includes DAWG data from 1990-1999. NERC also constructs winter and
summer projections. California’s energy requirements comprise part of the
Western Systems Coordinating Council (WSCC) projections.
1996 Summer Assessment
1997 Summer Assessment
1998 Summer Assessment
1999 Summer Assessment
2000 Summer Assessment
1996/1997 Winter Assessment
1997/1998 Winter Assessment
1998/1999 Winter Assessment
1999/2000 Winter Assessment
Utility Published Reports and Filings
City of Los Angeles, Department of Water and Power
Official Statement, LADWP Electric Plant Revenue Bonds, Issue of 1993
City of Los Angeles, Department of Water and Power Annual Report 97/98
City of Los Angeles, Department of Water and Power Annual Report 98/99
Pacific Gas and Electric 1996 Annual Report
Pacific Gas and Electric 1997 Annual Report
Pacific Gas and Electric 1998 Annual Report
Pacific Gas and Electric 1999 Annual Report
Pacific Gas and Electric 1994 10-K (filed with Security and Exchange
Commission)
Pacific Gas and Electric 1997 10-K (filed with Security and Exchange
Commission)
Pacific Gas and Electric 1998 10-K (filed with Security and Exchange
Commission)
Pacific Gas and Electric 1999 10-K (filed with Security and Exchange
Commission)
Sacramento Municipal Utility District Fingertip Facts 1990-91
Sacramento Municipal Utility District Fingertip Facts 1992
Sacramento Municipal Utility District Fingertip Facts 1994
Sacramento Municipal Utility District Fingertip Facts 1995
Sacramento Municipal Utility District Fingertip Facts 96/97
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 7 9
Sacramento Municipal Utility District Fingertip Facts 97/98
Sacramento Municipal Utility District Fingertip Facts 1999/2000
Sacramento Municipal Utility District Fingertip Facts 2000-2001
Sacramento Municipal Utility District Official Statement for Electric Revenue
Refunding Bonds, 1992 Series C
Sacramento Municipal Utility District Draft Five-Year Business Plan for
Distribution Services, 2000.
San Diego Gas and Electric 1994 10-K (filed with Security and Exchange
Commission)
San Diego Gas and Electric 1995 10-K (filed with Security and Exchange
Commission)
San Diego Gas and Electric 1996 10-K (filed with Security and Exchange
Commission)
San Diego Gas and Electric 1997 10-K (filed with Security and Exchange
Commission)
San Diego Gas and Electric 1998 10-K (filed with Security and Exchange
Commission)
San Diego Gas and Electric 1999 10-K (filed with Security and Exchange
Commission)
San Diego Gas and Electric 2000 10-K (filed with Security and Exchange
Commission)
San Diego Gas and Electric 1996 Annual Report
San Diego Gas and Electric 1997 Annual Report
San Diego Gas and Electric 1998 Annual Report
Other Documents
California Energy Commission’s Database of California Power Plants
2000 PL ANTS .XLS, current as of June 22, 2000.
The CEC power plants database lists all power plants in California one-tenth (0.1)
megawatt or larger by name, capacity, date placed in service, the county where
each is located, the owning company, the operating utility, contacts with
telephone numbers., address, phone number and associated information.
California ISO worksheet of outages and emergencies
OUTAGES.XLS
Declared Staged Emergencies for 1998, 1999, and 2000
California ISO publishes its outages online in PDF format.
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
280
Appendix D
Glossary of Terms
Electric system terminology has been defined throughout when
introduced. This glossary serves as a reminder for later uses of these sometimes
arcane terms.
AMR Automated meter reading, a technology approach that replaces
old meters with sophisticated programmable meters that can be
read remotely, as well as provide realtime pricing and ancillary
services
CEC California Energy Commission
CPUC California Public Utilities Commission
CTC Competition (or competitive) transition charge, the method
used for collection of stranded costs during the transition
period.
curtailable
load Power available for shedding by a customer, retail or
commercial. Curtailable load can be controlled by the
customer or remotely by the utility, and increases demand
elasticity
DA Direct access, a form of retail wheeling where an end customer
selects a generation supplier, whose energy is delivered by the
area’s monopoly UDC
DOE U.S. Department of Energy
DSM Demand Side Management
EE. Energy Efficiency
EOB Electricity Oversight Board, oversees the PX and ISO. FERC
has recommended disbanding this board
energy
efficiency Practices that reduce the amount of energy consumed,
especially without altering the output. For instance, a compact
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
281
ESP
FERC
IOUs
IPP
ISO
kW
kWh
MW
NERC
NRC
0 0 M
PBR
PG&E
premium
power
PURPA
PX
QF
reliability
RTO
SAIDI
florescent fixture may consume 18 watts of energy but
produce the equivalent light as a 60 watt incandescent bulb.
Energy service provider. In CPUC reports, the provider
involved in a direct access transaction with the formally
vertical UDC
Federal Energy Regulatory Commission
Investor-owned utilities
Independent power producer
California Independent System Operator
Kilowatt, one thousand watts
Kilowatt-hour
Megawatt, one million watts
North American Electric Reliability Commission
Nuclear Regulatory Commission
Out of market, a term for purchases by the ISO from external
parties
Performance-based ratemaking, used by San Diego Gas and
Electric
Pacific Gas and Electric Company
An ESP service to assure a higher stander power quality than
that guaranteed by the local monopoly UDC
Public Utility Regulatory Policies Act of 1978
California Power Exchange, the wholesale power pool
qualifying facility; independent power producers with
favorable, established contracts to supply generation to the
IOUs
Adequacy to supply electric demand at all times despite
planned and unplanned outages of system facilities, and
security to withstand sudden disturbances. (NERC)
Regional Transmission Organization
System Average Interruption Duration Index, an outage metric,
time per customer
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.
2 8 2
SCE
SONGS
stranded
costs
UDC
wscc
WSPP
Southern California Edison Company
San Onoffe Nuclear Generating Station
The term describing the purported loss of value to assets built
under the traditional rate-of-retum regulatory model.
Compensating utilities for these costs was accomplished in
California by the collection of the CTC
Utility distribution company. The remnant of the former
vertical monopoly, theoretically without responsibility for
either generation or transmission, reconstructed to manage the
low-voltage distribution of electricity to consumers
Western Systems Coordinating Council
Western Systems Power Pool
R eproduced with perm ission of the copyright owner. Further reproduction prohibited without perm ission.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Internal venturing in public agencies

PDF

Biopharmaceutical strategic alliances: Interorganizational dynamics and factors influencing FDA regulatory outcomes

PDF

Business political influence on agency regulatory decision making: The efficacy of corporate political activity in Taiwan's CATV license awarding

PDF

An analysis of third -party regulatory audit systems for medical device safety

PDF

Androgyny and transformational leadership: Effects of gender and sex -role identity in the collectivistic context of Taiwan, R.O.C.

PDF

Demand -side obstacles to girls' education in developing countries

PDF

A comparative examination of the impact of business -government relations on labor market reform in Egypt and Mexico, 1975--1995

PDF

Determining risk propensity of government program managers for high risk /high payoff projects

PDF

Detecting the effects social and business pressures on small California trucking firm tax compliance

PDF

Conversion of health care organizations from non -profit to for -profit status

PDF

Assessing United States information assurance *policy response to computer -based threats to national security

PDF

Determining acceptable seismic risk: A community participation-based approach

PDF

Hospital conversions: The California experience

PDF

By sea: The port nexus in the global commodity network (the case of the West Coast ports)

PDF

A financial audit model for entrepreneurial governments

PDF

Evaluation of professional services consultants in rural government

PDF

Charter schools' influence on public school administrators' innovative behaviors

PDF

Dependence on foreign labor, quality of education and unemployment in the GCC countries: In search of solutions

PDF

Essays on technological evolution and financial returns to innovation

PDF

Getting 'how' and 'why' straight: A critical discourse analysis of the National Partnership for Reinventing Government's ideological discourse on information and communication technologies

Asset Metadata

Creator Horiuchi, Catherine Miller (author)

Core Title Electric system restructuring and system reliability

School Graduate School

Degree Doctor of Public Administration

Degree Program Public Administration

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Economics, Commerce-Business,Energy,OAI-PMH Harvest,Political Science, public administration

Language English

Contributor Digitized by ProQuest (provenance)

Advisor Graddy, Elizabeth (committee chair), Cicchetti, Charles J. (committee member), Clayton, Ross (committee member)

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c16-98893

Unique identifier UC11338238

Identifier 3027725.pdf (filename),usctheses-c16-98893 (legacy record id)

Legacy Identifier 3027725.pdf

Dmrecord 98893

Document Type Dissertation

Rights Horiuchi, Catherine Miller

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the au...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus, Los Angeles, California 90089, USA