Distributed Energy Generation

Home , Brittle Power, Duke Energy

DISTRIBUTED ENERGY GENERATION SYSTEMS BASED ON RENEWABLE ENERGY AND NATURAL GAS BLENDING: NEW BUSINESS MODELS FOR ECONOMIC INCENTIVES, ELECTRICITY MARKET DESIGN AND REGULATORY INNOVATION

Joseph Nyangon

A dissertation submitted to the Faculty of the University of Delaware in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Energy and Environmental Policy

Fall 2017

DISTRIBUTED ENERGY GENERATION SYSTEMS BASED ON RENEWABLE ENERGY AND NATURAL GAS BLENDING: NEW BUSINESS MODELS FOR ECONOMIC INCENTIVES, ELECTRICITY MARKET DESIGN AND REGULATORY INNOVATION

Joseph Nyangon

Approved: ______John Byrne, Ph.D. Professor in charge of the dissertation on behalf of the Advisory Committee

Approved: ______Syed Ismat Shah, Ph.D. Interim Director of the Energy and Environmental Policy Program

Approved: ______Babatunde A. Ogunnaike, Ph.D. Dean of the College of Engineering

Approved: ______Ann L. Ardis, Ph.D. Senior Vice Provost for Graduate and Professional Education

I certify that I have read this dissertation and that in my opinion it meets the academic and professional standard required by the University as a dissertation for the degree of Doctor of Philosophy.

Signed: ______John Byrne, Ph.D. Professor in charge of dissertation

Signed: ______J. Mack Wathen MBA Member of dissertation committee

Signed: ______Steven Cohen, Ph.D. Member of dissertation committee

Signed: ______William Latham III, Ph.D. Member of dissertation committee

ACKNOWLEDGMENTS

This dissertation represents a crystallization of my research journey in the domains of distributed utilities and renewable electrcity integration originally conceived almost six years ago. It began when I was a graduate student at Columbia

Universitys’s School of International and Public Affairs (SIPA). A fortunate involment with the university’s Earth Institute practicum program featuring lectures by Professors Jeffrey Sachs, Steven Cohen, and Vijay Modi of mechanical engineering department provided my first exposure to how changes in business practices and technological innovations are reshaping the electricity sector. To my complete surprise, that exposure radically undermined some of my basic conceptions about the utility landscape and the interconnected business model innovation, technological advancements, customer preferences, and market forces that are driving this transformation. Whatever their pedagogic utility, those conceptions drawn from a misconception that distributed utilities were costlier and thus could not compete effectively with the centralized utility network did not at all fit the enterprise that historical data was already displaying. The result was a dramatic shift in my career trajectory, a shift from approaching this problem as a technological efficiency dilemma, to an interdisciplinary policy, engineering, and economics challenge. Without any doubt the ideas set forth in this dissertation substantially benefited from discussion and critique involving many scholars who are not readily identifiable by name. I sincerely thank my resourceful and indefatigable committee members who helped me to crystallize issues at the outset of this study. My primary advisor, and

iv mentor at the University of Delaware, Center for Energy and Environmental Policy, Professor John Byrne planted in me a curious search for scholarly direction. He has supported my work since our first meeting and set me to reading works of prominent theorists of political economy, politics of energy transitions and post-modernity notably Jacques Ellul, Thomas S. Kuhn, Lewis Mumford, David Ricardo, Joseph A. Schumpeter, Langdon Winner, and Joan Martinez-Alier. In addition, the long discussions, helpful suggestions, and enlightened discourse with Professor Byrne have strengthened the narrative of this dissertation substantially, shaped my conception of what the business incentives of distribution utilities ideas can be, and continue to provide a powerful—indeed, inspirational—model for me. Thank you for your continued support, insightful guidance, and patience over the past four years. Professor William Latham introduced me to advanced econometric modeling of energy systems and provided valuable insights in this area which helped to sharpen the analysis and conclusions. J. Mark Wathen’s comments polished my dissertation a lot. Because he understands how challenging energy transformation is to the electric industry than I do, I could adjust my discussion of the results and policy implications to reflect the raw reality of these issues to utility business. Professor Steven Cohen agreed to join my committee despite his busy schedule at the helm of the management of Columbia University’s Earth Institute. He has been a mentor since we met during my studies at Columbia, and my thinking on regulatory innovation—and on so much else—has benefited immensely from our conversations over the years. Finally, I acknowledge the privileged consultations and advice which I have shared with the remaining faculty throughout my four years here at CEEP—an institution that have helped give form to my thought.

v I owe my girlfriend, Jen, a great debt of gratitude. She has both challenged and inspired me in equal measure and her keen vision and intellectual energy are without equivalent. As the youngest in a family of many firsts and strong academic performance, my parents, brothers, and sisters have been a locus of endless love, creativity, and unwavering support and encouragement as they lead by example. Finally, I would like to thank all my colleagues at CEEP especially Job Taminiau, Mayuri Utturkar, Kathleen Saul, Joohee Lee, Jeongseok Seo, Jiajing (Athena) Bi, Nabeel Al Abbas, Soojin Shin, Benjamin Attia, and everyone I met along the way who have been a vital source of support and encouragement.

vi TABLE OF CONTENTS

LIST OF TABLES ...... xii LIST OF FIGURES ...... xiv LIST OF ACRONYMS AND ABBREVIATIONS ...... xix ABSTRACT ...... xxv

Chapter

1 INTRODUCTION ...... 1

1.1 Statement of Research ...... 1 1.2 Evolution of Electric Power Sector in the United States ...... 3

1.2.1 Signposts of Distributed Electricity Generation Expansion ...... 5 1.2.2 Emerging Challenges of Distributed Generation on Bulk Power System ...... 9

1.3 The Case for a More Distributed Energy Future ...... 20 1.4 Research Design and Methodology ...... 21 1.5 Organization of Chapters ...... 22

2 TRANSFORMING THE U.S. ELECTRICITY SYSTEM: BUSINESS MODEL INNOVATION ...... 25

2.1 Background ...... 25

2.1.1 Overview of Utility Sector in the United States ...... 25 2.1.2 Disruptive Triggers in the Utility Industry ...... 32 2.1.3 Distributed Generation Narratology: The Détente for Natural Gas and Renewable Energy Blending ...... 36 2.1.4 Benefits of Gas-Renewable Energy Partnering ...... 42

2.1.4.1 Carbon Mitigation ...... 49 2.1.4.2 Energy System Integration ...... 50 2.1.4.3 Synergistic Benefits of NG-RE Hybrid Power Generation Opportunities ...... 52

2.2 The Battle Over Centralization of Economic and Physical Control of Electricity Generation ...... 57 2.3 A Brief History of the Dominant U.S. Energy Policy Paradigm ...... 59

2.3.1 Defending the Centralized Energy Policy Approach in an Increasingly Complex and Changing Environment ...... 62

vii 2.3.2 Exclusion of Alternative Voices ...... 63 2.3.3 Weakening Role of States ...... 65 2.3.3 Loss of Alternative Voices ...... 66

3 EVALUATION OF TRENDS DRIVING POWER SECTOR TRANSFORMATION IN THE UNITED STATES ...... 75

3.1 Key Drivers of Change ...... 75

3.1.1 The Shifting Energy Generation Mix; Power System is Decarbonizing and Becoming More Distributed ...... 81

3.1.1.1 Rapid and Unprecedented Expansion of Intermittent Renewable Electricity in the U.S...... 81 3.1.1.2 Growth in Natural Gas Production and Low Cost Shale ...... 89

3.1.2 Flat to Declining Electricity Demand Growth ...... 93 3.1.3 Aging Infrastructure Imperatives ...... 93 3.1.4 Innovations in Data, Intelligence, and System Optimization .... 94 3.1.5 Revenue and Investment Challenges ...... 94 3.1.6 Evolving Customer Engagement ...... 102 3.1.7 Diverse Participation in Power Markets and Changing Utility Business Models ...... 104 3.1.8 Increasingly Integrated Power System and Interactions with Other Sectors ...... 104 3.1.9 Energy Security, Resilience, and Reliability Objectives ...... 105 3.1.10 Climate Change and Environmental Concerns over Air Emissions ...... 108

3.1.10.1 Impacts on Electricity Generation ...... 108 3.1.10.2 Impacts on Electricity Transmission and Distribution ...... 110 3.1.10.3 Impacts on Electricity Demand and Consumption . 111

3.2 Theoretical Framework ...... 113

3.2.1 Business Model Definitions ...... 113 3.2.2 New Business Models and Innovation ...... 118 3.2.3 Transformation Through Shared Visions ...... 125 3.2.4 Utility-Centric Business Model ...... 128 3.2.5 Customer-Centric Business Model ...... 132

3.2.4.1 The Sustainable Energy Utility (SEU) ...... 138

viii 3.3 Hamel Business Model Theory ...... 140

3.3.1 Characteristics of Hamel Business Model Concept ...... 144 3.3.2 Four Major Components of Hamel Business Model Framework ...... 146

3.4 Towards a New DER Electricity Service Vision: Détente for Distributed Energy Resources ...... 149

4 BLENDING FAST-FLEXING RENEWABLE GENERATION SYSTEMS AND FLEXIBLE NATURAL GAS TECHNOLOGIES FOR A CLEAN ELECTRICITY SYSTEM OF THE FUTURE...... 152

4.1 Theoretical Approaches and Tools ...... 154

4.1.1 Matching Demand and Supply Instantaneously ...... 154 4.1.2 California ISO’s Duck Curve ...... 156

4.2 Integrated Assessment Models ...... 159 4.3 Data Description, Statistics, and Assumptions ...... 163

4.3.1 Load and Fast-flexing Renewable Electricity Data ...... 163 4.3.2 Flexible-Baseload Natural Gas Generation Data ...... 166 4.3.3 Other Baseload Fossil Electricity Generation Data ...... 171

5 EVALUATION OF POLICY AND ECONOMICS OF INTEGRATING DISTRIBUTED UTILITIES ...... 181

5.1 Policy Platforms for Renewable Generation ...... 181

5.1.1 Renewable Feed-in Tariffs ...... 181 5.1.2 State RPS Policies ...... 183 5.1.3 Net metering ...... 189 5.1.4 Public Benefits Fund ...... 192 5.1.5 Interconnection Standards ...... 196

5.2 Economic and Environmental Control Variables ...... 197

5.2.1 League of Conservation Voters ...... 197 5.2.2 Average Electricity Price ...... 198 5.2.3 Electricity Import Ratio ...... 204 5.2.4 Energy Intensity and Per-capita Energy-related CO2 Emissions ...... 206 5.2.5 Per-capita Real GDP ...... 209

ix 5.3 Estimating the Basic Vector of Policy and Economic Controls ...... 210

6 ESTABLISHING A DISTRIBUTED UTILITIES MODEL FOR ESTIMATING THE EFFECTS OF FLEXIBLE NATURAL GAS TECHNOLOGIES ON RENEWABLE GENERATION ...... 222

6.1 Overview of the Methodology Framework ...... 222 6.2 Model Establishment ...... 222

6.2.1 Model Description ...... 222 6.2.2 Dynamic Econometric Model ...... 232 6.2.3 Estimation Techniques of the Model ...... 234

6.3 Empirical Results and Discussion ...... 237 6.4 Parameter Estimations of Policies and Control Variables ...... 246

6.4.1 Effects of Flexible-Natural Gas Technologies on Fast-flexing Renewable Energy Diffusion ...... 247 6.4.2 Effects of Fast-flexing Renewable Energy Technologies on Flexible-Natural Gas Diffusion ...... 255 6.4.3 Effects of Generation Technologies (Other Than Natural Gas) on Renewable Energy Diffusion ...... 260

6.4. Successful Business Models to Promote NG-RE Blended Power Generation ...... 264

6.4.1 Non-Wires Alternatives Model ...... 266 6.4.2 Utility as a Smart Integrator ...... 274 6.4.3 Electric Services Operator Model ...... 275

7 A ROADMAP FOR DELIVERING NG-RE HYBRID POWER GENERATION AND DISTRIBUTED RESOURCES: CASE STUDY OF NEW YORK ...... 285

7.1 Evaluating the REV Docket: The Détente for Utilities and DER ...... 285

7.1.1 From a Centralized and Incentive-driven Model to a Distributed System Platform ...... 287 7.1.2 Strategic Resources ...... 300 7.1.2 Customer Interface ...... 307 7.1.3 Value Network ...... 310 7.1.4 Core Strategy ...... 311

7.2 Enabling Higher Penetration of DER in New York ...... 312

x 7.3 Beyond Utility 2.0: A Shift to Utility 3.0 and Energy Democracy ...... 315 7.4 Five Pillars of Utility 3.0 (Energy Democracy) Framework ...... 325

8 CONCLUSION AND POLICY RECOMMENDATIONS ...... 326

8.1 Summary of Conclusions ...... 326 8.2 Barriers to Existing Policies and Mechanisms ...... 332

8.2.1 Market Barriers ...... 333 8.2.2 Financial Barriers ...... 334 8.2.3 Economic Regulation and Policy Barriers ...... 334

8.3 Recommended Policies and Regulatory Mechanisms ...... 335

8.3.1 Develop Long-term Efficient Price Signals and Incentives for DER...... 336 8.3.2 Create Joint Innovation Programs Across States to Share Best Practices ...... 337 8.3.3 Create Explicit Policy Incentives for Long-term Innovation ... 337 8.3.4 Enhance Market Transparency by Enabling Participation in Long-term Capacity Markets ...... 342

8.4 Climate Change Challenges ...... 343 8.5 Next Generation Utilities and System Flexibility ...... 343

REFERENCES ...... 347

Appendix

A. STATE NET METERING CAPACITY ...... 378 B. ELECTRICITY IMPORT RATIOS, BY STATE, 2001-2015 ...... 379 C. PANEL GENERALIZED METHOD OF MOMENTS DATA ANALYSIS ...... 380

C.1 Electric Power Sales, Revenue, Customers, Service Type, and Ownership in NYS in 2015 ...... 380 C.2 Tabulation of Electric Power Ownership in NYS in 2015 ... 385 C.3 Panel Generalized Method of Moments Output Results ...... 386

D. PERCENTAGE CHANGE IN ENERGY INTENSITY ...... 412 E. SUPPLEMENTARY FIGURES AND DATA ...... 414 F. COPYRIGHT PERMISSION LETTER ...... 418

xi LIST OF TABLES

Table 2.1 Power System Pathway Origins ...... 35

Table 2.2 Optimized Diverse Electricity Portfolio of Natural Gas and Renewables ...... 45

Table 2.3 Matrix of Comparative Benefits of Natural Gas and Renewable Energy ...... 53

Table 2.4 Value Proposition of NG-RE Hybrid Power Generation Synergies ...... 55

Table 2.5 Emerging Pathways of Power System Transformation ...... 73

Table 3.1 Ten Trends Transforming the Electric Power Sector ...... 79

Table 3.2 Potential Market Design Adaptations Needed to Address Supply- related Challenges at High Renewable Energy Penetration ...... 99

Table 3.3 Selected Exemplars Interpreting Business Models as Formal Conceptual Representations ...... 114

Table 3.4 The Business Model Evaluation ...... 115

Table 3.5 Evolution of Solar PV Business Models ...... 118

Table 3.6 Utility-side Vs. Customer-side Business Model ...... 134

Table 5.1 Top EFFRPS Values for Each States with an RPS Target ...... 188

Table 5.2 State Level Public Benefits Fund Actions ...... 193

Table 6.1 Descriptive Statistics for Annual U.S. National Observations .. 224

Table 6.2 Descriptive Statistics for Monthly State Level Observations .... 227

Table 6.3 Descriptive Statistics of Observations, By State ...... 230

Table 6.4 Appropriate Lag Structure for FFRET and FBNGT Technologies ...... 237

Table 6.5 Pooled Regression Results, Share of Renewable Installed Capacity ...... 239

xii Table 6.6 Estimated Results of Renewable Installed Capacity Using Popp et al. (2011) Specification ...... 248

Table 6.7 Re-estimated Results of Renewable Installed Capacity with Smaller Number of Lags ...... 253

Table 6.8 Empirical Results of Share of Flexible-Baseload Natural Gas Installed Capacity ...... 256

Table 6.9 Empirical Results of Additional Share of Other Technologies . 262

Table 6.10 Non-wires Alternatives Projects ...... 267

Table 6.11 Alternative Utility Business Models Proposed or Implemented by States ...... 277

Table 7.1 Studies Reviewed that Relate to Evaluation of the REV Business Model ...... 291

Table 7.2 Application of Hamel Business Model to Conventional Energy Utility ...... 299

Table 7.3 Top 25 Utilities in New York State, By Revenue and Service Type, 2015 ...... 302

Table 7.4 Tabulation of Ownership of NYS Utilities, 2015 ...... 308

Table 7.5 Distributed Utilities Framework of Utility 2.0 Business Model 319

Table 8.1 Tabulation of Utility Incentives ...... 339

Table 8.2 Summary of Policy Implications ...... 345

xiii LIST OF FIGURES

Figure 1.1 Primary Energy Production by Source, 1949-2016 ...... 7

Figure 1.2 Survey of Obstacles to Evolution of Utility’s Business Model ... 17

Figure 2.1 Evolutionary Landscape of Disruptive Triggers in the Utility Industry ...... 31

Figure 2.2 The Impact of Megatrends and Disruption Dynamics on the Power Sector ...... 35

Figure 2.3 Blueprint of a Fully Integrated DER System ...... 41

Figure 2.4 U.S. Primary Energy Consumption by Source and Sector, 2015 43

Figure 2.5 Average Capacity Factors of Natural Gas Combustion Turbines by NERC ...... 44

Figure 2.6 Existing and New Generating Resources for Phase III (MW) Moderate Federal Carbon Policy ...... 48

Figure 2.7 Example of Interrelationships among Areas of Energy Integration ...... 52

Figure 2.8 PJM’s Hourly Demand Bid Data for 2016 ...... 71

Figure 3.1 Ten Trends Driving Power System Transformation in the United States ...... 78

Figure 3.2 Cumulative Installed Renewable Capacity by Scenario ...... 83

Figure 3.3 Historical and Projected Solar PV Capacity by Sector, 2008 - 2020 ...... 84

Figure 3.4 U.S. PV installation forecast by segment, 2010-2021 ...... 85

Figure 3.5 Contracted Utility PV Pipeline in 2016 ...... 86

Figure 3.6 U.S. Polysilicon, Wafer, Cell, and Module Prices ...... 88

Figure 3.7 Henry Hub Spot Prices, SUNIDX and S&P GSCI Indices, 2000 – 2017 ...... 90

Figure 3.8 Taxonomy of Distributed and Renewable Energy Resources. .... 92

xiv Figure 3.9 Projections of Utility-sector Revenue Erosion in the United States Due to Distributed Energy Resources ...... 97

Figure 3.10 Total Electricity Retail Sales ...... 98

Figure 3.11 U.S. Natural Gas Proved Reserves and Marketed Production, 1971–2013 ...... 107

Figure 3.12 Future Market and Business Models ...... 116

Figure 3.13 Smart Integrator: Utility as Network Integrator ...... 121

Figure 3.14 Energy Service Utility (ESU) Business Model ...... 122

Figure 3.15 Two Generic Utility Business Models in the Electricity Value Chain ...... 131

Figure 3.16 New PV Business Models Focused on System Ownership and Control ...... 133

Figure 3.17 Customer-owned and Utility-controlled Value Network ...... 136

Figure 3.18 Utility-controlled and -owned Value Network ...... 137

Figure 3.19 Components of Hamel Framework for Business Concept Innovation ...... 145

Figure 4.1 Major Relationships in the Retail Energy Ecosystem ...... 155

Figure 4.2 California’s Duck Curve Showing Steep Ramping Needs ...... 158

Figure 4.3 Top-10 Solar Generating States in 2016 ...... 164

Figure 4.4 Project Sunroof County-Level Coverage for 2017 ...... 165

Figure 4.5 Top 10 Cities with Most Solar Potential ...... 165

Figure 4.6 Fast-flexing Renewable Energy Share of Installed Capacity, 2001-2016 ...... 169

Figure 4.7 Flexible-Baseload Natural Gas Share of Installed Capacity, 2001- 2016 ...... 170

Figure 4.8 Other Baseload Share of Installed Capacity, 2001-2016 ...... 172

xv Figure 4.9 Scatter Plot of FBNGT and FFRET Shares, ...... 173

Figure 4.10 Monthly Normalized Relationship between FBNGT and FFRET for Individual States, 2001 to 2016 ...... 174

Figure 4.11 Scatter Plot of BLFT and FBNGT Shares, 2001 to 2016 ...... 176

Figure 4.12 Scatter Plot of BLFT and FFRET Shares, 2001 to 2016 ...... 177

Figure 4.13 Monthly Normalized Relationship between FBNGT and BLFT for Individual States, 2001 to 2016 ...... 178

Figure 4.14 Monthly Normalized Relationship between BLFT and FFRET for Individual States, 2001 to 2016 ...... 179

Figure 5.1 Distribution of Renewable Portfolio Standards and Goals ...... 184

Figure 5.2 Summary of Solar Policy Actions in 2016 ...... 185

Figure 5.3 States with Mandatory Net Metering Rules ...... 190

Figure 5.4 Comparison of Total Installed New Metering Capacity in AZ, CA, MA, and NY in 2016 ...... 191

Figure 5.5 Comparison of Total Energy Sold Back in AZ, CA, MA, and NY ...... 192

Figure 5.6 States with Public Benefits Fund Policies and Incentives ...... 196

Figure 5.7 States with Interconnection Standards Laws ...... 197

Figure 5.8 Average Price of Electricity to Ultimate Customers for Residential Users and by State, 2013 to 2016 ...... 199

Figure 5.9 Average Price of Electricity to Ultimate Customers for Commercial Users and by State, 2013 to 2016 ...... 200

Figure 5.10 Average Price of Electricity to Ultimate Customers for Industrial Users and by State, 2013 to 2016 ...... 201

Figure 5.11 Average Price of Electricity to Ultimate Customers for Transportation Users and by State, 2013 to 2016 ...... 202

Figure 5.12 Average Price of Electricity to Ultimate Customers for All Sectors and by State, 2013 to 2016 ...... 203

xvi Figure 5.13 Percentage Change in Annual Average Electricity Prices, Price Benchmarks by State, 2010 to 2015 ...... 204

Figure 5.14 Import Ratios for Each State in 2015 ...... 205

Figure 5.15 Percentage Change in Import Ratios for Each State between 2001 and 2014 ...... 207

Figure 5.16 Energy Intensity by State, 2014 ...... 208

Figure 5.17 Energy Intensity by State ...... 209

Figure 5.18 Effective State-RPS Policy Stringency ...... 211

Figure 5.19 FITs Policy Stringency ...... 212

Figure 5.20 Net Metering Policy Stringency ...... 213

Figure 5.21 Interconnection Standards Policy Stringency ...... 214

Figure 5.22 Public Benefits Fund Policy Stringency ...... 215

Figure 5.23 LCV Score Control Variable ...... 216

Figure 5.24 Average Electricity Price Control Variable ...... 217

Figure 5.25 Electricity Import Ratio Control Variable ...... 218

Figure 5.26 Per-capita Energy-related CO2 Emissions, by State ...... 219

Figure 5.27 Per-capita Real GDP by State ...... 220

Figure 6.1 The Monthly Solar and Wind Generation Capacity in 10 States, 2001-2016 ...... 242

Figure 6.2 Monthly Accumulation Capacity of Solar and Wind Generation, by State, 2001-2016 ...... 243

Figure 6.3 The Annual Accumulation Capacity of Solar and Wind Generation in 10 States, 2001-2016 ...... 259

Figure 6.4 NGCC Capacity Factors by State ...... 264

Figure 6.5 Comparison of Capacity Generation of Natural Gas and Renewable Energy Generation, 2001-2016 ...... 265

xvii Figure 6.6 The DSPP Model ...... 273

Figure 6.7 The DSO Model ...... 274

Figure 7.1 Electric Utilities Service Territories ...... 301

Figure 7.2 Overview of NYS Electric Industry Participants ...... 305

Figure 7.3 Number of Utilities, by Ownership in 2015 ...... 306

Figure 7.4 Revenues, Sales and Customer Count of Major Utilities in NYS, 2015 ...... 309

Figure 7.5 Number of Customers by Ownership, 2015 ...... 310

Figure 7.6 New York Control Area Load Zones ...... 314

Figure 7.6 Status of Utility 2.0 Structural Change to Utility 3.0 ...... 316

Figure 7.7 The Rules and Principles of Utility 3.0 DER Model ...... 325

xviii LIST OF ACRONYMS AND ABBREVIATIONS

ACC Arizona Corporation Commission ARPA-E Advanced Research Projects Agency-Energy ASCE American Society of Civil Engineers BCA Benefit Cost Analysis BLFT Other Baseload Fossil Technologies BPU New Jersey Board of Public Utilities CAISO California Independent System Operator CAPEX Capital Expenditures CCGT Combined Cycle Gas Turbine CCS Carbon Capture and Sequestration CEC California Energy Commission CEEP Center for Energy and Environmental Policy CES New York Clean Energy Standard CHGEC Central Hudson Gas and Elec Corporation CHP Combined Heat and Power Systems

CO2 Carbon Dioxide Emissions COS Cost-of-service CPP Clean Power Plan CPUC California Public Utilities Commission CPX California Power Exchange DER Distributed Energy Resource DG Distributed Generation DOE United States Department of Energy

xix DR Demand Response DRAM California's Demand Response Auction Mechanism DSIPs Distribution System Integration Plans DSIRE Database of State Incentives for Renewables and Efficiency DSM Demand-side Management DSO Distributed System Operators DSP Distributed System Platform EE Energy Efficiency

EERSs Energy Efficiency Resource Standards EFFRPS Effective Renewable Portfolio Standards EIA United States Energy Information Administration ELECTRICPRICE Average electricity price ENERGSITY State’s Energy Intensity EPS OECD Environmental Policy Stringency (EPS) Database ERCOT Electric Reliability Council of Texas ESMs Earning Sharing Mechanism ESO Electric Service Operator FBNGT Flexible Baseload Natural Gas Technologies FERC Federal Energy Regulatory Commission FFRET Fast-flexing Renewable Energy Technologies FIT Feed-in Tariff FTRs Firm Transmission Rights FUA Power Plant and Industrial Fuel Use Act of 1978 GDP Gross Domestic Product

xx GHG Greenhouse Gas Emissions GMM Generalized Method of Moments ICTs Information and Communications Technologies IEA International Energy Agency IGCC Integrated Gasification Combined Cycle IMPORTRATIO Electricity Import Ratio INTERCONSTAND Interconnection Standards IOT Internet of Things

IOUs Investor Owned Utilities IPCC Intergovernmental Panel on Climate Change IPMVP International Performance Measurement and Verification Protocols IRP Integrated Resource Planning ISO Independent System Operator ISO-NE Independent System Operator New England ITCs Investment Tax Credits kW kilowatt LCOE Levelized Cost of Energy LCVSCORE League of Conservation Voters Scorecard LFCR Lost Fixed Cost Recovery LIPA Long Island Power Authority LMP Locational Marginal Pricing LOP Law of One Price LRAM Lost Revenue Adjustment Mechanism

xxi MBEs Market Based Earnings MISO Midcontinent Independent System Operator MW Megawatts MWh Megawatt hours NCUC North Carolina Utilities Commission NERC North American Electric Reliability Corporation NETMETER Net Metering NG Natural Gas

NG-RE Natural Gas Renewable Energy Hybrid Power Generation NGPA Natural Gas Policy Act of 1978 NMPC Niagara Mohawk Power Corporation NREL National Renewable Energy Laboratory NWA Non-Wires Alternatives NYISO New York Independent System Operator NYPA New York Power Authority NYPSC New York State Public Service Commission NYS New York State NYSEG New York State Electric and Gas Corporation NYSERDA New York State Energy Research and Development Authority O&M Operations and Maintenance OPEX Operational Expenditures PBR Performance Based Regulation

PERCAPITACO2 Per-capita Energy-related CO2 Emissions PERCAPITAGDP Per-capita Real GDP

xxii PG&E Pacific Gas and Electric Company PJM Pennsylvania-New Jersey-Maryland Interconnection PPL Pennsylvania Power and Light PSO Power Systems Optimizer PSRs Platform Service Revenues PUBENFUND Public Benefits Fund PUCN Public Utilities Commission of Nevada PUPA Public Regulatory Policies Act of 1978

PV Photovoltaic QER Quadrennial Energy Review R&D Research and Development RE Renewable Energy REV Reforming the Energy Vision RG&E Rochester Gas and Electric Corporation RGGI Regional Greenhouse Gas Initiative RPS Renewable Portfolio Standards RTO Regional Transmission Organization S-REITs Solar Real Estate Investment Trusts S&P GSCI S&P Goldman Sachs Commodity Index SCE Southern California Edison Company SDG&E San Diego Gas and Electric Company SEF Sustainable Energy Funds (Pennsylvania) SEIA Solar Energy Industries Association SEU Sustainable Energy Utility

xxiii Sha!"#$%acityFBNGT Capacity Share in FBNGT Sh$!"#$%acityFFRET Capacity Share in FFRET SRECs Solar Renewable Energy Credits SUNIDX MAC Global Solar Energy Stock Index Totex Total Expenditures TRECS Tradable Renewable Energy Credits TSO Transmission System Operator UPSC Utah Public Service Commission

WEO World Energy Outlook

xxiv ABSTRACT

Expansion of distributed energy resources (DERs) including solar photovoltaics, small- and medium-sized wind farms, gas-fired distributed generation, demand-side management, and energy storage poses significant complications to the design, operation, business model, and regulation of electricity systems. Using statistical regression analysis, this dissertation assesses if increased use of natural gas results in reduced renewable energy capacity, and if natural gas growth is correlated with increased or decreased non-fossil renewable fuels demand. System Generalized Method of Moments (System GMM) estimation of the dynamic relationship was performed on the indicators in the econometric model for the ten states with the fastest growth in solar generation capacity in the U.S. (e.g., California, North Carolina, Arizona, Nevada, New Jersey, Utah, Massachusetts, Georgia, Texas, and New York) to analyze the effect of natural gas on renewable energy diffusion and the ratio of fossil fuels increase for the period 2001-2016 to policy driven solar demand. The study identified ten major drivers of change in electricity systems, including growth in distributed energy generation systems such as intermittent renewable electricity and gas-fired distributed generation; flat to declining electricity demand growth; aging electricity infrastructure and investment gaps; proliferation of affordable information and communications technologies (e.g., advanced meters or interval meters), increasing innovations in data and system optimization; and greater customer engagement. In this ongoing electric power sector transformation, natural gas and fast- flexing renewable resources (mostly solar and wind energy) complement each other in several sectors of the economy.

xxv The dissertation concludes that natural gas has a positive impact on solar and wind energy development: a 1% rise in natural gas capacity produces 0.0304% increase in the share of renewable energy in the short-run (monthly) compared to the long-term effect estimated at 0.9696% (15-year period). Evidence from the main policy, environmental, and economic indicators for solar and wind-power development such as feed-in tariffs, state renewable portfolio standards, public benefits fund, net metering, interconnection standards, environmental quality, electricity import ratio, per-capita energy-related carbon dioxide emissions, average electricity price, per-capita real gross domestic product, and energy intensity are discussed and evaluated in detail in order to elucidate their effectiveness in supporting the utility industry transformation. The discussion is followed by a consideration of a plausible distributed utility framework that is tailored for major DERs development that has emerged in New York called Reforming the Energy Vision. This framework provides a conceptual base with which to imagine the utility of the future as well as a practical solution to study the potential of DERs in other states. The dissertation finds this grid and market modernization initiative has considerable influence and importance beyond New York in the development of a new market economy in which customer choice and distributed utilities are prominent.

xxvi Chapter 1

INTRODUCTION

If utilities are going to survive, they have to figure out how they are going to become much more capital efficient…This change requires migration of utilities away from thinking about their compensation being based upon deploying capital and instead thinking about compensation arrangements that are geared toward creating more capital efficiency.

Richard Kauffman, Chairman of Energy and Finance for New York in the administration of New York Governor Andrew Cuomo and Chairman of the New York State Energy Research and Development Authority (NYSERDA) Board, May 19, 2016.

1.1 Statement of Research

This research addresses three specific questions. First, the United States’ installed capacity of natural gas (NG) and renewable energy (RE) for power generation is steadily increasing. Some believe this growth could exacerbate power fluctuations and create significant supply-demand imbalances in energy markets as renewable electricity from solar and wind can be intermittent, variable, and uncertain, and natural gas prices tend to be volatile. This situation could further undermine grid resiliency, reliability, flexibility, and affordability of power generation system if these power sources continue to be used independently to generate electricity (Qolipour et al., 2016). This dissertation attempts to answer the question whether the evolving

1 energy sector in which RE and NG are prominent might cause these issues that some fear. A detailed study of the top ten states leading the United States in solar energy based on the 2016 rankings by the industry’s main trade group, the Solar Energy Industries Association (SEIA), was examined for its feasibility to address these concerns. Second, a shift towards distributed energy generation systems is already occurring (Acikkalp, Aras, and Hepbasli, 2014; Nyangon, Byrne, and Taminiau, 2017). Analysis of literature focusing on major technological changes and economic market development of NG and RE power generation technologies indicates emergence of multi energy hybrid power generation systems such as microgrids that incorporate a natural-gas-renewable-energy power generation paradigm. Considering this apparent shift towards a distributed generation (DG) future where consumers and emerging technologies support more optimized grid utilization, this study assesses if co-development of RE and NG markets has been complementary or conflicting in the ten states: Arizona, California, Georgia, Massachusetts, Nevada, New Jersey, New York, North Carolina, Utah, Texas. For this purpose, a dynamic econometric model that uses key policy, economic, and environmental indicators as well as control variables was developed to analyze the effects of NG on RE diffusion. Furthermore, utility business models that promote a blended NG-RE power generation system on both the supply-side and demand-side are offered to maximize economic and social benefits of joint development of these power sources (Biresselioglu, et al., 2016). Third, due to existing technical, economic, and policy barriers, there are shortcomings of the blended NG-RE-based DG electric power framework to achieve a maximization of the economic and social benefits. The dissertation investigates the

2 possibility of a blended DG system framework and assesses the effects of NG on RE diffusion. This requires investigation of the improved system configuration of the hybrid power systems, analysis of resource availability, formulation of a NG-RE power generation modeling framework including load forecasting, assessment of supply-demand imbalances, system optimization, and interpretation of the optimization result. Furthermore, the design of the envisaged NG-RE blended market requires adoption of an effective economic dispatch system and co-development of both NG and RE power generation mixes. To address these issues, the dissertation conducts a case study of a leading effort to find a blended market strategy, the New

York Public Service Commission’s grid and market modernization initiative called Reforming the Energy Vision (REV). REV seeks to reinvent the state’s energy regulatory structure to modernize the grid and introduce a new market economy with more customer choice and distributed energy resources.

1.2 Evolution of Electric Power Sector in the United States Electric power systems in the United States are experiencing a rapid and unprecedented transformation. A powerful confluence of structural, technological, and socio-economic factors is driving this change. Distributed technologies (e.g., distributed generation, energy storage, flexible demand, and advanced power electronics) are competing in the emerging electricity market and, as a result, are putting pressure on utilities and regulators to redesign a modern electricity system that reflects more capital-efficient options for the provision of electricity services are reflected (Funkhouser, Blackburn, Magee, and Rai, 2015; Sioshansi, 2016; U.S. Department of Energy, 2017b). The evolution of the sector is likely to destabilize the century-old government-regulated, vertically integrated, monopoly business model

3 that is the energy utility. For instance, in New York, California, Illinois, Massachusetts, and North Carolina, these forces are combining to transform the power systems at a rate faster than would have been possible two decades ago. In a 2016 scoping study of the future of power systems in the United States, McKinsey & Company write as follows:

For utilities, transformations can yield productivity improvements, revenue gains, better network reliability and safety, enhanced customer acquisition and retention, and entry into new business areas. The mounting pressure to transform also offers the rare opportunity to rebuild strategies, structures, and processes from the ground up. (Booth, Mohr, and Peters, 2016)

Similarly, in December of 2016, after more than two years of primary research, and quantitative modeling, and analysis of factors that are currently driving this change in power systems, the MIT Energy Initiative’s Utility of the Future study noted the following:

Distribution networks and system operations are now at the heart of modern electricity markets. Just as restructuring transmission network utilities was essential to create a level playing field for competition at the bulk power system level, regulators and policy makers today must carefully reconsider the roles and responsibilities of distribution utilities. The time has come to address incentives and create structures for the efficient provision of electricity services by a diverse range of conventional and distributed-energy resources and network assets. (Pérez-Arriaga and Knittel, 2016, p. 200)

This study defines distributed generation1 as small generating facilities which meet the electricity needs of a network user, supports economic operations of the

1 Distributed energy generation influences the trend of the distributed network by utilizing cost-effective efficiency and distributed energy resources to provide complex, integrated energy services and to reduce capital investment risk.

4 existing distribution network, are closer to the user, and are compatible with the environment (Alanne and Saari, 2006). Examples of distributed generation systems include solar and wind power systems, microturbines, micro-hydro generators, biomass systems, small-scale natural-gas-fired systems, etc. (Newcomb, Lacy, and Hansen, 2013). Complementarities of these distributed systems are growing spanning economic, technical, environmental, and political considerations (Lee et al., 2012). These opportunities will expand with the continued growth of renewable electricity resources, and with increasing recognition of domestic natural gas as a resource that is both less carbon-intensive resource than other fossil fuels and is domestically abundant and complementary to renewable energy resources (Pless et al., 2016).2

1.2.1 Signposts of Distributed Electricity Generation Expansion Integrating distributed energy resources (DERs) for power generation has grown dramatically in the last decade in the United States. From 2008 to 2016, the United States more than tripled its electricity production by wind power, and the amount of solar generation increased more than twenty-fold (U.S. Energy Information Administration, 2017d). According to the United States Energy Information

Administration (EIA), renewable energy resources—Including solar, wind, geothermal, hydropower, and biomass—generated 615 million megawatt-hours of

2 Opportunity for renewable energy and natural gas partnership exists in nine key areas, namely: development of hybrid technologies; R&D of energy-system integration; power sector market design; comparative analysis of alternative transportation pathways, (v) enhanced quantitative tools and models; public policy objectives; portfolio approach to research and development; joint myth-buster initiatives; and optimized cross-sectoral utilization of energy resources (Lee et al., 2012; Lee, Zinaman, and Logan, 2012).

5 electricity in 2015, representing an upsurge of 83% from 2010.3 Figure 1.1 shows the ongoing changes in primary energy production in the United States from 1949 to 2016 (U.S. Energy Information Administration, 2017f). Additionally, the United States plans to increase its share of renewables, beyond hydropower, in the power generation mix to 20% by 2030 (Ross and Damassa, 2015). A 2012 detailed study by SunShot Initiative4 shows that the United States could achieve 300 GW of solar photovoltaic (PV) installed capacity by 2030, and 630 GW by 2050 (Fthenakis, 2015).

3 Net generation for all sectors from all renewable energy sources, including all utility- scale, wind, all utility-scale solar, geothermal, biomass, wood and wood-derived fuels, other biomass, and other renewables solar. Total renewable energy generation (only solar and wind) in 2010, 2011, 2012, 2013, 2014, 2015, and 2016 were 95864, 121996, 145147, 176878, 210580, 229753, and 283092 thousand MWh, respectively. Source: (U.S. Energy Information Administration, 2017a).

4 SunShot Initiative is a national collaborative program administered by the U.S. Department of Energy (DOE) Solar Energy Technologies Office. It focuses on making solar energy cost-competitive with other forms of electricity by the end of the decade.

Source: U.S. Energy Information Administration (2017a)

Figure 1.1 Primary Energy Production by Source, 1949-2016

A confluence of trends is driving this transformation of electricity systems in the United States. These factors include the growing penetration of distributed generation—especially of cost-competitive, variable, renewable energy sources mostly solar PV and wind energy (and more recently, energy storage)—consumer preferences (Feldman et al., 2016), a more progressively aggressive state-wide demand-side management (DSM), energy-efficiency policy schemes (Barbose et al., 2013),

7 reversed energy-demand growth patterns, and flat to declining load growth (Nadel and Young, 2014). Other trends include proliferation of advanced information and communications technologies (ICTs), efficient technology options capable of unlocking device-level measurement and control at large scale (Rogers et al., 2015), and the increasing deep decarbonization of the grid system that is part of the global mitigation efforts under the 2015 Paris Agreement on climate change. Other issues driving change in the U.S. power system include cost-effective shale-gas developments that bring this low-cost gas into the market (Growitsch and Stronzik, 2014; Weiss et al., 2013), and the current infrastructural deficiencies. In 2017, the U.S. energy sector received an overall grade of D+ on infrastructure report card offered by the American Society of Civil Engineers (ASCE) (ASCE, 2017).5 Similarly, the Edison Foundation Institute estimates national level costs over 2010- 2030 at $582 billion in nominal terms (Chupka et al., 2008). At the same time, maintenance and system upgrade costs continue to expand as states issue new regulatory guidelines aimed at improving safety. For instance, the New York Department of Public Services estimates its required state-wide infrastructure replacement investments between 2014-2024 at approximately $30 billion (NYPSC, 2014). Finally, the increased interconnectedness of electricity with other critical transportation and communication infrastructure systems and the growing security threats from extreme weather and cyber and physical terrorism have also contributed to the transformation of power systems (Ward, 2013; and Campbell, 2015).

5 The 2017 Infrastructure Report Card estimates the cumulative investment electricity gap—including generation facilities and transmission and distribution infrastructure— between 2016-2015 to be $177 billion (ASCE, 2017).

1.2.2 Emerging Challenges of Distributed Generation on Bulk Power System Large influxes of DER capacity present several challenges to the design and operation of power systems and point toward a significantly different electricity market in the coming decade (Kind, 2013). As the energy sector transitions “from the traditional vertical structure of deterministic centralized production” into a more horizontal structure that is increasingly variable, service-oriented, “customer intimate,” and distributed in terms of productions and operations (Loock, 2012; Resnick Institute, 2012), regulators, utility managers, investors, policymakers, and other stakeholders are pondering over a pertinent question: How will these disruptions affect grid operations and the provision of electricity services in the future? Of the above factors, (i) intermittent renewable electricity or fast-flexing renewable electricity technologies (FFRET, i.e., wind and solar PV), and (ii) natural gas are the two most disruptive drivers of this unprecedented change in the power systems—although it looks like variable renewables are going to play a significant role in coming decades. This transformation is from a centralized-generation, top- down power system6 to a distributed generation system with renewables-natural gas partnerships constituting a cornerstone of the growing penetration of the market. Considering the highly disruptive nature of FFRET, low natural gas prices, and the other factors discussed above, this study evaluates economics of low-carbon futures in the U.S. electricity market through hybrid NG-RE solutions.

6 The centralized fossil fuel power generation paradigm is based on a business model that seeks to steadily increase electricity sales typically from an expanded asset base of large, centralized generation units and traditional delivery infrastructure system.

Transformation of the power sector presents new challenges for utilities and is profoundly affecting the industry’s existing business models (U.S. Department of Energy, 2015).7 This transformation has triggered changes in electricity production, transmission, distribution, and consumption (Schleicher-Tappeser, 2012; Small and Frantzis, 2010). Of paramount concern are two characteristics of FFRET: (i) intermittent production power generation, and (ii) very low operating costs (Aggarwal and Harvey, 2013). In a scoping study of distributed energy resources (DERs) conducted in 2012 (including distributed generation and storage, demand response, and electric vehicles), the Resnick Sustainability Institute at Caltech noted the following:

Historically the electric power system was operated as a load-following system. Loads were variable but predictable, generation was dispatchable, and there was no significant amount of bulk energy storage in the power system, so generation resources were operated through periodic dispatches that roughly aligned supply with demand and allowed automatic closed-loop controls to adjust generation to precisely match load (Resnick Institute, 2012).

Low-cost shale gas is driving a widespread switch from coal-to-gas leading to reduction in energy-related emissions throughout the United States (Wang and Krupnick, 2015; Nyangon, 2015b; Hefley and Wang, 2015). However, to meet the

2015 Paris Agreement on climate change targets, additional policy measures,

7 This transformation of the power sector is characterized a change from the old paradigm of centralized fossil fuel power generation (i.e., based on nuclear energy, coal, and other forms of fossil fuels) to DER generation (i.e., anchored on renewable energy and some form of natural-gas-fired generation to provide backup to the variable production from wind and solar power plants).

10 regulatory innovation, and business model innovation are required urgently.8 For instance, to address the variability of some DERs such as solar rooftops, the industry needs to position aspects of its power-delivery chain to maximize customer engagement in a tough managerial, regulatory, and financial environment. Otherwise the industry may be forced to undertake painful measures such as the premature shutdown of gas-fired power plants and unnecessarily high costs for customers to meet these targets. Booming, low-cost natural gas offers unique complementary benefits for mitigating the intermittency of renewables and the volatility concerns of each resource (e.g., fast-start capabilities for natural gas and low price volatility of renewables) (Pless et al., 2016; and Levi, 2013, p. 27-35). Rising natural gas production also offers a promising transition to a low-carbon economy (because of the potential for coal to natural gas substitution), and it militates against a lock-in of high-carbon energy sources through business-as-usual processes (Bruckner et al., 2014). Moreover, aging coal plants and low natural gas prices are making natural gas a more attractive generation fuel than coal. With increased expansion of unconventional gas reserves, the contribution of natural gas to reshaping the power sector is enormous and is steadily growing (See Chapter 3 for a business case for blending renewable energy and natural gas). The pathway to transformation is not progressing at the same pace

8 Besides concerns over carbon emissions, a shift away from coal-fired electricity generation has public health benefits. A study by the government of Ontario in Canada “estimated that shifting away from coal would reduce some 333,660 related illnesses and more than 700 deaths related to coal pollution to fewer than 6 deaths and only 2,460 illnesses. Put into monetary terms, the “coal switch” was estimated to save US$4.4 billion per year in health, environmental, and financial damages along with US$95 million in displaced operating and maintenance costs” (Sovacool, 2017).

11 nationally because it is highly sensitive to the regulatory policy, technical, political, and economic factors of each local situation. This study does not seek to predict a power system of the future. Rather, it identifies the factors that are driving change in the industry: i.e., the key barriers that might impede the dynamic evolution of the power sector, such as skewed incentives.

By using New York’s REV initiative as a case study, this dissertation offers a framework for market reform and regulatory innovation based on a comprehensive system of supportive business models and efficient electricity-market design. This framework blends renewable energy and natural gas to facilitate an effective outcome by continuously generating innovation in the power system regardless of how policies or technology objectives develop in the future. While the optimality paradigm that guided the development of the twentieth-century power system has generated significant inertia in the sector, a Utility 3.0, distributed-generation framework (an economically competitive and customer-centric platform) must facilitate an efficient and cost-effective coordination of economic activity without disruption. This study addresses the following hypotheses:

• Flexible natural-gas-fired generation systems facilitate diffusion of fast- flexing renewable energy generation systems such as solar and wind.

• The business model as a “unit of analysis” provides a sound method for conceptualizing the purpose of the business and its value proposition. Therefore, redesigning the utility business model towards one that focuses on developing DERs (i.e., one that promotes distributed NG-

RE hybrid power generation framework9) advances a customer choice service model, in part by integrating sustainability into the core modus operandi of the utility business.

A DER service model that blends NG and RE creates the opportunity to meet the growth-in-energy demand and tackle some of the modern dilemmas of the century- old, conventional utility—especially system performance. Any energy resource, either distributed or centralized, can provide electricity services. But combining NG and RE improves flexibility in siting, operations, and system performance, and it unlocks the contribution of resources that already exist in the power system (e.g., digital controls and power electronics, flexible demand, electric vehicles, and other innovations that may facilitate more ubiquitous distributed deployment of these resources or storage technologies).

In 2016, the International Energy Agency’s World Energy Outlook (WEO) forecasted broad transformations in the U.S. energy landscape over the next decade, with natural gas and renewable energy—mostly wind and solar—replacing coal which has dominated the market for the past 25 years (International Energy Agency, 2016). Nonetheless, FFRET sources, energy storage, and current demand represent only a small fraction of the resource mix. In 2015, coal, natural gas (dry), oil and natural gas liquids, nuclear, and renewable energy accounted for 20%, 32%, 28%, 9%, and 11%,

9 NG-RE hybrid framework focuses on promoting bulk energy generation systems, system integration, colocation, wholesale power market design, increased coordination, joint financing, portfolio approach to research and development, and optimized long-term and cross-sectoral utilization of natural gas and renewable energy resources.

13 respectively. Indeed, despite remarkable growth, solar and wind resources, only accounted for only 2.5% of the total primary energy production in the United States in 2015 (U.S. Energy Information Administration, 2017i). If the ongoing evolution in the power sector were to stop today—including electricity market design and operations, and business model innovation, and how it is regulated—the facts outlined above could be ignored as premature and marginal. Nonetheless, industry analysts, utilities, network agents and consumers believe that these facts indicate the industry is entering, or has entered, a new phase of substantial upheaval. This opinion is shared across the utility industry. Audrey Zibelman, the former chairperson of the New York Public Service Commission—the electricity regulatory body responsible for formulating far-reaching regulatory proceedings to make New York’s power system cleaner and more resilient, under the Reforming the Energy Vision initiative wrote as follows:

When informed by adequate information and pricing and enabled by platforms coordinated by utilities, DER markets can drive greater system efficiencies, facilitate the integration of variable renewable resources both in front of and behind the meter, and reduce overall energy bills… Placing the customers’ interests in total bill management, including reliance on DERs, at the center, rather than the fringes of the utility’s operating and business models, means that third- party and customer capital and market risk need to be added dimensions to how utilities meet their monopoly service functions. By allowing DER providers to contribute services and capital that result in greater value, innovation, and DER penetration onto the system, utilities’ capital requirements and associated returns from traditional cost-of-service regulation may be reduced, and utilities will necessarily incur additional expenses to accommodate these changes. (Zibelmen, 2016)

Similarly, Theodore F. Craver Jr., CEO of Edison International10 notes that:

It would be foolish to dismiss the potential for major changes in the utility-business model. (Amusa, 2013)

Zibelman and Craver are not alone in observing that business model innovation and DER investment should produce net benefits for both the participating customer and network agents. Furthermore, innovation in DER business model should support the transformation of the sector into one that is more consumer-centric, cost-efficient, and environmentally sound. Therefore, it is important—not only to utility regulators but also to the consumers, investors, and other power sector stakeholders—to understand the potential changes to the existing utility business models and the need for the market to prepare for the new DER-centric business models that may emerge. A survey of 515 electric utility executives in 2016 in the United States found that 97% believe that the utility business model needs to change, with most respondents affirming potential revenue opportunities in DERs. These opportunities include partnership opportunities with third-party providers and rate-based investments through a regulated utility. These executives were unsure, however, about how to design a DER-value-based business model to capture these opportunities (Utility Dive, 2016). According to the survey, the greatest obstacle to the evolution of the existing utility’s business model is the regulatory model and the attendant negative implications for the regulatory compact (that is, the agreements made between

10 Edison International is one of the largest public utility holding company in the United States that owns Southern California Edison (the second largest utility in the United States by revenue), and Edison Energy.

15 regulators and utilities regarding rate of return on investment, as illustrated in Figure 1.2.

Transforming the Nation’s Electricity System: The Second Installment of the Quadrennial Energy Review (QER), one of the defining strategy documents produced by the Obama administration, sums up the benefits of these trends to the industry, which include driving the restructuring of many major incumbent utilities, spurring the launch of many new ventures in DSM and customer engagement (e.g., Opower, Noesis Energy, SCIEnergy, BuildingIQ, C3 Energy, etc.), mergers and acquisitions, and initial public offerings:

States and utilities are exploring new distribution utility-business models while the private sector is providing new products and services to consumers. In the past decade, the electricity industry has seen a large increase in the number of businesses focused on providing electricity-related products and services outside of traditional utility- business models. These businesses have found opportunities to provide value to customers through innovative technologies, novel business models, and supportive state and Federal policy decisions—they are also changing the role of some ratepayers from passive consumers of electricity to informed shoppers and producers of electricity and related end-use services (U.S. Department of Energy, 2017b).

Source: Utility Dive (2016, p. 8.)

Figure 1.2 Survey of Obstacles to Evolution of Utility’s Business Model

Some envision that the pathways to new business models—especially that of realizing high penetration of renewable electricity systems—will continue even under existing policies and regulations. They envisage a future in which centralized resources will become relics of the past—the DER business model will replace the centralized-energy model (Lehr, 2013). In this evolution, NREL has identified six sets of actionable priorities needed to advance a distributed NG-RE hybrid power generation business model (Cochran et al., 2014):

• Create new metrics for clean energy investment that clearly define goals (including mechanisms for dealing with stranded power generation assets) for rate-payer protection, reliability, and investor risks.

• Adopt new pricing mechanisms that accurately reflect the value of electricity services (i.e., demand response, demand-side management, electricity supply, and delivery) rather than commodities to expand opportunities for renewable energy and natural gas synergies.

• Incentivize modernization of the generation fleet by replacing existing assets that have positive cash flows, encourage utilities to retire inefficient coal plants, and switch to natural-gas-fired generation plants.

• Explore new utility-business models beyond the current model premised on energy sales and regulated returns on investments to include emerging business opportunities that occur through distributed generation, demand response, on- site generation, and third-party-owned-and-operated energy systems (Kind 2013; Lehr 2013; Newcomb, Lacy, and Hansen, 2013).

• Expand public dialogue about the complementarity of renewable energy and natural gas to include different visions for customer-specific energy services and the new potential roles and responsibilities for consumers, businesses, regulators, and policymakers in the new-services business model.

• Explore a new range of hybrid electricity-generation business opportunities (e.g., micro-grids) to provide examples of how to deal with revenue erosion

and “death spiral” concerns in different conditions and scenarios—for example, through technological innovations, regulatory policy changes, institutional contexts, and new consumer demands.

Finally, while some industry analysts see these changes to the utility business model occurring in years to come, some of the largest utilities in the United States are

18 acting today. Exelon, a hybrid company, completed a $6.8 billion merger with Pepco in 2016, thereby creating Mid-Atlantic’s largest gas and electricity provider, while Duke is expanding into the higher-growth, natural gas and pipeline sectors with a deal for Piedmont (FERC, 2014; Polson, 2015). Coupled with a shrinking number of grid- connected customers, there is a growing negative concern among many utilities in the United States about the impacts of emerging distributed technologies on their business models. For example, New York state is pioneering a new regulatory model for its electric utility companies that is expected to herald the distributed energy system of tomorrow through the REV initiative. These changes aim to align the role of regulated utilities in the ownership, management, and operation of electric delivery systems via increased adoption of “disruptive technologies” (e.g., shareholder-incentive and lost- revenue mechanisms) and utility financial interests (Satchwell and Cappers, 2015). In Maryland, the public service commission initiated its grid modernization with the specific objective to transform electric distribution into seven major areas. These areas include rate design options for electric vehicles, improved valuing of costs and benefits of DERs, maximizing advanced metering benefits, energy-storage valuation, restructuring the interconnection processes of DERs, better distribution system planning, and considerations for retail choice issues that include protecting that interests of low-income customers (Maryland Public Service Commission, 2016). These efforts in New York and Maryland fit alongside ongoing proceedings in California, Minnesota, Hawaii, and elsewhere, where regulators are rethinking the use of DERs to defer costly investments, to bring more and varied energy technologies and providers into the market, and to experiment with new business models. For

19 example, California’s demand-response auction mechanism (DRAM)11 allows for a broad range of demand responses, electric vehicles, and customer battery-storage methods to compete in the wholesale markets (FERC, 2016). Other states, notably Hawaii, have shown how to add flexible natural gas generation along with energy storage to integrate and bring intermittent resources online quickly (Trabish, 2016a).

In New York, given the state’s high retail peak demand—estimated to be approximately 75% greater than the average system load and considering that 9% of the power is lost in transmission—these reforms have the potential to unlock investments in the entire energy market, thereby enabling network users and agents to accommodate variable distributed resources and customer-sited resources (e.g., as distributed system platform providers) and push for cleaner and more affordable electricity. With wind and solar-energy resources reaching grid parity across much of the nation and reshaping centralized fossil fuel-based grids, the number of states launching the broad modernization efforts of the utility business model will increase as more utility regulators tackle the evolution of the power sector.

1.3 The Case for a More Distributed Energy Future This study examines the development of the electricity sector over the past two decades under complex and different state regulatory regimes, taking the two hypotheses—the United States’ increased expansion of recoverable natural gas

11 California Public Utilities Commission established DRAM mechanism for third party providers to support demand response outside of utility programs part of an effort to bifurcate utility demand response programs into demand- and supply-side resources. Thereafter, the mechanism seeks to integrate demand response resources into the California Independent System Operator’s (CAISO) markets by 2018.

20 reserves and its growing contribution to the decarbonization of the energy system, and the growth of renewables factors—into consideration. The following four central questions arise given the industry consensus that utility-scale renewables, DERs that combine natural gas and renewables, demand-side solutions, and integrated transmission networks are all important to achieving flexible energy demand, sustainable energy options, and an optimal fuel mix in the American utility sector: How can a more compelling NG-RE power generation paradigm (that integrates infrastructure network, revenue models, customer interface, business models, organizational logic and mandate, risk management, and value proposition in improving communication with consumers and efficient operational boundary regime) be created and positioned to meet the challenges being experienced in the transformation of power systems towards a ‘sustainable’ and ‘smarter’ future? What regulatory barriers hinder the emergence of new business models to promote greater efficiencies in electricity generation so that the advantage of a decentralized, clean, and distributed grid system is realized? Are there leapfrog opportunities, and what are the emerging alternative business models for achieving these goals?

1.4 Research Design and Methodology The main questions related to developing the NG-RE hybrid distributed energy generation roadmap and the transition to a cleaner power sector are as follows: 1. To facilitate NG-RE hybrid power generation system, what key business models are best suited for the needs of the rapidly evolving electricity sector?

o What are the core components and characteristics of business model in the context of the electric power systems?

o Using proposals for the New York’s REV program as a case study, what are the characteristics of policy and business models that can facilitate distributed utilities that have high penetration of RE and NG?

o What are the implications of high RE and NG distributed generation systems to the electric grid of the future—a future in which demand- side resources and energy “network users and agents” play an active role in the investments and operations of the power sector?

o What is the taxonomy of customer-centric business models that will inform electricity systems of tomorrow? 2. Will power systems across the nation transform incrementally, or will evolutionary leaps occur, and what key pillars (i.e., rules and principles) of the electricity systems are required to facilitate an efficient transformation

process—both in terms of regulatory innovation and customer-sided value network—over the coming decade and beyond?

1.5 Organization of Chapters This study develops as follows. Chapter 2 evaluates the utility sector in the United States. It surveys the dominant centralized fossil-fuel energy policy paradigm. Then, it explores drivers of change for the power systems of the future. In addition, the chapter explores a broad spectrum of business models for facilitating distributed utilities that exist today—in particular, smart integrator and energy-services utilities including their economic benefits (i.e., behind-the-meter net generation resources) and support for electric grid reliability objectives. Chapter 3 discusses the background of the utility sector and surveys drivers of change for the utility of the future in the United States. The section discusses the core

22 dimensions of the Hamel business model framework, including the customer interface, core strategy, strategic resources, and value network. Chapter 4 describes the conceptual approaches of blending that this study uses to assess the effects of flexible natural gas technologies on fast-flexing RE power generation. The basic rationale underpinning this objective is to achieve a seamless transition from a primarily central station-based grid to a bi-directional grid system that integrates both dispatchable and non-dispatchable loads (solar and wind, and natural gas assets), storage, ancillary services such as electric vehicles, and capacity markets. Based on the evaluation of business model concepts presented in the chapter, a trifecta of taxonomy-generation technologies that integrates NG and RE assets is developed and analyzed: fast-flexing RE generation, flexible-baseload NG generation, and other baseload fossil generation technologies. Chapter 5 identifies and analyzes key policy and economic instruments that have been used to support RE investments and diffusion. It also describes the data requirements, variables, and model constraints, and it presents the dataset used in the research. Chapter 6 offers empirical analysis of this study. It details the methodological approach this study uses to estimate the effects of flexible NG power generation on

RE electricity diffusion based on data from the top ten, SEIA-ranked, solar-producing states using a dynamic, system generalized method of moments (system-GMM) econometric model estimator. Based on the evaluation of drivers of the transformation of the electric power sector presented in Chapter 3, a dynamic econometric model is used to assess policy and economic instruments. Chapter 6 also discusses assumptions and constraints applied in the development of the system-GMM model. The results of

23 the ten-states dynamic-panel model for the study period 2001 to 2016 are presented based on an in-depth analysis of the interrelations between different energy generation technologies, policies, and other control variables such as the real electricity import ratio, GDP per capita, energy intensity and per capita energy related carbon dioxide emissions. The quantitative analysis performed in Chapter 6 is relevant to the qualitative analysis undertaken in chapters 4, 5 and 7.

By using New York’s grid and market modernization initiative, that is, the REV case study, Chapter 7 assesses the value of distributed energy generation framework that supports an efficient and evolving power system. It explains the core dimensions of the NG-RE hybrid business model and innovations in energy governance (or Utility 3.0). It also describes policy variables and constraints, including regulatory policy and market demand, and it presents the different assumptions and dimensions used in the development of the model. This chapter discusses in detail the organization and structure of the “utility of the future” (through the New York’s REV proceedings and its vision), the challenges and opportunities these proposals pose for the utility industry, and ultimately the impact these reform efforts will have in other states. This framework promotes the emergence of a customer-centric view of electricity service and metrics based on a steady increase in

DER resources in New York. It presents an analysis of what the grid will have to offer to customers empowered by these DERs to attract more customers. Finally, Chapter concludes the study by presenting the findings, recommendations, limitation, and suggestions for further research to address the changing demands of a variety of utility customers efficiently while providing an optimal network for all.

Chapter 2

TRANSFORMING THE U.S. ELECTRICITY SYSTEM: BUSINESS MODEL INNOVATION

This chapter begins by discussing the dominant, centralized, fossil-fuel energy- policy paradigm. Then it surveys drivers of change for the utility of the future. It also presents the current utility-business models for the provision and consumption of electricity services for the following reasons: to contextualize how DERs affect the design and operation of power systems in the United States, and to create prescriptions for solutions to the ongoing evolution of the electricity sector that are governed by market principles and distributed-utility objectives. This chapter takes the key elements of the electric power industry discussed in the previous chapter and places them in a more theoretical framework: that of disruptive business model innovation.

Adding this element emphasizes the utility’s current managerial dilemma on disruptive innovation and the impact of a capabilities mismatch between the new business models and the incumbent, asymmetric incentive systems in the utility industry.

2.1 Background

2.1.1 Overview of Utility Sector in the United States

The bulk of electricity generation in the United States was until recently almost entirely within the sphere of utilities. This situation changed dramatically with the drastic price reduction of solar PV and natural gas (Nyangon, Byrne, and Taminiau, 2017). The desired attributes of an efficient power system include low carbon intensity, ready availability, easy accessibility, cost-effectiveness, reliability, resilience, and the embedding of smart, real-time digital control and services.

However, because states have different policies, regulatory environments, and business models, it is unlikely that a true transformation will occur everywhere in the nation. For example, the grid in New York and California and expanded opportunities for engagement between network users and agents could drive innovation and animate market participation and greater investment in the sector. This scenario could bring disparate interests into the partnership to develop common industry interests on the demand side of the grid, ultimately affecting the provision and consumption of electricity services and the best interests of the customer (Byrne, Taminiau, Seo, Lee, and Kim, 2016; Fthenakis, 2015). Additionally, because solar PV allows network users of any size to produce their power, thereby facilitating a “new bottom-up control logic” in the design and operation of power systems, PVs promise to be the most disruptive energy technology in the industry (Schleicher-Tappeser, 2012). This characteristic of RE is what Chariton and Markides (2002) define as “a way of playing the game that is both different from and in conflict with the traditional way” (p. 56). Many studies find that the growing penetration of DERs threatens the current electric-utility-business models (Frantzis,

Graham, Katofsky, and Sawyer, 2008; Nimmons and Taylor, 2008; Schoettl and Lehmann-Ortega, 2011). Policymakers, network and regulating agencies, and industry analysts have warned that unless the industry adopts new business models, the ongoing power-sector transition will lead to a massive loss of market share, revenues, and profits. For instance, Nimmons and Taylor (2008) note that, “although the utility to date has been generally reactive to state requirements (e.g., net metering, RPS, and standardized interconnection), the industry is expected to become proactive in the distributed PV market as it is pushed to key stakeholder status.”

Suffice it to say that once solar PV reaches significant market penetration, the concerns for grid modernization, safety, and potential “death spiral” or revenue erosion will drive the utilities involved in the provision of electricity services to create value for the network users and agents (Trabish, 2016b). For the large-scale adoption of new distributed technologies, a business model is an important analytic tool as the technology itself (Teece, 2010). In this regard, the ongoing changes in the U.S. electricity system in the create a fundamental business model conundrum for utilities. As explained in the previous section, deployment of DERs is occurring amidst increased natural gas production; changing generation mix; increasing consumer choice; low load growth; the proliferation of new technologies, services, and market entrants; and declining costs for renewable energy.

The National Renewable Laboratory’s Renewable Electricity Future Study estimates that it is feasible to produce 80% of America’s power from renewables in 2050 (Mendelsohn, Lowder, and Canavan, 2012). The MIT Energy Initiative’s Utility of the Future study presents a framework for proactive regulatory, policy, and market reforms that are designed to enable the dynamic evolution of the structure of the electricity industry as a “market platform, network provider, system operator, and a data hub” (Pérez-Arriaga and Knittel, 2016). These factors, coupled with shifting consumer and business model demands for delivering DERs, are essential to climate- change mitigation and have the potential to provide immediate, significant reductions in carbon-dioxide emissions in the U.S. power sector (Newcomb, Lacy, and Hansen, 2013). Amidst this dramatically changing energy landscape, President Obama unveiled the “Climate Action Plan” in June of 2013, which directed his administration

27 to initiate an interagency QER study to help match federal energy policy with the nation’s economic, security, and climate goals (U.S. Department of Energy, 2015). In 2015, the Obama Administration also announced the beginning of the implementation of the Partnerships for Opportunity and Workforce and Economic Revitalization (POWER+) initiative to help communities impacted by the ongoing transition away from the coal economy as more utilities shift to RE and natural gas (Nyangon, 2015a). The administration provided a $4 billion incentive fund in the 2016 budget to encourage states to make faster and deeper cuts in carbon emissions from electricity generation required under the Clean Power Plan (CPP). It also provided a $7.4 billion investment in sustainable electricity-generation technologies (U.S. Department of Energy, 2017b). Additionally, because of regulatory, policy, and market reforms intended to promote energy efficiency, U.S. electricity consumption flattened between 2005 and 2014, and total energy consumption declined by nearly 2%.12 Hitherto, the Unites States is embarking on a massive restructuring of its electricity supply within the next 30-40 years, with increased investment in distributed generation included in the resource mix of the future. Undoubtedly, these changes require a multidimensional framework that integrates all resources, be they distributed or centralized, to contribute to the efficient provision of electricity services and other public and environmental objectives.

12 Electric power sector total primary energy consumption is the sum of the power sector consumption values for renewable energy, fossil fuels, and nuclear electric power plus electricity net imports. The data used in this analysis were obtained from the Energy Information Administration’s (EIA) monthly energy review (U.S. Energy Information Administration, 2016a, 2016b) and U.S. Census Bureau’s Population Estimates from www.census.gov/popest/.

The United States does not have a single legal framework for the deployment of DERs (and, more specifically of renewable energy). A variety of federal and state agencies oversees energy-related policies, and renewable electricity generation is managed through state-level RPS13 (Barbose, 2014). With current demand growth close to zero, a variety of electric utilities in the United States are implementing strategic options, such as building distribution and transmission networks, mergers and acquisitions, and increasing portfolio investment in utility-scale solar to boost future earnings. Their strong balance sheets, low interest rates, and an appetite for renewable tax credits are enabling potential acquisitions of smaller renewable-generation companies to expand capacity that could boost earnings. For instance, YieldCos14 was supposed to help, but it has lost momentum as a financing vehicle. In this changing energy market, smaller utilities are selling to deep-pocketed buyers, especially from overseas, while large-cap utility companies continue to acquire regulated assets and divest merchant-generation units. The result is a pure-play regulated entity like the merchant divestitures created by Pennsylvania Power and Light (PPL) and Duke Energy (Polson, 2015).

13 A total of thirty-eight states have RPS or some kind of goal for renewables and almost all these states have met their targets for 2013 (Barbose, 2014). An RPS is a percentage mix of renewable energy (i.e., solar PV, wind, biomass, geothermal, tidal, and hydropower) that every electricity company in the state must include in their total fuel batch.

14 YieldCos include public solar ownership funds such as publicly quoted equities. New solar financing structures aim to make solar project investments more liquid by allowing developers to tap the capital markets (Srinivasan, Reddy, and Reddy, 2016). These options include project bonds, solar real estate investment trusts (S-REITs), and others.

By the end of 2016, there were nearly 4,000 utility entities in the United States. Only a few large players dominated the industry, however, operating under different regulatory changes and corporate structures and serving different customers. Investor- owned utilities controlled 72% of overall U.S. electricity sales with this percentage skewed toward the largest players (Satchwell and Cappers, 2015). The ten major integrated utilities—including Duke Energy, Southern Company, Entergy, NextEra, Dominion, and Xcel—controlled 38% of the integrated portion of the market (U.S. Energy Information Administration, 2017a, 2017b, 2017c). With this fact in the market and the prevailing operating economics increasingly becoming a powerful driver for the future growth of distributed generation and distributed technologies in general, the growth of variable renewable energy sources has the potential to create a significant threat to utilities’ business models in the future. Figure 2.1 depicts the interacting forces of business model innovation, policy, regulation, global market forces, consumer behavior, and technological innovation triggering the change in the utility industry in the United States.

Figure 2.1 Evolutionary Landscape of Disruptive Triggers in the Utility Industry

This study proceeds by first elucidating the six converging triggers that are accelerating the rate of deployment of DERs today. It then offers a framework for positioning the market towards proactive regulatory reform, which includes improvements in incentivizing distribution utilities, pricing of electricity services, and recommendations for electricity market and power sector design. This framework is intended to respond effectively to a variety of uncertain changes now underway in the

31 sector and can facilitate the emergence of an efficient portfolio of distributed utilities in various parts of the country to meet the needs of a fast-evolving electricity sector. The benefits of NG-RE blended distributed-generation model are then analyzed and the implications of providing electricity services in the distribution system from these resources explored in the study’s conclusion.

2.1.2 Disruptive Triggers in the Utility Industry As observed in the historic shift in telecom services from the infrastructure delivery of commodity dial-tone voice telephony of the last century to a broad range of personalized services, disruption has become a prevailing and uncompromising theme of technological innovation in the 21st century. Emerging revolutionary platforms for growth—such as the Internet of Things (IoT)15 and transactive energy— have sprung up to drive disruptive changes in recent decades, thus shifting value across the value chain and creating revenue implications for incumbent players and new market entrants (Kotter, 2014, p. 48, 99). Finally, industry stakeholders, particularly utilities, are provided with a blueprint—the energy cloud playbook— for proactively preparing and managing their organizations to maneuver around the energy cloud disruption and to position for long-term success. Disruption of power systems affects both the demand and supply sides of the business. In his book, The Fourth Industrial Revolution, Klaus Schwab wrote as follows:

15 Fereidoon Sioshansi defines IoT as an “expansive concept of billions of intelligent devices, networked to trade data in the background, and to execute commands intelligently, providing unimagined new levels of personalization and value, the currencies of our 21st century tech- driven economy and society” (Sioshansi, 2016).

Multiple sources of disruption trigger different forms of business impact. On the supply side, many industries are seeing the introduction of new technologies that create entirely new ways of serving existing needs and significantly disrupt existing value chains. Examples abound. New storage and grid technologies in energy will accelerate the shift towards more decentralized sources. The widespread adoption of 3D printing will make distributed manufacturing and spare-part maintenance easier and cheaper. Real-time information and intelligence will provide unique insights on customer and asset performance that will amplify other technological trends. Disruption also flows from agile, innovative competitors who, by accessing global platforms for research, development, marketing, sales, and distribution, can overtake incumbents faster than ever by improving the quality, speed or price at which they deliver value. Major shifts on the demand side are also disrupting business: Increasing transparency, customer engagement and new patterns of consumer behavior (increasingly built upon access to mobile networks and data) force companies to adapt the way they design, market and deliver existing and new products and services. (Schwab, 2016, p. 51-52)

While seemingly more insulated from competitor-led innovation, the highly- regulated utility industry is no exception to the rule of disruptive change. It therefore accelerates the need to leverage new capabilities to challenge old ways of doing things in different value chains. But this disruptive change is materially different in the age of platforms, technological innovation, and business model innovation. In today’s world, incumbent utilities have no time to plan and execute their plans, and they must start retooling differently. This fact is especially true when the confluence of six forces— policy shifts, technological innovation, regulation, changing global-market forces, business model innovation, and changing consumer behavior—undermines long- standing value chains. Table 2.1 details examples of the disruptive dynamics. Utility consumers care about the question of how they can enjoy the benefits from innovation created by dominant utility-platform firms without letting this dominance get out of control (Gawer, 2009, p. 3). And then, there are the strategic questions which investors

33 and managers care about: Which capabilities are needed and when should utilities open up their technologies and processes, and when should they focus on proprietary strategies? As a result, it is not a question of whether the existing utility-business models need to change; it is rather a question of what new forms they will take, in what contexts, and how rapidly the utility industry will have to alter course. The pace of change will be different in each state or regional market and each specific situation. For example, technologies that can slice through preexisting layers of regulatory processes and business models to directly connect customers to the electricity services they seek could grow faster, gaining traction across global markets. In what follows, this study describes the six forces through which disruption is having an impact and where network users and agents should assess their strategies (Figure 2.2). The utility industry currently faces a perfect storm of all the six areas (PwC, 2014).

Source: PwC (2014, p. 3)

Figure 2.2 The Impact of Megatrends and Disruption Dynamics on the Power Sector

Table 2.1 Power System Pathway Origins Disruptive Example of the Disruption Dynamics Triggers Policy • Carbon mitigation: Cap and trade, Clean Power Plan, EU Emissions Trading Scheme, COP21. • Flexibility: Promotion of distribution system operators, support for energy storage, support for international interconnection. • RE promotion: Renewable Portfolio Standards (RPS), Renewable Energy Directive.

• DER adoption: Net metering, feed-in tariffs, Solar Renewable Energy Credits (SRECs). Regulation • Changing utility regulatory models: Incentive-based regulation (e.g., United Kingdom’s Revenue = Incentives + Innovation + Outputs (RIIO), New York’s REV). Global • Sustainability: Marketplace differentiation and brand Market awareness. Forces • Accessibility: More options available to greater share of end-use customers. Business • Affordability: Declining cost of ownership for solar PV, Model energy storage, and other demand-side technologies. Innovation • Integration: Pairing of complementary disruptive technologies (e.g., solar + storage). Technological • Digitalization: Lowering the barrier for entry for innovative Innovation solutions. • Networking and data analytics: Harnessing distributed computing and data across the grid. Consumer • Control: More customers demanding control over their Behavior electricity usage and spending, and when and what type of power they buy. • Choice: More customers want the ability to purchase electricity from clean energy sources or self-generate and sell that power back to the grid.

2.1.3 Distributed Generation Narratology: The Détente for Natural Gas and Renewable Energy Blending Plenty of RE sources such as solar and wind can be converted into electricity using technologies that emit no greenhouse gases. Natural gas offers a cleaner alternative to coal or petroleum with the potential to deliver immediate, significant reductions in carbon dioxide emissions from the power sector—if newly discovered

36 economically recoverable gas reserves in the United States can be produced responsibly. The appeal of synergies of renewable energy and natural gas goes beyond carbon mitigation in the power sector (Nyangon, 2017). The development of renewable energy and natural gas systems promote energy security and energy independence, contribute to economic growth, and lead to a diversified resource base, sometimes independently and sometimes collectively. To date, however, the challenge of investing in renewable energy sources is that they are dispersed, intermittent, and non-dispatchable. A significant switch to a renewable energy system also requires substantial initial capital investment in infrastructure and technical innovations. In most of the world today, and certainly in the United States, the main barrier to large-scale energy transition remains the too-narrowly constrained energy policies and laws around problems with electrons, fuel, carbon, technologies, and the cost of technologies (Miller et al., 2015). Energy systems remain deeply entangled in broad patterns of economic and geopolitical arrangements, and they are often viewed largely as a techno-economic problem. Taminiau et al., (2014) and Miller et al., (2015) explain that large-scale, sustainable energy transition should not just be viewed as a techno-economic change but also a socio-energy systems design development. In this way, utility policy systems, including business models, strategies, and technical solutions, are in practice largely a problem of socio-energy system design and the attendant socio-technical innovations inherent therein. Furthermore, the need forward- looking energy policies, environmental regulations, or market designs continue to dominate markets and investment decisions at various levels and will impact electric grid reliability and electricity costs.

Miller et al., (2015) describe the disaffection and détente for hybrid business models in Arizona owing to too-narrowly constrained business models. They accordingly propose a reconceptualization of energy-policy imagination—of a kind that integrates social considerations more effectively into energy analyses and decision-making:

Solar energy has become the subject of deep conflicts among Arizona’s political institutions, elites, and publics. These conflicts have revealed that the choice confronting the state is not simply whether or not to “go solar”—the state’s utilities will easily achieve their 15% targets under the Commission’s renewable portfolio targets—but which model of solar-based socio-energy system to choose going forward. The diverse policies that have promoted solar energy in Arizona have catalyzed the emergence of at least six distinct models of socio-energy systems design within the solar industry, each of which has strong advocates, a distinct vision of the future, and a record of accomplishment of successful implementation. [What the Arizona solar energy developments demonstrate is that] energy policy choices have the potential to create vastly different societies. Advocates of distributed energy have long made this argument, but the challenge is much deeper. Major differences exist across these diverse models of socio- energy systems, including: the cost of produced energy; the geographies of energy construction; the sources of capital that invest in them; the financial beneficiaries of projects; those who pay for projects and bear risks; the future viability of utility business models; and the collective, emergent patterns of energy behavior among publics and energy users. (p. 32-33).

If the threat to utility-business models from distributed energy is as significant as some in the utility industry now fear, investors are likely to suffer significant investment losses. In the context of drivers of this evolution—in particular, the shale gas and oil boom, the deployment of renewables to address climate change, and the development of alternative fuel sources (including hybrid energy systems and electric vehicles)—a reframing of various dimensions of energy systems is particularly salient

38 for these energy-policy choices. This requires a reconceptualization of energy-policy design and governance to include structural reforms that establish financial independence between distribution-system operation, planning functions, and competitive market activities. For example, integrating NG-RE hybrid power generation into processes and practices for electricity market design—namely by “socializing energy policy, systematizing energy policy, publicizing energy policy,” and adequately reforming regulations governing the power networks to be as innovative as the businesses to which they are applied (Miller et al., 2015). It also requires formulating a blueprint of a fully integrated DER business model that capitalizes on both renewable energy and natural gas to access new revenue streams, wholesale power market opportunities, and energy services that offer customers resiliency, reliability, and reduced costs (Lee et al., 2012). Such a blueprint would enable utilities to assess their progress toward DER integration while working with customers, third parties, market operators, and regulators toward the full integration of DER processes across operations, energy markets, and existing strategic-planning mechanisms currently in use, like integrated resource planning (IRP). Furthermore, to manage this transformation process and objective, Lee et al., (2012) identified nine opportunities for blending natural gas and renewable energy markets:

• Development of hybrid technologies • Research and studies of energy system integration • Power sector market design • Comparative analysis of alternative transportation pathways • Enhanced quantitative tools and models

• Public policy goals • Portfolio approach to research and development • Joint myth-buster initiatives • Optimized long-term and cross-sectoral utilization of energy resources

Figure 2.3 shows key processes in a DER system which integrates the opportunities listed above by blending renewable energy and natural gas assets.

Source: Navigant (2016, p. 8)

Figure 2.3 Blueprint of a Fully Integrated DER System

2.1.4 Benefits of Gas-Renewable Energy Partnering In recent years, lower natural gas prices have raised concern about its impact on the increased penetration of renewable energy (Hefley and Wang, 2015; Cochran, Zinaman, Logan, and Arent, 2014; Lee, Zinaman, and Logan, 2012; Lee et al., 2012; Harvey, 2012). Claims of the golden age of natural gas threatening the future development of renewable energy have increased as abundant and cheaper natural gas dominates the energy discourse. However, this concern is unwarranted. Using weekly Henry Hub-linked natural gas spot prices and utility PV system prices from 2010 to 42

2015, Nyangon, Byrne, and Taminiau (2017) modelled these two markets for convergence to examine the extent of spot-market integration and the speed with which market forces move the two energy prices toward equilibrium. They concluded that complementary use of the technologies is likely. While price convergence is not likely to occur in these two markets soon, distinctive, complementary benefits of each resource (e.g., fast-start capabilities for natural gas and low-price volatility for PV) offer opportunities that expand market demand for both. The idea that natural gas and renewable energy resources complement each

other in different sectors of the economy—including power markets, residential and commercial, industrial and transportation sectors—is also evident from specific government analyses and energy-market forecasts (Weiss, Bishop, Fox-Penner, and Shavel, 2013). Figure 2.4 shows interdependencies of the renewable energy and natural gas markets for all the major end-use sectors, providing a useful impetus for a deeper understanding of the nexus between these vital resources.

Source: U.S. Energy Information Administration (2016a), Monthly Energy Review (April 2016), Tables 1.3, 2.1-2.6: Total = 97.7 quadrillion British thermal units (Btu)

Notes: 1 Does not include biofuels that have been blended with petroleum—biofuels are included in “renewable energy”. 2 Excludes supplemental gaseous fuels. 3 Includes less than -0.02 quadrillion Btu of coal coke net imports. 4 Conventional hydroelectric power, geothermal, solar/PV, wind, and biomass. 5 Includes industrial combined heat and power (CHP) and industrial electricity-only plants. 6 Includes commercial CHP and commercial electricity-only plants. 7 Electricity-only and CHP plants whose primary business is to sell electricity, or electricity and heat, to the public. Includes 0.2 quadrillion Btu of electricity net imports not shown under “source.” Sum of components may not equal to total due to independent rounding.

Figure 2.4 U.S. Primary Energy Consumption by Source and Sector, 2015

As a fast-ramping resource that is relatively easy to turn on and off, natural- gas-fired power plants (especially combustion turbines) are well-suited for meeting peak electricity load and backing up and smoothing out intermittent renewables. Because peaking units tend to run infrequently for short periods, they provide the capacity needed to help electric systems meet peak demand. Figure 2.5 shows average capacity factors of natural gas combustion turbine plants by the North American Electric Reliability Corporation (NERC) for 2012. This is also known as plant's capacity factor: i.e., the average hourly unit output of all combustion turbines divided

44 by the maximum possible output from these units. Table 2.2 details issues in natural

gas and renewables markets and related complementary benefits.

Source: U.S. Energy Information Administration (2013)

Figure 2.5 Average Capacity Factors of Natural Gas Combustion Turbines by NERC

Table 2.2 Optimized Diverse Electricity Portfolio of Natural Gas and Renewables Issue Resource Considered Complementary Gas pipeline constraints Natural gas Renewables (e.g., wind and solar) do which can affect fuel not experience market-based fuel supply and supply concerns as natural gas transportation generation. Resource variability and Renewable energy Renewable energy sources can only dispatchability be dispatched within the limits of resource availability. Natural gas can be dispatched flexibly, offering more capacity for system reliability.

45 Fuel price volatility and Natural gas Renewables have zero fuel costs and

generating costs relatively fixed generating costs. Cost competitiveness Renewable energy Natural gas plants have low upfront costs and overall levelized cost of energy (LCOE) which can complement the limited cost- competitiveness of renewables in the near-term for greater levels of deployment. Utilities with Natural gas Modularity of some renewable predominantly natural energy technologies offers valuable gas capacity may face flexibility in deployment and can also compliance costs if RPS hedge risks from future policy targets are not met uncertainty. Transmission planning Renewable energy Significant overlaps between high and costs wind energy potential and natural gas production regions offer opportunity for joint development thus accelerating investment in new transmission network.

Environmental Natural gas Renewables complement regulations and concerns environmental benefits of highly (MATS, CSAPR, etc.) efficient natural gas generation plants (i.e., when emissions reduction benefit of natural gas eventually plateaus as long-term emission thresholds become increasingly stringent). Uncertain state and Renewable energy Long-tern competiveness of natural federal incentives gas plants does not require major new local or federal policy or incentives.

46 Sources: Nyangon, Byrne, and Taminiau (2017); Lee (2012)

Stakeholders in the utility industry continue to debate how two such conflicting views can co-exist and what would be the net impact of high levels of distributed energy generation systems that blend both natural gas and renewable energy in power generation (Papaefthymiou and Dragoon, 2016). In terms of the interaction between

natural gas and renewable energy prices, Fletcher (2009) offered a “Goldilocks” theory in which market prices are “neither too hot nor too old” to explain how to stimulate the market development of two different energy sectors. Goldilocks theory holds when the right natural gas price leads to the highest level of renewables output: i.e., when there is an effective partnership between, for example, solar energy and natural gas. Ideally, this range should establish the growth of both markets by fostering investment

incentives and macroeconomic equilibrium for producers in both markets “without creating too powerful a feedback effect on consumer economies and without overly endangering one producer” (Fletcher 2009). In other words, the Goldilocks principle holds in the market when, “the idea that when natural gas prices are low, solar energy growth declines because solar looks expensive to consumers. Conversely, when

natural gas prices are high, electricity as a whole becomes less affordable, then consumers become less receptive to installing solar because they see it as an added

expense” (Walton, 2013). The question that continues to raise concern in the sector is this: When natural gas prices fall below Goldilocks parameters, could this degrade the investment profile of renewables, especially utility-scale PV? Using the Phillips-Sul convergence test, Nyangon, Byrne, and Taminiau (2017) modelled these two markets to investigate whether LCOE of solar PV and natural gas electricity generation in the United States

47 have converged. They concluded that there has been no price convergence or market

integration at the national level but that some level of integration exists at regional and state levels that require additional testing in future research. Using power systems optimizer (PSO) modeling of utility-scale PV and natural gas grid additions and price convergence, Shavel et al., (2014) simulated how renewable energy and natural-gas- fired electricity generation systems would develop on the ERCOT system through 2032 (Figure 2.6). Weiss et al., (2013) also examined interactions between the natural gas and renewables markets in ERCOT, both in the short and in the long-term.

Source: Shavel et al., (2014, p. 69). Table VII-5.

Figure 2.6 Existing and New Generating Resources for Phase III (MW) Moderate

Federal Carbon Policy

Both Shavel et al., (2014) and Weiss et al., (2013) conclude that, in the short run, low natural gas prices are unlikely to affect investment in rooftop PV or utility- scale solar due to the absence of fuel costs. However, in the long run these studies emphasize the complementary relationship between solar and natural gas and simulate the expected displacement of existing coal-fired generation. Renou-Maissant (2012) used both co-integration analysis and time-varying parameter models to analyze gas

prices of six western European countries for the period 1991–2009 (Renou-Maissant, 2012). She applied Kalman filter analysis to test whether natural gas prices in France, United Kingdom, Italy, Belgium, Germany, and Spain were converging. She found evidence of an ongoing process of convergence of industrial natural gas prices in the six EU countries since 2001 and concluded that there is a strong integration of these markets in continental Europe, except for Belgium. This study shows several complementary attributes and potential for greater partnerships in these markets from the perspective of electricity portfolio and market

49 design—even though some other studies have treated natural gas and renewable

energy sectors as direct competitors and adversaries. For instance, joint platforms of dialogue and collaboration between natural gas and renewable energy industries that are required to define and frame current and future policy questions in the power sector exist in carbon mitigation, system integration, and synergistic and hybrid technology opportunities including power-sector market design.

2.1.4.1 Carbon Mitigation The power sector in the United States faces the dual challenge of addressing transmission congestion and integrating intermittent renewable energy sources. The challenges spawn a highly fragmented and complex power system with divided responsibilities, lopsided investments, and inefficient coordination. Several factors have contributed to this state of affairs: uncoordinated deregulation and electricity market restructuring, growth of DERs, distributed generation and a utility-business model that does not account for disparities in net metering. Added to these issues is

the growing concern about climate change—in particular, about the energy-related emissions in the power sector. Energy-related emissions and associated health and

environmental consequences create additional costs for utilities and the economy. Gas-fired power generation partners better with intermittent renewable generation from solar and wind projects than coal-fired power plants do; hence, given the growing supplies from shale energy, utilities that retool and switch their generation model to hybrid gas-renewables rather than coal-fired plants are likely to optimize their carbon-mitigation potential. Considering these developments, Scott and Bernell (2015) propose a collaborative governance structure that brings together the major utility stakeholders

50 around which power sector operations are organized to address these challenges and to

identify potential carbon-mitigation opportunities. This governance approach is one

that is capable of “steering rather than top-down directing” the energy sector. For instance, they propose integrating key regional transmission organizations—including using the structure of the Regional Greenhouse Gas Initiative (RGGI) and NERC—to implement critical market policies at local and system levels. These policies include policies for reducing carbon and increasing use of renewable resources, electric-grid reliability, grid resiliency in light of growing system complexity and climate change, increase penetration of DERs, and energy-efficiency improvement.

2.1.4.2 Energy System Integration Renewable electricity and natural gas-fired power generation will both play an important role in the decarbonization of the power sector. However, optimizing energy systems across multiple pathways and scales requires better understanding and identification of the benefits of integrating these energy forms. Studies of energy integration, however, show that there is no one-size-fits-all approach. Each state should develop its combination of policies, market designs, and system operations to

achieve the desired system reliability and flexibility (Weiss et al., 2013; O'Loughlin et al., 2012; and Newcomb et al., 2013). Figure 2.7 illustrates areas of energy-system integration that are needed to support better coordination and planning of generation, transmission, and performance. The synergies, interactions, and trade-offs inherent at the nexus of distributed energy generation systems can inform the discussion on system integration, adaptation to new energy paradigms, and partnering options of RE and NG that are viewed much more as part of climate change solution.

51 Figure 2.7 illustrates areas of energy system integration needed to support

better coordination and planning of generation, transmission, and performance. The synergies, interactions, and trade-offs inherent at the nexus of distributed energy generation systems can form a subset of this wider discussion on system integration, adaptation to new energy paradigms, and partnership that lead to efficient electricity markets that are viewed much more as part of climate change solution.

Source: Modified from Cochran et al. (2012)

Figure 2.7 Example of Interrelationships among Areas of Energy Integration

2.1.4.3 Synergistic Benefits of NG-RE Hybrid Power Generation Opportunities

Table 2.3 compares NG and RE power generation systems. For instance, renewable energy electricity generation exhibits higher capital costs and zero fuel costs while natural-gas-fired electricity generation exhibits higher variable fuel costs.

Table 2.3 Matrix of Comparative Benefits of Natural Gas and Renewable Energy Factor Solar Wind Natural gas Capital cost Relatively high but Moderate, some Low, fairly stable overall declining. fluctuation. LCOE. Project life 25 to 30 years 20 to 25 years 20 years (IGCC) Fuel cost None None Variable (gas price volatility) but currently low. Fixed and Low variable costs. Low variable costs. Higher variable costs. variable O&M 53

LCOE Wide cost Fallen dramatically Shale boom in the U.S. spectrum, fallen i.e., onshore wind; has pushed down gas dramatically. highly project prices. Economics of specific. Combined Cycle Gas Turbine (CCGT) plants very attractive. Capacity Low (varies by Low (growing 20- High (42.5%). factor region, 13-22%). 40%). Output Energy output is Energy output is Dispatch flexibility. power variable, somewhat variable, mostly predictable. predictable Carbon Very low Very low Relatively clean-burning impact fossil fuel, carbon impact of natural gas less than half that of coal. Environmen Some opposition to Some opposition to Opposition to shale gas tal and siting of large siting; no extraction, especially social projects for combustion, fewer hydraulic fracturing. concerns ecosystem reasons; conventional no combustion pollutants, low water emissions, no water use. use for PV.

Resource Diverse, but best in Diverse, but often Relatively diverse for distribution Southwest region of far from load unconventional supplies; United States. centers. less so for conventional.

Sources: Lee (2012); Salvatore, J. (2013).

Additionally, co-locating renewable energy and natural gas plant generation can boost capacity factors, grid reliability, resiliency of the transmission network, and value proposition of the investment (Table 2.4). 54

Table 2.4 Value Proposition of NG-RE Hybrid Power Generation Synergies Synergy Natural Gas Renewable Energy Customer (utility offtaker) Power More efficient planning of More efficient transmission planning Improved planning generation transmission strategies Improved operator control (for (bulk energy) For hybrid systems, lower net Improved wholesale power market (i.e., hybrid system) emissions (e.g., CO2, NO2) and firm capacity) Lower transmission expenses lower marginal costs Expanded access to financing and Energy portfolio with reduced 55 Increased financing and associated associated increased generation

systemic risks increased production Improved resource assessments Industrial Increased market for concentrating Expanded market for natural gas as Improved fuel price hedge due Sector solar power (CSP) technologies industrial use increases to fuel diversification offered by industrial customers Potential for pipeline expansion and Diversified revenue streams in Improved integration of on-site decreased risk of congestion from the power market solar PV expanded use of long-term contracts. Benefits from Renewable Expansion of distributed generation Better cooperation with RE firms up Energy Certificates (RECs) could help clarify tariff structures and below the value chain leading to market opportunities and interconnection issues expanded CHP market. Improved capacity to “island” on-site power system during a wider power outage

Residential Better opportunities for load Increased demand in built environment Reduced energy costs Sector shifting to follow RE variability Potential for natural gas demand Increased security and Integrated energy system allows RE response where needed (i.e., Northeast) reliability to contribute to ancillary services, Diversified business growth Potential to participate in DR energy, and capacity markets as opportunities market firm capacity Insulation from electricity and gas price fluctuations Commercial Increased visibility and marketing Potential partnership with RE leasing Lower-priced, reliable, hedged Sector to retail customers firms to provide integrated energy power and energy procurement

56 Reduced integration costs of RE due system strategy

to load shifting Expanded natural gas infrastructure for Additional revenue via power Potential for licensing intellectual car refueling market property and to compete in Potential for natural gas demand Potential for added synergistic electricity market response system relationships in refueling and recharging Transportation Increased demand for renewably- Potential for utilities to rate-base gas Reduced fuel cost and Sector sourced gas stations. diversified fuel risk Opportunity to use existing gas Potential for fueling with low/zero Reduced fuel cost relative to infrastructure carbon sources gasoline

Sources: Cochran et al. (2012) and (2013)

2.2 The Battle Over Centralization of Economic and Physical Control of Electricity Generation

Among many changes that are occurring in the sector, electricity market designs, subsidies, regulations, technological and business model innovations, and power sector restructuring have all contributed to driving the transformation of the power system in the U.S. For instance, the major institutional adaptations—ranging from cost recovery, ex post rate determination, and regulatory approvals of investments—that have occurred in the power sector over the past 50 years to remove entry barriers to wholesale power markets were mostly prompted by technological innovations in electricity generation technology, resulting in further institutional change and economic value creation (Tomain, 2016; Kiesling, 2009). These institutional changes are incomplete and ongoing, however, and will result in a less centralized, top-down model when fully implemented. For the past century, the dominant paradigm in power systems has been centralized economic and physical control. The ideas, framework, and business model strategies presented and synthesized in this study compose a different paradigm: distributed economic and physical coordination based on natural gas and renewable energy blending through price signals, integrated intertemporal wholesale and retail markets, transactions, and contracts. The severity of wholesale market design challenges requires better integration of DERs in the power system, creation of a level playing field for all technologies (especially renewables and natural gas), and greater market flexibility. Enabled by diverse, small-scale generating systems, demand response, energy storage, and other DERs combined with micro-grids, and federated system of distributed system

operators (DSOs), DGs provide network agents and users with responsibility for the provision of reliable, real-time electricity services of the power systems under their operational control. Nevertheless, the concept of a “distributed control” paradigm has been interpreted variously as the power sector strives for reliability (Martini and Kristov, 2015), flexibility for districts’ smart-energy solutions (Sepponen and Heimonen, 2016), and operation of many generators and controllable loads (Bauknecht et al., 2007) through which overall transformation of the network structure is perceived and solutions are sought. While the increasingly affordable and ubiquitous information and communication technologies and the growing digitization of power systems have made decentralized coordination of the power system a reality, this decentralization is different from the “distributed control” concept, as suggested by power systems engineers. As Kiesling notes:

Distributed control in that context means using distributed technology to enhance centralized control of a system. Decentralized coordination is a paradigm in which distributed agents themselves control part of the system, and in aggregate their actions produce order; emergent order. Technology makes this order feasible, but the institutions, the rules governing the interaction of agents in the system, contribute substantially to whether or not order can emerge from this decentralized coordination process (2009, p. 2).

The challenge of a complex system like the power system in the United States is in coordinating the economic and physical requirements of the system (Sioshansi, 2016). The efforts to ensure that coordination and physical control of utility assets are distributed around the power grid represents a key paradigmatic shift in management of distribution utilities. Network users, network providers, and market agents must find new ways to operate in distributed market platforms. Properly assigning

responsibilities for network providers, market platforms, and system operators is therefore critical to an efficient, well-functioning electricity sector. To create a level playing field for the competitive provision of electricity services, the regulation of distribution utilities will require improvement. Improvement in management of grid assets is thus needed to facilitate development of more efficient distribution utility- business models. Second, the only way to establish a level playing field for all DERs and achieve efficient operation and planning in the distributed system network is to dramatically minimize potential conflicts of interest by various participants (i.e., by improving regulated charges and pricing structure such as electricity rates or tariffs.

2.3 A Brief History of the Dominant U.S. Energy Policy Paradigm

A logical starting point for a discussion of the changing U.S. energy policy— from a dominant paradigm defined by centralized economic and physical control based on power-systems engineering and natural monopolies to a more efficient, distributed energy generation model—is to ask, why did an emergent order16 that promotes distributed utilities and “value-creation” take so long to form in first place (Kiesling, 2009)? If energy policy is characterized by a judicious process by which a given entity (often governmental) provides energy development and market solutions—including provision and consumption of electricity services, through legislation, incentives to investment, and other public policy techniques—why

16 In a complex electricity power system, emergent order can take many forms, including achieving economic efficiency, optimized transmission congestion, longer- term resource adequacy, short-term reliability such as keeping the lights on, and improved flexibility in provision and consumption of electricity services (Kiesling, 2009, p. 2-3, 165).

has it not worked? Put differently, if better utilization of existing assets in the distribution system (i.e., renewable energy and natural gas partnerships to create hybrid technologies; R&D of energy-system integration, power-sector market design, and optimized cross-sectoral utilization of energy resources) and smarter energy consumption hold greater potential for cost savings efficiency, why is it underappreciated by network providers and system operators? As Amory Lovins wrote four decades ago in his seminal piece, “Energy Strategy: The Road Not Taken?,” we have for the past century chosen the “hard paths” over the “soft paths” in technology, market design, and policy (Lovins, 1976). The ideas presented and discussed in this study compose a different paradigm: offering decentralized physical and economic coordination through price signals, integrated wholesale and retail markets, contracts, transactions, and digital communication technologies. So why has it taken existing institutions so long to adapt to the soft path: distributed, simple, modular energy systems? Why have policymakers and regulators pursued the hard path: centralized, complex, capital-intensive energy systems? The answer turns out to be complex, and it is both a question of the historical nature of the electricity industry in the United States and institutional design. As Kiesling notes:

Grounded in neoclassical natural monopoly theory, institutions embody four principal components: control or entry, price fixing, prescription of quality and conditions of service, and the imposition of an obligation to serve. The hallmarks of such regulation continue to be price determination based on historic cost recovery, an obligation for the regulated firm to serve all customers who request service, and a legal entry barrier that excludes potential competitors from offering some or all of the services in the regulated firm’s value chain (2009, P. 3-4).

The key institutional adaptions that have taken place in electricity regulation— including federal legislative and regulatory approval of prudent investments and changes made to remove barriers to both wholesale and retail power markets—have largely been impelled by innovations in electricity generation (Kiesling, 2009). However, this institutional change and organizational mechanisms for industry innovations is incomplete and ongoing (Ming and Xie, 2016; Pacheco, York, and

Hargrave, 2014). Four separate institutional gaps exist in the current “hard path” of institutional design for the promotion of energy innovation in the U.S. electricity market (Bonvillian and Weiss, 2009):

• Lack of a strong translational research program in the Department of Energy (DOE) or elsewhere to support breakthrough research in energy innovation. Although the Advanced Research Projects Agency-Energy (ARPA-E), authorized in the America Competes Act of 2007, plays some of this role, at least in principle, its effectiveness at implementing institutional programs within DOE, in a connected and integrated manner with commercialization in mind, is hampered by staffing and funding challenges (Adee, 2007; Burks, Hoehn, and Van Over, 2009).

• Insufficient funding of commercial large-scale, engineering-intensive technologies such as carbon capture and sequestration (CCS) to determine their environmental performance, technical feasibility, costs, and safety.

• A third gap concerns investments—in improved manufacturing capacity, technology, and processes—with manufacturing cost-cutting and production scale-up in mind. This includes manufacturing capacity in energy efficiency technologies and solar PV (such as crystalline silicon (c-Si) modules that

currently account for approximately 90% of global manufacturing capacity, and the cadmium telluride (CdTe) that makes up the remainder of the global manufacturing capacity) (Davidson et al., 2014; Metz et al., 2014).

• The fourth gap is in the lack of “collaborative technology-roadmapping” technique by government, private industry, and academic experts to support strategic technology planning and examine issues associated with each

“technology element and its possible and preferred evolution” pathways (Bonvillian and Weiss, 2009, p. 55).

The implication of the dominant “hard path” paradigm that has defined the power system for the past century is fourfold: a) narrowed-down focus of energy- policy choices in a world that is increasingly complex and complicated; b) exclusion of alternative voices; c) weakening role of the state relative to the federal government; and d) loss of information for cogent analysis of the energy sector. These issues are discussed next to contextualize the evolution from centralized to distributed-power systems in the United States. Later chapters review the efficacy and cost-effectiveness of existing and proposed implementations of policies, regulations, and market designs proposed under a distributed power-system model.

2.3.1 Defending the Centralized Energy Policy Approach in an Increasingly Complex and Changing Environment

The narrowing down of energy policy objectives and significant investments in centralized economic and physical control began in 1824 when the United States Supreme Court determined that the power over interstate commerce would be solely regulated by the federal government (MacDonnell, 1989). Some of these earlier

regulations targeting the use of navigable waters for power generation were motivated by “efficiency” concerns, as conservationists promoted a greater role for nonpolitical scientists and planners in determining the efficient use of the nation's water-power resources (Cole, 1986). This efficiency paradigm is different from the modern efficiency model of doing more with less, as it was centered on the preservation of natural local resources by restricting the use of water from rivers, dams, and causeways. Basin-wide development has, as a result, served multiple purposes—such as electric power generation, irrigation, and navigation—as a key demand of the efficiency paradigm. Progressive conservationists consider this prevailing, narrowed-down energy- policy regime to be inefficient. Unfortunately, the earlier focus on efficiency encountered major obstacles and was not very successful for the following reasons. First, the idea of managing complex and multiple objectives as part of the policy means that the administration of the policy would require greater management and manpower resources. Yet, not enough personnel were available to implement this broad-based agenda. The National Waterways Commission was set up to support this energy policy focus, but it did not take off. Second, the idea of rolling out a comprehensive plan of this magnitude conflicted with personnel requirements; hence, to manage and sustain this complexity, the policy-goals and implementation agenda also had to be curtailed, which translated into the narrowing down of the overall policy objectives.

2.3.2 Exclusion of Alternative Voices Exclusion of alternative voices in the U.S. energy sector is best exemplified by policy responses in the post-1973 Arab oil embargo. The blockade promoted the

energy-independence objective as rich economies, especially the United States, realized just how vulnerable they were to the whims of the oil-exporting countries or to potential disruptions in the flow of oil—whether through the Strait of Hormuz or Malacca—or to broader countries in the Middle East and beyond. As a result, President Richard Nixon announced “Project Independence,” which sought to achieve energy-sufficiency by 1980—with references to the 1940s Manhattan Project and the 1960s man-on-the-Moon—to rally the country towards achieving energy independence. Following the oil embargo and the 1979 Iranian revolution, the initial response of many oil-exporting countries (including the United States) was entirely focused on supply-side production. The knee-jerk reaction was to decide how best to ration limited supplies of oil among the needy and who should stockpile how much to cushion a sudden disruption of future supplies. A minority decided to address the demand rather than the supply-side of the equation. California, for instance, established the California Energy Commission (CEC) to formulate its future energy policy with a major focus on energy efficiency, setting a firm foundation for energy- efficiency objectives in the State that endures to this day. President Jimmy Carter, sitting by the fireplace wearing a sweater and encouraging the American public to turn down their thermostats, called energy conservation “the moral equivalent to war.” These events, however, led to the stifling of decentralized coordination of the power sector and the exclusion of alternative voices. First, the oil embargo was a game changer. It signified for the first time the importance of protecting the nations’ energy supplies and energy security on a global scale. It also contributed to the exclusion of alternative voices concerning an integrated energy-policy system, because regulators, network providers, and market

reforms were focused on enhancing federal subsidies for fossil fuels at the expense of all resources, especially IRE. Second, it raised the significance of energy efficiency, albeit briefly through President Carter's sweater speech. But the overall focus on the provision of electricity services from DERs was narrowed down because some utilities worried that putting more focus on energy efficiency could eliminate investments in the centralized power-supply system, except for routine maintenance and upgrades.

2.3.3 Weakening Role of States Third, Cole (1986) argued that the narrowing down of energy policy to the

“hard path” paradigm led to the weakening of the role of states relative to the federal government. A consequence of the growth of the federal government’s role in the power sector is the result of a high degree of lobbying and growing vested interest at the federal level either for or against certain energy legislations and policies. Despite some adaptive changes over the past decades—such as performance-based rate making to provide firms with incentives for cost minimization—Cole (1986) and Macdonnell (1989) argue that states were weakened in their ability to nurture robust institutions to facilitate electrification and prevent the exercise of market power by vertically integrated monopolist players. The result is misaligned energy policy motivations at the state level as network consumers became disengaged from critical decisions on energy management and investments in the major power-sector infrastructure. The existing policy and sociotechnical systems of this period also contributed to the expansion of centralized, complex, capital-intensive power systems.

2.3.3 Loss of Alternative Voices

Finally, this expansionist “hard path” of technology and policy choices in the electricity market led to a loss of alternative voices in the development of all energy resources across the value chain. Technological change and business model innovation, however, have prompted changes in both regulatory institutions and revenue models, resulting in non-uniform incremental disaggregation of vertically- integrated power systems in different parts of the U.S. (Kiesling, 2009). It stifled the emergence of a level playing field for all resources and cost-effective electricity pricing systems, and it slowed down the building of multi-stakeholder participation in the provision of electricity services. At the same time, the need for proactive reform remained clear. But the lack of alternative voices especially in developing DERs slowed down the transformation agenda of the power systems and the pursuit of forward-looking, “state-of-the-art” regulatory tools, including engineering-based reference network model and incentive compatible menu of contracts to equip regulators for an evolving and uncertain electricity landscape (Pérez-Arriaga and Knittel, 2016). It also weakened institutional capacity and even heightened risks of energy investments. Because multiple network users, providers, and agents engage the power system at different points of the value chain—without a well-regulated structure that rewards cost-saving investments and operations that align utilities” business incentives with the continual pursuit of novel solutions—the result is a power system that is still largely reactive rather than proactive. Network users, mainly residential, know little about the electricity infrastructure, market, and options that provide their electricity services. This dearth of information may be understandable given fixed prices, tariffs and the rates network consumers have paid for many decades, but the lack of their full

participation in the market has slowed down the emergence of an integrated and comprehensive power system that effectively promotes better utilization of supply- side and demand-side resources. The Enron scandal (McLean and Elkind, 2003) and the California electricity crisis of 2000–2001 (Sweeney, 2002) reinforce the riskiness and danger of hasty institutional change in this industry. Before restructuring, three major, vertically- integrated utilities dominated California’s electricity market, i.e., the Pacific Gas and Electric Company (PG&E), the Southern California Edison Company (SCE), and the San Diego Gas and Electric Company (SDG&E). These together supplied roughly

70% of the state’s electricity (Byrne, Wang, and Yu, 2005). Following the adoption of a more DER regulatory policy environment, the California-regulated electricity- market model implemented the following features: (a) institutionalized competition through the unbundling of vertically integrated and investor-owned utilities (IOUs), retail wheeling, open access, “deregulated” entry of new generating capacity, establishment of wholesale market institutions such as California Power Exchange (CPX) and California Independent System Operator (CAISO), and restrictions on long-term contracts; and (b) new stringent pricing rules and other controls such as restrictions on market-based pricing, retail price discounts and freezes, transmission charges and firm-transmission rights (FTRs), performance-based regulation (PBR), no capacity payments, and no regulatory reserve margin policy (Sueyoshi and Goto, 2013). The concern of a hurriedly implemented regulatory praxis as the example of the California electricity crisis shows has substantial opportunity costs. First, the outcomes of this type of policy shift present organizational complexity to network

agents and regulators. California state regulators reached a sobering conclusion that— despite their best policy, finance, and energy market intentions—electricity prices in the state remained 50% above the national average (Byrne et al., 2005). Second, the crisis was characterized by very high wholesale prices, unprecedented rolling blackouts over an extended period, and escalating financial problems for utilities, consumers, and taxpayers. Third, by not reducing retail and wholesale market entry barriers and “rethinking regulatory institutions to enable them to adapt to technological change and not serve as a barrier to its deployment,” the crisis slowed FERC in its push for organized wholesale markets. This in turn slowed down efficient integration of dispatch and pricing of generation assets from different locations with the allocation of scarce transmission capacity, to the entire country (Kiesling, 2009). As a result, the industry relinquished a portion of its potential benefits from competition, innovation, and retail choices. Public-policy interest in California and nationally began to shift to grid modernization and the expansion of transmission and distribution networks through reforms such as, (a) increased financial investments intended to improve remote monitoring and control of high-voltage transmission networks, (b) bottom-up stimulation of investment in local distribution networks intended to improve remote monitoring and communication networks, and (c) increased deployment of smart-grid and smart-metering technology. However, three key post-crisis issues continue to be major obstacles to the California market and elsewhere. These are volatility in wholesale electricity prices, the challenge of demand-supply imbalances such as capacity obligations and diversification of resources, and the continued threats of power outages. Regulatory approaches can promote inefficiency and inhibit

innovation by imposing a uniform requirement on decision makers who have different capabilities, cost and pricing requirements, and benefits associated with innovation and economic dynamism. As Shen et al. (2014) note:

Price-based DR [demand response], reduces the system inefficiency created by time-invariant pricing. If the price differentials between different time periods are significant, customers would most likely respond to the price signals with significant changes in energy use, reducing their electricity bills if they adjust the timing of their electricity usage to take advantage of lower-priced periods and/or avoid reducing load when prices are higher…For a reliable and cost-efficient integration of variable generation output, adequate resources are needed to respond in real-time to any imbalances on the grid. DR is such a resource. Increased availability of dispatchable DR, especially the one with fast response capability, can help address the challenge of regulating the power grid by smoothing out the peaks and valleys associated with the grid variability… new electricity policy reforms, power market changes, and smart-grid technology advancement has dramatically changed [DR’s] role from merely an emergency load response to serving multiple functions as an economic measure of helping flatten the load when electricity price is higher” (2009, p. 817, 821, 823).

At the same time, these electricity-policy reforms and institutional-design changes have made planning and administration of the policies more complex and demanding, such that some demand-side resources have failed to transition successfully to the new market structure. For instance, direct load-control17 efforts involving small energy consumers have been terminated due to challenges associated

17 Direct load control is available to residential or small commercial customers. It refers to programs utilities offer remotely to customers such as turning off customer’s electrical equipment (e.g., air conditioning units) on short notice (Shen et al., 2014).

with metering and load aggregation,18 the need for expensive system changes to recognize their impacts, and other considerations such as lack of personnel and administrative resources to oversee their implementation (Eisen, 2012; Faruqui et al., 2010; US. Department of Energy, 2013). Furthermore, thermal energy-storage devices have not proven economically viable in the new market due to their limited differentials in the cost of electricity between on- and off-peak periods, despite targeted reforms to promote energy-storage technologies (Kost et al., 2013).

Despite some adaptive changes over the past two decades—such as the creation of PJM Interconnection,19—the challenges associated with distortion of merit order in the energy market continue to persist (Bresler et al., 2013). Following FERC orders 880, 888 and 889—which led to deregulation of traditional utility companies and to the subsequent setting up of an independent system operator (ISO) to manage functions of the electric grid that were previously undertaken by vertically integrated utility companies—PJM was born. PJM became an ISO in 1997 and a regional transmission organization (RTO) in 2002 (Sioshansi, 2013). The transition to a nodal pricing structure—that is, locational marginal pricing (LMP) aimed to better assign all congestion costs to the operators responsible for

18 Load aggregation is the process by which individual energy users come together in an alliance to leverage and secure more competitive prices when soliciting bids from energy suppliers than they might otherwise receive working independently.

19 PJM Interconnection is the world’s largest wholesale electricity market which has been operating since 1997 to provide network agents and users with systems and rules that ensure a fair and efficient market operations. It is an RTO and operates a competitive wholesale electricity market which manages the high-voltage electricity grid for all or parts of Maryland, Michigan, New Jersey, North Carolina, Ohio, Pennsylvania, Tennessee, Delaware, Illinois, Indiana, Kentucky, Virginia, West Virginia, and the District of Columbia.

creating the congestion—came with FERC in place. However, a well-functioning energy market requires both an efficient supply-side and a demand-side. The demand- side of the PJM wholesale energy market is underdeveloped (Aldina and Soden, 2013; FERC, 2010). This underdevelopment is a key consideration for maintaining an offer cap in the PJM interconnection market (Sioshansi and Pfaffenberger, 2006). As a result, despite all the well-intentioned efforts in applying energy policy as regulation to create a well-functioning market structure through PJM, price volatility, supply- demand imbalances, and power outages remain major challenges to the market. Figure 2.8 shows variations in hourly demand bid data for 2016 (PJM, 2017).

Figure 2.8 PJM’s Hourly Demand Bid Data for 2016

The extent of price variability in PJM network depends on three key factors: a) the magnitude of the demand shift between peak and off-peak periods, b) the elasticity of supply over the relevant range, and c) the price elasticity of demand (Bresler et al.,

2013). Additionally, a major challenge of the PJM’s LMP structure is that transmission losses are not borne by generators while firm access rights are rewarded rather than penalized. This leads to distortion in the merit order in the market. There are also reasons to be skeptical about the ability of regulators to administer regulatory measures in the market effectively and efficiently. Outside of the LMP-based energy markets, scheduling and payment for energy is based on the contract path even though the associated actual energy deliveries flow in the path of least resistance (Bhattacharyya, 2011, p. 498). Without prudent and effective management of the capacity market, a profitable game can be created and congestion within the LMP market can be affected; hence, the need arises for further administrative requirements in creating pricing points which can successfully recognize the location of generation and the physical path of flows. Thus, although the PJM market exemplifies the fact that competitive wholesale power markets can work if properly designed and supported by appropriate regulatory structures, challenges inherited from the dominant paradigm of centralized economic and physical control of power systems continue to bedevil the interconnection system. These challenges concern how to mitigate market power, how to increase the level of demand response, how to provide appropriate incentives for investment in the transmission infrastructure, how to implement scarcity pricing, how to manage seams with other RTOs and non-market areas, and how to incorporate voices of different stakeholder groups and achieve broad-based engagement with network generators and

users. Solutions to these challenges require the application of a broad-based and integrated electricity policy and institutional design that is principally based the on creation of competitive capacity markets and a range of more distributed technologies. The experience of the California electricity crisis in 2000 and 2001 exemplifies an industry that is the backbone of our modern, technology-rich lives and that is also one of the most “technologically backward in the modern economy” (Kiesling, 2011). Although regulatory systems can attempt to streamline operations of the electricity value chain and reduce transaction costs (through service improvements, customer- driven improvements, and new product offerings), so long as the centralized physical assets remain, relics of the dominant “hard path,” may not fully emulate certain product attributes that are valuable to the network users and the market. As Kiesling notes:

Economic growth and technological change have brought the electric industry and its regulation to a crossroads. Technological change from outside the industry has prompted changes in both regulatory institutions and business models, leading to the incremental disaggregation of the vertically-integrated firm in some regions of the U.S. and not in others (2009, P. 4).

Table 2.5 Emerging Pathways of Power System Transformation

Present Status Adjacent Status Vertical Integration • Next generation performance-based • Little or no power market restructuring regulation • Utility as a single-buyer • Clean energy restructuring Restructured Market • Mainstreaming the DSO model

• Intermediate / high levels of power market restructuring Independent system / market operator Low Energy Access • Bottom-up coordinated grid • Unreliable, limited or no access to expansion electricity • Bundled community energy • Can occur in restructured or vertically planning integrated market settings

Table 2.5 summarizes the emerging pathways of viable models for power system transformation: vertical integration, restructured markets, and low energy access environments. The next section details the changes that are occurring in the power sector to contextualize many of these issues in the face of constant and pervasive change— particularly in the intersection of technological change, policy and regulations, and market designs. While this study does not try to predict which drivers, technologies or business models will prevail, it is necessary to identify unnecessary barriers and distortionary economic incentives that presently hinder the efficient growth of distributed technologies if we are to understand a framework that will facilitate an efficient power sector regardless of how business models or technologies or regulatory objectives develop in the future.

Chapter 3

EVALUATION OF TRENDS DRIVING POWER SECTOR TRANSFORMATION IN THE UNITED STATES

This chapter discusses the background of the utility sector and surveys the key drivers of change for the utility of the future in the United States. The section also discusses the core dimensions of the Hamel business model framework, including the customer interface, core strategy, strategic resources, and value network, as an example of an approach for understanding these drivers and possible future utility sector changes.

3.1 Key Drivers of Change

Even as early as 1970, some scholars in the power industry believed that distributed technologies could play a major role in the design and operation of the power systems of the future and in balancing demand and supply resources. At that time, however, the potential for distributed generation was primarily evident in CHP power generation—a development that motivated the 1978 Public Utility Regulatory Policy Act in the United States and allowed independent power producers to enter the market. Amory Lovins, for instance, predicted the eminent arrival of a more distributed and renewable future (Lovins, 1976), but it would take nearly four decades before this seemed a realistic possibility. Schweppe (1978) imagined a power system in which demand actively participates in the provision of critical electricity services. This vision is fast becoming a reality today. A convergence of circumstances that are driving fundamental changes in the power sector is discussed in detail next. These ten key trends are as follows:

1) renewable energy cost reduction and the growing penetration of distributed generation; 2) reversed energy-demand growth patterns and flat to declining electricity demand forecast marked by more progressively aggressive state-wide demand-side management and energy-efficiency policy schemes due to rise in distributed generation and demand response; 3) innovations in data, intelligence, and system optimization—a proliferation of advanced information and communications

technologies (ICTs) and “intelligent efficiency” technology options capable of unlocking device-level measurement and control at large- scale; 4) increasingly integrated power system and interactions with other sectors (i.e., the growing interconnectedness of electricity with other critical transportation and communication infrastructures, which enhances the importance of electricity in modern economic development); 5) energy security, resilience, and reliability goals, including positive externalities of shale gas development in the United States such as low natural gas prices relative to other fuels (Nyangon, 2015b); 6) energy access and aging infrastructure imperatives; 7) revenue and investment challenges across the value chain: supply, transmission, delivery, and demand; 8) diverse participation in power markets and changing utility business models;

9) evolving customer engagement; and 10) climate change and environmental concerns at the local and global levels over air emissions—increasing decarbonization of the energy system as part of global mitigation efforts. Likewise, the growing security threats from extreme weather events add more pressure for better planning practices.

To provide background to the following discussion of the potential power system of the future, Figure 3.1 and Table 3.1 show the complex set of dynamic features that drive power-system evolution.

Figure 3.1 Ten Trends Driving Power System Transformation in the United States

Table 3.1 Ten Trends Transforming the Electric Power Sector Trend Example Key Drivers Threats Shifting energy Natural gas Renewable portfolio standard mandates Competitive process in recovery generation mix; growth Resource mix integrating DER. objectives power system is Renewables Technological advancements Stranded assets decarbonizing growth and becoming Abundant and cheap domestic natural gas Stranded costs for other fuel sources more distributed Retirements of Grid reliability coal and nuclear Aging assets generators Relative economic of natural gas Ability to meet demand requirements

79 Environmental mandates (coal) Losing share of generation capital base

Siting concerns and requirements (nuclear) Flat to declining Low GDP growth Flat to declining electricity retail sales demand growth Moderating population growth Limited or declining sources of new demand Improvements in energy efficiency of buildings and industry A shift to less energy-intensive industries

Aging Distribution and Investment in distribution system Lower customer satisfaction infrastructure transmission Transmission investment Lack of transmission investment imperative investment Reliability challenges

Innovations in Physical security Rapid growth of Internet of Things and Unproved smart grid benefits data, and cybersecurity Digitization (information-intense Reliability interruptions intelligence, and economy) Reputation risk to utilities if system Increased focus on grid security optimization unaddressed Rising cyber threats

80 Revenue and Falling revenue LCOE reductions

investment Grid investment challenges Evolving Prosumers Technology improvements Lower customer satisfaction customer Micro-grids Decreasing volumetric utility sales engagement Diverse Electric vehicles Tax and production incentives Decreasing utility sales participation in Energy storage Third party financing / ownership power markets and changing Growth in distributed generation utility business Storage adoption models

Increasingly Electric vehicles Regulatory mandates Added complexity in grid operations integrated power Vehicle-to-grid Technology improvements Customer disintermediation system and systems interactions with Infrastructure availability Use as a behind-the-meter generation other sectors Building’s Federal and state incentives thermal storage Relative economics compared with capacity conventional options (electric vehicles)

Energy security, Physical security Smart-grid technology Unproved smart grid benefits resilience, and and cybersecurity Increased focus on grid security Reliability interruptions

81 reliability objectives Rising cyber threats Reputation risk to utilities if unaddressed Climate change Paris Agreement Emphasis on green technologies and High environmental composition of concerns over air emissions from emissions the energy sector

3.1.1 The Shifting Energy Generation Mix; Power System is Decarbonizing and Becoming More Distributed

3.1.1.1 Rapid and Unprecedented Expansion of Intermittent Renewable Electricity in the U.S.

The rapid growth of distributed resources is occurring against a backdrop of cost reductions and the ongoing transition to a more renewable and intermittent generation mix (Feldman et al., 2016). Utility-scale PV production in the United States has grown rapidly in the last decade. In the second quarter of 2015, the U.S. installed an additional 1,393 MW of solar PV, thereby bringing the market up to 2,722 MW in the first half (GTM Research and SEIA, 2015). According to GTM Research and the Solar Energy Industries Association (SEIA) (2017), by 2021, nearly 30 states in the U.S. will be home to over 100 MWdc annual solar markets, with 20 of those states generating more than 1 GWdc of operating solar PV (SEIA/GTM Research, 2017). This is a significant change considering that it was in 2015 when, cumulative operating PV capacity nationally eclipsed the 20 GW mark for the first time (including all types of PV market segments—utility, residential, and commercial). SEIA estimates that new utility PV installations brought on-line on an annual basis reached 10 GW in 2016, which is up 29% and 24% over 2015 and 2014, respectively, with growth occurring in all the three segments but most rapidly in the residential market (SEIA/GTM Research, 2017). With these ongoing shifts in distributed generation, the utility PV market continues to be the primary driver of installation growth in the U.S. solar market. It accounts for 3.2 GW of installations that came online in the third quarter of 2016 (or 77% of the total PV capacity installed) and 52% of the capacity

installed in the second quarter of 2015 across market segments (SEIA/GTM Research, 2017; GTM Research and SEIA, 2015). Given the extension of the federal investment tax credit (ITC) to 30% through 2019 for third-party-owned, residential, non-residential, and utility PV projects under Section 48 of the tax code, NREL forecasted that the tax-credit extensions will accelerate solar and wind deployment through the 2020s with an aggregate renewables capacity of slightly over 450 GW, as illustrated in Figure 3.2 (Mai, Cole, Lantz, Marcy, and Sigrin, 2016):

Extending federal [renewable energy] tax credits, as enacted in the Consolidated Appropriations Act of 2016, can boost [renewable energy] deployment through the early 2020s. However, longer-term deployment effects are less certain because deployment drivers including future natural gas prices, and [renewable energy] cost reductions could play a more substantial role in the 2020s and beyond. More rapid [renewable energy] growth—driven by the tax credits—can result in significant cumulative CO2 emissions reductions.

Similarly, GTM Research and SEIA (2017) writes as follows:

Without question, the extension of the federal ITC ranks as the most important policy development for U.S. solar in years. Between 2016 and the end of the decade, the ITC extension will spur nearly 20 GWdc of additional PV capacity, positioning U.S. solar to remain a double- digit gigawatt annual market heading into the next decade. The extension of the federal ITC has enabled nearly 6 GWdc of utility PV to spill over into 2017, providing utilities with necessary breathing room to interconnect the 8 GWdc pipeline of utility PV currently under construction. By 2019, U.S. solar is expected to resume year-over-year growth across all market segments. And by 2021, 30 states in the U.S. will be 100+ MWdc annual solar markets, with 20 of those states home to more than 1 GWdc of operating solar PV

The above statements imply that distributed generation allows for unexamined synergies between natural gas and solar to launch a compatible market development of the two sources. 2016 was another record year for the growth of the U.S. solar PV market, with the residential segment adding a more-than 2 GW annual mark for the first time ever, and by a wide margin. According to GTM Research/SEIA (2015), the utility PV market grew by 16% and the non-residential PV market grew by 8%. The former market segment accounts for more than 50% of all installations brought online at the end of 2015. Under the assumptions of the EIA Annual Energy Outlook Reference case for base natural gas price shown in Figure 3.2, RE capacity driven by the federal ITC will expand and peak at 53 GW in 2020 (Mai et al., 2016).

Figure 3.2 Cumulative Installed Renewable Capacity by Scenario

Figure 3.3 Historical and Projected Solar PV Capacity by Sector, 2008 - 2020

Figure 3.3 shows historical and projected solar PV capacity for 39 states plus Washington, D.C., and their market segment through 2021.20 Figures 3.4 and 3.5 show

20 Notes: Total U.S. PV capacity additions are based on GTM Research and SEIA (2010 - 2015), IREC's data collection, and LBNL's Tracking the Sun Database. Note that GTM Research and SEIA's definition of utility-scale PV capacity differs from LBNL both in size thresholds and treatment of project phase completion.

the forecast of PV installation by segment for 2010-2021 and the contracted-utility PV projects21 in development (totaling 19.4 GW).22

Figure 3.4 U.S. PV installation forecast by segment, 2010-2021

21 The continued growth of utility-scale PV in the U.S. in the second quarter of 2015 marked the ninth consecutive quarter in which PV added at least half a gigawatt. For the first time in 2014 utility PV became an economically competitive energy resource for meeting peak power needs GTM Research and SEIA, 2015).

22 Procurement of 40% of the 16.6 GW utility PV pipeline in development is primarily due to economic competitiveness of solar in comparison to fossil-fuel alternatives.

Figure 3.5 Contracted Utility PV Pipeline in 2016

In 2015, a GTM Research/SEIA (2015) market-research analysis ranked California, North Carolina, Nevada, New York, and Massachusetts as the top five U.S. states for distributed energy generation. Solar PV installations in New Jersey ranked seventh behind Arizona in PV installation capacity. In addition, a growing number of states have redesigned their electricity markets to accommodate the growth of DERs— notably California (through AB 327 and its offshoots) and New York (through the

REV initiative discussed in Chapter 4). Other states such as Hawaii, Maryland, Massachusetts, and Minnesota are at different levels of formulating or implementing similar transformational processes to support their solar markets. Considering 2017 and beyond, residential, and non-residential PV markets are both expected to grow year-by-year, but the overall U.S. solar market will drop by approximately 4% on an annual basis in 2017 (SEIA/GTM Research, 2017). In addition, because of economic factors related to PV installations (such as net metering), residential solar will continue to depend heavily on the level of prevailing retail electricity rates (Darghouth, Wiser, and Barbose, 2016). Other economic considerations, such as component pricing, will be determined in the market by excessive component supply versus demand. For instance, in the third quarter of 2016, SEIA-modeled costs fell from $1.49/W and $1.80/W in the second quarter of 2015 to $1.09/Wdc and $1.21/Wdc for fixed-tilt utility PV and one-axis tracking technology, respectively (SEIA/GTM Research, 2017; GTM Research and SEIA, 2015). This represents a quarterly systems price reduction of 6.8% and 6.9% for 2016. The fundamentals of the solar PV industry remain quite favorable, with strong end-user demand, stable polysilicon, and solar panel pricing (see Figure 3.6), and improving profitability among solar manufacturers.

Figure 3.6 U.S. Polysilicon, Wafer, Cell, and Module Prices

Notes: Polysilicon and PV components prices fell from the second quarter of 2014 to the second quarter of 2015. Increased inventory and seasonally weak demand primarily drive significant polysilicon price reduction. Weak market demand levels and pressure from buyers looking for low- prices for module continued over the same period affecting wafer and cell prices in the second quarter of 2015 and is likely to continue in the first quarter of 2016. Data source: SEIA/GTM Research (2017)

3.1.1.2 Growth in Natural Gas Production and Low Cost Shale Domestic natural gas production in the United States is on the rise due to technological advancements that has improved shale gas extraction, and this growth is projected to continue into the 2030s. After natural gas, wind is the second-largest contributor of new capacity (Mai et al., 2014; U.S. Energy Information Administration, 2017a, 2017d). Figure 3.7 shows a strong relationship between the fall in natural gas price and fall in the MAC Solar index.23 This has raised concerns whether solar can flourish as gas prices continue to fall amid a glut of shale gas on the market. For instance, the data series in Figure 3.7 shows that solar PV and natural gas prices were quite widely spread during early 2008, began to fall and converge in early 2011, but started to rise and diverge after mid 2011 until the end of 2016 (Bloomberg, 2016; MAC Solar Index, 2017; S&P Dow Jones Indices, 2017). These changes also reflect market trends given that the installed grid-connected PV capacity was approximately 1600 MW24 in 2011 (inclusive of all types of PV), which represents a 74% increase over the 918 MW capacity installed in 2010 (NREL, 2010; EIA, 2012b). Including all types of PV, the cumulative installed capacity grew from 200 MW to 3.5 GW between 2000 and 2011 (Mendelsohn et al., 2012). Between 2008 and 2016, average annual natural gas prices fell by 35% from 5.63 to 3.68 ($ per Million Btu), while the MAC Global Solar Energy Index (SUNIDX) price dropped by nearly 79% (Bloomberg, 2016). However, the SPGSCI

23 MAC Solar Energy Index tracks globally-listed public companies that specialize in providing solar energy products and services.

24 In March 2012, SEIA and GTM reported 1.1 GW of installed capacity for utility- scale PV alone in the U.S. in 2011. However, definition of utility-scale used in their analysis is different from NREL’s (i.e., capacity > 5 MW).

natural gas index price increased by 14% over the same timeframe. Figure 3.7 shows annual SUNIDX and S&P GSCI price indices and Henry-hub-linked natural gas spot prices (for which a complete data set from 2000 to 2016 is available).25

Data sources: Bloomberg (2016)

Figure 3.7 Henry Hub Spot Prices, SUNIDX and S&P GSCI Indices, 2000 – 2017

25 All information for S&P GSCI natural gas index spot prices prior to its launch date are back-tested. Back-tested performance, which is hypothetical and not actual performance, is subject to inherent limitations because it reflects application of an index methodology and selection of index constituents in hindsight (S&P Dow Jones Indices, 2017). The inception of the MAC Solar Index was 31-March-2008 with a base of 1000. The historical data prior to the index inception, i.e., from 31-March-2005 to 3-March-2008, is backcasted simulated data (MAC Solar Index, 2017).

As discussed in the previous sections, natural gas-fired electricity generation plants are increasingly conflated with distributed renewable electricity plants. Figure 3.8 illustrates the overlap and taxonomy of these resources. While many RE generation systems can be deployed in both a distributed or centralized form, this study focuses on DERs that are both renewable and nonrenewable (natural gas) and their potential role in the evolving distributed utilities, which include small- and micro-cogeneration, backup turbines, and gas-fired systems.

Note: Region 4 in the graphic represents the primary area of this study

Figure 3.8 Taxonomy of Distributed and Renewable Energy Resources.

3.1.2 Flat to Declining Electricity Demand Growth Since the 1950s, U.S. electricity consumption has gradually declined each decade due to moderating population growth, energy efficiency improvements, technological advancements in appliance manufacturing, and decoupling of the economy to less energy-intensive industries (Nadel and Young, 2014). In addition, the power system is becoming more distributed. Coupled with state-focused policy schemes such as DSM and demand response, negative load growth is projected to continue. Negative load growth is coupled with negative policy schemes, both of which are projected to continue (Barbose, Galen L., Goldman, Charles A., Hoffman, Ian M., Billingsley, Megan, 2013).

3.1.3 Aging Infrastructure Imperatives

The current electricity infrastructure is increasingly challenged by transformations with respect to aging infrastructure and capacity and cyber and physical threats (Overton, 2015; Campbell, 2015; U.S. Department of Energy, 2017b). The increasing interdependence of power systems and interactions with other sectors—such as water, finance, telecommunications, transportation, and emergency- response systems—may exacerbate the vulnerability in this infrastructure. The American Society of Civil Engineers gave the U.S. energy sector a D+ rating (ASCE,

2013). Similarly, the Edison Foundation estimates that the national level vulnerability costs of energy infrastructure to be over 2010-2030 at $582 billion, in nominal terms (Chupka et al., 2008). In New York, the Department of Public Services estimates the cost of replacing its aging electric-transmission-and-distribution infrastructure over the next decade (to meet the projected energy demand) at approximately $30 billion—i.e.,

double the $17 billion invested in the state’s grid over the past decade (NYPSC, 2014).

3.1.4 Innovations in Data, Intelligence, and System Optimization

Innovations in data, intelligence, and system optimization—including a proliferation of advanced ICTs and “intelligent-efficiency” technology options that are capable of unlocking device-level measurement and control at large-scale—has in part facilitated the decentralization of electricity resources (Rogers et al., 2015). Increased digitalization of the power system has facilitated the computation and communication of the value of electricity services with finer temporal and spatial granularity. The result is that this has enabled energy demand to become more price responsive. The value of price-responsive demand is vast and growing for both consumers and utilities. It facilitates avoidance of energy payments during high-price conditions, integrates a price/quantity demand relationship into short-term operations, and reduces the frequency and magnitude of energy-scarcity events, thereby improving the reliability and quality of electricity supply. With the growth of DERs—particularly with low- cost shale gas, and relatively fixed generating costs of solar and wind—digitization is facilitating a more active management of networks, thereby eroding the passive network management model of the centralized-fossil-fuel paradigm (Pérez-Arriaga and Knittel, 2016).

3.1.5 Revenue and Investment Challenges

The transformation toward power systems of the future has been a goal of the U.S. for several decades, and it is being shaped by investments in demand-side resources (energy efficiency, demand response, and distributed generation) and

renewable energy. Renewable energy investment in the United States (dominated largely by solar power) increased by 19% in 2015 to US$ 44.1 billion—the largest increase since 2011 (Sawin, Seyboth, and Sverrisson, 2016). The pressure for integration of policy, finance and investment, and markets to support the roll-out of the power system of the future is felt by existing power-sector stakeholders across the value chain: generation, transmission, distribution, and demand (Byrne et al., 2016). For instance, supply-related revenue challenges are already being experienced in the midcontinent independent system operator (MISO) system at high renewable energy penetration—specifically at the bulk-power and wholesale market level—as abundant solar and wind energy is integrated into the market at zero marginal cost, thereby resulting in reduced utilization of existing conventional thermal-electric power generators (nuclear, natural gas, coal) (Zhang and Giannakis, 2016). At the level of customer retail, the expansion of energy-efficiency programs and the increased capacity of solar PV generation to reduce retail electricity consumption is raising revenue concerns for some utilities and impacting perceptions of investment risks and creditworthiness (Neme and Cowart, 2014; Zinaman et al., 2015). This revenue challenge is occurring in Arizona, New York, North Carolina, California, Colorado, and a few other states with growing PV generation, thereby adding pressure on utilities to adopt innovative mechanisms to incentivize efficiency in production systems and flexibility while allowing generators to recover their fixed costs. Strategic planning for the diversification of energy mix and utility choices, however, requires explicit recognition and response to both large-scale, nationwide challenges. However, it also requires cognizance of idiosyncratic local and regional obstacles and opportunities. For instance, utilities across the country exhibit distinctive

manifestations of the so-called “death spiral” (the cycle of eroding market share to distributed energy “prosumers” that raises the costs of remaining utility customers, thereby leading to accelerated market losses) (Athawale and Felder, 2016; CAISO, 2013; Felder and Athawale, 2014; Graffy and Kihm, 2014).

Nationwide, the dreaded utility “death spiral” argument is substantial. Some analysts believe that it is near or already here. For instance, according to Accenture (2014), projections of utility-sector revenue erosion—which are based on reduced load due to the rapid growth of distributed generation and gains in energy efficiency— could reach $48 billion annually by 2025 (Figure 3.9). However, considering the profiles, distributed energy generation systems—including potential policy, finance and energy market challenges and regulatory policies (e.g., RPS, and tariff designs)— differ substantially across states and even cities (Ribeiro et al., 2015; York et al., 2013). Thus, the challenge of the “death spiral” will be felt differently across the nation’s utilities. Similarly, the limitations of the aging infrastructure—due, for instance, to different rates of growth, different electricity-revenue and retail sales, differing geography, or varying levels of research and development (R&D) commitments—varies substantially by region and state (U.S. Department of Energy, 2015). Figure 3.10 shows growth in historical total electricity revenue, retail sales, real GDP, net energy generation (GWh) in PJM region, and revenue from electricity sales.

Notes: The U.S. residential and commercial demand excludes distributed generation— i.e., refers to electricity purchased from the grid and does not include self-generation. Under the alternative scenarios, the Accenture analysis suggests significant shortfall in potential revenues as purchase of electricity from the grid reduces with increasing distributed energy generation. The greatest risk to revenues and demand are in the period to 2025 in the perfect storm scenario with 300 TWh of potential loss (Data source: (Accenture, 2014).

Figure 3.9 Projections of Utility-sector Revenue Erosion in the United States Due to

Distributed Energy Resources

Notes: Total electricity retail sales measured in million kilowatt-hours, real gross domestic product (billions of chained 2009 dollars, annually, not seasonally adjusted), PJM RTO historical net energy (GWh), and revenue from electricity sales to ultimate customers (thousand dollars).

Figure 3.10 Total Electricity Retail Sales

Table 3.2 summarizes market design adaptations required to address shifts in market dynamics as renewable energy generation expands, end-user demand strengthens, polysilicon and solar-panel pricing stabilizes, and the return on investment for solar-power systems reaches grid parity. Power markets organized by both PJM interconnection and ISO New England (ISO-NE) now conduct forward- capacity auctions (Schatzki and Hibbard, 2013). As Gottstein and Schwartz note, these forward capacity auctions,

Permit a wide range of demand-side resources to compete with supply- side resources in meeting the resource adequacy requirements of the region. The response of demand-side resources in the PJM and ISO-NE auctions is impressive, and their participation is clearly demonstrating that reducing consumer demand for electricity is functionally equivalent to — and cheaper than — producing power from generating resources for keeping supply and demand in balance. (2010, P. 3)

Table 3.2 Potential Market Design Adaptations Needed to Address Supply-related Challenges at High Renewable Energy Penetration Encourage cost-effective Short-term: Long-term: Scaling up energy efficiency measures Upgrading electric investment in new capacity grid operations to resources to unlock market ensure flexibility flexibility • Incentivize more efficient • Upgrade • Develop tools to better production and system scheduling, forecast net demand and operations dispatch, and the value of various • Account for transmission weather forms of flexibility losses in delivered prices forecasting • Invest in low cost through, e.g., LMP • Consolidate generating options to • Allow energy efficiency balancing areas increase flexibility of and other demand resources • Promote more existing generation to participate in capacity dispatchability of • Adapt forward markets variable investment mechanisms renewable to capture the full range production of resource capabilities • Co-optimize • Adopt forward capacity energy and markets for specific reserves to system services improve the • Create forward markets effectiveness of for a time shifting scarcity pricing services • Expand the role of demand response

• Expand automated • Open day-ahead • Encourage new market measurement and markets for entrants wherever verification (M&V) plans existing ancillary possible and consistent consistent with IPMV26 services and with overall market protocols to deliver reliable begin to qualify structure savings new ancillary • Set standard capacity services values for a menu of standard efficiency measures • Consider location-specific efficiency measures as an alternative to transmission

Sources: Gottstein and Schwartz (2010); Neme and Cowart (2014)

As demand-side resources permeate capacity markets, it is becoming increasingly apparent that existing market designs are inadequate to deliver the amount of generating capacity required for reliability needs. For instance, competitive markets have not elicited sufficient investment in plant capacity to meet resource- adequacy requirements in some parts of the United States—as was expected in the post-electric industry restructuring of the mid-1990s (Sweeney, 2002). Economic regulation has been a pinnacle of the electricity industry, as sub-additive firm costs and economies of scale create natural monopoly activity in electricity distribution. A

26 The International Performance Measurement and Verification Protocols (IPMVP) offers best practices for verifying energy efficiency, water efficiency, and renewable energy projects and to assess building performance. It permits use of four measurements methods: partially measured retrofit isolation/stipulated measurement, retrofit isolation/metered equipment, whole facility/regression, and calibrated simulation.

100

key distinction in the regulatory strategy of the United States and in that of other nations begins to emerge upon reviewing the federal government’s actions pursuant to their distribution-utilities goals. Distributed renewable energy and efficiency mandates are set for federal buildings and installations only. Efficiency standards (such as fuel- economy targets and other programs within the Department of Energy such as the Energy Star program) are adjusted for industries in which costs can be passed to network users. And investments are allocated for market-based activities (creation, support, manipulation, incentives) that are more reward than punishment. It is rare to observe direct economic regulation that is not passed on to network users eventually. As a result, socializing or averaging the distribution-network costs to address distributional concerns is an increasing challenge, as disruptive energy efficiency and distributed technologies permeate capacity markets.27 As the scale of participation of energy efficiency and distributed generation increases in capacity markets, utilities are actively pursuing efficient power-system operations. For example, dispatch of power systems is increasingly based on the marginal cost to generate “smart” appliance programs through credit enhancements, purchase agreements, and rebates. These programs can help diffuse peak loads, respond to pricing signals, and lower market-clearing prices, with potentially significant economic benefits to consumers. Various utilities are also employing on- bill financing, which is a utility-serviced loan guaranteed by the utility payment and

27 Socializing the cost of infrastructure has limits and remains a tariff-design challenge to utilities because it sends inefficient economic signals in the market and negates the “equal-tariff-for-all” principle. For instance, although the quality of service standards to rural customers is usually subsidized, the quality of service is not exactly the same as what is offered to urban customers.

101

included on the utility bill for both efficiency upgrades and renewable systems. In addition, utilities and the private sector are engaging in innovative power purchase agreements that allow a group of residents to collectively “purchase” a solar array under contract to buy power over the life of the equipment. The federal government has been doing this for military and other federal housing with great success, and it is becoming more common in residential neighborhoods. Nonetheless, the institutional challenges remain significant at different levels in many parts of the country. For example, the rapid growth of distributed energy- generation systems has resulted in substantial grid-integration problems that risk stalling further progress unless continuing reforms in business models are expanded to cover many parts of the United States. RAP (2013) and Hogan (2013) recommend a two-part (energy + capacity) pricing structure and additional network investment in the delivery of energy that is based on three main objectives: (a) improving power- system operations and the integration of solar PV and wind by clarifying grid- company responsibilities and accelerating certain power-sector reforms; (b) improving renewable energy policies to lower cost and improve efficiency; (c) blending of natural gas and renewable energy sources more effectively through the creation of a flexible, least-cost mix of new conventional and renewable generation.

3.1.6 Evolving Customer Engagement Amidst the profound shift underway in the power sector, consumer choice and preference is a major driver of change. Greater customer engagement is directly driving investment trends in the consumption and provision of electricity services— for example through smarter homes, electric vehicles, distributed-generation, and energy-efficient appliances (U.S. Department of Energy, 2017b). As more customers

102

adopt distributed-energy resources—fuel cells, rooftop solar PV, small wind-power systems, hybrid power systems (e.g., solar hybrid and wind hybrid systems), onsite energy storage, electric vehicles, combined heat power, and energy management systems, or a combination of these technologies—to enhance system reliability, minimize cost, reduce environmental impact, and meet other performance expectations—they are benefiting from an unprecedented level of choice that was previously deemed impossible. To unlock the full potential of these choices, utilities in various parts of the country are reinventing themselves—by implementing a menu of contracts to incentivize efficiency and to pursue performance-based incentives and innovations in network investment and management—to facilitate these choices and thereby create a clean, reliable energy future (Jenkins and Pérez-Arriaga, 2017; Southern California Edison, 2016). The path suggested in this study outlines a set of utility capabilities, regulatory requirements (quality and economic regulation), and charges for electricity services that will be vital to facilitating these preferences and willingness to pay in an efficient, reliable, and affordable power system. Direct consumer participation in power markets holds the potential to change existing electricity market designs as distributed energy resources increase variability on the supply-side (Cossent and Gómez, 2013). Greater customer engagement in power markets across residential, commercial, and industrial sectors has been shown to be technically feasible, for example in the PJM network (PJM, 2013). As this engagement on the distribution grid (whether through digital communications, DERs, sensors, control systems, digital “smart” meters, greater customer participation, etc.) becomes economically viable and socially routine, the rest of the power sector faces both technical and policy challenges of co-optimizing

103

supply and demand dynamically as well as the need to embrace opportunities for the delivery of energy services.

3.1.7 Diverse Participation in Power Markets and Changing Utility Business Models

Across the country, participation in power-market ecosystems is diversifying. In emerging economies, where many power sectors are state-owned, there is a push to boost investment by opening the sector to more independent power producers (e.g., Mexico, South Africa). In more mature power sectors, this trend is taking the shape of greater direct consumer-power market participation, from generation to demand response. These diversification trends, driven by a need to respond to legacy challenges, will add a layer of complexity to power systems of the future.

3.1.8 Increasingly Integrated Power System and Interactions with Other Sectors The growing interconnectedness and interdependence of electricity and other key sectors with critical transportation and communication infrastructures—such as natural gas networks, heat, communications, and transportation—has increased the importance of electricity in modern economic development. Electricity is a key cornerstone of modern economies. As the second installment of Quadrennial Energy Review notes (U.S. Department of Energy, 2017b):

Electricity is essential for supporting and sustaining industrial output, government, emergency services, interdependent critical infrastructures, and the U.S. national security apparatus. These critical infrastructures include physical and information infrastructures that are required for communications, transportation, and almost every other element of economic and social activity. Even though it is essential to the economy, lifeline networks, emergencies, and the national security apparatus, electricity—unlike oil—cannot be stored at scale.

104

With renewables and natural gas gaining prominence in the energy mix of many countries, their interactions with other sectors has become significant in the provision of affordable, accessible, and reliable electricity supply. Modernization of energy infrastructures—through the growth of clean, smart, and resilient systems and services—is enabling new architectures to stimulate new economic transactions, greater energy efficiency and new consumer services, thereby creating demand for an enhanced integrated grid that can facilitate this transition.

3.1.9 Energy Security, Resilience, and Reliability Objectives

Energy security, resilience, and reliability concerns—including, positive and negative externalities of shale gas development in the United States (such as low natural gas prices relative to other fuels)—continue to preoccupy local and national discourses (Growitsch and Stronzik, 2014; Weiss et al., 2013). The U.S. natural gas market has a long history of deregulation, regulation, re-regulation, and cost-based ratemaking design associated with interstate gas pipeline transportation and end-user tariffs. The Natural Gas Policy Act (NGPA) of 1978 established a framework for deregulating interstate wellhead prices. The NGPA also unified two-tier pricing that created dual markets between federally regulated and intrastate natural gas (Wang and Krupnick, 2015, p. 7). The signing of the Natural Gas Wellhead Decontrol Act in July of 1989 ended the remaining price controls under NGPA (as gas supply contracts expired or were renegotiated or by January 1993, whichever came first). Significant change in the industry has also resulted from the enactment of the Public Regulatory Policies Act (PURPA) of 1978 and repeal of certain sections of the Power Plant and Industrial Fuel Use Act (FUA) of 1978.

105

Under FUA, restrictions were lifted for major fuel-burning facilities, such as those of industrial customers; hence, existing power plants could burn gas as a baseload fuel while new gas facilities were obligated to develop the capacity to use coal or another fuel. As a result, non-utility power generation grew significantly under PURPA, and applications for large-capacity plants increased. Added to these developments, in 1985, the U.S. government eased imports to increase competition in natural gas markets. At the federal level, the natural gas industry is regulated in several areas. FERC regulates pipeline rates and interstate pipeline construction. Rates approval covers payments for storage, rate design and services such as gas inventory charges. FERC regulates rate-based calculations, and states regulate end-user prices and gas pipeline construction, while the federal government regulates long-term import and export arrangements. The federal government regulates long-term import and export arrangements.

106

Figure 3.11 U.S. Natural Gas Proved Reserves and Marketed Production, 1971–2013

Historically, the U.S. natural gas market has grown alongside domestic gas production (Figure 3.11). In many cases, this growth has been associated with construction and expansion of the gas-pipeline network. As a result, the industry expanded rapidly from the 1920s to a mature market, reaching a 27% share of the total electricity generation in the United States in 2013 (EIA, 2014a). As of December, 2014, more than 10,000 small producers supplied nearly one-third of total gas production in the United States. Other major oil and gas companies supplied the rest— excluding imports (mainly from Canada and Mexico), which provided about 7% of the total supply. Broadman and Montgomery (1983), Joskow (1997, 2012) Costello

107

(2006), Fine et al. (2011), and Groat and Grimshaw (2012) have documented the changing regulatory policies regarding the unconventional natural gas market since the 1970s and the impact of these changes on reliability and grid development. These policies are the basis for the currently expanding market. Like the natural gas market, the growth of renewable energy is occurring in parallel with a mounting focus on energy security, resilience, reliability, and decarbonization of power systems objectives—objectives which have become more urgent since the Paris Agreement on climate change.

3.1.10 Climate Change and Environmental Concerns over Air Emissions

3.1.10.1 Impacts on Electricity Generation

The impacts of current and anticipated climate change on the energy sector are substantial. They affect electricity generation, transmission, and distribution, thereby threatening energy security, reliability, and power quality (Ward, 2013). Changes in temperature and precipitation patterns are altering the amount of river runoff—both in terms of seasonal patterns and total river flow—thereby impacting the effective generation capacity of hydroelectric power (Noreña et al., 2009). Temperature increase is modifying the solar energy-generation efficiency of photovoltaic cells, while increasing cloud cover reduces solar generating capacity (Cutforth and Judiesh, 2007; Bull et al., 2007). Increased intensity and frequency of storms is also affecting wind resources, with higher wind speeds but more variable wind patterns (Kaygusuz, 2009); and higher waves are disrupting production capacity and affecting the structural integrity of offshore wind turbines (Musial and Ram, 2010). Offshore oil and gas infrastructure and low-lying coastal facilities are vulnerable to weather disasters—

108

resulting from floods, drought, and sea-level rise expected to become more common and intense due to climate change (Karl et al., 2009), coastal erosion (Cayan et al., 2009), and other extreme events such as wave height caused by intense hurricanes (Energo, 2006) that compromise oil and gas production or even lead to shutdowns. In addition, changes in cultivation regimes such as biological productivity, incidence of plant pests and land use patterns are affecting biofuel and biomass generation as the temperature tolerance of cultivated biofuels is tested more regularly (Persson et al., 2009).

The efficiency of thermoelectric power plants—including fossil-fuel-fired plants, biomass-fired plants, and nuclear power facilities—depends on significant volumes of cooling water. Changing climate parameters such as ambient air and water temperature, pressure, flooding, and droughts could affect the technical efficiency of thermoelectric power plants and their reliability, and could also affect energy supplies from natural gas, geothermal, coal, biomass, and nuclear technologies (Hammer et al., 2011; UNDP, 2007). Higher water temperatures and decreased river runoff due to climate change could force power plants with open-loop cooling to scale down their operations, as demand surges for cooling water and peak power—as witnessed during the European heatwave in 2003 (Curtis, 2009: 429) and the near record heatwave in the United States in 2006 (Schwartz et al., 2008, p. 20). Additionally, temperature change could alter the conversion efficiency of concentrated solar-power generation, while higher water demand for solar‐thermal power, hydroelectric and nuclear-energy generation could create water stress in dryer climates (IPCC, 2011). Moreover, changes in crop productivity and disease distribution due to climate change could affect the availability of raw materials (for example, sugarcane bagasse) for thermal

109

generation, and biomass-based thermal-power generation in general (Schaeffer et al., 2008). Finally, due to concerns over greenhouse-gas emissions (GHG), there has been an increased push to accommodate both grid-tied and stand-alone renewable energy resources (such as wind turbines and PV arrays) via intermittent generation in existing grids (Vazquez et al., 2010). However, climate change could exacerbate intermittency and dispatchability concerns by lowering production efficiency and adding a significant amount of uncertainty to renewable energy output and planning decisions.

3.1.10.2 Impacts on Electricity Transmission and Distribution Electricity transmission and distribution lines could become energy sinks during a heatwave and could become less efficient because of the additional resistance induced. Increased resistance in the transmission infrastructure could also cause them to heat up and stretch, while low wind speeds and warmer ambient temperatures could prevent lines from cooling sufficiently, thereby increasing their sag and the potential for service disruption (Ackerman and Stanton, 2008, p. 12). In recent summers on the West Coast and Northeastern United States, large-scale blackouts have been attributed to transmission lines sagging in the heat (Godden and Kallies, 2012). Furthermore, severe climate disruptions—such as extreme winds, flooding, heat waves under climate change, excessive icing on overhead lines, lightning strikes, and meteorologically induced factors such as downed trees—could cause power transmission and distribution networks to fail and thereby increase the vulnerability of the existing network to frequent power outages (and millions of dollars in repair costs) (Rosenzweig et al., 2005; Montoya et al., 2013). As the aging electricity infrastructure needs replacement and system upgrades become necessary to meet peak demand, storm-caused power outages and direct

110

damage from climate-induced natural disasters (strong winds, heavy precipitation, high temperatures, and floods) could increase repair costs and traditional utility investments in the transmission and distribution assets (Peters et al., 2006). Furthermore, if changes in the hydrologic response of water basins intensify due to climate change, this could in some cases increase the vulnerability of the riparian infrastructure (such as power transmission systems and pipelines) to river siltation and other erosion processes (Ebinger and Vergara, 2011). Moreover, the increased frequency, duration, and spatial and temporal extent of storm-caused transmission- power outages might lead to prohibitively expensive insurance premiums. They might even cause commercial insurance carriers to stop writing such policies altogether, thereby causing significant economic impacts on the affected utilities and their customers (Johnson, 2010). Investment in mega-scale, regionally integrated transmission networks such as those planned under the Desertec Industrial Initiative in North Africa and the Middle

East (Lilliestam et a., 2011: 3380) or the Ivanpah solar facility in California’s Mojave Desert (Cardwell et al., 2014)—or even under large-scale nuclear, CCS, and other large-scale wind-technology-as-policy projects (Ansolabehere et al., 2003; Byrne et al., 2006, p. 9-11)—might also help to mitigate climate impacts.

3.1.10.3 Impacts on Electricity Demand and Consumption Increased cooling and heating demand in warmer and colder climates has a significant impact on energy demand, as “electricity demand is most sensitive to changes in summer climate, whereas heating-fuel demand is most sensitive to changes in the winter climate” (Hammer et al., 2011, p. 268). Therefore, when peak power demand exceeds available supply capacity during periods of hot weather, this might

111

result in power outages. Changing climate parameters such as warmer temperatures might increase energy demand for cooling and increase the number of cooling degree days (Wilbanks et al., 2007). Climate change impacts would vary across regions, with temperate regions experiencing reduced energy demand for heating while tropical areas experience increased energy consumption for cooling. Seasonal and regional shifts in electricity demand for heating and cooling therefore would likely affect electricity consumption, thereby leading to increased adoption of air conditioning and alternate fuel consumption (Scott and Huang, 2007). Rising temperatures and heatwaves could negatively affect the cooling efficiency of power plants and their production capacity due to above-normal water temperatures, which in turn might lead to higher water demand (Jowit and Espinoza,

2006; Hekkenberg et al., 2009). Generally, under a “no-regret” adaptation strategy, cooling demand could be offset by energy-efficiency options such as cooling towers and passive solar-building designs (Chan et al., 2010). Energy efficiency has proven to be a low-hanging fruit, because of the energy savings and cost-savings benefits derived from the projects—notwithstanding the accuracy of the impacts of climate- change projections (Mansanet-Bataller et al., 2008). Finally, future energy prices, increased economic activity, technology innovation, household income levels, and regional population changes might also significantly increase electricity demand, which could in turn increase demand for gas (Scott and Huang, 2007; Ruth and Lin, 2006).

112

3.2 Theoretical Framework

3.2.1 Business Model Definitions

The business model offers a valuable unit of analysis for innovation and can be positioned as an analytic tool to deconstruct options for the sustainable evolution and subsequent diffusion of business practice in the utility industry (Loock, 2012; Pätäri and Sinkkonen, 2014). There is no universally accepted definition of business model concept; however, several authors have proposed definitions in their publications.

Osterwalder, Pigneur, Clark, and Smith (2010) define a business model as “the rationale of how an organization creates, delivers, and captures value” while Chesbrough and Rosenbloom (2002) explain a business model as “the heuristic logic that connects technical potential with the realization of economic value.” Shafer et al. (2005) define a business model as “a representation of the underlining core logic and strategic choices for creating and capturing value within a value network.” Table 3.3 shows a sample of definitions uncovered by Zott et al. (2010). As an analytic tool, the business model concept has been widely used in analyzing investor preference (Loock,

2012), performance of an energy service company (ESCO) (Pätäri and Sinkkonen, 2014), micro-generation solutions (Provance, Donnelly, and Carayannis, 2011), the renewable energy market (Okkonen and Suhonen, 2010), energy-efficiency programs

(Behrangrad, 2015), evolution of energy utilities (Richter, 2012), and penetration of distributed generation (Funkhouser, Blackburn, Magee, and Rai, 2015). The business model concept has been widely tested in practice in the energy sector. Business model conceptualization is complex and can be approached from many directions.

113

Table 3.3 Selected Exemplars, Interpreting Business Models as Formal Conceptual Representations Author (Year) Definition Papers Citing the Definition Osterwalder, A business model is “the rationale of how an Miller, Richter, Pigneur, Clark, organization creates, delivers, and captures and O’Leary and Smith value” (2015); Richter (2010) (2013) Abdelkafi and A reinforcing feedback loop between the Newman, Chang, Täuscher (2016) created value to the customers, the value Walters, and Wills captured by the firm, and the value to the (2016) natural environment. Hamel (2000) “A Business Concept is a radical innovation López, Gutiérrez, that can lead to new customer value and and Gómez (2008) change the rules of the industry” (p. 66). [...]The business model is “nothing else than the business concept implemented in practice” (p. 66). Shafer et al. A business model is “a representation of the (2005) underlining core logic and strategic choices for creating and capturing value within a value network” (p. 202). Casadesus- “Business Model refers to the logic of the Pätäri and Masanell and firm, the way it operates and how it creates Sinkkonen (2014) Ricart (2010) value for its stakeholders” (p. 196)

In other words, various studies have analyzed electricity-market transformation in United States and have proposed key changes needed to align traditional electric- utility regulatory and business models with increased adoption of “disruptive technologies” such as solar PV and other renewables (Byrne and Taminiau, 2015; Byrne et al., 2015; Aggarwal and Harvey, 2013; Hanelt, 2013; Lehr, 2013; Newcomb,

114

Lacy, and Hansen, 2013; Wiedman and Beach, 2013; Costello and Hemphill, 2014; EPRI, 2014; Fox-Penner, 2010; Satchwell et al., 2011; Satchwell et al., 2014; Lacy et al., 2012; Graffy and Kihm, 2014; and Hanser and Van Horn, 2014). The majority of stakeholders in the utility industry are reevaluating existing regulatory and business models in the context of today's environment of increasing distributed-energy resource penetrations and other potent new pressures—particularly relatively flat utility sales growth and increased investment in DG. Osterwalder et al., (2010) definition of the business model has been widely tested in practice because the defining a business model is complex and can be approached from many directions. Common components of the “business model” (Table 3.4) include the value chain, value propositions, target markets, competitive strategy, revenue-generation models, customer interface, and a value network or infrastructure (Chesbrough, 2007; and Chesbrough and Rosenbloom, 2002).

Table 3.4 The Business Model Evaluation Business Model Pillar Description Customer interface Comprises the overall interaction with the customer. Consists of customer relationship, customer segments, and distribution channels. Value proposition What added values will the business offer for resource providers, project developers, technology vendors, community served, and other potential partners? Includes the bundle of products and services that creates value for the customer and allows the company to earn revenues. Infrastructure Describes the architecture of the company's value creation.

115

It includes assets, know how, and partnerships. Revenue model Represents the relationship between costs to produce the value proposition and the revenues that are generated by offering the value proposition to the customers.

Source: Osterwalder and Pigneur, 2012

The business model concept thus provides a valuable mechanism for structured deconstruction and analysis of the changing landscape of the utility industry (Loock, 2012). Figure 3.12 illustrates recent disruptive technologies and elements, thereby showing that the power sector has reached a “golden age” of utility reinvention—an inflection point where its future direction is much less predictable.

Source: Modified from Frantzis et al. (2008)

Figure 3.12 Future Market and Business Models

116

Utility-business models and regulatory paradigms in the U.S. electricity industry have been based on two fundamental characteristics: profit maximization and profit motivation (Satchwell and Cappers, 2015). In this regard, using the business model concept as an analytic device to build generic blueprints helps to deconstruct and understand core business functions (i.e., including policy, finance and market factors) (Boehnke and Wüstenhagen, 2007). Utility managers can use the business model to describe a spectrum of attributes for different regulatory approaches and design to implement, operate, control, and optimize their business (Boehnke and

Wüstenhagen, 2007; Johnson, 2010; Richter, 2013). As a management tool, a business model represents a profit-maximization strategy that is enabled by regulation or markets. It can also function as a blueprint for value addition and innovation (Baden- Fuller and Morgan, 2010; Fox-Penner, 2010). In this regard, profit-earning business activities that develop around utility programs—such as investment tax credits (ITCs), solar renewable energy credits (SRECs), state rebates, net metering, or interconnection standards—are business models. Table 3.5 depicts the evolution of solar PV business models.

117

Table 3.5 Evolution of Solar PV Business Models PV Supply (zero Third-party ownership Full integration (2nd generation) and operation (1st generation) generation) • Business models focus • Third parties drive • Business models on manufacturing, business models (i.e., promote integration of supply and installation as project developers PV electricity of PV systems and owners of PV generation in the supply systems, resulting in: and distribution • End-user is the owner infrastructure. o Reduced hassle • Utility is largely but complexity • A spectrum of passive, providing net for end-users. attributes for different metering and regulatory approaches. standards/simplified o Better access to interconnection, but finance options • Business models otherwise, unaffected. o Leveraging of emerge with variations current incentive of system (e.g., structure (e.g., for ownership, operation, commercial and control. building • Utilities become deeply applications) involved in PV utility- • Utility gradually takes scale generation on a facilitation role as • Solar PV products PV market share supply-chain becomes grows. “commoditized”

3.2.2 New Business Models and Innovation Business model innovation as a term remains largely unspecified in the current academic literature. Chesbrough (2010) notes that business model innovation is less a matter of superior foresight but of trial and error and ex-post adaptation. McGrath

(2010, p. 254) suggests that business model information entails “business model

118

experimentation,” while Sosna et al. (2010) understand it as a strategic renewal mechanism for organizations facing changes in their external environment. In this study proposal, business model innovation is presented as the development of new organizational forms for the creation, delivery, and capture of value. Complicating matters, different electric utilities in the United States have different starting points, value propositions, customer expectations (across customer classes), priorities, and they vary with respect to state regulations, economic and market frameworks. And like any significant journey, utility business plans and models need the flexibility to adapt to new information and changes in regulations at all levels of government (Lightner and Widergren, 2010). How can utilities meet these conflicting expectations in an uncertain environment? Fox-Penner (2010) has offered a solution through a “two-and-a-half- business model” innovation as an alternative. The half refers to a smart integrator scenario in which the utility operating the power grid does not own or sell the power delivered by the grid (Figure 3.13). Consequently, the power-generation system and the power grid (including its information and control systems) are community-owned (e.g., a community micro-grid). The advantage of community-owned generation is in its potential for economies of scale, because hundreds to thousands of individual customers participate in the project by installing distributed-generation plants (e.g., solar PV or wind turbines). They use the set standards, but the financing and administration is handled separately (say, by a utility). First, the smart integrator has well-developed analytic capabilities to ensure that the grid can deliver all the power demanded. It also has a

“green dispatch mechanism” to determine when and how to use low-carbon energy

119

sources such as solar, wind, and hydroelectric power. The obligation of the utility to ensure that the local grid delivers the power demanded (or responds to restore power) remains the same. Second, the smart integrator has a “highly secure but maximally open platform for information, price, and control signals” (Fox-Penner, 2010, p. 176). This feature ensures that it responds strategically to the regulatory regimes in diverse markets by integrating information for accounting, billing, and settlement systems to accommodate the more complicated pricing and payment options offered.28 And finally, a smart integrator can monitor new technological developments and figure out how they can make alternative energy sources more viable.

28 The value proposition of a Smart Integrator is to maximize the value of its infrastructure (e.g., smart grid network) by opening it to all service providers. However, it is unclear how the traditional business model will need to change to motivate the utility to play this role

120

Source: Goldman et al., (2013)

Figure 3.13 Smart Integrator: Utility as Network Integrator

121

Related to the smart-integrator utility-business model is the energy-services utility (ESU). In the ESU29 business model (Figure 3.14), the utility moves away from a purely asset and commodity-driven entity to become a service and value-added enterprise in which profit achievement hinges on the services provided to customers (Lehr, 2013).

Source: Goldman et al., (2013)

Figure 3.14 Energy Service Utility (ESU) Business Model

29 ESU is a plausible extension of the Smart Integrator model. However, ESUs provide a wide array of energy services than provided under the traditional utilities model. Examples of the ESU business model include piloting programs by Arizona Public Service (the largest electric utility in Arizona) and Southern California Edison (SCE).

122

To realize the power-delivery-balancing-reliability maintenance role of the smart integrator, Fox-Penner (2010) argues that a practicable new utility-business model must consider the feasibility of creating different triads of structure, regulation, and business models to facilitate a transformation to a sustainable business-innovation future. This process comprises a variety of innovations: • joint construction of generation and transmission projects in which different segments cooperate to finance and build generation and transmission assets and then share in their ownership, operations, and

benefits;

• growth of diversified independent transmission companies; • diversified generation (e.g., natural gas and renewables) beyond the basics of the utility business (such as hybrid RE systems, poly- generation units fueled by different generation sources, or zero-net energy systems);

• use of subsidiaries to speed up clean energy diversification (e.g., NextEra (a subsidiary of Florida Power and Light Company, the nation’s largest wind-plant owner), and MidAmerican (a subsidiary of the holding company, Berkshire Hathaway, the largest utility-wind owner and a recent entrant in the wind- and solar-developer market); and

• increasing use of utility consortia that expand member utilities’ service offerings beyond the provision of electricity (e.g., utilities that provide a variety of services to cooperative customers).

123

In addition to the above, a practicable utility-business model must also consider the feasibility of creating different structure, regulations, and transition strategies to facilitate a transformation to a “smarter” future (Fox-Penner, 2010). By combining the business model concept with innovation and organization theory, Chesbrough (2010) identifies barriers to business model innovation as “novelty, lock- in complementarities and efficiency.” This approach is quite robust and constitutes a sound “unit of analysis” for this study, because innovation research has been applied in different industries to analyze the consequences of radical technological changes for incumbent firms (Richter, 2013). For instance, applying the concept of disruptive innovation and the theory of organizational ambidexterity through product differentiation (e.g., electricity services) and integration of services to position the industry for the future (Duncan, 1976; and Christensen, 2006). As Christensen and Bower (2009) point out, disruptive technologies are hardly employed directly in established markets such as in electric utilities because of concerns that they could destroy the value of existing competencies; they are instead positioned to transform the architecture of the market in the medium and long term. For utility companies, the problem of employing disruptive technologies lies not in the technology itself but rather in their inability to commercialize new technologies outside of their own current business model (Chesbrough, 2007). The challenge for utilities in providing value and services is in establishing exactly what their future business model will look like, because the utility industry is capital-intensive. Consequently, electric utilities that implement disruptive technological innovations also need to engage in business model innovation (Hansen, Große-Dunker, and Reichwald, 2009).

124

3.2.3 Transformation Through Shared Visions Rising distributed-energy generation offers a promising large-scale transition to a low-carbon power sector, a cleaner economy, and continued transformation of the utility systems (Tan, Hassan, Majid, and Abdul Rahman, 2013). With increased expansion of recoverable gas reserves, the contribution of natural gas to low-carbon energy transition in the United States is huge and has steadily grown. A key benefit of the expanding role of natural gas and investment in related gas infrastructure in the United States is in its relatively low emissions profile compared to other hydrocarbons. Given the growing needs for decarbonization in the power sector and abundant, reliable, and clean energy sources, how can more compelling policy initiatives, business models, technology solutions, and innovative financing mechanisms be created so that these two vital forms of energy work in greater concert together? What barriers hinder the emergence of new integrated business models and utility rate structures, such as joint transmission investments, colocation, and NG-RE blended power generation and deployment? Plenty of RE sources such as solar and wind can be converted into electricity by using technologies that emit no greenhouse gases. Additionally, natural gas offers a cleaner alternative to coal or petroleum with the potential to deliver immediate, significant reductions in carbon-dioxide emissions from the U.S. power sector—if newly discovered economically recoverable gas reserves (mostly in North America) can be produced responsibly. The appeal of the synergies of RE and NG goes beyond carbon mitigation in the power sector; these synergies increase energy security and energy independence, contribute to economic growth, and lead to a diversified resource base. To date, however, the challenge of investing in RE sources is that they are dispersed, intermittent, and non-dispatchable. On the other hand, when it comes to

125

natural gas, electric utility managers and regulators worry that increased gas consumption could lead to overreliance on natural gas and potential lock-in (Fox- Penner, 2010). A significant switch to distributed generation systems also requires substantial initial capital investment in infrastructure, technical innovations, and transaction costs. In the context of the ongoing transformation of the power system, a repositioning of policy-finance-market mechanism is required towards distributed generation (Miller et al., 2015). The industry also needs to adopt flexible and innovative business models that capitalize on both RE and NG to optimize the benefits of both forms of energy: access to new revenue streams, wholesale power market opportunities, and energy services that offer the customer resiliency, reliability, and reduced costs (Lee et al., 2012). To manage this process of transformation in the power sector toward a more sustainable production future, seven opportunities exist for the natural gas and renewable energy partnership to jointly develop vibrant and robust hubs of integrated research and development, information exchange, planning, and decision-making (Cochran, Zinaman, Logan, and Arent, 2014; Lee, Zinaman, and Logan, 2012; Lee et al., 2012; Zinaman et al., 2015): • optimized long-term and cross-sectoral utilization of RE and natural gas resources through joint research aimed to ascertain which industry and technology pathways represent opportunities for better utilization of the country’s diverse energy resources across sectors and time scales;

• development of hybrid technologies (e.g., hybrid concentrating solar power and natural-gas-fired power-generation systems, biogas and natural gas co-

126

fired combined-cycle gas turbines, natural gas-powered compressed-air energy storage) to capture the respective benefits and to minimize the drawbacks of individual technologies;

• electricity market design; • enhanced quantitative tools and models to better incorporate cross-sectoral impacts, particularly those arising from natural gas;

• more public-policy dialogue and analysis aimed to better understand the potential roles of natural gas and RE in enhancing energy diversity, economic

prosperity, and climate change mitigation;

• pursuing a portfolio approach to support RE and energy-efficiency complementarity opportunities to decarbonize the energy sector; and

• joint initiative collaboration between natural gas and RE to dispel popular myths and inaccurate beliefs about each industry.

Despite competition that has been associated with both natural gas and distributed-energy resources, there exists a strong complimentary relationship between the two energy resources (Weiss, Bishop, Fox-Penner, and Shavel, 2013). From a public-policy perspective, O'Loughlin et al. (2012) detail the growing desirability of natural gas for electric generation in the United States: Natural gas is a relatively cleaner, low carbon fuel, is abundant and inexpensive, and is a conveniently reliable and low-cost energy source for peaking generation by utilities to support integration of renewables such as solar and wind. However, heavy reliance on gas could expose power customers to price shocks when gas prices spike. For this reason, there is need for a more consistent framework for depicting utility-profit maximization and profit

127

achievement if we are to align public policy, regulatory policy, and utility-business model objectives (Satchwell and Cappers, 2015). Formulation of new business models for RE resources that are aimed to help them reach step-change milestones in key disruptive technologies—grid parity of solar DG, lower cost and mass-scale storage solutions, vibrant and secure micro-grids, and ubiquitous behind-the-meter devices—has been addressed by several studies (See, e.g., (Frantzis, Graham, Katofsky, and Sawyer, 2008; Richter, 2012). Two generic business models have been identified from these studies. These have been classified into two broad categories: utility-side, and customer-side business models.

3.2.4 Utility-Centric Business Model Recent studies and energy-market surveys continue to show that renewable energy has the greatest potential for disrupting the current energy system in the United States (Richter, 2013). This position is supported by demand for a utility-side business models for DG, which requires power utilities to find new market-design approaches to the production, transmission and sale of electricity derived from reliable and cleaner energy sources (Small, 2010). For example, Richter (2013) sees a utility-side RE business model comprising “large-scale projects with a capacity between one and some hundred megawatts” and considers the principal technologies for this application to include “on- and offshore wind energy, large-scale photovoltaic systems, biomass and biogas plants, and large-scale solar thermal energy like concentrated solar power.” Byrne et al. (2016) notes that the linchpin for large-scale deployment of RE technologies is an economically sustainable business model. The value proposition in the utility-side business model for RE is the bulk generation of electricity (Nimmons and Taylor, 2008). The utility sector provides

128

essential services to consumers, including grid stability, resilience, support for DG, and sustaining tension gaps. From a utility’s point of view, the advantages of distributed generation in the U.S. include the following: more flexibility to site and operate than fossil-based technologies, distributed generation benefits from federal and state subsidies, and energy-efficiency incentives which reduce the upfront capital and operations and maintenance (O&M) costs (Sioshansi, 2016). Thus, in the long- term, investment in distributed generation provides better policy incentives for utilities. For example, the customer interface of distributed-generation projects consists of power-purchase agreements on a business-to-business level rather than on a relationship with the end user. Like most conventional power plants, electricity production and distribution from these distributed generation sources follows a similar logic—although the latter usually has less generation capacity. In this regard, electric utilities have two principal concerns when it comes to providing electricity services to the end-customer. First, the electricity must get to the customer reliably and safely. Second, power must be delivered efficiently for the utility to remain profitable. This puts pressure on the utility to minimize electrical grid losses during the transfer from the generation site to the customer. Utility-side business models, concepts, components, technologies, and systems must take these factors into consideration. The question then becomes the following: As the percentage of end-customers generating their own electricity from clean energy sources increases, what key policy, market and business concerns should utilities be aware of? An increase in the number of utility customers poses key salient challenges to the U.S. power grid:

129

• optimal deployment of expensive assets, • need for diversification of generation, • demand response, • grid stability, • grid modernization, grid access to address intermittency and non- dispatchability concerns,

• tariff implementation, • fair equalization of costs associated with the renewable electricity generation across the country,

• revenue-decoupling strategies, and • geographical dispersion and a fragmented, distributed generation market.

Some of these challenges can be addressed through deployment of ‘smart’ technologies at the utility-side to control and monitor grid operations, billing, and tariff management. However, utilities first need to identify in which part of the value chain in Figure 3.15 they should engage before developing new business models to respond effectively to the challenges above. The result is a more integrated electric utility system between the end-consumers and electric generators.

130

Source: Adapted from Frantzis et al. (2008)

Figure 3.15 Two Generic Utility Business Models in the Electricity Value Chain

To sum up, there are four basic business model elements on the utility-side: value proposition, customer interface, infrastructure, and revenue model (Table 3.6). However, large-scale deployment of renewable energies is changing the traditional electricity value chain. It is changing how electricity is produced, distributed, and sold to customers. The economic implications of these changes and other new energy technologies on the utility-side must deal with their potential for disruptions at every step of the conventional electric-energy value chain.

131

3.2.5 Customer-Centric Business Model Growing demand for affordable, reliable, domestically sourced, and low- carbon electricity driven in part by evolving public policy priorities has led to an increasing share of variable RE composition in electricity generation (Cochran, 2013). As a result, new opportunities for addressing the variability of renewables are being reviewed and strengthened, particularly at the customer-side, through business model and technological innovations that enable dispatchable demand response and DG. However, market design remains a primal challenge in merging these opportunities to create incentives and compensate providers justly for attributes and performance that guarantee a reliable and secure electric grid system. Cochran et al. (2013) examine a suite of wholesale power-market designs currently in use on the customer-side to safeguard the reliability of electricity supply, security, and flexibility in a landscape of significantly variable RE. They also assess various considerations required to ensure that an inclusive wholesale market designs cognizant of emerging technologies, such as demand response, DG, and distributed storage. These technologies play a substantial part in meeting electricity needs, meeting greenhouse gas reduction targets, and increasing energy security (Amor et al., 2010). As third-party ownership of these RE sources increases, utilities could lose their market share and profit margins, thereby necessitating reconsideration of their role in the electricity value chain. Figure 3.16 presents an example of a business model that is focused on strengthening system ownership and control as the level of utility involvement and complexity in management of resources increases.

132

Source: Modified from Frantzis et al. (2008)

Figure 3.16 New PV Business Models Focused on System Ownership and Control

One way in which the threat of revenue erosion could be turned into an opportunity to create and capture value for utilities is through customer-side RE business models. This business model encompasses distributed generation from sources such as solar PV, micro-wind turbines, and micro-combined heat and gas- power systems by using small-scale systems that are located close to the point of consumption (Onovwiona and Ugursal, 2006). There are various forms of distributed generation. In the context of private homes, distributed electricity generation is also referred to as residential generation. In corporate settings, the business model is

133

described as “contracting.” In this study, however, the customer-side business model refers to small-scale systems for private customers and small- to medium-sized business in the range of a few kilowatts to about 5 MW. Consequently, because of differences in the capacity of the customer-side business model compared to that of utility-side projects, the two models follow a very different logic in the value chain.

Table 3.6 Utility-side Vs. Customer-side Business Model Customer-side business model Utility-side business model Better customer relationship needed Relationship towards customers to develop new value propositions. remains mainly unchanged.

Changes in customer segments. Customer segmentation leads to New channels are needed. increased customer base and “eco” price premium earnings.

nterface Customer hosts energy generation system and shares the benefits with Channels mainly remain the same the utility. Electricity as commodity.

Customer I Customer Long-term customer relationship. Customer does not host energy generation systems. Customer pays per unit. Shift from commodity delivery to Bulk generation of electricity energy service provider. supplied to the grid—same value New value propositions needed for proposition.

roposition the market. Additional energy related services and value for the customer through more environmentally friendly Value P Value production.

134

Large number of small-scale assets. Small number of large-scale assets. Generation close to point of Centralized generation.

consumption. Experienced in large-scale No experience with development infrastructure projects. and operation of small-scale Partnerships with project developers projects.

Infrastructure and suppliers. Partnerships with system suppliers and local installers. Revenue from direct use, feed-in Revenues through feed-in of and/or from services. electricity.

High transaction costs reduce profit Economies of scale from large

odel margins. projects and project portfolios. New revenue models are needed. Revenue models are available. Electricity cost structure becomes Electric cost structures are in favor

Revenue M Revenue more complex due to many small of utilities experiences with large- investments instead of few large scale infrastructure financing. investments.

Source: Frantzis et al. (2008); Richter (2013).

Table 3.5 shows that both utility-side and customer-side business models follow different logics of value creation. The latter is based on many small projects, while the former is centered on a small number of large projects. Hence these models present different challenges and opportunities for the development of utility models. A customer-side business model30 (Figure 3.17) works best where high penetration of

30 In this model, the utility still pays for value-added products and services, and then recovers these costs through traditional rate-making proceedings. Source: (Frantzis, Graham, Katofsky, and Sawyer, 2008).

135

distributed generation or aggressive demand response pose serious grid-control and operations issues.

Source: Modified from Frantzis et al. (2008).

Figure 3.17 Customer-owned and Utility-controlled Value Network

Unlocking greater distributed generation value requires a business model that involves utilities directly in the ownership and control of assets and in the monetization of value of these assets, as illustrated in Figure 3.18. In this arrangement, utilities continue to execute their core competency roles: i.e., asset ownership and operation (Frantzis et al., 2008) In the case of states with RPS and a solar “carve-out,”

136

the SRECs would go to the utility because they have greater control and can bundle these SRECs into a green pricing program and sell them to other parties.31

Source: Adapted from Frantzis et al. (2008)

Figure 3.18 Utility-controlled and -owned Value Network

31 Due to additional policy, finance and market changes required to permit utility control and ownership, this business model may evolve more slowly than Fig. 4. It also requires distributed generation to exist on a sufficient scale for utility control and ownership to have significant impact.

137

3.2.4.1 The Sustainable Energy Utility (SEU) The Sustainable Energy Utility (SEU) was developed by the University of Delaware’s Center for Energy and Environmental Policy (Byrne, Martinez, and Ruggero, 2009). As a business model, SEU offers a fundamental reorientation of sustainable-energy business practices. It is a community-based model of development that is constructed around energy conservation and community-scale renewables. It aims to permanently lower the use of source materials, water, and energy to address concerns about climate change, rising energy prices, inequity of energy availability, and a lack of community governance of energy development (Byrne et al., 2009). The State of Delaware and the District of Columbia were the first to implement SEU in the United States. New versions of the model have been implemented by Sonoma County in California and the Commonwealth of Pennsylvania. Assessment of the model for application in international markets has been carried out in the City of Seoul (South Korea)32 and the City of Thane (India) (Gopal, 2013). The SEU business model expands beyond the traditional utility perspective to offer a “total solution” that matches performance to customer needs. It differentiates itself strongly from conventional utility products by establishing a practical and creative capitalization strategy for the negawatt-hour33 as opposed to the kilowatt- hour. The dimension of differentiation, therefore, is that the SEU strategy considers

32 Archives of the Seoul International Energy Conference available at: http://env.seoul.go.kr/archives/34476

33 Amory Lovins coined the term, “negawatt” in 1989 to capture the unit of energy that is not used (Lovins and Lovins, 1982). The term is still being used today and is still alive in current references to energy efficiency as a resource. The fundamental idea behind the negawatt idea is that energy efficiency can displace supply-side resources, investments in energy efficiency should be integrated into utility resource planning.

138

“the use value of a product, instead of its exchange value, as the source of its added value, therefore selling function-based services instead of products” (Steinberger, van Niel, and Bourg, 2009, p. 368). A key component of this capitalization strategy is converting the thoroughly accepted “energy efficiency gap” – i.e., the difference between what is a socially desirable level of energy efficiency investment and what is actually observed notion that not using energy is much cheaper than using energy and that the very large potential for energy use reductions is not being exploited (Sioshansi, 2013) – into an accessible value for its customers. The SEU strategy has several key competencies, including skill, knowledge, and the ability to creatively unlock the vast energy efficiency potential of the utility industry. In this regard, the SEU strategy seeks to: a) solve the problem of navigating the legal, financial, and policy labyrinth to unlock existing energy-savings potential (Byrne and Kurdgelashvilli, 2003); b) raise sufficient levels of capital investment by pooling together the pledged monetary savings from program participants and selling this resource on the capital markets to invest in energy efficiency intervention measures (Byrne and Taminiau, 2016). “By 2020, Delaware’s SEU anticipates that up to 93 percent of its revenue ($56.2 million) will be self-sustaining through shared savings

agreements and REC sales.” (Sioshansi, 2013, p. 536); and c) provide a range of contractual engagements that raise investor confidence, guarantee energy and monetary savings, pre-quality local and national ESCOs, standardize and facilitate pooled financing, and arrange detailed monitoring and verification that serves a diagnostic function.

139

3.3 Hamel Business Model Theory This section discusses an example of a business model for a market-based reform-view that contextualizes contemporary failures of a lock-in in one-dimensional view of the traditional model. It aligns business practices with an evolving set of regulatory and policy practices. There are many competing views about the future of the electric industry. Distributed resources and the transactive grid are described with words like distributed system operator, decentralized, disruptive, and distribution marginal pricing. As a result, it is not surprising that most industry futurists have noticed growing support for a new or revised demand-responsive, climate-friendly, information-centered, electric-system design paradigm. Among many others, business model innovation (BMI) is one candidate paradigm. According to Gary Hamel (2000) business model innovation:

Is the capacity to imagine dramatically different business concepts or dramatically new ways of differentiating existing business concepts (2000, p. 66).

In other words, business model innovation is meta-innovation, in the sense that it transforms the basis for competition by introducing more strategic variety into an industry or domain. As a construct, a business model provides a coherent framework for describing a firm’s organizational and financial architecture. Chesbrough and Rosenbloom (2002) explain that a business model can thus be imagined as “a mediator between a technology and economic value creation” to improve rather than jeopardize operations such as grid reliability. The landscape of the electric industry is changing very quickly due to new and rapid integration of intelligent end-use device technologies, increased proliferation of distributed-energy resources, growth in bulk-

140

power renewable generation, and increased prosumer participation. In this time of rapid and dramatic change, practitioners in the electric industry are opting for new utility-business models and rate-making and regulatory framework to visualize and construct a road map for mitigating operational challenges at both bulk-power and distribution levels to the mutual benefit of utilities, prosumers, and system operators. Business models in this regard start from the premise that the only way to out-do, out- smart and out-innovate competitors is to construct a business model so unlike what has come before that it leaves traditional competitors in a “gut-wrenching quandary” (Hamel, 2000: 69). In this context, business model innovation appears to be a new paradigm, as it is entirely about variety and targets all related components of the business model. Therefore, at the heart of a high-performance innovation system lies a capacity to “first identify, then deconstruct and reconstruct business models” to conform to the core values of the prevailing business model innovation paradigm in critical aspects—such as distribution-level locational pricing or locational marginal pricing and distribution (LMP+D) and transactive energy markets. It appears that contemporary scholars are noticing a sharp difference in the relationship between business model and business strategy that is obscuring a common emphasis on the importance of added value to value capture. Amit and Zott (2001) explain that a careful business model design can help firm managers to capture value, because business models may facilitate transactions in novel ways that do, in fact, end up creating added value. In contrast, others like Magretta (2002) believe that consideration of business models cannot help managers capture value, because while the business model ensures value creation, it does not necessarily create added value (i.e., something that competitors are not offering). As Hall and Roelich (2016) note in

141

their analysis of new business models which rely on more complex value propositions than the incumbent utility model:

A geographically constrained utility has a finite customer base, were that customer base to adopt deep retrofit, there would be few opportunities to compensate the loss in value capture by expanding market share. As such, the municipal utility model is compatible with better PPAs for generators, incentivizing demand side services, and relocalising energy value, but may fail to drive significant energy efficiency gains (p. 294).

A firm’s business model is acutely germane to its ability to capture value, because it is through its business model that the enterprise exercises its bargaining ability. To create and capture this value, it is vital for utilities to define clear strategies for operationalization of transaction cost economics. Ronald H. Coase and Oliver Williamson offer pertinent insights. As the degree of market uncertainty increases, the utility industry must devise new business models and organization of transaction costs, as Williamson (1979, p. 254) explains:

Whenever investments are idiosyncratic in nontrivial degree, increasing the degree of uncertainty makes it more imperative that the parties devise a machinery to “work things out”-since contractual gaps will be larger and the occasions for sequential adaptations will increase in number and importance as the degree of uncertainty increases. This has special relevance for the organization of transactions with mixed investment attributes. Two possibilities exist. One would be to sacrifice valued design features in favor of a more standardized good or service. Market governance would then apply. The second would be to preserve the design but surround the transaction with an elaborated governance apparatus, thereby facilitating more effective adaptive, sequential decision making.

142

On the contrary, Coase views these transaction costs as real costs confronting relevant economic agents. Coase ([1938] 1973, p. 116) acknowledges difficulties in recognizing and calculating these transaction costs:

It goes without saying that within the business organization information must be made available which enables these variations to be estimated. Before tackling the practical problem of how this information is to be obtained and presented, there are certain analytical difficulties which need to be faced. These difficulties centre around the fact that costs and receipts cannot be expressed unambiguously in money terms since courses of action may have advantages and disadvantages which are not monetary in character, because of the existence of uncertainty and also because of differences in the point of time at which payments are made and receipts obtained.

This means that business model design should not focus solely on products or technology as the starting point of innovation. Rather, it should encapsulate the entire process on how a firm will deliver benefits to customers and how to capture a portion of the value of these benefits delivered (Chesbrough and Rosenbloom, 2002; Taminiau et al., 2017; Carayannis et al., 2014). The key to designing a successful business model therefore lies in achieving clarity about value proposition and value capture.

Boons and Lüdeke-Freund (2013), Hall and Roelich (2016) and Teece (2010) examined the relevance of business model innovation to create and capture value in nine building blocks: customer value proposition, customer relationships, key partners, key activities, key resources, channels, customer segments, cost structure, and revenue stream (equivalent to value capture). As Burger and Weinmann (2016) illustrates:

Value creation of the Utility 2.0 occurs in the context of integrated services. For example, the pure energy savings potential of a residential customer is so low that it hardly justifies any investment into smart devices. However, if the technology is combined with assisted living systems, for example, for elderly or disabled people, or with automated house protection systems against illegitimate intrusion and theft, it offers a benefit that becomes attractive for customer segments other

143

than technology-affine, wealthy early adopters, who would be the primary target group (p. 315).

The concept of business model innovation is used in this chapter to describe a more comprehensive approach to integrated resources planning that goes beyond incremental innovation. Research in this stream emphasizes the importance of learning and experimentation (Hayashi, 2009), creativity (Teece, 2010), analogical reasoning (Enkel and Mezger, 2013), and cognitive reframing in new business model generation

(Mikhalkina and Cabantous, 2015). First, Hamel’s original consideration of business concept innovation is explained. Next, it is demonstrated how Hamel’s notion of business concept innovation applies to the electricity market by offering a definition of business model in the utility industry (Chesbrough and Rosenbloom, 2002; Loock, 2012; Zott and Amit, 2008). Criteria are suggested for differentiating between the major components of the business model and those of the subcomponents. Finally,

Hamel’s concept related to the emerging ecosystem of energy economics for NG-RE hybrids summarized as the “uberization of energy”—with connotations of a physical distribution and market “platform model” that integrates buyers, suppliers, and the market, with expected synergies accruing to the platform providers and participants.

3.3.1 Characteristics of Hamel Business Model Concept Hamel’s (2000) comprehensive business model framework is used here to evaluate a prominent utility-business model in the United States—New York’s REV model—to illustrate potential changes that await the energy utility actors. Applying a wrong business model will likely result in modest success even with good governance and strong leadership (Teece, 2007). Chesbrough and Rosenbloom (2002) identified six key functions of a business model: namely, value proposition, revenue generation

144

mechanism(s), value chain, value network, target market, and a competitive strategy (see also Chesbrough, 2007). To identify and examine key aspects of moving towards a service-based business model and the connections with service innovation and new service development, Kindström (2010), for example, adopted this classification while Shafer et al. (2005) categorized the four often-cited business model components: strategic choices, creating value, capturing value, and the value network. Hamel’s (2000) framework, which is applied here, incorporates most of these general features, and therefore provides a holistic perspective from which to study the U.S. electricity market and, in particular, the New York utility-revenue model (Figure 3.19).

Source: Modified from Hamel (2000, p. 94)

Figure 3.19 Components of Hamel Framework for Business Concept Innovation

145

3.3.2 Four Major Components of Hamel Business Model Framework

Hamel’s business model34 comprises four major components (i.e., core strategy, strategic resources, customer interface, and value network), three bridge components (i.e., customer benefits, configuration, and company boundaries), and several factors that determine the profit potential (i.e., efficiency, uniqueness, fit, and profit boundaries). The first major component, a core strategy, is the essence of how a firm chooses to compete, whereas the sub-element, the business mission, captures the overall objective of the strategy or what the business model is designed to accomplish or deliver. According to Hamel, the business’ mission defines a firm’s decisions such as the “value proposition,” “strategic intent,” “purpose,” “audacious goals,” and overall performance objectives of the firm. When a company changes its business mission, this does not necessarily result in business-concept innovation. The product/market scope, Hamel says, defines where the firm competes (i.e., the firm’s competitive arena). For instance, it determines its customers, geographies, and product segments. In this regard, a firm’s definition of product/market scope can be a source of business-concept innovation—especially when it is entirely different from that of traditional competitors (Hamel, 2000). Finally, a basis for differentiation captures how the firm competes differently from its competitors. For instance, a firm seeks answers to questions such as the following: How have competitors differentiated

34 The Hamel’s framework provides that for a firm to be an industry revolutionary, it needs to develop a “wealth potential” i.e., strategy on how it’s going to inject innovation into each component of the business model. There are four factors needed to determine the wealth potential of any business concept: extent to which the business concept is efficient in delivering customer benefits, unique, degree of fit among the elements of the business connect, and extent it exploits profit boosters that have the potential to generate above-average returns (Hamel, 2000).

146

themselves in the U.S. electricity market (e.g., in designing new utility revenue models such as through platform service revenues, rate design, and customer energy data usage)? Are there other dimensions of market-oriented revenue model differentiations that could be explored? In what aspects of the energy service (e.g., rate design) has there been the least differentiation? How could differentiation be increased in some of these dimensions (e.g., by implementing opt-in rate initiatives such as time-of-use rates or smart home rates)? And have differentiation opportunities been diligently sought in every dimension of the business model? Hamel’s second major component, strategic or unique firm-specific resources, constitutes a source of competitive advantage. Fundamentally transforming the resource base (say through increased renewable electricity generation or increased and responsible exploitation of new natural gas resources) for competition can be a source of business-concept innovation. A successful business model thus creates its own intellectual hegemony. Success turns a business model into the business model. Strategic resources embody core competences—in other words what the firm knows— and comprise skills and unique capabilities. Strategic assets comprise what the firm owns. They are rare and valuable things other than know-how and can include brand, patents, infrastructure, proprietary standards, and customer data. A prudent firm-wide use of strategic assets can lead to business-concept innovation. According to Amit and Schoemaker (1993), asymmetry in the resources a firm controls and discretionary managerial decisions about resource development and deployment can be sources of sustainable economic rent. On the other hand, core processes illustrate what people in the firm do: They are methodologies and routines used in translating competences, assets and other inputs into value for customers. A reconfiguration of central

147

components and core processes in the business model can be a basis for business- concept innovation Zhao, Pan, and Lu, 2016). The third major component is the customer interface. It comprises four elements: 1) fulfilment and support which relates to market access (i.e., how the firm reaches the market and it includes channels, customer support, and service levels); 2) information and insight which refer to the knowledge that is collected from

customers and utilized for their benefit. It also refers to the ability of an organization to extract insights from this information—insights that can be used to design new targeted products for customers. 3) relationship dynamics refers to the nature of interaction between the firm (producer) and the customers. For instance, how easy is it for the customer to interact with the firm, and what feelings do these exchanges invoke on the part of the customer? 4) Finally, the pricing structure specifies the options for charging the customers for the services rendered (i.e., flat-rate charges or charges for time e.g., TOU).

The fourth component of the business model is the value network that surrounds the firm and includes suppliers, partners, and coalitions that complement and strengthen the organization’s own resources. Suppliers typically reside “up the value chain” from the producer (Hamel, 2000). The configuration of activities is a bridge component that links an organizations’ core strategy with its strategic resources. Configuration specifies both unique ways in which competences, assets, and processes

148

are combined and interrelated to support a chosen strategy and how those linkages are managed. Intermediating between the core strategy and customer interface is another bridge component—the customer benefits—which comprises the bundle of benefits that is essentially being offered to the customer. Company boundaries refers to the decisions that have been made regarding what the firm does internally from what it contracts out to the value network. Four factors define the wealth potential of any business model. Efficiency determines the extent to which the value customers assign the benefits delivered exceeds their production costs. Uniqueness demonstrates the level of convergence among business models in terms of conception and execution in ways that are valued by customers: The greater the convergence among business models, the lower the potential for above-average profits. Fit means that all the elements of the business model are consistent and mutually reinforcing and that all the parts work together for the same end goal. Finally, profit booster(s) include increasing returns, competitor lock out, strategic economies, and strategic flexibility.

3.4 Towards a New DER Electricity Service Vision: Détente for Distributed Energy Resources Over the last decade, the concepts of business models and business model innovation have gained considerable interest in both the business press and the academic literature. In the context of research into the energy business, Giordano and Fulli (2012) analyze how new business models might leverage technological and business synergies, foster investments, and shift business value to electricity services in line with the goals of energy-conservation efficiency and sustainability. Pätäri and Sinkkonen (2014) analyze the viability of a business model for energy-service

149

companies (ESCOs) based on energy-performance contracting. Loock (2012) focuses on why a “customer intimacy” business model that proposes the best energy services within the renewable energy industry might succeed over business models that propose lowest price or best technology. Okkonen and Suhonen (2010) also apply the business model concept to the field of renewable energies, whereas Provance, Donnelly, and Carayannis (2011), analyze the roles of politico-institutional and socio- institutional dynamics in the choice of business models for the micro-generation energy industry. Bouncken and Fredrich (2016) evaluate the potential of business model innovation in alliances from a value architecture perspective for firm level financial performance and concluded that longer alliance duration decreases the chances of value capture. The bigger the firm size, the more business units and products hence better leverage of the new configuration of value chain activities through the alliance. Larger firms typically have greater visibility in the market and better bargaining power than smaller firms such that they can bargain for higher returns from the alliances. Finally, Richter (2012) reviews the current state of the literature on utility-side and customer-side business models in the context of utilities for renewable energies, including the advantages in terms of revenue potential and risk avoidance. Engelken et al. (2016) review the emerging field of research on business models for renewable energies in both developing and industrialized countries to foster the diffusion of profitable renewable energy businesses. Thus, it seems that business model and business model innovation concepts have been applied extensively within the field of renewable energy, but not within the field of distributed energy-generation (based on

150

renewable energy and some form of natural-gas-fired generation to provide backup to the variable production from wind and solar power plants). However, if not designed and managed properly, increased use of DERs can have a negative impact on existing distribution grids. These impacts can include voltage rise, cable thermal limits, and reverse power flow (Manfren, Caputo, and Costa, 2011; Morvaj, Evins, and Carmeliet, 2016). On the contrary, as explained earlier, distributed energy systems can have a positive impact on the grid if integration is properly managed, such as by decreased fossil-fuel usage, enhanced grid stability and protection against electric system failures through use of local renewables and optimal system operation. At the same time, a new wave of technologies developed to manage DER is presenting both opportunities and challenges for vendors and utilities. Navigant Research estimates that global DER management technology revenue will increase by 109% annually from $194.3 million in 2016 to nearly $2.1 billion in 2025, tripling from 124 GW to 373 GW. This projected growth of distributed energy systems presents economic, technical and environmental benefits and drawbacks.

151

Chapter 4

BLENDING FAST-FLEXING RENEWABLE GENERATION SYSTEMS AND FLEXIBLE NATURAL GAS TECHNOLOGIES FOR A CLEAN ELECTRICITY SYSTEM OF THE FUTURE

Following the introduction to theoretical and methodological issues in the first two chapters, this chapter provides a more detailed and comprehensive account of the role of natural-gas-fired power generation in supporting renewable energy investments and diffusion. The challenges of grid integration, with increased diffusion of variable renewable energy technologies and grid-connected devices, appliances, and building energy management systems have been analyzed with a view to improving policy integration, data collection, modeling, and engagement with stakeholders. Various studies have evaluated the cost-competitiveness of RE and NG power generation, such as those by Vasilis Fthenakis,35 Michael Hogan,36 April Lee, Owen Zinaman, Jeffrey Logan, Morgan Bazilian, Douglas Arent, and Robin L. Newmark (2012),37 and William E. Hefley and Yongsheng Wang (2015).38 Other studies with a focus on economics and electricity grid-infrastructure development—e.g., by Ignacio Pérez-

35 Fthenakis, V. (2015). Considering the total cost of electricity from sunlight and the alternatives. Proceedings of the IEEE, 103(3), 283-286.

36 Hogan, M. (2013). Aligning power markets to deliver value. The Electricity Journal, 26(8), 23-34.

37 Lee, A., Zinaman, O., Logan, J., Bazilian, M., Arent, D., and Newmark, R. L. (2012). Interactions, complementarities, and tensions at the nexus of natural gas and renewable energy. The Electricity Journal, 25(10), 38-48.

38 Hefley, W. E., and Wang, Y. (Eds.). (2015). Economics of unconventional shale gas development: Case studies and impacts. Cham: Springer International Publishing.

152

Arriaga39 and Peter Fox-Penner (in a book40 and a journal article41)—have addressed the need for upgrading critical infrastructure (i.e., the electricity grid) to mitigate the impacts of severe weather events and the considerable intermittency and non- dispatchability of the most cost-competitive renewable energy sources such as wind and solar. This chapter begins by shedding some light on how natural gas and renewable energy integration has historically been handled. This topic is not new to utility system operators and network agents because, unlike fossil-fuel sources, electricity generation from the most promising renewable energy sources is intermittent and non- dispatchable, and peak supply periods often do not coincide with peak demand.42 In the previous chapter, the analysis identified which factors are driving change in the electric power systems—especially the rapid growth of distributed systems, which are occurring against a backdrop of cost reductions and the ongoing transition to a more renewable and intermittent generation mix. Based on this analysis of the ten driving factors, a dynamic econometric model is developed to characterize the effect of

39 Pérez-Arriaga, I. (2011). Managing Large Scale Penetration of Intermittent Renewables. Cambridge, MA: MIT Energy Initiative Symposium. Massachusetts Institute of Technology.

40 Fox-Penner, P.S. (2010). Smart power: Climate change, the smart grid, and the future of electric utilities. Washington: Island Press.

41 Zarakas, W. P., Sergici, S., Bishop, H., Zahniser-Word, J., and Fox-Penner, P. (2014). Utility investments in resiliency: Balancing benefits with cost in an uncertain environment. The Electricity Journal, 27(5), 31-41.

42 Renewable energy options require a significant amount of backup capacity. Wind and solar power plants output varies between seasons; wind speeds vary significantly from day to day while cloud coverage affects solar output.

153

flexible natural gas capacity on fast-flexing renewable electricity generation capacity. The econometric model incorporates the flexible NG capacity; policy instruments, including renewable portfolio standards, feed-in systems, net metering, and controls variables such as electricity importation ration, energy integrity, etc.

4.1 Theoretical Approaches and Tools

4.1.1 Matching Demand and Supply Instantaneously

Power system operators are required to match demand and supply instantaneously at all times by dispatching generation, storage, or demand-response resources. For a successful outcome, network operators need to address three main issues: improved system operation, system-friendly renewable energy deployment, and investment in additional flexible resources. For instance, ensuring that system operations conform to well-established best practices through better coordination of information exchange between various power-system actors is a no-regret, low-cost option (Codognet, 2004). A conceptual view of the major relationships in the retail- energy ecosystem consists of DSO operations, wholesale market economics, DER economics, transmission and distribution investments, customer investments and cost and utility rates (Figure 4.1).43 Depending on the existing local conditions and relationships with concerned parties, this could impact the overall improvements in customer economics and policy effectiveness of economic tools (such as feed-in tariffs, DER incentives, real-time pricing, and operations of the DSO capacity

43 The thin blue arrows show the interrelationship between the variables at any given time while the red arrows indicate that one variable influences the rate of change of another variable.

154

markets) (Masiello and Aguero, 2016; Winkler et al., 2016). At higher distributed- energy penetration rates, the benefits of improved operations such as system-friendly renewable-deployment practices and optimal utilization of flexible resources (flexible generation, demand-side response, grid infrastructure and storage) are magnified and cost-effective (Pérez-Arriaga and Knittel, 2016).

Figure 4.1 Major Relationships in the Retail Energy Ecosystem

To date, however, back-up generation capacity has been provided by fossil- based electricity generation systems. Fossil-based power generation plants comprise fast-reacting natural-gas-fired technologies (flexible-baseload henceforth) and other baseload fossil-electricity generation systems. Baseload technologies, such as coal and

155

nuclear power plants and other low-efficiency generation technologies) have slow reaction times and high capital costs; thus, they do not take part in the balancing market because ramp constraints prevent them from responding at short notice. On the other hand, flexible-baseload technologies, including most gas-generation technologies (e.g., natural-gas-fired combined-cycle and natural-gas-fired combustion turbines) are associated with mid-merit order effect, low capital costs, modularity, and quick ramp-up times. They are therefore particularly suitable for meeting peak power demand and mitigating the variability of renewable energy-generation technologies.44 Variations in power output increase the need for flexible generation capacity in power systems. Even if this variation can sometimes be predicted, this does not eliminate the need for fast-ramping resources.

4.1.2 California ISO’s Duck Curve As discussed in the previous chapter, an important dimension of the DER ecosystem is its impact on wholesale market economics: i.e., on the high variability of some DERs (such as PVs) which make it difficult to correlate them exactly with peak load. As a result, increasing penetration of DERs in power markets such as California has created what is widely known as a “duck curve.” For example, in California, this is identified by relatively low net loads during 12:00–15:00 and relatively high net loads during 06:00–09:00 and 18:00–21:00 (CAISO, 2014). This has necessitated significant ramping of thermal generation in the evening in California, as is shown in

44 Variable renewable energy technologies such as solar and wind continue to experience significant growth, and are characterized by volatile, partially unpredictable, and mostly non-dispatchable power output.

156

Figure 4.2. This price-reduction effect, also known as the merit-order effect, is characterized by very high levels of variable renewable energy penetration in California (Gil and Lin, 2013; Woo et al., 2014).45 For instance, as generators with higher variable costs (thermal generators) are displaced by zero-variable cost- generation systems (renewable energy sources), wholesale market prices are instantaneously reduced during periods when solar or wind generation are available. As Woo et al. (2016) write:

Wholesale electricity prices are inherently volatile due to: (a) daily fuel-cost variations, especially for natural gas, which is widely used by combustion turbines (CT) and combined-cycle gas turbines (CCGT) in North America; (b) hourly weather-sensitive demands with intra-day and inter-day fluctuations, which must be met in real time by generation and transmission already in place; (c) planned and forced outages of electrical facilities; (d) hydro conditions for systems with significant hydro resources; (e) carbon-price fluctuations affecting thermal generation that uses fossil fuels; (f) transmission constraints that cause transmission congestion and generation re-dispatch; and (g) lumpy capacity additions that can only occur with long lead times (p. 300).

While the merit-order effect potentially benefits electricity consumers by reducing electricity prices and monthly bills, its impact on investment incentive for the natural-gas-fired combustion turbines and combined-cycle gas turbines can be negative because the attractiveness of these flexible-baseload technologies is greatly reduced by the development of fast-flexing generation supply (e.g., wind and solar PV

45 California generates substantial renewable energy because of its ambitious renewables energy programs (e.g., an RPS of 50% by 2020, feed-in tariff, and net energy metering)

157

systems), as Traber and Kemfert (2011) documented in a simulation study of the German electricity market (cf., the regression analyses of Woo et al., 2015, for California and Texas).

Source: CAISO, 2014

Figure 4.2 California’s Duck Curve Showing Steep Ramping Needs

Bravo et al. (2016), Fanelli, Maddalena, and Musti (2016), Hirth (2015), Karakaya (2016), Mai et al., (2014), Olkkonen and Syri (2016), Sinn (2016) have identified variability in generation as a significant barrier to the integration of both wind and solar resources. To date, given their relatively low penetration in the U.S. power sector, the integration of solar and wind sources has not required major changes

158

in system operations, electricity market design, and regulatory innovation. However, as renewable-generation buildout grows in key states such as California and New York under the adopted aggressive-renewables portfolio standards (RPS of 50% by 2030 in both states), drastic changes in the grid system are needed. As the penetration of solar and wind has increased in recent decades, peak-load generation technologies

(e.g., gas turbines) and other load-following “mid-merit” generation technologies (e.g., combined-cycle gas turbines), have been used to compensate for variability.

4.2 Integrated Assessment Models The economic literature on the electricity market, regulatory innovation and energy deployment has largely overlooked the issue of blending renewable energy and different fossil-based technologies (notably natural gas). A number of contributions focus on the key barriers and skewed regulatory incentives that presently impede renewable energy investment and deployment, and they offer a framework for regulatory and market reform that is based on a comprehensive system of efficient economic signals. Jenkins and Pérez-Arriaga (2017) investigate the fact that the increasing penetration of distributed resources creates greater uncertainty for regulators and electricity distribution utilities, and they consider how to address these challenges by using state-of-the-art regulatory tools designed to overcome information asymmetries, manage uncertainty, and align incentives for utilities to cost-effectively integrate distributed-energy resources. Jenner et al. (2013) show that deployment of solar PV has been driven by favorite feed-in tariff (FITs) policies. Lee and Zhong (2015) discuss how net-metering policy promotes small-scale renewable energy deployment, particularly for locally distributed power generation.

159

In a fast-rate, integrated assessment of price convergence of solar PV and natural gas market in the United States, Nyangon, Byrne, and Taminiau (2017) investigate the relationship between these two markets over the years 2007-2015 via an LCOE approach to monitor common trends and market integration. But this analysis does not touch upon the possible role of other generation technologies, especially baseload fossil share.46 The law of one price (LOP), which is recognized as the theoretical foundation for determining prices of homogeneous products traded in geographically separated markets, views the concept of an integrated market “as a situation in which natural gas prices are comparable across the country after accounting for transport costs.” It also implies that there is greater degree of market efficiency that allows for trade and free movement of consumers in one part of the country with suppliers or producers in other regions (Nyangon, Byrne, and Taminiau, 2017). However, while we conditioned our analysis on the share of renewable energy (solar PV) and natural gas generation (aggregate level), we did not distinguish between the roles of different natural-gas-fired and fossil-based technologies (flexible- baseload vs. other baseload) and other renewables (such as wind) to assess the implication of these technologies in an integrated fashion. Also, in this case, the analysis does not control for policy controls (such as RPS, tax incentives, certificates and FIT) and the possible interaction between investments in renewable energy and (fast-reacting flexible) baseload-generation technologies.

46 Nyangon, J., Byrne, J., and Taminiau, J. (2017). An assessment of price convergence between natural gas and solar photovoltaic in the U.S. electricity market. Wiley Interdisciplinary Reviews: Energy and Environment. Vol 6, Issue 3.

160

Additionally, because they do not control for the effect of regulatory policies (such as feed-in tariffs, utility quota obligations, net metering, obligations and mandates, or SRECs), these studies cannot provide any insights into the risks associated with different policy mechanisms. As Lee and Zhong (2015) wrote:

Policymakers enact fiscal and monetary policies that have direct or indirect and positive or negative effects on profitability. Therefore, risk management of the energy policy has been identified as one of essential elements of fiscal policies. National policies for promoting a promising renewable energy environment are mainly focused on regulatory policies, fiscal incentives and public financing. Although different countries enact different renewable energy policies, the majority of policy objectives aim to enhance the diversification of electricity generation mixes, to increase renewable energy involvement, to reduce reliance on fossil fuels, to enhance competitiveness of renewable energy sources, and to reduce carbon emissions or various combinations thereof. The rationale behind the policies is usually either to increase the return or decrease the risks for investors. In regard to risk reduction, policies can be grouped into policy de-risking instruments and financial de-risking instruments. Policy de-risking instruments aim to remove the underlying barriers that are the root causes of risks, while financial de-risking instruments aim to transfer the risks that investors present to public actors. As it is impossible to eliminate all risks through policy de-risking or to transfer all risk through financial de-risking, additional financial incentives are usually complemented to shift a commercially unattractive investment opportunity to a commercially attractive one. Therefore, an unstable energy policy would definitely affect the return and create uncertainties for investors (p. 782).

This analysis contributes to this strand of literature by exploring a blended relationship between renewable and other natural-gas-fired and fossil-based technologies (flexible-baseload vs. other baseload) by using capacity rather than consumption or production data. In this manner, the analysis captures, as purely as possible, market investment decisions and the effects of policy, weather conditions,

161

and technology-performance controls, since capacity informs investments without being confounded by these forces. The study also contributes to this strand of literature by providing the first state-level empirical analysis of the interaction of renewable-generation technologies with other flexible natural-gas-fired and other baseload-oriented technologies in the top-10 solar-producing states ranked by cumulative capacity installed through 2016 (Solar Energy Industries Association, 2016).47 In addition, the specifications of the dynamic econometric model developed in this study take into account a rich dynamic structure of investments in power generation over the years 2001-2016, thereby improving on the Phillips–Sul price- convergence test of solar PV and natural gas market integration applied in Nyangon,

Byrne, and Taminiau (2017). Furthermore, the dynamic empirical analysis improves the linear technological model of renewable energy diffusion, which is assumed in Popp, Hascic, and Medhi (2011). This study further contributes to the integrated- assessment modeling literature by providing insights into the historical interaction of renewable energy (particularly solar and wind capacity) and fossil-based generation technologies (natural gas), from a state-level bottom-up approach. It can inform both short-run and long-run calibration of such constraints and estimates.

47 Solar Energy Industries Association (SEIA) ranking of the top-10 solar-producing states in 2016 includes the number of megawatts installed per state and number of houses powered per megawatt of solar added. The rankings also include “remixed” based on the number of solar jobs, solar capacity installed, solar capacity per capita, and how much each state climbed in the overall rankings (2016).

162

4.3 Data Description, Statistics, and Assumptions As shown in Figure 4.3, the top-10 solar-generating states in 2016 were California, North Carolina, Arizona, Nevada, New Jersey, Utah, Massachusetts, Georgia, Texas, and New York. Together, these states account for a total of 35,026 MW cumulative capacity (Solar Energy Industries Association, 2016). The data sample includes these ten states for the period 2001 to 2016.48 Figures 4.4 and 4.5 show area-wide solar potential for portions of 50 states and Washington D.C. It shows the same for the top 10 cities with most solar potential in the United States, based on estimates provided by Google Project Sunroof (Conkling, 2017).

4.3.1 Load and Fast-flexing Renewable Electricity Data

The dependent variable of the dynamic econometric model is the percentage of net installed electricity capacity of renewables (solar and wind), which is represented

RENEWABLE as (!"#acity (i, t) ) over total electricity capacity ($%&al!"#acity',&) in state i, time t, which reflects the investment decision as purely as possible, expressed as follows:

>?@?ABCDE /01023450 7898:-.;<= (")*!"#"+'&,-. = (4.1) FG.8H 7898:-.;<=

48 These states also account for the majority of solar investment and production over the period considered.

163

Figure 4.3 Top-10 Solar Generating States in 2016

164

Figure 4.4 Project Sunroof County-Level Coverage for 2017

Source: Google Project Sunroof (Conkling, 2017)

Figure 4.5 Top 10 Cities with Most Solar Potential

165

Renewable energy and alternative-fuel technologies include hydroelectric power, geothermal, solar, wind, wood biomass, ethanol, biodiesel, and waste biomass according to the U.S. Energy Information Administration database (2017g). Mature technologies such as hydroelectric generation are excluded from the calculation of renewable energy capacity since it is a mature technology for which most of the natural endowment has already been exploited (Popp, Hascic, and Medhi, 2011). Moreover, hydroelectric facilities are also able to dispatch extremely quickly and are often used to meet peak demand. Therefore, hydroelectric facilities do not share the same characteristics and limitations as natural gas turbines (a common dispatchable source which can be ramped up in minutes) and renewable energy technology such as solar thermal power plants (which can be designed to be dispatchable on roughly equivalent timeframes to natural gas turbines). Besides, since the main argument of this study is that there are distinctive complementary benefits of a renewables-gas blended model (e.g., fast-start capabilities for natural gas; low price volatility for solar; flexible grid, etc.), biomass and other wood shares (which are storable according to the definition of renewable energy capacity) are also excluded from the analysis to test the robustness of the results.

4.3.2 Flexible-Baseload Natural Gas Generation Data EIA Form EIA-923 (Power Plant Operations Report), Form EIA-860 (Annual Electric Generator Report), and Form EIA-860M (Monthly Update to the Annual Electric Generator Report) distinguish between the following natural-gas-generation technologies: natural-gas-fired combined cycle, natural gas-fired combustion turbine, steam turbine, and internal combustion engines (U.S. Energy Information Administration, 2017b). These technologies can be split into flexible-baseload

166

resources which compensate for renewable variability (i.e., natural-gas-fired turbines and natural-gas-fired combined-cycle plants) and baseload fossil technologies (e.g., internal combustion, diesel, steam turbines, and other fossil technologies). The definitions of flexible-baseload natural gas technologies (or FBNGT) and other baseload fossil technologies (or BLFT) adopted by Verdolini, Vona, and Popp (2016) was applied in the analysis to distinguish between installed capacity along these two dimensions, and to test the argument that FBNGT technologies compensate for the variability of “fast-flexing renewable energy” (FFRET) technologies:49

We define [flexible-baseload] as the sum of Gas Turbines and Combined Cycle, as these are mid-merit technologies often used to address peak load. Conversely, we define [baseload] as Internal combustion/diesel; Steam; and Other type of generation. These are technologies which are generally characterized by lower efficiency levels and slower ramp up times (p. 8).50

Figures 4.6, 4.7 and 4.8 show the development of the shares of installed capacity for FFRET, FBNGT, and BLFT for the top-10 SEIA-ranked solar-producing states in 2016. FFRET increased substantially between 2006 and 2014—especially in California, Texas, Arizona and New York-at different rates across the states. A sharp increase in renewable capacity in California between 2012 and 2014 characterizes the

49 In the analysis, coal and nuclear plants were also included in the BLFT share while FFRET consists of solar and wind energy technologies.

50 U.S. Energy Information Administration. (2017b). Electric Power Monthly with Data for January 2017; Table 6.7.A. Capacity Factors for Utility Scale Generators Primarily Using Fossil Fuels, January 2013-January 2017. United States: U.S. Energy Information Administration (EIA) a breakdown based on the capacity factors for utility scale generators primarily using fossil fuels (coal, natural gas, petroleum), January 2013-January 2017.

167

coming-into-force of a number of friendly legislations: AB 327 (2013), which enabled California Public Utilities Commission to establish procurement requirements for renewables; CPUC’s decision 10-05-018 (2011), which authorizes the use of tradable renewable energy credits (TRECS) (capped at 25% of a utility's RPS requirement and a price of $50) for RPS compliance; AB 2514 (2010), which allows utilities to procure energy-storage systems, and AB 327 (2013), which requires investor-owned utilities to make net metering programs available until the utility reaches its net metering capacity or on July 1, 2017, whichever comes first (U.S. Department of Energy, 2017a). Furthermore, this period coincides with the approval of the New York State

Energy Research and Development Authority’s (NYSERDA) petition to establish and fund the operations of New York Green Bank.

168

Figure 4.6 Fast-flexing Renewable Energy Share of Installed Capacity, 2001-2016

169

Figure 4.7 Flexible-Baseload Natural Gas Share of Installed Capacity, 2001-2016

A more uniform increase of flexible-baseload natural gas generation technologies at the national level characterized the start of the shale gas boom in the United States (Hefley and Wang, 2015; Murtazashvili, 2015). Texas stands out, with very high FBNGT capacity in early 2000s—a trend that remains almost constant throughout the study period, as aging steam turbines and internal combustion engine plants are replaced with combined cycle and combustion turbine technologies. This

170

trend is consistent with the analysis performed by Shavel et al. (2013) to examine the interaction of natural gas and renewable energy markets in Texas. It is also consistent with a simulation of several scenarios of potential future grid expansion of the ERCOT system:

Natural gas and renewables both play substantial roles in ERCOT and together provide all new generation. In the absence of continued or enhanced policy supports, we did find that natural gas generation would be the primary addition of choice though 2032, even with significant declines in the price of wind and solar power. If gas prices remain very low and current wind plants retire, the share of energy from renewables in Texas might decline slightly by 2032. However, across the more likely scenarios we analyzed, wind and solar grow from their current 10% generation share to levels between 25 and 43%. Natural gas-fired generation provides all of the remaining incremental generation, adding 12 to 25 GW of new combined-cycle capacity. As expected, the mix of new gas and renewables generation is sensitive to the price of natural gas, cost declines in wind and solar power, and tax and transmission policies. Changes in these factors can cause significant shifts in the mix of future installations, leading to a wide range of plausible generation shares for wind, solar, and natural gas (p. 3).

FBNGT capacity in North Carolina, Utah, Georgia, New Jersey, and California increased significantly over the study period, while there was a sharp drop in Arizona and Nevada in 2011 and 2014, respectively.

4.3.3 Other Baseload Fossil Electricity Generation Data

The share of BLFT declined over time in all 10 states and nationally, albeit at different rates. Utah had the highest BLFT capacity, averaging 175,235 MWh, and remained almost constant throughout the study period. Utah is a state in contrast: it generates more electricity than it consumes, three-fourths of its net electricity generation came from coal, and the state is a net power supplier to neighboring states.

171

Some states such as North Carolina, Georgia and Arizona experienced a rather sharp decline of BLFT capacity. Figures 4.9, 4.10, and 4.11 substantiate these facts.

Figure 4.8 Other Baseload Share of Installed Capacity, 2001-2016

A scatter plot of flexible-baseload natural gas generation capacity and fast- flexing renewable electricity capacity (Figure 4.9) indicates a strong positive

172

correlation between the two variables. However, eye-balling inspection of the data shows a moderately positive correlation between FBNGT- and FFRET-electricity capacity, implying that flexible natural gas share, and solar and wind capacity are correlated; hence, corr (FBNGT share, FFRET share) ≠ 0. The combined relationship between FBNGT and FFRET for the ten states (2001-2016) is almost a one-to-one increase (i.e., 0.707%), as presented in Figure 4.10.

Figure 4.9 Scatter Plot of FBNGT and FFRET Shares, 2001 to 2016

173

Figure 4.10 Monthly Normalized Relationship between FBNGT and FFRET for

Individual States, 2001 to 2016

174

Figures 4.11 and 4.12 indicate a negative correlation between flexible-baseload natural gas and other baseload technologies. Figures 4.11 and 4.12 also indicate a negative correlation between other-baseload and fast-flexing renewable technologies. For example, in Figure 4.11, when the share of FBNGT generation is low, the BLFT fuel share is high. Similarly, in Figure 4.12, the BLFT fuel share and FFRET (solar and wind) capacity are inversely correlated. The relationship between FBNGT and BLFT (negative), and BLFT and FFRET (negative) for the ten states is presented in Figures 4.13 and 4.14, respectively.

175

Figure 4.11 Scatter Plot of BLFT and FBNGT Shares, 2001 to 2016

176

Figure 4.12 Scatter Plot of BLFT and FFRET Shares, 2001 to 2016

177

Figure 4.13 Monthly Normalized Relationship between FBNGT and BLFT for

Individual States, 2001 to 2016

178

Figure 4.14 Monthly Normalized Relationship between BLFT and FFRET for

Individual States, 2001 to 2016

179

To investigate the relationships between FFRET and FBNGT capacity, a regression model was developed using the share of the flexible-baseload natural-gas-

K4LMF fired technology ((")* !"#"+'&,-,. ) as the explanatory variable of interest. The other variables explored previously that affect the level of renewable energy generation capacity in state i at time t—such as market regulations and environmental policies—were also included in the regression. For example, public polices related to environmental policy and market regulations analyzed include feed-in tariffs, tax credits, emission targets, and investment incentives that are specifically designed to accelerate the level of solar- and wind-energy generation capacity in state i at time t.

Schallenberg-Rodriguez (2017), Jenkins and Pérez-Arriaga (2017), Byrne et al. (2016), Lee and Zhong (2015), and Popp et al. (2010) have extensively reviewed the positioning of policy, finance, economics, markets, or a combination thereof in dramatically different configurations to support infrastructure-scale design, investment, and deployment of solar PV. As such, this analysis includes policy– market–finance interaction as a key determinant of the FFRET infrastructure-scale investment market.

180

Chapter 5

EVALUATION OF POLICY AND ECONOMICS OF INTEGRATING DISTRIBUTED UTILITIES

This chapter evaluates the policy effectiveness of economic instruments that have been used to support distributed generation and deployment. The chapter begins by analyzing and evaluating the main policy, environmental, and economic instruments for solar PV and wind-power development in the top-ranked solar- producing states of the United States. As mentioned in the previous chapter, the evaluation was conducted in ten states over the period 2001-2016. The instruments and control variables considered include feed-in systems, state-RPS policies, a public benefits fund, net metering, interconnection standards, the League of Conservation Voters, electricity import ratio, per-capita energy-related carbon dioxide emissions, average electricity price, per-capita real GDP, and the energy intensity of the economy.

5.1 Policy Platforms for Renewable Generation

5.1.1 Renewable Feed-in Tariffs The analysis considered feed-in tariffs (FITs) and quota systems (green certificates or RPS): the two most commonly applied policies supporting renewable energy deployment (Ritzenhofen, Birge, and Spinler, 2016; Ragwitz and Steinhilber, 2014; Sarasa-Maestro et al., 2013). The effectiveness of these policy instruments in addressing new challenges for distribution regulations (such as high transaction costs, insufficient market liquidity, and a lack of access to low-interest capital) was analyzed. Regarding integration of finance, markets, and policies, data relevant to the power sector was sourced from the U.S. Department of Energy (DOE) Database of

181

State Incentives for Renewables and Efficiency (DSIRE), from the OECD Environmental Policy Stringency (EPS) database, and from the U.S. Energy

Information Administration (Botta and Koźluk, 2014; U.S. Department of Energy, 2017b; U.S. Energy Information Administration, 2017a).51 In line with previous findings of Schallenberg-Rodriguez (2017), FITs, tax incentives, and RPS mechanisms have been applied by most states to develop renewable energy technologies. FIT policy guarantees a fixed remuneration to renewable generation in the power market. These feed-in systems are particularly effective in supporting renewable energy technologies when a low-level of risk for investors is required—i.e., from an early stage of development to market competitiveness, and stimulate market innovation, manufacturing capacity, and technological leaning. In 2013, California, Hawaii, Maine, Oregon, Rhode Island, Vermont, and Washington introduced FITs for some renewable energy technologies.

Another form of FIT—offering long-term, wholesale electric-energy contracts to eligible generators—that has been implemented in the above-mentioned states and in Alabama, Kentucky, Mississippi, North Carolina, Tennessee, Virginia, Florida,

Georgia, Indiana, Michigan, New York, Texas, and Wisconsin is the “electricity provider program” (voluntary offerings) (Schallenberg-Rodriguez, 2017).

51 DSIRE is the most comprehensive database of information on incentives and policies that support renewable energy and energy efficiency in the U.S. The OECD EPS database has information on 15 different environmental policy instruments which includes market and non-market instruments. EIA’s Table 5 contains information on electric power industry generation of all the states by primary energy source.

182

5.1.2 State RPS Policies RPS mechanisms promote the right market view policy and conditions for developing renewable energy projects by addressing the cost of service regulations in terms of “prudence of inputs.” These regulations make it challenging for electric utilities to acquire solar- and wind-energy assets and to deliver improved performance (Jenkins and Pérez-Arriaga, 2017). Therefore, an RPS mechanism is suitable for encouraging renewable energy development to a certain level, while FIT policies should be applied to encourage the development of renewable energy sources (Abolhosseini and Heshmati 2014). As of April of 2017, 30 states 52 and the District of Columbia have adopted mandatory RPS targets for renewable electricity generation (Figure 5.1), and 23 states have active and binding Energy Efficiency Resource Standards (EERSs) for electricity. States that have actively created and implemented such electricity resource standards and other supporting regulatory policies have seen increased growth in renewables and efficiency. State and electric-utility solar policies continue to undergo changes. In 2016 alone there were nearly 212 solar policy actions. These actions occurred across 47 states plus the District of Columbia. Of these, 73 were related to net metering, 71 to residential fixed-charge and minimum-bill increases, and 20 to solar-valuation or net-metering studies, as shown in Figure 5.2 (North Carolina Clean Energy Technology Center, 2017).

52 The following states have passed RPS rules: AZ, CA, CO, CT, DE, HI, IA, IL, MA, MD, ME, MI, MN, MO, MT, NC, NH, NJ, NM, NV, NY, OH, OR, PA, RI, TX, VA, VT, WA, and WI.

183

Source: N.C. Solar Center at N.C. State University, Database of State Incentives for Renewables and Efficiency (DSIRE, 2017a) (accessed April 2017).

Figure 5.1 Distribution of Renewable Portfolio Standards and Goals

184

Source: (North Carolina Clean Energy Technology Center, 2017, p. 5). Note:

The total “number of States/Districts” is not the sum of the rows, as some states have multiple actions.

Figure 5.2 Summary of Solar Policy Actions in 2016

The underlying assumption of this evaluation is that FITs are particularly effective in supporting small and new producers (or in some cases prosumers) in the power market. On the contrary, RPS rules are more effective in promoting in-state renewable energy development in general, as they require a percentage of energy generated or sold to come from renewable energy sources (such as wind, solar, or geothermal energy) (Ritzenhofen, Birge, and Spinler, 2016; Yin and Powers, 2010).

185

However, a number of RPS studies have concluded that these policies have a varied impact on the share of renewable-electricity capacity in the fuel mix of the states where they were enacted (Carley, 2009; Michaels, 2008; Shrimali and Kniefel, 2011; Wiser, Namovicz, Gielecki, and Smith, 2007). According to Bowen and Lacombe, RPS policies,

Vary considerably on a number of characteristics, including the stringency of the percentage requirement, how much generation capacity is required to meet the standard, and what types of renewables are allowed to meet the requirements. This variability in the standard has posed difficulties for determining whether the policies have been effective in promoting adoption of renewable technologies. (2017, p. 179) Following Yin and Powers (2010), a new measure, defined as effective RPS (EFFRPS) in each state, was introduced in the model to account for these differences and to measure the strength of each state’s RPS policies. The EFFRPS variable may differ from the nominal RPS targets provided for in the state law. It measures the percentage of new renewable generation beyond the existing capacity. Therefore,

EFFRPS represents the “incremental percentage requirement” and measures the strength of the RPS to incentivize additional generation from the renewable energy sources, and was calculated as follows:

0Z-[.-\] 7898:-.; <^ NOOPQ(-. = R%S'T"U-. V !%W*)"X*-. − (5.1) 0H_:.`-:-.; a8H_b<^

In this equation, Nominal refers to the target RPS requirement in state i in year t. The data for the nominal RPS were obtained from the Database of State Incentives for Renewables and Efficiency (January of 2017 data release) (DSIRE, 2017). Coverage refers to the percentage of sales that is covered by the RPS at time t, based

186

on which electric utilities (or load-servicing entities) are required to meet the standard in specific state i. The coverage was calculated using EIA-861based on utility definitions in the DSIRE database. Electricity sales are taken from the EIA Detailed State Data (U.S. Energy Information Administration, 2017h). The term Electricity Sales refers to the total electricity sales in the state when the RPS standard became operational. Lastly, Existing Capacity is the amount of electricity generated in each state that is allowed by the RPS policy to be counted toward the RPS law at the time the law became effective in the state. DRISE database lists the definitions of renewables that qualify to be counted toward the standard and the capacity, and generation data was sourced from the EIA detailed state data (DSIRE, 2017; U.S. Energy Information Administration, 2017d; U.S. Energy Information Administration, 2017e). However, state mandates differ significantly from each other with respect to percentages, timelines, sources that count as “renewable,” and enforcement mechanisms. While some states—Arizona, Massachusetts, Vermont, and Montana— only allow generation from new renewable assets to be “counted” toward their RPS standard, New Mexico, New Hampshire, Maine, Delaware, North Carolina, New York, Virginia, Oregon, and Washington allow generation from some, but not all existing renewable electric-generating units. For example, New York’s Clean Energy Standard (CES) states (DSIRE, 2017b) the following:

The implementation plan provides detailed eligibility criteria for new RES resources based on size, geographical location, energy delivery requirements, and date of operation. In addition to new generation facilities, facilities that have performed significant upgrades, facilities that are repowered, or facilities that have been relocated might also qualify if the meet certain NYSERDA eligibility requirements. As of February 2017, the PSC has not determined if net metered Distributed Energy Resources would be eligible to generate Tier I RECs.

187

This variation presents a challenge in calculating the eligible existing RE capacity and the value of EFFRPS. Based on Equation 5.1, Table 5.1 summarizes the EFFRPS values calculated for the top-ten SEIA-solar-ranked states in 2016 and other states with RPS and capacity standards. However, it is important to note that the vast majority of these RPS goals, especially the RPS legislations implemented between 1999 and around 2007, were not passed for decarbonization53 but rather for diversification away from natural gas owing to reliability and volatility concerns associated with natural gas on electricity prices.

Table 5.1 Top EFFRPS Values for Each States with an RPS Target State RPS Target EFFRPS State RPS Target EFFRPS Arizona 15.0%* 8.7% New York 50.0% 0.17% California 50.0% 20.15% Nevada 25.0% 5.4% Georgia 0.0% 0.0% North Carolina 12.5%† 11.9% Massachusetts 22.1% 20.8 Texas 8.7%** 3.2% New Jersey 22.5% 18.1% Utah 20.0% 20.0% Other States State RPS Target EFFRPS State RPS Target EFFRPS Alabama 0.0% 0.0% Missouri 15.0% 10.5% Alaska 0.0% 0.0% Montana 15.0% 10.9% Arkansas 0.0% 0.0% Nebraska 0.0% 0.0%

53 The author would like to thank anonymous referee for pointing out this issue.

188

Colorado 30.0%† 18.0% New 24.8% 17.4% Hampshire Connecticut 27.0% 15.6.0% New Mexico 20.0%† 13.8% Delaware 25.0% 24.3.0% North Dakota 10.0%* 9.8% Florida 0.0% 0.0% Ohio 25.0%* 11.0% Hawaii 40.0% 33.3% Oklahoma 15.0%* 15.0% Idaho 0.0% 0.0% Oregon 25.0%† 7.5% Illinois 25.0% 22.5% Pennsylvania 18.0%* 13.7% Indiana 0.0% 0.0% Rhode Island 16.0%* 14.5% Iowa 1.2.0%† 0.0% South Carolina 0.0% 0.0% Kansas 20.0% 12.0% South Dakota 10.0%* 9.7% Kentucky 0.0% 0.0% Tennessee 0.0% 0.0% Louisiana 0.0% 0.0% Vermont 20.0%* 20.0% Maine 40.0% 8.2% Virginia 15.0%* 6.1% Maryland 20.0% 14.9% Washington 15.0% 11.2% Michigan 10.0% 6.6% West Virginia 25.0%* 24.8% Minnesota 31.5% 21.9% Wisconsin 10.0% 4.7% Mississippi 0.0% 0.0% Wyoming 0.0% 0.0%

* RPS standard that includes fossil fuels † RPS standard varies by the size of the electric utility ** Capacity standard Source: Adapted from Bowen and Lacombe (2017)

5.1.3 Net metering Besides the impacts of the stringency of state-level RPS, three additional policy instruments that are commonly implemented by states to encourage renewable energy development were also controlled for. Net metering (NETMETER) which is

189

designed to incentivize small, customer-sited generation, is another policy designed to encourage development of renewable electricity. It is also a binary variable and equals one if, in each year, there is an existing net metering law on the books. As shown in Figure 5.3, as of April of 2017, 39 states and the District of Columbia had mandatory net-metering rules that support renewable energy.

Figure 5.3 States with Mandatory Net Metering Rules

Data from the Monthly Electric Power Industry Report (Form EIA-861M) shows that the total PV installed net metering capacity increased by 21.1%, 39.7%, 40.1%, 54.2% in Arizona, California, Massachusetts, and New York, respectively, from 2015-2016. Figures 5.4 and 5.5 show the monthly distribution in 2016 of total-

190

capacity installations (MW) and the total energy sold back to the electric utilities (MWh) for PV and wind in four states: namely Arizona, California, Massachusetts, and New York.

Figure 5.4 Comparison of Total Installed New Metering Capacity in AZ, CA, MA, and

NY in 2016

191

Figure 5.5 Comparison of Total Energy Sold Back in AZ, CA, MA, and NY

5.1.4 Public Benefits Fund Another type of policy instrument tested is the public benefits fund (PUBENFUND). Public benefits funds are state-level programs developed through the electric-utility restructuring process to support renewable energy deployment and energy efficiency initiatives (Cheng and Yi, 2017). This program is maintained by the state public-utility commissions and is commonly supported through a charge to all consumers on electricity consumption or payments from the utilities themselves. Public benefits funds are also frequently referred to as a system-benefits charge. It is

192

also a binary variable and equals one if, in each year, a state maintains the fund as part of its policy mandate for supporting the portfolio of renewable electric-generating resources. It is zero otherwise. Table 5.2 summarizes PUBENFUND actions for the period 2000-2015.

Table 5.2 State Level Public Benefits Fund Actions State Action Taken California Eligible technologies include: geothermal electric, solar thermal electric, solar photovoltaics, wind (all), biomass, hydroelectric, municipal solid waste, landfill gas, tidal, wave, ocean thermal, hydroelectric (small), anaerobic digestion, and fuel cells using renewable fuels. Rates vary by utility and customer type. For example, renewables = ~1.6 mills/kWh; efficiency = ~5.4 mills/kWh; and research, development, and demonstration = ~1.5 mills/kWh. Connecticut The state spends nearly $120 million annually in furtherance of public benefits fund programs for renewable technologies. Delaware Created in 1999, the Delaware Green Energy Fund allows municipal utilities to opt out of the state’s RPS compliance schedule if they contribute to the fund for investor-owned utilities or create their own green energy fund with an equal surcharge. It provides up to 50% of installation costs for solar PV, fuel cells, solar water heating, and wind turbine systems. Hawaii Eligible renewable and other technologies include solar water heat and solar space heat. For 2009 and 2010, the public benefits fund includes 1% of projected total utility revenue; for 2011 and 2012, 1.5% of projected total utility revenue; and for 2013-onwards, 2% of projected total utility revenue, including revenue tax.

193

Illinois A legislation to restructure the electric industry in Illinois created a separate public benefits funds in 1999 to support renewable energy and residential energy efficiency. The Renewable Energy Resources Trust Fund supports renewable and other technologies including solar thermal electric, solar thermal, solar PV, wind, biomass, hydroelectric, fuel cells using non-renewable fuels, geothermal electric, wind (small), fuel cells using renewable fuels. The surcharge varies by customer class: $0.05 per month for residential electric service; $0.05 per month for residential gas service; $0.50 per month for nonresidential electric service with less than 10 MW of peak demand; $0.50 per month for nonresidential gas service; $37.50 per month for nonresidential electric service with at least 10 MW of peak demand; and $37.50 per month for nonresidential gas service taking at least 4 million therms of gas during the previous calendar year. Massachusetts The Massachusetts Renewable Energy Trust Fund supports solar PV, wind, biomass, geothermal electric, solar thermal electric, hydroelectric, municipal solid waste, combined heat and power, fuel cells using renewable and non-renewable fuels, landfill gas, tidal, wave, ocean thermal, anaerobic digestion. The state charged $0.0005 per/kWh in 2003 and in each following year. Minnesota Under the Xcel Energy's Renewable Development Fund eligible renewable and other technologies include solar PV, wind, biomass, hydroelectric, CHP, anaerobic digestion, and fuel cells using renewable fuels. Between 2008-2013, the total of $19.5 million annually was awarded; in 2013, approximately $24.5 million; and in 2014, a total of $42 million. Montana Montana’s Universal System Benefits Program was established in 1997 and eligible renewable and other technologies include solar water heat, solar space heat, geothermal electric, solar thermal electric, solar thermal process heat, solar PV, and wind. The surcharge rate is based on 2.4% of electric utilities' 1995 revenue.

194

New Jersey New Jersey’s societal benefits charge supports investments in Class 1 renewable energy and energy efficiency projects. The eligible technologies include solar water heat, geothermal electric, solar thermal electric, solar PV, wind, biomass, hydroelectric, CHP, landfill gas, tidal, wave, wind, anaerobic digestion, and fuel cells using renewable fuels. The state’s per-kWh surcharge varies yearly by funding target. New York New York's system benefits charge was established in 1996 by NYPSC. It supports energy efficiency, education and outreach, research and development, and low-income energy assistance in commercial, industrial, investor-owned utility, local government, nonprofit, municipal utilities, residential, cooperative utilities, schools, state government, federal government, multifamily residential, and institutional sectors. The annual collection targets for each utility are established by the PSC. Pennsylvania Pennsylvania administers four sustainable energy funds (SEF): (i) the Metropolitan Edison Region SEF which is administered by the Berks County Community Foundation; (b) the Sustainable Development Fund, in Southeastern Pennsylvania PECO's service territory which is administered by The Reinvestment Fund; (c) the West Penn Power SEF which is administered by The Energy Institute of Penn State University, in partnership with Energetics, Inc; and the Sustainable Energy Fund of Central Eastern Pennsylvania. The fund varies by utility territory. Vermont Eligible renewable energy technologies include solar PV, wind, biomass, CHP, fuel cells using renewable and non- renewable fuels, anaerobic digestion, micro-turbines, and other distributed generation technologies.

Source: DSIRE (2017)

As of April 2017, 21 states maintain a public benefits fund that supports renewable energy (Figure 5.6).

195

Figure 5.6 States with Public Benefits Fund Policies and Incentives

5.1.5 Interconnection Standards The last type of policy controlled for is interconnection standards (INTERCONSTAND), which are standards that facilitate the contracting process for customer-sited renewable energy generation. INTERCONSTAND is a binary variable. It equals one in a state where there is a codified interconnection standard to facilitate installation of customer-sited renewable energy. As of April of 2017, approximately 37 states and the District of Columbia had enacted such laws (Figure 5.7) (DSIRE, 2017).

196

Figure 5.7 States with Interconnection Standards Laws

The data for NETMETER, PUBENFUND, and INTERCONSTAND were taken from (DSIRE, 2017).

5.2 Economic and Environmental Control Variables

5.2.1 League of Conservation Voters

Besides these policy variables, two social and economic variables that are thought to have an impact on the development of renewable energy technologies or the adoption of RPS policies were tested. These include scores from the League of

197

Conservation Voters (hereafter, LCV)54 and electricity price. LCV scores, LCVSCORE, is defined as the average scores of environmental preferences. In this formulation,

LCVSCORE was used as a control variable—i.e., to control for environmental feeling among the state’s residents including utilities, investors, and other regulators. LCVSCORE acts as a proxy for the preference of environmental policies in the state. In states with stronger environmental policies, there is also likely to be a higher demand for renewable energy development (Yin and Powers, 2010). Data for LCVSCORE was sourced from the LCV scorecard database (LCV, 2017). Using the LCV score as a measure of environmental preference is common in the studies of RPS policies and designs (Bowen and Lacombe (2017); Carley (2009); Shrimali and Kniefel (2011); Wiser, Barbose and Holt (2011); Yin and Powers (2010)). An average score of all legislators in a state, in each year, was taken to derive the average state score. It is expected that a state’s share of renewable generation and its environmental score should exhibit a positive relationship.

5.2.2 Average Electricity Price Average electricity price, ELECTRICPRICE, is an important indicator of energy demand and was also controlled for. Shrimali et al., (2012) indicate that average electricity price may have an ambiguous relationship with renewable

54 The League of Conservation Voters has been publishing the National Environmental Scorecard about the most important environmental legislation and the corresponding voting records of all members of Congress since 1970. The scores cover the most important issues of the year, including energy, climate change, public health, public lands and wildlife conservation, and spending for environmental programs.

198

generated energy. In states with high electricity prices, investment in renewable energy resources could face more resistance—especially if consumers are required to pay the additional costs of developing renewables. On the other hand, high electricity prices may also imply a more cost-competitive outlook for renewable prices. Figures 5.8 to 5.12 show the average price of electricity to ultimate customers by end-use sector and by state from 2013 to 2016.

Figure 5.8 Average Price of Electricity to Ultimate Customers for Residential Users and by State, 2013 to 2016

199

Figure 5.9 Average Price of Electricity to Ultimate Customers for Commercial Users and by State, 2013 to 2016

200

Figure 5.10 Average Price of Electricity to Ultimate Customers for Industrial Users and by State, 2013 to 2016

201

Figure 5.11 Average Price of Electricity to Ultimate Customers for Transportation

Users and by State, 2013 to 2016

202

Figure 5.12 Average Price of Electricity to Ultimate Customers for All Sectors and by

State, 2013 to 2016

The national average power price benchmark fell 0.29% to 10.41¢/ kWh in 2015, down 3.67% from the year before. The big decline in average prices in 2015 occurred in New York (5.97%), Georgia (4.09%), Texas (2.68%), Nevada (2.57%), and New Jersey (-1.51%). As shown in Figure 5.13, at the other end of the movement, no decline in prices occurred in all the ten states in 2013 and 2014. Massachusetts and

203

Utah had the highest price increase of 10.1% to 16.90¢/ kWh and 9.96% to 7.84¢/ kWh in 2015 and 2012, respectively.

Figure 5.13 Percentage Change in Annual Average Electricity Prices, Price

Benchmarks by State, 2010 to 2015

5.2.3 Electricity Import Ratio

A state’s dependence on other states for its energy needs was also controlled for through the use of the electricity import ratio, IMPORTRATIO. This variable was

204

calculated from the difference between electricity sales and generation as a share of sales. The data for calculating IMPORTRATIO was taken from the EIA Detailed State Data (U.S. Energy Information Administration, 2017b, 2017d, 2017h). The import ratio illustrates the interdependency of states in the electricity grid. It is positive if electricity sales exceed a state’s generation capacity, and negative if a state exports electricity. States with large fossil resource potential for use in electricity generation, thus heavy exporters, have negative import ratios and lower renewable generation shares. Figure 5.14 illustrates the electricity import ratios for each state in 2015.

Data Source: U.S. Energy Information Administration, 2017b, 2017h

Figure 5.14 Import Ratios for Each State in 2015

205

High importing states such as California, New York, Massachusetts, Vermont, Idaho, Indiana, Ohio, Delaware, Maryland, and South Dakota have greater incentives to subsidize renewable energy and thereby increase their generation portfolio (Yin and Powers, 2010). Nevertheless, exporting states such as Montana, Wyoming, Utah, Arizona, New Mexico, Illinois, Connecticut, Nebraska, and New Hampshire may also have incentives to develop a diverse energy mix that includes renewables (e.g., a lot more wind and solar, more energy storage and demand response) potentially for sale outside the state. Furthermore, low-exporting states such as Nevada, Texas, Iowa, and Kansas may have both significant fossil-resources potential and significant renewable potential. Therefore, these contradictory indicators in the correlation between import ratio and renewable generation across different states make it difficult to express any prior expectation for this relationship.

5.2.4 Energy Intensity and Per-capita Energy-related CO2 Emissions Next, a state’s energy intensity (taking the logarithm of amount of energy consumed per unit of economic output or, specifically, Btu per chained 2009 dollar of GDP, ENERGSITY, as reported by the U.S. Energy Information Administration (2017h) was also controlled for. Figure 5.15 shows the percentage change in the energy intensity for each state between 2001-2007, 2008-2014, and 2001-2014.

206

Figure 5.15 Percentage Change in Import Ratios for Each State between 2001 and

2014

The top-ten states with the highest rates of CO2 emissions per capita in 2014, namely Wyoming (22,840 Btu per chained 2009 dollar of GDP), Louisiana, West

Virginia, North Dakota, Montana, Alabama, Alaska, Mississippi, and Arkansas (all in the 11-19,000 Btu per dollar range), and Kentucky (10,790 Btu per dollar) also had higher energy-intensity values. Massachusetts, New York, California, Maryland, and

New Jersey had some of the lowest rates of emissions per capita—between 2 and 4,500 Btu per dollar—and the least percentage change in energy intensity between 2001 and 2014 (at -0.26%, -0.23%, -0.25%, -0.30%, and -0.10%, respectively). The

207

2001, 2008, and 2014 national energy-intensity averages were 7,600, 6,700, and 6,200 Btu per dollar of GDP. Figures 5.16 and 5.17 show the energy intensity by state (thousand Btu chained 2009 dollar of GDP).

Source: U.S. Department of Labor. (2017)

Figure 5.16 Energy Intensity by State, 2014

208

Figure 5.17 Energy Intensity by State

5.2.5 Per-capita Real GDP

Lastly, the state’s per-capita real GDP (chained 2009 dollars), PERCAPITAGDP, as reported by the U.S. Bureau of Economic Analysis (BEA, 2017), was also controlled for. Per-capita real GDP by state measures the economic well-being of a state. The District of Columbia, Massachusetts, New York, and Connecticut had the highest per-capita real GDP in 2016, ranging from $160,472,

$65,545, $64,579, and $64,511, respectively (BEA, 2017). California’s and New

209

Jersey’s per-capita real GDP was $58,619 and $57,084, which is nearly 14% above the national average ($50,577). Mississippi had the lowest per-capita real GDP in 2016 ($31,881): 37% below the national average. We expect that higher per-capita real GDP levels will be associated with higher levels of renewables, as these states are more likely to invest in renewable energy generation despite the higher energy costs associated with these projects.

5.3 Estimating the Basic Vector of Policy and Economic Controls PERCAPITACO2 and PERCAPITAGDP, unlike EFFRPS and FITs, PUBENFUND, NETMETER, and INTERCONSTAND, LCVSCORE, ELECTRICPRICE, IMPORTRATIO, ENERGSITY, are not positively associated with a higher share of renewable generation; thus, they are expected to have smaller effect on their deployment and diffusion. Hence, the basic vector of these policy controls POLICY can be expressed as follows:

QOde!f-.gh =

[PQ(-.gh; Oe$-.gh; QklNROkRm-.gh; NETMETER-.gh; eR$NP!sR($tRm-.gh; … ] (5.2)

Following Bowen and Lacombe (2017), the optimal lag between the dependent KK/0F (/_\_x8yH_ a8`_) variable ((")*!"#"+'&,-. ) and all the variables in the POLICY vector was captured by using use a moving average of the policy variables between &

− 1 and & − 3. Figures 5.18 to 5.27 show the evolution of the policy variables (EFFRPS, and the two socio-economic variables (i.e., ELECTRICPRICE and LCVSCORE) in the sample for the ten states.

210

Figure 5.18 Effective State-RPS Policy Stringency

The use of the RPS policy mechanism began in early 2000, and it increased progressively after 2004, but with some heterogeneity across states. While California, Arizona, and New York score very high on the RPS policy instrument, Utah and North

Carolina score rather low.

211

Figure 5.19 FITs Policy Stringency

Seven states—including California, Hawaii, Maine, Oregon, Rhode Island, Vermont, and Washington—introduced FIT policy for renewable energy technologies in 2013. They also introduced a variant of the FIT program known as “electricity provider program” (voluntary offerings) (Schallenberg-Rodriguez, 2017). New York, Texas, North Carolina, Georgia, Alabama, Kentucky, Tennessee, Indiana, Michigan, Virginia, Florida, Wisconsin, and Mississippi have also implemented the electricity provider program. National FITs increased steadily from 2006 to 2013, indicating that more states relied on this policy instrument to develop their renewable

212

energy capacity. However, of the ten states studied, only California, Georgia, and New York had FIT laws that support renewable energy.

Notes: CA, NJ, NY, NV, UT, ID, KS, KY, ME, MO, MN, and OR have discretionary aggregate net metering caps, which means that regulators are authorized to adjust the cap. 1 denotes states with net metering rules in a given year, and 0 otherwise.

Figure 5.20 Net Metering Policy Stringency

213

Figure 5.21 Interconnection Standards Policy Stringency

214

Figure 5.22 Public Benefits Fund Policy Stringency

215

Figure 5.23 LCV Score Control Variable

Massachusetts, New York, California, and New Jersey have consistently maintained very high LCVSCORE values. A positive relationship between a state’s share of renewable electricity and a state’s environmental scores is assumed. Of the above ten states analyzed, Utah and Texas registered the lowest environmental scores although Texas realized a 50% increase in its environmental scores between 2014 and 2016.

216

Figure 5.24 Average Electricity Price Control Variable

The U.S. average retail electricity price per kWh in 2013, 2014, and 2015 was ¢10.10, ¢10.44, ¢10.41, respectively.

217

Figure 5.25 Electricity Import Ratio Control Variable

218

Figure 5.26 Per-capita Energy-related CO2 Emissions, by State

219

Figure 5.27 Per-capita Real GDP by State

Per-capita real GDP, average electricity price, and per-capita energy-related

CO2 emissions were included to capture overall economic well-being, expectations about future electricity demand, and other demand-side factors related to the states or their economy size that were not captured by the policy indicators. Irrespective of the technology considered (FFRET, FBNGT, or BLFT technologies), all these forces are expected to increase the demand for new power generation. Additionally, while none of the ten states analyzed invested in large-scale storage technologies for power

220

production during the sample period, it is likely that these options and low-cost electricity-storage technologies (batteries) may influence investment decisions. Therefore, a dummy variable was included in the model to control for the technological evolution of energy-storage and smart-grids technologies. This was built following a method recommended by Bowen and Lacombe (2017). The dummy equals one if a state had a mandatory RPS policy and zero if a state’s RPS is non- mandatory. However, inclusion of this dummy variable had little effect on the results, and it was insignificant.

221

Chapter 6

ESTABLISHING A DISTRIBUTED UTILITIES MODEL FOR ESTIMATING THE EFFECTS OF FLEXIBLE NATURAL GAS TECHNOLOGIES ON RENEWABLE GENERATION

This chapter reports the results of the econometric modeling.

6.1 Overview of the Methodology Framework The indicators used for this study are divided into three major categories: policy, economic, and environmental variables. In addition, a number of control variables are applied to test the robustness of the dynamic econometric model. A system-generalized method of moments (or system GMM) estimator is performed to reveal the dynamic relationships of the indicators in the model by using different combinations of indicators and control variables. The estimates for the period 2001- 2016 are reported for the sample of ten states.

6.2 Model Establishment

6.2.1 Model Description This section explains the empirical strategy designed to illuminate the econometric issues which characterize the identification of the effect of flexible natural gas technologies on renewable capacity. Tables 6.1 and 6.2 summarize descriptive statistics of variables of interest across the nation and the top-10 solar- producing states. Overall, the share of renewable generation capacity is still very low (2.8% and 2.3% when biomass is included or excluded from the calculation, respectively). Nevertheless, this share spans from 0.5 to 8.7%. For the ten states only, renewable capacity stands at 2.3% and 1.6% when including or excluding biomass

222

from the calculation, respectively. The share of FBNGT capacity is approximately five times higher than renewables on average, and the maximum penetration rate is 22%, or two and a half times that of renewable electricity capacity.

223

Table 6.1 Descriptive Statistics for Annual U.S. National Observations Variable Observations Mean Median SD Min Max Probability Data Source Share of FFRET Capacity 192 2.7559 2.0779 2.2108 0.4826 8.7132 4.00E-06 (excl. hydro, waste) Share of FFRET Capacity (excl. hydro, waste, biomass 192 2.3048 1.6302 2.1493 0.1191 8.118 3.00E-06 / wood)

224 Share of FFRET Capacity

(excl. wind, hydro, waste, 192 0.1585 0.0255 0.2871 0.0008 1.1355 biomass / wood) U.S. Energy Information Share of FFRET Capacity Administration (excl. Solar, hydro, waste, 192 2.1483 1.6106 1.9393 0.1171 7.226 1.90E-05 (2017d) biomass / wood) Share of FBNGT Capacity 192 13.6287 13.1405 3.5823 7.4580 22.1263 7.30E-03 Share of BLFT Capacity 192 55.7455 57.3591 5.4257 38.8658 64.2746 7.00E-05 Share of NUCLEAR 192 19.7037 19.8720 1.1458 16.6722 22.3499 2.34E-01 Capacity Share of HYDRO Capacity 192 6.6564 6.5029 1.2527 4.1482 10.3156 1.27E-02

Share of WOOD AND WOOD DERIVED FUELS 192 0.9640 0.9678 0.0750 0.8030 1.1368 2.88E-01 Capacity Share of OTHER 192 0.4510 0.4345 0.0682 0.3329 0.6326 2.76E-03 BIOMASS Capacity Policy Controls FIT 192 1.0000 1.0000 0.7520 0.0000 2.0000 2.48E-04 Botta and Kozluk (2014); EFFRPS 192 1.4250 1.6000 0.7363 0.0000 2.8000 5.15E-02 DSIRE (2017) Other Policy and Socio-Economic Controls 225

NETMETER 192 0.8802 1.0000 0.3256 0.0000 1.0000 PUBENFUND 192 0.9427 0.0000 0.2330 0.0000 1.0000 (DSIRE, 2017) INTERCONSTAND 192 0.6875 1.0000 0.4647 0.0000 1.0000 (DSIRE, 2017) LCVSCORE 192 48.6563 47.0000 4.9902 43.0000 61.5000 LCV (2017) U.S. Energy Information ELECTRICPRICE 192 9.0996 9.5350 1.2133 6.9000 11.0300 1.79E-04 Administration, (2017d) U.S. Energy Information IMPORTRATIO 192 -0.1011 -0.1008 0.0095 -0.1193 0.0845 9.23E-02 Administration, (2017d)

Energy Market Indices Control US-SUNDX Index 192 347.5939 278.8300 311.8078 65.7300 1548.200 US - S&P GSCI Index 192 450.9645 447.0381 166.0257 166.9670 862.8090 9.34E-03 Bloomberg (2017) US - NGUSHHUB (Henry 192 4.8688 4.2629 2.3652 1.6164 14.8423 Hub) Additional Confounding Factors U.S. Energy Information PERCAPITACO2 192 18.4603 18.5125 1.5259 16.1623 20.2297 0.0000 Administration,

226 (2017d)

PERCAPITAGDP 192 47824 48062 50577 1637 50577 4.11E-02 BEA (2017)

U.S. Energy Information ENERGSITY 192 6.7026 6.6500 0.5300 6.0000 7.6000 5.91E-04 Administration, (2017d)

Table 6.2 Descriptive Statistics for Monthly State Level Observations Variable Observations Mean Median SD Min Max Data Source Share of FFRET Capacity (excl. 1920 2.3451 1.4177 3.1431 0.0000 24.4268 hydro, waste) Share of FFRET Capacity (excl. 227 1920 1.6195 0.2093 2.9820 0. 0000 22.7901 hydro, waste, biomass / wood)

Share of FFRET Capacity (excl. wind, hydro, waste, biomass / 1920 0.4808 0.0035 1.3399 -9.24E-06 12.5151 wood) U.S. Energy Information Share of FFRET Capacity (excl. Administration (2017d) Solar, hydro, waste, biomass / 1920 1.1386 0.0228 2.3616 0. 0000 17.2375 wood) Share of FBNGT Capacity 1920 20.3484 20.8024 11.8754 0.0589 51.1373 Share of BLFT Capacity 1920 48.8801 49.0126 19.8360 14.5655 97.9096 Share of NUCLEAR Capacity 1920 20.5041 20.7224 15.4913 0.00 64.4233 Share of HYDRO Capacity 1920 5.7003 3.1946 6.4514 0.0084 35.8044

Share of WOOD AND WOOD 1920 0.7136 0.2503 0.9681 0. 0000 7.0417 DERIVED FUELS Capacity Share of OTHER BIOMASS 1920 0.7257 0.1692 0.9167 0.0000 4.7513 Capacity

FIT 1920 0.0849 0.00 0.1316 0.0000 0.4250 Botta and Kozluk EFFRPS 1920 2.7281 0.9958 4.1987 0.0000 20.5955 (2014); DSIRE (2017) NETMETER 1920 0.7250 1.0000 0.4466 0.0000 1.0000 (DSIRE, 2017) PUBENFUND 1920 0.4000 0.0000 0.4900 0.0000 1.0000 (DSIRE, 2017) INTERCONSTAND 1920 0.6938 1.0000 0.4611 0.0000 1.0000 (DSIRE, 2017) 228 LCVSCORE 1920 49.8022 43.1500 27.6778 1.1700 98.4500 LCV (2017)

U.S. Energy Information ELECTRICPRICE, log 1920 2.3133 2.2793 0.2927 1.6501 2.8893 Administration (2017d) U.S. Energy Information IMPORTRATIO 1920 -0.0458 0.0158 0.2641 -0.6522 0.4287 Administration (2017d) U.S. Energy Information PERCAPITACO2, log 1920 2.6998 2.6559 0.3413 2.0924 3.4123 Administration (2017d)

PERCAPITAGDP, log, 1920 10.7990 10.7705 0.1492 10.5436 11.0905 BEA (2017)

U.S. Energy Information ENERGSITY 1920 5.7908 5.7909 2.2413 2.7000 12.7000 Administration (2017d)

Generally, the distribution of FFRET investments is skewed to the left more than that of FBNGT technologies. On the other hand, the bulk of generation over the sample period was provided by BLFT technologies (at approximately 56%). The share of nuclear, hydro-, wood and wood-derived fuels, and biomass accounted for 19.7%, 6.7%, 1%, and 0.5%, respectively. California, Texas, Nevada, and New York are leaders in non-hydro renewable investments and deployment, and they generally have high RPS, FITs, LCV scores; low entry barriers; and a well-developed wholesale electricity market. It is also important to note that, for this study, eight states including Arizona, California, Massachusetts, New Jersey, Nevada, New York, and Texas are above the median in both renewable investments and in deployment of flexible- baseload natural gas generation technologies (e.g., natural-gas-fired combined cycle and natural-gas-fired combustion turbine). California and Texas lead in both. Table 6.3 shows summary statistics for all variables by state.

229

Table 6.3 Descriptive Statistics of Observations, By State US State AZ CA GA MA NC NJ NV NY TX UT Average Share of FFRET Capacity (Solar and Wind) 0.94 5.37 0.06 0.55 0.33 0.3 1.45 1.39 5.08 0.73 2.3 Share of FFRET

230 Capacity (solar Only) 0.77 1.74 0.06 0.34 0.33 0.28 1.22 0.02 0.03 0.03 0.16

Share of FFRET Capacity (Wind Only) 0.17 3.64 0.21 0.02 0.23 1.36 5.05 0.7 2.15

Share of FBNGT Capacity 15.06 31.07 10.43 31.12 5.79 20.81 34.97 18.78 28.34 7.1 13.63 Share of BLFT Capacity 48.8 24.45 58.99 47.73 55.75 26.66 51.84 28.92 56.17 89.49 55.75 Share of NUCLEAR Capacity 28.2 14.47 25.61 13.96 32.44 50.22 30.55 9.6 19.7

Share of HYDRO Capacity 6.61 14.88 2.4 2.6 3.67 0.04 6.21 18.67 0.23 1.68 6.66 Share of WOOD Capacity 0.07 1.83 2.74 0.39 1.49 0 0.39 0.23 0.96 Share of OTHER BIOMASS Capacity 0.04 1.24 0.13 2.87 0.19 1.43 0.02 1.1 0.12 0.11 0.45

FIT 0.11 0.07 0.08 1 EFFRPS 2.81 4.91 5.96 1.13 4.42 2.2 0.05 1.73 4.15 1.43 231 NETMETER 3.23 3.08 1.85 2.73 1.64 1.47 2.32 1.52 0.45 0.88

PUBENFUND 0.97 0.97 0.94 0.97 0.94

INTERCONSTAND 0.31 0.94 0.31 0.75 0.81 0.88 0.94 0.69 LCVSCORE 33.17 74.28 28.32 92.62 39.81 71.99 45.03 80.88 17.63 14.3 48.66 ELECTRICPRICE 8.92 13.17 8.5 13.49 8.14 12.67 9.34 14.84 8.98 6.81 9.1 IMPORTRATIO -0.49 0.22 0.02 0.27 0.03 0.18 -0.06 0.04 -0.17 -0.51 -0.1 CO2 Emissions Per Capita 15.3 10.17 16.83 11.18 15.08 13.44 16 9.61 25.72 24.78 18.46 GDP Per Capita 39988 52969 44369 60194 43796 55959 47555 59864 48528 42008 47823 Energy Intensity 6.86 3.63 6.73 3.24 6.01 4.57 5.4 3.28 10.34 7.85 6.7

6.2.2 Dynamic Econometric Model As explained earlier, it is assumed that the percentage of FFRET capacity

--./0 (!"#$%"&"'()*+, ) is a function of the percentage of FBNGT capacity -1230 (!"#$%"&"'()*+, ), as well as of the policy variables, and of all other controls. Furthermore, given that capacity is a stock variable at any single point in time during the study period, a dynamic econometric model that consists of variants of the share of renewable capacity, flexible natural gas capacity, and policy variables was considered. Following the same approach used by Popp et al, (2011) and Verdolini et al., (2016), the dynamic econometric model can be written as follows:

--./0 8 --./0 -1230 !ℎ"#$%"&"'()*+,, = 9:; 7 !ℎ"#$%"&"'()*+,,<9 + βShareCapacity+,,<; + KLMNO%P+,,<; + QR+,,<; + S+ + S, +∈+,,

(6.1) Where:

• !h"#$%"&acityFFRET denotes capacity share in fast-flexing renewable technologies (solar and wind);

• !h"#$%"&acityFBNGT is the capacity share in flexible-baseload natural gas technologies;

• POLICY is the basic vector of all the policy controls and includes both market-based and non-market-based instruments, such as RPS, FITs, public benefits fund, and government tax incentives;

• S+ denotes state-fixed effects and control for time-invariant factors (e.g., persistent institutional related factors identical to the state governance system);

232

• S, denotes time-fixed effects control for inter-temporal trends that are uniform across states (e.g., economic cycle);

• ε is a random error term denoting the share of FFRET that is left unexplained by the model (i.e., the model residuals) at time t; and

• X is the vector of the other relevant control variables influencing renewable energy diffusion, including an index for the size of the economy (per-capita GDP) and cross-state electricity importation ratio.

Equation 6.1 accounts for the rich dynamics of electricity- generation capacity. Furthermore, the use of more than one lag shows the rich and technology-specific dynamics of both RE and NG technologies. This approach also helps to account for both the short- and long-term effects of the variables of interest— in particular, the policy variable. Following Abid (2016); and Popp et al. (2011), I assumed that the dependent variable in the empirical specification is in first difference (i.e., s = 1 and ρ = 1)12, meaning which means that there have been dynamic changes have been occasioned by technological changes and market innovation. For example, over the study period the role of technology and innovation in natural gas markets—especially in the use of information and communication systems to facilitate operations of gas- spot markets in the form of day-ahead and online electricity markets— has substantially grown. Billing, load management, real-time metering and quality control are now commonly offered to customers under deregulation and are also integral parts of the integrated spot market.

233

6.2.3 Estimation Techniques of the Model A panel-data estimation technique was used. The estimation procedures are discussed below. The panel data set used in this study is composed of monthly data from 2001 to 2016 for the top-ten SEIA-ranked solar states of 2016. The total number of observations for the cross-sectional and time-series data is 1920 for each variable.

The dynamic model specification was estimated using the “system” generalized method of moments (system-GMM) estimator proposed by Arellano and Bover (1995). This method has two principal advantages.

First, least-squares-based inference methods—for example, fixed effects or random effects estimation—is susceptible to the potential endogeneity of some explanatory variables that result from measurement errors, risk of omitted variables, and simultaneity bias (Baltagi, Bun and Sarafidis, 2015; and Asiedu and Lien, 2011). To address this concern, an appropriate linear, dynamic panel-data model was chosen that allows for endogenous covariates, for instance, in the empirical analysis of policy interventions (i.e., RPS, FIT, or a public benefits fund to support renewable investments at the state level). These policy variables are simultaneously determined with the outcome variable of interest and are strictly not exogenous (Besley and Case, 2000). As Verdolini, Vona, and Popp (2016) noted,

It is well known in the literature that a simple within-transformation fails to provide accurate estimates in dynamic panels. This bias is due to the mechanical correlation between the within-transformed error term and the right-hand side variables, and it decreases with 1/\, where \ is the number of periods considered. While the debate regarding the best estimator for dynamic panels is still open, the system-GMM estimator has gained some consensus especially in the case of highly persistent series… Compared to the difference-GMM estimator proposed by Arellano and Bond (1991), the system GMM approach mitigates the weak instrument problem using moment conditions both for the equation in levels and in first-differences.

234

The interrelations between economic, economics, environment, and society presented in Equation 6.1 involve dynamic adjustment processes. Therefore, as is the common practice, the lagged values of the dependent variable, the covariates, and sometimes both were included in the empirical specification. The basic justification for this approach was to instrument the lag terms of the dependent variable with lags and lagged differences to adequately characterize the economic dynamic adjustment in the model. Therefore, the use of a system-GMM dynamic panel makes it possible to control for both the time-specific and individual effects of the explanatory variables. Also, where one or more lags of the dependent variable is included in the explanatory variable, the method bears through the problem of the endogeneity of variables. GMM estimation methods have been employed extensively with a relatively minimal set of statistical assumptions in various ways—including in energy economics (e.g., relationship between energy consumption and economic growth) (Burnett and Madariaga, 2017), environmental economics (e.g., relationship between economic growth and pollution) (Abid, 2016), health economics (e.g., health expenditures, organization of health care, aging, addiction, insurance) (Ssozi and Amlani, 2015), international economics (e.g., effects of trade policy and economic integration)

(Baltagi, Bun, and Sarafidis, 2015), macroeconomics and development economics (e.g., economic growth, transition economics and effectiveness of foreign aid) (Burnett and Madariaga, 2017), labor economics (e.g., minimum-wage effects and labor supply), and industrial organization (e.g., mergers and acquisitions, evaluation of competition policy).

235

Second, long-term policy planning drives increased investments and deployment of renewable energy-generation systems to address climate change and achieve energy independence. The relationship between a diversified energy portfolio and the need to reduce energy dependence and carbon emissions has gained the attention of scholars and the public alike after the two global oil crises in the 1970s. As a result, these variables are expected to be highly persistent. However, as discussed above, if the explanatory variables are persistent, then their lagged levels are weak instruments of the subsequent first differences (Arellano and Bover, 1995; Blundell and Bond, 1998). This problem presents a fundamental challenge to determining the optimal number of lags, as the rich dynamics of renewable energy capacity demands a choice of a suitable lag length for the lagged terms of the FFRET variable in Equation 6.1. Following Burnett and Madariaga (2017), Ssozi and Amlani (2015), and Acemoglu et al. (2014), a practical approach that includes only the lagged dependent variable alongside state-and-time fixed effects was applied to estimate the optimal number of lags: i.e., by adding lagged terms until the additional lag is statistically insignificant. Consequently, to test the effect of the dynamic behavior, an endogenous, lagged, one-period variable was introduced in the explanatory variables including IMPORTRATIO, PERCAPITACO2, PERCAPITAGDP, and ENERGSITY.

Table 6.4 shows the appropriate lag structure for FFRET and FBNGT technologies.

236

Table 6.4 Appropriate Lag Structure for FFRET and FBNGT Technologies excl. excl. excl. hydro, excl. hydro, hydro, waste, hydro, waste, Dependent waste biomass / waste biomass, Flexible-Baseload Natural Gas Variable wood wood Technologies Share of FFRET Capacity Share of FBNGT Capacity Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 Dependent 0.805*** 1.085*** 0.952*** 0.876*** 0.961*** 0.909*** 0.908*** Variable, t-1 (0.033) (0.049) (0.051) (0.051) (0.007) (0.025) (0.025) Dependent -0.385*** -0.012** -0.048** 0.053*** 0.022 Variable, t-2 (0.049) (0.071) (0.069) (0.025) (0.034) Dependent -0.293** -0.039 0.034 Variable, t-3 (0.051) (0.069) (0.025) Dependent -0.297 Variable, t-4 (0.052) Observations 543 543 543 543 1810 1810 1810 R-Squared 0.9652 0.9691 0.9716 0.9745 0.9606 0.9607 0.9608 Adjusted R- 0.9458 0.9517 0.9555 0.9602 0.9559 0.9561 0.9561 Squared Instruments 197 195 196 196 192 193 194

6.3 Empirical Results and Discussion

Using variations of Equation 6.1, results are presented in Table 6.5. The estimates show a sizable degree of persistence in renewable energy dynamics, confirming the need for a rich dynamic model to estimate the effects. When only one or two lags are considered, renewable capacity (i.e., FFRET including only solar and wind capacity and FFRET with biomass capacity) behaves like a random walk with a

237

drift, as shown in Table 6.5. However, when a third lag is added, the dynamics of renewable capacity become less persistent. According to Table 6.5, our best specification is Model 3, which shows that the short-term effect of FBNGT (i.e., flexible natural-gas-fire technologies such as combined cycle and combustion turbine) capacity on FFRET (i.e., solar and wind) capacity is very small at 0.0304 compared to the long-term effect estimated at 0.9696. The data was read from dynamic panel-data sets and analyzed at a monthly time step. If this time step is changed to annual, the above effect is likely to change. Distinguishing between these two effects by using a general dynamic framework thus becomes a vital policy consideration that is required to support investments in DER development. For this reason, a dynamic DER framework is a vital policy consideration for long-term market development, as is the case in this analysis.

238

Table 6.5 Pooled Regression Results, Share of Renewable Installed Capacity Dependent Variable: excl. hydro, excl. wind, excl. solar, excl. hydro, excl. hydro, excl. hydro, excl. hydro, waste, biomass, hydro, waste, hydro, waste, Share of FFRET waste waste waste waste Capacity wood biomass, wood biomass, wood Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 Model 7 0.970*** 0.954*** 1.039*** 1.034*** 0.884 *** 1.006*** 0.859*** Dependent Variable, t-1 (0.051) (0.051) (0.055) (0.055) (0.054) (0.056) (0.056) 239 -0.004** -0.013** -0.043** -0.002** -0.002** -0.038 0.012 Dependent Variable, t-2 (0.072) (0.071) (0.075) (0.076) (0.073) (0.079) (0.069) -0.325 -0.297 -0.263 -0.309 -0.133 -0.070 -0.331*** Dependent Variable, t-3 (0.051) (0.051) (0.053) (0.018) (0.053) (0.056) (0.051) Share of FBNGT Capacity, 0.0304 0.0302 0.013 -0.019

t-1 (0.018) (0.018) (0.006) (0.015) Share of FBNGT Capacity, -0.058

t+1 (0.022) Share of FBNGT Capacity, -0.023

t+2 (0.028) Share of FBNGT Capacity, -0.035

t+3 (0.028)

Share of FBNGT Capacity, 0.029

t+4 (0.061) 4.658 4.167** FIT (moving average) (1.264) (01.44) 0.328 0.298 0.270*** 0.286*** 0.205 0.078 0.098* EFFRPS (moving average) (0.0037) (0.054) (0.057) (0.056) (0.061) (0.026) (0.0.039) NETMETER (moving 0.915 0.789 0.446 0.508 0.441 0.046 0.285 average) (0.172) (0.250) (0.330) (0.323) (0.032) (0.137) (0.254) PUBENFUND (moving -1.593 -1.164

240 average) (0.758) (1.204)

INTERCONSTAND 0.006 0.062

(moving average) (0.183) (0.282) LCVSCORE (moving 0.017 0.014

average) (0.014) (0.015) ELECTRICPRICE -0.0255 0.196

(moving average) (0.078) (0.085) 0.444 -2.544*** -2.599*** 4.631*** -0.583*** -2.608*** IMPORTRATIO, t-1 (2.762) (3.614) (3.358) (3.425) (1.485) (2.563) 0.342 1.690 1.749 -0.400 0.551 0.515*** PERCAPITACO2, log, t-1 (0.740) (0.808) (0.790) (0.888) (0.329) (0.636)

-0.401 -1.819*** -2.222*** -2.275*** -1.433*** 3.317*** PERCAPITAGDP, log, t-1 (2.275) (2.265) (2.212) (2.226) (1.002) (1.757) Share of Nuclear Capacity, 0.104 0.019 0.024 -0.009 0.008 0.0007

t-1 (0.015) (0.018) (0.018) (0.016) (0.007) (0.014) -0.105 -3.660*** -3.460*** -5.428*** -1.809 1.052*** ENERGSITY, log, t-1 (1.977) (2.139) (2.078) (2.109) (0.934) (1.599) Observations 543 543 543 543 531 543 543 Instruments 191 196 195 195 194 195 195

241 R-Squared 0.9702 0.9715 0.9730 0.9696 0.9741 0.9792 0.9433

Durbin-Watson Stat 2.1800 2.1877 2.1476 2.1074 2.0773 1.9444 2.1783 -3.20 -0.11 20.15 23.72 35.33 16.35 -37.78 Combined Coefficient (0.74) (26.22) (25.71) (25.11) (25.19) (11.45) (19.72) Adjusted-Squared 0.9569 0.9555 0.9580 0.9529 0.9593 0.9677 0.9120 Standard errors (clustered at state level) are in parenthesis. Asterisks denote P-values at 99% (***), 95% (**), and 90% (*).

Figure 6.1 The Monthly Solar and Wind Generation Capacity in 10 States, 2001-2016

242

Figure 6.2 Monthly Accumulation Capacity of Solar and Wind Generation, by State,

2001-2016

Figures 6.1 and 6.2 reveal monthly generation capacity and monthly accumulation capacity of solar and wind in the 10 case-study states, respectively, thereby indicating that massive amounts of wind power entered into the power

243

markets in New York, Texas, Utah, and California earlier than solar did. However, Arizona, Georgia, New Jersey, Nevada, and North Carolina recorded significant amounts of solar power compared to wind. Solar electricity generation has exhibited a high growth rate in recent years because of favorable a policy, finance, and market environment—and particularly because of the decreased production costs of solar. Moving to a more detailed explanation of the models, Model 1 presents a basic specification where only FITs, RPS, net metering, public benefits fund, interconnection standards, average score on the League of Conservation Voters Scorecard, and average electricity price are considered as determinants of deployment of fast-flexing renewable technologies. Six main conclusions can be drawn from this model, which shed light on the research questions of this study and on results reported in the literature. First, in line with recent studies on the policy effectiveness of economic instruments for wind and solar power development (Butler and Neuhoff, 2008; Byrne et al., 2016; Byrne et al., 2017; Kim et al., 2017; and Li et al., 2017), reducing entry barriers promotes deployment of renewable energy technologies. Second, both FITs and RPS—the two most commonly applied policy instruments—have significant policy effects on solar and wind-power development (e.g., by addressing high financial and transaction costs that constrain the procuring of distributed generation in conventional energy markets, lack of access to low-interest capital, and insufficient market liquidity). However, most studies generally conclude that FITs have a greater effect than RPS (Li et al., 2017). But while such academic endeavors about the policy effectiveness of RPS, FITs, power purchasing agreements, capital grants, tax incentives, and preferential loans, and the appropriate sequence of application of these

244

incentive measures to support renewable energy deployment are needed, an effort to synthesize these policy virtues into integrated platforms tailored to their context is more vital to realizing a sustained commitment to the infrastructure-scale development of distributed utilities (Byrne et al., 2016). Third, effects estimates of interconnection standards (INTERCONSTAND) and the League of Conservation Voters Scorecard (LCVSCORE) were generally positive, but small. Fourth, a state’s LCVSCORE has little effect on renewable share in that state; for example, it exhibits less than 1.5% additional renewable generation for each LCVSCORE score per MWh. Fifth, effects estimates of public benefits fund (PUBENFUND) and electricity prices (ELECTRICPRICE) were both negative but smaller for the latter. Sixth, the fixed-effects estimate of RPS policy indicates that a

1%-point increase in the state’s RPS policy produces approximately 3.8%-point increase in a state’s renewable generation share. This result confirms that measuring the effectiveness of RPS policies should take into consideration “spatial dependence” among states and should not be limited only to the effects on RE generation.55 However, empirical evidence in this respect is rather mixed. While Bowen Lacombe (2017) and Heeter et al. (2015) find that stronger RPS laws do have a positive impact on renewable generation outside those states, while other contributions analyze the effectiveness of RPS policies but not of their significant spatial dependence in electricity markets (Shrimali et al., 2012; Yin and Powers, 2010). The difference between the results put forward in these previous studies and this study relate to

55 According to Bowen and Lacombe (2017, and Lesage and Pace (2009), spatial dependence arises when the values reported in one spatial location depends on the values in neighboring locations or a region.

245

several factors which I am unable to test directly, including the choice of econometric modelling approach, the choice of states and geographical focus of the analysis (e.g., ISO/RTO/PJM region as opposed to individual states), and the choice of policy proxies, timeframe, and control variables. Model 2 presents a detailed specification in which policy variables were considered and the vector of relevant control variables added—namely, the electricity import ratio (IMPORTRATIO), and the logarithms of per capita, energy-related carbon-dioxide emissions (PERCAPITACO2), per-capita real GDP

(PERCAPITAGDP) and the state’s energy intensity of the economy (ENERGSITY). In the context of this study, the direct effect coefficient of the policy variables and the controls have a specific economic meaning. For instance, the direct effect coefficient of EFFRPS, FIT, NETMETER, PUBENFUND, and INTERCONSTAND variables is the effect of each state’s policy on its own renewable generation. This interpretation does not take into consideration the marginal effects of these laws on renewable generation in surrounding states within a specific ISO/RTO region or across the regions. Furthermore, while the short-term effects of both the policies and market regulation appear small, it is difficult to distinguish their long-term direct and indirect marginal effects on renewable generation.

6.4 Parameter Estimations of Policies and Control Variables Parameter estimations of all the policies and control variables for solar and wind power production are listed in the preferred specification presented in Model 3.

Alongside the main variable of interest, !h"#$%"&acityFFRET, Model 3 includes two key policy variables (RPS and net metering), !h"#$%"&acityFBNGT, and all the controls applied in the study. Model 3 shows that the coefficients of electricity import ratio (-

246

2.544), per-capita real GDP (-1.819), and energy intensity of the economy (-3.660), all significantly and negatively influence the development of renewable energy technologies. These results indicate that solar- and wind-power development will be discouraged by a state’s negative electricity-importing ratio (i.e., if a state consumes less electricity than it generates), by the economic growth required, and by the state’s energy intensity of the economic structure implying importance of energy on the economy. The results also indicate that flexible natural gas capacity favors solar- and wind-power investment conditional on the state’s electricity import ratios, per-capita real GDP, energy intensity of the economy, the presence of nuclear generation, and other covariates.

6.4.1 Effects of Flexible-Natural Gas Technologies on Fast-flexing Renewable Energy Diffusion

The overall coefficient of the model (20.15) is positive and statistically significant, thereby indicating that natural gas has a positive impact on solar- and wind-power development, conditional on all other covariates. For example, a 1% rise in FBNGT capacity produces a 0.0304% increase in the share of FFRET. These modeled, long-run point estimates presented here are consistent with the insights from recent technical assessments made by other practitioners. For instance, using the

Phillips-Sul convergence test and the LCOE cost-estimation approach, Nyangon, Byrne, and Taminiau (2017) showed that complementary use of solar PV and natural gas technologies is already underway. These results also confirm the insights of Weiss et al. (2013), who indicate that, while low natural gas prices are unlikely to affect investment in solar PV in the short-term due to the absence of fuel costs, in the long- run, the complementary relationship between these two markets will become

247

dominant. Again, the estimated short-term effects of the RPS policy variable is small (0.056), but its high persistence over the study period yields significantly larger long- term effects. Popp et al. (2011) developed a similar econometric model but used the first- difference transformation of renewable capacity as the dependent variable. To test the robustness of these results, Table 6.6 presents the estimation of Model 3 using the specification of Popp et al. (2011) presented in Equation 6.1:

;;<=> A ;;<=> 3!"#$%"&"45678,: = BCD @ !"#$%"&"45678,:EB + ;GHI> βShareCapacity8,:ED + KLMNO%P8,:ED + QR8,:ED + S8 + S: +∈8,: (5.4)

The short-term effects of FFRET and FBNGT capacity is the same as those obtained using Equation 6.1 (0.0304). However, our dynamic econometric model has the advantage of allowing for assessment of both short- and long-term effects.

Table 6.6 Estimated Results of Renewable Installed Capacity Using Popp et al. (2011) Specification excl. hydro, waste, Dependent Variable: excl. hydro, waste Δ Share of RENEWABLE biomass / wood (FFRET) Capacity Model 1 Model 2 Dependent Variable, t-1 0.03*** (0.057) 0.027*** (0.059) Dependent Variable, t-2 -0.034 ** (0.079) -0.058** (0.082) Dependent Variable, t-3 -0.258** (0.056) -0.237*** (0.057) Share of FBNGT Capacity, t-1 0.079* (0.019) 0.076* (0.019) EFFRPS (moving average) 0.299 (0.053) 0.286 (0.061)

248

NETMETER (moving average) 0.449 (0.256) 0.324 (0.351) IMPORTRATIO, t-1 -5.315 (2.247) -6.689 (3.842) PERCAPITACO2, log, t-1 1.976 (0.815) 2.169 (0.858) PERCAPITAGDP, log, t-1 -1.049 (2.388) -1.082 (2.403) Share of Nuclear Capacity, t-1 0.041 (0.017) 0.043 (0.019) ENERGSITY, log, t-1 -2.011 (1.429) -2.262 (2.257) Observations 543 543 Instrument Rank 193 195 R-Squared 0.5575 0.5620 Coefficient 7.358 7.676 Durbin-Watson Stat 1.7541 1.7711

Note: Standard errors are in parenthesis. Asterisks indicate p-values of <0.1 (***), <0.05 (**), and <0.01 (*). All models include month dummies.

In Model 4, alongside hydro and waste, biomass56 was also excluded from the specification presented in Model 3. Unlike wind and solar energy based distributed generation units which are non-dispatchable, biomass plants (e.g., traditional generators, integrated gasification combined cycles, internal combustion engines, or combined heat and power) are considered dispatchable (Tanwar and Khatod, 2017).57 In addition, biomass-based distributed generation units can be burned and often co- fired by mixing with coal and gas to provide a constant and predictable source of

56 Biomass incorporates wood-fueled co-generation and biogas.

57 Unlike non-dispatchable generation systems, dispatchable renewable distributed generators can be stored and burned when needed.

249

power generation. Hence, including biomass in the definition of renewable energy could result in biased estimates (Kai et al., 2014). To address this concern, Model 3 specification was re-estimated by subtracting biomass capacity from the definition of the dependent variable to develop Model 4. The result in Model 4 shows a small decline in the effect of FBNGT capacity on renewable energy (i.e., from 0.0304 to 0.0302). In the long-run, this effect is close to unity. Therefore, the effect of FBNGT on biomass does not influence renewable energy deployment. In Model 5, the robustness of the results was tested by replacing

;;<=> !"#$%"&"45678,:ED with leads of FBNGT capacity, from t+1 to t+4. The objective of this additional step is to test the accuracy of the aggregate effects of FBNGT capacity compared with the individual ones by using lags of our variable of interest (flexible-baseload natural gas technologies). Models 1 to 4, implicitly assumed that agents form their expectations according to an adaptive rule; that is, the model is a function of both past realization and past expectations. This implies that there is a collective rationality of the estimated effects instead of individual rationality (Chow, 2011). In this regard, the positive significant coefficients of FBNGT signifies that agents are forward looking. However, this would weaken the GMM estimation strategy that is applied in this study. Reassuringly, leads t+1 to t+3 of the coefficients associated with FBNGT in Model 5 are negative and insignificant (i.e., -0.058, -0.023, and -0.035) and lead t+4 is not significant (0.029). This confirms that the model in Equation 6.1 is well-defined. To test the robustness of the model further, individual simulations of the effects of flexible natural technologies on solar power (Model 6) and wind power (Model 7) development were developed. The cumulative effect of past solar- and

250

wind-power capacity on their current generation capacity is 0.987 and 1.019, respectively. This implies that as in models 3 and 4, the short-term effect of FBNGT capacity on solar is very small but significant in the long-term. For wind power, the short-term effect of FBNGT on wind and the overall coefficient is negative (-37.78, and -0.0195, respectively), indicating again that unlike solar, a state’s high reliance on natural gas could induce a negative impact on wind power development. This insight is significant in two respects, especially when viewed in the context of whether wind power development should continue to benefit from tax incentives. First, while the cost of wind has reduced significantly and is lower than some fossil energy, such as oil and natural gas, its output power is variable and unpredictable (Lazard, 2016). Wind’s physical properties—its variability, location relative to demand centers, and forecast uncertainty—create technical challenges for existing systems (Li et al., 2017). Also, unlike solar electricity systems, higher wind penetrations often create economic conflicts with thermal generators, tied to cost recovery (Dong, 2012). For instance, the main economic instrument used to incentivize wind energy development is FITs. However, FITs only fix the price but not the quantity of the wind power generated. When the price of FIT is significantly high, wind generators are likely to increase installed capacity. As a result, the generators are forced to reduce the operating hours and such a move could degrade a “Goldilocks range” of natural gas and wind prices (Nyangon, Byrne and Taminiau, 2017). This physical and economic displacement of thermal generators takes place predominantly through the dispatch merit-order because wind has low marginal costs compared to solar. Second, wind power development is less impacted by serious financial deficits because it has relatively fixed generating costs.

251

Nevertheless, the estimated results of models 1 through 7 presented above do not resolve the endogeneity concerns of the long-term effect of natural gas on renewable energy diffusion. For instance, time-varying shocks ranging from changes in policy and their effectiveness and changes in economic and technology-related variables can have significant effects on both natural gas and renewable energy investments. However, following Burnett and Madariaga (2017), for panel data sets with short-time dimensions as in this study, the two-step system estimator applied here effectively resolves any potential biases created from the correlation between the state-

;;<=> level fixed effects, S8, the lagged !"#$%"&"45678,:ED values, and the endogeneity of other explanatory variables. This approach thus resolves a common criticism of system GMM methodology concerning a potential bias that arises when there are “too many instruments” compared with the number of observations, such that instruments overfit the endogenous regressors and thereby weaken the Hansen J test of the instruments’ joint validity (Roodman, 2009). Therefore, following Roodman (2009), the number of lags used to build the instruments and the instrument variables in the model specification were reduced to focus only on relevant policy variables and controls. Considering that the number of observations, N, in our model is small (10 states), the p-values of the tests for second-order serial correlation linked to the

Hansen J tests are unbelievably good. Table 6.7 shows the re-estimated results with a smaller number of lags and instrument variables of models 1 and 2 (i.e., with and without biomass) and models 3 and 4 (i.e., when the monthly fixed effects are replaced with a time trend to calculate possible values of the Hansen’s test).

252

Table 6.7 Re-estimated Results of Renewable Installed Capacity with Smaller Number of Lags excl. hydro, waste, excl. hydro, excl. hydro, waste, Dependent Variable: excl. hydro, waste Share of FFRET biomass, wood waste biomass, wood Capacity Model 1 Model 2 Model 3 Model 4 1.039*** 1.063*** 1.036*** 1.046*** Dependent Variable, t-1 (0.055) (0.055) (0.044) (0.044) -0.043 -0.057*** -0.104*** -0.091*** Dependent Variable, t-2 253 (0.075) (0.074) (0.061) (0.062)

-0.263 -0.279 -0.205 -0.232 Dependent Variable, t-3 (0.053) (0.053) (0.041) (0.043) Share of FBNGT Capacity, 0.0317** 0.0306** 0.005 0.0036** t-1 (0.017) (0.017) (0.012) (0.012) 0.270** 0.283** 0.242* 0.244** EFFRPS (moving average) (0.057) (0.055) (0.041) (0.039) NETMETER (moving 0.440** 0.498 0.219** 0.2198 average) (0.330) (0.322) (0.215) (0.209) -3.313 -3.143 4.117 3.991 IMPORTRATIO, t-1 (3.278) (3.205) (2.269) (2.212)

1.736 1.765 0.079 0.098 PERCAPITACO2, log, t-1 (0.803) (0.784) (0.580) (0.565) -2.029 -2.382 3.219 3.206 PERCAPITAGDP, log, t-1 (2.226) (2.174) (1.418) (1.383) Share of Nuclear Capacity, 0.021 0.024 -0.009 -0.006 t-1 (0.017) (0.017) (0.013) (0.012) -3.759 -3.542 -0.030 -0.006 ENERGSITY, log, t-1 (2.130) (2.068) (1.422) (0.012) Observations 543 543 543 543 254 Instruments 195 195 15 15

R-Squared 0.9729 0.9697 0.9585 0.9539 Adjusted R-Squared 0.958 0.953 0.9575 0.9528 J-Statistic 12.7754 14.008 20.5989 22.0877 Hansen J Statistic 6937 7606 11185 11994 Hansen Critical Probability 3.51E-04 1.82E-04 6.00E-06 3.00E-06

Note: Standard errors are in parenthesis. Asterisks indicate p-values of <0.1 (***), <0.05 (**), and <0.01 (*). Models 1 and 2 include month dummies while Model 3 and 4 include a quadratic time trend.

Two major conclusions can be drawn from the results presented in Table 6.7. First, except in Model 3, the coefficients associated with the variable of interest

00123 !"#$%"&"'()*+,-./ are roughly similar to those in Table 6.6. As a result, this shows that a higher persistency in the panel data set of renewables makes it hard to compute a reasonable long-term effect. Second, the combined effect of the lagged terms is slightly larger and closer to one than in Table 6.7. Lastly, the hypothesis that the instrumental variables are not correlated with the error terms (Hansen J test) and the hypothesis that the residuals have no second-order autocorrelation (Arellano–Bond tests) are both realized.

6.4.2 Effects of Fast-flexing Renewable Energy Technologies on Flexible-Natural Gas Diffusion

Thus far, the results presented above have focused on the effect of FBNGT capacity on FFRET development and diffusion. However, these model specifications assume that there is no effect of fast-flexing renewable electricity generation on natural gas diffusion. In addition, the formulation may result in an assumption that the drivers of natural gas and renewable energy investments discussed in Chapter 3 are separate and different from each other. To test and study these determinants, Equation 6.1 was reformulated with the share of FBNGT capacity as the dependent variable.

Table 6.8 presents estimated result of the determinants of FBNGT diffusion,

0?@A3 using both lags and leads of the dependent variable ShareCapacity+,-./ .

255

Table 6.8 Empirical Results of Share of Flexible-Baseload Natural Gas Installed Capacity Dependent Variable: Model 1 Model 2 Model 3 Model 4 Share of FBNGT Capacity 0.790*** 0.784*** 0.801*** 0.772*** Dependent Variable, t-1 0.032 (0.039) (0.018) (0.174) -0.082 Share of FFRET capacity, t-1 (0.042) -0.268 Share of FFRET capacity, t+1 (0.088) -0.155 Share of FFRET capacity, t+2 (0.122) 0.157 Share of FFRET capacity, t+3 (0.085) -1.272 -0.302 FIT (moving average) (3.400) (3.845) -0.305 0.056 0.019 0.004 EFFRPS (moving average) (0.114) (0.121) (0.028) (0.030) -0.912 0.799 NETMETER (moving average) (0.485) (0.629) 2.784 PUBENFUND (moving average) (3.242) INTERCONSTAND (moving 0.678 -0.022 average) (0.826) (0.766) -0.033 -0.025* -0.024** -0.016*** LCVSCORE (moving average) (0.038) (0.039) (0.010) (0.009) ELECTRICPRICE (moving 0.224 0.182 average) (0.215) (0.222) 14.261 -1.802 -1.895*** IMPORTRATIO, t-1 (7.858) (1.234) (1.224)

256

-4.369 -4.782** -4.285* PERCAPITACO2, log, t-1 (2.148) (1.189) (1.268) -2.469* 0.934** 1.945* PERCAPITAGDP, log, t-1 (6.184) (1.609) (1.647) 0.057 0.084 0.065 Share of Nuclear Capacity, t-1 (0.044) (0.019) (0.019) 3.869* 0.846 -0.633** ENERGSITY, log, t-1 (5.389) (1.800) (1.863) Observations 543 543 1810 1780 Instruments 190 194 199 198 R-Squared 0.9708 0.9716 0.9649 0.9664

Note: Standard errors of the share of FBNGT installed capacity are clustered at the level of state and are presented in parenthesis. Asterisks indicate p-values of <0.1 (***), <0.05 (**), and <0.01 (*). All models include month dummies.

Tables 6.8 summarizes results of different variations of Equation 6.1, with flexible-baseload natural gas capacity as the dependent variable. The value of coefficients of the leads and lags of the dependent variable are not significant, leading to the rejection of the null hypothesis that renewable energy capacity affects FBNGT diffusion. Similarly, the measure of environmentalism (LCVSCORE), a proxy of environmental policies, has no effect on the change in FBNGT capacity. These facts point to the little attention investors in natural gas market have paid to environmental policies and the installed capacity of renewable energy. While there is a general understanding that deep decarbonization of the electricity system has multiple economic benefits beyond those of environmental sustainability, the pace of deployment of critical components of decarbonization strategies (such as zero-carbon

257

variable energy resources—e.g., nuclear, natural gas, solar, wind, hydropower, biomass, and geothermal sources) and the deployment of clean-energy and DER technologies remains slow compared to the challenge. Moreover, apart from addressing the central challenge of climate change and mitigating its impacts, other environmental concerns critical to the power sector, including how to reduce conventional pollutants, manage the energy-water nexus, and mitigate land-use change and the likely impacts of siting electricity generation, transmission, and distribution assets, coordination of all these activities poses a serious challenge (Nyangon, Alabbas, and Agbemabiese, 2017; U.S. Department of Energy,

2017b). This natural-gas-renewables-partnership also emphasizes a “asymmetric” logic of complementarity between flexible-baseload natural-gas-fired technologies and fast- flexing renewables, with the former supporting investments and deployment for the later, but not vice versa. There are potential complementarity benefits of developing these energy sources together. However, this would be difficult to address in an appropriate manner with the data at hand. For instance, baseload natural-gas-fired units have fast-start capabilities and high-capacity factors in the medium and long term; but they also exhibit high price volatility, unlike solar PV. These characteristics make solar and natural gas intrinsically complementary.

In the context of a long-run decarbonization strategy, reliability, and efficiency in the electricity sector, the result in Table 6.9 also highlights salient facts about natural gas and renewable energy sources in the electricity sector. First, the main drivers of flexible-baseload natural gas investments are the electricity importation ratio, growth in population and CO2 emission rate (per capita energy-related carbon- dioxide emissions), size of the economy (per-capita real GDP), and electricity

258

intensity (size of energy intensive sectors in the economy). As shown in Figure 6.3, per-capita energy-related CO2 emission is declining at the state level and nationally. This decline is largely because of the decreased use of coal and the increased use of natural gas for electricity generation.

Figure 6.3 The Annual Accumulation Capacity of Solar and Wind Generation in 10

States, 2001-2016

259

Second, the results show that the FBNGT market is more volatile than the renewable energy market. This market volatility is not limited to prices but also affects other areas, including employment, and changes in commodity prices. For instance, between January of 2015 and August of 2016, the oil and natural gas extraction industry lost approximately 28,000 jobs or 14% of its workforce (U.S. Department of Labor, 2017). These job dislocations not only impact the immediate families directly; they also extend to the communities and towns that depend on proximate infrastructure and services like roads and schools.

6.4.3 Effects of Generation Technologies (Other Than Natural Gas) on Renewable Energy Diffusion

To compare the effect of other technologies on renewables other than FBNGT, shares of the capacity of hydro (HYDRO), coal (COAL), and other baseload technologies—including internal-combustion engines, diesel generators, steam turbines, and other fossil-based technologies (BLFT)—were added on the right-hand side of the base specification, one at a time. The reason for adding the shares of capacity of each technology one at a time is that these variables exhibit high collinearity (i.e., all shares sum to one). Therefore, HYDRO was added to the base model first (models 1 and 2), followed by BLFT (models 3 and 4), and lastly by

COAL capacity (models 5 and 6). Empirical results of the estimates are shown in Table 6.9. These technologies play different roles in the power system. Hydropower is one of the oldest carbon-free baseload and flexible forms of electricity generation that is used to provide a range of essential grid reliability services (e.g., high ramping capability to compensate fluctuations in demand and supply, spinning and supplemental reserves, reactive power and voltage support, and black-start capability).

260

On the other hand, the objective for perturbing our base model by adding coal- and oil-fired power-generation capacity was to test the assumption that only flexible natural-gas-fired technologies support renewable energy investment.

Additionally, based on the Energy Information Administration’s business-as- usual estimates, by 2040, 74% of grid-connected power in the United States will be sourced from natural gas, coal, and nuclear (U.S. Energy Information Administration, 2016a). Therefore, flexible natural-gas-fired generators could be a game changer as the efficiency of these technologies improves, thereby accelerating the transformation in the electric-grid system and affecting the need for traditional baseload in the long run. Overall, Table 6.9 confirms the previous findings: including additional shares of variable technologies to the base specification does not alter capacity estimates associated with the flexible, natural-gas-fired plants (i.e., the combined cycle used in peaking and baseload, and the combustion turbine used mostly as peaker plants). The results also corroborate the earlier conclusion that the feedback effect between fast- flexing renewables (i.e. wind and solar) and capacity in other electricity-generation technologies does not exist. It is recommended that new tools and operational solutions for managing distributed energy generation systems including DER business models needs to be considered to provide new interconnections required in managing grid flexibility and the increased complexity associated with intermittent renewable electricity and controllable load.58

58 Low-or zero-carbon-emitting and high capacity factor generation plants mostly associated with natural gas-fired units can be optimized in the grid to reduce reliability risks.

261

Table 6.9 Empirical Results of Additional Share of Other Technologies excl. hydro, excl. wind, excl. hydro, excl. hydro, excl. hydro, waste, excl. hydro, hydro, waste, waste, biomass Dependent Variable: waste waste biomass / waste biomass / / wood Share of RENEWABLE wood wood (FFRET) Capacity HYDRO Capacity BLFT Capacity COAL Capacity Model 1 Model 2 Model 3 Model 4 Model 5 Model 6 1.038*** 1.049*** 1.051*** 1.074*** Dependent Variable, t-1 1.062*** 1.067*** (0.055) (0.055) (0.057) (0.057) (0.055) (0.055) 262 -0.040** -0.042** -0.034** -0.049** Dependent Variable, t-2 -0.056** -0.057** (0.075) (0.076) (0.075) (0.076) (0.075) (0.076) -0.0262 -0.279 -0.263 -0.279 -0.269 -0.286 Dependent Variable, t-3 (0.053) (0.053) (0.053) (0.053) (0.052) (0.053) Share of FBNGT Capacity, 0.023* 0.025* 0.033 0.031*** 0.050*** 0.049** t-1 (0.023) (0.022) (0.019) (0.018) (0.021) (0.021) Share of HYDRO capacity, -0.011 -0.007

t-1 (0.018) (0.017) Share of BLFT capacity, t- 0.010 0.005

1 (0.018) (0.017) Share of COAL capacity, t- 0.033 0.029

1 (0.018) (0.017) 0.252*** 0.270*** 0.253*** 0.275*** 0.242*** 0.258*** EFFRPS (moving average) (0.065) (0.063) (0.065) (0.063) (0.059) (0.057) NETMETER (moving 0.415* 0.481* 0.417* 0.489* 0.349* 0.415* average) (0.333) (0.325) (0.335) (0.326) (0.333) (0.326)

-3.282 -3.119 -2.684 -2.629 -1.717 -1.793 IMPORTRATIO, t-1 (3.281) (3.209) (3.622) (3.546) (3.632) (3.552) 1.602 1.670 1.566 1.671 1.112 1.201 PERCAPITACO2, log, t-1 (0.833) (0.815) (0.837) (0.817) (0.864) (0.844) -2.021 -2.377 -1.800 -2.214 -3.001** -3.299** PERCAPITAGDP, log, t-1 (2.228) (2.176) (2.267) (2.214) (2.343) (2.288) Share of Nuclear Capacity, 0.019 0.024 0.029 0.028 0.043 0.044 t-1 (0.018) (0.018) (0.024) (0.024) (0.022) (0.021) -3.918*** -3.657*** -3.828 -3.556 -4.096 -3.874 ENERGSITY, log, t-1 (2.148) (2.087) (2.159) (2.102) (2.145) (2.084) Combined Coefficient 23.32 26.17 19.84 23.69 33.48 35.90 Observations 543 543 543 543 543 543 263 R-Squared 0.9730 0.9697 0.9730 0.9697 0.9732 0.9700

Notes: Standard errors of the share of FBNGT installed capacity are clustered at the level of state and are presented in parenthesis. Asterisks indicate p-values of <0.1 (***), <0.05 (**), and <0.01 (*). All models include month dummies.

6.4. Successful Business Models to Promote NG-RE Blended Power Generation From the above econometric analysis of the determinants of investments in variable renewables and the role of flexible natural gas technologies in a sample of the top-10 solar-producing states over the years 2001-2016, I conclude the following. First, as shown in Figures 6.4 and 6.5, states in which natural-gas-fired combined- cycle (NGCC) generators had capacity factors of between 25%-58% and a FBNGT capacity of at least 5% were more likely to invest in renewable energy generation, other factors being equal. As previous results indicate, the short-run effects of FBNGT technologies on FFRET capacity are low, but they are significant in the long-term (nearly a one-to-one increase or 0.707%).

Data Source: (U.S. Energy Information Administration, 2016b)

Figure 6.4 NGCC Capacity Factors by State

264

Figure 6.5 Comparison of Capacity Generation of Natural Gas and Renewable Energy

Generation, 2001-2016

Second, the analysis supports the conclusion that flexible-baseload natural-gas- fired technologies have supported investments in and deployment of fast-flexing renewables by hedging against variable supply and by providing dispatchable backup capacity. The integration of intemmittent renewables increases the need for fast- ramping natural gas-fired electricity generation and system flexibility options such as demand response and energy storage. The analysis presented here highlights a highly complementary relationship between variable RE and flexible NG-power generation systems, thereby calling attention to the need for integrated resource planning of these

265

vital energy forms and DER-centric utility-rate structures such as joint transmission investments, co-location, and co-installation of variable RE and flexible NG for efficient grid management.

6.4.1 Non-Wires Alternatives Model Analysis of the ten states shows that it is easy for traditional IOUs, especially in deregulated jurisdictions—to maintain grid reliability either on their own or through cooperative ventures such as voluntary pools—as generation assets are separated from either transmission or distribution functions. To mitigate the effects of increased information asymmetries that is likely to result during this structural separation process, policymakers, and regulators must therefore consider adopting non-wires alternatives (NWA) business model and other innovative tools to identify the new efficient frontier for investment in network operations, and maintenance of the electric grid assets. As noted by Jenkins and Pérez-Arriaga (2017, p 4):

Regulators need remuneration mechanisms that align incentives for utilities to both efficiently accommodate DERs and take advantage of new capabilities to reduce network costs and improve service quality. This requires equalizing incentives for savings in both capital expenditures (CAPEX) and operational expenditures (OPEX) so that utilities will, for example, pursue cost-effective active system management or “non-wires” solutions to improve costs and performance. Finally, as uses of the distribution network evolve, regulators will need to manage greater uncertainty, including both benchmark and forecast errors. Non-wires solutions have the potential to improve grid reliability, defer capital infrastructure, and reduce customers’ electric bills. Table 6.10 shows examples of non- wires opportunities to create such change as part of the efforts at the Consolidated Edison to build and improve transmission as a profit center through integration of smart grid technologies.

266

Table 6.10 Non-wires Alternatives Projects Participating Estimated Project Project Name Project Description State Period Utilities or RFP Type/Capacity organizations Timing Reference Promotes geo-targeting as a source of non-system Estimated net benefits charge funding. lifetime Seeks to increase energy electric savings 267 savings yield per program in 2017, 2018, dollar spent. It focusses on and 2019 are lifetime energy savings per 11,174 GWh, dollar spent (rather than first 11,476 GWh, New Jersey year savings). Increase the and 11, 849 Clean Energy Applied Energy use of private capital for GWh, New Programs AEG Group (AEG) clean energy investments. respectively. Jersey (NJCEP) (2017)

Seeks to identify, evaluate, and deploy a portfolio of customer-side solutions and Designed to non-traditional utility-side reduce the Brooklyn Queens solutions (e.g., DER storage overload of Consolidated Demand System, conservation feeders by 69 Edison NWA Management voltage optimization, fuel MW by New Company of Currently BQDM Project (BQDM) cell and solar panels) summer 2018. York 2016 New York, Inc. Underway (2016) This project runs from 2017-2021 and its goal is to

268 Coldenham / defer a distribution feeder New Distribution upgrade, with 1 MW of load Load Relief Central Hudson Mar 2017 York Feeder Upgrade relief needed by 2019. (NWA Central Solutions are currently Dec. Currently Hudson being evaluated. 2019 Underway) (2017) The goal of the Boothbay Designed to pilot is to create a hybrid Maine Power address energy solution that includes Reliability efficiency, Central Maine investments in a Maine 2016 Program—The demand Power Co. transmission system and Boothbay pilot response, and procurement of non- GridSolar DG challenges. transmission resources. (2016)

Nov 2014 Merritt Park / An equivalent of 1 MW of New May (NWA Central Distribution load relief is needed by Load Relief Central Hudson York 2019 Currently Hudson Feeder Upgrades 2019. Underway) (2017) Northwest An equivalent of 10 MW of Nov 2014 Corridor / load relief is needed by New May (NWA Central Load Relief Central Hudson Transmission 2019 to successfully defer a York 2019 Currently Hudson Upgrade transmission upgrade. Underway) (2017) Solicitation for the project occurred in the fourth

269 quarter of 2014. To Philips Road / New May successfully defer a new Load Relief Central Hudson Nov 2014 Substation York 2018 substation, an equivalent of (NWA Central 5 MW of load relief is Currently Hudson needed by 2018. Underway) (2017) This NWA project seeks to provide demand side management for transmission and New Load Relief distribution system for load York Consolidated Primary Feeder relief. A reduction of 4 MW Edison Con Relief - Columbus from the network by May Company of Summer Edison Circle summer 2021is expected. 2017 New York, Inc. 2021 (2017)

Is expected to provide demand side management for transmission and distribution system load New Consolidated Load Relief relief. An estimated York Edison and reduction of 7.1 MWs from Orange & Con Primary Feeder the network by summer May Rockland Summer Edison Relief - Hudson 2021 is projected. 2017 Utilities 2022 (2017) Explores forward-looking solution to the complexity

270 and implications of Consolidated Relief of peak New incorporating DER into the Edison, and Ongoing. loadings York Primary Feeder distribution system to Southern Solicitation Con Relief - address distribution loading Q3 California expected Edison Williamsburg issues. 2017 Edison by Q3 2017 (2017)

The goal of the proposed distribution system Central Hudson business model Gas & Electric, is to create a Consolidated Supplemental common Edison, NY Distributed System standard State Electric & Implementation between Gas, Niagara (SDSIP) Plan, Load relief and reliability. utilities in Mohawk Case 16-M-0411 Power quality. Conservation determining Power, National 271 In the Matter of voltage reduction. the suitability Grid, Orange & Distributed System Resiliency. Damage failure. of a project for Rockland Implementation Asset condition. New a non-wires Utilities, Plans. Joint business model and service alternatives New Rochester Gas SDSIP Utilities Report upgrade. solution. York 2016 & Electric (2017)

Water Street Substation is in New Brooklyn and it will need Load Relief Con York Water Street approximately 4 MW of Q4 Consolidated Edison Cooling Project load relief in the year 2019. 2017 Edison 2019 (2017)

As variable RE resources expand to play a larger role in power generation, new transmission and distribution lines will be required to connect wind and solar installations to the existing grid. The demand for new transmission and distribution lines will become vital for the development of a variety of distributed energy resources. To the extent that a utility divests generation assets and is only responsible for managing transmission and/or distribution functions, it becomes a “wires only” company. The wires-only utility business model can be classified into two main categories: a high voltage transmission system operator (TSO) and a lower-voltage distribution company or distribution system operator (DSO). In this model, RTO/ISOs will continue to manage capacity markets to ensure that enough electricity is available to meet the growth in market demand. Currently, transmission lines are owned by private utilities under terms set by RTO/ISOs as established by each regional organization and FERC. Within this structure, TSO can be an active participant, providing transmission lines and services, load-balancing, frequency modulation, load dispatch, and reliability. To ensure the system remain in balance and operate reliably, a clean interface between TSO and DSO is required. Therefore, the TSO are likely to remain responsible for aggregating enough electric capacity to be sold to customers as well as maintain the high voltage portion of the grid.

On the other hand, DSOs will play a more significant role—obtaining power from the regional TSO and other power providers and maintaining responsibility for customer demand and reliability of the local grid. The DSOs directly connect to the end users. A variation of DSO is evolving in New York to create an environment crucial for maintaining supplies, grid reliability, security, and environmental obligations. This model is also called the Distribution System Platform Provider or

272

DSPP. As discussed in the next chapter, the department of Public Service Commission (PSC) will oversee the utility through outcome-based incentive regulation while DSPP will own and operate all distribution utility assets in the New York. Figure 6.6 and 6.7 show DSPP and DSO models, respectively.

Source: NYPSC (2014); Kauffman and Zibelman (2015)

Figure 6.6 The DSPP Model

273

Source: NYPSC (2014); Zibelmen (2016)

Figure 6.7 The DSO Model

6.4.2 Utility as a Smart Integrator The wires-only TSO and DSO closely represents one segment of the traditional electric utility-transportation. Another model of the utility of the future that is emerging is known as a smart integrator. The smart integrator will be more diverse in provision of its products and services while operating in regulating in regulated market

274

and other competitive markets. The hallmark of a smart integrator model is that its revenue will be decoupled from electricity sales and it will be expected to fulfil energy efficiency and other environmental obligations. The primary business rationale for the smart integrator is to bring innovative technologies to the energy system to satisfy the multiple goals involved with a clean power future. At the firm-level, the smart integrator creates partnerships between utilities and innovative energy firms to bring new technologies and survives online through new business practices and processes. While traditional utilities would continue to either generate or transport electricity or both, the smart integrator will facilitate those transactions. The smart integrator can also facilitate adoption of new regulatory regime and interconnections between new technologies and new generation sources into the systems and the existing grid. This would be particularly vital in states that have high fossil generation composition such as Texas and Utah.

6.4.3 Electric Services Operator Model The electric services operator (ESO) most closely resembles the traditional investor-owned utility. ESO will preserve and extend and extend core capabilities of generating and delivering electricity, including identifying new technologies and exploring a variety of business opportunities to succeed in the market. The ESO will retain aspects of vertical integration and its business responsibility will be to provide electricity at a low-cost, reliable manner to customers within a large service territory. However, while the ESO may own generation and other assets, it will also be required to open access and purchase or transmit power from a variety of providers. In this regard, the regulation governing the operations of an ESO will differ from that of the investor-owned utility. The ESO will offer new services and products that are properly

275

priced and aligned with regulatory incentives, including incorporating DG and DER resources. In this regard, in other markets, ESOs are likely to maintain their vertically integrated role in the power sector but the focus of regulation will be on rewarding performance as opposed to rewarding increased sales. The ESO model will be recognizable to state regulators, and therefore more attractive because the ESO continues to own the generation assets that provide its supply but can also purchase or transmit power generated by others attached to its grid. However, although the ESO “still has an incentive to build its rate base through new capital projects, the nature of investments [will shift] significantly from bulk power production and delivery to investments that promote smart and resilient grids, centralized management of DERs and price-responsive consumers” (Cochran, et al., 2014; Zinaman et al., 2015). A variation of this model is likely to emerge in states like Georgia, Arizona, Massachusetts, Nevada, Texas, New Jersey, and Utah. Table 6.11 summarizes alternative business models for utilities that have been identified for the ten states discussed above. These proposals range from mechanisms which attempt to decouple sales from a utility’s revenues to those that support increased energy efficiency initiatives and programs.

276

Table 6.11 Alternative Utility Business Models Proposed or Implemented by States

Type of Mechanism the Docket/Order Addresses to State Utility Component Utility Rate Case or Order Promote NG-RE Blended Power Generation Docket No. G-01551A-10- Southwest Gas applies for a revenue increase of about 0458 approved by the $73.2 million, based on a proposed capital structure Arizona Corporation consisting of 52.3% common equity and 47.7% long- Commission (ACC) on term debt. Customer interface: pricing structure (revenue 277 December 13, 2011 (ACC, The Docket proposed a partial revenue decoupling 2011). model); information and mechanism consisting of a lost fixed cost recovery insight. (“LFCR’) component and a weather component.59 This provision would ensure a full revenue decoupling Value network: partners mechanism from electricity sales for Southwest Gas. such as investor-owned utilities Docket No. E-01345A-11- 0224 (for Arizona Public The ACC approved a lost revenue adjustment Core strategy: business Service Company) approved mechanism (LRAM) for the Arizona Public Service mission; product / market AZ on May 24, 2012. Company and for Tucson Electric Power Company. scope.

59 The LFCR component allow the utility to recover lost base revenues attributable to achievement of the Commission’s required annual energy saving as per the adopted Gas Utility Energy Efficiency Standards (A.A.C. R14-2-2501 et seq.)

Docket No. E-01933A-12- Value network: partners 0291 (for Tucson Electric e.g., investor-owned Power Company) approved utilities in June 2011. In 2001, California passed Customer interface: Section 739.10, requiring pricing structure (revenue that California Public model); information and Utilities Commission insight. Section 739.10 proposes revenue decoupling for (CPUC) resume revenue Pacific Gas & Electric, Southern California Edison, Value network: partners decoupling. and San Diego Gas & Electric. e.g., investor-owned

278 utilities Decision 13-09-023 that Strategic resources: core

was approved in September Provision for new energy savings and performance competencies; strategic CA 2013. incentive (ESPI) for investor-owned utilities. assets; and core processes. Georgia Code (O.C.G.A. § Customer interface: 46-3A-9) pricing structure (revenue The Code authorizes utilities to recover costs and an model); information and GA “additional sum” for approved programs.60 insight.

60 The Commission approved an additional sum of 8.5% of actual net benefits of electricity savings in 2013. This amount applies when electric utilities achieve 50% or more of energy savings. Currently, Georgia Power has 12 certified energy efficiency programs.

In the 2010 IRP, the Commission approved seven energy efficiency programs including five residential and two commercial programs. Georgia Power currently has 12 certified energy efficiency programs.

DPU Docket 07-50-A was The Docket target revenue decoupling for all gas and Customer interface: approved in July 2008. electric utilities. The Massachusetts DPU has also pricing structure (revenue approved decoupling plans for National Grid Electric model); information and Company (DPU 09-39), National Grid Gas Company insight. (DPU 10-55), Bay State Gas Company (DPU 09-30) Value network: partners and Western Massachusetts Electric Company (DPU e.g., investor-owned

279 10-70). utilities

DPU Order 11-120-A The Docket recommends amends shareholder Customer interface: incentives structure for electric and gas utilities. The pricing structure (revenue shareholder incentive is based on a combination of model); information and elements: benefit-cost analysis, energy savings, and insight. MA market transformation results. Public Utilities Commission Strategic resources: core of Nevada (PUCN) Docket competencies; strategic No. 14-10018 The Docket provides for first recovery of lost assets; and core processes. revenues from DSM programs for NV Energy (parent company of Nevada Power and Sierra Pacific Power Value network: partners NV Companies). e.g., investor-owned utilities

PUCN Docket No. 07- The Docket allows gas utilities to decouple their 06046 and Nevada Admin. profits from their sales within one year of the approval Code 704.953 (2008) of their energy efficiency programs. The rules also Customer interface: specify a revenue-per-customer system for pricing structure (revenue determining utility revenues to recover fixed costs. model); information and Under the Docket and the Code, electric gas utilities in insight. Nevada can choose to either implement decoupling or use a performance. SB 358 (2009) The SB 358 was passed in 2008 and the legislation removed financial disincentives faced by the utilities.

280 Docket No. 14-10018 In 2010, the PUCN approved a Lost Revenue Adjustment Mechanism for electric utilities. In 2015,

the PUCN proposed Docket No. 14-10018, offering a new multiplier method for lost revenue by the electric utilities. New Jersey Board of Public The Dockets replaced existing weather normalization Strategic resources: core Utilities (BPU), Order BPU clauses with a conservation incentive program in order competencies; strategic Docket Nos. GR05121019 to capture variations in gross margins of weather and assets; and core processes. NJ and GR05121020 of 2016 customer usage.61

61 Weather normalization clauses mitigate the financial effects of weather on utilities and their customers.

Docket Nos. Eo07030203 In New Jersey, utilities are not permitted to collect lost and Eo10110865. revenues. The New Jersey Office of Clean Energy (OCE) selected market managers to run energy efficiency programs in the state. The market managers Strategic resources: core were eligible to receive a performance incentive. In competencies; strategic 2011, OCE recommended (in Docket Nos. assets; and core processes. Eo07030203 and Eo10110865) elimination of performance incentive and a significant reduction of budgets for market managers. Order Cases 03-E-0640 and Customer interface:

281 06-G-0746 of 2007. Requires electric and gas utilities to file proposals for pricing structure (revenue revenue decoupling mechanisms in ongoing and new model); information and

rate cases. insight. NYPSC (2008) In 2008, NYPSC established incentives for energy Strategic resources: core efficiency programs for electric utilities under the competencies; strategic Energy Efficiency Portfolio Standard (EEPS) assets; and core processes. proceeding. Case 14-M-0101 Strategic resources; In 2014, NYPSC initiated the REV initiative to assess Customer interface; Value potential for major changes to the state’s utility network; and Core NY regulatory structure. Strategy North Carolina Utilities NCUC approves incentives for public utilities to adopt Strategic resources: core Commission (NCUC) and implement new energy efficiency programs competencies; strategic NC statute through a DSM and/or energy efficiency rate rider. assets; and core processes.

Docket No. E-7, Sub 1032 Customer interface: pricing structure (revenue Approved in October 2013, the Docket provides a cost model); information and recovery mechanism to Duke Energy—in the form of insight. a shared savings model offering recovery of program costs, lost revenues (up to 36 months), and a 11.5% Value network: partners portfolio performance incentive. e.g., investor-owned utilities Docket No. E-2, Sub 931 Strategic resources: core (2015) competencies; strategic assets; and core processes. The Docket provides Duke Energy a shared savings

282 model for recovery of program costs (for up to 36 Value network: partners e.g., investor-owned

months of net lost revenues). utilities Docket No. E-22, Sub 464 Value network (partners (2015) e.g., IOUs); Strategic resources: core competencies; strategic assets; and core processes. The Docket provides a program for cost recovery to Dominion North Carolina Power, up to 36 months of Customer interface: net lost revenues, and a performance incentive of 8% pricing structure (revenue for DSM and 13% for energy efficiency programs. model); information and insight.

Docket Nos. G-9, Sub 499 (2005) and G-9, Sub 550 Customer interface: (2008); Public Service pricing structure (revenue Company of North Carolina The dockets approved and facilitate revenue-per- model); information and Docket No. G-5, Sub 495 customer decoupling program for Piedmont Natural insight. (2008) Gas. SB 1972 In 2009, Texas considered SB 1972 seeking to Customer interface: decouple utilities’ profits from their sales but the legislation did not pass. pricing structure (revenue PUCT Substantive Rule model); information and §25.181 Rule §25.181 proposed a shared benefit incentive for insight. 283 TX investor-owned utilities. Utah Public Service Value network: partners Commission (UPSC) Utah does not have any decoupling mechanism is in e.g., investor-owned place for electric utilities. utilities Electric Service Schedule In 2003, UPSC approved Electric Service Schedule Strategic resources: core 193 (2003) 19362 to improve customer efficiency services, cost competencies; strategic effective energy efficiency and load management assets; and core processes. UT programs, managed by PacifiCorp.

62 Schedule 193 is a balancing account mechanism that streamlines funding of revenue programs collected outside of general rate case proceedings.

UPSC Docket No. 05-057- In October 2006, UPSC approved Questar Gas to Customer interface: T01 (2006) implement a Conservation Enabling Tariff (CET) and pricing structure (revenue demand-side management pilot program.63 Questar's model); information and CET is a form of decoupling electricity revenues from insight. electricity sales. HJR 9 Strategic resources: core HJR 9 created incentives to increase energy efficiency competencies; strategic and conservation. assets; and core processes.

284

63 Under the CET tariff structure, the revenues received by Questar are based on the number of customers rather than gas usage by customers. This is considered to be a form of decoupling. On June 24, 2009, the Pilot Program was extended to operate until December 31, 2010 (PSC Docket No. 05-057-T01, October 2006).

Chapter 7

A ROADMAP FOR DELIVERING NG-RE HYBRID POWER GENERATION AND DISTRIBUTED RESOURCES: CASE STUDY OF NEW YORK

Chapter 3 examines four major components of the Hamel business model framework (i.e., core strategy, strategic resources, customer interface, and value network). Each of these components has several subcomponents that interact with other subcomponents. Positioning the business model as the unit for analysis provides a robust and multi-dimensional tool for evaluating the suitability of new proposals for electric utilities and energy governance. Using the Hamel business model framework, this chapter assesses the utility business model innovation of New York’s Reforming the Energy Vision across the four major objectives: the role of core strategy, strategic resources, customer interface, and value network in improving communication with consumers and operational boundary of utilities in the new utility-business model regime and beyond.

7.1 Evaluating the REV Docket: The Détente for Utilities and DER

In 2014, New York Governor Andrew Cuomo and the state’s utility regulator (NYPSC) initiated the REV program: an ambitious project that seeks to fundamentally transform New York state’s electric power sector from a primary central model by integrating DERs into the planning and operation of electric distribution systems

(Sahasrabhojanee et al., 2016). The REV docket has two tracks: “Track One” focuses on the development of DER markets and their utility as the distributed-system

285

platform (DSP) providers while “Track Two” focuses on reforming utility-ratemaking practices and revenue streams in the state to accommodate the proposed DSP provider model. For instance, in its REV “Track Two” white paper (“Ratemaking and Utility- business models” of 28 July, 2015), the staff of the NYPSC recommends (NYPSC, 2015a) the following:

Combination of financial incentives that consist of new MBEs opportunities, practical adjustments to conventional ratemaking methods, and concrete targets with new positive-only, symmetrical, and bidirectional earnings impacts. This combination allows early gains around overall cost reduction as well as continued assurance that public policy goals are met….and instills the broad-based confidence that REV requires, place the State firmly on the path to industry modernization, and provide the Commission the transparency necessary to determine how best to adjust the regulatory formula as the market matures and less regulatory intervention is needed.

The REV program will be implemented over a period of years through the mutual efforts of industry, customers, non-governmental advocates, and regulatory partners. The initiative encourages regulatory changes that promote energy efficiency, demand response, increased storage capacity, and renewable energy resources. These reforms seek to empower end-users by providing more choices and by fostering improvement in the performance of the power sector across six policy objectives

(NYPSC, 2015b, p. 4):

• system-wide efficiency, • system reliability and resiliency, • enhanced customer bill knowledge and tools to support effective management of the total energy bill,

• market animation and leverage of customer contributions,

286

• fuel and resource diversity, and • the reduction of carbon emissions.

This means that New York is, “removing market barriers and bridging market gaps that have historically impeded the clean energy sector from benefiting from technological innovations” (Kauffman and Zibelman, 2015). The major impact of the REV program has been to encourage the integration of solar- and wind-energy generation into the existing utility’s power network. Therefore, this evaluation focuses on the regulations and directives specified by the NYPSC and the guidelines and directives released by the major power utilities in the state [e.g., Consolidated Edition, Long Island Power Authority (LIPA), Niagara Mohawk Power Corporation (NMPC), New York Power Authority (NYPA), New York State Electric and Gas Corporation (NYSEG), Central Hudson Gas and Elec Corporation (CHGEC), Orange and Rockland Utility Inc., and Rochester Gas and Electric Corp (RG&E)] to explore the characteristics, nuances, structure, and approaches that exist in them.

7.1.1 From a Centralized and Incentive-driven Model to a Distributed System Platform

Retail peak electricity demand in New York is approximately 75% greater than the average system load and nearly 9% of power generated in the state is lost in transmission (NYPSC, 2015). Essential investment needed over the next 10 years to replace the state’s aging infrastructure to meet currently projected energy demand is projected at $30 billion (NYPSC, 2015c). The REV initiative is thus a primary response strategy intended to make distribution planning more transparent and better integrated. For instance, it proposes to have electric distribution companies act as

287

distributed system platform (DSP) providers that coordinate distributed energy markets and the state’s transition to a cleaner, more resilient, and more affordable grid (NYPSC, 2015, p. 32).

The REV model foresees a “transactive grid” in which “consumers and other parties can take full advantage of every type of energy resource—on both sides of the meter” (Zibelman, 2016). Key to this ambition is to modify the traditional regulatory model and realign utility interests with consumer interests: Utilities will be provided with the opportunity to share in the savings associated with efficiency increases (Zibelman, 2016). Two price-signal processes play a critical role in this regard. First, REV establishes benefit-cost analyses as a foundational procurement tool to determine DER deployment (Zibelman, 2016). Perhaps chosen due to its regulatory familiarity and apparent simplicity (Felder and Athawale, 2016), the benefit-cost analysis is to work in tandem with the multi-year distribution-system integration plans (DSIPs) – developed by the utilities under the REV approach—to assure a fair, open and value- based decision-making process (Zibelman, 2016). The benefit-cost approach will be applied in four key categories of utility expenditures (Zibelman, 2016).64 The benefit- cost approach will be applied in four key categories of utility expenditures (Zibelman, 2016):

• Investments in DSP capabilities, • Procurement of distributed energy resources through competitive selection, • Procurement of distributed energy resources through tariffs, and

64 How the benefit cost tool can be best applied—or whether other tools should be used instead – is still under discussion (Felder and Athawale, 2016). DSIP can be thought of as an integrated resource plan (IRP) but at the distribution level.

288

• Energy efficiency programs. The second key tool is to use locational marginal pricing (LMP) principles to determine the full value of distributed resources. Application of LMP principles can help distinguish what configuration of distributed resources and systems yield the overall best value for the system and consumers (Zibelman, 2016). In terms of a repurposed DER policy, market development, innovation in value strategy, development of a benefit-cost analysis framework in coordination with the DSIPs, and investment in community-choice aggregation programs, the REV business model shares some of these characteristics with other ambitious and successful initiatives, particularly the German Energiewende initiative (Binder and Foster, 2016). New York is not alone in its efforts to integrate distribution planning with DERs to achieve optimal systems efficiencies. Parallel regulatory actions have been proposed in California (California Public Utilities Commission, 2014), Hawaii (HPUC, 2017), Massachusetts (Commonwealth of Massachusetts, 2017), and Minnesota (e21 Initiative, 2014). However, the REV model represents the most promising Utility 2.0 business model, as it, at least, challenges two fundamental components of the conventional model: (i) the assumption that electricity demand is inelastic, and (ii) the notion that economies of scale make a centralized generating model the most economical way to meet power needs (Brooks, 2015).

As a result, positioning the “business model” as the unit for evaluating the REV initiative provides a platform for understanding the fundamental reorientation of the structure of the power sector to address the challenges of energy sustainability, energy economy, climate change, and social development. As part of the literature review for this chapter, 36 studies that attempt to quantify the value of the integrated REV model

289

by using a business model unit-analysis approach across the six policy objectives discussed above were examined. Table 7.1 shows a sample of the studies and the reports reviewed.

290

Table 7.1 Studies Reviewed that Relate to Evaluation of the REV Business Model

REV Goals for NY Authors Title Publisher Year Utility 2.0 Issues Distributed Grid Coverage

Information asymmetry, System-wide efficiency; On the radicality of New The capital expenditure bias, market animation and Astoria, R. York’s Reforming the Electricity and Hope regulatory leverage of customer (2017) Energy Vision Journal 2017 regime contributions 291 Fuel and resource diversity; system U.S. Chapter 2: Maximizing reliability and resiliency; Department Economic Value and U.S. system-wide efficiency; of Energy Consumer Equity (p. 22- Department of Electricity markets, reduction of carbon (2017b) 78) Energy 2017 reliability, and resilience emissions Order on Net Energy Metering Transition, System reliability and Phase One of Value of resiliency; system-wide Distributed Energy efficiency; reduction of Resources, and Related NY Public carbon emissions NYPSC Matters (15-E-0751, 15-E- Service Net metering, value of (2017b) 0082) Commission 2017 DERs

Chapter 2. Distribution utilities and Sioshansi, F. Value of an Integrated Academic 2016 their place in the integrated System-wide efficiency (2016) Grid Press grid model. Market animation and The Innovation Platform Chapter 5. Infrastructure services, Sioshansi, F. leverage of customer Enables the Internet of Academic 2016 personalization, and value (2016) contributions through Things Press creation innovation platform REVing Up the Energy Vision in New York: IEEE Power System reliability and Zibelman Seizing the Opportunity to New revenues for new and Energy 2016 resiliency; system-wide 292 (2016) Create a Cleaner, More Innovative services Magazine efficiency Resilient, and Affordable Energy System Phased optimization of the Felder and Optimizing New York's benefit-cost analysis Utilities DERs, DSPs, and Benefit- Athawale Reforming the Energy 2016 methodology applied to Policy Journal Cost Analysis framework (2016) Vision all the REV policy objectives IEEE A Comparison of NYS International Sahasrabhoja Utilities’ Approaches to System reliability and Conference Distributed generation nee et al. Integrate Distributed 2016 resiliency; system-wide on Smart integration (2016) Energy Resources and the efficiency Energy Grid Penetration Over Time Engineering

Utility of the Future: An System-wide efficiency; Pérez- MIT Energy Initiative Massachusetts Energy storage for end- reduction of carbon Arriaga et al. response to an industry in Institute of user and system co- emissions. (2016) transition Technology 2016 optimization The REVolution yields to System-wide efficiency; a more familiar path: New The Rate base, capital market animation and Makholm, York’s Reforming the Electricity expenditure bias and the leverage of customer J.D. (2016) Energy Vision (REV) Journal, 2016 information asymmetry. contributions Co-Optimization of Power and Reserves in Dynamic

293 T&D Power Markets with

Nondispatchable Marginal-cost-based Renewable Generation IEEE dynamic pricing of Caramanis et and Distributed Energy Publications electricity services, al. (2016) Resources Database 2016 wholesale power markets System-wide efficiency New York Adopts New Carson and Revenue Model for Utility Financial Kreilis Electric Utilities under Incentives, Customer (2016) REV EnerKnol 2016 Interests System-wide efficiency Order Establishing the NY Public Investments in DSP NYPSC Benefit Cost Analysis Service Benefit Cost Analysis capabilities; procurement (2016d) Framework (14-M-0101) Commission 2016 (BCA) Framework of DERs

Order Authorizing the Clean Energy Fund Framework (14-M-0094, NY Public NYPSC 10-M-0457, 07-M-0548, Service Reduction of carbon (2016b) 03-E-0188, 13-M-0412) Commission 2016 DERs emissions Order Authorizing Utility- Administered Energy Efficiency Portfolio Budgets and Targets for NY Public Reduction of carbon NYPSC 2016-2018 (Jan. 22, 2016, Service Utility energy efficiency emissions; System-wide

294 (2016c) 15-M-0252) Commission 2016 portfolio efficiency

Order Resetting Retail Energy Markets and System reliability and Establishing Further NY Public Monitoring and regulation resiliency; system-wide NYPSC Process (15-M-0127, 12- Service of retail efficiency (2016e) M-0476, 98-M-1343) Commission 2016 energy markets

Order Authorizing Enable deployment of Framework for DERs and increase the System-wide efficiency; Community Choice NY Public benefits of retail market animation and NYPSC Aggregation Opt- Out Service competition to all leverage of customer (2016a) Program (14-M-0224) Commission 2016 customers contributions

Agüero & The Conceptual description of Khodaei Roadmaps for the Utility Electricity Utility of the Future the six policy objectives (2015) of the Future Journal 2015 Roadmaps of the REV model Fuel and resource diversity; system Tao, Conference reliability and resiliency; Bahabry, and on Systems Smart grid, sustainable system-wide efficiency; Cloutier Customer Centricity in the Engineering energy, customer utility reduction of carbon (2015) Smart Grid Model Research 2015 service model emissions The Impact of Distributed 49th Hawaii

295 Energy Resources on International Distributed System Fuel and resource

Tabors, He, Incumbent Utilities: A Conference Platforms, Distributed diversity; market and Birk Case Study of Long on System Locational Marginal animation and leverage of (2016) Island, New York Sciences 2015 Prices, and DERs customer contributions Utility 2.0 characteristics System-wide efficiency; Institute for and the transition to market animation and Farrell Beyond Utility 2.0 to Local Self democratized energy leverage of customer (2014) Energy Democracy Reliance 2014 governance (Utility 3.0) contributions

Energy Service Companies and Energy Performance Contracting: System-wide efficiency; Is there a need to renew market animation and Pätäri and the business model? Journal of leverage of customer Sinkkonen Insights from a Delphi Cleaner ESCOs and energy contributions; System (2014) study Production performance reliability and resiliency

296

Besides the studies and regulatory orders considered above, the following studies have applied business model unit analysis in the energy sector in general:

• explorative analysis of the generic business model revealed investor

preference for “customer-intimacy” business models over lowest-cost or best-technology models (Loock, 2012); • regulatory reform of the distribution of renewable energy benefits needs to be in accordance with, or drive the change of, existing utility- business models to retain and achieve nation-wide objectives for further distributed-energy deployment (Barbose, et al., 2016);

• regulatory reform of the distribution of renewable energy benefits needs to be in accord with, or drive the change of, existing utility- business models to retain and achieve nation-wide objectives for further distributed-energy deployment (Barbose, et al., 2016);

• the role of external politico-institutional and socio-institutional dynamics in the formation and success of business model options is as a co-authoring agent in addition to internal decision-making (Provance, Donnelly, and Carayannis, 2011);

• identification and analysis of “community-solar” business models as an alternative deployment strategy for solar energy that could mitigate or circumvent current concerns of negative impacts of distributed

photovoltaic deployment for utility revenues and equity distributions of subsidies (Funkhouser, Blackburn, Magee and Rai, 2015); and

• the proliferation of demand-side management models and options in terms of transaction characteristics, renewable energy correlation, and

297

load control shows demand response diversification but energy efficiency complexity (Behrangrad, 2015).

Positioning the “business model” as the unit for analyzing utility-sector transformations provides a viable perspective for unlocking significant benefits at different levels. For instance, a study conducted by the Brattle Group for the New York Independent System Operator (NYISO) determined that (Newell and Faraqui (2009),

dynamic pricing can provide substantial benefits in New York State by reducing total resource costs, lowering customer market costs, and improving economic efficiency. With estimated market-based cost savings in the range of $171 million to $579 million per year, the benefits to electric consumers can be significant, especially when technology serves to facilitate demand response and energy conservation.

A critical challenge in the analysis of Utility 2.0 candidates (such as the REV docket) is therefore to construct a comprehensive analytical framework that can compare business model options across the entirety of the energy-utility spectrum. As discussed in Chapter 3, the pitfall of ambiguous definitions and overlapping terminology needs to be avoided in such a framework. Table 7.2 offers a four-part, multi-dimensional, Hamel analytical framework that tries to deliver on such a set of definitions and characteristics that can be positioned in the evaluation of the REV docket. These dimensions extend beyond business model innovation in the electricity- market sector. These dimensions attempt to account for the increasing focus on performance-based utility operation, the relationship dynamics that accompany such a shift (Kushler et al., 2006; Nowak, et al., 2015; Selviaridis and Wynstra, 2015), and

298

the apparent requirement to move to a servitization system—as mandated by system reliability and resiliency, system-wide efficiency, and the climate change challenge (Barnett et. al., 2013; Carley, 2012; Steinberger et al., 2009).

Table 7.2 Application of Hamel Business Model to Conventional Energy Utility Component Subcomponents Definition Application to conventional utility Core Depicts the architecture Centralized, large-scale, Strategic competencies of the utility value production. Long- resources Strategic Assets creation. It includes distance transmission and strategic assets, know distribution. Prohibitive Core Processes how, core processes and cost for duplication, competencies. Customer Fulfilment and Involves the overall Consumers of electricity, interface support interaction with the monthly billing, short- Information and customer including term relationships, insight customer relationship, distant and standardized. customer segments, Bulk generation of Relationship distribution/fulfilment dynamics electricity, commodity- support and channels, focused Pricing structure and the pricing (revenue model) (revenue) structure Value Encompasses added Low cost of electricity at network values utility company high volume, guaranteed Suppliers or business offers for service Partners resource providers or “Just and reasonable” Coalitions suppliers, project prices developers, technology vendors, community Shareholder return served, and other potential partners.

299

Business The utility’s capacity to Limited flexibility: Core mission change course in the maintains complex Strategy Product / face of potential system of market scope existential business interconnections and model risks. This generation capabilities Basis for capacity is influenced with second-to-second differentiation by the flexibility and management across complexity of both the (state) borders. For business model but also example, extreme the infrastructure it reduction in sale volume operates. could initiate “death spiral” (Athawale and Felder, 2016).

Table 7.2 offers a brief example application of the various dimensions to the existing utility business model (Column 4). The above business model options are discussed below to expand the application of the analytical framework to the REV docket.

7.1.2 Strategic Resources New York State (NYS) has various types of electric utilities. These include investor-owned private utilities, retail-power marketers, NYS-owned public authorities, and municipal utilities. These can be grouped into two major service types: bundled and delivery. Figure 7.1 and Table 7.3 show electric-utilities service territories (NYPSC, 2017a) and a list of the top-25 major utilities in the state in 2015 ranked by revenue and service type. The NYPSC regulates all the utilities in the state and establishes a set of guidelines that includes comprehensive information for developers about the protection and integration of DERs.

300

Source: NYISO (2017)

Figure 7.1 Electric Utilities Service Territories

On the other hand, the New York Independent System Operator (NYISO)—a non-profit organization set up by NYS—provides an open platform by which utilities can procure energy from energy suppliers. NYISO administers the state’s wholesale electricity markets and provides reliability planning for bulk-electricity grid power. It also administers New York State’s wholesale electricity markets. On the other hand, FERC has the jurisdiction to regulate wholesale electricity rates, hydroelectric licensing, and interstate electricity sales (FERC, 1987). Under FERC Order 745 about demand-response pricing, FERC regulates wholesale product tariffs by an independent

301

system operator (ISO) such as NYISO—including those that integrate DERs into wholesale markets (Zibelman, 2016).

Table 7.3 Top 25 Utilities in New York State, By Revenue and Service Type, 2015 Revenues Sales Customers

Service Thousand Utility Name Ownership MWhrs Count Type Dollars Consolidated Edison Bundled Investor Owned 4,757,588.0 20,206,464 2,545,762 LIPA Bundled State 3,326,347.9 18,151,255 1,115,541 Consolidated Edison Delivery Investor Owned 3,313,064.0 36,828,725 851,992 NMPC Bundled Investor Owned 1,598,810.0 13,042,081 1,288,882 Direct Energy Retail Power Business Energy Marketer 1,052,583.0 14,317,230 14,980 NYPA Energy State 1,008,484.0 18,578,384 824 NYSEG Bundled Investor Owned 789,556.0 6,748,513 669,017 NMPC Delivery Investor Owned 659,514.1 21,394,469 358,956 Constellation Retail Power NewEnergy, Inc Energy Marketer 580,389.2 8,215,486 2,178 CHGEC Bundled Investor Owned 417,835.4 2,553,909 247,746 ENGIE Retail Power Resources Inc. Energy Marketer 362,597.0 5,216,435 693 RG&E Bundled Investor Owned 335,252.0 2,605,567 282,915 Consolidated Retail Power Edison Sol Inc Energy Marketer 331,486.0 3,910,908 47,534 NYSEG Delivery Investor Owned 291,363.4 8,900,387 215,321

302

Orange and Rockland Utility Inc Bundled Investor Owned 271,010.0 1,463,572 138,109 Direct Energy Retail Power Services Energy Marketer 237,623.9 2,333,353 198,911 RG&E Delivery Investor Owned 222,431.0 4,503,779 90,897 Constellation Energy Services Retail Power NY, Inc. Energy Marketer 190,923.3 3,076,884 13,484 Ambit Energy Retail Power Holdings, LLC Energy Marketer 186,648.4 1,890,178 183,180 Orange and Rockland Utility Inc Delivery Investor Owned 179,718.7 2,564,756 90,312 Hudson Energy Retail Power Services Energy Marketer 131,698.9 1,470,254 21,979 LIPA Delivery State 123,240.2 1,774,183 3,563 Liberty Power Retail Power Corp. Energy Marketer 112,233.6 1,331,204 19,195 AP Holdings Retail Power LLC Energy Marketer 107,158.0 401,659 5,029 Central Hudson Gas and Electric Corporation Delivery Investor Owned 100,436.0 2,523,407 54,687

Source: (U.S. Energy Information Administration, 2017) Reliable, affordable and increasingly clean power are transmitted, stored, and distributed throughout an infrastructure that spans thousands of miles of interstate and

intercity pipelines, transmission lines, and natural gas storage facilities. The state’s

strategic resources and energy assets are owned, operated, and regulated by a variety

303

of private and public entities (Figure 7.2). The functions provided by this complex infrastructure operating 24 hours a day, 365 days a year—with the high reliability, longevity and high capital costs associated with the deployment of this infrastructure—create a path-dependency in which current utility-business models either enable or constrain future energy-system development (Kauffman and Zibelman, 2015). The resulting electric-utility landscape that manages the flows of all these energy sources has experienced consolidation to the point at which, in 2015, a

“baker’s dozen” of three holding companies (Consolidated Edison, Long Island Power Authority, and Niagara Mohawk Power Corporation) representing 2.4% of all integrated utilities controlled 49% of utility revenues (U.S. Energy Information Administration, 2017c). Figure 7.3 shows distribution of the number of utilities by ownership in 2015. Figure 7.3 shows distribution of the number of utilities, by ownership in 2015.

304

305

Figure 7.2 Overview of NYS Electric Industry Participants

Figure 7.3 Number of Utilities, by Ownership in 2015

The REV model seeks to address infrastructural challenges raised in chapters 1 and 2. Using advanced data analytics, the REV docket aims for a shift to low-carbon decentralized generation and a multi-directional flow of data and energy. It is positioned to achieve what we might call the Infrastructure to Services Transition: i.e., the evolution of infrastructure for commodity delivery to support greater personalization of value—new purposes, new platforms, enabled new infrastructure, and new apps (services) (Cooper, 2016). In New York, nearly double the $17 billion

306

invested over the past decade is needed to replace the State’s aging electric transmission and distribution infrastructure between 2015 and 2025 to meet currently projected energy demand (NYPSC, 2015).

7.1.2 Customer Interface The REV docket emphasizes improved customer choice and lower cost of service for consumers. For instance, it underscores enhanced customer knowledge and tools for effective management, market animation and leverage of rate-payer contributions, system-wide efficiency, fuel and resource diversity, system reliability and resiliency, and carbon-emissions reduction (NYPSC, 2015; NYPSC, 2015). Conventional energy utilities compete by establishing an effective utility-consumer relationship (Byrne and Taminiau, 2016) that is characterized by standard, billing- based interactions that are impersonal, distant, and standardized (Hannon et al., 2013). This distance is partly a result of the fiduciary obligation to the owners of the conventional energy utility—its shareholders (Byrne and Taminiau, 2016). A second aspect of the distance is that conventional energy utilities interfere with the consumer only in a limited fashion, as they do not go “beyond the meter” (e.g., the behavioral stipulations of energy use are limited) (Hannon et al., 2013). Table 7.4 shows a tabulation of utilities ownership in 2015.

307

Table 7.4 Tabulation of Ownership of NYS Utilities, 2015 Cumulative Cumulative Ownership Count Percent Count Percent Behind the Meter 12 9.68 12 9.68 Cooperative 1 0.81 13 10.48 Investor Owned 15 12.10 28 22.58 Municipal 12 9.68 40 32.26 Retail Power Marketer 81 65.32 121 97.58 State 3 2.42 124 100.00 Total 124 100.00 124 100.00

Although the conventional energy utility, through demand-side management processes, offers incentives to consumers to monitor their energy-use trends, the consumer is typically responsible for managing such changes (e.g., switching to high- efficiency appliances). And in contrast to multi-decade power-purchase partnerships, supply contracts provided by the conventional energy-utility model are short-term and, as such, provide flexibility to the consumer to switch providers (Hannon et al., 2013). Of the 124 utilities analyzed in 2015, investor-owned utilities in NYS accounted for about 12%, representing 71% of customers (Figures 7.4 and 7.5). Investor-owned utilities operate under conditions of a guaranteed rate of return set by the NYPSC. The cost structure of the utility is determined by its focus on large-scale asset investment, pursuit of economies of scale, and long-term infrastructural commitment. The cost-of- service (COS) model is a core feature of the conventional utility (Burr, 2007;

McDermott, 2012). The model entails the “rate-making process … i.e., the fixing of

‘just and reasonable’ rates, [and] involves a balancing of the investor and the consumer

308

interest” (Federal Power Commission v. Hope Natural Gas Co. 320 U.S. 591, 603, 1944). This has also been called the “end result doctrine,” according to which “the aim of regulation is to preserve the balance of the original bargain between investors and customers” (McDermott, 2012).

Figure 7.4 Revenues, Sales, and Customer Count of Major Utilities in NYS, 2015

309

Figure 7.5 Number of Customers by Ownership, 2015

7.1.3 Value Network

The business model of the conventional utility pursues asset-based expansion, and, through its commodity-focused strategy, increased sale of products and additional value delivered to shareholders. Satchwell and Cappers (2015) suggest that:

Under traditional COS regulation, a utility is motivated to solve system reliability and customer access issues by investing capital instead of maximizing the value it can extract from existing assets. The goal of the conventional utility, as such, can be conceptually positioned at one end of a profit-motivation spectrum: The “motivation to build incremental assets

310

for the primary purpose of expanding its rate-base” (Satchwell and Cappers, 2015). Because regulators reward or punish utilities for taking actions to achieve certain public-policy goals and to maintain “just and reasonable revenues,” this model faces a wide array of challenges—especially in a DER framework. So-called “incentive regulation,” however, establishes the working conditions of the utility. Within these conditions, “[g]iven any set of regulations, utilities participate in actions which most benefit their principal constituencies—shareholders and management—while meeting the requirements of the regulations” (Lazar, 2011). For instance, because the principal constituency of the investor-owned utility is its shareholder base, the utility must grow its customer-base through value addition to delivering shareholder value.

7.1.4 Core Strategy Established utilities such as RG&E, NYSEG, Orange and Rockland, Central

Hudson, and Consolidated Edison—which exhibit a large asset base, a commitment to implementing a specific business model, and stable service—have in certain situations displayed limited flexibility to update to a new business model. Response to new risks or market-environment changes depend on adaptations in the regulatory environment

(Nyangon, Byrne, and Taminiau, 2017). Nevertheless, signaling modern society’s dependence on the viability of the energy utility, such changes have happened repeatedly throughout the lifetime of the U.S. electricity timeline—in particular, in the de-regulation and transition stages (Utility 2.0 business model) (Brocks, Nyangon and Taminiau, 2016; McDermott, 2012). The energy transition to distributed wind- and solar-electricity generation, for instance, elevates business model concerns about the potential need for higher electricity rates or cost-shifting to non-solar customers, reduced utility-shareholder

311

profitability, reduced utility-earnings opportunities, and inefficient resource allocation (Barbose, et al., 2016). Among the responses proposed to address these concerns in NYS are the following: to reduce compensation to customers who have installed distributed energy, to facilitate higher-value distributed-energy deployment, to gain utility ownership and financing of distributed photovoltaics, or to align utility profits with the deployment of distributed energy (Barbose, et al., 2016). However, as Hess (2015) and Bayulgen and Ladewig (2016) caution, policymakers and utility stakeholders should ensure that these strategies and innovations are not construed as attempts at regime preservation rather than adaptation. In other words, REV is altering the political and economic structure of the utility landscape in NYS by changing the obligations of stakeholders. Distributed utilities, for instance, act as DSPs. To further elaborate, a DSP is defined as “an intelligent network platform that will provide safe, reliable, and efficient electric services by integrating diverse resources to meet customers and society’s evolving needs. The DSP fosters broad market activity that monetizes system and social values, by enabling active customer and third party engagement that is assigned with the wholesale market and bulk power system” (NYPSC, 2015). Furthermore, ESCOs, which currently provide only commodity service in New York state, are encouraged under the REV framework to offer more classes of service.

7.2 Enabling Higher Penetration of DER in New York

Figures 7.6 (a) and 7.6 (b) show NYS control area load zones.

312

D A - WEST B - GENESE C - CENTRL D - NORTH E - MHK VL E F - CAPITL G - HUD VL H - MILLWD I - DUNWOD F J - N.Y.C.

313 K - LONGIL

E A C B G H

J K

(a)

CLINTON

FRANKLIN

ST LAWRENCE E ESSEX

JEFFERSON F LEWIS WARREN

HAMILTON WASH- ING- C TON

HERKIMER SARATOGA OSWEGO B ONEIDA FULTON

314 SCHE- MONTGOMERY RENS- NEC- SELAER TADY WAYNE ALBANY ONONDAGA ORLEANS MONROE MADISON SCHO- HARIE NIAGARA OTSEGO ONTARIO COLUM- CAYUGA GENESEE GREENE BIA LIVINGSTON SENECA CHENANGO YATES CORTLAND WYOMING A ERIE TOMPKINS DELAWARE SCHUYLER ULSTER G BROOME TIOGA DUTCHESS

STEUBEN CHEMUNG ALLE- GANY SULLIVAN CATTARAUGUS PUTNAM H CHAUTAUQUA E C ORANGE B WESTCHESTER G ROCKLAND SUFFOLK BRONX NASSAU I NEW YORK NEW YORK CONTROL AREA K RICHMOND QUEENS KINGS LOAD ZONES J

(b)

Figure 7.6 New York Control Area Load Zones

7.3 Beyond Utility 2.0: A Shift to Utility 3.0 and Energy Democracy Many utilities still have an incentive to increase electricity sales. Figure 7.6 illustrates the position of state regulatory regimes based on two fundamental factors: (i) revenue decoupling for sales, and (ii) structural separation. Structural separation represents the degree to which power generation, transmission and distribution, and retail sales are unbundled into independent entities (Felder and Athawale, 2016). This is an important component of the Utility 2.0 attributes—especially the four components of the utility-business model discussed above (Brocks, Nyangon and Taminiau, 2016). The lower-left quadrant depicts utilities operating in a Utility 1.0 business model. As explained in Chapter 2, large and centralized electricity systems both in terms of technology and ownership characterizes the Utility 1.0 business model. As Farrell (2014, p.7) writes,

Utility 1.0 is a business model for an electricity system entirely owned by the electric company, from power plant to transmission and distribution network to the meter on the building. The system is based on large power plants that capture the economies of scale in producing power from fossil and nuclear fuels.

315

Source: Adapted from Farrell (2014)

States analyzed in Chapter 6.

Other states

Figure 7.6 Status of Utility 2.0 Structural Change to Utility 3.0

Figure 7.6 shows that states that have implemented revenue decoupling (in the right quadrant) have also implemented a wide range of innovative and forward- looking regulatory models to facilitate greater utility-service offerings. However, states such as Florida, Utah, North Dakota, Iowa, Tennessee, and Nebraska that have had poor experience to date with DER and energy-efficiency incentives for utility shareholders tend to be more cautious about performance-based ratemaking. As a result, a diverse set of approaches emerges that is characterized by a disparate pattern

316

of innovation in the utility sector. Some states are close to the Utility 3.0 business model (energy democracy), while others are still reeling from the challenges of Utility 1.0. Experiences of New York with decoupling policies have shown that revenue decoupling can reduce the pressure to increase sales and allow customers to share in efficiency gains, and provides incentives to utilities to build new power plants. As noted by the NYPSC staff (NYPSC, 2014, p. 47) :

Rate-of-return regulation, using an annual rate case cycle, provides very little incentive to the utilities to improve performance. Benefits of any efficiency gains are reflected in the next year’s rate case. [Rate-of- return] regulation may also encourage the utility to over-invest in capital spending, because earnings are directly tied to rate base. For the same reason, utilities are rewarded for the inefficiencies in the bulk and distribution systems that require capital spending to build for unmanaged peak loads…extending the term of rate plans can encourage utilities to seek innovation and efficiencies since the longer period between rate cases allows utilities the opportunity to keep or share savings. Long term plans typically contain earnings sharing mechanisms (ESMs). ESMs allow utilities to keep a portion of earnings in excess of the allowed return while requiring a portion of the over- earnings to be passed back to customers. These mechanisms allow customers to share in efficiency gains achieved during the terms of the rate plans.

Regardless of where each state in Figure 7.5 may be heading, effective transition strategies are required to mitigate tension in developing new power plants, transmission lines, risk to third-party providers of DER services, and in realizing DER public-policy goals. For example, while some new electricity-grid system is required to allow centralized wind- and solar-power plants to supply renewable electricity from far-flung location to cities, proponents of Utility 2.0 and beyond often criticize such projects as investments in a centralized transmission system that is a hallmark of Utility 1.0. Ultimately, provision of utility services, roles, personalization of value,

317

and customer-centric paradigm towards a prosumer role will naturally lend themselves to economies of scale (e.g., colocation and joint transmission of distributed generation) and thus to some form of regulatory systems. Considering the DER market development issues that New York and other states such as Vermont have tried to address during the transition to their new regulatory and utility business models (e.g., differences in market structure at the state level, utility asset ownership, planning and operational responsibilities, openness of utility networks, regulatory processes, utility role in providing value-added services, assessing and ensuring customer benefits, leveraging experience, and changes to cost-of-service regulation), a combination of Utility-2.0 policies65 and energy democracy 66 (Utility 3.0) policies are required to realize the full benefit of this envisaged clean-energy paradigm. Table 7.5 summarizes the contribution of the New York REV model to the Utility 2.0 transition and potential utility-innovation strategies towards energy governance (Utility 3.0).

65 Utility 2.0 include a mix policies such as increased integration of DERs, revenue decoupling, separate oversight of energy programs and integration of distribution and transmission planning.

66 Utility 3.0 or energy democracy policies include net metering and feed-in tariff. Utility 3.0 adds two principal factors to Utility 2.0: local control by customers and equitable access.

318

Table 7.5 Distributed Utilities Framework of Utility 2.0 Business Model NY REV Model (Utility 2.0) Energy Governance Innovation Strategies (Utility 3.0) • REV addresses electricity grid infrastructure related • Provisions for charging and access rule challenges. changes. • Emphasize Infrastructure to Services Transition, i.e., • Reliability and network infrastructural-related support for greater personalization of value—new services such a personalization and value.

319 purposes, new platforms, enabled new infrastructure, and • Application of advanced data analytics, and

new apps (services). multi-directional flow of data and energy for

• In New York, an estimated $30 billions of investment in improved service delivery. the State’s aging grid infrastructure is needed in the next • Energy democracy governance on decade (2015-2025). infrastructure and utility data. • NY Department of Public Service Case 14-M-0101 • Deeper engagement with third parties through Proceeding on Motion of the Commission in Regard to seminars and periodic meetings. Strategic resources Strategic Reforming the Energy Vision, proposed 4/24/14 (NYPSC, 2014) and order adopted 2/26/15 outlines the framework. • Time limited innovation stimulus open to both utilities and third parties. • 2015 NY State Energy Plan outlines how to enact REV model. • NYSERDA’s Clean Energy Fund provides $5B in new green energy investment over 10 years, starting in 2016.

• NYPA’s programs lead by example. • NYPSC, a government agency, wrote REV • REV has two tracks: o Track 1: examines the role of distribution utilities in promoting EE, load management, DER, consumer control, and wholesale market issues; considers whether distribution utilities should serve as DSPs o Track 2: regulatory and ratemaking changes • PSC’s REV Docket promotes greater consumer choice 320 in energy use.

• Distribution utilities act as DSPs. • Comprehensive, multi-criteria price control

across DER spectrum. • Covers six key utility policy objectives: market animation and leverage of ratepayer contributions, system-wide • Consideration for revenue structure such as efficiency, fuel and resource diversity, system reliability base revenue, revenue adjustment for and resiliency, and reduction of carbon emissions. rewards/penalties, and uncertainty Core strategy strategy Core mechanisms. • DSPs67 provide pricing structures.

67 DSPs act like “mini-ISOs” situated between the NYISO and consumers. DSPs will provide pricing structures by using localized automatic systems to balance production and load in real time to allow for more DER integration.

• Efficiency treated like part of utility revenue requirement, • Combines operating expenditure and capital not a dedicated surcharge. expenditure (CapEx) into total expenditure or “Totex”77 • PSRs68 and MBEs69 replace EIMs70

• Modified clawback71 mechanisms to encourage third party interactions

321 68

PSRs (platform service revenues) are revenues that utilities, in their capacity as DSP providers, will earn from market participants.

69 New MBEs (market based earnings) could include value added from services such as an online portal for customers, transaction/ platform access fees, optimization/scheduling services that add value to DER, energy services financing, engineering services for micro-grids, etc.

70 EIMs (earning impact mechanisms) are monetized. Performance incentives. They could be used for peak reduction, EE, customer engagement and information access, affordability, and interconnection. Different EIMs do not have to have the same directionality. They should be established on a multi-year basis.

71 “Clawback” refunds unspent amounts of utilities’ capital budget to consumers. This can be revised so the money that would have been spent on a project can be retained if DER supplants the need for project. Clawback could also be modified so that utilities are indifferent to whether it is spent by them or a third party.

77 “Totex” is total expenditures. Under totex, that capital and operating expenditures are treated as equivalent and recovered under the same formula. The formula sets a ratio of “slow” money to “fast money”. The “slow” money is included in the RAV and the “fast” money is recovered on an annual basis.

• ESMs72 tied to performance index. • Provision for third parties to appeal to PSC or Competition Commission regarding price • Scorecards73 to evaluate non-monetized measurements. control before instituted78 • 3-year rate plans (opt in for 5). • Should be innovation driven79 • Value of DER calculated as LMP+D74 • Create fund targeting small scale innovation • Increased encouragement of TOU75 rates projects in the utility value chain. • Set balanced efficiency inventive rates.

322

72 ESMs (earning sharing mechanism) allow utilities to retain earnings above a baseline return on equity. Beyond that level, earnings are shared between utilities and customers. At higher levels, savings are dedicated entirely to consumers.

73 Scorecards measure performances that do not have any direct earnings impacts. Scorecards are proposed for system utilization and efficiency, distributed generation, EE, and dynamic load penetration, carbon reduction, opt-in TOU rate efficiency, market development, MBE use, customer satisfaction, customer enhancement, and conversion of fossil fuel end uses.

74. The “value of the D” is the benefit that should be produced to the customer in terms of total cost avoided or reductions to the distribution system by DER. The value of the “D” is not established, while LMP is.

75 Time of use rates

78 Any party can make a price control modification request to GEMA during the final proposals stages. GEMA, the gatekeeper, determines whether this modification request should be referred to the Competition Commission.

79 Companies should compete for partial finding outside of the price control framework.

• Market power concerns over distribution utility acting as • Share risk through symmetric efficiency DSP. uncertainty mechanisms. • Concerns over whether MBEs can replace traditional • Provide the revenue system to cater for rise in utility earnings. demand or volumes of activity. • Combine financial incentives such as new MBEs, ratemaking adjustments, concrete targets with positive, symmetric, and bidirectional earnings impacts.76 • Each utility submitted a Benefit-Cost- Analysis (BCA). • DPS recommends financial incentives such as new MBEs

323 to simplify access to DSP platform and to offset impact of

DSP capital by sharing platform costs, adjustments to conventional rates, new positive only symmetrical bidirectional earnings impacts.

• Emphasize enhanced customer-centric paradigm (e.g., • Besides customer-centric paradigm, seeks to

face knowledge and tools for effective management). improve satisfaction, reliability, safe network services, better connection terms, social

Inter • 43 different programs offered in NY Plan. Customer Customer

76 Staff recommends an opt-out data exchange because integrating DER requires standardized time stamped energy usage information. A single entity to operate a data exchange is being considered.

• Transactions take place in a nonlinear manner. obligations, and meeting the set environmental

targets. • Distribution network operators should submit and publish realistic business plans with demonstrable value to consumers. • Remove market barriers to enable a dynamic clean energy • Promote energy democratic principles e.g., economy at a scale to create opportunities and growth encourage utility stakeholders and network while protecting the environment. companies to play a full role in achieving a sustainable energy sector and deliver long term • DSP’s interact between consumers, sellers of products, and value in network services for current and

324 NYISO to create a market pricing platform that allows future consumers. monetization and exchange of resources such as DER,

DSM, EE, storage. • Emphasize customer satisfaction, reliability, safe network services, better connection terms, etwork

n environmental integrity, etc. • Ensure transparent, upfront price control Value Value framework (i.e., sets out what outputs network companies need to deliver upper limit on allowed return, symmetrical incentives). • Increase involvement of network companies and non-network parties in the energy decisions.

7.4 Five Pillars of Utility 3.0 (Energy Democracy) Framework

Figure 7.7 illustrates key components of energy policies and technologies that will play a major role in the electricity system of the future. It also illustrates how they overlap with Utility 2.0 and 3.0 (energy democracy). Achieving the promise of both Utility 2.0 and Utility 3.0 means repositioning the principles, structure, and policies that govern grid participation and accessibility. This requires substantial changes in the current utility-business models, based on five key pillars: flexibility of the grid system to encourage capture of economic opportunities across the utility-value chain, efficiency, low-carbon goal, local control, and equitable access.

Source: Adapted from Pérez-Arriaga and Knittel (2016)

Figure 7.7 The Rules and Principles of Utility 3.0 DER Model

325

Chapter 8

CONCLUSION AND POLICY RECOMMENDATIONS

This chapter concludes the dissertation by summarizing the research conducted in the preceding chapters. Answers to the two research questions that were introduced in Chapter 1 are presented in this chapter on the basis of the discussions provided in Chapters 2 to 7. These answers represent the conclusion of this study, which attempted to test whether natural gas is antagonistic to renewable energy development in the United States. As the study revealed, natural gas has a positive impact on renewable energy diffusion. This chapter also recommends policy and regulatory innovation mechanisms for distributed energy generation development in the United States.

8.1 Summary of Conclusions This chapter is divided into two parts. The first (Section 8.1) summarizes the foundational work of the study provided in Chapters 1 to 4—assessment of the transformation of electricity market design mechanisms (i.e., replacing centralized decisions with competitive distributed processes to meet electricity demand in an efficient, secure, and environmentally acceptable manner. This section also examines the role of natural-gas-fired standby generators to provide reliable power (so called balancing markets) which are intimately related to short–term markets, in order to maintain grid stability. While Section 8.2 reviews barriers to existing policies and mechanisms for DER developmet, the second part of the summary (Section 8.3) covers Chapters 5 to 7, which outlined the current technical, regulatory, and policy framework in the top-ten solar generating states, along with existing supply- and demand-side policies. Specifically, the focus of Section 8.3 is to identify the needed policy and market design changes to achieve long-term hybrid NG-RE power

326

generation capacity and clean energy support mechanisms. This information is applied as a baseline for developing a comprehensive policy framework and recommendations to enhance the compatibility of these mechanisms with existing market factors to minimize associated market distortions. A major focus of this dissertation is improving renewable energy diffusion and flexibility performance through joint-development and deployment with flexible natural gas technologies—both for new-build plants and for retrofits of existing natural gas facilities. The literature analysis conducted as part of this study show that the use of information and communication technologies to facilitate operations of gas- spot markets in the form of day-ahead and online electricity markets has grown substantially in the ten states studied—with the growth in solar generation capacity and increased deployment of flexible natural-gas-fired technologies. Due to the ongoing electricity market design changes identified in Chapter 3 such as new revenue models and business model innovation, advanced billing, load management, real-time metering and quality control are now commonly offered to customers throughout the United States as integral parts of the integrated electricity market development. In addition, improvements in ramping capabilities, start-up times, and part-load behavior in the ten states studied revealed that these practices are continuing with declining cost of natural gas and are being undertaken in parallel with more moderate full-load efficiency improvements. Electricity importation ratio, population, per capita energy- related CO2 emissions, per-capita real GDP, and electricity intensity are the main drivers of investments in flexible natural-gas-fired technologies such as NGCC (see

Section 5.2). Between 2000 and 2014, per-capita energy-related CO2 emissions progressively declined at both the state level and nationally. Nevada, Georgia, North

327

Carolina, Massachusetts, Texas, New York, Utah, and Arizona had some of the highest rates of CO2 emissions decline per-capita at -44.71%, -34.81%, -32.62%, - 29.71%, -27.94%, -26.53%, -25.50%, and -21.99%, respectively. The decline in the national average rate of CO2 emissions per-capita between 2000 and 2014 was - 36.40%. California and Arizona had the lowest rate of decline in per-capita energy- related CO2 emission during the same timeframe at -17.97%, and -19.02%, respectively. This drop is largely due to the increased use of low-cost natural gas for electricity generation throughout much of the U.S., driving a widespread switch from coal-to-gas.

This dissertation established a dynamic econometric model for the U.S.’ electricity sector (i.e., system Generalized Method of Moments or System-GMM). This model is a useful optimization tool for evaluating the dynamic relationship of the indicators in the model and for aligning fragmented policies in different states with overall economic and environmental objectives. The estimates for the period 2001- 2016 were reported for the sample of the ten states selected based on their solar generation capacity in 2016 by SEIA. A set of control variables were tested for the expected impacts. The dynamic econometric model was used to evaluate the effects of natural gas on renewable energy development. The results revealed a positive and statistically significant (20.15) effect of natural gas on solar- and wind-power development, conditional on all other covariates (see Section 6.4.1). For instance, a 1% rise in natural gas capacity produces approximately 0.0304% increase in the share of renewables in the short-term (monthly) and nearly 0.9696% over the 15-year period. There were no significant effects of renewable energy capacity on natural gas development tested. In this regard, potential intrinsic complementary benefits (e.g.,

328

fast-start capabilities of and high capacity factors for natural gas which also exhibit high price volatility, unlike solar) were difficult to address in an appropriate manner with the data at hand. These characteristics make distributed electricity, especially solar and natural gas intrinsically complementary. Natural gas burns more cleanly than other hydrocarbons hence its position as a transition bridge fuel to low-carbon power future remains unchallenged. Despite the strong potential for low-cost natural gas supply and the growing potential for demand balancing capacity using traditional methods for mitigating power fluctuations observed in the study, the transition to sustainable energy systems will not occur automatically. The success of the utility sector at state, regional, and national levels will depend largely on sound sustainable business models and supporting policies and measures. As such, all the relevant institutions and regulatory systems must be proactive in enabling a policy framework that both promotes system flexibility to integrate distributed energy resources and encourage implementation of lower-carbon alternative to coal-fired generation. In the ten states studied, the number of policies designed to incentivize coal-to-gas switching measures has increased in recent years, with strong deployment of both gas-fired turbines and combined-cycle plants at the expense of coal. At the same time, however, existing tax incentives and other policies to support solar and wind energy development are still highly inadequate and uncertain. Continued advancement in technological innovation is therefore critical to improving efficiency and flexibility performance of natural gas- fired plants to support integration of fast-flexing renewables and serve as a short-term, lower-carbon alternative to coal-fired generators, while preventing long-term stranding of natural gas plants.

329

Overall, a shift towards distributed electricity systems and distributed utilities is unstoppable. Analysis of literature focusing on major technological advancements and economic market integration of natural gas and renewable energy markets indicates emergence of hybrid NG-RE power generation paradigm (see Section 6.4). However, in the absence of effective governance frameworks, regulatory innovation and electricity market design mechanisms to support this hybridization, a number of technical, institutional, financial, and capacity barriers continue to impeded this transition to distributed generation model and process. For this purpose, the results simulated in this dissertation using a system GMM model to assess the effects of NG on RE diffusion, including business models that promote a blended NG-RE power system on both the supply-side and demand-side, serve as the basis for proposing a variety of policy recommendation to foster an integrated NG-RE-based DG strategies and policies, as presented below. Analysis of the influence of policy and regulation on DER business models in the ten states indicates that utilities investing in DER currently do so, to some degree, in order to capture opportunistic value created by existing regulations and policies. The analysis in this dissertation shows that effective implementation of NG-RE hybrid power generation systems offer significant benefits to the United States’ power sector in the form of mitigating carbon emissions, supporting energy independence, contributing to economic growth, and reducing future energy shortages. Because natural gas power generation systems have gone through several stages of development and are relatively mature compared to solar and wind energy, joint development strategies and policies of these energy forms will guarantee the above benefits to utilities and consumers. For instance, these stages include cogeneration or

330

combined cooling and power (CHP), combined cycle generation (including CHP and trigeneration or combined cooling heating and power (CCHP), and CCGT plants80). On the other hand, wind and solar power generation both consist of three main stages of development, namely, offshore wind energy, onshore wind energy generation, and hybrid power generation (including wind-hydro power, wind-solar power, and wind- diesel power etc.); and solar PV, CSP, and heating and cooling systems, respectively. However, literature review and empirical analysis of the ten states do not imply that regulatory innovation and policies will continue to define and drive the DG business model indefinitely. Continued cost declines of solar and advances in technology may well create DER markets that drive utilities towards fundamentally new business models that are defined by, for example, delivery of transitory value and less characterized by regulatory and policy conditions. The dynamic econometric model and market analysis offer potential sustainable pathways for assessing joint development of a competitive hybrid NG-RE power generation market and the attendant benefits such as diversified energy market, environmental integrity, and energy security. These results are thus useful for energy planning and formulation of business practices that take into consideration market competitiveness of natural gas relative to alternative generation technologies in the electricity system, regional market conditions, and the market design innovations needed to improve grid integration of solar and wind deployment. International experience shows that countries that have successfully promoted competitive

80 CCGT efficiency now exceeds 60% and this is expected to improve to 65% over the next decade (International Energy Agency, 2017).

331

renewable energy and natural gas markets with coal, and technology-neutral competitive mechanisms such as carbon pricing and emission caps score high in relation to four essential building blocks: (1) a long-term vision that recognizes that natural gas being a source of carbon emissions, R&D goals and targets should gradually also focus on improving efficiency of gas-fired generators or promoting gas power generation with CCS, because unabated gas, just like coal, is very carbon- intensive in the long run, (2) improvement in regulatory innovation and policies to drive the structure of distributed generation business archetypes, such as solar PV business model archetypes or solar-plus-storage archetypes, to achieve these goals and targets, (3) sound governance structures and appropriate incentive mechanisms for ensuring the non-discriminatory provision of electricity markets, networks, and system operation, and (4) effective public service commissions and administrative processes for monitoring how these archetypes respond to the business environment as new regulations and policies take effect in the market (Perez-Arriaga and Knittel, 2016, p. 221).

8.2 Barriers to Existing Policies and Mechanisms The expansion of renewable energy in the U.S. would be impossible without support from state and federal policymakers. Supporting or restricting policies, which established market rules and conditions have shaped the ununiformed development of renewable energy markets across different states and regions. Although certain polices and measures have been implemented across states through long-term planning to improve energy policy design and implementation, business model innovation, and provision of customer-centric products and services, the market is still in need of

332

significant improvement in regulatory and market reform to eliminate key barriers and skewed incentives that presently impede the efficient evolution of the power sector. As highlighted earlier, natural gas and renewable energy markets interact and thus the associated technologies should be developed and deployed jointly. Therefore, if states are to transition cost-effectively to affordable, secure, and sustainable energy systems which feature more diverse energy sources and rely more heavily on NG and RE distributed electricity systems, it is essential that policymakers align consumer investment decisions with efficient power system outcomes. To be effective, these targets need to be backed by policies that ensure integrated market development; move away from a siloed, supply-driven perspective towards a systems perspective in decision making in order to increase efficiency and decrease system costs in the long- run. Deployment barriers of NG and RE blended power development can be classified into three main areas: market, financial, and economic regulation, and policy, as presented below.

8.2.1 Market Barriers Market related barriers are an obstacle for the development of renewable projects in many states. This problem is especially critical for small scale solar projects due to market risks related to suitability and reliability degree of the developers of renewable project developers. From another perspective, small scale renewables lack the necessary guarantees to obtain funding from private sector investors and require high financial resources for R&D. In addition, market barriers such as disparate market participation requirements and varying degrees of incomplete and asymmetric information across jurisdictions create difficulties for participation in distributed generation markets, relative to conventional centralized technologies. The

333

dramatic rise of low-cost shale gas and growing reliance on renewable energy for electricity generation offer complementarities spanning economic, technical, environmental, and institutional considerations and addressing these market barriers will support the emergence of sustainable energy future.

8.2.2 Financial Barriers The relatively high capital cost of renewable energy projects imposes risks that may deter their deployment by private sector investors. Regulators who evaluate viability of renewable projects from a least-cost perspective experience fundamental financial barriers. Although many states have implemented tax incentives and other policy incentives as part of the ensemble of economic signals, beyond the basic pricing structures and charges to ease these barriers, significant financial barriers remain. In the U.S. natural gas market, because of low upfront costs, when regulators deem prudent, fuel costs and associated price risks are largely passed on to consumers under mechanisms such as purchased gas adjustment (PGA). While low natural gas prices also lower overall levelized costs of energy and will continue to do so in the short term, the gas markets face financial barriers to the development of flexible natural gas pipeline service options, grid capacity expansion, investment in distribution network, and upfront investment in smart grid technologies.

8.2.3 Economic Regulation and Policy Barriers

As discussed in Chapter 3, the increasing integration of distributed resources is making the system operator role of distribution utilities increasingly complex. New responsibilities are being created to manage end-user data, and in deployment of innovative technologies such as distributed energy storage systems, electric vehicle

334

charging infrastructure, and advanced metering. Electric distribution companies and cloud services companies such as Opower (Oracle), Vivint, Honeywell, GE, Tesla, or Solar City are currently fulfilling these new roles by developing platforms where distribution network utilities and network users interact with grid connection process, make phone calls during supply interruptions, and implement metering and billing processes. For these reasons, regulations and policies must also be adapted to eliminate barriers, support balancing of electricity generation and load over larger geographic areas, allow distributed resources to access local markets and services, and facilitate interaction of these resources with market operators, suppliers, aggregators, and transmission or independent system operators.

8.3 Recommended Policies and Regulatory Mechanisms As discussed in Chapter 6, energy system integration and enhanced policy and regulatory mechanisms will bring new opportunities for optimization of distributed generation and increased efficiency in delivering electricity services. Long-term coordinated planning of the energy sector and joint infrastructure investment can incentivize consumers to play an active role in energy system management and enable distributed generation growth. Although some measures have already been widely implemented across the U.S., such as advanced metering infrastructure and smart appliances, a significant need remains for policy evaluation. Improving regulatory and business decision-making environment, including taxation costs, urban planning, international trade, and market innovation requires strong and consistent policy coordination. Currently, several policy and economic instruments have been implemented at the state level to increase the share of renewable energy in the energy mix. A review of the existing support policies for distributed resources (see Section

335

5.1) shows that most schemes are attempting to develop market-based mechanisms for demand-side management, standardization and technical performance of distributed energy systems; enhance public awareness of distributed resources; remove obstacles to grid connection; and reduce transaction and administrative costs. The following subsections outline a number of specific strategies and policies, including feed-in tariffs, tax incentives, investment subsidies, tendering, and quota-based tradable green certificates such as renewable energy certificates that different states have recommended to implement at the demand-side and supply-side. While each of these schemes offer certain advantages, consensus on which policy instrument may deliver the best results at the lowest costs remain varied. This is a crucial issue, especially if distributed energy generation is to gain a sizable share of the energy mix in future.

8.3.1 Develop Long-term Efficient Price Signals and Incentives for DER It is recommended that state governments should develop a vision for a sustainable energy future similar to the New York’s REV initiative. The central tenet of these initiatives would be to address multiple energy policy challenges, transform utilities from monoliths that produce and sell a single commodity to market makers that connect customers to a spectrum of electricity services, and track progress towards stated objectives. Defining long-term cost-efficient incentives that promote distributed utilities is essential component for meeting energy security, climate change and air quality objectives within states, as well as for ensuring the energy sector respond optimally to multiple challenges and attain set policy goals. This includes implementing forward-looking revenue trajectory that aligns utility incentives for cost- saving investments and operations with the remuneration structure.

336

8.3.2 Create Joint Innovation Programs Across States to Share Best Practices Cross-sector collaboration and innovative partnerships between distribution companies and among states need to be enhanced to achieve national sustainable energy objectives. The results of the System-GMM dynamic econometric model showed that natural gas is complementary to renewable energy. Joint innovation projects for natural gas and renewables would create market opportunities that benefit manufactures and users of renewable energy technologies while contributing to a more cost-effective transition to affordable, secure and sustainable energy systems. Collaboration with energy developers, public utility commissions, and local stakeholders is essential to building capacity and sharing best practices for diversifying utility revenue streams and promoting distributed generation.

8.3.3 Create Explicit Policy Incentives for Long-term Innovation

It is recommended that public service commissions should review and improve market incentives for distribution utilities to reward performance improvements, cost savings, and long-term innovation. Regulators should accelerate policy support for technology, finance, and market development at all stages of the innovation cycle. Many business leaders and distribution utilities urge the need for reevaluation of the existing policies and power sector’s structure to minimize conflicts of interest. With the growing distributed generation capacity in the U.S. that this dissertation demonstrates, states need to develop measurable and targeted policy incentives for distributed generation at all phases of innovation to facilitate both incremental and radical innovation. The proposed incentive mechanisms can be classified into two main categories: (1) quantity incentives, in which the government sets a target for the

337

quantity of distributed generation, and (2) price incentives, in which the government intervenes to set preferential output prices for distributed energy generation, and the market determines the quantity of distributed generation at the specified price. As discussed in Section 5.1, these mechanisms are implemented through feed-in tariffs, state RPS policies, public benefits fund, net metering, and interconnection standards. Feed-in tariffs is based on long-term purchase agreements and is the most popular mechanism used to support the development of new distributed generation projects. Initiatives such as the U.S. Department of Energy Staff Report to the Secretary on Electricity Markets and Reliability, the Quadrennial Energy review, the SunShot Initiative, Accelerate Energy Productivity 2030, Revolution Now, U.S. Energy Sector Vulnerabilities Report are key efforts to co-ordinate and promoting reliability and resiliency in the electric system. Table 8.1 summarizes key utility incentives in the ten states studied.

338

Table 8.1 Tabulation of Utility Incentives Rates / incentives Structure and offered Notes Participating Utilities Rocky Mountain Power (UT); Austin Energy (TX); Arizona Public Service Co. (AZ); Salt River Project (AZ); Tucson Electric Power Co. (AZ); Pacific Gas & Electric Co. (CA); San Diego Gas & Electric Co. (CA); Southern California Edison Co. (CA); Georgia Power Co. (GA); NSTAR Electric Company (MA); Duke Energy Progress (NC); Jersey Central Power & Lt Co. (NJ); Nevada Power 339 Three time-of- Co. (NV); Sierra Pacific Power Co. (NV); New York State Elec & Gas Corp day periods, two (NY); Rochester Gas & Electric Corp. (NY); Constellation Energy Services NY, Time-of-use option seasons Inc. (NY); Southwestern Electric Power Co. (TX). Austin Energy (TX); Arizona Public Service Co. (AZ); Salt River Project (AZ); Pacific Gas & Electric Co. (CA); San Diego Gas & Electric Co. (CA); Georgia Power Co. (GA); Town of Wellesley (MA); Duke Energy Progress (NC); Jersey Green choice Central Power & Lt Co. (NJ); Nevada Power Co. (NV); New York State Elec & option Gas Corp. (NY); TXU Energy Retail Co LP (TX); PacifiCorp (UT). Austin Energy (TX); Chicopee Electric Light (MA); New Braunfels Utilities (TX); Arizona Public Service Co. (AZ); Pacific Gas & Electric Co. (CA); Cobb Electric Membership Corp. (GA); Massachusetts Electric Co. (MA); Duke Energy Bill credit for Carolinas, LLC. (NC); Jersey Central Power & Lt Co. (NJ); Nevada Power Co. solar energy (NV); Long Island Power Authority (NY); Oncor Electric Delivery Company Residential solar generation LLC (TX); PacifiCorp (UT).

CPS Energy, San Antonio (TX); Austin Energy (TX); Pacific Gas & Electric Co. (CA); Arizona Public Service Co. (AZ); Salt River Project (AZ); Pacific Gas & Electric Co. (CA); Southern California Edison Co. (CA); San Diego Gas & Electric Co. (CA); Georgia Power Co (GA); Snapping Shoals El Member Corp. (GA); Massachusetts Electric Co. (MA); NSTAR Electric Company (MA); Duke Energy Carolinas, LLC. (NC); Duke Energy Progress (NC); NJ Clean Energy Program (NJ); Nevada Power Co. (NV); NYSERDA (NY); Niagara Mohawk Power Corp. (NY); Long Island Power Authority (NY); Consolidated Edison Co. Energy efficiency Extensive set of (NY); CenterPoint Energy (TX); Oncor Electric Delivery Company LLC. (TX); rebates programs City of San Antonio (TX); PacifiCorp (UT); City of Logan (UT).

340 Promotes Salt River Project (AZ); Austin Energy (TX); Pacific Gas & Electric Co. (CA); resource Georgia Power Co. (GA); Massachusetts Electric Co. (MA); Duke Energy

conservation in Progress (NC); Public Service Elec & Gas Co. (NJ); Consolidated Edison Co-NY new Inc. (NY); Oncor Electric Delivery Company LLC (TX); Brigham City Green building construction Corporation (UT). Arizona Public Service Co. (AZ); Austin Energy (TX); Pacific Gas & Electric Co. (CA); Duke Energy Carolinas (NC); Public Service Elec & Gas Co. (NJ); Nevada Residential demand Power Co. (NV); Consolidated Edison Co. (NY); City of San Antonio (TX); response PacifiCorp (UT). Austin Energy (TX); Arizona Public Service Co. (AZ); Pacific Gas & Electric Co. (CA); Duke Energy Carolinas, LLC. (NC); Nevada Power Co. (NV); Long Island Residential inside Five inclining Power Authority (NY); Oncor Electric Delivery Company LLC (TX); PacifiCorp city limits blocks (UT).

Residential outside Three inclining Austin Energy (TX); Pacific Gas & Electric Co. (CA); Duke Energy (NC); city limits blocks Nevada Power Co. (NV); PacifiCorp (UT). Two-tier CPS Energy, San Antonio (TX); Arizona Public Service Co. (AZ); San Diego Gas Residential service inclining block & Electric Co. (CA); Duke Energy Progress (NC); New York State Elec & Gas tariff rate Corp. (NY); PacifiCorp (UT). Austin Energy (TX); Arizona Public Service Co. (AZ); Pacific Gas & Electric Co. Provides (CA); Massachusetts Electric Co. (MA); Duke Energy Carolinas, LLC. (NC); Plug-in electric incentives for Nevada Power Co. (NV); Oncor Electric Delivery Company LLC (TX); vehicles charging stations PacifiCorp (UT). 341

8.3.4 Enhance Market Transparency by Enabling Participation in Long-term Capacity Markets

As suggested earlier in this dissertation, to allow small new distribution entrants to compete on a level playing field, regulators should remove information asymmetries and opportunities that promote market power abuse. As shown in the distributed utilities model (see Section 6.4.1), the short-term effect of natural gas on solar is very small but significant in the long-term. For instance, time-varying shocks ranging from changes in policy and their effectiveness as well as changes in economic and market factors can have significant effects on both natural gas and renewable energy investments. Where large market players have a significant advantage over smaller ones, these changes disproportionately affect small market players. Therefore, instead of relying on bilateral arrangements that favor large distribution utilities, market transactions should be concentrated in centralized market sessions. Centralized market sessions enhance market transparency because they favor liquidity and reveal information of market participants instead of relying on bilateral arrangements between distribution companies. Furthermore, policies and market mechanisms that allow all market players (large and small) to participate in real-time markets and self-manage their energy imbalances can enhance market transparency. For example, one system operator can find it advantageous to self-manage its energy imbalances (by summing up the positive and negative deviations of several power plants in all its portfolio) and settle only their net deviations in the balancing market, while another operator could avoid this aggregation by balancing within the same portfolio. Regulations and electricity market designs should leverage the opportunity brought by digitization, ICTs,

342

transaction models, and increased access to energy information to improve liquidity and transparency in distributed energy generation markets.

8.4 Climate Change Challenges As explained in Section 3.1.10, the increasing vulnerabilities of electricity systems to severe weather and climate change is substantial and growing. These impacts threaten electricity generation, transmission, and distribution networks. As a result, utilities are under significant pressure to develop adaptation strategies to respond to current and future climate impacts as well as regulatory pressure to mitigate. To meet these climate change challenges, there is need for expansion of distributed solar and wind power and improvement in coordination of distributed generation assets between consumers and other network agents to deliver sustainable economic and secure electricity supply. According to the The Second Installment of the Quadrennial Energy Review (QER), extreme weather is the leading cause of electric power outage events. For instance, all 12 of the large-scale disruptions to the grid in 2015 were weather related (U.S. Department of Energy, 2017b). Given that the problem of climate change influences virtually every aspect of utility business practices, it is going to be interesting to see how the institutional narrative of regulatory reforms in New York, California and elsewhere evolves in the coming years to foster decarbonized electricity systems, integrate new energy technologies, and support consumer cost minimization.

8.5 Next Generation Utilities and System Flexibility

This dissertation has concluded that technologies currently in the innovation pipeline that support distributed resources need strong policy support to facilitate the

343

next generation utilities of the future. Improving policies and regulation of distribution utilities at the state and regional levels is needed to enable development of more efficient distribution utility business models. This study has identified three emerging policy trends for the next generation utilities: (1) need for new policy approaches to support electricity market design and innovation, (2) desire to support a strong functioning energy sector governance structure (and related institutions) and diversity of investors and project sizes, and (3) push to integrate distributed electricity technologies into wholesale spot markets. It is in response to the demands for state of the art regulatory tools and policies, including an incentive-compatible menu of contracts and other related trends (see Section 3.1) that the distinctions between the traditional policy categories and the next generation regulatory innovations for an evolving and uncertain electricity landscape are beginning to break down. The recent price and quantitate incentive policy approaches (see Section 5.1) surveyed in this dissertation may therefore point the way toward the next generation of distributed utility business models and policies. Table 8.2 summarizes the policy implications of the statistical regression and empirical analysis results.

344

Table 8.2 Summary of Policy Implications

• To ensure centralized and distributed resources can jointly and collaboratively operate in the market, a long-term effieicnt price signal, incentives and policies that obligate utilities to invest in DERs needs to be implemented.

• To keep pace with the rapidly changing utility landscape, electricity distribution utilities need to invest in business model innovation programs by creating joint innovation programs across states as well as supporting research and development efforts and dissemination of best practices between network utilities.

• The electric grid is cyber and physical system and the proliferation of DERs will increase vulnetabilities of these systems. Moving forward new approaches to strengthening cybersecurity standards and privacy regulations will be required to address the influx of DERs and ICTs in electricity networks.

• To minimize information asymmetries and improve organization of the electricity market in a more distributed future, a more comprehensive implementation of policies that enhance market transparency by enabling equal participation in long-term capacity markets is recommended.

• Finally, policymakers and regulators must facilitate business practices that promote integration of all resources, be they centralized or distributed, for example, through adoption of IoT technologies for greater DER visibility, control, and achievement of other public objectives.

345

The above five policies and market innovations are critical to the development of a distributed future because limitations of aging energy infrastructure, decreasing costs of renewable energy technologies, increasing customer engagement, and the growing share of distributed electricity generation will continue to present challenges to the utility industry. Proponents of distributed resources and a diversified electricity mix point to its potential to address overdependence on traditional modes of electricity supply, stimulate opportunities for procuring cheaper and more affordable electricity, and to provide pathways for sustainable energy future. Collectively, these factors have and will continue to compel regulators and policymakers to look for new business models, market designs, and frameworks that are better adapted to these changing market circumstances. For example, several utilities across the country are also trialing hybrid policies that include a combination of traditional FITs and net metering or auctions, adoption of premium FITs for renewable electricity producers in spot market. Future research is therefore needed to address the limitations related to this dissertation such as these hybrid policy alternatives as electricity market design requirements and macroeconomic environment changes.

346

REFERENCES

Abdelkafi, N., & Täuscher, K. (2016). Business models for sustainability from a system dynamics perspective. Organization & Environment, 29(1), 74-96. Abid, M. (2016). Impact of economic, financial, and institutional factors on CO2 emissions: Evidence from sub-Saharan Africa economies. Utilities Policy, 41, 85- 94. Abolhosseini, S., & Heshmati, A. (2014). The main support mechanisms to finance renewable energy development. Renewable & Sustainable Energy Reviews, 40, 876-885. ACC. (2011). Arizona corporation commission and southwest gas corporation (docket no. G-01551A-10-0458). in the matter of the application of southwest gas corporation for approval of the establishment of just and reasonable rates. Retrieved from http://images.edocket.azcc.gov/docketpdf/0000133931.pdf Accenture. (2014). The Digitally Enabled Grid: How can utilities survive energy demand disruption? Accenture’s Digitally Enabled Grid program—2014 edition. United States: Accenture. Acikkalp, E., Aras, H., & Hepbasli, A. (2014). Advanced exergy analysis of an electricity-generating facility using natural gas. Energy Conversion and Management, 82, 146. Ackerman, F., and Stanton, E. A. (2008). Can Climate Change Save Lives? A comment on ‘Economy-wide estimates of the implications of climate change: Human health.’ Ecological Economics 66:1, 8-13. Ackermann, T., Anderson, G., & Söder, L. (2001). Distributed generation: A definition. Electric Power Systems Research, 57, 195-204. AEG. (2017). AEG Proposal to Provide NJCEP Program Administration and Management Services. New Jersey: Applied Energy Group (AEG). Aggarwal, S., & Harvey, H. (2013). Rethinking policy to deliver a clean energy future; The Electricity Journal, 26(8), 7-22. Agüero, J. R., & Khodaei, A. (2015). Roadmaps for the utility of the future. The Electricity Journal, 28(10), 7-17. Alanne, K., & Saari, A. (2006). Distributed energy generation and sustainable development. Renewable and Sustainable Energy Reviews, 10(6), 539-558. Aldina, R., & Soden, J. (2013). A guide to inform institutions about participation in PJM’s demand response programs.

347

Amit, R., & Schoemaker, P. J. H. (1993). Strategic assets and organizational rent. Strategic Management Journal, 14(1), 33-46. Amit, R., & Zott, C. (2001). Value creation in E-business. SMJ Strategic Management Journal, 22(6-7), 493-520. Amor, M. B., Lesage, P., Pineau, P., & Samson, R. (2010). Can distributed generation offer substantial benefits in a northeastern American context? A case study of small-scale renewable technologies using a life cycle methodology. Renewable and Sustainable Energy Reviews, 14(9), 2885-2895. Amusa, M. (2013). How the biggest U.S. utilities are planning for the future. Retrieved from http://www.utilitydive.com/news/how-the-biggest-us-utilities-are- planning-for-the-future/130302/. Arellano, M., & Bover, O. (1995). Another look at the instrumental variable estimation of error-components models. Journal of Econometrics, 68(1), 29-51. doi:10.1016/0304-4076(94)01642-D Arent, D., Arndt, C., Tarp, F., & Zinaman, O. (2017). Moving forward. The political economy of clean energy transitions (). GB: Oxford University Press. ASCE. (2013). The 2013 Report Card for America's Infrastructure. United States: American Society of Civil Engineers (ASCE). ASCE. (2017). The 2017 Infrastructure Report Card: Energy. Reston, VA: American Society of Civil Engineers (ASCE). Asiedu, E., & Lien, D. (2011). Democracy, foreign direct investment and natural resources. Journal of International Economics, 84(1), 99-111. Retrieved from http://www.sciencedirect.com/science/article/pii/S002219961000125X Astoria, R. (2017). On the radicality of New York’s reforming the energy vision. The Electricity Journal, 30(5), 54-58. Athawale, R., & Felder, F. (2016). Residential rate design and death spiral for electric utilities: Efficiency and equity considerations. In F. P. Sioshansi (Ed.), Future of utilities - utilities of the future: How technological innovations in distributed energy resources will reshape the electronic power sector [ISBN: 978-0-12- 804249-6] (1st ed., pp. 193-209). London: Elsevier. Bade, G. (2017). Zibelman exit interview: How the New York REV is paving the way for transactive energy. Retrieved from http://www.utilitydive.com/news/zibelman-exit-interview-how-the-new- york-rev-is-paving-the-way-for-transac/435252/

348

Baden-Fuller, C., & Morgan, M. S. (2010). Business models as models. Long Range Planning, 43(2-3), 156-171. Baltagi, B. H., Bun, M. J. G., & Sarafidis, V. (2015). Dynamic Panel Data Models. Amsterdam, The Netherlands: Amsterdam School of Economics. Barbose, G. (2014). Renewables Portfolio Standards in the United States: A Status Update. Washington, D.C: Lawrence Berkeley National Laboratory (LBNL). Barbose, Galen L., Goldman, Charles A., Hoffman, Ian M., Billingsley, Megan. (2013). The future of utility customer-funded energy efficiency programs in the USA: Projected spending and savings to 2025. Energy Efficiency Energy Efficiency, 6(3), 475-493. Barnett, N. J., Parry, G., Saad, M., Newnes, L. B., & Goh, Y. M. (2013). Servitization: Is a paradigm shift in the business model and service enterprise required? Strategic Change, 22(3‐4), 145-156. Bauknecht, D., Leprich, U., Späth, Philipp, Skytte, Klaus, & Esnault, B. (2007). DG-GRID: Regulating Innovation & Innovating Regulation. Energy research Centre of the Netherlands (ECN): Energy Intelligent Europe (EIE) Programme. Bayulgen, O., & Ladewig, J. W. (2017). Vetoing the future: Political constraints and renewable energy. Environmental Politics, 26(1), 49-70. BEA. (2017). Per capita real GDP by state (chained 2009 dollars, all industry total). Retrieved from https://www.bea.gov/iTable/index_regional.cfm Behrangrad, M. (2015). A review of demand side management business models in the electricity market. Renewable and Sustainable Energy Reviews, 47 Bengtsson, M., & Kock, S. (2000). "Coopetition" in business networks to cooperate and compete simultaneously (technology information). Industrial Marketing Management, 29(5) Besley, T., & Case, A. (2000). Unnatural experiments? estimating the incidence of endogenous policies. The Economic Journal, 110(467), 672-694. Retrieved from http://www.jstor.org/stable/2667771 Bhattacharyya, S.C. (2011). Energy economics: Concepts, issues, markets and governance (1st ed.). London; New York: Springer. Biresselioglu, M. E., Kilinc, D., Onater-Isberk, E., & Yelkenci, T. (2016). Estimating the political, economic and environmental factors’ impact on the installed wind capacity development: A system GMM approach. Renewable Energy, 96, 636- 644.

349

Biswas, I., Avittathur, B., & Chatterjee, A. K. (2016). Impact of structure, market share and information asymmetry on supply contracts for a single supplier multiple buyer network. European Journal of Operational Research, 253(3), 593- 601. Bloomberg. (2016). Bloomberg Terminal. Bloomberg professional. Retrieved from [Online] Blundell, R., & Bond, S. (1998). Initial conditions and moment restrictions in dynamic panel data models doi://dx.doi.org/10.1016/S0304-4076(98)00009-8. BNEF. (2016). New Energy Outlook 2016: Long-Term Projections of the Global Energy Sector. London, United Kingdom: Bloomberg New Energy Finance. Boehnke, J., & Wüstenhagen, R. (2007). Business models for distributed energy technologies - evidence from German cleantech firms. Unpublished manuscript. Bonvillian, W. B., & Weiss, C. (2009). Stimulating innovation in energy technology: Meeting the worlds urgent need for a cleaner and more efficient energy system will require a more effective strategy that incorporates technology neutrality, international collaboration, institutional change, and a more fine-tuned understanding of the innovation process; Issues in Science and Technology, 26(1) Boons, F., & Lüdeke-Freund, F. (2013). Business models for sustainable innovation: State-of-the-art and steps towards a research agenda. Journal of Cleaner Production, 45, 9-19. Booth, A., Mohr, N. & Peters, P. (2016). The digital utility: New opportunities and challenges. Retrieved from http://www.mckinsey.com/industries/electric-power- and-natural-gas/our-insights/the-digital-utility-new-opportunities-and-challenges Botta, E., & Koźluk, T. (2014). Measuring Environmental Policy Stringency in OECD Countries a Composite Index Approach. Paris: OECD Publishing. Bouncken, R. B., & Fredrich, V. (2016). Business model innovation in alliances: Successful configurations. Journal of Business Research, 69(9), 3584-3590. Bowen, E., & Lacombe, D. J. (2017). Spatial dependence in state renewable policy: Effects of renewable portfolio standards on renewable generation within NERC regions. The Energy Journal, 38(3) BP. (2016). Statistical Review of World Energy 2016. London, United Kingdom: BP. BQDM. (2016). Brooklyn Queens Demand Management Program (BQDM). Implementation and Outreach Plan, Case 14-E-0302 - Brooklyn Queens Demand Management Program. New York: Consolidated Edison Company of New York, Inc.

350

Bravo, D., Sauma, E., Contreras, J., de la Torre, S., Aguado, J. A., & Pozo, D. (2016). Impact of network payment schemes on transmission expansion planning with variable renewable generation. Energy Economics, 56, 410-421. Bresler, S., Centolella, P., Covino, S., & Sotkiewicz, P. (2013). Smarter demand response in RTO markets: The evolution toward price responsive demand in PJM. In F. P. Sioshansi (Ed.), Energy efficiency towards the end of demand growth (). Amsterdam: Elsevier Science/AP. Broadman, H. G., & Montgomery, W. D. (1983). Natural Gas Markets After Regulation: Methods of Analysis and research Needs. Washington, D.C.: Resources for the Future. Brocks, A., Nyangon, J., & Taminiau, J. (2016). Utility 2.0: A multi-dimensional review of New York's Reforming the Energy Vision (REV) and Great Britain's RIIO utility business models. Newark, DE: A report of the Energy & Environmental Policy Analysis (EEPA) Services Program, Center for Energy and Environmental Policy, University of Delaware. Brooks, C. (2015). The periodic table of the electric utility landscape: A series of visual tools for enhanced policy analysis. The Electricity Journal, 28(6), 82-95. Bruckner, T., Bashmakov, I. A., Mulugetta, Y., Chum, H., De la Vega Navarro, A, Edmonds, J., . . . Zhang, X. (2014). Chapter 7 - Energy systems. Contribution of Working Group III to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (pp. 511-597). Cambridge, U.K. and New York, NY, USA: Cambridge University Press. Burger, C., & Weinmann, J. (2016). European utilities: Strategic choices and cultural prerequisites for the future. In F. P. Sioshansi (Ed.), Future of utilities - utilities of the future: How technological innovations in distributed energy resources will reshape the electric power sector (pp. 303-322). Amsterdam; Heidelberg: Elsevier, Academic Press. Burks, S. K., Hoehn, M. A., & Van Over, C. J. (2009). Protecting IP rights in the context of ARPA-E. Intellectual Property & Technology Law Journal, 21(9) Burnett, J. W., & Madariaga, J. (2017). The convergence of U.S. state-level energy intensity. Energy Economics, 62, 357-370. Burr, M. T. (2007). Razing the regulatory compact. Arlington: Public Utilities Reports, Incorporated. Butler, L., & Neuhoff, K. (2008). Comparison of feed-in tariff, quota and auction mechanisms to support wind power development. Renewable Energy, 33(8), 1854-1867.

351

Byrne, J., & Kurdgelashvilli, L. (2003). The role of policy in PV industry growth: Past, present and future. In A. Luque, & S. Hegedus (Eds.), Handbook of photovoltaic science and engineering (). Hoboken, NJ: Wiley. Byrne, J., & Taminiau, J. (2016). A review of sustainable energy utility and energy service utility concepts and applications: Realizing ecological and social sustainability with a community utility. Wiley Interdisciplinary Reviews: Energy and Environment, 5(2), 136-154. Byrne, J., Martinez, C., & Ruggero, C. (2009). Relocating energy in the social commons: Ideas for a sustainable energy utility. Bulletin of Science, Technology & Society, 29(2), 81-94. Byrne, J., Taminiau, J., Kim, K. N., Lee, J., & Seo, J. (2017). Multivariate analysis of solar city economics: Impact of energy prices, policy, finance, and cost on urban photovoltaic power plant implementation. Wiley Interdisciplinary Reviews: Energy and Environment, 6(4). Byrne, J., Taminiau, J., Kim, K. N., Seo, J., & Lee, J. (2016). A solar city strategy applied to six municipalities: Integrating market, finance, and policy factors for infrastructure‐scale photovoltaic development in Amsterdam, London, Munich, New York, Seoul, and Tokyo. Wiley Interdisciplinary Reviews: Energy and Environment, 5(1), 68-88. doi:10.1002/wene.182. Byrne, J., Wang, Y., & Yu, J. (2005). Lessons from a Comparative Analysis of California and PJM Electricity Restructuring Models. Newark, DE: Center for Energy and Environmental Policy, University of Delaware. CAISO. (2014). What the Duck Curve Tells Us about Managing a Green Grid. California, United States: California Independent System Operator (CAISO). Campbell, R. J. (2015). Cybersecurity Issues for the Bulk Power System. Washington: Congressional Research Service. Carayannis, E. G., Grigoroudis, E., Sindakis, S., & Walter, C. (2014). Business model innovation as antecedent of sustainable enterprise excellence and resilience. Journal of the Knowledge Economy, 5(3), 440-463. Carley, S. (2009). State renewable energy electricity policies: An empirical evaluation of effectiveness. Energy Policy, 37(8), 3071-3081. Retrieved from http://www.sciencedirect.com/science/article/pii/S0301421509002250 Carley, S. (2012). Energy demand‐side management: New perspectives for a new era. Journal of Policy Analysis and Management, 31(1), 6-32. Retrieved from http://onlinelibrary.wiley.com/doi/10.1002/pam.20618/abstract

352

Carson, E., & Kreilis, J. (2016). New York Adopts New Revenue Model for Electric Utilities under REV. United States: EnerKnol.

Casadesus-Masanell, R., & Ricart, J. E. (2010). From strategy to business models and onto tactics. Long Range Planning, 43(2–3), 195-215.

Central Hudson. (2017). Central Hudson’s non-wires alternative initiatives. Retrieved from https://www.cenhud.com/workingwithus/non-wires-alternative-opportunities Cheng, Q., & Yi, H. (2017). Complementarity and substitutability: A review of state level renewable energy policy instrument interactions. Renewable and Sustainable Energy Reviews, 67, 683-691. Chesbrough, H. (2007). Business model innovation: It's not just about technology anymore. Strategy & Leadership, 35(6), 12-17. Chesbrough, H. (2010). Business model innovation: Opportunities and barriers. Long Range Planning, 43(2-3), 354-363. Chesbrough, H., & Rosenbloom, R. S. (2002). The role of the business model in capturing value from innovation: Evidence from Xerox corporation's technology spin-off companies. Industrial and Corporate Change, 11(3), 529. Chow, G. (2011). Usefulness of Adaptive and Rational Expectations in Economics. Christensen, C. M. (2006). The ongoing process of building a theory of disruption (dialogue on the effects of disruptive technology on firms and industries). Journal of Product Innovation Management, 23(1) Christensen, C. M., & Bower, J. L. (2009). Customer power, strategic investment, and the failure of leading firms., 146-172. Retrieved from http://www.econis.eu/PPNSET?PPN=588532010 Chupka, M. W., Earle, R., Fox-Penner, P., & Hledik, R. (2008). Transforming America's power industry: The investment challenge 2010-2030. Washington: The Brattle Group. Coase, R. H. (1973). Business organization and the accountant. In J. M. Buchanan, & G. F. Thirlby (Eds.), L.S.E. essays on cost (pp. 116). New York: New York University Press. Cochran, J. (2012). Integrating variable renewable energy in electric power markets best practices from international experience. Golden, CO: National Renewable Energy Laboratory (U.S.), Joint Institute for Strategic Energy Analysis., Clean Energy Ministerial.

353

Cochran, J. (2013). Market evolution: Wholesale electricity market design for 21st century power systems. Golden, CO.: National Renewable Energy Laboratory (U.S.). Cochran, J., Zinaman, O., Logan, J., & Arent, D. (2014). Exploring the Potential Business Case for Synergies Between Natural Gas and Renewable Energy. United States: Joint Institute for Strategic Energy Analysis (JISEA). Codognet, M. (2004). The Shipper as the Architect of Contractual Relations in Access to Natural Gas Networks. Tuscon, AZ: Proceedings of the 8th Annual ISNIE Conference. The International Society for New Institutional Economics. Cole, D. H. (1986). Reviving the federal power act's comprehensive plan requirement: A history of neglect and prospects for the future. Environmental Law, 16(3), 639. Commonwealth of Massachusetts. (2017). Grid Modernization and Integrating Distributed Generation (Order DPU 12-76-B). Boston, Massachusetts: Commonwealth of Massachusetts. Con Edison. (2017). Con Edison non-wire solutions. Retrieved from https://www.coned.com/en/business-partners/business-opportunities/non- wires-alternatives Conkling, J. (2017). Shedding light on solar potential in all 50 U.S. states. Retrieved from https://blog.google/products/maps/shedding-light-solar-potential-all-50-us- states/ Cooper, J. (2016). The innovation platform enables the internet of things. In F. P. Sioshansi (Ed.), Future of utilities. utilities of the future. how technological innovations in distributed energy resources will reshape the electric power (pp. 91-110). New York: Elsevier. Cossent, R., & Gómez, T. (2013). Implementing incentive compatible menus of contracts to regulate electricity distribution investments. Utilities Policy, 27, 28-38. Retrieved from //www.sciencedirect.com/science/article/pii/S0957178713000532 Costello, K. (2006). Overview of Natural Gas Economics, Market Structure and Regulation. The National Regulatory Research Institute (NRRI). Costello, K. W., & Hemphill, R. C. (2014). Electric utilities' 'death spiral': Hyperbole or reality? Electricity Journal, 27(10), 7-26. Darghouth, N. R., Wiser, R. H., & Barbose, G. (2016). Customer economics of residential photovoltaic systems: Sensitivities to changes in wholesale market design and rate structures. Renewable and Sustainable Energy Reviews, 54, 1459- 1469.

354

Davidson, C., James, T. L., Margolis, R., Fu, R., & Feldman, D. (2014). U.S. Residential Photovoltaic (PV) System Prices, Q4 2013 Benchmarks: Cash Purchase, Fair Market Value, and Prepaid Lease Transaction Prices. Golden, CO: USDOE Office of Energy Efficiency and Renewable Energy Solar Energy Technologies Office. Dong, C. G. (2012). Feed-in tariff vs. renewable portfolio standard: An empirical test of their relative effectiveness in promoting wind capacity development doi://dx.doi.org/10.1016/j.enpol.2011.12.014 DSIRE. (2017a). Incentives/policies for renewables and efficiency . Retrieved from http://www.dsireusa.org DSIRE. (2017b). New York clean energy standard. Retrieved from http://programs.dsireusa.org/system/program/detail/5883 Duncan, R. B. (1976). The ambidextrous organization: Designing dual structures for innovation. In R. H. Kilmann, L. R. Pondy & D. Selvin (Eds.), The management of organization (pp. 167-188). New York: North-Holland. e21 Initiative. (2014). Phase I Report: Charting a Path to a 21st Century Energy System in Minnesota. Minnesota: e21 Initiative. Eisen, J. B. (2012). Distributed energy resources, "virtual power plants," and the smart grid UR Scholarship Repository.

Engelken, M., Römer, B., Drescher, M., Welpe, I. M., & Picot, A. (2016). Comparing drivers, barriers, and opportunities of business models for renewable energies: A review. Renewable and Sustainable Energy Reviews, 60, 795-809. Enkel, E., & Mezger, F. (2013). Imitation processes and their application for business model innovation: An explorative study. International Journal of Innovation Management, 17(1) EPRI. (2014). Time and Locational Value of DER: Methods and Applications. Palo Alto, CA: Electric Power Research Institute (EPRI). Fanelli, V., Maddalena, L., & Musti, S. (2016). Modelling electricity futures prices using seasonal path-dependent volatility. Applied Energy, 173, 92-102. Farrell, J. (2014). Beyond Utility 2.0 to Energy Democracy. United States: Institute for Local Self Reliance. Faruqui, A., Hajos, A., Hledik, R. M., & Newell, S. A. (2010). Fostering economic demand response in the Midwest ISO. Energy, 35(4), 1544-1552. Felder, F. A., & Athawale, R. (2014b). The life and death of the utility death spiral. The Electricity Journal, 27(6), 9-16.

355

Felder, F. A., & Athawale, R. (2016). Optimizing New York’s reforming the energy vision. Utilities Policy, 41, 160-162. Feldman, D., Barbose, G., Margolis, R., Bolinger, M., Chung, D., Fu, R., . . . Wiser, R. (2016). Photovoltaic System Pricing Trends: Historical, Recent, and Near- Term Projections 2015 Edition. Washington, D.C.: National Renewable Energy Laboratory (U.S.), United States., Office of the Assistant Secretary of Energy Efficiency and Renewable Energy, United States., Department of Energy, Office of Scientific and Technical Information. FERC. (1987). FERC Order 500. Regulation of Natural Gas Pipelines After Partial Wellhead Decontrol.” FERC Stats and Regs Paragraph 30,761Federal Energy Regulatory Commission (FERC). FERC. (2010). Demand Response Compensation in Organized Wholesale Energy Markets. ( Lanham: Federal Information & News Dispatch, Inc. Retrieved from ProQuest Advanced Technologies & Aerospace Journals Retrieved from https://www.ferc.gov/CalendarFiles/20100913090259- Weishaar,%20Demand%20Response%20Supporters.pdf. FERC. (2014). Order authorizing proposed merger of Exelon corporation and Pepco holdings; Retrieved from https://www.ferc.gov/whats-new/comm- meet/2014/112014/e-1.pdf FERC. (2016). Assessment of Demand Response and Advanced Metering. (United States: Federal Energy Regulatory Commission. Retrieved from https://www.ferc.gov/legal/staff-reports/2016/DR-AM-Report2016.pdf. Fine, S., Fitzgerald, S., & Ingram, J. (2011). Potential impacts of environmental regulation on the US generation fleet. Washington, DC.: Edison electric institute. Fletcher, S. (2009). 'Goldilocks range' of prices. Oil and Gas Journal, 107(23) Ford, A. (1999). Cycles in competitive electricity markets: A simulation study of the western United States. (statistical data included). Energy Policy, 27(11) Fox-Penner, P.S. (2010). Smart power: Climate change, the smart grid, and the future of electric utilities. Washington: Island Press. Frantzis, L., Graham, S., Katofsky, R., & Sawyer, H. (2008). Photovoltaics Business Models. United States: USDOE. doi:10.2172/924651. Fthenakis, V. (2015). Considering the total cost of electricity from sunlight and the alternatives. Proceedings of the IEEE, 103(3), 283-286. doi:10.1109/JPROC.2015.2399052

356

Funkhouser, E., Blackburn, G., Magee, C., & Rai, V. (2015). Business model innovations for deploying distributed generation: The emerging landscape of community solar in the U.S. Energy Research & Social Science, 10, 90-101. Gawer, A. (2009). Platforms, markets and innovation. Cheltenham [u.a.]: Elgar. Gil, H. A., & Lin, J. (2013). Wind power and electricity prices at the PJM market. IEEE Transactions on Power Systems, 28(4), 3945-3953. Giordano, V., & Fulli, G. (2012). A business case for smart grid technologies: A systemic perspective. Energy Policy, 40(1), 252-259. Goldman, C. A., Satchwell, A., Cappers, P., & Hoffman, I. M. (2013). Utility business models in a low load growth/high DG future: Gazing into the crystal ball? Boise, ID: Lawrence Berkeley National Laboratory, Committee on Regional Electric Power Cooperation (CREPC)/State-Provincial Steering Committee (SPSC) Meeting. Gopal, S. (2013). Urban renewabilization in India through public-public partnerships: A feasibility study of applying the "sustainable energy utility" model to the city of thane. PhD thesis, Center for Energy and Environmental Policy, University of Delaware. Gottstein, M., & Schwartz, L. (2010). The Role of Forward Capacity Markets in Increasing Demand-Side and Other Low-Carbon Resources: Experience and Prospects. Montpelier, VT: The Regulatory Assistance Project (RAP). Graffy, E., & Kihm, S. (2014). Does disruptive competition mean a death spiral for electric utilities? Energy Law Journal., 35(1), 1-44. GridSolar. (2016). Boothbay Sub-Region Smart Grid Reliability Pilot Project & Docket No. 2011-138. Maine, USA: Central Maine Power Co., Request for Approval of Non-Transmission Alternative (NTA), Pilot Project for the Mid- Coast and Portland Areas. Groat, C., & Grimshaw, T. (2012). Fact-Based Regulation for Environmental Protection in Shale Gas. Austin: University of Texas: Energy Institute. Growitsch, C., & Stronzik, M. (2014). Ownership unbundling of natural gas transmission networks: Empirical evidence. Journal of Regulatory Economics, 46(2), 207-225. doi:10.1007/s11149-014-9252-0 GTM Research, & SEIA. (2015). U.S. Solar Market Insight Q2 2015.Market Insight Boston: Greentech Media and Solar Energy Industries Association.

357

Hall, S., & Roelich, K. (2016). Business model innovation in electricity supply markets: The role of complex value in the United Kingdom. JEPO Energy Policy, 92, 286-298. Hamel, G. (2000). Leading the revolution: How to thrive in turbulent times by making innovation a way of life. Boston, Mass.: Harvard Business School Press. Hannon, M. J., Foxon, T. J., & Gale, W. F. (2013). The co-evolutionary relationship between energy service companies and the UK energy system: Implications for a low-carbon transition. Energy Policy, 61, 1031-1045. Hansen, E. G., Große-Dunker, F., & Reichwald, R. (2009). Sustainability innovation Cube—A framework to evaluate sustainability-oriented innovations. International Journal of Innovation Management, 13(04), 683-713. Harvey, F. (2012). 'Golden age of gas' threatens renewable energy, IEA warns. The Guardian Retrieved from https://www.theguardian.com/environment/2012/may/29/gas-boom- renewables-agency-warns Hayashi, A. M. (2009). Do you have a 'plan B'? (intelligence)(strategy) (business models). MIT Sloan Management Review, 51(1) He, X., Delarue, E., D'haeseleer, W., & Glachant, J. (2011). A novel business model for aggregating the values of electricity storage. Energy Policy, 39(3), 1575- 1585. Heeter, J., Beiter, P., Flores-Espino, F., Hurlbut, D., & Liu, C. Quantifying the level of cross-state renewable energy transactions. Clean Energy States Alliance Webinar, 29. Hefley, W. E., & Wang, Y. (Eds.). (2015). Economics of unconventional shale gas development: Case studies and impacts. Cham: Springer International Publishing. Hess, D. J. (2016). The politics of niche-regime conflicts: Distributed solar energy in the united states. Environmental Innovation and Societal Transitions, 19(4), 42- 50. Hirth, L. (2015). The optimal share of variable renewables: How the variability of wind and solar power affects their welfare-optimal deployment. Energy Journal, 36(1), 149-184. Hogan, M. (2013). Aligning power markets to deliver value. The Electricity Journal, 26(8), 23-34. HPUC. (2017). Instituting a Proceeding to Investigate the Implementation of Reliability Standards for Hawaiian Electric Company, Inc., Hawaii Electric Light

358

Company, Inc., and Maui Electric Company, Limited. (Docket No. 2011- 0206). Hawaii: Hawaii Public Utilities Commission (HIPUC). Huang, L. (2013). Natural gas and electricity: A primer on energy markets International Energy Agency. (2016). World energy outlook 2016. Paris, France: International Energy Agency (IEA). doi:10.1787/weo-2016-en International Energy Agency. (2017). Energy technology perspectives 2017. Retrieved from http://www.iea.org/etp2017/summary/?utm_content=buffer279a6 IPMVP committee and others. (2012). International performance measurement and verification protocol: Concepts and options for determining energy and water savings, vol. I. tech rep. Retrieved from www.evo-world.org

Jenkins, J. D., & Pérez-Arriaga, I. J. (2017). Improved regulatory approaches for the remuneration of electricity distribution utilities with high penetrations of distributed energy resources. The Energy Journal, 38(3), 28-29. Jenkins, K., McCauley, D., Heffron, R., Stephan, H., & Rehner, R. (2016). Energy justice: A conceptual review. Energy Research & Social Science, 11, 174-182. Jenner, S., Chan, G., Frankenberger, R., & Gabel, M. (2013). What drives states to support renewable energy? Energy Journal, 33(2), 1-12. Johnson, M.W. (2010). Seizing the white space: Business model innovation for growth and renewal. Boston, Mass.: Harvard Business Press. Joskow, P. L. (1997). Restructuring, competition and regulatory reform in the U.S. electricity sector. Journal of Economic Perspectives, 11(3), 119-138. Joskow, P. L. (2012). Creating a smarter U.S. electricity grid. Journal of Economic Perspectives, 26(1), 29-48. Kai Zou, Agalgaonkar, A. P., Muttaqi, K. M., & Perera, S. (2014). An analytical approach for reliability evaluation of distribution systems containing dispatchable and non-dispatchable renewable DG units IEEE Transactions on Smart Grid, 5(6), 2657-2665. Karakaya, E., Nuur, C., & Hidalgo, A. (2016). Business model challenge: Lessons from a local solar company. Renewable Energy, 85, 1026-1035. Kauffman, R., & Zibelman, A. (2015). New York’s reforming the energy vision. Power. Kiesling, L. L. (2011). Regulation's Effect on Experimentation in Retail Electricity Markets. Department of Economics at Northwestern University: SSRN.

359

Kiesling, L.L. (2009). Deregulation, innovation and market liberalization: Electricity regulation in a continually evolving environment. New York: Routledge. Kim, K., Heo, E., & Kim, Y. (2017). Dynamic policy impacts on a technological- change system of renewable energy: An empirical analysis; Environmental and Resource Economics, 66(2), 205-236. Kind, P. (2013). Disruptive challenges: Financial implications and strategic responses to a changing retail electric business. Washington, D.C.: Edison Electric Institute. Kindström, D. (2010). Towards a service-based business model - key aspects for future competitive advantage. European Management Journal, 28(6), 479-490. Kindström, D., & Ottosson, M. (2016). Local and regional energy companies offering energy services: Key activities and implications for the business model. Applied Energy, 171, 491-500. Kost, C., Flath, C. M., & Moest, D. (2013). Concentrating solar power plant investment and operation decisions under different price and support mechanisms. Energy Policy, 61, 238-248. Kotter, J.P. (2014). Accelerate. Building strategic agility for a faster-moving world. Boston, MA: Harvard Business Review Press. Retrieved from https://www.epo.org/index.html Lazar, J. (2011). Electricity Regulation in the U.S: A guide. United States: Regulatory Assistance Project. Lazard. (2016). Levelized Cost of Energy Analysis 10.0. New York, USA: NewsRX LLC. LCV. (2017). National environmental scorecard. Retrieved from http://scorecard.lcv.org Lee, A., Zinaman, O., & Logan, J. (2012). Opportunities for Synergy Between Natural Gas and Renewable Energy in the Electric Power and Transportation Sectors. Golden, CO: Joint Institute for Strategic Energy Analysis (JISEA), National Renewable Energy Laboratory. Lee, A., Zinaman, O., Logan, J., Bazilian, M., Arent, D., & Newmark, R. L. (2012). Interactions, complementarities and tensions at the nexus of natural gas and renewable energy. The Electricity Journal, 25(10), 38-48. Lee, C. W., & Zhong, J. (2015). Financing and risk management of renewable energy projects with a hybrid bond. Renewable Energy, 75, 779-787.

360

Lehr, R. L. (2013). New utility business models: Utility and regulatory models for the modern era. The Electricity Journal, 26(8), 35-53. Lesage, J.P., & Pace, R. K. (2009). Introduction to spatial econometrics. Boca Raton [u.a.]: CRC Press. Levi, M.A. (2013). Power surge: Energy, opportunity, and the battle for America’s future. Oxford University Press. Li, S., Chang, T., & Chang, S. (2017). The policy effectiveness of economic instruments for the photovoltaic and wind power development in the European union. Renewable Energy, 101, 660-666. Lightner, E. M., & Widergren, S. E. (2010). An orderly transition to a transformed electricity system. IEEE Transactions on Smart Grid, 1(1), 3-10. Logan, J., Lopez, A., Mai, T., Davidson, C., Bazilian, M., & Arent, D. (2013). Natural gas scenarios in the U.S. power sector. Energy Economics Energy Economics, 40(2), 183-195. Loock, M. (2012). Going beyond best technology and lowest price: On renewable energy investors’ preference for service-driven business models. Energy Policy, 40(1), 21. Lovins, A. B. (1976). Energy strategy: The road not taken? Foreign Affairs, 55(1), 65- 96. Lovins, A.B., & Lovins, L. H. (1982). Brittle power: Energy strategy for national security. Andover, Mass.: Brick House Pub. Co. MAC Solar Index. (2017). MAC global solar energy stock index. Retrieved from http://www.macsolarindex.com MacDonnell, L. J. (1989). Federal interests in western water resources: Conflict and accommodation. (new challenges to western water law). Natural Resources Journal, 29(2), 390-411. Magretta, J. (2002). Why business models matter. Harvard Business Review, 80(5), 86-92. Mai, T., Cole, W., Lantz, E., Marcy, C., & Sigrin, B. (2016). Impacts of Federal Tax Credit Extensions on Renewable Deployment and Power Sector Emissions. (United States: USDOE Office of Energy Efficiency and Renewable Energy (EERE), Policy & Analysis; USDOE Office of Energy Efficiency and Renewable Energy (EERE), Wind and Water Technologies Office (EE-4W); USDOE Office of Energy Efficiency and Renewable Energy (EERE), Solar

361

Energy Technologies Office (EE-4S). Retrieved from http://www.nrel.gov/docs/fy16osti/65571.pdf. Mai, T., Mulcahy, D., Hand, M., Maureen, B., & Samuel, F. (2014). Envisioning a renewable electricity future for the united states. Energy, 65, 374-386. Makholm, J. D. (2016). The REVolution yields to a more familiar path: New York’s reforming the energy vision (REV). The Electricity Journal, 29(9), 48-55. Manfren, M., Caputo, P., & Costa, G. (2011). Paradigm shift in urban energy systems through distributed generation: Methods and models. Applied Energy, 88(4), 1032-1048. Markides, C. C., & Charitou, C. D. (2003). Responses to disruptive strategic innovation. MIT Sloan Management Review, 44(2), 55. Retrieved from http://search.proquest.com/docview/224963412 Martini, P. D., & Kristov, L. (2015). Distribution Systems in a High Distributed Energy Resources Future; Berkeley, Calif.: Lawrence Berkeley National Laboratory, Environmental Energy Technologies Division, United States Department of Energy (DOE). Maryland Public Service Commission. (2016). Maryland PSC launches proceeding on transforming electric distribution systems in Maryland. Retrieved from http://www.psc.state.md.us/wp-content/uploads/Electric-Distribution- Systems-Review_PC44.pdf Masiello, R., & Aguero, J. R. (2016). Sharing the ride of power: Understanding transactive energy in the ecosystem of energy economics. IEEE Power and Energy Magazine, 14(3), 70-78. McDermott, K. (2012). Cost of service regulation in the investor-owned electric utility industry - a history of adaptation. Washington D.C.: Edison Electric Institute. McGrath, R. G. (2010). Business models: A discovery driven approach. Long Range Planning, 43(2-3), 247-261. McLean, B., & Elkind, P. (2003). The smartest guys in the room: The amazing rise and scandalous fall of Enron. New York: Portfolio. Mendelsohn, M., & Feldman, D. (2013). Financing U.S. renewable energy projects through public capital vehicles: Qualitative and quantitative benefits. National Renewable Energy Laboratory (U.S.). Mendelsohn, M., Lowder, T., & Canavan, B. (2012). Utility-scale concentrating solar power and photovoltaics projects: A technology and market overview. Golden, Colorado: National Renewable Energy Laboratory (NREL).

362

Metz, A., Fischer, M., Xing, G., Yong, L., & Julsrud, S. (2014). International Technology Roadmap for Photovoltaic: 2013 Results. Germany: PVGroup. Michaels, R. J. (2008). A national renewable portfolio standard: Politically correct, economically suspect. The Electricity Journal, 21(3), 9-28. Mikhalkina, T., & Cabantous, L. (2015). Business model innovation: How iconic business models emerge. In C. Baden Fuller, & V. Mangematin (Eds.), Business models and modelling (1st ed., pp. 59-95). UK: Emerald Group Publishing Limited. Miller, C. A., Richter, J., & O’Leary, J. (2015). Socio-energy systems design: A policy framework for energy transitions. Energy Research & Social Science, 6(4), 29- 40. Ming, H., & Xie, L. (2016). Revenue adequacy of wholesale electricity markets with demand response providers; IEEE Publications Database, 1-5. Morvaj, B., Evins, R., & Carmeliet, J. (2016). Optimization framework for distributed energy systems with integrated electrical grid constraints. Applied Energy, 171, 296-313. Murtazashvili, I. (2015). Origins and consequences of state-level variation in shale regulation: The cases of Pennsylvania and New York. In W. E. Hefley, & Y. Wang (Eds.), Economics of unconventional shale gas development: Case studies and impacts. springer. (). Cham: Springer International Publishing. Nadel, S., & Young, R. (2014). Why is electricity use no longer growing? Washington, D.C. American Council for an Energy-Efficient Economy (ACEEE). Navigant. (2016). Navigating the Energy Transformation. United: Navigant Consulting, Inc. Neme, C., & Cowart, R. (2014). Energy Efficiency Participation in Electricity Capacity Markets – The U.S. Experience. Montpelier, VT: Regulatory Assistance Project (RAP). Nemet, G. F. (2006). Beyond the learning curve: Factors influencing cost reductions in photovoltaics. Energy Policy, 34(17), 3218-3232. Newcomb, J., Lacy, V., & Hansen, L. (2013). New Business Models for the Distribution Edge. e-Lab discussion paper. Rocky Mountain Institute, Electricity Innovation Lab.

363

Newman, R., Chang, V., Walters, R. J., & Wills, G. B. (2016). Model and experimental development for business data science. International Journal of Information Management, 36(4), 607-617. Nimmons, J., & Taylor, M. (2008). Utility solar business models. Emerging utility strategies & innovation. Washington, D.C.: Solar Electric Power Association (SEPA). North Carolina Clean Energy Technology Center. (2017). The 50 States of Solar; Q4 2016; Annual Review Executive Summary. United States: North Carolina Clean Energy Technology Center. Nowak, S., Baatz, B., Gilleo, M., Kushler, M., Molina, M., & York, D. (2015). Beyond carrots for utilities: A national review of performance incentives for energy efficiency.

Nyangon, J. (2015a). Obama’s budget proposals for clean energy and climate investments. Retrieved from http://freefutures.org/obamas-budget-proposals-for-clean-energy-and-climate- investments/ Nyangon, J. (2015b). The U.S. shale gas revolution and its implications for international energy policy. Green Monitor, 3(1), 184-190. Retrieved from https://www.researchgate.net/publication/311886479_The_US_Shale_Gas_ Revolution_and_its_Implications_for_International_Energy_Policy Nyangon, J. (2017). Carbon finance: How carbon and stock markets are affected by energy prices and emissions regulations (book review). The Energy Journal, 38(4) Nyangon, J., Alabbas, N., & Agbemabiese, L. (2017). Entangled systems at the energy-water-food nexus. In P. Rao, & Y. Patil (Eds.), Reconsidering the impact of climate change on global water supply, use, and management (pp. 144-165). United States: IGI Global. Nyangon, J., Byrne, J., & Taminiau, J. (2017). An assessment of price convergence between natural gas and solar photovoltaic in the U.S. electricity market. Wiley Interdisciplinary Reviews: Energy and Environment, 6(3), 1-20. NYISO. (2017). Power Trends 2017: New York’s Evolving Electric Grid. New York: New York Independent System Operator (NYISO). NYPSC. (2014). Reforming the energy vision: Staff report and proposal. Albany, New York: NYPSC. NYPSC. (2015a). Proceeding on Motion of the Commission in Regard to Reforming the Energy Vision (CASE 14-M-0101). Staff White Paper on Ratemaking and

364

Utility Business Models. Albany, NY: New York Public Service Commission (NYPSC). NYPSC. (2015b). Proceeding on Motion of the Commission in Regard to Reforming the Energy Vision, Order Adopting Regulatory Policy Framework and Implementation Plan, CASE 14-M-0101. Albany, NY: New York Public Service Commission (NYPSC). NYPSC. (2015c). Staff white paper on benefit-cost analysis in the Reforming Energy Vision Proceeding. Albany, NY: New York State Department of Public Service (NYPSC). NYPSC. (2016a). Order Authorizing Framework for Community Choice Aggregation Opt- Out Program (14-M-0224). Albany, NY: NY Public Service Commission (NYPSC). NYPSC. (2016b). Order Authorizing the Clean Energy Fund Framework (14-M-0094, 10-M-0457, 07-M-0548, 03-E-0188, 13-M-0412). Albany, NY: NY Public Service Commission (NYPSC). NYPSC. (2016c). Order Authorizing Utility-Administered Energy Efficiency Portfolio Budgets and Targets for 2016-2018 (15-M-0252). Albany, NY: NY Public Service Commission (NYPSC). NYPSC. (2016d). Order Establishing the Benefit Cost Analysis Framework (14-M- 0101). Albany, NY: NY Public Service Commission (NYPSC). NYPSC. (2016e). Order Resetting Retail Energy Markets and Establishing Further Process (15-M-0127, 12-M-0476, 98-M-1343). Albany, NY: NY Public Service Commission (NYPSC). NYPSC. (2017a). Distributed generation information. Retrieved from http://www3.dps.ny.gov/W/PSCWeb.nsf/All/DCF68EFCA391AD60852576 87006F396B?OpenDocument NYPSC. (2017b). Order on Net Energy Metering Transition, Phase One of Value of Distributed Energy Resources, and Related Matters (15-E-0751, 15-E- 0082). Albany, NY: NY Public Service Commission (NYPSC). O’Loughlin, M., Graves, F., Levine, S., Thapa, A., & Celebi, M. (2012). The Emerging Need for Greater Gas-Electric Industry Coordination. Washington, DC.: The Brattle Group. Ofgem. (2015). Non-Traditional Business Models: Supporting Transformative Change in the Energy Market. London, United Kingdom: Office of Gas and Electricity Markets.

365

Okkonen, L., & Suhonen, N. (2010). Business models of heat entrepreneurship in Finland. Energy Policy, 38(7), 3443-3452. Olkkonen, V., & Syri, S. (2016). Spatial and temporal variations of marginal electricity generation: The case of the Finnish, Nordic, and European energy systems up to 2030. Journal of Cleaner Production, 126, 515-525. doi://dx.doi.org.udel.idm.oclc.org/10.1016/j.jclepro.2016.03.112 Onovwiona, H. I., & Ugursal, V. I. (2006). Residential cogeneration systems: Review of the current technology. Renewable and Sustainable Energy Reviews, 10(5), 389-431. Organisation for Economic Co-operation and Development. (2000). Experience curves for energy technology policy. International Energy Agency. Paris: OECD/IEA. Osterwalder, A., & Pigneur, Y. (2012). Business model generation: A handbook for visionaries, game changers, and challengers. Journal of Product Innovation Management, 29(6) Osterwalder, A., Pigneur, Y., Clark, T., & Smith, A. (2010). Business model generation: A handbook for visionaries, game changers, and challengers. Hoboken, NJ: Wiley. Pacheco, D. F., York, J. G., & Hargrave, T. J. (2014). The coevolution of industries, social movements, and institutions: Wind power in the United States. Organization Science, 25(6), 1609-1632. Papaefthymiou, G., & Dragoon, K. (2016). Towards 100% renewable energy systems: Uncapping power system flexibility. Energy Policy, 92, 69-82. doi://dx.doi.org.udel.idm.oclc.org/10.1016/j.enpol.2016.01.025

Pätäri, S., & Sinkkonen, K. (2014). Energy service companies and energy performance contracting: Is there a need to renew the business model? insights from a Delphi study. Journal of Cleaner Production, 66, 264-271. Pérez-Arriaga, I. (2011). Managing Large Scale Penetration of Intermittent Renewables. Cambridge, MA: MIT Energy Initiative Symposium. Massachusetts Institute of Technology. Pérez-Arriaga, I. J., Rubio-Odériz, F., Gutiérrez, J. F. P., Ogando, J. A., & Marín, J. (1995). Marginal pricing of transmission services: An analysis of cost recovery. IEEE Transactions on Power Systems, 10, 65-72. Pérez-Arriaga, I., & Knittel, C. (2016). Utility of the Future: An MIT Energy Initiative Response to An Industry in Transition. Boston, MA: MIT Energy Initiative.

366

Pérez-Arriaga, I.J. (2013). Regulation of the power sector, power systems. London, United Kingdom: Springer. PJM. (2013). 2012 Economic Demand Response Performance Report: Analysis of Economic DR participation in the PJM wholesale energy market after the implementation of Order 745. Norristown, PA.: PJM Interconnection. PJM. (2017). PJM hourly demand bid data. Retrieved from http://pjm.com/~/media/markets-ops/energy/day-ahead/hourly-day-ahead- bid-data-2016.ashx Pless, J., Arent, D. J., Logan, J., Cochran, J., & Zinaman, O. (2016). Quantifying the value of investing in distributed natural gas and renewable electricity systems as complements: Applications of discounted cash flow and real options analysis with stochastic inputs. Energy Policy, 97, 378-390. Polson, J. (2015). Why growth-hungry U.S. utilities won't touch MLPs for pipes; Retrieved from https://www.bloomberg.com/news/articles/2015-10-28/why- growth-hungry-utilities-won-t-touch-mlps-for-their-gas-pipes Popp, D., Hascic, I., & Medhi, N. (2011). Technology and the diffusion of renewable energy. Energy Economics, 33(4), 648-662. Popp, D., Newell, R. G., & Jaffe, A. B. (2010). Energy, the environment, and technological change. In B. Hall H., & N. Rosenberg (Eds.), Handbook of the economics of innovation (pp. 873-937). Burlington: Academic Press. Provance, M., Donnelly, R. G., & Carayannis, E. G. (2011). Institutional influences on business model choice by new ventures in the microgenerated energy industry. Energy Policy, 39(9), 5630-5637. Qolipour, M., Mostafaeipour, A., Shamshirband, S., Alavi, O., Goudarzi, H., & Petković, D. (2016). Evaluation of wind power generation potential using a three- hybrid approach for households in Ardebil province, Iran. Energy Conversion and Management, 118, 295-305. Ragwitz, M., & Steinhilber, S. (2014). Effectiveness and efficiency of support schemes for electricity from renewable energy sources. Wiley Interdisciplinary Reviews: Energy and Environment, 3(2), 213-229. RAP. (2013). Recommendations for Power Sector Policy in China: Practical Solutions for Energy, Climate, and Air Quality. Beijing, China: Regulatory Assistance Project (RAP). Renou-Maissant, P. (2012). Toward the integration of European natural gas markets: A time-varying approach. Energy Policy, 51(3), 779-790.

367

Resnick Institute. (2013). Grid 2020: Towards a policy of renewable and distributed energy resources. Retrieved from http://scholar.aci.info/view/146877f971319ff0146/14e934a82380012000b Ribeiro, D., Hewitt, V., Mackres, E., Cluett, R., Ross, L. M., Vaidyanathan, S., & Zerbonne, S. (2015). The 2015 City Energy Efficiency Scorecard. Washington, American Council for an Energy Efficient Economy (ACEEE). Richter, M. (2012). Utilities’ business models for renewable energy: A review. Renewable and Sustainable Energy Reviews, 16(5), 2483-2493. Richter, M. (2013). Business model innovation for sustainable energy: German utilities and renewable energy. Energy Policy, 62, 1226-1237. Ritzenhofen, I., Birge, J. R., & Spinler, S. (2016). The structural impact of renewable portfolio standards and feed-in tariffs on electricity markets. European Journal of Operational Research, 255(1), 224-242. Rogers, E. A., Carley, E., Deo, S., & Grossberg, F. (2015). How information and communications technologies will change the evaluation, measurement, and verification of energy efficiency programs. Wash: American Council for an Energy-Efficient Economy. Roodman, D. (2009). A note on the theme of too many instruments. Oxford Bulletin of Economics and Statistics, 71(1), 135-158. Ross, K., & Damassa, T. (2015). Assessing the Post-2020 Clean Energy Landscape. Was: World Resources Institute. S&P Dow Jones Indices. (2017). S&P GSCI natural gas index. Retrieved from http://us.spindices.com/indices/commodities/sp-gsci-natural-gas Sahasrabhojanee, S. A., Sathe, J. S., Bujanovic, T., Ghosh, P. K., & Maxwell, W. H.A comparison of NYS utilities' approaches to integrate distributed energy resources and the penetration over time. 23-28. Sarasa-Maestro, C. J., Dufo-López, R., & Bernal-Agustín, J. L. (2013). Photovoltaic remuneration policies in the european union. Energy Policy, 55, 317-328. Satchwell, A., & Cappers, P. (2015). A framework for organizing electric utility regulatory and business models. The Electricity Journal, 28(8), 119-129. Sawin, J. L., Seyboth, K., & Sverrisson, F. (2016). Renewables 2016 Global Status Report.REN21 Secretariat. Schallenberg-Rodriguez, J. (2017). Renewable electricity support systems: Are feed-in systems taking the lead? Renewable and Sustainable Energy Reviews, 76, 1422- 1439.

368

Schatzki, T., & Hibbard, P. (2013). Assessment of the Impact of ISO-NE’s Proposed Forward Capacity Market Performance Incentives. Boston, MA: Analysis Group. Schleicher-Tappeser, R. (2012). How renewables will change electricity markets in the next five years. Energy Policy, 48(1), 64-75. Schoettl, J., & Lehmann-Ortega, L. (Eds.). (2011). Photovoltaic business models: Threat or opportunity for utilities? (1st ed.). Cheltenham, UK; Northampton, MA: Edward Elgar Publishing. Schwab, K. (2016). The fourth industrial revolution (First edition ed.). New York: Crown Business. Retrieved from http://www.econis.eu/PPNSET?PPN=848495179 Schweppe, F. C. (1978). Power systems 2000: Hierarchical control strategies: Multilevel controls and home minis will enable utilities to buy and sell power at real time rates determined by supply and demand. IEEE Spectrum, 15(7), 42-47. SDSIP. (2016). Supplemental Distributed System Implementation (SDSIP) Plan & Case 16-M-0411 In the Matter of Distributed System Implementation Plans. New York: Joint Utilities of New York. SEIA/GTM Research. (2017). U.S. solar market insight report 2016 Q4. Retrieved from http://www.seia.org/research-resources/solar-market-insight-report-2016-q4 Selviaridis, K., & Wynstra, F. (2015). Performance-based contracting: A literature review and future research directions. International Journal of Production Research, 53(12), 3505-3540. doi:10.1080/00207543.2014.978031 Sepponen, M., & Heimonen, I. (2016). Business concepts for districts energy hub systems with maximised share of renewable energy. Energy and Buildings, 124, 273-280. Shafer, S. M., Smith, H. J., & Linder, J. C. (2005). The power of business models. Business Horizons, 48(3), 199-207. Shavel, I., Fox-Penner, P., Weiss, J., Hledik, R., Ruiz, P., Yang, Y., . . . Zahniser- Word, J. (2014). Exploring Natural Gas and Renewables in ERCOT, Part III: The Role of Demand Response, Energy Efficiency, and Combined Heat and Power. Cambridge, MA: The Brattle Group. Shavel, I., Weiss, J., Fox-Penner, P., Ruiz, P., Yang, Y., Carroll, R., & Zahnise-Word, J. (2013). Exploring Natural Gas and Renewables in ERCOT Part II: Future Generation Scenarios for Texas. W: The Brattle Group and Texas Clean Energy Coalition.

369

Shen, B., Ghatikar, G., Lei, Z., Li, J., Wikler, G., & Martin, P. (2014). The role of regulatory reforms, market changes, and technology development to make demand response a viable resource in meeting energy challenges. Applied Energy, 130, 814-823. Shrimali, G., & Kniefel, J. (2011). Are government policies effective in promoting deployment of renewable electricity resources? Energy Policy, 39(9), 4726-4741. Shrimali, G., Jenner, S., Groba, F., Chan, G., & Indvik, J. (2012). Have state renewable portfolio standards really worked? synthesizing past policy assessments to build an integrated econometric analysis of RPS effectiveness in the US. DIW Discussion Papers, 1258 Retrieved from https://www.econstor.eu/handle/10419/68450 Simkins, B.J., & Simkins, R. E. (2013). Energy finance analysis and valuation, risk management, and the future of energy. Hoboken, N.J.: Wiley. Sinn, H. (2016). Buffering volatility: A study on the limits of Germany’s energy revolution. CESifo Working Paper, 5950 Retrieved from https://www.econstor.eu/handle/10419/144985 Sioshansi, F.P. (2010). Generating electricity in a carbon-constrained world. San Diego: Elsevier Academic Press. Sioshansi, F.P. (2013). Energy efficiency towards the end of demand growth. Amsterdam: Elsevier Science/AP. Sioshansi, F.P. (2016). Future of utilities - utilities of the future: How technological innovations in distributed energy resources will reshape the electric power sector. Amsterdam: Heidelberg: Elsevier, Academic Press. Sioshansi, F.P., & Pfaffenberger, W. (2006). Electricity market reform: An international perspective. Amsterdam; Boston: Elsevier. Small, F., & Frantzis, L. (2010). The 21st Century electric utility. Positioning for a low-carbon future. Boston, MA: Ceres Report. Solar Energy Industries Association. (2016). SEIA's top 10 solar states. Retrieved from http://www.seia.org/research-resources/top-10-solar-states Sosna, M., Trevinyo-Rodríguez, R. N., & Velamuri, S. R. (2010). Business model innovation through trial-and-error learning: The naturhouse case. Long Range Planning, 43(2–3), 383-407. Southern California Edison. (2016). The Emerging Clean Energy Economy: Customer- Driven. Modernized. Reliable. United States: Southern California Edison.

370

Sovacool, B. K. (2017). The history and politics of energy transitions: Comparing contested views and finding common ground. The political economy of clean energy transitions (pp. 16-35). GB: Oxford University Press. Srinivasan, S., Reddy, V. K., & Reddy, V. K. (2016). Towards a better understanding of renewable energy YieldCos. Renewable and Sustainable Energy Reviews, 65, 154-163. Ssozi, J., & Amlani, S. (2015). The effectiveness of health expenditure on the proximate and ultimate goals of healthcare in sub-Saharan Africa. World Development, 76, 165-179. Steinberger, J. K., van Niel, J., & Bourg, D. (2009). Profiting from negawatts: Reducing absolute consumption and emissions through a performance-based energy economy. Energy Policy, 37(1), 361-370. Sueyoshi, T., & Goto, M. (2013). A comparative study among fossil fuel power plants in PJM and California ISO by DEA environmental assessment; Energy Economics, 40, 130-145. Sweeney, J.L. (2002). The California electricity crisis. Stanford, Calif.: Hoover Institution Press: Stanford Institute for Economic Policy Research. Tabors, R. D., He, H., & Birk, M. (2016). The impact of distributed energy resources on incumbent utilities: A case study of long island, New York; Proceedings of the Annual Hawaii International Conference on System Sciences, 2016-March, 2576- 2583. Taminiau, J., Nyangon, J., Lewis, A. S., & Byrne, J. (2017). Sustainable business model innovation: Using polycentric and creative climate change governance. In Z. Fields (Ed.), Collective creativity for responsible and sustainable business practice (pp. 140-159). United States: IGI Global. Taminiau, J., Wang, Y., & Byrne, J. (2014). Drivers for change. In J. Byrne, & Y. Wang (Eds.), Green energy economies: The search for clean and renewable energy (1st ed.). Piscataway: Transaction Publishers. Tan, W., Hassan, M. Y., Majid, M. S., & Abdul Rahman, H. (2013). Optimal distributed renewable generation planning: A review of different approaches. Renewable and Sustainable Energy Reviews, 18, 626-645. Tanwar, S. S., & Khatod, D. K. (2017). Techno-economic and environmental approach for optimal placement and sizing of renewable DGs in distribution system. Energy, 127, 52-67. Tao, H. Y. S., Bahabry, A., & Cloutier, R. (2015). Customer centricity in the smart grid model; Procedia Computer Science, 44, 115-124.

371

Teece, D. J. (2007). Explicating dynamic capabilities: The nature and microfoundations of (sustainable) enterprise performance. Strategic Management Journal, 28(13), 1319-1350. Teece, D. J. (2010). Business models, business strategy and innovation. Long Range Planning, 43(2), 172-194. Tomain, J. P. (2016). A perspective on clean power and the future of US energy politics and policy. Utilities Policy, 39, 5-12. Trabish, H. K. (2016a). 17% of Hawaiian electric customers now have rooftop solar. Retrieved from http://www.utilitydive.com/news/17-of-hawaiian-electric- customers-now-have-rooftop-solar/413014/ Trabish, H. K. (2016b). How storage can help solve the distributed energy ‘Death spiral’: A new Navigant paper lays out three examples of how storage can facilitate the transition to a more distributed grid. U.S. Department of Energy. (2013). Benefits of demand response in electricity markets and recommendations for achieving them: A report to the United States Congress Pursuant to section 1252 of the Energy Policy Act of 2005. Washington, D.C.: U.S. Department of Energy. U.S. Department of Energy. (2015). Quadrennial Energy Review: Energy Transmission, Storage, and Distribution Infrastructure. Washi: United States Department of Energy (DOE). U.S. Department of Energy. (2017a). Renewables portfolio standard. Retrieved from http://programs.dsireusa.org/system/program/detail/840 U.S. Department of Energy. (2017b). Transforming the Nation’s Electricity System: The Second Installment of the QER, Quadrennial Energy Review. Washington, D.C.: U.S. Department of Energy (DOE). U.S. Department of Labor. (2017). Employment, hours, and earnings from the current employment statistics survey (national), all employees, thousands, seasonally adjusted. Retrieved from https://www.bls.gov/oes/current/oessrci.htm U.S. Energy Information Administration. (2013). Natural gas-fired combustion turbines are generally used to meet peak electricity load. U.S. Energy Information Administration (EIA) U.S. Energy Information Administration. (2014). What is U.S. electricity generation by energy source? Retrieved from http://www.eia.gov/tools/faqs/faq.cfm?id=427&t=3

372

U.S. Energy Information Administration. (2016a). Capacity factors for utility scale generators primarily using fossil fuels (table 6.7a) and capacity factors for utility scale generators not primarily using fossil fuels (table 6.7b). Retrieved from http://www.eia.gov/electricity/monthly/ U.S. Energy Information Administration. (2016b). U.S. Energy Information Administration, Monthly Energy Review (April 2016), Tables 1.3, 2.1-2.6. Unit: U.S. Energy Information Administration. U.S. Energy Information Administration. (2017a). Electric power industry generation by primary energy source, 1990 through 2014 united states (table 5). Retrieved from https://www.eia.gov/electricity/state/ U.S. Energy Information Administration. (2017b). Electric Power Monthly with Data for January 2017. Table 6.7.A. Capacity Factors for Utility Scale Generators Primarily Using Fossil Fuels, January 2013-January 2017. United States: U.S. Energy Information Administration (EIA). U.S. Energy Information Administration. (2017c). Electric power sales, revenue, and energy efficiency (form EIA-861). Retrieved from https://www.eia.gov/electricity/data/eia861/ U.S. Energy Information Administration. (2017d). Electricity data browser. net generation for all sectors. Retrieved from https://www.eia.gov/electricity/data/browser/ U.S. Energy Information Administration. (2017e). Energy-Related Carbon Dioxide Emissions at the State Level, 2000-2014. Washington, D.C.: U.S. Energy Information Administration. U.S. Energy Information Administration. (2017f). Existing nameplate and net summer capacity by energy source, producer type and state (EIA-860). Retrieved from https://www.eia.gov/electricity/data/state/ U.S. Energy Information Administration. (2017g). How many smart meters are installed in the united states, and who has them? Frequently Asked Questions. Retrieved from www.eia.gov/todayinenergy/detail.cfm?id=25492 U.S. Energy Information Administration. (2017h). Net generation by state by type of producer by energy source (EIA-906, EIA-920, and EIA-923). Retrieved from https://www.eia.gov/electricity/data/state/ U.S. Energy Information Administration. (2017i). Renewables and alternative fuels. Retrieved from https://www.eia.gov/renewable/data.php

373

U.S. Energy Information Administration. (2017j). Retail sales of electricity by state by sector by provider (EIA-861). Retrieved from https://www.eia.gov/electricity/data/state/ U.S. Energy Information Administration. (2017k). U.S. Energy Information Administration/Monthly Energy Review January 2017. Washington, D.C.: U.S. Energy Information Administration (EIA). U.S. Energy Information Administration. (2017l). Wind adds the most electric generation capacity in 2015, followed by natural gas and solar. Retrieved from www.eia.gov/tools/faqs/faq.cfm?id=108&t=3 Utility Dive. (2016). Utility Dive report: The State of the Electric Utility 2016. United States: Utility Dive. Verdolini, E., Vona, F., & Popp, D. (2016). Bridging the gap: Do fast reacting fossil technologies facilitate renewable energy diffusion? Nota Di Lavoro, Fondazione Eni Enrico Mattei, 51.2016 Retrieved from https://www.econstor.eu/handle/10419/145377 Walton, J. (2013). An empirical exploration of the effects of natural gas prices on solar energy growth Wang, Z., & Krupnick, A. (2015). A retrospective review of shale gas development in the united states: What led to the boom? Economics of Energy & Environmental Policy, 4(1). Ward, D. M. (2013). The effect of weather on grid systems and the reliability of electricity supply. Climatic Change, 121(1), 103-113. Weiss, J., Bishop, H., Fox-Penner, P., & Shavel, I. (2013). Partnering Natural Gas and Renewable Energy in ERCOT. Washington, D.C.: The Brattle Group. Williamson, O. E. (1979). Transaction-cost economics: The governance of contractual relations. The Journal of Law and Economics the Journal of Law and Economics, 22(2), 233-261. Winkler, J., Gaio, A., Pfluger, B., & Ragwitz, M. (2016). Impact of renewables on electricity markets? Do support schemes matter? Energy Policy, 93, 157-167. doi://dx.doi.org.udel.idm.oclc.org/10.1016/j.enpol.2016.02.049 Wiser, R., Barbose, G., & Holt, E. (2011). Supporting solar power in renewables portfolio standards: Experience from the united states. Energy Policy, 39(7), 3894-3905.

374

Wiser, R., Namovicz, C., Gielecki, M., & Smith, R. (2007). The experience with renewable portfolio standards in the united states. The Electricity Journal, 20(4), 8-20. Woo, C. K., Ho, T., Zarnikau, J., Olson, A., Jones, R., Chait, M., Horowitz., I., & Wang, J. (2014). Electricity-market price and nuclear power plant shutdown: Evidence from California. Energy Policy, 73, 234-244. Woo, C. K., Moore, J., Schneiderman, B., Ho, T., Olson, A., Alagappan, L., . . . Zarnikau, J. (2016). Merit-order effects of renewable energy and price divergence in Californian’s day-ahead and real-time electricity markets. Energy Policy, 92, 299-312. Woo, C. K., Zarnikau, J., Cutter, E., Ho, S. T., & Leung, H. Y. (2015). Virtual bidding, wind generation and California’s day-ahead electricity forward premium. The Electricity Journal, 28(1), 29-48. Xu, J., Li, L., & Zheng, B. (2016). Wind energy generation technological paradigm diffusion. Renewable and Sustainable Energy Reviews, 59, 436-449. Xu, J., Luo, N., Li, M., & Xie, H. (2017). A novel paradigm-oriented approach towards NG-RE hybrid power generation. Energy Conversion and Management, 145, 220-232. Yao, M. M., Wu, D., Wu, X. B., Wu, Z. H., & Chen, S. C. (2012). Business model innovation of modern service company: A value network perspective. Management of Technology (ISMOT), 2012 International Symposium On, 76-80. Yin, H., & Powers, N. (2010). Do state renewable portfolio standards promote in-state renewable generation? Energy Policy, 38(2), 1140-1149. York, D., Kushler, M., Hayes, S., Sienkowski, S., Bell, C., & Kihm, S. (2013). Making the Business Case for Energy Efficiency: Utility Performance with Supportive Regulation. Washington, D.C.: American Council for an Energy-Efficient Economy. Zarakas, W. P., Sergici, S., Bishop, H., Zahniser-Word, J., & Fox-Penner, P. (2014). Utility investments in resiliency: Balancing benefits with cost in an uncertain environment. The Electricity Journal, 27(5), 31-41. Zhang, Y., & Giannakis, G. B. (2016). Distributed stochastic market clearing with high-penetration wind power. IEEE Trans. Power Syst. IEEE Transactions on Power Systems, 31(2), 895-906. Zhao, X., Pan, W., & Lu, W. (2016). Business model innovation for delivering zero carbon buildings. Sustainable Cities and Society.

375

Zibelman, A. (2016). REVing up the energy vision in New York: Seizing the opportunity to create a cleaner, more resilient, and affordable energy system. IEEE Power & Energy Magazine, 14(3), 18-24. Zinaman, O., Miller, M., Adil, A., Arent, D., Cochran, J., Vora, R., . . . Pillai, R. K. (2015). Power Systems of the Future: A 21st Century Power Partnership Thought Leadership Report (Technical Report NREL/TP-6A20-62611). Golden, CO: National Renewable Energy Laboratory. Zolnowski, A., Weiß, C., & Böhmann, T. (2014). Representing service business models with the service business model canvas -- the case of a mobile payment service in the retail industry. 2014 47th Hawaii International Conference on System Sciences, 718-727. Zott, C., & Amit, R. (2008). The fit between product market strategy and business model: Implications for firm performance. SMJ Strategic Management Journal, 29(1), 1-26. Zott, C., Amit, R., & Massa, L. (2010). The business model: Theoretical roots, recent developments, and future research. Barcelona, Spain: IESE Business School, University of Navarra. Zott, C., Amit, R., & Massa, L. (2011). The business model: Recent developments and future research. Journal of Management, 37(4).

376

APPENDIX

SUPPLEMENTAY TABLES, FIGURES, AND INFORMATION

377

A. STATE NET METERING CAPACITY

ARIZONA CALIFORNIA NEW YORK

PV All Technologies PV Wind All Technologies PV All Tech State Net Energy Energy Energy Energy Capacity Capacity Capacity Capacity Capacity Capacity Capacity Metering Sold Back Sold Back Sold Back Sold Back (MW) (MW) (MW) (MW) (MW) (MW) (MW) Capacity (MWh) (MWh) (MWh) (MW) 2016M01 771 8182 772 8196 3989 11016 18 4118 11016 546 573 2016M02 781 10994 782 11013 4101 8328 18 4231 8328 566 593 378

2016M03 792 14953 793 14976 4213 16772 17 4345 16772 585 614 2016M04 808 16402 809 16428 4335 28020 17 4468 28020 605 633 2016M05 822 18015 822 18040 4445 26974 17 4579 26974 617 645 2016M06 834 14078 835 14102 4563 25957 17 4706 25957 632 660 2016M07 844 8591 844 8607 4657 27526 17 4802 27526 646 674 2016M08 856 9035 856 9054 4782 16477 17 4931 16477 662 690 2016M09 865 10006 865 10023 4884 16733 17 5035 16733 690 712 2016M10 874 11866 875 11885 5003 12166 17 5158 12166 706 735 2016M11 886 11430 886 11443 5139 9799 18 5297 9799 719 750 2016M12 899 11414 900 11431 5283 7849 18 5447 7849 735 766

B. ELECTRICITY IMPORT RATIOS, BY STATE, 2001-2015

AZ CA GA MA NC NJ NV NY TX UT US Average 2001 -0.44 0.20 0.00 0.27 0.01 0.19 -0.20 0.00 -0.17 -0.54 -0.10 2002 -0.50 0.22 -0.02 0.22 -0.01 0.17 -0.10 0.05 -0.20 -0.57 -0.11 2003 -0.47 0.21 0.00 0.13 -0.05 0.25 -0.10 0.04 -0.18 -0.59 -0.11 2004 -0.56 0.23 0.02 0.15 -0.01 0.28 -0.20 0.05 -0.22 -0.56 -0.12 2005 -0.46 0.21 -0.03 0.17 -0.01 0.26 -0.24 0.02 -0.19 -0.53 -0.11 2006 -0.43 0.18 -0.02 0.18 0.01 0.24 0.08 0.00 -0.17 -0.57 -0.11 379

2007 -0.47 0.20 -0.06 0.18 0.01 0.24 0.08 0.02 -0.18 -0.63 -0.10 2008 -0.57 0.22 -0.01 0.24 0.04 0.21 0.00 0.03 -0.16 -0.65 -0.10 2009 -0.52 0.21 0.02 0.28 0.07 0.18 -0.10 0.05 -0.15 -0.58 -0.10 2010 -0.53 0.21 0.02 0.25 0.06 0.17 -0.04 0.05 -0.15 -0.51 -0.10 2011 -0.44 0.23 0.09 0.32 0.10 0.16 0.06 0.05 -0.16 -0.42 -0.09 2012 -0.48 0.23 0.07 0.36 0.09 0.13 0.00 0.05 -0.18 -0.33 -0.10 2013 -0.50 0.23 0.07 0.40 0.03 0.13 -0.04 0.08 -0.14 -0.40 -0.09 2014 -0.47 0.24 0.07 0.43 0.04 0.08 -0.03 0.07 -0.12 -0.46 -0.09 2015 -0.46 0.25 0.05 0.41 0.04 0.01 -0.08 0.07 -0.15 -0.39 -0.08

C. PANEL GENERALIZED METHOD OF MOMENTS DATA ANALYSIS

C.1 Electric Power Sales, Revenue, Customers, Service Type, and Ownership in NYS in 2015 Service Revenues Utility Name Ownership Sales ($) Customers Type ($) Consolidated Edison Co-NY Inc. Bundled Investor Owned 4,757,588 20,206,464 2,545,762 Long Island Power Authority Bundled State 3,326,348 18,151,255 1,115,541 Consolidated Edison Co-NY Inc. Delivery Investor Owned 3,313,064 36,828,725 851,992 Niagara Mohawk Power Corp. Bundled Investor Owned 1,598,810 13,042,081 1,288,882 Retail Power Direct Energy Business Energy Marketer 1,052,583 14,317,230 14,980 New York Power Authority Energy State 1,008,484 18,578,384 824 New York State Elec & Gas Corp. Bundled Investor Owned 789,556 6,748,513 669,017 Niagara Mohawk Power Corp. Delivery Investor Owned 659,514 21,394,469 358,956 Constellation NewEnergy, Retail Power Inc Energy Marketer 580,389 8,215,486 2,178 Central Hudson Gas & Elec Corp. Bundled Investor Owned 417,835 2,553,909 247,746 Retail Power ENGIE Resources Inc. Energy Marketer 362,597 5,216,435 693 Rochester Gas & Electric Corp. Bundled Investor Owned 335,252 2,605,567 282,915 Retail Power Consolidated Edison Sol Inc Energy Marketer 331,486 3,910,908 47,534 New York State Elec & Gas Corporation Delivery Investor Owned 291,363 8,900,387 215,321 Orange & Rockland Utils Inc Bundled Investor Owned 271,010 1,463,572 138,109

380

Retail Power Direct Energy Services Energy Marketer 237,624 2,333,353 198,911 Rochester Gas & Electric Corp. Delivery Investor Owned 222,431 4,503,779 90,897 Constellation Energy Retail Power Services NY, Inc. Energy Marketer 190,923 3,076,884 13,484 Retail Power Ambit Energy Holdings, LLC Energy Marketer 186,648 1,890,178 183,180 Orange & Rockland Utils Inc Delivery Investor Owned 179,719 2,564,756 90,312 Retail Power Hudson Energy Services Energy Marketer 131,699 1,470,254 21,979 Long Island Power Authority Delivery State 123,240 1,774,183 3,563 Retail Power Liberty Power Corp. Energy Marketer 112,234 1,331,204 19,195 Retail Power AP Holdings LLC Energy Marketer 107,158 401,659 5,029 Central Hudson Gas & Elec Corp. Delivery Investor Owned 100,436 2,523,407 54,687 Retail Power Just Energy New York Corp. Energy Marketer 92,538 834,497 103,457 Retail Power Plymouth Rock Energy, LLC Energy Marketer 80,996 831,566 18,097 Retail Power IDT Energy, Inc. Energy Marketer 72,263 529,185 105,631 Noble Americas Energy Retail Power Solutions LLC Energy Marketer 70,902 1,105,559 175 Retail Power Viridian Energy NY LLC Energy Marketer 66,024 635,986 46,690 Major Energy Electric Retail Power Services Energy Marketer 64,558 574,842 41,988 Retail Power American Power & Gas Energy Marketer 61,772 471,327 30,844

381

Green Mountain Energy Retail Power Company Energy Marketer 57,780 418,941 60,908 Energy Services Providers, Retail Power Inc Energy Marketer 53,339 561,758 42,204 TransCanada Power Retail Power Marketing, Ltd. Energy Marketer 51,604 581,068 89 Family Energy, Inc. New Retail Power York Energy Marketer 49,629 491,374 58,614 Retail Power Bluerock Energy, Inc. Energy Marketer 44,579 599,631 10,627 XOOM Energy New York, Retail Power LLC Energy Marketer 44,025 340,075 41,129 Retail Power Energy Plus Holdings LLC Energy Marketer 43,111 276,411 30,986 Retail Power Agway Energy Services, LLC Energy Marketer 39,284 436,722 43,671 Retail Power Champion Energy Services Energy Marketer 38,621 481,736 188 Retail Power Kiwi Energy NY LLC Energy Marketer 37,962 299,379 58,641 Retail Power NOCO Electric Energy Marketer 37,902 679,349 18,093 Retail Power Great Eastern Energy Energy Marketer 33,726 370,565 12,002 Jamestown Board of Public Util Bundled Municipal 30,988 450,988 19,235 Village of Freeport - (NY) Bundled Municipal 29,889 252,818 14,686 Village of Solvay - (NY) Bundled Municipal 29,706 549,852 5,110 Retail Power MPower Energy LLC Energy Marketer 29,472 205,778 49,655 Energy Coop of New York, Retail Power Inc Energy Marketer 27,553 561,645 5,677

382

Retail Power Marathon Power LLC Energy Marketer 24,753 224,732 3,943 North American Power and Retail Power Gas, LLC Energy Marketer 24,483 183,923 18,814 Village of Rockville Centre - (NY) Bundled Municipal 23,178 209,460 10,791 NextEra Energy Services, Retail Power LLC Energy Marketer 22,604 253,475 542 Village of Fairport - (NY) Bundled Municipal 22,434 441,238 16,920 Accent Energy Holdings, Retail Power LLC Energy Marketer 19,878 142,129 19,371 Retail Power U.S. Energy Partners LLC Energy Marketer 19,639 387,781 4,450 Constellation Energy Power Retail Power Choice Inc. Energy Marketer 19,194 186,006 20,383 Retail Power Spark Energy, LP Energy Marketer 18,214 129,063 19,879 City of Plattsburgh - (NY) Bundled Municipal 18,205 459,091 9,975 Retail Power Texas Retail Energy, LLC Energy Marketer 17,837 305,568 1 Retail Power Public Power LLC (CT) Energy Marketer 16,774 185,986 11,876 Retail Power EnergyMark, LLC Energy Marketer 15,328 254,466 3,359 Iron Energy LLC dba Kona Retail Power Energy Energy Marketer 14,513 143,723 2,380 Retail Power HIKO Energy, LLC Energy Marketer 14,051 118,380 16,624 New Wave Energy Retail Power Corporation Energy Marketer 13,515 225,051 3,692 Town of Massena - (NY) Bundled Municipal 12,396 206,249 9,707 Steuben Rural Elec Coop, Inc Bundled Cooperative 10,119 87,683 6,271

383

Retail Power Alpha Gas and Electric, LLC Energy Marketer 9,827 67,336 11,333 Retail Power Robison Energy, LLC Energy Marketer 9,547 84,011 2,603 Lake Placid Village, Inc - (NY) Bundled Municipal 9,432 168,282 4,883 Retail Power Verde Energy USA Energy Marketer 8,785 62,672 9,796 Retail Power Astral Energy LLC Energy Marketer 8,394 72,589 4,069 Village of Arcade - (NY) Bundled Municipal 7,944 161,471 4,186 Retail Power Eligo Energy, LLC Energy Marketer 7,645 81,637 84,976 Columbia Utilities Power, Retail Power LLC Energy Marketer 7,618 35,960 5,736 Retail Power Agera Energy LLC Energy Marketer 7,110 76,633 1,381 SolarCity Corporation Bundled Behind the Meter 6,983 52,891 6,393 City of Salamanca - (NY) Bundled Municipal 6,352 119,959 3,666 EDF Industrial Power Retail Power Services (NY), LLC Energy Marketer 6,253 115,658 4 Retail Power Starion Energy NY, Inc. Energy Marketer 6,011 30,919 7,222 Pennsylvania Electric Co Bundled Investor Owned 5,648 46,715 3,406 Retail Power Entrust Energy Energy Marketer 5,583 53,149 4,600 Reliant Energy Northeast Retail Power LLC Energy Marketer 5,522 66,609 757 Retail Power Greenlight Energy Inc. Energy Marketer 4,732 42,623 5,013 Vivint Solar, Inc. Bundled Behind the Meter 4,432 29,407 3,808 Retail Power Ethical Electric, Inc. Energy Marketer 4,114 32,967 6,734

384

Retail Power UGI Energy Services, Inc. Energy Marketer 3,789 69,796 210 Stream Energy New York, Retail Power LLC Energy Marketer 3,757 46,630 5,720 Retail Power Residents Energy, LLC Energy Marketer 3,523 30,394 5,155 Retail Power Oasis Power, LLC Energy Marketer 3,388 21,061 6,001 SunPower Capital, LLC Bundled Behind the Meter 3,110 26,579 3,031 Sunrun Inc. Bundled Behind the Meter 3,084 19,308 2,439

Source: U.S. Energy Information Administration. (2017c)

C.2 Tabulation of Electric Power Ownership in NYS in 2015

Tabulation of Ownership Date: 07/10/17 Time: 17:54 Sample: 1 124 Included observations: 124 Number of categories: 6 Cumulative Cumulative Value Count Percent Count Percent Behind the Meter 12 9.68 12 9.68 Cooperative 1 0.81 13 10.48 Investor Owned 15 12.10 28 22.58 Municipal 12 9.68 40 32.26 Retail Power Marketer 81 65.32 121 97.58 State 3 2.42 124 100.00 Total 124 100.00 124 100.00

385

C.3 Panel Generalized Method of Moments Output Results

Dependent Variable: FFRET Method: Panel Generalized Method of Moments Date: 06/30/17 Time: 18:44 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRETBIOMASS(-1) FFRETBIOMASS(-2) FFRETBIOMASS(-3) @MOVAV(FIT,12) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) @MOVAV(PUBENFUND,12) @MOVAV(INTERCONSTAND,12) @MOVAV(LCVSCORE,12) @MOVAV(ELECTRICPRICE,3) Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C -3.200651 0.741882 -4.314231 0.0000 FFRETBIOMASS(-1) 0.956066 0.050223 19.03625 0.0000 FFRETBIOMASS(-2) -0.012530 0.070648 -0.177362 0.8593 FFRETBIOMASS(-3) -0.295844 0.050368 -5.873666 0.0000 @MOVAV(FIT,12) 4.433809 1.274458 3.478977 0.0006 @MOVAV(EFFRPS,12) 0.274539 0.038902 7.057159 0.0000 @MOVAV(NETMETER,12) 0.743936 0.176706 4.210010 0.0000 @MOVAV(PUBENFUND,12) -1.827216 0.762149 -2.397452 0.0170 @MOVAV(INTERCONSTAND,12 ) 0.056893 0.185627 0.306488 0.7594 @MOVAV(LCVSCORE,12) 0.015112 0.013680 1.104644 0.2701 @MOVAV(ELECTRICPRICE,3) 0.189716 0.080162 2.366647 0.0185

Effects Specification

Period fixed (dummy variables)

R-squared 0.971424 Mean dependent var 2.368756 Adjusted R-squared 0.955999 S.D. dependent var 3.783440 S.E. of regression 0.793632 Sum squared resid 221.7080 Durbin-Watson stat 2.180040 J-statistic 1.12E-20 Instrument rank 191

386

Dependent Variable: FFRET Method: Panel Generalized Method of Moments Date: 06/27/17 Time: 18:49 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRETBIOMASS(-1) FFRETBIOMASS(-2) FFRETBIOMASS(-3) @MOVAV(FIT,12) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) @MOVAV(PUBENFUND,12) @MOVAV(INTERCONSTAND,12) @MOVAV(LCVSCORE,12) @MOVAV(ELECTRICPRICE,3) IMPORTRATIO(-1) @LOG(PERCAPITAC O2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C -0.112844 26.21761 -0.004304 0.9966 FFRETBIOMASS(-1) 0.953534 0.050574 18.85423 0.0000 FFRETBIOMASS(-2) -0.013420 0.071025 -0.188949 0.8502 FFRETBIOMASS(-3) -0.296906 0.050688 -5.857568 0.0000 @MOVAV(FIT,12) 4.166051 1.460076 2.853312 0.0046 @MOVAV(EFFRPS,12) 0.297588 0.054186 5.491974 0.0000 @MOVAV(NETMETER,12) 0.789788 0.250124 3.157591 0.0017 @MOVAV(PUBENFUND,12) -1.636252 1.204247 -1.358735 0.1751 @MOVAV(INTERCONSTAND,12 ) 0.061986 0.281703 0.220038 0.8260 @MOVAV(LCVSCORE,12) 0.014064 0.014504 0.969690 0.3329 @MOVAV(ELECTRICPRICE,3) 0.196441 0.084938 2.312766 0.0213 IMPORTRATIO(-1) 0.444043 2.761531 0.160796 0.8723 @LOG(PERCAPITACO2) 0.341953 0.740480 0.461798 0.6445 @LOG(PERCAPITAGDP) -0.400935 2.275328 -0.176210 0.8602 NUCLEAR(-1) 0.014840 0.014438 1.027853 0.3047 @LOG(ENERGSITY) -0.104584 1.976661 -0.052909 0.9578

Effects Specification

Period fixed (dummy variables)

R-squared 0.971537 Mean dependent var 2.368756 Adjusted R-squared 0.955542 S.D. dependent var 3.783440 S.E. of regression 0.797742 Sum squared resid 220.8279 Durbin-Watson stat 2.187639 J-statistic 7.62E-10 Instrument rank 196

387

Dependent Variable: FFRETBIOMASS Method: Panel Generalized Method of Moments Date: 06/30/17 Time: 12:05 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRETBIOMASS(-1) FFRETBIOMASS(-2) FFRETBIOMASS(-3) FBNGT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO (-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C 20.15107 25.71461 0.783643 0.4338 FFRETBIOMASS(-1) 1.038763 0.054796 18.95677 0.0000 FFRETBIOMASS(-2) -0.042923 0.074992 -0.572371 0.5674 FFRETBIOMASS(-3) -0.263075 0.052816 -4.980990 0.0000 FBNGT(-1) 0.030426 0.018021 1.688332 0.0922 @MOVAV(EFFRPS,12) 0.270759 0.057445 4.713368 0.0000 @MOVAV(NETMETER,12) 0.446308 0.330503 1.350391 0.1778 IMPORTRATIO(-1) -2.544904 3.614400 -0.704101 0.4818 @LOG(PERCAPITACO2) 1.690983 0.807988 2.092833 0.0371 @LOG(PERCAPITAGDP) -1.819657 2.264822 -0.803444 0.4223 NUCLEAR(-1) 0.019360 0.018043 1.072996 0.2840 @LOG(ENERGSITY) -3.660622 2.139414 -1.711039 0.0880

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.972982 Mean dependent var 3.201509 Adjusted R-squared 0.958040 S.D. dependent var 4.103418 S.E. of regression 0.840548 Sum squared resid 246.5757 Durbin-Watson stat 2.147615 J-statistic 12.54096 Instrument rank 195 Prob(J-statistic) 0.000398

388

Dependent Variable: FFRET Method: Panel Generalized Method of Moments Date: 06/30/17 Time: 13:59 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRET(-1) FFRET(-2) FFRET(-3) FBNGT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO (-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

FFRET( -1) 1.062459 0.054697 19.42445 0.0000 FFRET(-2) -0.057891 0.075698 -0.764773 0.4449 FFRET(-3) -0.279460 0.052693 -5.303514 0.0000 FBNGT(-1) 0.029507 0.017814 1.656389 0.0985 @MOVAV(EFFRPS,12) 0.283272 0.055898 5.067672 0.0000 @MOVAV(NETMETER,12) 0.503793 0.322543 1.561939 0.1192 IMPORTRATIO(-1) -2.525739 3.535198 -0.714455 0.4754 @LOG(PERCAPITACO2) 1.728131 0.789480 2.188949 0.0293 @LOG(PERCAPITAGDP) -2.214668 2.210704 -1.001793 0.3171 NUCLEAR(-1) 0.023208 0.017641 1.315573 0.1892 @LOG(ENERGSITY) -3.464202 2.077033 -1.667861 0.0962 C 23.71866 25.08739 0.945441 0.3451

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.969752 Mean dependent var 2.368756 Adjusted R-squared 0.953025 S.D. dependent var 3.783440 S.E. of regression 0.820012 Sum squared resid 234.6747 Durbin-Watson stat 2.150536 J-statistic 13.83788 Instrument rank 195 Prob(J-statistic) 0.000199

389

Dependent Variable: FFRETBIOMASS Method: Panel Generalized Method of Moments Date: 06/29/17 Time: 17:53 Sample (adjusted): 2001M12 2016M08 Periods included: 177 Cross-sections included: 3 Total panel (balanced) observations: 531 2SLS instrument weighting matrix Instrument specification: C FFRETBIOMASS(-1) FFRETBIOMASS(-2) FFRETBIOMASS(-3) FBNGT(1) FBNGT(2) FBNGT(3) FBNGT(4) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO (-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C 35.32825 25.18710 1.402632 0.1616 FFRETBIOMASS(-1) 0.884114 0.054020 16.36646 0.0000 FFRETBIOMASS(-2) -0.029682 0.072521 -0.409285 0.6826 FFRETBIOMASS(-3) -0.133030 0.053406 -2.490909 0.0132 FBNGT(1) -0.058377 0.022612 -2.581710 0.0103 FBNGT(2) -0.023179 0.027873 -0.831573 0.4062 FBNGT(3) -0.034982 0.027747 -1.260769 0.2083 FBNGT(4) 0.029038 0.021509 1.350032 0.1779 @MOVAV(EFFRPS,12) 0.205213 0.061104 3.358446 0.0009 @MOVAV(NETMETER,12) 0.441286 0.321464 1.372738 0.1707 IMPORTRATIO(-1) 4.636159 3.425650 1.353366 0.1768 @LOG(PERCAPITACO2) -0.400144 0.887654 -0.450788 0.6524 @LOG(PERCAPITAGDP) -2.274771 2.225702 -1.022046 0.3075 NUCLEAR(-1) -0.009809 0.015717 -0.624065 0.5330 @LOG(ENERGSITY) -5.427853 2.109433 -2.573134 0.0105

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.974071 Mean dependent var 3.110300 Adjusted R-squared 0.959342 S.D. dependent var 3.996618 S.E. of regression 0.805868 Sum squared resid 219.5050 Durbin-Watson stat 2.072525 J-statistic 20.34731 Instrument rank 194 Prob(J-statistic) 0.000006

390

Dependent Variable: SOLAR Method: Panel Generalized Method of Moments Date: 06/29/17 Time: 21:20 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C SOLAR(-1) SOLAR(-2) SOLAR(-3) FBNGT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO (-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C 16.34501 11.44798 1.427764 0.1543 SOLAR(-1) 1.005764 0.056059 17.94130 0.0000 SOLAR(-2) -0.038083 0.079027 -0.481904 0.6302 SOLAR(-3) -0.070468 0.056944 -1.237493 0.2167 FBNGT(-1) 0.012763 0.006370 2.003661 0.0459 @MOVAV(EFFRPS,12) 0.077848 0.026498 2.937903 0.0035 @MOVAV(NETMETER,12) 0.046332 0.136801 0.338684 0.7351 IMPORTRATIO(-1) -0.582874 1.485439 -0.392392 0.6950 @LOG(PERCAPITACO2) 0.550737 0.329874 1.669537 0.0959 @LOG(PERCAPITAGDP) -1.433105 1.002020 -1.430216 0.1535 NUCLEAR(-1) 0.008483 0.007259 1.168558 0.2434 @LOG(ENERGSITY) -1.809343 0.934064 -1.937065 0.0535

Effects Specificat ion

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.979214 Mean dependent var 0.636070 Adjusted R-squared 0.967719 S.D. dependent var 1.916277 S.E. of regression 0.344298 Sum squared resid 41.37084 Durbin-Watson stat 1.944433 J-statistic 7.772331 Instrument rank 195 Prob(J-statistic) 0.005305

391

Dependent Variable: WIND Method: Panel Generalized Method of Moments Date: 06/29/17 Time: 21:26 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C WIND(-1) WIND(-2) WIND(-3) FBNGT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO(-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C -37.78543 19.71947 -1.916148 0.0562 WIND(-1) 0.859636 0.056085 15.32748 0.0000 WIND(-2) 0.012268 0.069830 0.175687 0.8606 WIND(-3) -0.331424 0.051140 -6.480681 0.0000 FBNGT(-1) -0.019466 0.015141 -1.285683 0.1994 @MOVAV(EFFRPS,12) 0.097826 0.039621 2.469044 0.0140 @MOVAV(NETMETER,12) 0.285317 0.253638 1.124901 0.2614 IMPORTRATIO(-1) -2.607790 2.563475 -1.017287 0.3097 @LOG(PERCAPITACO2) 0.515114 0.636300 0.809545 0.4188 @LOG(PERCAPITAGDP) 3.317485 1.756753 1.888418 0.0598 NUCLEAR(-1) 0.000661 0.014218 0.046508 0.9629 @LOG(ENERGSITY) 1.052575 1.599926 0.657890 0.5110

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.943339 Mean dependent var 1.7326 86 Adjusted R-squared 0.912005 S.D. dependent var 2.225145 S.E. of regression 0.660066 Sum squared resid 152.0550 Durbin-Watson stat 2.178331 J-statistic 13.59066 Instrument rank 195 Prob(J-statistic) 0.000227

392

Dependent Variable: FFRETBIOMASS Method: Panel Generalized Method of Moments Date: 06/30/17 Time: 14:49 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRETBIOMASS(-1) FFRETBIOMASS(-2) FFRETBIOMASS(-3) FBNGT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO(-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

FFRETBIOMASS( -1) 1.039872 0.054748 18.99375 0.0000 FFRETBIOMASS(-2) -0.042557 0.074983 -0.567554 0.5707 FFRETBIOMASS(-3) -0.263491 0.052805 -4.989865 0.0000 FBNGT(-1) 0.031764 0.017824 1.782074 0.0756 @MOVAV(EFFRPS,12) 0.270419 0.057436 4.708151 0.0000 @MOVAV(NETMETER,12) 0.440474 0.330275 1.333658 0.1832 IMPORTRATIO(-1) -3.313495 3.278202 -1.010766 0.3128 @LOG(PERCAPITACO2) 1.736927 0.802790 2.163612 0.0312 @LOG(PERCAPITAGDP) -2.029219 2.226345 -0.911458 0.3627 NUCLEAR(-1) 0.020763 0.017827 1.164705 0.2449 @LOG(ENERGSITY) -3.759081 2.130349 -1.764537 0.0785 C 22.47243 25.29885 0.888279 0.3750

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.972986 Mean dependent var 3.201509 Adjusted R-squared 0.958047 S.D. dependent var 4.103418 S.E. of regression 0.840482 Sum squared resid 246.5368 Durbin-Watson stat 2.149423 J-statistic 12.77546 Instrument rank 195 Prob(J-statistic) 0.000351

393

Dependent Variable: FFRET Method: Panel Generalized Method of Moments Date: 06/30/17 Time: 14:53 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRET(-1) FFRET(-2) FFRET(-3) FBNGT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO(-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

FFRET( -1) 1.063450 0.054642 19.46225 0.0000 FFRET(-2) -0.057592 0.075690 -0.760891 0.4472 FFRET(-3) -0.279826 0.052683 -5.311484 0.0000 FBNGT(-1) 0.030615 0.017611 1.738412 0.0830 @MOVAV(EFFRPS,12) 0.282964 0.055890 5.062876 0.0000 @MOVAV(NETMETER,12) 0.498957 0.322315 1.548039 0.1225 IMPORTRATIO(-1) -3.143396 3.204999 -0.980779 0.3274 @LOG(PERCAPITACO2) 1.765250 0.784333 2.250637 0.0250 @LOG(PERCAPITAGDP) -2.381447 2.173635 -1.095606 0.2740 NUCLEAR(-1) 0.024344 0.017425 1.397059 0.1633 @LOG(ENERGSITY) -3.542383 2.068329 -1.712679 0.0877 C 25.56444 24.68741 1.035525 0.3011

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.969755 Mean dependent var 2.368756 Adjusted R-squared 0.953030 S.D. dependent var 3.783440 S.E. of regression 0.819969 Sum squared resid 234.6497 Durbin-Watson stat 2.152098 J-statistic 14.00815 Instrument rank 195 Prob(J-statistic) 0.000182

394

Dependent Variable: FFRETBIOMASS Method: Panel GMM EGLS (Period random effects) Date: 07/01/17 Time: 22:22 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Swamy and Arora estimator of component variances Cross-section SUR (PCSE) standard errors & covariance (no d.f. correction) Instrument specification: C FFRETBIOMASS(-1) FFRETBIOMASS(-2) FFRETBIOMASS(-3) FBNGT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO (-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

FFRETBIOMASS( -1) 1.036632 0.063746 16.26184 0.0000 FFRETBIOMASS(-2) -0.104129 0.089179 -1.167641 0.2435 FFRETBIOMASS(-3) -0.205391 0.063324 -3.243498 0.0013 FBNGT(-1) 0.004535 0.013463 0.336829 0.7364 @MOVAV(EFFRPS,12) 0.242064 0.051195 4.728277 0.0000 @MOVAV(NETMETER,12) 0.219796 0.163913 1.340931 0.1805 IMPORTRATIO(-1) 4.117049 1.778536 2.314852 0.0210 @LOG(PERCAPITACO2) 0.078878 0.340010 0.231986 0.8166 @LOG(PERCAPITAGDP) 3.219439 1.293925 2.488119 0.0131 NUCLEAR(-1) -0.009296 0.011552 -0.804692 0.4214 @LOG(ENERGSITY) -0.030117 0.887726 -0.033926 0.9729 C -35.07325 14.50454 -2.418087 0.0159

Effects Specification S.D. Rho

Cross-section fixed (dummy variables) Period random 0.000000 0.0000 Idiosyncratic random 0.840288 1.0000

Weighted Statistics

R-squared 0.958525 Mean dependent var 3.201509 Adjusted R-squared 0.957506 S.D. dependent var 4.103418 S.E. of regression 0.845882 Sum squared resid 378.5085 Durbin-Watson stat 2.035789 J-statistic 20.59895 Instrument rank 15 Prob(J-statistic) 0.000006

Unweighted Statistics

R-squared 0.958525 Mean dependent var 3.201509 Sum squared resid 378.5085 Durbin-Watson stat 2.035789

Dependent Variable: FFRET Method: Panel GMM EGLS (Period random effects)

395

Date: 07/01/17 Time: 22:25 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Swamy and Arora estimator of component variances Instrument specification: C FFRET(-1) FFRET(-2) FFRET(-3) FBNGT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO (-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C -35.52220 15.66330 -2.267862 0.0237 FFRET(-1) 1.046417 0.044212 23.66826 0.0000 FFRET(-2) -0.090907 0.061980 -1.466707 0.1430 FFRET(-3) -0.232007 0.043335 -5.353820 0.0000 FBNGT(-1) 0.003675 0.012197 0.301335 0.7633 @MOVAV(EFFRPS,12) 0.244788 0.039687 6.167999 0.0000 @MOVAV(NETMETER,12) 0.218137 0.209604 1.040711 0.2985 IMPORTRATIO(-1) 3.992836 2.211740 1.805292 0.0716 @LOG(PERCAPITACO2) 0.098517 0.565641 0.174170 0.8618 @LOG(PERCAPITAGDP) 3.205080 1.383429 2.316765 0.0209 NUCLEAR(-1) -0.006227 0.012465 -0.499568 0.6176 @LOG(ENERGSITY) 0.170510 1.387943 0.122851 0.9023

Effects Specification S.D. Rho

Cross-section fixed (dummy variables) Period random 0.000000 0.0000 Idiosyncratic random 0.819758 1.0000

Weighted Statistics

R-squared 0.953958 Mean dependent var 2.368756 Adjusted R-squared 0.952826 S.D. dependent var 3.783440 S.E. of regression 0.821743 Sum squared resid 357.2134 Durbin-Watson stat 2.044299 J-statistic 22.08771 Instrument rank 15 Prob(J-statistic) 0.000003

Unweighted Statistics

R-squared 0.953958 Mean dependent var 2.368756 Sum squared resid 357.2134 Durbin-Watson stat 2.044299

396

Dependent Variable: FBNGT Method: Panel Generalized Method of Moments Date: 06/27/17 Time: 12:20 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 10 Total panel (balanced) observations: 1810 2SLS instrument weighting matrix Instrument specification: C FBNGT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) @MOVAV(PUBENFUND,12) @MOVAV(INTERCONSTAND,12) @MOVAV(LCVSCORE,12) @MOVAV(ELECTRICPRICE,3) IMPORTRATIO(-1) @LOG(PERCAPITAC O2) @LOG(PERCAPITAGDP) @LOG(ENERGSITY) Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C -14.97546 12.74150 -1.175329 0.2400 FBNGT(-1) 0.960891 0.006800 141.3177 0.0000 @MOVAV(EFFRPS,12) 0.027797 0.026033 1.067742 0.2858 @MOVAV(NETMETER,12) 0.042204 0.184454 0.228807 0.8190 @MOVAV(PUBENFUND,12) -0.915498 0.417257 -2.194088 0.0284 @MOVAV(INTERCONSTAND,12 ) -0.273184 0.186360 -1.465896 0.1429 @MOVAV(LCVSCORE,12) -0.001633 0.007680 -0.212604 0.8317 @MOVAV(ELECTRICPRICE,3) 0.142040 0.079410 1.788682 0.0739 IMPORTRATIO(-1) 0.533581 0.413724 1.289703 0.1973 @LOG(PERCAPITACO2) 0.318431 0.744377 0.427782 0.6689 @LOG(PERCAPITAGDP) 1.321494 1.275856 1.035770 0.3005 @LOG(ENERGSITY) -0.145949 0.679473 -0.214798 0.8300

Effects Specification

Period fixed (dummy variables)

R-squared 0.960622 Mean dependent var 20.69757 Adjusted R-squared 0.955974 S.D. dependent var 11.82996 S.E. of regression 2.482204 Sum squared resid 9969.039 Durbin-Watson stat 2.105389 J-statistic 3.98E-12 Instrument rank 192

397

Dependent Variable: FBNGT Method: Panel Generalized Method of Moments Date: 06/27/17 Time: 12:16 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 10 Total panel (balanced) observations: 1810 2SLS instrument weighting matrix Instrument specification: C FBNGT(-1) FBNGT(-2) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) @MOVAV(PUBENFUND,12) @MOVAV(INTERCONSTAND,12) @MOVAV(LCVSCORE,12) @MOVAV(ELECTRICPRICE,3) IMPORTRATIO(-1) @LOG(PERCAPITAC O2) @LOG(PERCAPITAGDP) @LOG(ENERGSITY) Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C -14.10 476 12.73419 -1.107629 0.2682 FBNGT(-1) 0.909910 0.024897 36.54769 0.0000 FBNGT(-2) 0.052904 0.024856 2.128448 0.0335 @MOVAV(EFFRPS,12) 0.025663 0.026024 0.986137 0.3242 @MOVAV(NETMETER,12) 0.041346 0.184253 0.224399 0.8225 @MOVAV(PUBENFUND,12) -0.888497 0.416995 -2.130713 0.0333 @MOVAV(INTERCONSTAND,12 ) -0.263302 0.186215 -1.413968 0.1576 @MOVAV(LCVSCORE,12) -0.001763 0.007672 -0.229810 0.8183 @MOVAV(ELECTRICPRICE,3) 0.137737 0.079349 1.735827 0.0828 IMPORTRATIO(-1) 0.520441 0.413319 1.259174 0.2081 @LOG(PERCAPITACO2) 0.328930 0.743583 0.442359 0.6583 @LOG(PERCAPITAGDP) 1.243131 1.274998 0.975007 0.3297 @LOG(ENERGSITY) -0.175597 0.678876 -0.258658 0.7959

Effects Specification

Period fixed (dummy variables)

R-squared 0.960732 Mean dependent var 20.69757 Adjusted R-squared 0.956070 S.D. dependent var 11.82996 S.E. of regression 2.479500 Sum squared resid 9941.187 Durbin-Watson stat 2.007214 J-statistic 4.03E-14 Instrument rank 193

398

Dependent Variable: FBNGT Method: Panel Generalized Method of Moments Date: 06/27/17 Time: 12:21 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 10 Total panel (balanced) observations: 1810 2SLS instrument weighting matrix Instrument specification: C FBNGT(-1) FBNGT(-2) FBNGT(-3) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) @MOVAV(PUBENFU ND,12) @MOVAV(INTERCONSTAND,12) @MOVAV(LCVSCORE,12) @MOVAV(ELECTRICPRICE,3) IMPORTRATIO(-1) @LOG(PERCAPITAC O2) @LOG(PERCAPITAGDP) @LOG(ENERGSITY) Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C -13.63139 12.73537 -1.070356 0.2846 FBNGT(-1) 0.908107 0.024924 36.43466 0.0000 FBNGT(-2) 0.021625 0.033708 0.641542 0.5213 FBNGT(-3) 0.034237 0.024929 1.373356 0.1698 @MOVAV(EFFRPS,12) 0.024641 0.026028 0.946737 0.3439 @MOVAV(NETMETER,12) 0.042434 0.184204 0.230364 0.8178 @MOVAV(PUBENFUND,12) -0.874672 0.417002 -2.097521 0.0361 @MOVAV(INTERCONSTAND,12 ) -0.257760 0.186208 -1.384260 0.1665 @MOVAV(LCVSCORE,12) -0.001804 0.007670 -0.235170 0.8141 @MOVAV(ELECTRICPRICE,3) 0.135361 0.079347 1.705942 0.0882 IMPORTRATIO(-1) 0.509127 0.413288 1.231893 0.2182 @LOG(PERCAPITACO2) 0.333355 0.743386 0.448428 0.6539 @LOG(PERCAPITAGDP) 1.200318 1.275030 0.941404 0.3466 @LOG(ENERGSITY) -0.190832 0.678780 -0.281140 0.7786

Effects Specification

Period fixed (dummy variables)

R-squared 0.960778 Mean dependent var 20.69757 Adjusted R-squared 0.956094 S.D. dependent var 11.82996 S.E. of regression 2.478821 Sum squared resid 9929.598 Durbin-Watson stat 2.010318 J-statistic 1.82E-12 Instrument rank 194

399

Dependent Variable: FBNGT Method: Panel Generalized Method of Moments Date: 07/01/17 Time: 11:06 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FBNGT(-1) @MOVAV(FIT,12) @MOVAV(EFFRPS, 12) @MOVAV(NETMETER,12) @MOVAV(INTERCONSTAND,12) @MOVAV(LCVSCORE,12) @MOVAV(ELECTRICPRICE,12) Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C 4.389498 3.065796 1.431764 0.1531 FBNGT(-1) 0.790367 0.032613 24.23493 0.0000 @MOVAV(FIT,12) -1.271693 3.400560 -0.373966 0.7087 @MOVAV(EFFRPS,12) -0.305924 0.114342 -2.675515 0.0078 @MOVAV(NETMETER,12) -0.911676 0.484697 -1.880920 0.0608 @MOVAV(INTERCONSTAND,12 ) 0.678048 0.826370 0.820514 0.4125 @MOVAV(LCVSCORE,12) -0.032634 0.037838 -0.862455 0.3890 @MOVAV(ELECTRICPRICE,12) 0.224499 0.215311 1.042676 0.2978

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.970878 Mean dependent var 20.30058 Adjusted R-squared 0.955285 S.D. dependent var 10.33042 S.E. of regression 2.184461 Sum squared resid 1684.470 Durbin-Watson stat 1.671356 J-statistic 1.46E-18 Instrument rank 190

400

Dependent Variable: FBNGT Method: Panel Generalized Method of Moments Date: 07/01/17 Time: 11:03 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FBNGT(-1) @MOVAV(FIT,12) @MOVAV(EFFRPS, 12) @MOVAV(NETMETER,12) @MOVAV(PUBENFUND,12) @MOVAV(INTERCONSTAND,12) @MOVAV(LCVSCORE,12) @MOVAV(ELECTRICPRICE,3) IMPORTRATIO(-1) @LOG(PERCAPITAC O2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C 30.45918 71.37729 0.426735 0.6698 FBNGT(-1) 0.783706 0.039060 20.06411 0.0000 @MOVAV(FIT,12) -0.302473 3.844806 -0.078670 0.9373 @MOVAV(EFFRPS,12) 0.056316 0.121364 0.464031 0.6429 @MOVAV(NETMETER,12) 0.799156 0.629375 1.269761 0.2050 @MOVAV(PUBENFUND,12) 2.784372 3.242477 0.858718 0.3911 @MOVAV(INTERCONSTAND,12 ) -0.022144 0.765812 -0.028916 0.9769 @MOVAV(LCVSCORE,12) -0.025555 0.039744 -0.643001 0.5206 @MOVAV(ELECTRICPRICE,3) 0.182102 0.221618 0.821694 0.4118 IMPORTRATIO(-1) 14.26199 7.858210 1.814916 0.0704 @LOG(PERCAPITACO2) -4.393236 2.148892 -2.044419 0.0417 @LOG(PERCAPITAGDP) -2.469428 6.184499 -0.399293 0.6899 NUCLEAR(-1) 0.057265 0.044161 1.296720 0.1956 @LOG(ENERGSITY) 3.869568 5.384967 0.718587 0.4729

Effects Specification

Period fixed (dummy variables)

R-squared 0.971563 Mean dependent var 20.30058 Adjusted R-squared 0.955837 S.D. dependent var 10.33042 S.E. of regression 2.170935 Sum squared resid 1644.823 Durbin-Watson stat 1.643338 J-statistic 6.43E-10 Instrument rank 194

401

Dependent Variable: FBNGT Method: Panel Generalized Method of Moments Date: 07/01/17 Time: 11:15 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 10 Total panel (balanced) observations: 1810 2SLS instrument weighting matrix Instrument specification: C FBNGT(-1) FFRET(-1) @MOVAV(EFFRPS,12) @MOVAV(LCVSCORE,12) IMPORTRATIO(-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C 5.012563 18.17 542 0.275788 0.7827 FBNGT(-1) 0.800808 0.018420 43.47398 0.0000 FFRET(-1) -0.082406 0.041892 -1.967080 0.0493 @MOVAV(EFFRPS,12) 0.016868 0.028793 0.585826 0.5581 @MOVAV(LCVSCORE,12) -0.024134 0.010059 -2.399129 0.0165 IMPORTRATIO(-1) -1.802318 1.234982 -1.459389 0.1447 @LOG(PERCAPITACO2) -4.782176 1.189365 -4.020782 0.0001 @LOG(PERCAPITAGDP) 0.934487 1.609616 0.580565 0.5616 NUCLEAR(-1) 0.084497 0.019241 4.391514 0.0000 @LOG(ENERGSITY) 0.846380 1.800890 0.469978 0.6384

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.964837 Mean dependent var 20.69757 Adjusted R-squared 0.960516 S.D. dependent var 11.82996 S.E. of regression 2.350689 Sum squared resid 8901.966 Durbin-Watson stat 1.894115 J-statistic 2.01E-11 Instrument rank 199

402

Dependent Variable: FBNGT Method: Panel Generalized Method of Moments Date: 07/01/17 Time: 11:16 Sample (adjusted): 2001M12 2016M09 Periods included: 178 Cross-sections included: 10 Total panel (balanced) observations: 1780 2SLS instrument weighting matrix Instrument specification: C FBNGT(-1) FFRET(+1) FFRET(+2) FFRET(+3) @MOVAV(EFFRPS,12) @MOVAV(LCVSCORE,12) IMPORTRATIO(-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C -3.829825 18.52648 -0.206722 0.8363 FBNGT(-1) 0.772605 0.017408 44.38174 0.0000 FFRET(1) -0.268631 0.088080 -3.049845 0.0023 FFRET(2) -0.155451 0.122526 -1.268716 0.2047 FFRET(3) 0.156836 0.085306 1.838515 0.0662 @MOVAV(EFFRPS,12) 0.003956 0.030292 0.130597 0.8961 @MOVAV(LCVSCORE,12) -0.015995 0.009971 -1.604203 0.1089 IMPORTRATIO(-1) -1.894899 1.224184 -1.547887 0.1218 @LOG(PERCAPITACO2) -4.285595 1.267716 -3.380565 0.0007 @LOG(PERCAPITAGDP) 1.945076 1.646153 1.181589 0.2375 NUCLEAR(-1) 0.064795 0.018909 3.426620 0.0006 @LOG(ENERGSITY) -0.633132 1.863371 -0.339778 0.7341

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.966358 Mean dependent var 20.63142 Adjusted R-squared 0.962168 S.D. dependent var 11.85237 S.E. of regression 2.305332 Sum squared resid 8407.627 Durbin-Watson stat 1.933876 J-statistic 6.10E-12 Instrument rank 198

403

Dependent Variable: FFRETBIOMASS Method: Panel Generalized Method of Moments Date: 07/01/17 Time: 22:18 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRETBIOMASS(-1) FFRETBIOMASS(-2) FFRETBIOMASS(-3) FBNGT(-1) HYDRO(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO(-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

FFRETBIOMASS( -1) 1.038029 0.054881 18.91429 0.0000 FFRETBIOMASS(-2) -0.040837 0.075103 -0.543741 0.5870 FFRETBIOMASS(-3) -0.262408 0.052882 -4.962090 0.0000 FBNGT(-1) 0.023266 0.022635 1.027911 0.3047 HYDRO(-1) -0.010879 0.017834 -0.610025 0.5422 @MOVAV(EFFRPS,12) 0.252474 0.064577 3.909636 0.0001 @MOVAV(NETMETER,12) 0.414860 0.333227 1.244975 0.2140 IMPORTRATIO(-1) -3.282527 3.281538 -1.000301 0.3179 @LOG(PERCAPITACO2) 1.602458 0.833196 1.923266 0.0553 @LOG(PERCAPITAGDP) -2.021324 2.228353 -0.907093 0.3650 NUCLEAR(-1) 0.019865 0.017903 1.109583 0.2679 @LOG(ENERGSITY) -3.918231 2.148164 -1.823990 0.0690 C 23.31509 25.35896 0.919402 0.3585

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.973015 Mean dependent var 3.201509 Adjusted R-squared 0.957971 S.D. dependent var 4.103418 S.E. of regression 0.841239 Sum squared resid 246.2735 Durbin-Watson stat 2.154775 J-statistic 12.43490 Instrument rank 196 Prob(J-statistic) 0.000421

404

Dependent Variable: FFRET Method: Panel Generalized Method of Moments Date: 07/01/17 Time: 22:20 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRET(-1) FFRET(-2) FFRET(-3) FBNGT(-1) HYDRO(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO(-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

FFRET( -1) 1.061986 0.054807 19.37690 0.0000 FFRET(-2) -0.056328 0.075833 -0.742796 0.4581 FFRET(-3) -0.279064 0.052773 -5.288016 0.0000 FBNGT(-1) 0.024604 0.022333 1.101700 0.2714 HYDRO(-1) -0.007634 0.017408 -0.438512 0.6613 @MOVAV(EFFRPS,12) 0.270438 0.062826 4.304519 0.0000 @MOVAV(NETMETER,12) 0.481190 0.325228 1.479550 0.1399 IMPORTRATIO(-1) -3.119828 3.209182 -0.972157 0.3316 @LOG(PERCAPITACO2) 1.670238 0.814592 2.050397 0.0411 @LOG(PERCAPITAGDP) -2.376651 2.176253 -1.092084 0.2756 NUCLEAR(-1) 0.023699 0.017507 1.353657 0.1767 @LOG(ENERGSITY) -3.656759 2.087090 -1.752086 0.0806 C 26.17102 24.75533 1.057187 0.2912

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.969772 Mean dependent var 2.368756 Adjusted R-squared 0.952921 S.D. dependent var 3.783440 S.E. of regression 0.820919 Sum squared resid 234.5201 Durbin-Watson stat 2.155472 J-statistic 13.96708 Instrument rank 196 Prob(J-statistic) 0.000186

405

Dependent Variable: FFRETBIOMASS Method: Panel Generalized Method of Moments Date: 07/01/17 Time: 22:24 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRETBIOMASS(-1) FFRETBIOMASS(-2) FFRETBIOMASS(-3) FBNGT(-1) BLFT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO (-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

FFRETBIOMASS( -1) 1.049087 0.05 7629 18.20419 0.0000 FFRETBIOMASS(-2) -0.041868 0.075083 -0.557625 0.5775 FFRETBIOMASS(-3) -0.263232 0.052865 -4.979335 0.0000 FBNGT(-1) 0.033044 0.018577 1.778793 0.0761 BLFT(-1) 0.010470 0.017985 0.582166 0.5608 @MOVAV(EFFRPS,12) 0.252761 0.065276 3.872199 0.0001 @MOVAV(NETMETER,12) 0.416876 0.334614 1.245839 0.2137 IMPORTRATIO(-1) -2.683539 3.622828 -0.740731 0.4594 @LOG(PERCAPITACO2) 1.565934 0.837238 1.870358 0.0623 @LOG(PERCAPITAGDP) -1.800061 2.267229 -0.793948 0.4278 NUCLEAR(-1) 0.029179 0.024672 1.182695 0.2377 @LOG(ENERGSITY) -3.827537 2.159947 -1.772051 0.0773 C 19.84435 25.74501 0.770804 0.4413

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.973009 Mean dependent var 3.201509 Adjusted R-squared 0.957963 S.D. dependent var 4.103418 S.E. of regression 0.841322 Sum squared resid 246.3222 Durbin-Watson stat 2.151883 J-statistic 12.23200 Instrument rank 196 Prob(J-statistic) 0.000470

406

Dependent Variable: FFRET Method: Panel Generalized Method of Moments Date: 07/01/17 Time: 22:28 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRET(-1) FFRET(-2) FFRET(-3) FBNGT(-1) BLFT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO (-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C 23.68900 25.12579 0.942816 0.3464 FFRET(-1) 1.067383 0.057343 18.61402 0.0000 FFRET(-2) -0.057427 0.075812 -0.757485 0.4493 FFRET(-3) -0.279663 0.052766 -5.300101 0.0000 FBNGT(-1) 0.030818 0.018355 1.678985 0.0941 BLFT(-1) 0.004975 0.017394 0.286019 0.7750 @MOVAV(EFFRPS,12) 0.274921 0.063099 4.356973 0.0000 @MOVAV(NETMETER,12) 0.489848 0.326468 1.500446 0.1344 IMPORTRATIO(-1) -2.629086 3.545796 -0.741466 0.4589 @LOG(PERCAPITACO2) 1.670833 0.817622 2.043527 0.0418 @LOG(PERCAPITAGDP) -2.214434 2.213939 -1.000223 0.3179 NUCLEAR(-1) 0.027943 0.024032 1.162726 0.2457 @LOG(ENERGSITY) -3.555753 2.101542 -1.691973 0.0915

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.969760 Mean dependent var 2.368756 Adjusted R-squared 0.952903 S.D. dependent var 3.783440 S.E. of regression 0.821079 Sum squared resid 234.6117 Durbin-Watson stat 2.152300 J-statistic 13.98150 Instrument rank 196 Prob(J-statistic) 0.000185

407

Dependent Variable: FFRETBIOMASS Method: Panel Generalized Method of Moments Date: 07/01/17 Time: 22:29 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRETBIOMASS(-1) FFRETBIOMASS(-2) FFRETBIOMASS(-3) FBNGT(-1) COAL(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO (-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

FFRETBIOMASS( -1) 1.050830 0.054991 19.10923 0.0000 FFRETBIOMASS(-2) -0.034109 0.074889 -0.455453 0.6491 FFRETBIOMASS(-3) -0.268690 0.052724 -5.096157 0.0000 FBNGT(-1) 0.050915 0.021034 2.420628 0.0160 COAL(-1) 0.032734 0.017559 1.864290 0.0631 @MOVAV(EFFRPS,12) 0.242482 0.059224 4.094293 0.0001 @MOVAV(NETMETER,12) 0.349789 0.333408 1.049131 0.2948 IMPORTRATIO(-1) -1.716810 3.632354 -0.472644 0.6368 @LOG(PERCAPITACO2) 1.111505 0.863718 1.286884 0.1990 @LOG(PERCAPITAGDP) -3.007852 2.343573 -1.283447 0.2002 NUCLEAR(-1) 0.042460 0.021812 1.946614 0.0524 @LOG(ENERGSITY) -4.096076 2.144628 -1.909924 0.0570 C 33.47725 26.58563 1.259223 0.2088

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.973239 Mean dependent var 3.201509 Adjusted R-squared 0.958321 S.D. dependent var 4.103418 S.E. of regression 0.837731 Sum squared resid 244.2239 Durbin-Watson stat 2.165007 J-statistic 10.04645 Instrument rank 196 Prob(J-statistic) 0.001526

408

Dependent Variable: FFRET Method: Panel Generalized Method of Moments Date: 07/01/17 Time: 22:30 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRET(-1) FFRET(-2) FFRET(-3) FBNGT(-1) COAL(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO (-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C 35.90187 25.95541 1.383213 0.1675 FFRET(-1) 1.074134 0.054949 19.54801 0.0000 FFRET(-2) -0.049344 0.075644 -0.652313 0.5146 FFRET(-3) -0.285736 0.052668 -5.425204 0.0000 FBNGT(-1) 0.048569 0.020830 2.331629 0.0203 COAL(-1) 0.029982 0.017169 1.746277 0.0816 @MOVAV(EFFRPS,12) 0.257756 0.057623 4.473172 0.0000 @MOVAV(NETMETER,12) 0.415489 0.325570 1.276188 0.2027 IMPORTRATIO(-1) -1.792606 3.552455 -0.504611 0.6142 @LOG(PERCAPITACO2) 1.201162 0.843568 1.423906 0.1554 @LOG(PERCAPITAGDP) -3.299859 2.288934 -1.441657 0.1503 NUCLEAR(-1) 0.044487 0.021378 2.080983 0.0382 @LOG(ENERGSITY) -3.874063 2.084193 -1.858783 0.0639

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.970007 Mean dependent var 2.368756 Adjusted R-squared 0.953287 S.D. dependent var 3.783440 S.E. of regression 0.817724 Sum squared resid 232.6982 Durbin-Watson stat 2.167284 J-statistic 11.47799 Instrument rank 196 Prob(J-statistic) 0.000704

409

Dependent Variable: FFRETDELLTA Method: Panel Generalized Method of Moments Date: 06/29/17 Time: 14:11 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRETBIOMASS(-1) FFRETBIOMASS(-2) FFRETBIOMASS(-3) FBNGT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO (-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coefficient Std. Error t-Statistic Prob.

C 7.357985 27.31580 0.269367 0.7878 FFRETBIOMASS(-1) 0.032741 0.057916 0.565320 0.5722 FFRETBIOMASS(-2) -0.033573 0.079674 -0.421378 0.6737 FFRETBIOMASS(-3) -0.258408 0.056058 -4.609688 0.0000 FBNGT(-1) 0.079769 0.018720 4.261209 0.0000 @MOVAV(EFFRPS,12) 0.298894 0.052951 5.644777 0.0000 @MOVAV(NETMETER,12) 0.448289 0.256315 1.748982 0.0812 IMPORTRATIO(-1) -5.315276 2.246474 -2.366052 0.0185 @LOG(PERCAPITACO2) 1.976474 0.814938 2.425306 0.0158 @LOG(PERCAPITAGDP) -1.049461 2.388243 -0.439428 0.6606 NUCLEAR(-1) 0.040568 0.017215 2.356525 0.0190 @LOG(ENERGSITY) -2.011027 1.428737 -1.407555 0.1601

Effects Specification

Period fixed (dummy variables)

R-squared 0.557577 Mean dependent var -0.001735 Adjusted R-squared 0.316828 S.D. dependent var 1.080566 S.E. of regression 0.893133 Sum squared resid 279.9882 Durbin-Watson stat 1.754130 J-statistic 11.57782 Instrument rank 193 Prob(J-statistic) 0.000667

410

Dependent Variable: FFRETDELLTA Method: Panel Generalized Method of Moments Date: 06/29/17 Time: 14:12 Sample (adjusted): 2001M12 2016M12 Periods included: 181 Cross-sections included: 3 Total panel (balanced) observations: 543 2SLS instrument weighting matrix Instrument specification: C FFRET(-1) FFRET(-2) FFRET(-3) FBNGT(-1) @MOVAV(EFFRPS,12) @MOVAV(NETMETER,12) IMPORTRATIO (-1) @LOG(PERCAPITACO2) @LOG(PERCAPITAGDP) NUCLEAR(-1) @LOG(ENERGSITY) FIT Constant added to instrument list

Variable Coeffici ent Std. Error t-Statistic Prob.

C 7.676135 27.26412 0.281547 0.7785 FFRET(-1) 0.027187 0.059439 0.457389 0.6477 FFRET(-2) -0.058202 0.082260 -0.707534 0.4797 FFRET(-3) -0.237210 0.057262 -4.142560 0.0000 FBNGT(-1) 0.076251 0.019358 3.938924 0.0001 @MOVAV(EFFRPS,12) 0.286143 0.060744 4.710614 0.0000 @MOVAV(NETMETER,12) 0.323591 0.350511 0.923197 0.3565 IMPORTRATIO(-1) -6.689993 3.841733 -1.741400 0.0825 @LOG(PERCAPITACO2) 2.169634 0.857928 2.528923 0.0119 @LOG(PERCAPITAGDP) -1.082256 2.402517 -0.450467 0.6527 NUCLEAR(-1) 0.043187 0.019170 2.252854 0.0249 @LOG(ENERGSITY) -2.262476 2.257147 -1.002361 0.3169

Effects Specification

Cross-section fixed (dummy variables) Period fixed (dummy variables)

R-squared 0.562093 Mean dependent var -0.001735 Adjusted R-squared 0.319927 S.D. dependent var 1.080566 S.E. of regression 0.891105 Sum squared resid 277.1298 Durbin-Watson stat 1.771057 J-statistic 14.97953 Instrument rank 195 Prob(J-statistic) 0.000109

411

D. PERCENTAGE CHANGE IN ENERGY INTENSITY

State % Change (2001-2007) % Change (2008-2014) % Change (2001-2014) AL -0.06 -0.05 -0.12 AK -0.19 -0.12 -0.36 AZ -0.14 -0.05 -0.13 AR -0.13 0.00 -0.14 CA -0.14 -0.09 -0.25 CO -0.03 -0.10 -0.15 CT 0.25 -0.12 0.02 DE -0.08 -0.08 -0.13 DC -0.26 0.03 -0.32 FL -0.13 0.00 -0.13 GA -0.02 -0.10 -0.15 HI 0.00 -0.01 -0.20 ID -0.14 0.07 -0.09 IL -0.03 -0.01 -0.02 ID -0.10 -0.07 -0.16 IO -0.03 -0.02 0.05 KS -0.05 0.01 -0.11 KY -0.06 -0.11 -0.18 LA -0.03 -0.03 -0.07 ME -0.12 -0.07 -0.18 MD -0.14 -0.15 -0.30 MS -0.11 -0.15 -0.26 MI -0.05 -0.05 -0.07 MN -0.03 -0.05 -0.08 MS -0.12 0.05 -0.14 MS -0.01 -0.04 -0.07

412

MO -0.01 -0.14 -0.15 NE -0.05 0.01 -0.01 NV -0.25 0.03 -0.21 NH 0.04 -0.11 -0.08 NJ -0.03 -0.05 -0.10 NM -0.08 -0.09 -0.20 NY -0.11 -0.15 -0.23 NC -0.11 -0.09 -0.20 ND 0.17 -0.30 -0.15 OH -0.04 -0.10 -0.13 OK -0.04 -0.10 -0.15 OR -0.21 -0.09 -0.30 PA -0.03 -0.05 -0.09 RI -0.20 -0.05 -0.19 SC -0.03 -0.07 -0.11 SD -0.05 0.08 0.10 TN -0.11 -0.12 -0.26 TX -0.23 -0.06 -0.30 UT -0.14 -0.11 -0.22 VT -0.09 0.19 0.09 VA -0.07 -0.07 -0.19 WA -0.08 -0.08 -0.17 WV 0.02 -0.14 -0.16 WI -0.10 -0.07 -0.15 WY -0.20 0.04 -0.23 U.S Average -0.10 -0.08 -0.19

413

E. SUPPLEMENTARY FIGURES AND DATA

Note: Signposts of the transformation underway in the U.S. electric power sector. The signposts of the unprecedented change in the power systems in the United States include the two most disruptive drivers: fast-flexing renewable resources (i.e., solar PV and wind), and natural gas. While natural gas and renewable electricity generation are on the rise as shown in the figure above, coal is declining, while petroleum and nuclear remain nearly constant.

414

Note: Intermittent renewable electricity is experiencing rapid and unprecedented expansion as shown in the two figures above.

415

416

Note: Emerging distributed energy generation attributes.

417

F. COPYRIGHT PERMISSION LETTER The work presented in this dissertation has been published elsewhere in a variety of articles and journals. Please find the copyright license agreement attached below.

418 Wiley Interdisciplinary Reviews: Energy and Environment

Published by Wiley (the "Owner")

Date: 2016-10-20

Contributor name: Joseph Nyangon

Contributor address: University of Delaware Energy and Environmental Policy 11 Academy Street, Graham Hall Newark Delaware United States 19716-5600

Manuscript number: WENE-519.R2

Re: Manuscript entitled An Assessment of Price Convergence Between Natural Gas and Solar PV in the U.S. Electricity Market (the "Contribution") for publication in Wiley Interdisciplinary Reviews: Energy and Environment (the "Journal") published by Wiley Periodicals, Inc. ("Wiley")

Dear Contributor(s):

Thank you for submitting your Contribution for publication. In order to expedite the editing and publishing process and enable the Owner to disseminate your Contribution to the fullest extent, we need to have this Copyright Transfer Agreement executed. If the Contribution is not accepted for publication, or if the Contribution is subsequently rejected, this Agreement shall be null and void. Publication cannot proceed without a signed copy of this Agreement.

A. COPYRIGHT

1. The Contributor assigns to the Owner, during the full term of copyright and any extensions or renewals, all copyright in and to the Contribution, and all rights therein, including but not limited to the right to publish, republish, transmit, sell, distribute and otherwise use the Contribution in whole or in part in electronic and print editions of the Journal and in derivative works throughout the world, in all languages and in all media of expression now known or later developed, and to license or permit others to do so. For the avoidance of doubt, "Contribution" is defined to only include the article submitted by the Contributor for publication in the Journal and does not extend to any

supporting information submitted with or referred to in the Contribution ("Supporting Information"). To the extent supporting information submitted with or referred to in the Contribution ("Supporting Information"). To the extent that any Supporting Information is submitted to the Journal for online hosting, the Owner is granted a perpetual, non-exclusive license to host and disseminate this Supporting Information for this purpose.

2. Reproduction, posting, transmission or other distribution or use of the final Contribution in whole or in part in any medium by the Contributor as permitted by this Agreement requires a citation to the Journal suitable in form and content as follows: (Title of Article, Contributor, Journal Title and Volume/Issue, Copyright © [year], copyright owner as specified in the Journal, Publisher). Links to the final article on the publisher website are encouraged where appropriate.

B. RETAINED RIGHTS

Notwithstanding the above, the Contributor or, if applicable, the Contributor’s employer, retains all proprietary rights other than copyright, such as patent rights, in any process, procedure or article of manufacture described in the Contribution.

C. PERMITTED USES BY CONTRIBUTOR

1. Submitted Version. The Owner licenses back the following rights to the Contributor in the version of the Contribution as originally submitted for publication (the "Submitted Version"):

a. The right to self-archive the Submitted Version on the Contributor’s personal website, place in a not for profit subject-based preprint server or repository or in a Scholarly Collaboration Network (SCN) which has signed up to the STM article sharing principles [ http://www.stm-assoc.org/stm-consultations/scn-consultation-2015/] ("Compliant SCNs"), or in the Contributor’s company/ institutional repository or archive. This right extends to both intranets and the Internet. The Contributor may replace the Submitted Version with the Accepted Version, after any relevant embargo period as set out in paragraph C.2(a) below has elapsed. The Contributor may wish to add a note about acceptance by the Journal and upon publication it is recommended that Contributors add a Digital Object Identifier (DOI) link back to the Final Published Version.

b. The right to transmit, print and share copies of the Submitted Version with colleagues, including via Compliant SCNs, provided that there is no systematic distribution of the Submitted Version, e.g. posting on a listserve, network (including SCNs which have not signed up to the STM sharing principles) or automated delivery.

2. Accepted Version. The Owner licenses back the following rights to the Contributor in the version of the Contribution that has been peer-reviewed and accepted for publication, but not final (the "Accepted Version"):

a. The right to self-archive the Accepted Version on the Contributor’s personal website, in the Contributor’s company/institutional repository or archive, in Compliant SCNs, and in not for profit subject-based repositories such as PubMed Central, subject to an embargo period of 12 months for scientific, technical and medical (STM) journals and 24 months for social science and humanities (SSH) journals following publication of the Final Published Version. There are separate arrangements with certain funding agencies governing reuse of the Accepted Version as set forth at the following website: http://www.wiley.com/go/funderstatement. The Contributor may not update the Accepted Version or replace it with the Final Published Version. The Accepted Version posted must contain a legend as follows: This is the accepted version of the following article: FULL CITE, which has been published in final form at [Link to final article]. This article may be used for non-commercial purposes in accordance with the Wiley Self-Archiving Policy [ http://olabout.wiley.com/WileyCDA/Section/id-828039.html ]. b. The right to transmit, print and share copies of the Accepted Version with colleagues, including via Compliant SCNs (in private research groups only before the embargo and publicly after), provided that there is no systematic distribution of the Accepted Version, e.g. posting on a listserve, network (including SCNs which have not signed up to the STM sharing principles) or automated delivery.

3. Final Published Version. The Owner hereby licenses back to the Contributor the following rights with respect to the final published version of the Contribution (the "Final Published Version"):

a. Copies for colleagues. The personal right of the Contributor only to send or transmit individual copies of the Final Published Version in any format to colleagues upon their specific request, and to share copies in private sharing groups in Compliant SCNs, provided no fee is charged, and further provided that there is no systematic external or public distribution of the Final Published Version, e.g. posting on a listserve, network or automated delivery.

b. Re-use in other publications. The right to re-use the Final Published Version or parts thereof for any publication authored or edited by the Contributor (excluding journal articles) where such re-used material constitutes less than half of the total material in such publication. In such case, any modifications must be accurately noted.

c. Teaching duties. The right to include the Final Published Version in teaching or training duties at the Contributor’s institution/place of employment including in course packs, e-reserves, presentation at professional conferences, in-house training, or distance learning. The Final Published Version may not be used in seminars outside of normal teaching obligations (e.g. commercial seminars). Electronic posting of the Final Published Version in connection with teaching/training at the Contributor’s company/institution is permitted subject to the implementation of reasonable access control mechanisms, such as user name and password. Posting the Final Published Version on the open Internet is not permitted.

d. Oral presentations. The right to make oral presentations based on the Final Published Version.

4. Article Abstracts, Figures, Tables, Artwork and Selected Text (up to 250 words).

a. Contributors may re-use unmodified abstracts for any non-commercial purpose. For online uses of the abstracts, the Owner encourages but does not require linking back to the Final Published Version.

b. Contributors may re-use figures, tables, artwork, and selected text up to 250 words from their Contributions, provided the following conditions are met:

(i) Full and accurate credit must be given to the Final Published Version.

(ii) Modifications to the figures and tables must be noted. Otherwise, no changes may be made.

(iii) The re-use may not be made for direct commercial purposes, or for financial consideration to the Contributor.

(iv) Nothing herein will permit dual publication in violation of journal ethical practices.

D. CONTRIBUTIONS OWNED BY EMPLOYER

1. If the Contribution was written by the Contributor in the course of the Contributor’s employment (as a "work-made-for-hire" in the course of employment), the Contribution is owned by the company/institution which must execute this Agreement (in addition to the Contributor’s signature). In such case, the company/institution hereby assigns to the Owner, during the full term of copyright, all copyright in and to the Contribution for the full hereby assigns to the Owner, during the full term of copyright, all copyright in and to the Contribution for the full term of copyright throughout the world as specified in paragraph A above.

For company/institution-owned work, signatures cannot be collected electronically and so instead please print off this Agreement, ask the appropriate person in your company/institution to sign the Agreement as well as yourself in the space provided below, and email a scanned copy of the signed Agreement to the Journal production editor. For production editor contact details, please visit the Journal’s online author guidelines.

2. In addition to the rights specified as retained in paragraph B above and the rights granted back to the Contributor pursuant to paragraph C above, the Owner hereby grants back, without charge, to such company/institution, its subsidiaries and divisions, the right to make copies of and distribute the Final Published Version internally in print format or electronically on the Company’s internal network. Copies so used may not be resold or distributed externally. However, the company/institution may include information and text from the Final Published Version as part of an information package included with software or other products offered for sale or license or included in patent applications. Posting of the Final Published Version by the company/institution on a public access website may only be done with written permission, and payment of any applicable fee(s). Also, upon payment of the applicable reprint fee, the company/institution may distribute print copies of the Final Published Version externally.

E. GOVERNMENT CONTRACTS

In the case of a Contribution prepared under U.S. Government contract or grant, the U.S. Government may reproduce, without charge, all or portions of the Contribution and may authorize others to do so, for official U.S. Government purposes only, if the U.S. Government contract or grant so requires. (U.S. Government, U.K. Government, and other government employees: see notes at end.)

F. COPYRIGHT NOTICE

The Contributor and the company/institution agree that any and all copies of the Final Published Version or any part thereof distributed or posted by them in print or electronic format as permitted herein will include the notice of copyright as stipulated in the Journal and a full citation to the Journal.

G. CONTRIBUTOR'S REPRESENTATIONS

The Contributor represents that the Contribution is the Contributor’s original work, all individuals identified as Contributors actually contributed to the Contribution, and all individuals who contributed are included. If the Contribution was prepared jointly, the Contributor has informed the co-Contributors of the terms of this Agreement and has obtained their written permission to execute this Agreement on their behalf. The Contribution is submitted only to this Journal and has not been published before, has not been included in another manuscript, and is not currently under consideration or accepted for publication elsewhere. If excerpts from copyrighted works owned by third parties are included, the Contributor shall obtain written permission from the copyright owners for all uses as set forth in the standard permissions form or the Journal’s Author Guidelines, and show credit to the sources in the Contribution. The Contributor also warrants that the Contribution and any submitted Supporting Information contains no libelous or unlawful statements, does not infringe upon the rights (including without limitation the copyright, patent or trademark rights) or the privacy of others, or contain material or instructions that might cause harm or injury. The Contributor further warrants that there are no conflicts of interest relating to the Contribution, except as disclosed. Accordingly, the Contributor represents that the following information shall be clearly identified on the title page of the Contribution: (1) all financial and material support for the research and work; (2) any financial interests the Contributor or any co-Contributors may have in companies or other entities that have an interest in the information in the Contribution or any submitted Supporting Information (e.g., grants, advisory interest in the information in the Contribution or any submitted Supporting Information (e.g., grants, advisory boards, employment, consultancies, contracts, honoraria, royalties, expert testimony, partnerships, or stock ownership); and (3) indication of no such financial interests if appropriate.

H. USE OF INFORMATION

The Contributor acknowledges that, during the term of this Agreement and thereafter, the Owner (and Wiley where Wiley is not the Owner) may process the Contributor’s personal data, including storing or transferring data outside of the country of the Contributor’s residence, in order to process transactions related to this Agreement and to communicate with the Contributor. By entering into this Agreement, the Contributor agrees to the processing of the Contributor’s personal data (and, where applicable, confirms that the Contributor has obtained the permission from all other contributors to process their personal data). Wiley shall comply with all applicable laws, statutes and regulations relating to data protection and privacy and shall process such personal data in accordance with Wiley’s Privacy Policy located at: http://www.wiley.com/WileyCDA/Section/id-301465.html.

[ X ] I agree to the COPYRIGHT TRANSFER AGREEMENT as shown above, consent to execution and delivery of the Copyright Transfer Agreement electronically and agree that an electronic signature shall be given the same legal force as a handwritten signature, and have obtained written permission from all other contributors to execute this Agreement on their behalf.

Contributor's signature (type name here): Joseph Nyangon

Date: 10/20/2016

SELECT FROM OPTIONS BELOW:

[ X ] Contributor-owned work

[ ] U.S. Government work Note to U.S. Government Employees A contribution prepared by a U.S. federal government employee as part of the employee's official duties, or which is an official U.S. Government publication, is called a "U.S. Government work", and is in the public domain in the United States. In such case, Paragraph A.1 will not apply but the Contributor must type his/her name (in the Contributor's signature line) above. Contributor acknowledges that the Contribution will be published in the United States and other countries. If the Contribution was not prepared as part of the employee's duties or is not an official U.S. Government publication, it is not a U.S. Government work.

[ ] U.K. Government work (Crown Copyright) Note to U.K. Government Employees For Crown Copyright this form cannot be completed electronically and should be printed off, signed in the Contributor’s signatures section above by the appropriately authorised individual and returned to

the Journal production editor by email. For production editor contact details please visit the Journal’s the Journal production editor by email. For production editor contact details please visit the Journal’s online author guidelines. The rights in a contribution prepared by an employee of a UK government department, agency or other Crown body as part of his/her official duties, or which is an official government publication, belong to the Crown and must be made available under the terms of the Open Government Licence. Contributors must ensure they comply with departmental regulations and submit the appropriate authorisation to publish. If your status as a government employee legally prevents you from signing this Agreement, please contact the Journal production editor.

[ ] Other Including Other Government work or Non-Governmental Organisation work Note to Non-U.S., Non-U.K. Government Employees or Non-Governmental Organisation Employees For Other Government or Non-Governmental Organisation work this form cannot be completed electronically and should be printed off, signed in the Contributor's signatures section above by the appropriately authorised individual and returned to the Journal production editor by email. For production editor contact details please visit the Journal’s online author guidelines. If you are employed by the Department of Veterans Affairs in Australia, the World Bank, the World Health Organization, the International Monetary Fund, the European Atomic Energy Community, the Jet Propulsion Laboratory at California Institute of Technology, the Asian Development Bank, or are a Canadian Government civil servant, please download a copy of the license agreement from http://olabout.wiley.com/WileyCDA/Section/id-828023.html and return it to the Journal Production Editor. If your status as a government or non-governmental organisation employee legally prevents you from signing this Agreement, please contact the Journal production editor.

Name of Government/Non-Governmental Organisation: ______

[ ] Company/institution owned work (made for hire in the course of employment) For "work made for hire" this form cannot be completed electronically and should be printed off, signed and returned to the Journal production editor by email. For production editor contact details please visit the Journal's online author guidelines. If you are an employee of Amgen, please download a copy of the company addendum from http://olabout.wiley.com/WileyCDA/Section/id-828023.html and return your signed license agreement along with the addendum.

Name of Company/Institution: ______

Authorized Signature of Employer: ______

Date: ______

Signature of Employee: ______

Date: ______