N Melissa M. Krueger Q Thomas L. Mumaw Thomas A. Loquvam Pinnacle West Capital Corporation 400 North 5th Street, MS 8695 Phoenix, Arizona 85004 Tel: (602) 250—3630 Fax: (602) 250-3393 E—Mail: Thomas.MumaW@pinnaclewest.com Melissa.Krueger@pinnaclewest.com Thomas.Loquvam@pinnaclewestcom Attorneys for Arizona Public Service Company BEFORE THE ARIZONA CORPORATION COMNIISSION COMMISSIONERS DOUG LITTLE, Chairman BOB STUMP BOB BURNS TOM FORESE ANDY TOBIN IN THE MATTER OF THE APPLICATION OF UNS ELECTRIC, INC. FOR THE ESTABLISHMENT OF JUST AND REASONABLE RATES AND CHARGES DESIGNED TO REALIZE A REASONABLE RATE OF RETURN ON THE FAIR VALUE OF THE PROPERTIES OF UN S ELECTRIC, INC. DEVOTED TO ITS OPERATIONS THROUGHOUT THE STATE OF ARIZONA, AND FOR RELATED APPROVALS. Cor“ . m DOCKET N O. E—O4204A- 15-0142 ARIZONA PUBLIC SERVICE COMPANY’S NOTICE OF FILING SURREBUTTAL TESTIMONY Arizona Public Service Company provides notice of filing the Surrebuttal Testimony of Ashley C. Brown, Ahmad Faruqui, Charles A. Miessner, and Cory Welch in the above-referenced matter. RESPECTFULLY SUBMITTED this 23rd day of February 2016. By: Melissa M. Thomas L. Mumaw Thomas A. Loquvam Attorneys for Arizona Public Service Company ORIGINAL and thirteen (13) copies of the foregoing filed this 23rd day of February 2016, with: Docket Control ARIZONA CORPORATION COMMISSION 1200 West Washington Street Phoenix, Arizona 85007 COPY of the foregoing mailed/delivered this 23rd day of February 2016 to: Jane L. Rodda Administrative Law Judge Arizona Corporation Commission 1200 W. Washin ton Phoenix, AZ 85 07 Janice Alward Legal Division Anzona Corporation Commission 1200 W. Washin ton Phoenix, AZ 85 07 Brian Smith Bridget Humphrey Legal Division Arizona Corporation Commission 1200 W. Washin ton Phoenix, AZ 85 07 Thomas Broderick Utilities Division Arizona Corporation Commission 1200 W. Washin ton Phoenix, AZ 85 07 Bradley S. Carroll UNS Electric, Inc. 88 East Broadway Blvd. Mail Stop HQE910 PO. Box 711 Tucson, AZ 85702 Chapman Sulp ur Springs Valley Electric Cooperative, Inc. 31 1 E. Wilcox Sierra Vista, AZ 85650 Steve W. Chriss Wal—Mart Stores, Inc. 2011 SE. 10th Street Bentonville, AR 72716-0550 C. Webb Crockett Patrick J. Black Fennemore Crai , PC. 2394 East Came back Road, Suite 600 Phoenix, AZ 85016 Attorneys for Freeport—McMoRan and AECC Jeffrey W. Crockett Crockett Law Group PLLC 2198 E. Camelback Road, Suite 305 Phoenix, AZ 85016-4747 Attorney for SSVEC Kenneth R. Saline K.R. Saline & Associates, PLC 160 N. Pasadena, Suite 101 Mesa, AZ 85201 Michael W. Patten Jason D. Gellman Snell & Wilmer L.L.P. One Arizona Center 400 E. Van Buren Street, Suite 1900 Phoenix, AZ 85004-2202 Attorneys for UNS Electric Rick Gilliam Vote Solar 1120 Pearl Street, Suite 200 Boulder, CO 80302 Meghan Grabel Osborn Maledon, PA. 2929 North Central Avenue Phoenix, AZ 85012 Attorney for AIC Garry D. Ha s The Law Of ices of Garry D. Hays, PC 2198 E. Camelback Rd., Suite 305 Phoenix, AZ 85016 Attorney for ADSA Katie A. Dittelberger Earthjustice 633 17th Street, Suite 1600 Denver, CO 80202 Attorneys for Vote Solar Kevin C. Higgins Ener y Strategies, LLC 215 State Street, Suite 200 Salt Lake City, UT 84111 Timothy M. Ho an Arizona Center or Law in the Public Interest 514 W. Roosevelt Street Phoenix, AZ 85003 Attorney for ACAA, Vote Solar, SWEEP & WRA Mark Holohan, Chairman AriSEIA 2122 W. Lone Cactus Drive, Suite 2 Phoenix, AZ 85027 Briana Kobor Vote Solar 360 22nd Street, Suite 730 Oakland, CA 94612 Eric J. Lacey Stone Mattheis Xenopoulos & Brew, PC 1025 Thomas Jefferson St., NW, 8th Fl. West Tower Washington, DC 20007-5201 Attorney for Nucor Craig A. Marks Craig A. Marks, PLC 10645 N. Tatum Blvd., Ste. 200-676 Phoenix, AZ 85028 Attorney for AURA Robert J. Metli Munger Chadwick, P.L.C. 2398 East Camelback Road, Suite 240 Phoenix, AZ 85016 Attorney for Nucor Jay I. Moyes Jason Y. Moyes Moyes Sellers & Hendricks Ltd. 1850 N. Central Ave., Suite 1100 Phoenix, AZ 85004 Attorneys for FPAA Cynthia Zwick Arizona Community Action Association 2700 N. 3rd Street, Suite 3040 Phoenix, AZ 85004 Vincent N itido Trico Electric Cooperative, Inc. 8600 W. Tangerine Road Marana, AZ 85653 Jill Tauber Chinyere A. Osuala Earthjustice Washington DC Office 1625 Massachusetts Avenue, NW, Suite 702 Washington, DC 20036—2212 Daniel W. Pozefsky Chief Counsel RUCO 1110 W. Washin ton, Suite 220 Phoenix, AZ 85 07 Pat Quinn AURA 5521 E. Cholla St. Scottsdale, AZ 85254 Court Rich Rose Law Group pc 7144 E. Stetson Drive, Suite 300 Scottsdale, AZ 85251 Attorney for TASC Lawrence V. Robertson, Jr. Attorney At Law 2247 E. Frontage Road, Suite 1 PO. Box 1448 Tubac, AZ 85646 Attorney for Noble Americas Gary Yaquinto Arizona Investment Council 2100 N. Central Avenue, Suite 210 Phoenix, AZ 85004 Ken Wilson Western Resource Advocates 2260 Baseline Road, Suite 200 Boulder, CO 80302 Ellen Zuckerman SWEEP Senior Associate 4231 E. Catalina Drive Phoenix, AZ 85018 Jeff Schlegel SWEEP Arizona Representative 1167 W. Samalayuca Dr. Tucson, AZ 85704—3224 Scott Wakefield Hienton & urry, P.L.L.C. 5045 N. 12 Street, Suite 110 Phoenix, AZ 85014—3302 Attorney for Wal-Mart Timothy J. Sabo Snell & Wilmer L.L.P. One Arizona Center 400 E. Van Buren Street Phoenix, AZ 85004 Attorney for Trico SURREBUTTAL TESTIMONY OF ASHLEY C. BROWN On Behalf of Arizona Public Service Company Docket No. E-04204A-15-0142 February 23, 2016 Table of Contents INTRODUCTION ............................................................................................................. .. 1 II. THE EFFECT UNSE’S PROPOSAL WILL HAVE ON LOW INCOME CUSTOMERS ................................................................................................................... .. 5 III. THE FUTURE OF SOLAR AFTER UNSE’S PROPOSAL ............................................ .. 9 IV. ARGUMENTS FOR DELAY AND INACTION SHOULD BE REJECTED ............... .. 26 UNSE’S PROPOSAL FOR A RENEWABLE CREDIT RATE IS AN APPROPRIATE WAY TO COMPENSATE CUSTOMERS FOR ENERGY EXPORTED TO THE GRID. .................................................................. ..................... .. 35 VI. 10 11 12 13 14 15 16 17 18 19 2O 21 22 23 24 25 26 27 28 CONCLUSION ............................................................................................................... .. 44 SURREBUTTAL TESTIMONY OF ASHLEY C. BROWN ON BEHALF OF ARIZONA PUBLIC SERVICE COMPANY (Docket No. E-04204A-15-0142) INTRODUCTION PLEASE STATE YOUR NAME, OCCUPATION, AND ADDRESS. My name is Ashley C. Brown. I am Executive Director of the Harvard Electricity Policy Group (HEPG) at the Harvard Kennedy School, at Harvard University. HEPG is a “think tank” on electricity policy, including pricing, market rules, regulation, environmental and social considerations. HEPG, as an institution, never takes a position on policy matters, so my testimony today represents solely my opinion, and not that of 10 the HEPG or any other organization with which I may be affiliated. 11 PLEASE DESCRIBE YOUR PROFESSIONAL QUALIFICATIONS. 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 I am an attorney with extensive experience in infrastructure, especially energy and regulatory matters. I served 10 years as a Commissioner of the Public Utilities Commission of Ohio (1983—1993), where I was appointed and re-appointed by Democratic Governor Richard Celeste. I also served as a member of the NARUC Executive Committee and as Chair of the NARUC Committee on Electricity. I was a member of the Advisory Board of the Electric Power Research Institute. I was also appointed by the US. Environmental Protection Agency as a member of the Advisory Committee on Implementation of Clean Air Act Amendments of 1990, where I served on the subcommittee charged with implementing emissions trading. I am also a past member of the Boards of Directors of the National Regulatory Research Institute and the Center for Clean Air Policy. I have served on the Boards of Oglethorpe Power Corporation, Entegra Power Group, and e-Curve, and as Chair of the Municipal Light Advisory Board in Belmont, MA. I serve on the Editorial Advisory Board of the Electricity Journal. I have been at Harvard continuously since 1993. During that time I have also been N Senior Consultant at the firm of RCG/Hagler, Bailly, Inc. and have been Of Counsel to the law firms of Dewey & LeBouef and Greenberg Traurig. I have also taught in training programs for regulators at Michigan State University, University of Florida, and New Mexico State University (the three NARUC sanctioned training programs for regulators), as well as at Harvard, the European Union School of Regulation, and a number of other universities throughout the world. I have advised the World Bank and the Inter-American Development Banks on energy regulation and have advised governments and regulators in more than 25 countries around the world, including 10 Brazil, Argentina, Chile, South Africa, Costa Rica, Zambia, Tanzania, Namibia, Ghana, 11 Mozambique, Hungary, Ukraine, Russia, India, Bangladesh, Saudi Arabia, Indonesia, 12 and the Philippines. I have written numerous journal articles and chapters in books on 13 electricity markets and regulation, and I am co-author of the World Bank’s Handbook 14 for Evaluating Infrastructure Regulatory Systems. 15 16 17 18 I hold a BS. from Bowling Green State University, an MA. from the University of Cincinnati, and a J .D. from the University of Dayton. I have also completed all work, except for the dissertation, on a Ph.D. from New York University. My current CV is provided as Attachment ACB-lSR. 19 20 21 22 TESTIFIED PREVIOUSLY YOU HAVE CORPORATION COMIVIISSION? BEFORE 25 26 27 28 ARIZONA No. I have testified, however, before FERC and various state commissions as well as before Congressional and state legislative committees. 23 24 THE ON WHOSE BEHALF DO YOU OFFER TESTIMONY? On behalf of the Arizona Public Service Company. 2 Q. WHAT IS THE PURPOSE OF YOUR TESTIMONY? A. The purpose of my testimony is to rebut objections to UNS Electric’s (UNSE) proposed 3 rate design for solar DG customers. This includes addressing issues discussed in the 4 testimony of TASC witness Mark Fulmer, Vote Solar witness Briana Kobor, and 5 Western Resource Advocates witness Ken Wilson about Retail Net Metering, the value 6 of solar, and what is in the long-term interests of solar as a technology and of the 7 Arizona economy. 8 9 My testimony will be organized as follows: 10 o I will begin by examining and refuting suggestions by witnesses Kobor and 11 Wilson that the proposed changes might somehow harm low income customers, 12 showing that what hurts low income customers is the current net energy metering 13 pricing system; 14 o In the next section of my testimony, I will address claims about the likely impact 15 of UNSE’s proposal on the future of solar power in Arizona. Solar witnesses 16 argue the proposed reform is bad for the future of solar in the state; I will show 17 that the proposed reform is in the true long-term interest of solar energy and 18 solar DG customers (as opposed to the short-term rent seeking of current big 19 solar DG installation firms, given their subsidy-based business model); and that 20 there is no reasonable ground for the argument that the reform will have an 21 overall negative impact on jobs. In fact, there is good evidence that the existing 22 policy has a significant negative impact on jobs and on the Arizona economy as 23 24 a whole; 25 o I will go on to examine various arguments for delay and inaction presented by 26 Ms. Kobor and Mr. Fulmer, showing why none of these arguments present an 27 adequate reason to continue with the existing unjust and inefficient net energy 28 metering system, especially given that the difficulty of changing this system increases the longer it is allowed to remain in place; and 0 Finally, I will refute arguments presented by witnesses Fulmer and Kobor against UNSE’s proposal for a renewable credit rate, and show that the UNSE proposal is not only reasonable, but introduces healthy market discipline to establishing fair compensation for DG energy. ARE THERE OTHER RECURRING PROBLEMS WITH THE ARGUNIENTS PRESENTED IN OPPOSITION TO UNSE’S PROPOSED SOLAR DG RATE REVISION THAT SHOULD BE HIGHLIGHTED? 10 Yes. There are three recurring flaws with the testimony. 11 First, a lack of evidence, and failure to assume any responsibility for proving, or even 12 establishing the plausibility of, their assertions. The testimony provided is often more of 13 the rhetorical ■ourish one might expect in a political campaign, rather than the type of 14 thoughtful and evidence-based analysis appropriate for a regulatory commission. 15 16 Second, one-dimensional thinking. Witnesses frequently selectively present one piece of 17 a whole picture, without acknowledging real effects that go in different, and often polar 18 opposite, directions. 19 Third, their arguments presuppose that solar DG, unlike any other resource in the state’s 20 portfolio, including other renewables, is entitled to be compensated at retail, rather than 21 wholesale rates, that it should be insulated from cost and market pressures that discipline 22 the prices of all other resources, and that any policy that delivers less than that 23 privileged position in the marketplace is unduly discriminatory. As a corollary here, 24 witnesses Ms. Kobor and Mr. Fulmer seem to assume that pricing and public policy in 25 regard to solar should be judged entirely by one criterion: how much solar DG is sold. 26 As I will argue below, they give no adequate reason why regulators should embrace 27 such a self-serving, myopic View of public policy. Legitimate regulatory objectives, 28 4 such as efficient markets, fair and reasonable prices for consumers, incentives for productivity and efficiency gains, enabling effective competition between resources, even the long term economic viability of solar DG, all seem to give way, in the testimony of Ms. Kobor and Mr. Fulmer, to the single minded objective of selling solar DG. II. THE EFFECT UN SE’S PROPOSAL WILL HAVE ON LOW INCOME CUSTOMERS VOTE SOLAR WITNESS BRIANA KOBOR EXPRESSES CONCERN THAT UNSE’S PROPOSAL MAY HURT LOW INCOME CUSTOMERS. PLEASE RESPOND. I disagree with the statement that Ms. Kobor makes on page 40 of her testimony. It is 10 11 unsupported and speculative. And upon review, it is clear that the opposite is true. The current rate design and net metering tariff overly—subsidizes rooftop solar, and, in the 12 aggregate, transfers wealth from less af■uent to more af■uent customers. 13 14 15 16 17 18 19 20 21 22 23 HOW DO ROOFTOP CUSTOMERS? SOLAR SUBSIDIES HURT LOW INCOlVIE Higher income customers are more likely to install rooftop solar, and all other customers, including low income customers, pay the subsidies in question in the form of higher rates.1 This is, in effect, a wealth transfer from lower income customers to higher income customers. All available analysis indicates that the cross-subsidies inherent in the current suite of net metering and volumetric rate design subsidies transfer wealth from low income customers to high income customers. A 2013 study by E3 Consulting of net metering in California found that the median income of net metering customers was 168% of the median California household income—and the system as whole was projected to see another $1.1 billion annually in costs by 2020—costs, which would 24 25 26 27 28 1 Low income customers lack the capital to invest in solar themselves, are less likely to live in a dwelling whose roofs they own, and, do not meet the stringent credit requirements solar DG lessors impose on customers. Moreover, even where low or fixed income households do own their own homes, many, particularly seniors, cannot accept the limitations solar lessors impose on selling their homes. Thus, in effect, low income people are almost systematically unable to participate in the solar DG market. It is the almost exclusive domain of more af■uent households. have to be borne by those (on average, poorer) households not participating in net metering.2 The Center for American Progress has also done some recent work on this issue, looking at median-income data in relation to solar installation patterns from Maryland, Massachusetts, and New York and (in a separate article) in California, Arizona, and New Jersey. Although the main conclusion they emphasize is that solar installations are not limited to areas with predominantly “rich” households, there is a clear and important pattern in the data they show of few or no solar installations among areas with the lowest income households, and relatively many among the highest— income households.3 10 11 12 13 14 15 16 In specific regard to Arizona, the Staff of the Arizona Corporation Commission itself found, based on a review of the locations of customer DG installations within APS service territory, that there “may be a tendency for DG systems to be located in areas of higher income” in their analysis of the APS net metering proposal.4 Low income customers are beginning to notice and object to the financial burden they are bearing to support better—off households. In my own state of Massachusetts, low income customers have recently filed a petition seeking relief from having to subsidize solar DG 17 customers.5 18 19 20 21 22 23 24 25 26 27 28 2 Energy and Environmental Economics, California Net Energy Metering Ratepayer Impacts Evaluation. Prepared for the California Public Utilities Commission by Energy and Environmental Economics (October 28, 2013). It should be acknowledged here that the cross-subsidy impact of net metering in California from lower—income to higher—income customers is strengthened considerably by Califomia’s tiered rate system, under which the highest-consuming customers have the greatest financial motivation to install solar DG systems. However, I note that Ms. Kobor opposes UNSE’s proposal to eliminate their highest rate tier (see Kobor Testimony at pp. 63—64), making the California cost-shift information relevant to this discussion. 3 Hernandez, Mari. “Rooftop Solar Adoption in Emerging Residential Markets.” Center for American Progress, May 29, 2014 and Hernandez, Mari, “Solar Power and the People: The Rise of Rooftop Solar Among the Middle Class.” Center for American Progress, October 21, 2013. 4 See Arizona Corporation Commission Docket No. E—01345A-l3-O248, September 30, 2013, memo titled “Arizona Public Service Company — Application for Approval of Net Metering Cost Shift Solution.” 5 Petition of the Low-Income Weatherization and Fuel Assistance Program Network to Apply G.L. c. 164, sec. 141, submitted to the Commonwealth of Massachusetts Department of Public Utilities, November 17, 2015. National Grid Rate Case, D.P.U. 15-155. As a former legal services lawyer, I find it troubling, to say the least, that we condone the continued existence of a tariff that consciously and deliberately forces lower income households to subsidize higher income households. I can see no justification for such an economically regressive policy. IS MS. KOBOR CORRECT THAT UNSE’S PROPOSAL MIGHT CREATE A SLIPPERY SLOPE, AT THE END OF WHICH LOW INCOME CUSTONIERS ARE HARMED? The slippery slope argument is a red herring. Indeed, the real harm to low income customers, as already noted, is by perpetuating net metering. In general economic \DO \]O\Ul theory, of course, cross-subsidies are best avoided, but there may be circumstances 10 where they cannot be. Thus, it is undeniable that some are embedded in tariffs, many of 11 them inadvertent and/or economically insignificant, but also some that result from 12 conscious policy decisions. Each one must be judged on its own merits and be narrowly 13 targeted to meet a clearly articulated policy objective, and to do so in a way that neither 14 asymmetrically in■ates profits to particular actors in the marketplace at the consumers’ 15 expense, unduly dilutes price signals, renders markets less efficient, nor provides 16 perverse incentives that discourage attainable productivity gains. Thus, cross subsidies 17 designed to assure universal service, such as those supporting rural electrification or 18 assisting low income households, support well-articulated policy objectives. They are 19 generally designed to avoid the pitfalls noted, and are subject to regulatory oversight and 20 review, as well as potential reformulation, to make certain that they continue to be 21 effective in changing circumstances and that they do not have adverse social effects. In 22 short, each cross—subsidy, and whether it needs to be retained and/or modified, stands on 23 its own. 24 25 Thus, net metering must stand or fall on its own merits, not in the context of other cross- 26 subsidies. The policy developed in a time when meters were dumb, energy price signals 27 were less precise and solar panels cost far more than they do today, when the tax 28 1 subsidies were less certain, where storage technology was just a dream, where the social 2 effects were largely unknown, and when solar DG market penetration was so small that 3 price 4 dramatically, and net metering needs to be reassessed in its own context, and with 5 reference to the standards I just mentioned. It can and should be done without regard to 6 what other cross-subsidies may or may not exist. The slippery slope mentioned by Ms. 7 Kobor simply does not exist. distortions were insignificant. All of those circumstances have changed 8 9 Q. WRA WITNESS KEN WILSON ALSO ARGUES THAT A DEMAND CHARGE COULD HARM LOW INCOME CUSTOMERS. IS HE CORRECT? 10 A. No. First, Mr. Wilson fails to recognize that demand charges do not increase rates. They 11 are revenue neutral since the demand costs are already embedded in tariffs. What 12 demand charges do is make those costs transparent, and by doing so, enable all 13 customers, low income included, to shape their demand in ways that can reduce their 14 bill. It also provides an opportunity for programs like LIHEAP6 to design their low 15 income subsidies to capture that increased opportunity for saving. 16 Second, Mr. Wilson argues generally that low income customers would be penalized 17 because their less efficient appliances cause higher loads.7 He offers no empirical 18 evidence to support this 19 users, having fewer electrical appliances, have smaller loads than other customers. The :(1) future of solar after UNSE’s proposal. seems more likely, in the aggregate, that low income 22 23 24 25 26 27 28 6 LII-IEAP — Low Income Home Energy Assistance Program. 7 Wilson Direct Testimony at p. 9. III. THE FUTURE OF SOLAR AFTER UNSE’S PROPOSAL TASC WITNESS MARC FULlVIER SUGGESTS THE “STIFLE” DG SOLAR; MS. KOBOR SAYS THAT IT CURTAIL FUTURE DG GROWTH.” WHAT, IN IMPACT OF THE PROPOSAL ON THE FUTURE OF UNSE PROPOSAL WILL WOULD “VERY LIKELY YOUR VIEW, IS THE SOLAR IN AZ? To assess the possible impacts of the UNSE proposal on the future of solar in Arizona, it is important to first disentangle four things that Mr. Fulmer and Ms. Kobor con■ate: the financial interests of solar installation companies, the financial interests of solar DG customers, the long—term prospects for solar DG in the electricity markets, and, of course, the public interest. 10 11 WHY IS IT IMPORTANT TO DISTINGUISH BETWEEN THE FINANCIAL AND THE COMPANIES OF SOLAR INSTALLATION INTERESTS FINANCIAL INTERESTS OF SOLAR DG CUSTOlVIERS? 12 Advocates for large DG solar installation companies, such as Mr. Fulmer, simplistically 13 assume that the interests of solar customers and solar companies are aligned with one 14 another. For example, in his December testimony, Mr. Fulmer raises the issue of the 15 impact UNSE’s proposed rate change will have on PV adoption. His testimony takes the 16 cost of solar DG installation as a given, and then asks what electricity price would be 17 needed to support investment in solar DG.8 The presumption is that if the end 18 customer’s payback is not rapid and guaranteed, then the proposed electricity price is 19 not adequate to support solar DG. 20 His focus on recovery of customer investment obscures the tension between the interests 21 of solar DG customers and solar installers. Prospective solar customers are looking for 22 cost effective means of meeting their need for electricity. Many of them also are 23 motivated by a public-spirited desire to increase the efficiency and “greenness” of the 24 electricity system as a whole. These aims, however, are not at all technology specific, so 25 going solar is but one option, not 26 27 28 8 Fulmer Direct Testimony at p. 15. option, as Ms. Kobor and Mr. Fulmer seem to I assume. It may not even be the most desirable option, and, as shown in this testimony 2 and elsewhere, it is probably never the most cost effective option. 3 4 Q. CAN YOU EXPLAIN WHERE THE INTERESTS OF LARGE ROOFTOP SOLAR COMPANIES DIVERGE FROM CUSTOMER INTERESTS? A. The interests of large rooftop solar companies and utility customers diverge on several 5 levels. In regard to solar leases, the most cost effective installation for customers may be 6 quite different than what is most profitable to the lessor. Solar DG companies, perhaps 7 lessors even more than vendors, routinely make false sales representations about their 8 products and benefits they offer, often retain the RECs and SRECs for their own use, as 9 opposed to the use the customer may prefer, and offer lease agreements whose terms are :(1) notoriously onerous for customers.9 12 However, the primary divergence of interests relates to the massive profits made by 13 these companies—profits that are effectively paid by utility customers. Solar DG profits 14 by using the highly in■ated and heavily subsidized net meter prices to shield themselves 15 from market pressure and having to pass on the declining cost of solar panels, whereas 16 the consumer interest is better served by having solar DG vendors/lessors be subject to 17 the marketplace or regulatory disciplines to which all other players in the energy market 18 are subject. Simply stated, it would be better if solar DG developers had to compete with 19 other forms of energy, rather than chasing just subsidies for themselves. 20 Thus, the customers’ energy efficiency objectives may require price signals and self21 generation options that con■ict with the objectives of solar DG developers. As seen in 22 Ms. Kobor’s and Mr. Fulmer’s testimony, solar DG providers oppose demand charges 23 and other types of pricing that would enable new energy service providers and vendors 24 to offer consumers products to reduce their energy bills. They are committed to :: maintaining barriers to new entrants who might offer valuable products and services that 27 9 A number of these practices are under active investigation by consumer protection agencies in a variety of jurisdictions across the US, including Arizona. 28 10 provide customers with more options, preferring to limit the diversity of products and services found in a robust market, so as to even further reduce the competition for solar DG. Solar DG interests also generally oppose pricing regimes that would make solar DG units more ■exible, more attuned to system requirements, and more efficient, because they prefer the simplicity of selling primitive products without having to respond to price signals for a better, more efficient product, something that would add considerable value for customers, but require a more modern and sophisticated business model than simply peddling simple installations at in■ated prices. Finally, consumers benefit from an efficient electricity market with competitive discipline, while solar DG 10 companies are self-interested in preserving artificially high prices for their products at 11 considerable cost to consumers, solar and non—solar alike. 12 Q. IT IS NORMAL FOR COMPANIES TO MAKE PROFITS, SO WHY IS IT APPROPRIATE TO SEPARATELY CONSIDER PROFITS MADE BY LARGE ROOFTOP SOLAR COMPANIES? A. The problem is two—fold: how the profits are derived and the magnitude of the profits 13 14 15 made by large rooftop solar companies. If the profits were derived from being more 16 efficient, making productivity gains, and reducing costs, those profits would have been 17 earned and deserved. For solar DG companies, however, as revealed in SolarCity’s most 18 recent 10K filing with the SEC10 (discussed below), the profits are derived by chasing 19 subsidies and cross subsidies, including having the price to beat, unlike every other 20 energy source, be retail rather wholesale. Beyond that, of course, since the price they 21 compete against is the bundled, monopoly retail rate,11 they have the advantage of being 22 paid a monopoly price without being subject to the discipline of cost based regulatory 23 oversight. 24 25 26 27 10 SolarCity Corp 10K, filed 2/24/15 for period ending 12/31/ 14, (available at http://files.shareholder.com/downloads/AMDA-14LQRE/144512701 1x0xSlS64590—15897/1408356/filing.pdf). 11 The retail price, they “compete” against, of course, includes compensation for a host of goods and services they do not provide, and re■ects the benefits for them that are associated with reallocating revenue responsibility of a significant share of system costs to non-solar customers. 28 11 1 The second issue is the magnitude of profits that, until now, have gone largely 2 unexamined. It seems that, so far, large solar DG companies have been quite effective in 3 preserving large margins for themselves when making solar DG installations. As shown 4 in the testimony of APS witness Cory Welch, filed concurrently with this testimony, 5 rooftop solar leasing companies obtain an average of 40% margins on each installation 6 in UNSE’s service territory.12 This is an astonishing return, particularly in the monopoly 7 context described above. Policy makers and regulators would be well advised to see the 8 issues of subsidies and cross subsidies in the context of such high profits provided to 9 rooftop solar companies. The UNSE proposal for pricing solar DG constitutes a very 10 reasonable way of restoring a marketplace discipline to the pricing and a fairer way of 11 allocating costs. 12 13 Q. HAS THE ISSUE OF ROOFTOP SOLAR COlVIPANIES’ PROFITS BEEN STUDIED ELSEWHERE? 14 A A recent study by MIT, The Future of Solar, provides a comprehensive overview of the 15 state of solar technology and of the industry as a whole and is very revealing.13 While 16 the authors of the study do not'look specifically at profit levels, they do examine how 17 prices that enable and encourage short term profit taking by solar DG companies have 18 an adverse impact on the long term economic viability of solar energy. As part of this 19 study, the MIT authors present a comparative analysis of the costs and prices associated 20 with solar installations, contrasting utility-scale and DG installations. MIT built up its 21 costs estimates from data on hardware costs, combined with surveys of installers and 22 data about wages, and included costs for customer acquisition, sales taxes, margin, and 23 general and administrative expenses—working to develop an estimate of costs that 24 would be sufficient to sustain the industry. Then they compared their bottom—up costs 25 estimates (per watt of installed capacity) with reported actual prices charged (per watt of 26 27 12 Surrebuttal Testimony of Cory Welch Attachment CJW — 28R. 13 See The Future of Solar Energy: An Interdisciplinary MIT Study. MIT (2015). https://mitei.mit.edu/system/files/MIT%20Future%200f%208olar%20Energy%205tudy_compressed.pdf 28 12 installed capacity). An interesting pattern emerged—for utility-scale PV, MIT’s bottom— up cost estimates were fairly closely matched to actual prices charged for the systems. This was not the case for distributed generation installation of solar. As expected, due to the loss of certain economies of scale, MIT’s bottom—up costs estimate was higher than for utility-scale solar. This finding is unsurprising. The surprise here is that MIT found a considerable gap between its bottom—up cost estimate and actual reported prices charged for solar DG systems. Here, average reported prices charged exceeded MIT’s cost estimate by an average of 50%—a considerable margin.14 The existing subsidy structure (including net metering, as well as federal tax incentives) makes it possible for 10 11 12 13 14 15 16 17 18 19 homeowners to pay these prices while still coming out marginally ahead economically— but the prices they are being charged do not re■ect the best possible value for consumers, and they provide no incentives for productivity and efficiency gains for solar DG. Indeed, they discourage such efforts. It is also re■ective of the fact that while the costs of solar panels themselves have been in dramatic decline, the prices for installing the units have been increasing, thereby enabling the solar DG developers to retain most, if not all, of the benefits of declining panel costs for themselves rather than passing them on to their customers. This is re■ected in a recent study by Lawrence Berkeley National Labs which found that out of six countries it compared to the US. (Germany, Japan, Italy, China, France, and Australia), only France had higher costs for installed 2O residential PV systems. 15 21 22 These developments may be in the short—term interests of residential PV system 23 developers, but it is not in the long-term interest of solar power, whose interests would 24 better served by a pricing regime that encouraged increased productivity, better 25 26 27 14 Id. p. 86. Residential and 15 Barbose, Galen and Naim Darghouth. Tracking the Sun VIII: The Installed Price Laboratory National Berkeley Lawrence States. Non-Residential Photovoltaic Systems in the United (August 2015). 28 13 compatibility with system requirements, and deployment of technology, such as storage and smart inverters, that would better secure solar DG’s place in Arizona’s energy resource portfolio.16 Moreover, it is rate design and retail net metering that enables solar DG lessors to retain most of the margin, and essentially pass on pennies on the dollar to solar DG customers, recovering the balance of their profits from taxpayer funded subsidies plus cross-subsidies paid by non-solar customers. No competitive market or properly regulated market would enable that to happen. Thus, UNSE’s use of large scale solar, procured in a competitive market, as the benchmark for pricing solar DG solves the problems and allows for more of the benefits of declining cost to be passed on to 10 customers. It might also be noted that the UNSE proposal also has the very positive 11 effect of not diverting capital from the more efficient large scale wind and solar 12 renewable energy to the less efficient solar DG, a likely development if net metering is 13 not replaced by a more rational pricing regime. 14 SolarCity’s 10K filing is also quite revealing about its motivations for opposing pricing 15 reforms: 16 17 18 Modifications to the utilities’ peak hour pricing policies or rate design, such as to a ■at rate, would require us to lower the price of our solar energy systems to compete with the price of electricity from the electric grid.17 (emphasis added) 19 This is the acknowledgement by the leading player in the solar DG that its business 20 model is fully dependent on being shielded from competition, hardly a virtue from the 21 standpoint of either consumers or the public interest. The actual cost, or even market 22 valuation, to provide service is not mentioned—it is all about charging as much as 23 possible, depending on the utility rates and existing incentives that will leave no 24 incentive for productivity and will provide marginal benefits, at best, for solar DG 25 26 27 16 See MIT Study. 17 SolarCity Corp 10K, filed 2/24/15 for period ending 12/31/14, p. 11 http://files.shareholder.com/downloads/AMDA—l4LQRE/1445 12701 1x0xS 1564590-15 897/1408356/filing.pdf). 28 14 (available at consumers. In short, net metering enables arbitrarily high prices for consumers, extraction of monopoly rents by solar DG vendors/lessors, and dim prospects for the future of solar DG. WHAT DO THESE LARGE PROFIT MARGINS, AND THE PRICING STRUCTURE UNDERLYING ROOFTOP SOLAR LEASES, MEAN WHEN ASSESSING UNSE’S PROPOSAL? It means that the subsidies supporting rooftop solar can be reduced. Indeed, with the recent Congressional extension of the Investment Tax Credit for solar, cross subsidies can be eliminated, without doing harm to increased market penetration by rooftop solar. 10 Indeed, with increased exposure to market risk, solar DG vendor/lessors would be compelled to reduce their prices and improve their products, developments that would 11 make solar DG much more attractive to the public. This, of course, would require the 12 13 14 solar DG industry to alter its business model of simple rent seeking and subsidy chasing to one of vigorous competition in the market place and to be willing to accept rates of return commensurate with its performance in the marketplace. 15 16 Based on the discussions of undisciplined profit margins above, advocates of net 17 metering are presenting the wrong analysis when they argue that eliminating the retail 18 net metering subsidy would make solar DG economically infeasible for end use 19 customers. With the large surplus margins shown by Mr. Welch and the MIT study built 2O into these prices (on top of normal business margins), there is a strong case to be made 21 that the same DG installations could be provided at considerably lower cost—a cost 22 which might well be affordable within the context of a revised tariff for solar DG 23 customers—and still be profitable for DG vendors/lessors. Indeed, with the UNSE 24 proposed reference price, both vendors and customers would be incentivized to improve 25 both efficiency and productivity, as the savings would accrue to them, but would be 26 earned, as opposed to being the gifts of a severely ■awed pricing methodology. 27 28 15 TASC claims to champion competition and oppose monopoly power, and thereby serve as the consumer’s champion in creating a competitive marketplace. In fact, the reality is exactly the opposite. Their advocacy of net metering in this proceeding and others around the country calls for perpetuation of an inefficient, highly in■ated, price not subject to any competitive or even cost based pressure, a price that can only survive in a non-competitive environment. In effect, they are seeking a market where they are free to sell their product to customers, but where those very same customers have little opportunity to see competitive or cost based pressure on the prices they are compelled to pay for either purchasing solar DG or having to pay the cross-subsidies inherent in net 10 metering. 11 12 13 14 15 16 17 Current rate design and retail net metering enables TASC members and other solar DG vendors/lessors to charge monopoly rents, subject only to the potential of competition from other DG vendors/lessors, who share the same self-interest of preserving arbitrarily high margins. They seek to preserve a business model in which customers are deprived of the pricing benefits associated with either competitive markets or cost based regulation. This is not in the interest of any utility customer, and in particular, the interests of UNSE’s customers. 18 19 20 WHAT ABOUT THE FUTURE OF SOLAR ITSELF? DO THE INTERESTS OF LARGE SOLAR DG INSTALLATION COlVIPANIES THREATEN THE FUTURE OF DISTRIBUTED SOLAR AS A COlVIPETITIVE ENERGY TECHNOLOGY? 21 22 23 24 25 26 27 28 Yes, the short term rent seeking business model of most, if not all, of these companies has created an unsustainable environment in which solar cannot ■ourish in the long run. In the short term, as noted, the current rate benefits the solar industry, because of the inherent wealth transfer from non—solar to solar customers, plus the wealth transfer from all customers to solar DG vendors/lessors discussed above. That is not a sustainable long-term strategy, particularly if significant expansions in solar DG adoption are hoped for. There are two reasons for this. The first is simply that the public appetite for paying 16 higher than necessary prices for goods and services is very limited, and their patience with that reality, once it becomes, as is inevitable, public, is not great. Second, and perhaps even more important in the long term, is the fact that solar, like every other form of energy, must constantly be improving its productivity and overall performance to remain competitive. Thus, to align that reality with the incentives for the solar DG industry, incentives for productivity and efficiency gains should be embedded in rates. \O O\IO\Ul- > Unfortunately, that is precisely the opposite of what occurs under current net metering 10 tariffs. Current tariffs provide absolutely no incentive to improve the performance of a 11 generating resource that already ranks last among renewables in efficiency and cost 12 effectiveness, both in terms of economic efficiency and as a tool for reducing carbon 13 emissions. Any money spent on improving the technology or on storage, under retail net 14 metering, goes to reduce profits, not enhance them. The arbitrarily high, ■at prices 15 permitted by current rate design and retail net metering simply do not incentivize 16 investing in improvements. Indeed, it does exactly the opposite. 17 MS. KOBOR AND MR. FULMER CLAIM THAT ADOPTING UNSE’S PROPOSAL WILL REDUCE THE NUMBER OF SOLAR JOBS. WHAT IS YOUR RESPONSE? 18 19 20 21 22 23 24 25 26 27 First, neither Ms. Kobor nor Mr. Fulmer offer any specific evidence or analysis . regarding UNSE’s service territory, both in terms of the number of current solar jobs or the precise effect of UNSE’s proposal on those jobs. In fact, in a recent proceeding before the Nevada Public Utilities Commission, TASC and others relied on the same Solar Foundation National Solar Jobs Census attached to Mr. Fulmer’s testimony to similarly claim that demand charges proposed by NV Energy would cause a massive reduction in solar jobs. Upon review by Nevada PUC Staff, however, the Nevada Commission rejected the Solar Foundation National Solar Jobs Census, stating that “the figures cannot be reasonably relied upon as an estimate of the number of solar jobs in Nevada or the number of jobs that could potentially be impacted by this Order.”18 The Solar Foundation report should be rejected here for the same reason: there is no Arizona, much less UNSE, specific data that can be relied upon in considering the impact of UNSE’s proposal on solar—related jobs. ARE THERE 'OTHER PROBLEMS WITH MR. KOBOR’S CLAIMS REGARDING SOLAR JOBS? FULMER’S AND MS. Yes, both involve a very one-dimensional look at the economic impact of solar that is severely ■awed and ultimately misleading. Advocates of subsidies for distributed solar 10 11 12 13 14 15 generation often point to the supposed economic benefits—particularly job creation—of rooftop solar installation.19 But claims about a positive impact on job creation are one— sided—they only count new jobs created in solar. They do not even bother to claim that net metering creates more jobs than competitively priced solar DG would. They simply look at solar DG jobs in a compete vacuum and without a real context. Perhaps more importantly, they fail to even consider the broader effect on the economy. 16 17 If the cost of electricity is higher, jobs are likely to be lost elsewhere in the economy— 18 there is no reason to assume that the net job impact of distributed solar power is 19 positive.20 In fact, a recent study by Tim James, Anthony Evans and Lora Mwaniki- 20 Lyman of Arizona State University used an Arizona-specific regional economic model 21 (a REMI model), balancing the costs of installed DG capacity (and related financing 22 23 24 25 26 27 18 Order, Application of Nevada Power Company d/b/a NV Energy for approval of a cost-of-service study and net metering tariffs, Public Utilities Commission of Nevada Docket Nos. 15-07041 & 1507042 at paragraph 229 (December 23, 2015). 19 There may be some question as to the quality of at least some of the jobs, the solar witnesses claim will be gained. In its most recent 10K filing with the SEC, SolarCity, the largest vendor/lessor of solar DG, disclosed that its employments practices are the subject of an active investigation by the US. Department of Labor. 20 The National Solar Jobs Census 2014 attached to Mr. Fulmer’s testimony is a good example of the one-sided solar jobs cheerleading genre. It touts solar employment in isolation from the rest of the economy. 28 18 1 costs) against what APS estimates to be the related savings on generation purchases and 2 generation capacity investment over thirty years (and their related customer savings), 3 based on different levels of investment in solar DG that might be made in the APS 4 service territory.21 This study models the complexity of judging the economic and job 5 impacts of a particular policy or subsidy—of course, there is an immediate positive 6 impact on some jobs from additional solar employment, but, over time, taking into 7 account the effects of lost spending power by consumers who have to pay more for their 8 electricity, the projected impacts on jobs and on the gross state product of the Arizona 9 economy are decidedly negative (for example, the model shows cumulative losses in gross state product, over time, in the multiple billions of dollars).22 10 11 Before leaving the topic of jobs, it is worth remembering that most solar panels sold or leased in the US. are manufactured in China. In all likelihood, more American jobs are 12 associated with other forms of generation. Q. IS THE NUMBER OF SOLAR INSTALLATIONS THE DETERMINATIVE STANDARD AGAINST WHICH UNSE’S PROPOSAL SHOULD BE JUDGED? A. . No, it is not. It is clear from their testimony that both Mr. Fulmer and Ms. Kobor believe 15 16 17 that there is only one dimension by which to judge public policy on solar DG pricing: 18 how much solar DG is sold or leased. The theory they seem to articulate is quite simple: 19 if a tariff provision results in more solar DG, that is good, and if there is any slowdown :(1) in solar DG’S rapid growth, that is bad as a matter of public policy.23 22 23 24 25 26 27 21 As detailed above, I am skeptical about how many of these savings will materialize. 22 Evans, Anthony, Tim James, and Lora Mwaniki-Lyman. “The Economic Impact of Distributed Solar in the APS Service Territory, 2016-2035.” Report, L. William Seidman Research Institute, W.P. Carey 2S3chool of Business, Arizona State University, February 16, 2016. (Attachment ACE-28R). See, e. g., Kobor Direct Testimony at p. 51. It is also curious that neither MS. Kobor nor Mr. Fulmer even bother to ask whether there might be a more efficient pricing methodology under which more solar DG would enter the market. They appear to be wed to the outdated, rent seeking business model of the interests for whom they are testifying to even contemplate new approaches that might advance solar DG out of its niche on the margins into the mainstream of Arizona’s energy resources. 28 19 Public policy and regulatory decision making cannot and must not be as one dimensional as Ms. Kobor and Mr. Fulmer would urge. The point of an electricity rate is to establish a just and reasonable rate, disciplined by the market and/or cost. Such a rate should enable solar DG to compete as a mainstream energy source (even assuming it retains the advantages of federal tax subsidies and renewable energy requirements). If a short—terrn decline in the growth of solar DG results, that is not necessarily a bad outcome for the world—especially if it means investment is instead being directed towards more cost effective technologies, such as improved solar DG products, utilityscale renewables, and/or innovations that allow for more efficient and more productive means of providing solar DG. 10 11 Q. HOW DO YOU RESPOND TO MS. KOBOR AND MR. FULMER’S CLAIMS ABOUT SOLAR INSTALLATIONS PLUMNIETING IN SALT RIVER WERE CHARGES DEMAND AFTER TERRITORY PROJECT’S IMPLEMENTED? A. I don’t think the level of DG installations in SRP’s service territory is at all indicative, or 12 13 14 15 even relevant, in considering the impact of demand charges on solar installation levels. 16 After SRP implemented a demand charge, Ms. Kobor reports that applications for the 17 DG program fell by 95%.24 Mr. Fulmer makes the same point, and provides a chart 18 showing monthly solar DG applications in SRP for three years before and nine months 19 after the change.25 The chart itself shows an interesting pattern not mentioned by Mr. 20 Fulmer or Ms. 21 approximately nine months before the rate change, including a huge spike in the month 22 prior to the change. (A rough estimate based on eyeballing the graph provided by Mr. 23 Fulmer suggests that monthly applications were about ten times the monthly average of 24 the first two years shown.) Kobor—there was a significant 25 26 27 24 Kobor Direct Testimony at p. 39. 25 Fulmer Direct Testimony at p. 17. 28 20 increase in applications in the l A look at the whole picture suggests a more nuanced and complex story than the one 2 told by Mr. Fulmer and Ms. Kobor. As one would expect, the imminent change to a 3 different rate formulation prompted a significant spike in demand—presumably, 4 customers with some intention of installing solar DG in the next couple of years were 5 highly motivated to get their applications in before the rate change. After this spurt of 6 activity, as one would expect, with demand temporarily exhausted, applications dropped 7 significantly. It remains to be seen what will happen once the system has absorbed the 8 demand spike that occurred right before the rate change. It is also important to note that 9 SolarCity has filed an anti-trust suit against SRP for its tariff reforms, alleging that these 10 changes have made it impossible to compete in SRP territory. Thus, SolarCity and 11 others in the solar DG industry have a powerful incentive to essentially boycott SRP 12 customers. To do otherwise would undermine their anti-trust case, since doing business 13 in SRP territory would effectively disprove their allegations of being unable to compete. 14 Thus, the drop in solar DG installations in SRP territory, assuming it is true, could well 15 be a self—fulfilling prophecy by the solar DG industry. 16 Indeed, that self—serving boycotting behavior was also evidenced by Nevada’s very 17 recent 18 Commissioners, and even the Governor, to restore net metering by suspending 19 operations in the state. Given the analysis of the large margins of profitability above, I 20 would suggest that such a move might simply open up the market to new, local 21 22 competitors. 23 Events in SRP’s service territory offer little useful information for another reason: after 24 SRP implemented demand charges, solar installers could literally walk across the street 25 and sell rooftop solar at immense profit margins (60% per system, according to Mr. 26 Welch’s study26) in APS’s service territory. It is not clear why rooftop solar companies 27 28 experience; namely, that large solar 26 Welch Surrebuttal Testimony, Attachment C! W — 28R. 21 installers will attempt to pressure would voluntarily accept lower profits by installing in SRP’s territory when they could continue receiving the (overly) rich subsidies available in APS’s territory with no incremental effort. WHY DO CURRENT RATE DESIGN AND RETAIL NET METERING MAKE THE FUTURE OF SOLAR UNSUSTAINABLE? In the long term, in order to be fully sustainable, solar energy needs to be fully competitive on both a price and qualitative basis. That means both that solar should be competitive on a price basis, independent of any subsidy, and that steps need to be taken to reduce the intermittent and off peak production characteristics of solar (e. g. link it to 10 storage, or use western rather than southern exposure in order to better align production 11 with peak demand). 12 Current rate design and retail net metering is exactly the wrong incentive. They simply 13 throw utility customer money at distributed solar in its most inefficient and primitive 14 form. Retail net metering not only fails to incent increases in productivity, but actually 15 discourages them. It makes solar artificially more profitable when companies refrain 16 from investing in technological development or taking other steps to improve 17 productivity. Under markets driven by competitive forces, by contrast, investments in 18 technological innovation increase the ability of companies to compete, and thus offer a 19 positive rate of return that justifies that initial investment. What is critical to understand 2o is that net metering, regardless of its profitability for solar DG vendors/lessors, is a 21 subsidy so poorly designed that it actually runs contrary to the long run economic 22 viability of distributed solar energy. 23 24 HOW MIGHT UNSE’S PROPOSAL AFFECT ENERGY IN ARIZONA? THE FUTURE OF SOLAR 25 UNSE’s proposal is an important, if not critical, first step to preserving the future of 26 solar in Arizona. It is true that the proposed rate reform is likely to compel solar 27 vendors/lessors to change their business model from chasing subsidies to competing 28 22 (something they are loathe to do because they are ■ourishing in a much too cozy environment at present). These companies will have to decide whether to change their model or not, and if not, some may well seek to move their model into other jurisdictions that may continue to shield them from market and regulatory pressures that provide greater opportunity for consumers. The world in which that is happening, fortunately, is changing, and the rooftop solar industry, just as every other segment of the energy sector, will have to become more competitive. If some players refuse to adapt, new players will emerge who will see business 10 11 12 13 14 15 16 opportunities and thrive on the challenge of well-functioning markets as opposed to extracting subsidies. In the long run, solar energy will have a much brighter future and will be better assured of finding its place in the mainstream of energy resources. The dire predictions of Ms. Kobor and Mr. Fulmer should be treated with a great deal of skepticism. CONTRARY TO THE TESTIMONY OF MS. KOBOR AND MR. FULMER, HOW MIGHT DEMAND CHARGES ACTUALLY HELP THE FUTURE OF SOLAR? 17 The demand charges proposed by UNSE provide price signals that will inevitably 18 enhance the productivity and efficiency of solar DG. What the fixed charge proposal of 19 UNSE does do is to promote overall system efficiency by tying rates and cost causality 20 more closely together so that marginal rates better re■ect actual marginal costs while the 21 fixed rates recover unavoidable fixed costs. This improves the price signals to 22 customers, reduces the degree to which cross subsidies are built into rates (including 23 those that ■ow from non—solar to solar customers), and makes the actual market value of 24 solar DG and energy efficiency more transparent. In short, the result of both the change 25 in fixed costs and the adoption of demand charges for solar DG customers is to insert the 26 disciplines of market and cost that have been lacking in the past. In specific regard to 27 28 23 1 solar DG, the UNSE proposals will provide price signals that will enable solar DG 2 installations to operate optimally for both the solar host and the system as a whole. 3 4 What is important is that customers cannot respond to signals to shape their load in more efficient ways unless they are given price signals to do so. A recent study by the Rocky 5 6 7 8 9 Mountain Institute (RMI) urges that demand rates be part of residential bills because they are inherent to the overall objective of energy efficiency.27 Indeed, RMI coins a phrase, “■exiwatts,” to describe the services and technology that exist to fill the business space demand charges will offer. A recent RMI blog post hails demand charges as an opportunity for new technologies, customer options, and reduced grid costs: 10 11 12 13 14 15 16 Demand charges are a promising step in the direction of more sophisticated rate structures that incent optimal deployment and grid integration of customer-sited DERs. A demand charge more equitably charges customers for their impact on the grid, can reward DG customers with bill savings, and opens up potential for an improved customer experience using load management tools. It can also benefit all customers through reduced infrastructure investment and better integration of renewable, distributed generation.28 Similarly, a joint statement by the National Resources Defense Council and the Edison Electric Institute endorses the use of demand charges.29 17 18 The Natural Resources Defense Council also supported demand charges in a recent 19 filing with the California Public Utilities Commission.30 Those positions, as well as 20 those taken by UNSE and the ACC Staff in this proceeding, are an excellent indication 21 that there are a wide variety of parties and interests who see demand charges as an 22 23 24 25 26 27 28 27 Lehrman, Matt. “Are Residential Demand Charges the Next Big Thing in Electricity Rate Design?” Blog Post, RMI Outlet (May 21, 2015). http://blogrmiorg/blog 2015 05 21 residential demand charges next big thing in electricity rate design 8 Id. 29 EEI and NRDC, “EEI/NRDC Joint Statement to State Utility Regulators,” February 12, 2014 (http://docs.nrdc .org/energy/files/ene_14021 101a.pdf). 3° Proposal of the National Resources Defense Council (NRDC) in Determining a Net Energy Metering Successor Standard Contract or Tariff, filed August 3, 2015 in Rulemaking 14—07-002 before the Public Utilities Commission of California. 1 important element of the efficient pricing of electricity. Demand charges work to smooth 2 customer demand, reducing spikiness, and increasing the utility’s ability to rely on more 3 efficient resources, rather than turning to its last—resort, less energy efficient sources of 4 generation. 5 One might expect that solar industry interests would see the value in price signals that 6 would enable customers to shape their load in a manner that is both economically 7 efficient and environmentally desirable. Instead of seeing the opportunities for solar 8 energy to help customers shape their loads in more beneficial ways, witnesses Mr. 9 Fulmer and Ms. Kobor oppose innovation and efficiency by defending and clinging to 10 an inefficient and outdated business model of chasing subsidies for the most primitive 11 and inefficient use of solar energy. Once again, it appears that Ms. Kobor and Mr. 12 Fulmer seek to protect the short term profitability of their clients instead of looking out 13 for the long term future of solar energy and for a more efficient and more :: environmentally friendly electricity marketplace. Q. ARE THERE OTHER EXAMPLES OF HOW TASC AND VOTE SOLAR’S POSITIONS WILL HARM THE FUTURE OF SOLAR? A. Mr. Fulmer and Ms. Kobor’s apparent opposition to “enabling technologies” that could 16 17 18 help customers manage their demand is another example of how their position 19 disregards the long-term interests of solar energy. Ms. Kobor’s argument is particularly 20 ironic. Her main objection is that such technologies are “uncommon, costly to 21 implement, and have not achieved widespread adoption.”31 It is ironic that she raises this 22 as an objection to changing net metering rates, since it is exactly these kinds of archaic 23 rates that make it hard for smart enabling technologies to take hold. Current net metering 24 rates provide zero incentive for customers to invest in such technologies. It is hardly fair :2 to complain that they have not been widely adopted while defending a rate structure that 27 28 31 Kobor Direct Testimony at p. 35. 25 prevents customers from realizing any savings by using these technologies. One of the important benefits of UNSE’s proposed revision to rates for DG customers would be that it should contribute to the development of these kinds of technologies. Indeed, it is fair to say that not only is their position contrary to solar, but by opposing tariff changes, such as demand charges, they are effectively precluding innovations that would enable providers of sophisticated energy services to enter the market. That is \l because such companies depend on price signals to optimize the use of “negawatts” and “■exiwatts.” Transparent and meaningful unbundled prices will enrich the marketplace, provide more options for consumers, and help to optimize the role that solar DG can 10 play. It is ironic that, instead of seeing that as an opportunity for the solar DG industry, 11 or a benefit for customers, the two witnesses argue for a “dumbed down” marketplace, 12 the effect of which is to reduce the amount of goods and services, restrict market entry 13 for otherwise valuable market participants, 14 optimizing solar DG. 15 16 17 18 and, ironically provide channels for IV. ARGUMENTS FOR DELAY AND INACTION SHOULD BE REJECTED HOW DO MR. FULMER AND MS. KOBOR ARGUE COMMISSION SHOULD DELAY OR TAKE NO ACTION RETAIL NET lVIETERING? THAT THE REGARDING 19 Mr. Fulmer and Ms. Kobor provide a variety of arguments that urge no action. These 20 include that (i) an unbiased “Value of Solar” study should be performed before any 21 action is taken; (ii) rooftop solar customers only account for 2% of cost shifting in 22 UNSE’s service territory and that other, larger, cost shifts exist; (iii) UNSE has not 23 performed specific studies on certain topics; (iv) rooftop solar is needed for UNSE to 24 achieve compliance with the distributed generation component of Arizona’s Renewable 25 Energy Standard; (v) UNSE’s proposal would violate Arizona net metering rules; and 26 (Vi) demand charges are not part of the modern trend in rate design. I address each in 27 turn. 28 26 Q. IS AN UNBIASED “VALUE OF SOLAR” STUDY NEEDED TO ADEQUATELY ASSESS UNSE’S PROPOSAL? A. No, it is not. MS. Kobor argues that UNSE needs to conduct a “full benefit/cost 1 2 3 analysis” (presumably, a “Value of Solar” Study), Stating that without such an analysis, 4 there is “no way to determine the current relationship between the retail rate and the 5 value of NEM exports, and thus no way to determine the reasonableness of the 6 Renewable Credit Rate.” If a cost Shift exists, Ms. Kobor says, there is no way to even 7 tell what direction it goes in!32 I beg to differ. This is like saying that unless I can give : the precise height of an elephant, I can’t say it is bigger than a horse. 10 The “value” of solar Simply does not need to be assessed before UNSE’S proposal is 11 acted upon. Rates are based on either a market or cost base, not some theoretical and 12 highly subjective notion of value. There have been a number of such studies in recent 13 years, which come to quite diverse and often con■icting conclusions. I do not believe 14 such studies can accurately determine what the prices should be, and I certainly do not 15 see any basis for pricing all energy sources based on cost and market, while solar DG, 16 alone, of all resources, is priced based on some consultant’s subjective assessment of 17 value. 18 Q. WHY DO YOU BELIEVE THAT THE VALUE OF SOLAR CANNOT BE ACCURATELY DETERMINED THROUGH AN UNBIASED STUDY? A. Assessing the future value of solar necessarily involves making arbitrary and subjective 19 20 21 determinations based on Speculation about future events as well as monetizing alleged 22 attributes of distributed solar, some of which may not, in fact, actually be attributes. It is 23 Simply impossible for one to conduct such a Study on an unbiased, much less accurate, 24 basis. 25 26 27 28 32 Kobor Direct Testimony at p. 27. 27 Ms. Kobor herself recognizes the fundamental problem, acknowledging the existence of competing 2013 APS studies finding drastically different values.33 She seems confident, however, that with Commission oversight, a cost/benefit analysis can be conducted that will produce a “reliable result,” and suggests use of a guidebook prepared by the Interstate Renewable Energy Council, citing the categories of benefits identified in the IREC report:34 a) avoided energy benefits b) system losses C) generation capacity 10 d) transmission and distribution capacity 11 e) grid support services 12 f) 13 g) security services 14 h) environmental services 15 financial services social services 16 j) 17 k) utility costs 18 1) customer costs decline in value for incremental solar additions at high market penetration distribution 19 transmission 20 transmission/distribution investment) 21 22 and (T&D) line loss reduction (avoided n) environmental benefits (emission mitigation costs) avoided purchased power/risk 23 P) avoided grid support 24 (l) economic development 25 26 27 33 See Kobor Direct Testimony at p. 27. 34 Ms. Kobor, pp. 27-28. Witness Mark Fulmer references many of these same elements in his Direct Testimony. 28 28 That Ms. Kobor considers the IREC Report to be an unbiased starting point proves more N than she intended. The IREC publication, rather than being a “best practices” guide, is an advocacy piece that simply lays out an outline for ways of arguing for cross subsidization of solar DG— without any developed evaluation methodology. Indeed, it is an attempt to find a rationale for the prices derived from retail net metering, which was never a carefully reasoned pricing regime, but was, rather, a default methodology that evolved for reasons that are no longer relevant in today’s electricity market. Not only does IREC offer no methodological assistance, it provides no basis whatsoever for monetizing the criteria 10 noted above. It is, in fact, little more than a laundry list of potential attributes solar 11 advocates can use to call for higher prices for solar DG, without ever offering a serious, 12 fact based rationale for the claims asserted. It suggests, for example, an examination of 13 the impact of solar DG on carbon reduction, but gives little guidance on how such an 14 effort should be undertaken, and, remarkably, never even suggests that one might 15 examine the cost effectiveness of solar DG in reducing carbon emissions compared to 16 such alternatives as energy efficiency, large scale solar, nuclear, and wind. Similarly, it 17 fails to even reference the fact that in order to assess the carbon effects of solar DG, one 18 needs to clearly identify what generating resources are being displaced (e. g. coal, 19 combined cycle) by solar DG when it is producing energy and what the impact of the 20 intermittent nature of solar DG is on dispatch, as well as the environmental impact, not 21 to mention economic efficiency, of ramping generation up and down to accommodate 22 the intermittent injection of solar DG energy into the system. 23 24 The point here is not that the IREC document, or any number of value of solar studies, 25 are incomplete and biased, although the IREC report clearly is, as are many value of 26 solar studies. Rather, it is that such studies are highly subjective, often quite arbitrary, 27 and, if reasonably complete, extraordinarily 28 29 complex (if the authors are truly disinterested analysts, as opposed to advocates with a point of View), and, to be done correctly, these studies require a great deal of time and expense. Moreover, the results, no matter how honestly derived, are always going to be highly subjective, full of debatable assumptions, and subject to severe criticism by any number of interest groups with an axe to grind or a point of View to advance. Perhaps most interesting is that such studies rarely even reference the historic reference points used for pricing, markets and COStS. There is a reason we rely on markets and market prices, and not “value analysis,” whenever possible. When we do use value analyses, it is important to keep these 10 11 12 13 14 15 16 17 18 limitations in mind. Indeed, electricity pricing in the US. has always been based on one of two highly disciplined foundations, cost (including avoided cost), and/or market. Every energy source in the country is priced on one of those foundations. Pricing one resource based on someone’s subjective View of “value,” while pricing every other resource based on a disciplined and systematic approach, is simply indefensible as a matter of public policy. The public deserves better than that. DO YOU AGREE WITH MS. KOBOR THAT THE EXISTENCE OF OTHER COST SHIFTS IS A REASON TO TAKE NO ACTION ON ROOFTOP SOLAR SUBSIDIES? 19 I do not agree with her at all. First, the fact that there may be other cross—subsidies in 20 rates is hardly a reason for refusing to remedy perhaps the most inefficient of cross 21 subsidies. This is particularly true where the remedy, as proposed by UNSE, is so simple 22 to put in place. 23 24 25 26 27 More importantly, Ms. Kobor misses the point almost completely. It may be true that other factors contribute more to UNSE’s revenue deficiency. The much bigger issue, and the one Ms. Kobor completely ignores, is that current rate design and retail net metering causes shifts, socially regressive ones, in cost allocation among customers. If solar DG customers are excused, as they are under current rate design, from having to 28 30 pay their fair share of the fixed and demand costs associated with the energy service provided to them, those costs do not disappear. Regulators are then left with just two options, either pass on those costs to non-DG customers, or, alternatively, force utilities to absorb those lost revenues, an outcome that is very likely to result in underinvestment in the grid (and may not be legally permissible or consistent with the regulatory compact). Neither of those outcomes is acceptable in any event. The real issue is avoiding that highly unfortunate choice, and the way to fix it is to require all customers, including rooftop solar customers, to pay their fair share of the system’s fixed and demand costs. 10 ll 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 Ms. Kobor’s testimony draws the wrong conclusion in reasoning that there is no good reason to single DG customers out as a class for special reformed rates. First, as noted by several witnesses, now including myself, in this proceeding, there are many reasons why UNSE is correct in making the tariff reforms it proposes. Revenue deficiency is only one of them, and perhaps not even the most important one. In specific regard to revenue deficiency, however, DG solar is expected to continue to grow in Arizona, and the more it grows, the bigger a burden the cross—subsidy will represent. Eventually, it will simply not be sustainable. Moreover, as can already be seen in this state and others, the politics of getting the tariffs right becomes increasingly difficult when more and more people are invested in a severely ■awed tariff that skews the prices in costly and economically perverse ways. It is best to get the prices right from the beginning so that when customers make their decision about whether or not to go solar, the price signals are correct and the costs and benefits to society are correctly aligned in the tariff formulation. Failing to do so means that every new solar DG installation represents a significant investment made based on expectations of continuing out—of—market and above cost pricing, thereby increasing the difficulty of ever reforming the policy. In fairness to both DG and non-DG customers alike, it is important to get rate signals 27 correct and sustainable as soon as possible. 28 31 Q. MS. KOBOR ASSERTS THAT UNSE HAS FAILED TO ADEQUATELY PROVE ITS CLAIMS. PLEASE RESPOND. A. Ms. Kobor challenges UNSE’S proposal that distributed generation has the potential to 1 2 3 seriously impact the grid and UNSE’s analysis of the incentive problems it creates for 4 customers, by complaining that UNSE has not provided detailed evidence that the 5 problems it fears are happening at scale today, complaining that “UNSE...relies on 6 broad national and regional studies, which may or may not apply to UNSE’s grid and 7 service territory.”35 She herself, however, provides no actual evidence that the facts are 8 on her side. She does not even present a plausible theory that would suggests the facts 9 might be on her side or explain why she thinks Arizona’s circumstances are not 10 adequately captured by regional or national studies. What, exactly, does she think is :: different about Arizona? She never bothers to explain. 13 In doing this, she is failing to meet her client’s burden of raising sufficient doubt as to 14 the applicant’s proposal to shift the burden of proof back to the utility. I was a regulator 15 for ten years. I have taught regulation in more than two dozen countries, and at many 16 institutions within the United States, including all three NARUC—approved training 17 programs, at Michigan State University, the University of Florida, and New Mexico 18 State University, as well as here at Harvard. The process that should be followed in 19 cases like these is clear. The utility makes a proposal, and parties who wish to rebut the 20 proposal, whether the staff of the Commission or outside parties, have the opportunity to 21 do so. They do, however, also have a burden of providing evidence to support their 22 rebuttal, and if they don’t meet that burden, there is nothing for the Commission to go 23 on. Witness Kobor has offered absolutely no probative evidence to meet that simple 24 burden. 25 26 27 28 35 Kobor Direct Testimony at pp. 16-17. 32 1 In fact, both Ms. Kobor and Mr. Fulmer complain that the extensive evidence and 2 analysis provided by UNSE is somehow inadequate but fail to offer any countervailing 3 evidence or even, in many cases, a strong theory suggesting why UNSE’s analysis 4 requires further support. What witnesses Mr. Fulmer and Ms. Kobor have provided here 5 is testimony making broad assertions with no evidentiary basis. Consequently, they have 6 failed what the law terms their burden of going forward with credible evidence 7 supporting their assertions. 8 It is ironic that all this effort to undermine UNSE’s analysis still fails to get to the heart 9 of the issue. Ms. Kobor devotes many pages to an argument that points to the 10 unsurprising fact that DG customers are still a relatively small part of the UNSE 11 customer base—therefore, the gross amount of cross subsidy they are receiving is 12 currently relatively small. It is not clear 13 motivation for the tariff reforms proposed by UN SE, but her largely exclusive focus on 14 it makes it appear that she has presented less than a complete understanding of all of the 15 16 issues at hand. Q. 17 18 Ms. Kobor chooses to focus on just one GENERATION DISTRIBUTED THE CONSIDER UNSE SHOULD REQUIREMENT IN ARIZONA’S RENEWABLE ENERGY STANDARD (RES) IN SETTING ITS RATE FOR SOLAR DG ENERGY? 19 No. Although Ms. Kobor warns that a change in the net metering rate might result in 20 UNSE failing to meet its distributed generation requirement in upcoming years,36 this is 21 not a consideration that it is appropriate to incorporate into rate design. Significantly, 22 she has also failed to make any showing that her assertion that implementing UNSE’s 23 rate proposals would cause UNSE to be out of compliance with its required quota of 24 DG. In effect, she is pleading for distorted pricing and incentives over the long term in 25 order to head off a specific, short—term problem which indeed may not turn out to be a 26 problem at all. This is not good ratemaking as expounded by Professor Bonbright! 27 28 36 Kobor Direct Testimony at p. 52. 33 Instead, if there is a need for additional solar DG to meet state requirements, that should be handled separately, helping to keep costs transparent and efficiency incentives in place. Moreover, the rule itself is absolutely within the discretion of the Commission to modify, eliminate, or waive. WHAT IS THE RELEVANCE OF MS. KOBOR’S ARGUMENT THAT THE PROPOSED RENEWABLE CREDIT RATE WOULD VIOLATE THE COMIVIISSION’S EXISTING NET ENERGY METERING RULES? Ms. Kobor makes this argument on pp 32—33 of her testimony. Without evaluating this argument substantively (I have not examined the question of whether the proposed rate violates the existing rule or not), it is improbable that this is an insurmountable obstacle, 10 11 12 13 should the Commission wish to approve the proposed rate. Whatever the existing rules are, they were established by the Commission. When I sat on the Public Utilities Commission of Ohio, we always had available the ability to repeal, modify, or waive commission-created rules where we believed the circumstances were such that the 14 action(s) was (were) both warranted and reasonable. 15 16 17 IS THERE ANY BASIS FOR MS. KOBOR’S ASSERTION ON PAGE 37 THAT “MOVElVIENT TOWARDS MANDATORY DEMAND CHARGES FOR ALL RESIDENTIAL CUSTONIERS IS IN NO WAY REFLECTIVE OF MODERN TRENDS IN RATEMAKING?” 18 None at all. I am Executive Director of a leading “think ta 19 20 21 22 23 24 25 26 ” on electricity policy, and the potential role of demand charges in rate design is a frequent subject of discussion. Recent proceedings in Ms. Kobor’s home state of California included significant debate over that issue. In fact, demand charges are a critical element in the movement in ratemaking toward unbundling prices, making prices more transparent, and providing customers with meaningful price signals in order to bring greater efficiency to the use of electric energy. Indeed, the position of the ACC Staff in this matter is a classic example of demand charges being at the center of regulatory thinking in the US. today. Suffice it to say that Ms. Kobor’s deeming the idea as not re■ective of modern trends in 27 ratemaking is both uninformed and profoundly mistaken. 28 34 WOULD YOU AGREE WITH MS. KOBOR’S ASSERTION THAT THE PROPOSAL TO REDUCE THE NUlVIBER OF CUSTOMER TIERS IS ALSO NOT REFLECTIVE OF “MODERN RATE DESIGN?” No. Ms. Kobor has a distorted view of “modernized rate design” if she believes that reduction of the number of rate tiers is not compatible with it.37 This issue has been vigorously debated within her own state very recently, and is a topic of debate across the country, indeed, around the world. UNSE’S PROPOSAL FOR A RENEWABLE CREDIT RATE IS AN APPROPRIATE WAY TO COMPENSATE CUSTOMERS FOR ENERGY EXPORTED TO THE GRID. 10 11 TURNING TO THE THIRD ELElVIENT OF UNSE’S PROPOSED RATE, IS THE MOST RECENT RENEWABLE PPA A REASONABLE BENCHMARK FOR SETTING THE RENEWABLE CREDIT RATE FOR COMPENSATING DG SOLAR PRODUCTION? 12 Yes, this approach is entirely reasonable. It uses a benchmark price established for 13 intermittent renewable energy by looking at the last arms-length transaction to purchase 14 intermittent renewable energy in the competitive bulk power market. By using the most 15 recently negotiated rate, the price recognizes that energy prices ■uctuate and does not 16 lock in a higher than market standard offer for solar DG. It does not over—compensate 17 distributed generation beyond levels of compensation offered to grid-scale renewables, 18 thereby averting the potential for diverting capital from a more efficient generator to a less efficient one. Indeed, by gearing the price paid for solar DG to that of a more 20 efficient resource, UNSE’ proposal has the very positive effect of incentivizing solar DG 21 to become more efficient and improve productivity. That incentive is completely lacking 22 in the existing retail net metering pricing model, and that is one of the problems that 23 cries out for reform of the type being proposed by the applicant in this proceeding. 24 Nonetheless, UNSE’s proposal is still generous in the sense that it is paying the same for 25 solar DG as it does for utility-scale solar, despite the fact that the latter is the more 26 efficient resource. Finally, the benchmark price is derived from transactions involving 27 37 Kobor Direct Testimony at p. 55. 28 35 1 energy resources that are, like solar DG, intermittent. Thus, UNSE’s proposal avoids 2 having to compare apples with oranges. The price point is one that is subject to market 3 discipline, recognizes ■uctuations in the wholesale market, and prevents a reallocation 4 of capital toward less efficient resources. 5 Q. WHAT ABOUT MR. FULNIER’S AND MS. KOBOR’S ARGUNIENT THAT THE PRICE OFFERED FOR DG SOLAR SHOULD BE HIGHER THAN THE PRICE OF UTILITY-SCALE RENEWABLES, BECAUSE DG SOLAR OFFERS MORE VALUE TO THE UTILITY? A. The argument has no merit. The argument they are making is a variation on the “value 6 7 8 9 of solar” theme discussed above. Ms. Fulmer and Ms. Kobor both suggest that rooftop 10 solar offers more value than utility—scale solar in their discussion of UNSE’s proposal to 11 compensate DG solar at the going wholesale market rate, as established by the latest 12 comparable PPA. The geographic diversity of solar DG systems is their main argument. 13 This, they suggest, will alleviate the intermittency of solar power. At the same time, Mr. 14 Fulmer argues, distributed solar systems do not have “the potential habitat, visual and 15 cultural impacts associated with utility—scale solar plants.”38 However, neither of the two 16 makes any effort to quantify this additional “value.” They do not even try to show that 17 the so called “value” they reference is equal to the full retail price of delivered 18 electricity. Because they do, as noted, cite anecdotal examples of the added “value” they 19 claim, I will address each of them. 20 Assessing the value of geographic diversity relative to intermittency. This argument 21 is one dimensional thinking—Mr. Fulmer and Ms. Kobor highlight one or two small 22 possible benefits of DG solar compared to utility—scale solar, while ignoring the 23 overwhelming, evidence—based, consensus that grid-scale PV generation is 24 efficient—not just for the obvious reasons of economies of scale, and the fact that grid— : scale plants are far more likely to have optimized panel placement and tracking, but also 27 28 38 Fulmer Direct Testimony at p. 4. 36 because a utility-scale solar plant, purpose-built to produce solar power, is more likely to be optimally situated in areas of peak sunshine.39 It is unlikely this argument about geographic diversity points to any significant advantage of solar DG over grid—scale solar. For one thing, it is based on the false assumption that grid—scale solar plants are limited to a single location. Utility—scale solar plants can take advantage of geographic diversity as well—and the potential for diversity is great, since utilities can purchase power from distant plants, as long as they are connected to the transmission grid. Rooftop solar, of course, by its very nature is 10 11 entirely concentrated within the narrower confines of a distribution utility. Thus, the claim of geographic diversity as a benefit of rooftop solar has no basis in fact. 12 Even if one accepts the dubious premise of greater geographic diversity in rooftop solar 13 systems, the claim that this outweighs the benefits of utility—scale solar does not hold 14 water. The previously cited Brattle Group study comparing grid—scale with rooftop solar 15 systems looks at this issue in the terms proposed by Mr. Fulmer and Ms. Kobor, pitting 16 the intermittency of a single utility-scale solar unit against an array of rooftop solar 17 units. While acknowledging the potential role geographic diversity could play in 18 reducing intermittency from the rooftop solar units, the Brattle analysis also looks at 19 other significant factors, noting that “Utility-scale systems that oversize the panel array 20 relative to inverter capacity will likely have a better profile (less variability) than any 21 given residential-scale system,” a factor that needs to be weighed against geographic 22 diversity, stating that the net impact on ancillary services needs is “difficult to 23 determine,” but noting grid-scale solar’s other advantages of “better location selection 24 25 26 27 39 Tsuchida, Bruce, Sanem Sergii, Bob Mudge, Will Gorman, Peter Fos-Penner, and Jens Schoene. Comparative Generation Costs of Utility-Scale and Residential-Scale in Xcel Energy Colorado’s Service Area. Brattle Group, 2015, p. 9. Available at http://brattle.com/systern/publications/pdfs/OOO/OOS/ 188/original/Comparative_Generation_Costs_of_Uti lity—Scale_and_Residential-Scale_PV_in_Xcel_Energy_Colorado's_Service_Area.pdf?1436797265. 28 37 1 (higher insolation), better controllability and visibility by the system operator, and being 2 able to provide downward ancillary services.” After reviewing other factors, such as the 3 higher capacity factor of utility-scale PV, the possible transmission loss reductions 4 associated with distributed PV, the Brattle study concludes that “[o]verall, inclusion of 5 these factors is likely to increase the cost difference between utility—scale and 6 residential—scale PV systems.”40 7 8 9 Assessing the value of “habitat, visual and cultural impacts.” In his November, 2015, testimony, Mr. Fulmer cites the discussion of these potential impacts in the DOE’s Suns/wt Vision Study as an argument for the greater value of solar DG as opposed to 10 utility-scale 11 However, his presentation of the issue is one-sided, neglecting the issues faced by distributed solar, ironically raised by the same DOE study he cites only a 12 few pages beyond the information he presents in his testimony: 13 14 15 16 owners of existing systems face potential challenges when growing trees or new structures on nei g hborin g P ro P ert shade their solar collectors. Given that there 1s no common—law right to sunhght 1n the United States, these issues present serious barriers to the adoption of solar energy42 17 The DOE report goes on to suggest some legal mechanisms that may allow neighbors to 18 navigate this issue through establishing “landowners’ rights to present and future 19 unobstructed direct sunlight” or through sale of easements, etc.—but they have 20 appropriately pointed out that the installation of rooftop solar is not without implications 21 to neighboring homes—particularly if it results in attempts to limit the growth of trees 22 on a neighbor’s property or their attempts to remodel or expand homes.43 What DOE is 23 24 25 26 27 40 Id. at pp. 35-36. 41 For those, like me, puzzled by the reference to “cultural impacts,” the DOE study writes that “con■icts may arise if development impacts cultural sites or interferes with US. Department of Defense (DOD) Vision Study,” Department of Energy, February 2012, at p. 171. These activities.” potential “cultural impact” problems seem to be readily addressed through proper site selection. 2 DOE SunShot Vision Study at 184. 43 For that matter, it may be worth considering the incentive solar panels might present to cut down existing shade trees. When tree shade is an issue, certainly, the calculus about the environmental costs 28 38 discussing is selective destruction of other people’s trees to accommodate solar DG. In 2 short, we may often run the real risk of losing the aesthetic, shade, and carbon offset 3 benefits provided by trees in order to accommodate solar DG. Thus Mr. Fulmer’s claim 4 that solar DG provides greater habitat value is highly dubious at best. Similarly in regard 5 to the claim of “visual” value, another well—known source of neighborhood con■ict 6 related to solar DG is the potential for glare from solar panels to adversely impact 7 neighbors. In short, visual and habitat impacts are not limited to utility-scale solar. 8 Q. MR. FULlVIER CRITICIZES UNSE’S RENEWABLE ENERGY CREDIT BECAUSE IT IS SET UP TO ADJUST EVERY YEAR. WHAT IS YOUR EVALUATION OF THIS ARGUMENT? A. Mr. Fulmer raises a number of concerns about the proposed compensation methodology; 9 10 11 however, not all his concerns seem to be consistent with each other. In part, his concern 12 seems to be that the deal being offered to rooftop solar customers is not good enough. 13 The proposal here is to tie rooftop solar compensation to a measure of the actual price of 14 renewable energy in the open marketplace in a given year—instead, Mr. Fulmer 15 suggests that rooftop solar customers should be given a twenty year price guarantee. He 16 asserts, “the prudent utility will look at its needs in the future and consider all the 17 options for meeting those needs in a least—cost fashion. . ..If you can take actions NOW 18 that can save ratepayers money (or reduce risk or meet some other planning goal) in the :: future, at higher costs today, they are likely the correct actions to take.”44 21 Mr. Fulmer’s stated concern for securing marginal efficiencies for ratepayers related to 22 getting the most bang for their bucks in the purchase of renewable power is inconsistent 23 with his main objection to the proposed pricing scheme: “Further, as proposed, Rider 11 24 will likely act more like a ratchet, ever going down. This obviously creates a problem 25 26 27 and benefits of solar panels becomes increasingly complex, expanding to include the question of how additional air conditioning power might be used to compensate for loss of shade. Fulmer D1rectTest1mony at p. 9. 28 39 for someone considering an investment in a fixed asset.”45 That is, Mr. Fulmer (correctly, in my opinion) anticipates that grid—scale renewable energy will get cheaper over time (keeping in mind that, as discussed above, that larger—scale renewables are more cost effective today).46 And, disturbingly, he views this as a problem, one that should be solved by committing utility resources to overpaying for DG renewable energy for decades into the future—when that same amount of money, if indexed to declining renewables costs, could buy ratepayers far more renewable energy per dollar under the UNSEproposal, and would at the same time provide rooftop solar providers with an incentive to be more efficient. 10 Q. WOULD THE RENEWABLE CREDIT RATE BE SUBJECT TO GAMING? A. Ms. Kobor raises the vulnerability of the proposed pricing system to gaming as a 11 12 concern,47 suggesting that the utility could manipulate its PPAs to artificially de■ate the 13 Renewable Credit Rate. 14 15 There is no system that is not at least theoretically vulnerable to some type of gaming. 16 Indeed, policing against gaming that is contrary to the public interest is an important part 17 of the raison d ’etre of the ACC. 18 reason to retain the severely ■awed system of retail net metering, which itself is already 19 being gamed by the solar DG industry.48 20 transparency. Through annual public filings, the Commission and the public will be able 21 to review the Renewable Credit Rate and address any concerns about gaming that may 22 23 24 25 26 27 Thus, the risk of gaming is simply not a sufficient What UNSE has proposed has the virtue of 45 Fulmer Direct Testimony at p. 7. 46 Curiously, Mr. Fulmer does not object to the fact that utilities, under net metering buy excess solar DG at the retail rate, something which also ■uctuates over time, but, based on solar DG industry marketing claims (and I received such a robo call very recently) that ■uctuation is upWard. Thus, Fulmer is essentially arguing not against uncertainty, but rather calls for a completely asymmetrical arrangement where solar DG gets the benefits of any upward price ■uctuation, but has no risk of downward ■uctuation in the marketplace. 47 Kobor Direct Testimony at p. 31. 48 SolarCity’s recent 10K filing clearly describes that its business model is built on chasing subsidies, a classic example of gaming, which is not in the public interest. See SolarCity Corp. 10K (2/24/15) at p. 38. 28 40 arise. Moreover, the Commission itself has the requisite skills and intelligence to monitor, identify, analyze, and remedy any adverse gaming that may occur. IS THE PROPOSAL MADE BY LON HUBER OF RUCO FOR THREE RATE OPTIONS FOR SOLAR DG CUSTOMERS A GOOD ALTERNATIVE RATE APPROACH? RUCO’s aim here as explained by Mr. Huber is laudable: RUCO would like to begin by ensuring that rooftop DG can be a neutral cost proposition for ratepayers as soon as possible. Once that milestone is reached RUCO would like to see DG be a net benefit to all ratepayers. Finally, the third milestone, RUCO would like to see a closer cost parity between wholesale grid-connected solar and rooftop solar. \] 10 11 12 13 14 I agree with these goals. However, the rate structure RUCO proposes would not get Arizona there “as soon as possible”—in fact, though it might progress slowly in that direction, it would never actually arrive, even after twenty years. I think we can fix the retail net metering rate problem faster than that! 15 Mr. Huber suggests three rate options, all of which, in my opinion, are inferior to the 16 UNS proposal. Among other issues, they don’t seem to fit together in a way that 17 constitutes consistent public policy. Although it appears from his testimony that he 18 understands the issues, the RUCO proposals do not seem to have translated into a 19 proposal that resolves the issues before the Commission. Here are the major problems 20 with RUCO’s proposal as I see them: 21 1. 22 23 24 25 26 27 28 The Non-Export Option. With this option, solar DG customers could stay on traditional rates—they would just have to agree not to export solar power to the grid. To me, this does not address the issues. To the extent that customers choose any of the current rates, in which fixed costs are recovered through variable energy use charges, the cross subsidy issues above are not addressed at all. Furthermore, the non—export provision is not the direction Arizona should go if it wants to optimize the potential contributions of solar power—the aim, rather, 41 should be to optimize the benefit to the consumer and the grid of investments in solar, by having fair rates that create proper incentives for efficient deployment and use of solar DG—not to discourage customers from maximizing the benefits of their investment in solar energy. The “TO Option. This proposed rate weakens some elements of the UNS proposal that is intended to correct inequities associated with solar DG’s participation in retail net metering. The minimum bill (alternative to fixed charge) increase is smaller ($12); and the proposed payment for exported power is higher. The proposal is to pay 8.5 cents/kWh for excess power exported to the 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 grid—considerably more than the going rate for utility—scale renewable energy. This higher rate is based on some very rough calculations of capacity savings undertaken by Mr. Huber. For the reasons of solar interrnittency I discussed above, in my opinion these capacity savings will not be realized by the utility or its customers. At the same time, the proposal includes an aggressively high TOU demand charge during summer peak hours (2pm-8pm)—at $19.50/KW, the charge is almost twice as high as the higher of the two demand charges proposed by UNS for customers whose demand exceeds 7 kW. This does provide a strong incentive to customers who choose this rate to minimize their consumption during these summer peak hours—I would note, though, that (keeping in mind the California duck curve chart discussed earlier) the long—run interest of the utility and of the growth of solar in Arizona is to encourage solar DG customers to use energy when the sun is shining, but consciously reduce their usage when the weather is cloudy and when the sun declines and sets. The 2pm-8pm TOU rate proposed is too blunt an instrument to promote this behavior. Overall, it is a little hard to tell what the impact of this proposed option would be, in terms of minimizing cross subsidies. The element of choice of rates endorsed by RUCO 27 28 42 1 would be the decisive factor in preserving cross-subsidies, since presumably only 2 customers who hope to benefit from such a rate would choose it. 3 3. RPS Bill Credit Option. This final (very generous) option addresses only the 4 export rate portion of the cross subsidy. Reforming the export rate is a step in the 5 right direction, but to be comprehensive, it must be coupled with rate design 6 reforms. Further, this RPS Bill Credit Option pursues this partial reform in a way 7 that would be better described as “glacial” than “gradual.” The renewable credit 8 rate would start at a generous 11 cents/kWh—a rate which customers could lock 9 in for twenty years. For new customers, the initial rate offer would slowly 10 decline as more solar capacity was added, reaching competitiveness with current 11 utility-scale renewable energy rates as late as 2025 (though potentially this could 12 happen earlier, depending on how fast growth in solar capacity occurs). 13 Throughout this period, customers could lock in these elevated per kWh 14 reimbursement rates for twenty years at a time. So it would be twenty to thirty 15 years before the utility was buying DG solar power at a competitive rate 16 (assuming, 17 further)—and even then, cross—subsidies associated with the use of solar to offset 18 a customer’s own consumption would not be addressed at all. While this option 19 has some merit in its structure and intent, it takes gradualism far beyond what the :(1) respected Professor Bonbright could have intended! meanwhile, that utility-scale renewable costs do not decline Q. DO YOU HAVE ANY COMMENTS ON THE ISSUE OF RATE GRADUALISM THAT THE RUCO PROPOSAL IS TRYING To ADDRESS? A. Yes. On the issue of rate gradualism, I must disagree with Mr. Huber’s assertion that 22 23 24 UNSE’S proposal violates this principle. Increasing the fixed charge from $10 to $20 25 (and now, as revised by UNSE, $15) is not “seriously adverse to existing customers” 26 and is well short of revolutionary. For most customers, this change will, if anything, be :: accompanied by a Slight reduction in the overall bill, as the burden of paying for cross 43 subsidies to other customers is eased. And the individual customers likely to experience the largest bill increases, DG customers, could for the most part protected by grandfathering. The major change here is to the rates prospective new DG customers will face, as they receive a more accurate signal about the value of the solar DG in which they are considering investing. VI. CONCLUSION DO YOU HAVE ANY CONCLUDING REMARKS? Yes. As I have explained in my testimony, the central issue for solar DG in this matter is 10 not a con■ict between the utility and solar energy providers. Rather, UNSE’s proposed rate change for solar DG customers is an attempt to enable effective competition 11 12 13 14 between different energy resources, in which solar DG will enjoy the “mainstream” energy role called for by Vote Solar witness Ms. Kobor. The new UNSE rate would provide powerful incentives to increase the productivity and efficiency of distributive solar generation, ones that envision long term financial sustainability for distributed 15 16 17 18 solar energy. This attempt is being strenuously opposed by a solar DG industry dedicated to preserving its special status as the beneficiary of arbitrarily high, out of market, non-cost based, heavily subsidized prices, resulting in higher profits for themselves at the expense of consumers, solar and non—solar alike, to the long run 19 detriment of the viability of distributed solar generation. 20 21 UNSE’s proposals before the Commission would not only remedy an urgent and 22 growing problem related to the subsidies of higher—income customers by lower-income 23 customers entailed by net energy metering, reduce the risk of depriving the grid of 24 needed investment, open opportunities for new service providers to enter the market to 25 provide services and products to enhance efficiency opportunities for customers, and 26 enhancing competitive markets forces in the power sector. Just as important, and most 27 ironically, it would also establish a pricing structure to enhance the long term prospects 28 44 1 of distributed solar generation far beyond the dubious prospects offered by the “chasing 2 subsidies” business model currently being used by most solar DG developers. 3 Q. DOES THIS CONCLUDE YOUR SURREBUTTAL TESTIMONY? A. Yes. Attachment ACB -18R 1 of 8 ASHLEY C. BROWN EXECUTIVE DIRECTOR HARVARD ELECTRICITY POLICY GROUP MOSSAVAR-RAHMANI CENTER FOR BUSINESS AND GOVERNMENT JOHN F. KENNEDY SCHOOL OF GOVERNMENT HARVARD UNIVERSITY 79 JOHN F. KENNEDY STREET CAMBRIDGE, MA 02138 617-495-0959 ashley_brown@harvard.edu http://www.hks.harvard.edu/hepg/brown.html Ashley Brown is an attorney. He is the Executive Director of the Harvard Electricity Policy Group at Harvard University’s John F. Kennedy School of Government. It is a leading “think tank” on matters related to electricity restructuring, regulation, and market formation. He has been an instructor in Harvard’s Executive program on “Infrastructure in a Market Economy,” at the World Bank Regulatory Training Program at the University of Florida, and at the European University’s Florence School of Regulation. Mr. Brown has also served as an arbitrator in matters relating to the evolution of competition in infrastructure industries. Before his current activities, Ashley Brown served as Commissioner of the Public Utilities Commission of Ohio, appointed twice by Governor Richard F. Celeste, ■rst for a term from April 1983 to April 1988 and for a second term from April 1988 to April 1993. As Commissioner, he was of ■ve members responsible for the regulation of the state’s electricity, telecommunications, surface transport, water and sanitation, and natural gas sectors. Prior to his appointment to the Commission, Mr. Brown was Coordinator and Counsel of the Montgomery County, Ohio, Fair Housing Center. From 1979-1981 he was Managing Attorney for the Legal Aid Society of Dayton, Inc. From 1977 to 1979 he was Legal Advisor of the Miami Valley Regional Planning Commission in Dayton. While practicing law, he specialized in litigation in federal and state courts, as well as before administrative bodies. He has served as an expert witness in litigation in the courts and administrative agencies In addition, Mr. Brown has extensive teaching experience in public schools and universities. EDUCATIONAL BACKGROUND 1968 1971 1977 1967 FAMILY CURRENT AFFILIATIONS Wife Daughter Daughter BS. MA. J .D. Bowling Green State University, Bowling Green, Ohio University of Cincinnati, Cincinnati, Ohio University of Dayton School of Law, Dayton, Ohio Doctoral Studies (all but dissertation) New York University, New York, New York Attended Universidade do Parana; Curitiba, Parana, Brazil as an exchange student Edith M. Netter Sara Mariasha Brown-Worsham Mariel Schaefer Brown Member, Editorial Advisory Board of The Electricity Journal Member, Editorial Board, International Journal of Regulation and Governance Member, Board of Directors, e-Curve Fellow, Centro de Estudios en Regulacion e Infraestructura, Fundacién Getulio Vargas, Rio de Jane’iro, Brazil Attachment A08 -1 SR 2 of 8 Member, Policy Committee, David Rockefeller for Latin American Studies, Harvard University Member, Brazilian Studies Committee, David Rockefeller Center for Latin American Studies, Harvard University Member, Advisory Board of Development Gateway Site, The World Bank Frequent speaker and lecturer on regulatory, infrastructure, and energy policy matters in North and South America, Europe, Africa and Asia. PREVIOUS AF F ILIATIONS Member, Board of Directors, Entegra Power Chairman, Town of Belmont Municipal Light Advisory Board Member, Board of Directors, Oglethorpe Power Corporation, Tucker, GA Member, Editorial Advisory Board of Electric Light and Power Vice-Chair, American Bar Association Committee on Energy, Section of Administrative Law and Regulatory Practice Chair, American Bar Association Annual Conference on Electricity Law Member, The Keystone Center Energy Advisory Committee Member, National Association of Regulatory Utility Commissioners Member, Executive Committee, National Association of Regulatory Utility Commissioners Chair, Committee Commissioners on Electricity, National Association of Regulatory Chair, Subcommittee on Strategic Issues, National Association of Regulatory Utility Commissioners Member, Great Lakes Conference of Public Utilities Commissioners Member, Great Lakes Conference of Public Utilities Commissioners Executive Committee Member, Mid-America Regulatory Conference Member, Board of Directors, The National Regulatory Research Institute Member, Advisory Council to the Board of Directors of the Electric Power Research Institute Member, US. EPA Acid Rain Advisory Committee Chair, Planning Section, National Govemors' Association Task Force on Electric Transmission Utility Attachment AGE -1 SR of 8 Member, the Keystone Center Dialogue on Emissions Trading Member, the Keystone Center Project on the Public Utility Holding Company Act of 1935 Member, The Keystone Center Project on State/Federal Regulatory Jurisdictional Issues Affecting Electricity Markets Member, Policy Steering Group, The Keystone Center Project on Electricity Transmission Member, Advisory Council of the Board of Directors of Nuclear Electric Insurance Limited Member, Advisory Council of the Consumer Energy Council of America Project on Electricity Member, Advisory Committee of the Consumer Energy Council of America Air Pollution Emissions Trading Project Member, National Task Force on Low Income Energy Utilization and Conservation Member, Board of Directors, Center for Clean Air Policy Member, National Blue Ribbon Task Force on Allocating the Cost of New Transmission Of Counsel, Dewey & LeBoeuf Of Counsel, Greenberg Tauris INTERNATIONAL EXPERIENCE Member, Board of Director, Entegra Power Group Member, U.S. Delegation of State Government Of■cials in the Center for Clean Air Policy/ German Marshall Fund Sponsored Exchange on Clean Air Issues to Germany, 1989 Member, U.S. Delegation to International Electric Research Exchange (IERE), Rio de Janeiro, Brazil, 1991 Consultant, Hungarian Ministry of Industry and Trade on Gas and Electric Regulatory policy, 1991-1992 Advisor to Ministry of Trade and Industry on Writing New Laws Governing Electricity, Natural Gas, and Regulation Consultant, SNE, Costa Rican Regulatory Agency, on Transmission Access Issues, 1992 Advisor on Development of Independent Power Producers and Transmission Access Consultant, World Bank Mission to Hungary Investigating the Financing of New Power Plants for MVM (Hungarian Electric Co.), 1992 Preparation of Background Materials in Preparation of a World Bank loan to the Hungarian Power Sector Attachment ACB -1SR 4 of 8 Member, U.S. Delegation, in Conjunction with the US. Department of Energy, to the Argentina and United States Natural Gas and Electricity Regulatory Meetings, 1992 Consultant, ENARGAS, the Argentine gas regulatory agency, 1992 Providing Training for ENARGAS Commissioners and Consultant, USAID India Private Power Initiative Program on the Introduction of Private Generation and Competition into the Public Sector, 1993 Preparation of a Report on Introducing and Promoting Private Investment in the Indian Power Sector Instructor, Regulatory Training Program of the National Regulatory Research Institute at Ohio State University and the Institute of Public Utilities at Michigan State University, Buenos Aires, Argentina, 1993 Providing Training to Commissioners and Sta■ of ENARGA S Consultant, The Province of Salta, Argentina on infrastructure regulation, 1996 Providing Training to Commissioners and Sta■" of the Regulatory Agency of the Province of Salta Consultant, USAID, Philippines Electric Sector Restructuring, 1994 Preparation of Analysis and Report on Restructuring the Philippine Power Sector Including the Attraction of Private Capital in Generation, and Introduction of Competition Consultant, USAID, Russian Electric Sector Restructuring, 1994 Preparation of Analysis and Report on Restructuring the Russian Power Sector Including the Attraction of Private Capital in Generation, and Introduction of Competition Participant, Harvard University’s East Asian Electricity Restructuring Forum, 1994-1995 Delivering a Series of Lectures in China, Indonesia, and Thailand on Reforming the Power Sector Consultant, Government of Ukraine on Electricity regulatory policy and industry restructuring, 1994-1995 Advisor to the National Energy Regulatory Commission on the Structure, Processes and Substance of Electricity Regulation Consultant, Government of Brazil on Electric Sector Restructuring, 1995-1996 Adviser to the Ministry of Mines and Energy on Various Issues Related to Privatization and Introduction of Competition in the Power Sector Consultant, Energy Regulatory Board of Zambia, 1997- 2001 Advisor to the Energy Regulatory Board on the Structure, Processes and Substance of Electricity Regulation Member, Brazil-US. Energy Summit, 1995-1996 Preparation of a Report and Lecture on the Options for the Regulation of a Restructured Brazilian Power Sector Consultant, Nam Power, the electric utility in Namibia, 1998-1999 Advisor on Development of Independent Power Project and on Restructuring of the Electric Distribution Sector Attachment ACB -18R 5 of 8 Consultant, Government of Indonesia on electricity regulation, 1999 Training Government and Industry Personnel on Electricity Regulation Consultant, Government of Mozambique on reform of the commercial code, 2000 Advisor on Reformation and Rewriting of the Commercial Code Instructor, South Asia Forum for Infrastructure Regulation, 1999-present Annual Training Regulatory Personnel from Five South Asian Countries Consultant, Government of Tanzania on electricity regulation, 2002 Advisor of Rewriting the Laws Governing Energy and Transport Regulation Consultant to Inter-American Development Bank on Sustainability of Sector Reform in Latin American energy markets, 2001 -2002 Preparation of a report and Analysis on the Sustainability of Power Sector and Regulatory Reform in Latin America, with Specific Focus on Colombia, Honduras, and Guatemala Consultant to Inter-American Development Bank, Brazilian Electric Restructuring, 2002 Preparation of A Report and Analysis on Problems in the Privatization and Market Reform on the Brazilian Power Sector Consultant to World Bank on Brazilian energy regulation, 2002-2004 Preparation of A Report and Analysis of Means for Improving Regulation of the Brazilian Power Sector. Consultant to the Brazilian Government on Redesign of Electricity Market, 2003 -2004 Advisor to Ministry of Mines and Energy on Electricity Market Design Consultant to Government of Dominican Republic on Electricity Regulation, 2004 Delivery of a Series of Lectures on Problems in Restructuring and Privatization in Dominican Power Sector Consultant to Eskom, South Africa, 2004-2005 Advisor on to Eskom on Restructuring of South African Electric Distribution Sector Consultant to World Bank on Regulation and Market Reform in Russian Power Sector, 2004-2005 Preparation of Report and Lecture on Regulatory Issues in proposed New Market Design of Russian Power Sector, and Attraction of Private Capital Consultant to Government of Guinea-Bissau on Infrastructure Regulation, 2005 Training Government and Industry Personnel on Infrastructure Regulation Consultant to the Government of Mozambique on Electricity Regulation, 2006-2007 Assisting in the Re-Establishment of the Electricity Regulatory Agency Consultant to the Government of Equatorial Guinea, 2007 Assisting in writing the country ’s electricity law Consultant to the Public Utilities Commission of Anguilla, 2008 Report on Funding Regulatory Agencies Languages: English, Knowledge of Spanish and Portuguese Attachment ACB -1 SR 6 of 8 PUBLICATIONS Brown, Ashley, Jillian Bunyan. "Valuation of Distributed Solar: A Qualitative View." The Electricity Journal. 27.10 (2014): 27-48. Brown, Ashley. "Power of Connections." The Indian Express, March 10, 2014. Brown, Ashley and Louisa Lund. "Distributed Generation: How Green? How Ef■cient? How Well Priced?" The Electricity Journal, April 6, 2013. Brown, Ashley, and Victor Loksha. "International Experience with Open Access to Power Grids: Synthesis Report." ESMAP Knowledge Series 016/13, November 2013. Brown, Ashley. "Concessions, Markets and Public Policy in the Brazilian Power Sector." September 5, 2012. Brown, Ashley. "Coming Out of the Dark: For one, prices must communicate to customers the actual cost of energy at the time of consumption." The Indian Express, August 8, 2012. Tolmasquim, Mauricio T. Power Sector Reform in Brazil. Preface by Dilma Rousseff. Afterword by Ashley C. Brown. 2012 Empresa de Pesquisa Energetica (EPE). Brown, Ashley, and Francesca Ciliberti-Ayres. "Development of Distributed Generation in the United States." A Report Prepared for Empresa de Pesquisa Energetica (EPE), November 20, 2012. Brown, Ashley, Steven Levitsky, and Raya Salter. “Smart Grid and Competition: A Policy Paper.” Prepared for the Galvin Initiative, July 28, 2011. Brown, Ashley. "Can Smart Grid Technology Fix the Disconnect Between Wholesale and Retail Pricing?" Vol. 24, Issue 1 (Jan/Feb. 2011): 1040-6190. Brown, Ashley, and Raya Salter. "Smart Grid Issues in State Law and Regulation." White Paper sponsored by the Galvin Electricity Initiative, September 17, 2010. Brown, Ashley, and James Rossi. "Siting Transmission Lines in a Changed Milieu: Evolving Notions ofthe 'Public Interest' in Balancing State and Regional Considerations." University of Colorado Law Review 81, no. 3 (Summer, 2010). Brown, Ashley. Infrastructure: The Regulatory and Institutional Dimension. June 2010. 36 pages. Brown, Ashley, James F. Bowe Jr., Julio A. Castro, and Sonia C. Mendonca (Dewey & LeBoequLP) The Financial Crisis and Implications for U.S.-Mexico Brazil’s New Natural Gas Law. Latin American Law and Business Report Volume 17, Number 5, May 31, 2009. 5 pages. Brown, Ashley and James Rossi. Siting transmission lines: evolving the “public interest” to balance state and regional considerations. Version 1.0, dated August 3, 2009. The paper was originally prepared for the National Renewable Energy Laboratory“ 5 Conference on Multistate Decision Making for Renewable Energy and Transmission: Spotlight on Colorado, New Mexico, Utah, and Wyoming, August 1 1, 2009, Denver, Colorado. 7 SR 8 Ashley. 18, A 66 2008. 23 2007. 2006. 2003), 23-34. 2003), 1: 1-11. 2003): 2002. 2002): 2002). 2002), 2: (2002), 29-40. 1998. 1 1996): 1993): 13 Attachment A03 -1 SR 8 of 8 Brown, Ashley C. “Electricity A■er the Energy Policy Act of 1992: The Regulatory Agenda.” The Electricity Journal, (January-February 1993): 33-43. Brown, Ashley C. “The Energy Policy Act of 1992: The Paradox Facing the States.” Public Utilities Fortnightly, Volume 131, Number 1 (January 1, 1993): 26-28. Brown, Ashley C. “Sunshine May Cloud Good Decision Making.” Volume 7, Number 2 (Summer 1992): 113-116. for Applied Research and Public Policy, O'Neil, Richard P. and Brown, Ashley C. “Privatization and Regulation of the Oil, Natural Gas, and Electric Industries in Hungary.” Energy Law Journal, Volume 13, Number 1 (1992): 25-42. Brown Ashley C. and Bamich, Terrence L. “Transmission and Ratebase: A Match Not Made in Heaven.” Public Utilities Fortnightly, Volume 127, Number 11 (June 1, 1991): 12-16. Brown, Ashley C. “State Public Utility Regulation and Title IV.” In The New Clean Air Act: Compliance and Opportunity, edited by Reinier Lock and Dennis P. Hakawik, 177-182. Arlington, Virginia: Public Utilities Reports, Inc., 1991. Brown, Ashley C. “The Overjudicialization of Regulatory Decisionmaking.” Natural Resources and Environment, Volume 5, Number 2 (Fall 1990): 15-16 Brown, Ashley C. “A Possible Tradeoff of Federal and State Transmission Jurisdiction.” Fortnightly, Volume 124, Number 10 (November 9, 1989): 21-23. Public Utilities Brown, Ashley C. “State Power Over Transmission Access and Pricing: The Giant Will Not Sleep Forever.” Public Utilities Fortnightly, Volume 124, Number 10 (November 9, 1989): 21 ~23. Brown, Ashley C. “The Balkans Revisited: A Modest Proposal for Transmission Reform.” The Electricity Journal, Volume 2, Number 3 (April 1989): 32-39. Brown, Ashley C. “Breaking the Transmission Logjam.” The Electricity Journal, Volume 1, Number 1 (July 1988): 14-19. Brown, Ashley C. “Percentage of Income Payment Plans: Regulation Meets Social Reality.” Fortnightly, Volume 119, Number (March 19, 1987): 9-12. Public Utilities Attachment ACB 28R 1 of 59 “’ WRCAREY SCHOOLUfBUSINESS ARIZONA STATE UNIVERSITY . sel d m n research institute THE ECONOMIC IMPACT OF DISTRIBUTED SOLAR IN THE APS SERVICE TERRITORY, 2016-2035 Dr. Tim James, Dr. Anthony Evans and Lora Mwaniki-Lyman L. William Seidman Research Institute W. P. Carey School of Business Arizona State University FINAL REPORT February 16, 2016 3 Attachment ACB - 28R 2 of 59 L. WILLIAM SEIDMAN RESEARCH INSTITUTE The L. William Seidman Research Institute serVes as a link between the local, national, and international business communities and the W. P. Carey School of Business at Arizona State University (ASU). First established in 1985 to serve as a center for applied business research alongside a consultancy resource for the Arizona business community, Seidman collects, analyzes and disseminates information about local economies, benchmarks industry practices, and identifies emerging business research issues that affect productivity and competitiveness. Using tools that support sophisticated statistical modeling and planning, supplemented by an extensive understanding of the local, state and national economies, Seidman today o??ers a host of economic research and consulting services, including economic impact analyses, economic forecasting, general survey research, attitudinal and qualitative studies, and strategic analyses of economic development opportunities. Working on behalf of government agencies, regulatory bodies, public or privately-owned firms, academic institutions, and non-profit organizations, Seidman specializes in studies at the city, county or state-wide level. Recent and current clients include: Glendale Community College Greater Phoenix Economic Council Arizona Commerce Authority (ACA) Arizona Corporation Commission (ACC) Arizona Department of Health Services (A DHS) HonorHealth Arizona Dept. Mines and Mineral Resources lntel Corporation Arizona Hospital and Healthcare Association iState lnc. Arizona in vestment Council The McCain institute Maricopa Community Colleges Maricopa integrated Health System Navajo Nation Div. Economic Development The Pakis Foundation Phoenix Convention Center Arizona Mining Council Arizona Public Service Corporation (APS) Arizona School Boards Association Arizona Town Hall Arizona 2016 College Football Championship Banner Health The Phoenix Philanthropy Group BHP Billiton Phoenix Sky Harbor international Airport The Boeing Company Protect the Flows The Boys Girls Clubs of Metro Phoenix Public Service New Mexico (PNM) The Central Arizona Project (CAP) Raytheon Chicanos Por La Causa Republic Services, lnc. The City of Phoenix Fire Department Rio Tinto CopperPoint Mutual Rosemont Copper Mine Curls Resources (Arizona) Salt River Project (SRP) De Menna Associates Science Foundation Arizona (SFAZ) I Dignity Health Tenet Healthcare 0 The Downtown Tempe Authority The Tillman Foundation Environmental Defense Fund Turf Paradise Epic Rides/The City of Prescott Valley ME TRO Light Rail Tenet Healthcare Twisted Adventures lnc. Vote Solar initiative Waste Management lnc. Yavapai County Jail District Excelsior Mining Executive Budget O?ice State of Arizona The Fiesta Bowl First Things First Freeport McMoRan Attachment ACB 28R of 59 Executive Summary 0 This study examines the economic impact of three distributed (rooftop) solar deployment scenarios in the APS service territory for the study period 2016—2035, including the legacy effects of each scenario throughout the (assumed) 30 year economic life of distributed solar systems.1 0 When considered in the round from a purely financial perspective, it concludes that all three potential distributed solar deployment scenarios will have a detrimental effect the State of Arizona and Maricopa County economies, all other things being equal. 0 Additional distributed solar is estimated to lower gross state product (GSP) by approximately $4.8 billion to $31.5 billion (2015 $), dependent on the scenario. 0 Additional distributed solar deployment is also estimated to result in the net loss of 16,595 to 116,558 job years’ private non-farm employment over the entire study period, dependent on the scenario. 0 Any benefits emanating from each scenario are at best temporary, only coincident with the timing of the solar installations, and quickly counteracted by their long-run/legacy effects. 0 In all three scenarios, the total amount of money paid by distributed generation and central station generation electricity consumers, 2016—2060, is greater than the amount which would have been paid had they all alternatively continued to draw electricity from the utility’s central grid. 0 That is, in each distributed solar scenario, electricity consumers as a whole will pay more for the same amount of electricity consumed, and therefore have less money to spend in other parts of the economy. 1 The study assumes that the cost of a 2035 distributed solar installation will only be paid off in full in 2065, thereby accounting for legacy effects. lfthe economic life ofan installation is less than 30 years, the negative economic consequences will be greater. Attachment ACB 28R 4 of 59 LITERATURE REVIEW 0 The study begins with a comprehensive literature review to assess state-of-the-art methods in economic impact analysis. 0 Seidman’s methodological approach is initially positioned in a 3 x 2 matrix classification of economic impact studies, illustrated below. Seidman’s 3 x 2 Classification of Economic Impact Models - COUNT GROSS PARTIAL GROSS GENERAL GROSS COUNT NET PARTIAL NET GENERAL NET Gross studies only consider the direct positive impacts of increased economic activity in a specific sector. 0 Net studies represent a more thorough form of economic modeling as they also account for the tradeoffs in the economy which result from incentivizing one specific sector. 0 Counts are usually survey-based or theoretical capacity installation quantifications of the number of direct employees within one specific sector. 0 Partial models consider the wider effects of levels of activity in one specific sector, including the indirect and induced effects ofthe direct change, but do not consider the feedback effects of changed levels of activity in that sector — for example, the effect on wages in the labor market. 0 General models offer the most comprehensive economy-wide analysis, taking into account all of the economic interconnections and feedback effects. They also yield the most significant Gross and Net impacts. ‘ l Attachment ACB 28R of 59 o A critique of fourteen contemporary solar economic impact studies identifies only one example of a general equilibrium analysis — that is, Cansino, Cardenete, Gonzalez and Pablo-Romero’s (2013) study of Andalusia. However, this is a gross, rather than net analysis, because the authors combine renewables and non—renewables as a single sector, thereby preventing any substitution between conventional and renewable forms of generation, and effectively only allowing for positive direct demand shocks in their modeling. 0 Nine of the fourteen critiqued papers adopt the partial model approach, but six of these are gross, rather than net, studies. Positioning Seidman’s Approach Relative to Fourteen Contemporary Economic Impact Studies , , , Partial ,, o Cansino et al. 2013 i o Pollin and Garrett- o AECOM, 2011 l & Jo o Loomis, Peltier, 2009 i Alderman ,2013 - ETIC, 2016 l Judson, o Motamedi 2012 Only positive negative impacts 0 VSI and Clean Energy Project Nevada, 2011 - VSI, 2013 o Comings et al. , 2014 l - NYSERDA, 2012 l - Alvarez et al., 2009 - Treyz et al., 2011 - Frondel etal., 2009 Net Both positive and l o Berkman et al.,2014 l negative impacts l i , o In the absence of an existing CGE model for the State of Arizona, and taking into account time and cost constraints, Seidman implements 3 Partial Net REMI analysis of solar deployment in the APS service territory, 2016-2035, as the next best alternative. ECONOMIC IMPACT ANALYSIS 0 The capital costs and financing implications of each distributed solar deployment scenario are first estimated by APS, validated by Seidman, and allocated by economic sector using NREL’s JEDI model for distributed solar installations throughout the supply chain in the State of Arizona. Attachment ACB 28R 6 of 59 0 APS also supplied data describing the financial impact of each solar deployment scenario on its operating cash flow, future central station generation investments, and retail electricity rates. 0 The changes in investment included in the economic impact model are: 0 0 The annual installed costs of distributed solar capacity, 2016-2035;2 and 0 APS’ deferred or avoided central station generation investments, 2016-2035. The long-term legacy costs of the investment included in the economic impact model are: 0 The customer financing costs of distributed solar installations, 2016-2060;3 and 0 Consumer electricity rate savings, due to the deferred or avoided central station generation, 2016-2060. 0 The results for each scenario take into account the direct, indirect and induced economic impacts of the distributed solar deployment, and the 30-year legacy effects reflecting the economic life of the solar installations and deferred central station generation. 0 Using an Arizona-specific REMI model, the economic impact ofthe low case scenario, which assumes 1,300 deC of nameplate distributed solar PV installations by 2035 in the APS service territory, is as follows :4 LOW CASE SCENARIO LState of Arizona Maricopa County Total Private NonFarm Employment 5 (Job Years) -16,595 -15,685 Gross State Product _ (Millions 2015 S) -$4,806.6 -$4,491.8 Real Disposable Personal Income (Millions 2015 $) -$1,787.3 -$1,862.4 APS assumes an initial $2.50 a watt. 3 Based on the assumed 30 year economic life of the distributed system, the customer financing costs of solar installations, 20162035, will not be completed until 2065. The REMI model used currently only provides economic impact estimates up to and including 2060, but Seidman does not believe that this will materially affect the conclusions in the analysis. If the economic life of an installation is less than 30 years, the negative economic consequences are in all probability greater than the estimates presented in this study. 4 Total effects for each economic measure may not tally clue to rounding—up. 5 Ajob year is equivalent to one person having a full-time job for exactly one year. Attachment ACB 28R 7 of 59 o If the low case distributed solar deployment scenario actually transpires, the State of Arizona is estimated to lose 16,595 job years ofemployment, plus over $4.8 billion gross state product, and $1.8 billion real disposable personal income (both 2015 S). o The low case distributed solar scenario therefore estimates negative impacts for all three economic impact measures assessed for the study period, including legacy effects, in the State of Arizona and Maricopa County. 0 The economic impact of the expected or medium case scenario, which assumes 5,000 MWdc of nameplate distributed solar PV installations by 2035 in the APS service territory, is as follows:6 EXPECTED CASE SCENARIO State of Arizona Maricopa County l Total Private NonFarm Employment 7 (Job Years) —76,308 -71,344 Gross State Product _ (Millions 2015 s) -$21,613.3 -$20,149.9 Real Disposable Personal Income _ , (Millions 2015 S) -$7,956.4 -$8,087.9 0 If the expected or medium case distributed solar deployment scenario actually transpires, the State of Arizona is estimated to lose 76,308 job years of employment, plus over $21.6 billion gross state product, and approximately $8 billion real disposable personal income (both 2015 S). o The expected or medium case distributed solar scenario’s negative impacts for all three economic measures are approximately 4.5 times greater than the low case scenario’s impacts in the State of Arizona for the 2016-2035 study period, including legacy effects. 0 The economic impact ofthe high case scenario, which assumes 7,600 dec of nameplate distributed solar PV installations by 2035 in the APS service territory, is as follows:8 5 Total effects for each economic measure may not tally clue to rounding-up. 7 Ajob year is equivalent to one person having a full-time job for exactly one year. 8 Total effects for each economic measure may not tally due to rounding-up. l Attachment ACB 23R 8 of 59 HIGH CASE SCENARIO State of Arizona Maricopa County Total Private NonFarm Employment (Job Years)9 —116,558 -108,857 Gross State Product _ _ (Millions 2015 S) —$31,454.4 -$29,346.7 Real Disposable Personal Income _ (MllllOl‘lS 2015 S) -$11,901.4 -$12,091.2 o If the high case distributed solar deployment scenario actually transpires, the State of Arizona is estimated to lose 116,558 job years of employment, plus $31.5 billion gross state product, and $11.9 billion real disposable personal income (both 2015 S). o The high case distributed solar scenario’s negative impacts for all three economic measures are 6.5 to 7 times greater than the low case scenario’s impacts in the State of Arizona for the 2016-2035 study period, including legacy effects. 0 The high case distributed solar scenario’s negative impacts for all three economic measures are also 46% to 53% greater than the expected or medium case scenario’s impacts in the State of Arizona for the 2016-2035 study period, including legacy effects. - Seidman’s APS study therefore clearly demonstrates that increased adoption of distributed solar generation represents a loss to the Arizona economy in the low, expected and high distributed solar deployment scenarios. This is because the overall cost of provision of electricity to the State of Arizona will rise when referenced against a base case where electricity continues to be provided by central station generation. 9 Ajob year is equivalent to one person having a full-time job for exactly one year. Attachment ACB 28R 90f59 Table of Contents Executive Summary 1.0 Introduction 1.1. Net Metering 1.2. Economic Impact Analysis 1.3. Study Overview 2.0 Economic Impact Assessment Methods 3.0 Evaluation Framework and Review of Fourteen Economic Impact Analyses 10 4.0 Economic Impact Analyses — Magnitudes & Preferred Modeling Methods 27 5.0 Economic Impact of Net Metering — Scenarios, Assumptions and Method 30 5.1. Scenarios and Assumptions 30 5.2. Study Method 32 5.3. Solar Deployment Scenarios 34 6.0 Simulation Results: Low Distributed Solar Deployment Scenario 35 7.0 Simulation Results: Expected Distributed Solar Deployment Scenario 38 8.0 Simulation Results: High Distributed Solar Deployment Scenario 41 9.0 Conclusions 44 Appendix 48 A.1. The REMI Model 48 A.2. Effects Not Incorporated into the Analysis 48 Glossary 49 Attachment ACB 28R 10 of 59 1.0 Introduction The purpose of this study is to calculate the total (net) economic impact of an APS distributed solar NEM program in Arizona up to and including 2035. 1.1. Net Metering Net metering (NEM) encourages consumers to invest in renewable energy technologies by crediting them for distributed generation at the same tariff they pay for purchasing centrally—generated power. Originating in Idaho and Arizona in the early 19805, this utility resource usage and payment scheme allows customer meters to effectively run backwards whenever their own generation is in excess of their level of consumption. Customers use their generation to offset their consumption over an entire billing period, and only pay for their net power purchase per month: that is, the amount of electricity consumed minus the amount of electricity generated. NEM credits are, de facto, based on current centrally-generated power tariffs. Some suggest that NEM unfairly passes on the fixed costs of building and operating a transmission grid used by participants to non-participating customers. This is because residential and small business’ utility rates volumetrically recover all costs, including those that are fixed. Advocates typically counter this criticism by arguing that NEM customers bring benefits to the grid that equal or exceed the fixed costs they avoid paying for through self-generation, includingjob creation and other economic impacts. NEM is currently available in Arizona for a wide range of distributed generation renewables, including solar PV, solar thermal, wind, biomass, biogas, hydroelectric, geothermal, combined heat and power, and fuel cell technologies. The Arizona Corporation Commission (ACC) has not set a firm kilowatt-based limit on system size capacity. it simply stipulates that a system size cannot exceed 125% of a customer’s total connected load or electric service drop capacity. There is also no aggregate capacity limit for net-metered systems in Arizona. However, each utility is obliged to file an annual report listing the net metered facilities and their installed capacity for the previous calendar year. Approximately 38,000 of APS’ current 1.2 million customer base have distributed solar. Attachment ACB 28R 11 of 59 1.2. Economic impact Analysis An economic impact analysis measures the effect of a policy, program, project, activity or event on a national, state or local economy, with particular emphasis on three types of effects or impacts. These are the direct, indirect and induced impacts: 0 0 Direct impacts include the initial capital investment when a business, policy or program is launched, and the people directly employed to manufacture a product, provide a service or deliver a program. Indirect impacts are the economic growth or decline resulting from inter-industry transactions or 0 supplier purchases, such as a distributed solar installation company’s purchase of solar modules. Induced impacts occur when the workers either directly or indirectly associated with an organization, policy or program spend their incomes in the local economy, when suppliers place upstream demands on other producers, and when state and local governments spend new tax revenues. The indirect and inducedeconomic impacts are second order expenditures and jobs created as a result of the initial "injection" of expenditure and direct jobs. For example, a utility employee hired to administer a NEM program would represent a direct job. Purchases made by a utility are indirect impacts; and the income that the utility or supplier companies’ employees spend in the local economy will in turn create revenues/income for a,variety of other businesses, generating induced effects. The second and later rounds of indirect and induced expenditure are not self-perpetuating in equal measure. Through time, they become smaller as more of the income/expenditures “leak” out of the examined economy.10 The cumulative effect of the initial and latter rounds of expenditure is known as the multiplier effect. There is no one "magic" multiplier estimate for every conceivable scenario. Due to the inter-linked nature of the State of Arizona’s economy and its links to the rest of the US. (and the world), the eventual ripple effects depend on numerous factors.11 A full understanding of the total impact that a specific energy policy will have on an economy is therefore rather more complex than just an extrapolation of direct impacts. 1° For example, in the form of savings, or payments on goods and services produced outside of the state. 11 In very simple terms, what matters is the size ofthe direct impact, where it occurs (that is, which county/state and which sector of the economy) and the duration of the impact. Attachment ACB 28R 12 of 59 1.3. Study Overview To help position APS’ service territory study and provide a context for its findings, Section 2 begins with an overview of economic impact modelling approaches to renewable energy, summarized in the form of a 3 x 2 matrix. Fourteen published analyses drawn primarily, but not exclusively, from the U.S., and additional insights from Canada, Germany, and Spain (listed in Table 1) are reviewed by Seidman in Section 3, with a particular focus on assumptions, methods and conclusions. Examining the varying magnitude of the employment and gross state product (GSP) impacts for each of the different types of study defined by the economic impact model matrix in Section 4, a clear rationale for Seidman’s approach to assess the economic impact of distributed solar deployment in the APS service territory is also provided. Sections 5 — 9 then examine the economic impact of three distributed (rooftop) solar deployment scenarios in the APS service territory for the study period 2016-2035 in the State of Arizona and Maricopa County. The analyses include the legacy effects of each scenario throughout the (assumed) 30 year economic life of the solar systems.12 Section 5 introduces the 3 solar deployment scenarios assessed for APS. These are: 0 A low case scenario, which assumes 1,300 MWdC of nameplate distributed solar PV installations by 2035 in the APS service territory, which will increase APS’ total number of distributed solar customers to approximately 150,000 accounts; 0 An expected or medium case scenario, which assumes 5,000 deC of nameplate distributed solar PV installations by 2035 in the APS service territory, which will increase APS’ total number of distributed solar customers to approximately 690,000 accounts; and 12 Based on the assumed 30 year economic life of the distributed system, the customer financing costs of solar installations, 20162035, will not be completed until 2065. The REMl model used currently only provides economic impact estimates up to and including 2060, but Seidman does not believe that this will materially affect the conclusions in the analysis. If the economic life of an installation is less than 30 years, the negative economic consequences are in all probability greater than the estimates presented in this study. Attachment ACB 28R 13 of 59 A high case scenario, which assumes 7,600 dec of nameplate distributed solar PV installations by 2035 in the APS service territory, which will increase APS’ total number of distributed solar customers to approximately 1,050,000 accounts. Table 1: Economic Impact Analyses Critiqued as Part of Current Study Geography California California Illinois __ Montana Montana Massachusetts Missouri U.S. , , , , ,_ Nevada New York Rhode Island Andalusia Germany Ontario Spain Title & Author(s) AECOM (July Economic and Fiscal Impact Analysis of Residential Solar Permitting Reform Vote Solar Initiative (April Economic and Job Creation Benefits of SB 43/AB 1014 Loomis, Jo and Alderman (December Economic Impact Potential of Solar Photovoltaics in Illinois Comings, Fields, Takahashi and Keith (June Employment Effects of Clean Energy Investment in Montana Energy and Telecommunications Interim Committee (January Quantifying the Economic Impacts of Net Metering in Montana Motamedi and Judson (March Modeling the Economic Impacts of Solar PV Development in Massachusetts Treyz, Nystrom and Cui (October Analyzing AMultireaipna' Vote Solar Initiative and Clean Energy Project Nevada Solar Jobs Now Proposal of andJobCreation Benefits NYSERDA (January New York Solar Study Berkman, Lagos and Weiss Distributed Generation Contracts Standard Program and Renewables Energy Fund: Jobs, Economic and Environmental Impact Study Cansino, Cardenete, Gonzalez, and Pablo-Romero Economic Impacts of Solar Thermal Electricity Technology Deployment on Andalusian Productive Activities: A CGE Approach Frondel, Ritter, Schmidt and Vance Economic Impacts from the Promotion of Renewable Energy Technologies The German Experience Pol/in and Garrett-Peltier Building the Green Economy: Employment Effects of Green Energy Investments for Ontario Alvarez, Jara, Julian and Bielsa (March Study of the Effects on Employment of Public Aid to Renewable Energy Sources Section 6 describes the simulation results for the low distributed solar deployment scenario. Section 7 presents the simulation results for the expected distributed solar deployment scenario. Attachment ACB 28R 14 of 59 Section 8 describes the simulation results for the high distributed solar deployment scenario. Conclusions are offered in Section 9. Attachment ACB - 28R 15 of 59 2.0 Economic Impact Assessment Methods There are a number of different approaches to an economic impact assessment. These are codified in Figure 1 below. Figure 1: Classification of Economic Impact Models COUNT GROSS PARTIAL GROSS GENERAL GROSS COUNT NET PARTIAL NET GENERAL NET Figure 1 illustrates two key distinctions among economic impact studies. The first distinction is between gross studies and net economic impact studies. Studies that are Gross in nature only consider the direct positive impacts of increased economic activity — in this case, solar generation. Net studies represent a more rounded form of economic assessment because they also account for the trade-offs in the economy which result from incentivizing one specific sector, such as the negative impacts on utilities and reduced spending and investment in other economic activities associated with increased solar activity. For example, a gross study might consider the positive effects of the installation of 100MW utility-scale solar on the level of economic activity alone, while a net study ofthe same installation would additionally allow for the negative economic impacts such as the decreased use of conventional forms of generation if these were displaced, and the net changes in residential, commercial and industrial energy bills. Consider also the installation of a distributed solar system by a homeowner. To meet a $30,000 cost of installation, the homeowner will forego spending the same $30,000 on something else, such as perhaps a new or refurbished swimming pool at their property. There are obviously positive economic effects associated with the homeowner’s investment in a distributed solar system, which would be captured in a gross economic study. However, in this example, there are also negative effects associated with the loss Attachment ACB 28R 16 of 59 of investment in the swimming pool, which are only ever considered alongside the positive benefits of the solar installation as part of a net study. Nine gross and five net studies are examined in Section 3. The gross studies are: 0 California: AECOM, 2011 0 California: Vote Solar Initiative, 2013 0 Illinois: Loomis, Jo & Alderman, 2013 0 Massachusetts: Motamedi 0 Montana: Comings, Fields, Takahashi and Keith (Synapse Energy Economics), 2014 0 Montana: ETIC, (2016) 0 Nevada: Vote Solar Initiative, 2011 o Andalusia: Cansino, Cardenete, Gonzalez and Pablo-Romero, 2013 0 Ontario: Pollin and Garrett-Peltier, 2009 Judson, 2012 The net studies are: Missouri & U.S.: Treyz, Nystrom and Cui, 2011 0 New York: NYSERDA, 2012 0 Rhode Island: Berkman, Lagos and Weiss (the Bratton Group), 2014 0 Germany: Frondel, Ritter, Schmidt and Vance, 2009 0 Spain: Alvarez, Jara, Julian and Bielsa, 2009 The second key distinction is between simple counts, partial (equilibrium) modeling, and macroeconomic (or general equilibrium) modeling. Counts are typically tallies of direct measures of economic activities, such as jobs, investments, or sales, without any attempt to capture the impacts ofthe inter-relationships with other economic sectors. As a result, counts can be more or less extensive in terms of their reach. Some just concentrate on counting the number of direct employees or assessing the level of sales within a specific economic sector, while others seek information about a sector’s entire supply chain. Counts can be made by surveys or by assessing theoretically the required inputs for the installation of defined amounts of solar capacity — for Attachment ACB - 28R 17 of 59 example, the first part of a JEDI model which estimates the number ofjobs created in the solar sector in a linear fashion based on the MW capacity of the solar installations. Studies examined in this report that use the counts method are: 0 Montana: ETIC, 2016 0 Germany: Frondel, Ritter, Schmidt and Vance, 2009 0 Ontario: Pollin and Garrett-Peltier, 2009 0 Spain: Alvarez, Jara, Julian and Bielsa, 2009 Partial models consider the wider effects of levels of activity in a specific economic sector, and are one of the most common commercial approaches in economic impact modeling. In contrast to counts, which generally assess the direct impacts of a change in the economy, partial models also consider the indirect and induced effects of the direct changes within a particular geography. The one drawback with partial models is that they do not consider the feedback effects of changed levels of an investment or economic activity such as, for example, the effect of large solar projects on wages in the labor market. Studies examined in this report that use the partial model method are: 0 California: AECOM 2011 0 California: Vote Solar initiative, 2013 0 Illinois: Loomis, Jo 0 Massachusetts: Motamedi & Judson, 2012 0 Missouri & U.S.: Treyz, Nystrom and Cui,, 2011 0 Montana: Comings, Fields, Takahashi and Keith (Synapse Energy Economics), 2014 0 New York: NYSERDA, 2012 0 Nevada: Vote Solar Initiative and Clean Energy Project Nevada, 2011 0 Rhode Island: Berkman, Lagos and Weiss (the Bratton Group), 2014 Alderman, 2013 General models consider the effects of levels of solar activity in an economy—wide context with reference to every economic interconnection and feedback effect. An example is computable general equilibrium (CGE) models. These model the entire economy and attempt to account for all ofthe impacts associated with a specific level of solar activity. Only one study examined in this report uses a general model to assess Attachment ACB 28R 18 of 59 impacts, due to the cost prohibitive nature of producing a CGE model for a state or a region. This is Cansino, Cardenete, Gonzalez and Pablo-Romero’s (2013) study of Andalusia. Figure 2 summarizes the studies examined in this report in terms ofthe method employed, and whether they consider positive impacts alone, or both positive and negative impacts. Figure 2: Classification of Studies Examined by Method o o , Pollin and GarrettPeltier, 2009 ETIC, 2016 0 o 0 o 0 Net Both positiveg_n_q negative impacts _._ o 0 . ._ ., Alvarez et al., 2009 Frondel et al., 2009 o 0 0 AECOM, 2011 Loomis, Jo & Alderman ,2013 Motamedi & Judson, VSI and Clean Energy Project Nevada, 2011 VSl, 2013 Comings et al- 201.4 NYSERDA, 2012 Treyz et al., 2011 Berkman et al., 2014 o Cansino et al. 2013 Attachment ACB 28R 19 of 59 3.0 Evaluation Framework and Review of Fourteen Economic Impact Analyses To objectively critique fourteen contemporary analyses of the economic impact of solar PV/renewables, Seidman uses the following questions as an evaluation framework: (a) What is the context for a study? (b) What are the study’s objectives? (c) Which geography is being studied? (d) What is the time-horizon of the study? (e) Which economic modeling tool is used? (f) What types of effects are modeled, with reference to Seidman’s 3 x 2 classification of economic impact models? (g) What are the key inputs and assumptions used in the modeling process, including the solar growth projection assumptions? (h) What are the key findings? The following tables in this Section provides Seidman’s assessment of each ofthe fourteen contemporary studies. Reference will also be made, where appropriate, when a particular study method is replicated in multiple geographies by the same authors. Attachment ACB 28R 20 of 59 _.,F99r!9m_ic andfiscal Impact Analysis of Residential _. ,. ._ Ju'v 2°11 ,_ Considers the impact of a 76% reduction in homeowner permitting costs for solar PV when scaled to the regional and state level, taking into account the projected growth in the industry through 2020. Objective(s) 0 Evaluate the economic and fiscal implications of a streamlined local government permitting system for installing residential solar PV. California Geography 2012-2020 Time Period lMPLAN Modeling Tool Type of Effects - This is a Partial Gross analysis, as it lacks detail on negative impacts considered. Examined o Considers a few more factors than the VSI reports, such as the initial down payment for a solar system which is positioned as a loss to homeowner savings and a gain to the solar industry. o It is at best a weak, borderline example of a net partial study as it does not: 0 Explicitly consider non—solar energy sector losses; 0 Take into account utility obligations from a transmission and distribution grid perspective in terms of savings, upgrades or modifications; 0 Quantity the impact of a reduction in the demand for centralized power generation due to increased distributed generation; 0 Remove the rebate dollars paid to homeowners and installers from the lMPLAN inputs; and 0 Consider the administrative costs associated with changing permitting rules. 0 Also questionably assumes that increased homeowner savings from reduced electricity bills will be spent in fullin-state. 0 Base case scenario uses California Solar lnitiative’s 2011 residential installation costs of Model Assumptions $6.97 per watt decreasing to $3.63 per watt by 2020. o Streamlined permitting would reduce annual costs by $0.38 per watt in 2020 (Le. from $6.10 per watt in 2011 to $3.25 per watt in 2020). Investment Tax Credit of 30% is assumed to continue through 2020. Average size of residential solar systems was 5.6 kW, 2012-2020. All solar systems will be purchased in California, albeit region unknown. Assumes solar in both cases will appeal to homeowners whose annual electricity bills would be reduced by at least 5% post-installation. Value of residential solar only impacts property taxes when the home is sold. Buyers will pay on average 3.6% more for solar PV homes. Growth 0 Projects 1,006,500 installations at 5 utilities’ service areas for current permitting, 2012Solar 2020; or an additional 131,500 installations for streamlined permitting. projection Assumptions 0 332 MW installed 2007-2011; 2,668 MW installed 2012-2020 without streamlined permitting (BAU case). Effects Scaled 0 Current permitting scenario assumes: per Year 0 73.5 job years created per total MW installed, amounting to 196,020 job years in total for the entire 2012-2020 period; (2015 5) 0 $1.24 million GSP per MW per year (2015 S); and o $69.70 per MW per year increase in additional sales tax, property tax, and payroll Title Auth°risi Background Attachment ACB 28R 21 of 59 Title Author(s) Background Similar Studies Objective(s) Geography Time Period Modeling Tool Type of Effects Examined Model Assumptions Solar Growth Projection Assumptions Effects Scaled per Year (2015 s) and Job creation Bene■ts of SB The Vote April 2013 SB43 and AB 1014 are two shared renewable pilot programs to enable residential renters and commercial customers to subscribe via PG&E, SCE, and SDGE to an offsite renewable project and receive a utility bill credit in return. VSl (2010) Colorado; o VSI (2011) Nevada; - VSI (2011) Iowa; and o The Solar Foundation (2013) Colorado. 0 Estimate the number of jobs created under SB 43/AB 1014, and the increased dollars that will subsequently circulate throughout the California _ California JEDI (based on IMPLAN l-O) version January 3, 2013 This is a Partial Gross analysis of two shared renewable programs. Study does not consider net job creation. It simply details the cumulative employment benefits of both proposed shared renewable programs, without taking into account the potential loss ofjobs in other energy sectors. State sales tax revenue and instate economic activity results are also exclusively considered from a shared renewable program perspective. Authors ignore the net changes that will in reality occur due to changes in other sectors of the state economy prompted by both programs, including the potential for higher energy bills. Crystalline Silicon —' fixed mount commercial; single axis tracking utility scale. For both pilots, study assumes the following local purchases: 0 100% of components for solar installations < 100 kW; 0 50% of components for 100 kW to 1 MW installations; and o 30% of components for installations > 1 MW. For both pilots, it also assumes the following local manufacturing: 0 10%-20% of components for installations < 1 MW; and 0 5-10% of components for installations > 1 MW. This amounts to 546 MW local total purchases for the implementation of both pilot schemes, and 91.5 MW to 183 MW local manufacturing. 2014—2016 construction period. 25 year operational phase. For SB 43, 53 MW installed in 2014, 161 MW installed in 2015, and 286 MW installed in 2016, resulting in a 500 MW pilot. For AB 1014,65 MW installed in 2014, 285 MW installed in 2015, and 650 MW installed SB 43 is estimated to have a gross jobs years/MW, MW per year, and $5,291 sales tax revenue per MW per year (2015 S). AB 1014 is estimated to have a gross jobs impact of 24.0 job years/MW, $175,000 GSP per MW per year, and $5,331 sales tax revenue per MW per year (2015 S). Attachment ACB 28R 22 of 59 Title Author(s) Background Objective(s) Geography Time Period Modeling Tool Type of Effects Examined Model Assumptions Solar Growth Projection Assumptions Effects Scaled perYear (2015 S) Potential of Solar Photovoltaics in Illinois Loomis, Jo and Alderman, December 2013 i0} Renewable Energy Commerce and Economic grant. Considers employment and output impacts for the construction and operations phases of 3 deployment scenarios, with 3 levels of Illinois 2014-2030 JEDI PV Model (PVS4.5.13) o This is a Partial Gross analysis. 0 It exclusively considers renewable (solar) sector impacts, including supply chain. I it does not consider corresponding impacts in other parts of the energy sector, or other economic sectors. 0 Installations profile: 0 10% residential (80% retrofits, 20% new construction); 0 10% small commercial; 0 20% large commercial; 0 60% utility-scale. o 100% local purchases: 0 Labor and soft costs (permitting and business overhead); and 0 Residential and small commercial materials and equipment. 0 All materials and equipment for large commercial and utility—scale installations are purchased 100% out-of—state. 0 Three levels of instate manufacturing per scenario — 0%, 5%, and 10%. 0 2,292 MW, 2714 MW, or 11,265 MW by 2030. - For all 3 scenarios at 10% in-state manufacture: o 12.2 gross job years per MW installed; 0 Approximately $107,000 GSP per MW per year (2015 S); and o labor income per MW_p_er year (2015 Attachment ACB 28R 23 of 59 Title Author(s) Background Modeling the Economic Impacts PV and Judson, March 28, 2012 (Unpublished REMI. commission for the New England Energy and Commerce Association Renewables and Distributed Generation Committee. Objective(s) - Assess the economic impact of the 0 Construction of 305 MW of solar PV, 2012-2018; and 0 Operation of solar PV installations, 2012—2025. Massachusetts Geography 0 2012-2018 construction; and Time Period c 2012-2025 operations. Modeling Tool REMI Type of Effects 0 Partial Gross study, which generically describes, but does not state, the value of inputs used.13 Examined 0 Energy cost savings are only considered savings perspective. Model I Combination of residential, commercial, and utility-scale solar installations, with Assumptions regional purchase coefficients of 0.629, 0.564, and 0.580 respectively. 0 Construction phase uses total investment after federal and state tax credit cost reduction, including some consumer consumption reallocation and production costs, along with consumer electricity price, and business electricity fuel cost changes. 0 Models locally supplied inputs as total construction spending. Consumer price of electricity, electricity fuel costs for businesses, and production cost to utilities are used to represent the energy cost savings; and analysis assumes no change to SREC Solar Growth 0 Additional 305 MW of PV, 2012-2018, taking total installation to 400 MW. Projection 0 Does not state the split between residential, commercial and utility—scale solar. Assumptions Effects Scaled 0 20.1 job years created per MW installed. per Year 0 Approximately $122,000 GSP per MW per year (2015 $). (2015 5) 0 Approximately $155,000 personal income per MW per year (2015 5). 13 Motamedi and Judson mention energy cost savings, implying some consideration of the negative economic impacts of solar deployment. However, their PowerPoint presentation does not include any obvious assessment of negative impacts, and the REMI output is not suggestive of their inclusion. As a result, Seidman has classified their approach as Partial Gross. Attachment ACB 28R 24 of 59 Title Author(s) Background Objective(s) Geography Time Period Modeling Tool Type of Effects Examined Model Assumptions A Multiregional Macroeconomic Framework for Analyzing Energy Policies Treyz, Nystrom and Cui, October 2011 and national economic study considering the local, and social impacts. Missouri’s RPS, excluding Compares effects of electricity price—cap mandate (Scenarios 1 and 2) and an alternative bond-funded cost-recovery strategy (Scenarios 3 and 4) to finance the subbing of wind and solar for coal. Missouri and the U.S. 0 Construction impacts (RPS implementation), 2011-2021. 0 Operational impacts, 2011—2035. 0 Partial Net study. 0 0 Baseline: No RPS implemented in Missouri. Scenario 1 = IOUs raise prices to statutory cap of 1% to recover low cost of subbing wind and solar for coal (cost fully recovered by 2023). Scenario 2 IOUs raise prices to statutory cap of 1% to recover high cost of subbing wind and solar for coal (cost fully recovered by 2025). Scenario 3 IOUs issue bonds with maturity of 15 years at 3.25% interest rates to raise funding needed for low cost infrastructure. Scenario 4 = IOUs issue bonds with maturity of 15 years at 3.25% interest rates to raise funding needed for high—cost infrastructure. In Scenarios 1 and 2: o 1% compound increase in commercial and industrial electricity prices; 0 1% compound increase in residential electricity prices, with lower disposable income corresponding consumption reallocation. In Scenarios 3 and 4: 0 Utilities issue bonds at bank prime rate of 3.25% per year for 15 years; 0 Impacts greater in the 20205 when consumers haVe to pay higher prices to pay off bonds, compared to 20105 when consumers pay the costs up front in Scenarios 1 and 2. In Scenarios 1—4: 0 Solar panel purchase and 0&M are treated as semiconductor manufacture exogenous final demand with corresponding consumption reallocation o IOU rebates accounted for in production cost and transfer payments; 0 Partial substitution of conventional electricity for solar electricity allows households to reduce conventional electricity consumption and expense, captured in consumption reallocation; and 0 Creation of a custom industry for commercial wind generation, to account for demands. RPS: Coal 66%, Wind 14.7%, Solar 0.3% and Other 20% from 2021 onwards. Coal declines from 81% of electric production in 2010 to 66% by 2021; wind and solar from 0% to 15%. Graphs rather than data tables are provided, creating difficulties for interpretation. A state RPS is assumed to cause a short-term decrease in local employment, real GDP and personal real disposable income per capita. Raising electricity prices is estimated to result in the loss of 4,000 to 5,000 job years by 2021 or 2025, before recovering to the same level as the 2010 baseline in 2031. A bond scheme is estimated to create an initial short term annual employment increase the trend reverses upon completion of the RPS in 2021, of up to 1,000 0 0 0 0 o o Growth Solar Projection Assumptions Scaled Effects per Year (2015 5) 0 0 0 0 0 o Attachment ACB 28R 25 of 59 decreasing by 2,000 to 3,000 jobs per year up until 2027, before recovering to a net decrease of 600-800 jobs by 2035. Real GDP would steadily decrease under the price-cap scenario, hitting a low of $350— $458 million loss in 2021 and 2025, before regaining some ground to a $102 million loss in 2035 (2015 $). The utility bond approach would have expand real GDP until 2021, peaking at $153$204 million in 2019, fading to a decrease of $306-$408 million in 2027, before picking up to a loss of 5153-244 million by 2035 (2015 S). Attachment ACB 28R 26 of 59 Employment Effects of Clean Energy Investment in Montana Comings, Fields, Takahashi and Keith (Synapse Energy Economics), 2014 Examines the employment impacts of hypothetical additions Montana’s _ energy portfolio. 0 Estimate employment impacts of construction and 0&M activities associated: Objective(s) o Large—scale wind; 0 Large—scale solar PV; 0 Small—scale solar PV (rooftop), and 0 Energy efficiency. Montana Geography 0 Installation of systems is assumed to take place in 2016-2017. Time Period _ 0 Assumes 20 years of system operation. model. JEDI NREL’s IMPLAN in conjunction with capacity data from Modeling Tool Type of Effects 0 Partial Gross study of direct, indirect and induced employment impacts. 0 Makes no attempt to consider net effects. Focused entirely on job impacts of solar Examined installation and 0&M spending and considers no other benefits of solar deployment. 0 Develops solar spending patterns associated with rooftop and utility-scale installations Model model with adjustments for local conditions. using NREL’s Assumptions 0 Estimates construction jobs in short-run and allocates them over 20 years together with 0&M to obtain a 20 year cumulative job impact per average MW deployed. Growth 0 No actual projections. Solar and 2 0 Uses NREL’s (2012) maximum hypothetical potential of 4,409 GW Projection GW rooftop solar PV for Montana. Assumptions Effects Scaled 0 Small PV — 9.2 job years per MW. , . per. PV;.EP M ,9 Per Year Title Author(s) Background Attachment ACB 28R 27 of 59 Title Author(s) Background Quantifying the Economic Impacts of Net Metering in Montana Energy and Telecommunications Interim Committee (ETIC), January 2016 Examines the historical economic development impact of net in 2014 and 2000—14 in Montana. Objective(s) 0 Evaluate economic development impacts of the installation of net metering systems in terms of the following benefits and costs: 0 Bill savings of net metering customers; 0 Residential property value increases; 0 Revenue generated by installations; 0 Employment from installations; 0 Value of avoided carbon emissions; 0 Costs of income tax credits; and 0 Universal System Benefits (USB) renewable energy and Research Development (R&D) allocations. Montana Geography 2000-2014 Time Period Modeling Tool Counts based on survey/modeling estimates from other states. Type of Effects 0 This is in fact not an economic impact study or a normal assessment of economic Examined development impacts. 0 It’s a partial Count Gross analysis that considers a limited set of costs and benefits associated with net metering system deployments. o The tax revenue estimates are unclear, incomplete and based on very general 355“ n■ptionsModel 0 Based mostly on Montana Renewable Energy Association (MREA) survey data. Assumptions 0 Uses NREL models to assess installation sales revenue based total installations each year but no specifics of the nature of the system(s) installed are given. 0 Employment outcomes are also based on survey work done by the Montana Environmental Information Center, Synapse Energy and the Sierra Club. 0 It is lacking in a number of aspects. It needs to: o Considerfu/l indirect and the induced impacts of net metering; 0 Use appropriate bespoke models for Montana reflective of local economic circumstances; and 0 Not rely on very general rule of thumb estimates for jobs, revenues and taxes generated as base data. 0 It double-counts historical property value and homeowner energy savings as separate benefits. Solar Growth 0 The extent of net metering systems installed in 2014 is stated as $4M (2014 S) but there Projection is no statement of the extent of system additions or their capacity between 2010 and Assumptions _ _ Effects Scaled 0 There is no statement of installed capacity per Year statement of GSP, employment or tax revenue. It is thus impossible to calculate a jobs impact per MW, GSP per MW per year, or sales tax revenue per MW. Attachment ACB - 28R 28 of 59 Title Author(s) Background Similar Studies Objective(s) Geography Time Period Modeling Tool Type of Effects Examined Model Assumptions Growth Solar Projection Assumptions Scaled Effects per Year (2015 5) Economic and Job Creation Bene■ts of the Nevada Solar Jobs Now Proposal of , , , ,, . Vote Solar Initiative and Clean Energy Project Nevada the economic impact of expanding Nevada’s DG solar market from 35 MW to 400 _ MW between 2011 and 2020. 0 VSI (2010) Colorado; 0 VSI (2011) Iowa; 0 VSI (2013) California; and The Solar Foundation (2013) Colorado. - Evaluate the economic, job benefits and tax impacts of expansion of and changes to the incentive structure of Nevada’s Solar Jobs Now proposal of 2011. Nevada 2011-2020 NREL’s Jobs and Economic Impacts (JEDI) model. 0 This is a very simplistic and rather opaque Partial Gross analysis since it lacks any consideration of the negative impacts of expansion. 0 It is biased in terms of its assessment of economic impacts since it does not: 0 Consider any non-solar energy sector losses; 0 Take into account utility obligations from a transmission and distribution grid perspective in terms of savings, upgrades or modifications; 0 Quantify the impact of a reduction in the demand for centralized power generation due to increased distributed generation; 0 Consider the economic impacts of rebate dollars paid to D6 homeowners and installers; 0 Examine the economic impacts of reduced spending on other categories of expenditure throughout the expansion phase from capital expenditures on DG solar systems; and 0 Consider the administrative costs associated with changing permitting rules. 0 Base assumptions are drawn from a JEDI model specific to Nevada. 0 Basic premise is a growth of 365 MW in residential and commercial DG solar. No specifics about system characteristics used in the JEDI model are outlined in the , 0 365 MW installed 2011—2020. 0 Over the period 2011-2020, The SolarJobs Now Proposal is estimated to have: 0 A gross jobs impact of 28.5 job years/MW; 0 $443,400 GSP per MW per year (2015 S); and ___§22,500 sales tax revenue per MW (2015 S). Attachment ACB 28R 29 of 59 Title Author(s) York Solar Study York State Energy Research & Development Authority (NYSERDA), January 2012 New York Act of 2011. Study required by The Evaluate the cost—benefits of increasing solar PV in NY to 5,000 MW by 2025. New York State 2013-2049 REMI Partial Net study. 0 Quantifies direct PV job impacts of each scenario, economy—wide net impacts, gross state product, retail rate impacts, and environmental impacts. 0 Economy-wide netjob analysis includes: 0 Positive impacts such as the creation of new PV jobs, and ratepayer savings when electricity prices are suppressed by PV output; and 0 Negative impacts, such as the cancellation of new power plants that are made unnecessary by the added PV capacity, or the additional costs of PV incentives, which reduce personal disposable income. 0 Net retail impact of PV deployment includes: 0 The above—market costs of PV; 0 Net metering costs; and 0 Savings generated by the suppression of wholesale electricity prices. 0 Net environmental impacts include: 0 Lower emissions via a reduction in the need for fossil fuel plants; and time. 0 Land use changes from rooftop to scenarios: Three 0 Model 0 Low Cost Scenario, using DOE SunShot goal for PV cost reduction, assuming Assumptions extension of the federal tax credit (FTC) through 2025; 0 Base Case Scenario, using a DOE survey and moderate reduction of FTC beyond 2016, plus costs of $2.5 million/MW for large-scale and $3.1 million/MW for smallscale installations; and 0 High Case Scenario, based on the national average annual PV system price decline ' over the past decade, with FTC reverting to a pre—federal stimulus level in 2016. o 5% of solar components are manufactured in NY; the rest are imported. o Incentive costs are recovered from ratepayers through their electricity bills. 0 Quantified benefits of the 5000 MW by 2025 goal include a wholesale price suppression assumption, a reduction in energy lost to transmission and distribution inefficiencies, a reduction or deferral of the need to upgrade the utility distribution system, avoided RPS compliance costs, and a monetized carbon value of $15 per ton. Growth o Achieve 5,000 MW solar PV deployment by 2025. Solar 0 Four policy options are analyzed to stimulate demand: Projection 0 Utilities obliged to purchase tradable solar renewable energy credits (SRECs) from Assumptions spot market, supported by a price floor mechanism to provide greater degree of revenue certainty; 0 Utilities manage a competitive procurement similar to CA in which they award longterm contracts to purchase renewable energy; 0 Residential and commercial small PV system rebates, and larger systems incentives, provided centrally via competitive bidding; and 0 Utilities incentives for larger projects through competitive long—term contracts, and a cents per kWh produced for smaller projects. Background 0bjective(s) Geography Time Period Modeling Tool Type of Effects Examined Attachment ACB 28R 30 of 59 Effects Scaled per Year (2015 S) o 0 0 created per dependent on scenario, 2013-2025. 700 economy—wide jobs net gain (low) or 750 to 2,500 economy-wide jobs net loss (base and high), 2013-2049. $15,760 GSP per 'MW per year gain (low), or $16,930 to $58,386 GSP per MW per year loss (base and high), 2013-2049 (2015 S). Attachment ACB 28R 31 of 59 Title Distributed Generation Standard Contracts Program and Renewables Energy Author(s) Background Berkman, Lagos and Weiss (The Brattle Group), 2014 0 Prepared for the Rhode Island Office of Energy Resources and Commerce as stipulated by the July 2013 Distributed Generation Standard Contracts (DGSC) Law. 0 Examine the potential economic, fiscal and environmental impacts of the Distributed generation Standard Contract (DGSC) and Renewable Energy Fund (REF) 20134—2038. Island 2014-2038 IMPLAN in conjunction with energy capacity planning and energy dispatch models 0 A Partial Net study in terms of its economic impact assessment. 0 Includes spending on installations as a gross addition to final demand. 0 Does not net out the associated purchase/leasing costs which would likely swamp installation spending. 0 Includes payments to DGSC/REF participants but no allows no countervailing reduction in non-DGC ratepayers’ spending. Costs to ratepayers are assessed but not included in the economic impact assessment. Assess central generation capacity and operating costs with a capacity planning and economic dispatch model. 0 Includes both wind and solar renewable energy. 0' Operational life span of renewable resources assumed to be 25 years. 0 Source metrics for with and without DGC and REF scenarios obtained from past studies. 0 Use secondary sources to assess central generation and capacity costs using approximations rather than primary modeling. o It is unclear how DGSC/REF capacity deletions/additions are assessed to affect central generation costs. 0 Three (assumed not forecast) scenarios above 2013 40 MW are assessed: o 160 MW (by 2019) with REF of $800,000 in solar installations; 0 200 MW (by 2019) with REF of $800,000 in solar installations; and REF of $1,600,000 in solar installations. 1,000 MW (by MW: per GSP annual 0 Average 0 160 MW DGC:$191,790 GSP per MW (2015 S); o 200 MW DGC: $182,216 GSP per MW (2015 S); and 0 1,000 MW DGC: $135,290 GSP per MW (2015 S). 0 Average annual job years per MW: 0 160 MW DGC: 1.53 jobs; 0 200 MW DGC: 1.465 jobs; and 0 1,000 MW DGC: 1.095 jobs. Objective(s) Geography Time Period Modeling Tool Type of Effects Examined Model Assumptions Growth Solar projection Assumptions Effects Scaled per Year (2015 5) Attachment ACB 28R 32 of 59 Title Economic Impacts of Solar Thermal Electricity Technology Deployment on Author(s) Background Cansino, Cardenete, Gonzalez and Pablo-Romero, 2013 Annals of Regional Science published paper estimating the impact on productive activities of increasing theproduction capacity of two types of solar thermal plant in Andalusia. Objective(s) 0 To quantify the gross direct and induced productivity impacts of a single parabolic trough solar collector power plant and a single solar tower plant for the Andalusian economy. 0 To also quantify the gross direct and induced productivity impacts of both types of solar thermal technology based on the addition of 789 MW installed capacity by 2013 to comply with the Sustainable Energy Plan for Andalusia (PASENER). Andalusia (Spain) Geography Time Period 0 2008—2013 installation; and 30 year estimated lifetime foreach plant. Modeling Tool Static computable general equilibrium (CGE) model, consisting of 27 productive activities in the Andalusian Type of Effects 0 General Gross study.14 Examined o Describes gross economic impacts by sector, based on an enlarged electricity sector and non—renewables and prevents any substitution. Model 0 Walrasian notion of competitive equilibrium, extended to include producers, Assumptions households, government, and foreign sectors. 0 The single representative consumer maximizes a Cobb—Douglas utility function. 0 Government maximizes a Leontief utility function. 0 Foreign sector is modeled as a single sector that includes the rest of Spain, the European Union, and the rest of the world. 0 Benchmark equilibrium scenario includes a perfect inelastic supply of capital and positive unemployment rate, and a fixed level of government and foreign sector activities which allows relative prices, activity levels, public deficit and foreign trade deficit to work as exogenous variables. - Equilibrium is defined as an economic state in which the representative consumer maximizes his utility, the 27—sector productive activities maximize their profits after taxes, and public revenue is equal to the payments to the different economic agents. 0 Does not consider if Andalusia’s gross output gains are at the expense of other states’ output e.g. from the crowding-out effect of power generation. Solar Growth 0 For the single plant analysis: projection o 50 MW parabolic trough plant with 624 collectors; and Assumptions 0 17 MW solar tower plant with 2,750 heliostats. 0 Estimated lifetime of each plant is 30 years. 0 For the PASANER scenario, to meet the 800 MW target by 2013 (789 MW additions), the model assumes 80% parabolic trough and 20% solar tower. Effects Scaled - Scenario 1 (single plant additions) is estimated to result in an economy—wide gross per Year productivity increase of 0.75% for the parabolic trough plant, ora 0.68% gross productivity increase for the solar tower plant. 0 Scenario 2 (PASANER) is estimated to result in an economy-wide gross productivity increase of 35.37% over the 30-year lifetime of the parabolic trough and solar tower plant additions (30.81% parabolic trough; 4.57% solar tower). 1“ Cansino et al. use a 27-sector CGE model that is a general modeling representation of the Spanish economy, allowing for both positive and negative feedback effects of increased levels of solar penetration in Andalusia. However, they model renewables and non—renewables as a single sector that does not allow for substitution between forms of generation, which means that they are effectively only allowing for positive direct demand shocks in their modeling. This is why Seidman classifies their approach as a General Gross model. Attachment ACB 28R of 59 Title Author(s) Background Objective(s) Geography Time Period Modeling Tool Type of Effects Examined Model Assumptions Solar Growth Projection Assumptions Effects Scaled per Year (2015 S) Economic Impacts from the Promotion of Renewable Energy Technologies German Experience Schmidt and Vance, Critically reviews cost and job implications of the Renewable Energy Sources Act (EEG) — the centerpiece of the German promotion of renewable energy. This guaranteed stable feed— in-tariffs (FlTs) for up to 20 years, and also favorable conditions for investments in green electricity production for the long—term. To demonstrate the impact of government—backed renewable incentives for stimulating the Germany 2000-2020 Non-Applicable 0 Count Net study which balances gross renewable sector gains with: o The losses that result from the crowding out of cheaper forms of conventional energy generation; and o The drain on economic activity precipitated by higher electricity prices, including a loss of consumer spending power, and lower total investments of industrial energy consumers. 0 Also notes that: 0 New green jobs are often filled by workers who were previously employed, leading to a further overestimate of gross jobs effects; 0 Energy security benefits of solar PV are undermined by reliance of imported fossil fuel sources to meet technological demand; and o Technological innovation is stifled via a subsidy that compensates an energy technology for its lack of competitiveness. 0 Assesses real net present cost of solar subsidies, based on the volume of solar generation, the FIT, and conventional electricity prices. 0 Specific net cost per kWh = difference between solar FIT and market prices at the power exchange. - Utility central station generation costs of 2-7 cents/kWh 0 Utilities obliged to accept delivery of power into their own grids from independent renewable producers o Solar—specific FlT of 50.62 cents/kWh paid by utilities in 2000 falling to 43.01 cents/kWh in 2009. I If solar subsidization ended in 2009, electricity consumers would still face charges until 2029. o Assumes 2% annual inflation. 0 Cost estimates for PV modules installed 2000-2008 are based on an overall solar electricity production of 96 billion kWh during 20 years of subsidization. 0 Germany had 5,311 MW installed PV capacity in 2008. 0 Net cost promoting Solar PV per MW installed: $3.18 million, 2000-2008 (2015 S).15 15 €2.2 million (2007 €) converted to USS at a rate of US$12 €0.7687. Attachment ACB 28R 34 of 59 Title Author(s) Background Objective(s) Geography Time Period Modeling Tool Type of Effects Examined Model Assumptions Growth Solar Projection Assumptions Scaled Effects perYear for Ontario , University of Massachusetts-Amherst study sponsored by the Green Energy Act Alliance, Blue Green Canada, and World Wildlife Fund (Canada). o Considers the employment benefits of two Ontario green investment agendas: 0 Baseline Integrated Power System Plan (IPSP): $18.6 BN investment over 10 years in conservation and demand management, hydroelectric, on-shore wind, bioenergy, waste energy recycling and solar power; and o Expanded Green Energy Act Alliance (GEAA): $47.1 BN investment over 10 years in , windend 10 years - Author-modified provincial l-O tables for Ontario, combined with national I-O tables for Canada to construct wind, solar, biomass and building retrofitting as industries in their own right. Also uses U.S. data (BLS 2007 Occupational Employment Survey) to determine which occupations are likely to be in high demand for each of the 8 renewable energy areas considered. Count Gross study, addressing employment. No comparison is made with alternative, non—green investments. Neither do they consider if a green investment program is the most effective way to generate jobs in the region. Uses three factors to establish relative employment effects of alternative green investments: 0 Labor intensity of spending -— that is amount spent on workers rather than land, energy, or materials; 0 Local content of spending; and o Wage rates. 3% of baseline IPSP spending is allocated on an annual basis to solar. allocated on an annual basis to solar. 16% of expanded GEAA 88 MW of solar energy supplied over 10 years for baseline IPSP. 1,738 MW of solar energy supplied over 10 years for expanded GEAA. 89.7 gross job years GEAA: 68.7 gross job years per MW installed. Attachment ACB 28R of 59 Title Author(s) Background Objective(s) Geography Time Period Modeling Tool Type of Effects Examined Model Assumptions Solar Growth Projection Assumptions Effects Scaled per Year of the Effects on Employment of Public Aid to Renewable Energy Sources Alvarez, Jara, Julian and Bielsa, March 2009 Universidad Rey Juan Carlos study part-funded by DG TREN (Energy & Transport) of the European Commission. To demonstrate the extent to which government support for green jobs in Europe has been economically counterproductive. Non-Applicable 0 Count Net study. - Compares average amount of subsidized investment needed to create a solarjob with the average amount of capital needed for a job in the private sector. 0 Also compares the average annual productivity that the solar job subsidy would have contributed to the economy had it not been consumed in public financing, with the average productivity of labor in the private sector that allows them to keep their job. The total subsidy to PV, wind, and hydro since 2000 is $36 billion. No additional solar plants have been constructed since December 2008. o $12.1 billion has been committed for PV generation, 2000-2008. - Assumes that Spain has installed 2,934 MW solar PV by 2008. o 0 For every renewable energy job financed by government, on average 2.2 jobs will be lost in the private sector. However, for every solar MW installed, 8.99 private jobs are destroyed as a result of related regimes- Attachment ACB 28R 36 of 59 4.0 Economic Impact Analyses — Magnitudes & Preferred Modeling Methods Gross (positive impact only) studies clearly produce higher estimates of the economic impacts of solar enhancements than net studies, as demonstrated by the studies reviewed in Section 3. It is also important to note that gross studies are uniformly positive, while net studies are generally negative in terms of divined economic impact. The principal effect of using a partial model approach rather than a count approach, or using a general (macroeconomic) modeling approach rather than a partial approach, is to reinforce the magnitude ofthe divined economic impacts. Thus, using a general (macroeconomic) model approach yields the most significant gross and negative studies. Figure 3 summarizes the magnitude of impacts by type of economic impact study, based on the studies critiqued in Section 3. Counts usually quantify the number of jobs. The Ontario Count Gross analysis reviewed in Section 3 estimated 68.7 to 89.7 gross (direct) job years are generated for every MW of wind and solar energy installed, which averages out at 69.74 for both renewable programs. The Spanish Count Net analysis reviewed in Section 3 estimates that 8.99 private jobs are lost through "green jobs” mandates, subsidies and related regimes, for every 1 MW of solar installed. Frondel et al. do not provide actual job counts for their German Count Net analysis. They simply conclude that ”...any result other than a negative net balance ofthe German PV promotion would be surprising” (p. 17), based on a per capita subsidy of $257,400 in 2008, the EEG’s crowding out effects, negative income effects and the unprecedented competition from cheaper Asian imports.16 Partial model estimates extend beyond a count to additionally estimate Gross State Product (GSP). The Partial Gross models reviewed in Section 3 estimated 5 to 73.5 gross job year gains per MW installed, and 15 Frondel et al. report that in 2006 and 2007, almost half of Germany‘s PV demand was covered by imports, most notably from Japan and China. Attachment ACB 28R 37 of 59 a GSP gain of$106,800 to $1.24 million per MW installed per year (2015 $). The AECOM study appears to be something of an outlier, as the gross job year estimate for the three other studies ranges from 5 to 24.9 job years per solar MW installed. Four of the studies in this section estimate GSP contributions of $106,800 to $176,354 GSP per MW per year (all 2015 $). The two exceptions, estimating significantly higher GSP contributions per MW per year are VSl (2011) in Nevada, and the AECOM study. NYSERDA’s Partial Net model estimates a 700 economy-wide net gain in job years for their low case scenario, but a 750-2,500 economy—wide net loss for job years for their base and high case scenarios. Similarly NYSERDA estimate a $15,760 GSP net gain per MW installed per year for their low case scenarios, compared to net losses of $16,930 to $58,386 per MW installed per year for their base and high case scenarios (all 2015 $). Treyz et al. only present graphs, rather than actual data, which appear to show a net negative loss in both job years and GDP, 2011-2035. Figure 3: Magnitude of Economic Impacts 0 70 gross job years per MW 0 -8.99 privatejobs per MW per year Gross Only positive o_r negative impacts Net Both positive g_n__d negative impacts Partial Models Range of 5 to 73.5 gross job years per MW. Range of $106,830 to $1.24 million GSP per MW per year. Range of +750 to — 2,500 net job years per MW, dependent on the scenario. Range of +$15,862 to GSP per MW installed per year, dependent on the scenario. 17 This is based on the PASENER target, 80% of which would be met by parabolic trough. 18 This is based on the PASENER target, 20% of which would be met by solar tower. General Models total $7,198 production per MW installed per year for trough parabolic installations.17 total $4,265 production per MW installed per year for tower solar installations.18 Attachment ACB - 28R 38 of 59 The General Gross model reviewed in Section 3 offers two solar-technology dependent estimates. These are a total gross productive increase of $7,075 per MW installed per year for parabolic trough; and $4,192 per MW installed per year for solar tower.19 Based on the 6-way matrix of economic impact studies initially presented in Section 2, the implementation of a General Net analysis of solar deployment in the APS service territory, 2016-2035 is the best methodological approach forthe current study. However, to the research team’s knowledge, a CGE model ofthis nature currently does not exist for the State of Arizona; and it would be cost prohibitive to test and develop a CGE model for the State of Arizona in a short time frame. As a result, the current study implements a Partial Net analysis of solar deployment in the APS service territory, 2016-2035, presented in Sections 5 8. Seidman expects the results presented in the subsequent Sections to be directionally correct, but possibly understated, compared to a General Net (CG E) approach. 19 This uses an IRS 2013 dollar—euro annual currency exchange rate of U551: €0.783. Source: IRS (2014), downloaded at www.irsgov/IndividuaIs/International-Taxpayers/YearIy-Average-Currency-Exchange-Rates. Value is then converted into 2015 s using the Bureau of Labor Statistics CPI Inflation Calculator. Attachment ACB 28R 39 of 59 5.0 Economic Impact of Net Metering — Scenarios, Assumptions and Method 5.1. Scenarios and Assumptions Three distributed (rooftop) solar deployment scenarios in the APS service territory are assessed for the study period 2016—2035, including the legacy effects of each scenario throughout the (assumed) 30 year economic life ofthe solar systems.20 The solar deployment scenarios assessed for APS are: 0 A low case scenario, which assumes 1,300 deC of nameplate distributed solar PV installations by 2035 in the APS service territory, which will increase APS’ total number of distributed solar customers to approximately 150,000 accounts; 0 An expected or medium case scenario, which assumes 5,000 MWdc of nameplate distributed solar PV installations by 2035 in the APS service territory, which will increase APS’ total number of distributed solar customers to approximately 690,000 accounts; and o A high case scenario, which assumes 7,600 dec of nameplate distributed solar PV installations by 2035 in the APS service territory, which will increase APS’ total number of distributed solar customers to approximately 1,050,000 accounts. Distributed solar deployment is assumed to take place throughout the period of study in each scenario — that is, up to and including 2035. Approximately 86% ofthe solar installations are assumed to occur in Maricopa County, 5% in Pinal County, and 9% in Yuma County in each scenario. The capital costs and financing implications of each solar deployment scenario is determined by examining the level of distributed generation as forecast by APS using generic assumptions about the costs of standard DG solar systems and financing parameters. NREL’s JEDI model for solar installations is used to 2° Based on the assumed 30 year economic life of the distributed system, the customerfinancing of solar installations, 20162035, will not be completed until 2065. The REMI model used currently only provides economic impact estimates up to and including 2060, but Seidman does not believe that this will materially affect the conclusions in the analysis. If the economic life of an installation is less than 30 years, the negative economic consequences are in all probability greater than the estimates presented in this study. Attachment ACB 28R 40 of 59 distribute the capital costs ofthe solar installations throughout the supply chain in the State of Arizona.21 Figure 4 summarizes the breakdown of the JEDI model’s solar system costs used in this analysis. This is based on national industry averages, and may not match Arizona’s experience exactly, but is nevertheless widely accepted as a reasonable approximation. Administrative and support services account for an estimated 40% of solar system costs. This probably includes general administrative costs associated with state government permitting and federal rebates, and also local administrative costs in the solar industry. Figure 4: JEDI Model Exogenous Final Demand Categories 4% 3% 11% Fabricated metal product manufacturing I Computer and electronic product manufacturing 8 Electrical equipment and appliance manufacturing Construction I Professional and technical services Administrative and support services Source: Authors’ Calculations APS has also supplied Seidman with an estimate ofthe financial impact of each solar deployment scenario on the utility’s operating cash flow, future central station generation investments, and electricity retail rates. Approximately 70% of the deferred or cancelled central station generation investments occurring under the three distributed solar scenarios are assumed to occur in Maricopa County, with the balance in Pinal County. The investment changes included in the economic impact model are: 0 The annual installed costs of distributed solar capacity, 2016-2035; and 21 NREL’s JEDl models are an open-source, Excel—based, user-friendly tools that estimate the economic impacts of constructing and operating power generation and biofuel plants at the local and state levels. To find out more about the JEDl models, see http://www.nrel.gov/analysis/jedi/aboutJedi.html Attachment ACB 28R 41 of 59 0 APS’ deferred or avoided central station generation investments, 2016-2035. The long-term legacy costs included in the economic impact model are: 0 The customer leasing costs of distributed solar installations, 2016—2060;22and - Consumer electricity rate savings, 2016-2060, from the study period’s deferred or avoided central station generation. The timeframe of three of these elements extends beyond the last year of deployment (2035). This is because there are legacy effects associated with the deployment of distributed solar. For example, any customer installing a distributed solar PV system will have to meet the financial costs of that system for up to 30 years after the system has been installed on their roof. A utility is also required to recoup any investment in central station generation investments via retail electricity rates over the lifetime of that investment — again, usually 30 years. The legacy effects are therefore accounted for in the analysis. The modelling elements are discussed in more detail in Section 5.2. 5.2. Study Method Given the absence of a CGE model for the State of Arizona, Section 4 recommended the implementation of a Partial Net analysis of solar deployment in the APS service territory, 2016-2035. As a result, this study makes use of an Arizona-specific version ofthe REMI regional forecasting model, updated at the Seidman Research Institute, to produce partial net estimates of the impact on the Arizona economy of changes in the economic environment in the state. REMI is especially useful when examining the economic impact associated with the launch or expansion of a new program, such as NEM, in a particular region, state or country. Through its dynamic modeling, REMI takes account of variations in the economic impact of a program as it moves from the establishment 22 Based on the assumed 30 year economic life of the distributed system, the customer financing costs of solar installations, 20162035, will not be completed until 2065. The REMI model used currently only provides economic impact estimates up to and including 2060, but Seidman does not believe that this will materially affect the conclusions in the analysis. If the economic life of an installation is less than 30 years, the negative economic consequences are in all probability greater than the estimates presented in this study. Attachment ACB 23R 42 of 59 to operations phase, and also shows how estimates can vary through time. These estimated impacts are the difference between the baseline economy and the baseline economy augmented with the level of solar deployment assumed under each scenario. As a result, the analysis measures the Arizona economy up to 2035 with and without the existence of the new solar rooftop program. The use of a county level model also enables a more detailed disaggregation of results to occur, estimating the "leakage" of economic impacts into other counties in Arizona. Due to its overall flexibility, REMI allows for the examination of a whole host of different scenarios — different businesses and/or different construction and operations phases — while simultaneously providing estimates that are consistent across projects. The method for estimating the economic impact involves four fundamental steps: 1. Prepare a baseline forecast for the state and county economies: This Business As Usual (BAU) case forecasts the future path of state and county economies based on a combination of an extrapolation of historic economic conditions and an exogenous forecast of relevant national economic variables. 2. Develop a program or policy scenario: This scenario describes the direct impacts that each distributed solar deployment scenario could generate in APS’ service territory. 3. Compare the baseline and policy scenario forecasts. 4. Produce the “delta” results: Differences between the future values of each variable in the forecast results estimate the magnitude that each distributed solar deployment scenario could have on the state or county economies, relative to the baseline. The baseline or counterfactual scenario employed in this study assumes that there are no additions to the current stock of distributed solar installations over the period 2016-2035 in APS’ service territory. One consequence ofthis counterfactual scenario is that APS would need to add to both its central generation and transmission capacity, to cope with the increased load within its territory over the period. To cover the capital costs of the enhanced capacity and all subsequent operations and maintenance costs, APS would typically need to increase utility revenues over a 30-year period from the date of each investment. In isolation, this would manifest as a reduction in consumer spending, because utility customers would Attachment ACB 28R 43 of 59 collectively need to pay more for these new investments, and is also accounted for in the current study, up to and including 2060. In reality, some ofthis increased revenue will be provided by population growth which is creating the additional demand for new generation, and some will be offset by lower revenues for depreciating existing investments over time. 5.3. Solar Deployment Scenarios Three distributed solar deployment scenarios are analyzed in this study. To represent the effects of increased penetrations of distributed solar, three key changes are included in the current study for the 2016—2035 time horizon. These are: 0 The capital costs expended on rooftop solar systems purchased or leased by distributed generation customers, which are assumed to yield 20 years of construction-based benefits on the Arizona economy; 0 The financial payments made by utility customers for leased solar systems for the economic life of their assets. This represents a reduction in spending on other goods and services and, as such, a likely reduction in economic activity in Arizona; and o The reduction in revenue requirement for APS as a result of decreased net investment in centrally generated power. This represents a loss to the Arizona economy due to the reduction in central station generation construction and employment, offset by savings on fuel, 0&M and financing costs over time. Each scenario is modeled over a 20—year timeframe, starting in 2016 and ending in 2035, to estimate the employment, gross state product (GSP), and real disposable personal income (RDPI) for the State of Arizona and Maricopa County. However, there are also legacy effects associated with solar deployment and the deferral or cancellation of central station generation investments, which occur in the years immediately following an installation and last for the economic life of the solar installations. These legacy effects are therefore also included in the cumulative 2016-2035 estimate provided for each assessed economic measure, expressed in 2015 dollars (2015 23 The legacy effects for any 2035 distributed solar installations should last until 2065, to reflect the economic life of the system. The current REMI model is unable to provide estimates after 2060, but Seidman does not believe that this will materially affect the conclusions in the analysis. If the economic life of an installation is less than 30 years, the negative economic consequences are in all probability greater than the estimates presented in this study. Attachment ACB 28R 44 of 59 6.0 Simulation Results: Low Distributed Solar Deployment Scenario The low case scenario assumes that over $1.5 billion is invested in new distributed solar installations by 112,000 customers between 2016 and 2035, and the net deferral or cancellation of $85.5 million central station generation investments up to and including 2065 (all nominal S).24 Table 2 estimates the total employment impacts of the low case distributed solar scenario for the period 2016—2035. These are full—time (or equivalent) annual employment changes, applicable to all sectors and industries apart from government and farm workers. They include employees, sole proprietors and active partners, but exclude unpaid family workers and volunteers. The data is expressed in job years. The label "job year” is important and should not be simplified or abbreviated to "job". A "job year” is defined as one person having a full-time job for exactly one year. This means, for example, that one employee holding the same position at the same organization throughout 2016—2035 will account for 20 job years, but also only represent 1 job. Table 2: Total Private Non-Farm Employment impacts 2016-2035 (including Legacy Effects to 2060) Geography State of Arizona Maricopa County Source: Authors’ Calculations Job YearsZS -16,595 -15, 685 Table 2 suggests that the low case distributed solar scenario could have a negative employment impact of 16,595 full-time (or equivalent) job years in the State of Arizona throughout the 2016-2035 period of study, including any legacy impacts up to 2060. This legacy effect accounts for the fact that the true effects of the distributed solar deployment are only experienced in full after the period of study (20162035), consistent with the economic life of each solar installation.26 In Maricopa County, there is a negative employment impact of 15,685 job years for the study period as a whole (including subsequent legacy effects). 24 This simply reflects a deferral from the base case. 25 A job year is equivalent to one person having a full-time job for exactly one year. 25 The legacy effect should continue up to and including 2065. However, REMI currently does not allow for any analysis beyond 2060. If the economic life of an installation is less than 30 years, the negative economic consequences are in all probability greater than the estimates presented in this study. Attachment ACB 28R 45 of 59 Table 3 summarizes the industry sectors impacted the most by the low case distributed solar scenario. Table 3: Statewide Employment Impacts by Industry Sector (Job Years)27 Sector Forestry, Fishing, and Related Activities , , Construction Manufacturing Transportation ,,and Warehousing . , Finance and Insurance Real Estate and Rental and Leasing Professional Services Management of Companies and Enterprises Administrative and Support Educational Services and Recreation Accommodation and Food Services Administration Total Net Change in Job Years Total Number of Job Years Lost in Non-Solar Industry Sectors29 Source: Authors’ Calculations _, ,_ , ,. Total Job Years, 2016-206028 -2 -639 -2,549 -385 -548 -3,102 —514 -845 -3,505 -89 -440 -406 —1,348 -16,595 22,042 The table suggests that administrative and support services could benefit from the low case distributed solar scenario in terms of employment created. However, all other sectors are estimated to experience job losses, resulting in the total net estimate of 16,595 job years lost statewide. The administrative gain probably originates to a large extent from the permitting of solar installations, and also business support functions within the solar industry. The sectors estimated to experience the biggest job losses (expressed 27 A job year is equivalent to one person having a full—time job for exactly one year. 23 Total job years may not tally due to rounding—up. 29 This is a summation ofthejob years lost in non-solar industry sectors negatively impacted by the deployment of new distributed solar, 2016-2035. Attachment ACB 28R 46 of 59 in cumulative job years) during the study period in rank order are professional; scientific and technical services; health care and social assistance; retail trade; the construction industry; and utilities. Table 4 estimates the cumulative gross state product (GSP) and real disposable personal income impacts (RDPI) associated with the low case distributed solar scenario for the period 2016-2035. Table 4: Total Gross State Product (GSP) and Real Disposable Personal Income Impacts (RDPI) 2016-2035 (including Legacy Effects to 2060) Geography State of Arizona Maricopa County Source: Authors’ Gross State Product Millions (2015 $) -$4,806.6 -$4,491.8 Real Disposable Personal Income Millions (2015 S) -$1,787.3 -$1,862.4 Table 4 shows that in aggregate terms during the study period 2016—2035, and including legacy effects, total GSP could be cumulatively lower by over $4.8 billion (2015 S) in the State of Arizona. This includes an estimated $4.5 billion GSP lost in Maricopa County (2015 $). Table 4 also shows that in aggregate terms during the study period 2016-2035, and including legacy effects, RDPI is estimated to be cumulatively lower by almost $1.8 billion (2015 S) in the State of Arizona. This includes an estimated fall in RDPI of over $1.86 billion in Maricopa County (2015 The employment, GSP, and RDPI losses associated with the low distributed solar deployment scenario are valid, because the total amount of money paid by distributed generation and central station generation electricity consumers over the relevant time period (which extends beyond 2035) is greater than the amount which would have been paid had they all instead continued to draw electricity from the utility’s central grid. In short, electricity consumers are paying more for the same amount of electricity consumed under the low distributed solar deployment scenario, and therefore have less money to spend in other parts ofthe economy. 3" Some of Maricopa County’s estimated losses in RDPI will be offset by minor gains in other counties, thereby resulting in a negligibly smaller loss for the State as a whole. Attachment ACB 28R 47 of 59 7.0 Simulation Results: Expected Distributed Solar Deployment Scenario The expected or medium case scenario assumes that approximately $8.9 billion in total is invested by 650,000 customers in distributed solar installations between 2016 and 2035, and the deferral or cancellation of $194 million central station generation investments (all nominal 5).31 Table 5 estimates the total employment impacts of the expected or medium case distributed solar scenario for the period 2016-2035. These are full-time (or equivalent) annual employment changes, applicable to all sectors and industries apart from government and farm workers; and the data is again expressed in job years. Table 5: Total Private Non-Farm Employment Impacts 2016-2035 (including Legacy Effects to 2060) Geography State of Arizona Maricopa County so’u;c'é.- Job Years32 -76,308 Table 5 suggests that the expected or medium case distributed solar scenario would have a negative employment impact of 76,308 full-time (or equivalent) job years in the State of Arizona for the 2016-2035 period of study, including any legacy impacts up to 2060. This legacy effect accounts for the fact that the true effects of the distributed solar deployment are only experienced in full after the period of study (2016-2035), consistent with the economic life of each solar installation.33 In Maricopa County, there is a negative employment impact of 71,344 job years throughout the study period (including subsequent legacy effects). Table 6 summarizes the industry sectors impacted the most by the expected or medium case distributed solar scenario. 31 This simply reflects a deferral from the base case. 32 Ajob year is equivalent to one person having a full-time job for exactly one year. 33 The legacy effect should continue up to and including 2065. However, REMI currently does not allow for any analysis beyond 2060. If the economic life of an installation is less than 30 years, the negative economic consequences are in all probability greater than the estimates presented in this study. Attachment ACB 28R 48 of 59 Table 6: Statewide Employment Impacts by Industry Sector (Job Years)34 Sector Forestry, Fishing, and Related Activities Mining Utilities ,. , , ., , Wholesale Trade Retail Trade Transportation and Warehousing , Real Estate and Rental and Leasing and Technical Services Management of Companies and Enterprises Administrative and Support Services Educational Services Health Care and Social Assistance Arts, Entertainment, and Recreation Accommodation and Food Services OtherServices, except Public Total Net Change in Job Years Total Number of Job Years Lost in Non-Solar Industry Source: Authors’ Calculations Total Job Years, 2016—206035 -18 -2,563 -7,709 . , , . . -1,504 -2,691 -15,762 -2,472 ., , -4,948 -14,366 -361 29,025 —2,336 —18,026 -2,231 -76,308 105,333 The table again suggests that administrative and support services alone could benefit from the expected or medium case distributed solar scenario in terms ofjob years’ employment created. However, all other sectors are estimated to experience job losses, resulting in the total net estimate of 76,308 job years lost statewide. The administrative gain again probably originates to a large extent from the permitting of solar installations and business functions within the solar industry. The sectors estimated to experience the biggest job losses (expressed in cumulative job years) during the study period in rank order are health care and social assistance; retail trade; professional; scientific and technical services; the construction industry; and utilities. 34 Ajob year is equivalent to one person having a full-time job for exactly one year. 35 Total job years may not tally due to rounding-up. 36 This is a summation ofthejob years lost in non-solar industry sectors negatively impacted by the deployment of new distributed solar, 2016-2035. Attachment ACB 28R 49 of 59 Table 7 estimates the cumulative gross state product (GSP) and real disposable personal income impacts (RDPI) associated with the expected or medium case distributed solar scenario for the period 2016-2035. Table 7: Total Gross State Product (GSP) and Real Disposable Personal Income Impacts (RDPI) 2016-2035 (including Legacy Effects to 2060) Geography State of Arizona Maricopa Source: Authors’ Calculations Gross State Product Millions (2015 S) -$21,613.3 -$20,149.9 Real Disposable Personal Income Millions (2015 $) -$7,956.4 l —$8,087.9 Table 7 shows that in aggregate terms during the study period 2016-2035, and including legacy effects, total GSP could be cumulatively lower by over $21.6 billion (2015 S) in the State of Arizona under the expected or medium case scenario. This includes an estimated $20.1 billion GSP lost in Maricopa County (2015 5). Table 7 also shows that in aggregate terms during the study period 2016—2035, and including legacy effects, RDPI is estimated to be cumulatively lower by approximately $8 billion (2015 $) in the State of Arizona. This includes an estimated fall in RDPI of almost $8.1 billion in Maricopa County (2015 $).3‘7 The employment, GSP, and RDPI losses associated with the expected distributed solar deployment scenario are valid, because the total amount of money paid by distributed generation and central station generation electricity consumers over the 2016-2060 time horizon is greater than the amount which would have been paid had they all continued to draw electricity from the utility’s central grid. In short, electricity consumers are paying more for the same amount of electricity consumed under the expected distributed solar deployment scenario, and therefore have less money to spend in other parts of the economy. 37 Some of Maricopa County’s estimated losses in RDPI will be offset by minor gains in other counties, thereby resulting in a negligibly smaller loss for the State as a whole. Attachment ACB 28R 50 of 59 8.0 Simulation Results: High Distributed Solar Deployment Scenario The high case scenario assumes that approximately $13.4 billion is invested by approximately 1 million customers in distributed solar installations between 2016 and 2035, and the deferral or cancellation of $194 million central station generation investments (both nominal S).38 Table 8 estimates the total employment impacts of the high case distributed solar scenario for the period 2016-2035. These are full-time (or equivalent) annual employment changes, applicable to all sectors and industries apart from government and farm workers; and the data is again expressed in job years. Table 8: Total Private Non-Farm Employment Impacts 2016-2035 (including Legacy Effects to 2060) Geography State of Arizona Job Years39 ~116,558 Source: Authors’ Calculations Tab e8 suggests that the high case distributed solar scenario could have a negative employment impact of 116,558 full-time (or equivalent) job years in the State of Arizona for the 2016-2035 period of study, including any legacy impacts up to 2060. This legacy effect accounts for the fact that the true effects of the distributed solar deployment are only experienced in full after the period of study (2016-2035), consistent with the economic life of each solar installation.40 In Maricopa County, there is a negative employment impact of 108,857 job years throughout the study period (including subsequent legacy effects). Table 9 summarizes the industry sectors impacted the most by the high case distributed solar scenario. 38 This simply reflects a deferral from the base case. 39 Ajob year is equivalent to one person having a full-time job for exactly one year. 4" The legacy effect should continue up to and including 2065. However, REMI currently does not allow for any analysis beyond 2060. If the economic life of an installation is less than 30 years, the negative economic consequences are in all probability greater than the estimates presented in this study. Attachment ACB 23R 51 of 59 Table 9: Statewide Employment Impacts by Industry Sector (Job Years)“1 Sector Total Job Years, 2016-2060"2 —30 —3,496 -10,632 Forestry, Fishing, and Related Activities Mining Utilities ring Wholesale Trade Retail Trade and , Warehousing _ , Real Estate and Rental and Leasing Professional and Technical services Management of Companies and Enterprises Administrative and Support Services Educational Services Health Care and Social Assistance Arts, Entertainment, and Recreation Accommodation and Food Services Services, except Public Administration Total Net Change in Job Years Total Number of Job Years Lost in Non-Solar Industry Sectors“ Source: Authors’ Calculations ,_ -2,074 —4,318 -25,645 -3,847 -7,892 -20,701 -538 45,650 -3,898 ~29,486 -3,668 -116,558 162,208 Consistent with the previous two scenarios, the table suggests that administrative and support services could benefit alone from the high case distributed solar scenario in terms of job years employment created. The administrative gain again probably originates to a large extent from the permitting of solar installations, and also business support functions within the solar industry. All other sectors are estimated to experience job losses, resulting in the total net estimate of 116,558 job years lost statewide. The sectors estimated to experience the biggest job losses (expressed in cumulative job years) during the study period in rank order are health care and social assistance; retail trade; professional; scientific and technical services; the construction industry; and other services (excluding public administration). 41 Ajob year is equivalent to one person having a full-time job for exactly one year. 42 Total job years may not tally due to rounding-up. 43 This is a summation ofthe job years lost in non-solar industry sectors negatively impacted by the deployment of new distributed solar, 2016-2035. Attachment ACB - 28R 52 of 59 Table 10 estimates the cumulative gross state product (GSP) and real disposable personal income impacts (RDPI) associated with the high case distributed solar scenario for the period 2016-2035. Table 10: Total Gross State Product (GSP) Impacts 2016-2035 (including Legacy Effects to 2060) Geography State of Arizona Gross State Product Millions (2015 5) —$31,454.4 Real Disposable Personal Income Millions (2015 S) —$11,901.4 Source: Authors’ Calculations Table 10 shows that in aggregate terms during the study period 2016-2035, and including legacy effects, total GSP could be cumulatively lower by $31.5 billion (2015 S) in the State of Arizona under the high case scenario. This includes an estimated $29.3 billion GSP lost in Maricopa County (all 2015 S). Table 10 also shows that in aggregate terms during the study period 2016—2035, and including legacy effects, RDPI is estimated to be cumulatively lower by $11.9 billion (2015 $) in the State of Arizona. This includes an estimated fall in RDPI ofalmost $12.1 billion in Maricopa County (2015 S).44 The employment, GSP, and RDPl losses associated with the high distributed solar deployment scenario are valid, because the total amount of money paid by distributed generation and central station generation electricity consumers over the 2016-2060 time horizon is greater than the amount which would have been paid had they all continued to draw electricity from the utility’s central grid. In short, electricity consumers are paying more for the same amount of electricity consumed under the high distributed solar deployment scenario, and therefore have less money to spend in other parts of the economy. 44 Some of Maricopa County’s estimated losses in RDPI will be offset by minor gains in other counties, thereby resulting in a negligibly smaller loss for the State as a whole. Attachment ACB 28R of 59 9.0 Conclusions The goal of this study is to assess the impact of three distributed solar deployment scenarios in the APS service territory on economic activity in the State of Arizona and Maricopa County. The results of the analysis are influenced to an extent by the choice of economic impact model implemented. Economic impact analyses can generally be classified in one of 6 ways, represented in Figure 5. Figure 5: Seidman’s 3 x 2 Classification of Economic Impact Models COUNT GROSS PARTlAL GROSS GENERAL GROSS PARTIAL NET GENERAL NET l COUNT NET _ Gross studies only consider the direct positive impacts of increased economic activity in a specific sector, whereas Net studies represent a more thorough form of economic modeling as they also account for the trade-offs in the economy which result from incentivizing one specific sector, Counts are usually survey-based or theoretical capacity installation qua ntifications ofthe number of direct employees within a specific economic sector, which can extend to that sector’s entire supply chain. Partial models consider the wider effects of levels of activity in a specific economic sector, including the indirect and induced effects of the direct sectoral change. Frequently assessed via input—output models such as IMPLAN and REMI, partial models do not consider the feedback effects of changed levels of activity in a specific sector, such as the effect of large solar projects on wages in the labor market. General models offer the most comprehensive economy-wide analysis, taking into account all of the economic interconnections and feedback effects. Of the fourteen contemporary solar economic impact studies critiqued by Seidman, only one uses a general equilibrium model. This is Cansino, Cardenete, Gonzalez and Pablo-Romero’s (2013) study of Andalusia. However, this is a gross, rather than net analysis, because the authors combine renewables and non-renewables as a single sector, thereby preventing any Attachment ACB 28R 54 of 59 substitution between conventional and renewable forms of generation, and effectively only allowing for positive direct demand shocks in their modeling. The principal effect of using a Partial model approach rather than a Count approach, or using a General modeling approach rather than a Partial approach, is generally to reinforce the magnitude ofthe divined economic impacts. Thus, using a General model approach yields the most significant Grass and Net impacts. However, to the research team’s knowledge, a CGE model currently does not exist for the State of Arizona; and it would be cost prohibitive to test and develop a CGE model for the State of Arizona in a short time frame. Seidman has therefore implemented a Partial Net REMI analysis of solar deployment in the APS service territory, 2016—2035, for the current study. This is the next best alternative from a methodological standpoint; and it is consistent, for example, with the approach taken by Berkman, Lagos and Weiss (2014), NYSERDA (2012), and Treyz et al. (2011), critiqued in Section 3. Figure 6 positions Seidman’s approach relative to the fourteen critiqued studies Figure 6: Classification of Seidman’s 2016 Approach for APS Relative to Fourteen Contemporary Economic Impact of Solar/Renewables Studies , o o Gross Only positive o_r negative impacts Nat Both positive negative impacts . ,_ , Pollin and Garrett— o Peltier, 2009 o ETIC, 2016 o 0 o o Alvarez et al., 2009 Frondel et al., 2009 AECOM, 2011 Loomis, Jo & Alderman ,2013 Judson, Motamedi 2012 _, o , ,. Cansino et al. 2013 ' VSI and Clean Energy Project Nevada, 2011 i VSI, 2013 Comings etal. , 2014 NYSERDA, 2012 Treyz et al., 2011 Berkman et al., 2014 i SEFPMAN 2915, The economic impacts of all three distributed solar deployment scenarios are assessed in terms of private non—farm employment, gross state product, and real disposable personal income. Attachment ACB - 28R 55 of 59 The study clearly demonstrates that increased adoption of distributed solar generation represents a loss to the Arizona economy as a whole in all three scenarios. This is because the overall cost of provision of electricity to the State of Arizona will rise when referenced against a base case where electricity continues to be provided by central station generation. If the low case distributed solar deployment scenario actually transpires, the State of Arizona is cumulatively estimated to lose: 0 16,595 job years private non—farm employment; 0 Over $4.8 billion gross state product (2015 S); and o $1.8 billion real disposable personal income (2015 S). This takes into account both the solar installation study period (2016-2035) and the legacy effects ofthose installations to reflect the estimated 30 year economic life of the solar systems and deferred central station generation.45 If the expected or medium case distributed solar deployment scenario actually transpires, the State of Arizona is cumulatively estimated to lose: 0 76,308 job years private non-farm employment; 0 Over $21.6 billion gross state product (2015 S); and 0 Almost $8 billion real disposable personal income (2015 S). This also takes into account both the solar installation study period (2016-2035) and the legacy effects of those installations, to reflect the estimated 30 year economic life ofthe solar systems and deferred central station generation. If the high case distributed solar deployment scenario actually transpires, the State of Arizona is cumulatively estimated to lose: 45 The legacy effects of any 2035 distributed solar installation or deferred central station generation will continue until 2065. However, the REMI model used currently only provides economic impact estimates up to and including 2060, but Seidman does not believe that this will materially affect the conclusions in the analysis. If the economic life of an installation is less than 30 years, the negative economic consequences are in all probability greater than the estimates presented in this study. Attachment ACB 28R 56 of 59 0 116,558 job years private non-farm employment; 0 Approximately $31.5 billion gross state product (2015 S); and 0 $11.9 billion real disposable personal income (2015 S). This again takes into account both the solar installation study period (2016-2035) and the legacy effects of those installations, to reflect the estimated 30 year economic life of the solar systems and deferred central station generation. The implications of these findings are potentially far—reaching, as they challenge a sometimes expressed claim that an aggressive distributed solar initiative will have a significant positive impact on the state and county economies in the State of Arizona. In short, and wholly based on the financial implications of solar installations from a customer, utility and supplier perspective, this study estimates that any benefits emanating from the three distributed solar deployment scenarios are at best temporary and only coincident with the timing of those solar installations. This is because the lasting legacy effects of each distributed solar scenario, which reflect the economic life ofthe installed systems and deferred central station generation, are negative. That is, in all three scenarios, the total amount of money paid by distributed generation and Central station generation electricity consumers over the relevant time period (2016-2060) is greater than the amount which would have been paid had they all alternatively continued to draw electricity from the utility’s central grid. In each distributed solar scenario, electricity consumers as a whole are being asked to pay more for the same amount of electricity consumed, and therefore have less money to spend in other parts of the economy. Thus, when considered in the round from a purely financial perspective, the economic impact of all three potential solar deployed scenarios in the APS service territory are estimated to have a detrimental effect on both the State of Arizona and Maricopa County economies, all other things being equal. Attachment ACB 28R 57 of 59 Appendix A.1. The REMI Model REMI is an economic-demographic forecasting and simulation model developed by Regional Economic Models, Inc. REMI is designed to forecast the impact of public policies and external events on an economy and its population. The REMI model is recognized by the business and academic community as the leading regional forecast/simulation tool available. Unlike most other regional economic impact models, REMI is a dynamic model that produces integrated multiyear forecasts and accounts for dynamic feedbacks among its economic and demographic variables. The REMI model is also an "open" model in that it explicitly accounts for trade and migration flows in and out of the state. A complete explanation of the model and discussion of the empirical estimation of the parameters/equations can be found at www.remi.com. The operation of the REMI model has been developed to facilitate the simulation of policy changes, such as a tax increase for example, or many other types of events — anything from the opening of a new business to closure of a military base to a natural disaster. The model's construction includes a large set of policy variables that are under the control ofthe model's operators. To simulate the impact of a policy change or other event, a change in one or more of the policy variables is entered into the model and a new forecast is generated. The REMI model then automatically produces a detailed set of simulation results showing the differences in the values of each economic variable between the control and the alternative forecast. The specific REMI model used for this analysis was Policy Insight Model Version PI+ version 1.7.2 of the Arizona economy (at the county level) leased from Regional Economic Models Inc. by a consortium of State agencies, including Arizona State University, for economic forecasting and policy analysis. A.2. Effects Not Incorporated into the Analysis No major financial impacts were left out. Attachment ACB 28R 58 of 59 Glossary Gross State Product (GSP): The dollar value of all goods and services produced in Arizona for final demand/consumption. Job Year: A job year is equivalent to one person having a full-time job for exactly one year. Real Disposable Personal Income: The household income that is available to be spent after tax payments. Technically speaking, real disposable personal income is the sum of wage and salary disbursements, supplements to wages and salaries, proprietors’ income, rental income of persons, personal dividend income, personal interest income, and personal current transfer receipts, less personal taxes and contributions for government social insurance. Attachment ACB 28R 59 of 59 C W. SCHOOL ofBUSINESS ARIZONA STATE UNIVERSITY seidman institute L. WILLIAM SEIDMAN RESEARCH INSTITUTE 660 S MILL AVENUE, SUITE 300 TEMPE AZ 85281-4011 Tel: (480) 965 5362 Fax: (480) 965 5458 www.5eidmaninstitute.com \O O\]O\ SURREBUTTAL TESTIMONY OF AHMAD FARUQUI 10 On Behalf of Arizona Public Service Company “ 12 Docket No. E-04204A-15-0142 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 February 23, 2016 Table of Contents 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 I. INTRODUCTION ............................................................................................................. ..1 II. OVERVIEW AND ORGANIZATION OF SURREBUTTAL TESTIMONY ................. .. 1 III. THE APPROPRIATENESS OF DEMAND CHARGES AS A PRICE SIGNAL ........... .. 2 IV. THE ROLE OF DEMAND CHARGES IN PROMOTING ADVANCED ENERGY TECHNOLOGIES ............................................................................................................ .. 5 V. CUSTOMER UNDERSTANDING AND ACCEPTANCE OF DEMAND CHARGES ....................................................................................................... ............... .. VI. THE IMPACT OF DEMAND CHARGES ON BILLS AND ELECTRICITY CONSUMPTION ............................................................................................................ .. 13 VII. TRANSITIONING TO DEMAND CHARGES ............................................................. .. 14 VIII. CONCLUSION ............................................................................................................... .. 15 SURREBUTTAL TESTIMONY OF AHMAD FARUQUI ON BEHALF OF ARIZONA PUBLIC SERVICE COMPANY (Docket No. E-04204A—15-0142) INTRODUCTION PLEASE STATE YOUR NAME, JOB TITLE, BUSINESS ADDRESS AND PARTY FOR WHOM YOU ARE FILING TESTIMONY. My name is Ahmad Faruqui. I am a Principal with The Brattle Group. My business address is 201 Mission Street, Suite 2800, San Francisco, California 94105. I am ■ling testimony on behalf of Arizona Public Service Company. HAVE YOU PREVIOUSLY TESTIFIED IN THIS PROCEEDING? 10 Yes, I ■led Direct Testimony on December 9, 2015. 11 12 SUMMARY AND ORGANIZATION OF SURREBUTTAL TESTIMONY 13 WHAT IS THE PURPOSE OF YOUR SURREBUTTAL TESTIMONY IN THIS PROCEEDING? 14 The purpose of my Surrebuttal Testimony is to rebut some of the points made in the 15 direct testimony of several intervenors in this proceeding, including, TASC witness 16 Fulmer, Vote Solar witness Kobor, WRA witness Wilson, and RUCO witness Huber. In 17 addition, I will comment on some of the points raised by Staff witness Solganick. 18 19 20 21 22 23 24 25 26 27 28 PLEASE SUMMARIZE YOUR SURREBUTTAL TESTIMONY. Several intervenors have mischaracterized demand charges and three-part rates in general. Demand charges are an appropriate price signal that closely relates the design of the rate to the costs it is recovering. Through this close alignment with costs, in addition to improving economic ef■ciency and equity/faimess, three-part rates will provide an incentive for customers to adopt emerging energy management technologies that reduce power system costs for all customers. Customers are likely to be able to understand the concept of demand and respond to a demand charge by reducing their maximum demand through behavioral changes or adoption of the aforementioned technologies. In fact, while some customers’ bills will go down and others will go up with the new rate design, demand charges will provide all customers including those with limited income - with three opportunities to reduce their electricity bill: First, by managing their demand, second by conserving energy, and third by shifting usage to offpeak periods. In my Surrebuttal Testimony, I elaborate on these points. HOW IS YOUR SURREBUTTAL TESTIMONY ORGANIZED? My Surrebuttal Testimony is organized around the following issues: the appropriateness \]O\UI of demand charges as a price signal; the role of demand charges in promoting advanced energy technologies; customer understanding and acceptance of demand charges; the impact of demand charges on bills and electricity consumption; the impact of three-part 10 rates on customer bills; and how to make the transition to demand charges. 11 12 ARE YOU SPONSORING ANY ATTACHMENTS TO YOUR SURREBBUTTAL TESTIMONY? 13 No. 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 III. THE APPROPRIATENESS OF DEMAND CHARGES AS A PRICE SIGNAL DO YOU AGREE WITH THE ASSERTION THAT DEMAND CHARGES ARE NOT AN APPROPRIATE PRICE SIGNAL? No. As I indicated throughout my Direct Testimony, three-part rates, which include a demand charge as well as a ■xed charge and an energy charge, do a much better job of re■ecting the cost structure of generating and delivering electricity than two—part rates, which recover costs almost entirely through a volumetric charge.1 Two—part rates overcollect costs from larger-than-average customers and under-recover costs from smallerthan-average customers. Not only do three-part rates improve equity in rate design; they also encourage technological innovation by incentivizing the adoption of newly emerging energy technologies and by bringing about changes in energy consumption behavior that lead to more ef■cient use of power grid infrastructure and resources. ‘ See Decision No. 51472 (Oct. 21, 1980). 1 Q. WHY ARE THREE-PART RATES AN IMPROVEMENT OVER TWO-PART RATES? 2 A. Some intervenors have argued that while three-part rates may do a better job of 3 re■ecting costs in the short run, they do not do so in the long run.2 They argue that two- 4 part rates send a better price signal than three part rates.3 Precisely the opposite is true. 5 In the long run, transmission, distribution, and generation capacity costs are directly 6 driven by peak demand. 7 capacity in the long run. Reductions in demand reduce the need for new capacity. By 8 virtue of being tied speci■cally to a measure of a customer’s maximum demand, 9 demand charges capture this relationship between demand and infrastructure investment Thus, increases in demand translate into a need for more 10 11 requirements. 12 The View that three-part rates are an improvement in rate design is supported by the 13 testimony of ACC Staff. Staff witness Solganick, for instance, indicates that rate design 14 should recognize the concepts of customer, demand, and energy costs, and the time-and 15 season—differentiated nature of these costs.4 Staff witness Solganick further notes that 16 three-part rates are the norm for medium and large commercial and industrial (C&I) 17 customers, and thus set a precedent in Arizona.5 Indeed, in much of the country, three- 18 part rates are the norm for commercial and industrial customers and have been the norm 19 for the better part of the past century. 20 21 Q. SOME INTERVENORS WHO OPPOSE THE INTRODUCTION OF A DEMAND CHARGE HAVE PROPOSED ALTERNATIVE RATE DESIGNS. PLEASE SUMMARIZE YOUR UNDERSTANDING OF THOSE PROPOSED DESIGNS. A. Two alternative rate designs are mentioned in intervenor testimony. One is a minimum 22 23 bill, in which each customer would pay a ■xed minimum amount for electricity each 24 25 26 27 28 2 3 4 5 Direct Direct Direct Id, p. Testimony of Mark F ulmer, p. 19. Testimony of Briana Kobor, p. 34-35. Testimony of Howard Solganick, p. 10. 13. month, even if their net consumption (net of any self-generation) was very low.6 The second is a time-of-use (TOU) rate in which the volumetric charge varies by time of day, with a higher price during peak hours of the day and a lower price during off-peak hours.7 DO YOU AGREE WITH EITHER OF THESE ALTERNATIVE DESIGNS? No, I do not believe either of these is a suitable replacement for the three-part rate that has been proposed by UNSE. These alternatives will not solve the cost-shift issue that is \O O\]O\ 1o 11 12 13 14 attendant to the two-part design that is currently in place. Neither will they adequately re■ect cost of service or incent adoption of new technology. I expand on these points below. WHY IS A MINIMUM BILL NOT A SUITABLE ALTERNATIVE TO A THREE-PART RATE WITH A DEMAND CHARGE? There are two reasons why it is not a suitable alternative. 15 First, minimum bills must be set at a very high level in order to suf■ciently recover 16 capacity costs from rooftop solar customers. Second, minimum bills by themselves do 17 not reward reductions in demand or improvements in load factor. 18 incentive to do either. Thus, there is no 19 20 WHY IS A TOU RATE NOT A SUITABLE ALTERNATIVE TO A THREEPART RATE WITH A DEMAND CHARGE? 21 Since infrastructure costs do not vary with electricity consumption, they cannot be 22 recovered adequately through a volumetric (kWh) rate, TOU or otherwise. From a cost- 23 causation standpoint, the most ef■cient way to represent kilowatt-based costs is through 24 a kilowatt-based charge, i.e., through a demand charge. That is why demand charges are 25 part of the standard tariff for most commercial and industrial customers. 26 27 28 6 Direct Testimony of Kenneth Wilson, p. 1 1-12. 7 161., p. 13, Fulmer Direct, p. 23. 1 TOU rates are an appropriate method for recovering energy costs if they vary by time- 2 of-use but not for recovering capacity costs. Thus, they are a good complement to a 3 demand charge; not a substitute. Offering a rate with both a demand charge and a time- 4 varying energy charge may be the best option. 5 IV. THE ROLE OF DEMAND CHARGES IN PROMOTING ADVANCED ENERGY TECHNOLOGIES 7 Q. WOULD THE INTRODUCTION OF DEMAND CHARGES ADOPTION OF DISTRIBUTED ENERGY RESOURCES (DER)? 8 A. A change in rate design will affect the economics of DER. Adoption of the technologies 6 IMPACT 9 is driven in part by their economic attractiveness,8 thus the inclusion of demand charges 10 in rate design should affect their adoption levels. Some intervenors have suggested that 11 demand charges would curtail the adoption of distributed generation (DG), rooftop solar 12 in particular.9 However, this technology-speci■c perspective takes too narrow a view on the impacts of demand charges on energy technology adoption. 15 Q. WHY IS A FOCUS ON THE IMPACT OF DEMAND CHARGES ON ROOFTOP SOLAR PV TOO NARROW OF A PERSPECTIVE? 16 A. A three-part rate will foster technological innovation by encouraging customers to adopt 17 technologies that enable peak demand reductions and also reduce their energy 18 consumption. Examples of such technologies include battery storage, smart thermostats, 19 demand controllers, and energy information displays. The use of these technologies to 20 reduce demand will not only reduce customer bills but will also reduce the utility’s costs, thus bene■tting all customers. 23 In the same vein, introducing a demand charge and reducing the volumetric charge 24 would decrease the economic attractiveness of energy technologies that cannot provide 25 26 27 8 Other factors beyond economics also drive consumer buying decisions, such as a desire to be “green” or to have more control over energy consumption or simply to buy the newest technologies. 9 Kobor Direct, p. 4, F ulmer Direct, p. 17. 1 energy savings during those peak hours when the energy reductions are most valuable to 2 the system. This simply means that the three-part rate structure is encouraging adoption 3 of those technologies that are most bene■cial to the power grid and to customers. It is 4 important to take this broader View of energy technologies to avoid overstating the 5 importance of one particular option that may not be the most bene■cial. 6 Q. HOW WILL VEHICLES? A. WRA witness Wilson has suggested that ownership of an electric vehicle (EV) would 7 8 9 DEMAND CHARGES IMPACT OWNERS OF ELECTRIC lead to a “peakier” consumption pro■le, resulting in an increase in the cost of charging 10 the vehicle with a demand charge.10 This is not necessarily correct. 11 charged during nighttime hours when the household is otherwise using relatively little 12 electricity, then charging the electric vehicle would not necessarily create a new peak. 13 Additionally, smart charging equipment would allow the EV owner to manage his or her 14 charging to reduce the possibility of setting a new peak. If the vehicle is 15 It is also important to note that, while off-peak load building is bene■cial to the power 16 system by reducing average costs, the simultaneous charging of several EVs on a 17 capacity constrained feeder could lead to a need for distribution system upgrades. 18 Demand charges should help to send a price signal that encourages ■atter load pro■les ;: throughout the day and reduces this possibility. 21 V. CUSTOMER UNDERSTANDING AND ACCEPTANCE OF DEMAND CHARGES 22 Q. DO YOU AGREE WITH THOSE INTERVENORS WHO HAVE SUGGESTED THAT CUSTOMERS WILL NOT UNDERSTAND DEMAND CHARGES? A. No, 1 don’t agree with TASC witness F ulmer and WRA witness Wilson, for example, 23 24 25 who have suggested that residential customers will not be able to understand demand 26 27 28 10 Wilson Direct, p. 10. 1 charges.”’12 2 understand demand charges, if they are explained properly to them. I believe there are several reasons why customers will be able to 3 4 5 Q. WHY DO YOU BELIEVE DEMAND CHARGE? A. First, 117,000 customers of Arizona Public Service (“APS”) have elected to take service THAT CUSTOMERS CAN UNDERSTAND A 6 on voluntary demand charges. APS has been offering these rates to its residential 7 customers since the very early 1980s. 8 advanced metering infrastructure (“AMI”), there is evidence that customers were able to 9 comprehend the notion of demand and recognize the bene■ts of being on such a rate. In other words, long before the advent of 10 Second, just about every customer has encountered the concept of electricity demand in 11 daily life, perhaps without knowing what demand was. It would be hard to ■nd a 12 residential customer who is not familiar with a light bulb. When buying or installing a 13 light bulb, the customer had to choose a bulb that would project a certain amount of 14 light. It was then that the customer would have encountered the power of the bulb 15 expressed in watts.13 The wattage would have been expressed as 40 watts, 60 watts, 75 16 watts or 100 watts (or their equivalent, if the bulb was a compact ■uorescent or LED 17 bulb). 18 and 150; or 100, 200 and 250. Thus, it would be dif■cult to ■nd a customer who has not 19 encountered the concept of watts. Further, if the customer had purchased a high wattage 20 hair dryer and a high wattage electric iron, and decided to run both at the same time, 21 they may have tripped the circuit breaker, requiring a trip to the garage or basement to 22 reset it after one of the two devices had been unplugged. 23 through which customers would have become familiar by experience with the concept of 24 25 demand or capacity. 26 27 28 Some wattages would have been higher, for three-way bulbs, such as 50, 100, ” F ulmer Direct, p. 18. 12 Wilson Direct, p. 5. 13 Watts is the industry-accepted unit of power or demand. 7 That is yet another way Third, customers do not need to know the precise de■nition of a kilowatt in order to be able to respond to a demand rate. Simple messages encouraging customers to avoid the simultaneous use of electricity-intensive appliances can convey this concept in easy-tounderstand terms without even using the word “kilowatt.” Fourth, all of this would apply even with greater force to customers with rooftop solar. They would have encountered the concept of watts (or kilowatts) once again when they purchased or leased their solar panels since that is the measure in which the size and cost of the panels are expressed. Demand rates for rooftop PV customers, therefore, would \DO \]O‘\ convey prices in terms in units that they are already familiar with. 10 11 Q. DO YOU AGREE THAT A DEMAND CHARGE WILL FUNCTION AS AN ADDITIONAL FIXED CHARGE FOR MOST CUSTOMERS? A. No, I do not agree with Vote Solar witness Kobor who states that a demand charge will 12 13 “likely function as an additional ■xed charge for most residential and small commercial 14 customers because they lack the tools and understanding to effectively respond to the 15 demand charge price signal.”14 Demand charges are signi■cantly different than ■xed 16 charges in that customers can, and are likely to, reduce their demand charge by lowering 17 their demand. Conversely, customers are not able to take any behavioral actions to 18 reduce their ■xed charges. 19 20 Q. HOW ARE DEMAND CHARGES DIFFERENT THAN FIXED CHARGES? 21 A. Importantly, demand charges vary with a customer’s demand for electricity. Customers 22 '23 24 with high maximum demand will be charged more than customers with low maximum demand. The result is that customers are charged in a manner that is proportional to their use of the power grid. Fixed charges, on the other hand, charge each customer the 25 26 27 14 Kobor Direct, p. 36 1 same amount regardless of their use of the power grid. Referring to a demand charge as 2 a ■xed charge ignores this important distinguishing feature of demand charges. 3 Additionally, with a demand charge, customers have the ability to reduce their bill by 4 changing the way they consume electricity. 5 under the customer’s control. 6 that customers respond to demand charges by reducing their maximum demand. 7 Additionally, the Direct Testimony of APS witness Miessner stated that 60% of a 8 sample of APS’s customers on a three-part rate reduced their demand after switching to 9 the three-part rate, with those who actively manage their demand achieving demand Fixed charges, on the other hand, are not In my Direct Testimony, I cited four studies that found savings of 10% to 20% or more.15 12 Q. DO THE INTERVENORS OFFER ANY EVIDENCE THAT CUSTOMERS CANNOT UNDERSTAND A DEMAND CHARGE? 13 A. TASC witness F ulmer and Vote Solar witness Kobor cite a study in California by Hiner 14 and Partners (“Hiner”) to suggest that customers could not understand a demand 15 charge.16 However, for several reasons this study does not support the conclusion that customers cannot understand a demand charge. 18 Q. WHY IS IT INCORRECT TO USE THE HINER STUDY TO SUPPORT A CONCLUSION THAT CUSTOMERS CANNOT UNDERSTAND A DEMAND CHARGE? A There are several problems with Mr. F ulmer’s and Mr. Kobor’s use of the Hiner study in 19 20 21 their testimony. 22 claim that customers will not understand demand charges: “The survey found ‘Possible 23 that concept was confusing and respondents did not understand that it varies based on In his testimony, Mr. Fulmer states the following in support of his 24 25 26 27 15 Direct Testimony of Charles Miessner, p. 7. 16 F ulmer Direct, p. 18, Kobor Direct, p. 36. 28 9 kW demand levels, which made demand charges appear low relative to monthly Service fee ’”17 Based on careful inspection of the study, at no stage was customer understanding of demand charges even investigated. Mr. Fulmer selectively quotes extracts from commentary by the study authors and presents this information as a ■nding of the survey. The reality is that in a conjoint analysis, investigating relative preferences for various rates, the study found that the existence of demand charges was relatively unimportant in rate plan selection. Rather the presence of a “monthly service fee had more influence on rate choice than any other attribute,” followed by “the price per kWh 10 associated with different rate structures rather than by the rate structure itself.”18 11 To explain this result, the study authors speculated, “[It is] possible that [the] concept [of 12 demand charges] was confusing and respondents did not understand that it varies based 13 on kW demand levels, which made demand charges appear low relative to [a] monthly 14 service fee. ”19 15 This is speculative commentary, not fact, and only one possible explanation of many for why demand charges seemed to have little impact on rate plan 16 selection. 17 18 Additionally, Mr. Fulmer indicates that the survey identi■ed the following as surveyed 19 customers’ preferred features in a solar rate: “57% stated save money, 39% said simple, 20 and 34% said ‘■ts my habits and lifestyle.’”20 Mr. Fulmer fails to mention, however, 21 that the same study says that the advantages of demand charges are that they can, “save 22 money (through changing behavior), gives control over the 23 listed “confusing” as a negative attribute of all four of the rates examined in the study--a Moreover the study 24 25 26 27 17 18 ‘9 2° 2‘ F ulmer Direct, p. Hiner & Partners, Id., slide 22. Fulmer Direct, p. Hiner & Partners, 18. Inc. “RROIR Customer Survey — Key Finding,” April (2013), slide 18. 19. “Final Report: Solar (NEM) Rate Preferences Survey Results” (June 2015), slide 8. 28 10 feed in tariff; a demand charge, a solar capacity charge and a panel rate (where you are billed by the size of circuit panel for delivery).22 In fact when one looks at how customers rated the four plans on simplicity (“Does not require a lot of effort to understand how my energy use will affect my bill.”), there is very little variation in the results.23 Twenty-eight percent of customers found the feed in tariff plan (which involves only kWh) to be simple, 26% found the installed capacity charge and the panel rate to be simple and 24% found the demand charge to be simple.24 In sum, the Hiner study does not prove that customers would not understand or not be interested in a rate with a demand charge. The assertion by the intervenors that it does is an unsupported generalization. 11 IS ANY OTHER EVIDENCE OFFERED TO SUGGEST THAT CUSTOMERS WOULD NOT WANT OR UNDERSTAND A DEMAND CHARGE? 13 Vote Solar witness Kobor claims that APS’s 10% enrollment level in its voluntary three- 14 part rate is evidence that customers do not want three-part rates.25 15 dif■cult to get customers to voluntarily sign up for new energy programs. 16 demand response programs have participation of around 10%. When new rate designs 17 are introduced on a voluntary basis they rarely achieve enrollment levels in excess of 18 20% to 30%.26 19 typically must be offered on a mandatory or default basis to achieve signi■cantly higher 20 enrollment levels.27 This It is inherently been the experience with time-varying rates. Many New rates 21;: 22 23 23■i,s■des22,26,30,34. 241d 25 25 Kobor Direct, p. 38 26 For information on residential demand response program participation, see FERC reports on advanced 5 265 metering and demand response: http://www.ferc.gov/industries/electric/indus-act/demandresponse/dem-res-adv—metering.asp 27 Ahmad Faruqui, Ryan Hledik, and Neil Lessem, “Smart by Default,” Public Utilities Fortnightly, August 2014. 28 24 11 1 Q. IS CUSTOMER ACCEPTANCE THE ONLY CRITERION THAT SHOULD BE CONSIDERED WHEN EVALUATING THE MERITS OF A THREE-PART RATE? A. While customer acceptance of and satisfaction with the new rate design is certainly a 2 3 4 consideration, it is not the only criterion that should be taken into account. It is only one 5 of the ten Bonbright criteria. In fact, if customer acceptance were the only principle that 6 mattered, one could argue that customers should simply be given free electricity, as they 7 would certainly be more satis■ed with free electricity than with paying for it. Rather, as 8 I discussed throughout my direct testimony, a demand rate which more closely aligns 9 the structure of the rate with underlying costs improves fairness in rate design and can 10 have significant bene■ts for customers. Factors such as cost causation, equity/faimess, 11 and the impact of the rate on emerging energy technology adoption are all critical 12 considerations beyond a having only narrow focus on customer acceptance. 13 14 Q. REGARDING THE ISSUE OF EQUITY, DO YOU AGREE WITH THOSE INTERVENORS WHO HAVE SUGGESTED THAT MANY CROSS-SUBSIDIES ARE EMBEDDED IN CURRENT RATES AND THAT WE THEREFORE SHOULD NOT FOCUS ON JUST ADDRESSING A CROSS-SUBSIDY RELATED TO THE ADOPTION OF DISTRIBUTED GENERATION? A. No, I do not agree with the argument made by AURA witness Alston that it is 15 16 17 inappropriate to address DG-related cross-subsidies without addressing all subsidies 18 embedded in today’s rates.28 19 Therefore, the introduction of a demand charge does more than just address the DG- 20 related cross subsidy. It also addresses the subsidization of customers with peak (i.e., 21 costly) consumption patterns by those with ■at (i.e., bene■cial) consumption patterns. 22 If deployed to all customers, this amounts to the removal of a signi■cant cross-subsidy :: and one not just limited to DG-related issues. First, demand charges better align rates with costs. 25 26 27 28 28 Direct Testimony of Thomas Alston, p. 3 12 Second, while DG market penetration may be relatively low today, it is the trajectory of adoption that matters. PV costs continue to decline rapidly, and Congress just extended the income tax credit by ■ve years. It is better to get in front of this issue now, before it becomes a bigger problem and while grandfathering of current rooftop solar customers under the current rate is still a feasible policy option. VI. THE IMPACT OF CONSUMPTION DEMAND CHARGES ON BILLS AND ELECTRICITY SOME INTERVENORS SUGGEST THAT DEMAND CHARGES WILL INCREASE BILLS FOR LIMITED INCOME CUSTOMERS. IS THERE ANY EVIDENCE TO SUPPORT THIS CLAIM? 0 0 \10\ I have not come across any empirical evidence from the intervenors which shows that 11 limited income customers will be made worse off overall with a three-part rate. 12 Whether or not a three-part rate will cause a customer’s bill to increase or decrease is 13 not determined by the customer’s income or even their monthly consumption. The bill 14 impact is driven by the customer’s load factor. If the customer has a high load factor 15 (i.e., a relatively ■at electricity consumption pro■le), then his or her bill is likely to 16 decrease. If the customer has a low load factor (i.e., a “peaky” consumption pro■le), the 17 bill is likely to increase. A common mistake is to equate the impact of a demand charge 18 on a customer’s bill with the customer’s total monthly consumption. Demand charges 19 do not automatically increase bills for small users, because a small user could have a 20 higher load factor than the class average. 21 22 COULD DEMAND CHARGES PROVIDE LIMITED INCOME CUSTOMERS WITH AN OPPORTUNITY TO REDUCE THEIR ELECTRICITY BILL? 23 Yes, demand charges could provide limited income customers with a new opportunity to 24 save money on their electricity bills. Whereas the existing two—part rate provides only 25 one opportunity to save money - by reducing one’s total consumption - the three-part 26 rate also provides an opportunity to save through both reductions in total consumption 27 and reductions in maximum demand. Certain actions which would not provide material 28 13 bill savings under the two—part rate, such as staggering the use of electricity intensive appliances, would yield a bill reduction under the three—part rate. WILL DEMAND CHARGES REDUCE THE INCENTIVE TO CONSERVE ENERGY? I don’t agree with WRA witness Wilson’s and TASC witness Fulmer’s suggestion that the lowering of the volumetric rate will reduce the incentive to conserve.29 As I indicated earlier in my Surrebuttal Testimony, a three—part rate gives customers an additional option for reducing their electricity bill - they can reduce total consumption and/or maximum demand. Rather than removing a customer’s incentive to conserve, 10 demand charges encourage a different and/or more efficient type of conservation 11 is, conservation at peak times when it is most valuable to the power system. 12 Direct Testimony, APS witness Miessner indicates that customers on APS’s demand 13 rate have not only reduced their demand but their total electricity consumption as well.30 14 Demand charges are therefore not implicitly going to impair energy ef■ciency efforts; 15 they will simply guide those efforts toward the most bene■cial ef■ciency initiatives. that In his 16 VII. 17 DO YOU AGREE WITH STAFF, WHO CUSTOMERS NEED EDUCATION AND TRANSITION TO A THREE-PART RATE? 18 19 education and information about demand charges in order to successfully make the 21 transition to a three-part rate. I agree with this and believe that a transition plan should 22 be developed when making signi■cant changes to rate design in order to facilitate a 23 smooth introduction of the new rate. 24 26 27 HAS SUGGESTED THAT INFORMATION IN THE Yes. Staff witness Solganick indicates throughout his testimony that customers will need 20 25 TRANSITIONING TO DEMAND CHARGES Q. WHAT DO YOU CONSIDER TO BE IMPORTANT ELEMENTS OF A RATE TRANSITION PLAN? 29 Wilson Direct, p. 9, F ulmer Direct, pp. 21 and 24 3° Miessner Direct, pp. 7-8. 28 14 The transition plan should be tailored to the speci■c needs of the utility and its customers. As such, it will vary from one jurisdiction to the next. In other words, there is not a “one size ■ts all” approach to the rate transition. Still, there are several examples of elements that I would consider to be useful options to consider in the plan. As I described above, one is a customer education plan that includes the provision of general information about the new rate and opportunities to mitigate potential bill \IO\UI impacts, as well as targeted outreach and education for those customers who are most likely to experience bill changes under the new rate. 10 VIII. CONCLUSION 11 WHAT DO YOU CONCLUDE, INTERVEN OR TESTIMONY? 12 The intervenors’ objections to the three-part rate are not supported by evidence. For 13 several reasons, I believe that a three-part rate would be a signi■cant improvement over 14 the current two-part rate. It would do a much better job of re■ecting the cost structure 15 of generating and delivering electricity. The rate would simultaneously improve 16 economic ef■ciency while promoting equity and fairness in rate design -arguably the 17 two most important principles in rate design. It would provide customers with new 18 opportunities to save money on their electricity bills. Finally, it would foster innovation 19 by improving the economics of a range of emerging energy technologies that can reduce 20 demand and, as a result, infrastructure costs. BASED ON YOUR REVIEW 21 22 23 DOES THIS CONCLUDE YOUR SURREBUTTAL TESTIMONY? Yes, it does. 24 25 26 27 28 15 OF THE SURREBUTTAL TESTIMONY OF CHARLES A. MIESSNER On Behalf of Arizona Public Service Company Docket No. E-04204A-15-0142 February 23, 2016 Table of Contents INTRODUCTION ............................................................................................................. .. 1 II. SUMMARY ...................................................................................................................... .. 1 III. MISSTATEMENTS ABOUT APS DEMAND RATES .................................................. .. 3 IV. INTERVENORS’ OBJECTIONS TO DEMAND RATES ARE UNFOUNDED ........... .. 4 STAFF’S RECOMMENDATION FOR CUSTOMER EDUCATION ON THREEPART RATES ................................................................................................................. .. 21 VI. \DO \]OLI4>10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 UN SE’S PROPOSAL FOR A SOLAR PURCHASE RATE ......................................... .. 23 VII. CONCLUSION ............................................................................................................... .. 23 SURREBUTTAL TESTIMONY OF CHARLES A. MIESSNER ON BEHALF OF ARIZONA PUBLIC SERVICE COMPANY (Docket No. E-04204A-15-0142) INTRODUCTION PLEASE STATE YOUR NANIE AND BUSINESS ADDRESS. Charles A. Miessner, 400 North Fifth Street, Phoenix, Arizona 85004. DID YOU PROVIDE TESTIMONY EARLIER IN THIS DOCKET? Yes, I provided Direct Testimony on behalf of Arizona Public Service Company (APS). DID YOU REVIEW THE REBUTTAL TESTIMONY OF UNSE AND THE DIRECT TESTIMONY OF STAFF AND OTHER INTERVENORS? 10 Yes, I did. 11 12 WHAT IS THE PURPOSE OF YOUR SURREBUTTTAL TESTIMONY? 13 The purpose of my Surrebuttal Testimony is to respond to certain assertions and claims 14 made by other intervenors that relate to my Direct Testimony and to assess their 15 recommendations made in this proceeding. 16 buy—back rate for customers with rooftop solar and Staff’s proposal for a transition 17 period to three—part rates. I will also address UNSE’s proposal for a 18 19 II. SUMMARY 20 WILL YOU PLEASE SUMMARIZE YOUR SURREBUTTAL TESTIMONY? 21 In my Surrebuttal Testimony, I respond to the objections to three-part demand rates 22 made by various intervenors. Specifically, I address five key points: 23 ' Demand charges; 24 Minimum bills; 25 Time of use (TOU) rates; 26 27 28 ' Complexity for residential customers; and Impacts on energy efficiency. Demand charges. Contrary to the objections made by TASC and Vote Solar, APS believes demand charges appropriately re■ect the cost of service and recover the right amount of revenue. Minimum bills. Minimum bills are not a viable alternative to three—part demand rates as asserted by TASC, WRA, and other intervenors. Minimum bills would have to be both very large and tiered for small, medium, and large homes to have any beneficial effect. Specifically, the minimum bill would have to be significantly larger than the amounts 10 11 12 typically discussed by these parties, $10 to $25, to address the cost recovery issue for any but the smallest energy users. For example, minimum bills for medium-sized and large-sized homes would likely need to be in the range of $70 to $150 per month. 13 Time-of-Use (TOU) rates. Two-part time—of—use (“TOU”) energy rates are not a viable 14 alternative to three-part demand rates as asserted by TASC, WRA, 15 intervenors. 16 designs, still do not and cannot adequately re■ect cost of service because they inherently 17 recover infrastructure costs through variable kWh charges. What’s more, neither 18 minimum bills nor TOU kWh rates incent customers to invest in home technologies that 19 actually reduce the utility’s infrastructure costs. While TOU kWh rates provide a higher 20 incentive to reduce energy during the on-peak hours, if this response is not consistent 21 throughout the entire month, these rates will have either a limited impact or no impact at 22 all on the demand-related infrastructure costs. and other TOU kWh rates, while having advantages over other two-part kWh rate 23 24 25 Complexity for residential customers. I rebut the contention made by TASC, WRA, and Vote Solar that residential customers will not be able to understand or manage demand. 26 27 28 As discussed extensively in my Direct Testimony, APS’s existing three-part demand rate has shown just the opposite; a significant number of APS customers have already accepted the demand-rate concept and can and do manage their demand. 2 Impacts on energy efficiency. I also respond to claims made by TASC, WRA, and other intervenors that three-part demand rates will hurt energy efficiency. These parties claims are wrong. A demand rate will incent energy efficiency programs that are refocused on opportunities that can reduce both energy and demand, which would be much more valuable to the electrical system and other customers because they could result in both savings in future fuel costs and savings from avoiding or deferring the need to build additional power plants. In addition, APS supports the general direction proposed by UNSE to replace net metering with a solar purchase rate arrangement for the excess power. While the best 10 rate is avoided cost, the purchase rates proposed by UNSE could provide a reasonable 11 method to value excess energy exported to the grid. Certainly, the purchase rate should 12 be no higher than the purchase price for grid—scale PV solar power. 13 14 For these reasons, APS recommends that the Commission approve three-part demand 15 rates for UNSE residential customers. 16 Staff and agreed to by UNSE in their rebuttal testimony. 17 supports Staff’s recommendation for a transition period combined with customer 18 education. APS further supports adoption of UNSE’s net metering proposal. This is consistent with the proposals made by In that respect, APS also 19 20 21 22 III. MISSTATEMENTS ABOUT APS DEMAND RATES DOES APS HAVE EXTENSIVE EXPERIENCE WITH THREE-PART DEMAND RATES FOR RESIDENTIAL CUSTONIERS? Yes. As detailed in my Direct Testimony APS has several decades of experience with 23 residential demand rates. We currently have more than 117,000 residential customers 24 on a demand rate, which is approximately 11% of our total residential customer base. In 25 addition, all of our business customers served under our small to extra—large rates are 26 billed with a demand charge. 27 28 HAVE APS’S RESIDENTIAL DEMAND RATES BEEN SUCCESSFUL? Yes. While Vote Solar witness Kobor considers an 11% adoption rate to be low and indicates that not many customers desire a demand rate, it is actually quite impressive for a rate that is voluntary and competed with as many as four residential two—part TOU rates (now just two). In fact, APS has the highest participation in residential demand rates in the country. Vote Solar claims that APS’s current three—part demand rate does not produce any demand reduction — the participants are only high—use customers who naturally save on the rate. Both parts of this assertion are untrue. I have provided substantial information in my Direct Testimony concerning the demand and energy 10 reductions achieved by customers currently on the demand rate, which are significant. 11 See also APS’s response to RUCO Data Request 1.6, which is attached as Attachment 12 CAM-lSR and incorporated into my Surrebuttal Testimony by this reference. 13 14 IV. 15 INTERVENORS’ OBJECTIONS TO DEMAND RATES ARE UNFOUNDED WHAT OBJECTIONS DO SOME INTERVENORS HAVE TO THREE-PART DEMAND RATES? 16 Certain intervenors testified that they object to three—part rates for residential customers 17 because they have asserted: 18 Demand charges do not re■ect long—run cost of service; 19 ' Demand charges will somehow recover too much revenue because of customer 20 diversity; 21 Minimum bills and TOU rates are better alternatives; 22 Customers won’t understand or manage demand; and 23 ' Demand charges will disincent energy efficiency. 24 25 Q. DO DEMAND CHARGES REFLECT COST OF SERVICE? 26 A. Yes. 27 28 In fact, demand charges correct the misalignment between a customer’s cost of service and their bill inherent in two-part energy rates that rely on a monthly service charge and kWh energy charges to recover the utility’s infrastructure investment necessary to serve the home. I included a detailed explanation of this issue in my Direct Testimony in this proceeding and therefore will refer the reader to that testimony rather than repeating that information here. However, the highlights are: 0 A significant portion of the cost to serve residential customers is comprised of infrastructure investments; 0 These costs are indisputably driven by the kW demand of the home, not the monthly kWh consumption; and o 10 A kW demand charge is the most appropriate and accurate way of recovering these costs. 11 12 13 14 15 16 17 18 19 20 21 22 BUT AREN’T ALL COSTS VARIABLE IN THE LONG RUN? Not in the sense that some intervenors are referring. Utilities typically face higher fuel costs, higher customer-related costs like meters and billing systems, and higher infrastructure costs as they build new power plants, transmission lines, and distribution facilities over time. However, once an infrastructure investment is made it is not variable over its useful life, which can be several decades. Therefore, this notion of ■uctuating costs over time can lead to a somewhat common misconception about fixed versus variable costs articulated by TASC witness Fulmer in his direct testimony. Fulmer claims that kWh charges, which he correctly characterizes as a variable charge, are a better re■ection of cost of service because all costs are variable in the long run. Hence variable charges for variable costs. 23 The problem with this reasoning is that the increases in customer-related costs and 24 infrastructure investment are not driven by increased kWh consumption, but rather the 25 increased number of customers and increased kW demand, respectively, over time. All 26 costs and charges can and do change over time, but that doesn’t change the fundamental 27 need for a direct nexus between rate design and cost of service. 28 Increased fuel and variable operating and maintenance costs (“0&M”), which vary with kWh production, should be recovered through higher kWh charges; increased customer—related costs should be recovered through higher monthly service charges; and increased infrastructure costs should be recovered through higher demand charges. WILL DEMAND CHARGES OVERRECOVER INFRASTRUCTURE COSTS DUE TO CUSTOlVIER DIVERSITY AS CLAIMED BY WRA? No. That’s not how the rate making process works. APS has been billing residential and business customers on demand rates for decades and has established those charges in numerous rate cases. The diversity issue was appropriately re■ected and adjusted for 10 as a part of this process. There simply is no double counting. 11 12 13 14 15 16 Let me explain with an example. The monthly demand for APS’s residential customers, i.e., the hour when the electrical load in each home is at its peak, typically occurs between 5 pm. and 9 pm. in both summer and winter months, although there are always exceptions to this usual pattern. However, the precise hour of this demand will vary for each home — some will peak at 6 p.m., others at 7 pm. and so on. That is what is called diversity. 17 18 Consider a simple example with 5 residential customers, each with an individual 19 monthly peak demand of 6 kW for their home and lower kW loads in other hours. The 20 sum of the individual demands for the 5 homes would be 30 kW (6 kW times 5 21 customers). As illustrated in Table 1 below, in this example the combined load of all 5 22 homes, taken as a group, peaks at 7 p.m., which only adds up to 26 kW because not all 23 of the homes are that their maximum load at that time. This combined maximum hourly 24 load for the group is referred to as the class peak. In addition, suppose the utility also 25 serves some business customers and the combined peak for all customer classes occurs 26 at 6 p.m., which is referred to as the system peak. The residential load in this example is 27 23 kW at the system peak hour. 28 Table 1. Illustrative Hourly kW loads for 5 Customers Customer 5:00 PM kW 1 2 3 4 5 Total 4 4 4 6 4 22 System Peak 6:00 PM kW 4 4 5 5 5 23 Class Peak 7:00 PM kW 6 6 6 4 4 26 8:00 PM kW 4 4 4 4 6 22 Individual Demand kW 6 6 6 6 6 30 10 11 12 13 14 15 So what’s the issue? There’s a mistaken notion that because the utility bills each of the 5 customers according to their individual demand, the 6 kW, even though the infrastructure costs are driven by the combined peak of all the homes at the time of the system peak, the 23 kW, they will over recover these costs — they will charge for an aggregate of 30 kW, but only incur costs for 23 kW. 16 17 IS THIS NOTION CORRECT? 18 No, it’s not, and for two reasons. First of all, not all of the utility’s infrastructure costs 19 are driven by the system peak. For example, the costs for some equipment like the pad 20 mounted transformer in front of the home are driven by the diversified peak of the 21 homes served off of that transformer (5 homes, and 26 kW in this illustrative example), 22 not the system peak. Similarly, the cost for the local substation that serves the home is 23 driven by the neighborhood or class peak, not the system peak. Infrastructure costs that 24 are farther “upstream” from the home and serve a much wider group of customers, such 25 as power plant costs, are primarily driven by the system peak. 26 27 2,8 Second and more importantly, irrespective of the cost drivers, diversity does not result in any double counting or over recovery. This is because of the way the demand charge is derived in the rate-making process. PLEASE EXPLAIN. In the rate-making process the demand charge is calculated in two steps: (1) the allocation of demand—related costs to a specific customer class and (2) the derivation of the monthly demand charge using the allocated costs and the class billing determinants (i.e. the total customer kW that will be billed each month). 10 In the first step, the utility’s infrastructure costs are allocated or assigned to each customer class based on the cost drivers discussed above. This step establishes the total 11 12 13 14 15 infrastructure cost to be recovered through rates from the specific class. Continuing with the same example of 5 residential customers, assume that the various infrastructure costs associated with the three cost drivers (individual demand, class peak, and system peak) are as shown in Table 2 below. These costs are allocated to the residential class according to the kW specific to each driver: the cost related to the kW of the individual 16 homes is allocated at 30 kW (5 homes at 6 kW peak demand per home); the cost related 17 18 to the class peak kW is allocated at 26 kW; and the costs related to the system peak are allocated at 23 kW. 19 2O 21 22 23 24 25 26 27 28 This allocation process results in $3,780 per year of total infrastructure costs to be recovered in rates from residential customers. Importantly, this allocated cost fully re■ects the appropriate level of diversity for each type of cost. Table 2. Illustrative Cost Allocation and Demand Charge Derivation Step 1. Cost Allocation 00 10 11 12 13 Cost Driver Individual Demand Class Peak System Peak Total Unit Cost $-kW 30 40 80 Units kW 30 26 23 Allocated Cost 900 $ 1,040 $ 1,840 $ 3,780 $ Step 2. Derivation of Demand Charge Total Allocated Cost Total Billed kW (12 months of kW) Monthly Demand Charge ($-kW) Revenue from Demand Charge $ $ $ 3,780 360 10.50 3,780 (30 kW X 12 months) ($3,780 360 kW) ($10.50 X 360 kW) 14 15 In the second step of the process, the monthly demand charge is derived by dividing the 16 $3,780 annual cost by the total kW that will be billed for all 5 customers, which in this 17 case is 360 kW (5 customers times 6 kW per month times 12 months). 18 demand charge is $10.50 per kW. This step also ensures that there is no double counting 19 or overrecovery of costs because the kW used to derive the charge (30 kW per month, 20 360 kW per year) are the same kW that will be billed customers. 21 utility’s annual revenue from the demand charge in this example is $3,750 ($10.50 per 22 kW times 30 kW per month times 12 months), which is exactly the same as the 23 infrastructure costs allocated to the residential class. 24 25 26 27 The resulting In other words, the In recap, the cost allocation process ensures that the right amount of costs are recovered in rates by appropriately re■ecting the customer load diversity in the allocated costs for each customer class. The rate derivation process ensures that the demand charge will be designed to recover this allocated cost by calculating the charge using the same 28 9 “undiversified” kW that customers will be billed on — and that the utility must have the local infrastructure in place to support. Expanding the example from five customers to a million complicates the math a bit, but would not change the result. WHY IS A MINIMUM BILL NOT A VIABLE ALTERNATIVE TO A THREEPART DEMAND RATE? A minimum bill is not a viable alternative to a three—part demand rate both from a costof-service perspective and a practical standpoint. Conceptually, a minimum bill would 00 have to address two potential situations — a customer whose kWh and kW drop significantly in a particular month, e.g., the customer is absent from the house for two 10 weeks out of the month, and a situation where a customer’s kWh drops significantly, but 11 their kW remains at or near the normal level for the home. 12 A sufficiently sized minimum bill could be somewhat useful for the first situation where 13 both the kWh and kW usage drop significantly. In this case the minimum bill could help 14 recover infrastructure costs that are: (1) sized for and reserved for a home’s electrical 15 service; (2) are not recovered through the monthly service charge; (3) are not used by 16 the customer or otherwise paid for in the absent month; and (4) cannot be used to 17 temporarily serve other customers, and charged to them. 18 19 In this case the customer isn’t paying for or using the demand—related facilities in the 20 vacation month. Nonetheless, the unused facilities may not be available to serve another 21 customer because they are not suitably fungible, or the facilities are needed 22 customer in a subsequent month and therefore cannot be shifted to someone else, or the 23 absence occurs in a month with low system loads. 24 needed to serve anyone else. serve the Therefore, the facilities are not 25 In this case, where a customer significantly, but temporarily, reduces their kW demand, 26 a large minimum bill could help pay for these infrastructure facilities, e.g. substations, 27 wires, poles, transformers, and power plants. 28 10 However, the minimum charge couldn’t be one-size—fits-all. Undoubtedly, the dedicated but unused facilities will be much higher for a large home versus a small apartment. For example, a large home with a monthly bill of $500 may have $350 of infrastructure costs per month (e.g. 70%), while a small apartment with a $60 bill may have $40 of monthly infrastructure costs. A minimum facilities charge of $30, the $20 service charge, plus $10 for the dedicated OLA-wa demand—related facilities may be reasonable for the small apartment, but it does not \] come close to recovering any reasonable portion of the $350 infrastructure cost for the large home. Therefore, any minimum bill would have to be tiered to the normal kW demand for the 1o 11 12 13 14 15 home — a higher minimum bill for the higher kW needed to serve the larger home and a lower minimum for the lower kW needed to serve the small apartment. In addition, the levels of minimum bills would have to be significantly higher than the amounts currently proposed by solar companies, residential advocates, or other proponents of this concept. Finally, because the minimum bill would need to be tiered by the home’s kW, the minimum bill concept provides little to no advantage over a demand charge with a 16 minimum billed kW. 17 18 19 2o 21 WHAT IF A CUSTOMER REDUCES THEIR MONTHLY KWH, BUT NOT THEIR KW DEMAND? The minimum bill concept is even more troubling when the customer significantly reduces the kWh energy but not the kW demand for their home. customer continues to consume the monthly kW, but not pay for it if they are served 22 under a two-part rate. Thus, there are no unused facilities that could theoretically be 23 used to serve another home. 24 In this case the As a result, no realistic minimum bill concept could adequately replace a demand charge for recovering these infrastructure costs. 25 26 27 28 11 1 This point is demonstrated through the following example, which is typical for 2 customers with solar generation. The assumptions and results are provided below in 3 Table 3. 4 5 Table 3. Minimum Bill Example for Typical Residential Customer Medium Size Home : 8 9 10 11 12 13 14 15 16 Typical Monthly Electrical Load Cost of Service Bill 2-part rate Bill 3—part rate Minimum Bill kW Used kWh Used Service Charge Demand Charge Energy Charge Total 6 6 6 1200 1200 1200 20 20 20 $36 0 $36 $72 $108 $72 $128 $128 $128 $25 $20 $20 $20 $36 0 $36 $18 $27 $18 $74 $47 $74 $25 Same Demand, 75% Lower Energy 300 6 Cost of Service 300 6 Bill 2-part rate 300 6 Bill 3-part rate Minimum Bill 17 18 19 20 21 22 23 24 This home consumes 6.0 kW demand and 1,200 kWh of energy per month. The customer’s monthly bill is $128 per month without taxes or adjustor rates. This is based on a $20 service charge, $6 per kW demand charge and $0.06 per kWh energy charge for the demand rate and $20 service charge and $0.09 per kWh for the two—part rate, which are similar to the charges proposed by UNSE. The $128 bill is the same whether it’s computed under a two-part energy rate or a three—part demand rate, and the cost to provide service is also $128 per month. 25 26 But what happens when the customer reduces their monthly kWh energy consumption 27 by 75%, from 1,200 to 300 kWh, and their monthly demand usage remains at 6.0 kW. 28 12 The cost of service is assumed to be reduced from $128 to $74, if the $0.06 kWh rate truly re■ected variable energy costs, because the utility avoided $54 in variable costs ($0.06 * 900 kWh) such as fuel and variable O&M. For APS’s residential customers this variable cost of energy service would be closer to $0.04 per kWh, not $0.06. But because UNSE provides most of its generation through power purchases rather than their own power plants, the higher variable cost number might not be unreasonable. The bill under the three-part demand rate is also reduced to $74. Thus, the rate structure is aligned with the cost of service, and the UNS customer received the right price signal \O for reducing the kWh energy, namely, the variable cost of energy service. However, the 10 bill under the two—part energy rate is reduced to $47, which is significantly less than the 11 cost of service. 12 customer received $0.09 per kWh to reduce their energy consumption when the variable 13 cost was only $0.06 per kWh; while the customer still “demanded” (i.e. needed) 6.0 kW 14 of capacity at some point during the billing month, and the utility needed to have the 15 infrastructure in place to meet that demand, irrespective of the overall reduction in kWh. The two part rate is not aligned with cost of service because the 16 17 What does a minimum bill do to correct the deficiency in this scenario? The answer — nothing. 18 19 20 21 22 23 A minimum bill of $25, which is around the high range proponents are discussing, has absolutely no effect at all. In fact, even for this medium size home the minimum bill would have to be at least $47 a month $27 variable cost for the kWh still consumed by the home 25 26 27 to make any contribution towards the infrastructure costs that are still used, but not paid for. In this case, the correct minimum bill for a medium—size home would be $74, the amount resulting from a three-part demand rate. 24 the $20 service charge plus the The minimum amount would include $20 for the monthly service costs, $36 for the demand-related infrastructure costs, and $18 for the fuel costs associated with the 300 kWh of usage, which would be included in the minimum. As in the first case, the minimum bill would have to be tiered to the home’s kW load, and 28 13 would have to be at a level that is significantly higher than that proposed by proponents of the concept, to be effective at all in recovering infrastructure costs. Figure 1. Minimum Bills Likely Needed for Small, Medium and Large Homes Versus Levels Proposed by Proponents Needed for Large Home Needed for Medium Home 10 Needed for Small Home 11 Discussed by Proponents high range 12 Discussed by Proponents low range 13 3 Z $0 $20 $40 $60 i 1 $80 $100 $120 $140 $160 14 15 16 17 18 19 20 21 22 23 24 25 26 27 In summary, a minimum bill, as proposed by various intervenors is not a viable alternative to a three-part demand rate to recover infrastructure costs. To be effective and fair, the minimum bill could not be one—size—fits—all — it would have to be tiered to the usual demand needed to serve each home. It would also have to be much higher than the service charge and the variable costs for any kWh consumed in the month to have any contribution towards fixed cost recovery. The minimum bill would also have to distinguish between customers that have both low kWh and low kW usage in a given month from those that reduce their kWh energy consumption but not their kW demand. In either case, the levels of minimum bill amount would have to be much higher than even the highest range discussed or proposed by proponents of the concept. Minimum bills of $30 for small homes, $70 for medium—size homes, and $150 for large homes would be the range of possibilities for a minimum bill to be a viable alternative to a three—part demand rate. 28 14 UNDER A MINIMUM BILL, DOES A CUSTOlVIER HAVE AN OPPORTUNITY TO REDUCE THEIR BILL BELOW THE MINIMUM AMOUNT? No. A minimum bill acts like an “adder” to the basic service charge. A customer would not have any opportunity to reduce their bill below this amount. In this example, if the minimum bill was $74 per month for a medium-size home, which includes 300 kWh of consumption, the customer would have no opportunity to reduce the bill below the $74 minimum. In contrast, under a three—part rate the customer would have an opportunity to substantially reduce the bill, in that a demand rate affords a customer who does in fact reduce their demand to reduce their bill — the very alignment and price signal that is \DO \IO\Ul 10 desired and not attainable with a minimum bill. It is also worth noting that a minimum bill does not send any‘effective price signal for customers to invest in home energy 11 technologies 12 13 14 ARE TOU ENERGY RATES A BETTER ALTERNATIVE THAN THREE-PART DEMAND RATES? No. WRA, and TASC assert that two-part TOU energy rates are a better alternative to 15 three-part demand rates. 16 They assert that the TOU rates are easier for customers to understand and are as effective in recovering infrastructure costs as the demand rates. 17 But this is incorrect. 18 19 PLEASE EXPLAIN. 2O TOU rates can have an important role in aligning rates with costs, but by themselves 21 they are not a viable alternative to a three-part demand rate, which can also have a time- 22 of-uSe structure. While TOU kWh rates have advantages over other two-part kWh rate 23 structures, they still do not and cannot adequately re■ect cost of service because they 24 inherently recover infrastructure costs through variable kWh charges. 25 TOU kWh rates provide an incentive to reduce energy during the on—peak hours, which 26 is helpful from a resource perspective, if this response is not consistent throughout the 27 28 15 Even though entire month, it will have minimal impact on the demand—related infrastructure costs necessary to serve the home. For example, if a customer reduces their kWh energy usage during half of the on—peak hours in a month, let’s say every other day, the utility would not likely be able to reduce the infrastructure investment needed for the home very much, if at all — certainly not anywhere near 50%. Just because the customer reduces their electrical demand on Monday, but requires the usual electrical demand on Tuesday, doesn’t mean the utility can permanently downsize the grid infrastructure, such as transformers, poles, wires, and other equipment needed to serve their home. In this example, the utility would not be 10 able to downsize the grid at all. 11 However, the custOmer would be overcompensated through the avoided higher on-peak energy charges. 12 13 Theoretically, these sporadic energy reductions could have some beneficial impacts at 14 the system peak level, which in turn drives utility’s power plant capacity costs. But this 15 would require enough participants in the TOU rate with a sufficient diversity of sporadic 16 energy reductions at different times of the day, and days of the year, to provide a 17 combined energy reduction that is somewhat more consistent across the utility’s critical 18 peak hours. However, even in this hypothetical case, the combined diversified impact 19 on system peak would almost certainly be far less than the sum of the impacts for the 20 individual homes. 21 22 In any event, a two-part TOU energy rate would not be likely to incent the type of technology or electrical appliance choices that are focused on reducing the home’s 23 electrical infrastructure requirements. 24 25 26 27 For example, some electrical appliance choices, such as instantaneous water heaters, may use a high level of demand, but lower kWh’s during the month, compared to alternatives. These devices can reduce the utility’s fuel costs to serve the home, but also require significantly more infrastructure investment from the utility. Two-part kWh rates, including TOU rates, would incent the customer to 28 16 invest in the latter choice, with the higher demand but lower energy requirements. Other home energy technologies would also be incented to focus on, and be rewarded for, reducing on—peak energy but not necessarily demand, under a two-part TOU rate. From a customer’s perspective, a three-part TOU rate provides three ways to save on their bill — shifting kWh energy usage to off—peak hours, reducing overall kWh energy usage, and reducing the on—peak demand, while a two—part TOU rate only provides the first two ways to save, as illustrated in Figure 2 below. In addition, these three ways to save are much better aligned with the utility’s cost of service, which creates a win—win 10 situation where the customer’s bill savings result in similar utility cost savings. 11 Figure 2. 12 13 14 Demand Rates — A New Way to Slice the Pie $100 $100 15 17 Lower Cost for electricity VARIABLE (Total Energy Usage) 16 VARIABLE (Total Energy Usage) 18 New way to cover grid costs 19 20 21 22 23 FIXED Ways to Save ' Lower Your Overall Usage Ways to Save ' Lower Your Overall Usage ' Stagger the use of major appliances ' Run appliances during non-peak usage times Consider home automation products 24 25 WHAT ABOUT APPLYING THE DEMAND ONLY TO ON-PEAK HOURS? 26 Some parties including Staff, RUCO, and WRA have proposed that demand charges 27 should only apply to on—peak hours for residential customers. 28 17 APS believes this argument involves the inherent tension between theoretical precision and practical application. In general, APS supports applying the demand charge to the on-peak hours And this opinion is for residential customers, but only under certain circumstances. driven more by practical considerations than by theoretical precision. PLEASE EXPLAIN. All utility charges are a mix of theoretical precision and practical application. The perfect charges, whether they are demand charges, energy charges or something else, would likely be too complex and expensive to implement. For example, the best demand charge would likely be two demand charges — an untimed (non—time—of—use) 10 demand charge to recover distribution costs that vary with the size of the home and are 11 driven by the neighborhood peak, rather than the system peak, and an on—peak demand 12 charge to recover power plant infrastructure costs that are driven by system peak hours. 13 This could also be accomplished through separate on—peak and off-peak demand charges 14 that would differ by the power plant capacity costs. 15 16 However, APS believes that this structure would be complex, at least initially, and 17 therefore 18 Therefore, APS believes that an on—peak demand charge is a viable option for 19 recovering the costs of both power lines and power plants if the following conditions 2O are met: (1) the monthly service charge recovers the grid costs from the meter, point of 21 delivery, service drop to the home, and the distribution transformer (along with the other 22 customer-related costs such as meter reading, billing, and customer care); and (2) the on- 23 peak period is defined to include the hours that typically drive the design peaks for 24 residential feeders and substations as well as the system—peak hours that drive power 25 plant costs. more appropriate for business customers than residential customers. 26 As a practical matter, a uniform un-timed demand charge could work because residential 27 customers largely peak on the system peak. 28 18 Under this scenario an un—timed demand charge would reasonably re■ect the infrastructure costs for each home, without unduly complicating the bill. SHOULD THE MONTHS? DEMAND CHARGE ONLY BE APPLIED TO SUMMER No. The demand charges could be somewhat higher in the summer months compared with the winter, but APS does not recommend only applying the demand charge to one month or one season. From a theoretic standpoint, as discussed above, not all infrastructure costs are driven by system peak months or hours. For example, the distribution grid, while typically sized for summer load, is needed to serve homes 10 throughout the year. And while it may be cost justified to apply a significant portion of 11 power plant costs only to the core summer months, from a practical standpoint 12 residential customers already face high summer bills, and this option would exacerbate 13 that issue. 14 WILL THREE-PART DEMAND RATES HURT ENERGY EFFICIENCY? 15 No, it will refocus energy efficiency and turn it into a better resource for the utility and 16 provide means by which customers can exercise greater control over their utility bill 17 through demand management. 18 19 PLEASE EXPLAIN. 20 As discussed in my Direct Testimony, APS believes that home technology investments 21 can be an important resource for meeting future power needs — if the investments are 22 properly incented and focused on reducing the costs for both building and running 23 power plants in the future. 24 programs on investments that reduce operating costs, such as fuel and variable O&M, 25 but not the costs of the power plants themselves, which are more significant. A three- 26 part demand rate incents home technologies and energy efficiency investments that can 27 reduce both of these costs. Currently, the two—part kWh rates focus energy efficiency 28 19 Concerning the customer’s potential bill savings, a three-part demand rate will have a lower kWh charge compared with a two-part kWh rate. But, under the three—part demand rate, the demand charge will provide the customer an additional incentive and opportunity to save on their bill. Therefore, APS believes that three-part demand rates can result in viable opportunities for energy efficiency programs, and those programs will have a much higher value to the electric system because of the potential increased utility cost savings compared with those that are primarily focused on energy savings. WHAT WERE VOTE SOLAR’S REMARKS EFFICIENCY AT MIN GUS HIGH SCHOOL? CONCERNING ENERGY 10 Vote Solar witness Kobor testified that Mingus Union High School District (“Mingus”) 11 was harmed because APS implemented a demand charge to their bill after the customer 12 invested more than $1 million in energy efficiency. Because the estimated bill savings 13 from the energy efficiency project was apparently targeted at reducing kWh energy and 14 not kW demand, the actual bill savings were much lower than expected, which reduced 15 the investment’s net benefits. 16 17 DO YOU AGREE WITH VOTE SOLAR’S TESTIMONY REGARDING THE ENERGY EFFICIENCY INVESTNIENTS MADE BY MIN GUS? 18 No. 19 20 21 22 23 24 25 26 27 28 PLEASE EXPLAIN. In 2013 and 2014, Mingus implemented energy efficiency projects apparently targeted at reducing kWh energy and not kW demand. As a result, the actual bill savings were much lower than expected, which reduced the investment’s net benefits. along with all business customers with loads greater than 20 kW Mingus — has been subject to a three-part demand rate for decades. Therefore, contrary to Vote Solar’s testimony the demand charge was in place many years before the energy efficiency investment took place. An adjustment to how the demand charge is calculated during a low—load month was approved in our last general rate case and effective in July 2012, which was also 20 well before the energy efficiency measures were installed. Unfortunately, it appears that either Mingus or their third-party vendor may have miscalculated the anticipated savings from its investment in energy efficiency. APS is sympathetic to our customer for this situation. We know that utility bills are especially important to our schools because of relatively tight overall funding and limited control over significant portions of their operating budgets such as teacher’s salaries. As such, we also know it is critical that the investments they make in energy efficiency, solar and other technologies produce sufficient savings in utility bills or other 10 operating costs to justify their cost. 11 While APS does not know precisely how the savings miscalculation occurred for the 12 project, a plausible reason is that Mingus (or their third-party vendor) may have 13 overestimated the savings by dividing the total bill by the monthly kWh to get an 14 average savings per kWh, rather than calculating specific expected savings for the 15 demand and energy components of the bill, as they should have. Had the estimated bill 16 savings included both the demand and energy components, the actual bill savings would 17 have been more in line with or even surpassed their expectations. 18 estimated bill savings from the reduction in monthly kWh were overstated compared to 19 the actual bill reduction. As a result, the 2O 21 22 STAFF’S RECOMMENDATION FOR CUSTOMER EDUCATION ON THREE— PART RATES 23 DO YOU AGREE WITH STAFF THAT CUSTOMER EDUCATION IS AN IMPORTANT PART OF IMPLElVIENTING THREE-PART RATES? 24 Yes. 25 their bill, how the bill is calculated and the actions they can take to save and mitigate 26 potential impacts from rate changes. This education can emphasize that under a three- 27 part time—of—use rate, for example, the customer would have three ways to save on their APS believes that it is important to educate customers about all components of 28 21 bill — lower their overall monthly energy use (or kWh), shift energy to the off—peak hours, and lower their monthly peak usage (kW demand) during a specific on—peak period. Customer education would also focus on the use of modern technologies such as home energy monitors and controls, smart thermostats, advanced air-conditioners, battery storage, and smart inverters. These devices help manage and reduce peak usage, thus allowing customers to better manage and reduce their bills. 10 11 12 13 14 15 16 DOES THIS EDUCATION SUGGESTED BY TASC? HAVE TO BE HIGHLY COMPLICATED AS No. Not at all. We have found that the demand charge concept and strategies to save can be, and should be, explained very simply for the general customer group. Customers don’t need an energy engineer in their home, as suggested by TASC in a Salt River Project proceeding, to understand either. Additional detailed information can be made available on the utility’s website, or through other education channels, for the customers that are interested in learning about further specific details. 17 DOES APS HAVE ANY EXAMPLES OF THIS EDUCATION MATERIAL? 18 Yes. Attachment CAM—28R is a sample draft customer education piece we are currently 19 developing as we contemplate proposing an expansion of our three—part rate program. 20 The piece is not yet complete. 21 examples of education concepts. See also APS’s response to RUCO Data Request 1.6, 22 which is attached as Attachment CAM-38R and incorporated into my Surrebuttal 23 Testimony by this reference. 24 three—part rate program. However, APS thought it could be helpful to provide This material includes information used for our current .25 26 27 28 22 VI. UNSE’S PROPOSAL FOR A SOLAR PURCHASE RATE WHAT DOES UNSE PROPOSE CONCERNING A SOLAR PURCHASE RATE? UNSE proposes that excess power from rooftop solar that flows back to the grid be purchased by the utility at a solar purchase rate and credited on the bill each month. Their proposed solar purchase rate re■ects the purchase price from a large grid—scale solar plant, which would be revised from time to time. WHAT IS APS’S POSITION? APS believes that the best rate for excess power would be an avoided cost rate. UNSE’s proposed purchase price from a grid-scale solar plant could be reasonable, but 10 should be the maximum considered for purchase of excess generation from rooftop 11 solar. 12 13 VII. CONCLUSION 14 PLEASE SUMMARIZE YOUR CONCLUSIONS. 15 APS recommends that the Commission approve three—part demand rates for UNSE 16 residential customers. This is consistent with the proposals made by Staff and UNSE in 17 their Rebuttal Testimony. APS also recommends that the Commission adopt Staff’s 18 recommendation for a transition period combined with customer education. APS further 19 recommends adoption of UNSE’s net metering proposal. 20 21 DOES THIS CONCLUDE YOUR SURREBUTTAL TESTIMONY? Yes. 22 23 24 25 26 27 28 23 Attachment CAM-1 SR Page 1 of 9 RESIDENTIAL UTILITY CONSUMER OFFICE'S FIRST SET OF DATA REQUESTS TO ARIZONA PUBLIC SERVICE COMPANY IN THE MATTER REGARDING UNS ELECTRIC RATE CASE DOCKET NO. E—O4204A-15-0142 DECEMBER 22, 2015 RUCO 1.2: APS’S Residential Three—Part Demand Charge Based Rates — On page 7, line 22 of APS witness Charles A. Miessner’s rate design direct testimony he states that “We looked at a sample of customers that switched from an energy—only time—of—use rate to the three-part demand rate and found that about 60% of those customers saved on their demand and energy. We also found that those who actively manage their demand have achieved demand savings of 10% 20% or more. On average, customers on the three-part rate reduce their monthly demand by 3% to 4% depending on the season. These customers also tend to save on their on-peak and monthly kWh usage after switching to the three-part rate." Based on that statement please answer the following questions: a. Please state the methodology that APS employed to select its sample. b. Please specify the number of residential customers under this plan that were used in APS’s sample? c. Please provide the worksheet and criteria used to justify the statement that “60% of residential customers that switched from a time of use plan to the APS residential three-part demand rates saved." d. Please identify the 40 percent of the sample that did not save, and reasons why they did not save given APS’s criteria. Response: e. Please provide your calculations, criteria, and supporting documentation to support the statement “We also found that those who actively manage their demand have achieved demand savings of 10% - 20% or more." f. Please provide your calculations, criteria, and supporting documentation to support the statement “On average, customers on the three-part rate reduce their monthly demand by 3% to 4% depending on the season. These customers also tend to save on their on-peak and monthly kWh usage after switching to the three—part rate." a. Information about the sample and the selection method is provided in the first page/tab of Attachment AP515766. Witness: Charles Miessner Page 1 of 2 RESIDENTIAL UTILITY CONSUMER OFFICE'S FIRST SET OF DATA REQUESTS TO ARIZONA PUBLIC SERVICE COMPANY IN THE MATTER REGARDING UNS ELECTRIC RATE CASE DOCKET NO. E-04204A-15-0142 DECEMBER 22, 2015 Response to RUCO 1.2 (continued): b. The total study size was 977 customers, constituted all customers meeting the criteria. Attachment CAM-18R Page 2 of9 which c. The summary information is provided in AP815766. cl. The summary information for the customers that did not save under a demand rate is included in AP515766. Typically these customers did not save under a demand rate because their on-peak demand was relatively high in relation to their overall energy consumption and it appears they did little or nothing additional to manage their electrical usage patterns. e. As shown in the attachment, the top 20% (most successful) savers reduced their bills by 10% to 20% or more under the demand rate. f. As provided in the attachment, the average demand reduction for the sample was 3% to 4% while the top 20% reduced their monthly demand by roughly 24% on average. Witness: Charles Miessner Page 2 of 2 Attachment CAM-1 SR Page of 9 ARIZONA PUBLIC SERVICE COMPANY Residential Demand Rate Analysis Background: Analysis performed in 2015 The purpose of the study was to assess the Impact ofa three-part demand rate on demand, energy, and monthly bills for residential customers. The study isolated the demand chage impact by comparing the same customer before and after switching to a three-part rate. Since the three—part rate was a time-of—use rate, APS compared customers moving from a two»part TOU rate with similar on-peak hours. The study specifically compared the two-part Rate ET—Z with the three-part Rate ECT—Z, both having on-peak hours of 12 noon to 7 pm weekdays. Sampling Frame: Phoenix Metro customers Switched from ET-2 to ECT-Z in 2013 Had 12 months billing data in 2012 and 2014 Resided in same home for the three year period Total sample size 977 customers Adjustments: Load data was normalized for temperature and humidity for summer months. Winter months were not adjusted because correlation factors between load and weather were very low. APS15766_Demand Rate Analysis.xlsx Background Page 4of9 &BiImpacl ts Load ACAM-SRt achme1nt mwmmwmmmmwwwmmwmmmmmm APRatSA1n57a6ly_sDeim.axnldsx (5%623-217)34)483)-(7%.30%9-02%39%3.5%)94) 10%(41-344)34)210)-19%(.4%687%26.5%)07) 15%(3(12-86)39)247)(1.1%6-3%29%4.)2%35) 20%(312-64)117)-46)14%.(7%-363%26%2.)1%67) 25%(89) (32—11-58)69)-1.(4%—25413%%8.)01%5) 30%((76)1-096)820)-1(.7%11%4—19%1%5.)6%1)35%(99) (48) (51) (-04-8(3.%9-71%7.)4%68) 40%((66) (41) (38) (-03-6(2.%-4310.%0%)38) 55%(31) (25) (6) (-010%4-(.%219.%0%)28) 60%(12) 7(190%0-1%2-(.129%%2.%)88) 65%2(4)60.0%-11%(1-%67.%45) 70% (96) 1-062)69-.(5%84715.%)4%06) (40) 45%(11) (29) (-02-5(.1%62-51%9.)1%43) 50%(78) 688600.3%1%4%3%(21—S%4.64) 75%37(4)0.0%1%4%(3—6%17.65)80%181251560.8%4%9%6%(5—37.4%9) 85%200451550.8%7%9%(70%1.01) 90%14452920.6%9%5%12%(9-31.1%1) 95%256631931.11%10%16%7.23%82 151916635302.25%34%22%33%%4118%.43 kWhTot%COnOfukWOskWh%TotOntOfonf-mkW%—aPChef%-rlsakP—lakPnkge CSPUBLAOREMIZPVOAICNEAY RAnalDeRatesmideanyetiasdlis skWcbydur%tmorhastaui■nmitenhgdsre kW, ckWh, andThefbimoinrshetrwoangetosarinutelmocth-iepnlyagrt LoadCh(SNWhutomer1amemp,nBilSitzdhgeurimetyrl)er (70) A(37) (32) (v-0e2S—3.r(a392-g1%9.1).e91%%88) ~23% -19% -11% —14% —13% -8% -9% -9% -5% -6% -4% ofPage 95 Bimpacl ts ILoad 1%0.77 5.4%20 9.8%10 (27.63) (25.31) (13.58) (18.44) (16.23) (10.51) (10.56) (13.28) (6.04) (7.40) (5.18) (-7.%61) (-3.%20) (-4.2%0) (-1.6%0) (0%0.26) 111%3.41 (-76.%13) CAM-At achme1nt SR 13mWinter 2% -1% -4% 0% 2% 3% 4% 4% 6% 13% 17% -26% -18% -7% -10% -9% -6% -3% -S% -5% —2.2% 5ChkWOn%-aPnkge 3% 1% 1%1 2% 3% 10% 17% -19% 1%-1 -6% —10% -12% -5% -9% -5% -1% -2% -1% -2% -1% -2.0% kWh%Off-Pk 0% 1% 5% 0% 8% 10% 8% 16% 10% 14% 26% 32% 1.7% -29% -18% -10% -13% -9% -2% -1% -3% kWh%On-Pk 1% 4% 2% 12% 4% 5% 13% 19% 0% 0% -21% -12% -7% -10% -12% -4% -8% —4% -2% -2% -1.3% kWhTot%al (1.2) (0.9) (0.3) (0.5) (0.4) (0.3) (0.1) (0.3) (0.3) 0.1 (0.2) 0.0 0.1 0.1 0.2 0.2 0.3 0.6 0.8 (0.11) OnkW-Pk (5) 25 7 104 27 3o 98 (6) 163 (20) (182) (115) (66) (108) (125) (46) (92) (54) (24) (12) (17) OfkWhf-Pk (61) (45) (23) (32) (221 (5) (3) (9) 4 kWhOn-Pk (5) (1) 12 45 23 137 53 58 151 231 (16) (242) (159) (88) (140) (147) (52) (94) (63) (22) (18) kWhTotal %Customers CSPUBLAOERMIPZVAOICNCYEA AnalDeRRatesmidaeynestdials modurckWby s%thrsanuangemtihfnesrdg N(No WeLoadChoWirmanltizhaetgiorne) 5% 10% 15% 20% 25% 30% 35% 40% 45% 50% 55% 60% 65% 70% 75% 80% 85% 90% 95% 100% Average ARatPSn15a7l6ys_Diem.xalnsdx ofPage 96 &B'mpacts Load CAM-ASRt achme1nt ARatPSn15a76ly_sDiem.xanlsdx Notes: aandE1.dtxajcuxlsedtei.nrsg kWhOfOnTot%Ckwu5kW%OnsChfto-mP—aekrlPsnkge CSPUBLAOERMIZPVOAICNEYA AnalDeRRatesmideaynetsidails modurcbys%kwthrsauntmifnhegsdre 1ChAnLoadAnBinaunagulael (5%-14237%(47)30)-365%82).1%4%0.12%7)8) (89) (10%-321(02)-2413)6%2%.5%1%5.33%5)9) (81) (15%-21(—137)-356)2%8%.11%8.08%49%)6) ((75) 20%43%246%10-(152)477).2%0.3%98%5)6) (25%-2105(—1152)33%97)2%.5)4%1%7.81)%9) (41) ((83) 30%1-0—6%7(924)-2.1%8.64%0)6) (97) (26) (35%-—6%05(-27.1%4.42%1)2) ((37) (71) (19) (50%-0(34-2‘1.9%8.1%8)9) (16) (31) (9) (7) (55%—1%-0(—22-1.8%7.1%)23) (5) 40%-10S—6%(8-513)2.1%9.53%1)7) (23) (75) (8)(45%-04%1(3-6.%97.4%)73) (50) (14) (6) (60%10%-0(2-1.8%5.1%2)5) 65%700.0%2%(1%1—510.3%3) 70%5614433%0.2%(-641.9%3) 75%131210.1%3%0%4%(-1620.%92) 80%159291300.9%7%10%5%(-13.%15) 85%12736910.7%8%6%(-15.%30) 90%101406%610.10%4%(-16.%8) 95%2045814612%0.14%11%15%8.4%946 100%3751172581.23%33%20%26%216%75.42 (43) (29) (14) (A—2-03v2-1e.9248.r03a%1)g5e0) DatLoada Page 7of9 ACAM-SRt achme1nt APRatS1n5a76ly_sDiem.xanlsdx (MoSMAvgWiuNamoAynv-—tLAvgMoOehFnArcolauypt)rcniltdohrly CSPAOREUMIZBPVOLAICNEAY RAnDRatesimdaelnyetisadlis dkwcby96smthrusaotrnim■nhegdsre D(TRaehimrea-otnfepduasret) 2014)(LEcoaCyleaTn-dZarar kToO%CkWTuOfWnsfkToOto—-OfmfWPnkWOatheS-rkWilnAsfuPahm-kPnltkuearl 5%13459374.149788,3..71247370,4.5349%37%43%01719 10%43215.199963,4.8143169,4.86539%41235910%4 15%252617.206,4.1366625.50,34635%38%7028931%6 256617.221,024.71%394,'06.3427,9534817%3 25727.15217881,4.16394%06.92,41%632%37%7523485072 3587217.2340939,4.216%41054,6.7329517013%2 2358117.215,59214.216398%5,6.17643%333%438%8297057 4637218.0254,5.519446%361,6.27943%34659271%5 4260217.5217,9944.371%24106.5,1735%49832860% 2563818.,0255415.8%4473,506.18,33%4367093% 2562017.,23345.18%4262,706.84,138%3082450%8 2657117.,01968854.216%38404,15.8636%0321970% 2662418.,523699845.17%2430,6.283434%037%6296%2 2763118.,022495835.17%42880,6.5733%437%03285%7 276168.1231,597044.18%42420,6.184.38%321594%1 80%264618.,24055.813,443406.19,39%364125760%8 2869318.,5274615.9%248484,197.504,32791458632% 90%26498.1,23599354.81442236,6.841%535%38%9473610 95%671218.,26055.81,346660797.133%4,37%572017%5 216548.0,2821625.9468%4,0157.264,73512845%7 2A5937.1,11229v,99434.e41127r146.6,a38%3g598e273%2 DatLoada 8of9 Page SRCAM-At achme1nt ARatPnSa15l7y6s_ieDxmlasndx L(FANMoAvgWioanMScvaun-tmAoydaephrOtliyc)rlt 346.39419694.2107.85,5792375%%127,5438197 36.440515.24497.,725664810%37,262%84162 40%35%36.14475.22988.,66524915%0%8,172659308472 3417.4695.2548.8,6832591%3,506598%325 33%38%46.14505.2388.,2661425%80%27,95347523 346.14519865.2398.,60663238%,72304857%48 38%41%35%6.42414.2187.,630237049,615328%143 347.483125.2638.95,70326014%2,7804219%35 346.42415.2178.7,3063224%5,1769306%79 40%34%37.46615.25278.,679250%490%3,862754936 346.43415.2227.82645,450%2,8149071%56982 3439015.9024.1967.583,95260%60%,8372956302 37%6.40%34%42315.2188.,62923860,079154280%319 340%6.14.4149332047.62,62337%13,9628059%70 346.41219634.2157.95,6092387%126,9057412%86 33%346.14149574.20777.,2621380%%19,746381549 346.44815.24888.648,.249%13,8725463%875 346.40214.20696497.5975,23%70,163479%041 346.40819544.20787.2608,379%1,82356091%42 39%45%34%5.13514.2149796.,4882156,09408%37 CSPOAEMURPIBVZALOCNIYEA AnDRatResmaidleynstdiasl mdkwcby%sohturnamthin■segrde (TRaEimwen-oeft—-rpugsayer)t 2012)(LcETayolean-arZdar AOfWikWSOnkTonuWOfkOnmfTo-%CuWtPfeah-skPrtlohamkelrs 346.42515.2257.8,626240%A5,72v4193e86r9a4ge BiMontlhly 9ofPage SRAt achment CAM—1 ARatPnSa15ly76si_eD.xmlasnxd $$ $$ 5S $$S S$ $5$ 5$ $$ $$ $$ (TRatEni(mTDwRateim-oehm-frepougafse-nyurd)ptsea)rt 11BiMoAET-vBiMoAecenlrvtahZnelrgt-yhaezglye CSPUBLOEAMRPIVZAIONCYNCEA AnalDeRatResmidyaeenstdisal modurkWcby%shtnurmahtinsfeirgde AnWiS%CAnunWimsSut%Comenaemuslrtomaeslre s 1$25%91S67535%.047.1096185 15210%9361574510%.0961358.0972 25115%25917415%1.32570.31965 2$1320%$1059423720%1.896714.35086 2$125%21$95057425%.3125976.1356 25130%0281$64730%1.5842.180986 $1235%91$72635%.1359628. 49028 $21340%41$62940%.67189.368250 15245%9$71502845%1.498.17517 2$150%31$942650%.568.01298 2$155%02$81543555%.17684.73419 12860%16502760%.30967.35742 2165%02719265%.1604782.5934 12970%71342870%1.5963481.9860 1275%9172875%1.5762439.18067 1280%917562980%.683417.059241 2185%38214085%.3289147.539152 1290%971462990%.165278.243671 12995%736120.95%1756. 2976 1271326094100%.371425%7.91742 210A2713v928.1e5A436r29.va803eg71r0age Notes: andaEt1.dxcjuelsdt.oinrgs 10f2 ACtAacMhm-2e8nRt Cover drumasRev4} tratesrfiv—ateis howloyoeeneryou nweryourergyb,ilandHersisaruamogvmeepl.stieons; Softmausegjor‘aeprltiahnec s 1) (dreAheatwatleycC,terreicr, ) Whenpwaitodoluntaoinsundrtiebtiler.y tafevhdonecdiyouaroetneokninerernig. Runadurpntoiusmlna-nipesnacegeagsk 2) Setdtoidelscyclyourhronawuashnaerye andpool put tyour ipump monaer Cohomeapunrtosmidaeutirocnts 3) Ptrhoandlecgormanadotrsmtlaebsrle rgrtoepeak aarwaydyouruseucataege wigiittsSetheavoughtcyouroiuntdg. andtpdoaalhraomdeutlecrst, formayou.naging Lowerovyouruseralge 4) Takeaofbasidevnseaorgnmyt-eacfgi ency andtitofflpouirpogolrn-htasminss,g fandwhenlaesetaravinsyour oom,ting tfhdegrhieinetarmhsgwousmheretaeser andlintowisLEDs, htoandwwernietcehrin,g tEnerAnaonlourtokisauilnyrgvzeyr , fjeunaaw.mste 2of CAtaMch—m2e8nRt 2 Page drRev u4} mastraetr—ifvatels ofStasihunolusmrenatpsvioecearis’fgete sheishermajalnotrSt80awiwhiatuponnipayelniaveclneoc.egrs bithiwiJraher20yomxaspionimimoslucvaletnaltl, Why? WiMore Who Save Money?l It’Jtshalispeak Btpeak lsgroehyourcauweresveuataisneggser, OfWakW0.0nsha04efr TkWOn0.W'Tel vis1on STEVE Stprwho putSonaHTJeiwho iscOEAoVMve,mornErIcM'SisetEryned and cocmowihome nvmfeocneritsmnOnkwhce,s fand.worhomeroOvkwOn1.kWfworrmeo3k.nmk take an exampl Letata;’s e the usitathsey’amenrge timojoneummestntae,inf look ~ l greateston h'to75'sherACTur79°downto0onkw0,DrnokWnys.er eCookOnkdl ivnx0srCookodmnernesr. s PekWUskw3.7.aoff5lkgoeadswTVHasthone TVHasthon-e Tums ACdown ft.1,b(tdurhEeyaoxasq.onarvs8pmue00niemcpdnslaergsl) RESIDENTIAL UTILITY CONSUMER OFFICE'S FIRST SET OF DATA REQUESTS TO ARIZONA PUBLIC SERVICE COMPANY IN THE MATTER REGARDING UNS ELECTRIC RATE CASE DOCKET NO. E-04204A-15-0142 DECEMBER 22, 2015 Attachment CAM-38R Page 1 of 10 RUCO 1.6: APS'S Residential Three-Part Demand Charae Based Rates — What programs, tools, or customer support does APS provide for customers on their demand charge based rates? Please explain how APS currently educates customers who switch to a demand charge based rate? Does APS see their education strategy changing over time? Response: APS customers receive information on available rate plans, including the Company’s three-part residential demand rate, in several ways. Customers calling APS requesting to set up a new account are asked a set of questions about lifestyle choices and property characteristics designed to assist both the customer and the call center associate to determine which rate plan might be the most beneficial to the customer (for example: Do you have gas service at your home? Does your home have a pool? Are you at home during daytime hours?). Call center associates also provide a description of how APS’s threepart residential demand rate works when an existing customer contacts APS through the call center to inquire about changing rate plans. In addition, customers that request additional information will be directed to aps.com or, if the customer cannot access the website, a letter can be sent that explains each of the available residential rate plans (this letter is attached to this response as AP515759). In each of the Company’s customer service offices brochures are available to residential customers that discuss each of the available rate plans to assist customers in determining which plan might be appropriate for their circumstances. This brochure is attached as APSlS760. The above information is also available at aps.com for customers to review at any time. In addition, to assist customers in choosing an appropriate rate plan, the Company offers a rate comparison tool for existing customers that will use the customer's actual energy usage history to provide an overview of each available rate plan. This information is accessed through each customer’s individual online account. Screenshots of this tool are attached as APSlS761. APS provides customers with information regarding rate plans and rate plan options on an ongoing basis and will continue to assess new education strategies. Witness: Charles Miessner Page 1 of 1 Attachment CAM-38R Page 2 of 10 January 6, 2016 JOE & MARY SMITH 400 N 5TH ST PHOENIX AZ Dear Joe Mary Smith: Thank you for contacting APS regarding our residential service plan options. We recognize that when it comes to energy usage, different people have different needs. That's why we offer several electric service plans — so you can ■nd the one that is most convenient for your lifestyle and saves you the most money Please see the detailed information that is included in this letter for more information Additionally, you can ■nd complete information on our Web site at www.ags.comfrates. and can also perform a comparison and ■nd out about other options such as Green Choice rates. If you have any questions regarding this or need additional information, please feel free to contact our Customer Care Center at (602)371—7171 or (800)253—9405. Associates are available 24-hours—a-day. Or, visit us online at wwwapscom We appreciate your business and the opportunity to serve you. Sincerely, 5W Bernard APS Customer Care Center APS1 5759 Page 1 of 4 Attachment CAM-38R Page 3 of 10 Standard This plan helps those who use less energy save money. It doesn't make any difference what time of day electricity is used. This plan may be best for you it: a 0 You generally use 1 000 kWh or less each month due to the size of your home and type of appliances. You live in a home, mobile home, condominium or apartment that is 1,100 square feet or less. You do not have a swimming pool or spa that is electrically heated. Here's how it works: 0 In the summer (May-October) you are billed at different costs per kilowatt hour (kWh) depending on your energy usage; a The ■rst 400 are billed at about 97¢ c The next 400 are billed at about 13.7¢ c. The next 2,200 are billed at about 16.3¢ a All remaining kWh are billed at about 17.4¢ - ln the winter (November-April) the cost is about 94¢ per kWh used. Time Advantage 7 p.m.-Noon This plan is best for those who have minimal energy usage during on-peak periods (Noon-7 p.m., Monday-Friday), especially during the summer months. This plan may be best for you if: 0 o a a - You generally use L000 kWh or more each month due to the size of your home and type of appliances. You are not home during the day or have low daytime energy use. You are able to use your dishwasher, dryer, washer and range more during off-peak hours. You are able to operate your major electric appliances such as the water heater, pool pump and spa heater during off—peak hours. You have a programmable thermostat or can set your air conditioning to a warmer temperature during on-peak hours. Here's how it works: a The plan is billed on an off—peak and on-peak basis. - Off-peak hours are weekdays from T pm. to Noon and all day Saturday and Sunday, as well as six major holidays". Electricity used during off-peak hours is billed at a lower rate. - On-peak hours are Monday through Friday from Noon to 7 pm. and are billed at a higher rate. - in the summer (May-October) you are billed at about 24.49% per kWh used on-peak, and 61¢ per kWh used off—peak. - In the winter (November—April) you are billed at about 19.8¢ per kWh used on-peak, and 61¢ per kWh used off-peak. .:. Note: The APS meter readers must have safe, unassisted access to physically touch the meter each month. APS1 5759 Page 2 of 4 Attachment CAM-38R Page 4 of 10 Combined Advantage 7 p.m.-Noon This plan is best for those who have minimal energy usage during peak periods (Noon-7 p.m., MondayFriday). You can save on your bill by adjusting when you use energy and how much you use at one time. This plan works just like the Time Advantage 7 p.m.-Noon plan, with one main difference: a This plan has a Demand component, which is the largest portion ofthe bill and is billed in addition to the charge for on-peak and off-peak kilowatt hours used. -:- The Demand (kW) is the one on-peak 60-minute period of the billing cycle when you use the most electricity. This plan may be best for you it: 0 You are able to spread out your use of major appliances during on-peak hours, so you are not using them all at once. 0 You are able to operate your major electric appliances such as the water heater, pool pump and spa heater during off—peak hours. Here's how it works: - The plan is billed on an off-peak and on-peak basis, with an on-peak demand component. .:- Off-peak hours are weekdays from 7 pm. to Noon and all day Saturday and Sunday, as well as six major holidays“. -:- On—peak hours are Monday through Friday from Noon to 7 pm. and are billed at a higher rate. :- The Demand is the one on-peak 60—minute period of the billing cycle when you use the most electricity. It is also the largest component of the bill. - The chart below shows the costs for the different timeframes and components of the bill: - Summer (May-October billing cycle) Cost Demand charge (kW) On-peak kWh used Off-peak kWh used $13404 per kW 8.845¢ per kWh 4.363¢ per kWh Winter (November-April billing cycle) Cost Demand charge (kW) On-peak kWh used Off-peak kWh used $9.203 per kW 5.815¢ per kWh 4.273¢ per kWh To save money, it is important for you to use more energy on weekends and weekday mornings before Noon or evenings after 7 pm, since electricity used during off-peak hours costs less. It is also important to limit the number of appliances that you use at one time during on-peak hours in order to minimize the demand charge. .2. Note: The APS meter readers must have 3a fe, unassisted access to physically touch the meter each month. APS15759 Page 3 of 4 Attachment CAM-38R Page 5 of 10 Time Advantage Super Peak i' p.m.-Noon This plan is best for those who can signi■cantly limit energy usage during peak (Noon-T pm, MondayFn'day} and Super Peak periods {3 p.m.—6 p.m. Monday-Friday. summer only). A smart meter must be installed at your home to select this rate. This plan works just like the Time Advantage Torn-Noon plan= with one main di■'erence: You are able to use more of your energy during off-peak hours. and can signi■cantly limit your energy use during the summer between the hours of3 pm. and 6 pm, Monday through Friday in the billing months of June through Aug‘u st. Here's how it works: i The plan is billed on an off-peak and ion-peak with a Super peak period in the summer billing months of June through August. c- Off-peak hours are weekdays from F pm. to Moon and all day Saturday and as well as six major holidays". :- On-peak hours are Monday through Friday from Noon to ? pm. and are billed at a higher rate. c- Super-peak hours are Monday through Friday from pm. in the billing months of June through August and are billed at the most expensive cost per kWh on this plan. I The chart below shows the costs per kWh for the different timeframes: Off-peak Nov-Apr Per kWh May, Sep, Oct IP‘er kWh ? All day Sat-Sun and six holidays On-peak tMon-Fri) Noon-3 p.m. p..m.-eB pm. 6 p.m.-?' p.m. - . . Nov-Apr Per kWh 19.8254; Jun-Aug Per kWh . . May, Sep, Oct IP'er kWh Jun-Aug Per kWh 19.82591 To save it is important for you to use more energy on weekends and weekday mornings before Noon or evenings after T p.m.. since electricity used during off-peak hours costs less. “Holidays for T p.m.-floon [Rate Plans: New Year's Day [January 1] Memorial Day {last Monday in May} Independence Day tJuEIy 4] Labor Day [first Monday in September} Thanksgiving Day [fourth Thursday in November} Christmas Day [December 25in NOTE: If these holidays fall on a Saturday, the preceding Friday will be o■-peak- If they fall on a the following Monday will be off-peak. APS15759 Page 4 of 4 PaM21ge Page 6of10 ACtAaMch-m3e8nRt toans lifyour APS Service Plfit estyle Oaps.corn/plans APSPeRTiembaket OurPeRTiepmsbraogektram dhelteandsyoiganevpurgdy Byptiuamhrnniocqpseautyine.g apcyrmeotagurnumle.y byredyunosreciungry peakteiwvmemenhtse—nsrtigy diheattsmgahnedst. wifdWe186trepiosmoeglnaekt feStJrhvoueptmnuesgbhr. Ewitpvnotaelkntecls wheonoorlkiednays . Efwi■vlheaovcestulrhnets b2aeptwn.emdn. Foraindfobritmoantuiolt TPRcaloiepmbruaogerktalm, (36atm7use10tr-o3—26Ph)o0enix) (685a9oro—r0te8ha14es)8r. TiASuPedvmapnetakgre 7Noom.on Ifrecyduanosecnurgey 2bydph0ormsuoe%ruiapneskgr of36dandwpm.puemrkidna.gys. tJandAhumorongsuesht ofodheyfuo-npreiunagsrky tfbeplCbhmaoyenaiosrndutyer. iftplhyaiosun: k1Usem.Wor0aocnehrt0. dUswisyhawosehuienr.g mdandarmcyhonierg.e odhfuo-preinsa.gk Usama(pywloiaejnutcers hpooldespuaptrmeinapr.g) ohfo-upersa.k Htphroaegmvoesatble ACstorycaweormntuer dtandaeumnp-reiantugrke shuopeur-pse.ak CAodmvbainetadge 7Nopm.—on Thifbiplsemuyanisoenrtugy od(7tiNftump-ooreminsag.k FwMdayalanderoidkn-ealdys, Chsiftploehlnyidsaocsyentur).: 1ekUsmormo.Wa0ocnehrt0. dUswisyhawosehuinrg. manddarmcyhoniergr. ohdfu-preinsag.k odUshfemuo-nrpeinragsky tandamaipmaurolsieajnorces h(poolwespauptmerap.) Htphroaegmvoemstable ACtsetorcawyormnuer dhotenum-preinastug.rke ofCamastuyaosgnjeur adophnusool-iarpunecasgk ustynotihmosueltianm'erogeusly. TiAdmvanetage No.7on ifbiplThsemuyanoiserntugy o(7tdNoiftmpm.u-opreinsagk dayalwFManderoikdn-ealdys. ifChspltoehylnidsaoscyuentr:). 1ekusmor.Wa0ocnerht0 dmuauoenrscdie'nhtrgy po7(tneNp-roi.meodan.sk FMroidn-ay). wdUsisyahwsoaheuinrg. manddamrcyhoniergre. ohdfu-preinsa.gk aemaoUsplthciatejrnoicers hspoolandweuaspactemhr.p ohdfesu-aprtiensarg.k tpHhroeag100°/o UNSE Territory: 40% r Return Ccst UllS Baseline PV Production PPA , TPO Project Return APS Territory: 60% UNSE Territory: 40% Return Coat UllS Baseline PU Production PPA Figure 6. Impact of Locational Factors on Solar TPO Project Return, 201525 Figure 7 plots the observed solar TPO lease rates in each of six jurisdictions in AZ and CA (represented by the green dots) on the same graph as what lease rates would be if instead solar TPO providers achieved a benchmark 40 percent project return in those jurisdictions, accounting for locational differences (represented by the red dots). These red and green dots on Figure 7 correspond with the red and green dotted lines in Figure 6, respectively. The positive difference between the observed solar TPO lease rates and the TPO lease rates at 40 percent project return, shown as the green shaded area in Figure 7, represents an opportunity for solar TPO providers to achieve “additional return" in those service territories. As is evident in Figure 7, solar TPO project returns increase with increasing utility rates, which cannot be accounted for by variations in locational factors. In other words, calculated project returns vary by utility and are positively correlated with the utility rates. 25 Prior to retroactive bonus depreciation. Attachment CJW-ZSR 14 of 18 3- ,3: 2015 No Bonus Depreciation Bill Savings . . Additional . . . . . . . . . Return . . . .. ~: ,- F, 40% Project Return l 2016 Bonus Depreciation included o . . 3m . Additional Return . . . . . .... 40% Project Return 0 U Figure 7. Project Value Analysis across six utility service territories in AZ and CA Navigant conducted this analysis for lease rates in 2015 and 2016. We found that in four out of the six utility service territories analyzed, SolarCity, for example, increased their lease rates in 2016. This occurred despite declining system costs and favorable policy re-introducing the 50 percent bonus depreciation allowance. The chart above clearly illustrates that solar TPO providers have headroom in many jurisdictions, including UNSE’s service territory, to reduce solar TPO rates while still achieving project returns at or above those achieved in UNSE’s service territory in 2015 (when lease rates were lower, and when bonus depreciation had not yet been re-introduced, as is discussed in further detail in the next section). 2.4.5 impact of Policy Figure 8 shows how the ITC and bonus depreciation policy impact project returns in UNSE and APS territories for various solar TPO lease prices. Throughout 2015 bonus depreciation did not exist for solar systems. However, in December 2015, bonus depreciation was reintroduced and retroactively applies to all 2015 projects. 25 In Figure 8, the red line reflects policy in place during 2015, which has been replaced by current policy (blue line) as of December 2015 and applies retroactively to 2015 projects. Following the favorable bonus depreciation change, solar TPO project returns increased significantly. For example, if lease rates were held constant at $0.087/kWh, project return in UNSE service territory for systems installed in 2015 would have retroactively increased from 40 percent to 60 percent. Similarly, solar TPO providers in APS’s service territory experienced project return increases from 60 to 110 percent for systems installed in 2015 due solely to the re—introduction of bonus depreciation. Simultaneously, UNSE customers have seen increases in lease rates from 2015 to 2016. These lease rate increases are consistent with multiple residential solar players announcing plans to raise lease prices at the end of 2015.31 As shown in Figure 8, UNSE customers have seen a 9 percent increase in solar TPO lease rates, representing a further project return increase from 60 percent in 2015 to 80 percent in 2016. UBS, Global Research — “SolarCity Corp, Getting a Bigger Policy Boost", 16 December, 2015 Attachment CJW-ZSR 15 of 18 ln contrast, the purple line reflects previously anticipated 2017 policy 10 percent ITC and no bonus depreciation. Before these recent policy changes, solar companies would have had to compete along the purple line as of Jan 15‘, 2017, yet now they are operating along the blue line. UNSE ~ Policy impact on Project Return Policy in Place During 2015 Current Retroactive 2015 Previous Anticipated Tax Policy l7 LT} J"1.JS l l ) a!' n Lll i) APS Policy impact on Project Return Current Policy Retroactive 2015 30080 ‘TC Ho 50090 tn l l Bonus Depreciation Policy Place D unn, in2015 9 ,_ Anticipated Tax Policy '23 3: —33 Figure 8. Incentive Impact on Project Return, APS and UNSE Service Territories The analysis above suggests that the combined impacts of the re-introduction of bonus depreciation and the increase of lease rates from 2015 to 2016 offer headroom for solar TPO providers to reduce lease rates and adjust to changing rate structures while still enjoying the same project returns achieved in 2015. For instance, in 2015 in UNSE‘s territory, SolarCity, the leading solar TPO provider, could earn a project return of 40 percent with solar TPO prices set at $0.087/kWh. With the re-introduction on bonus depreciation, this should permit SolarCity, the leading solar TPO provider in UNSE service territory, to earn 40 percent return with lease rates of about $0.075/kWh, which differs substantially from current observed lease rates of $0.095/kWh. The headroom available in other service territories appears to be even greater, based on our analysis indicating that service territories with higher offset rates tend to have larger project returns. The above analysis is presented in a slightly different format below in Table 3. Attachment CJW-ZSR 16 of 18 Table 3. Policy Impact of Project Returns, 2015 and 201627 2015 policy in place through Dec 2015 . 2015 Solar Lease Rate ($lkWh) 0.087 40% 60% 0.105 2015 retroactive change to bonus depreciation in place after Dec 2015 . 2015 Solar Lease Rate ($lkWh) 0.087 60% 110% 0.105 2016 policy in place after Dec 2015 . 2016 Solar Lease Rate ($lkWh) 0.095 80% 110% 0.105 Navigant notes that project return calculations can be sensitive to certain input assumptions. Since project returns grow exponentially as lease rates increase (see Figure 8), this sensitivity is most notable when lease rates and corresponding project returns are high. The robustness of this analysis is in its comparative nature, such that minor uncertainties in inputs are applied equally across all jurisdictions, and across comparative policy and lease price changes. As a result, the conclusions of this analysis are driven primarily by the relative values of the calculated project returns across service territories and over time. Furthermore, Navigant makes no assertions regarding whether any individual project return is deemed to be acceptable, too high, or too low. Although these calculated project returns are high, we note that we have made several conservative assumptions in our analysis that would actually tend to understate, rather than overstate, true project returns. These conservative assumptions include: Cost of debt: Our analysis used a cost of debt of 6 percent throughout the analysis. Some sources indicate that this cost of debt could be as low as 5 percent.28 Lease term and residual value: The analysis uses a 20 year contract term with no residual value for contract renewal and no residual value for the system at the end of life. The typical system life is longer than 20 years and the system is expected to have a residual value at the end of the lease term. o Markup assumed for the ITC and depreciation basis: We used a 35 percent markup on system cost to calculate the value of the system for the purpose of lTC and system depreciation benefits. This value is also known as the fair market value (FMV). Using FMV as the basis for tax credits and depreciation benefits would effectively result in a solar TPO developer reporting a system value of $3.74-3.87/W-DC to the Internal Revenue Service, which is still lower than observed The ability of PV providers to system sales prices that typically range from 27 Project returns are in■uenced by several key factors including: installed system cost, lTC, bonus depreciation, accelerated depreciation. UBS Solar, US Alternative Energy YieldCos, 4Q15 Playbook: Giving Solar ‘Credit,’ January 2014. 29 Deutsche Bank Market Research, SolarCity, Analyst Day Recap, December 15, 2015. 30 “A Survey of State and Local PV Program Response to Financial Innovation and Disparate Federal Tax Treatment in the Residential PV Sector", Lawrence Berkeley National Laboratory, June 2015 31 SolarCity 2015 Analyst Day, December 15 2015. 1 Attachment CJW-ZSR 17 of 18 markup cost to something more akin to a price, or system value, when calculating tax credits and depreciation is a key driver in the favorable economics for solar TPO providers. :a_ Key findings include the following: Navigant’s research indicates that solar TPO providers choose to operate in jurisdictions where they can maximize their return by undercutting utility offset rates.34 Solar TPO providers appearto be tracking utility rates and pricing accordingly, evidenced by higher observed lease prices in jurisdictions with higher utility rates. These higher lease prices cannot be fully accounted for by variations in system cost, solar production, and tax rate (locational factors). Navigant’s analysis found that solar TPO providers’ project returns vary by utility service territory, with higher project returns calculated in service territories having higher utility offset rates. Federal incentives such as the Investment Tax Credit (ITC), accelerated depreciation, and bonus depreciation have a significant impact on project return. The solar TPO business model is able to maximize the benefits of these federal incentives, which are amplified considerably by the TPO’s ability to use a system “value”, which is higher than the system cost, as the basis for the tax credit and asset depreciation. Navigant’s research found that despite continuing declines in solar system costs and favorable policy decisions (e.g., re-introduction of bonus depreciation), lease rates have recently increased in certain locations, consistent with public disclosures from leading solar players and indicating higher project returns for solar TPO providers. In 2015, UNS Electric, lnc. (UNSE) solar TPO providers experienced an estimated 40 percent project return, which is expected to increase to around 80 percent in 2016, due to the lease rate increase from $0.087/kWh to $0.095/kWh between 2015 and 2016 and the re-introduction of the 50 percent bonus depreciation allowance (see Figure 8 on 13). We conclude that solar TPO providers have headroom to adjust to some changes in rate structures while maintaining project returns. 32 “Evaluating Cost Basis for Solar Photovoltaic Properties”, US. Treasury Department. v of Solar Generating Assets", Solar Energy Industries Association, value of a customer’s bill reduction for each kWh generated by the ($lkWh) customer’s solar system. in other words, it is the amount of their bill that is “offset” for each kWh generated (hence the term). Attachment CJW-ZSR 18 of18 APPENDIX A. Financial Assumptions System Spec'f'cat'ons Asset life/investment horizon (Years) Installed cost ($NV-DC) Total asset size (kW) Annual capacity factor (%) Annual degradation (%lyear) Fixed O&M ($lkW-year)* Fixed 0&M escalator Financing Cost of equity Cost of debt Percentage of cap structure — equity Percentage of cap structure — debt Debt amortization period (Years) Residual Value Target Debt Service Coverage Ratio Taxes and Federal income tax lncentwes State income tax Investment Tax Credit De reciation t e p yp Discounting convention System Cost Markup for Tax and Depreciation State incentives Local incentives Other Lease rate Lease escalation rate *O&M costs include all 0&M components as well as inverter replacement. 20 Varies by location 7.00 Varies by location 0.50%lyear 20.00 1.90% Model output 6.00% Model output Model output 20 $0.00 1.30 35.00% CA: 8.84%; AZ: 6.00% 30.00% MACRS, Bonus where applicable Mid-year37 35.00% None None (SMUD: $500/system) Varies by location 2.90% 35 National Renewable Energy Laboratory, US. Residential Photovoltaic (PV) System Prices, Q4 2013 Benchmarks: Cash Purchase, Fair Market Value, and Prepaid Lease Transaction Prices, Oct. 2014. 35 National Renewable Energy Laboratory, Distrib uted Generation Renewable Energy Estimate of Costs, ~ w,Accessed February1,2016. 37 A mid-year discounting convention is a standard assumption about when cash flows occur throughout the year for the purposes of a discounted cash flow analysis. The problem with an end-of—year discounting convention is that it discounts the future value too much. It assumes that the entire cash ■ow for a given year comes at the very end of that year, and therefore should be discounted accordingly. This is often inaccurate, since cash flows typically occur in each month of the year. The mid-year discounting convention better represents the time-value of these monthly cash ■ows than an end-of-year convention. The mid—year convention assumes that all the cash comes in halfway through the year, which averages out the time differences between the individual monthly cash ■ows.