UNITIL ENERGY SYSTEMS, INC. DIRECT TESTIMONY OF PAUL M. NORMA1~D MARGINAL COST STUDY AND RATE DESIGN

UNITIL ENERGY SYSTEMS, INC. DIRECT TESTIMONY OF PAUL M. NORMA1~D MARGINAL COST STUDY AND RATE DESIGN New Hampshire Public Utilities Commission Docket No. DE -0 031

TABLE OF CONTENTS 1. INTllODUCTION 1 II. MARGINAL COST STUDY OVERVIEW OF MARGINAL COST STUDy Ill. RATE DESIGN R.c\TE DESIGN INFORMATION CLASS REVENUE TARGETS INDIVIDUAI. RATE DESIGNS IV. PROPOSED TARIFF CHANGES LIST OF SCHEDULES PMN-l PMN~ PMN 3 w PMN- Qualifications of Paul M. Normand Marginal Cost Study Rate Design PMN-3-1 Development of Class Revenue Targets PMN-3- PMN-3-3 PMN-3- Design ofrates Revenue Proof Typical Bill Impact Analyses Discussion of Costing Methodology 03

NHPUC Docket No. DE -0 Testimony ofpaul M. Nomlalld Page 1 of I. INTRODUCTION Q. Please state your name, position and business address. 3 A. My name is Paul M. Normand. I am a principal with the firm of Management Applications Consulting, Inc. ("MAC"). This company provides consulting services to the utility industry in such field as loss studies, econometric studies, cost analyses, rate design, expert testimony, and regulatory assistance. MAC's headquarters is located at Rocky Drive, Suite 1, Reading. Pennsylvania 0. Q. A. What is your educational background? 1graduated from Northeastern University in 1, with a BachelorofScience Degree and a Master of Science Degree in Electrical Engineering-Power System Analysis. I have I I attended various conferences and meeting concerning engineering and cost analysis. 1 1 1 Q. A. What is your professional background? I was employed by the Massachusetts Electric Company in the Distribution Engineering Department while attending Northeastern University. My principal areas ofassignment included ne\v service, voltage conversions, and system planning. Upon graduation from Northeastern University, Ijoined Westinghouse Electric Corporation Nuclear Division in Pittsburgh, Pennsylvania. In that position, I assisted in the procurement and economic analysis of electrical/electronic control equipment for the nuclear reactor system. In 1, ljoined Gilbert Associates as an Engineer providing consulting services in the rate and regulatory area to utility companies. I was promoted to Senior Engineer in 1, Manager of the Austin office 10, and Director of Rate Regulatory Service in 11. 033

NHPUC Docket No. DE -0 Testimony of Paul M. Normand Page of 1 In June, ]3, J left Gilbert to fonn a separate consulting finn and I am now a principal and President of Management Applications Consulting, Inc. My principal areas of 3 concentration have been in loss studies, economic analyses. and pricing. Q. A. I 1 Have you testified in support of any cost studies that you participated in or performed? Yes. I have testified about such studies before the following regulatory agencies: the Maine Public Utility Commission. the Public Utility Commission of Texas, Illinois Commerce Commission, Ncw Hampshire Public Utilities Commission, New Jersey Board of Public Utilities. New York Public Service Commission, Pennsylvania Public Utility Commission, the Massachusetts Department of Public Utilities, the Kentucky Public Service Commission, the Arkansas Public Service Commission, the Public Service Commission of Louisiana, the Public Utilities Commission of Ohio, the Public Service Commission ofmissouri, the Delaware Public Service Commission, the Maryland Public Service Commission, the Indiana Utility Regulatory Commission, the North Carolina Utilities Commission, the Kansas Corporation Commission, and the Federal Energy Regulatory Commission. 1 Q. 1 A. 1 Could you please briefly discuss your technic experience? I have perfonned numerous accounting and marginal cost of service studies, time differentiated bundled and fully unbundled cost studies for both electric and gas utilities since 10. I have also used such studies in the design and presentation of detailed rate proposals before regulatory agencies. My additional experience has been in the area ofunaccounted [or loss evaluations for electric and gas utilities 03

NHPUC Docket No. DE -0 Testimony of Paul M. NOlmand Page 3 of 1 ".J for over twenty-four years. These studies include a detailed review of each system and the calculation of appropriate recovery factors. Additional information about my background and experience is contained in Schedule PMN-I. Q. A. 11 Please summarize your testimony. My testimony \vill discuss the marginal cost of service study that \vas performed by MAC that is used as the basis for the rates proposed in the filing by Unitil Energy Systems, Inc. ("Unitil Energy" or "Company") in this proceeding. The marginal cost study provides the basis for determining the level of revenues to be recovered from each class of service as well as component costs that are used to design rates. I am also responsible for the development ofclass revenue requirements using the results of the marginal cost study. The marginal cost study identifies the changes in costs associated with changes in the number of customers, level of sales, and capacity requirements placed on Unitil Energy's distribution system. Finally, I am responsible [or detennining the levels of revenue to be recovered from each rate class and the design of distribution rates for each customer class. 1 1 Q. 1 A. Please outline the organization of your testimony and schedules. My testimony consists ofthree sections beyond this one. Each section discusses the cost and rate study and analyses conducted as a part oftills filing and is supported by schedules displaying a summary ofthe study results as well as key data employed in 03

NHPUC Docket No. DE lo-oss Testimony of Paul M. Normand Page of the calculations. These schedules are numbered PMN- though PMN-. Schedule PMN-l details my qualifications and experience.... -> The second section of my testimony summarizes the calculations incorporated into the Company's marginal cost of service study and presents its results as Schedule PMN-. The marginal cost to serve each class pwvides the basic measure of each class's proportionate share ofthe total costs to serve. Furthermore, the marginal cost study provides a guide to the development of the facilities charges included in the proposed rates. 11 The third section ofmy testimony relates to rate design. In this section, I employ the marginal cost study's results, after making an equi-proportional adjustment to class marginal costs to reconcile with Uniti1 Energy's distribution service revenue requirements. The proposed revenue targets and detailed rate design calculations are also included in Schedule PMl\'-3 ofmy direct testimony. 1 1 1 finally, in Schedule PMN-, I present a detailed explanation of some of the more technical aspects of the marginal cost study and the assumptions underlying my results. 03

NHPUC Docket No. DE -0 Testimony of Paul M. Normand Page ofl 1 II. 3 Q. A. S 1I MARGINAL COST STUDY Overview of Marginal Cost Studv Please summarize the objectives of a marginal cost study. The marginal cost study is provided in Schedule PMN-. A marginal cost study provides an estimate ofthe additional cost ofproviding an additional unit ofservice. These cost estimates are utilized as a benchmark or reference in initially setting rates to the extent allowed by further considerations for rate continuity, intra-class equity, and customer impact. The use of marginal costs in ratemaking tends to result in a level and pattern of prices that promotes economically rational consumption decisions, and thereby promotes an efficient allocation of society's resources. Sending customers accurate price signals regarding the costs that will result from their consumption decisions furthers efficiency. Customers, in tum, will be able to make informed decisions on their use of utility service. Q. A. 1 1 1 How is a marginal cost study used in the rate design process? Follo\ving the precedent established by the New Hampshire Public Utilities Commission ("'the Commission") in DG 00-03, the marginal cost study is used to establish revenue levels and prices for each rate class on the basis ofmarginal costs, adjusted using the Equi-Proportional Method ("EPM") to recover the allowed revenue requirements. The proposed total system distribution service revenue requirements are established at the adjusted test year levels. Distribution service marginal costs by class (which differ from the revenue requirement) are then 03

NHPUC Docket No. DE -0 Testimony of Paul M. Nonnand Exhihit PMN- Page of adjusted to match the distribution system total revenue requirements on a pro-rata.., j basis using the EPM. These resulting scaled marginal costs by class and cost component become the theoretical targets for the design ofdistribution service rates. Q. A. I I Could you provide an overview ofthe methodology you employed in developing your marginal cost study? I have computed the marginal distribution costs to serve each ofunitil Energy's rate classes based on test year costs. I have used regression and engineering techniques to estimate the hypothetical distribution costs ofserving an increment ofcustomer load, including the unit costs ofadding distribution plant facilities as well as the additional costs for O&M. These distribution capacity costs were measured in terms ofdollars per peak day kw. I have also used engineering estimates to identify the investment in services and meters and added corresponding O&M expenses necessary to serve a new customer. From these factors, I have developed the annual revenue requirements to serve each of the Company's rate classes. These costs are stated in tenns of customer and demand charges. A discussion of the methods I employed in the 1 marginal cost study is more fully described in Schedule PMN-. 1 1 TIle marginal study excludes all production and transmission costs as they are irrelevant to the design of distribution rates based on the Conunission's current requirements. 03

NHPUC Docket No. DE -0 Testimony of Paul M. Normand ExJlibit PMN- Page of Q. A. What were the results of the marginal cost study? Schedule PMN-, Table, tabulates the long-run marginal costs to serve each 3 customer class. In addition, this table calculates the revenues that would be generated if the Company were to introduce full marginal cost-based pricing and if customers were to continue to consume as they have in the past (Table, line ). Schedule PMN-, Table, provides a summary of marginal costs on a unit cost basis. Finally, Schedule PMN-. Table, presents the EPM adjustment to restate the unit marginal costs to a level that will recover the distribution service revenue requirements target (line ). t 1 III. RATE DESIGN Rate Design Information Q. A. Please describe your cost summary. Schedule PMN-, Table, presents the infonnation from the marginal cost study and current revenue. The schedule derives the theoretical revenue target for each 1 1 1 class and compares it with the proposed revenue requirements. Each component of each c1ass's rates is scaled upward or dovvtiward so that the total revenues derived from all classes at the calculated rates will produce the overall system revenue requirement proposed by Uniti! Energy. These costs are then employed as the basis for designing rates subject to certain limitations ofcustomer bill impacts and revenue.. mcrease maximum or caps. 03

NHPUC Docket No. DE -0 Testimony of Paul M. Nonnand Exhibit PM~- Page of 1 Q. 3 A. Class Revenue Targets Please describe how you established class revenue targets. My revenue target calculations are shown on Schedule PMN-3. This schedule, consisting offour sub-schedules was used both to establish class revenue targets and to design rates. Bill impact considerations for customers limited the maximum base rate increases so that no rate class received in excess of % ofthe system average increase. In other words, since the Company is seeking an overall distribution 1I system increase of.% (see Schedule PMN-3-1. page 1 of. line ), the m.aximum amount any class's rates could be increased is approximately 3% (i.e., % x.%). The differences between the assigned revenue requirements and the maximum level of revenues allowed under the restriction described above were summed and then allocated on a pro-rata revenue basis to the classes whose rate increases were not affected by the limitation on the level of increase.' If that 1 1 1 redistribution of revenue resulted in any class exceeding its maximum allowed increase, the unrecovered revenue requirement for such classes would again be allocated to the remaining classes unaffected by the rate caps. Fortunately, only one class. Rate 0, was affected by this constraint in the initial allocation. The subsidy (Rate 0 revenue shortfall) produced by establishing these revenue targets was then allocated equally to the remaining uncapped C&I rate classes. These calculations are shown on Schedule PMN-3-1, page 1 of, line. \ The revenue target for outdoor lighting was established by increasing the existing revenue levels by the system average increase. Additional revenue was not allocated to this class. 030

NHPUC Docket No, DE -0 Testimony of Paul M, Nonnand Page of Q..., j A. Why did you select a revenue cap of %? I have selected this rate cap based on several considerations. First and foremost, I considered bill impacts. Since this rate cap is being applied to the design of distribution rates and since distribution rates are normally much less than half the customer's total energy bill, overall bill impacts will be much lower than the maximum distribution rate increase. Second, I examined the relative difference 1I between costs to serve and current revenue levels. At the present time, the residential rate class provides approximately one-half ofthe Company's distribution revenues but accounts for approximately two-thirds ofits cost to serve. My experience in the industry suggests that the % cap is not unusual and has been found to be reasonable by other regulatory authorities and in the Company's last Settlement Agreement approved by the Commission in Order No.,. Finally, a continued gradual movement towards costs is essential to reduce existing interclass subsidies and improve efficient pricing for all customer classes. 031

NHPUC Docket No. DE lo-oss Testimony ofpaul M. Nonnand Page of Individual Rate Designs Q. "' A. -' How did you approach the design of individual rates? Once the revenue targets were established for each rate class, detailed rate design was conducted, as shown on Schedule PMN-3-. The rate design process was guided by two general principles - moving rates toward the marginal costs to serve and providing some level ofrate stability for customers by controlling bill impacts. 11 L3 Marginal costs to serve include two types ofcost -costs that vary with the number of customers and costs that vary with the peak day demands ofcustomers. In essence, the utility must construct a distribution system capable ofhandling the anticipated loads of customers under varying demand requirements. These capacity costs are incurred regardless ofthe volumes consumed by customers throughout the test year. Therefore, it is more appropriate to recover these costs through a fixed or facilities charge, rather than a volumetric or kwh charge. The very bottom oftable ofthe marginal cost study, Schedule PMN-, shows the total marginal costs for each class expressed in terms of dollars per month per customer. 1 1 Q. 1 A. How did you determine your proposed customer charges? Using the rate revenue targets, the rate design process simply became a matter of raising customer charges to the limits imposed by rate stability and bill impact considerations. For rates D, G, G-kWh meter, and G-Quick Recovery Water Heating, the proposedcustomercharges were increased by approximately 0% from 03

NHPUC Docket No. DE -0 Testimony of Paul M. Nonnand Page 11 of their existing levels which is greater than the overall average system increase cap of 3 3% but moves more appropriately towards the calculated marginal coste;. The G I customer charge (secondary) was decreased by approximately % while the G I customer charge (primary) was increased by approximately %. Both were set based on the adjusted marginal unit cost. Q. A. 11 How did you design the kwh and kwikva rates for each rate class? Having already established the marginal revenue targets for each class, I subtracted the revenues derived from the proposed customer charge increases to compute the remaining revenues to be recovered through either the demand or kwh charges as appropriate for each rate. I simply made a pro rata adjustment to the existing charges to achieve the desired revenues from each class. For the residential Rate D, I maintained the existing 0 kwh blocking structure along with the $.00 per kwh pricing differential in order to mitigate the proposed impact ofthe increase somewhat for lower usage customers. 1 Q. 1 A. 1 How did )'OU design the Outdoor Lighting Rates? The proposed rate design consisted of adjusting all fixed rate components for each fixture charge by an equal percentage based on the class targeted revenue requirement percent increase level. 033

NHPUC Docket No. DE -0 Testimony of Paul M. Nonnand Page of Q. A. Have you provided a summary of the rates you designed? Yes. a table summarizing the rates is provided on page 1 of Schedule PMN-3-. 3 Q. A. Did you verify that the proposed rates match the distribution revenue requirements? Yes, the proposed rates were multiplied by the appropriate test year billing determinants to provide a revenue proofas shown on Schedule PMN-3-3. Q. 11 A. Have you analyzed the impact of the proposed rates on the customers within each rate class? Yes, I have prepared typical bill impact analyses for each rate class. showing the impact on the customer as a function of monthly use. I've prepared these 1 comparisons on Schedule PMN-3-. The monthly usage ranges and percentage of bills are displayed vertically on these pages along the left side and the bilt calculations with revenue differences shown in the right columns. The charges included in the present and proposed bill impacts are sho\\-ti in the lower part ofeach rate. 1 1 Q. For purposes of calculating the bill impacts resulting from the new rates proposed by the Company, what assumptions did you make regarding the supply cost portion of customers' bills? 03

NHPUC Docket No. DE -0 Testimony of Pau1M. Normand Page 0f A. For comparative purposes. the Company used the supply cost based on its most 3 recent approved costs for May as shown on the bottom portion of each biu companson, IV. Q. PROPOSED TARIFF CHANGES Have you prepared revised tariffs? A. Yes, I have. A red-lined version of the existing tari ffs that have been changed to R reflect the revised customer, kwh, and kw charges as appropriate for each tariffare provided in Volume I ofthis filing under the tab labeled 'Tariffs". A clean version ofthe proposed tariffs is included under this tab as well. 11 Q. A. Is the Company proposing any changes the terms and conditions of its tariffs, other than the revisions noted above? No. The Company is proposing no other changes to its existing approved tcnns and conditions. 1 Q. 1 1 Has the Company modified its existing outdoor lighting tariffto reflect NlI RSA -E:3 which establishes the "New Hampshire Dark Sky Policy" to encourage outdoor lighting efficiency at the municipal level, and NH RSA -E: which requires the establishment of electric utility rates for partial night use of outdoor lighting systems? 03

NHPUC Docket No. DE -0 Testimony of Paul M. Normand Page of A. Yes, the Company has made several changes to its existing outdoor lighting tariffto....j incorporate the requirements of NH RSA Chapter -E. The changes are similar to those proposed by PSNH in DE 0-03 for its midnight outdoor lighting service option. Q. A. 11 Please describe the midnight service option. Under the proposed midnight service option. a customer can receive partial night lighting service. In order to receive service under the midnight service option, the existing all-night photocell which turns the luminaire on approximately one-half hour after sunset to one-halfhour before sunrise ",viii be removed and replaced with a photocell capable ohurning the luminaire on hom approximately one-half hour after sunset to midnight. Q. A. 1 1 1 Please describe the rates under the midnight service option. The distribution rates under the midnight service option are the same as ratcs under the all-night service option because the fixture (excluding the photocell), lamp and maintenance costs do not change under the midnight option. The distribution rates are fixed monthly charges for each luminaire. The monthly kwh per luminaire has been lowered, however, ret1ecting the number of hours from one-half hour after sunset to midnight. The monthly kwh is applied to all energy based charges (i.e. External De1iver,Y Charge, Stranded Cost Charge, System Benefits Charge, Default Service, and Electricity Consumption Tax). 03

NHPUC Docket No. DE J0-0 Testimony ofpaul M. Normand Page of Q. -..) Are customers required to pay any costs up-front before they can receive service under the midnight service option? A. Yes. Since the additional equipment and installation cost associated with the new photocell are not reflected in the distribution charges under the midnight service option, customers requesting midnight service are required to pay for these costs prior to the installation of the new photocell. the tariti. These requirements arc provided in Q. 11 A. 1 1 1 Have you proposed any other changes to the existing lighting tariff? Yes, Unitil Energy currently offers mercury vapor lamps as a "white" light source for street and customer lighting. As ofjanuary L 0, in accordance with the Energy Policy Act of 0. mercury vapor lamp ballasts are no longer manufactured or imported for sale in the United States. With this filing, Unitil Energy has proposed new tariff rates for the introduction of metal halide CMH"') lighting in order to provide its customers an alternative lighting source that resembles mercury vapor. This proposal provides another option for existing customers when their mercury vapor ballast fails, for customers requesting conversions, and for new customers or new lighting installations. Q. Please describe how you have designed rates for this option? 03

NHPUC Docket No. DE -0S Testimony of Paul M. Nonnand Page of A. The detail costing for the MH lights have been provided in the v,'orkpapers of this..,.) filing. The underlying costs arc based on the estimated plant costs, associated loading costs related to the material and installation costs, and an appropriate levc1ized fixed charge rate for these fixtures. Q. A. In your opinion, are the costs and rates employed in your analyses just and reasonable? Yes, the costs and the proposed idtes are just and reasonable. Further, the rates are not unduly discriminatory. Finally, the rates represent a careful balancing ofthe costs of providing electric distribution service with customer impact concerns. 11 Q. A. Does this conclude your testimony? Yes, it does. 03