The URL can be used to link to this page

Home My WebLink About2003-10-28-UTIL-min.pdf Electricity Distribution Costs Comparisons of Urban and Suburban Areas Prepared by Paul Chernick and Patrick Mehr, Lexington Electric Utility Ad-hoc Committee (10/28/03) 1 Objective A typical investor-owned utility ("IOU") serves a large territory that includes urban areas and surrounding suburban (or even rural) areas. Each IOU charges the same rates to customers throughout its service territory, regardless of location. In discussions about the formation of new municipal electric utilities ("munis"), speculations have been advanced about the effect on cities if suburban towns municipalize and cease to be served by the IOU 1 If distribution costs are lower in suburban areas, and those communities depart, the IOU's average costs, and hence its rates, will have to increase.2 If this "suburban flight" scenario were likely, and if distribution costs were lower in suburban areas, it would legitimately be a matter of considerable concern for city governments. Whether municipalities will choose to form their own utilities will depend on many factors, including the municipality's level of frustration with its IOU and the extent to which other issues preoccupy the municipal government. The critical factor in the suburban-flight scenario is the assumption that most suburban towns are less expensive to serve than most cities. Electricity distribution costs are driven by the type of infrastructure necessary to serve the local customer base. Certain features of the distribution infrastructure may, all other things equal, drive electricity distribution costs up or down. Specifically, compared to the suburbs, urban areas tend to have the following characteristics: 1 Forty-one munis already exist in Massachusetts. 2If municipalization simply allowed the communities with inherently lower costs to avoid the averaging of their costs with those of inherently more expensive communities, it might have little real benefit to the Commonwealth as a whole. A similar process, in which a competitor picks off the low- cost customers of a utility that is required to charge equal prices to all customers, is called cherry- picking. This is not a close parallel to municipalization, in which the "cherries" would be picking themselves, but the same terminology has been used by opponents of municipalization. 1 A larger portion of the distribution infrastructure is underground, in vaults and conduit under streets and sidewalks, rather than overhead, on poles along streets. 2. Construction and maintenance of the infrastructure is more difficult and expensive due to high building density, lack of room for setting up construction equipment, more congested streets, and higher real-estate costs for substations and other facilities. 3 The load density is higher, that is, each mile of distribution circuit serves more customers and delivers greater amounts of electricity In urban areas, item (1) increases distribution investment (but reduces maintenance cost, since underground systems do not require tree-trimming and are less vulnerable to weather and accidents), item (2) tends to increase distribution costs, while item (3) tends to decrease costs, all compared to suburban areas. As the various characteristics of the distribution infrastructure drive costs in different directions, it is not clear whether electricity distribution costs would tend to be higher in urban or in suburban areas. In principle, this question could be answered for the greater Boston area by computing NSTAR's cost of distribution service separately for urban and suburban areas. Such an analysis would require that NSTAR track its investments and operating costs for each community it serves (or at least for groups of similar communities). So far as we know, NSTAR has not tracked that information. Absent the necessary cost information specific to the urban and suburban communities NSTAR serves, this paper presents five alternative approaches to determining whether electricity distribution costs are generally higher in urban than in suburban areas. 2. Comparisons Data and Analysis 2.1 Consolidated Edison's Analysis of Urban and Suburban Costs We are aware of only one published direct comparison of distribution costs in urban and suburban areas. In 1982, the staff of the New York Public Service Commission looked at the differences in costs between the two areas served by Consolidated Edison: New York City and Westchester The study found that "A significant reason for the differences" in the cost of serving these areas was "the cost of the distribution system per unit of electricity sold. Although the underground system in New York City is expensive, the number of kWhs delivered per dollar of investment is very high. This means that expenses [are low] relative to Electricity Distribution Costs:Comparisons of Urban and Suburban Areas Lexington Electric Utility Ad-hoc Committee 10/28/03 Page 2 of 12 revenues. Westchester, which is more suburban and rural in nature, has a low population density This increases the unit costs. "3 In the study, Westchester was • 11.8% of Consolidated Edison's sales, but • 12.6% of the peak load, • 14.3% of distribution plant investment, • 14.9% of services and meters investment, and • 14.5% of distribution expenses. This study suggests that the cost of distribution for New York City, the most urban of urban areas, is lower than the cost of providing service to its suburbs. 2.2. Comparison of Rates for NsTAR's Subsidiaries NSTAR's electric service territory consists of three subsidiaries: • Cambridge Electric Lighting (CELCo): serves only Cambridge; totally urban, with a very high population density (about 30% higher than Boston's). • Boston Edison (BECo): serves the city of Boston, several smaller cities (e.g., Chelsea, Waltham) and the western suburbs, a mix of urban and suburban areas. • Commonwealth Electric (CommElec): serves part of Southeastern Massachusetts and Cape Cod, mostly suburban (or even rural) areas, with a small urban area (New Bedford). We compared NSTAR's residential rates for each of the three subsidiaries (from www.nstaronline.com), for customers with annual usage ranging from the average usage in Cambridge (400 kWh/month) to the average usage in the BECo and CommElec territories (about 600 kWh/month).4 Exhibit 1 shows the results of our comparison of both the "delivery" 3 Quotes are from "A Study of the Comparative Costs of Electric Service in Westchester County and the City of New York," Berak, F., Arnett, H. Nadel, J, New York Department of Public Service Power Division, Case 28157 January 29, 1982, pp. 4-5. 4 Commercial and industrial loads and customer mix vary more widely between utilities, and the rates for those customers are more complex, making comparisons of rates more difficult and less meaningful. Electricity Distribution Costs:Comparisons of Urban and Suburban Areas Lexington Electric Utility Ad-hoc Committee 10/28/03 Page 3 of '? rates—all charges except supplier services (NsTAR's costs of purchasing power for its customers)—and distribution charges (just the customer and distribution charges). Distribution costs for small, medium and large residential customers are the highest in NSTAR's least-urban area (Commonwealth Electric) and the lowest in NsTAR's most-urban area (Cambridge). Boston Edison distribution costs are in between, consistent with that service territory being a mix of urban and suburban areas. Exhibit 1 Comparison of NSTAR Company Residential Rates Cambridge Boston Commonwealth Electric Light Edison Electric Service territory Cambridge Boston, New Bedford, only suburbs Plymouth, Cape Cod Type of area urban mix urban, mostly suburban to suburban rural Customers per mile of line 46 30 18 % distribution underground 71% 34% 20% Rate (residential)* 01 Al 32—Annual Delivery service charges* Customer (per month) $6.87 $6.43 $3.73 Distribution (per kWh) $0.02434 $0.03900 $0.04524 Transition (per kWh) $0.00308 $0.01813 $0.02717 Transmission (per kWh) $0.02366 $0.00733 $0.00534 Energy Conservation (per kWh) $0.00250 $0.00250 $0.00250 Renewable Energy(per kWh) $0.00050 $0.00050 $0.00050 Delivery cost(0/kWh) 400 kWh/month customer 7 1 8.4 9.0 600 kWh/month customer 6.6 7.8 8.7 Delivery cost(customer&distribution char es, it/kWh) 400 kWh/month customer 4.2 5.5 5.5 600 kWh/month customer 3.6 5.0 5.1 Electricity Distribution Costs:Comparisons of Urban and Suburban Areas Lexington Electric Utility Ad-hoc Committee 10/28/03 Page 4 of 12 2.3. PEPCo's Washington DC and Maryland Jurisdictions We sought other examples, comparable to the NSTAR situation, in which an urban utility is surrounded or adjacent to a suburban utility We looked first at the most densely populated cities reported in the 2000 Census. Other than Cambridge, we did not find any cities served by an IOU that did not serve at least the suburbs, and often other cities and rural areas as well. Exhibit 2 summarizes the results for the highest-density cities. Exhibit 2: Highest-Density Cities in US 2000 Census City Area Density Population (mit) (pop/mi ) Utility Serves (New York City, NY 8,008,278 303.3 26404 ConEd NYC &Westchester (Paterson city NJ 149,222 8.4 17765 PSE&Glbig chunk northern NJ (San Francisco, CA 776,7331 46.7 16632 PG&Ellarge &varied area !Jersey City NJ 1 240,055 14.9 16111 PSE&G1big chunk northern NJ Cambridge, MA 101,355 6.4 15837 CELColonly Cambridge Daly City CA 103,621 7.6 13634 PG&Ellarge &varied area Chicago, IL 2,896,016 227 1 12752 CommEdIChicago &suburbs (Santa Ana, CA 337,9771 27 1 12471 SCEIServes cites, suburbs, rural lInglewood, CA 1 112,5801 9.1 12371 SCEIServes cites, suburbs, rural (Boston, MA 1 589,1411 48.4 12172 BEColBoston+ &suburbs lEl Monte, CA 115,965 9.6 12080 SCEIServes cites, suburbs, rural Hialeah, FL 226,419 19.2 11793 FP&Llmost of FL (Newark, NJ 273,546 23.8 11494 PSE&Glbig chunk northern NJ (Philadelphia, PA 1,517,550 135.1 11233 PEColPhiladelphia&suburbs Yonkers, NY 196,086 18.1 10833 ConEd1NYC &Westchester !Norwalk, CA 103,298 9.7 10649 SCEIServes cites, suburbs, rural !Miami, FL 362,4701 35.7 10153 FP&Llmost of FL lElizabeth, NJ 120,5681 12.2 9883 PSE&GIbig chunk northern NJ (Berkeley, CA 102,743 10.5 9785 PG&Ellarge &varied area !Providence, RI 173,618 18.5 9385 Narragansettlmost of RI (Washington, DC 572,059 61 4 9317 PEPCo1DC & MD suburbs None of these cities quite matches our criteria, but Washington comes close. While Potomac Electric Power (PEPCo) serves both Washington and the Maryland suburbs, DC and Maryland separately set rates for their areas, based on (among other things) the distribution investment and expenses in each jurisdiction.5 We obtained PEPCo's basic residential rate schedules and other data by jurisdiction from PEPCo's web site (www.pepco.com). As shown in 5 The Virginia suburbs are served by Virginia Power, which serves most of that state. Electricity Distribution Costs:Comparisons of Urban and Suburban Areas Lexington Electric Utility Ad-hoc Committee 10/28/03 Page 5 of 12 Exhibit 3, the residential distribution rates for the urban part of the service territory are substantially lower than those for the suburban portion. Exhibit 3: PEPCo Urban and Suburban Rates Service area Washington DC Maryland suburbs Sq miles 68 572 Population served 572,000 1 450,000 Density (population/sq mile) 8,412 2,535 Type of area urban suburban Summer Winter Summer Winter Residential rates: (June-Oct) (Nov-May) (June-Oct) (Nov-May) Distribution service charge Minimum charge, including first 30kWh ($/month) $0.47 $0.47 kWh charge, next 370 kWh (0/kWh) 0.945 0.945 kWh charge, in excess of 400 kWh (0/kWh) 2.845 1.942 Customer ($ per month) $5.54 $5.54 kWh charge,first 800 kWh (0/kWh) 3.112 1.953 kWh charge, in excess of 800 kWh (0/kWh) 3.112 1.503 Transmission service charge Minimum charge, including first 30kWh ($/month) $0.15 $0.15 kWh charge, in excess of 30 kWh (0/kWh) 0.521 0.521 kWh charge, first 800 kWh (0/kWh) 0.5141 0.514 kWh charge, in excess of 800 kWh (0/kWh) 0.514 0.514 Residential bill (855 kWh/month average, half in summer)(0/kWh) Distribution only 1 74 3.17 Distribution &transmission I 2.26 3.68 2.4 Seattle and Environs We also looked for suitable pairs involving large, but less dense, cities. Exhibit 4 lists those cities, down to populations of 300,000, or about half that of Boston. Most of these are not useful for our purposes, because the city is served by an IOU that also serves large surrounding suburban and rural areas (e.g., Houston, Phoenix, San Diego and Dallas), or because the city is served by a municipal utility, but the neighboring suburbs are served by an IOU with a mix of urban and suburban, and sometimes rural, loads (e.g., Sacramento, Austin, San Antonio). In southern California, both Los Angeles and several smaller nearby communities are served by separate munis, but the smaller cities are about as densely settled as Los Angeles (with Electricity Distribution Costs:Comparisons of Urban and Suburban Areas Lexington Electric Utility Ad-hoc Committee 10/28/03 Page 6 of 12 comparable customers per mile of line and percentages of underground distribution), and are therefore not really good proxies for suburban towns in Massachusetts.6 Exhibit 4 Largest Cities in 2000 Census Excludes Cities in Exhibit 2. City Population Area(m12) Density(pop/mit) Utility Los Angeles, CA 3,694,820 469.1 7876 Muni Houston, TX 1,953,631 579.4 3372 Phoenix, AZ 1,321,045 474.9 2782 San Diego, CA 1,223,4001 324.3 3772 Dallas, TX 1 188,580 342.5 3470 San Antonio,TX 1 144,646 407.6 2808 Muni Detroit, MI 951,2701 138.8 6854 San Jose, CA 894,943 174.9 5117 Indianapolis, IN 781,870 361.5 2163 Jacksonville, FL 735,6171 757 71 9711Muni Columbus, OH 711 470 210.31 3383ICSP Austin, TX 656,562 251.51 2611 Muni Baltimore, MD 651 1541 80.81 8059IBG&E Memphis,TN 650,100 279.3 2328IMuni Milwaukee, WI 596,974 96.1 62121 El Paso, TX 563,662 249.1 22631E1 Paso Elec (also serves rural load) Seattle,WA 563,374 83.9 6715IMuni Denver CO 554,636 153.4 3616IPSCo Nashville-DavidsonTN 545,524 473.3 11531Muni Charlotte, NC 540,828 242.3 2232IDuke Fort Worth, TX 534,6941 292.5 18281 Portland, OR 529,121 134.3 3940IPacifiCorp & Portland General Electric Oklahoma City OK 506,132 607 8341 Tucson, AZ 486,6991 194 7 2500IAPS New Orleans, LA 484,674 180.6 2684INOPSI Las Vegas, NV 478,434 113.3 42231NV Power Cleveland, OH 478,403 77.6 61651Muni &CEI Long Beach, CA 461,5221 50.4 91571SCE Albuquerque, NM 448,607 180.6 2484IPNM Kansas City MO 441,545 313.5 1408KCP&L Fresno, CA 427,6521 104 41 4096 Virginia Beach, VA 425,257 248.3 1713 6 We actually looked at all US cities with a population of at least 100,000. Electricity Distribution Costs:Comparisons of Urban and Suburban Areas Lexington Electric Utility Ad-hoc Committee 10/28/03 Page 7 of 12 -i i City Population Area(mit) Density(pop/mi2)1Utility Atlanta, GA 416,474 131 7 31621 Sacramento, CA 407,0181 97.2 4187 Muni Oakland, CA 399,484 56.1 7121 PG&E Mesa, AZ 396,375 125 31711 Tulsa, OK 393,0491 182.6 2153 Omaha, NE 390,007 115.7 3371 Omaha PPD Minneapolis, MN 382,618 54.9 69691 Colorado Springs, CO ! 360,890 185.71 1943 Muni St. Louis, MO 348,189 61.9 5625 Wichita, KS 344,284 135.8 25351 Pittsburgh, PA 334,563 55.6 6017 Arlington, TX 332,9691 95.8 3476 !Cincinnati, OH 331,285 78 42471 !Anaheim, CA 328,014 48.91 67081Muni Toledo, OH ! 313,6191 80.61 3891 !Tampa, FL 1 303,4471 112.11 2707 We considered Tennessee and Nebraska, because they are both served predominantly by publicly-owned utilities, which generally have smaller and more homogeneous territories.? Nebraska did not turn out to be useful, because the public power districts serve large areas. For A example, the largest Nebraska city, Omaha, is served by the Omaha Public Power District, which also serves 13 counties (and hence does not seem to be an urban utility). Tennessee provided more useful information, which we discuss in the next section. Otherwise, the only example we found in the data was that of Seattle, which is served by a muni and has a total population and population density similar to Boston's, while the neighboring Snohomish County (a largely suburban area) is served by a public utility district, which is essentially a special-purpose muni. As summarized in Exhibit 5, we computed the supply costs (purchased and generated power) for each of the utilities, and subtracted those supply costs from the residential and overall rates charged respectively by Seattle and Snohomish County to approximate their respective distribution costs (see Exhibit 3).8 We found that distribution rates, as approximated in this analysis, were higher in suburban Snohomish County than in urban Seattle, both for residential customers and overall. 7 Nebraska utilities are all public; Tennessee has one small IOU 8 We do not provide similar comparisons for commercial, industrial and "other" customers because those classes of customers are too varied. Electricity Distribution Costs:Comparisons of Urban and Suburban Areas Lexington Electric Utility Ad-hoc Committee 10/28/03 Page 8 of 12 Exhibit 5. Comparison of Costs, Seattle and Vicinity Year 2001 Snohomish Utility Name Source Seattle Muni County PUD Type of area urban suburban Summer Peak (MW) EIA-861 1,234 901 Winter Peak(MW) EIA-861 1 755 1,286 Net Generation (MWh) EIA-861 3,929,875 612,783 Purch Util (MWh) EIA-861 3,905,212 7,022,820 Exchng-Net(MWh) EIA-861 2,902,435 -95,678 Total Source (MWh) EIA-861 10,617 126 7,539,925 Sales to Consumers(MWh) EIA-861 8,990,171 6,184,545 Frnshd W/O Chrg (MWh) EIA-861 78,740 Total Production Expenses ($1,000) EIA-412 403,151 369,121 Total Production Plant ($1,000) EIA-412 538,554 334,654 Carrying Cost of Prodn plant (%/yr) assumed 5% 5% Cost of purchased &produced power net of resale revs (0/kWh) Lex EUC 3.90 3.86 Retail Revenue ($1,000) EIA-861 503,437 395,501 Resale Revenue ($1,000) EIA-861 75,333 146,285 Rates(by class of customer) minus wholesale power analysis Residential Revenues ($1,000) EIA-867 178,129 225,850 Residential sales (MWh) EIA-861 2,973,916 3,215,167 Residential customers (#) EIA-867 321 422 246,204 MWh/year per residential customer 9.3 13.1 Residential rate (0/kWh) 5.99 7.021 minus cost of power (0/kWh) 2.09 3.16 Total Revenues ($1,000) EIA-867 500,886 395,501 Total sales (MWh) EIA-867 8,974,215 6,184,545 Total customers (#) EIA-861 354,556 271 193 Total rate (0/kWh) Lex EUC I 5.58 6.391 minus cost of power net of resale revs (0/kWh) Lex EUC 1.68 2.53 2.5. Tennessee Municipal Utilities There are 47 munis in Tennessee with more than 5,000 customers; each muni serves one or more cities, and typically some surrounding areas as well. All those munis purchase their wholesale electricity primarily from the Tennessee Valley Authority at similar prices. Therefore, differences in rates should be driven primarily by differences in distribution costs. Given this wealth of data, we decided to look at the effect of customer density on distribution costs (as Electricity Distribution Costs:Comparisons of Urban and Suburban Areas Lexington Electric Utility Ad-hoc Committee 10/28/03 Page 9 of 12 reflected in rates) over the range of customer densities. We measured density as the ratio of the number of residential customers per mile of distribution circuit each muni serves.9 Exhibit 6 is our plot of the munis' residential rates as a function of customers per mile of distribution circuits, including the linear trend line through the scatter plot. Both the scatter plot and the trend line indicate that costs are lowest in the more urban service areas and increase as the service area becomes more suburban or rural. Exhibit 6: Rates and Density of Tennessee Municipal Utilities Tennessee munis 00 4. 4 r~4-4 A f C ' • € 6.50 d t • :. .. . C o U 4 5.50 5.00 15.00 25.00 35.00 45 00 55.00 # residential customers/total miles of distribution circuits 9 We used data from the Utility Data Institute's DarWin database for 1996. Electricity Distribution Costs:Comparisons of Urban and Suburban Areas Lexington Electric Utility Ad-hoc Committee 10/28/03 Page 10 of 12 3 Conclusions All five analyses presented above are consistent with the conclusion that electricity distribution costs are lower in urban areas than in suburban areas. • A full cost-allocation study with geographical accounting data for distribution costs found that distribution was less expensive in New York City than in its suburbs. • Among the three NSTAR subsidiaries, the least urban (Com Elec) has the highest distribution rates, the most urban (Cambridge) has the lowest. • The more urban portion of the PEPCo territory (Washington DC) has lower residential distribution rates than the suburban (Maryland) portion. • Seattle City Light has lower delivery costs for residential customers and on average than the Snohomish PUD, which serves the surrounding suburban area. • Among the Tennessee munis, all of which have essentially the same power supply, residential rates are lowest for the munis with the highest density of customers per mile of distribution circuits. These analyses used all the viable approaches we could think of, short of performing a geographic cost-allocation study for BECo. They compare different parts of the same IOU, different IOUs owned by a holding company, and different publicly-owned utilities. They use an analysis by a regulatory agency, rate schedules, and accounting data. They use data from the 1980s, 1990s, and 2001, as well as current rates. They all reach the same conclusion. The suburban-flight scenario rests on an assumption that is not only wrong, but backwards. If any group of municipalities have an economic incentive to form municipal electric utilities, it seems to be the large cities, not the suburbs.10 Therefore, there is no reason to believe that making municipalization feasible in Massachusetts would disadvantage the larger cities, or force them to pay higher rates. To the contrary, an urban area such as Boston would benefit in three different ways from enactment of Bill 111468. 10We expect that inherent cost differences between communities are unlikely to drive municipalization. Some mix of low-cost and high-cost communities, urban, suburban and rural, may become seriously interested in municipalizing, depending on how dissatisfied they are with their IOU Electricity Distribution Costs:Comparisons of Urban and Suburban Areas Lexington Electric Utility Ad-hoc Committee 10/28/03 Page 11 of 12 • By making practical the option to form a muni, Bill H1468 will create a form of competition for electricity distribution, thereby leading the IOU to provide better service and lower rates everywhere, including in urban areas. • Just as for a suburban community, the economics of forming a muni for a city may turn out to be sufficiently attractive to warrant the effort, with possibly better service and substantial savings on their electricity bills for residents, businesses and the City itself. • If a typical suburban town buys out its portion of the IOU's infrastructure to form a new muni, the remainder of the IOU's service territory, including the urban areas, will likely see a reduction in the IOU's rates, since the IOU's average distribution costs will decrease once the IOU no longer has to serve that high-cost suburb. Electricity Distribution Costs:Comparisons of Urban and Suburban Areas Lexington Electric Utility Ad-hoc Committee 10/28/03 Page 12 of 12