Investigation to Establish Guidelines for Service Quality Standards for Electric Distribution Companies and Local Gas Distribution Companies Pursuant to G.L. c. 164, ' 1E

DTE 99-84

INTRODUCTION

The Department of Telecommunications and Energy ( A Department @ ) opened this proceeding A to develop guidelines for service quality standards to be included in the performance-based regulation ( > PBR = ) plans @ gas and electric companies file pursuant to G.L. c. 164, ' 1E ( A' 1E @ ). October 29, 1999 Notice of inquiry ( A NOI @ ), p. 1. The cited statute authorizes the DTE to adopt A rules and regulations to establish and require performance based rates @ for gas and electric utilities. It further requires that the Department, when A promulgating such performance based rate schemes . . . [shall] establish service quality standards @ for each gas and electric company. ' 1E(a). The legislation authorizes the Department to impose penalties on companies that fail to meet any established service quality standards, in an amount not to exceed 2 percent of the company = s transmission and distribution ( A T&D @ ) revenues.
In prior proceedings (e.g. Boston Edison Company, DTE 96-23 (1998); Cambridge Electric Light Company et al., DTE 97-111 (1998); Eastern-Essex Acquisition, DTE 98-27 (1999)), the Department noted the importance of establishing service quality standards as companies restructure or merge. The Department also stated its preference to address service quality issues in a generic proceeding in order to treat companies evenhandedly. This past summer, the Department formally announced its intent to open the present proceeding. Eastern-Colonial Acquisition, DTE 98-128, at 16, n. 20 (1999).
In this proceeding the Department seeks comments on two broad topics: the design of service quality plans and of the penalty mechanisms that would apply to companies that fail to meet service quality standards.

INTEREST OF THE UWUA

The Utility Workers Union of America, AFL-CIO ( A UWUA @ or A Union @ ) is pleased to offer comments in this proceeding. As the legislature emphasized in passing the Electric Industry Restructuring Act (St. 1997, c. 164, ' 1(h), A reliable electric service is of utmost importance to the safety, health and welfare of the commonwealth = s citizens. @ This proceeding is an essential step to make sure that A electric industry restructuring . . . enhance[s] the reliability @ of our electric system. Id.
UWUA is a national union that represents 50,000 utility workers across the company, 5,000 of those in Massachusetts. UWUA = s members work to provide consumers across the country with a safe, reliable and affordable supply of gas and electricity. As the people who inspect manholes and substations, fix gas line leaks, and answer customers = telephone calls, UWUA = s members have the experience and knowledge to recognize the value to consumers of reliable supply and the measures that must be taken to protect against service declines in a restructured environment.

1. LEGAL STANDARD

As noted in the Introduction, the Restructuring Act authorizes the Department to promulgate rules under which electric and gas companies file performance based rate plans. G.L. c. 164, ' 1E(a). The law also requires the Department, A in promulgating such performance based rate schemes @ to establish quality of service standards for each company. Those standards must, at a minimum, include the following:

customer satisfaction
service outages
distribution facility upgrades
repairs and maintenance
telephone service
billing service
public safety

In addition, the Department must establish benchmarks for employee staff levels and employee training programs. ' 1E(a). A gas or electric company that makes a performance based rate filing after the effective date of the Restructuring Act cannot reduce its staffing levels below November 1, 1997 levels, unless the reductions are included in any applicable collective bargaining agreement or the company demonstrates to the Department that the reductions will A not adversely disrupt service quality standards. @ ' 1E(b). Companies must file annual reports with the Department each year, comparing actual performance in the prior year to the standards. G.L. c. 164, ' 1E(c). The Department may levy penalties up to 2 percent of a company = s T&D revenues for failure to meet service quality standards. Id.
The Department has thus far required only some electric and gas companies to adopt several of the quality of service measures specified in ' 1E, either in restructuring or acquisition orders. See, e.g., Boston Gas Company, DPU 96-50 (1996)(setting measures for response to odor calls; employee lost time accidents; telephone response time; meeting scheduled appointments; DTE complaint cases; customer bill adjustments; and on-time meter reads); Eastern-Essex Acquisition, DTE 98-27 (1998)(establishing procedures for setting similar measures); Boston Edison Company, DTE 96-2 at 47-56 (1998)(setting measures for outage duration, customer satisfaction, and line losses). Some companies have not yet adopted any of the ' 1E measures; no company has yet adopted standards for distribution facility upgrades or repairs and maintenance.
The legislature separately mandated the adoption of service quality measures in G.L. c. 164, ' 1F(7). Under that section:

The department is authorized and directed to oversee quality and reliability of service and to require that quality and reliability are the same as or better than levels that exist on November 1, 1997. . . . The department is authorized and directed to promulgate rules and regulations to establish service quality standards for each distribution, transmission and gas company, including, but not limited to, standards for universal service, customer satisfaction, service outages, telephone service, billing service, and public and employee safety.

Section 1F(7) differs from ' 1E in three important aspects. First, under ' 1F, the Department is unconditionally A directed @ to A establish service quality standards, @ whether or not performance based rate rules have also been adopted. Second, the Department must insure that quality of service does not decline below November 1, 1997 levels. This is important because many companies have cut staffing significantly since 1997. Service quality may have declined as a result. Third, ' 1F requires the department to adopt a standard for universal service.

1. THE QUALITY OF SERVICE IS AT RISK

The Massachusetts utility industry is undergoing major change, driven not only by the 1997 Restructuring Act, but also by economic forces that affect companies across the country. Electric companies have divested themselves of generating assets; electric and gas companies are engaging in a wave of mergers and acquisitions;1 and virtually every company has reduced its staffing levels in preparation for a more competitive environment. In a recent survey of its twelve largest locals, the Union found that staffing levels are generally down 20% to 30% since 1991, across the country. The two largest Massachusetts locals (Local 369 and Local 387, both at Boston Edison Company) are down 33% and 27%, respectively, from 1991 staffing levels. Union members here and elsewhere consistently note several problems connected to staff cuts and restructuring: their companies perform inspections less frequently; necessary but non-emergency repairs are deferred, sometimes indefinitely; while the average age of the workforce is increasing, retiring workers are often not replaced; and some companies are cutting back on training programs for new employees. Workers are worried about system reliability, their own safety and the safety of the public.
The gas and electric infrastructure is aging and, in states like Massachusetts with the oldest systems, in need of continuous maintenance.2 Distribution system upgrades are needed not only to replace old and defective equipment but also to keep up with increasing loads on pole-top transformers, feeder circuits, and substations.3 Recent incidents in Massachusetts and elsewhere demonstrate the perils of cutting back too deeply on maintenance efforts and staffing levels.
On March 4, 1998, a gas explosion in Attleboro killed two people, injured seven others, destroyed one home and damaged 68 others. A city backhoe that was excavating a drain line hit a Bay State Gas line in an area marked as A NO GAS @ by an employee of Bay State = s outside locating service. DTE Incident Report, 57-59 George Street, Attleboro, MA, March 4, 1998. The inexperienced employee the contractor hired was unaware that there could be gas service in a house that had no outside indication of gas supply. A Statement of Brian McCarthy @ (Attachment A of these Comments)( A McCarthy Statement @ ). Companies decide to use outside locating services rather than in-house employees to save money. This type of cost-cutting will only increase in the restructured environment.
In Chicago, Commonwealth Edison = s system experienced major outages this past summer, one of which shut down the federal court and the downtown Loop. Over 100,000 customers lost power over the course of the summer. On September 15, 1999, the company issued a major investigative report and a plan to address weaknesses in its transmission and distribution system. The report noted:

[W]hile ComEd = s inspection programs seemed appropriate, there were only imperfect mechanisms in place to ensure execution [of repairs]. . . It is not clear, from a review of the records, how often inspections were actually performed, and the inspections that were performed may have been too passive, too cursory, to truly maintain the system .


Additionally . . . ComEd needs to ensure better follow-up on maintenance requests. While virtually all T&D emergencies are dealt with immediately, there appear to be altogether too many deficiencies which, had they been identified and addressed sooner, would not have become critical in the first place. . . .[R]outine maintenance requests . . . were rarely tracked to ensure follow-up.

A Blueprint for Change: Executive Summary for the Investigative Report By Commonwealth Edison, at A-11 (September 15, 1999).
Fortunately, no Massachusetts utility experienced similar outage problems this summer, but the quoted description of underlying maintenance problems reads to many the Union = s Massachusetts members as a description of their own systems. As in Chicago, Massachusetts companies respond well to emergencies and the problems that the public sees. But routine maintenance is quite often deferred. A Statement of Phillip A. Trombly, @ Attachment B ( A Trombly Statement @ ).4 Inspection cycles have doubled or tripled and critical equipment is often in poor condition when eventually inspected. Trombly Statement. There are neither sufficient numbers of workers nor adequate management systems to follow up on repairs identified through routine inspections. Trombly Statement The problems range from poles that are condemned by workers but not replaced; to load tap changers that are inoperable, affecting proper voltage levels; to uninspected transformers that pose a serious risk of exploding. Trombly Statement.
UWUA firmly believes that setting service quality standards will have a significant and beneficial impact on company practices. Managers translate these measures into goals for employees and into changed maintenance and customer service practices. See McCarthy Statement. The Department can protect the reliability, safety and quality of the utility system through proper design of service quality measures.

1. THE DEPARTMENT MUST GATHER ADEQUATE INFORMATION FROM COMPANIES BEFORE ESTABLISHING FINAL RULES OR GUIDELINES

The Department = s procedural schedule includes a round of initial comments, due December 3; a round of reply comments, due December 22; and the possibility of further proceedings that A may provide for technical sessions and/or hearings. @ NOI, at 4.
UWUA encourages the Department to gather adequate information from all utility companies before establishing final regulations or guidelines. At present, only some of the state = s gas and electric companies have any of the ' 1E service quality measures in place. No company has measures for all of the categories included in G.L. c. 164, '' 1E and 1F(7). Companies currently track the information needed to set service quality measures in somewhat inconsistent fashion, and some companies do not track the information at all. See, e.g., Eastern -Essex Acquisition, DTE 98-27, at 35-39 (1998)(discussing differences between Boston Gas and Essex Gas data collection systems). The Department will have difficulty establishing consistent measures and an even playing field for all companies if it does not gather additional information.5
Gathering information will also allow the Department to develop standards that will be easier to implement and that are more likely to improve service quality and reliability. For example, if all gas companies first provide information on the percentage of odor calls that are answered in one hour, the Department will more easily be able to set an achievable goal that will treat companies fairly yet raise the bar for any poorly performing companies. Similarly, if all companies report on their current staffing levels, turnover rates, and internal staffing benchmarks, the Department can more readily implement the statutory requirement of setting employee benchmark levels. In order to set a universal service standard, the Department should gather more information about the number of each company = s low-income customers who are served on the low-income tariff, who receive weatherization services, who make payments under a deferred payment plan, or who have been disconnected.
UWUA includes as Attachment D a list of questions it believes each company should address in further proceedings in this docket.

1. DESIGN OF SERVICE QUALITY PLANS/SPECIFIC MEASURES


1. Introduction: Types of Actions the Department Can Take

The Department can take different types of actions that will improve reliability, safety, and customer service. Depending on the area of concern, the Department may:

C Set performance standards now, as to measures for which adequate information is available. Some standards may be enforced with financial incentives, while others may simply establish norms (e.g., for inspections and maintenance), be used in storm performance reviews, or serve as triggers for customer guarantees;

C Prescribe the information that utilities must file in the near future, to allow cross-company comparisons and the future setting of standards;

C Require regular reporting of data that the utilities already collect. Reporting alone may improve utility performance, by concentrating management = s attention on problem areas;

C Require additional data collection and reporting;

C Impose requirements for improved data-collection systems, to allow the Department (and the utility) to monitor performance, and to be used as the basis of future standards and incentives;

C Implement customer guarantees.

Below, UWUA recommends a variety of these actions. No single tool is adequate to carry out the statutory mandate (G.L. c. 164, ' 1F(7)) that service shall not decline below 1997 levels.

2. Form of standards

The Department may implement standards in these forms:

C Data collection and reporting.

C Penalties (and potentially rewards1) for aggregate performance

C Customer guarantees of service, implemented as changes in the utility = s terms and conditions.

While the first two categories are self-explanatory, customer guarantees are a relatively new concept in utility regulation. In contrast to the broad effect of the aggregate penalties under G.L. c. 164, ' 1E, which compensate all ratepayers with a small rate reduction if service is inadequate overall, a customer guarantee results in direct payments to the customers affected by specific failure in the utility = s service. Aggregate penalties and rewards are determined annually or periodically, in rate cases or other proceedings before the Department, but customer guarantees take effect immediately, giving the utility prompt feedback and incentive to maintain service quality.
For example, in connection with its merger with Scottish Power, PacifiCorp proposed, and its five state commissions accepted, a set of eight customer guarantees. Under those guarantees, if PacifiCorp is not timely in showing up for an appointment, restoring service, or responding to a complaint, it will pay the customer $50. Customer guarantees are also offered by all UK electric distributors.
Customers who are suffer from missed appointments, for example, would receive direct compensation for their inconvenience and wasted time. Missed appointments and inadequate response to customer inquiries are frequent and often irritating problems of dealing with large organizations. The customer guarantee payments should make the worst-affected customers feel better, far more than aggregate penalty payments diluted over an entire customer base.
In addition, the payments under customer guarantees would make inadequate customer service very conspicuous within the company. While the financial effect would likely be minor, the fact that a check must be written will increase the responsibility of the entire organization that delivers the service, from the service person who showed up late, to the dispatcher who handled the scheduling, to their supervisors, up to the managers who set the budgets for performing these functions. The experience in the UK is that customer guarantees work well and achieve their goals quickly. Utilities initially made significant customer guarantee payments, but these fell to negligible levels within a couple of years.
Customer guarantees also operate with little Department oversight. If utilities properly inform their customers of the existence of the customer guarantee program, the operation of the guarantees can be left largely to utilities and their customers, with the Consumer Division mediating occasional disputes.

3. Subjects for Incentives

Performance standards should be developed to focus on the following four broad areas:

C T&D reliability,

C Commodity losses,

C Quality of customer service, and

C Employee and public safety

Performance standards can target:

C Outcomes (including interruption frequency and duration).

C Process (including inspection, maintenance and sizing standards).

C Inputs (including staffing, equipment inventory, employee training).

These approaches can be tied together. For example, the Department might give the utility wide latitude regarding staffing and maintenance standards (that is, process), if reliability (an outcome) does not decline, but might put the utility on notice that reduced staffing and maintenance will trigger penalties if this contributes to slower response to storm damage.
None of the recommendations made in these comments are intended to eliminate or reduce existing performance standards that have been adopted in any of the Department = s restructuring or merger orders.

1. Performance Standards for T&D Reliability Outcomes

Reliability performance standards for electric utilities should cover the following areas:

C System average interruption duration index (saidi),

C System average interruption frequency index (saifi),

C Momentary average interruption frequency index (maifi)

C Performance of the utility = s worst-performing circuits,

C The time required for supply restoration for worst-affected customers.

Standards based on SAIDI and/or SAIFI have been adopted repeatedly by the Department, in Boston Edison Company-Commonwealth Acquisition, DTE 99-19 (1999), Boston Edison Company, DTE 96-23 (!998), Eastern Edison Company, DTE 96-24 (1997), and others merger or restructuring cases.
A standard for worst-performing circuits is appropriate to ensure a measure of equity across each utility = s service territory, and to prevent the neglect of circuits with chronically poor performance. All customers in a rate class pay the same rate for T&D service, and they should all receive reasonable value for the bills they pay. Even if the utility = s average performance is adequate, it may be seriously short-changing those customers who live in areas in which repairs and upgrades are more difficult and who are not particularly vocal or well-organized to demand better service. Data obtained through discovery in DTE 96-23 indicate that for Boston Edison, at least, worst-performing circuits tend to be concentrated in inner-city residential areas.2
Power quality, including voltage stability, short-term (e.g., 6-cycle) voltage sags, voltage spikes, frequency stability, and harmonics, is another important area. Localized power quality problems might be appropriately addressed through a customer guarantee, although an aggregate system-wide standard may also be useful.
There are two important requirements for effective implementation of reliability performance standards: a clear operational definition of an outage and accurate data.

1. Definition of Outage

For operational purposes, the definition of an outage should exclude the two extremes from the data: momentary outages of less than one minute and major events.3 Momentary outages should be excluded because, individually, they impose a fairly insignificant cost to most customers. Major events increase the instability of the performance measure and as a result, their inclusion distorts inter-year comparisons. For example, the outcome of an unusually large storm in any given year will skew that year = s frequency of outages when without the storm, utility performance could be well within the normal range.
Outages should be classified as excludable major events if they meet one or more of the following criteria:

C Major events like storms resulting in simultaneous multiple equipment failures. Events should be excluded based on the number of separate incidents or failures, not simply because the localized outcome is severe. For example, a major storm that takes out a large number of poles and lines simultaneously should be excluded. A single substation failure that interrupts service to many customers should not be considered a major event.

C Weather conditions that interfere extraordinarily with utility response to outages. For example, in this fall = s rainstorms, flooding and persistent high winds prevented Florida Power and Light crews from working on overhead lines for over a day. This increased the duration of outages. Large snow drifts from a winter storm can also hinder repairs. If this type of weather does not cause an exceptional number of outages, the event may be included in saifi, but excluded from saidi.

C Generation and transmission ( A G&T @ ) problems beyond the control of the distribution company or its affiliates. Outages in the regional generation supply, or on the transmission system of unaffiliated utilities, can result in disconnection of distribution customers, due to insufficient capacity, voltage, or other problems.

Meaningful measurement of utility performance requires that these excludable events be strictly defined. The precise numerical definition that is appropriate will depend on the extent of the variation in outage data and is likely to differ among utilities.
Excluded events should not be ignored: momentary outages should be addressed through a separate performance standard and any major event excluded from a utility = s performance measure should be subject to Department review. Regardless of the seriousness and cause of the outage, utility actions before and after the event affect the outcome. Therefore, events that relate to storms, G&T failures and other major events, the Department should conduct an investigation, on some periodic basis, that would consider the following factors:

C Emergency preparedness (e.g., availability of trucks and equipment, inventory levels, arrangements for obtaining additional resources, emergency training),

C Preventive inspections and maintenance steps taken before the major event,

C Cost control,

C Service-restoration efforts.
A utility should not be allowed to recover storm costs without a favorable Department review of the company = s preparedness and response.

2. Data Collection and Reporting for T&D Reliability

Performance cannot be measured meaningfully unless the utility = s data-collection system consistently reports data that fairly represent performance. The utility needs accurate data to improve system reliability and the Department needs accurate data to establish reasonable benchmarks, to measure compliance with performance standards, and to provide effective guidance to the utilities.
The informal and manual systems generally used to tabulate outages and generate reliability statistics are not very accurate. For example, when Scottish Power converted to an automatic recording system, with full information about that number of customers disconnected by each outage, it found about 25% more customer outage minutes than under its previous system. Based on its initial review of the PacifiCorp outage reporting system, Scottish Power estimated that PacifiCorp had under-reported its outage frequency (saifi) by 80%, and its outage duration (saidi) by 20%.
Furthermore, informal and manual systems are prone to human error, and cost-cutting is likely to lead to further under-reporting of outages. For example, the accuracy of a system that depends on employees entering outage data will suffer as workforce reductions leave employees with less time to record data. UWUA = s Massachusetts members note that staffing cuts at have in fact eroded the ability of utilities to maintain adequate outage records.
To improve the accuracy of the outage data, the Department should require each utility to develop a data-collection system that:
C Includes a database of the number of customers served by each feeder, between each pair of switches, and on each line transformer, so that identification of the failed equipment will directly determine the number of customers affected. This database will improve dispatch of line crews. Dispatchers will be able to identify immediately the number of customers affected by each incident.

C Automatically transfers outage reports, from the scada system, crew dispatch, call center, or other systems, directly to the reliability database, so that the time of interruption and reconnection are as accurate as possible.
In addition, the Department should require utilities to audit their existing outage reporting systems, locate any outage data that has been reported but not compiled and enter data manually where necessary.
The electric utilities should report annually to the Department on:
C their progress in improving data collection,

C reliability results with the new data-collection and tracking mechanism,
C the differences in the data reported under old and new collection systems.

To facilitate comparisons across utilities, data should be reported separately for overhead and underground systems, and for network and radial systems. Most utilities already track data in this level of detail. Data should also be available for each feeder circuit.

3. Recommendations for T&D Reliability Standards

Pending improvements in utility outage data collection, the Department should
C Enforce all reliability performance standards currently in effect.
C Impose interim saidi and saifi standards for electric utilities that do not current have any. For 2000, the baseline for each interim standard should be equal to the average of the utility = s own data for 1993-1997. For 2001, the baseline should be updated to 1994-1998, unless the 1998 data would reduce the baseline, as this would violate the mandate of G.L. c. 164, ' 1F(7) that service quality cannot decline below 1997 levels. In late 2001 or 2002, the Department should revisit the saidi and saifi standards, based on the data then available.
Data on momentary outages (for the industry or individual utilities) do not appear to be sufficient to set standards at this time. The utilities should improve data collection and file data on the frequency of momentary outages during 2001.
In addition, the Department should require each utility to implement customer guarantees. Every time the utility fails to meet one of these standards, it should pay the affected customer at least $50. For some extended failures, the payments would be higher.
C To address the worst-performing circuits, each utility should identify the five feeders that had the highest average SAIDI over the last three years, the five feeders with the highest average SAIFI, the five with the highest average MAIFI, and the two highest by the composite measure: SAIDIF/SAIDIA + SAIFIF/SAIFIA + MAIFIF/MAIFIA, where F is for the feeder and A is system average.4 The utility should be required to bring the measure at least half way to system average within two years, or credit each customer on the feeder $50 in the third and each subsequent year.

The worst circuits of 1998 can now be identified, and should be improved by calender year 2001. Once the 2001 data are available, customer guarantees would be paid (early in 2002) for any circuits that have not improved sufficiently. In a month or so, the utilities should be able to determine the worst circuits of 1999, which should be improved by calendar year 2002. Early in 2003, customer guarantees would be paid for those circuits from 1998 or 1999 that had not reached their improvement targets in 2002. No circuit should be counted in more than one year = s obligation simultaneously.


C . If the customer is disconnected because of an equipment failure in the T&D system, the Company should restore the customer = s service within 12 hours, or credit the customer $50, plus another $5 for each additional hour of interruption.


C . If the utility needs to turn the customer = s power supply off for planned maintenance work or testing, the customer should be given at least 2 days = notice, or pay the customer a $50 credit for the unnecessary inconvenience.

Once the utilities have implemented improved data-collection systems, the Department should reconsider the existing aggregate reliability performance standards, and expand them on a statewide basis. As more accurate actual data are compiled, the utility performance should be subject to thorough public review. Simply maintaining the initial level of reliability indicated by the improved data may not be sufficient. The utilities = cost-cutting activities may have already increased the frequency and duration of outages. To identify deteriorating service, the Department should compare indices across companies and over time. Some differences between utilities may be reasonable, due to differences in customer density, percentage of underground distribution, and severity of weather. On the other hand, differences may indicate that some utilities have failed to provide service quality that their peers in the state and region have provided at reasonable rates.
In general, performance benchmarks should be set somewhere between the industry norm and the average of the utility = s past performance. In the longer term, the benchmark should be set at the better of the industry average and the company = s own historical performance. The benchmark might shift over time to reflect changes in industry or utility performance. For example, the benchmark could be recalculated as the running five-year average for the utility or the industry. However, since ' 1F(7) of the Act requires that reliability not deteriorate below the level provided in November 1997, the pre-1998 averages should be used as a floor, below which the standards are never allowed to fall.6

2. Performance Standards for Commodity Losses

Losses (line losses for electric utilities, unaccounted-for commodity for gas utilities) are somewhat unique in that the value of increased or decreased performance is easy to measure. Each kWh or MMBtu that is delivered to the utility, but not to customers, has a clearly-defined economic cost: the price of the lost commodity. Whether the commodity is provided by the utility, or by a third-party marketer, the costs of the losses will eventually be borne by the utility = s customers.
Losses are subject to the control of the utility. Electric utilities have many options for decreasing losses, including:

C Installing more and larger conductors at all voltages.

C Using low-loss transformer cores.

C Increasing distribution voltages.

C Improving power factors with capacitors and similar techniques.
Similarly, gas utilities can reduce losses by:

C Reducing leakage from lines.

C Replacing vented equipment (e.g., measurement devices) with non-venting equipment.
Whenever a utility operates under a rate cap, it will have a financial incentive to avoid the expenditures that reduce losses and no incentive to reduce losses that are borne by ratepayers. To even out the incentives, the utility must be required to share in any increase or decrease in the loss factor. This incentive can be provided through a number of mechanisms, such as:

C Holding the loss adjustment charged to ratepayers (or their energy suppliers) constant during the period of a rate cap, with the utility paying the costs of any increase in the loss percentage or retaining the savings from any decrease.

C Setting a target loss percentage, based on the past performance of the utility and the industry, and allowing the utility to charge customers or marketers for a weighted average of the target and actual loss percentage.
Losses vary with the mix of customer loads, since small customers (and electric customers served at lower voltages) tend to have higher losses than large customers. Reported losses also reflect under-metering of gas by small meters at low temperatures, and the differences between delivery and meter-reading dates. A cold December will tend to increase reported losses for the year, since the sendout is reported in December but the meters are not all read until January. Therefore any target loss percentage should be adjusted for weather and changes in customer mix.
Each utility should be required to file its annual average commodity loss for each of the last ten years, and adjust those data, if feasible, for weather and customer mix. Any PBR plan should either hold the loss recovery percentage constant between rate reviews, or compute the annual loss allowance as a weighted average of a target percentage (based on the experience of the utility and the entire industry) and the actual losses.

3. Performance Standards for Customer Service Results

1. Customer service quality standards should include the following areas:

C Customer-satisfaction, based on targeted surveys of municipalities, large customers and residential customers who have recently contacted the company.

C Number of and responsiveness to customer complaints.

C Phone center performance (e.g., percentage of calls answered live within a given number of seconds).

C Installation and repair appointments (e.g., number or percent of appointments missed, average time from order to ordinary install or repair).

C Billing (e.g., number of corrected bills, time to investigate high bills or other billing complaints, involuntary disconnections per 1,000 customers).

C Percentage of on-cycle meter reads.

C For gas companies, response to odor and leak calls, with differentiation by severity (such as the distinction between Type I and Type II calls noted in the Department = s questions).
In Boston Gas Company, DPU 96-50 (Phase I), Bay State Gas Company, DTE 97-97, Eastern-Colonial Gas Acquisition, DTE 98-128 and Boston Edison-Commonwealth Acquisition, DTE 99-19, the Department adopted the following customer service performance measures for gas and electric utilities: (a)calls answered within a specified time; (b)appointments met on the same day as requested; (c)complaints reported by Department = s Consumer Division; (d)amount of adjustments made to customers = bills, and (e) on-cycle meter reads.

1. Recommendations:

1. The Department should continue in force the customer service measures implemented thus far and require companies that have not yet implemented customer service performance standards to do so by a date certain. G.L. c. 14, ' 1F(7) requires the Department to establish service quality measures for customer satisfaction, telephone service and billing service regardless of whether a company has filed for PBRs. Enough information is currently available to permit the Department to set reasonable benchmarks in the near future.
Each company should have standards for the following performance measures:

C percentage of bills based on actual meter readings.

C number of involuntary disconnections per thousand customers.

C corrected bills per thousand customers.

C phone-center performance, particularly the number of calls answered within a given number of seconds.5

C number of customer complaints to the Department per thousand customers, focused on complaints that require Department intervention (and excluding simple inquiries).

1 customer satisfaction, as measured by surveys of commercial/industrial customers, municipalities and residential customers who have had recent contact with the company.7

In addition, the Department should require the companies to offer its customers guarantees of satisfactory service. Every time the utility fails to meet one of these service standards, it will pay the affected customer at least $50. The standards listed below were proposed by PacifiCorp and accepted in five states. The Department should adopt them for Massachusetts utilities:

1) Appointments We will keep all mutually agreed appointments with the customer, whether over the phone or in writing. Beginning in the year 2001 we will offer the customer a morning appointment, between 8 AM and 1 PM, or an afternoon appointment, between 12 Noon and 5 PM.


2) Switching On the Customer = s Power Upon customer request we will activate the power supply within 24 hours provided no construction is required and all government requirements are met.


3) Response to Bill Inquiry If the customer has a question about the electric bill we will investigate and respond to the customer = s inquiry within 15 business days.8


4) Problems with the Customer = s Meter If the customer suspects there is a problem with the meter we will investigate and report back to the customer within 15 business days.

5) Power Quality Complaints If the customer notifies us about a problem with the quality of electric supply we will either initiate an investigation within 7 days or explain the problem in writing within 5 business days.

Each electric utility should adopt these standards, or apply to the Department for modifications that reflect company-specific factors. Each gas utility should adopt standards 1 B 4, with minor modifications to reflect gas terminology.

4. Performance Standards for Employee and Public Safety Results

The Department should adopt the following four safety standards.

C Lost-Time Accident Rate standard, based on the number of lost-time accidents per 200,000 hours worked by company employees; and

C Lost-and-Restricted Hours Rate standard, based on the Lost Time Accident hours plus the Restricted Work hours per 200,000 hours worked by company employees.

C First-Party and Third-Party Damage to Underground Lines, as incidents per year.

C Odor and Leak Response Time standard (for gas companies), based on the percentage of calls to which the utility arrives on the scene within one hour. Data should be recorded separately for response times to serious Class I and less-urgent Class II odor calls.
The first two standards are modeled after the NEES standards filed in DTE 99-47. The latter two standards are based on provisions in place for Boston Gas, Bay State and/or Essex Gas Companies, in DTE 96-50, 97-97, and DTE 98-27, respectively.

5. Process Performance Standards

The Department must make sure that companies follow the proper procedures for inspecting and maintaining their systems. It can do so by specifying inspection and maintenance procedures; by reducing rates to customers when utilities do no follow procedures and service disruptions result; or by treating failure to comply with process standards as prima facie evidence of imprudence in assessing storm damage. The Department can also adopt customer guarantees so that customers are paid when equipment that was not properly inspected or maintained fails.
Performance standards could include utility compliance with established procedures and schedules for inspection of all appropriate equipment, including such tasks as:

C Physical testing poles, towers, and supporting equipment for structural integrity.

C Visual inspection of transformers, switching and protective devices, regulators and capacitors, conductors and cables, and company-owned street lights, to check for obvious external hazards and physical defects.

C Less-frequent but detailed inspection of distribution equipment with written reports on equipment condition.

C Gas safety inspections, including those required by CMR 101.06.

C Testing and replacing of meters on a timely basis.

C Monitoring of loads of installed equipment.

Maintenance procedures and policies can also be addressed through performance standards. For example, the Department could establish standards for:

C Equipment stocking, to ensure that inventory is sufficient for timely replacements, even with multiple outages.

C Tree-trimming schedules.

C Identifying equipment that should be replaced or reinforced before failure, considering the extent of current and projected loading, vulnerability of the equipment to future overloads, and line loss reduction.

The trend among many Massachusetts utilities has been to reduce the stock of spare equipment available. This is most easily demonstrated for line transformers. Attachment C is a graph of the ratio of the number of line transformers in stock to the number in service, for each of the larger Massachusetts electric utilities, for the years 1989 B 1998, from the utilities = FERC Form 1 filings, page 429. The declines in the number of spares has been dramatic for Massachusetts Electric and Boston Edison; the trend for Eastern Edison has been gradual but steady over the last six years; and WMECo = s stock has been at very low levels from the start of the electronically-available data in 1993.6 The data in Attachment

C are consistent with the experience of UWUA members that BECo = s stock of transformers, especially those sized for residential neighborhoods, has not been sufficient to insure prompt replacement of failed transformers, especially during summer heat waves.

Performance standards can also address equipment sizing standards and procedures, such as:

C Physical loading requirements for overhead lines and poles, potentially resulting in shorter spans, heavier messenger wire, better-quality conductors, and better guying.

C Electric loading requirements for new equipment, such as transformers, to reduce overloading and premature failure.

C Algorithms for determining the extent, frequency and duration of overloads on line transformers, and replacement of overloaded transformers.

In determining appropriate process performance standards, the Department should review the utility = s existing standards, compliance with those standards, and the relative stringency of those standards compared to other utilities. The Department should review any changes that the utility has made to its standards to determine if they provide net benefits to ratepayers or are just a cost-cutting measure in response to restructuring.

a. Recommended Department Action on Process Performance Standards
The Department should establish guidelines for inspection, maintenance, repairs and upgrades on a utility-specific basis. These guidelines should reflect each company = s most stringent practices in the past 10 years, except where the Company is able to demonstrate that less-stringent practices in the ratepayers = interest. To provide a basis for setting these guidelines, the Department should require the utilities to provide information on current and past practices including:
C Company manuals on inspections, maintenance and system design.

C Data on equipment inventory levels, with an explanation of why inventory levels for any key equipment have declined in recent years.

C National standards.

C The types of loading data the company collects; a detailed explanation of when a line, transformer or substation is considered overloaded; and company timelines for replacement/upgrading of overloaded equipment.

C Information on the frequency of reliance on mutual aid agreements, including the number of occasions, the number of crews utilized for each event and the nature of the precipitating event, and

C Use of outside contractors for inspection, maintenance, testing.7
In addition, the Department should institute a reporting and recording requirement, similar to that implemented in California. In California, every utility is required to file a short and concise annual report of its A corrective maintenance program. @ The report includes the percentage of each equipment category (e.g., poles) inspected on schedule, the condition of the facilities inspected, the percentage in need of corrective action, a schedule for correction, the percentage corrected on schedule and an explanation for delays in corrective action. In addition to this report, the California Commission requires utilities to maintain and make available on request detailed records of equipment problems identified on inspection and corrective actions.9 Until the Department has established inspection and maintenance guidelines, the reports should be based on the utilities = current inspection schedules. Once there are new guidelines, these reports can be used to monitor compliance.
At present, the UWUA does not recommend direct penalties for noncompliance with process guidelines. However, noncompliance should trigger customer guarantees if:

C extended outages are tied to a failed transformer, switch, etc., that was not properly inspected, or not repaired after a failed inspection,

C extended outages are due to fallen pole that was not properly inspected, or replaced after a failed inspection, and

C any delay in replacement of failed equipment is due to insufficient stock.

1 EMPLOYEE STAFFING LEVELS AND TRAINING

G.L. c. 164, ' 1E(a) provides that the Department A shall include benchmarks for employee staff levels and employee training programs @ as part of any set of service quality standards. G.L. c. 164, ' 1E(b) further provides:

In complying with the service quality standards and employee benchmarks established pursuant to this section, a distribution, transmission or gas company that makes a performance based rate filing after the effective date of this act shall not be allowed to engage in labor displacement or reductions below staffing levels in existence on November 1, 1997, unless such are part of a collective bargaining agreement or agreements between such company and the applicable organization or organizations representing such workers, or with the approval of the department following an evidentiary hearing at which the burden shall be upon the company to demonstrate that such staffing reductions shall not adversely disrupt service quality standards as established by the department herein.

UWUA believes that this staffing level mandate must be applied by function or department within each company, not to total staffing levels. The obvious intent of the legislature is to protect against staffing reductions that directly affect quality of service: cutbacks in the number of customer service representatives, those involved in resolving billing disputes, repair personnel, line crews, cable splicers, etc. In the unlikely event that a particular company has, for example, reduced its customer service staff while increasing its human resources department, ' 1E still applies and the company should be required to demonstrate that the smaller customer service staff will not result in poorer service.
In order to develop staffing benchmarks, the Department must gather more information about staffing levels, employee turnover, employee training, and any benchmarks the companies themselves employ to determine staffing levels. (UWUA includes sample questions that companies should answer, in Attachment D, under the heading A Employee Staffing and Training Information. @ ) Any company that has reduced its staffing below November 1, 1997 levels, in any function or department that directly affects system reliability, response to customer inquiries, meter reading, emergency response, or billing disputes, must be required to demonstrate either that the staff reductions are consistent with any relevant collective bargaining or that they will not disrupt quality of service.
In order to carry out the statutory mandate for setting benchmarks for employee training programs, the Department = s guidelines should require each company to file a report that includes a description of the training programs that existed on November 1, 1997; that are in place as of the time the report is filed; and, if training programs have been scaled back since 1997, a demonstration that any training cutbacks will not impact quality of service. Training benchmarks are important because staffing levels in key operational departments have been declining steadily for the past ten years at many Massachusetts companies. The smaller workforce must be highly trained to insure the reliability, safety and quality of gas and electric service.

1 RESPONSES TO DTE = S QUESTIONS
UWUA has addressed above many of the Department

= s questions posed in its November 5, 1999 Memorandum. UWUA includes additional responses below.

Question 1 (regarding uniformity of performance standards): The number and types of measures should be uniform, although some measures will be specific to gas companies (e.g., response to odor calls) or to electric companies (e.g., frequency and duration of outages). Every company must have a standard for telephone response time, for customer satisfaction, for universal service, and for all of the standards in G.L. c. 164, '' 1E, 1F(7). The actual numerical goal (e.g., responding to 90% of the calls in 20 seconds) should be uniform to the extent possible, but certain standards must be based on the company = s own geography, circumstances, and customer base. For example, an electric company with a more suburban or rural customer base may have different outage frequencies and durations than a more urban company with a large proportion of customers served underground.
Some measures may have varying numerical targets at the outset but a uniform goal over time. For example, the Department might approve higher and lower telephone response time goals for companies that presently have different equipment in place, but over time the Department should require all companies to have up-to-date equipment that would allow each of them to respond to calls in the same amount of time. Similarly, it is unlikely that any two companies presently use identical customer satisfaction surveys and therefore the numerical goals for customer satisfaction may have to vary. In the long run, however, the Department should require every gas company and, separately, every electric company to use the same survey and comply with the same numerical goal.
The advantages of uniform numerical goals are that they create a level playing field for all companies and treat customers equally, regardless of where they live.

Question 2 (regarding the use of the measures adopted in Boston Gas Company, DPU 96-50 as the model for gas company measures) and
Question 3 (regarding the use of the measures adopted in Boston Edison Company, DTE 99-19 as the model for electric company measures)
The Boston Gas model and Boston Edison model are too limited to apply to all gas and electric companies. There are no measures for distribution facility upgrades, repairs and maintenance, or universal service, and no benchmarks for employee staffing levels and training, all of which are required by the Restructuring Act.
Regarding DPU 96-50, each of the measures included in that case is an important measure. In its comments, above, UWUA proposes modifications to several of the types of measures included in DPU 96-50. As one example, UWUA believes targeted surveys of municipalities, commercial/industrial customers, and customers who have recently contacted the company are more valuable than general customer surveys, as the latter are more a reflection of the company = s public relations efforts than the former. UWUA proposes a broader range of billing and customer service measures, including number of corrected bills and time to respond to billing complaints, and also proposes a set of customer guarantees. Finally, UWUA proposes that there should be a standard for commodity losses. (See preceding comments for specific proposals).
Regarding DTE 99-19, UWUA considers all of the measures in that case worthwhile but its proposals, above, would modify some of the 99-19 measures and add new ones. UWUA proposes measures for momentary outages, for worst-performing circuits, and for losses. UWUA also proposes operational definitions for A outages @ and A major events @ to count those outages that have a measurable impact on quality of service and which are not due to forces outside of company control. (See preceding comments for specific proposals).
Question 4 (regarding the statistics compiled by the Consumer Division): Complaints and calls to the Department = s Consumer Division are important to consider because they can reflect the failure of the company to respond adequately to consumer inquiries or complaints. A single caller to the Department may represent 100 other dissatisfied customers who were too timid to call the Department or unaware of their right to do so. A small number of complaints can indicate a serious customer service problem. However, not every customer who calls the Consumer Division is a disgruntled customer who has already tried to resolve a problem directly with the company. Some customers call the Consumer Division first and others call simply to get information. To the extent the Department = s guidelines include any measure based on Consumer Division statistics, the Department must insure that all calls are logged and categorized in a consistent manner.

Question 5 (regarding SAIDI/SAIFI definitions): UWUA proposes above that companies count outages that are one minute or longer when calculating SAIDI or SAIFI. The definition itself should not vary from company to company, although the numerical benchmark may need to vary (see question 1, above).

Question 6 (regarding including or excluding severe weather events): Major system events (including severe weather and loss of generation supply outside the control of the company) should generally not be included in calculating SAIDI or SAIFI. These major events must be strictly defined, as discussed in the comments above. The duration of an major outage, however, may well depend on a number of factor = s within the company = s control: prior inspection and maintenance of poles and lines, equipment inventory policies, staffing levels, emergency planning and system design. The Department must insure that the exclusion of major events does not mask deficiencies and shortcomings that are within the company = s control to fix. The Department must adopt inspection and maintenance guidelines and employee staffing benchmarks in order to protect customers from unduly prolonged outages. (See discussion of A Process Performance Standards, @ above.)

Question 6 (sic --2nd #6)(regarding outage measures other than SAIDI and SAIFI) and
Question 7 (regarding worst-performing circuits): UWUA proposes measures for momentary outages and for worst-performing circuits and also proposes customer guarantees for customers who lose power for an extended period of time, due to local distribution system equipment failure. (See discussion above). Customers are increasingly concerned about momentary outages because they disrupt or damage computers, timers, industrial process equipment, and a range of electronic devices. The standard for worst-performing circuits addresses problems that do not affect enough customers to significantly change the SAIFI or SAIDI numbers but which can cause substantial harm or loss to the customers affected. UWUA believes that most companies already collect data from which the worst-performing circuits can be identified. If the data is not available, the Department should order companies to collect it and numerical goals can be set at a later date.

Question 8 (regarding the method for establishing the proper benchmark): Whenever possible, benchmarks should be set based on average statewide performance. This will require all companies to move their performance up to the industry norm10 and provide appropriate incentives for lower-ranked companies to improve. In the short run, however, sufficient, comparable data may not be able to develop statewide averages. (For example, one company may track 20-second telephone response rates; another may track 30-second response rates; and a third company may not adequately track this data). In that event, the benchmarks should be based on each company = s historical data. The historical data cannot include post-1997 performance data if that performance is worse than existed in 1997 because of the mandates of G.L. c. 164, ' 1F(7)(the Department shall A require that quality and reliability are the same as or better than levels that exist on November 1, 1997. @ ) A benchmark cannot reflect any decline in service that has occurred since passage of the Restructuring Act as this is the very ill ' 1F(7) seeks to prevent.
Statewide averages would also be inappropriate for those measures which cannot reasonably be set on a uniform numerical basis. For example, some companies may have lower SAIFI and SAIDI targets due to their unique geography and distribution systems (see question 1, above). The numerical targets should then be based on the company = s own history. The Department, however, should revisit the question of which measures are set on a uniform numerical basis sometime in the year 2001 or early in 2002. Once data are collected on a uniform basis for a few years, the Department may conclude that standards which at first varied from company to company should in fact be uniform. For example, the Department may conclude, after reviewing a few years of consistently-reported SAIFI and SAIDI data, that any variations do not appear to correlate with geography, type of distribution system, or other company-unique factors and may in fact reflect different levels of commitment by management to minimizing outages.

Question 9 (regarding use of data from other industries to set benchmarks): The Department, the companies and all interested parties would benefit from knowing how other industries perform in terms of telephone response time, complaint resolution and other functions that are not unique to gas or electric supply. Many industries have dramatically improved their telephone response times and offer the types of customer guarantees (e.g, for missed appointment) that UWUA proposes above. UWUA, however, questions the extent to which this information is available, especially in a format that would allow for cross-industry comparisons. The Department may wish to direct companies to include this type of information in their PBR/quality of service plan filings.

Question 10 (regarding penalties/rewards): UWUA believes that penalties are appropriate, at the levels specified in G.L. c. 164, ' 1E (2 percent of transmission and distribution revenues), and are essential to create proper incentives for management. Companies should at least be allowed to propose rewards for performance that is clearly above the both that company = s own historical average performance and the industry average. This will avoid rewarding poor performing companies simply for moving towards average performance or rewarding a better-than-average company simply for maintaining past performance. The Department, however, should be cautious about approving rewards and should require the company to demonstrate that the benefits to customers of improved performance substantially outweigh the size of any reward.
Question 11 (regarding the allocation of penalty amounts): The dollar amount of penalties (or rewards) should vary with the benefit or harm of improving or declining service. Companies should be free to propose maximum penalty amounts that vary from measure to measure. For example, customers as a whole almost certainly value a decrease in outages more than they value a reduction in the number of billing adjustments, and the penalty amounts could vary accordingly.
UWUA questions whether the A total penalty @ should be A allocated @ at all. The statute precludes the Department from imposing a penalty in any year that exceeds 2 percent of revenues. If a company, at the outset, allocates this 2 percent cap across (for example) five different service quality measures, the maximum it will pay in any one category is .4 percent of revenues. The statute, however, does not compel this small of a cap. The Department could set the dollar amount of each penalty based on the perceived benefit or harm to customers of improving or declining performance in each service quality category, or based on the Department = s conclusion that a penalty of a particular amount creates the appropriate incentive for management. To the extent that the sum of all of the individual penalties might exceed the 2 percent cap, it is unlikely (and almost impossible) that a company = s performance in any year would decline in each and every category to the worst possible level. If this were to occur, the Department would simply reduce the total penalty to the 2 percent maximum.11

Question 12 (Linearity of the penalties): In general, UWUA believes that penalties (or rewards) should change more rapidly as performance degrades further from the target performance, but reserves its right to file reply comments on determining the level of performance that would result in the maximum penalty (or reward).

Question 13 (regarding rewards): See questions 10 to 12.

Question 14 (regarding staffing levels): For companies that make PBR filings, G.L. c. 164, ' 1E(b) requires the Department to set benchmarks for staffing levels based on November 1, 1997 staffing levels, unless either of the two exceptions contained in the statute applies. As discussed in its general comments above, staffing levels must be monitored by department or function if they are to provide meaningful benefits to customers.

I:\Clients\UWU.PBR\99-84\fincomments.uwu