Cost of Service Study & Cost of Service Study & Rate Design

March 27, 2014Jason N. Rauch, Ph.D.Maine Public Utilities Commission

Agenda

• Background theory and purpose of proper cost allocation

• Structure of Cost of Service Study– cost allocation and classification of system costs

• energy, demand, customer

• Embedded costs vs. Marginal costs

• Maine’s experience– Electricity supply market vs. delivery monopoly– Cost of service study information and mechanics– Process of assessing cost of service study– Issues arising in cost of service studies

• Practical considerations– Study results vs. actual rates 2

Bonbright rate design principles

• Dr. James C. Bonbright was on the faculty of the Columbia University School of Business from 1919 to 1960. He published the widely cited book ''Principles of Public Utilities” in 1961.

• Bonbright’s principles are often summarized as three objectives • Bonbright’s principles are often summarized as three objectives – Revenue requirement – Fair apportionment of production costs among customers – Optimal efficiency

• The Public Utility Regulatory Policy Act of 1978 (PURPA) added: – Conservation of energy by users – Efficient use of facilities and resources by utilities – Equitable rates to consumers

3

Bonbright rate design principles (1961)

• The related, practical attributes of simplicity, understandability, public acceptability, and feasibility of application

• Freedom from controversies as to proper interpretation • Effectiveness in yielding total revenue requirements under the fair return

standard standard • Revenue stability from year to year • Stability of the rates themselves, with a minimum of unexpected changes

seriously adverse to existing customers • Fairness of the specific rates in the apportionment of total costs of service

among the different customers • Avoidance of undue discrimination in rate relationships • Efficiency of the rate classes and rate blocks in discouraging wasteful use

of service while promoting all justified types and amounts of use: – a. In the control of the total amounts of service supplied by the company, and, – b. In the control of the relative uses of alternative types of service

4

Purpose of Cost of Service Study

• Which cost categories have to be recovered?- REVENUE REQUIREMENTS

-capital costs

-operation and maintenance costs-return on investment-return on investment

• What are the cost allocation rules and tariff components?- COST ALLOCATION

-embedded cost vs. marginal cost

- TARIFF DESIGN-cost causation

-rate stability

5

General principles for cost allocation

• Assets used individually shall be charged individually (in other words, assign costs to customers who cause a utility to incur them)

“Causality Principle“

• Assets used jointly shall be covered by contributions from all users

“Common Good Principle“

6

Cost allocation concepts - customer classes

• Customer classes defined by

– End-user type

– Size (kW)

– Voltage (one phase vs. three phase)

• Generally define residential, commercial, industrial classes, which may be separated into further classes based upon size and/or voltage

7

Cost allocation concepts – embedded vs. marginal cost

• How should the revenue requirements be allocated to network users?

– Based on embedded cost principle

also referred to as “average” or “allocated” cost

– Based on marginal cost principle

8

Cost allocation and tariff design

• Steps in a Fully Distributed Cost of Service Model

1. Allocation of revenue requirements to cost centers (Where did cost occur?)

2. Allocation of costs from cost centers to products / customer groups (Who should pay and how the individual network usage is determined?)

3. Tariff Design (What is the chargeable quantity and how many charges should be paid?)

9

Step 1 - Allocation of Revenue Requirement to cost centers

• Revenue Requirement = Sum of different cost categories

- Direct cost allocation to cost centers where cost - Direct cost allocation to cost centers where cost directly attributable

- Where cost no directly attributable: use of Allocation Factors

• Allocation Factors should reflect economic causality and can be based on:

– Gross or net asset values of assets allocated directly to cost centers– Number of employees allocated directly to cost centers

10

Step 2 - Allocation of costs to customers

• Demand dependent cost - allocation factors based on:

– Coincident Demand Peak

– Non – Coincident (Individual Demand Peak)– Non – Coincident (Individual Demand Peak)

– Mixed Approach

• Energy dependent cost - allocation based on energy consumption

• Customer dependent cost - allocation based on number of meters (application of weighting coefficient whenever appropriate)

11

Step 3 - Tariff design - possible tariff components

• · Demand charge ($/kW)

• · Energy charge ($/kWh)

• · Fixed charge ($)• · Fixed charge ($)

12

Efficient price = marginal cost

• Marginal cost is the cost to serve an additional unit –it may have little relation to embedded (average) costs

• Pricing above marginal cost sacrifices consumer surplus (difference between the value consumers would have gotten from consuming more, and what they would have paid for it).

• Pricing below marginal cost wastes resources (the cost of the excess energy exceeds its value to consumers).

13

Embedded Cost Pricing

• based on the existing total cost allocated to units (energy or demand) transported through networks.

• Advantages• Advantages

– costs easily identified

– revenue requirements coverage and financial viability of network service providers ensured

• Disadvantages

– does not create proper efficiency signals! (violation of cost causation principle)

14

Marginal Cost Pricing

• based on future cost necessary to transport one kW or kWh through networks.

• Advantages

– reflects the major economic rules and creates proper

efficiency signals

• Challenges– derivation of marginal costs can be controversial

– additional mechanisms to ensure coverage of revenue requirement might be necessary (e.g., Ramsey Pricing)

15

Marginal Pricing is Complex

• Marginal cost varies by hour (or more often), location, and voltage level of service.

• Marginal cost is uncertain, so prices need to be adjusted in real time to maximize efficiency.

• If prices vary by hour, price changes must be communicated to • If prices vary by hour, price changes must be communicated to consumers and hourly consumption must be recorded.

• Charging marginal cost doesn’t necessarily provide the right amount of revenue.

“Far better an approximate answer to the right question, which is often vague, than an exact answer to the wrong question, which can always be made precise.”-John W. Tukey

16

Rates that time vary according to marginal costs improve price signal (example)

Old rates Efficient marginal cost based rates

Energy charges $ per kWh $ per kWh

Summer 0.10

Peak 0.25

Off-peak 0.12

Winter 0.10

- Improved price signal provides more compensation to solar generation, incentivizing solar based upon electricity infrastructure cost causation and economic efficiency (instead of subsidies, which may also be relevant to account for external economic benefits, but are more difficult to quantify accurately). 17

Peak 0.10

Off-peak 0.05

Service charge $ per month $ per month

As exists or per study results 5.00 4.36

Minimum bill to match revenue requirement

15.00

Totals

Summer revenue per month 85.00 134.16Winter revenue per month 59.00 42.61

Average monthly revenue 65.50 65.50

Short Run Marginal Cost (SRMC) Pricing

• Determine future cost necessary to transport one kW or kWh through networks, given fixed capital stock (investments are not allowed ). Short Run Marginal Costs include network losses and network congestion cost.network congestion cost.

• Advantages– creates proper short term efficiency signals

• Disadvantages– does not address locational signals in the network infrastructure

– revenue raised by SRMC pricing not sufficient to cover cost– pure SRMC pricing could generate significant price fluctuations– allocation rules for congestion and transmission losses can be controversial– does not reflect that for transmission and distribution infrastructure, most costs are long-run

18

Long Run Marginal Cost (LRMC) Pricing

• determine future cost necessary to transport one kW or kWh through networks, given flexible capital stock (investments are allowed).

• Advantages– creates proper long term efficiency signals

• Challenges– revenue raised by LRMC pricing may not be sufficient to cover revenue requirement costs

– determination of LRMC can be controversial

19

Proper allocation of marginal costs minimizes cross-subsidies

• Cross-subsidies

– Distort customers’ decisions about consumption and investmentinvestment

– Reduce overall social economic welfare

– May be difficult to sustain if customers have supply choices

20

Maine Cost of Service Study Practice

• For generation, no cost of service study needed– Market sets prices (for energy and capacity), but…

• Capacity prices (kW) are available but not largely passed through to customersthrough to customers

• Challenge is to provide efficient kWh price signals

– Time-of-use pricing, critical peak pricing, real-time pricing

• For delivery, cost of service study follows the efficient cost causation principles of long-run marginal costs– In theory, only demand (kW) and fixed customer charges for

delivery-only transmission and distribution utilities that exist in Maine

21

Supply Rate Options (kW charges)

• ISO-NE market tracks customer capacity obligations via ICAP (Installed Capacity) tags

• The ICAP tag is the individual customer’s annual peak demand contribution • determined by the one hour in the year when the total grid demand is highest

• Large industrial customers often pay a separate capacity charge• Either through Competitive Electricity Provider (CEP) or Standard Offer (SO)

service.

• Medium and small commercial and residential customers are usually charged only in kWhs

• Capacity costs are rolled into the retail kWh prices charged by the CEP or SO service.

• Less efficient pricing signal, but realities of implementation (social and technological) have so far prevented passing through generation capacity charges

22

Supply Rate Options (KWh charges)

• Flat Rate

– Does not reflect time-varying nature of electricity supply costs

• Time of Use Rate

– Rate is time-varying to correlate to system peak, but the prices – Rate is time-varying to correlate to system peak, but the prices are static and not related to real-time system needs

• Dynamic Rate

– The rate is time-varying and adjusted in (near) real-time to address current system needs

• Critical peak pricing (CPP)

• Variable peak pricing (VPP)

• Real-time pricing (RTP)

23

Price Risk-Reward Trade-off for kWh Electric Rates

PTR: Peak-time rebatePTR: Peak-time rebate

TOU: Time-of Use

CPP: Critical Peak Pricing

VPP: Variable Peak Pricing

RTP: Real-time Pricing

24

Status of Maine Supply Pricing Options

• Competitive Electricity Suppliers– May provide dynamic rates to customers, but if so, are

presently only available to large customers

• Standard Offer (default supply service)– Default energy supply service procured by Maine PUC– Default energy supply service procured by Maine PUC– A Time-of-Use option is now available for Central Maine Power

(CMP) customers, leveraging capabilities of smart meters

• Challenges– While external infrastructure (smart meters) is installed to

provide for more real-time pricing options, the back-end billing systems (software) requires upgrading at CMP and EmeraMaine.

25

Delivery Cost Allocation(information needed)

• Demand-related costs– Transmission

• Lines• Substations

– Distribution– Distribution• Upstream Lines• Substations• Trunkline Feeder Primary Line

• Design demand-related costs– Distribution

• Local Primary Line• Line Transformer• Local Secondary Line

• Customer-related costs– Meters– Customer services

26

Marginal Cost Study

• Estimates cost of providing an additional unit of service

• Costs are forward looking– E.g., CMP’s latest marginal cost study has a 2012 test year

escalated to 2014

• Distribution substations (and above)– Utilize coincident peak demand allocation

• Local primary lines (and below)– Utilize non-coincident peak (individual peak) demand

allocation

27

Allocation of Delivery Costs for Efficient Rates

• Efficient rates would mirror the marginal costs of the utility infrastructure components, with charges set equal to marginal costs for each component

– Demand-related costs = time differentiated monthly demand charge

– Customer-related costs = fixed monthly customer charge

– Design demand-related costs allocation is more controversial

• Might be all allocated to fixed customer charge, all to a demand charge based upon annual peak demand, or might be parsed between the fixed customer charge and monthly demand charge

28

Need to adjusting marginal cost rates to satisfy revenue requirement

• Often exists a shortfall between rates set at marginal costs (MC) and the revenue requirement

– (adjust rates toward the fair-return price at f, the intersection between demand (D) and average total costs (ATC))

29

Approaches to rate adjustment

• Ramsey Pricing– Adjust marginal costs rates in proportion to customer class

economic elasticities

• High economic efficiency• High economic efficiency

• May be perceived as unfair

• Accurate data difficult to get

• Equiproportionality– Adjust each rate class equally

• Often perceived as having greater fairness

• Less economically efficient

30

Importance of accurate metering and data for cost of service study

• Accurate calculation of marginal costs requires information to estimate incremental costs of infrastructure components infrastructure components

• Varying rates require a meter that can record usage in each pricing period

• Accurate metering and billing provides the correct pricing signal so customers can make optimal decisions with regard to consumption and investment

31

Process of assessing cost of service study

• Utility submits cost of service study according to the tenants of Maine rate case precedents (long-term marginal cost analysis) in a rate case that involves rate redesign– not all rate cases include rate design

• PUC Staff and other intervening parties conduct discovery on • PUC Staff and other intervening parties conduct discovery on the study – All data and analysis pertinent to the study is provided or available for

review– May or may not involve expert consultants hired by the PUC and/or

intervening parties

• Staff and parties provide critiques and/or additional analysis based upon their review

• Utility is given a chance to review and accept or rebut critiques and analysis

32

Central Maine Power (CMP) Rate Case - Cost of Service Study Issues

• Time-varying nature of upstream distribution and transmission costs?

• For local distribution facilities (primary lines, line transformer, and secondary line (service drop)), are these more demand-related or fixed customer cost related?

• What is the proper method of deriving marginal costs? (Office of the Public Advocate vs. CMP)

33

Practical considerations – legacy (study results vs. actual rates)

• For a delivery-only utility, cost of service study shows rates should only be kW and customer charges

• In Maine, legacy of vertical integration of electric utilities has resulted in kWh charges remaining on tariff rates set to recover transmission and distribution costs.

– Maine transmission and distribution utilities divested generation assets around 2000, yet rate design still includes kWh charges in pending rate cases before the Maine PUC.

– Need to balance rate stability and customer familiarity with movement of rate toward optimal pricing signals.

34

Practical considerations – rate stability

• Smart meters now enable tracking of individual coincident and non-coincident factors and conducting demand charges– although still data management challenges– although still data management challenges

• Yet moving towards kW charges instead of kWh charges will invariably make some customers better off and others worse off based upon the individual customer usage characteristics– Challenge is how to move toward kW charges while

maintaining rate stability and customer equity

35

Practical considerations – customer familiarity

• Many smaller customers not familiar with demand charge concept

– How can just a few hours of use during the year dictate the – How can just a few hours of use during the year dictate the value of the demand charge? What about all those other hours?

– Perceptions of• Randomness

• Unfairness

• Additional cost

• Necessitates customer education

36

How do you implement efficient rates for all types of customers?

• Calculate marginal costs• Develop time-differentiated marginal cost prices• Adjust the marginal cost prices to hit the class revenue

target and achieve other ratemaking objectivestarget and achieve other ratemaking objectives• Incorporate real-time features as much as possible

(critical peak pricing, RTP, direct load controls)• Predict short-term consumer response• Reformat bills and educate consumers (and utility and

regulatory staff) about how to control their bills under the new rates.

• Monitor customer response (and its affect on marginal costs) and adjust rates as necessary.

37

Conclusions

• Efficient rate design should be the first step taken to improve economic and energy efficiency.

• It does not make sense to use expensive, subsidized command-and-control mechanisms to change electricity consumption if more efficient price signals can do most of the job, at lower cost.

38

Conclusions (continued)

• Regulated tariffs for delivery-only utilities should cover network costs

• Cost reflectiveness of prices crucial for proper economic signals (long-run marginal costs)

• Balance between normative economic principles and political / social factors is often necessary

39

Acknowledgements

• Thank you to Dr. Konstantin Petrov and Dr. Katja Keller of KEMA Consulting, Germany, Dr. Hethie Parmsano of NERA, USA and Dr. Ahmad Faruqui of The Brattle Group, USA for presentation materials.The Brattle Group, USA for presentation materials.

• Jason.Rauch@maine.gov

40

Questions?