insights for rate design using a retail gas utility’s rate filing 12 2-13

DRAFT 3 (12-3-13): NOT FOR ATTRIBUTION WITHOUT PRIOR WRITTEN CONSENT OF THE AUTHOR

Practical Insights for Existing Utility Retail Pricing and Revising Retail Pricing By

Robert J. Procter, Ph.D.

I. Paper Overview

Pricing (ratemaking) for a regulated retail utility provides a fascinating window

into the overlapping arenas of economics, public policy, and legal requirements. In this

paper, arguments are summarized that both support and provide insights into rate

regulation for the practitioner with little or no background in this subject matter.

Additional insights are gained into the overlap between economics, public policy and

legal requirements by examining how one utility rate proposal1 weaves together

arguments from these three broad areas to support its rate proposal.

This author has attended numerous meetings, workshops, and conferences

where some participants have little or no grasp of the most basic issues in this subject

matter and speak about the need to ‘de-regulate the markets’ to let innovation flourish. I

anticipate that the focus on rate setting as an critical element of regulation will continue

into the future as this industry continues to see advances in grid modernization,

movements to integrate products such as transactive energy, and increasing

telecommunications requirements to sustain and improve utility operations continue

advancing.

While the case filed by NWN is the ‘straw man’ for this paper, the issues raised

herein extend to both public utility commission review and deliberations of both

electric and gas utilities generally. Section I examines the role of the Commission as

that has been expressed in administrative rule, statute, and legal decisions. It also

contains an overview of selected actions by The Federal Energy Regulatory

1


Commission (FERC) that help provide insight into the role policy plays when

developing a rate design. Section II borrows aspects from regulatory economics and

the economic theory of the firm that form an analytical footing for the remainder of the

paper. This includes a short overview of the relationship between determining a rate

design and rate setting, once a rate design has been proposed by the utility. Section III

examines in more depth several issues that arise about rate design and rate setting using

material presented in Section II. Specifically, six explicit or implicit assumptions in

NWN’s testimony supporting its LRIC Study are critiqued. Finally, Section IV will

present a selected set of conclusions and recommendations.

I. Legislative Action and Court Rulings Define The Role of the Commission

How the legislature and courts have previously opined on the issues of fair and

just rates bears on the review of NWN’s approach to defending its rate proposal. This

short overview of legislative direction and court rulings helps to define how much

latitude NWN (or any utility rate making under Commission jurisdiction) has to define

both a proposed rate design and set of rates. It also helps to define the necessary scope

of testimony used to argue affirmatively in support of a proposed rate design and set of

rates.

Starting with the Commission’s Mission Statement, it describes its role as one

focused to "Ensure that safe and reliable utility services are provided to consumers at

just and reasonable rates”2[emphasis added]. This statement raises a question about

how one determines if a set of proposed rates meets this standard.3 Referring to Oregon

law, the “The legislature has authorized the PUC, in regulating the rates of public

utilities within its jurisdiction, to ‘make use of the jurisdiction and powers of the office

2


to protect [utility] customers, and the public generally, from unjust and unreasonable

exactions and practices and to obtain for them adequate service at fair and reasonable

rates. The commission shall balance the interests of the utility investor and the

consumer in establishing fair and reasonable rates.’”4 The justices go on to note that

“...rates are fair and reasonable for the purposes of this subsection if the rates provide

adequate revenue both for operating expenses of the public utility *** and for capital

costs of the utility, with a return to the equity holder that is: (a) Commensurate with the

return on investments in other enterprises having corresponding risks; and (b) Sufficient

to ensure confidence in the financial integrity of the utility, allowing the utility to

maintain its credit and attract capital.”5

The ruling goes on to note that since the Legislature’s direction to the

Commission is quite broad on this issue and the Court’s review of rates previously

established through Commission Orders is typically very limited. It also notes that

even if an aggrieved party files suit and the Court remands the Commission’s Order, the

Commission is free to review the case using its existing delegated authority, unless the

remand has specifically limited its discretion. Interestingly, the ruling also notes, “…as

to ratemaking by the [Oregon] PUC, the [Oregon] Supreme Court has noted that it is

the legality of the end result of the ratemaking process, and not the legality of each

calculation or input used during that process, that controls.”6 7 [Emphasis added]

How does one distinguish between unfair and unduly discriminatory, on the one

hand, from fair and reasonably discriminatory, on the other? In Oregon, that

determination lies primarily with the Commission. Oregon courts have acknowledged

that the Commission is acting within its legal authority if it balances the competing

3


objectives embedded within a set of rates, even if one or more aggrieved party believes

that those rates are unfair, unjust, and unduly discriminatory. This grants a good deal

of latitude to the utility to determine a set of rates as long as the utility is able to mount

a persuasive defense that their proposed rates, along with their proposed rate design, are

just and reasonable and not unduly discriminatory.

While The FERC has no authority over the setting of retail rates by the

Commission (or any other state’s utility commission), it has adopted various solutions

to the policy question of what amount of fixed cost should be recovered using a

customer charge versus what amount of fixed cost should be recovered using a

commodity charge.8 The FERC rightly notes that the issue of what amount of fixed cost

is recovered using a commodity charge has implications for the total cost of service for

various customer classes, and customers within a class. They also rightly note that the

total cost of service of a given customer will depend on that customer’s load factor. As

a result, the amount of fixed cost included in a commodity charge will have differential

impacts on various customers reflecting differential load factors between those

customers.9

The FERC first assigned all fixed costs to the commodity charge and over the

years shifted between all fixed costs being assigned to the customer charge to having a

portion of them included in the commodity charge and a portion assigned to the

customer charge. It’s interesting to note that these various approaches to allocating

fixed costs were sometimes supported on the basis of whether peak use or annual

consumption was paramount in planning, while at other times a particular allocation of

fixed costs between the variable and the fixed charges was justified on the basis of how

4


they supported a particular policy goal, such as increasing consumption. This short

overview of The FERC approaches illustrates that in the rate design and rate setting

policy arena, such factors as how the existing system is currently operating and what

policy goals are in play have been crucial considerations when allocating fixed costs

between a fixed charge and a commodity charge.

II. An Overview of Economic Theory for Rate Setting and Judging Fairness

Rate setting refers to determining the overall pattern of prices (rates) that are

contained in a utility’s rate schedules or tariffs. As such, that pattern also reflects the

utility’s cost allocation and rate design. Rate design includes what ‘billing factors’ are

used for residential versus industrial rates, for example, as well as the structure of rates

within one segment of the utility’s customer base. For example, are rates flat across

hours of the day, day of the week, week of the month, and/or month of the year? Will

there be a price levied that reflects the cost the utility faces when deliveries are at a

peak? Will there be a fixed charge per customer? What will be the pattern of charge

per unit of consumption, otherwise known as the volumetric charge? Rate design refers

to the process of translating costs that are allowed recovery from a given customer class

through rates into specific prices.10

Since utilities have relatively high fixed costs as a proportion of total costs, rate

design should include at least two components: a fixed component and a variable

component. The fixed component would recover some portion of the utility’s fixed

costs and the variable component would recover the utility’s variable costs and possibly

some portion of its fixed costs.

5


Often, there are multiple variable components. For example, for an electric

utility it’s not uncommon to find a fixed customer charge combined with a variable

charge for capacity, referred to as a demand charge, and a second variable charge for

energy, referred to as an energy charge. For a gas utility, it’s also not uncommon to

find a fixed customer charge per meter, and a commodity charge measured in dollars

per therm of gas delivered. There may also be a separate demand charge.

Now, turning to micro-economic theory (MET) among the management

decisions addressed include optimal pricing, organizing production, and optimal input

combinations. The figure below, titled ‘Monopoly,’ illustrates the economic goal of

regulation. Economic regulation of retail electric or gas utilities has two primary goals:

lower price from Pm to Pf and increase output from Qm to Qf. While the retail electric or

gas utility must stand ready to serve whatever amount of consumption its customers

desire at the prices contained in its rate tariffs, this diagram still illustrates these two

primary goals of economic regulation.11

6


Note that when a large portion of total costs are fixed costs, the utility under-

recovers revenues if it proposes rates by equating the proposed rate to marginal cost,

P=MC, (Pr in the graph). Yet, in MET, students are trained that the valid pricing and

production rule is to set production where P=MC, in both the short-run and the long

run. This result is illustrated in the graph where P=MC lies below the ‘fair return price’

of P=AC. Yet, in the case of the regulated firm, such a decision leads to a cost under-

recovery. This presents a dilemma to the regulator. Rates set using MC under-recover

costs while rates set using average costs (AC) send price signals that do not reflect costs

for any incremental change in consumption. Therefore, the regulator must either use a

separate charge to recover many or all of these fixed costs, and/or set rates using

average cost pricing (ACP). P=AC is often the solution to this revenue under-recovery

problem, and we will return to this issue in Section III.

One very important part of MET, and one that is a significant element in rate

setting in regulated industries, is the issue of cost causation. Assuming that a retail gas

utility sets rates using costs (rather than some form of market-based rates), how various

expenses are treated becomes crucial to assessing if the rates are just and reasonable.

At first blush it might appear that cost based rates constrain the options for rate setting,

and support a conclusion that cost based rates are inherently fair and just (since they are

cost based). However, we will see that the reality is quite a bit more complex. The

complexity is driven, in part, by how the following issues are resolved: (A)

Determining what costs are fixed and what costs are variable; (B) Determining what

fixed costs are included in the rate base; (C) Determining marginal costs; and (D)

Determining how to assign ‘joint’ costs.

7


Regarding ‘fair and just,’ Section I summarized existing legislative and legal

guidance on this issue, which at the retail level in Oregon, kicks the issue back to the

Commission for resolution. As will be discussed later, some of the testimony filed by

Commission staff argues that assigning costs to the party causing those costs achieves

the dual goals of efficiency and equity. This approach to defining equity will also be

addressed.12 We will also examine how it is that fair and just rates or prices, what in

economics is referred to as equity lies outside the scope of the MET

A. Determining what costs are fixed and what costs are variable.

By convention, fixed costs are those that are impossible to change in the

short-run and variable costs are those that can be changed ‘quickly’ in the short

run. By convention, MET defines the long run as that point in time when all

costs are variable. As a result of these conventions, the issue of what costs are

fixed and what costs are variable overlaps with the next one, how to determine

what is a short run versus a long run cost. How these two issues are resolved

will significantly affect the Long-Run Incremental Cost (LRIC) study and

therefore also affect cost allocation and rate design.

In MET, inputs to production and their associated costs are separated

into fixed and variable categories. The reality is it’s not as straightforward as

one might imagine based on the simple examples used in illustrating the

concepts of short run costs and long run costs. It isn’t as simple because there is

no clear dividing line between them. Each input used to produce and deliver a

kWh of electricity or therm of gas will likely have different lengths of time over

which they are fixed. For example, replacing one power pole or one line switch

8


due to failure of the equipment are examples of capital equipment that may be

treated as a variable cost and included in short run variable cost (SRVC).13

Some examples of fixed costs include the existing transmission and

distribution system pipes or power lines, gas compression stations, generating

plant, central office building, fixed maintenance and operating expenses, back

office support systems including mainframe computers and existing human

capital. Some examples of variable costs include fuel for the existing power

plant, gas in the existing pipelines, some maintenance and repair expenditures

associated with the existing system.

For a given rate filing, constructing the utility’s short run total cost

schedule (SRTC) is developed using the costs associated with the resources

used in production and delivery that are fixed in the short run (SRFC) plus those

variable costs (SRVC) that pertain to its rate filing. In contrast, the long-run

total cost schedule (LRTC) should be constructed using the costs of those inputs

used to produce and deliver a product that can only be changed in the long run.

B. How short-run costs are distinguished from long-run costs

Recall that on a close reading of economic theory, a business’s LRTC

are those costs that are associated with the amount of each input used to produce

and deliver the needed quantity of kWh of electricity or therms of gas in the

long run. What is tricky is identifying the long run. As was mentioned above,

there is no clear demarcation between the short run and the long run. We can

say that the long run is when all inputs are variable. Identifying that point in

time, or span of time, for an actual analysis isn’t straightforward. Some crafting

9


of a long run scenario is the usual solution. While future costs may be used for

LRTC, historical costs may also be used when they are believed to reflect

expected future costs.

One additional complication is how to deal with the utility’s historical

costs. Recall that in the long run, if all costs are variable, fixed costs, by

definition, are non-existent. Yet, in the context of rate setting for a retail

electric or gas utility, there are historical costs that must be recovered the rates.

For example, returning to the ‘Monopoly’ graph, the amount of the firm’s

historical costs, or what is also sometimes called sunk costs, is approximated by

the area (Pf – Pr)*Qf.14

C. Determining how to assign ‘joint’ costs

Joint costs can be a particularly challenging area when determining the

total cost of a given segment of the utility’s production and/or delivery process.

They often also pose a challenge when assigning costs to different products

and/or user groups, such as between different groups of customers (e.g.,

residential, large commercial, industrial, street & area lighting, etc.). Economic

theory recommends that these types of costs, as with any cost, be assigned to the

various functions and customers based on the extent to which that function

and/or group of customers give rise to that cost. This is a principle enunciated

in microeconomic theory irrespective of market structure. While that direction

sounds quite clear, putting it into operation can be another matter entirely,

which we will explore later in this paper.

10


D. Determining Marginal Cost

Before we can determine marginal cost, MC, we need to identify total

cost. Total cost can be determined separately for SRTC and for LRTC. A

simple form for LRTC associated with the production and delivery of a kWh of

electricity or a therm of gas is in formula (1) below,

(1) LRTC(Qi) = P1*X1 + P2*X2 + . . . + PN*XN

Where,LRTC(Qi) = Long Run Total Cost at output Qi,Pn = Price per unit of Input Xn used to produce output Qi,Qi = Output quantity i,Xn = Quantity of Input Xn required to produce Qi.

Long-run average and marginal costs are then derived using this LRTC

schedule. If long-run total cost is LRTC(Qi), then long-run average cost,

LRAC(Qi), equals LRTC(Qi)/Qi. In turn, long-run marginal cost, LRMC(Qi),

equals [d(LRTC(Qi)]/d[Qi]. Take note that output, Qi is in the denominator of

LRMC. There is no requirement that d[Qi] > 0. It can either be positive,

0<d[Qi], or negative, d[Qi]<0. This is an important observation because it opens

the possibility to determine MC, for either SRMC or LRMC, using decrements

in sales. Therefore, we can determine a LRMC for the last unit of output

without that last unit of output serving sales growth, as is required in NWN’s

testimony. Assuming a discontinuous cost curve, LRMC(Qi) is approximated

using Long Run Incremental Cost, LRIC,

(2) LRIC(Qi) = | [LRTC(Qi) - LRTC(Qi-j)]/[Qi –Qi-j ] |

11


where, Qi can either be > Qi-j or < Qi-j. In (2), LRIC(Qi) depends on the slope of

the LRTC curve, which depends on the slope of the business’ production and

delivery functions under fixed input prices.15

As we have just seen, load growth, either by adding customers or

increasing consumption per customer, is not a necessary condition to be able to

estimate LRIC(Qi).16 Additionally, for the business that produces multiple

products that are delivered to various customer groups, there essentially is a

corresponding LRTC, LRAC, and LRMC for each product and customer group

combination.17

Distinguishing short-run costs from long-run costs does not depend on

whether a power or gas transmission or distribution line is being installed to

meet new customers, or growth among existing customers, or replacing old or

outdated equipment, or making upgrades to existing equipment, or replacing

existing equipment due to wear and tear. Capital expenditures that lie outside

the scope of replacing a defective part (as was addressed above) should be

included in long-run total cost irrespective of whether they arise from expansion

of the existing infrastructure to meet growth or are needed simply due to wear

and tear to maintain service quality.18

III. Economic Theory and the NWN Initial Testimony

As we have seen, arriving at a set of rates is based on more than economic

theory. Section II summarized several key issues that help to determine a set of

rates once we have a rate design. When the analyst moves from theory that is

largely based on the principle of cost causation, and into the practice of developing

12


a set of rates, the goals and objectives of the business become quite significant.

While cost causation remains a guiding principle within MET, in practice cost

causation should be based largely on policy guidance from the executive committee

that should be informed by analysis. Let’s now turn to examining arguments made

in the NWN rate filing and compare those arguments to the framework laid out in

the previous sections.

There are six sub-sections below and each sub-section addresses a different

explicit or implicit assumption in NWN’s arguments. Each section’s title represents

one of the assumptions. Each section’s content is a critique of that assumption.

A. The MC Study does not reflect actual costs

Included in NWN’s initial filing was a LRIC study and supporting

testimony (Feingold testimony).19 Early in Feingold’s testimony, he argued,

“Marginal cost studies do not reflect actually incurred costs, but rely on

estimates of the expected changes in cost associated with changes in utility

service.”20 Recall that section II (D) illustrated that MC could be estimated

using the business’s existing LRTC data at a given level of existing output,

denoted Qi. Doing so reflects its current production and delivery capabilities.

As a result, requiring that MC be forward looking and based on estimated costs

unduly limits the scope of the utility’s MC studies.

MC calculated on future costs, are also useful in rates designed to help

limit consumption before design day requirements are reached by signaling to

customers the incremental cost of that added consumption. Pricing using a MC

calculated from the existing production and delivery functions does provide a

13


signal to the market about the cost of system expansion when the expansion

costs are reasonably approximated by the business’s existing LRTC.

B. The MC Study should exclude equipment repair and replacement costs

Feingold argues that including distribution mains replacement costs in

MC becomes relevant when “…new customers are added to the system…

[which] may increase design day requirements above…[what] existing facilities

can serve…”21 Feingold’s argument that there should be no cost included in the

LRIC associated with transmission and distribution (T&D) when consumption

by existing customers lies below peak delivery capability buries a policy issue

in the analytics of the LRIC study. Rather, existing customers should face some

cost associated with meeting peak system use in order to assure the existing

system is being used effectively. Existing customers should also confront costs

incurred to maintain service quality.

C. Uses must be in conflict for common (joint) costs to arise

Feingold argues, “Common costs occur when the fixed costs of

providing service to one or more classes, or the cost of providing multiple

products to the same class, use the same facilities and the use by one class

precludes the use by another class” [Emphasis added]. When a good or service

is bought and sold in a market, like electricity and gas are, common costs arise

when the use of a particular facility, say a distribution line, by one class (or one

customer within a class) does not preclude its use by another class (or another

customer in that same class). It is non-exclusivity in use that makes a given cost

a common cost, not exclusivity in use as Feingold argues.

14


D. The Stand Alone Cost test can be used to find cross-subsidies

Allocating joint product costs is a well-traveled road in regulatory

economics. As such, there is no reason to delve into the pros and cons of one

method versus another method any further. Rather, the overlap between joint

cost allocation and subsidy free pricing is one worth exploring. Feingold argues

that a subsidy free price is one where the price of the service exceeds MC but

lies below the stand-alone cost (SAC) of that service. In this writer’s

professional opinion it is highly unlikely that a panel of experts would find a

real-world rate schedule that is absolutely free of all cross-subsidization.

There is quite a bit of literature on using the SAC test for determining

the presence or absence of any cross-subsidies of either products or customers.

One such paper is titled, “Against the Stand-Alone-Cost Test in U.S. Freight

Rail Regulation.”22 Author Pittman calls the use of SAC into question with

railroads on two points: First, while the companies face a revenue adequacy

constraint, they are not constrained to earn zero economic profit. Second, other

businesses must able to enter the industry and offer all or a sub-set of services to

the railroad’s customers.

Regarding the first point, he says, “…once firm-wide economic profits

exceed the estimated cost of capital … that fact is (obviously) not the same as a

regulatory constraint on company profits.” In retail gas and electricity

regulation, while the companies are provided an opportunity to earn a specified

rate of return (ROR), they may earn a ROR above or below the allowed level

embedded in a set of rates. Earnings in excess of the allowed ROR are

15


economic profits. These quasi-rents will likely not attract entry at least in part

due to the regulatory impediments to entry and exit. Setting aside any legal

issues surrounding a third-party working directly with the utility’s existing

customers (which can be significant), entry and exit into the utility business is

anything but free.

Since the SAC methodology requires developing a proxy utility in order

to have a reference for cost comparisons, Heald argues that it is virtually

impossible to construct such a benchmark able to determine if a subsidy exists.23

Reviewing various methods to assess whether cross subsidy is an issue, Heald

argues against the use of SAC for several additional reasons above and beyond

what Pittman argues. While Pittman focused on the violation of key

assumptions that underlie SAC, Heald focuses on the complexity inherent in the

SAC methodology. First, the cost functions of the existing and alternative

technologies are required. Second, the SAC method is data intensive. Third,

asymmetric information between the existing business, regulators and potential

entrants makes accurate SAC testing virtually impossible. Fourth, comparing

each output of the existing firm to the possible cost of each product produced

separately by potential entrants can result in very different conclusions partly

depending on how rapidly technology is changing and the strength of economies

of scope and scale enjoyed by the incumbent firm.24

His conclusions rightly include the observation that cost allocation is

partly technical and partly political. More to the point, he argues that efforts to

find technical solutions to this problem of determining if subsidies exist, where

16


they are, and how large they are will only generate frustration. The crucial issue

for Heald is the challenge in developing comparable cost data. In his words,

“Without comparable cost data, the cross subsidy problem cannot be

satisfactorily addressed.”25

Jamison echoes Heald’s conclusion noting, “…it is infeasible for

regulators to establish subsidy-free prices with any degree of confidence.”26

Jamison also supports Heald arguing that to develop subsidy-free prices requires

the regulator to know “…the utility's cost function, its competitors' cost

functions, their competitors' cost functions, and so on until all combinations of

products which could have economies of joint production and that could be

affected by the utility's prices, have been considered.”27

In a separate paper, Ralph notes that Faulhaber (who authored a seminal

article on the topic of cross-subsidization) demonstrated that the test for

subsidy-free prices “…must be applied to all possible groupings of consumers

(or products) as well as to each individual consumer, since each individual may

cover their incremental costs, and yet some group of consumers may not…”28 It

is little wonder that both Jamison and Heald were less than sanguine about the

ability of a regulatory body to determine whether or not a set of proposed prices

contained any cross-subsidies.

Falhaber himself felt the need to weigh in on the cross-subsidy debate

with a short paper, which he begins by noting a tendency for analysts and

researchers to incorrectly apply some of his principles that underlie his approach

to testing for cross-subsidization. After presenting a simple example, he

17


reiterates a crucial conclusion from his 1975 paper that, “both the SAC and the

IC [incremental cost] tests must be applied not only to each service individually,

but to all possible groups of services.”29 He then hammers home the point that

applying these tests to individual services in isolation, which he notes has

tended to be the case in the regulatory arena, is a fatal error and cannot be

considered a reasonable approximation, or ‘good enough’ approach.

E. Economic theory requires that fixed costs be recovered using a fixed charge

Feingold turns to MET to find justification for his argument that all

fixed costs should be recovered using the fixed customer charge and all variable

costs should be recovered in the commodity charge. What is rarely noted is that

when setting P=MC, results in P=MC=AC, in equilibrium. This means that the

MET supports the recovery of fixed cost using a price that captures both fixed

and variable costs.30 This conclusion is the opposite of what Feingold argues.

What this means is that in the theory of monopolistic markets, setting the price

of a good to include some amount of fixed costs leads us to set a price where the

MR=AC, which is completely within the scope of MET. Finally, recall that

Section I contained a short review of various approaches to fixed cost allocation

by the FERC that illustrated the role policy objectives have played in their

various solutions to allocating fixed cost between a fixed versus a variable

charge.

F. When no cross-subsidy exists, the rates are just and reasonable

We have previously seen that “just and reasonable rates” is an important

consideration in rate design and rate setting. In Section I, it was noted that in

Oregon determining what rate design and set of rates are just and reasonable

18


rests almost exclusively with the Commission. We also saw that the FERC

allows rates to be discriminatory as long as they are not unduly discriminatory.

It was also noted that the load shape of different customers that face the same

tariff is a critical factor when assessing how a set of proposed rates actually

affects those customers. Further, Feingold implicitly argues that rates should be

considered just and reasonable when costs have been assigned to the product

and group causing those costs. When this standard is met, it also means that

there is no cross-subsidy.

Let us consider two competing rate designs in light of the important

standard of just and reasonable. The two alternatives are crafted to highlight the

inherent subjectivity of the just and reasonable standard. For simplicity, assume

there are the following two competing rate designs,31

Rate Design One: 100 percent of allowable fixed costs are recovered using

the customer charge.

Rate Design Two: 50 percent of the allowable fixed costs is recovered using

the customer charge. The remaining 50 percent of

allowable fixed costs are added to all allowable variable

costs and recovered using the variable (commodity)

charge.

Also assume there are two customer groups, A and B. Group A owns a 1,000

sq. ft. house, and Group B owns a 3,000 sq. ft. house. We do not need to know

the level of the fixed and variable prices to make several inferences about their

impact on existing customers from these two competing rate designs. When

19


Rate Design One (RD1) is compared to Rate Design Two (RD2), RD1 (a)

implicitly subsidizes the gas use of customer group B with the 3,000 sq. ft.

house at the expense of customer group A with the 1,000 sq. ft. house;32 (b)

encourages greater gas usage; (c) reduces the incentive for Energy Efficiency,33

and (d) reduces the utility’s near term revenue recovery risk. Even if there is

agreement about cost allocation, a reasonable observer may argue that RD1

discriminates against Group A since it subsidizes the usage of gas for space

heating by Group B. Conversely, Group B might argue that RD1 is unfair since

it penalizes them relative to RD2 simply because they can afford to purchase a

larger home.34

Fairness here goes beyond any debate about whether the resulting rates

are or are not subsidy free. They may be subsidy free and still be argued to be

unfair. Customer group A is more likely to be represented by a consumer

advocacy group that may also argue that RD1 is discriminatory because it

results in an equivalent fixed charge to each residential customer even though

they have very different degrees of ability to pay. It will also likely be argued

that since RD1 encourages greater consumption, future fixed cost will likely be

higher than would be the case under RD2 due to that higher usage, and Group A

should not be penalized by Group B’s choices. If it can be shown that Group

B’s usage led to higher costs on a per unit basis, Group A would likely also

argue they are also being penalized due to the profligate behavior of Group B.

While these arguments are potentially endless, they point to likely

arguments claiming undue discrimination that go beyond debates about whether

20


or not a set of prices are subsidy free. Ultimately, these debates are often

resolved either in a settlement acceptable to most or all parties to the rate case,

or by the Commission in the absence of such a settlement. Arguments about the

inherent fairness of one of these two rate designs, i.e., the absence of undue

discrimination, are positional, not factual.

Regarding the fuel choice decision (either in new construction or in

existing construction where a furnace replacement is contemplated), as the fixed

charge fraction of the average monthly bill increases, the incentive to stay with

gas heating or install gas furnace in new construction will decline, assuming all

else remains constant. These are several reasons why the rate design should be

a matter of business and public policy. As we have seen, for a given level of

costs to be recovered in rates, varying the rate design will affect customers

differently and will send differing signals to the market.

Oregon PUC Staff initial testimony sponsored by Dr. George Compton

(Staff) directly addresses the issue of fairness. Staff does note that when cost

causation remains murky, the method to use that retains fairness requires

assigning those costs to the different products and customer classes based on

benefits received.35 This method too represents a judgment that society prefers

this assignment rather than other possible assignments.

While such a rule may at first blush appear objective, it is not. There is

nothing about such a rule that is inherently more or less fair than a competing

rule that assigns costs proportional to customer’s ability to pay. One example of

21


such an assignment might be a lifeline residential utility rate using customer

income data to qualify for such a rate.

IV. Conclusions

Considering the significance of the issues raised herein, the analysis

contained in this article, and its conclusions should be relevant to other

regulated utilities, commissions, and interveners in states other than Oregon.

Further, this paper also illustrates the complexity involved in addressing

proposals to alter the existing rate regulation. No doubt, some will point to this

complexity as prima facia evidence for the need for reform. That is in the eye

of the beholder. It is my hope that expositions such as are contained in this

paper will at least help elevate those discussions. Finally, there are a number of

arguments made in NWN’s initial testimony that are called into question in this

paper. A series of specific conclusions follow.

1. Even though this paper examined a narrow set of specific issues that affect

debates on a rate proposal, it will hopefully help illuminate the complex

issues involved with proposals to alter retail rate regulation. Advocates of

retail de-regulation should familiarize themselves with these types of basics

of existing regulation in order to better argue their proposals.

2. Policy choices, such as, how fixed costs are recovered, should be explicitly

identified as policy choices rather than attempting to frame them as

analytical issues that have a technical solution. Identifying them as policy

choices should be accompanied by explicit analysis that illustrates the

ramifications of each policy alternative.

22


3. If a rate proposal is submitted that requests recovering fixed cost in a fixed

charge, that proposal should be analyzed in comparison to other possible

approaches to recovering fixed costs. Additionally, adjusting the utility’s

allowed rate of return to reflect differing amounts of revenue recovery risk

should be part of that evaluation.

4. Fairness is a moral issue not an analytic issue. As such, it is important to

address such issues as who decides what rates are fair and identify what

rule(s) is (are) used to make this determination.

5. Reducing the commodity charge and increasing the customer charge does

not result in greater conservation incentives.36 Economic theory supports a

charge that reflects the costs that arise at the point in time when the decision

has to be made. If a city is considering charging for parking, economic

theory supports charging per use rather than offering a monthly pass.

Offering a monthly pass (a fixed payment) encourages greater use of

parking than will be the case with a fee paid per visit for a specified length

of time, all else held constant.

6. The marginal costs (MC) in a LRIC Study may be forward looking, though

they need not be forward looking. They may also be based on historical

data from existing production and delivery technologies. This allows

incremental costs may be determined even when there is no expansion in

deliveries. Identifying a MC in the absence of sales growth is one important

factor in signaling the costs of such growth to the market prior to its

occurrence.

23


7. Allocating common costs is more art than science. No objective criteria

exist to use to reach an unimpeachable allocation of common costs between

one or more customers, or customer groups, and/or products.

8. The stand alone cost test cannot be relied on to identify cross subsidies.

Further, assuming that all parties were to agree that no cross subsidy exists,

rate case parties can be expected to have differing positions about rate

fairness.

9. Including replacement costs in LRIC study is important since those costs are

incurred to sustain service quality.

24


End Notes

25

1 UG 221NORTHWEST NATURAL GAS COMPANY NW NATURAL’S Application for a General Rate Revision, December 11, 2011.2 Oregon Public Utility Commission, Oregon.gov, http://www.puc.state.or.us/Pages/about_us.aspx3 What follows is based largely on “GEARHART v. PUBLIC UTILITY COMMISSION OF OREGON,” Frank GEARHART; Patricia Morgan; Kafoury Brothers, Inc.; and Utility Reform Project, Petitioners, v. PUBLIC UTILITY COMMISSION OF OREGON and Portland General Electric Company, Respondents.08487; 09093; A140317. Argued and submitted Feb. 03, 2012. -- February 06, 2013. See: http://www.publications.ojd.state.or.us/docs/A140317.pdf4 Ibid, at 85.5 Ibid.6 Ibid., at 94.7 This ruling by the Oregon Supreme Court appears to follow the same ruling previously made by the U.S. Supreme Court in a 1944 case, FPC versus Hope Natural Gas Company. See: Deloitte Center for Energy Solutions, “Regulated utilities manual A service for regulated utilities,” February, 2004, p. 7.8 A concise summary of the various approaches and rationales appears in the previously noted FERC Cost-of-Service Manual, pp. 30-33.9 Ibid, p. 29.10 The term ‘regulatory body’ is meant to be inclusive. It applies to state utility commissions with jurisdiction over investor-owned utilities (IOUs) as well as regulation of consumer-owned utilities (COUs) by their customers, or board of directors, or city councils, and so forth. While there are some differences in some details of determining the amount of cost to be used to develop the rates, the alternatives discussed in this article also apply to COUs. 11 If the industry was competitive, price would be even lower and output even higher, but that isn’t relevant to the case of the retail electric or gas utility.12 Within economic theory, one must turn to a specialized body of research and writing called Welfare Theory to find a significant body of work that addresses the issue of fairness from a broader, societal, perspective. Outside the narrow confines of economic theory, writings on issues of fairness abound. Judging fairness involves making a moral judgment. Determining that one or another set of rules or set of outcomes is or is not fair requires that I make a moral judgment. For example, some citizens find the growing income and wealth disparity in the U.S. morally fair while other find it morally repugnant. While economists have involved themselves in assessing economic impacts of such disparity, and have been involved in research on social and political as well as economic aspects of income and wealth disparity, reaching any conclusion about fairness, equity, requires making a moral judgment. This is equally true for rate setting in regulated industries. 13 The existing power lines and gas transmission lines are typically examples of short run fixed cost (SRFC) and the cost to replace a power pole damaged in an auto accident, for example, can be treated as a short run variable cost (SRVC). 14 Ideally, we would estimate FC by the difference between SRTC and SRVC at the optimal level of output. Using the formula specified in a reasonable approximation when variable costs are not provided.15 Why this is important will become clear later in this paper.16 It is also not a sufficient condition because it may be possible to meet that growth with the existing fixed inputs and therefore no incremental cost arises from that growth. However, if existing resources are inadequate to meet load growth, and new investments are required whose life extends beyond the proposed rate period (for example), those new investments should be included in long-run total cost. 17 In this case, the LRTC for the business will be the sum of all LRTC schedules for each combination of product and customer group. 18 Note that investments due to wear and tear are made to avoid a decrement in Qi. Therefore, in this case, d[Qi]<0.

http://www.puc.state.or.us/Pages/about_us.aspx

19 Exhibit 1100 – Direct Testimony – Long-Run Incremental Cost Study / Rate Design, Direct Testimony of Russell Feingold, NWN/100.20 Ibid, pp. 5 - 6.21 Ibid, pg. 6.22 Russell Pittman ,”Against the Stand-Alone-Cost Test in U.S. Freight Rail Regulation,” EAG 10-1 CA April 2010.23 David Heald, “Contrasting Approaches to the ‘Problem’ of Cross subsidy,” Management Accounting Research, 1996, pp. 53-72.

24 Ibid, p. 58.25 Ibid, p. 69.26 Mark A. Jamison, “Theory and Application of Subsidy-Free Prices,” from Industry Structure and Pricing: The New Rivalry in Infrastructure, Kluwer Academic Publishers, 1999, p. 140.27 Ibid.28 Eric Ralph, “Cross-subsidy: A Novice’s Guide to the Arcane,” Duke University, July 27, 1992, p. 15.29 Gerald R. Faulhaber, “Cross-Subsidy Analysis With More Than Two Services,” The Journal of Competition Law & Economics, August, 11, 2002, p. 442.30 From the standpoint of cost recovery risk, cost recovery risk is reduced using the method proposed by Feingold. However, that argument does not appear in his direct testimony. 31 Several other simplifying assumptions are: The utility servers only one state and has no unregulated entities; Agreement has been reached about how both allowable fixed and variable costs have be allocated between the two customer groups; There are only two products, the capability to deliver gas represented by the billing factor used for the fixed charge, and the delivery of gas represented by the billing factor for the variable charge; Group A has a peakier load factor than Group B; All gas delivers are to homes; Residences are grouped into two groups with an average residence size of either 1,000 sq. ft. (Group A) or 3,000 sq. ft. (Group B); All residences are built to the same energy code specifications and are alike in every other respect that affects gas consumed for space heating; and, gas is only used for space heating. 32 The term ‘subsidizes’ is used to denote one argument that rate case parties will argue that goes beyond the more rarified debate about subsidy free pricing presented earlier.33 ‘Conservation’ and ‘Energy Efficiency’ are two different factors that affect consumption.34 The point of this exercise was to provide a simple example of how the assignment of fixed cost between a fixed versus a commodity charge will likely have differential impacts on different customers within a given customer group and/or between customer groups. 35 Staff testimony Compton/15-Compton/16.36 See: Feingold, pp. 63-64.

insights for rate design using a retail gas utility’s rate filing 12 2-13

Business