20161216Morrison Direct with Exhibits 108-111.pdf

BEFORE THE ~. r·· r EI\! D f\\: lri...1 --
2C:S /EC 16 Ph I: 36
) IN THE MATTER OF INTERMOUNTAIN
GAS COMPANY'S APPLICATION TO
CHANGE ITS RATES AND CHARGES
FOR NATURAL GAS SERVICE.
) CASE NO. INT-G-16-02
)
)
)
___________ )
DIRECT TESTIMONY OF MICHAEL MORRISON
IDAHO PUBLIC UTILITIES COMMISSION
DECEMBER 16, 2016
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Q. Please state your name and address for the
record.
A. My name is Mike Morrison. My business address
is 472 West Washington Street, Boise, Idaho.
Q.
A.
By whom are you employed and in what capacity?
I am employed by the Idaho Public Utilities
Commission (Commission) as a Staff Engineer.
Q. Please give a brief description of your
educational background and experience.
A. I received a Bachelor of Science degree in
Chemical Engineering from the University of Southern
California in 1983, a Master of Science degree in
Mechanical Engineering from the University of Idaho in
2002, and a Doctor of Philosophy in Geophysics from Boise
State University in 2014. I have been a registered
professional engineer in Idaho since 1998. I have
completed graduate level courses in statistical methods,
experimental design, and mathematical inversion.
attended the Electrical Utility Basic Practical
Regulatory Program offered by New Mexico State
University's Center for Public Utilities.
I
Between 1988 and 2009, I held a number of
engineering positions at Micron Technology, Inc.
seven years, I was responsible for implementing
For
statistical methods and systems throughout the Company,
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
1
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
and I developed and taught numerous courses in industrial
statistics, data analysis, multiple regression,
experimental design, reliability modeling, and
statistical process control.
I began work at the Idaho Public Utilities
Commission in 2014.
Q. What is the purpose of your testimony?
A. I will address the Company's cost-of-service
methodology, and the Company's proposed base rate revenue
requirement allocation among its customer classes.
will also address the Company's adjustments to its
billing determinants, including its proposed weather
I
normalization adjustments and adjustments to customer
counts. I will conclude my testimony with a discussion
of the impact that the Company's proposed Rate-of-Return
will have on its line extension policies.
Please summarize your testimony. Q.
A. I will begin with a discussion of the Company's
cost-of-service methodology: The Company did not conduct
a load study, excluded two rate classes, and is proposing
cost allocators that are both novel and inappropriate.
The Company's methodology will not result in a fair
allocation of the Company's revenue requirement among its
rate classes and, absent a load study, it is not possible
to develop a suitable alternative revenue allocation
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
2
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
methodology. I propose that the Company, Commission
Staff, and the Company's stakeholders hold workshops in
order to develop a suitable load study and cost-of­
service methodology. Until a satisfactory cost-of­
service study is completed, I propose that the Company's
revenue requirement be allocated in proportion to the
normalized revenue currently being collected from each
rate class.
I will also propose an alternative to the
Company's weather normalization adjustments. The
Company's adjustments were obtained using selected
coefficients from a model that is unstable. My
adjustments, on the other hand, are based on a whole
model that is much more robust than the Company's model.
Additionally, my analysis of the Company's GS-1
subclasses is more detailed than the Company's.
I will conclude by discussing the relationship
between the Company's Section C service line and mains
extension policies and its rate base. These policies
were last updated over 30 years ago, and currently
specify a 12.5% Rate-of-Return on the Company's
investment in line and mains extensions. I will
recommend that the Company's service line and mains
extension policies be updated to reflect the Rate-of­
Return authorized by the Commission.
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
3
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
THE COMPANY'S COST-OF-SERVICE STUDY
Q.
A.
What is the purpose of a cost-of-service study?
A cost-of-service study allocates the Company's
revenue requirement to the Company's rate classes in
accordance with the principle of cost causation. The
cost causation principle states that costs should be
borne by the class that causes them to be incurred.
Costs incurred in the service of a single class, or its
individual members, should be allocated directly to that
class; however, because much of the Company's plant
serves multiple classes, a cost-of-service study is
necessary to determine the fraction of costs that are
caused by each class.
Q. What is a load study, and why is it a necessary
component of a cost-of-service study?
A. A load study determines peak usage, by class,
of system components that cannot be directly allocated.
This information is used to develop allocators that are
required by the cost-of-service study in order to
allocate shared plant costs, O&M costs, and some
administrative costs, to each class.
In general, plant equipment is designed to meet
the maximum load that will be placed on individual pieces
of plant equipment, so costs are caused by the need to
meet system peak. An obvious example is a transmission
CASE NO. INT-G-16-02
12 /16 /16
MORRISON, M. (Di)
STAFF
4
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
main, which provides gas to all customers in an area. A
load study is required to determine the fraction of the
main that is being used by each class at system peak.
Without such peak consumption information, it isn't
possible to allocate these costs fairly in a manner that
is consistent with the principle of cost causation.
Q. Which allocators are derived from data obtained
by a load study and how are they used?
A. Two general types of allocators are developed
using load study data: Coincident peak (CP) allocators
and non-coincident peak (NCP) allocators. Coincident
peak allocators are calculated using each class' share of
system peak demand, and are used to allocate plant that
is used simultaneously by multiple classes. Coincident
peak allocators are most often used to allocate the costs
of transmission and storage facilities.
By definition, the system will only experience
a single annual peak day, and it is possible for test
year system peak to occur on an extra-ordinary day that
does not truly represent how each class causes the
Company to incur costs. To avoid a potentially unfair
allocation of transmission and storage costs, CP
allocators are usually determined using data from
multiple peak days. A common CP allocator, the 3CP
allocator, is calculated using information from the
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
5
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
system's three highest monthly peak days.
Non-Coincident Peak allocators are used to
allocate distribution plant, with the rationale that
large segments of the distribution plant are designed
primarily to meet the loads of particular classes. For
example, distribution plant serving residential areas
will supply gas primarily to residential customers, and
distribution plant serving an industrial park will
probably serve a combination of commercial and industrial
needs. Because these local components of the
distribution system no longer serve all customer classes,
it would be improper to allocate these costs based on
their share of system peak demand. Instead, distribution
plant is usually allocated using a NCP allocator computed
using each class's own peak.
When usage is mixed in large portions of the
distribution system, it may also be appropriate to use a
peak and average allocator. Such an allocator is
particularly appropriate for large distribution mains,
which often serve diverse needs. As its name suggests,
the peak and average allocator is created by combining an
allocator based on each class's contribution to system
peak with an allocator based on each class's share of
average system consumption. Peaks and averages are
usually weighted using system load factor.
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
6
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Q. Please describe the Company's load study, and
the allocators that it uses to assign transmission and
storage costs to its rate classes.
A. The Company did not conduct a load study.
Instead, the Company devised allocators from monthly
billing data. Somewhat misleadingly, the Company refers
to these as "Peak Day Allocators."
To be clear, the Company created these
allocators by combining peak day information from its
approximately 150 industrial and transportation
customers, who are equipped with meters capable of
recording daily demand, with the monthly billing
information from its approximately 340,000 Residential
and general service customers who are not so equipped.
To create its allocator, the Company subtracted the peak
usage of its industrial and transportation customers from
its measured system peak, and then allocated the
remaining quantity in proportion to the remaining class'
billed consumption. Thus, for the Company's three core
classes, allocation is effectively based on average
January consumption rather than peak day consumption.
Without the results of a load study, it is not
possible to accurately assess the Company's proposed
allocator; however, it is likely that the Company's
allocator, which combines actual January 1st peaks of some
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
7
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
classes with the average daily January consumption of
other classes, will unfairly allocate costs.
I also note that the Company assumed a peak day
of January 1, 2016. This is a holiday, and it is quite
possible that some of the Company's industrial and
transportation customers were not operating at full
capacity on this day, so that plant-related costs could
be unfairly shifted to the Company's residential and
general service customers. As I explained earlier, the
Company's test year consists of actual data for January
through June, and forecast data for the months July
through December. In previous years, the Company's
system peak has occurred on various days in December and
January, so it is not actually possible for the Company
to know that January 1, 2016 will be its test year peak
day.
The Company is replacing its current Encoder
Receiver Transmitter (ERT) meters with ERT meters capable
of recording hourly consumption information from each of
its customers. A relatively small number of these meters
could be used to obtain the peak information needed to
develop accurate CP and NCP allocators.
Q. What are customer-related costs, and how are
they typically classified and allocated?
A. Customer-related costs are costs that increase
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
8
1
2
3
4
5
6
7
8
9
10
11
12
13
14
nee 6
16
17
18
19
20
21
22
23
24
25
incrementally with the addition of each new customer, and
that are not directly related to the increased
consumption or demand that the customer might place on
the system. Such costs are almost always identifiable
with a particular customer, so it is appropriate to
directly allocate Customer-related costs to that
customer's class. The costs of meters are typically
classified as customer-related, because they are clearly
identifiable with individual customers, and because the
Company incurs these costs whether or not the customer
consumes any gas.
Q. What are distribution services, and how does
the Company propose allocating the costs of distribution
services and other related costs?
A. Distribution services are the pipes used to
connect customers to the Company's distribution mains.
Because each distribution service serves one customer,
distribution services are properly classified as
customer-related costs, and should be directly allocated
to that customer's class. Nearly one-quarter of the
Company's rate base, or $149,255,628, is included in
Federal Energy Regulatory Commission (FERC) Account 380,
distribution services. Another $83,527,017 of the
Company's rate base is included in accounts related to
expenses incurred connecting customers to its system,
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
9
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
e.g. meters, regulators, ERT devices, and equipment
installations (FERC Accounts 381 through 385).
I agree with the Company's proposal to classify
its distribution services and related accounts as
customer-related costs, but disagree with its method for
allocating them among its customer classes. Rather than
directly allocating these costs to the appropriate
classes, the Company proposes allocating these costs
using a weighting scheme that is based on the costs of
meters used by each class. The Company refers to this as
its "Meters Study."
Why is this methodology inappropriate? Q.
A. Of course, it would be best for the Company to
allocate these costs directly. Unfortunately, the
Company has not maintained the records that are necessary
for direct allocation. The Company explained in its
response to Production Request No . 202 that
"Intermountain's accounting records have never provided
the functionality to track plant and accumulated
depreciation by rate class for FERC accounts 380, 381,
382, 383, 384, and 385." See Exhibit No. 108.
We note that meter costs vary widely, with the
meters used by some GS-1 customers costing more than 16
times as much as baseline residential meters, and the
meters used by T-5 customers costing almost 38 times as
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
10
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
much as baseline residential meters. The Company's
"Meters Study" is an appropriate methodology for
allocating meter costs, but the Company provided no
evidence supporting the notion that meter cost is
indicative of the costs of providing a pipe from the
Company's distribution main to a customer.
Absent the necessary cost accounting
information, the next best methodology for allocating the
costs of distribution services would be to use an
allocator based on the relative costs of installing
distribution services for each class. Similarly, we
would expect the costs of regulators to be allocated in
proportion to the costs of regulators used by each class,
for the costs of ERT devices to be allocated in
proportion to the costs of ERT devices used by each
class, and so-on. The information required to create
these allocators is readily available through the system
that the Company uses to estimate line extension costs .
Q. Please explain the Company's proposal to
classify a portion of its distribution mains (FERC
Account 376) as customer-related costs.
A. Because it has not conducted a load study, the
Company does not have the class peak information that is
typically used to determine appropriate allocators.
Instead, the Company proposes classifying 47.16% of its
CASE NO. INT-G-16-02
12 /16/16
MORRISON, M. (Di)
STAFF
11
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
$164,694,644 distribution mains rate base as customer­
related, and the remaining 52.84% as demand-related. The
Company's rationale for its Customer/Demand
classification is that its distribution mains serve two
functions: Connecting customers to the Company's gas
supply system, and supplying customers at peak.
Therefore, the Company argues, a portion of the costs of
its distribution mains should be allocated in proportion
to the number of customers in each class.
Q. Why do you disagree with the Company's proposal
to classify a portion of its distribution mains plant as
customer-related?
A. As discussed earlier, classification of
customer-related plant costs is generally limited to
those incremental costs that can be identified with
individual customers. By definition, mains serve
multiple users, and so it is not appropriate to classify
any portion of the mains as customer-related costs.
Also, as discussed earlier, the costs of connecting
customers to the distribution system is already captured
in the $149,255,628 of its distribution services that are
classified as customer-related costs.
Under the Company's allocation method,
$310,452,639 of the Company's $596,065,557 plant-in­
service (52%) would be classified as customer-related,
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
12
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
and not based on allocators that accurately reflect the
way that each class causes the Company to incur costs.
Q. Do you have any other concerns with the
Company's proposal to classify a portion of its
distribution mains as customer-related costs?
A. Yes. Even if I accepted the premise that a
portion of distribution mains should be classified as
customer-related, the Company's methodology for
determining the customer-related portion is incorrect.
The Company used the minimum intercept method to
determine its 47.16%/52.84% split. In this method, the
Company used regression modeling to determine the cost of
a hypothetical, zero-capacity system. The costs of this
system were then classified as customer-related. The
balance of the system's actual costs were then classified
as demand.
In order to determine the cost of a
hypothetical, zero-capacity system, it would be necessary
to use capacity as a modeling factor; however, the
Company used nominal pipe diameter, rather than capacity,
in its regression model. Nominal pipe diameter is a poor
proxy for capacity: For a given operating pressure, the
square of pipe diameter (cross-sectional area) is a much
better measure of the pipe's carrying capacity than
nominal diameter. Furthermore, the proper regression
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
13
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
methodology requires that that each pipe type be weighted
by the number of feet. The Company used no weighting
method, so the Company 1 s 137,841 feet of nominal 1.25
inch diameter steel pipe receives as much weight as its
6,612,833 feet of nominal 2.00 inch diameter steel pipe.
Another concern is related to the quality of
data used by the Company to develop its regression
models. Determining a zero system cost requires project
level information for all data used in the analysis. The
Company explained that in 2013, it adopted a new IT
System, and that project level electronic data was not
available for prior years. Nevertheless, the Company
used aggregated annual data for the years 1959 through
2015 in its analysis. Thus, the Company 1 s analysis was
not able to distinguish between mains that are still in
service, those that have been fully depreciated, and
those that have been retired. A substantial portion of
the data included in the Company 1 s analysis represents
the valuation of systems acquired by the Company, rather
than the actual costs of installing this pipe. For
example, the Company explained that most, or possibly
all, of the dollar value of 3.5 million feet acquired in
1959 is based on the book values of assets acquired when
the Company was formed. This mixture of actual costs and
book valuation is not appropriate for determining the
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
14
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
cost of a zero-capacity system. Unfortunately, the
Company lacks sufficiently detailed data to enable the
Company to properly perform the sort of analysis and
classification that it proposes.
Q. Please explain the use of plant-in-service in
the Company's cost-of-service methodology.
A. As noted earlier, the Company does not maintain
the records necessary to determine net plant-in-service,
either by account or by customer class, so the Company's
cost-of-service model allocates gross, rather than net,
rate base. Gross rate base includes every item of
unretired plant, including depreciated plant,
Contributions in Aid of Construction, and other plant not
paid for by the Company's investors.
After allocating gross plant-in-service, the
Company prorates the Company's revenue requirement in
proportion to each class's share of gross plant-in­
service.
Why is this a problem? Q.
A. The purpose of a cost-of-service study is fair
allocation of the Company's revenue requirement amongst
its rate classes. This includes an opportunity to earn a
fair Rate-of-Return on the Company's own investment in
plant; however, only a fraction of the Company's gross
plant-in-service represents Company investment. To a
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
15
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
greater or lesser degree, each plant account includes
Company investment, depreciated plant, and customer
contributions. Thus, the use of gross plant-in-service
introduces factors other than the Company's actual
inv estment into the Company's allocation methodology .
Q. Which classes did the Company exclude from its
cost-of-service study ?
A. The Company excluded its two interruptible snow
melt classes, IS-Rand IS-C. The Company has stated that
these classes are small, and has included their
consumption and rev enues in its analysis of its
residential and general service classes. The Company
also provided no information about its schedule H-1,
Ketchum/Sun Valley Area Hook-up Fee.
Q. Why should the Company's interruptible snow
melt classes have been included in the Company's cost-of­
serv ice study ?
A. The Company proposes apply ing the same rates to
its interruptible snow melt classes as it does to
customers in the corresponding firm residential and
general service classes. The Company 's interruptible
snow melt classes were established by Commission Order
No. 31089 in rate Case No . INT-G-09-03. In that case,
the Company explained that by interrupting these
customers' consumption during times of peak demand, it
CASE NO. INT-G-16-02
12 /16 /16
MORRISON, M. (Di)
STAFF
16
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
would be able to defer the costs of upgrading its system
to meet system peak. The Company also explained that
because the load factor of snow melt applications differs
substantially from that of other uses, separate classes
were justified in order to assure that customers with
snow melt systems were not subsidized by other customers.
Because their consumption can be curtailed by
the Company, interruptible snow melt customers do not
cause the Company to incur the costs of meeting
incremental increases in peak load. In its response to
Production Request No. 102, the Company stated that "The
snow melt tariffs have allowed customers to continue to
add this equipment without necessitating the investment
in millions of dollars of capacity upgrades to serve the
snow melt load under peak day conditions." See Exhibit
No. 109.
Given these benefits, their differing load
characteristics, the Company's ability to curtail their
consumption, and their existence as a separate class, the
fundamental principle of cost causation requires that
interruptible snow melt customers and other customers be
treated as separate classes for the purposes of cost
causation and ratemaking. Although the interruptible
snow melt classes represent a fairly small portion of the
Company's sales, they were deemed sufficiently important
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
17
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
for the Company to separate them into their own rate
classes in 2009.
Q. What are y our recommendations regarding the
Company's interruptible snow melt schedules, IS-Rand
IS -C?
A. As noted earlier, the Company did not conduct a
load study, and did not include information about its
interruptible snow melt classes in its cost-of-service
study, so it is not possible to allocate costs to these
classes fairl y based on the cost causation principle .
recommend that the IS-Rand IS-C classes be included in
the load study and cost-of-service workshops that I am
proposing. Until then, I recommend that costs be
allocated to these classes in proportion to the
normalized revenue currently being collected from these
two classes . My proposed allocation is summarized in
Ex hibit No. 110.
WEATHER AND CONSUMPTION NORMALIZATION
I
Q. What is the purpose of weather and consumption
normalization?
A. Weather normalization adjusts test year gas
consumption to the level that would hav e been consumed in
an average weather year . For e xample, if the test year
were cooler than normal, we would e x pect gas consumption
to be greater than it would have been in an average
CASE NO . INT-G-16-02
12 /16 /16
MORRISON, M. (Di)
STAFF
18
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
weather year. Weather normalization is a very important
part of consumption normalization, which adjusts
consumption for changes in other factors, such as changes
in the numbers of customers in each rate class.
Q. Describe the Company's weather normalization
methodology.
A. The Company created consumption models for its
RS-1, RS-2, and GS-1 classes using statistical multiple
regression, and then used the results of these models to
estimate per-customer consumption for each class. Each
month's result was then multiplied by the forecast number
of customers for that month. Recall that the Company's
test year is a hybrid that uses actual data for the
months January through June, and predicted data for the
months July through December. For the months January
through June, the Company adjusted actual consumption
data; however, for the months July through December, the
Company predicted monthly consumption using its model.
The Company's models use weighted monthly
heating degree days (HDD) as a weather variable, a trend
variable, and an autoregressive term as predictors. A
peculiar feature of the Company's model is its use of
different weather coefficients for different months.
Q. Do you agree with the Company's regression
methodology?
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
19
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
A. No. Although the Company's models were created
using autoregressive terms, these terms were not included
in the calculation of monthly consumption. The general
effect of this omission was to underestimate the
Company's monthly consumption estimates.
The use of autoregressive terms in a predictive
weather normalization model is inappropriate. One
obvious problem is that they can cause the model to
become grossly unstable when used to make predictions
beyond the time period of the data used to create the
model. I repeated the Company's test year calculations
using the Company's full model, including its
autoregressive terms. The resulting consumption
estimates were unrealistically high: The Company's GS-1
model predicted test year consumption that was 79% higher
than its GS-1 customers actually consumed.
Unfortunately, the Company 's solved the model
instability problem by omitting the autoregressive terms
from its monthly calculations, resulting in
underestimates of annual consumption.
There is also a technical problem with the use
of autoregressive terms for weather normalization:
Although models containing autoregressive terms are
appropriate for some control and signal processing
applications, their use in weather normalization violates
CASE NO. INT-G-16-02
12 /16 /16
MORRISON, M. (Di)
STAFF
20
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
regression's fundamental independence assumption.
The Company's regression model assumed a linear
relationship between Heating Degree Days and consumption;
however, as we can see in Figure 1, the relationship is
clearly non-linear.
Fig. 1: RS-2 Consumption per Customer
180
160
140
Vl 120 E ~ CJ l(X) .c r-
-0 80 ~ e HS··02 ·-60 a)
40 •
20
()
0 200 4(XJ 600 S(X) l(X)O 1200 1400
HDD
Q. Please describe your approach to weather
normalization modeling.
A. I used standard regression techniques of the
sort taught in intermediate college statistical methods
courses. These methods are robust, well characterized,
and in common use. They are also grounded in a
mathematical foundation that makes them amenable to
objective evaluation. I used Ramsey and Schafer's 1997
textbook, The Statistical Sleuth, as a reference.
Like the Company, I used multiple regression to
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
21
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
develop monthly consumption models for each class.
Unlike the Company, I analyzed each of the Company's GS-1
subclasses (small commercial, large commercial,
compressed natural gas, and irrigators) separately, and
then aggregated the results. The consumption patterns of
these subclasses differ sufficiently to warrant separate
treatment.
I used the Company's weighted heating degree
days as a weather predictor.
year as a covariate.
I also included calendar
The Company's method for forecasting system
growth is based on historical relationships between the
numbers of active building permits and customer growth in
each portion of its service territory. This approach is
reasonable, and I accepted Company's RS-1, RS-2, and GS-1
forecast totals; however, because the Company did not
forecast GS-1 growth by subclass, I allocated the growth
projected by the Company to its small commercial and
large commercial subclasses.
I used backward/mixed stepwise regression to
develop models of per-customer consumption for the
Company's RS-1 and RS-2 customers, as well as separate
models for each of its GS-1 subclasses. My criteria for
inclusion/exclusion from the model was a change in
adjusted R2 = 0.02. Non-linearity in the relationship
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
22
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
between consumption and heating degree days was modeled
using quadratic and cubic curvature terms.
The resulting models were simpler, but much
more robust than those developed by the Company. With
the obvious exception of the irrigation subclass, the
number of monthly heating degree days was an excellent
predictor of consumption for the RS-1, RS-2, and GS-1
classes.
As mentioned earlier, the Company's model used
different consumption coefficients for each month.
Although this approach is not inherently incorrect, it
results in an unnecessarily complicated model. The
Company's RS-2 model uses 15 different factors. The RS-2
model developed as I have described uses just four
factors.
Q. How do your adjustments compare with those
proposed by the Company?
A. Overall, my normalized consumption estimate for
the Company's core customers (RS-1, RS-2, and GS-1) is
approximately 2.15% greater than that proposed by the
Company . My estimates for the residential RS-1 and RS-2
classes are 1.54% and 2.12% greater, respectively, than
the Company's. My estimate for the aggregated GS-1 class
is 2.38% greater than that proposed by the Company.
Overall, my estimate of test year gas operating revenue
CASE NO. INT-G-16-02
12 /16 /16
MORRISON, M. (Di)
STAFF
23
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
is 2.02% greater than the Company's. My estimate of the
Company's base rate revenue (gas operating revenue less
the cost of gas) is 2.62% greater than the Company's
estimate. I have summarized my proposed rate
determinants in Exhibit No. 111. A summary of my
proposed revenue requirement allocation can be found in
Exhibit No. 110.
THE COMPANY'S LINE AND MAINS EXTENSION POLICIES
Q. How are the Company's service line and mains
extension policies affected by the Company's Rate-of­
Return?
A. The Company's line and mains extension policies
are discussed more fully by Staff witness Farley. The
allowance represents the maximum investment that the
Company is allowed to make when connecting a new customer
to its system. As such, it represents a significant
fraction of the Company's distribution plant-in-service.
As I have explained, the Company's accounting system does
not distinguish between the Company's investment and
customer contributions to those accounts, so it is
imperative that the Company's tariff accurately reflect
the Rate-of-Return authorized by the Commission.
Currently, the tariff is predicated on a 12.5% Rate-of­
Return, and the Company has not updated its tariffs to
reflect its proposed Rate-of-Return. As explained by
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
24
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
Staff witness Farley, the Company has resisted Staff
requests to update the formulae in its line and mains
extension policies to reflect its proposed Rate-of­
Return.
Q. What are your recommendations regarding the
Company's line and mains extension policies?
A. The Company's service line and mains extension
policies should be updated immediately to reflect the
Commission authorized Rate-of-Return.
CONCLUSIONS AND RECOMMENDATIONS
Q. Please summarize your recommendations regarding
the Company's cost-of-service study.
A. Because the Company did not conduct a load
study, it was unable to develop system peak allocators
that would permit a fair allocation of the Company's net
plant-in-service to its rate classes. Furthermore, the
Company excluded its interruptible snow melt customers
from its cost-of-service model. I recommend that the
Company be required to conduct a load study that includes
all of its classes.
The Company proposes that the costs of
services, meters, regulators, ERTs, and related accounts
be allocated using factors derived from relative meter
costs. This is inappropriate. The costs of service
lines should be allocated using the relative costs of
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
25
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
providing service lines to each class, the costs of
regulators should be allocated using relative regulator
costs, and so on. Absent the cost accounting information
required to directly allocate these costs, acceptable
allocators could be developed using the Company's records
of line extension costs.
The Company is also proposing an inappropriate
classification of over 47% of its mains as customer-
related. Typically, most distribution-related costs that
cannot be directly allocated are allocated using non­
coincident peak, or peak and average allocators.
Cumulatively, the Company is proposing that 52% of its
$407,663,702 distribution-related rate base be classified
as customer-related.
I propose that the Company, Staff, and other
stakeholders convene a series of workshops in order to
develop a load study and update the cost-of-service study
in order to incorporate improved allocators obtained from
the load study and from analysis of the actual costs
incurred by the Company when connecting new customers.
All schedules, including the Company's interruptible snow
melt and Ketchum/Sun Valley area hookup schedules, should
be included in the process.
Q. Please summarize your recommendations regarding
the Company's weather normalization methodology.
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
26
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
A. The Company's weather normalization adjustments
were derived using autoregressive models that include
autoregressive terms. Although these terms were used to
model the data, they were not used in the Company's
adjustments to January -June test year adjustments, or
in its July -December forecasts. As a result, the
Company's estimates underestimate the weather normalized
consumption of its RS-1, RS-2, and GS-1 classes.
I used a simpler and more robust modeling
methodology. My proposed billing determinants are
summarized in Exhibit No. 111.
Q. What are your recommendations regarding the
Company's Line and Mains extension policies?
A. The Company's Service Line and Mains extension
policies should be updated immediately to reflect the
Commission authorized Rate -of-Return. Additionally, as
explained by Staff witness Farley, the factors and
assumptions used in the derivation of these schedules
were last updated in 1986. I propose that these
schedules be updated in a separate rate case.
Q. Do you have any other observations and
recommendations?
A. Yes. The Company explained to Staff that the
Company had considerable difficulty obtaining data
necessary for cost allocation and weather normalization.
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
27
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
As noted earlier, the Company does not keep the
information necessary to determine costs, depreciation,
or customer contributions by class for any of its plant
related FERC accounts. This relatively coarse level of
information is adequate for determination of the
Company's overall revenue requirement, but is inadequate
for best practice revenue requirement allocation. There
was also considerable difficulty obtaining the data
necessary to fully evaluate the Company's weather
normalization methodology and its mains study. The
Company explained that because of system upgrades in 2002
and 2013, detailed data was unavailable. I recommend
that the Company adopt data retention policies that
assure that this information is available when needed for
ratemaking purposes.
Q.
A.
Does that conclude your testimony?
Yes, it does.
CASE NO. INT-G-16-02
12/16/16
MORRISON, M. (Di)
STAFF
28
REQUEST NO. 202: Please provide a breakdown of Capital Expenditures in FERC
accounts 380, 381, 382, 383, 384, and 385 and by existing rate class.
RESPONSE TO REQUEST NO. 202:
Please see the CD file labeled "PR #202 Plant Act Bats YTD Sep 2016.xlsx "for the
breakdown of actual account balances for January through September 2016. For theforecasted
months of October -December 2016, please refer lo file "Gas Plant in Service.xlsx" submilled
in response to Production Request No. 178.
I ntermountain 's accounting records have never provided the fimctionality to track plant
and accumulated depreciation by rate class for FERC accounts 380, 381, 382, 383, 384, and
385. Since this type of direct assignment was not possible, lntermountain chose lo use an
industry accepted praclice lo allocate these accounts to the various customer classes.
Because I ntermounlain can associate its meters in service with a customer class, the
Company was able to determine the number of each type of meter currently installed for each
rate class. lntermountain then valued those meters at the current replacement cost/or each
meter. The total current cost of the meters for each class was calculated and then averaged by
l the number of meters in each class. That average meter cost was indexed and then multiplied by
annual customers to create a weighted customer allocator. The weighted customer allocator
was split into Group 1 and Group 2 categories as outlined in the response to Production Request
No. 200. Th e calculation of the weighted customer allocator was included in response to
Production Request No. 33, "PR 33 2015 Meter Stud_v-CONFIDENTJAL.xlsx ". These
allocators were used to assign the FERC accounts 380-385 to lntermountain 's customer classes.
Record Holder: Mike McGrath. 208-377-6000
Location: 555 S Cole Rd. Boise. ID 83707
Sponsor/Preparer: Ted Dedden. 208-377-6000
RESPONSE OF JGC TO NINTH PRODUCTION REQUEST OF COMMISSION STAFF Exhibit No. I 08
Case No. INT-G-16-02
M. Morrison, Staff
12/16/16
REQUEST NO. 102: Please provide any workpapers showing the Company's cost
allocation methodology for its interruptible snow melt schedules, IS-Rand IS-C.
RESPONSE TO REQUEST NO. 102:
The snow melt schedules provide an operational tool for lntermountain to maximize its
system efficiency and to keep costs low for all customers. Snow melt customers have a very low
load factor and tend to use gas during peak usage periods. The snow melt tariffs have allowed
customers to continue to add this equipment without necessitating the investment in millions of
dollars of capacity upgrades to serve the snow melt load under peak day conditions. When the
tariffs were established, the IS-R (Residential) customers were priced the same as RS-2 customers,
and IS-C (Commercial) customers were priced the same as GS-1 customers. Since no incremental
infrastructure has been added to serve these customers on a peak day, lntermozmtain believes it is
still appropriate to price their usage based on the new RS (Residential) or GS-1 (Commercial)
rates. Therefore, the /S-R customers and usage have been included with RS and the JS-C customers
and usage have been included with GS-1 for cost allocation purposes.
Record Holder:
Location:
Sponsor/Preparer:
Mike McGrath, 208-377-6000
555 S Cole Rd, Boise. ID 83707
Lori Blattner. 208-377-6000
RESPONSE OF IGC TO FOURTH PRODUCTION REQUEST OF COMMISSION STAFF.
Exhibit No. I 09
Case No. INT-G-16-02
M. Morrison, Staff
12/16/16
Staff Allocation of Revenue to Existing Classes
Test Year Ending December 31, 2016
Normalized Test Year Revenue at Current Rates Revenue Requirement
Base Rate Gas Operating
Class Base Rate Fraction Revenue Base Rate
RS-1 $10,953,290 0.128473 $29,282,269 $11,417,980
RS-2 43,290,919 0.507768 138,209,206 45,127,524
15-R 20,158,015 0.000307 76,710,182 27,259
GS-1 26,149 0.236438 97,025 21,013,213
15-C 1,891 0.000022 6,791 1,971
LV-1 407,862 0.004784 2,852,315 425,165
T-3 833,713 0.009779 791,175 869,083
T-4 8,909,434 0.104501 8,364,283 9,287,414
T-5 675,936 0.007928 649,238 704,612
Total $85,257,209 1.000000 $256,962,485 $88,874,221
Staff Allocation of Revenue to Proposed Classes
Test Year Ending December 31, 2016
Normalized Test Year Revenue at Current Rates Revenue Requirement
Base Rate Gas Operating
Class Base Rate Fraction Revenue Base Rate
RS $54,244,209 0.636242 $167,491,476 $56,545,504
15-R 20,158,015 0.000307 76,710,182 27,259
GS-1 26,149 0.236438 97,025 21,013,213
15-C 1,891 0.000022 6,791 1,971
LV-1 407,862 0.004784 2,852,315 425,165
T-3 833,713 0.009779 791,175 869,083
T-4 9,585,370 0.112429 9,013,521 9,992,027
Total $85,257,209 1.000000 $256,962,485 $88,874,221
Notes:
1. Base rate revenues exclude PGA costs of gas.
2. $88,874,221 revenue requirement from Witness Terry's Exhibit 103.
3. Base Rate Revenue requirement obtained by multiplying Revenue Requirement
by Class Base Rate Fraction.
Exhibit No. 110
Case No. INT-G-16-02
M. Morrison, Staff
12/16/16
Determinant
No. Bills
Total Therms
No. Bills
Total Therms
No. Bills
Total Therms
No. Bills
Total Therms
-~ n tn ~. ~ & 0: ~ ~ &
----0 Z -· 0: 3. ~ z
t/) t::l 0 o L. -o?-i-~ ~ 6:: ~ P:, I ..-~,.....
0 9' ...., 0
+:-N
Staff's Estimate of Billing Determinants
For the Company's Test Year Ending December 31, 2016
Residential Classes (RS-1, RS-2, Proposed RS, and IS-R)
Jan Feb March April May June July Aug
RS-1
67,871 67,909 67,792 67,488 67,149 66,759 66,821 66,905
7,003,940 5,924,110 4,254,150 2,920,936 1,580,978 541,104 190,702 116,131
RS-2
238,657 239,101 239,620 239,913 240,209 240,504 241,150 241,746
34,076,866 29,124,512 21,722,795 15,930,489 10,122,151 5,612,327 4,065,840 3,747,025
Proposed RS Class
306,528 307,010 307,412 307,401 307,358 307,263 307,971 308,651
41,080,805 35,048,621 25,976,945 18,851,425 11,703,130 6,153,430 4,256,543 3,863,156
IS-R: Residential Interruptible Snow Melt
81 82 82 84 84 85 85 85
44,420 23,236 12,318 1,933 1,631 1,083 383 376
Sep Oct Nov Dec Annual
66,996 67,167 67,274 67,357 807,488
217,491 904,678 3,453,779 5,864,316 32,972,314
242,300 242,944 243,367 243,796 2,893,307
4,200,713 7,227,533 18,564,090 29,609,318 184,003,658
309,296 310,111 310,641 311,153 3,700,795
4,418,204 8,132,210 22,017,869 35,473,634 216,975,972
85 85 85 85 1,008
731 1,190 13,007 37,089 137,397
Staff's Estimate of Billing Determinants
For the Company's Test Year Ending December 31, 2016
Commercial Subclasses (GS-10, GS-11, GS-12, GS-20, and GS-60)
Determinant Jan Feb March April May June July Aug Sep Oct Nov Dec Annual
GS-10
No. Bills 73 73 75 76 75 73 70 70 72 77 73 75 882
Total Therms 648,680 538,049 457,258 336,198 271,623 191,624 151,493 247,673 344,058 311,609 458,263 612,727 4,569,254
Block 1 145,350 141,332 133,681 116,481 97,594 60,711 44,433 68,831 102,101 101,377 182,555 154,927 1,349,374
Block 2 322,460 263,988 216,727 159,335 125,263 92,425 68,517 110,979 165,588 118,442 192,763 303,531 2,140,017
Block 3 180,870 132,728 106,849 60,382 48,765 38,488 38,543 67,863 76,369 91,790 82,945 154,268 1,079,862
GS-11
No. Bills 32,038 32,032 32,009 31,952 31,897 31,828 31,896 31,949 31,996 32,082 32,140 32,192 384,011
Total Therms 18,829,887 15,875,869 11,646,852 8,469,977 5,444,505 3,308,721 2,715,707 2,647,654 2,757,495 4,000,591 9,764,693 15,868,037 101,329,989
Block 1 4,219,228 4,170,201 3,405,008 2,934,566 1,956,218 1,048,275 796,527 735,815 818,299 1,301,525 3,889,889 4,012,208 29,287,760
Block 2 9,360,353 7,789,332 5,520,279 4,014,192 2,510,825 1,595,875 1,228,247 1,186,374 1,327,123 1,520,619 4,107,409 7,860,675 48,021,302
Block 3 5,250,306 3,916,336 2,721,565 1,521,219 977,463 664,571 690,933 725,466 612,073 1,178,447 1,767,396 3,995,153 24,020,927
GS-12
No. Bills 3 3 3 3 3 3 3 3 3 3 3 3 36
Total/Block 3 169 104 351 1,487 1,714 182 522 649 349 1,454 231 221 7,433
GS-20
No. Bills 68 69 65 61 62 62 62 62 62 62 62 62 759
Total Therms 711,914 627,077 474,136 367,282 297,798 238,633 209,598 203,689 211,494 253,389 398,970 555,961 4,549,940
Block 1 159,519 164,718 138,616 127,251 106,999 75,604 61,476 56,608 62,762 82,436 158,935 140,574 1,335,497
Block 2 353,893 307,669 224,727 174,067 137,334 115,099 94,796 91,270 101,787 96,313 167,822 275,411 2,140,188
Block 3 198,502 154,690 110,793 65,964 53,464 47,931 53,326 55,811 46,945 74,640 72,213 139,976 1,074,256
GS-60
No. Bills 1 0 0 1 8 19 15 19 18 17 4 1 103
Total Therms 0 0 0 0 1,959 15,499 20,291 14,352 9,074 1,030 5,864 3 68,071
Block 1 0 0 0 0 677 3,675 2,566 3,662 2,556 1,030 5,864 3 20,034
Block 2 0 0 0 0 1,282 11,824 14,416 10,689 6,517 0 0 0 44,729
Block 3 0 0 0 0 0 0 3,308 0 0 0 0 0 3,308
;:::; $'. n m
--.... . Pl &
;: $'. ~ -· :::::oz:!. °' 3. 9 z
"' -0 0 z . --o?....,-
Po) [./JI _... cro ..... o-(1) p) I
N~.-
0 °' --+, I
.i::. B
-~nm ~. ~ & 0: ~ en er
;:::; 0 Z ;:;:
0\ :l. ~ z ~ z ~ '"O?-,-
p:, I -
(JQ ~O-ro PJ I w '"t>-'"-+\ 0\ 0 I ...., 0
+:-Iv
Determinant
No. Bills
Total Therms
Block 1
Block 2
Block 3
No. Bills
Total Therms
Block 1
Block 2
Block 3
Jan Feb March
32,183 32,177 32,152
20,190,651 17,041,099 12,578,596
4,524,097 4,476,251 3,677,305
10,036,705 8,360,989 5,961,733
5,629,848 4,203,859 2,939,559
7 8 8
6,148 3,407 1,761
958 713 478
2,746 1,862 1,283
2,444 832 0
Staff's Estimate of Billing Determinants
For the Company's Test Year Ending December 31, 2016
Commercial Classes (GS-1 and IS-C)
April May June July Aug Sep Oct Nov Dec Annual
All GS-1 (Includes GS-10, GS-11, GS-12, GS-20, and GS-60)
32,093 32,045 31,985 32,046 32,103 32,151 32,241 32,282 32,333 385,791
9,174,944 6,017,599 3,754,659 3,097,610 3,114,016 3,322,470 4,568,073 10,628,022 17,036,949 110,524,687
3,178,298 2,161,489 1,188,265 905,003 864,916 985,718 1,486,367 4,237,243 4,307,713 31,992,665
4,347,594 2,774,704 1,815,222 1,405,976 1,399,311 1,601,016 1,735,374 4,467,994 8,439,618 52,346,235
1,649,052 1,081,406 751,172 786,632 849,789 735,736 1,346,332 1,922,785 4,289,619 26,185,787
IS-C
8 8 9 9 9 9 9 9 9 102
501 26 20 0 0 10 179 1,247 2,711 16,010
307 26 20 0 0 10 179 1,247 1,973 5,911
194 0 0 0 0 0 0 0 738 6,823
0 0 0 0 0 0 0 0 0 3,276
Determinant
No. Bills
Total Therms
No. Bills
Total Therms
Block 1
Block 2
Block 3
No. Bills
Total Therms
Block 1
Block 2
Block 3
No. Bills
Demand
Therms
Overrun
-~ n tT1 ~. ~ & a:~ C', &
;:::; 0 Z ;:;: °' 3. :::> z
"' :'.:j 0 0 L, .
'"t:1? -l -P, I -(Jq ~o-C', P, I ~ ~ a:
0 I H) 0 ~ N
Staff's Estimate of Billing Determinants
For the Company's Test Year Ending December 31, 2016
Industrial and Transportation Customers (LV-1, T-3, T-4, and T-5)
Jan Feb March April May June July Aug Sep Oct Nov Dec Annual
LV-1
18 18 19 18 18 18 18 18 18 18 18 18 217
611,474 535,067 552,455 477,142 468,888 448,284 437,000 464,425 468,975 543,275 654,775 655,800 6,317,560
T-3
6 6 6 6 6 6 6 6 6 6 6 6 72
4,085,789 4,002,252 3,497,018 3,561,492 3,829,801 3,558,066 2,802,052 2,884,657 3,909,847 4,767,780 3,889,548 3,987,949 44,776,251
639,350 685,324 557,405 602,777 672,831 643,754 600,355 659,069 945,777 965,580 756,115 579,743 8,308,080
283,603 303,253 240,030 265,352 295,551 287,260 227,348 253,449 369,721 359,926 291,154 242,375 3,419,023
3,162,835 3,013,674 2,699,583 2,693,364 2,861,419 2,627,053 1,974,349 1,972,138 2,594,349 3,442,274 2,842,279 3,165,831 33,049,148
T-4
82 82 83 83 84 84 85 85 82 82 80 81 993
30,097,020 27,212,795 26,163,694 19,612,914 19,185,335 16,081,079 16,263,500 16,014,439 18,009,572 23,104,566 25,438,908 27,452,850 264,636,672
10,579,501 10,081,742 10,038,853 8,951,397 8,649,594 8,070,863 7,950,293 7,934,993 8,317,770 9,004,045 10,237,085 10,931,399 110,747,535
8,422,485 8,263,323 8,557,216 7,257,099 6,857,044 5,562,670 5,454,414 6,090,375 6,440,697 7,012,691 7,443,625 8,599,123 85,960,760
11,095,034 8,867,730 7,567,626 3,404,418 3,678,697 2,447,546 2,858,792 1,989,072 3,251,105 7,087,830 7,758,199 7,922,327 67,928,377
T-5
13 13 13 13 13 13 13 13 13 13 13 13 156
44,685 44,685 44,685 44,685 44,685 44,685 44,685 44,685 44,685 44,685 44,685 44,685 536,220
1,345,688 1,261,503 1,338,859 1,287,306 1,321,419 1,278,537 1,326,885 1,275,235 1,282,450 1,325,435 1,286,550 1,325,485 15,655,352
437,768 245,505 402,400 458,228 423,688 566,829 211,770 179,130 378,160 198,840 481,640 136,850 4,120,808
CERTIFICATE OF SERVICE
I HEREBY CERTIFY THAT I HAVE THIS 16TH DAY OF DECEMBER 2016,
SERVED THE FOREGOING DIRECT TESTIMONY OF MICHAEL MORRISON, IN
CASE NO. INT-G-16-02, BY MAILING A COPY THEREOF, POSTAGE PREPAID,
TO THE FOLLOWING:
MICHAEL P McGRATH
DIR -REGULA TORY AFFAIRS
INTERMOUNTAIN GAS CO
PO BOX 7608
BOISE ID 83707
E-MAIL: [e-mail]
BRADMPURDY
ATTORNEY AT LAW
2019 N 17TH STREET
BOISE ID 83702
E-MAIL: [e-mail]
CHAD M STOKES
TOMMY A BROOKS
CABLE HUSTON LLP
1001 SW 5TH AVE STE 2000
PORTLAND OR 97204-1136
E-MAIL: [e-mail]
tbrooks@,cablehuston.com
BENJAMIN J OTTO
ID CONSERVATION LEAGUE
710 N 6TH STREET
BOISE ID 83702
E-MAIL: [e-mail]
PETER RICHARDSON
GREGORY MADAMS
RICHARDSON ADAMS PLLC
515 N 27TH STREET
BOISE ID 83702
E-MAIL: [e-mail]
[e-mail]
RONALD L WILLIAMS
WILLIAMS BRADBURY
1015 W HAYS ST
BOISE ID 83702
E-MAIL: [e-mail]
EDWARD A FINKLEA
EXECUTIVE DIRECTOR
NW INDUSTRIAL GAS USERS
545 GRANDVIEW DR
ASHLAND OR 87520
E-MAIL: e:[e-mail]
ELECTRONIC ONLY
MICHAEL C CREAMER
GIVENS PURSLEY LLP
E-MAIL: [e-mail]
F DIEGO RIV AS
NW ENERGY COALITION
1101 8TH AVENUE
HELENA MT 59601
E-MAIL: [e-mail]
SCOTT DALE BLICKENSTAFF
AMALGAMATED SUGAR CO LLC
1951 S SATURN WAY
STE 100
BOISE ID 83702
E-MAIL: [e-mail]
CERTIFICATE OF SERVICE
KEN MILLER
SNAKE RIVER ALLIANCE
PO BOX 1731
BOISE ID 83701
E-MAIL: [e-mail]
LANNY L ZIEMAN
NATALIE A CEPAK
THOMAS A JERNIGAN
EBONY M PAYTON
AFLOA/JA-ULFSC
139 BARNES DR STE 1
TYNDALL AFB FL 32403
E-MAIL: lanny.zieman. l @us.af.mil
[e-mail]
[e-mail]
[e-mail]
ANDREW J UNSICKER MAJ USAF
AFLOA/JACE-ULFSC
139 BARNES DR STE 1
TYNDALL AFB FL 32403
E-MAIL: [e-mail]
CERTIFICATE OF SERVICE