20150601Knox Direct.pdf

DAVID J. MEYER VICE PRESIDENT AND CHIEF COUNSEL FOR
REGULATORY & GOVERNMENTAL AFFAIRS AVISTA CORPORATION P.O. BOX 3727 1411 EAST MISSION AVENUE SPOKANE, WASHINGTON 99220-3727
TELEPHONE: [phone] FACSIMILE: [phone] [e-mail]

BEFORE THE IDAHO PUBLIC UTILITIES COMMISSION

IN THE MATTER OF THE APPLICATION ) CASE NO. AVU-E-15-05 OF AVISTA CORPORATION FOR THE ) AUTHORITY TO INCREASE ITS RATES )
AND CHARGES FOR ELECTRIC AND ) NATURAL GAS SERVICE TO ELECTRIC ) DIRECT TESTIMONY
AND NATURAL GAS CUSTOMERS IN THE ) OF STATE OF IDAHO ) TARA L. KNOX )

FOR AVISTA CORPORATION
(ELECTRIC ONLY)

I. INTRODUCTION 1
Q. Please state your name, business address and 2
present position with Avista Corporation. 3
A. My name is Tara L. Knox and my business address 4
is 1411 East Mission Avenue, Spokane, Washington. I am 5
employed as a Senior Regulatory Analyst in the State and 6
Federal Regulation Department. 7
Q. Would you briefly describe your duties? 8
A. Yes. I am responsible for preparing the 9
electric regulatory cost of service model for the Company, 10
as well as providing support for the preparation of 11
results of operations reports, among other things. 12
Q. What is your educational background and 13
professional experience? 14
A. I am a graduate of Washington State University 15
with a Bachelor of Arts degree in General Humanities in 16
1982, and a Master of Accounting degree in 1990. As an 17
employee in the State and Federal Regulation Department at 18
Avista since 1991, I have attended several ratemaking 19
classes, including the EEI Electric Rates Advanced Course 20
that specializes in cost allocation and cost of service 21
issues. I am also a member of the Cost of Service Working 22
Group and the Northwest Pricing and Regulatory Forum, 23
which are discussion groups made up of technical 24
Knox, Di Page 1 Avista Corporation
professionals from regional utilities and utilities 1
throughout the United States and Canada concerned with 2
cost of service issues. 3
Q. What is the scope of your testimony in this 4
proceeding? 5
A. My testimony and exhibits will cover the 6
Company’s electric revenue normalization adjustment to the 7
test year results of operations, the proposed Load Change 8
Adjustment Rate to be used in the Power Cost Adjustment 9
mechanism, and the electric cost of service study 10
performed for this proceeding. A table of contents for my 11
testimony is as follows: 12
Description Page 13
I. Introduction 1 14
II. Electric Revenue Normalization 3 15
III. Proposed Load Change Adjustment Rate 7 16
IV. Electric Cost of Service 10 17
18
Q. Are you sponsoring any exhibits in this case? 19
A. Yes. I am sponsoring Exhibit 13 composed of 20
three schedules. Schedule 1 details the calculation of 21
the proposed Load Change Adjustment Rate, Schedule 2 22
includes a narrative of the electric cost of service study 23
process, and Schedule 3 presents the electric cost of 24
service study summary results. 25
Knox, Di Page 2 Avista Corporation
Q. Were these exhibit schedules prepared by you or 1
under your direction? 2
A. Yes, they were. 3
4
II. ELECTRIC REVENUE NORMALIZATION 5
Q. Would you please describe the electric revenue 6
normalization adjustment included in Company witness Ms. 7
Andrews’ pro forma results of operations? 8
A. Yes. The electric revenue normalization 9
adjustment represents the difference between the Company’s 10
actual recorded retail revenues during the twelve months 11
ended December 2014 test period, and base rate retail 12
revenues on a normalized (pro forma) basis. The total 13
revenue normalization adjustment increases Idaho net 14
operating income by $4,056,000, as shown in adjustment 15
column 2.07 on page 7 of Ms. Andrews Exhibit No. 12, 16
Schedule 1. 17
The revenue normalization adjustment consists of four 18
primary components: 1) re-pricing customer usage 19
(adjusted for any known and measurable changes) to base 20
tariff rates presently in effect, 2) adjusting customer 21
load and revenue to a 12-month calendar basis (unbilled 22
revenue adjustment), 3) weather normalizing customer usage 23
Knox, Di Page 3 Avista Corporation
and revenue, and 4) eliminating the provision for earnings 1
sharing associated with the 2014 earnings test. 2
Q. Since these elements are combined into a single 3
adjustment, would you please identify the impact of each 4
component? 5
A. Yes. A breakdown of the four components of the 6
revenue normalization is as follows: 7
1. The re-pricing of billed usage including the 8 elimination of adder schedule revenue and 9 related amortization expense (Schedule 59 10
Residential Exchange Credit, Schedule 91 Public 11 Purpose Tariff Rider, Schedule 95 Optional 12 Renewable Power and Schedule 97 BPA Settlement 13 Adjustment)1 results in a reduction to net income 14 of $103,000.. 15
2. The re-pricing of unbilled calendar usage and 16 elimination of unbilled adder schedule revenue 17
and expense results in a reduction to net income 18 of $87,000.2 19 3. The weather adjustment reduces net income 20
$393,000. 21 4. Finally, the elimination of the 2014 earnings 22
sharing (customer share) results in an increase 23 to net income of $4,639,000. 24
The combined impact of these four elements is an 25
increase to net income $4,056,000. 26
Q. Please briefly summarize the electric weather 27
normalization process. 28
1 Municipal Franchise Fee and Power Cost Adjustment revenues and related expenses are eliminated in separate adjustments.
2 The unbilled adjustment consists of removing December 2013 usage billed in January 2014 from the 2014 test year, adding December 2014 usage billed in January 2015 to the 2014 test year, and re-pricing the
net usage at present base rates.
Knox, Di Page 4 Avista Corporation

A. The Company’s electric weather normalization 1
adjustment calculates the change in kWh usage required to 2
adjust actual loads during the 2014 test period to the 3
amount expected if weather had been normal. This 4
adjustment incorporates the effect of both heating and 5
cooling on weather-sensitive customer groups. The weather 6
adjustment is developed from a regression analysis of ten 7
years of billed usage per customer and billing period 8
heating and cooling degree-day data. The resulting 9
seasonal weather sensitivity factors (use-per-customer-10
per-heating-degree day and use-per-customer-per-cooling-11
degree day) are applied to monthly test period customers 12
and the difference between normal heating/cooling degree-13
days and monthly test period observed heating/cooling 14
degree-days. 15
Q. Have the seasonal weather sensitivity factors 16
been updated since the last rate case? 17
A. Yes. The factors used in the weather adjustment 18
are based on regression analysis of monthly billed usage-19
per-customer from January 2004 through December 2013, 20
which is the most recent completed analysis. 21
Q. What data did you use to determine “normal” 22
heating and cooling degree days? 23
Knox, Di Page 5 Avista Corporation
A. Normal heating and cooling degree days are based 1
on a rolling 30-year average of heating and cooling 2
degree-days reported for each month by the National 3
Weather Service for the Spokane Airport weather station. 4
Each year the normal values are adjusted to capture the 5
most recent year with the oldest year dropping off, 6
thereby reflecting the most recent information available 7
at the end of each calendar year. The calculation 8
includes the 30-year period from 1985 through 2014. 9
Q. Is this proposed weather adjustment methodology 10
consistent with the methodology utilized in the Company’s 11
last general rate case in Idaho? 12
A. Yes. The process for determining the weather 13
sensitivity factors and the monthly adjustment calculation 14
is consistent with the methodology presented in Case No. 15
AVU-E-12-08. 16
Q. What was the change in kWhs resulting from 17
weather normalization for the twelve months ended December 18
2014 test year? 19
A. Weather was warmer than normal throughout 2014, 20
except for February, which was colder than normal. The 21
summer months of July and August were particularly hot. 22
Since electric usage is impacted by both heating and 23
cooling, weather normalization required an addition to 24
Knox, Di Page 6 Avista Corporation
usage for warm weather during the winter (partially offset 1
by the February 2014 arctic outbreaks) and a reduction to 2
usage for the hot summer. These offsetting impacts 3
resulted in a relatively small annual weather adjustment 4
even though the monthly variations were volatile. 5
Overall, the adjustment to normal required the 6
addition of 340 heating degree-days during the heating 7
season,3 and the deduction of 199 cooling degree-days 8
during the summer season.4 The annual total adjustment to 9
Idaho electric sales volumes was a reduction of 9,000,496 10
kWhs, which is approximately 0.3% of billed usage. 11
The electric system monthly weather adjustment 12
volumes were provided to Company witness Mr. Johnson as an 13
input to the Pro Forma Power Supply analysis. 14
15
III. PROPOSED LOAD CHANGE ADJUSTMENT RATE 16
Q. What is the Load Change Adjustment Rate? 17
A. The Load Change Adjustment Rate (LCAR) is part 18
of the Power Cost Adjustment (PCA) mechanism that prices 19
the change in power supply-related costs associated with 20
the change in actual retail loads, from the retail loads 21
3 The heating season includes the months of January through June and October through December. 4 The summer season includes the months of June through September. June is included in both seasons because both heating load and cooling
load fluctuations occur during the month.
Knox, Di Page 7 Avista Corporation

that were used to set the PCA base costs. The LCAR 1
determination process for all Idaho investor-owned 2
utilities was established in IPUC Case No. GNR-E-10-03, 3
Order No. 32206, which was approved on March, 15, 2011. 4
Q. How is the rate determined? 5
A. The proposed LCAR is determined by computing the 6
proposed revenue requirement on the production and 7
transmission costs contained within Ms. Andrews’ Idaho 8
electric pro forma total results of operations. The 9
production/transmission revenue requirement amount is then 10
divided by the Idaho normalized retail load used to set 11
rates in order to arrive at the average production and 12
transmission cost-per-kWh embedded in proposed rates. 13
This amount is then multiplied by the proportion of 14
production and transmission costs classified as energy-15
related in the cost of service study. 16
Q. Do you have an exhibit schedule that shows the 17
calculation of the proposed LCAR? 18
A. Yes. Exhibit No. 13, Schedule 1 begins with the 19
identification of the production and transmission revenue, 20
expense and rate base amounts included in each of Ms. 21
Andrews’ actual, restating, and pro forma adjustments to 22
results of operations. The “2016 Pro Forma Total” on line 23
Knox, Di Page 8 Avista Corporation
35 at the bottom of page 1 shows the resulting production 1
and transmission cost components. 2
Page 2 shows the revenue requirement calculation on 3
the production and transmission cost components. The rate 4
of return and debt cost percentages on Line 2 are inputs 5
from the proposed cost of capital. The normalized retail 6
load on Line 10 comes from the workpapers supporting the 7
revenue normalization adjustment. Line 11 represents the 8
average total production and transmission cost-per-kWh 9
proposed to be embedded in Idaho customer retail rates. 10
Lines 12 and 13 are values taken from the cost of service 11
study report titled “Functional Cost Summary by 12
Classification at Uniform Requested Return” representing 13
total costs at unity. Line 12 shows the amount of 14
production and transmission costs classified as energy 15
related, while Line 13 shows the total production and 16
transmission costs in the study. 17
The resulting 2016 LCAR on Line 14 is $0.02399 per 18
kWh or $23.99 per MWh. The 2017 LCAR is $25.99 per MWh 19
and the calculation is shown on Exhibit No. 13, pages 3 20
and 4 of Schedule 1. The calculation of the LCAR will be 21
revised based on the final production and transmission 22
costs, and rate of return, that are approved by the 23
Commission in this case. 24
Knox, Di Page 9 Avista Corporation
IV. ELECTRIC COST OF SERVICE 1
Q. Please briefly summarize your testimony related 2
to the electric cost of service study. 3
A. I believe the Base Case cost of service study 4
presented in this case is a fair representation of the 5
costs to serve each customer group. The Base Case study 6
shows Residential Service Schedule 1 provides less than 7
the overall rate of return under present rates. All of 8
the other service schedules provide more than the overall 9
rate of return under present rates to varying degrees. 10
Q. What is an electric cost of service study and 11
what is its purpose? 12
A. An electric cost of service study is an 13
engineering-economic study, which separates the revenue, 14
expenses, and rate base associated with providing electric 15
service to designated groups of customers. The groups are 16
made up of customers with similar load characteristics and 17
facilities requirements. Costs are assigned or allocated 18
to each group based on (among other things), test period 19
load and facilities requirements, resulting in an 20
evaluation of the cost of the service provided to each 21
group. The rate of return by customer group indicates 22
whether the revenue provided by the customers in each 23
group recovers the cost to serve those customers. The 24
Knox, Di Page 10 Avista Corporation
study results are used as a guide in determining the 1
appropriate rate spread among the groups of customers. 2
Schedule 2 of Exhibit No. 13 explains the basic concepts 3
involved in performing an electric cost of service study. 4
It also details the specific methodology and assumptions 5
utilized in the Company’s Base Case cost of service study. 6
Q. What is the basis for the electric cost of 7
service study provided in this case? 8
A. The electric cost of service study provided by 9
the Company as Exhibit No. 13, Schedule 3 is based on the 10
twelve months ended December 31, 2014 test year pro forma 11
results of operations presented by Ms. Andrews in Exhibit 12
No. 12, Schedule 1. 13
Q. Would you please explain the cost of service 14
study presented in Exhibit No. 13, Schedule 3? 15
A. Yes. Exhibit No. 13, Schedule 3 is composed of 16
a series of summaries of the cost of service study 17
results. The summary on page 1 shows the results of the 18
study by FERC account category. The rate of return by 19
rate schedule and the ratio of each schedule’s return to 20
the overall return are shown on Lines 39 and 40. This 21
summary was provided to Company witness Mr. Ehrbar for his 22
consideration regarding rate spread and rate design. The 23
Knox, Di Page 11 Avista Corporation
results will be discussed in more detail later in my 1
testimony. 2
Pages 2 and 3 are both summaries that show the 3
revenue-to-cost relationship at current and proposed 4
revenue. Costs by category are shown first at the 5
existing schedule returns (revenue); next the costs are 6
shown as if all schedules were providing equal recovery 7
(cost). These comparisons show how far current and 8
proposed rates are from rates that would be in alignment 9
with the cost study. Page 2 shows the costs segregated 10
into production, transmission, distribution, and common 11
functional categories. Line 44 on page 2 shows the target 12
change in revenue which would produce unity in this cost 13
study. Page 3 segregates the costs into demand, energy, 14
and customer classifications. Page 4 is a summary 15
identifying specific customer-related costs embedded in 16
the study. 17
The Excel model used to calculate the cost of service 18
and supporting schedules has been included in its entirety 19
both electronically and in hard copy in the workpapers 20
accompanying this case. 21
Q. Given that the specific details of this 22
methodology are described in the narrative in Exhibit No. 23
13, Schedule 2, would you please give a brief overview of 24
Knox, Di Page 12 Avista Corporation
the key elements and the history associated with those 1
elements? 2
A. Yes. Production costs are classified to energy 3
and demand in this case based on the system load factor. 4
The Company proposed this approach in Case No. AVU-E-11-5
01. While the Company chose to use the traditional 6
replacement-cost-based peak credit methodology in the last 7
Idaho case (Case No. AVU-E-12-08), we believe that the 8
system load factor method is preferable, as I discuss 9
later in my testimony. 10
Transmission costs are classified as 100% demand and 11
allocated by the average of the 12 monthly coincident 12
peaks. This methodology is the same treatment as the last 13
Idaho case (Case No. AVU-E-12-08) and reflects the 14
methodology accepted in the Settlement in Case No. AVU-E-15
10-01. 16
Distribution costs are classified and allocated by 17
the basic customer theory5 accepted by the Idaho Commission 18
in Case No. WWP-E-98-11. Additional direct assignment of 19
demand-related distribution plant has been incorporated to 20
reflect improvements accepted by the Commission in Case 21
No. AVU-E-04-01. 22
5 Basic customer cost theory classifies only meters, service lines from the distribution system to the customer’s premise, and street lights as customer-related plant; all other distribution facilities are
considered demand-related.
Knox, Di Page 13 Avista Corporation

Administrative and general costs are first directly 1
assigned to production, transmission, distribution, or 2
customer relations functions. The remaining administra-3
tive and general costs are categorized as common costs and 4
have been assigned to customer classes by the four-factor 5
allocator accepted by the Idaho Commission in Case No. 6
AVU-E-04-01. 7
Q. Does the Company’s electric Base Case cost of 8
service study follow the methodology filed in the 9
Company’s last electric general rate case in Idaho? 10
A. Yes, with one exception. The peak credit 11
methodology used for classification of production costs 12
into energy-related and demand-related categories is 13
different from the traditional peak credit determination 14
presented in the last Idaho general rate case. 15
Q. What is the Company proposing in this case with 16
regard to the peak credit methodology? 17
A. In this case the Company is proposing to use the 18
system load factor to determine the proportion of the 19
production function that is demand-related.6 This peak 20
credit ratio is then applied uniformly to all production 21
costs. This is the same method the Company proposed in 22
6 One minus the load factor equals the demand percentage or peak credit
ratio.
Knox, Di Page 14 Avista Corporation

Case No. AVU-E-11-01 that was derived from ideas developed 1
through cost of service workshops held at the Idaho 2
Commission in February 2011 and September 2012. 3
Q. What do you believe are the benefits of using 4
the system load factor to determine the peak credit ratio? 5
A. There are several benefits to the system load 6
factor approach for identifying the demand-related 7
proportion of production costs: 1) It is simple and 8
straightforward to calculate; 2) it is directly related to 9
the system and test year under evaluation; and 3) the 10
relationship should remain relatively stable from year to 11
year. 12
Q. How was the peak credit methodology determined 13
and applied in past studies? 14
A. In the Company’s cost of service studies prior 15
to 2010 and the 2012 case (AVU-E-12-08), Avista’s electric 16
system resource costs were classified to energy and demand 17
using a comparison of the replacement cost per kW of the 18
Company’s peaking units, to the replacement cost per kW of 19
the Company’s thermal and hydro plants (separately). This 20
analysis created separate peak credit ratios applied to 21
thermal plant and hydro plant. Fuel and load dispatching 22
expenses were classified entirely to energy, and peaking 23
plant related costs were classified entirely to demand. 24
Knox, Di Page 15 Avista Corporation
Q. What is the net effect of the proposed change in 1
the peak credit method? 2
A. The net effect of this change is to increase the 3
overall production costs that are classified as demand-4
related. Using the prior method, approximately 31.28% of 5
total production costs would be classified as demand-6
related. Under the proposed method, 37.93% of total 7
production costs are classified as demand-related. In 8
this circumstance, costs are shifted toward the low load 9
factor residential and small commercial class, and away 10
from all the other classes. However, the impact on the 11
cost study results is relatively minor. The shift in 12
costs at unity were less than 1% of present revenue for 13
all schedules except Schedule 25P where costs at unity 14
were reduced by 1.5%. 15
Q. Did the Company use recent load research for 16
demand-related cost allocations in the electric cost of 17
service study in this case? 18
A. Yes. The Company contracted with DNV-GL7 to 19
develop hourly load estimates by rate class for cost of 20
service demand allocation purposes. The study was 21
completed in December 2014 with the final report provided 22
7 The DNV-GL Group (Det Norske Veritas, Germanischer Lloyd) is an industry leader in load research for the energy sector, providing comprehensive services from initial study planning to analysis and
final reporting.
Knox, Di Page 16 Avista Corporation

in January 2015. The new load research study, included 1
with the Company’s workpapers in this filing, utilized 2
sample metering in place from the last load study 3
(discussed in Case No. AVU-E-10-01, the Company’s 2010 4
general rate case) augmented by additional sample sites 5
added during the intervening years. The study is based on 6
data collected over the period July 1, 2013 through June 7
30, 2014. 8
Q. Did the 2014 load study show any major changes 9
in usage across the customer classes? 10
A. No, other than the change from a purchase and 11
sale contract to self-generation for Schedule 25P, there 12
were no major changes. The study did capture the impact 13
of schedule shifting from Schedule 21 to Schedule 11 that 14
occurred several years ago and the residential class shows 15
a slightly higher contribution to the peaks than in the 16
2009 study. In general the results were consistent with 17
prior study results. 18
Q. What are the results of the Company’s electric 19
cost of service study presented in this case? 20
A. Illustration No. 1 below shows the rate of 21
return and the relationship of the customer class return 22
to the overall return (relative return ratio) at present 23
rates for each rate schedule: 24
Knox, Di Page 17 Avista Corporation
Illustration No. 1: 1
Customer Class Return Ratio
Paper Schedule 25P 9.20% 1.41
As can be observed from the above table, Residential 2
service Schedule 1 shows under-recovery of the costs to 3
serve them. The Extra Large General service Schedule 25, 4
the Pumping service schedule (31/32) and the Lighting 5
service schedules (41-49) are slightly over, but very near 6
unity. The General, Large General and Extra Large 7
General-Clearwater Paper service schedules (11/12, 21/22, 8
and 25P) show over-recovery of the costs to serve them. 9
The summary results of this study were provided to Mr. 10
Ehrbar for consideration in the development of proposed 11
rates. 12
Q. Does this conclude your pre-filed direct 13
testimony? 14
A. Yes. 15
Knox, Di Page 18 Avista Corporation