20040220Yankel Direct.pdf

pcrFI\J,\1-"-,,
! 1 1::-0
~..
"""'I rn
'"'
r,'! lUd;~ ' en ._- u Hll
' . " ,
i G
jr\LITli~S co iSSIOf~
BEFORE THE IDAHO PUBLIC UTILITIES COMMISSION
IN THE MATTER OF THE APPLICATION
OF IDAHO POWER COMPANY FOR
AUTHORITY TO INCREASE ITS INTERIM
AND BASE RATES AND CHARGES FORELECTRIC SERVICE
CASE NO. IPC-03-
IDAHO IRRIGATION PUMPERS
DIRECT TESTIMONY OF
ANTHONY 1. Y ANKEL
February 20 2004
TABLE OF CONTENTS
IDAHO POWER'S GENERAL RATE CASE
Case No. IPC-O3-
Testimony of Anthony J. Yankel
Page
INTRODUCTION ................................................
TEST YEAR EXPENSES
... ......................................
ALLOCATION OF EXPENSES TO CLASSES... . .. .
. . . . . . . ..
.. . . . . . . . 18
Overview..................................................
Modelingproblems..........................................
MethodologyProblems.......................................
DataProblems.............................................
RateDesign...............................................
SUMMARYOFRECOMMEDATIONS... ...........................
INTRODUCTION
PLEASE STATE YOUR NAME, ADDRESS, AND EMPLOYMENT.
I am Anthony J. Yanke!. I am President of Yanke I and Associates, Inc. My
address is 29814 Lake Road, Bay Village, Ohio, 44140.
WOULD YOU BRIEFLY DESCRIBE YOUR EDUCATIONAL
BACKGROUND AND PROFESSIONAL EXPERIENCE?
I received a Bachelor of Science Degree in Electrical Engineering from Carnegie
Institute of Technology in 1969 and a Master of Science Degree in Chemical Engineering from
the University ofIdaho in 1972. From 1969 through 1972, I was employed by the Air
Correction Division of Universal Oil Products as a product design engineer. My chief
responsibilities were in the areas of design, start-up, and repair of new and existing product lines
for coal-fired power plants. From 1973 through 1977, I was employed by the Bureau of Air
Quality for the Idaho Department of Health & Welfare, Division of Environment. As Chief
Engineer of the Bureau, my responsibilities covered a wide range of investigative functions.
From 1978 through June 1979, I was employed as the Director of the Idaho Electrical Consumers
Office. In that capacity, I was responsible for all organizational and technical aspects of
advocating a variety of positions before various governmental bodies that represented the
interests of the consumers in the State of Idaho. From July 1979 through October 1980, I was a
partner in the firm of Yanke I, Eddy, and Associates. Since that time, I have been in business for
Yankel, DI
Irrigators
myself. I am a registered Professional Engineer in the states of Ohio and Idaho. I have
presented testimony before the Federal Energy Regulatory Commission (FERC), as well as the
State Public Utility Commissions ofIdaho, Montana, Ohio , Pennsylvania, Utah, and West
Virginia.
ON WHOSE BEHALF ARE YOU TESTIFYING?
I am testifying on behalf of the Idaho Irrigation Pumpers Association
Irrigators
WHAT IS THE PURPOSE OF YOUR TESTIMONY IN THIS PROCEEDING?
The purpose of my testimony is to propose an adjustment to the expense levels
that Idaho Power has filed in this case and address data and other problems with the Company
class cost-of-service study.
PLEASE SUMMARIZE YOUR TESTIMONY.
A. 1. With respect to the Company s filed 6-month actual, 6-month budgeted non-fuel
expenses, I found that there was no removal of inappropriate expenses or
normalization. The Company s historic data shows multiple instances where
expenses are abnormally high one month, only to be reversed in a following period.
As a result, the reliance on specific monthly data is not reflective of overall costs. I
Yankel, DI
Irrigators
recommend that the Company s filed non-fuel expenses be reduced by $5 794 724 in
order to reflect the levelized trend in these costs over the previous five years.
2. With respect to class cost-of-service, I address a host of problems, starting with the
fact that the Company s cost-of-service study produces erroneous and unreliable
results.
3. I address the allocation methodology proposed by the Company for generation and
transmission costs. I recommend the continued use of an unweighted 12-CP method
as is used for jurisdictional allocation purposes.
4. I next address the lack of ability to review in any meaningful way the Company s load
research data that is contained in a proprietary program to which none ofthe parties in
this case have access. This is followed by a discussion of the mismatch between using
actual 2002 peak load data to define demand responsibility while using lower
revenues, based upon 2003 normalized energy. After this I address specific problems
or oversights with the way some distribution costs are allocated to Irrigators.
5. Although I address a number of problems, I cannot provide an alternative allocation
method for each. As a result of those areas where a specific allocation method could
be developed, a rate of return for the Irrigation class was calculated that is above the
iurisdictional average. My ultimate recommendation with respect to class cost-of-
service is that far more study needs to be done, and until that is done, there should be
an even percentage increase to all classes. I further address the rate design of
Irrigation customers and recommend that no more than the average rate increase be
placed upon the demand and customer charges for the Irrigation class.
Yankel, DI
Irrigators
TEST YEAR EXPENSES
WHAT IS YOUR UNDERSTANDING OF THE COMPANY'S NON-POWER
O&M EXPENSE DATA (ACCOUNTS 500-935) FILED IN THIS CASE?
The Company filed 6-months of "actual" data and 6-months of budget data.
Subsequently, the Company provided 9-months of actual data (January-September 2003) and 3-
months of budget data (October-December 2003) in response to a Staff data request. It is my
understanding from Exhibit 26 that there were no adjustments to this data for normalizing any of
the non-power O&M expenses involved.
Q. IS THIS AN ACCEPTABLE PROCEDURE?
A. No. First
, "
actual" data is more accurate and reliable to use than budget data because
it is known and measurable. Budget data that is at the end of the year is even less reliable
because it is further removed from the time the budget was actually made. Second, and more
importantly, there were no adjustments made to either the budget or "actual" data that would
normalize/remove the impact of non-routine events. There is no assurance that all of these
expenses are reflective of normal, ongoing costs.
Q. HAS THE COMMISSION EVER ADDRESSED THE ISSUE OF NOT RELYING
UPON HISTORICAL TEST-YEAR DATA?
Yankel, DI
Irrigators
A. Yes. In Idaho Power s last rate case the Irrigators questioned the use of the historical
expense data. After reviewing the evidence in the last case, the Commission stated:
Although fair questions were raised by Irrigators regarding the increases in
1993 O&M expenses, no evidence was presented to show these expenses were
improperly or artificially inflated, nor was it demonstrated that they would be
less in the future. The Commission has traditionally relied on historical test-
year data adjusted for specific known and measurable adjustments and has
with the exception of weather and stream-flow sensitive revenue and expenses
rejected adjustments to historical data based strictly on statistical analyses. We
find no reason to change that policy in this case. Accordingly, we will not
adjust the 1993 O&M expenses.
Q. DO YOU ALSO TAKE ISSUE WITH THE "ACTUAL" COSTS FILED BY THE
COMP ANY?
A. Yes. There are essentially two problems. First, although these expenses are labeled
as "actual", in fact they appear to be merely "as-booked" expenses that may reflect estimates or
in some other way be reversed at a later time. Second, just because an expense is "actual" that
does not mean that it is reflective of normal operations.
Q. PLEASE GIVE AN EXAMPLE OF WHY THE USE OF UNADJUSTED
ACTUAL" EXPENSES MAY NOT BE APPROPRIATE FOR RATE MAKING PURPOSES.
A. Figure 1 depicts "actual" expenses by month for Account 500 (Operation and
Supervisory Engineering-Steam) for the five years prior to the test year.
1 Order No. 25880 at page 6.
Yankel, DI
Irrigators
$300 000
$200 000
$100 000
$( 100 000)
Figure 1
Acct. 500 (1998-2002)
C\I C")C")-.;tex)
..-
C\IC\I
Month
Note that although there was a steady decline in these expenses, there were significant
fluctuations in individual monthly data around this decline. Depending upon which 6 , or 12-
month period one chooses, the outcome could be very different.
Figure 2 depicts the same data for Account 500 with the addition of the six months of
actual data" that was filed by Idaho Power in this case.
Yanke1, DI
Irrigators
Figure 2
Acct. 500 (5-Years plus 6-months filed)
500 000
000 000
$500 000
I'-
("')
L()
C\I
("')
oq-
L()
L()oq-
I'-
("')("')
$(500 000)
Month
Although the data for the first 60 months listed on Figure 2 is exactly the same as that
found on Figure 1 , these graphs look dramatically different because the magnitude of the
expenses listed in the first 6-months of the test year are significantly different than that witnessed
over the previous five years. Although there were wide fluctuations from month to month during
the previous five years, these fluctuations become pale in comparison to the increases that were
found in the "actual" data as originally filed by the Company. The June 2003 expense for
Account 500 could be considered "off the chart" and yet, there was no adjustments made for
abnormal expenses.
Figure 3 depicts the same data for Account 500 with the addition of 3 months of "actual"
data as obtained from the Company s response to Staff Request 85.
Yankel, DI
Irrigators
Figure 3
Acct. 500 (5-years plus 9-months actual)
500 000
000 000
$500 000
$(500 000)
$(1 000 000)
$(1 500 000)
$(2 000 000)
I'-
("')
Months
The addition of 3 more months of "actual" data introduces further fluctuations into these
expense values with an extremely large negative entry, followed by a second negative entry. To
a large extent, the negative entries in August and September 2003 erase the impact of the
extremely high positive entries in June and July 2003. However, note that the Company does not
use these three additional months in its case.
Q. HAS THE COMPANY OFFERED ANY EXPLANATION REGARDING THE
DRAMATIC SWINGS IN THE DATA FOR ACCOUNT 500?
Yankel, DI
Irrigators
A. Yes. The Company s response to Irrigator Request 22 addresses these data swings.
First, the Company stated that the declining trend is driven by decreases in labor costs. With
respect to the large fluctuations that started in June 2003 , the Company offered the following:
The actual charges for Account 500 through June 2003 include a large entry in
June for the Bridger Plant. This entry, for $1 134 000 represents a temporary
accrual of annual overhaul costs. This accrual has been reversed in the
following months. Following this reversal, the annual charges to Account 500
for Bridger should be similar to the levels experienced in 2001 and 2002.
Accordingly, the "actual" costs used by the Company are "as booked" by the Company and may
be considered "actual", but they are subject to change/reversal in a later period. Thus, looking at
a limited timeframe the Company s "actual" costs at times do not reflect ongoing expenses.
Q. PLEASE GIVE A SECOND EXAMPLE OF WHY THE USE OF UNADJUSTED
ACTUAL" EXPENSES MAY NOT BE APPROPRIATE FOR RATE MAKING PURPOSES.
A. Figure 4 depicts actual expenses for Account 502 (Steam Expense) for the 5-years prior
to the test year, the 6-months of "actual" data filed by the Company, plus 3 additional months.
Y ankel, DI
Irrigators
Figure 4
Acct 502 (5-years plus 9-months actual)
000,000
500 000
000 000
500,000
CD
..-..-
(500 000)
000,000)
500 000)
..- CDCD
Months
As with Account 500, it can be seen that although there is a relatively consistent level of
expense associated with Account 502, there can also be huge variations around this level. This
variation can be seen to be as great as $1 500 000 per month.
As with Account 500, even when using "actual" data, the use of a different 6 , or 12-
month period can produce very different results with respect to what level of expense is used for
ratemaking purposes. As can be seen from Figure 4, the "as filed" expenses for Account 502
have some months that are significantly higher than the overall trend. If the next three months of
actual data are included (the last three data points on Figure 4), it greatly reduces the impact of
what appears to be abnormally high expenses for April, May, and June 2003. However, these
last three months are not included in the Company s filing.
Yankel, DI
Irrigators
Q. WHAT EXPLANATION HAS THE COMPANY OFFERED REGARDING THE
DRAMATIC SWINGS IN THE DATA FOR ACCOUNT 502?
A. The Company s response to Irrigator Request 21 addresses these data swings. The
explanation is similar to that offered regarding the swings in the data for Account 500.
FERC Account 502 contains the costs of operating the boiler and
associated equipment at each of Idaho Power s three jointly owned steam plants.
The costs in this account are based on the amounts reported to Idaho Power by the
operating partner at each plant. The significant components of the expenses
booked in Account 502 are chemicals to maintain the quality of the water used to
produce steam and the labor costs to operate the boiler. The chemical costs vary
with changes in generation, while the labor costs remain fairly constant. For
accounting purposes, we estimate O&M expenditures for the month and then
adjust those estimates to the actual amounts in the following month when the
actual amounts are known. For simplicity, only a few accounts are used for
estimating purposes: Account 502 is one of those accounts. This estimating
process can cause year-to-year variations.
In 2000, Account 502 increased $1 284 718 over 1999. All of this
increase can be attributed to the Bridger Plant, which increased $1 302 415 over
1999. The increase at Bridger is due to an increase in actual charges of$577 000
a reversal of accruals of $214 000 made in 1999 that did not recur in 2000, and an
increase in the estimate of expenses for December that was $506 000 greater than
the corresponding period in1999.
In 2002 Account 502 decreased $1 751 360 - Bridger decreased
150 000 and Valmy decreased $602 000. Bridger s decrease is the result ofa
decrease in actual charges of $436 000 and a decrease in the amount estimated for
December 2002 of $804 000 over December 2001. The decrease at Valmy is due
to actual charges for 2002 being $508 000 less than 2001 charges, while the
reversal of the prior months estimate in January of 2002 was $97 000 greater than
the corresponding period in 2001.
Once again, just because a particular level of expense is "booked" in one month, does not
mean that it will not be reversed in the future. Additionally, some of the expenses "booked" may
in fact be estimates that are to be trued up at a later time. Obviously, there are wide variations in
Yankel, DI
Irrigators
the level of expenses from year to year, even in an account such as 502 that only contains the
cost of operating (not maintaining) the boilers.
Q. ARE THERE EXPENSES CONTAINED IN ACCOUNTS 500-935 THAT DO NOT
REFLECT NORMAL EXPENSE?
A. Yes. Unfortunately, the Company has not identified in its filing (nor has it made any
adjustment) to reflect abnormal expenses. One such area of abnormal expenses can be found in
Account 536 (Water for Power). Figure 5 illustrates the monthly expenses "booked" to this
account for the five years prior to the test year as well as the first 9-months of the test year.
Figure 5
Acct. 536 (5-years previous plus 9-months actual)
000 000
$900 000
$800 000
$700 000
$600 000
$500 000
$400 000
$300 000
$200 000
$100 000
I'-
T'""
I'- C"')C"') C"') ~
Month
I'-L()
L()
L()C"')
T'""
Yankel, DI
Irrigators
Not only from Figure 5, but also from the response to Irrigator Request 24, it can be seen
that this Account 536 contains a lot of abnormal activity:
The decreasing trend in Account 536 expenses since 1997 can be traced to
three main factors: refinanced American Falls debt obligation, conclusion of the
amortization ofFERC headwaters benefits, and reduced incentive royalty
payments for Milner.
Idaho Power is realizing lower interest expenses related to the American
Falls debt obligation. This obligation was refinanced with a variable rate
instrument in 2000 and interest rates have been very low recently.
An assessment for use of Federal water impoundments (headwater
benefits) was instituted in the late 1980's. Retroactive FERC headwater benefits
through 1988 were computed and paid in 1990. Based on IPUC Order 23224, this
cost was amortized over 120 months. In 2000, the amortization period finished
and expense dropped accordingly.
Drought conditions throughout Idaho in recent years have affected
streamflows and generation at Milner. Since 2000, generation has fallen below
the threshold requiring payment of incentive royalties to Milner Dam Inc.
Account 536 expenses in the test year are higher than the preceding two
years due to the cloud seeding program. The cloud seeding program was started
in late 2002, but the test year (2003) includes expenses for a full year of
operation. Cloud Seeding Program costs more than offset the expense reductions
previously noted.
Obviously, the "norm" for this account is anything but consistency. Cloud seeding is not
a regular activity and should not be included as a normal test year expense. Likewise, including
the full incentive royalty payments for Milner Dam Inc. should possibly be reflected.
The problem is defining a level of expense that can be considered normal for test year
purposes. A possible solution would be to define an overall level of expense that is "normal"
without specifying the exact amount that must be included in each FERC account for the exact
expense for every item on the Company s books.
Yankel, DI
Irrigators
Q. WHAT IS YOUR RECOMMENDATION FOR THE RATE MAKING
TREATMENT OF NON-POWER COST ACCOUNTS 500-935 EXPENSES SO THAT THEY
ARE REFLECTIVE OF COSTS FOR RATE MAKING PURPOSES?
A. As pointed out above
, "
as booked" values may not be as reflective of expenses that
one would wish to use for rate making purposes. Additionally, "as booked" expenses may
contain expenses that are not reflective of normal operations. These abnormalities in "
booked" costs can best be smoothed out by using an historic trend for these expenses.
In order to do this, I have compiled the data for each month of the previous five years for
all non-power costs accounts (excluding Accounts 501 , 547, 555, and 557). Figure 6 illustrates
the summation of these accounts and the trend that results. It should be remembered that these
are still "as booked" values so the major fluctuations that have been demonstrated above are still
included in the monthly data. The purpose of this analysis is not to remove the individual
fluctuations, but to define the overall trend that exists in spite of individual monthly variations.
Yankel, DI
Irrigators
$25 000 000
$20 000 000
$15 000 000
$10,000 000
$5,000,000
Figure 6
All Non-Power Costs Test
Year
~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~ ~
Month
Q. HOW DO THE ACTUAL EXPENSES INCURRED DURING THE FIRST NINE
MONTHS OF THE TEST YEAR (AS WELL AS THE REMAINING 3-MONTHS OF
BUDGET DATA) COMPARE WITH THIS TREND ANALYSIS?
A. As can be observed from Figure 6, the test year data and the trend data are generally
in agreement. Exhibit 301 contains a numerical comparison ofthe values derived from the trend
analysis with the budgeted and actual data filed by the Company. As can be seen from column
B" of Exhibit 301 , the "as filed" expense data (6-months actual and 6-months budget) are
794 724 or about 2.7% greater than that derived by the trend analysis. The data in column
C" of Exhibit 301 indicates that if the "as filed" data were updated to 9-months actual and 3-
months budgeted data, then these expenses would exceed the trend by $4 788 244 or 2.2%.
Yankel, DI
Irrigators
Q. WHAT ARE YOUR RECOMMENDATIONS WITH RESPECT TO MAKING AN
ADJUSTMENT TO THE "AS FILED" TEST YEAR EXPENSES?
A. I recommend that the "as filed" values (actual and budget) not be used for rate
making purposes. As demonstrated above, even the Company s actuallbooked data is subject to
estimation and reversal. The trend analysis over the last five years gives a much clearer, long-
term picture of the expenses that are being incurred on a normal basis. I recommend that at a
minimum the Company s "as filed" non-power cost expense data be reduced for rate making
purposes by $5 794 724.
PLEASE EXPLAIN HOW YOU ARRIVED AT YOUR PROPOSED $5.
MILLION REDUCTION IN EXPENSES.
A. Using the 5-years of previous monthly data, I let the computer do a trend analysis.
Graphically, this is represented by the straight line on Figure 6. The computer analysis also gave
me an equation that represented this line. Using this equation, I calculated the amount of
expense that would be expected if the trend over the previous five years were expected to
continue. Note, calculated expense for each month increases as the trend over the past 5-years
has been increasing. The results for the individual months of the test year from the trend
analysis are listed on Exhibit 301.
Q. WHY DO YOU CONSIDER THE ABOVE ADJUSTMENT TO BE A
MINIMUM"
Yankel, DI
Irrigators
A. Although I reviewed far more accounts than I have addressed in my testimony, I did
not make a detailed review of all of the Company s accounts. There could very likely be
accounts where data was included in my trend analysis that was higher than normal or not
normal, like cloud seeding. In fact, from my experience, I assume that many additional (non-
normal) costs were incurred during the drought conditions since 2000 that are included in my
trend analysis. For example, the Company s response to Irrigator Request 21 addresses the
swings in Account 502 expenses and it states in part:
The decrease in expense in 1995 from 1994 of$840 596 is a result of decreased
generation at each of the plants primarily due to good hydro conditions, resulting
in reduced chemical costs.
Obviously, the drought conditions ofthe last few years may have greatly added to the
Company s non-power cost expenses, and thus, added expenses to this trend analysis.
Yankel, DI
Irrigators
ALLOCA TI ON OF EXPENSES TO CLASSES
Overview
Q. PLEASE GIVE AN OVERVIEW OR REALITY CHECK OF WHAT THE
COMPANY IS PROPOSING IN SUPPORT OF THE PROPOSED HUGE AND
DISPROPORTIONATE RATE INCREASE FOR THE IRRIGATION CLASS
Idaho Power claims that it is attempting to prevent "rate shock" by limiting the rate
increase to the Irrigation class to "only a 25% increase" compared to the 67.1 % increase that the
Company s cost-of-service study claims is needed from these customers. Ifthe Company s cost-
of-service results were accurate, this translates into a rate of 62 mills kWh for Irrigators and 59
mills per kWh for Residential customers . Thus, Residential customers would be paying less for
electricity than Irrigation customers, in spite of the fact that Irrigation customers:
1. Are large users at a single location, with lower distribution and customer
related costs per kWh;
2. Do not even use the secondary distribution system;
3. Generally have a much higher load factor (ratio of average use to non-
coincident use); and
4. Generally have a much higher coincident factor (ratio of average use to
coincident peak use).
A similar comparison can be made with the rates that are shown by the Company s cost-
of-service study for the Schedule 9-Secondary customers. The Company s cost-of-
service study shows that these customers should only be paying 39 mills compared to the
62 mills shown for the Irrigation customers. Thus, Irrigators would be paying almost
2 Company Exhibit 41 page 1 line 237.
Yankel, DI
Irrigators
60% higher rates for service in spite of the fact that when compared to the Schedule 9-
Secondary customers Irrigators:
1. Use approximately the same level of energy per customer;
2. Do not even use the secondary distribution system (as is recognized in the
Company s cost-of-service study);
3. Generally have a higher load factor (ratio of average use to non-coincident
use); and
4. Generally have a higher coincident factor (ratio of average use to
coincident peak use).
It is absurd to suggest that Irrigators should pay 60% more for energy than the Schedule 9
Secondary customers as well as more than Residential customers. According to the Company
2002 FERC Form 1 the Irrigators are paying an average of 51 mill per kWh-they would be very
happy to pay the 39 mills per kWh that the Company s cost-of-service study suggests the
Schedule 9-Secondary customers should pay. This clearly demonstrates there are major flaws in
the Company s cost allocation methodology and/or data.
Q. ARE THERE OTHER RATES THAT YOU CAN POINT TO THAT COULD SHED
ADDITIONAL LIGHT UPON THE ILLOGICAL RESULTS THAT ARE PRODUCED BY
THE COMPANY'S COST-OF-SERVICE STUDY?
A. Yes. This Commission is well acquainted with PacifiCorp as it regulates that utility'
service area in Idaho. I have spent the last 15 years working for the Committee of Consumer
Services (The Committee) in Utah on various PacifiCorp cases. The Committee represents
residential, small commercial, and irrigation interests in Utah. I have testified on behalf of these
rate classes as a witness for the Committee. There have been a number ofPacifiCorp rate cases
Yankel, DI
Irrigators
in Utah over the last 15 years. These have provided regular and thorough review ofPacifiCorp
allocation methodology, models, and data. Additionally, various Task Forces have been
established where such things as data analysis and load research data were thoroughly reviewed
in a non-rate case setting by technical people representing various parties. The net result of all of
this activity has been:
A clear understanding that the data that is being incorporated into the cost-of-
service study is reliable;
A reasonable assurance that most of the data being utilized is reflective of cost
causation; and
A recognition that there are limits to the use and validity of some of the data
especially the load research data.
After all of this review, the rates in Utah for Residential (R1) customers averaged 67
mills, while for Irrigation (R24) customers they averaged 43 mills, i.e. Thus, Residential rates in
Utah are about 60% more expensive than the Irrigation rates in Utah. This is not a discrepancy
in rates that needs to be changed over time. There is an agreement in place between the various
parties in Utah that reviewed PacifiCorp s cost of service data and load research data.
exchange for not needing to collect further load research data for the Irrigation class in Utah, the
Irrigation customers will simply be given the average system rate change in all future cases.
Thus, the relative difference between Irrigation rates and Residential rates in Utah will remain
relatively unchanged with Irrigation rates being significantly below Residential rates.
Idaho Power has not undergone this level of scrutiny. As I will discuss later in my
testimony, there are major gaps in the ability of the Company s cost-of-service study to produce
results that can or should be relied upon for setting rates.
Yankel, DI
Irrigators
Q. IS GROWTH IN IRRIGATION LOAD FUELING THE NEED FOR A RATE
INCREASE?
A. No. Idaho Power projects the need for significant plant additions over the next ten
years. During the next three years alone, it estimates that its capital expenditures will be $675
million . The Company projects an annual growth rate over the next 10 years of2.3% per year
This annual growth rate is broken down as follows:
Residential 2.4%
Commercial 4.1 %
Industrial 2.4%
Irrigation 0.4%
Even this small projected growth in the Irrigation load in light of current proceedings before the
Idaho Department of Water Resources and the Snake River Basin Adjudication Court which may
result in an actual decline in irrigation from groundwater in order to satisfy prior surface water
rights. It is also my understanding that Idaho Power projected zero growth in normalized
Irrigation load between 2002 and 2003. Clearly, the Irrigators are not primary drivers in the need
for additional capital expenditures. There is no basis from a marginal cost or forward looking
perspective to significantly increase rates to Irrigators.
Q. HAVE YOU REVIEWED THE BASIS FOR THE COMPANY'S RATE SPREAD
PROPOSAL?
3 Keen testimony at page 27.
4 *Idaho Powers 2002 IRP page 9.
Yankel, DI
Irrigators
A. Yes. I have reviewed the Company s cost-of-service computer model and its various
methodologies for functionalizing, classifying, and allocating costs, as well as the specific data
sources used to supply the data for the model.
Q. HAVE YOU FOUND ANY PROBLEMS WITH THE VARIOUS COMPONENTS
THAT WENT INTO DEVELOPING THE COMPANY'S PROPOSED RATE SPREAD?
A. Yes, I have found major problems with virtually all components the Company used to
develop its rate-spread proposal. Most of the problems are not class specific but universal to all
classes or multiple classes. By reason of these widespread problems, it is impossible to define
cost causation in any accurate and reliable way. Accordingly, the only fair and reasonable thing
to do is to spread any rate increase evenly to all customer classes. If the Commission is
interested in adjusting rates between the customer classes, a separate procedure should be
initiated to address these technical problems.
Modeling Problems
Q. HAVE YOU DISCOVERED ANY PROBLEMS WITH THE COMPANY'S COST-
OF-SERVICE COMPUTER MODEL IN THIS CASE?
A. Yes, the Company s computer model is fatally flawed and cannot be relied upon to
give accurate results, even if one were to agree with all allocation methodologies and data inputs.
I am not speaking here of the data used or a difference of opinion on how to allocate costs. I am
speaking of computational or logic errors within the program that cause it to produce counter-
Yankel, DI
Irrigators
intuitive results. An example of such a counter-intuitive result would be to have a class ' rate of
return increase or decrease when it should not have, given input data changes. The lack of an
accurate computer model obviously affects all customer classes.
Q. HAVE YOU ATTEMPTED TO FIND THE ERROR OR ERRORS IN THE
COMP ANY'S COMPUTER MODEL IN ORDER TO CORRECT THIS PROBLEM?
A. I have made such an attempt, but the Company s model is far from "user friendly
and such a task is well beyond the capability of most intervenors or the Staff to address. I have
been doing cost-of-service studies for over 25 years and I have never been faced with such a
computer model that was impossible to follow (except in the last case where Idaho Power used
essentially this same model). A cost-of-service model should be no more than a huge collection
of simple additions, subtractions, multiplications, and divisions. The Company s cost-of-service
model is little better than a "Black Box
When I discovered in the course of my investigation in this case that the Company
model was not producing reliable results, I began to look into the model further. Exhibit 30 is
the Company s jurisdictional cost-of-service study in this case and it allocates costs by FERC
account in a traditional and understandable manner. Exhibit 39 consists of 42 pages and it is
where the Company demonstrates how it allocated costs to various classes. As can be seen on
page 3 through 40 of that exhibit, the Company s class cost-of-service study does not allocate
costs based upon FERC account, but by functional categories that were developed in Exhibit 37.
One would think that this two-step procedure should be a small complication, but one
that can be easily followed. In reality, it is far more complicated. The Company also provided a
Yankel, DI
Irrigators
copy of its class cost-of-service model in electronic form so that the formulas and logic could be
readily followed. At this point I began to run into the "Black Box . Although the Company
computer model produced the 42 pages that are contained on Exhibit 39, only 14 of these 42
pages can be found in the computer model. I am told that the model somehow writes over itself
so that it produces some information like page 3 (allocation of plant in service) and then
immediately destroys it. It is impossible to see how the computer links the various data in the
different cells together in order to produce the results that are presented.
The complete lack of transparency of this model does all parties a disservice. One can
only proceed on faith that the model is accurately defining the cost causation as prescribed by the
methodologies and data advocated by various parties. It is completely inappropriate to say that
Irrigation customers, or any group of customers should have their rates increased by "X" amount
and that one has to take the increase based upon faith. When the model begins producing
erroneous results, one must abandon all faith in the model.
Q. HAVE YOU DEVELOPED A COST-OF-SERVICE MODEL IN THIS CASE?
A. Yes, I have developed a straightforward cost-of-service model to utilize in place of
the Company s Black Box model. I have made this study available to all parties as an electronic
file as one of my workpapers. I classified and allocated all costs in the same manner as
suggested by the Company s Jurisdictional cost-of-service study. Exhibit 302 contains a
comparison of the rate of return for each class as calculated by the Company s class cost-of-
service model (Column A) and those calculated under my "Basic" model (Column B). I would
Yankel, DI
Irrigators
not expect the results of these two models to be the same. In spite of this expectation, the results
of both models are very similar.
Q. WHAT DOES THIS COMPARISON DEMONSTRATE?
A. For the Irrigation class, the Company s model calculated a rate of return of -
while my model (using all of the Company s assumptions) placed their rate of return at -0.42%.
For the Residential class, the Company s model calculated a rate of return of5.62% while my
model (using all of the Company s assumptions) placed that rate ofretum at 5.20%. For all
customer classes, the difference between rates of return between my model and that offered by
the Company were similar.
Q. IS THE COMPANY'S CLASS COST-OF-SERVICE COMPUTER MODEL THE
ONLY THING THAT NEEDS TO BE TAKEN ON FAITH WITH RESPECT TO DEFINING
A RATE SPREAD IN THIS CASE?
A. No, unfortunately, the Company s load research data is also not subject to scrutiny.
The Company s Load Research data is contained in a proprietary computer program known as
Lodestar. It is my understanding that only one person in the Company knows how to operate this
program with an additional person acting as backup. I have asked for the raw load research data
to be provided in a usable electronic format and have been given the data as an output to
Lodestar only. This means that I would need to manually transcribe hundreds ofthousands
data points by hand into a usable Excel or Access format. It is not that the Company is refusing
Yanke1, DI
Irrigators
to supply the information, it's that the Lodestar program that the Company has chosen to process
this data cannot be operated, scrutinized or verified by Staff and intervenors, rendering it another
Black Box.
Q. WHY IS IT IMPORTANT THAT THE DATA BE AVAILABLE IN A STANDARD
PC FORMAT?
A. The Company s load research data is the foundation upon which the all important
demand allocators are based. Unlike other data used in a rate case, the load research data is a
sample that is purportedly designed to accurately reflect the population as a whole. The sample
design for the load research program should generate reliable statistical results. However, a good
sample design does not insure acceptable results. If the actual sample is not a true reflection of
the population in general, then the data is useless. It is imperative that the Staff and intervenors
have an opportunity to review the data gathered to insure that it is reflective of the general
population.
Until data can be provided in a format that is usable to Staff and intervenors, it will
simply be data that is provided in a "Black Box" that we are asked to accept on faith. Unlike the
cost-of-service study that I was able to develop, there is no such possibility to develop an
alternative to the Company s load research data. All parties in this case are forced to use the
Company s load research data without the opportunity to verify its validity. Because these "all
important" demand related factors are not verifiable, they should not be used to shift cost
responsibilities.
Yankel, DI
Irrigators
Methodology Problems
Q. WHAT ALLOCATION METHODOLOGY DO YOU RECOMMEND FOR
ALLOCATING DEMAND RELATED GENERATION AND TRANSMISSION COSTS TO
CUSTOMER CLASSES IN THIS CASE?
A. I recommend the use ofthe 12-Coincident Peak (12-CP) method. This method has
been used in Idaho for reviewing the allocation of generation and transmission costs between
jurisdictions and between customer classes for approximately 25 years. Although no specific
method has been endorsed for class cost-of-service purposes, the 12-CP appears to be the only
accepted and utilized method for allocating these costs between jurisdictions.
Q. DOES THE COMPANY USE THE 12-CP ALLOCATION METHOD IN THIS
CASE FOR DEMAND RELATED GENERATION AND TRANSMISSION COSTS?
A. Idaho Power does use a 12-CP method for allocating generation and transmission
demand related costs on a iurisdictional basis. However, the Company only puts up the pretense
of using this method for class cost-of-service purposes. Because the Company claims the need to
weight" each month's coincident peak value for class cost-of-service purposes, and because it
weights" most months by ", it effectively only uses a 5-CP method for allocating generation
costs and a 3-CP method for allocating transmission costs. In the last rate case (Case No. IPC-
94-5) the Company s development of its allocation factors stopped here. In this case, it appears
that this "weighting" approach produced an unrealistic result, so the Company is only using half
the weighted value and half the unweighted value to ultimately produce an allocator. The fact
Yankel, DI
Irrigators
that the Company needs to temper the results of its own marginal weights confirms how
inappropriate the results are.
The use of two different methods of allocating the same costs (between jurisdictions and
classes) can and does cause mismatches in who pays for what costs. For example, Idaho Power
has a significant space-heating load and the Idaho jurisdiction has allocated costs based upon that
space-heating load as well as other loads. However, according to the Company s 3-CP method
0" transmission costs are allocated to space-heating customers in the marginal weighting
process, in spite of the fact that the whole jurisdiction has allocated costs based, in part, on
space-heating usage. This is neither just nor reasonable.
Q. DOES THE WEIGHTING USED BY IDAHO POWER FIT THE LOAD PROFILE
PLACED UPON THE COMPANY'S GENERATION RESOURCES?
A. No. The graph below lists the average System peaks over the last four years.
500
000
500
~ 2 000
:E 500
000
500
Figure 7
Load plus Maintenance Requirement
Average 1999-2002
Load
Maintenance
.---.
.-A
.. ---.
Yankel, DI
Irrigators
As can be seen from Figure 7 , the monthly peak loads are all of similar magnitude . The lowest
average peaks occur in April and they are 2/3rd the level of the peak month. Maintenance is
performed throughout the year, with some valley filling in April and May, but also during the 5-
months that the Company designated as having capacity deficits. As a matter of fact, the
Company has averaged more maintenance at the time of the November and December (two of its
capacity deficit months) than it does in January, February, March, and September (non-deficit
months). There is nothing in this data that supports the Company s use of only a 5-CP method
for allocating generation costs and a 3-CP method for allocating transmission costs under its
marginal cost" weighting scheme.
Q. THE COMPANY CLAIMS6 THAT ITS MARGINAL COST WEIGHTING
ACTORS ARE BASED UPON THE CAPACITY DEFICITS FOUND IN ITS 2002
INTEGRATED RESOURCE PLAN (IRP) FOR THE SUMMER MONTHS OF JUNE, JULY
AND AUGUST, AS WELL AS THE WINTER MONTHS OF NOVEMBER AND
DECEMBER. IS THAT WHAT THE COMPANY'S IRP DEMONSTRATES?
A. No. In fact, the Company s IRP7 only addresses resource deficits during four months
and not five months.These months are June, July, November and December-August is not
included.
Exhibit 303 contains two graphs from page 29 of the Company s 2002 IRP. These
graphs represent the 50th Percentile of Water and Load. For purposes of developing marginal
5 Numerical values listed on Exhibit 305.
6 Brilz testimony at page 15.
7 See for example the 2002 IRP at pages 3 , 4, 6, and 28.
Yankel, DI
Irrigators
costs in this case, the Company only focused on the 5-year timeframe of 2003-2007 8. From
Exhibit 303 it can be seen that the only energy deficits found during this timeframe in the 2002
IRP occurred in December 2006 and November and December 2007. It also shows regular peak
capacity deficits during June and July with some deficits in November and December of this
timeframe.
The Company s 2002 IRP also addresses a 70th Percentile Water and Load condition.
Graphs associated with conditions that are not expected to be exceeded 70% of the time (a
probability of being exceeded 30% ofthe time) can be found on Exhibit 304. Under the 70th
percentile condition there are more energy deficits in both the summer as well as the winter
months. There is a slight energy deficit at the end of this period associated with August, but it is
pale in comparison to the other months and much smaller than the September energy deficit.
With respect to a capacity deficit under the 70th percentile condition, August only shows one
small deficit in the 4th year, and often has more of a surplus9 than other months.
From all of the above, it can be concluded that the Company misinterpreted its own 2002
IRP. That IRP clearly and repeatedly stated that Idaho Power had capacity deficits in only the
four months of June, July, November, and December.
Q. THE COMPANY CLAIMS 10 THAT ITS TRANSMISSION RELATED
MARGINAL COST WEIGHTING FACTORS ARE BASED UPON DEFICITS FOUND IN
ITS 2002 INTEGRATED RESOURCE PLAN (IRP) FOR THE SUMMER MONTHS OF
8 Response to Federal Executive Agencies Request I-d at page 6.
9 In 2007 under the 70th percentile condition August has more of a surplus than January, February, and
September.
10 Brilz testimony at page 16.
Yankel, DI
Irrigators
JUNE, JULY, AND AUGUST. IS THAT WHAT THE COMPANY'S IRP
DEMONSTRATES?
A. No. Exhibit 306 is a copy of page 33 from the Company s 2002 IRP that depicts the
transmission deficit under the 50th and 70th percentile conditions. As can be seen from both
graphs, the only predicted transmission deficits during the 2003-2007 timeframe occur in July. I
can only assume that the Company chose to use three months to allocate transmission costs
because it felt that a 1-CP method would look too ridiculous.
Basically, the entire notion of placing "0" weight on certain months that do not show up
as a deficit in the Company s IRP is absurd. The purpose of the IRP is to identify resource
shortfalls and to plan to correct for them. There will always be resource shortfalls over the
horizon that need to be addressed. That does not mean that times with no shortfalls
demonstrated should be given free electric service.
Q. IS THE COMPANY'S TREATMENT OF ITS ENERGY COSTS APPROPRIATE?
A. The Company s proposed treatment/weighting of energy costs is far more appropriate
than its treatment of demand related generation and transmission charges, even though the
Company used only the weighted "marginal cost" values and did not split the allocators 50:50
with the "unweighted" usage. In spite of the winter energy deficits shown under the 50th
percentile condition and the summer/winter deficits shown under the 70th percentile condition
the Company proposed to use the energy in each month to allocate costs. This is consistent with
what the Company should have done with demand related generation and transmission costs.
Yankel, DI
Irrigators
However, the use of marginal costs as weighting factors is not appropriate. The energy
costs contained in any cost-of-service study are the average costs and not the marginal costs. It
is improper to spread the entire energy usage of a class over the cost of the last or marginal unit
produced. The vast majority ofthe energy consumed by all ofIdaho Power s customers is either
hydro (at zero fuel cost) or Company-owned thermal (with minimal fuel cost per kWh). It is not
proper to allocate these average fuel costs based upon the price of the last unit of generation.
Additionally, this provides a mismatch with the PCA rates that are based on an average cost per
kWh.
Q. WHAT ARE YOUR OVERALL RECOMMENDATIONS WITH RESPECT
ALLOCATING GENERATION AND TRANMISSION RELATED COSTS FOR CLASS
COST OF SERVICE PURPOSES?
A. I recommend that the 12-CP method be used for all generation and transmission costs.
I recommend that no "marginal" weighting factors be applied to either the demand or energy
component of these costs. In this manner the class allocation process would be consistent with
the jurisdictional allocation procedures that assigned these overall costs to Idaho Jurisdiction in
the first place. Additionally, the use of un weighted costs is far more reflective of cost causation
than the weighted allocation factors proposed by the Company.
Q. USING YOUR COST-OF-SERVICE STUDY, WHAT IS THE RESULT OF USING
UNWEIGHTED ALLOCATION FACTORS FOR SPREADING GENERATION AND
TRANSMISSION COSTS?
Yankel, DI
Irrigators
A. Using my cost-of-service study and simply removing the weightings that the
Company has applied has a major impact upon the rate of return calculated for the Irrigation
customers. As can be seen from Exhibit 302 column ", the rate of return for the Irrigators
goes from -0.42% in my "Base" cost-of-service study with no changes to any ofthe Company
methods or data to 2.11 % or halfway from where it was to the overall system average rate of
return. This is an extremely significant change and demonstrates the impact that the Company
weighting factors have on the Irrigation class. No other rate schedule experienced anywhere
near the magnitude of this impact by changing the weighting factors.
Data Problems
Q. BEYOND THE COMPUTER MODELING PROBLEMS AND THE
METHODOLOGICAL PROBLEMS, DOES THE COMPANY'S COST ASSIGNMENT TO
CUSTOMER CLASSES SUFFER FROM ANY ADDITIONAL PROBLEMS?
A. Yes. The Company s cost assignment procedures suffer additionally from critical
input data problems. If the input data is wrong, it does not matter how accurate the computer
model or how appropriate the methodology. It would be unjust and unreasonable to make
determinations of relative cost responsibility based upon inaccurate information.
Q. PLEASE GIVE AN EXAMPLE OF THESE DATA PROBLEMS THAT CAUSE
DISTORTIONS IN THE OVERALL COST RESPONSIBILITY THAT IS ASSIGNED TO
VARIOUS CUSTOMER CLASSES.
Yankel, DI
Irrigators
A. One example can be found in the raw data that is used to define the monthly
coincident peak usage for Irrigators as well as other customer classes. Basically, the load
research data that the Company asks us to accept "on faith" is a reflection of actual 2002 usage
and not normalized usage. However, the energy and associated revenue that is used for rate
making purposes in this case is based upon normalized usage. Everyone knows that the 2002
summer was very hot and dry compared to normal. Thus, the Irrigators are faced with being
credited normalized revenue, but charged with demand levels that are well above normal.
Q. IS THERE A CORRELATION BETWEEN OVERALL CONSUMPTION AND
DEMAND FOR THE IRRIGATION CLASS AS A WHOLE?
A. Yes, there is a strong correlation for the Irrigation class as a whole. Admittedly, on a
single month basis there may be very little correlation for an individual Irrigator that has one
piece of equipment and a fixed demand that does not vary ifhe operates 1 hour or 744 hours in
the month. However, load research and cost allocation looks at the class as a whole. Irrigators
will not turn-on their pumps under given conditions or will postpone operation until the
beginning of another billing cycle if conditions are right. On an aggregated basis this results in a
strong correlation between kWh usage and kW demand levels as can be seen from the following
graph:
Yankel, DI
Irrigators
000 000
800,000
600 000
400,000
200,000
Figure 8
Irrigation In-Season
Billing and Average Demand
Billing Demand
Year
As can be seen from Figure 8 , the overall billing demand for the Irrigators is highly correlated to
the overall energy usage. If the Company is going to normalize energy and revenue for rate
making purposes, then it must also normalize the demands that are assigned to the various
customer classes.
Exhibit 307 details the results ofthis normalization ofIrrigation demand based upon the
exact same method the Company used for establishing coincident demand, except starting with
normalized energies as opposed to 2002 energy levels. As can be seen from Exhibit 307, there is
a difference of 9.1 % between the sum of the 12-coincident peaks that are normalized and those
that are not normalized. If one were to compare the difference between normalized values and
actual 2002 coincident demand values for Irrigation customers for only the 5-coincident peaks
that the Company used in its filing, then that difference grows to 16.2%.
Yankel, DI
Irrigators
Q. ARE THERE OTHER NORMALIZING ADJUSTMENTS THAT SHOULD BE
MADE TO THE COINCIDENT PEAK DEMAND FOR OTHER CLASSES?
A. Yes, there are two other types of normalizing adjustments that need to be made. First
there were several other rate groups (Residential, Schedule 7, Schedule 9-Primary and Schedule
Secondary) where energy levels and revenues were normalized, but 2002 actual coincident
demands were used to establish class coincident peaks in the Company s cost-of-service study.
These schedules also need to have their demands normalized. The calculation 1 1 for normalizing
each of these schedules is presented in Exhibit 308 through Exhibit 311.
The second normalization adjustment that needs to be made is to normalize the non-
coincident peak data that is developed in the same manner by the Company for these same
schedules. The maximum non-coincident peak data is used in the Company s cost-of-service
study for purposes of allocating demand related distribution plant. The maximum non-
coincident peak (based upon normalized usage) for Irrigation, Residential, Schedule 7, Schedule
Primary, and Schedule 9-Secondary are highlighted in Column M of Exhibit 307 through
Exhibit 311.
Q. WHAT IMP ACT DOES USING NORMALIZED DEMAND DATA (TO BE
CONSISTENT WITH USING NORMALIZED ENERGY AND REVENUE DATA) HAVE ON
THE RATE OF RETURN FOR THE IRRIGATORS AND OTHER CLASSES?
11 Based upon the same methodology as used by Idaho Power except starting with nonnalized demands.
Yankel, DI
Irrigators
A. As can be seen from Exhibit 302 Column ", by using normalized demands 12 the
rate of return for the Irrigation customers is increased to 3.50% with the jurisdictional average at
97%. This adjustment brings the Irrigation rate of return up another 25% from where the
Company s cost-of-service study placed it compared to the jurisdictional average. Again, the
impact of using normalized demand values upon the Irrigation customers is far greater than on
the other classes.
Q. ARE THERE OTHER DATA PROBLEMS THAT YOU HAVE IDENTIFIED?
A. Yes. There are a number of problems with the classification of some distribution
plant as proposed by the Company. Some ofthese classification problems are found on page 12
of the workpapers of Company witness Brilz that I have included in my testimony as Exhibit
312. Near the bottom of that exhibit is a listing of the amount of plant that is considered to be
associated with Primary Distribution and the amount of plant that is considered related to
Secondary Distribution. As can be seen from Exhibit 312, there is $135.8 million classified as
Primary related Account 364 (Poles, Towers, and Fixtures) with $14.0 million (8.4%) classified
as Secondary related. Exhibit 312 also shows $72.7 million classified as Primary related
Account 365 (Overhead Conductors & Devices) with $14.7 million (16.9%) classified as
Secondary related.
The first thing that seems odd about these values is that the relative amount of Secondary
Conductors and Devices is twice that ofthe Poles and Towers. In response to Irrigator Request
41-c the Company stated:
12 These demand and energy values are unweighted.
Yankel, DI
Irrigators
If a pole or tower carried both primary and secondary lines, it would be
assigned to the primary component. In this situation, the pole or tower would
need to be constructed to the standards necessary to support primary lines.
such, it would be classified as primary.
Thus, the Company s classification scheme for Account 364 completely ignores any
Secondary responsibility for dual-purpose poles. Additionally, all Poles and Towers are booked
to Account 364, even if they are part of a Service Drop to a specific customer
Q. IS THIS TREATMENT OF ACCOUNT 364 COSTS APPROPRIATE?
A. No. Obviously, this procedure is bias toward placing more plant costs into the
Primary component. Unfortunately, there is insufficient information available to correct this
bias.
Q. ARE THERE OTHER APPARENT PROBLEMS WITH THE CLASSIFICATION
OF COSTS FOR ACCOUNTS 364 AND 365?
A. Yes. The Response to Irrigator Request 42 provides a breakdown of "Pole Miles
and "Wire Miles" by Primary and Secondary components. The breakdown was as follows:
Secondary
359
659
Primary
075
596
Pole Miles
Wire Miles
In and of itself, these values would be accepted at face value. However, the rest of the
Response to Irrigator Request 42 places these values into question. It states that "Pole Miles
Yankel, DI
Irrigators
refers to the distances between poles, which is understandable. It further states that Wire Miles
is the distance between poles times the number of wire phases, which also makes sense. The
response goes on to give an example:
For example, a 100 foot span of three-phase conductor would equal 100 feet
of pole miles and 400 feet of wire miles (i., three wires plus the neutral or
four times the pole distance).
When one applies this definition and example to the "Pole Miles" listed in the response
one would expect the "Wire Miles" to be significantly larger than those provided. Even if the
entire distribution system was 2-phase, this would mean 3 "Wire Miles" for each "Pole Mile
yet the total "Wire Miles" provided are closer to 2 times the "Pole Miles" than they are to "
times. Within the limits of this case, it has not been possible to delineate the source of this
discrepancy.
Q. GIVEN THE DISCREPANCIES YOU HAVE POINTED OUT WITH RESPECT TO
THE BASE DATA ASSOCIATED WITH ACCOUNTS 364 AND 365, WHAT ARE YOUR
RECOMMENDATIONS?
A. Significant shortcomings exist and the data and the development of the numbers that
go into the Company s cost-of-service case require much greater review and analysis. Account
364 and Account 365 costs need to be further investigated and a reliable set of numbers needs to
be produced.
For purposes of this case, I recommend that 16.86% of Account 364 costs be classified as
Secondary related. This is the same percentage as is classified as Secondary related for Account
13 See Response to Irrigator Request 41-e.
Yankel, DI
Irrigators
365. My general feeling is that this percentage is low, but some number has to be picked and
16.86% at least moves Account 364 costs in the right direction.
Q. ARE THERE ANY OTHER PROBLEMS WITH DISTRIBUTION RELATED
COSTS?
A. Yes. Account 366 deals with Underground Conduit and Account 367 deals with
Underground Conductors. As can be seen from Exhibit 312 63.1% of Underground Conduit
(Account 366) is classified as Secondary, while only 19.6% of Underground Conductor (Account
367) is classified as Secondary. Because one would expect Secondary Conduit to carry
Secondary Conductors, one would expect that the ratio of Secondary Conduit to be similar to that
of Secondary Conductor. As seen from Exhibit 312, the ratio of Secondary Conduit is over 3-
times that of Secondary Conductor. No explanation is provided for this wide discrepancy, which
needs to be investigated. Because of the lack of additional information, I have made no
adjustment to the classification that the Company has proposed for these accounts.
Q. SHOULD IRRIGATORS BE ALLOCATED ANY ACCOUNT 366 OR ACCOUNT
367 COSTS?
A. No. Irrigators do not use underground conduit and conductors so they should not be
allocated any of these costs. In addressing this issue the Company stated in Response to Irrigator
Request 58-a:
Idaho Power s current property accounting records do not identify
investments in poles, overhead conductors, underground conduit, or
Yankel, DI
Irrigators
underground conductors by customer class nor by customer characteristics
such as rural versus urban location, number of customers per feeder, average
number ofline miles per customer in each class, etc. Because this specific
data is not available, Idaho Power has historically allocated all of the
investments in FERC Accounts 364, 365, 366, and 367 using class non-
coincident peak demand (or coincident group peak demand) methodology
described in the National Association of Regulatory Utility Commissioner
Electric Utility Cost Allocation Manual. ... The methodology does not
recognize class-specific characteristics or demographics. Should it be
desirable to recognize class-specific characteristics such as rural versus urban
location or average number of line miles per customer in each class, a
methodology other than the non-coincident peak demand methodology would
need to be developed to recognize that certain classes utilize, or do not utilize
the components of the distribution system is used by the irrigation class while
at the same time recognizing that a larger portion 0 the overhead system is
used by the irrigation class on a per-customer basis than by the other classes
of customers. (Emphasis added)
It is not appropriate to assign costs to Irrigators for plant that they do not use because
this specific data is not available" and the Company believes that it is under-allocating costs in
another area. I agree that non-coincident peak demand is widely used for distribution plant
allocation and I am not taking exception to the overall allocation method. However, just like
Schedule 9-Primary customers are not allocated costs with respect to secondary lines (because
they do not use them), Irrigators should not be allocated costs associated with underground
distribution equipment. This does not destroy the integrity of the non-coincident peak method; it
just recognizes that some customer groups do not use certain equipment.
I recommend "zero" underground distribution plant being allocated to the Irrigation class.
If the Company feels that "specific data is not available" to properly allocate distribution costs
then this specific data should be gathered in the future.
Q. ARE THERE ANY OTHER PROBLEMS WITH THE COMPANY'S DATA THAT
WOULD TEND TO OVER-ALLOCATE COSTS TO THE IRRIGATION CLASS?
Yankel, DI
Irrigators
A. Yes. Irrigators typically have a Service Drop that goes from the Line Transformer on
a Pole, down the Pole, and ending at a Meter on the same Pole. I do not have an exact number
for this distance, but I would estimate it to be a maximum of 30 feet. When Idaho Power
developed its allocation factor for Services, it computed the cost based upon an assumed Service
Drop length of70 feet . Response to Irrigator Request 44-c stated:
The Company does not maintain the detailed data on the length of the service
drop for each customer and therefore cannot calculate the actual average
service drop length by customer class. However, based on the expertise ofthe
Company s agricultural representatives and their knowledge of the
Company s customers, it is believed that no service drop for irrigation
customers exceeds 120 feet or would be less than 10 feet and on average
would be approximately 60 feet in length. (Emphasis added)
Although I disagree with the Company s ultimate estimate that the Service Drop for
Irrigators would average 60 feet, there are two things upon which we agree. First, the Company
does not have good data upon which to allocate Services. Second, the average length of a
Service for an Irrigation customer is less than what was used to allocate costs to Irrigators.
I am not making a specific recommendation to change this allocation factor in this case
but I recommend that this be another one of those areas that gets serious attention in the future.
Q. WHAT IS THE OVERALL RATE OF RETURN FOR IRRIGATORS AND OTHER
CUSTOMER CLASSES, BASED UPON THE ADDITIONAL CHANGES YOU HAVE
RECOMMENDED TO THE DISTRIBUTION DATA?
14 See Response to Irrigator Request 44-a.
Yankel, DI
Irrigators
A. Although I have pointed out a number of problem areas in the allocation of
distribution costs, I have only recommended two changes:
1. Increase the classification of Secondary related Poles and Towers (Account 364)
up to the same level as used for Secondary related Overhead Conductors (Account
365); and
2. Assign "zero" cost responsibility to Irrigation customers with respect to
Underground equipment (Accounts 366 and 367).
Based upon making only these two additional adjustments to the distribution data, the
rate of return for each customer class is demonstrated on Exhibit 302 Column ". The rate of
return for the Irrigation class becomes 5.11 % compared to a jurisdictional average of 4.97%.
Once again, these two adjustments had a significant impact upon the Irrigation customers while
having a much smaller impact upon other customer classes.
Q. WHAT ARE YOUR OVERALL RECOMMENDATIONS WITH RESPECT TO
COST ALLOCATION IN THIS CASE?
A. I recommend an across the board increase to all customer classes. This is based upon
two factors. First in spite of all of the data problems, all the major customer groupings have a
rate of return that is relatively near the system average. Second because of all of the data
problems, no customer grouping should be given a disproportionate increase or decrease.
Idaho Power has not had its rates and supporting cost-of-service data reviewed for 10
years. All customer groups deserve to have their rates based upon true cost-of-service. The
Yankel, DI
Irrigators
Company s cost-of-service study does not rise to this level. To insure that that happens in the
future the necessary effort should be required.
Rate Design
Q. IS THE RATE DESIGN PROPOSED BY THE COMPANY APPROPRIATE FOR
THE IRRIGATION CUSTOMERS?
A. No. The rate design proposed by the Company places far too much emphasis upon
fixed charges and not enough on energy charges. Demand charges are designed to recover
demand costs as well as keep peak demand in check by giving a customer a chance to save
money by keeping his demand as low as practical. It is well known that individual Irrigators
have only one appliance (a pump) to turn either "off' or ". In this context, pricing demand
to keep peak load in check is somewhat meaningless. Once an Irrigator turns on his pump, his
demand charges will be the same if he operates 1 hour or 744 hours in the month.
I recommend that the demand and customer component of Irrigator s rates be increased
by no more than the average rate increase. In this manner, more emphasis (charges) will be
placed on energy consumption, the only commodity that Irrigation customers can control.
Yanke1, DI
Irrigators
SUMMARY OF RECOMMENDATIONS
Q. PLEASE SUMMARIZE THE IRRIGATORS RECOMMENDATIONS TO THE
COMMISSION IN THIS CASE.
A. 1. The Company s overall rate request should be reduced by at least $5 794 724 in
order to reflect more realistic O&M expense figures on a going forward basis.
2. The Company s cost-of-service study and its load research data must be put in an
electronic format that is understandable and transparent to all parties so that they
can be fully scrutinized.
3. A comprehensive study needs to be made of the data that is used as input to the
cost-of-service study.
4. Any rate increase should be spread evenly between the customer classes until
such time as better information is available. If a disproportionate rate increase
were to occur in this case, the corrections that I have proposed should be
incorporated into any such analysis.
5. The customer charge or demand charge for Irrigators should not exceed the
system average rate increase. This properly places monetary impact upon the one
cost component over which Irrigators have control.
Q. DOES THIS CONCLUDE YOU DIRECT TESTIMONY?
A. Yes.
Yankel, DI
Irrigators