20040628Decoupling Report.pdf

IDAHO POWER () C' .--'" F \lr-t '. l.... t.; :,,-,-," I' ~;- :ojJ',~ . '1 "-~'- ii-.-, L LJ .,...........,~ l~-_. -~~.. ~'~ ,.2n ilqJUr~ 2-5 'lrl :9-An IDACORP CompanyMichael J. YoungbloodRegulatory Affairs RepresentativePricing and Regulatory Services i Lr -" ~ "1 '.j C': U d L J C-' r-, r T J~ c: . :",~f,~~,J :f').;~l. , I t i,,- -...J . .. I . 'vv:JUlft [phone]FAX [phone]myou n9 b ood~i d ahopower. comJune 22 2004Ms. Lisa D. NordstromDeputy Attorney GeneralIdaho Public Utilities Commission472 W. Washington StreetO. Box 83720Boise, 10 83720-0074Dear Ms. Nordstrom:As stated in IPC-O4-, in response to the NW Energy Coalition s petition to the Idaho PublicUtilities Commission to investigate financial disincentives to investment in energy efficiency!Idaho Power Company agreed to circulate a "white paper" analysis on the subject. Enclosed isthe white paper Decouplingfor Idaho Power Conlpany, which was developed for the Companyby Eric Hirst. Mr. Hirst is an independent consultant who has written much on decoupling previous years , and is very well respected within the electric industry. Mr. Hirst has workedwith the PUC in the past and most recently worked with Idaho Power facilitating interaction
with the public advisory committee during the development of the Company s 2004 Integrated
Resource Plan.
Decoupling is the process of severing the link between a utility s kWh sales and its recovery
revenues to cover fixed costs. Often this link is cited as the reason utilities are not more
proactive in DSM or conservation measures. Idaho Power undertook development of this
white paper to see if and how decoupling might work here in Idaho and in conjunction with the
Company s Power Cost Adjustment (PC A).
The report is based on 2003 data used in the Company s most recent general rate case. Mr.
Hirst reviewed previous decoupling experiences in California ! Maine , New York , Oregon , and
Washington. He also analyzed the effects various forms of recoupling might have on Idaho
Power customers and shareholders over the next three years.
If you have any questions regarding the report , please contact me at 2882.
MJY
Enclosures
DECOUPLING FOR IDAHO POWER COMPANY
Eri c Hirs t
Consulting in Electric-Industry Restructuring
Bellingham, Washington
March 30, 2004
Prepared for
Idaho Power Company
Boise, Idaho
Mike Youngblood, Project Manager
CONTENTS
1. INTR 0 D U CTI ON
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. 3
2. CURRENT SITUATION
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. 4
3. COLLECTION OF FIXED COSTS THROUGH VARIABLE RATES
. . . . . . . . . . . .
. 5
4. POSSIBLE RECOUPLING MECHANISMS
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. 8
5. ANALYSIS OF RECOUPLING MECHANISMS FOR IPC
. . . . . . . . . . . . . . . . . . . .
. 10
6. IPC D EC 0 UPLIN G MOD EL RES UL TS .. 12
BASE CASE
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. 12
REVENUE PER CUSTOMER RECOUPLING
. . . . . . . . . . . . . . . . . . . . . . . . . .
INFL A TI ON RE CO U P L IN G .. 0 .. 16
FORECAST-LOAD-GROWTHRECOUPLING .. 17
EFFECTS OF DSM PROGRAMS
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
7. CONCLUSIONS
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. 18
APPENDIX A: PAST EXPERIENCE WITH DECOUPLING .. 21
APPENDIX B. DETAILS ON RECOUPLING WORKBOOK
. . . . . . . . . . . . . . . . . . . .
. 24
- ,
1. INTRODUCTION
Decoupling severs the link between a utility s kWh sales and its recovery of revenues
to cover fixed costs. Advocates of energy-efficiency programs favor decoupling because
cutTent ratemaking practices collect substantial revenues for fixed costs through a utility
energy charge ($/MWh). As a consequence, utility programs that improve customer energy
efficiency create tension between the interests of customers (whose bills go down) and
shareholders (whose earnings decline).
Although decoupling may be motivated by the desire to expand electric-utility energy-
efficiency programs, its effects are broader. That is , decoupling will affect customer bills and
rates, as well as utility revenues, even if no utility DSM programs are implemented.
During the early 19908, various forms of decoupling were deployed in Maine, New
York, California, and Washington. During the mid-1990s , these efforts were largely abandoned
as utilities and state regulators anticipated a restructured, competitive electricity industry,
although Oregon began decoupling in the late 1990s. Recently, California reinstituted
decoupling. Appendix A provides details on the states ' experiences with decoupling. Readers
interested in additional background on decoupling should see the references by Carter; * Eta
Stoft and Belden;# Hirst;~ Moskovitz, Hanington and Austin;t and Nadel , Reid and Wolcotc
Decoupling involves two major steps. The first is the policy decision to break the link
between sales and revenues. The second, analytically more difficult , step is to recouple utility
revenues (more precisely, revenues to cover fixed costs) to something other than actual kWh
sales. Decoupling also involves other issues, such as:
whether to decouple for all or only some rate classes
whether to recouple on a class-specific or system-wide basis,
whether to apply the decoupling-induced rate adjustments to energy charges only or to
both energy and demand charges, and
S. Carter, "Breaking the Consulnption Habit: Ratemaking for Efficient Resource Decisions/' The Electricity
Journal 14(10), 66- 74~ December 2001.
J. Eto, S. Start and T. BeldenThe Theory and Practice of Decoupling, LBL-34555, Lawrence Berkeley
Laboratory, Berkeley, CA, January 1994.
E. Hirst, Statistical Recoupling: A Ne\v Way to Break the Link Between Electric-Utility Sales and Revenues
ORNL/CON-372, Oak Ridge National Laboratory, Oak Ridge, TN , September 1993.
D. Moskovitz, C. Harrington and T. Austin
, "
Weighing Decoupling vs Lost Revenues: Regulatory
ConsiderationsThe Electricity Journal 5(9), 58- 63, Novelnber 1992.
S. N, Nadel, M. W. Reid and D, R. Wolcott (editors), Regulatory Incentivesfor Dem-and-Side Management
American Council for an Energy-Efficient Economy, Washington , DC , 1992.
the frequency with which rates are adjusted for decoupling.
The next section describes the cutTent (2003) situation that Idaho Power Company (IPC)
faces with respect to recovery of its fixed costs. Section 3 focuses on class-specific rate
structures and how they affect recovery of fixed costs. Section 4 briefly reviews alternative
ways to recouple utility revenues to something other than energy sales. Section 5 explains the
analytical method developed to examine alternative recoupling mechanisms for IPC , with
additional details in Appendix B. Section 6 presents model results. And the final section
summarizes the results, findings, conclusions, and recommendations from this study.
2. CURRENT SITU TI 0 N
This paper focuses on (and deals only with) the following rate classes: Residential
(Schedule 1), Small General (7), Large General (9), Large Power (19), and Irrigation (24).
Together, these five classes account for 99% of IPCt s 2003 proposed revenue requirement.
Based on information from the CUITent IPC rate case, 56% of the 2003 cost-of-service
revenue requirement covers fixed costs ($303 million of the $541 million total), with the
remaining 44% for variable energy costs ($237 million for fuel purchased power , and variable
operations and maintenance at generating stations). * As shown in Fig. 1 , the fixed-cost (FC)
component is greatest for Schedule 7 (70%) and smallest for Schedule 19 (36%); this difference
is probably a consequence primarily of differences in load factors among classes. This suggests
that the net -revenue-loss problem associated with utility energy-efficiency programs might be
greatest for the Small
General class of customers.
Figure 1 also shows
fixed costs as a share ofproposed revenue
requirements. Because of
the large proposed cost
shift from the inigation
class to the other classes
(25% of the inigation cost
of service), the share of
revenue requirement from
fixed costs is much greater IPCFixad Co,ts
for this class than the share Fig.
800/0
ca: 700/0
LL 600
LLIa:: 50%
c:c
:I:
en 400/0
t; 300
200/0
LLI
u:: 1 DO
EB % of Costs
0% of Revenue Requirement
Residential Small General large General Large Power Irrigation
Percentage of 2003 costs and proposed revenue
requirement from fixed costs , by rate class.
1 assume that the only variable costs IPC experiences are for energy production.
of total costs. * The effects of the shift from cost of service to revenue requirements is much
smaller (about 5%) for the other four classes. The remainder of this paper uses proposed
revenue requirements as the basis for calculating and adjusting fixed costs.
Table provides key statistics, based on the 2003 rate case , for each customer class. The
Residential class accounts for just over half of the company s total fixed costs. Normalizing the
fixed costs for each class by the number of customers in each class shows substantial
differences, ranging from $420/customer for Small General to $206 000 for Large Power. The
difference between the proposed energy charge and variable energy cost is greatest for Small
General ($40/MWh) and smallest for Large Power ($3/MWh), with an average of $ 16/MWh.
Table Fixed- and variable-cost characteristics of IPC rate classes
Rate Class
Total
Fixed costs, million $153.13.54.21.60.303.
Fixed costs as percentage63.69.46.36.60.56.of total cost
Fixed costs as percentage60.66.43.34.80.56.of revenue requirement
Fixed costs/customer
, $
457420 186 206 278 253 756
Variable cost, $/MWh21.22.21.19.24.21.
Energy charge, $/MWh51.62.26.22.35.37.
The 2003 cost of service for class 24 is $100.million but the proposed revenue
requirement is only $75.4 million, a 25% reduction.
3. COLLECTION OF FIXED COSTS THROUGH VARIABLE RATES
The relative importance of decoupling for different rate classes depends on the
relationship between fixed and variable costs (Fig. 1) and the rate design for that class
(discussed here). Rates for classes 1 and 7 include per-customer and energy charges , while
those for the other classes also include several demand charges.
The assumption that all of the class 24 fixed costs are to be recovered from the proposed rates implies that
the energy charge for this class is much too low. Thus, the substantial subsidy of class 24 costs make the results
presented here suspect for that class.
- #
To keep this discussion from becoming too complicated and to focus on the issues rather than the details the
Schedule 9 and 19 subclasses (Secondary, Primary, and Transmission) are combined into one average class. Similarly,
the demand charges are aggregated for each class into one average charge.
For schedule 1, 74% of the class-specific fixed costs are collected through the energy
charge (top of Fig. 2), amounting to $113 million for 2003 (bottom of Fig. 2). For Schedule 7,
the percentage affixed costs collected by the energy charge is almost as high (71 %), but IPC'
exposure is much lower ($10 vs $113 million) because Schedule 7 accounts for less than 10%
of the revenues of Schedule
Interpreting the rate schedules for the other three classes is more complicated because
of their demand charges. Should these demand charges be considered variable or fixed? That
, do they vary with energy (volumetrically) or are they fixed? The answer is probably class
and charge specific and likely falls part way between 100% variable and 100% fixed. * For
example, the peak demand for Schedule 9 customers may have a large weather-sensitive
component, in which case summer demand (MW) and summer energy consumption (MWh) are
likely to be highly correlated. On the other hand, demand for Schedule 19 customers might be
dominated by industrial processes, which are independent of weather. If these processes are
either on or off, demand will be largely independent of energy sales. This issue is complicated
by the fact that the proposed rate schedules include on- and off-peak demand charges as well
as basic (12-month average) demand charges.
To some extent, the treatment of demand charges is an empirical issue. We could
analyze historical data by rate class to detennine how tightly coupled (Le., colTelated) energy
sales and demand are. To some extent, this is a policy issue: deciding whether to adjust rates
for decoupling through energy charges only or through energy and delnand charges.
If the revenues collected through demand charges are largely independent of energy
sales, then energy-efficiency programs aimed at Schedules 1 , 7 , and 24 have much greater
effects on FC recovery per kWh of energy saved than do such programs aimed at Schedules 9
and 19 (top of Fig. 2). Weighting each class by its contribution to total revenue shows the
importance of IPC' exposure to FC losses from each class. Clearly, the Residential class ($113
million, bottom of Fig. 2) is the most important, and Large Power ($3 million) is the least
important. Overall, 58% ($177 million) of IPC' s FC revenues are collected through energy
charges, and an additional 25% ($76 million) is collected through demand charges.
On the other hand, if the revenues from demand charges are proportional to those from
energy charges, all five customer classes create exposures of70% or more. Indeed , in this case
more than 90% of fixed costs are collected through variable charges for Schedules 9, 19, and
24. Overall, $252 million of fixed costs are collected through energy and demand charges
accounting for 47% of IPC revenues.
In the long run (say, 10 to 20 years), all costs are variable.
100
:is
(/)
a: w
LL. e,:,
UJ a:
I- c(
(/) J:0 UU w
...JW m
LL II:LL c:c
;:::..
0 eco upl i ng 0 ala
120
100II:LL en
..........
-En-
s:: a:
CI:
:I:
.- (.)
en ..JI-
en 0 -0 ~
:::-
DecouplingData
Fig. 2.
ResidentialSmall GeneralLarge General
JC;~
""~';."~~~~.....-~,,,~
fh~"~f~:::
~~f~g9\
" .':.;.. . ",;'.; ~\;'-~.., ";):
id.l!tVfl"'I.t-.
. ,;
i~~!,
)t~*-~
' ~:
t~"
,~ ." ,;:"~ ,
h 'fi- . "
,~.
~i~~ t;;:n~~l*~~.
. ~t~
" "
Large Power Irrigation
IbJ Energy and Demand Charges Are Variable
D Energy Charge Are Variable
?i~";';;"'t':
1 '
~;~~,~\~~;;
:::~-OO~'
~$l~'!bf~:if~;$;~~~-r""i;
~~;&~~t,~;
fit
- :~:~
:;\jr t1.~,
~..",.
. l' ,
~~.,'"
t;;,))
'\~:~'\~:~&"
Large Power IrrigationResidentialSmall GeneralLarge General
Collection of fixed costs through variable charges (energy plus demand or
energy only) by rate class. The top chart shows the percentage of fixed costs
collected through variable charges, and the bottom chart shows the year
2003 dollar amounts collected through variable charges.
Figure 3 presents
this information in yet '""' 35
another way. This figure ~
shows the net revenue loss
(the loss in FC recovery)
to IPC per MWh of energy
reduction. * Again, results UJ are shown for two cases:
demand changes are I-
proportional to energy Z
changesand demand
changes are independent
of energy changes. On a De~DuplingDataper MWh basisthe Fig. 3.
company is most exposedto energy-efficiency
programs aimed at the Residential and Small General classes, with losses of $27 and $36 per
MWh. At the other end of the spectrum, if demand-related revenues are independent of energy
sales, the losses for the Large General and Large Power classes are only $3 and $1 per MWh.
Averaged over all five classes, the company would lose $16 for every MWh reduction in sales.
CI Energy and Demand Charges
Are Variable
0 Energy Charges Are Variable
"f;\"i\Ji~
~~0it!
\jJj
'f,~,g~
~. ~~~ . )~,~ ~ )- :
~~t
,.., ~'. "
!i'I 'l!?i'c E
~,~:;.,.~;
i~~;
~tIff~FJ
\\'.
i!i."'~"".
, "
~H!fli;~
7'o:g;kt~,.
. "
1ffi'b;'N?;b~I/'
tb~~~'\i1;~;;~
~..
/?fi""",
Residential Small General Large General Large Power Irrigation
Loss of fixed -cost revenues per MWh of sales
reduction by rate class.
These results suggest that, if IPC decides not to implement decoupling for all rate
classes, it might focus initially on schedules 1 and 7. Because the residential class accounts for
more than half of IPC' s fixed costs and residential customers pay for much of their fixed costs
through the energy charge, IPC' s earnings losses are quite high , both in absolute terms and on
a per MWh basis. Although Schedule 7 accounts for only 4% of IPC's fixed costs , its energy
charge of $62/MWh is the highest of all rate schedules.
4. POSSIBLE RECOUPLING MECHANISMS
Decoupling mechanisms, of necessity, recouple utility revenues to something other than
sales. Possible recoupling mechanisllls include explicit attrition adjustments intended to track
the detenninants affixed costs (e., the cost of capital), the number of utility customers (which
seems most applicable to distribution costs), inflation (perhaps with a productivity offset), the
determinants of electricity sales, or some other Inecbanism. A key policy issue here is whether
recoupling should focus on tracking fixed costs (which seems the most reasonable but could
The numbers shown in Fig. 3 are based on the proposed rate structures, while those in Table 1 are based on
actual costs. The only substantial discrepancy occurs for lITigation customers; Figure 3 shows a net revenue loss of
$26.3/MWh while Table 1 shows only $lO.7/MWh.
be quite complicated) or on some proxy for sales (consistent with the traditional treatment of
fixed costs). A third option is to agree upfront on the level of allowed fixed costs for a few
years and to then have frequent rate cases. The Oregon PUC chose this approach in the mid-
1990s for decoupling mechanisms implemented by PacifiCorp and FOE , with rate cases to be
held every two years.
Two statistical analyses of data from several utilities showed little connection between
changes in a utilitys fixed costs and its electricity sales:
In the long-run the relationship between (fixedJ cost and customer growth
is stronger or no worse than the corresponding relationship between costs
and sales.
The short-term analysis of year-la-year changes in sales vs. base costs
shows no statistically significant relationship. Yet ... the assumed
existence of a strong correlation between these two factors is the
foundation of traditional sales-based regulation.
Similarly, Eta, Stoft, and Belden wrote
, "
Relying on 25 years of aggregate financial
statistics from 160 investor-owned utilities, we find that one-year changes in load or numbers
of customers are both poorly-coITelated with changes in nonfuel costs. Hence , the proponents
of RPC (revenue per customer decouplingJ are COITect in arguing that RPC does no worse than
traditional ratemaking in tracking nonfuel costs (indeed, we find it does slightly better).
These analyses show that decoupling replaces one set of factors unrelated to the
determinants of fixed costs with another set of factors unrelated to those costs. Decoupling, on
average, should have no positive or adverse effect on a utility s opportunity to recover its fixed
costs. On a year to year basis, decoupling might (or might not) stabilize FC recovery.
- ,
c. Mamay and G. A. Comnes
, "
Californias ERAM Experience " Chapter 3 in Regulatory Incentives for
De/nand-Side Managelnentedited by S. M. Nadel, M. W. Reid, and D. R. Wolcott, 39-, American Council for an
Energy-Efficient Economy, Washington, DC, 1992.
D. Moskovitz and G. B. Swofford
, "
Revenue-per -Customer DecoupJing," Chapter 4 in Regulatory Incentives
for Den1.and-Side Managementedited by S. M. Nadel, M. W. Reid , and D. R. Wolcott, 63- 77 , American Council for
an Energy-Efficient Economy, Washington, DC, 1992.
J. Eto, S. Stoft, and T. BeldenThe Theory and Practice of Decoupling, LBL-34555 , Lawrence Berkeley
Laboratory, Berkeley, CA, January 1994.
5. ANALYSIS OF RECOUPLING MECHANISMS FOR IPC
developed an Excel
workbook to quantify the effectsof different recoupling
mechanisms on customer
electricity bills and rates and on
IPC revenues. The workbook
calculates the interactions
between a particular recoupling
mechanism and alternative
forecasts of the number
customers, peak demand, and
energy sales. These analyses use
data for 2003 from the IPC rate
case to simulate results for 2004, Fig. 4.
2005, and 2006 (Fig. 4).
INPUTS
2003 Rate Case
2004 IRP Forecasts
2004 - 2006
Base Case
PARAMETERS
Recoupling Mechanism
Alternative Forecasts
~ Recoupling
""
AnalysIs
Resu Its
Diagram of recoupling model.
The workbook is set up to test three forms of recoupling:
Revenue-per customer (RPC) decoupling, in which the amount of allowed FC recovery
is based on the number of customers each year. This method can be implemented on a
class-specific basis or on an aggregate basis (across the five rate classes) each year.
Inflation, in which the amount of allowed FC recovery is increased each year according
to the overall inflation index based on Gross Domestic Product (GDP).
Forecast growth, in which the amount of allowed FC recovery is predetermined on the
basis of the IRP forecasts of number of customers , electricity sales, and peak demand
for each year. Combined with the rate structures proposed in the 2003 rate case , these
forecast values determine the amount ofFC revenues expected to be collected each year.
Table 2 shows the forecasts prepared for the company s 2004 IRP used to simulate these
three recoupling mechanisms. Over the 4-year period from 2003 to 2006 , growth is highest for
forecast revenue (7.3%) and lowest for inflation (6.%). Because of the relative magnitudes
these forecasts, decoupling on the basis of forecast load growth will yield more revenue to
cover IPC's fixed costs than would RPC decoupling, which , in turn , would yield more revenue
than would use of the GO P inflation factor.
Other forms of recoupling might be feasible, but have not yet been incorporated into the workbook or tested.
Table 2.Year-to-year growth for three IPC recoupling mechanisms
Revenue per GDP Forecast
customer inflation revenue
2004024 020 027
2005023 020 023
2006023 021 021
2004 to 2006071 061 073
The workbook considers two foffi1s of recoupling: (1) all five rate classes face the same
changes in energy and demand charges because of decoupling, or (2) recoupling is done on a
class-specific basis. In the latter case, some classes could face rate increases at the same time
other classes face rate decreases. Although this might be hard to explain to the public , class-
specific decoupling might be more equitable because it considers separately the contribution
from each class to FC recovery.
Finally, the workbook adjusts rates in one of two ways: (1) energy and demand charges
or (2) energy charges only. This distinction is irrelevant for classes 1 and 7 (Residential and
Small General) because these two classes do not face demand charges. Customers in the three
other rate classes with high load factors would prefer a mechanism that adjusted both energy
and demand charges, while customers with low load factors would favor adjustments to only
the energy charge.
Appendix B contains additional detail on this workbook. The workbook contains many
assumptions necessary to conduct the calculations and to focus on the essentials rather than the
details. The key assumptions include:
All year-to-year changes in variable energy costs are recovered through the Power Cost
Adjustment (PCA) clause.
None of the transmission and distribution costs are variable; all of these costs are fixed.
The schedule 9 and 19 subclasses (Secondary, Primary, and Transmission) can be
combined into single classes to simplify the present analyses.
The various demand components (basic, summer , and nonsummer) can similarly be
combined into one demand component (and charge) for each relevant schedule (9 , 19,
and 24).
The basic demand component varies from year to year with the IRP forecasts of average
peak monthly demand (average of the 12 monthly peaks) each year.
The summer and nonsummer demand components vary from year to year with the IRP
forecasts of maximum monthly demand (maximum of the 12 monthly peaks) each year.
Only five rate classes are considered here (1 , 7,9, 19 , and 24); the other classes (which
together, account for only of IPC's revenues) are ignored.
The decoupling rate adjustments occur without any lag (i.e., in the same year the costs
change). That is, this analysis ignores the complications of balancing accounts and after-
the-fact trueups that would affect rates in subsequent years.
The decoupling mechanisms considered here are all weather-normalized. That is
they-unlike current ratemaking-compensate the company for its fixed costs on the
basis of normal weather conditions.
6. IPC DECOUPLING-MODEL RESULTS
BASE CASE
The base case is defined as the situation forecast for the 2004 IRP in terms of annual
growth in the number of customers, peak demand, and energy use for each customer class. The
effects on customers and on IPC's FC recovery is exactly as expected , based on the three-year
growth in the three recoupling mechanisms.
With forecast recoupling, there are no adjustments (by definition); i.e., actual growth
in customers, energy, and demand match expected growth in these factors. Company losses
(and customer bill reductions) are greater with inflation recoupling than with RPC recoupling,
Table 3 and Fig. 5 show the effects of these two decoupling mechanisms on each rate class
when decoupling is implemented on a class-specific basis and when it is implemented in
aggregate (last column in Table 3). The results show both percentage and absolute changes in
customer bills (and IPC FC revenues), demand charges , and energy charges. (Because classes
Similarly, customer payments for fixed costs are weather normalized. For example , if the weather one year
is extreme, the company will collect (and consumers will pay) less money for transmission and distribution with
decoupling than it (they) would under traditional ratetnaking. Adding a weather-adjustment component to a recoupling
mechanism is feasible but complicates the calculations. Doing so would require use of the IPC computer models that
weather adjust sales for each customer class and development of assumptions on "actual" weather (heating and cooling
degree days) in future years.
1 and 7 do not have demand charges, these numbers are always zero.Annualized changes are
one-third the 3-year totals presented here.
Table 3.Base-case results (3-year changes in electric bills and rates relative to case
with no decoupling) for RPC and inflation recoupling, 2004 to 2006a
Rate Class Aggre-
Totalb gate b
Revenue-per-customer recoupling
% Electric Bill
$ Electric Bill32010581539 2009 1593 2694 801
(thousand $)
% E/D Charges - 3 .
Energy Charge
(~/kWh)
Demand Charge
($/kW -month)
Inflation recoupling
Electric Bill - 0 .
$ Electric Bill299915783637 1280 1813 7681 - 7 681
(thousand $)
% E/D Charges
Energy Charge
~IkWh)
Demand Charge - 0 .
($/kW-month)
Re suIts for forecast recoupling are not shown because it is the base case.
These percentage and dollar changes are the same as those IPC would experience in its
recovery of fixed costs.
All the results shown in this section apply the same percentage change to energy and demand charges. It would
be possible (and the Recoupling model is set up) to adjust energy charges only. It is not possible to adjust demand
charges only because classes 1 and 7 pay no demand charges.
oes
en
...I ..J
II: c:r:
:I:
:! (J (C0 C 0
~ Z ~en ct
:) ~
(J
Z w 0- C N
----
W )-
c( wJ: Z(.) w
Recoupling
U'J ...J ...J W
OJ a:
a:: c:r:W J:
(J CD0 C g
~ Z C'\ICJ) ct ::J
-=: ~
(J -= 0
Z W 0- C C'\I
W ~
CJ C)
a::cz: w:I: Z
(.)
Recoupling
Fig. 5.
ecoupiingMetric Per-CustomerRecou ling
I;~~~~~'i/~'
rl~
1;:,cjO,
:,:'),
1~t~r~!\
ills nergy/Dem and harges
Total
RA TE CLASS
Recoupling Metric: Inflation Recoupling
~;fi~"
~~::
';i,~r~ifiw,'L""
~,;"..,.
ij,WM~l
;;~:;:::-:;~
1.:,
~:~~%~~i?,
ILl Bills 0 Energy/Demand Charges
-10
Total
RATE CLASS
Three-year effects of two recoupling mechanisms on customer bills and
energy Idemand charges by rate class. With RPC decoupling, IPC collects
$2.million less than it would with no decoupling mechanism. With
Inflation decoupling, IPC collects $7.7 million less over this 3-year period.
Under these base-case conditions, the forecast load growth recoupling
mechanism yields no changes in customer bills or rates.
The effects are much greater for inflation decoupling than for RPC decoupling because
the assumed growth in inflation is lower than the assumed growth in the number of customers
(6.1 v 7.% over the 3-year analysis period). With class-specific recoupling, customer bills (and
IPC FC recovery) are cut by $2.7 million with RPC decoupling and by $7.7 million with
inflation decoupling, compared with the base-case recovery of fixed costs (absent any
decoupling mechanism) of$946 million over the 3-yearperiod. These reductions represent 0.
and 0.45% of total customer bills for this 3-year period.
The effects of the two mechanisms under base-case conditions are greatest for Class 7
but result in bill and rate increases for class 24. * The percentage changes in the energy and
demand charges are greater than those in overall bills because customer bills increase under
base-case conditions and because the customer charge is unaffected by decoupling.
Although there are substantial differences in the results between the two recoupling
mechanisms, among rate classes when implemented on a class-specific basis , and between the
total and aggregate results for RPC decoupling, these effects are all small. For example, the 3-
year effect on customer bills is well under 1 percent. The effects on rates , although larger in
percentage terms, are also smalL
I next tested each of the three recoupling mechanisms against different growth rates for
customers, demand, and energy. The results of these analyses are discussed below , separately
for each of the three recoupling mechanisms.
REVENUE PER CUSTOMER RECOUPLING
Because this recoupling mechanism is based on one component of customer bills (the
monthly customer charge), the results differ according to differences in growth rates among the
three billing components (customers, demand, and energy).
As noted above, the base case results when all classes are treated the same (aggregate)
are quite different than when the classes are treated separately. The effects are much larger for
the class-specific recoupling, presumably because of the large differences among classes in the
fixed-cost-per-customer amounts, ranging from $420 for class 7 to $206 000 for class 19 , and
because the results for class 24 (and sometimes for class 1) are of the opposite sign than those
of the other classes and the aggregate.
Appendix Table A -3 shows results for cases in which one or more of the billing
determinants is increased by %/year for all three years , six cases in all. In addition , the table
shows these results relative to the base-case results, the focus of this discussion.
As noted earlier, the results for Schedule 24 are suspect.
The results, for both customers and IPC, are symmetrical about the base case. That is
increasing, say, energy use by l%/year over its base-case values has exactly the same effects
but with the opposite sign of decreasing energy use by %/year relative to the base case. This
symmetry applies to the two other recoupling mechanisms also.
Increasing (or decreasing) the growth rates for all three billing determinants by the same
amount has the same effects on FC recovery as does the base case. If growth in the number of
customers is higher (lower) by %/year than in the base case , FC revenues are higher (lower)
by 50/0independent of whether decoupling is class specific or aggregate. Customer bills
increase most for class 19 (0.8%) and least for class 9 (0.3%) with the class-specific application
of this recoupling mechanism.
Increasing demand and/or energy growth, while leaving customer growth unchanged,
lowers FC revenues. The results are much more sensitive to changes in energy use than to
changes in peak demand, probably because classes 1 and 7 have no demand charges.
The effects of changes in any of these three factors are additive. For example , the effects
of increasing peak demands by %/year plus the effects of increasing electricity use by %/year
are the same as the effects of increasing both demand and energy by %/year.
INFLA TI ON REC 0 U PLIN G
Inflation recoupling is completely independent of the three billing detenninants. As with
RPC, the effects of changes in customer, demand, and energy growth are symmetrical around
the base case. That is, increasing growth in the number of customers, peak demand, or energy
use have the same effects, but with the opposite sign , as do decreasing growth in these three
factors.
Unlike RPC, the effects of inflation recoupling are the same regardless of whether it is
implemented in aggregate or on a customer-specific basis. Also unlike RPC , the effects on each
customer class are similar. Specifically, none of the six cases analyzed shows a difference in
the direction of effect across customer classes. For example , increasing all three growth rates
by %/year leads to a reduction in customer bills that ranges from -3% for class 9 to -
for class 19, with an average of - 0.6%.
Table A -4 shows results for the same set of cases discussed above for RPC , in which
one or more of the billing detenninants is increased by %/year for all three years. Changes
in energy growth rates have a much larger effect than do changes in demand , which , in turn
have a larger effect than do changes in the number of customers. The effects of changes in the
three factors are additive.
FO RECAST - LOAD-GR OWTH RECOUPLING
Forecast recoupling depends on changes in all three billing determinants. Comparing
the right-hand sides of Tables A-4 and A-5 shows that the effects of forecast recoupling,
relative to the base case, are identical to those for inflation recoupling.
As with the other two mechanisms, the results are symmetrical around the base case.
Similarly, the effects are additive across all three billing determinants.
EFFECTS OF DSM PROGRAMS
When the only change from base-case conditions is slower growth in energy sales (and
perhaps peak demand), the companys collection ofFC revenues increases (as intended) by the
same amount regardless of the recoupling mechanism in place. If demand growth is unaffected
by the assumed IPC DSM program (i., its only effects are on energy sales), the decoupling
adjustment is smaller (as expected, because revenue collection through demand charges is
unaffected). Table 4 shows the effects on IPC FC recovery for DSM programs that cut energy
and demand by %/year (i., 1 % in 20042% in 2005 , and 3% in 2006) and programs that cut
energy use only. * The effects of even such a large and effective DSM program on IPC revenues
are very small, less than of base revenues over this 3-year period. In these cases
decoupling works exactly as intended to ensure the company suffers no loss in FC revenue
because of reductions in energy use or peak demand.
Table 4.
Reductions in
Increase in IPC fixed-cost recovery (relative to base case) associated with
reductions of 1 % per year in energy use or energy use and demand
Increase in IPC fixed-cost recovery 2004- 2006
million $ PercentageEnergy only 11 0.Energy and demand 16 0.
IPC fixed-cost revenue for the 3-year period 2004- 2006 in the base case is $946
million.
The reductions in energy sales and demand described above , relative to the base case
lead to a 90/0 increase in customer electricity bills and a 3% increase in energy and demand
charges over this 3-year period. As shown in Fig. 6, the percentage rate increases are highest
for classes 7 and 24 and lowest for classes 9 and 19.
The same results would obtain for such reductions in energy and demand regardless of the mo6vation for the
energy and demand cuts.
oi5
-J LU
c: c:(W :I::2: U f,Q 3
0 c 8
C\len c:( I::::) "C .c;
~ ~ ~
w )-
" "
a:::c:( w:J: Z t)
?fl.
Recoupling Metric: Load Growth Recoupling
(J Bills0 Energy/Demand Charges
.._, '
~~"i
l!Oh'im,
~t~Iii;"",!,
ii'r.,c;
. ~'
4-"
~; . .'. ~
~f.~
w,~':,)8~it: '
~t-
;&~~"~
~wl'?'-
,..
iJi;;'~:i;;,
" ";",,""~
~~~im.
~r~;
mhr.
lPi).~ ,',"
,$0;,:\-'1,~~:J
Total
R QoouplingRATE CLASS
Fig. 6.Effects of 1 % per year reductions in energy use
and peak demands for three years on electricity
bills and rates, relative to the base case.
7. CONCLUSIONS
Current electric-utility rate making, as practiced in most jurisdictions throughout the
United States, collects substantial revenues to recover fixed costs from variable energy charges.
This practice makes little econonric sense. Specifically, a utility s ability to recover its prudently
inculTed fixed costs depends on factors that are (a) unrelated to those costs and (b) largely
outside its controlincluding economic and population growth in its service area, which , in turn
affect energy sales.
This long-standing quirk in ratemaking unintentionally, but unavoidably, penalizes
utilities that encourage their customers to use electricity more efficiently. Thus, utilities face
a clear disincentive to help their customers improve energy efficiency.
Decoupling is a mechanism that breaks the link between electricity sales and utility
revenues. To implement decoupling, utility revenues need to be recoupled to some other
factor(s). This recoupling is necessary to ensure that the utility has an opportunity to recover
its fixed costs. However, many of the factors considered for recoupling-such as the number
of customers, inflation, or forecast revenues-may have no more logical connection to fixed
costs than does kWh sales.
Although decoupling is intended to remove the penalties in existing ratemaking for
utility DSM programs, its effects can be much broader. That is , depending on the recoupling
method chosen, utility revenues (and, therefore, customer rates and bills) can vary from year
to year independent of a utilitys DSM programs.
Decoupling is a zero-sum effort. If the company is paid more money to cover its fixed
costs (good for IPC), consumers will, unavoidably, pay more for transmission and distribution
services (bad for consumers). The reverse is also true.
The amount of the decoupling adjustment each year depends on how far from actual
conditions the recoupling mechanism is. For example , if recoupling is tied to inflation and the
actual growth in billing determinants differs substantially from inflation for that year , the
decoupling adjustment will be large. If the year-la-year changes in the number of customers,
peak demand, and energy sales yield changes in non- PCA revenues very different from the
inflation rate, the decoupling adjustment will be much larger than if the inflation rate and actual
revenues move together. Thus, decoupling does not necessarily stabilize FC recovery nor does
it make such recovery more predictable than traditional ratemaking.
Preparation of this paper was motivated by the advocacy of decoupling by the Natural
Resources Defense Council and the Northwest Energy Coalition.Cavanagh proposes that the
Idaho PUC allow the company and other interested parties three to six months to develop
design recommendations for the Commissions consideration." These recommendations are
to consider the recoupling mechanism, separate v combined treatment of rate classes , weather-
normalization of the recoupling mechanism, and the frequency with which true-ups are to
occur. Cavanagh suggests there is ample "analysis and experience" to support a workable
mec hani sm.
I agree with Cavanagh that such a mechanism can be developed. Indeed , this paper
examined three such alternatives. The larger questions , in my view , are:
Does decoupling make sense to IPC at this time? IPC's DSM programs cuITently
operate at a very modest level, yielding only small effects on energy use. The 2004 IRP
might propose additional, stronger programs. But those programs are likely to focus on
reductions in summer peak demand more than on year-round energy efficiency. As
such, the new programs may have little effect on IPC's kilowatt-hour sales.
What unintended effects might decoupling have? Although decoupling would
completely sever the link between energy sales and utility revenues , it can and will
affect utility revenues for other reasons. In particular , the combination of a recoupling
Indeed, regulators in Maine and Washington abandoned decoupling in the mid-1990s largely for reasons
independent of the utilities' energy-efficiency programs. Decoupling in both states led to large rate increases because
of a slowdown in the economy (Maine) or high power costs (Washington).
R. CavanaghDirect Testilnony of Ralph CavanaghCase No. IPC-033-, before the Idaho Public Utilities
Commission, February 20, 2004.
mechanism and large changes in the factors affecting that mechanism could yield
nontrivial year-to-year changes in IPC revenues and , therefore , in customer bills and
rates.
Given the uncertain answers to these two questions , I recommend that IPC maintain an
open mind about decoupling. Specifically, I suggest the company accept Cavanagh'
suggestion and form a decoupling collaborative to work on these issues at the conclusion of the
cutTent rate case. Hopefully, this paper will serve as useful background for that collaborative.
There is no way to know what IPC' s actual fixed costs and FC recovery would be in the
future. They might be higher (or lower), more (or less) predictable, and more (or less) stable
than without decoupling. Absent detailed information on expected fixed costs and the
determinants of these costs, function by function , the potential benefits of decoupling with
respect to revenue predictability and stability remain unknown.
From a theoretical perspective, the recoupling mechanism should be tied to factors that
directly affect a utility fixed costs. Such factors are surely function specific , with different
factors affecting fixed costs for generation, transmission , and distribution. Developing such a
mechanism could be time consuming and complicated (as evidenced by the Electric Revenue
Adjustment Mechanism used in California from the early 1980s through the early 1990s).
Absent such a detailed understanding of utility fixed costs and their determinants , recoupling
uses mechanisms that relate to fixed costs no better than do kilowatt-hour sales, the current
approach to ratemaking.
My bottom line, based on past experience and the analyses presented here , is that
decoupling is likely to have only modest effects on IPC revenues and customer bills. It could
have slightly larger effects on the energy and demand rates for particular customer classes
depending on the specifics of the recoupling mechanism.
C;\D a ta \Docs \IPC\IPCDec our ling Report. wpd
APPENDIX A: PAST EXPERIENCE WITH DECOUPLING
This brief discussion is divided into three parts, the first dealing with decoupling during
the mid-1980s to early 19908, the second covering the Oregon decoupling collaborati yes in the
early- to mid-1990s, and the third dealing with decoupling implemented after the Western
electricity crisis of 2000/2001.
MID-1980s TO EARLY 1990s
California was the first state, in 1981, to implement a decoupling system , called the
Electric Revenue Adjustment Mechanism (ERAM) (Marnay and Comnes 1992). Once every
three years, the California pue set rates for each of the state s utilities in a general rate case.
The rate-case process, based on a future test year, included a determination of the amount of
money the utility could collect for its fixed costs. The ERAM mechanism was used to ensure
that for the years between rate cases the utility collected the carTect amount of money to cover
these costs.
The PUC used attrition mechanisms to determine the amount of money the utility could
collect each year. Financial attrition adjusted for changes in the utility s cost of capital. These
adjustments were handled in annual proceedings that set interest rates and return on equity for
all the California utilities.
Operational attrition adjusted for changes in operating costs , such as wage rates and the
costs for certain materials. These costs were adjllsted on the basis of price indices.
Finally, rate-base attrition adjusted for changes in the utility ratebase. These
adjustments were based primarily on forecasts of capital expenditures developed during the
general rate cases.
During the first decade of operation, ERAM had very small effects on utility rates and
volatility.
New York, during the late 1980s and early 1990s , used decoupling mechanisms similar
to Californias ERAM.
Washington and Maine adopted decoupling mechanisms in 1991 (Washington Utilities
and Transportation Commission 1992; Maine PUC 1993). Neither state used the California
approach. Instead, these states adjusted allowed fixed costs on the basis of growth in the
number of electricity customers.
The mechanisms adopted in Washington and Maine were used for only a few years. The
commissions abandoned decoupling because of substantial rate increases. These rate increases
had nothing to do with the utilitys DSM programs. In Washington power-supply costs (which
were part of the decoupling mechanism) increased sharply, which led to decoupling-related
price increases. In Maine, slower than expected economic growth led to rate increases.
MID-1990s
POE (1993) and PacifiCorp (1993) conducted decoupling collaboratives , in response
to an order from the Oregon PUC. The PGE collaborative proposal included the following
steps:
Establish base revenues using a 2- year test period
Establish monthly revenue benchmarks and incremental power cost estimates
Restate actual sales and revenues as if normal weather had occurred
Implement decoupling rate adjustments every six months
Amortize decoupling adjustments over 18 months
Spread decoupling adjustment among customer classes using the rate spread adopted
by the PU C in the 1991 general rate case.
In March 1995, the Oregon PUC adopted the PGE collaborative mechanism. The
following year, the PUC declined to adopt a decoupling mechanism for PacifiCorp. However
in 1998, the PUC ordered PacifiCorp to adopt an Alternative Form of Regulation that applied
decoupling only to the distribution function.
In 2001, PGE (Lesh 200 1) proposed a distribution-only decoupling mechanism for
residential and small nonresidential consumers only. The mechanisms would apply on a per
customer basis. The PUC rejected the POE proposal.
EARLY 2000s
During the past two years, the California PUC , in response to state legislation , has
reintroduced decoupling for the California utilities (Bachrach and Carter 2004). Southern
California Edison culTently has a decoupling mechanism in place for distribution costs only,
using a revenue-per-customer approach. The company proposed to add fixed-generation costs
to a new decoupling mechanism, using ERAM-like mechanisms. PG&E proposed to decouple
fixed costs for distribution and generation using an inflation index. SDG&E proposed a
revenue-per-customer mechanism.
As of now, decoupling operates in California and in Oregon only. While other states
may be considering decoupling, none has such mechanisms in place.
SUMMARY
Four states adopted decoupling mechanisms during the mid-1980s through early 1990s.
These experiences suggest the following lessons. The California ERAM mechanisms worked
as expected and yielded very small rate adjustments. However , these mechanisms can be
complicated, and the annual mini-rate cases required for implementation can be contentious.
The Washington and Maine experiences show that decoupling can have effects that go well
beyond those related to utility DSM programs. In particular, nontrivial changes in other factors
included in the decoupling mechanism (power-supply costs in Washington and changes in the
trend of per-customer electricity use in Maine) can lead to politically unacceptable rate
Increases.
The Oregon experience during the mid-1990s included different decoupling mechanisms
for PGE and PacifiCorp. More recently, the California PUC is , once again, implementing
decoupling, and other states are considering such mechanisms.
Although the initial decoupling experiments were reasonably well documented
(especially Californias), that is not the case for the more recent experiments. In particular , I
had a tough time finding (and understanding) information on the Oregon and recent California
experiences. Perhaps more important, I could find no study on the effects and effectiveness of
decoupling on utility DSM programs. As a consequence , we have no idea what the practical
effect, if any, is of decoupling on a utilitys incentive to run cost-effective programs.
REFEREN CES
D. Bachrach and S. Carter
, "
Status of California s Policy Efforts to Eliminate Utilities
Disincentive to Invest in Energy Efficiency and Distributed Generation " Natural Resources
Defense Council, San Francisco, CA, February 27 , 2004.
P. G. Lesh 2001
, "
Advice No. 01-, Distribution Decoupling Adjustment " Letter to Oregon
PUC, Portland, OR, March 19.
Maine Public Utilities Commission 1993Order Approving Stipulation Docket Nos. 90-085-
et aI., Augusta, ME, February 5.
C. Marnay and G. A. Comnes
, "
Californias ERAM Experience " Chapter 3 in Regulatory
Incentives for Demand-Side Managementedited by S. M. Nadel , M. W. Reid , and D.
Wolcott, 39 - 62, American Council for an Energy-Efficient Economy, Washington , DC , 1992.
PacifiCorp 1993Report of PacifiCorp Decoupling Collaborative Portland , OR , May.
Portland General Electric 1993Decoupling Collaborative Final Report Portland , OR , April.
Washington Utilities and Transportation Commission 1992 First Supplemental Order
Rejecting Tariff Filing; Authorizing Refiling, Docket No. UE-920630 , Olympia W A
September 24.
APPENDIX B. DETAILS ON RECOUPLING WORKBOOK
The Recoupling workbook contains three sheets: 1&0, Base , and Calc. The top part of
the first sheet (1&0, which stands for inputs and outputs) contains all the user inputs , while the
bottom part contains the decoupling results. The user inputs include class-specific or aggregate
growth rates (%/year relative to the base case discussed below) for the number of customers,
peak demand, and electricity sales. In addition, the user specifies which of the three forms of
recoupling to use, whether results are calculated on a class-specific basis or in aggregate , and
whether differences between actual and allowed fixed cost-recovery are collected or refunded
through energy and demand charges or through energy charges only.
The bottom part of 1&0 contains results for the particular decoupling case chosen (left-
band side) as well as the base case (right-hand side). Decoupling results (all of which are
presented relative to the no-decoupling base case) include:
Percentage and dollar changes in annual electric bills
Changes in IPC recovery of fixed costs
, *
and
Percentage and actual changes in energy (~IkWh) and demand ($/kW-month) charges.
The Base sheet contains information from the 2003 rate-case filing, in particular data
from Brilz exhibits 42 and 43; see Table A-I. These data include characteristics of each rate
class (number of customers, basic demand, summer and nonsummer demand, and summer and
nonsummer energy use); proposed rate structures for each class; year 2003 revenues for each
customer class based on the proposed rate structures~ and the fixed and variable costs for each
class. Table 1 summarized these results for each rate class.
In addition, the Base sheet contains the company s IRP forecasts for 2004 , 2005 , and
2006 of the number of customersmaximum monthly demand, annual average of the maximum
monthly demands each year, and electricity sales for each of the five rate classes, as well as the
overall inflation rate. Table A-2 shows these results.
These two sets of inputs are combined to calculate base-case results on class-specific
and total revenues, including recovery of fixed costs.
The Calc sheet calculates decoupling results given the inputs provided in 1&0. These
results, for 2004, 2005and 2006include the number of customers , the three demand
components, annual energy use, revenues collected from retail customers , revenues collected
for fixed costs (i., those not collected through the PCA), and allowed FC recovery (based on
the form of recoupling selected in 1&0).
Changes in IPC recovery of fixed costs are equal in magnitude and opposite in sign to the changes in annual
customer electric bills.
Table A-1. Inputs to Recoupling workbook from 2003 rate case
Rate Class Total or
vera~
Rate Class Characteristics
# of customers334917152 076 105 684 397 934
Basic 737 530 267
~ SSummer 399 997 040 436E ~Nonsummer 709 908 515 131
Total20 ,845 434 555 834
..c:Sum mer932072475800 214 505 668 1 ,226 233 532 662e' 3:Nonsummer 3209321 196860 2214 213 1 473,156 312 462 406 012
~ ~
Total 4141393 265335 3014 427 1 978 824 538 695 10 938 674
2003 Proposed Idaho Rates
Customer, $/month10.10.24.500.
"'0 Basic 1 .
3;: aSummer
E ~ Non-summer
Q.) Y:J-
...
Average
..c:Summer06140729 0290 0249 0326OJ ~ 3:Non-summer04910583 0252 0212 0457c:: ~
-...
Total05190620 0262 0221 0353 0373Y7-
Fixed Cost Percentages
of total costs63.69.46.36.60.56.
of requested rev req60.66.43.34.80.56.
2003 Proposed Revenues (thousand $)
Customer, 1 858 957 628 374 51 ,008
Demand40 087 379 416 882
Energy214787463 961 773 306 408,291
Total254977321124 006 780 096 534 180
Costs, thousand $
Variable101888832 595 41 , 197 893 233 406
Fixed153089489 411 21 ,583 58,203 300 775
Total254977321124 006 780 096 534 180
Variable, $/MWh024602570231 0208 0090 0213
:ixed-Cost Revenue/Customer457.419.186 206 278 253 755.
Table A-2. Base-case rowth rates
O/aearfrom IRP
Rate Class Total or
24 Averag~
Customers
2004
2005
2006
Cumulative
Maximum MW
2004
2005
2006
Cumulative
Average MW
2004
2005
2006
Cumulative
Sales
2004
2005
2006
Cumulative
Price Deflator
YearPCWGDP Inflation/yr
2003127
2004149
2005171
2006195
Cumulative061
Ta
b
l
e

A
.
.
3.

I
P
C

D
e
c
o
u
Re
s
u
l
t
s
:

R
e
v
e
n
u
e
-
er
-
c
u
s
t
o
m
e
r

De
c
o
u
,
2
0
0
4
.
.
.

2
0
0
6
Ra
t
e

C
l
a
s
s
To
t
a
l
Ag
g
r
e
g
.
C,

D
,
E

G
r
o
w
t
h

R
a
t
e
s
,
%
/
y
e
a
r

=

0
%
E
l
e
c
t
r
i
c

B
i
l
l
$
E
l
e
c
t
r
i
c

B
i
l
l
32
0
05
8
53
9
20
0
9
15
9
3
26
9
4
80
1
%
E
/
D

C
h
a
r
g
e
s
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e
Di
f
f
e
r
e
n
c
e
s

f
r
o
m

0
,
0
B
a
s
e

C
a
s
e
Ra
t
e

C
l
a
s
s
,
D
,
E

G
r
o
w
t
h

R
a
t
e
s
,

%
/y
e
a
r

=

1
To
t
a
l
Ag
g
r
e
g
.
%
E
l
e
c
t
r
i
c

B
i
l
l
$
E
l
e
c
t
r
i
c

B
i
l
l
32
3
06
9
55
5
20
3
0
16
0
9
27
2
1
80
9
%
E
/
D

C
h
a
r
g
e
s
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e
,
D
,

E

G
r
o
w
t
h

R
a
t
e
s
,
%
/
y
e
a
r

=

1
,
%
E
l
e
c
t
r
i
c

B
i
l
l
0.
4
8
0.
4
4
0.
4
6
0.
4
6
$
E
l
e
c
t
r
i
c

B
i
l
l
38
6
7
75
4
13
6
3
33
6
4
51
3
7
70
5
0
35
4
7
30
4
15
6
2
64
7
17
7
1
78
3
1
78
5
0
%
E
/
D

C
h
a
r
g
e
s
1.
5
8
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e
C,

D
,

E

G
r
o
w
t
h

R
a
t
e
s
,
%
/
y
e
a
r

=

0
/0

El
e
c
t
r
i
c

B
i
l
l
0.
4
8
$
E
l
e
c
t
r
i
c

B
i
l
l
32
0
05
8
28
1
1
25
9
4
10
7
5
50
6
9
31
7
5
12
7
1
58
5
51
8
23
7
5
23
7
5
%
E
/
D

C
h
a
r
g
e
s
0.
4
6
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e
,
D
,
E

G
r
o
w
t
h

R
a
t
e
s
,
%
/
y
e
a
r

=

0
,
%
E
l
e
c
t
r
i
c

B
i
l
l
0.
4
8
$
E
l
e
c
t
r
i
c

B
i
l
l
32
2
4
37
3
18
4
6
20
9
1
35
6
81
7
8
62
8
5
35
4
4
31
5
30
6
12
3
7
54
8
4
54
8
4
%
E
/
D

C
h
a
r
g
e
s
1.
4
2
0.
4
8
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e
,
0
,
E

G
r
o
w
t
h

R
a
t
e
s
,

c
/o
/
y
e
a
r

=

0
,
%
E
l
e
c
t
r
i
c

B
i
l
l
0.
4
8
0.
4
7
$
E
l
e
c
t
r
i
c

B
i
l
l
32
2
4
13
7
3
31
1

7
26
7
6
16
2

-
10
5
5
3
86
5
9
35
4
4
31
5
15
7
8
66
7
75
5
78
5
8
78
5
8
E/
D

Ch
a
r
g
e
s
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e
Ta
b
l
e

A
-
4.

I
P
C

D
e
c
o
u
p
l
i
n
g

R
e
s
u
l
t
s
:

I
n
f
l
a
t
i
o
n

D
e
c
o
u
p
l
i
n
g
,

20
0
4

-

2
0
0
6
Ra
t
e

C
l
a
s
s
19

24

To
t
a
l

A
g
g
r
e
g
.
C,

D
,

E

G
r
o
w
t
h

R
a
t
e
s
,
%
/y
e
a
r

=

0
%
E
l
e
c
t
r
i
c

B
i
l
l

-
37

-
39

-
91

-
$
E
l
e
c
t
r
i
c

B
i
l
l

-
29
9
9

-
15
7
8

-
36
3
7

-
12
8
0
%
E
/
D

C
h
a
r
g
e
s

-
33

-
72

-
81

-
$
E
n
e
r
g
y

C
h
a
r
g
e

-
07

-
54

-
07

-
$
D
e
m
a
n
d

C
h
a
r
g
e

0
.
00

0
.
00

-
05

-
,
D
,
E

G
r
o
w
t
h

R
a
t
e
s
,

%
/
y
e
a
r

=
1
1
J

1
%
E
l
e
c
t
r
i
c

B
i
l
l

-
98

-
05

-
$
E
l
e
c
t
r
i
c

B
i
l
l

-
78
0
6

-
20
1
5

-
53
7
1
%
E
/
D

C
h
a
r
g
e
s

-
3.
4
4

-
11
.
04

-
$
E
n
e
r
g
y

C
h
a
r
g
e

-
18

-
69

-
$
D
e
m
a
n
d

C
h
a
r
g
e

0
.
00

0
.
00

-
19
6
6
C,

D
,
E

G
r
o
w
t
h

R
a
t
e
s
,
%
/
y
e
a
r

=

1
%
E
l
e
c
t
r
i
c

B
i
l
l

-
53

-
58

-
95

-
$
E
l
e
c
t
r
i
c

B
i
l
l

-
42
6
2

-
17
0
0

-
37
9
4

-
12
9
9
%
E
/
D

C
h
a
r
g
e
s

-
1
.
89

-
40

-
94

-
$
E
n
e
r
g
y

C
h
a
r
g
e

-
10

-
58

-
08

-
$
D
e
m
a
n
d

C
h
a
r
g
e

0
.
00

0
.
00

-
06

-
82

-
18
1
3

-
76
8
1
2.
4
8

-
09

-
06

-
76
8
1
Di
f
f
e
r
e
n
c
e
s

f
r
o
m

0
,
0
B
a
s
e

C
a
s
e
Ra
t
e

C
l
a
s
s
1
9

66

-
43

-
34

-
43
7

-
17
3
4

-
68
6

-
17
9
8
32

-
31

-
01

-
2.
4
7
14

-
03

-
02

-
00

-
03

-
02

-
To
t
a
l

A
g
g
r
e
g
.
56

-
94
6
2

-
94
6
2
82

-
07

-
04

-
01

-
1
5

-
71
4
3
01

-
00

-
00

-
01
5
17
1
4
3
32
8
12
4
07
0
48
0
7
17
7
0
92
8
4
92
8
4
26
3
12
2
15
7
-
1

60
3
60
3
,
0
,
E

G
r
o
w
t
h

R
a
t
e
s
,
%
/y
e
a
r

=

0
,
%
E
l
e
c
t
r
i
c

B
i
l
l
$
E
l
e
c
t
r
i
c

B
i
l
l
29
9
9
15
7
8
49
0
9
18
6
4
12
9
5
00
5
5
00
5
5
12
7
1
58
5
51
8
23
7
5
23
7
5
%
E
/
D

C
h
a
r
g
e
s
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e
j
D
,
E

G
r
o
w
t
h

R
a
t
e
s
,
%
/
y
e
a
r

=

0
,
%
E
l
e
c
t
r
i
c

B
i
l
l

-
82

-
87

-
98

-
$
E
l
e
c
t
r
i
c

B
i
l
l

-
65
4
3

-
18
9
3

-
39
4
3

-
13
6
2
%
E
/
D

C
h
a
r
g
e
s

-
88

-
10
.
37

-
03

-
$
E
n
e
r
g
y

C
h
a
r
g
e

-
15

-
64

-
08

-
$
D
e
m
a
n
d

C
h
a
r
g
e

0
.
00

0
.
00

-
06

-
57
6
13
1
6
5
13
1
6
5
35
4
4
31
5
30
6
12
3
7
54
8
4
54
8
4
C,

D
,

E

G
r
o
w
t
h

R
a
t
e
s
,
%
/
y
e
a
r

=

0
%
E
l
e
c
t
r
i
c

B
i
l
l
0.
4
4
0.
4
8
$
E
l
e
c
t
r
i
c

B
i
l
l
65
4
3
18
9
3
52
1

5
19
4
6
15
5
3
9
15
5
3
9
35
4
4
31
5
57
8
66
7
17
5
5
78
5
8
78
5
8
%
E
/
D

C
h
a
r
g
e
s
10
.
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e
Ta
b
l
e

A
-
5
.

I
P
C

D
e
c
o
u
Re
s
u
l
t
s
:

F
o
r
e
c
a
s
t

G
r
o
w
t
h

De
c
o
u
,
2
0
0
4

-

2
0
0
6
Ra
t
e

C
l
a
s
s
To
t
a
l

A
g
g
r
e
g
.
,
D
,
E

G
r
o
w
t
h

R
a
t
e
s
,

ol
e
/
y
e
a
r

=

0
%
E
l
e
c
t
r
i
c

B
i
l
l
$
E
l
e
c
t
r
i
c

B
i
l
l
%
E
/
D

C
h
a
r
g
e
s
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e
Di
f
f
e
r
e
n
c
e
s

f
r
o
m

0
,
0
B
a
s
e

C
a
s
e
Ra
t
e

C
l
a
s
s
,
D
,
E

G
r
o
w
t
h

R
a
t
e
s
,
%
/
y
e
a
r

=

1
To
t
a
l
Ag
g
r
e
g
.
%
E
l
e
c
t
r
i
c

B
i
l
l
0.
4
3
56
0
$
E
l
e
c
t
r
i
c

B
i
l
l
48
0
7
43
7
17
3
4
68
6
79
8
94
6
2
94
6
2
48
0
7
43
7
73
4
68
6
79
8
94
6
2
94
6
2
%
E
/
D

C
h
a
r
g
e
s
2.
4
1
83
8
$
E
n
e
r
g
y

C
h
a
r
g
e
06
9
$
D
e
m
a
n
d

C
h
a
r
g
e
03
8
Cj

D
,

E

G
r
o
w
t
h

R
a
t
e
s
,
%
/y
e
a
r

=

1

,
%
E
l
e
c
t
r
i
c

B
i
l
l
$
E
l
e
c
t
r
i
c

B
i
l
l
12
6
3
12
2
15
7
16
0
3
16
0
3
26
3
12
2
15
7
60
3
60
3
%
E
/
D

C
h
a
r
g
e
s
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e
C,

D
,
E

G
r
o
w
t
h

R
a
t
e
s
,
%
/y
e
a
r

=

0
,
%
E
l
e
c
t
r
i
c

B
i
l
l
$
E
l
e
c
t
r
i
c

B
i
l
l
12
7
1
58
5
51
8
23
7
5
-
2
3

7

5
12
7
1
58
5
51
8
23
7
5
23
7
5
%
E
/
D

C
h
a
r
g
e
s
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e
,
D
,
E

G
r
o
w
t
h

R
a
t
e
s
,

ol
a
/
y
e
a
r

=

0
%
E
l
e
c
t
r
i
c

B
i
l
l
0.
4
4
0.
4
8
$
E
l
e
c
t
r
i
c

B
i
l
l
35
4
4
31
5
30
6
23
7
54
8
4
54
8
4
35
4
4
31
5
30
6
23
7
54
8
4
54
8
4
%
E
/
D

C
h
a
r
g
e
s
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e
C,

D
,
E

G
r
o
w
t
h

R
a
t
e
s
,

%
/y
e
a
r

=

0
%
E
l
e
c
t
r
i
c

B
i
l
l
0.
4
4
0.
4
8
0.
4

7
$
E
l
e
c
t
r
i
c

B
i
l
l
35
4
4
31
5
15
7
8
66
7
17
5
5
78
5
8
78
5
8
35
4
4
31
5
15
7
8
66
7
17
5
5
78
5
8
78
5
8
%
E
/
D

C
h
a
r
g
e
s
$
E
n
e
r
g
y

C
h
a
r
g
e
$
D
e
m
a
n
d

C
h
a
r
g
e