HomeMy WebLinkAbout20190109Avista to Staff 11-22.pdfAvista Corp.
1411 East Mission P.O.Box3727
Spokane. Washington 99220-0500
Telephone 509-489-0500
TollFree 800-727-9170
4Wtsrfr
Gorp,RilC E IVT D
i8l9 J&t{ -9 AH 9: t+3
January 8,2019
Idaho Public Utilities Commission
472W. Washington St.
Boise, lD 83720-0074
Attn: Diane Hanian
Re: Production Request of Commission Staff in Case No. AVU-E-18-12
Dear Ms. Hanian,
Enclosed is Avista's response to IPUC Staffs production request in the above referenced docket.
Included in this mailing are the original and two paper copies of Avista's response to production
request: Staff ll-22. The electronic version of the responses were emailed on 0110812019.
Also included is Avista's CONFIDENTIAL response to PR_llC. These responses contain
TRADE SECRET, PROPRIETARY or CONFIDENTIAL information and are separately
filed under IDAPA 31.01.01, Rule 067 and233, and Section 9-340D,Idaho Code. Due to their
voluminous nature, they are being provided in electronic format only, via thumb drive, under a
sealed separate envelope marked CONFIDENTIAL.
If there are any questions regarding the enclosed information, please contact Paul Kimball at
(509) 495-4584 or via e-mail at paul.kimball@avistacorp.com.
Sincerely,
Paul Kimball
Mgr., Compliance & Discovery
Enclosures
JiJLIC
MF,4lSSlOt;
AVISTA CORPORATION
RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION:
CASE NO:
REQUESTER:
TYPE:
REQUEST NO.:
IDAHO
AVU-E-18-12
IPUC Staff
Production Request
Staff- 12
DATE PREPARED
WITNESS:
RESPONDER:
DEPARTMENT:
TELEPHONE:
1212912018
N/A
Amber Gifford
Energy Efficiency
(s09) 49s-2896
REQUEST:
Exhibit 1, Section 5.5 indicates that the Company's fuel efficiency program realization rate
was 62%o. What steps are the Company taking to improve the accuracy of the methodology
used to estimate this program's savings?
RESPONSE:
Avista has updated its Fuel Efficiency program's savings values based on Nexant's
2016-2017 Electric Impact Evaluation Report.
AVISTA CORPORATION
RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION:
CASE NO:
REQUESTER:
TYPE:
REQUEST NO.:
IDAHO
AVU-E-18-12
IPUC Staff
Production Request
Staff- l3
DATE PREPARED: 1212912018WITNESS: N/A
RESPONDER: Amber Gifford
DEPARTMENT: Energy Efficiency
TELEPHONE: (s09) 49s-2896
REQUEST:
What is the increase in the number of Therms of natural gas consumed by customers who
participated in the Company's fuel efficiency program (Converted their HVAC and other
appliances to Natural Gas)? What is the average net savings, in dollars, of customers who
participated in this program?
RESPONSE:
As is reported in Nexant's 2016-2017 Natural Gas Impact Evaluation, the increase to the
number of therms used by residential customers due to the Company's Fuel Efficiency
Program in 2017 was 82,948 (verified gross)r . For the 2016 program year, the increase was
350,976 therms (unverified - adjusted reported gross)2
Avista does not record the billing impact of a conversion on a per customer basis, however,
the Company could calculate an estimated annual net savings value based on an average of
electric and natural gas rates for a particular period of time.
I ldaho 2016-2017 Natural Gas Impact Evaluation, Table 5-15
2 Idaho 2016 Annual Conservation Report, Table 3-7
AVISTA CORPORATION
RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION:
CASE NO:
REQUESTER:
TYPE:
REQUEST NO.:
IDAHO
AVU-E-18-12
IPUC Staff
Production Request
Staff- 14
DATE PREPARED
WITNESS:
RESPONDER:
DEPARTMENT:
TELEPHONE:
1212912018
N/A
Amber Gifford
Energy Efficiency
(s09) 4es-2896
REQUEST:
Exhibit 1, Section 5.7 indicates that the Company's Shell program realization rate was 27%.
What steps are the Company taking to improve the accuracy of the methodolory used to
estimate this program's savings?
RESPONSE:
Avista will begin using Regional Technical Forum (RTF) savings values in 2019 for all shell
measures in the residential prescriptive portfolio.
AVISTA CORPORATION
RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION:
CASE NO:
REQUESTER:
TYPE:
REQUEST NO.:
IDAHO
AVU-E-18-12
IPUC Staff
Production Request
Staff- l5
DATE PREPARED: 1212912018WITNESS: N/A
RESPONDER: Lynn Roy
DEPARTMENT: Nexant, Inc.
TELEPHONE: (303) 792-8668
REQUEST:
Exhibit 1, Equation 3-2 includes the variables "event" and the interaction term, "treat X
event." Neither of these is described in Table 3-5. Please provide a description for these
variables. Also, please explain why the variable "treat" is not included in the model.
RESPONSE:
The below response has been provided by Avista's 3'd party EM&V evaluator, Nexant, Inc.
Equation 3-2 contains a typo. The correct regression specification is in the revised section below
and highlighted in yellow. The word "event" has been replaced with the term "post" which is the
post-treatment indicator variable. The "treatment" variable is part of an interaction term
("treatXpost") which is equal to 1 for the treatment group during the post-treatment period, and 0
otherwise. In other words, the term is equal to 0 at all time periods for the control group and during
the pre-treatment period for the treatment group. The methods described here do not apply to
Nexant's evaluation of Oracle's Home Energy Reports program, so the reference to that program
is stricken. The evaluation methodology for the Home Energy Reports program is described in
Section 5.
3.4.4.2 Ex Post Estimation Method - Revised with edits indicated in yellow
After the comparison groups for treatment customers were selected and validated, energy impacts
were estimated using a difference-in-differences (DiD) methodology for the Shell, HVAC, and
FuelEflrciency@programS(theLowIncomeandHomeEnergyReports
programs used a participant pre/post billing analysis, see Sections 3.4.4.3 and 5.8, respectively).
Impacts are estimated as the difference in average consumption between treatment and comparison
customers in each month, with the slight difference between the two groups on the pre-treatment
year removed. This calculation controls for residual differences in load between the groups that are
not eliminated through the matching process, thus reducing bias.
The DiD analysis can be done by hand using simple averages or by using panel regression
analysis. Customer fixed effects regression analysis allows each customer's mean consumption to
be modeled separately, which reduces the standard error of the impact estimates without changing
their magnitude. Additionally, panel regression easily facilitates calculation of standard errors,
confidence intervals, and significance tests for load impact estimates that correctly account for the
correlation in customer loads over time.
The model specification for estimating load impacts is shown in Equation 0-1 and Table 0-l
provides detail for each model variable. The model was estimated separately for each month.
Variable Description
Equation 0-1: Monthly Energy Savings Model Specification
daily-ronsurnptionr: a * laost*BheaE(post, * u, * e
Table 0-1: Description of Energy Savings Model Regression Variables
d,aily -c ott suntption i Per customer consumption (kWh or therms) for customer i
Mean consumption for all customers
The coefficient on the post-treatment indicator variable
post Equal to I for the post-treatment period and 0 for the same month
in 2015
DiD estimator of the treatment effect (the impact in kWh or
therms)
keatXpost Interaction of treatment and post variables, equal to I for the
post-treatment period for participants and 0 otherwise
The customer fixed effects variable for customer i
The error term
In Equation 0-1 the variable daily_consatnptioa, equals electricity or gas consumption during
the time period of interest, which would be each month of the post-treatment period. The index I
refers to each individual customer. The estimating database contained electricity and gas
consumption data during the pre- and post-treatment periods for both treatment and matched
comparison group customers. The variable post is equal to I for months after installation and a
value of 0 for the same month in 2015. The treatXposl term is the interaction of treat and post and
its coeff,rcient f is a differences-in-differences estimator of the treatment effect that makes use of
the pre-treatment data. The primary parameter of interest is p, which provides the estimated
energy impact of the rebate programs during the relevant period. The parameter a is equal to mean
daily consumption for each customer for the relevant time period (e.g., monthly). The vi term is the
customer fixed effects variable that controls for unobserved factors that are time-invariant and
unique to each customer.
This was estimated for each month of 2016 and 20ll separately. Impacts are estimated on a
per-customer basis. Reference consumption is equal to observed treatment consumption plus the
estimated impact.
a
Y
F
rri
a
AVISTA CORPORATION
RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION: IDAHO
CASE NO: AVU-E-I8-12
REQUESTER: IPUC StaffTYPE: Production Request
REQUEST NO.: Staff - 16
DATE PREPARED: 1212912018WITNESS: N/A
RESPONDER: Lynn Roy
DEPARTMENT: Nexant, Inc.
TELEPHONE: (303) 792-8668
REQUEST:
Regarding Exhibit 1, Equation 5-2: What steps were taken to guard against multicollineaity?
RESPONSE:
The below response has been provided by Avista's 3'dparty EM&V evaluator, Nexant, Inc.
No steps were taken to guard against multicollinearity in estimating the model specified in
Equation 5-2 because multicollinearity is not a problem in the lagged dependent variable (LDV)
model. There are only two predictors in the model - prior electricity consumption and the
treatment variable. The treatment variable has been randomly assigned and is therefore not
correlated with prior consumption.
AVISTA CORPORATION
RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION:
CASE NO:
REQUESTER:
TYPE:
REQUEST NO.:
IDAHO
AVU-E-18-12
IPUC Staff
Production Request
Staff- 17
DATE PREPARED: 1212912018WITNESS: N/A
RESPONDER: LynnRoy
DEPARTMENT: Nexant, Inc.
TELEPHONE: (303) 792-8668
REQUEST:
Regarding Exhibit 1, Equation 5-2: What was the reason for incorporating the double sum,
II'o-,.,t=:. r=l in the model?
RESPONSE:
The below response has been provided by Avista's 3'dparty EM&V evaluator, Nexant, Inc.:
Nexant estimated monthly energy savings for two calendar years. The double sum allowed us to
compactly express these terms in one expression, covering two years, rather than two expressions
- one for the first year of the evaluation, and one for the second year of evaluation.
AVISTA CORPORATION
RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION
CASE NO:
REQUESTER:
TYPE:
REQUEST NO.:
IDAHO
AVU-E-18-12
IPUC Staff
Production Request
Staff- 18
DATE PREPARED
WITNESS:
RESPONDER:
DEPARTMENT:
TELEPHONE:
t212912018
N/A
Lynn Roy
Nexant,Inc.
(303) 792-8668
REQUEST:
Regarding Exhibit 1, Equation 5-2: Please define the index variables t, y, and n.
RESPONSE:
The below response has been provided by Avista's 3'dparty EM&V evaluator, Nexant, Inc.:
The variable "t" indexes the month of the year, the variable o'y" indicates the year, and "n" indexes
the number of years under evaluation.
AVISTA CORPORATION
RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION:
CASE NO:
REQUESTER:
TYPE:
REQUEST NO.:
IDAHO
AVU-E-18-12
IPUC Staff
Production Request
Staff - l9
DATE PREPARED
WITNESS:
RESPONDER:
DEPARTMENT:
TELEPHONE:
1212912018
N/A
Lynn Roy
Nexant, Inc.
(303) 7e2-8668
REQUEST:
Regarding Exhibit l, Equation 5-2: How are the effects of weather incorporated into the
model for kWhity.
RESPONSE:
The below response has been provided by Avista's 3'd pa.ty EM&V evaluator, Nexant, Inc.
Avista's Home Energy Report program was implemented as a randomized control trial (RCT),
whereby Avista customers eligible for participation in the program were randomly assigned to
either treatment or control conditions. Customers assigned to the treatment condition received the
Home Energy Reports, while customers assigned to the control condition were not. Assuming the
assignment of customers to treatment and control status was random, the treatment effect on
electricity usage, as estimated by the LDV model, can be interpreted as net of any change in energy
usage due to weather. Thus, there is no need to model usage as a function of weather, in fact the
strength of the RCT approach lies in eliminating the need to model electricity usage as a function
of weather. The electricity savings for behavioral programs such as Avista's Home Energy
Reports are generally too small to be detectable using a weather modeling approach; an RCT is
required to estimate Home Energy Report electricity savings with an acceptable degree of
precision.
AVISTA CORPORATION
RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION: IDAHO
CASE NO: AVU-E-I8-12
REQUESTER: IPUC StaffTYPE: Production Request
REQUEST NO.: Staff - 20
DATE PREPARED:
WITNESS:
RESPONDER:
DEPARTMENT:
TELEPHONE:
t212912018
N/A
Lynn Roy
Nexant,Inc.
(303) 7e2-8668
REQUEST:
Exhibit 1, Table 5-21 explains that o is a coefficient on the billing month t, post-period year
indicator variable that is left-out due to collinearity. Please explain how this differs from the
intercept that would normally be included in a lagged dependent variable (LDV) model.
RESPONSE:
The below response has been provided by Avista's 3'dpa.ty EM&V evaluator, Nexant,Inc.:
Table 5-21 of Exhibit I should not have conditioned the description of Fo as omitted from
estimation due to collinearity. The LDV models were estimated using Stata statistical software; the
Stata regression output for the 2013 and 2016 cohorts is shown below in Figures 1 and 2,
respectively, and confirm that no variables were omitted due to collinearity.
Figure l: Regression Output - 2013
Progran: Opof,er Behaviorial erogran (Home Energy eeports)
rdaho Electric: 2O1l cohort
source I ss df i{5+------------
Mode
nesi dua
.rota'l
Humber of obs = 1514102F(112'1513989) -19566. oOProb>F - 0.OOO0R-squared - O.591aAdj R-squared - O.5914Root MSE - 15.445
rt.rt 522759970 a12 4667499.743611645201513989 238. 551615
avg-dai1y--e I
!ml#treat0ent641 1647 L
coef. std. Err t Dltl [95X conf. rnterval]
643 1644 1645 1646 1@7L
648 1649 1
650 1651 1652 1653 1654 1655 1656 1
657 1658 1659 1660 1
661 1662 1663 1664 1
665 1666 1667 L
668 1669 1670 I67L !67? 7
673 7674 a675 1676 1677 7-678 7.679 7680 1681 1682 1
681 1684 1685 1686 1687 1
6E8 1689 1690 169L 1692 1691 1
694 1695 1
-. 1108631
-.3987121-.47t44L-.5555372-.71t?973-1. 022308-1. 1516E3-. 9219373
-.8088417-. 5673261
-.529987-.4520659-.535E852-.7002598-. 6006658-.581E468-.7285415-.9645496-1.287911
-1. 132946
-.8791212-.6872546-.6690006-.5061554-. 5905715
-.6132217-. 5514078
-.613299{-.88176E6-L.2?)7)4-1.678238-t.777977-1.516285-1.250375-1.092369
-.8857696-.8788168-.8841299-.7951595-.7885804-.8281S7
-1.063952-1.55549
-1.701514-1.475895-1.150869
-1.017644-.7 54917 3
-.6962691-.7819115-.8618011-.9011054-.9773718-1.193719-1.728594
.1805795.1789957.t747977.178592
.1785013.1747292.1747043
.1786866.179051
.1816511.1825436
.1912421.19126E8
.a9147%
.1,97.5?42.1917268
.1917431.1920264
.1919708.1903714
.1900617.L90721
.1921989.1985099
.1984673
.1987615.1990692
.199405.1995615.L997379.1998101
.2000014.2001153
.2007469
-70??829
. 207 5096.?07444.2076154
.207637.2078061
.2080199.2081318
.2081951.2085077
.2086047
.2090598.?70577
.2151807.215229.?L54434.2157065
.2160028.2161152
.2161501.216709
0.4690.0260.008
0. ooz
0. ooo0.0000.0000.000
0. oo00.002
0. oo4
0.018
0. oo5
0.0000.0020.0020.0000.0000.000
0. o00
0.0000.000
0.0010.0110.0010.0020.0050.0020.000
0. o000.000
0.0000.000
0.000
0. o00
0.0000.0000. ooo
0. o000.0000.000
0. ooo
0. o000.000
0.0000.0000. oo00.0000.001
0. 0000.0000.0000. 000
0. ooo0.000
-.4847928-.7491574-.8228985-.9055713-1. 08114q
-L- f726rt
-1. 501937
-L. ?74157-t -1,5977 6-. 9233605
-.887766)-.8268938
.2210666-.0479069-.1219E16-. 2055011
-.3814159-.6720045-.80142E5
- .5737L78
- .4579079-. 2112921
-.L722478-.o77?38L-. 1610049
-.3249667
-.225285-. 206068E
- . 3527 3L7
-.5881844-.9116553-.759E209-. 5066067-.3134447
-.2S1SO54-. 1172828
-.2015845-.223656-. 163219
-.22?4725-. 4906149
- .8322547-1.286617-1. 385982
-1. 124065
-.8569177
-.6959014-. 479058-.472226-.4771,7t7-. 3883981-. 381287 5-.42M45?-.6560204-1.147041
-1.292866-1.0670t7-.7411193-.6050279-.3311705
- .27 44277
-.3596499
- .4390717
-.4777474-. 55?7 542
-.7696801
-1.303852
-o.72-2,2)-2.66-1.11-4.10-5.72-6.44-5.L7-4.52-?.L2-2. 90
-2.36
-2. EO
-3. 66
-3. 14-3.03-3.80-5.O2-6.7\-5.95-4,63-1.60-3. 48-2.55-2. 98-3.09-2.74-3. 08-4.42-6.13
-8. 40
-8.89-7 .58-6.2)-5.40
-4 .27-4.74-4. 26
-3.83-3,79-3.98
-.9107655-1.075553-.9760466
- .9576247
-1. 104 3 51-1.340915-1,664168
-1.506072-1. 2 51636
-1. 061065
-1. 046096-.895428-. s7q5625
-L.OOZ7A7-.9435766-1. 004126-1.277902-L,6L5?L4
-2.069859
-2.7-69973-1.908504-1.643832-]-.4a8837-t.29248L-1.285408-1.291088-L.20232L-1.195873-1.235909-L.47t88)-1.963937
-2.L10202-1.884753-1.560619-1.43026-1.176664-1.118111-7-704t73-t-28l574-t.324463-1.400989-L-6L7758-2.151336
-5.11
-7 .46
Cohort
-8.16-7. 08-5.50-4.83-3.51-3.24-3.61
-4. OO
-4.17-4.5?-5.52-7.98
88392d49L1514101. 583.7949t2
yn
61264)
644645
646@7
648649
650
651
652
653
654
655
656
657658659660
661
662
663664
665666
667
668
669670
67L67?
673
674
675676
677678
679680
681
682683
684
685686
687
688
689
690691
6926936v695
696
_cons
.5369213
-.15934614.9027547.7427648.9522956.8800656.6599048. 5384194.37289
-.4663182-L. )52947
-1.486806
. 0432504
-1.546391L.2420341.8815184.381t112.882114
.5625188
-1.154102-1.713329-3.758042-2.47)741,1.14088
1. 518388
-2. O345281. 195 52
2.5194042.1942862.70A]-721.57322L
-.8181855-2.246638-5- 204896
-1.391399-2. 4862
-7.679171
-2.9,18?781.6240642. 5546981.0534656.9188510.050786.1015917.348694
-.9083125-1.148916-7.07706.4442 551
-.4939993. 548011
5. 317 532.69324
-9765)74
4. 138812
.2063855
.2059647
. 2061 52
.2060816.20tr43
.2068428.206837.2068165.208044/.2084266
.21340/+4.2t31754
.2135441.2135067
.2136886.2L3676V
.2140516.2144389
.2135321.2111281
.7L3?9?2.2141948
.27.77272.2176614.747 V4
.21811?4
. ?183467
.2184507.218732E.2191735
. 2193 584.2192264
.2L94204.220?4!2
.22128.2212039
. ?233291.?237799.2233481.2234796
.2?377.67.2212784.224)776
.2242195.2742947
.2250917.2278537
.2278162.2280282.2282177.2?84593
.2285696.2288781
.2295103.1o8579
-2. 1615.5123.26LL.77
4. 0413.41
0. oo9o.439
0. ooo
0. ooo0.0000.000
0. ooo0. ooo
0. ooo0.025
0. oo00.0000.8390. ooo
o. ooo
o.000
o. ooo0.000
0. oo80. ooo
0.000o. ooo
0. ooo
o. ooo
0.000
o. oooo. ooo
o. oooo. ooo
0. oooo. ooo
0. ooo
o. ooo
o. ooo
o. ooo
0. oo00.000o.000o. ooo
0. o00
0. ooo
0. ooo
0. oo0
0. ooo
0. ooo
0. oooo,000
0. o000.0510.030
0. ooo
0. ooo
0. oo0
o. o(}0
o. ooo
.1124]-29
-.56301014. 498703
7.1388558.5476736.474666.254578.131065
1.96513-.874t,/7r-L.77].2t3-1.905015-. ]7 52887-1.964857.82]2L131.462779
3.9637772.462021.L440232
-L.572026-2.t31-17 5
-4.L77856-2. 600469.9142673
1. 091297
-2. 462021
.767 5685
2.091-249L.76557A?.271,6t.L43286
-1.248062-2.674694-5.636565-3.829021-2.923672-3.066889-1.3558991.186309
2.116685.6149894
6,47 9Z7L9.611009
5.662129a.909077
-L.3494U-1.595502-1.523611-.0026724-.94115543.1002194.8695412.244646.47670513.514028
.94L4297.244)3745.3068048.1466819.156916
7 .28547
7 . 0652978.9437724.780651
-.0578293-.9146822-1.068598.4617895-1.7?79261.6608564.3003174.8028453. 302607
.9810545
- .7 )6574
-1.295284-3.318228-L.7470L31-.7 67493L.945479
-1.6070351.623472
?.947 562. 6229953.7-307452.003156
-.3887093-1. E1E581
-4.7 7 3227
-2.953774-2.048728-2.49L454-2.4806582. 061819z,99Z7L1. 491941
7 ,35842910.49055
6. 5410542.748)L2
-.467L405-.7023308-.6105092.8911826
-.04664261.9957845.7655183.141833
a.37 637
4 - 7 4)6)6
7--o.71.)7.4t.
33.)2.41.21.-7.-6.-6.0.-7.5.18.
20.13.z.-5.-8.-17.-9.6.6.-9.5.11.10.12.7.-3.-10.-2t.-15.-11..-11.-13.7.11.4.
30.44.
27.10.-4.-t.-4,1.
60
7778
57l626
?o29
02
24
3197
20z4
81
7.7
4844
63
4Z
03
54
9E
1697l348
5303
32177t
25
63
19
1477
o7274t7L
8579
2L47
04
047)
95
1 ag-kwh-d .7098484
10. 404 58
. o005606
.1479198
1266.21
70.34
0. ooo
0.000
.7087496
10.11.466
.770947t
10.69449
Figure 2: Regression Output -2016 Cohort
Prograrn: Opo*er Eehaviorial lrogram (Home Energy Reports)
Idaho E'lectric: 2016 cohort
Source I ss dt i{5
uodel IResidual I-------------+-rotal I
91963200.5 4441410951.1308719 21355e7. Z8L34.707796
uumber of obs *
F( 44,108719) =1Prob > F
R-squaredrdj n-squared -Root MSE135394154308763 438.505111
1o87645912.69o. oooo
0.6940o.694011. 585
avg-dai1y--e I coef. std. Err t P>lrl [95x conf. tnterval]
ymftreatment675
676
677
678
6V9
680
681
682
683684
685
686
687
688
689
690
691
692
691
694695
696
1
1
1
L
1
1
1
1
1
7
1
1
1
1
1
1
1
1
1I
1
1
15.867.)7
-6.513. 56
15. 6012.6012.919. 504.2221.25
ym
676
677674
679
680681
682683
684685
686
687
688
689
690
691
692
693694
695
696
-.1813235-.2850918-.4)29992-.4128735-.1968818-.2038867-.4145327-.6387068
- .7 ?59327
-. 5273983
- .47445)7
-. 40137 5
-.5414156
- . s9777 66
-.5615527
-.448L677-.1836467
-.2125296-.5819736
-.8419786-, 8920167
-1.585628
- .40172)9
-3.148851
-3.036841-. 0068179.5107458
.506625
. 5011728
5. OS97l98.30778/.6.7171014.520L67?.97LOM1.385474
-1.232861.67457452.9705882.4076t92.47625
1.830959.8t92275.431689
.1879522
.1892642
1798387
18112 54
18284 l3
18476141863642
19041121911765
19221821929902
.191826.19492
1961101
19750931991881
2006694
u 021163
20119642046931.20578
1161333
t804577
.1811311821696
18305041838196
184 5018
185164 3
18659511868671
1867 567
18687 51
L87176?
188030 51887346
18971631903667
19104 591917528
.192V671940194
2 5 56208
o. l1lo.115o.0180.025
o.2910.278
0.029
0. o01
0. ooo
o. 0060.014
0. o380.005
o. o02
o. oo40.024o.160
o.291
o. o04o. ooo
o. o000. ooo
-.5318023-.uoo924-,79r]669
- .7749546
-.5621523-, 5722676
-.7854452
-1. 011907
-1.101025-.90111405-.8s2709-.7812685-.92t4532-. 9821468-. 9486653-.838s707-.5769531
-.6086719-.9406247
-1.2437.72-L.29534-2.20524
- .7 57 4L59
-1. 504261
-3.1942E2-.3656114.1504645
. 14 50027
.13286414.7340187.94153
6.3510634. 1 518972.601752
1. 016939
.7.7at552
. 0699088
-.o7463L5
-.0506924.1683848.1644943
-. 0435802
-.2655062
-.1508402-.1506562-. 0961984
-.0214815-.159178-.2L34064-.1,74440L
- . 0577 647.2096397
.1836126
-.7A11225-.4407455-. 4886937
-.9660155
-. 0500318-2.793446-7.679404.35a9357
.87rAZ7.8682471.86448165.46546
8.6740197.083138
4.8864373.3382561.754008
-.46294641. 0464133.f43702?.7A2064
2.8520812.208777
1.1995395.932699
-1. 01
-L- 57
-2.37-2.23
-1. 06-1.08-2- 19
-3. l5
-3.7 I
-2.V4-2.46-2.07-2.78
-3. 05-2.84-2.25-o.92
-1. 05-2.86-4.11-/r.33-5. OZ
-2.24-L7.)7-16.65-0. 047.78
2.7 52.70
77.3344.4615.9724-19
025
o00000970
o05
o06
o07
000o00
000
000
000
ooo
ooo
o00
ooo
000000
ooooo0
000
o.
0.0.0.
0.
0.0.o.0.
0.o.
o.0.
0.0.
o.
0.0.
0.0.
0.
-1.642775.3077359
1 ag-kwh-d_cons
.90122982.160679
?. 597 47 52.03)475
2. 10042
1. 4 53141.43897.574.93068
o011797
1126102
761.9716.29 .89891771.9007660.000
o. oo0
.90354192.420591
AVISTA CORPORATION
RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION:
CASE NO:
REQUESTER:
TYPE:
REQUEST NO.:
IDAHO
AVU-E-18-12
IPUC Staff
Production Request
Staff - 2l
DATE PREPARED:
WITNESS:
RESPONDER:
DEPARTMENT:
TELEPHONE:
1212912018
N/A
Lynn Roy
Nexant, Inc.
(303) 7e2-8668
REQUEST:
Regarding the analysis of the Company's Home Energy Reports Program (Section 5.8). Did
either the Company or Nexant perform a simple, two-sample comparison of kWh consumed
by each group? If so, please provide the results of the analysis with workpapers.
RESPONSE:
The below response has been provided by Avista's 3'dpa.ty EM&V evaluator, Nexant, Inc.:
Nexant did perform such comparisons for the 2013 and 2016 Cohorts, they are shown below in
Tables I and 2,respectively.
Table 1: Difference in Means t-test Values - 2013 Cohort
44.23 0.43
Table 2: Difference in Means t-test Values - 2016 Cohort
37.60 1.06
These two-sample comparisons were conducting using Stata statistical software. The annotated
Stata code that generated the statistics shown in Tables I and2 is shown below in Figures 3 and 4,
respectively. Figures 3 and 4 also show the commands that generated Tables 5-19 and 5-20 in
Nexant' s evaluation report.
44.29 -0.79
Control Average Daily
Usage: Pre period
Treatment Average Daily
Usage: Pre period P-value (95%)GriticalValue (t)
37.52 o.29
Treatment Average Daily
Usage: Pre period Critical Value (t)ControlAverage Daily
Usage: Pre period P-value (95%)
Figure 3: Workpaper for Difference in Means t-test Values - 20'|.3 Cohort
* 4. GeaeraEe ValidaEioo Checks
local state 'Idaflo"].ocal fue]. 'Eleccrici
// Bashington or Idaho
/ / ELectri-c or Gas
rA. Va]-idaEe preEreatlleaE equJ,valence for 2013 cohort
use ilitodified Dat.a/02-Calendarized_'st,aEe r_'fuel'_Brllinq_Data.dta", clear
keep if cohort : 2013 t 5yu. 4- lm(2013, 5) E lrrtr >- W(2012. 6)
ttesE avgtaily_usagtc, byttreatDe[t) // Geaeraces NexaDE Heeora*dua Tabl.e :.
collapse (mean) avg daily usage, by(:p.treatDenE)
reshape lIide avg_daily_usage, i (Im) j (treaEnenc)
browse // Generaces I'lextnt' Reporc Table 5-19
Figure 4: Workpaper for Difference in Means t-test Values - 2016 Cohort
*E. ValidaEe preEreatEeBt equivaleace for 2016 cohort
use ilHodified Daca,fo2_Cale*darized_Idaho Electri-c_Bi1tr-ing_DaEa.dEa', clear
keep if coborE - 2016 & lirB <= lm(2016, 3)
tcest avg_dai1y_usage, by(Ereat[ent) /l Generates NexanE Herorandum lab1e 2
col.3-apse (nean) avg daily usage, by(ip treaEmerlt)
reshape uide avg_daily_usage, i {lm} j (EreatmeEt}
browse .// GeaeraEes hlexa.lrt neport Tab1e 5-20
AVISTA CORPORATION
RESPONSE TO REQUEST FOR INFORMATION
JURISDICTION: IDAHO
CASE NO: AVU-E-I8-12
REQUESTER: IPUC StaffTYPE: Production Request
REQUEST NO.: Staff -22
DATE PREPARED
WITNESS:
RESPONDER:
DEPARTMENT:
TELEPHONE:
1212912018
N/A
Amber Gifford
Energy Efficiency
(s0e) 49s-2896
REQUEST:
Please provide the technical manual(s) used to estimate DSM energy efficiency savings for
2016 and20l7.
RESPONSE:
Please see StaflPR_22 - Attachment A for 2016 unit energy savings and Staff PP.-22 -
Attachment B for unit energy savings.