1. Maximum Economically Achievable Electricity Savings from Unconstrained Investment in Energy Efficiency 2012 - 2031 REVISED ANALYSIS January 19, 2011

2. Contents Overview Approach Results Methodology Revisions to Analysis Presented on December 20, 2010 Major Drivers of Long-Term Changes in Opportunities Issues 2

3. Overview 3

4. DRP Context VELCO unconstrained savings forecast constitutes a “boundary case” in the Public Service Board’s DRP process It is the first of four efficiency resource plans to be developed, analyzed, and compared with the maximum achievable analysis 4

5. Flat budget to acquire all economically achievable potential across 20 years Ramp up across 5 years to achieve 3% savings, relative to annual energy use Status quo budgets adjusted for inflation Additional Scenarios 5

6. PSB DRP process created after VELCO engaged VEIC to conduct the analysis for unconstrained potential Necessitated expansion of scope of work to include updating costs and savings estimates over time for key efficiency technologies, applicable across DRP scenario—especially lighting History 6

7. Approach 7

8. Operating definition: The maximum amount of cost-effective electricity savings that can be achieved, as quickly as possible, by deploying the most aggressive program strategies available in all major residential and business markets. Maximum Economically Achievable Efficiency Savings 8

9. IF avoided costs are constant or rising in real terms, all else being equal, AND the sooner and faster maximum achievable savings are achieved, THEN the greater the resulting net benefits.* Maximum achievable penetration rates are the result of informed professional judgment about what’s possible with the most aggressive intervention strategies in each market. *This is automatically true, due to discounting Guiding Principles 9

10. General Approach Same basic approach as analysis used in 2009 Forecast 20: Integration with VELCO 2010 Forecast Updated efficiency technology characterizations Two-stage analysis of measures / measure bundles Conceptual program design Market penetration rates 10

11. Integration with VELCO 2010 Forecast Preliminary 2010 sales and peak demand forecast (Itron) VEIC worked closely with Itron to coordinate energy sales baseline energy intensities with savings analysis baselines This time Itron prepared a forecast adjusted for effects of continued EEU investments of $40 million annually throughout the period Along with an unadjusted forecast, without the effects of any continuation of Vermont DSM investment in efficiency 11

12. Development of 90 / 10 Scenario Indicates expected summer peak demand savings under extreme weather conditions Assumes that the 90 / 10 DSM savings have the same percentage increase over the 50 / 50 DSM savings as the 90 / 10 forecast has over the 50 / 50 forecast 12

13. Updated Characterizations of Efficiency Technology Re-assessed future changes in baseline Characterized additional measures (including solar residential hot water) 13

14. Two-stage Analysis of Measures / Measure Bundles Assess the cost-effectiveness of efficiency measures or measure bundles Predict maximum penetration rates with most aggressive conceptual program designs specified for each market Same approach as with original Forecast 20 and to be conducted in the analysis of other efficiency resource plans in the DRP process. 14

15. Avoided Costs Used in this Analysis From 2009 Regional Analysis of Avoided Costs (Synapse Energy Economics) Expressed at same level as VELCO forecast 15

16. Conceptual Program Design Program design incorporates incentives covering 100% of installed or incremental costs, depending on market. Exceptions where experience suggests customers will adopt measures with less than 100% cost reimbursement 16 Example: Incentives for large C&I customer retrofits Program will “buy down” customer investment costs to 1 year, enabling customers to fund their contribution to efficiency investments out of operating budgets. Example: Incentives for large C&I customer retrofits Program will “buy down” customer investment costs to 1 year, enabling customers to fund their contribution to efficiency investments out of operating budgets.

17. Market Penetration Rates Retrofit Markets Annual pace and acceleration in market penetration is discretionary and constrained only by infrastructure readiness. Lost Opportunity Markets Replacements: Pace depends on natural turnover of existing equipment stock New Construction: Pace depends on number of homes and business facilities 17

18. Residential Existing Homes Retrofit Piggyback on fossil heating retrofit program designed to retrofit the entire Vermont housing stock across 20 years Direct installation of all cost-effective electric efficiency measures Whole-house re-lighting Early replacement of appliances, AC Electric water heating conversion to solar or natural gas Products Target all existing homes not reached through the retrofit programs Lighting 18

19. Commercial & Industrial Existing Facilities Retrofit Plan on achieving 80 – 90% participation among targeted eligible customers in as short a time as possible Ramp up from current levels within 3 years to a maximum sustainable rate to reach that target, backing down gradually throughout the rest of the 20-year horizon Lost Opportunity Terminal market penetration rates approach 80 – 90% for all but infra-marginal (not most efficient) alternatives The EEU could ramp up to these maximum rates in 3 – 5 years 19

20. Improvements in Residential Efficiency Analysis ( over 2009 Forecast 20) Lighting -In F20, program savings for standard spiral CFLs ended in 2014 (after first phase of EISA regulations). Based on recent evaluations and negotiations with DPS, CFL (standard and specialty) assumed to continue (with the EISA compliant baseline shift) until 2019. -SSL lighting re-characterized -Recessed down light added Appliances - New CEE tiers added / adjusted 20

21. Improvements in Residential Efficiency Analysis ( over 2009 Forecast 20) Other additions -Air sealing and insulation measures for electric heat -Cooling savings from shell measures -Solar hot water as an efficiency measure Separated low-income from existing homes Matched RNC lighting and product assumptions to products program 21

22. Improvements in C&I Efficiency Analysis (over 2009 Forecast 20) Lighting -New measures characterized for SSL for: down-lighting, screw-in, linear fluorescent replacement, refrigerated case light fixtures -Revised assumptions for LED costs and efficacy changes over time, based on revised Multi-Year Program Plan for SSL R&D (DOE EERE, March 2010) -New measures characterized for reduced-wattage T8 lamp and lamp / ballast 22 Example Existing T12 lighting is replaced due to EISA phase-out of T12 lamps, and lower-wattage / lumen replacements are suitable, rather than straight T8 lamp / ballast replacements. Example Existing T12 lighting is replaced due to EISA phase-out of T12 lamps, and lower-wattage / lumen replacements are suitable, rather than straight T8 lamp / ballast replacements.

23. Improvements in C&I Efficiency Analysis (over 2009 Forecast 20) HVAC measure characterizations updated, future adjustments made for changing federal standards Due to advances in efficient technologies, new retrofit opportunities assumed to become available for more recently installed equipment,10 to 15 years out Assumed that measure costs would increase gradually over time (in real dollars), based on an expectation of a gradual increase in the cost of savings due to increasing baselines. 23

24. Additional Savings— Other Sources not Considered Other sources of cost-effective efficiency savings that were not considered in this analysis—or in Itron’s forecast—that would increase maximum achievable efficiency savings: Accelerating the adoption of more efficient standards for buildings and equipment New, undiscovered, unidentified technologies emerging beyond 2020 Changes in the load profiles of equipment and appliances with the introduction of real-time pricing and usage displays 24

25. Results 25

26. The Big Picture: 2012 - 2031 2,317 GWh / year 390 MW / year Summer Peak 390 MW / year Winter Peak 26 Societal TestElectric System Test PV of Net Benefits ($M)$2,113$1,318 Benefit-Cost Ratio2.421.86

27. Statewide Incremental Savings 27

28. Statewide Cumulative Savings 28

29. Residential Incremental MWh Saved, by End Use 29

30. Residential Incremental MWh Saved, by End Use 30

31. Residential Cumulative MWh Saved, by End Use 31

32. Residential Cumulative MWh Saved, by End Use 32

33. Residential Incremental Summer Peak MW Saved, by End Use 33

34. Residential Incremental Summer Peak MW Saved, by End Use 34

35. Residential Cumulative Summer Peak MW Saved, by End Use 35

36. Residential Cumulative Summer Peak MW Saved, by End Use 36

37. C&I Incremental MWh Saved, by End Use 37

38. C&I Incremental MWh Saved, by End Use 38

39. C&I Cumulative MWh Saved, by End Use 39

40. C&I Cumulative MWh Saved, by End Use 40

41. C&I Incremental Summer Peak MW Saved, by End Use 41

42. C&I Incremental Summer Peak MW Saved, by End Use 42

43. C&I Cumulative Summer Peak MW Saved, by End Use 43

44. C&I Cumulative Summer Peak MW Saved, by End Use 44

45. Zonal Forecasts: Cumulative Annual GWh Savings, Summer Peak MW Savings, Winter Peak MW Savings 45

46. 46 Vermont Potential Load Zones

47. 47 Zonal Savings AB

48. 48 Zonal Savings C D

49. 49 Zonal Savings EF

50. 50 Zonal Savings G H

51. 51 Zonal Savings IJ

52. 52 Zonal Savings KL

53. 53 Zonal Savings MN

54. 54 Zonal Savings OP

55. Benefit–Cost Analysis 55

56. Program PV of Societal Benefits ($M) PV of Electric Energy System Benefits ($M) Residential Residential New Construction$174$24 Retail Products$1,108$851 Existing Homes$218$126 Low Income (SF and MF)$18$10 Sub-Total Residential$1,518$1,012 Commercial & Industrial C&I New Construction$326$280 C&I Efficient Equipment$873$796 C&I Retrofit$881$765 Sub-Total C&I$2,081$1,841 EVT Core Supporting Services-- Portfolio of Programs$3,599$2,852 Benefits 2012 – 2031 56

57. Program PV of Societal Costs ($M) PV of Electric Energy System Costs ($M) Residential Residential New Construction$64 Retail Products$208$291 Existing Homes$76$118 Low Income (SF and MF)$10$17 Sub-Total Residential$359$491 Commercial & Industrial C&I New Construction$145$118 C&I Efficient Equipment$322$232 C&I Retrofit$358$391 Sub-Total C&I$825$741 EVT Core Supporting Services$303 Portfolio of Programs$1,487$1,535 Costs 2012 – 2031 57

58. Societal Net Benefits & Benefit-Cost Ratios 58 Program PV of Net Benefits ($M) Benefit-Cost Ratio Residential Residential New Construction$1102.71 Retail Products$9005.33 Existing Homes$1422.86 Low Income (SF and MF)$71.73 Sub-Total Residential$1,1604.23 Commercial & Industrial C&I New Construction$1812.25 C&I Efficient Equipment$5512.71 C&I Retrofit$5232.46 Sub-Total C&I$1,2562.52 EVT Core Supporting Services*($303)- Portfolio of Programs$2,1132.42 * Because societal costs are negative, BCR is meaningless

59. Electric System Net Benefits & Benefit-Cost Ratios 59 Program PV of Net Benefits ($M) Benefit-Cost Ratio Residential Residential New Construction($40)0.38 Retail Products$5602.92 Existing Homes$91.07 Low Income (SF and MF)($7)0.58 Sub-Total Residential$5212.06 Commercial & Industrial C&I New Construction$1632.38 C&I Efficient Equipment$5633.42 C&I Retrofit$3741.96 Sub-Total C&I$1,0992.48 EVT Core Supporting Services*($303)- Portfolio of Programs$1,3181.86 * Because societal costs are negative, BCR is meaningless

60. Budget Projections 2012 – 2021 60

61. Budget Projections 2022 – 2031 61

62. Methodology 62

63. Residential Market Analysis 63 “Bottom-up” Approach Estimate per measure savings each year Market categories: Efficient Products (EP), Existing Homes (EH), Low Income (LI), New Construction (RNC) EH & LI - Market penetrations = Program homes X percentage with unit X percentage ready for retrofit EP – Market penetrations = Estimated stock turnover and new stock for EP X percentage influenced by program, minus EH, LI, RNC penetrations where appropriate RNC – Market penetrations= Program homes X percent with unit X percent influenced by program

64. “Bottom-up” Annual Savings Calculation Retrofit Measures: Per-measure savings XHomes in program XPercentage homes with unit XPercentage ready for replacement (cost effective) =Annual measure maximum achievable potential 64 Efficient Products: Per-measure savings XStock mortality (existing stock / measure life) +New stock (existing stock in Year X – existing stock in Year X-1) X Percentage influenced by program =Annual measure maximum achievable potential Residential New Construction: Per-measure savings X Homes in the program X Percent of home with unit X Percent influenced by program =Annual measure maximum achievable potential

65. Retrofit Example Heat Pump DHW Early Replacement (2012) 65 Annual Max Achievable Savings Potential = 176 MWh ParameterDescriptionValueResult Per Measure Savings (kWh) 1914 #homesNumber of homes participating in HPwES program 15002,871,000 %electric DHWPercentage of homes with Electric DHW 0.31890,010 %ready% of electric DHW with cost effective potential for retrofit 0.34302,603 %HPDHW% of retrofits assumed to HP (as opposed to Natural Gas or Solar DHW) 0.58175,509

66. Efficient Products Example Clothes Washer (CEE Tier 3) (2013) 66 Annual Max Achievable Savings Potential = 686 MWh ParameterDescriptionValueResult Stock Mortality# existing units needing replacing each year: (# Homes x % Homes with CW) / Measure Life = (294,500 * 0.92) / 14 =19,458 19,458 New StockNew units purchased each year (due to new homes and more homes having unit): = Units in 2013 – units in 2014 =(294500 * 0.92) – (293000 * 0.92) = 1,412 20,870 % influenced by program 0.245,009 In program penetration Divide out spillover (influenced but not in program) and subtract EH, LI and RNC penetrations = (5009/1.2) – 216 – 202 – 69 3,686 Per Measure Savings (kWH) 186 Total Savings Potential Penetration x Per Measure Savings=3686 * 186685,596

67. Residential Market Technologies 39Efficient Product measures 43 Existing Homes measures 14 Low Income measures 30 Residential New Construction measures 67

68. C&I Market Analysis 68 “Top-down” Approach Forecast sales and estimate potential % savings Determine cost by energy saved Break up analysis by market categories -New Construction -Existing facilities Retrofit opportunities Equipment opportunities Track eligible stock to avoid double-counting Interior lighting savings adjusted for “cooling bonus” or “heating penalty”

69. “ Top-Down” Savings Calculation Building end-use MWh sales per year XApplicability factor XFeasibility factor XTurnover or retrofit-not-complete factor XSavings factor XPenetration factor =Annual measure maximum achievable potential 69

70. Example: High-Performance T8 vs. Standard T8 70 Annual Maximum Achievable Savings Potential = 90.7 MWh

71. C&I Market Technologies 87efficiency measures 10 different commercial building types and street lighting 2,253 combinations analyzed (technology / building type / market) 71

72. C&I Existing Market Characteristics 72

73. C&I New Construction Market Characteristics 73

74. Revisions to Analysis Presented on December 20, 2010 74

75. F20 Unconstrained Analysis REVISIONS VEIC accidentally retained the 2007 avoided costs used in the 2009 Forecast 20 analysis in the unconstrained savings forecast presented on December 20. The updated avoided costs approved by the Board in December 2009—and used by Efficiency Vermont in 2010 for program implementation—are substantially lower than the 2007 values mistakenly used in the original unconstrained analysis. VEIC has revised the unconstrained analysis using the 2009 avoided costs. 75

76. F20 Unconstrained Analysis REVISIONS Reductions are especially significant in avoided generation capacity costs. VEIC corrected this error by re-screening all efficiency measures. Most measures are still cost-effective at lower avoided costs. Others either are no longer cost-effective throughout the analysis period or they become cost-effective in later years. VEIC also applied the revised real discount rate of 5.6%, recommended by the DPS (changed from 5.7%). 76

77. F20 Unconstrained Analysis REVISIONS VEIC adjusted measure penetrations accordingly by removing measures no longer cost-effective, or postponing and lowering market penetration rates for measures that become cost-effective in later years. VEIC re-screened all programs and entire portfolio for cost-effectiveness, and calculated new electricity savings. Effects of lower avoided costs on measure cost- effectiveness and electricity savings are significant, particularly in the residential sector. 77

78. Annual Avoided Costs— Energy and Capacity 78 Difference between 2009 and 2007

79. Cumulative Effects: Change in Energy and Peak Demand Savings 79

80. Effects of Change in Societal Net Benefits 80

81. Effects of Change in Electric System Net Benefits 81

82. New avoided costs rendered some residential measures cost-ineffective in early years. Zero penetrations for the following measures and years: 2012 – 2019: Efficient Products heat pump DHW 2012 – 2015: Efficient Products, RNC, Existing Homes and Low-Income heat pump dryers 2012 – 2018: Existing Homes solar DHW 2012 – 2024: Existing Homes heat pump DHW Residential Sector Changes 82

83. Residential Sector Changes During the re-screening process we revised characterizations for two measures: Products Program heat pump DHW, changed load shape of from “DHW insulation” to “DHW fuel switch” Electric heat air sealing and insulation measures— previously broken down between high, medium, and low— were consolidated to an average air sealing and insulation measure 83

84. C&I Sector Changes Ground Source Heat Pumps -In the 5 building types where these were previously found to be cost-effective, they now fail in the first few years, but pass in Year 5. Penetrations were adjusted to start in Year 5. "High Efficiency residential-size refrigerator" now fails in warehouses. "Booster water heat for dishwashing" in Lodging now fails until Year 10, so it is excluded. The above measures have relatively small savings. Adjustments made only a small difference in the overall C&I maximum achievable forecasted savings. 84

85. C&I Sector Changes Several measures now do not pass cost- effectiveness screening for the first few years in a specific building type. But they almost pass, and thus remain in the measure mix since the measure still passes when aggregated across building types. Total Resource Costs change is due to the updated real discount rate. C&I results also reflect a change in the “Vermont Other Sales” forecast. 85

86. Summary of Changes in Residential 86

87. Change in Residential Incremental MWh Saved, by End Use 87

88. Change in Residential Incremental Summer Peak MW Saved, by End Use 88

89. Change in Residential Incremental Winter Peak MW Saved, by End Use 89

90. Summary of C&I Changes 90

91. Change in C&I Incremental Energy (MWh) Saved, by End Use 91

92. Change in C&I Incremental Summer Peak MW Saved, by End Use 92

93. Change in C&I Incremental Winter Peak MW Saved by End Use 93

94. Major Drivers of Long-term Changes in Opportunities 94

95. Codes & Standards: Short Term DOE reviewing ASHRAE 90.1-2004 ASHRAE 90.1-2007 available now Vermont standards (9 V.S.A §2791) include: -Metal halide lamp fixtures -State regulated incandescent reflector lamps -Residential furnaces and residential boilers 95

96. Codes & Standards: Long Term DOE seeking 1 – 2% annual efficiency improvement (20 – 40% change by 2028) Potential adoption of regional standards Changes in impact from: -New and / or expanded standards and service -Better enforcement of codes -Expanded codes 96

97. Savings Due to Codes & Standards 97

98. Decline of Savings from Compact Fluorescent Lamps 2012 – 2020 First tier of new federal lighting standard Incentives for specialty CFLs only (dimmers, three-way, etc.) After 2020 Second tier of federal lighting standards CFLs are baseline 98

99. Rise of Solid-state Lighting LEDs and Organic LEDs “New Frontier” -Highly efficient (eclipsing current technology) -Extremely long life Barriers exist, but are constantly shifting -Cost (high but falling) -Compatibility (new screw-in lamps coming to market) 99

100. Rebatable CFLs vs. SSL 100

101. Savings from CFLs vs. SSL 101

102. Issues 102

103. Baseline Household Lighting Energy Use, After 2020 Inconsistency persists between post-2020 baseline household lighting energy use intensities: Steep drop in lighting efficiency savings forecast because CFLs become baseline Itron EUIs continue gradual decline implying continued prevalence of incandescent 103

104. Baseline Household Lighting Energy Use, After 2020 Two options: VEIC develops an adjustment to be applied to post-2020 sales forecast when applying unconstrained residential maximum achievable savings Itron departs from EIA projections and builds discontinuity into its post-2020 sales peak demand forecasts 104

105. Avoided Costs Matter The inadvertent use of the 2007 avoided costs forced an early analysis of the sensitivity of efficiency measure cost-effectiveness to changes in avoided costs Lower avoided energy, and especially capacity costs: Rendered some measures cost-ineffective throughout the forecast period Delayed the point at which some measures become cost- effective Ultimately lowered the amount of cost-effectively achievable savings over the next 20 years, especially in the residential sector. 105