Bonneville Power Administration
|BPA Home EE Home Reports and Publications Program Evaluation Index Residential Evaluation|
EVALUATION OF BONNEVILLE'S 1991 LONG-TERM RESIDENTIAL WEATHERIZATION PROGRAM
PREPARED BY SYNERGIC RESOURCES CORPORATION
This report was prepared by Synergic Resources Corporation (SRC) for the Bonneville Power Administration (Bonneville). The report documents the activities and findings of an impact evaluation of Bonneville's Long-Term Residential Weatherization Program (RWP) in 1991.
The study had three objectives, namely to:
To achieve these objectives, Bonneville selected a sample of 29 utilities that participated in the program in 1991. It also set minimum sample sizes for each utility for each type of customer to be analyzed: 1991 RWP participants, 1989 RWP participants, and RWP non-participants. The latter two groups were used as comparison groups in a standard quasi-experimental research design. The sums of the minimum sample sizes for these three customer types across the 29 selected utilities was 665 1991 RWP participants, 999 1989 RWP participants, and 507 RWP non-participants.
SRC staff visited each of the 29 selected utilities. During each visit, random samples of specific types of customers were selected. Detailed program data were obtained for all sampled 1991 RWP participants. In addition, monthly (or bi-monthly, depending on the utility) billing data for the periods October 1989 - September 1990 (i.e., the "pre-treatment" year or period) and October, 1991 - September, 1992 (the post-treatment period or year) were obtained for all sampled 1991 RWP participants, 1989 RWP participants, and RWP non-participants.
Following data collection, the billing histories for each of the sampled households were weather-normalized using the Princeton Scorekeeping Method (PRISM). Valid PRISM estimates of Normalized Annual Consumption (NAC) were estimated for 788 1991 participants, 1,208 l989 participants, and 697 non-participants. Two alternative methods were used to identify households with poor PRISM results, to be dropped from the savings analysis. The first method, labeled the optimal case selection method, utilized the PRISM NACs for 768 1991 participants, 1,178 1989 participants, and 667 non-participants. The second method, labeled the traditional case selection method, utilized the PRISM NACs for 585 1991 participants, 851 l989 participants, and 431 non-participants. Estimates of the difference between the NACs in the pre-and post-treatment years (called DNAC) was then computed for each household.
The estimates of average annual net savings for the entire region were (corresponding standard errors are in parentheses):
The estimate of 2,445 kWh estimated using the optimal case selection method and non-participants as the comparison group was selected as the most appropriate estimate of net savings for the l99l program year. The standard error for this estimate indicates that there is a 90 percent probability that the actual average annual net savings for the l99l program year is between 2,064 kWh and 2,826 kWh, or plus-or-minus 381 kWh (16 percent).
Using this definition of the comparison group, the population estimates of average net first-year savings for the l99l program were 2,646 kWh for participants in Climate Zone 1 (roughly Western Oregon and Washington); 692 kWh in Climate Zone 2 (roughly Eastern Oregon and Washington and Western Idaho); and 1,343 in Climate Zone 3 (roughly Eastern Idaho and Western Montana). The estimate for Climate Zone 1 is not significantly different, at the 90 percent confidence level, from the estimates for the other two Climate Zones, due to very large standard errors for the estimates for Climate Zone 2 and 3.
The population estimate of average program costs is $2,751 per participant (1991 $). Bonneville incurred $1,537 of these costs; utilities, $19; and participants, $1,196. Average costs in the Climate Zones 1, 2, and 3 were $2,845, $2,069, and $2,097, respectively.
The estimated levelized costs of savings produced by 1991 program activities (in 1993 mills/kWh saved) were:
Consistent with the selection of the net savings estimate derived using non-participants as the comparison group and cases selected using the optimal case section method, the estimate of 64 mills/kWh was selected as the most appropriate estimate of the levelized cost of savings for the 1991 program. Bonneville incurred 39 mills/kWh of the levelized cost, and participants 25 mills/kWh. The estimated levelized cost varied tremendously across regions -- 61 mills/kWh in Climate Zone 1, 175 mills/kWh in Climate Zone 2, and 92 mills/kWh in Climate Zone 3. In addition, the estimated levelized cost varied substantially depending on the types of measures installed -- 50 mills/kWh for participants installing no glazing measures, 70 mills/kWh for participants installing both glazing and non-glazing measures, and 133 mills/kWh for participants installing only glazing measures.
The findings of this study reverse a trend of previous evaluations, which found declining savings over time and increasing levelized costs. The estimated net savings for the 1991 program year are similar to the estimates for the 1985, 1986, and 1988 program years.
Need help finding something? Send us a message, or give us a call at 800-622-4519.