ESA Multifamily Segment Study Status Update
Prepared for:
Mary O’Drain, PG&E
and
The ESA Multifamily Segment Study Team
Prepared by:
Anne West
Teri Duncan
Eric Rambo
The Cadmus Group, Inc.
Dulane Moran
Jane Peters, PhD
Research Into Action
April 20, 2013
This page left blank.
TABLE OF CONTENTS
Introduction .................................................................................................................................................. 1
Task 1: Project Initiation Meeting and Research Plan .............................................................................. 3
Research Plan ........................................................................................................................................ 3
Task 1 Next Steps .................................................................................................................................. 3
Task 2: Gather California Multifamily Housing Data Relevant for Low-Income Customer Programs ...... 4
Analyzing Census Data .......................................................................................................................... 4
Organizing IOU Program and Customer Data ....................................................................................... 5
Building Owner and Operator Survey ................................................................................................... 6
Sampling Plan ........................................................................................................................................ 9
Examine Comments to ESAP Proceedings .......................................................................................... 13
IOU Staff Interviews ............................................................................................................................ 15
Task 2 Next Steps ................................................................................................................................ 15
Task 3: Catalog Existing Multifamily Energy-Efficiency Programs .......................................................... 16
Task 3 Next Steps ................................................................................................................................ 17
Task 4: Review and Evaluate Multifamily Programs and Research Relevant for Low-Income Customers
................................................................................................................................................................ 17
Task 4 Next Steps ................................................................................................................................ 18
Task 5: Identify and Assess Alternative Program Designs and Delivery Strategies ................................ 18
Current Program Components ............................................................................................................ 18
Task 5 Next Steps ................................................................................................................................ 18
Task 6: Identify Financing and Funding Options ..................................................................................... 19
Develop List of Relevant Financing Options........................................................................................ 19
Review Results from Other Study Tasks and Other Studies ............................................................... 19
Task 6 Next Steps ................................................................................................................................ 20
Task 9: Conduct Public Workshop to Discuss Research and Recommendations ................................... 20
Task 9 Next Steps ................................................................................................................................ 20
Appendix A: Supporting Material for Census Tract Analysis and GIS Mapping ...................................... 21
This page left blank.
INTRODUCTION
This status update memo summarizes Phase 1 research activities through March 2013, conducted for
the ESA Multifamily Low-Income Segment Study. The following sections discuss research conducted by
Cadmus and Research Into Action (together, the Cadmus team), and specify progress made to-date. We
also list the next steps to complete Phase 1 deliverables. The Phase 1 deliverables were described in the
ESA Program Multifamily Segment Study Research Plan filed February 25, 2013. Separate electronic files
of supporting materials are included as attachments to this report.
The primary activities in this Study are designed to meet the CPUC ALJ Decision’s (D.12-08-044) research
objectives. These activities and areas of focus are:






Gather California Multifamily Housing Data Relevant for Low income Customer Programs
Catalog Existing Multifamily Energy Efficiency Programs Relevant for Low Income Customers
Review and Evaluate Multifamily Programs and Research Relevant for Low Income Customers
Identify and Assess Alternative Program Designs and Delivery Strategies
Identify Financing and Funding Options
Conduct Public Workshops
Phase 1 Highlights
The Cadmus team is progressing toward meeting the Phase 1 deliverables, completing the following
tasks. The team:












Submitted the draft Research Plan
Conducted the first public workshop and received comments from stakeholders
Interviewed the ESA and MF program managers at all four IOUs
Cataloged and created a summary table of public comments to the ESAP proceedings
Submitted data requests to the four IOUs and received most of the data
Analyzed American Community Survey (ACS) summary data, PUMS and PUMA datasets
Using ArcGIS, mapped low-income, multifamily households by census tract
Estimated the number of low-income multifamily households by county and IOU service
territory
Conducted a preliminary literature review to identify which programs to include in the catalog of
nationwide programs serving the multifamily sector and particularly low-income residents
Began discussions about cost-effectiveness criteria used in the ESA Program
Compiling datasets to develop the sample frame for the building owner and manager surveys
Completed a preliminary draft of the building owner and manager survey
Summary of Next Steps
Chapters for each of the key activities listed above will be delivered as they are completed, anticipating
delivery of chapters for the Phase 1 report by May 1. Cadmus will then compile the individual chapters
into a Phase 1 draft report and circulate to the Study Team for comments. The draft final report for
Phase 1 activities will be delivered by June 14, 2013.
ESA Multifamily Segment Study Status Update / April 20, 2013
1
Table 1 summarizes the next steps needed to complete the tasks that correspond to each of the
research activities for Phase 1. Following the table, further detail about the progress made for each
Phase 1activity to-date is provided.
Table 1. Summary of Next Steps to Complete Research Tasks for Phase 1
Task
Task 1: Project
Initiation Meeting and
Research Plan
Next Steps


Task 2: Gather
California Multifamily
Housing Data
Task 3: Catalog Existing
Multifamily EnergyEfficiency Programs
Task 4: Review and
Evaluate Multifamily
Programs and Research
Relevant for LowIncome Customers
Task 5: Identify and
Assess Alternative
Program Designs and
Delivery Strategies








Continue reviewing documents related to ESA and multifamily proceedings.
Prepare summary memo describing the approach that will be implemented in
Phase 2.

Organize collected data to incorporate it with data collected by Research into
Action through Task 3.
Review and highlight key information from the program staff interviews
conducted at all four IOUs.
Continue to compile data regarding California’s programs by reviewing
recommendations from ESA proceedings, findings from additional studies, and
research results from other study tasks, specifically Task 2 and Task 3.
Compile preliminary program design strategies, options, and suggestions for
discussion with the study team.
Continue discussions about cost-effectiveness with the IOUs
Work with the task 2 staff leads to develop questions about financial decision
for the building owner/operator surveys.



Task 6: Identify
Financing and Funding
Options
Task 9: Conduct Public
Workshop to Discuss
Research and
Recommendations
Based on input from public workshop and comments posted to the
Energydataweb.com site, Cadmus will submit revised draft research plan to the
Study Team.
Summarize current participation by census tract based on analysis of IOU
participant data.
Continue to track incoming customer and participant data.
Compare program participation levels to the low-income multifamily household
estimates by census tracts to identify underserved areas with high levels of
estimated potential participants.
Continue to catalog programs.
Prepare progress status update for the catalog of multifamily energy-efficiency
programs.
Prepare Bibliography.
Prepare catalog draft for submittal to Study Team April 15.



Conduct a second public workshop to review the findings from Phase 1 of the
study, to be held after the Phase 1 draft report is completed and submitted to
the Study Team in June 2013.
ESA Multifamily Segment Study Status Update / April 20, 2013
2
Task 1: Project Initiation Meeting and Research Plan
The Cadmus team joined the study team—Mary O’Drain, Pacific Gas and Electric (PG&E); Carol Edwards,
Southern California Edison (SCE); Brenda Gettig, San Diego Gas & Electric (SDG&E) and Southern
California Gas (SoCalGas); and Tory Francisco, California Public Utilities Commission (CPUC)—for the
project initiation meeting in San Francisco on January 18, 2013. Representatives from the California
independently owned utilities (IOUs) and the CPUC Energy Division also attended the meeting.
During the meeting, Cadmus presented an overview of the scope, research approach, and activities for
the study. The presentation included key project milestones and their associated timelines. Meeting
participants also identified and discussed critical issues for consideration during the study.
Following the meeting, Cadmus prepared and delivered a memo of the meeting’s minutes, documenting
the key points to be considered for the research plan, action items and next steps, and items open for
further discussion.
RESEARCH PLAN
The Cadmus team developed and submitted an outline for the research plan on February 6, 2013, for
approval by the study team. The first draft was submitted on February 8. In response to comments from
the study team, we submitted a revised draft plan on February 21.
Based on direction from the study team, Cadmus created a new and more accessible version of the plan
for the public workshop. In this version, we streamlined some of the tables, explained the phases of the
activities, and moved the more lengthy details to an appendix.
The revised draft for the public workshop was submitted on February 25. The Energy Division posted
this draft to energydataweb.com for review by public stakeholders in advance of the workshop.
Comments on Research Plan from Public Stakeholders
During the first public workshop, held in San Francisco on March 5, 2013, Cadmus received feedback on
the draft research plan from the stakeholders in attendance. Additional public comments were posted
to energydataweb.com.
TASK 1 NEXT STEPS

Based on input from public workshop and comments posted to energydataweb.com,
Cadmus will submit a revised draft research plan to the study team. We prepared an
appendix with comments from each provider and will address comments in additional
appendices.
ESA Multifamily Segment Study Status Update / April 20, 2013
3
Task 2: Gather California Multifamily Housing Data Relevant for Low-Income
Customer Programs
ANALYZING CENSUS DATA
In January and February, Cadmus gathered and reviewed publicly available data to identify
characteristics of low-income households in multifamily dwellings. In the process, we found that one
source—American Community Survey (ACS) summary data, which is available as pre-defined tables at
the census tract level—could not be used directly to determine the intersection between low-income
inhabitants who also live in multifamily buildings. Tabulating this intersection is possible only by
estimating census tract households using proportions from the more aggregate Public Use Microdata
Sample (PUMS).
Unlike the census tract data, PUMS data can be directly manipulated as individual household records;
however, these data are identifiable only for a larger geographic area, the Public Use Microdata Area
(PUMA). This restriction is imposed to ensure the confidentiality of respondents. Whereas a census tract
may comprise as few as 2,000 residents, a PUMA comprises approximately 100,000 residents.
For each PUMA, we calculated the percentage of households meeting the low-income criterion of 200%
of federal poverty guideline and residing in buildings with five or more units. Our approach was as
follows:
The PUMS datasets for each state include housing- and person-level records. For example, a five-person
household has one record in the housing dataset and five records in the person dataset; a serial number
ties the five-person records to the one-household record. Cadmus defined a household as meeting the
low-income criteria if one person in that household earned less than 200% of the income defined by the
federal poverty guideline.
We counted the number of households meeting both multifamily and low-income criteria for each
PUMA. This value was divided by the total number of multifamily households within that PUMA to
obtain the percentage of multifamily households below 200% of the federal poverty guideline. This
percentage represents the conditional probability that a household meets both the low-income and the
multifamily criteria. We excluded records for group or institutional quarters.
The ACS five-year data summaries provide counts of housing units by number of units in the building.
There are also summaries of the number of people at different income-to-poverty index levels; but,
because these data are available only in pre-populated summaries, it is not possible to map these counts
of people to counts of households.
Therefore, Cadmus determined that the best method for estimating the number of low-income,
multifamily households was to use the conditional probability computed at the PUMA level of the
intersection of low-income and multifamily household status. We applied this proportion to the count of
multifamily households within each census tract to obtain an estimate of the number of low-income and
multifamily households within that tract.
ESA Multifamily Segment Study Status Update / April 20, 2013
4
For 2011 U.S. Census data, census tracts are not nested within PUMAs. Cadmus determined1 the
percentage of each census tract overlapped by a PUMA and used this proportion to adjust population
counts2. These proportions, along with the PUMA-level conditional probabilities, were multiplied by the
census tract level counts of multifamily households. These results in turn were summed by census tract,
in order to obtain the number of low-income, multifamily households within each census tract.
Appendix A provides a map of the results described above, as well as summaries of the census tract
estimates by county and IOU service territory.
ORGANIZING IOU PROGRAM AND CUSTOMER DATA
Data Tracking
Cadmus set up a secured FTP site for IOU program administrators to transfer data that we will use to
analyze past and current trends in program participation. These data are stored in an encrypted file
network and accessible only to the data-analysis team. Each transferred file is catalogued in a
spreadsheet to track data requests and receipts.
Status of Data Received
Cadmus produced a memo for the four IOUs that provided detailed information about the data needed
for the study. In February, we began receiving participant data for the Multifamily Energy Efficiency
Rebate Program (MFEER) and the Energy Savings Assistance Program (ESA) and began reviewing the
data. All four IOUs have provided MFEER and ESA participant data to Cadmus.
Cadmus has screened the MFEER and ESA participant datasets for completeness and all appears to be
complete at this time. All participant data received to date has been cleaned and integrated into a SQL
database for further analysis.
Cadmus has received data requested in item 2 of the data request memo, which is account data flagged
as a multi-family account, from SCE. PG&E is in the process of transferring this data and Cadmus is
making further inquiries with SCG and SDG&E. This information will be used to identify the population of
low income multifamily buildings in each IOU service territory.
Table 2 lists the data provided by each IOU.
Table 2. Status of Participant and Account Data Received
Data Received? (Y/N)
Data requested
GIS data layer of each utility territory for
GIS mapping (e.g., shape files).
SCE
PG&E
SDG&E
SoCalGas
Y
Y
Y
Y
Notes
Received from
Energy
1
Using ArcGIS’ ‘Union’ tool
2
Assumes uniform distribution of population. Dasymetric mapping was outside the scope of this analysis.
ESA Multifamily Segment Study Status Update / April 20, 2013
5
Data Received? (Y/N)
Data requested
Customer account data multi-family
accounts (dwellings of 5 or more units)
with any flags indicating a common area
account and CARE tariff.
ESA (LIEE) participant dataset covering last
3 years.
ESA (LIEE) participant dataset from 2009
process evaluation.
MFEER participant dataset covering last 18
months.
Datasets for other programs serving
multifamily customers and low income
multifamily buildings.
SCE
PG&E
SDG&E
SoCalGas
Y
In
Progress
In
Progress
In
Progress
Y
Y
Y
Y
In
Progress
Y
In
Progress
In
Progress
Y
Y
In
Progress
Y
In
Progress
In
Progress
Notes
Commission
Cadmus is
following up with
PG&E, SDG&E,
and So Cal Gas
Cadmus will
compare results
of the current
study to the
2009 study.
Y
In
Progress
Cadmus is
following up with
utilities. Expect to
obtain results of
the DRA dataset
for each utility.
BUILDING OWNER AND OPERATOR SURVEY
In January and February, Cadmus gathered and reviewed multiple public data sources of low-income
property owners and managers. These sources, for which we determined there is some overlap, are:




Multiple Assistance and Department of Housing and Urban Development (HUD) Section 8
Contracts Database
U.S. Department of Agriculture (USDA)affordable rental property list
California Tax Credit Allocation Committee List
HUD’s Low Income Housing Tax Credit list
The data provided by the IOUs will include information to identify common areas and, potentially,
contact information for building owners and managers. These data will be included in the sample frame.
Cadmus is drafting the survey to gather information about the property, property owner, and tenant
characteristics. Cadmus is incorporating questions from the MFEER process evaluation survey and
developing additional questions as needed to cover areas of interest.
List Sources
Cadmus researched several ways to identify a comprehensive list and a sample frame from which to
complete 300 surveys with property owners and managers. We have researched both public and
ESA Multifamily Segment Study Status Update / April 20, 2013
6
purchased lists. The public lists will help us assess the market characteristics among properties with low
income tax credits, rental subsidies, and other affordable housing funding or subsidies. The lists that we
could purchase will augment the study by providing a comprehensive picture of the multifamily housing
market.
ACS data indicate there are a small percentage of condominiums owned by low income families. Since
the numbers are small, they will be difficult to identify. We will contact stakeholder groups to request
information about these and other small groups. We will not develop a separate survey strata for low
income owners of condominiums. However, we will pay particular attention to identifying low income
owners of condominiums. We may interview a small number outside of the parameters of the random
sample (if they are not picked up through random sampling). Additional surveys will not be a statistically
significant stratum, but will provide contextual data for this research.
The public lists are administered by different sources but there is some overlap among the data. Projects
funded with low income housing tax credits (LIHTCs) typically have at least one lender and usually
multiple lenders, and many times these are public lenders. Some of the LIHTC projects located in rural
California have USDA funding, loans, and/or rental subsidies. HUD Section 8 is a rental subsidy program,
not a lender, but many properties with LIHTCs can receive rental subsidies. Any HUD Section 8 and USDA
properties with LIHTCs will overlap on the lists.
Public lists are summarized below:

Low Income Housing Tax Credit (LIHTC). The California Tax Credit Allocation Committee (CTCAC)
administers two LIHTC programs: one federal and one state. Both programs were authorized to
encourage private investment in affordable rental housing for households meeting certain income
requirements. Projects funded with LIHTCs from HUD typically have at least one lender, usually
multiple lenders, and many times these lenders are public lenders. Because of this there is some
overlap between the state tax credit list and the federal tax credit list.

State tax credit. This list is maintained by the CTCAC and contains active projects that it is
monitoring for program compliance and any projects that have been awarded a reservation of tax
credits and are being constructed or rehabilitated. The list contains 3,486 records and 3,237 records
have telephone numbers. The list can be found at http://www.treasurer.ca.gov/ctcac/history.asp

Federal tax credit. This list is maintained by HUD and contains properties that are included in the
LHITC program. There is one administrator of funds in each state. In California, the administrator is
the CTCAC. There are 1,843 individual records on the list and 1,770 have telephone numbers. The
list can be found at http://www.huduser.org/portal/datasets/lihtc.html

Multifamily Assistance and Section 8 Contracts Database. HUD provides rental subsidies for eligible
tenant families. This list contains properties where rents of some residential units are subsidized by
HUD. There is some overlap between this list and the LIHTC lists (state and federal) because HUD
Section 8 is a rental subsidy program, not a lender, but many LIHTC properties have public rental
subsidies, including HUD Section 8 subsidies. The list contains 1,750 records and 1,748 have
ESA Multifamily Segment Study Status Update / April 20, 2013
7
telephone numbers.
http://portal.hud.gov/hudportal/HUD?src=/program_offices/housing/mfh/exp/mfhdiscl

California Rural Development Projects. This list of multifamily rentals in rural California is
maintained by USDA. There is some overlap between lists because rentals receive USDA funding and
may have received a low income housing tax credit and/or Section 8 rental subsidy. The list contains
500 records and 286 of them have telephone numbers. This list was obtained from Lorna Lorea at
USDA.

California Housing Finance Agency (CalHFA).This list is maintained by the California Housing Finance
Agency (CalHFA) and contains records for both HUD Section 8 (described above) and non-Section 8
rental properties. The non-Section 8 records do not have CalHFA-administered Section 8 contracts.
These CalHFA properties are offered to low income residents. There are approximately 450 records
available and about half are designated as low income non-Section 8 subsidized rentals. The list can
be found at http://www.calhfa.ca.gov/multifamily/rental
Possible purchased list sources are:

Property Investor and Owner List. This list contains multifamily rental properties pulled from tax
and deed records of real estate property lists. There are 80,000 records but only 12,183 have
telephone numbers. The records can be appended with the year the property was built, the unit
range size, the loan amount, and the value of the property.

Property Management Companies. This list contains multifamily rental properties. There are 8,644
records with telephone numbers.
Cadmus submitted an electronic public records request for a list of all multifamily rental properties for
which California has property tax records. This request was made through the California Public Records
Act (Government Code section 6250 et seq.). We received a reply from the Disclosure Office stating that
this information is generally maintained at the county level.
We requested these records from a small sample of counties (Marin, Los Angeles, and Alpine) and then
determined this was not a feasible way to acquire all tax records for multifamily housing in California.
We discovered records are available but there is a delivery fee and the timeframe may be several weeks
depending on the county. Property owners are not required to include phone numbers so the lists
contain no phone numbers. Each list contained different information for each property.
Marin County stated the list could be provided but would cost $112. The list would include property
address where known, number of units, owner name, mailing address, and property value.
Los Angeles County sent a cost proposal of $1,362 for the records. It could send all of the same
information as Marin County. In addition, it could provide the year property was built, number of
buildings on the property, and some building characteristics such as square footage and number of
rooms.
ESA Multifamily Segment Study Status Update / April 20, 2013
8
Alpine County said it did not have any records for multifamily housing with five or more units.
Survey
Cadmus drafted a survey that asked questions about property, ownership, and tenant characteristics.
We incorporated survey questions from the MFEER process evaluation we conducted last year and
added questions about other areas. The survey will address the following topics:









Ownership and property management structure
Property characteristics including number of units and number of buildings
Decision-making characteristics including purchases for new and replacement equipment in
the individual units and in common areas
Awareness of energy-efficiency programs and motivation to participate
Program participation
Current, past, and future energy-savings activities
Building characteristics such as number of buildings, number of units on the property, age of
property, and primary fuel type used to heat and cool the property
Tenant characteristics such as turnover rates, rental subsidy use, utility payment structure,
and energy assistance program participation
Additional program needs
SAMPLING PLAN
For our survey of owners and operators of multifamily properties, we plan to use a two-dimensional
design, two strata defined by market rate versus assisted housing and three strata defined by the size of
the property management company operating a particular property. Ordinarily, stratification is reserved
for a situation where we can reduce overall variance within the sample by sampling separate subpopulations—which we do not expect here. Stratification is also used where there is interest in
comparing sub-populations with one another and some of these sub-populations are small. This is the
case on both dimensions of this research plan. Both assisted housing and large operating companies are
rare enough that we would not expect to draw an adequate sample from simple random selection. Yet,
we are quite interested in differences in decision-making and other characteristics of both of these subpopulations. Therefore, we do not propose to stratify by climate zone or property tenure because we
expect these to be broadly distributed enough that some respondents will fall into any likely category
we would devise.
Figure 1 shows an overview of the sampling plan, which we describe in detail, below.
ESA Multifamily Segment Study Status Update / April 20, 2013
9
Figure 1. Overview of the Sampling Plan
Sub-Sector Strata
The sub-sector strata differentiate between market rate housing and assisted housing. To target our
survey to properties known to serve the limited income residential market, a list of properties that have
received housing assistance provides certain access to appropriate respondents. Moreover, housing that
has participated in assistance programs may be systematically different than housing that has not. We
expect, however, that a sizeable portion of the multifamily housing market serving low-income
households will not have participated in any assistance programs. To capture this sub-sector, a more
general sample frame is needed. That will be provided by the market rate stratum. We expect to
complete 150 surveys in each stratum, for a total of 300 completed surveys.
Our sampling frame for the assisted housing stratum will be the database built from lists described
above. We will randomly sample from the unique addresses in the database. Since some of the
component lists contain household contact information, rather than property owner or manager
contacts, we will match addresses with account information provided by the utilities, matching
residential accounts with common area accounts for the same address. These common area accounts—
for hall lighting, laundry facilities, etc.—provide contact information for the property owners and
managers.
For the market rate stratum, we will draw from purchased lists. (We may also use utility account
information for common area accounts. We will obtain a complete list of common area accounts from
utilities.) We will geocode the service addresses to identify their location. We will sample service
addresses randomly from census tracts where, based on our census data research, we estimate
approximately 66% of multifamily households are limited income households. This represents the top
quintile in limited income saturation among multifamily households. We will select from high saturation
areas because neither the purchased lists nor utility account data provide information about the income
status of property tenants. During the initial survey screening process we will assess whether a given
ESA Multifamily Segment Study Status Update / April 20, 2013
10
property is appropriate to our research and include only those that serve low income tenants, where the
maximum income is 80% of area (county) median income. Where the saturation of low income
multifamily housing is higher, this screening process is much more efficient.
By selecting from concentrations of low-income multifamily housing there is some possibility that the
sample will be biased in unknown ways. That is, there may be differences in owners’ and managers’
decision-making in places where low-income properties are more concentrated compared to areas
where low-income properties are less concentrated.
Counteracting this potential for bias, we note that a high concentration of low-income multifamily
households relative to all multifamily households does not indicate a concentration of low income
multifamily households relative to all households. In fact, our approach does not bias the sample
relative to all households. The concentration of low-income multifamily housing is approximately
randomly distributed across the total numbers of households, as illustrated in Figure 2. The average
concentration of low-income households relative to all households among our target census tracts is
only 13%. Thus, while we cannot rule out any bias from our approach, it will not yield only properties
within dense clusters of low-income multifamily housing.
Figure 2. Number of Low-Income Multifamily Households by Total Number of Households
for Census Tracts where More Than 65% of Multifamily Households are Low Income Households
Size Strata
We expect decision-making about multifamily properties to be closely related to the size of the
companies that own and operate the properties. We also expect that larger companies are more
difficult to collect information from in our survey. Yet larger companies manage far more properties
than smaller ones. Thus, it is critical that our sample capture a cross section of company sizes.
ESA Multifamily Segment Study Status Update / April 20, 2013
11
As we contact property owners and managers to screen for the survey, we will ask them how many
properties are managed by their company. Within each of the sub-sector strata, we will complete
surveys for each of three size strata until we have reached a quota of 50, for a total of 150 completes
across the three size strata. We will define the size strata to capture the smallest 45% of companies, the
middle 45% of companies relative to size, and the largest 10% of companies. Based on research
conducted for the MFEER process evaluation, and subject to additional refinement, these three
categories represent companies: managing no more than 25 units, more than 25 but no more than 250
units, and 250 units or more.
Generalizing to the Population of Low-Income Multifamily Properties
In some cases, the analysis will be directed toward differences between the six strata. For these, there
are no particular difficulties presented in using the data. Where interest centers on only one of the
strata or where we intend to generalize to the entire population of properties, we will use postweighting of the results to account for the sample design. For instance, although our design will gather
50% of the data from properties receiving assistance, only approximately 12% of multifamily properties
actually receive assistance. These properties are over-represented in the sample and must be
discounted for the combined analysis. Conversely, the market-rate stratum represents 88% of the
properties but only 50% of the sample; thus these are under-represented in the sample and their effect
on the combined analysis must be inflated. Weighting the results based on population proportions will
allow the combined estimate to accurately reflect the relative prevalence of the two sub-sectors.
Assisted Housing Weight = 0.12
Market Rate Housing Weight = 0.88
Thus, the market rate housing data is weighted approximately nine times as heavily as the assisted
housing data, reflecting the population ratios. The equal stratum size, however, provides enough data
for assisted housing properties that they can be evaluated separately or compared to market-rate
properties.
To properly apply these weights, we require an accurate estimate of the true percentage of properties
receiving assistance. Our current estimate is based on data from the National Housing Preservation
Database (NHPD) and the American Community Survey (ACS). The NHPD is produced by the Public and
Affordable Housing Research Corporation (PAHRC) and the National Low Income Housing Coalition
(NLIHC). The NHPD incorporates all available data on federally subsidized housing properties, including
nine separate funding streams. The NHPD indicates 361,973 assisted housing units in the State of
California within buildings of 5 or more units. The ACS indicates 2,983,028 total multifamily housing
units in the State of California within buildings of 5 or more units. We will look for corroborating
information to establish the final design weight, prior to final data analysis.
We will use the same approach to adjust for the size strata.
Small Companies = 0.45
Medium Companies = 0.45
Large Companies = 0.10
ESA Multifamily Segment Study Status Update / April 20, 2013
12
To combine the strata for an overall estimate—assuming the dimensions are uncorrelated—we simply
cross-multiply the two sets of percentages.
Assisted Housing Small Company Weight = 0.12 * 0.45 = 0.054
Assisted Housing Medium Company Weight = 0.12 * 0.45 = 0.054
Assisted Housing Large Company Weight = 0.12 * 0.10 = 0.012
Market Housing Small Company Weight = 0.88 * 0.45 = 0.396
Market Housing Medium Company Weight = 0.88 * 0.45 = 0. 396
Market Housing Large Company Weight = 0.88 * 0.10 = 0.088
By applying these weights we can combine data from the six strata and accurately reflect survey findings
for the entire population of properties.
Survey Firm
We have researched several California firms that conduct surveys to assess their pricing structure, staff
availability, capacity, and level of experience. The next step is to compare the firms by cost and
experience and determine which one is best suited for this project.
EXAMINE COMMENTS TO ESAP PROCEEDINGS
The team reviewed the comments to the State of California’s ESAP proceeding relevant to the
multifamily sector, classified the comments according to topic, and created a reference tool for further
research as needed.
Cadmus reviewed the Administrative Law Judge (ALJ) decision document, “D1208044 On Large InvestorOwned Utilities' 2012-2014 Energy Savings Assistance (ESA) (Formerly Referred to as Low Income Energy
Efficiency or LIEE) and California Alternate Rates for Energy (CARE) Applications” Related to Proceedings:
A1105017; A1105018; A1105019; A1105020. We conducted a search for the decision document and
proceedings information, downloaded all of the related documents, and filtered out 59 comments and
documents from the CPUC website.
We created a reference tool, cataloging all 59 comment documents by date, document type, entity filed
by, author of document, description, and the direct link to the full electronic document. Cadmus and
Research into Action staff created a list of the stakeholders that presented comments; these included
the CPUC, stakeholders, and the IOUs.
At this initial stage, we created a summary table that organized the comments as proposed, supporting,
or opposing each of the main themes in the ALJ decision document and in a memo sent directly to the
Multifamily Study research team from the National Consumer Law Center (NCLC), California Housing
Partnership Corporation (CHPC), and National Housing Law Project (NHLP).
Cadmus noted the number of times each stakeholder posted a comment to the CPUC website for this
decision and associated proceedings to compare with the summary in the decision document (pages
137-164). We also cross-referenced the CPUC ALJ decision to compare the decision to the most often
occurring themes in the comment document and the memo provided to the research team by NCLC,
ESA Multifamily Segment Study Status Update / April 20, 2013
13
CHPC, and NHLP. We provided some preliminary information during the public meeting on March 5,
2013. Stakeholders and number of comments are listed in Table 3. Distribution of support is shown in
Table 4.
Table 3. Number of Comments Posted by Stakeholder on CPUC Website
Stakeholder Group
Number
of
Comments
Stakeholder Group
Number of
Comments
Association of California Community
and Energy Services
5
National Consumer Law Center
2
Black Economic Council
3
National Housing Law Project
1
Brightline Defense Project
1
Natural Resources Defense Council
2
California Housing Partnership
Corporation
2
Niagara Conservation Corporation
1
California Large Energy Consumers
Association
1
Opower
1
California Public Utilities Commission,
Division of Ratepayer Advocates
3
Pacific Gas & Electric Company
4
Center for Accessible Technology
3
San Diego Gas & Electric Company
7
Energy Efficiency Council
3
Southern California Edison Company
8
Green For All
2
Southern California Gas Company
7
La Cooperativa de Campesina
4
The East Los Angeles Community
Union
5
Latino Business Chamber of Greater
Los Angeles
3
The Energy Efficiency Council
1
Maravilla Foundation
5
The Greenlining Institute
3
National Asian American Coalition
3
The Utility Reform Network
4
ESA Multifamily Segment Study Status Update / April 20, 2013
14
Table 4. Prevalence of Support for Key Proposals/Strategies from the
Decision Document Section 3.10.1 – 3.10.5.6
Number of
Supporters
Proposal
Proposal
Number of
Supporters
Establish "Single Point of Contact"
14
Updated marketing approach to
multifamily homes
7
Assistance to MF owners for central
heat & hot water systems (like HUDDOE/WAP)
12
Value of housing subsidies NOT
counted as income
6
Full integration of ESAP with other EE
Programs (MIDI/EUC/MFEER)
10
Model successful low-income MF EE
programs in other states
4
MF segment underserved; barriers to
entry in ESA for MF
10
Simplify Owner Authorization
(Property Owner Waiver) forms;
coordinate across IOUs
2
"Expedited Enrollment" or
"Categorical Eligibility"
8
Make ESA Program "Neighborhood
Approach" more effective
1
Adopt "whole house," performancebased approach
8
No "carve out" of funds for
investors/owners of deed restricted
MF
2
TELACU multi-phase pilot
8
Set per unit and per building caps on
ESA program assistance
1
IOU STAFF INTERVIEWS
Cadmus developed a set of key questions in an interview guide for IOU program managers and delivered
the interview guide to the study team on February 19. The interviews focus on ESA program delivery,
coordination between multiple programs, integration of multifamily programs, finance, and other issues
brought up by the program staff. The information gleaned from the interviews will provide important
details and insights to help inform all of the research activities of this study.
The Cadmus team conducted interviews with PG&E staff on February 20 and 21, with SCE staff on March
11, with SDG&E on March 18, and with SCG on March 20. Information critical to the study will be
summarized from the interview notes and applied specifically to the development of tasks 4 and 5.
TASK 2 NEXT STEPS

Summarize, review, and highlight key information gleaned from the program staff
interviews conducted at all four IOUs.

Conduct a more detailed review and analysis for each of the comments posted in response
to the decision (D.12-08-044) d as well as the posted comments from the March 5, 2013
public meeting.

Summarize current participation levels by census tract based on analysis of utility
participant data.
ESA Multifamily Segment Study Status Update / April 20, 2013
15

Geocode past program participants to identify underserved areas with high levels of
estimated potential participants.

Segment census tracts into priority levels and create demographic profiles (for example,
year structure was built, age, language, ethnicity, percent of multifamily and single family).

Continue to track incoming customer and participant data from each of the IOUs and followup to collect additional data as needed.
Task 3: Catalog Existing Multifamily Energy-Efficiency Programs
Research Into Action conducted a literature review to identify programs for inclusion in the catalog of
programs serving the multifamily sector, both nationwide and in California. Through this review,
Research Into Action identified 27 programs in California that focus on low income households,
multifamily households, or both. The team also identified 43 multifamily programs in states other than
California.
In a Microsoft Excel document, Research Into Action developed a framework for the program catalog.
This framework builds on the list of topics for inclusion in the catalog listed in the Research Plan,
breaking each element into components that will allow for easier comparison between the included
programs. A list of the program components to be cataloged is attached to this memo. The catalog
framework also provides a structure for tracking the sources of information on each program.
Research Into Action is in the process of collecting details about the programs identified for inclusion in
the catalog. These details are drawn from secondary information sources ranging from ACEEE reports, to
program filings, to monthly and annual program reports, to evaluation reports, to program websites. To
date, the team has reviewed 36 sources of data relevant to programs in California and 36 additional
sources of data relevant to programs nationwide. A bibliography of these sources is attached to this
memo.
The two catalog tasks were approached somewhat differently. For the national review, analysts
specifically filtered out exclusive low-income programs such as WAP or rate assistance programs.
Because these programs are common low-income program models that do not necessarily address
multifamily buildings the research team focused instead on multifamily programs, ideally those with
specific outreach or augmented services designed to reach low-income residents.
The California catalog task had a somewhat different scope. Because ESA, LIHEAP, CARE and FERA were
included in our review there are programs in the California catalog that focus on low-income residents
but do not have explicit multifamily components. The research team will consider this context in the
comparison and review tasks that follow this initial data gathering.
The research team continues to gather information for the catalog for all of the identified programs. We
will catalog each program as completely as possible given the information provided in publicallyavailable secondary sources. Given the limitations of these sources, it will not be possible to fill in every
ESA Multifamily Segment Study Status Update / April 20, 2013
16
catalog field for every program identified, particularly for programs administered outside of traditional
energy efficiency organizations, which often provide little publically available information.
The research team has found it particularly difficult to obtain detailed information related to program
cost effectiveness, including budgets, expenditures, and energy savings, as well as any adjustments and
assumptions underlying these figures. In some cases, these data may be reported for a program
portfolio as a whole, rather than specific to multifamily or low income program components.
In addition, detailed information about implementation approaches and eligible measures is also
difficult to ascertain from publically available information– for example, determining precisely which
measures are included in ‘weatherization’ offerings. Finally, catalog information for programs that were
recently proposed or launched will necessarily be less detailed than information for more established
programs since these programs have had little time to generate results and are less likely to have
publically available evaluation reports.
Ultimately, during Phase I, it may be necessary to compare some programs at a more general level
rather than the detailed level the research team initially attempted. The detailed program summaries
that the research team will create in Phase II will draw on interviews with program managers and more
in-depth document review to fill in information on leading programs.
TASK 3 NEXT STEPS

Continue to catalog programs.

Status update of progress for the catalog of multifamily energy-efficiency programs.

Prepare bibliography.
Task 4: Review and Evaluate Multifamily Programs and Research Relevant for
Low-Income Customers
Research Into Action is developing a detailed understanding of the landscape in which programs
targeting the low-income multifamily sector operate in California. To this end, the team reviewed:




The ESA decision (D.12-08-044) with a focus on the key issues related to the multifamily sector.
CPUC guidance on the Energy Upgrade California Multifamily program.
The California Energy Efficiency Strategic Plan.
The Multifamily Home Energy Retrofit Coordinating Committee (MF HERCC)’s report, Improving
California’s Multifamily Buildings.
Staff also reviewed comments to these proceedings relevant to the multifamily sector and created a list
of organizations that had submitted comments for future contact. The team classified the comments by
topic and created a reference tool for further research as needed.
Research Into Action and Cadmus staff also spoke with representatives from the National Consumer Law
Center and the California Housing Partnership, who shared their views on the ESA proceeding.
ESA Multifamily Segment Study Status Update / April 20, 2013
17
TASK 4 NEXT STEPS

Continue reviewing documents related to ESA and multifamily proceedings.

Prepare summary memo describing the detailed approach that will be implemented in Phase 2.
Task 5: Identify and Assess Alternative Program Designs and Delivery
Strategies
CURRENT PROGRAM COMPONENTS
To lay the groundwork for program design considerations and options, Cadmus leveraged findings from
process evaluation reports and program materials provided by the IOUs and found on the Internet.
Using data collected on the LIEE/ESA, MFEER, and MIDI programs, Cadmus compared the similarities and
differences in critical program components among these three programs, as well as the different ways
these programs are implemented across the four IOUs.
Cadmus compared the following program features:









Implementation strategies and delivery mechanisms
Hard-to-reach markets and the strategies for overcoming barriers reaching these markets
Program-sponsored training opportunities
Marketing tactics including leveraging trade ally marketing efforts
Energy education for participating customers
Program requirements including eligibility qualifications and treatment of previous
participants
Measure offerings by IOU, climate zone, and sector
Data tracking processes
Quality assurance and quality control protocols
The supporting documentation for the comparison described above is included as an electronic
attachment to this report. (ESA Program Design Components, Excel Worksheet)
TASK 5 NEXT STEPS



Organize the data Cadmus collected so that it can be easily incorporated with the data
collected by Research into Action through task 3. Specifically, we will map our findings to
the program features of multifamily energy-efficiency programs (particularly those for lowincome customers) operating in California and nationally. This will provide comparisons
across programs and IOUs.
Review and highlight key information gleaned from the program staff interviews conducted
at all four IOUs.
Continue to compile data regarding California’s programs by reviewing recommendations
from ESA proceedings, findings from additional studies, and research results from other
study tasks, specifically task 2 and task 3.
ESA Multifamily Segment Study Status Update / April 20, 2013
18

Compile preliminary program design strategies, options, and suggestions for discussion with
the study team. This will allow us to more fully develop alternative program designs and
delivery strategies, providing suggestions for a long-term program strategy.
Task 6: Identify Financing and Funding Options
DEVELOP LIST OF RELEVANT FINANCING OPTIONS
Cadmus will leverage the work completed as part of task 3 and summarize existing financing options
after task 3 is completed.
Currently, the IOUs are awaiting a final decision from the Energy Division approving the financing pilot
designs for 2013-2014. We anticipate one of these pilots will be targeted to the multifamily sector and
will be included in our list of existing finance options.
Once Cadmus listed financing solutions (after task 3 is completed), we will talk to the IOU financing
product managers (Al Gaspari, PG&E; Jim Dodenhoff, SCE; and Frank Spasaro, Sempra) and CHPC to see
if they are aware of any additional solutions available to affordable housing owners. If Energy Division
provides referrals to agencies that provide funding and financing, we will also interview those contacts
(up to three).
Understand ESAP Cost-effectiveness
Cadmus staff interviewed Mary O’Drain, PG&E, in March regarding cost-effectiveness calculations for
ESAP. We developed an interview guide for this discussion; our findings indicate the following:





Cost-effectiveness for ESAP includes two tests, modified participant and utility cost test.
Non-energy benefits, such as water savings, are included where quantifiable.
The cost-effectiveness approaches are consistent statewide.
Since ESAP currently does not include significant owner co-pay components, much less
owner financing, additional discussion with the cost-effectiveness working group will be
needed to determine how property owners’ use of financing will affect cost-effectiveness
for ESAP.
PG&E indicated both owner and renter financing solutions are on the table for consideration
at this time.
REVIEW RESULTS FROM OTHER STUDY TASKS AND OTHER STUDIES
To determine what type of financing solution(s) should be proposed, we have to understand if there is a
need to further segment the low-income multifamily sector. For instance, in completely master-metered
properties, split incentives are usually not a concern. Another consideration might be if a commercial
building also has existing obligations that can make taking out a loan difficult. We expect some of the
existing tasks in this study (e.g., task 2) to provide insights on these issues.
In addition, Cadmus conducted focus groups with multifamily property owners and representatives in
March as part of the PG&E Financing Pilot Support Research project. The results are still being compiled.
ESA Multifamily Segment Study Status Update / April 20, 2013
19
Cadmus notified the PG&E Financing Products manager (refer to CWA 2500739126) about our
multifamily sector characterization work and indicated the multifamily sector focus groups Cadmus is
conducting in March could result in information useful to both projects. Cadmus staff met internally to
discuss the two projects, the overlap in target customer populations, and questions to be asked during
the focus groups.
SCE and Sempra Utilities have expressed interest in additional focus group work after the financing final
decision approving the pilot designs and may want Cadmus to conduct additional focus groups with
multifamily customers.
TASK 6 NEXT STEPS

Conduct additional interviews to discuss cost-effectiveness approaches used by the four IOUs.

Work with the task 2 staff leads to develop questions about financial decision for the building
owner/operator surveys.
Task 9: Conduct Public Workshop to Discuss Research and Recommendations
The Cadmus team conducted a public workshop on March 5 in San Francisco to present the research
plan to stakeholders. Our objective was to review the study plan with the stakeholders, solicit their
feedback, and request their assistance in gathering relevant data to inform the study.
We introduced the project and outlined the key tasks in the work plan in a PowerPoint presentation.
Attendees were asked to review the draft work plan posted on the Energydataweb.org site and to
provide any detailed or formal comments through Website comment process.
Attendees were encouraged to ask questions and make comments during the presentation. An open
discussion followed the presentation. Cadmus prepared a summary of the discussion after the
workshop, identifying key points for consideration in the study.
TASK 9 NEXT STEPS

Conduct a second public workshop to review the findings from Phase 1 of the study, to be held
after the Phase 1 draft report is completed and submitted to the study team in June 2013.
ESA Multifamily Segment Study Status Update / April 20, 2013
20
Appendix A: Supporting Material for Census Tract Analysis and GIS Mapping
Figure A-1 below shows low-income, multifamily households by census tract in the Bay Area, as well as
by county for the whole state (inset map).
ESA Multifamily Segment Study Status Update / April 20, 2013
21
Figure A-1. Low-income Multifamily Households in Bay Area
ESA Multifamily Segment Study Status Update / April 20, 2013
22
Table A-1 below shows the estimated number of low-income, multifamily households at the county
levels.
Table A-1. Estimated Number of Low-income Multifamily Households by County
County
Alameda County
Alpine County
Amador County
Butte County
Calaveras County
Colusa County
Contra Costa County
Del Norte County
El Dorado County
Fresno County
Glenn County
Humboldt County
Imperial County
Inyo County
Kern County
Kings County
Lake County
Lassen County
Los Angeles County
Madera County
Marin County
Mariposa County
Mendocino County
Merced County
Modoc County
Mono County
Monterey County
Napa County
Nevada County
Orange County
Placer County
Plumas County
Riverside County
Sacramento County
San Benito County
San Bernardino County
San Diego County
San Francisco County
San Joaquin County
San Luis Obispo County
San Mateo County
Census
Tracts
361
1
9
51
10
5
208
8
43
199
6
31
31
6
151
27
15
9
2,346
23
56
6
21
49
4
3
94
40
20
583
85
7
453
318
11
369
628
197
139
54
158
ESA Multifamily Segment Study Status Update / April 20, 2013
Estimated LowIncome MultiFamily Households
66,142
425
371
7,031
333
464
26,152
517
3,091
32,150
429
4,246
5,735
351
16,425
2,777
1,273
691
573,898
1,819
8,377
169
2,110
5,727
98
3,007
13,547
3,933
1,839
110,224
7,920
465
54,292
57,940
572
49,494
146,923
60,874
18,457
7,639
20,910
23
County
Santa Barbara County
Santa Clara County
Santa Cruz County
Shasta County
Sierra County
Siskiyou County
Solano County
Sonoma County
Stanislaus County
Sutter County
Tehama County
Trinity County
Tulare County
Tuolumne County
Ventura County
Yolo County
Yuba County
Total
Census
Tracts
90
372
53
48
1
14
96
100
94
21
11
5
78
11
174
40
14
8,057
Estimated LowIncome MultiFamily Households
16,965
56,141
7,423
4,629
54
1,239
11,095
12,710
11,423
3,086
1,463
155
6,804
1,001
18,900
10,394
1,728
1,484,048
Table A-2 and Table A-3 below show the estimated number of low-income, multifamily households by
IOU service territory.
Table A-2. Estimated Low-income Multifamily Households by Electric Service Territory
Electric Utility
Pacific Gas & Electric Company
San Diego Gas & Electric
Southern California Edison
Total
Estimated Low-Income
Multi-Family Households
398,602
152,273
417,447
968,322
Table A-3. Estimated Low-income Multifamily Households by Gas Service Territory
Gas Utility
Pacific Gas & Electric Company
San Diego Gas & Electric
Southern California Gas
Total
Estimated Low-Income
Multi-Family Households
459,040
146,921
819,097
1,425,059
ESA Multifamily Segment Study Status Update / April 20, 2013
24