Part A. Justification
A1 Circumstances That Make the Collection of Information Necessary
The Housing Choice Voucher (HCV) program is the federal government’s largest low-income housing assistance program. The HCV program currently serves approximately 2 million households nationwide. It is administered federally by the U.S. Department of Housing and Urban Development and locally by approximately 2,400 local, state, and regional housing agencies, known collectively as public housing agencies (PHAs). Funding for the HCV program is provided entirely by the federal government. The funding that PHAs receive includes the housing subsidy itself, plus administrative fees to cover the costs of running the program. In fiscal year 2012, HUD will spend approximately $1.35 billion on administrative fees to PHAs for the HCV program.
When the voucher program was first implemented in the 1970s, the system for reimbursing PHAs for the costs of program administration was loosely based on empirical evidence. Over time, however, the system for estimating and allocating fees has become more complex and—in some ways—more arbitrary, as HUD and Congress have tried to balance fairness with cost savings, while trying to avoid large year-to-year swings in funding for PHA staffs.
The Housing Choice Voucher Program Administrative Fee Study is designed to evaluate the amount of funding needed to administer a high-performing and efficient HCV program based on direct measurement of the work actually performed by voucher administrators and detailed collection of overhead and non-labor costs associated with the program. The study will measure and identify the tasks performed by PHA staff to meet program requirements, to assist voucher holders in finding and renting suitable housing in a timely way, and to ensure that a broad range of affordable rental housing throughout the community is available to voucher families. The study will estimate a total administrative cost per voucher for each PHA in the study, as well as costs for the major tasks involved in voucher administration. Ultimately, the findings of the study will be used to inform the development of a new formula for allocating HCV program administrative fees.
The study is limited to PHAs with high-performing and efficient HCV programs. The reason is that the purpose of the study is to provide cost information to inform the development of a new administrative fee formula, and HUD wants to model administrative costs only at those HCV programs that are high performers. The study is not intended to provide guidance on how standard and low performing agencies could improve performance or efficiency. If that was the study goal, the sample would require looking at lower performing programs as well.
The study is proceeding in phases. The first phase, the reconnaissance phase, involves developing the study design and identifying candidate sites for the study. This phase of the study is now largely complete, although we are continuing to identify study candidates. We identify study candidates through site visits and file reviews of PHAs with a track record of high-performance on SEMAP, the program that HUD uses to assess HCV program performance. The site visits and file review activities are designed to determine whether these PHAs meet the criteria for inclusion in the study. We obtained OMB approval for the initial reconnaissance visits on March 18, 2011 (OMB Control No. 2528-0267) and for a second round of reconnaissance visits on March 27, 2012 (OMB Control No. 2528-0283).
The next phase of the study is a pretest phase to test the data collection methodology for the full national study. We obtained OMB approval for the pretest on February 10, 2012 (OMB Control No. 2528-0267). The pretest involves collecting data on the costs of HCV program administration from four PHAs across the country that operate high performing and efficient HCV programs of different sizes. Data collection includes measuring the staff time spent on the program using random moment sampling, as well as collecting and validating PHA data on overhead costs and non-labor costs of program administration. Data collection for the pretest was completed in mid-June 2012, with analysis of pretest findings taking place in June and July 2012.
The final phase of the study is the full national study, for which OMB approval is currently being sought. The full national study will involve the collection of data on the costs of HCV program administration from a national sample of 50 to 60 high performing and efficient PHAs using the data collection methods tested and refined through the pretest. The final sample size for the full study will depend on the availability of funding, but will be between 50 and 60 PHAs. Data collection for the full study is expected to begin in November 2012 and continue through early 2014. The results will be used to develop costs estimates for HCV administration and recommendations for a new administrative fee formula for the HCV program in 2014.
A2 How and by Whom the Data Will Be Used
A2.1 Project Overview
The goal of the project is to collect detailed and accurate information on the costs of doing all the tasks associated with administering the HCV program, that is, all the labor and non-labor costs that the administrative fee is supposed to cover. While PHAs report the costs of their HCV programs overall, this is not done at a task level. We also learned through the first round of reconnaissance site visits that the reported costs do not always represent actual costs. Staff who work on multiple housing programs might spend more or less time on the HCV program relative to other housing programs (such as public housing) than is represented by the share of staff costs that are “charged” to the HCV program.
Another reason for doing primary data collection on costs (versus relying on reported cost data) is that PHA functions that serve all the agency’s programs, such as accounting and human resources, may be charged to the HCV program in different ways. PHAs are not required to calculate what share of these overhead costs are actually used by the HCV program, but that is what is needed in order to arrive at a complete cost of operating the HCV program, which includes the staff time (and associated personnel costs) spent on the program’s front-line activities, the non-labor costs associated with these front-line activities, the staff time spent on overhead activities that support the HCV program, and the non-labor costs associated with these overhead activities. In sum, in order to estimate costs for the individual tasks involved in operating the program, and in order to isolate costs attributable to HCV program administration (versus the other activities a PHA may perform), we need to conduct detailed cost data collection and analysis of the type proposed for this study.
We will conduct this detailed cost data collection and analysis for a sample of 50-60 PHAs with high-performing and efficient HCV programs. The PHAs will be selected randomly from among all PHAs with SEMAP high performer status in three of the past four years and that are confirmed via site visits, which go somewhat beyond SEMAP, to meet the study criteria for high-performance. In selecting the random sample of PHAs for consideration for the study, we stratified by size in order to ensure that the final sample covers a range of program sizes. By virtue of being randomly selected, the final sample is expected to cover a range of different PHA types and market conditions, although no stratification was done for these characteristics.
The cost data collection will produce estimates of the overall cost per voucher of administering the program at each PHA and estimates of the cost of conducting key tasks required for HCV program administration. We will adjust for differences across PHAs in labor market costs and then will run a regression analysis of the overall cost per voucher using a variety of PHA characteristics—such as HCV program size, urban/rural location, client demographics, and staff caseloads—to isolate cost drivers. We will use these cost drivers to develop a fee formula for all PHAs that is based on the average costs per voucher for the sample PHAs and takes into account the key factors that make costs higher or lower for PHAs. For example, if serving a high share of clients with disabilities is found to have a significant effect on costs, the PHA’s share of clients with disabilities could be a factor in determining how much administrative fee it receives per voucher.
The information on the costs of different administrative tasks will be used to inform HUD and the industry on which aspects of the program cost the most to administer and whether there may be opportunities to achieve greater efficiencies. If we find little variation across PHAs in the cost of doing a given task, it is likely that there is little that can be done to change that cost. On the other hand, a wide variation would call for further analysis of PHA characteristics and practices associated with higher or lower costs and potential recommendations for program changes or changes in administrative practices.
We will also use the data collected through the study to answer related questions of interest to HUD and the housing industry, including: is there a minimum number of vouchers needed to operate the HCV program on administrative fees alone, and what would be the appropriate fees/costs for Family Self-Sufficiency (FSS) coordinators.
A2.2 Purpose of the Data Collection
We are requesting OMB approval for four data collection activities:
Measurement of PHA staff time via random moment sampling (RMS) at 50 to 60 PHAs.
Cost data collection through analysis of financial documents and interviews with PHA staff at 50 to 60 PHAs.
Transaction count data collection at 50 to 60 PHAs.
Telephone interviews with up to 130 PHAs with less than 250 vouchers.
The purpose of each data collection activity is described below.
Measurement of PHA Staff Time per Activity via RMS
At the 50 to 60 PHAs participating in the main study, all staff who work on any of the core functions of the HCV program will be asked to record the time they spend on HCV program tasks and sub-tasks over a two month period (approximately 40 working days) using RMS. (Note that at PHAs with over 100 HCV program staff we will sample 100 staff members to complete the RMS). Each staff member will be given a smart phone for the two-month period and will be asked to respond to multiple “notifications” from the phone during that time (12-15 per day). The notifications guide the staff person through a succession of touch-screens in order to identify the specific task the person was working on at a particular point in time.
Each PHA staff participating in RMS receives customized sampling of notifications to reflect his or her work pattern. Prior to the start of data collection, the study team loads information on every PHA staff’s specific work schedule by day of week and hours in a day on web site. Almost any schedule configuration is possible. The notifications are then distributed across specified time blocks within the staff person’s particular work schedule. If a staff person states that he or she occasionally works late or on weekends, the phone is triggered to sample “odd” hour windows as well. Odd hours include evenings until 10:00 PM as well as weekends. However, if a particular PHA staff person does not respond to an “odd hour” notification within three hours it gets cleared from the phone.
Appendix A shows the succession of questions and screens involved in responding to a notification. Depending on what activity a staff person is doing, responding to a notification could involve as few as two screens (or clicks) or as many as six screens (or clicks). Under any scenario, it takes a person less than a minute to respond to a notification. Appendix A also shows the staffing chart that we will ask PHAs to complete to identify the staff to be included in the RMS data collection.
The purpose of the data collection is to obtain estimates of the proportion of time each staff works on different HCV program tasks, as well as work outside the program and time off. The goal is to use this sampling technique to arrive at accurate estimates of the amount of time each staff spent on a given activity over a two month period. This estimate can then be multiplied by the hourly cost of that staff (salary plus benefits) to arrive at a total labor cost for that activity during the two month period. We then use the information obtained from the PHA on how many times the activity occurred over the two month period to arrive at a labor cost per activity. The data on overhead costs and non-labor costs obtained through PHA interviews and review of financial information can be then added to the labor cost to arrive at an estimate of the total cost per activity.
The reason for collecting the information on time spent per activity using RMS versus timesheets or traditional time and motion direct observation is that we determined (through beta-tests conducted during the reconnaissance phase of the study) RMS to be the most cost effective method of achieving the study’s goals.
First, we wanted to collect information on a number of subtasks associated with HCV program administration, as well as tasks associated with the FSS program, and with special voucher programs. Overall, we expect to generate time (and cost) estimates for approximately 10-15 separate activities or subtasks. This level of detail is difficult to collect using a timesheet, which most staff would complete at the end of the day. The timesheet would need to include all of these tasks as potential “rows,” as at some PHAs staff perform all (or most) of these functions, as well as separate columns to record the portion of time spent on special voucher types versus the regular tenant-based vouchers. This would be very difficult to achieve with a paper timesheet given the number of rows and columns, and we were concerned about using a web-based system that would require staff to log into a separate system each day to record their time. Also, some staff, such as inspectors, are not in the office every day, so being able to report their time on a handheld device is convenient.
Second, we determined that we needed to collect data over a long period of time (at least two months) in order to capture rare events. Some activities, such as intake, do not happen very frequently, especially at smaller agencies, so the data collection period needs to be long enough to pick up such activities. Feedback from the reconnaissance phase suggests that doing timesheets for two months would be much more burdensome than doing RMS. Doing a traditional time and motion study in which PHA staff are shadowed by observers over a two month period, would be prohibitively expensive.
Third, much of the work that HCV program staff does involves working on the computer or speaking by phone or in person to a client. We determined in the reconnaissance phase that it is not easy to do a time and motion study on this type of work, as the observers would need to interrupt workers periodically to ask them what they were doing. In addition, “multi-tasking” would not easily be captured.
The phone used for RMS data collection is an LG Vortex running Android 2.2. The custom phone application programmed for this study is the only capability available to the PHA respondent (web browsing, email, phone, and open text messaging are not available). The custom app only supports the specific HCV data collection and direct messaging between the PHA staff user and the study team. The messaging allows the study team to communicate with a respondent who may require assistance. There are numerous features in the app that allow PHA staff users to verify their responses, see old responses, and perform “multi-response” answering when they have been out of the office for a prolonged period. The phone’s GPS capability will not be activated for this study because it is not needed.The phones will be preprogrammed with all study notifications at the start of the data collection period. The phones will also store all user responses, streaming the responses to a web server in real-time via 3G data connectivity. If 3G data connectivity is lost at any time, the device continues to store the user responses but does not stream them to the server. The user has no change in experience. Once connectivity is restored, all the responses are streamed to the server. The phones are configured to easily store all eight weeks of data collection.
HUD owns the smartphones and will receive them once data collection is complete.
Cost Data Collection
We will ask each PHA in the study to provide us with a detailed budget for the HCV program, detailed financial statements for the most recently completed fiscal year, and salary and benefits information for HCV program staff and overhead staff for the most recently completed fiscal year and for the time measurement period. The reason for collecting cost data is that we need accurate information on the labor costs for the staff involved in RMS (salaries and benefits for the data collection period) in order to estimate the labor cost associated with the activities recorded through RMS. We also need accurate data on overhead costs and on non-labor program costs in order to create an estimate of the total costs associated with a given HCV activity. We will also ask for auditors’ reports and will obtain administrative data from HUD’s Voucher Management System for each PHA. We will use this information to develop preliminary estimates for labor and non-labor costs associated with HCV frontline activity, for the labor and non-labor costs associated with overhead functions performed for the HCV program, and for the labor and non-labor costs associated with the HCV program’s share of overhead functions performed (such as payroll) for all programs the PHA administers.
Once we have developed these preliminary cost estimates, we will interview the PHA’s financial staff and HCV Director to determine whether any adjustments need to be made to capture accurately the overhead and other costs associated with the HCV program. For example, a PHA may charge all legal expenses as overhead expenses, when in fact the legal fees may be directly attributable to a program, such as the development of a project the PHA undertook. In this case we would re-categorize the legal fees as a direct expense associated with the development project.
We plan to conduct most of the PHA interviews in person, but some interviews will take place by telephone. As part of the pre-test we tested both the in-person and phone methods. We determined that for small programs and those with simple finances data collection can be carried out by phone. We will send a list of documents required and data elements to be collected to the PHAs in advance of the interviews so that we can make the most efficient use of the interview time. Appendix B shows the materials requested from the PHAs in advance and the interview questions to be asked of PHA financial staff.
Transaction Count Data Collection
In order to create estimates of the time and cost per activity—for example, the time and cost per inspection—we need to know how many times that activity happened during the RMS data collection period. We will ask the PHA to generate these “transaction counts” prior to the start of RMS data collection for the previous 12 months and again at the end of the data collection period for the two-month RMS period. Appendix C presents the list of transaction counts that we expect to request.
Telephone Interviews with Small HCV Programs
About 15 percent of high-performing HCV programs administer 100 vouchers or less. It is not feasible to measure the time that staff spend on specific HCV tasks at programs with less than 100 vouchers because they tend to have very few staff (typically less than three) and relatively little program activity over a two-month period; thus, through RMS we would not obtain a sufficient number of moments to obtain reliable estimates of the percent of time spent on each activity
We will not attempt to measure staff time per activity using RMS for PHAs with fewer than 100 vouchers. Instead, we will use telephone data collection to measure the staff costs associated with administering the program – the share of time that each staff spends on the program and his/her salary or benefits – and the other costs associated with administering the program.
We plan to conduct telephone interviews with up to 130 high-performing HCV programs with less than 250 vouchers. (A copy of the interview guide is provided in Appendix D.) The reason for including agencies in the 101 to 249 voucher size range is that we want to analyze whether there is a minimum number of vouchers below which the costs incurred by the PHA to administer a voucher exceed the amount of HCV administrative fee received. Based on the number and range of tasks involved in administering the HCV program, the cost of labor for staff able to do these tasks, and the current administrative fee levels, we estimate the minimum size for an HCV program to be between 70 and 140 vouchers, if the program is to be supported by administrative fees alone.
The information collected through the telephone interviews with small HCV programs will be used to arrive at a per voucher administrative cost for these agencies that is more accurate than that currently available through HUD’s VMS because it provides a better estimate of the staff time spent on the program (versus the staff time charged to the HCV program) and also monetizes in-kind services (such as office space) provided to the HCV program at reduced cost or free of charge.
A2.3 Who Will Use the Information
HUD will use the results of the study to quantify the costs of administering the HCV program, to develop a new formula for allocating administrative fees in the HCV program, and to answer related questions about the cost of operating the FSS program and the minimum number of vouchers for an HCV program to operate on HCV administrative fees alone. The information collected through the study will not be used to provide guidance on how agencies could improve HCV program performance.
A2.4 Instrument Item-by-Item Justification
Exhibit A-1 describes the target respondents, content, and reason for inclusion for each data collection activity that involves individuals: time measurement via RMS, cost data collection, transaction count data collection, and telephone interviews with small programs. Copies of the data collection instruments are provided as Appendices.
Exhibit A-1. Item-by-Item Justification of Data Collection Instruments
Data Collection Activity |
Data Collection Instrument(s) |
Respondents, Content, and Reason for Inclusion |
Time Measurement via RMS |
RMS Screen Flow (Appendix A)
|
Respondents: All PHA staff involved in HCV front-line activities: eligibility and intake; lease up; ongoing occupancy; inspections; monitoring, supervisory, and accounting activities; supportive services; customer service; community/owner relations; and HCV staff meetings. Content:
Reason: Time measurement via RMS will be used to develop estimates of the staff time spent on each HCV program activity as well as staff time spent on the HCV program as a whole. The data will be used to develop estimates of overall and per task costs of operating the HCV program. |
Cost Data Collection |
Cost Data Collection (Appendix B) |
Respondents: Select PHA staff with detailed knowledge of HCV and agency costs, including the HCV Director, the Director of Finance, and other PHA staff or consultants (e.g., a fee accountant) as needed. The specific individuals to be interviewed will be determined in consultation with each PHA Content:
Reason: It is necessary for estimating program costs to have accurate and up-to-date information on the number of vouchers allocated and under lease, including special program vouchers, as well as the number of participants in the PHA’s FSS program. We also need to collect information on HCV program costs and overhead costs from the PHA in order to account for the full cost of voucher administration. Information on the share of overhead costs used by the HCV program is not readily available in PHA financial and administrative data and therefore must be obtained through a combination of document review and interviews with PHA finance staff. Finally, we will collect information on the impact of recent cuts to the HCV administrative fee in order to understand how program costs have changed over time. This information could be used to inform the overall cost estimates produced for each PHA. |
Transaction Counts Collection |
Transaction Count Data Collection (Appendix C) |
Respondents: Select HCV staff needed to run reports on HCV program transactions from the PHA’s management information system. Content:
Reason: We need to know the volume of program activities during the time measurement period and the previous 12 months in order to make estimates of the time and cost “per activity.” For example, we need to know the number of inspections that took place over the data collection period in order to estimate the time and cost per inspection. |
Telephone Interviews with Small Programs |
Small Program Interview Guide (Appendix D) |
Respondents: 1-2 staff at up to 130 PHAs with small HCV programs (< 250 vouchers). Content:
Reason: We need to estimate the total cost of administering the HCV program as a stand-alone program in addition to the share of costs currently borne by other programs or entities for small HCV programs not included in the time measurement sample. These costs will be analyzed as part of the development of the new administrative fee formula and will be used to inform the question of whether there is a minimum size for the HCV program to be operated on HCV administrative fees alone. We also need to understand the impact of reductions in HCV administrative fees to be able to assess whether the current program costs reflect cuts that have been made and that could result in reductions in the quality of service delivery. |
A3 Use of Improved Technologies
Among the data collection activities for which OMB clearance is being sought, the time measurement data collection via RMS makes use of improved technologies. RMS data collection will be done through a specially designed smart phone provided to PHA staff by the study team. The smart phone methodology has several advantages, including:
The ability to be carried around with PHA staff at all times.
The ability to for the research team to monitor the responses of PHA staff in real time, helping to ensure completeness and accuracy.
The ability to collect very detailed information on the types of work that PHA staff are engaged in, using multiple activity and sub-activity categories, and allowing staff to link their work to different household and voucher types.
Interviews with PHA staff will be conducted in person or by telephone and with minimal use of technology. To the extent possible, we will email PHA staff materials in advance of the interviews that they can use to prepare and that will reduce time that staff need to spend being interviewed.
A4 Efforts to Avoid Duplication
HUD is not undertaking any similar studies of staff time or HCV program costs.
A5 Involvement of Small Entities
The study will include some small nonprofit organizations and small units of government. We will work closely with the staff of small agencies to make sure that the data collection is done most efficiently and with the least burden on staff. We have a representative from a small PHA on the study’s Expert and Industry Technical Review Group who has reviewed the data collection approach and instruments.
A6 Consequences of Less Frequent Data Collection
The data collection will occur one time only for each PHA in the full study. The data collection for this study will not be repeated for the PHAs involved in the pretest of the data collection approach.
We do not expect any additional data collection beyond the current request.
The proposed data collection activities are consistent with the guidelines set forth in 5 CFR 1320.6 (Controlling Paperwork Burden on the Public—General Information Collection Guidelines). There are no special circumstances that require deviation from these guidelines.
A8 Consultations Outside the Agency
Draft versions of the data collection materials were reviewed in November 2011 by the Expert and Industry Technical Review Group (EITRG) created for the study. The EITRG consists of PHA staff, housing researchers, housing industry representatives, and time measurement experts. The Expert and Industry Technical Review Group will also review the results of the pretests that are currently underway.
In accordance with the Paperwork Reduction Act of 1995, the Department of Housing and Urban Development published a 60-Day Notice of Proposed Information Collection for Public Comment: Data Collection for Full Housing Choice Voucher Program Administrative Fee Study on February 7, 2012. The Department also provided to OMB a notice for publication in the Federal Register announcing the 30-day notice for public comment on the proposed data collection.
A9 Payments to Respondents
A9.1 Proposed Compensation for PHAs Participating in Time Measurement Study (50 to 60 PHAs)
In pretesting the time measurement data collection for the study, we compensated PHAs using a two part structure that recognized (1) the burden of responding to interview questions and assembling financial and cost data, which is roughly equal for all participating PHAs; and (2) the burden of participating in the RMS data collection, which will vary substantially across PHAs depending on how many staff the PHA has working on the HCV program. First, the four pretest PHAs received $2,800 to offset the costs associated with assembling financial and other HCV administrative data and for participating in three sets of interviews with the study team. Second, the PHAs received a payment equivalent to $300 times the number of staff participating in the RMS data collection.
We propose to provide this same level of compensation to the 50 to 60 PHAs that will participate in the time measurement part of the study. As was done for the pretest, we propose to make two payments to participating agencies: one payment of $2,800 at the start of RMS data collection (after the cost data and first set of transaction counts are provided) and a second payment equal to $300 per staff once all data collection activities are complete. As described further below, the compensation is important for ensuring a high rate of participation in the study and full PHA cooperation in the data collection effort.
The compensation (both the $2,800 and the per-staff amount) will be provided to the agency and not to individual agency staff. Abt Associates will sign an MOU with each agency as a condition of providing the payment. The MOU will state that “the compensation is expected to be used to defray personnel costs due directly to this information collection.” A copy of the MOU is provided in Appendix E.
A9.2 Justification for Compensating for PHAs Participating in Time Measurement Study and Rationale for Compensation Amount
Justification for Compensating PHAs
Recent research has argued that direct measurement through observation or surveys such as random moment sampling (RMS) is an important element of cost analysis given the limitations of administrative data.1 But direct measurement has not been used extensively for policy analysis outside the health care field. The data collection proposed for this study is highly innovative for housing research, and if executed effectively, could be an important contribution to the field.
It is critical to compensate PHAs financially for their participation in the study in order for the data collection to be effective. The purpose of the compensation is to offset the personnel costs incurred by the PHA from devoting staff time and resources to assembling financial and other administrative data for the study team, responding to interview questions, and (most important) reporting multiple times a day on their work activities for a full two months via RMS. Given the length of the data collection period and the repeated interruptions to the work of PHA staff that RMS entails, it is important that the PHA leadership and HCV program managers incorporate the study requirements into their staff’s workload rather than the study becoming a source of conflict. The time measurement data collected will not be meaningful unless all HCV staff at a PHA (or the sample of staff for larger programs) participate in RMS consistently (answering all 12-15 notifications per day) and over the full data collection period (40 days).
One reason that compensating agencies is important is that the amount of funding that PHAs receive from HUD for operating the HCV program has been effectively cut by approximately 25 percent since the study started due to administrative fee proration. We learned during the reconnaissance phase that some agencies have laid off staff, some have opted not to fill staff vacancies, and some have consolidated office space. These cost reduction measures have resulted in higher caseloads for line staff and more stressful working conditions.
At a hearing of the Senate Committee on Banking, Housing and Urban Affairs on April 26, 2012, HUD Secretary Shaun Donovan expressed concerns about the impact of the funding cuts on PHA operations: “First of all, I am very concerned about the level of administrative funding for PHAs. We have seen really what has been unheard of; more than 10 agencies around the country have actually refused, actually turned back to us vouchers that help homeless veterans find housing. We have more than a dozen agencies that have turned back their entire voucher programs just in the first few months of this year because of the substantial cuts that we took last year just in the administrative funding…”
Given that agencies have been operating under reduced administrative funding in fiscal years 2011 and 2012 and many say they are understaffed, asking senior HCV staff to spend a few days collecting financial information for the study and being interviewed by the study team, and asking frontline program staff to spend up to 15 minutes per day for 40 days (a total of 10 hours per staff person over the course of the study) responding to RMS (in addition to the time needed to be trained) constitutes a substantial imposition and one that will require the PHA to make up the time elsewhere.
Without compensation, many PHAs will be unwilling to participate in the study, which comes at a critical time for supporting future requests for HCV administrative funding. Through the first and second reconnaissance phase we have made contact with more than 130 PHAs to invite them to participate in the study. One of the first questions that PHAs ask about the study is how they will be compensated for the time they devote to the data collection effort. In the first phase of the reconnaissance, we had to contact a total of 90 agencies to get 59 to agree to be part of the study, and in the second reconnaissance phase thus far we have contacted 43 agencies thus far to get another 30 to agree to participate. The overwhelming reason that agencies give for declining to participate is that they do not have the staff time available to participate in the study.
Furthermore, the PHAs participating in the pretest phase of the study said that the compensation was very important to the success of the study, even if did not fully cover the staff costs incurred. As one of the agencies in the pretest put it, “In this climate of reduced admin funding, the compensation was an important consideration in deciding whether to participate in the study.”
Finally, the importance of compensating agencies for their participation in the study has also been an underlying theme in the meetings of the Expert and Industry Technical Review Group (EITRG) created for the study. The EITRG has consistently indicated that it would not be fair or efficient to expect PHAs to commit the substantial staff time necessary for this study without compensation.
In sum, we need to ensure a high participation rate among the high-performing PHAs we contact in order to ensure that the overall study has validity. Also, because of the extensive screening process required for eligibility in the study—which involves on-site data collection by the study team—it is extremely costly to the study for agencies that were validated in the reconnaissance phase of the study to decline to participate in the main study. Based on our experience recruiting agencies for the reconnaissance phase, feedback from the pretest, and input from the EITRG, we will not achieve a high rate of participation among in the study without compensation.
Rationale for the Compensation Amount
The level of compensation—$2,800 per PHA plus the equivalent of $300 per staff participating in RMS—was developed based on the amount of staff burden and the approximate staff cost associated with that burden. As shown in Exhibits A-2 and A-3, the level of compensation is equivalent to approximately $40 per hour for the time spent by the supervisory staff who will be assembling financial data and being interviewed and approximately $25 for the time spent by the line staff involved in random moment sampling. This amount would not fully cover the cost of the staff’s time, which would be higher than $40 or $25 including benefits, and is consistent with how studies routinely compensate survey respondents.2
Exhibit A-2. Activities and Proposed Compensation for Overall Study (Per PHA)
Activity |
Number of staff |
Time per Activity |
Total Time/PHA |
Cost per PHA @$40hr. |
Time spent preparing for and assembling financial data, information on program staffing, and producing transaction counts. |
2 per agency (may include fee accountant). |
12 hrs. of work per staff (fewer if additional staff participate). |
24 hrs. |
$960 |
Time spent meeting / speaking with study team staff |
2 per agency. |
24 hrs. of work per staff. |
48 hrs. |
$1,920 |
Total |
|
|
72 hrs. |
$2,880 |
Proposed compensation per agency for costs incurred from overall study |
|
|
|
$2,800 |
Exhibit A-3. Activities Required of All Staff Participating in RMS (Per PHA)
Activity |
Time per Response |
Number of Responses |
Total Time per Staff |
Cost per Staff @$25/hr. |
Training for PHA staff |
2 hrs. |
Once at start of data collection |
2 hrs. |
$50 |
Time measurement (RMS) |
1 min. |
12-15 “surveys” per day for 40 days = 600 minutes |
10 hrs. |
$250 |
Total |
|
|
12 hrs. |
$300 |
Proposed compensation per agency for costs incurred from staff participation in RMS |
|
|
|
$300 x the number of participating staff |
Payments to public agencies to offset the costs of study participation are not unprecedented, but usually take the form of grants. For example, for the recent Effects of Housing Vouchers on Welfare Families study, HUD entered into grant agreements with the participating PHAs and made payments to them of approximately $35,000 to $40,000 to offset administrative burden of participating in the evaluation. Compensating PHAs in the form of grants would not be cost-effective for this study because of the number of PHAs involved (50 to 60 in the main study) and the expense to the contract of administering these individual grants.
We do not think the proposed compensation amounts are so large as to place undue pressure on agencies to participate in the study, as they do not fully cover the staff costs associated with study participation. To quote from one PHA Executive Director on the EITRG: “The numbers you are suggesting seem eminently reasonable to me. Helping you with the study is not a part of any of their job descriptions, so they are displacing other work to help out.” The PHAs participating in the pretest assessed the compensation amount as follows: “The per person RMS data collection compensation seems appropriate. I don’t believe the one-time stipend fully compensates the housing authority for the costs of preparing the financial data and transaction counts. Even so, the overall amount of compensation is adequate to encourage accurate data collection.”
A9.3 Compensation for Participation in Telephone Interviews with Small Programs
We do not plan to compensate agencies for participating in the telephone interviews with small programs because of the lower level of effort required from the PHAs. Based on the pretest, we expect PHAs to spend 4 to 6 hours preparing for the interview. In addition, the interview itself will take about 2 hours to complete, with another hour of follow-up possible. The total burden for participating PHAs is therefore expected to be about 8 hours.
A10 Arrangements and Assurances Regarding Confidentiality
HUD’s contractor, Abt Associates, takes seriously the responsibility to protect the subjects they interview. Abt Associates’ Institutional Review Board (IRB) conducted an informal review of the project in December 2010 and determined that the project does not require further review by the IRB because it does not meet the definition of research under the federal human subject regulations. The purpose of the project is to develop an administrative fee formula, rather than to create generalized knowledge.
Participating PHAs will be notified that the information collected will be used for this study only and not for any other purpose.. In particular, the introductory letter and scripts will state that to ensure confidentiality, no personal identifying information from any interviews or other data collection conducted by Abt with PHA staff will be used in any report or data provided to HUD. The letter and scripts will also state that information collected will not be attributed to a single PHA or PHA staff person but will be presented in the aggregate and analyzed across a range of variables. Finally, the letter and scripts will also state that in the unlikely event that the study uncovers an area where the program is not being operated according to the regulations, HUD will notify the agency but will not take further action unless the agency continues to operate the program in violation of the regulations. Section 502 (g) of the Housing and Urban Development Act of 1970, 12 USC 1701z–2 provides the statutory basis for HUD to provide for confidentiality in its research activities.
A11 Sensitive Questions
The data collection instruments prepared for this study do not contain any sensitive questions, although detailed financial information will be collected. Interviewers will be trained to be sensitive to respondents’ concerns and to remind respondents that none of the information they provide to the research team during any phase of the study will harm or count against their agency in any HUD performance assessment or funding decisions.
A12 Estimate of Annualized Burden Hours
Exhibit A-4 provides information on the estimated time necessary to complete the data collection for the HCV Administrative Fee Study. Total burden for data collection for the study is estimated at 19,340 hours.
Exhibit A-4. Respondent Burden
|
A |
B |
C |
D |
E |
F |
G |
Data Collection Activity |
Number of Respondents |
Average Burden per Respondent (Minutes) |
Total Burden (Minutes) (A*B) |
Number of Responses per Respondent |
Total Respondent Burden (Minutes) (C*D) |
Total Burden per Respondent (Minutes) (B*D) |
Total Respondent Burden (Hours) (E/60) |
Preparation for RMS Data Collection |
60 (1 staff per site, 60 sites) |
60 |
3,600 |
1 |
3,600 |
60 |
60 |
Training for RMS Data Collection |
1,200 (20 staff per site, 60 sites) |
120 |
144,000 |
1 |
144,000 |
120 |
2,400 |
RMS Data Collection |
1,200 (20 staff per site, 60 sites) |
15 |
18,000 |
40 |
720,000 |
600 |
12,000 |
RMS Monitoring |
60 (1 staff per site, 60 sites) |
60 |
3,600 |
8 |
28,800 |
480 |
480 |
Cost Data Collection |
120 (2 staff per site, 60 sites) |
1,200 |
144,000 |
1 |
144,000 |
1,200 |
2,400 |
Transaction Count Data Collection |
120 (2 staff per site, 60 sites) |
480 |
57,600 |
1 |
57,600 |
480 |
960 |
Interviews with Small Programs |
260 (2 staff per site, up to 130 sites) |
240 |
62,400 |
1 |
62,400 |
240 |
1,040 |
Total |
|
|
|
|
|
|
19,340 |
A13 Estimated Record Keeping and Reporting Cost Burden on Respondents
There is no cost to respondents other than the time required to prepare for the interviews, complete the interviews, and participate in the time measurement activities.
A14 Estimated Cost to the Federal Government
The total contract amount for the full study is $5.1 million. Of this total, approximately $3.7 million will be used for the data collection activities described in this request.
A15 Reasons for Changes in Burden
This submission to OMB is a new request for approval; there is no change in burden.
A16 Tabulation Plan, Statistical Analysis, and Study Schedule
A16.1 Tabulation Plan and Statistical Analysis
We will use the information collected through the time measurement and PHA interviews to produce, for each PHA, an estimate of the total staff time per HCV program activity for regular HCV vouchers and select special programs. We will also estimate the total cost of that staff time taking into account salaries, benefits, overhead costs, and other direct costs. We will use the transaction count information to estimate the cost per voucher (or applicant) per year. The tabulation plan includes the following elements:
Estimating the cost per activity (e.g., the cost per annual inspection). The cost per activity is a function of the time it takes to conduct the task, the salary of the staff person doing the task, and the other costs (overhead and non-labor costs) allocated to the task. The total time per activity is calculated as the percent of RMS (which is equal to the percent of time) spent on each activity multiplied by the total number of hours worked. The estimate of the time per activity is derived by calculating the total time spent on the activity divided by the number of times the activity was conducted during the data collection period. In order to estimate the cost per activity, the time is multiplied by the fully loaded labor cost (salaries loaded with fringe, overhead and other costs) of the staff members performing the activity.
Estimating the total cost for each activity per unit under lease. This is the total cost for the activity for all times the activity is conducted per recipient per year. Some activities are generally only conducted once per recipient per year (e.g., annual recertification). But other activities may be conducted more than once for each unit under lease, such as interim recertifications, unit inspections, or calling people from the waiting list. In the case of inspections the total cost of the activity per unit would be the cost of inspecting all units associated with a recipient per year. The total cost for the activity will depend on the cost per activity and on the number of times the activity is conducted for each recipient. (In the inspection example, the number of inspections may vary depending on the quality of the housing stock, on the thoroughness of the inspection, and on agency policies regarding the number of times a unit can be inspected and whether the agency allows owner self-certification of repairs). Because all the agencies that are included in the cost study run high performing programs, we assume that their procedures are acceptable, so that the fact that different agencies conduct different numbers of inspections per unit is not a concern.
The total administrative cost per unit under lease. This will depend on the total costs of each activity and on the mix of activities in each agency. The total administrative cost per unit is the measure that will be used in the development of the formula. Again, because we assume that all agencies included in the study run high performing programs, we can use their costs to develop the formulas.
The analysis of time and cost per activity will include both descriptive tables and a model that seeks to isolate the role of program, local, and participant characteristics on administrative costs.
The descriptive tables will present the average and range of time and costs for each activity overall and for specific subtasks and for specific types of agencies/markets/vouchers/voucher holders. The level of disaggregation in presentation will depend on the level of disaggregation available in the data. For example, if the time measurement data can support it, we may be able to estimate the time per annual recertification for family households and for elderly households, in addition to estimating the average time overall per annual recertification. Similarly, if the data can support it, we may be able to estimate time per activity by type of voucher.
Analysis tables may look something like Exhibit A-5 for estimates of time per activity. Similar tables will be produced for estimates of costs per activity. Each table will include panels for each activity included in the analysis. Similar tables may present time and costs per activity by type of voucher. Because we expect that each PHA will serve households of all types, we expect to have observations for each of the 60 PHAs participating in the RMS data collection, for a total of 60 observations for each row – one observation per PHA.
Exhibit A-5. Time per Activity Overall and by Household Type
Activity |
Mean Time (minutes) per Activity |
Median Time (minutes) per Activity |
25th Percentile |
75th Percentile |
Conduct eligibility determination |
||||
Overall |
|
|
|
|
Elderly household |
|
|
|
|
Disabled non-elderly household |
|
|
|
|
Non-disabled family |
|
|
|
|
Homeless |
|
|
|
|
Conduct annual recertification |
||||
Overall |
|
|
|
|
Elderly household |
|
|
|
|
Disabled non-elderly household |
|
|
|
|
Non-disabled family |
|
|
|
|
Homeless |
|
|
|
|
In addition to providing descriptive tables, we will also use the data on cost per activity to identify factors that are associated with costs. This will be done using a regression model that estimates cost per activity as a function of PHA, recipient, and market characteristics. While the formula development model that estimates total costs per voucher as a function of PHA, recipient and market characteristics will rely only on data that are available for all PHAs, the models of costs for specific activities can use variables that are available only for study sites (such as the data collected from the interviews). This is because the cost per activity models are being used to try to estimate cost drivers, while the formula model needs to be applied to all programs—which means it must include only variables available for all programs.
Exhibit A- 6 presents a preliminary list of variables that might be included in the regression analysis of total costs, which will serve as a basis for a HCV administrative fee formula. Inclusion in the formula will be determined based on which of these variables have the largest impacts on program costs. Variables believed to have a substantial impact on program costs and that can be collected across all programs will be included in the regression analysis. In considering which variable to include in the regression analysis for the full cost study, we will need to be mindful of the relationship between the number of variables in the regression and the sample size needed to be able to detect effects. Also, the regression analysis can only include variables that are available for all programs, so that costs can be estimated for the sites not in the study.
Exhibit A-6. Program, Participant, and Market Characteristics that Could Affect Program Costs and Are Available for All PHAs
Issue/Area |
Variables |
Variable Categories |
Number and types of programs
|
|
|
Organizational structure |
|
|
Staffing |
|
|
Local environment |
|
|
Local housing market |
|
|
Local population characteristics |
|
|
The primary analysis of the telephone interviews with small programs will focus on the overall costs of operating a voucher program with fewer than 250 vouchers. This information will be used to determine how to adjust the HCV administrative fee formula for small PHAs and how the cost per voucher recipient compares to larger PHAs. Additional analysis will provide descriptive information on how smaller PHAs are able to operate the program under the current administrative fee structure. We will also use the data collected to analyze whether there is a minimum number of vouchers below which it is not feasible to manage the program on the administrative fee alone.
The telephone interviews with small programs will gather information on the total personnel costs of operating the voucher program, including contractor and temporary staff in addition to PHA staff. The interviews will also collect detailed information on the overhead costs and direct costs of the program, taking data reported to HUD’s Voucher Management System as a starting point. We will start with an overview of average total costs, personnel costs, and other costs by various program sizes of small PHAs, such as in Exhibit A-7. For comparison, costs for larger programs will also be presented. The table will have the following number of observations by row:
PHAs with 1-100 vouchers: 65 observations – from the interviews with small programs
PHAs with 101-249 vouchers: 70 observations – 65 from the interviews with small programs and 5 from the RMS data collection.
PHAs with 250+ vouchers: 55 observations – from the RMS data collection.
We will also analyze information on total costs per voucher recipient by program characteristics, voucher types, client characteristics, and local factors such as housing market characteristics and labor costs. This analysis will help determine whether small PHAs have different costs based on factors such as whether they are a stand-alone or combined program, whether they have special purpose vouchers, the clients they serve, and the local conditions in which they operate. Detailed cost per activity information will not be available for these agencies.
The costs per recipient of operating small programs will also be compared to the costs for larger programs. This information will be useful in determining how to incorporate small PHAs into the HCV administrative fee formula.
The sample sizes for the RMS component of the study and the telephone interviews with small programs are designed to achieve the maximum reliability of estimates within the available budget. The initial study design called for 30-40 PHAs participating in RMS data collection, but HUD determined that this was not a sufficient number an expanded the sample to 60 PHAs in the RMS study, plus 100-130 PHAs in the small program interviews. We expect a sample of 55-70 PHAs to provide reliable estimates as described in Section B.2.2.
Exhibit A-7. Costs per Recipient Overall and by Size of Small PHA Program
Activity |
Mean Cost per Voucher |
Median Cost per Voucher |
25th Percentile |
75th Percentile |
Total Costs |
||||
All PHAs < 250 vouchers |
|
|
|
|
PHAs 1 to 100 Vouchers |
|
|
|
|
PHAs 101 to 249 Vouchers |
|
|
|
|
PHAS with 250+ Vouchers |
|
|
|
|
Personnel Costs |
||||
All PHAs < 250 vouchers |
|
|
|
|
PHAs 1 to 100 Vouchers |
|
|
|
|
PHAs 101 to 249 Vouchers |
|
|
|
|
PHAs with 250+ Vouchers |
|
|
|
|
Direct and Overhead Costs |
||||
All PHAs < 250 vouchers |
|
|
|
|
PHAs 1 to 100 Vouchers |
|
|
|
|
PHAs 101 to 249 Vouchers |
|
|
|
|
PHAs with 250+ Vouchers |
|
|
|
|
A16.2 Study Schedule
Under the current study schedule, the data collection will be conducted over a 70-week period, starting in November 2012 and continuing through February 2014. This schedule assumes that OMB approval is received by September 30, 2012.
A17 Expiration Date Display Exemption
All data collection instruments will prominently display the OMB control number and the expiration date for OMB approval.
A18 Exceptions to Certification
This submission describing data collection requests no exceptions to the Certificate for Paperwork Reduction Act (5 CFR 1320.9).
1 Smith, Mark W. and Paul G. Barnett, “Direct Measurement of Health Care Costs,” Medical Care Research and Review, Vol. 60 No. 3 (Supplement to September 2003), 74S-91S.
2 Household survey efforts typically pay respondents $20-$25 for completing a 30-45 minute survey online or by phone, which translates to $40-$50 per hour of the respondent’s time. Examples of recently-approved household surveys include the CDBG Disaster Assistance Study survey ($25 for 45 minutes) and the Housing Counseling Outcomes Study survey ($20 for 30 minutes).
Part
A: Justification ▌pg.
| File Type | application/msword |
| File Title | Abt Single-Sided Body Template |
| Author | Abt Associates Inc. |
| Last Modified By | Marina L Myhre |
| File Modified | 2012-11-19 |
| File Created | 2012-11-19 |