Revised July 25, 2003
ESTIMATING THE SHORTFALL IN REQUESTED VOUCHER FUNDING FOR FISCAL YEAR 2004
By Barbara Sard and Will Fischer
PDF of report
HUD Response To Center Report Fails To Refute Finding That Administration's Budget Request Would Result In Large Reduction In Number Of Low-Income Families With Housing Vouchers
If you cannot access the files through the links, right-click on the underlined text, click "Save Link As," download to your directory, and open the document in Adobe Acrobat Reader.
This paper explains the data sources and methodology used to estimate the shortfall in voucher funding for fiscal year 2004 under the Administration’s budget request and the FY 2004 appropriations bill approved by the House VA-HUD Appropriations Subcommittee on
July 15, 2003. These sources and methods are used in two Center analyses: “New HUD Data Show Families Will Likely Lose HousingVouchers if Congress Approves President’s Budget Request,” available on the internet at https://www.cbpp.org/7-11-03hous.htm, and “Funding Level Approved by House Subcommittee Would Reduce, but Not Eliminate, Shortfall In HousingVoucher Funding,” available on the internet at https://www.cbpp.org/7-21-03hous.htm.
Projections of voucher utilization rates, per-unit costs, and total program costs in these analyses are based on data that HUD collected from 2,449 state and local housing agencies on or before
April 9, 2003concerning the operations of their voucher programs in the six months from August 2002 through January 2003. We rely on other data sources to make certain adjustments to the data collected from individual agencies, as discussed below.
Projecting the Leasing Rate of Authorized Vouchers
In making projections of the average number of vouchers (or “units”) leased during fiscal year 2003 and the number under lease in October 2003, we assume that the average monthly rate of change in the number of units leased for each individual agency from the first quarter of that data (August 2002 – October 2002) to the second quarter of the data (November 2002 -
January 2003) continues through October 2003. For example, if an agency leased an average of 200 units per month from August 2002 through October 2002 and an average of 203 units per month from November 2002 through January 2003, the agency’s number of units leased would have increased by three units from the first quarter to the second quarter. If this rate of increase (three units each quarter or one unit each month) continued, this agency would have 213 units under lease by October 2003.
The only exceptions to this method were applied to agencies that leased (or were projected to lease) more than their contracted number of vouchers. (Each state or local agency administering housing vouchers has one or more contracts with HUD for a specified maximum number of vouchers. We relied on housing agencies’ reports of their contracted number of vouchers in each month in the HUD data.) For purposes of calculating the average number of vouchers leased during federal fiscal year 2003, if either the agency’s number of units leased during the portion of fiscal year 2003 covered by the HUD data (October 2002 –
January 2003) or the projected average number of units leased for the full fiscal year was above the agency’s number of contracted units, the agency was counted as having only the contracted number of units leased. In other words, no leasing rate above 100 percent for an individual agency was included in our projected leasing rate for fiscal year 2003. This was done because the fiscal year 2003 appropriations act does not permit agencies to receive annual renewal funding for more than their contracted number of vouchers. (Agencies are permitted to use their reserve funds to support such additional vouchers while they reduce the number of households in their programs through normal turnover.) Similarly, for purposes of projecting the number of units under lease in October 2003, agencies whose projected number of vouchers leased was above their number of contracted vouchers were counted as having only the contracted number of units leased.
We made one additional adjustment in the projected leasing rates for agencies that, in one or more months of the reported data, had leased more than their contracted number of vouchers but reduced their leasing rate after that point. Agencies that temporarily lease more units than they are authorized to administer can prevent their number of units leased from exceeding 100 percent of their authorized vouchers over a full year if they reduce their utilization rates below 100 percent in subsequent months. It is unlikely that such declines from over 100 percent to under 100 percent represent long-term trends that will continue once the number of vouchers in use has been brought into balance with the authorized number. To avoid artificially extending such trends, our analysis projected the number of vouchers under lease in October 2003 at agencies that had been reducing the number of vouchers under lease (after leasing at a level in excess of their authorized level during the August 2002 though
January 2003 data period) as being the higher of the number of units calculated for these agencies for October 2003 using the rate-of-change methodology described above or the average monthly number of units projected to be under lease at these agencies for fiscal year 2003 as a whole. Here, as well, no agency was counted as having more than 100 percent of its contracted number of units under lease in October 2003.
It is important to note that we assume no further increase in voucher utilization after the first month of fiscal year 2004. This is a conservative assumption. It is likely that agencies would be able to use a portion of the approximately 95,000 vouchers that we assume will remain unleased throughout fiscal year 2004 if they had funds available to do so.
Estimating Voucher Costs
We estimated per-voucher costs for February through September 2003 by projecting forward each agency’s per-voucher costs (including both housing assistance payments made by agencies to landlords and administrative fees that agencies earn from HUD for each month a voucher is leased), based on the national rate of change in per-voucher costs from the first quarter of the HUD data (August 2002 – October 2002) to the second quarter of the data (November 2002-
January 2002). Units leased beyond an agency’s monthly contract level were excluded from the calculation of this national per-unit cost trend. This method projected an increase in per-voucher costs of 5.4 percent during the last eight months of fiscal year 2003. Using this method, we calculated a national average per unit cost of $6,691 for fiscal year 2003. (This estimated increase in costs per voucher for the last eight months of fiscal year 2003 is consistent with the increase in HUD’s fair market rents, which determine how much rent a voucher can cover, between fiscal year 2002 and fiscal year 2003. The national average FMR rose by 6.3 percent from fiscal year 2002 to fiscal year 2003. The methodology for computing the national average FMR from HUD’s FMR data is described below.)
To reflect expected changes in rental costs between fiscal year 2003 and fiscal year 2004, per-unit costs for 2004 then were calculated by adjusting the 2003 national average estimated cost per-unit, based on the percentage increase nationally between the FMRs for 2003 and HUD’s proposed FMRs for 2004. (HUD issued the proposed FMRs for 2004 on
May 28, 2003; see 68 FR 31870.) This was done by calculating a weighted average state FMR for each state for both 2003 and 2004 by weighting the local FMRs based on the number of rental households in the FMR area, according to the 2000 census. We then determined the national FMR for each year by weighting the state FMRs by the number of vouchers under contract to PHAs in each state. For purposes of these calculations, the total number of vouchers in each state is based on HUD data for individual housing agencies from the HUDCAPS system, accessed November 25 -27, 2002, and available at https://pic.hud.gov/pic/haprofiles/haprofilelist.asp. This comparison yields a 2.5 percent annual increase in fiscal year 2004 average rental costs compared with 2003, for a national average per-voucher cost in fiscal year 2004 of $6,870.
Leasing and Cost Rates for Vouchers Not Reported in the HUD Data
The HUD data used in our analysis were reported to HUD by 2,449 housing agencies with a total of 1,942,206 contracted vouchers in
January 2003. (The January contracted level exceeded the December level by about 21,500 vouchers. The increase is likely due to HUD’s contracting for a portion of the incremental and tenant protection vouchers funded in the fiscal year 2002 appropriations act.) Omitted from these data are approximately 124,000 vouchers under contract to approximately 100 additional agencies. We determined the number of omitted vouchers by comparing the 1,942,206 contracted vouchers reported to the estimated total of 2,066,000 authorized vouchers in fiscal year 2003. (This estimate does not include the additional 43,000 tenant protection vouchers newly funded by the fiscal year 2003 appropriations act, which was enacted after the period covered by the HUD data. These tenant protection vouchers are discussed below.) The 124,000 omitted vouchers break down into two groups: the 94,955 vouchers administered by 16 agencies in the Moving to Work demonstration (MTW) and 29,000 “other” missing vouchers likely administered by the remaining non-reporting agencies. The steps used to estimate leasing and cost rates for the MTW and other missing vouchers are discussed below.
In a cover memo accompanying the data, HUD stated that the data did not include reports from agencies in the Moving to Work Demonstration (MTW). We compared the list of agencies authorized to participate in MTW with the agencies included in the HUD data and identified 16 MTW agencies omitted from the data compilation. For these 16 agencies, we considered all of their 94,955 contracted vouchers to be in use and requiring funding, as they receive funding on this basis under their MTW contracts. (We used other available data to determine the number of contracted vouchers for these agencies and their funding level in fiscal year 2004.)
For the remaining 29,000 authorized vouchers for which no data were available, we applied the national average leasing rate and national average cost per voucher derived from the data HUD reported (without the additional MTW vouchers). That is, for fiscal year 2003, we assumed that 95.7 percent of the missing vouchers, or 27,753 vouchers, would on average be leased during the year at an average cost of $6,691 per voucher (including administrative fees). For fiscal year 2004, we assumed that the leasing rate of 96.6 percent that we projected from the HUD data as described above, or 28,014 additional vouchers, would be leased in October 2003, at the average FY 2004 projected per voucher cost (without the MTW vouchers) of $6,870.
Determining the Funding Required to Renew Vouchers in Use in Fiscal Year 2004
In addition to projecting forward the leasing and cost trends for the more than 1.9 million vouchers covered by the HUD data and adding leasing and cost figures for the 124,000 omitted vouchers, as described above, we made two other adjustments to determine the number of vouchers requiring renewal funding in fiscal year 2004 and their likely cost.
We assumed that half of the 43,000 tenant protection vouchers initially funded by Congress in the fiscal year 2003 appropriations act would be leased by October 2003 and require renewal funding for fiscal year 2004. HUD awards tenant protection vouchers to housing agencies when such vouchers are needed to replace other federal housing assistance that is no longer available. No information is currently available concerning awards of newly-funded tenant protection vouchers. (HUD usually does not publish notice of the funding awards until months after the additional vouchers have been awarded to local agencies.) It is likely that HUD has already contracted with housing agencies to administer some of these vouchers and that agencies have issued them to families that have begun to use them to cover their rent payments. Most tenant protection vouchers are issued to tenants who can use them in the previously federally-subsidized apartments in which they currently reside, enabling the lease-up process to occur quickly. Our assumption that 21,500 of the fiscal year 2003 tenant protection vouchers will be leased in October 2003 could be too low or too high. While it may overstate the number actually leased in October, it is likely to understate the number leased on average during fiscal year 2004.
For the cost of the 21,500 additional tenant protection vouchers that we include in our projection of vouchers leased in October 2003, we use the national average cost per voucher for 2004 derived from the HUD data, $6,870 (without the additional MTW vouchers). Because we do not know which agencies will receive these vouchers, we could not use agency-specific cost data.
Finally, we reduced the total number of vouchers likely to be in use in October 2003 (2,020,000) and their total fiscal year 2004 cost ($13.88 billion) by the number and estimated cost of the vouchers that do not need renewal funding because they continue to be funded under multi-year appropriations still available in fiscal year 2004. The Congressional Budget Office’s March 2003 baseline indicates that 9,501 authorized vouchers are in this category. Lacking information on which agencies administer these multi-year-funded vouchers, we assumed they had the same average cost of all other fiscal year 2004 vouchers ($6,871, including administrative fees).
To estimate the shortfall in renewal funding requested by the Administration, we subtracted the $13.81 billion needed, based on these calculations of the number of vouchers likely to be in use in October 2003 and their likely cost, from the Administration’s request of $12.55 billion. (Endnotes 1 and 7 to the paper analyzing the shortfall in the Administration’s budget quest explain how we determined that the Administration’s budget results in a request for $12.55 billion for voucher renewals.) The resulting shortfall is $1.26 billion.
To estimate the shortfall in renewal funding that would be provided under the bill approved by the House subcommittee, we subtracted the $13.81 billion in funding needed from the $13.23 billion that the House bill would make available for voucher renewals. (See endnote 1 of the paper concerning the House bill.) The resulting shortfall under the House bill is $583 million.
Tables 1 and 2 summarize the elements of our analysis for fiscal years 2003 and 2004, respectively. (Note that some of the figures provided are rounded, so calculations used to determine the average cost and percentage of vouchers in use cannot be reproduced precisely from figures in tables.)
Fiscal Year 2003 Voucher Estimates
Number of Vouchers in Use
Percent of Contracted Vouchers in Use
Average Cost (including fees)
Total Cost of Vouchers in Use Based on Individual Agency Costs Available
Data from 2,449 Agencies Compiled by HUD
Used varying totals for individual months averaging 1,936,288
16 MTW Agencies
Other Missing Vouchers
Vouchers with Multi-Year Funding
Total Renewal Funding Needed
Fiscal Year 2004 Voucher Estimates
Number of Vouchers in Use in October 2003
Percent of Contracted Vouchers in Use
Average Cost (including fees)
Total Cost of Vouchers in Use Based on Individual Agency Costs Available
Based on Data from 2,449 Agencies Compiled by HUD
16 MTW Agencies
Subtotal Based on Available Data
Other Missing Vouchers
FY 03 Tenant Protection
Vouchers with Multi-Year Funding
Total Renewals and Renewal Funding Needed
 These data were only available in paper format and were inputted manually to enable them to be used for agency-by-agency analysis. As a result, data entry errors may have occurred in a small number of cases. When the data are totaled across all agencies for each category of data in each month, the totals deviate from the totals provided in the paper copies of the data by between 0.000% and 0.027%. In some cases, agencies may have made errors in reporting their data. Where it appeared that the agency had erroneously overstated the number of vouchers leased, we used the authorized number of vouchers as the number leased.
Agencies reported fees earned based on HUD’s fiscal year 2002 fee schedule. HUD published the FY 2003 fee schedule on
May 6, 2003(68 FR 24078), after the data analyzed in this paper were reported to HUD. HUD will make appropriate increases in fees earned during federal fiscal year 2003 at the end of agencies’ fiscal years. It also is possible that HUD will rely on provisions of the fiscal year 2003 appropriations act to reduce the fees earned by some agencies that are determined to have “excess” fees remaining from prior years. A lawsuit challenging these provisions was filed on July 11, 2003, so it is unclear whether or when they will take effect.
 Administrative fees accounted for less than 1 percent of the dollar increase in per voucher costs that occurred between August 2002-October 2002 quarter and the November 2002-January 2003 quarter.
 After adjustment for a portion of the vouchers missing from this analysis funded on a basis other than actual costs, as discussed below, we estimate the national average per-voucher cost for fiscal year 2003 at $6,692 and for fiscal year 2004 at $6,871.
 We estimated that a total of 2,066,000 vouchers were authorized in 2003 (without new FY 2003 tenant protection vouchers) as follows. In its budget documents for FY 2003, HUD states that 2,077,332 vouchers expired and needed renewal in FY 2003. HUD staff subsequently informed us that this number overstated the number of renewals required by about 24,000 vouchers. See
Barbara Sardand Will Fischer, “President’s Budget Requests Insufficient Funding for Housing Vouchers in 2004,” Center on Budget and Policy Priorities, revised April 24, 2003, available on the internet at < http://www.centeronbudget.org/3-27-03hous.pdf>.) We reduced the number of expiring vouchers by 30,000 based on: 1) this representation regarding the overstatement of 24,000 vouchers; and 2) to account for approximately 6,400 expiring Section 8 moderate rehabilitation subsidies for which renewal funding was requested as part of overall voucher funding in fiscal year 2003. (Housing agencies did not include these so-called Mod. Rehab. subsidies in the data reported to HUD, and for fiscal year 2004, HUD has requested funding for these subsidies in a new Project-based Rental Assistance account.) To this adjusted number of expiring vouchers (approximately 2,047,000), we added the 18,827 vouchers that, according to the Congressional Budget Office March 2003 baseline, were funded under multi-year contracts in 2003.
 Some MTW agencies receive voucher funding under a different formula than other agencies, while others receive funding under the same rules. Those that receive funding under different rules were not required to submit data under the process HUD used to collect data for the April 9 compilation.
 The 16 MTW agencies omitted from the HUD data reports are Tulare County (CA),
State, New Haven, Delaware , Lawrence/Douglas (KS), Chicago , Cambridge (MA), Lincoln (NE), Keene (NH), High Point (NC), Portage (OH), Portland (OR), Louisville , Philadelphia , Pittsburgh and Vancouver (WA). Based on review of these agencies’ MTW documents, other information available from HUD and conversations with agency staff, we confirmed that it is likely these 16 agencies receive funding for all of their contracted vouchers (and not only the number in use as tenant-based assistance), adjusted each year for inflation rather than changes in actual costs. We derived these agencies’ number of contracted vouchers from HUD’s website at https://pic.hud.gov/pic/haprofiles/haprofilelist.asp in July 2003, and made the assumption that all of these vouchers had been awarded by the beginning of fiscal year 2003. We adjusted the annual budget authority required by the MTW agencies based on the applicable annual adjustment factor (AAF) determined by HUD. Because no AAFs are currently available for fiscal year 2004, we adjusted budget authority from fiscal year 2003 to fiscal year 2004 using the lower of the 2003 AAF or the percentage increase in the applicable proposed fair market rent. Seattle
 We calculated the total number of vouchers likely to be under lease in October 2003 by adding the figures for each of the categories of vouchers according to the applicable methods discussed above. See the first column of Table 2 below, “Number of Vouchers in Use in October 2003.”
 See footnote 8 and the last column of Table 2, “Total Cost of Vouchers in Use Based On Individual Agency Costs Available.”
 This figure represents the average number of contracted vouchers in fiscal year 2003 based on the HUD data plus other sources as explained above. If instead the January 2003 reported level of contracted vouchers were used to derive the total, an additional 5,918 vouchers would be included for a total of 2,047,294, consistent with note 5.
 This column applies to October 2003, the first month of the fiscal year. To be conservative, our analysis assumes no further increase in utilization during fiscal year 2004.
 Based on HUD’s estimate that 2,106,233 vouchers expire in FY 2004 and CBO’s figure of 9,501 vouchers with multi-year funding, the total number of contracted vouchers in FY 2004 should be 2,115,773, or about 6,600 more than our calculation shows. HUD may have assumed that more incremental or tenant protection vouchers funded in fiscal year 2002 would have been awarded in fiscal year 2003 and need renewal in fiscal year 2004 than this table reflects. To date, HUD has announced the award of only 16,460 of the 25,900 fiscal year 2002 incremental vouchers funded by Congress. It is also possible that HUD’s FY 2004 expiring voucher figure continues to include Section 8 moderate rehabilitation subsidies, despite HUD’s proposed transfer of these subsidies to a new project-based rental assistance account. A further possibility is that more vouchers are unaccounted for than we estimated in the “missing voucher” category for both fiscal years 2003 and 2004. If the explanation is the last of these, our estimates of renewal funding needed and the shortfalls in the Administration’s budget request and the House bill would be approximately $40 million too low.
 The numbers in this column would produce a total number of expiring vouchers that is about 9,000 fewer than the number estimated by HUD. Our analysis uses HUD’s figure of 2,106,233 expiring authorized vouchers in fiscal year 2004.
 This figure is slightly higher than the one used to determine that there was a shortfall of $1.26 billion in voucher renewal funding under the Administration’s request, due to rounding.