APPENDIX: The Basis for the Estimate that the Budget Would Support 370,000 Fewer Vouchers in 2010
February 18, 2005
This analysis concludes that the level of funding proposed in the President's budget for low-income housing vouchers in 2010 would support 370,000 fewer vouchers that year than are being funded in 2005. The Administration may argue that this conclusion is unjustified because the President's budget does not provide specific information about the proposed funding level for the housing voucher program in 2010 and because our estimate of the reduction in the number of vouchers does not rely on the Administration's baseline estimate for the housing voucher program. While such assertions would be correct, they would not be relevant to the validity of this analysis, which uses appropriate methods to determine the amount of funding for housing vouchers in 2010. These methods are fully consistent with the President's budget proposals for discretionary spending in 2010 and with the amount that would be needed to provide the same number of vouchers in 2010 as are being provided in 2005.
Funding Assumed in the President's Budget for Housing Vouchers in 2010
For the first time since 1989, the President's budget provides no information about proposed funding levels for individual discretionary programs or budget accounts for years beyond the coming fiscal year (2006). The budget does, however, propose overall levels for discretionary funding for each year through 2010 (as well as statutory caps to limit appropriations to these levels). In addition, information provided in the budget, in conjunction with supplementary information on the budget that the Office of Management and Budget has provided to the Congressional budget committees, shows the amount that the budget assumes for each budget "function" (large groupings of programs with similar policy ends) and each budget "subfunction" (smaller groupings of programs within each budget function) for each year through 2010. 
In this analysis, we have taken the priorities for the allocation of funding within each budget subfunction that the President's budget sets for 2006 and assumed these priorities would be maintained through 2010. In other words, the share of the funding in a budget subfunction that would go to each program in that subfunction is assumed to remain the same in 2007-2010 as the share that the President's budget proposes for each program in 2006. (The Administration may argue this does not represent Administration policy, but in the absence of any information from the Administration about individual program funding after 2006, it is the most appropriate way to assess how the changes the budget shows in funding levels for each subfunction after 2006 would affect the programs within the subfunctions. The Administration may argue that a particular program would be more favored after 2006 than in 2006, but if so, the Administration would need to acknowledge that the share of funding going to another program within the same subfunction would have its share of the subfunction reduced.)
The Administration also may contend that overall funding levels shown for discretionary programs by budget function and subfunction for years after 2006 are meaningless because they are simply "formulaic." The Administration made that argument last year when confronted with the cuts in discretionary programs in years from 2006 through 2009 that were shown in the supplementary OMB budget documents issued in conjunction with the President's fiscal year 2005 budget. A memorandum sent last spring from the Director of OMB to federal agency heads, however, indicated that this argument was not valid. As the Washington Post reported last May, the OMB memorandum specifically directed agency heads to adhere to the discretionary program levels for fiscal year 2006 shown in the OMB budget documents when agencies prepared their budget for 2006. 
At this early date, no such "smoking memo" has been discovered that would disprove an Administration claim that the function and subfunction discretionary funding levels reflected in the new budget documents are merely "formulaic" and do not represent Administration policies. Such a claim, however, does not stand up well under scrutiny. The amounts that the new OMB budget documents show for each function and subfunction do not appear to represent any discernible formula.
For instance, the level of discretionary funding that the budget documents show for the Natural Resources and Environment function (function 300) drops by 2 percent (in nominal terms) between 2006 and 2010, but the funding level shown for General Science, Space, and Technology (function 250) increases by 5.8 percent. Even within a single function, the path of funding varies. Funding for Conduct of Foreign Affairs (subfunction 153) declines by 4 percent between 2006 and 2010, while funding for International Development and Humanitarian Assistance (subfunction 151) climbs by 18 percent. Perhaps most tellingly, funding for Other Advancement of Commerce (subfunction 376) suddenly more than doubles (from $3.4 billion to $7.3 billion) between 2009 and 2010, presumably to accommodate the increased funding the Census Bureau will need to carry out the 2010 census. It is hard to imagine a formula that could produce these results.
Amount Needed in 2010 to Fund the Same Number of Vouchers as in 2005
After estimating the level of funding assumed in the President's budget for the housing voucher program in 2010, we calculated the effect that this funding level would have on families that rely on vouchers for affordable housing. We did not use OMB's baseline projections for this purpose because, in the case of housing vouchers, the baseline does not accurately reflect the cost of maintaining the current level of activity. There are two reasons for this, both related to the baseline rule that tells OMB and the Congressional Budget Office to set baseline discretionary funding levels for future years by assuming that the same level of funding provided for each program in the current year will be provided for that program in future years, with an adjustment only to account for overall inflation.
The first reason that the OMB baseline is not the appropriate standard to use here relates to how the baseline treats rescissions of funds appropriated for a prior year but never spent. The baseline amount for a program in a future year is the net level of funding for that program in the current year, adjusted for inflation. For example, if $2 billion is appropriated for a program in 2005 and a $1 billion rescission of funds provided for that program in a prior year also is enacted in 2005 the net funding for the program in 2005 will be considered to be $1 billion. Yet $2 billion will actually be made available for program activities, and $2 billion, adjusted for inflation, is the amount that will be needed in future years to maintain the same level of services under the program. If the operations of a program are such that rescissions of roughly the same amount can be made every year, using a baseline projection that effectively assumes that the same amount of unused funding from prior years will be rescinded in every future year will not present a problem; the future rescissions will partially offset the future appropriations needed to maintain program services and will bring the program's net funding down to the level that the baseline assumes. But if future rescissions are not possible because there will be no unused funds from prior years left to rescind, then this way of setting the baseline will lead to an understatement of the cost in future years of maintaining services in the program.
The housing voucher program represents a case where the baseline rule under which rescissions made in the current year are mechanically assumed to be repeated in every future year leads to an understatement of future funding needs. The net funding level in 2005 for Section 8 housing assistance (including both the voucher program and a separate "project-based" housing assistance program) is $18.5 billion, but this includes a rescission of prior-year funding of $1.6 billion. In other words, the actual amount made available for Section 8 in 2005 is $20.1 billion, but the net funding level is considered to be $18.5 billion because of the rescission of $1.6 billion of unused prior-year funds. OMB's baseline simply projects the $18.5 billion, with an adjustment for inflation, into future years.
This means that the baseline, which is supposed to represent the level of funding needed to continue providing the same services in future years, implicitly assumes that large rescissions will be made in every future year. However, according to the President's 2005 budget, no unspent Section 8 funds from prior years will be available to rescind after 2008. Without the rescission that the OMB baseline implicitly assumes will occur again in 2010, the $20.4 billion in net funding that the OMB baseline includes for Section 8 in 2010 will fall well short of the amount needed to support the same number of vouchers as in 2005.
The second problem with using the conventional baseline in assessing voucher funding needs in 2010 is that the baseline rules call for using two standard measures of inflation in projecting discretionary spending needs. The employment cost index is used to inflate amounts used for federal personnel costs. The GDP price index is used to inflate all other amounts, including funding for the housing voucher program. The Administration estimates that the GDP price index will increase at an average rate of 2.1 percent a year between 2005 and 2010. According to CBO, however, the average cost of a housing voucher will go up somewhat faster than that, at an average annual rate of about 2.7 percent. The reason for this difference stems largely from the growing gap between market rents and the incomes of low-income families.
Because OMB follows the baseline rules in assuming that large rescissions of prior-year funding can be achieved in every year and also uses the GDP price index to inflate 2005 funding levels, its baseline projection of funding for the housing voucher program in 2010 substantially understates the level of funding that actually will be needed in 2010 to provide the same number of vouchers as in 2005.
This analysis uses an estimate of the amount needed in 2010 that is calculated in a way much more appropriate for the housing voucher program. To do that, we calculated how much it would cost to provide the same number of vouchers in 2010 as were funded in 2005. Based on an analysis of HUD lists of the funding levels provided to state and local housing agencies in 2005 and a separate HUD database that contains data submitted by housing agencies on the number of vouchers they administer that were in use and the cost of those vouchers, we estimate that HUD provided housing agencies with renewal funding for 1,949,458 vouchers for fiscal year 2005 and that the average cost per voucher (including the cost of the voucher subsidy, but not the fee paid to the state or local agency administering the voucher to cover administrative costs) at which these vouchers are being funded in 2005 is $6,805. HUD estimates that an additional 23,813 new vouchers will be issued in 2005 to households that lose assistance under other federal housing assistance programs. These "tenant-protection" vouchers will be funded out of a separate tenant protection voucher account in 2005, but after 2005, they must be funded under the regular voucher program.
To calculate the level of funding needed to provide the same number of vouchers in 2010, we took the $6,805 per-voucher cost in 2005 and inflated it by 14.1 percent (about 2.7 percent per year), the Congressional Budget Office estimate of the rate at which costs per voucher will grow from 2005 to 2010. This results in a 2010 cost per voucher of $7,764. We then multiplied the 23,813 tenant protection vouchers that will be issued in 2005 by 96 percent, which is the percentage of authorized vouchers in use in 2004. We added the result (22,860 vouchers) to the 1,949,458 vouchers that otherwise will be funded in 2005. This produces a total of 1,972,318 vouchers that will need support in 2010 to maintain the same number of vouchers.  We multiplied the 1,972,318 vouchers by the estimated per-voucher cost of $7,764 in 2010, producing a total funding need of $15.313 billion. This is the estimate of the funding needed in 2010 to support the same number of vouchers as are being funded in 2005.
This funding level was compared to the amount that the Administration's budget is estimated to include for voucher subsidies in 2010 (based on the funding level that the budget documents show for the housing assistance subfunction in 2010 and the proportion of funding in that subfunction that would go for voucher subsidies in 2006 under the President's budget). The difference between the $15.313 billion funding need and this amount is the shortfall in voucher funding in 2010 under the budget. We divided the shortfall by the projected cost of $7,764 per voucher in 2010 to estimate the number of vouchers being funded in 2005 for which funding would not be available in 2010.
 See Sharon Parrott, Isaac Shapiro, David Kamin, and Ruth Carlitz, "Unpublished Administration Budget Documents Show Domestic Cuts Would Significantly Reduce Funding for Most Public Services," Center on Budget and Policy Priorities, revised February 14, 2005. This analysis, as well as backup data and a detailed description of the methodology.
 See, David Kamin, Richard Kogan, and Bob Greenstein, " Administration Memo Confirms Plans for Budget Cuts in Many Domestic Discretionary Programs in 2006," Center on Budget and Policy Priorities, Revised June 1, 2004.
 Each year (even when adequate funding is available) some authorized vouchers are not used, often families issued a voucher are unable to find an apartment where they can use their voucher. Because of inadequate funding, the 1,949,458 renewal vouchers funded in 2005 is significantly below both the number of vouchers actually in use according to the latest HUD data used in our analysis (covering May-July 2004) and 96 percent of authorized vouchers requiring renewal, so we did not further reduce this number based on CBO's estimated 2010 utilization rate of 96 percent.