Sharp Expansion of HUD’s “Moving-To-Work” Demonstration Raises Serious Concerns
A proposal before Congress would sharply expand HUD’s Moving-to-Work (MTW) demonstration. Unless important limitations are added, this expansion would reduce the number of families receiving housing assistance by shifting funds out of the Section 8 housing voucher program. It also would expose more low-income families to risky policies than is necessary to test innovative approaches, and allow local policies to diverge to a degree that could weaken housing assistance programs.
The proposal is part of the Section 8 Voucher Reform Act (SEVRA), which overall is an important, well-crafted measure containing significant improvements to the voucher program and other federal rental assistance programs; the House Financial Services Committee approved the bill in July 2009. SEVRA includes a provision that would allow up to 80 state and local housing agencies to participate in the Moving-to-Work demonstration program (compared to 30 today) and rename it the Housing Innovation Program (HIP).
Despite its name, MTW is not focused primarily on supporting employment. MTW allows HUD to grant sweeping waivers of a wide range of federal statutes and regulations to agencies that administer voucher and public housing programs in order to test experimental policies. It also allows HUD to establish special funding formulas for MTW agencies and to permit them to shift funds between the voucher program and public housing.
SEVRA’s MTW provision would constitute a fundamental and far reaching change to federal housing policy, as the expanded demonstration could affect close to 1 million vouchers and public housing units — about 30 percent of the total nationally. Proponents argue that expanding MTW would not only help test innovative policies, but also support revitalization of public housing and provide housing agencies with needed flexibility. These are worthy goals, and some current MTW agencies have implemented promising, responsible initiatives to pursue them. A sweeping MTW expansion, however, could cause considerable harm, and each of these goals could be achieved through less disruptive means.
- MTW has caused fewer families to receive assistance per dollar of federal funding and has shifted funds away from housing agencies that are not in the MTW demonstration. Using special MTW funding agreements and the flexibility that MTW gives them to transfer funds from vouchers to public housing, MTW agencies have added resources to renovate and replace public housing developments and for other affordable housing development efforts.
At some agencies, however, these added investments have come at the expense of helping as many needy families as possible with available funding. MTW agencies left unused about 30,000 vouchers in 2008 that they could have used to assist needy families because they shifted the funds to other purposes. These shifts did not result in an offsetting increase in families assisted with other forms of assistance. Overall, the five largest MTW agencies assisted about eight families per $100,000 in federal funding they received, while non-MTW agencies on average helped about 15 families with the same amount. If new agencies admitted to MTW under the proposed expansion allocated funds in the same manner as current agencies, tens of thousands of additional families could be left without assistance.
In addition, the relatively rich subsidies that MTW agencies receive have often resulted in lower funding for non-MTW agencies.
- MTW allows HUD to waive key program rules that make housing assistance programs more effective and protect low-income families. Research has shown that vouchers are highly effective in reducing homelessness and housing instability — hardships that have serious harmful effects on the development of young children. However, MTW permits agencies to implement policies that risk scaling back these benefits, including rent increases and time limits that cut off assistance even for working-poor families who cannot afford housing on their own.
- A comprehensive national policy would preserve public housing more effectively. Preserving high-quality public housing should be a federal priority. But policymakers can best achieve it through a comprehensive national approach that includes added resources and policy changes, not through special agreements with selected local housing agencies. In the past year Congress has already taken important steps in this direction, providing $4 billion to renovate public housing in the February 2009 economic recovery package and full (or nearly full) funding for public housing operating subsidies in 2010 appropriations legislation.
- Where added flexibility is appropriate, Congress should provide it through specific changes applying to all agencies, not sweeping deregulation for a select few. Added flexibility in some elements of the housing assistance programs would enable agencies to better meet local needs without weakening protections for low-income families or creating major program differences from one community to another. When new flexibility is appropriate, Congress should provide it to all agencies, rather than limiting it to those in MTW.
- A more modest expansion of MTW would be adequate to test innovative policies. The sweeping MTW expansion in SEVRA is far larger than is needed to test policy alternatives and could strain HUD’s capacity to conduct effective evaluations. The 1 million units that could be affected by experimental policies under the MTW expansion are more than 200 times as many as were involved in Jobs Plus or Moving-to-Opportunity, two rigorous, multi-city housing policy demonstrations that succeeded in generating a wealth of concrete policy findings. It is valuable to permit policy experimentation, but where policies that carry risks for vulnerable families are involved, experiments should not be carried out at a scale that far exceeds what is needed to generate useful findings.
As SEVRA moves forward, Congress should narrow the MTW expansion to a scope appropriate for a demonstration by reducing the number of agencies that can participate or limiting the number of families affected. In addition, it should restrict transfers of voucher funds and take other measures to reduce the demonstration’s risks and ensure that it results in useful policy findings.
SEVRA Would Expand MTW Sharply
Congress authorized MTW by statute in 1996. Today 30 agencies participate, and the 2009 and 2010 appropriations bills grant HUD authority to admit six more. Congress included some important safeguards in these bills, limiting the size of the agencies that HUD could admit and prohibiting special arrangements that would provide the new MTW agencies with added funding, but otherwise left most rules governing MTW unchanged.
In addition to permitting HUD to admit as many as 80 agencies to MTW, SEVRA would rename the demonstration the Housing Innovation Program (HIP) and place new requirements on participating agencies in areas such as evaluation, transparency, and tenant protections.  Of the 80 agencies, 60 would be “full HIP” agencies, with broad flexibility to operate outside regular federal statutes and regulations. Another 20 would participate in a more limited “HIP-lite” component and would be prohibited from changing rent rules or imposing time limits or work requirements.
SEVRA would place no limit on the size of agencies that can be admitted to HIP, except for a vague requirement that HUD admit “both large and small” agencies. Fewer than 1 percent of agencies participate in MTW today, but they administer about 11 percent of vouchers and public housing units. If the new agencies admitted are of approximately the same size as current MTW agencies, close to 1 million vouchers and public housing units — or nearly 30 percent of the total nationally — could be subject to the expanded demonstration.
This would pose considerable risks for low-income families, as well as for agencies that are not admitted to MTW. While proponents have put forward several rationales for such a large MTW expansion, each could be promoted through less disruptive means, and none justifies the sweeping expansion that SEVRA contains.
Special MTW Arrangements Not the Best Way to Address Public Housing Funding Needs
MTW has provided substantial added resources to housing agencies to renovate or demolish and replace public housing. This has occurred in large part because HUD has established special arrangements that provide some MTW agencies with more funding than they would receive under the regular voucher funding formula, and has permitted agencies to transfer funds out of their voucher programs and use them in public housing or other affordable-housing development efforts.
As a result of long-term underfunding, more than 165,000 public housing units were lost and not replaced from 1995 to 2008. This underfunding now appears to be easing considerably, but public housing still faces a backlog of close to $20 billion in unmet capital needs (and perhaps considerably more). Public housing developments play a major role in meeting the housing needs of the most vulnerable Americans, and preserving high-quality public housing is an important policy goal. But MTW expansion is not the appropriate way to achieve it.
Often, special funding arrangements have provided added resources for MTW agencies by diverting funds from agencies that do not participate in MTW. In addition, shifting funds from the voucher program to public housing and other housing development efforts has resulted in fewer needy families receiving housing assistance than could have been helped with available funds.
Added Funding to MTW Agencies Has Lowered Funding for Non-MTW Agencies
Most agencies receive voucher funding under formulas established each year in appropriations legislation. Most MTW agencies, however, are funded under special formulas established through agreements with HUD.
The details of MTW funding calculations are not always made public, so it can be difficult to determine precisely why agencies receive such high funding levels. The most significant factor, however, appears to be that funding for MTW agencies is often determined by multiplying a per-unit funding level by each agency’s total number of authorized vouchers. By contrast, non-MTW agencies receive funding based on costs only for the share of their vouchers actually in use during the preceding year. Non-MTW agencies on average used 93 percent of their vouchers in 2008, so a typical agency would receive a considerable boost if it were funded based on total vouchers rather than vouchers in use. In fact, most of the largest MTW agencies are using significantly fewer than 93 percent of their vouchers so they benefit even more from this funding method.
In addition, MTW agencies often are permitted to accumulate unlimited funding reserves, while HUD and Congress at times recapture reserves from non-MTW agencies or reduce funding for non-MTW agencies with excess reserves. In 2008 and 2009, Congress directed HUD to “offset” (that is, deduct) a portion of the reserves that an agency held at the start of the year from the agency’s voucher funding; agencies lost $720 million in 2008 funding and $750 million in 2009 as a result. This further tilted the funding system in favor of MTW agencies, which were largely exempt from the offsets. (Even without such offsets, however, MTW agencies with special funding arrangements would have received 40 percent more funding per authorized voucher in 2008 than other agencies.)
The added funding for MTW agencies sometimes has come at the expense of other agencies’ ability to fund their vouchers. In three of the last five years, appropriations have fallen short of the amount for which agencies were eligible under the voucher formula. When this occurs, HUD reduces funding for all agencies on a pro-rata basis, and consequently each additional dollar that MTW agencies receive directly reduces funding levels for other agencies below the amount they are due.
Some MTW agencies also receive public housing operating funds through special formulas. For the most part, these formulas base funding levels on the amounts that agencies received at the time they entered MTW and exempt agencies from major changes to the public housing operating formula that were made in 2007. The 11 agencies with special MTW operating fund formulas were eligible for 53 percent more funding per unit in 2008 than other agencies, on average.
Public housing operating subsidies were underfunded each year from 2003 to 2009, and HUD cut funding levels for all agencies by proration. Consequently, to the extent that special funding arrangements provided more funds to MTW agencies, they reduced the funding levels for non-MTW agencies. It is unclear, however, how much of the added funding that MTW agencies received went to these agencies because of the special arrangements and how much resulted from other factors. The public housing operating fund formula bases subsidy levels on a complex set of variables (including operating costs for privately owned housing in the same region, the age of an agency’s developments, the type of buildings in the developments, and the population served), and it is difficult to accurately estimate how much funding most MTW agencies would have received under the regular formula.
MTW Funding Transfers Have Reduced the Number of Families Receiving Assistance
Despite receiving added voucher funds, some MTW agencies have left large numbers of their authorized vouchers unused and shifted the funds to other purposes (or accumulated them as reserves). From 2005 to 2008, MTW agencies used more than $1 billion in voucher funds for purposes other than voucher assistance. In 2008 alone, agencies shifted about $400 million in voucher funds in this manner. Agencies left about 30,000 vouchers unused in 2008 that they could have used with funding they had available.
A substantial portion of the funds shifted out of the voucher program were used to maintain, renovate, or replace public housing. Agencies likely felt considerable pressure to make such transfers due to the deep underfunding of public housing. In addition, an agency may be particularly reluctant to allow public housing developments to fall into disrepair, since an agency may believe it would face greater public criticism for broken windows or boarded-up units at public housing projects than for the higher rent burdens and longer waiting lists that would result from shifting funds out of the voucher program.
But shifting large amounts of voucher funds to renovate or replace public housing or for other permitted purposes generally does not assist additional needy families (or does not assist a sufficient number of additional families to offset the reduction in the number of families being served with vouchers). HUD does not release data on the total number of families that MTW agencies assist, but an analysis of reports submitted by the five largest MTW agencies — which accounted for 86 percent of the vouchers left unused in 2008 — indicates that those agencies actually reduced the number of families they assisted in public housing during the same period that they left vouchers unused.
This does not mean that these agencies wasted their funds, which often went to worthwhile public housing revitalization efforts or other services or activities benefitting low-income people rather than direct rental assistance. But it does mean that these agencies were less cost effective than other agencies in using federal rental assistance funds to meet the fundamental goal of helping needy families afford housing. Each year, tens of thousands of families go without housing assistance who would have been helped if MTW agencies used their voucher funds for voucher assistance, or if HUD distributed the funds to other agencies through the regular voucher formula rather than shifting them to MTW agencies as a result of special arrangements.
The magnitude of this problem could rise substantially under the MTW expansion in SEVRA. As noted above, the 2009 and 2010 appropriations acts specifically prohibited new MTW agencies from receiving extra funding as a result of their participation. SEVRA does not contain such a limitation and places only modest restrictions on transfers of voucher funds to other purposes. If new agencies admitted to the demonstration leave vouchers unused at the same rate as current agencies, SEVRA’s proposed expansion of MTW could leave 45,000 additional families without voucher assistance.
A Comprehensive National Policy Is the Best Way to Preserve Public Housing
Public housing funding needs are national in scope and should be addressed through a comprehensive national policy. Congress has recently taken significant steps in this direction. The American Recovery and Reinvestment Act of 2009 provided $4 billion in additional public housing capital funds (on top of the regular annual appropriation Congress provided that year), and Congress has approved sufficient appropriations for fiscal year 2010 to fully (or nearly fully) fund public housing operating subsidies for the first time since 2002.
More resources will be needed, but these should be distributed through existing formulas or through new mechanisms made available to agencies on a competitive basis. HUD should not provide them to a minority of agencies through obscure MTW funding arrangements that often reduce funding for other agencies. Nor should individual agencies provide them by transferring voucher funds and thereby leaving needy families without assistance.
Some MTW agencies have supported revitalization of public housing through mechanisms that enable an agency to borrow greater amounts from private sources than is possible under regular program rules, in addition to receiving added federal funding. These mechanisms include, for example, converting public housing subsidies to “project-based” vouchers. It would make sense to expand use of such policies (subject to restrictions to ensure they do not harm low-income families), but it would be more effective to enable all agencies to undertake them than to expand the size of MTW.
Targeted, National Changes Preferable to the Blanket Flexibility Provided by MTW
Some proponents of large-scale MTW expansion contend it is needed to allow agencies to implement a wide variety of policies of their own choosing, not just to test specific innovative policies. Such local policies, the argument goes, would be better suited to the needs of local communities and thus make the rental assistance programs more effective.
However, providing local flexibility with respect to certain housing assistance program rules would likely do more harm than good. And while providing local flexibility with regard to various other rules would be desirable, MTW generally is not the best way to accomplish that.
MTW Would Permit Changes to Rules That Have Helped Make Programs Effective
Many of the federal statutes and rules governing the rental assistance programs provide important protections for low-income families or help ensure that federal funds are used efficiently and effectively. Some rules could be streamlined or otherwise improved. However, giving local agencies broad flexibility to alter basic program rules risks causing significant harm by undermining the consistent national standards that simplify the programs’ operations and allowing agencies to tinker with rules that provide important protections for needy families.
For example, MTW grants housing agencies broad latitude to alter the rules for determining rent payments for housing assistance recipients (which now generally require tenants to pay 30 percent of their income after certain deductions). SEVRA’s MTW provision places some limits on the sharpest rent increases but would still allow major changes. The current rent formula, however, has proven highly effective. It ensures that the poorest families receive adequate subsidies to enable them to meet their basic needs; research shows that housing vouchers provided to families with children sharply reduce homelessness, overcrowding, and housing instability, while enabling some of the most vulnerable families to afford necessities like food and dental care. The formula also avoids giving somewhat better-off families larger subsidies than they need.
Critics have argued that because rents under the voucher program rise as a family’s income increases, the rent rules may discourage work. But the most rigorous study of this issue to date found no significant negative effect on earnings or employment.  This suggests that any potential work disincentive is offset by the various forms of support that vouchers provide for work, such as enabling a worker to move closer to a job opportunity or to avoid an eviction that would have made it difficult to keep a current job.
Some aspects of the rules for calculating incomes and rent are excessively complex and could be streamlined, but this would require only modest changes within the existing framework, and a number of changes of this nature are contained in SEVRA’s non-MTW provisions. It is possible that more substantial changes that could be tested under MTW would improve the rent rules as well, but many of the possible alternatives also could weaken them.
For example, some MTW agencies have implemented flat rents that do not vary based on family income. This may encourage work to some degree, but at the risk of charging low-income families more rent than they can afford (which could make vouchers less effective in areas such as fighting homelessness) and providing higher-income families with more subsidy than they need (which would waste federal funds that could have been used to help needier families). Other MTW agencies have raised the share of income that tenants pay as rent from 30 percent to 35 percent, which reduces costs and allows agencies to help more families but also could discourage work and result in some added hardship for families. Other substantial changes would carry similar tradeoffs.
Existing rules that prevent agencies from imposing time limits also play a key role in making the voucher program effective. MTW, in contrast, permits agencies to impose time limits, even for working-poor families that cannot afford housing on their own. This could weaken the powerful impact that rental assistance has in preventing homelessness and housing instability, and could expose a substantial number of families to serious hardship.
Both the existing MTW statute and the SEVRA expansion require agencies to establish hardship policies to protect families from the harshest effects of policy experiments. This requirement, however, is vague, and hardship policies typically require struggling families to understand sometimes complex program rules and take the initiative to seek relief. Moreover, the limited data available suggest that only a fraction of the families needing hardship exemptions actually receive them (or may even know such exemptions are available and how to apply for them). For example, the Tulare County (California) Housing Authority’s MTW program has established strict time limits that have caused more than 1,000 families to lose housing assistance and rent rules that require some families to pay substantially more than they would under the regular rules. As of 2009, only 62 families had requested hardship relief.
Patchwork of Local Rules Would Complicate Program Administration
In some cases, the existence of consistent national rules itself provides important benefits by simplifying the operations of the rental assistance programs. The rent rules provide an example; they apply with only limited variations to all of the major HUD rental assistance programs and cover all of the private owners and nearly all of the housing authorities that administer the programs. This consistency is strength of the current system.
Most importantly, consistent rent rules make it easier for voucher holders to exercise their right to move with their voucher anywhere in the country where a voucher program operates, not just within the jurisdiction of the agency that issued the voucher. This portability option is a crucial underpinning of the voucher program’s effectiveness in helping low-income households.
For example, a worker who is laid off but finds a new job in a different county can use his or her voucher to move to an apartment within commuting distance of the new employer. Similarly, a victim of domestic violence can flee an abuser, an elderly person or person with a disability can move closer to a needed caregiver, and a family with children can move to an area with better-quality schools — all without losing their vouchers. It makes little sense to require low-income people like these to sift through widely varying local rent rules to determine which localities they could move to without seeing the value of their assistance change sharply.
Maintaining common rules from one locality to another also provides important consistency for the many private entities that participate in the housing assistance programs in more than one locality (including property owners who rent to voucher holders, developers who use housing subsidies to back construction or rehabilitation of affordable housing developments, and lenders who finance such developments). Under the current system, these entities can easily assess the value of a housing subsidy in a given locality.
Finally, it is far easier for HUD to monitor and enforce a single set of national rules. HUD reports that the amount of errors in tenant rent determinations fell by 65 percent from 2000 to 2007, in significant part due to HUD initiatives to provide technical assistance to agencies and monitor their performance. Such efforts would be far more difficult if a large number of agencies set their own rules and the rules varied from area to area.
Allowing widespread local flexibility in rent rules would replace a system that works well with a complex patchwork of local alternatives whose impact is unknown. It would be wiser to test alternatives through small, rigorously evaluated experiments; if these experiments identify changes that would clearly improve on the current system, Congress could enact those at the national level.
Where Flexibility Is Appropriate, All Agencies Should Have It
Local flexibility is desirable in some aspects of the rental assistance programs, and housing agencies already have many options under current law. For example, they can target vouchers on particular subgroups (such as working-poor families or people who are homeless) and enter into “project-basing” agreements that use vouchers to support development of new affordable housing. Such flexibility allows the programs to better respond to local needs and priorities without fostering fundamental program differences that would make it difficult for a family to move or a lender or owner to operate in multiple communities.
There are strong arguments for expanding agencies’ flexibility in certain areas. MTW expansion is not the appropriate way to do this, however, since even under the broad expansion in SEVRA, the demonstration would cover only a small minority of agencies. Expansion would provide those agencies with more deregulation than is needed or beneficial, while doing nothing to help other agencies.
It would be more effective to enact specific legislative changes that provide added options to all agencies. SEVRA would take this approach in many areas. For example, it would allow agencies to conduct housing quality inspections every two years instead of annually and to set aside a larger portion of their vouchers for project-based assistance. It may be beneficial to provide agencies with some options beyond those now in SEVRA, but Congress should do this by expanding SEVRA’s non-MTW provisions or through other legislation applying to all agencies.
Sweeping MTW Expansion Is Much Larger than Needed to Test Innovative Policies
As with any government program, the voucher and public housing programs should evolve over time as lessons are learned and circumstances change. To guide this process, it is useful to permit agencies to test innovative policies. Such demonstrations, if rigorously and independently evaluated, can provide information about the effectiveness of various alternatives that can help policymakers determine which ones to implement nationally.
The MTW expansion permitted under SEVRA, however, exceeds the scope needed for research purposes. Jobs Plus and Moving-to-Opportunity, two recent multi-city HUD demonstration programs that were rigorously evaluated and generated a wealth of policy lessons, involved about 5,000 families each, including the control groups. SEVRA’s MTW expansion could affect as many as 1 million vouchers and public housing units — 200 times as many as Jobs Plus or Moving-to-Opportunity, and many more than are needed to test even a wide range of innovative policies.
An expansion of the magnitude proposed in SEVRA would not only be unnecessary to test innovative policies but could actually impede that goal by making it difficult for HUD to conduct a meaningful evaluation. The most effective way to measure the impact of experimental policies such as rent changes, work requirements, and time limits will usually be to randomly select some families to be subject to those policies and assign otherwise-identical families to a control group. In addition, evaluators should carefully track families even after they leave housing assistance, since a study focused on current participants will miss much of the impact over time of policies such as time limits.
Rigorous evaluations of this type will generate far more meaningful findings because they can distinguish the effects of the policies being tested from those of other factors. (For example, they can distinguish increases in family earnings that are the result of an experimental policy from those that stem from unrelated factors such as trends in the regional economy.) But rigorous evaluations require more resources. A very large expansion of MTW could force HUD to carry out numerous inexpensive, superficial evaluations that leave key policy questions unanswered, rather than concentrating its resources on a smaller number of high-quality evaluations.
Though MTW agencies administer about 380,000 voucher and public housing units and are implementing alternative policies in a range of program areas, the demonstration has generated few hard policy findings to date because it has not been rigorously evaluated. HUD has begun to require MTW agencies to report more information on program activities and outcomes, and the 2010 appropriation bill required the department to carry out an evaluation of MTW. These efforts will likely produce some useful information, although it is unclear whether they will result in a rigorous evaluation that will generate meaningful policy findings. SEVRA’s MTW provisions contain a more specific requirement for evaluation of MTW agencies, and could be strengthened further to explicitly require rigorous evaluation.
Some expansion of MTW likely will be needed as well, since there will probably be some important policy alternatives that current MTW agencies are not interested in testing. The six new agencies that HUD is permitted to admit under the 2009 and 2010 appropriation bills will provide considerable opportunities for added experimentation. And there may be some value in expanding MTW further in SEVRA to test still more policies.
But Congress should limit the size of the expansion to minimize the number of families exposed to untested policies. For example, Congress could cap the number of public housing and voucher units that can be affected (particularly under the “full HIP” component, which allows testing of the riskiest policies) or reduce the number of agencies that can be admitted. In addition, HUD should have the authority to ensure that participating agencies test the types of policies where research is most needed.
The goals of testing innovative policies, providing added flexibility for housing agencies in certain policy areas, and meeting funding needs in public housing are all reasonable and important. SEVRA’s sweeping expansion of MTW, however, is not the best way to achieve these goals. It would reduce the number of low-income families receiving housing assistance by diverting funds from the voucher program, expose many more families to risky policies than is needed to test and evaluate innovative approaches, and allow policies to diverge from one community to another to a degree that could weaken housing assistance programs.
There are less disruptive and more effective ways to promote each of these goals. As SEVRA moves forward, Congress should limit the scope of the MTW expansion to a scale appropriate for a demonstration, restrict the transfer of voucher funds, ensure that HUD has adequate authority to focus the demonstration on the policies where experimentation would be most beneficial, and mandate a rigorous evaluation. Where deregulation is appropriate, Congress should make it available to all agencies nationwide, not a select minority of agencies participating in MTW. And policymakers should address the challenge of preserving public housing through a comprehensive national approach rather than special funding arrangements with MTW agencies.
 For additional information on SEVRA, including its MTW/HIP provision, see Will Fischer, “SEVRA Housing Voucher Reform Bill Would Update and Streamline Program,” Center on Budget and Policy Priorities, September 22, 2009, . For a detailed side-by-side comparison of the House bill and current law, see the table https://www.cbpp.org/sites/default/files/atoms/files/9-22-09hous-appendix.pdf.
 For additional information on unmet funding needs in public housing, the importance of addressing them, and possible strategies for doing so, see Will Fischer and Barbara Sard, “Preserving Safe, High-Quality Public Housing Should Be a Priority of Federal Housing Policy,” Center on Budget and Policy Priorities, October 8, 2008.
 This comparison includes both voucher subsidy funds and funds provided to cover the cost of administering vouchers. It includes new vouchers issued in 2008, except those issued through the Veterans Affairs Supportive Housing program. Sufficient data are not yet publicly available to make the same comparison for 2009, but the available data suggest that MTW agencies with special funding again received more than 40 percent more funding per voucher than other agencies.
 These estimates are based on HUD data on the amount of voucher subsidy funds provided to MTW agencies and the amount that the agencies spent on voucher subsidies. Some MTW agencies receive voucher administrative funding (which for non-MTW agencies is provided through a separate budget account) and subsidy renewal funds together in single funding stream. In these cases, we estimated the amount that was intended as administrative funding and deducted it from the agency’s funding level before calculating the amount of funds unspent. An earlier CBPP analysis estimated that $300 million in voucher funds were shifted in 2008 and at least 25,000 vouchers left unused as result. The revised figures used here reflect additional information about agencies’ funding levels.
 The five largest agencies are Atlanta, Baltimore, Chicago, Philadelphia, and Washington, DC. These calculations include all families listed in agency reports as receiving housing assistance, including those participating in housing assistance initiatives outside the regular voucher and public housing programs. The funding levels include voucher subsidy funds, administrative fees, public housing operating funds, and public housing capital formula grants, but exclude HOPE VI grants. The reports cover agency fiscal years, which may end in the last month of any calendar quarter. The data here are based on reports for years ending from September 2008 to June 2009; to permit comparison with calendar year 2008 data for non-MTW agencies, funding levels at MTW agencies were adjusted for inflation.
 MTW agencies are required to continue to assist “substantially the same” number of families as they would have had they not combined or shifted voucher and public housing funds, but HUD has interpreted this requirement loosely and does not appear to have enforced it consistently.
 Agencies newly admitted under the SEVRA MTW expansion would be allowed to begin shifting funds between vouchers and public housing only if they used at least 95 percent of their vouchers or spent 95 percent of their voucher funds in the preceding year. But once an agency met this threshold for one year, it would be allowed to continue shifting funds for as many years as it remained in the demonstration.
 The bill states that agencies cannot adopt rent rules that would result in “families making substantially higher rent payments than would customarily be made by families of comparable income” under the regular rules. But the bill does not indicate how high an increase must be to be considered “substantial,” how broad a range of income is considered “comparable,” or whether the standard would apply to each family or to an average of the families assisted by an agency (which would allow large increases for individual families). HUD has interpreted several vague or ambiguous provisions of the MTW statute to provide agencies with broad flexibility.
 Gregory Mills, et al., “Effects of Housing Vouchers on Welfare Families,” prepared by Abt Associates for the HUD Office of Policy Development and Research, 2006.
 Mills, et al., 2006.
 HUD Office of Policy Development and Research, “Quality Control for Rental Assistance Subsidy Determinations: Final Report for FY 2007,” 2008.