You are here

HUD Seeks Significant Improvements to “Moving to Work” Demonstration, But Additional Changes Needed

January 21, 2015

The Department of Housing and Urban Development (HUD) is preparing to extend through 2028 the Moving to Work (MTW) demonstration, which waives program rules and sets special funding formulas for 39 of the state and local agencies administering the Housing Choice Voucher and public housing programs.  HUD intends to address some key MTW shortcomings by reducing shifts of rental assistance resources to other purposes, distributing funding more fairly between MTW and non-MTW agencies, and requiring more rigorous evaluation of certain MTW activities.  Important details of the extensions will need strengthening, however, to ensure that they actually benefit low-income families; added changes are also needed in some areas that HUD’s plans don’t cover. 

Despite its name, MTW is a broad deregulatory initiative rather than one focused on employment.  It allows HUD to waive most of the main statutory and regulatory provisions that govern vouchers and public housing, establish special funding formulas for MTW agencies, and permit agencies to shift voucher and public housing funds between the programs or to other uses. 

MTW agencies operate under agreements with HUD that delineate the scope of waivers, establish responsibilities of agencies and HUD, and set funding formulas.  The current agreements run through 2018, at which point the agencies’ participation in MTW will end unless the agreements are extended.  HUD and MTW agencies are negotiating a decade-long extension now — three years before the agreements expire — to address agencies’ concerns that it would be more difficult to obtain private investment in MTW-funded housing development projects if the agreements aren’t extended until they are close to expiration. 

Since its implementation in 1998, MTW has resulted in some useful innovation but has also exhibited major shortcomings.[1]  According to the most recent available description of HUD’s plans, HUD intends to use the extensions to address a number of these shortcomings and plans various changes to the current agreements.[2]  Substantial adjustments are needed to some aspects of HUD’s plans to ensure that they bring about real benefits, however, and improvements are needed in some additional areas where HUD has not said that it plans changes.

  • HUD plans to end preferential public housing funding formulas for some MTW agencies, which deprive non-MTW agencies of funds to maintain developments in decent condition.  Instead, HUD plans to provide equitable funding to all agencies that operate public housing.
  • HUD plans to require agencies to achieve a “voucher utilization rate” of 90 percent or greater.  If agencies had to use at least 90 percent of their voucher funds for vouchers — rather than shifting large amounts to other purposes, as some MTW agencies do today — more than 20,000 additional low-income families could receive rental assistance at no added federal cost.  But HUD’s plan appears too ill-defined and weakly enforced to achieve a significant, sustained increase in the number of families receiving rental assistance.
  • HUD is considering raising the total number of families that agencies must assist through their MTW programs.  But HUD may undermine this requirement by allowing agencies to count families that receive very shallow assistance.  
  • HUD plans to require added evaluation of new MTW policies that pose the greatest risks for low-income families, likely including major rent changes, restrictions on where families can live, time limits, and work requirements.  HUD has yet, however, to fill in key details needed to assess whether its plan will result in meaningful evaluations that generate useful information about the policies’ impact.   
  • HUD’s plan would do virtually nothing to encourage MTW agencies to give low-income families access to neighborhoods with good schools and low crime.  Research shows that living in lower-poverty neighborhoods can have major benefits for children.[3]  Expanding housing choice is one of MTW’s three statutory goals, but many MTW agencies have done little to help families move to a wider range of neighborhoods.  Some have used their flexibility under MTW to restrict housing choice.

The decisions HUD makes now regarding MTW’s extension have long-term implications.  HUD has broad discretion over whether and how to extend agencies’ participation in MTW, but once the extensions through 2028 are signed, the agreements will give agencies legal rights that limit HUD’s authority to require further improvements for another decade. 

Moreover, the impact of the extension agreements could reach beyond the 39 agencies now in MTW.  HUD views the extensions as laying the groundwork for adding other agencies to MTW if Congress were to authorize such an expansion.[4]   Several legislative proposals have called for major MTW expansions.[5]  The extension terms for today’s MTW agencies wouldn’t automatically apply to new agencies, and some provisions could need adjustment based on new legislative requirements, but HUD would likely seek to enter agreements with new MTW agencies that are as consistent as possible with the extensions.

Providing Equitable Public Housing Funding


One important improvement HUD plans to make through the extension agreements would end alternative formulas used to set public housing operating funding for 11 MTW agencies.[6]  HUD plans to fund these agencies (after a transition period to allow them to adjust) under the same formula used for other MTW agencies and non-MTW agencies.

The special MTW formulas for these 11 agencies are far more generous on average than the regular operating fund formula.  In 2013, agencies funded through alternative formulas were eligible for $7,900 per public housing unit — 91 percent more than the average for non-MTW agencies and the 28 MTW agencies without special formulas (see Figure 1).  A portion of this difference may be explained by factors such as the age of developments and the incomes of residents, which affect operating subsidy needs, but much of the difference appears to reflect the terms of the MTW formulas.

In recent years, Congress has usually underfunded the public housing operating fund, resulting in pro rata reductions for all agencies.  The added funding provided to the 11 MTW agencies makes these prorations deeper than they otherwise would be.  If MTW agencies with alternative formulas received the same average funding as other agencies, they would have been eligible for $260 million less in 2014.[7]  HUD would then have had sufficient funds to provide all agencies 94 percent of the funding for which they were eligible, instead of the 89 percent they actually received.  

The deep underfunding of public housing has left some agencies with substantially less funding than they need to operate their developments.  This can compel agencies to raise utility charges and other fees on vulnerable low-income residents or to cut back on security or maintenance, potentially exposing residents to safety threats and causing higher federal costs down the road.  Ending preferential MTW formulas after a reasonable transition period and thereby making more funding available for the most severely underfunded agencies would reduce the need for such steps.

Encouraging Agencies to Use Voucher Funds to Assist More Needy Families

HUD plans new requirements to encourage housing agencies to assist more families through the voucher program.  As described in available documents, however, these measures could have little long-term impact.  Some MTW agencies use far fewer vouchers than they could with available funds, largely because MTW agencies receive nearly all of their voucher funds through block grant formulas, which adjust voucher subsidy and administrative funding from year to year based on inflation, regardless of how many families an agency assists or how it uses its funds.  By contrast, non-MTW agencies receive subsidy funding based on the cost of the authorized vouchers they used in the previous year adjusted for inflation, and administrative funds based on the number of their vouchers in use.[8] 

MTW Agencies Use Far Fewer Vouchers Than They Could with Available Funds

These federal funding policies create a much stronger incentive for non-MTW agencies to issue vouchers to low-income families than MTW agencies.  Data on voucher use show the impact of these incentives.  Non-MTW agencies leased 99.5 percent of the vouchers they were authorized and funded to use in 2013, the last year for which complete data are available.[9]  In contrast, the 35 agencies that participated in MTW throughout 2013 used only 86 percent of the vouchers they could have supported with their 2013 funding.[10]  (See Figure 2.)


Overall, MTW agencies shifted more than $350 million in voucher funds to other purposes in 2013 or left the funds unspent.  MTW agencies can use transferred funds to provide rental assistance to additional families through public housing or other programs, but this hasn’t occurred on a scale anywhere near sufficient to offset the loss of vouchers, and the bulk of the funds haven’t been used to expand rental assistance.

MTW agencies have used voucher funds for a wide range of purposes, including renovating public housing, developing new housing, providing services to housing assistance recipients, supplementing agency budgets for administration and operations, or leaving funds unspent and accumulating large reserves.  Some expenditures appear to have achieved positive outcomes, but there has been no rigorous assessment of whether they promote the goals of the housing assistance programs cost-effectively.  Indeed, available evidence strongly suggests that some transfers do not improve cost effectiveness.  For example, studies by the Government Accountability Office (GAO) and others have found that developing new affordable housing is generally less efficient than providing vouchers to help low-income families afford modest housing of their choice in the private market.[11]  

There is no evidence that transfers of MTW funds for services or other purposes pay for themselves (for example, by increasing tenant incomes and rent payments or freeing up subsidies to assist new families by enabling recipients to afford housing on their own).[12]  Any transfers that do not achieve those outcomes or provide housing assistance cost-effectively through other programs will exacerbate the already severe unmet need for assistance.  HUD estimates that in 2011, some 8.5 million renter households that had incomes below half of the median income in their area and didn’t receive rental assistance either lived in substandard housing or paid more than half their income for housing.[13]

HUD Plans New Voucher Utilization Requirements, But Further Improvements Needed

HUD has identified low voucher utilization at MTW agencies as a significant problem, noting that “it is critical that the scarce resources appropriated by Congress for housing assistance be used to house families.”[14]  To further this goal, HUD has said it will extend MTW agreements only for agencies that achieve a 90 percent “voucher utilization rate” in some year before 2018 and will require a 90 percent utilization rate on an ongoing basis after the agreements are extended.  If all MTW agencies had raised the share of their voucher funds they spent on vouchers to 90 percent in 2013, about 23,000 additional families would have received assistance.  Yet a sustained increase of this magnitude in effective assistance for low-income families would be unlikely under the planned MTW reforms HUD has described, for three reasons.

  • HUD plans to allow agencies to meet the utilization requirement by spending 90 percent of their voucher funds on qualifying expenditures or using 90 percent of the vouchers they are authorized to administer.  The latter option could encourage agencies that wish to shift funds to other purposes to provide very shallow subsidies, enabling them to use 90 percent of their authorized vouchers with much less than 90 percent of their funds.  Very shallow subsidies, however, will often be too low to bridge the gap between the cost of decent, stable housing and the amount that the lowest-income families can afford to pay.
  • HUD plans to allow spending on developing “affordable housing” to count as voucher spending for the utilization requirement.  Since development expenditures tend to be less cost effective over the long run than direct rental assistance, spending funds on development will assist fewer families, on average.  In addition, development projects may reduce rents only modestly (if at all) below market rents, so they could be much less effective than direct rental assistance in providing stable housing for the neediest families. 
  • HUD plans a one-time adjustment to voucher funding in 2019 for agencies with low utilization in 2018, but after that, MTW agencies would be allowed to revert to a block grant formula with no incentive for higher utilization.[15]  Thus, after 2019 the voucher funding formula for MTW agencies wouldn’t be based on their actual prior-year voucher utilization and expenditures — the approach that has successfully encouraged high utilization at non-MTW agencies.  Instead, HUD apparently would rely on administrative sanctions to enforce the 90 percent utilization requirement. 

Identifying violations and determining sanctions would require case-by-case decisions that would strain HUD’s administrative capacity.  HUD would likely give agencies a period of time to return to compliance before imposing any sanctions, and agencies might take political or legal action to weaken or delay penalties further, dampening any incentive to maintain a high voucher utilization rate.  As a result, there would be a serious risk that agencies wishing to shift funds to other purposes would comply with the requirement in 2018 (to avoid losing funding) but then allow utilization to decline.   

HUD should determine compliance with the 90 percent utilization requirement based solely on the share of the agency’s voucher subsidy funds spent on voucher subsidies or other direct rental assistance — not development, services, administration, or other purposes.  HUD also should determine the bulk of MTW agencies’ voucher subsidy funding throughout the term of the extensions the same way it does for all other agencies:  based on utilization in the previous year.  MTW agencies could continue to receive 10 percent of their funds as a block grant that they could use for other purposes without penalty, but HUD should reduce their funding proportionately if an agency’s utilization rate falls below 90 percent.

Improving Enforcement of “Substantially the Same” Requirement

HUD plans to take another step to require agencies to assist more families by strengthening implementation of a statutory requirement that MTW agencies assist “substantially the same” number of low-income families as they would without the flexibility to transfer voucher and public housing funds to other purposes.  Congress included this requirement in the legislation authorizing MTW in 1996, but HUD didn’t establish a process to determine if agencies met this requirement until 2013, and the current policy has two important flaws. 

  • HUD has implemented the requirement by assessing whether an agency assists substantially the same number of families as it did when it entered MTW (adjusted for new voucher awards and removal of public housing units).  This “baseline” is often well below the number of families the agency could assist if it didn’t transfer funds, mainly because some MTW agencies had large numbers of idle vouchers and vacant public housing units when they entered MTW but received funding for those units under the funding formulas then in place.
  • HUD counts nearly any income-eligible family that receives any type of MTW-funded housing-related benefit as “assisted,” even if the assistance is very small.  For example, the current definition would count as “assisted” any eligible family living in a building that received even a small MTW-funded subsidy during construction, for as long as the building remains as affordable housing.  As a result, agencies could sharply cut the number of families with substantial rental assistance while “assisting” many other families with small shares of their funding, then use the remaining funds for other purposes. 

HUD intends to improve the process for setting the baseline for the “substantially the same” requirement.  One possible option, HUD has indicated, is setting the baseline to reflect the number of families each agency could assist with its funding.  This would be a significant improvement.

HUD, however, hasn’t indicated how it will define “assistance” in the MTW extension agreements.  One promising model is provided by the Affordable Housing and Self-Sufficiency Improvement Act (AHSSIA), a bill proposed in 2012 that would have expanded MTW, added new protections to address some of the demonstration’s shortcomings, and enacted other reforms to federal rental assistance programs.  Under a draft of AHSSIA circulated by the House Financial Services Committee leadership in April 2012, families would only be counted as assisted if they receive rental assistance that reduces their rent burden to levels that on average are comparable to those of families that receive assistance under the regular rent rules.[16]   

Requiring Rigorous Evaluation of Major Policy Changes

Reports from the Urban Institute, HUD, and GAO have found that MTW has generally failed to generate conclusive results showing whether the alternative policies that agencies have implemented have been effective.  A 2010 HUD report and some other assessments have highlighted “best practices,” but GAO noted that “in most cases, the practices chosen were based on the opinions of HUD or contracted staff and largely involved anecdotal (or qualitative) data rather than quantitative data.”[17] 

This is partly because MTW wasn’t designed as an experimental evaluation, in which randomly selected families receive housing assistance under alternative policies and are compared to otherwise similar families who receive assistance under regular program rules.  Random assignment is not necessary for evaluating relatively modest changes, such as streamlining the processes for determining tenant incomes or conducting housing quality inspections, where the risks of adverse consequences for low-income families are limited and the impact on administrative costs is direct and easy to measure.  But random assignment is a standard feature of successful demonstrations that test more substantial policy changes; otherwise it is very difficult to determine those policies’ actual effects.

For example, some MTW agencies require non-elderly, non-disabled recipients of housing assistance to work as a condition for continuing to receive subsidies.  Such policies may increase work to some degree among the small minority of rental assistance recipients who are not elderly, disabled, already working, or subject to work requirements under another program.[18]  But they may also result in homelessness or other severe hardship for families that cannot find jobs, as well as higher administrative costs from monitoring and enforcement.  Without an experimental evaluation, it is difficult to determine whether subsequent changes in employment, hardship, or costs stem from the work requirement or other factors, such as economic conditions or changes in the agency’s caseload.

The extension of MTW agreements gives HUD the opportunity to make the demonstration more effective in testing alternative policies.  HUD intends to require more rigorous evaluation of major new policies not permitted in the regular voucher and public housing programs, which it indicated could include time limits, work requirements, major rent changes, and restrictions on housing choice.  This is an important step, but two details will be critical: 

  • Evaluations should be sufficiently rigorous to draw meaningful conclusions about the effects of MTW policies.  HUD should require random assignment evaluations for major policy changes — including those noted above that HUD identified and other policies that risk substantial harm to low-income families — unless it determines in a particular case that another approach would measure a policy’s impacts equally or more accurately.
  • HUD should make clear that if federal funds are unavailable to support a rigorous evaluation, agencies will not be permitted to implement the specified policies unless they identify other resources to fund a rigorous evaluation meeting HUD’s standards.  This will give agencies interested in testing policies a strong incentive to seek the needed resources from foundations and other sources. 

HUD should also require agencies, to the extent feasible, to evaluate major policy changes they have already adopted.  HUD appears to have targeted new policies for more rigorous evaluation because it is more difficult to experimentally evaluate policies that have been in place for a period of time (since this would require shifting some families back to the regular public housing and voucher rules).  Some policies adopted shortly before the extension agreements are signed, however, may be early enough in the implementation process that a rigorous, controlled evaluation is still feasible.  Even when policies have been in place for an extended period, non-experimental evaluation and monitoring — such as tracking outcomes for families terminated from assistance due to time limits or work requirements — could provide useful information.

Supporting Housing Choice

One of MTW’s three statutory goals is to “increase housing choices for low-income families.”  Broadening housing choice — in particular, helping low-income families with children gain access to low-poverty, high-opportunity neighborhoods — is an important policy goal.  A growing body of evidence indicates that living in high-poverty neighborhoods can impair children’s cognitive development, school performance, and mental and physical health.  By contrast, poor children who live in low-poverty neighborhoods and consistently attend high-quality schools perform better academically than those who do not.

Many MTW agencies, however, have done little to promote choice, and some have even used their expanded flexibility to limit the choices available to low-income families.  For example, voucher holders normally can use their vouchers in any community with a voucher program.  This “portability” right allows families to move closer to a job or a desired school, even if these opportunities are in the suburbs and the voucher was issued by a central city housing agency.  More than a dozen MTW agencies, however, have proposed (and in most cases obtained) waivers from HUD allowing them to restrict voucher portability.

HUD lists mobility restrictions as one of the policy types for which it may require more rigorous evaluation.  HUD should not, however, grant waivers permitting such restrictions even if they are evaluated.  Housing choice is a fundamental characteristic of the voucher program, and HUD is legally obligated to support fair housing opportunities.  Moreover, the primary benefit of restricting portability would go to local agencies spared the work of administering the option, not to families.  HUD has proposed a regulation making portability easier to administer for all agencies; this is a far better approach than allowing MTW agencies to scale back portability rights.

HUD should also establish criteria to identify MTW agencies whose voucher holders (or recipients of all forms of rental assistance) are excessively concentrated in high-poverty areas and require them to develop plans to use their MTW flexibility to help assisted families access a broader range of neighborhoods.  These plans could include targeted outreach to landlords and property managers in high-opportunity communities, housing search assistance, funds for security deposits, use of project-based vouchers in low-poverty neighborhoods, and voucher subsidy caps tied to neighborhood rents so they allow the use of vouchers in a wide range of neighborhoods. 

HUD should also require the identified agencies to use a portion (perhaps 5 percent) of any voucher funds they shift to purposes other than rental assistance to support the plan.  For example, an agency that shifts the full 10 percent permitted under the voucher utilization requirement described above would have to use 0.5 percent for activities that broaden housing opportunities, leaving no more than 9.5 percent for other purposes.[19]


The extension agreements offer a pivotal opportunity for HUD to strengthen federal rental assistance.  Rental assistance sharply reduces homelessness and housing instability — problems with long-term adverse effects on education and health outcomes — but fewer than one in four eligible families receive assistance due to funding limitations.  Well-designed extension agreements could help tens of thousands of added low-income families afford housing using available funds, improve the quality of assistance for families now served by MTW agencies, spur effective research on alternative rental assistance policies, and establish a policy framework that could reduce the potential adverse effects if MTW is expanded to additional agencies.  A similar opportunity to improve the demonstration will not occur for another decade.

HUD has outlined goals that would move the demonstration in the right direction, such as requiring higher voucher utilization and more rigorous evaluation.  It’s critical, however, that HUD strengthen the extension provisions to generate meaningful lessons about the effectiveness of alternative policies and deliver concrete benefits to the vulnerable low-income people the public housing and voucher programs serve.

End notes:

[1] See Will Fischer, “Sharp Expansion of HUD’s MTW Demonstration Raises Serious Concerns,” Center on Budget and Policy Priorities, January 22, 2010,

[2] Department of Housing and Urban Development, “Summary of Moving to Work (MTW) Extension Principles and Updated Terms,” November 24, 2014.

[3] Barbara Sard and Douglas Rice, “Creating Opportunity for Children: How Housing Location Can Make a Difference,” Center on Budget and Policy Priorities, October 15, 2014,

[4] HUD, 2014, p. 1.  

[5] Senator David Vitter (R-LA) introduced a bill on January 7, 2015 requiring HUD to add 250 agencies to MTW; Representative Steve Stivers (R-OH) introduced a bill in June 2014 permitting HUD to add ten agencies.   In April 2012, the House Financial Services Committee’s leadership included an MTW expansion plan in a draft of the Affordable Housing and Self-Sufficiency Improvement Act that would have allowed HUD to expand MTW to cover more than 1.3 million voucher and public housing units — over a third of the total — while adding measures to seek to limit the adverse consequences.

[6]  The Chicago, New Haven, Delaware, Oakland, Lawrence-Douglas County (KS), Cambridge, Baltimore, Philadelphia, Pittsburgh, Portland (OR), and Seattle housing agencies currently have alternative public housing operating fund formulas. 

[7] The detailed data needed to determine how much operating funding MTW agencies would be eligible for if they were funded under the regular formula are not publicly available, so the actual impact could be somewhat higher or lower.

[8] Both MTW and non-MTW agencies also have their funding for existing vouchers raised or lowered on a pro rata basis if Congress appropriates more or less overall voucher funding than agencies are due, and receive added funding if they are awarded new vouchers.

[9] The share of funded vouchers in use at both MTW and non-MTW agencies was somewhat higher than normal in 2013 because voucher funding was cut deeply below the 2012 level.  Agencies consequently needed to use a larger share of their funds to continue assisting the families they assisted the previous year — and many agencies had to reduce the number of families they assisted because of inadequate funding.  In 2012, when (as in most years) voucher funding rose modestly and was sufficient to cover most or all vouchers in use, MTW agencies used 83 percent of the vouchers they could have used with their funds and non-MTW agencies used 98 percent.  Complete data for 2014 are not yet available, but since funding rose significantly above the 2013 level, the share of funds in use likely fell at both groups of agencies.

[10] The 86 percent figure is the share of all of their funded vouchers, including vouchers above their “authorized caps,” that MTW agencies used in 2013.  Congress permits MTW agencies, but not other agencies, to use vouchers above the number they are authorized to administer, and some MTW agencies receive sufficient funds to do so.  In 2013, MTW agencies used 89 percent of the somewhat smaller number of vouchers that were covered by available funds and were below their authorized caps.

[11] Government Accountability Office, “Federal Housing Assistance: Comparing the Characteristics and Costs of Housing Programs,” GAO-02-76, January 2002,

[12] Barbara Sard, “Most Rental Assistance Recipients Work, Are Elderly, or Have Disabilities,” Center on Budget and Policy Priorities, July 17, 2013,, pp. 8-9 and 16-18.

[13] Department of Housing and Urban Development, Worst Case Housing Needs 2011: Report to Congress, August 2013,

[14] HUD, 2014, p. 1.  

[15] In 2019, HUD plans to continue funding for agencies with 2018 utilization at or above 90 percent at their full block grant funding level adjusted for inflation, but to reduce funding for agencies with utilization below 90 percent to equal their 2018 utilization level adjusted for inflation.  Since HUD only plans to grant extensions to agencies that achieve 90 percent utilization at some point from 2014 through 2018, it appears that this funding adjustment would only apply to agencies that achieve 90 percent utilization before 2018 but then drop below 90 percent in 2018.

[16] Some refinements of the AHSSIA definition would be needed, including limiting an exception for families subject to experimental rent rules and prohibiting agencies from counting units that receive deep rental assistance through non-MTW funds.

[17] Government Accountability Office, “Moving to Work Demonstration: Opportunities Exist to Improve Information and Monitoring,” GAO-12-490, April 2012, p. 21.

[18] In 2010, the most recent year for which data are available, 88 percent of households with HUD rental assistance were elderly, disabled, currently or recently working, or likely subject to work requirements under Temporary Assistance for Needy Families (TANF) or another program.  See Sard 2013, note 12 above.

[19] HUD could further encourage MTW agencies to help families move to high-opportunity areas by requiring them to use 90 percent of their voucher administrative funding for either administration of voucher-funded rental assistance or activities that help families use that assistance.  Currently MTW agencies can use unlimited amounts of administrative funds for purposes unrelated to their voucher programs.