Emergency Savings through Housing Interventions

Remarks of Chris Herbert, Joint Center for Housing Studies, Harvard University

At first blush, this panel of papers appears to have a strong central theme as all three present examples of emergency savings initiatives built around housing programs. Yet, the programs are distinct in fundamental ways, aimed at households in very different financial circumstances and designed to address very different savings goals. Perhaps the common lesson that can be drawn from these examples is that expanding opportunities for low-income people to save to meet unexpected financial needs and opportunities will by necessity employ a range of approaches designed to meet a range of needs.  In part, the need for diverse approaches reflects the fact that among the key barriers to a greater degree of savings are behavioral issues, including whether individuals are motivated to save and whether they are given nudges and incentives to follow through on the intention to save. An effective way to address these behavioral obstacles is to imbed savings opportunities into existing programs to take advantage of moments when people are likely to be motivated to save and also to leverage institutional arrangements to create mechanisms for savings.

The first two papers in this panel present examples of how homeless and homeownership programs may offer such opportunities. The final paper, detailing an approach to generating savings for renters out of more efficient management (or ‘rental equity’ in the parlance of the paper), is different in both the method used to generate savings and in the purpose of these savings. Rather than rely on participants setting aside their own income, the program provides an opportunity for residents to earn additional income by participating in efforts to reduce the cost of managing the housing development where they live. This program is essentially addressing the fundamental challenge of fostering savings among low-income households with little slack in their household budgets to support savings. Given the high cost of providing housing services even modest cost savings for the development can represent significant financial assistance for individual households without straining their budgets.

Specific comments and observations on the three papers follow.  I conclude with a few general observations on the papers.

Emergency Savings for Shelter Residents: Modifying New York City’s Mandated Savings Initiatives

The goal of this planned effort is to foster savings among homeless families with earned income in New York City’s shelter system in order to facilitate their move out to permanent housing and to achieve longer-run financial stability. The program is designed to address several flaws in existing mandates for this population to save money monthly, which has experienced low take up rates despite being mandatory. The program is built around three innovative features to address what are perceived to be the chief obstacles to greater participation in savings efforts by these clients.

One of the main obstacles to greater participation by working families is the lack of confidence that moving into permanent housing is financially feasible given how tight their household budgets are. To address this, the program includes weekly peer support group meetings that are built into the existing casework program flow. The weekly meetings incorporate discussions with former shelter residents that have made the transition to permanent housing to provide peer examples of success. The meetings also provide an opportunity for residents to personalize their savings goals (such as identifying furniture they would like to obtain for their new home) to move the savings effort  beyond simple mandatory compliance and more a goal that the residents have bought into.  The peer group also provides a forum where they publicly commit to working toward a savings goal as a further form of motivation.

A second feature of the program is to provide easy options for making deposits, by allowing for automatic deposits from other accounts and providing deposit forms with monthly statements. The accounts would also include more accessible and transparent information on savings achieved through better designed statements and online or telephone systems. Importantly, the savings would also be accessible immediately upon move out to pay for the costs associated with this move, rather than at some point after move out which is current practice in other initiatives.

Finally, the program would also provide bonuses for regular weekly savings behavior, which would include both perquisites while in the shelter (such as relaxation of residency rules or coupons for local businesses) and the accrual of bonus points towards a one-time payment of $250-$500 upon move out. As designed, these bonuses would be tied to making some contribution—however small—on a weekly basis with the goal of creating a habit of savings. Points would be earned in a given month if the resident saved some amount each week of that month. The final cash bonus would be a fixed amount based on the number of points earned rather than the amount saved to reward the routine savings behavior rather than the level of savings.

In my view, the program has a number of notable strengths. First, it starts with a clear recognition of many of the key barriers to savings among the client population, including many behavioral aspects. Importantly, the nature of the program also builds on the existing operations and incentives of the ongoing shelter system. The peer support group would be facilitated by existing case workers and simply be an extension of existing case work, while the savings effort would build on an existing mandate for savings among working families. As a result, implementation costs do not seem excessive. If successful, the program has the clear potential to pay for itself as the savings from even a slightly shorter stay in the shelter system would generate substantial savings.

A key question for the pilot will be whether the program offers sufficient incentives for shelter residents to voluntarily participate.  As the paper notes, moving out of shelter it is not necessarily the personal goal of many residents. Despite being mandated, the existing savings requirement is reported to have a very low take up rate. Once residents agree to participate, the peer support groups may create more buy in to the savings goal, but residents must first agree to participate. The program does offer some ongoing bonuses for regular savings behavior, such as later ‘lights out’ allowances and the like. As the program is implemented it is likely that the program will have to test the attractiveness of various carrots for drawing residents into the program. The incentives may have to include both greater upfront bonuses as well as higher payouts upon move out to make that goal seem more attainable.

Another program design question is whether the proposed structure for earning rewards is too strict in some ways and too lax in others.  Fostering a weekly savings habit is clearly an important goal, but since failure in the early stages of building a new habit is common, there may need to allow for some incentive to get back on the wagon after a week of savings is missed. For example, perhaps points could be earned for each week that savings are added, but an additional bonus is awarded when there are no missed weeks.

On the other hand, the program’s feature of not setting any standard for the amount of savings may be too lax. Residents could game the system by making a token deposit to earn credits. More points ought to be awarded for coming closer to meeting the stated savings goal. Furthermore, if the ultimate goal of the program is to facilitate move out, it seems logical that the monthly savings goal should be tied to the estimated monthly rent payments required for permanent housing. The paper does not go into these specifics, but it does seem like this is a necessary step to create the confidence that residents need that move out will be sustainable. Rewarding residents for any savings, however small, will not encourage progress toward this goal.

Given the high cost of shelters, the program would seem to offer the potential for being self-financing through savings from earlier move outs. Some understanding of the savings associated with earlier move outs would be useful for considering what level of funding might be available from program savings both for administrative costs and for bonus payments. Specifically, how much earlier would residents have to move out to make the program break even, and is that a reasonable goal? Given the low costs of the program and the potential substantial programmatic savings, it would seem like even slightly shorter shelter stays would pay for the program.

“Save at Home”: Building Emergency Savings One Mortgage Payment at a Time

The goal of this proposed program is to address the need for emergency savings that is common among new homebuyers who often tap out savings in purchasing their homes, face higher housing and other living costs after purchasing, and tend to underestimate the likelihood of experiencing a financial shock in the future. While the emergency savings fund, referred to as a Mortgage Reserve Account (MRA), could potentially be used to address a range of needs, the primary purpose is to provide a cushion that will help homeowners meet their monthly mortgage obligations and avoid a default. In short, an MRA is a savings account that is funded through monthly contributions made as part of routine mortgage payments that would be available to meet emergency financial needs. The savings account could also be seeded with funding at closing to provide an immediate cushion while monthly contributions build up over time.

While the program could be incorporated through a variety of mortgage lending channels, the paper notes that state housing finance agencies (HFAs) provide a ready-made system for such a program as it would build on their existing lending programs. The primary option for collecting the savings contributions would be to adapt existing loan servicing systems that are already used to devote a portion of monthly payments to fund escrow accounts for property taxes and insurance. The MRA would simply represent another line item in this monthly bill collected through the existing loan servicing system. While any lender could adapt servicing systems for this purpose, state HFAs have the additional advantage of incorporating housing counseling into many of their lending programs, providing an opportunity to educate borrowers on the benefits and operations of MRAs. HFAs also offer reduced interest rates, down payment assistance, and more flexible underwriting, all of which could be used as carrots for inducing borrowers to participate in this program.

There are a number of significant strengths of the proposed program. The existing servicing infrastructure provides a ready-made opportunity for collecting monthly contributions in an efficient manner that will overcome the behavioral barriers to following through on savings intentions. There is also good reason to believe that take up rates would be relatively high both because borrowers should be receptive to the need to protect the significant investment they are making in their home while the HFA lending system also provides a number of significant carrots that to further induce borrower participation. Finally, the wave of foreclosures from recent years has made lenders wary of extending loans to borrowers with limited financial means, and particularly those with small down payments and limited savings. The MRA would provide a mechanism for reducing lending risk and could help to expand borrowing options for these borrowers.

However, while the general framework of the proposed program is straightforward, the paper lays out the many complex programmatic details that need to be worked through in implementing the idea. In assessing both the appeal and feasibility of the proposal it becomes obvious that the ‘devil is in the details.’ A critical issue is what the process will be for accessing savings in the account. At one extreme the borrower could have unlimited access to use the funds as they see fit, relying on the segregated nature of the funds to encourage homeowners to preserve their use for true emergencies. At the other extreme, withdrawals could be limited to predefined hardship situations requiring documentation and approval by a third party. The choice of which process to use depends on the balance to be achieved between ease and cost of administration, the benefits of the savings as perceived by the borrower, and the likelihood that the savings will be available to prevent mortgage default.

Another challenging question is what the target savings amount should be and how long it should be maintained. Setting the savings goal will depend in part of how the goal of the program is defined. If the ultimate goal is to protect against mortgage default, then something on the order of 3 to 6 months of monthly payments may be appropriate. Alternatively, if the fund is intended to address unexpected costs for home repairs or medical emergencies than a larger fund may be needed.

One appealing aspect of the program is that depending upon how complex the process is for accessing funds, there may be little administrative cost to running the program. The one aspect of the planned intervention that might require more ongoing costs is monitoring of monthly payments and intervention when they are not made. The paper describes a ‘light touch’ approach where monitoring is automated and follow up contact may be made by email or phone calls. Still, such a system will entail costs in design and management. Therefore, another set of important questions are what these costs are likely to be, what are options exist for funding them, and how significant the benefits of these additional features are likely to be.

Given the range of options for structuring the program and the implications of these choices for both the rate of take up by borrowers and the program’s success in warding off mortgage default, it would be valuable to have a range of HFAs implement pilots to gain experience with different program configurations. So the ultimate question regarding this proposal is what types of incentives might be needed for HFAs to be willing to invest in piloting a program of this nature?

Expanding Economic Opportunity through Renter Equity

The proposals presented in the first two papers of this panel were embedded in housing programs because the fundamental aim of the interventions was to take advantage of the participants’ change in housing situation to provide motivation for participating in the program and also to generate savings specifically aimed at remedying a housing need. In contrast, the final paper in this panel presents a savings program embedded in a housing program, but it neither aims to use housing needs as a motivating factor for program participation nor seeks to generate savings for a specific housing purpose. Instead, the program’s goal is to leverage opportunities for lowering the cost of managing rental housing as a source of savings for low-income residents. Bit beyond its aim of fostering savings, the program also has the goal of helping to improve living conditions in the rental development and the surrounding community through greater resident engagement.

The paper presents actual experience running the savings program since 2002 in three assisted housing developments in the same neighborhood in Cincinnati. The program seeks to reduce the cost of managing the development by reducing tenant turnover (which incurs costs for lost rent and unit repairs), fostering greater responsibility among tenants for maintaining property conditions, and giving residents tasks to complete that would otherwise have to be paid for. Residents received credits for timely rent payments, participation in regular resident engagement meetings, and completion of work assignments around the property.  To discourage turnover, the credits increase in value with the length of residency and may only be cashed out after five years. However, the tenants can borrow against the credits beginning six months after they move in, with the loan cap rising from one month’s rent during the first year to two month’s rent through the fifth year of residence. After that residents can either withdrawal what they need in cash or borrow up to 80 percent of their balance.

The strength of this approach is that it provides low-income households with an opportunity to essentially generate additional income that can be tapped for emergencies as needed. The efforts required to generate the savings are fairly easy to accomplish, including timely rent payment, attending regular resident meetings, and completing simple jobs. The program has the added benefit of enhancing community engagement and potentially improving the quality of life in the development and surrounding area.

The goal of the paper is to document the approach that has been developed in Cincinnati in the hope of spreading the concept to other rental housing developments. While it appears to have been quite successful in Cincinnati, there are questions about what conditions are needed to spread this concept elsewhere.  One means of helping to facilitate adoption of the program would be to provide more detailed information on the costs of running the program, including the savings realized from reduced turnover, the reduction in reduction in project management costs, and the value of renter credits that have accrued. The Cincinnati example notes that a charitable gift was used to establish a reserve fund for the renter equity credits as a backstop for bank financing for the project. A key ingredient for replicating the program will be to understand the extent of financial obligations that are incurred and what the risks of having to tap reserves would be. If the program is to be brought to any scale the reliance on charitable contributions would have to be minimized and so the financial viability of the model absent subsidies needs to be documented.

A similar concern exists regarding the loan option offered to residents. In the Cincinnati case the loan fund is run by local, socially conscious investors who do not charge interest and rely on a modest loan fee to pay for administrative costs. The loan fund is a vital component of the program, providing residents with a means of tapping their savings during the first five years of residency to meet emergency needs. But if these funds have to rely on investors that do not expect to earn interest on their funds it will be hard to replicate. If the renter credits are backed by a reserve fund, it would seem possible to design a loan product that is made affordable by the availability of the guarantee provided by the reserve fund.  Partnering with a local community bank or loan fund to develop an approach that covers the cost of funds would be an important step in allowing the program to be adopted elsewhere.

One way in which the program seems like it could be tailored to promote the long term financial health of the residents is to incorporate some form of financial counseling, coaching or peer support groups to address broader savings goals.  As structured, the program does not attempt to encourage any additional savings or to improve financial habits among residents that would enable them to make more timely rental payments or to build up savings outside of the program. While these efforts would add costs to the program, or could extend the benefits it produces.

Finally, given the reliance on greater resident engagement and the use of resident labor, the program would seem to represent a substantial change in management approach. For this reason it would also seem that the program would be most likely to be adopted by mission-driven organizations that are significantly motivated to provide opportunities for enhancing the financial well-being of their tenants and have substantial financial capacity to manage the funds needed to back the renter savings accounts. What is likely to be needed to test whether this model can be spread outside of the Cincinnati developments where it was incubated is to engage with some of the larger, more sophisticated owners of assisted housing to see if they would be willing to implement a pilot.

Concluding Observations

The case statement for the Salon identified six key factors that proposals would be evaluated against. The first is scale, including whether the program holds the potential for reaching a significant number of individuals and be widely adopted and whether it is likely to have a lasting impact on savings behavior over time. Two of the other dimensions identified are inextricably linked to the question of scale, including whether existing organizations have the capacity to launch the strategy and whether the cost of implementing the approaches can be covered with sustainable resources. By these dimensions both the homeless and homeownership programs presented at the Salon hold real promise. Both build on existing institutional capacity of organizations operating in these areas and require little additional costs to implement. The main factor these proposals need to overcome is evidence of effectiveness since neither has yet to be implemented. Hopefully, the experience in New York with the Emergency Savings Program will yield some insights into how best to structure this effort and whether enhanced savings behavior by participants can yield savings sufficient to cover the costs of providing support and financial incentives to save. Similarly, if one or more state HFAs implement the mortgage reserve account idea it would go be of great value in testing this proposal and helping to validate its potential for fostering savings while expanding lending options for low-income homebuyers.

While also holding great promise, the rental equity concept faces more significant challenges in addressing issues of scale, institutional capacity, and costs and sustainability. The experience in Cincinnati appears to have had some significant success, generating $143,000 in savings for low-income renters over a 10-year period. However, the program has benefited from charitable contributions to backstop both the renter savings and the loan program that seems so critical to the program’s effectiveness. The unique nature of tenant engagement may also require a unique set of skills and motivations by property owners to implement. But given the limited options for supporting savings among low-income renters and the creative way the program generates income to support these savings the proposal deserves consideration for expanding to other sites. If the program is to be brought to scale, the feasibility of supporting the renter savings credits absent subsidy needs to be documented. Likewise, the associated loan program needs to be retooled to generate sufficient income for lenders to be willing to offer this product.  But both of these necessary steps may be feasible. Perhaps a large, well-capitalized, national organization with a mission of providing affordable housing that serves as a platform for upward mobility among low-income residents would be interested in adapting this concept in some of their developments to test whether the concept would work outside of the unique circumstances of the Cincinnati developments where the idea has already had some success.