Transitioning to Homeownership: Asset Building for Low- and Moderate-Income Households

Abstract

This paper assesses the asset building of households that take part in shared-equity homeownership (SEH) models. The contribution of this paper is a comparison of outcomes for households participating in shared-equity programs to other low- and moderate-income households who rent or own properties without restrictions on appreciation. We matched participants in SEH programs to households with similar characteristics from the Panel Study of Income Dynamic (PSID) over the 1997–2017 period. The findings indicate that in real terms, median SEH homeowners accumulated about $1,700 in housing wealth annually or around $10,000 during their holding period. This amount is lower than the $2,100 median annual gain in home equity experienced by similar PSID owners but statistically and economically significantly larger than the $16 in annual gain experienced by similar PSID renters. The findings provide evidence that households participating in SEH programs experienced positive, but modest, wealth gains that were slightly lower than homeowners in unrestricted units but substantially higher than renters.

Introduction

Homeownership is associated with positive household outcomes, including well-evidenced physical, psychological, and financial benefits (Dietz & Haurin, 2003; Herbert et al., 2013; Rohe & Lindlad, 2013). Financial benefits are grounded in the central role that homeownership plays in the United States (Kuebler, 2013). Wealth building is a key benefit of homeownership and occurs when the home appreciates, or when additional value from capital improvements is gained. Furthermore, homeownership functions as a forced savings mechanism through paying down the principal balance on the mortgage. In cross sectional data, there is a strong association between homeownership and wealth. The median net worth of homeowners was $255,000 in 2019 compared to $6,300 for renters (2019 Survey of Consumer Finances). The role of homeownership in wealth accumulation is particularly vital for lower income households since it is often their primary vehicle for asset building (Galster & Santiago, 2008; Herbert & Belsky, 2008). For homeowners with gross household incomes below $60,000, which is approximately the median household income, the median share of their net worth from home equity is 72 percent compared to 35 percent for households with income of $60,000 or more (2019 Survey of Consumer Finances).

However, the benefits of homeownership are not equally distributed and people of color and lower income households overall do not experience the same gains as higher income and white households (Cortes et al., 2007; Flippen, 2004; Freeman, 2005; Reid, 2005). For example, among homeowners with gross household incomes below $30,000 in 2019, which is about 50 percent of median household income, median home equity was $78,000 compared to $146,000 for households that earn $60,000 or more (2019 Survey of Consumer Finances). Furthermore, people of color do not benefit from homeownership to the same extent as white households; in 2019 wealth from homeownership for white households was $130,000 compared to $66,800 for Black or African American homeowners and $95,000 for Hispanic or Latinx homeowners (2019 Survey of Consumer Finances). Extant research, which is explored in further detail in the literature review section, describes the conditions (e.g., loan terms, interest rate, rate of appreciation, etc.) under which asset building from homeownership is the most or least effective, particularly for low- and moderate-income households.

In addition to the unequal distribution of wealth accumulation among homeowners, homeownership remains out of reach for many low-income households. Access to homeownership is impacted by borrowing constraints: wealth, income, and credit quality (Acolin et al., 2016; Barakova et al., 2003; Ehlenz, 2019; Jacobus & Davis, 2010) and the large wealth gap has only grown in the post Great Recession recovery (Acolin et al., 2019). Barriers to entry due to increased costs of living and lack of affordable housing options have priced many households out of the market. Rising costs make it difficult to save for a down payment. Some renters may be discouraged from applying for a mortgage due to a lack of information and misconceptions about how much is required for a down payment or concern about personal debt such as student loans (Acolin et al., 2016).

Non-traditional programs, such as shared equity homeownership (SEH) programs, are an important tool in asset building to provide entry into homeownership for people of color and lower income households who otherwise would not become homeowners (Jacobus & Davis, 2010; Thaden et al., 2013; Ehlenz, 2018). These programs aim to create long-term affordability of homes with a one-time community investment, by producing units at a reduced price that are purchased by low- and moderate-income households with restrictions on the resale value to ensure they remain affordable for future lower-income homebuyers (Wang et al., 2019).

SEH programs include models that aim to share the risks and advantages of homeownership while offering three key benefits: long-term affordability, wealth creation, and community stabilization and displacement prevention (often in response to gentrification) (Carlsson, 2019; Temkin et al., 2013; Thaden et al., 2013). The main types include community land trusts, limited-equity cooperatives, and deed-restricted units. Although the prevalence of SEH models has increased, alongside research about their impacts, there continues to be a gap in our understanding of the extent to which SEH models support wealth building for lower income homeowners. A prominent critique remains that, by design, resale restrictions limit the amount of equity SEH homeowners capture (Diamond, 2009; Carlsson, 2019; Ehlenz, 2019). This in turn impacts the household’s ability to build wealth. Yet, this criticism does not typically consider the amount of wealth accumulated if households remained renters, instead focusing on comparison to market-rate purchases—something which is often not feasible to lower-income buyers, particularly in high-cost markets.

In this paper we present a comparison of outcomes for households participating in shared equity programs to other low- and moderate-income households who rent or own properties without restrictions on appreciation. In doing so, we assess wealth building of households that enter SEH, guided by the objective of establishing asset building among participants in shared equity programs. Few empirical studies currently exist on the performance of SEH programs despite their size and increasing popularity in recent years. To our knowledge, this is the first study that compares wealth accumulation over the last two decades for participants in different types of shared equity programs to outcomes of renters and homeowners with similar characteristics. We compare annual home equity accumulation for SEH homeowners compared to annual home equity accumulation for homeowners and other wealth accumulation for renters in the Panel Survey of Income Dynamics (PSID) using propensity score matching and regression controlling for individual characteristics. The results indicate that, while homeowners in the PSID have an advantage with regard to wealth accumulation, SEH homeowners accumulate home equity statistically and economically significantly more than renters accumulate wealth. Due to sample limitations, we are not able to provide estimates for different SEH program types or for different racial and ethnic groups.

We begin by providing a brief overview of SEH which is followed by a discussion of the existing literature on determinants of wealth accumulation from homeownership, and SEH and wealth accumulation. We then describe our data and methods to estimate change in housing wealth by SEH owners and compare it to change in housing and non-housing wealth by PSID owners and renters. Finally, we conclude with our analysis and the discussion of the study’s results showing the potential of SEH programs in supporting wealth building for low- and moderate-income households.

Background on Shared Equity Homeownership

The original core tenets of shared equity models—those of community control, affordability, land stewardship, and shared equity—found their roots in Indigenous knowledge, working class movements, and movements for racial justice (Carlsson, 2019). For example, one of the first community land trusts was established in 1969 in southern Georgia to create access for Black farmers who faced barriers to land ownership (Carlsson, 2019). While recent work has pointed to the ways that some SEH programs have reduced the emphasis on community control since the 1980s, many of the other characteristics remain integral to SEH models: the housing is owner-occupied, equity is shared, resale is restricted, and in the case of CLTs, a stewardship organization oversees the land (Davis, 2017; DeFilippis et al., 2018).

Shared-Equity Homeownership programs and resale restrictions

There are different types of shared equity models, the most common of which are community land trusts (CLTs,) limited equity housing cooperatives (LECs), and deed restricted housing units. In CLTs, the homeowners own the home and lease the land through a long-term lease for a minimal fee from the community land trust, essentially removing land costs for residents. CLTs have members who vote for a board of directors, which are composed of CLT homeowners, public officials, and residents of the surrounding area (Carlsson, 2019). Importantly, CLTs understand land as a public asset, not a private good, which guides the principles of community control (Choi et al., 2017). However, it can be difficult for a community to establish a CLT and start acquiring properties because it requires the organization to have enough capital to acquire and maintain ownership of the land, widespread support of surrounding neighbors and public officials, private lenders that are open to providing mortgages for homes on leased land, and subsidies that remain in the home and cannot be recaptured (Davis, 2017; Thaden, 2012). As of 2018, it was estimated that there were about 225 CLTs nationwide providing over 12,000 units and this number has been growing in recent decades (Thaden, 2018).

Limited-equity housing cooperatives consist of low-income residents who own shares in a corporation that owns the deed to the building in which they live. LECs are generally formed through the conversion of rental apartment buildings (either market rate or subsidized), with most of them located in New York City. The co-op owns the building through a blanket mortgage and residents are simultaneously shareholders, leaseholders, and voting members of the corporation that own the building (Carlsson, 2019). The structure of LECs mean that they have a lower-entry threshold than CLTs since no capital is needed for an organization to purchase the land, but simultaneously they have less of an impact on affordability since land costs are covered by being divided between the owners. LECs can also be difficult to sustain because there is not necessarily an organization overseeing the long-term management of the program (Ehlenz, 2018). As of 2018, estimates indicated that about 167,000 units remained available in LECs mostly formed in the 1960s and 1970s (Thaden, 2018).

Another form of SEH programs that has been growing in popularity in recent decades is deed-restricted housing. Deed-restricted housing units are generally produced as an outcome of inclusionary zoning regulations that require a certain share of for-sale units to be reserved for households below a certain income threshold (generally for terms of at least 30 years). When the builders provide these units, they provide a form of affordable ownership options that are integrated as part of a larger, generally market rate, development that contributes to mixed-income communities (Thaden & Wang, 2017). While comprehensive figures on the number of deed-restricted homeownership units are lacking, in a recent survey Thaden and Wang (2017) find that at least 50,000 deed-restricted units have been produced with the majority being constructed within the last two decades.

Shared equity programs commonly include resale restrictions to enable long-term affordability. In particular, resale restrictions ensure that the initial subsidy offered through the shared equity program outlasts one participant/household. The resale restrictions limit how much an owner is able to receive when selling. In most cases, programs have specific provisions to ensure owners are able to recoup the value of capital improvements during their ownership period (for example, if they add an expansion or make substantial renovations).

Different formulas are used including: the appraisal-based formula which accounts for changes in the value of the structure; indexed formula based on changes in area median income or consumer price index (CPI); and fixed rate formulas based on the holding length. These formulas may contribute to dissociating the resale value from the market value, smoothing volatility in market prices for participants while maintaining affordability over the long run. For example, Champlain Housing Trust’s CLT program uses an appraisal-based approach in which owners receive 25 percent of the estimated appreciation based on market appraisals at purchase and resale. Other programs adjust the percent of appreciation the seller receives based on duration (longer stays are associated with a higher share of appreciation). (Burlington Associates, NA). The appraisal-based approach exposes households to some housing market risk since their appreciation will depend on local market conditions.

In the case of participants in programs with indexed formulas, appreciation is based on the change in the index over the holding period. For example, OPAL Community Land Trust uses a formula based on the CPI calculated by the Bureau of Labor Statistics for the Seattle-Tacoma-Bremerton area in which it is located. To illustrate, over the 1999–2012 period, an owner would experience 38.1 percent appreciation based on the indexed formula compared to a 93.3 percent increase in appraised value over the same period (OPAL, NA).

Programs that adopt a fixed rate formula have the most predictability for households since they know upfront what their rate of appreciation will be as it is independent of market dynamics. For example, Homestead CLT applies a 1.5 percent annual growth rate so that after 10 years, a household who purchased a home for $200,000 would sell it for $232,108 (Homestead, NA).

The generosity of the resale formula can lead to more or less appreciation going to the participants but more generous formulas would require ongoing subsidies to maintain the affordability of the unit. By design, program participants are expected to experience lower levels of appreciation on average than if they owned unrestricted units. However, resale formulas may smooth market fluctuation and offer more stable returns, particularly during housing price declines.

Strengths and challenges facing SEH models

The greatest strengths of SEH models are three-fold: they preserve long-term affordability, increase access to homeownership for households often excluded from the market, and provide opportunities for wealth accumulation. Shared equity programs have been successful in increasing access to homeownership to low-income households, addressing wealth gaps, and reducing foreclosures (Jacobus & Davis, 2010; Temkin et al., 2010). Notably, the Champlain Housing Trust, a SEH program which developed 424 single-family homes and condos between 1984 and 2008, demonstrated overwhelming rates of residential stability. Occupancy, use, and resale controls remained for 97 percent of these units with only nine foreclosures over 25 years (Davis & Stokes, 2009). In some cases, affordability actually improved upon resale: in the Champlain Housing Trust, the average home was affordable to a household at 56.6 percent of the area median income (AMI), but on resale it was affordable to a household at 53.4 percent of the AMI (Davis, 2017).

Further, initial subsidies offered for SEH projects benefit future generations of homeowners. Research has demonstrated that many SEH homeowners are able to purchase market-rate housing after their initial SEH purchases which makes a subsidized property available for a new lower-income household (Davis & Stokes 2009; Temkin et al., 2010). This stability also contributes to the overall sustainability of SEH programs (Temkin et al., 2010). In addition to individual household benefits, research has demonstrated community benefits of SEH programs. SEH models can play a role in stabilizing neighborhoods, reducing displacement, and limiting speculation of property values, thereby slowing impacts of gentrification (Choi et al., 2017). Shared equity programs also result in quality housing and long tenures (Temkin et al., 2013). The provision of long-term affordability, while simultaneously providing access to homeownership for low-income households, is what makes SEH models particularly unique (Ehlenz & Taylor, 2019). Moreover, the SEH sector is increasingly serving a more diverse set of households in terms of race and ethnicity (Wang et al., 2019). Together, long-term affordability and access to homeownership contribute to the third strength of SEH programs: they offer opportunities for wealth accumulation. SEH programs, community land trusts in particular, facilitate sustainable wealth building for households (Thaden et al., 2013).

However, SEH models are not without challenges and are often critiqued for three key reasons. First, SEH programs, particularly CLTs, require substantial initial capital to be established, often in the form of subsidies, which has limited the scalability of their model and constrains their impact (Green & Hanna, 2018). Second, SEH programs face policy and government barriers, with a lack of supportive policies at the national level to provide mortgage credit to these non-traditional forms of ownership and lack of supportive zoning policies at the local level. Third, and one we address in this paper, critics point to the limitations on household wealth building through resale restrictions, particularly within a longer historical context of excluding wealth building opportunities to people of color through redlining practices (Ehlenz & Taylor, 2019). The home’s appreciation at resale is identified as one of the principal drivers of wealth accumulation from owning SEH units (Jacobus, 2007; Ehlenz & Taylor, 2019). The literature on the determinants of wealth accumulation from homeownership and the impact of shared equity models on wealth accumulation is explored in the next sections.

Determinants of Wealth Accumulation from Homeownership

Different factors and conditions have been shown to make homeownership a more or less effective tool for asset building, especially as it pertains to people of color and lower-income households. These include the rates of appreciation of the units, the housing price cycle, the forced saving mechanism associated with repaying mortgage principal based on financing and refinancing options, the ability to capture tax benefits, and home maintenance and capital improvements.

Appreciation occurs when the value of the home increases over time. Local housing market factors impact the price cycle and home appreciation. For example, in areas that have a restricted housing supply due to development regulations, as demand outpaces supply and drives down home inventory, prices will increase. Macroeconomic factors may also impact rates of home appreciation. The Great Recession caused by the housing bubble, and subsequent burst and collapse of the financial markets, led to significant depreciation in home values. Thus, the timing of when the household enters and exits the market is essential (Di et al., 2003; Di et al., 2007). However, the timing of entry and exit may matter more for some households than others. Newman & Holupka (2016) demonstrate macroeconomic conditions for first-time white homeowners have a direct impact on gains and losses of net worth during periods of expansion and decline, but for first-time Black homeowners, total net worth declined regardless of economic conditions.

The duration (length of time the home is owned) also impacts the cost of homeownership. If the home is sold before the gain in appreciation is greater than closing costs,¹ the homeowner experiences a negative return on their investment. There is evidence that lower income households experience shorter holding periods and may sell more often before experiencing sufficient appreciation levels to cover closing costs (Belsky & Duda, 2002). Longer holding periods lead to more wealth accumulation by moderating the cyclical nature of housing markets, particularly for lower-income households since they lack opportunity to invest elsewhere (Di et al., 2003; Di et al., 2007). Longer holding periods also allow a larger share of the principal to be paid down, particularly since interest payments dominate the early years of a mortgage. Unfortunately, lower-income households tend to have higher rates of residential mobility and own a home for shorter periods of time due to financial insecurity. Reid (2004) found that more than half of low-income households exited homeownership within five years of their first purchase whereas only about one-fourth of higher-income households exit within that time frame.

Increased rates of appreciation clearly expand wealth accumulation, however, moderate- and low-income homebuyers do not necessarily experience the same gains as higher income households do from appreciation (Belsky et al., 2005; Bostic & Lee, 2008; Newman & Holupka, 2016; Rappaport, 2010). Higher valued homes, and the neighborhoods they are in, build greater appreciation, while highly segregated minority neighborhoods stifle housing appreciation (Flippen, 2004). Galster & Santiago (2008) found that homes owned by white households appreciated twice the rate of homes owned by Black and Latinx households. Higher-income white households report greater home values and experience some of the largest gains in housing wealth appreciation from forced savings and annual appreciation, compared to people of color and lower-income households (Boehm & Schlottmann, 2004). In some cases, stagnant appreciation of lower priced homes negates the asset building potential from homeownership altogether (Reid, 2004; Shlay, 2006).

The amount available for a down payment can impact the type of loan and financing terms a borrower may be able to secure. People of color and lower-income households are less likely to have sufficient resources for large down payments (Bostic & Lee, 2008). Often, these households do not get help from extended family or have limited access to other resources for the downpayment to enter homeownership (Hall & Crowder, 2011). When households have little or no equity in their home, they are at more risk of default when a downturn in a market occurs (Galster & Santiago, 2008). The financing terms of the loan impacts wealth accumulation. Higher interest rates, higher origination fees, and riskier loan products, such as interest-only and negative amortization loans, depress asset building potential from principal payments that represent an important forced saving mechanism (Galster & Santiago, 2008; Grinstein-Weiss et al., 2013). Lower-income households are more likely to have higher-cost and riskier loans, which are predominant in lower-income minority communities (Belsky et al., 2005; Bostic & Lee, 2008; Rappaport, 2010). Low-income homeowners are also less likely to take advantage of refinancing, which can substantially reduce costs associated with homeownership when interest rates fall (Belsky et al., 2005). In 2019, only about 18 percent of homeowners with income below $60,000 had a primary mortgage that was refinanced compared to 34 percent for households with income of $60,000 or more (2019 Survey of Consumer Finances).

Low-income homeowners are less likely to benefit from the tax benefits provided to homeowners. In particular, they are rarely in a position to take advantage of deductions for mortgage interest payments and real property taxes since the standard deduction often exceeds the total amount from itemized deductions (Belsky et al., 2005; Herbert & Belsky, 2008; Rappaport, 2010). Home maintenance and capital improvements in a property protects the investment and can increase a home’s value. The condition and age of the home subsequently impacts the cost of homeownership. Lower income households are more prone to purchasing older homes that often require substantial maintenance since they can only afford low value homes. Often these homeowners do not have the disposable income needed for unexpected repairs, preventive maintenance, and capital improvements, which can deflate the home’s value. Higher income households are more likely to make capital improvements and undertake expansion projects (Baker & Kaul, 2002; Mendelsohn, 1977). Van Zandt et al. (2011) show that it is common for low-income homeowners to face unexpected costs and home repairs they are not able to afford, affecting their ability to maintain and keep their home. Furthermore, Boehm & Ihlanfeldt (1986) found that homeowners in neighborhoods with dilapidated structures were less likely to engage in maintenance and improvement expenditures.

Overall, the extant literature on wealth accumulation indicates asset building is generally positively and significantly associated with homeownership. The average net wealth of owners is 2.2 times that of renters (Di et al., 2003). Although lower-income households do not accumulate as much wealth as higher-income households, they experience higher returns from owning than renting (Turner & Luea, 2009). Herbert et al. (2013) estimate low-income households accumulate approximately $10,000 each year compared to low-income renters. Even during a period when home values are depreciating, lower income homeowners experience greater gains in wealth accumulation than renters (Grinstein-Weiss et al., 2013).

Homeownership does not always lead to more wealth than renting, especially for lower income and minority households (Boehm & Schlottmann, 2004; Galster & Santiago, 2008). Wainer & Zabel (2020) found that lower-income households who entered homeownership between 2001– 2007 experienced little to no gain in wealth accumulation by 2013 compared to renter households, although this was an exceptional period marked by an historical housing boom and bust at the national level.

Shared Equity Homeownership and Wealth Accumulation

SEH programs may moderate some of the adverse effects that stifle wealth accumulation for lower income homeowners. For example, over 91 percent of shared equity homeowners in a study by Temkin et al. (2010) remained homeowners after five years, well above the national 50 percent norm for low-income, first-time homeowners. This is complemented by risks such as foreclosure or mortgage defaults being mitigated by the stewardship organization, that can work to delay those events, facilitate short-sales, provide direct grants or loans, and offer homeownership counseling to identify potential loan modifications (Carlsson, 2019; Davis & Stokes, 2009; Thaden and Rosenberg 2010). Shared equity homeowners also resell for similar reasons, and at similar rates, as other homeowners, and many shared equity homeowners purchase a market-rate property after selling their shared-equity home (Temkin et al, 2010). Further, in some shared equity models, the buyer’s mortgage is reduced through a subsidy provided by the stewardship organization (Thaden et al., 2013). Do these benefits offset the financial impact of resale restrictions?

As discussed above, resale restrictions in SEH programs often limit the resale price of the home. These restrictions serve to preserve the long-term affordability of the home for generations of low-income homeowners. However, there remains significant debate about the impact of resale restrictions on low-income households participating in shared equity programs.

Several studies have compared individual wealth accumulation for different types of shared equity homeownership models (e.g., Davis, 2017; Davis & Stokes, 2009; Jacobus & Davis, 2010; Jacobus & Maxwell, 2019; Temkin et al., 2013; Wang et al., 2019). Some of these studies have evaluated resales from only one program (Davis & Stokes, 2009; Jacobus & Davis, 2010; Jacobus & Maxwell, 2019),² while other studies have compared multiple shared equity models and affordable homeownership programs (Davis, 2017; Jacobus, 2007; Temkin et al., 2013; Wang et al., 2019).³ All these studies demonstrate that, on average, at the time of resale, shared equity homeowners recoup their downpayment, earn a return on their investment, and generate more wealth from homeownership than they would have received if they stayed renting based on modeling the cost of renting, which deflates the argument that there is no financial gain captured from shared equity models.

Home price appreciation and net equity gain for different types of shared equity homeownership models vary based on a variety of factors such as date the property was purchased and sold, the holding period, original purchase price, location of the home, and the resale formula. On average, the annualized internal rate of return was over 25 percent for a homeowner reselling after five years (Jacobus, 2007; Jacobus & Davis, 2010); however, the range varies substantially based on the program from about 6.5 percent to 60 percent (Davis, 2017; Temkin et al., 2013). Furthermore, Wang et al. (2019) show that not only the program but the time period in which the homeowner sold the home mattered. The median gross appreciation during the housing boom period (2001–2006) was 3.3 percent, during the housing bust period (2007–2012) was 1.7 percent, and during the housing recovery period (2013–2018) was 0.5 percent. Despite the variation due to some of the factors referenced above, these studies consistently demonstrate that for the average household, the combined wealth accumulation over the holding period is approximately $13,000 to $16,000 (Davis, 2017; Davis & Stokes; Jacobus, 2007; Jacobus & Davis, 2010; Wang et al., 2019).

These foundational studies highlight that shared equity homeownership models generate wealth accumulation for moderate- and low-income homebuyers. Though, a key question remains: how do outcomes for shared equity homeowners compare to similar first-time home buyers? Theodos et al. (2019) addresses this question by using a similar methodology employed in the collection of wealth studies – difference-in-difference and propensity score matching approaches – but looks at financial health rather than wealth accumulation. Short-term financial health (credit scores and credit utilization rate as well as debt types and amount) and loan performance is evaluated for nine shared equity programs and similar first-time homebuyers in the same metropolitan region.⁴ S hared equity homeowners take out smaller mortgages, have lower monthly payments, and are less likely to have additional financing on their homes compared to similar first-time homebuyers (Theodos et al., 2019). There were no differences found regarding mortgage delinquencies which suggest that shared equity homeowners perform just as well on their mortgage as similar homeowners (Theodos et al., 2019). Expanding on this line of inquiry, our study specifically examines asset building and wealth accumulation outcomes for SEH models compared to similar low- and moderate-income households who rent or own properties without restrictions on appreciation.

Data and Methods

Two main datasets are used for this study to compare wealth accumulation by SEH homeowners relative to similar owners and renters over the 1997–2019 period. Information about SEH homeowners is obtained from the HomeKeeper National Data Hub and information about low-income renters and owners is obtained from the Panel Survey of Income Dynamics (PSID) (2020).

The HomeKeeper data is managed by Grounded Solutions Network, a national organization that supports SEH programs, and data is entered by individual participating organizations with detailed information about the property transaction and the household purchasing and selling the housing unit. Data from organizations with a range of SEH programs are included. The data is described in more details in Wang et al. (2019). For this study, we restrict the sample to households that purchased and resold a home between 1997 and 2019, with information about the purchase and resale price along with the amount of the mortgage balance at origination and resale in order to estimate total and annualized housing equity gain over the holding period.⁵ For SEH households, annual housing wealth gain is calculated based on the effective resale price minus the effective purchase price plus the difference between the mortgage balance at resale and at origination and the holding period.⁶ This provides a sample of 1,177 households. Table 1 provides a breakdown of the units included in the analysis. As shown, about three quarters of units are part of CLT programs, with the rest mostly split between LECs and units with deed restrictions. The units are present in all census regions with an over-representation in the Northeast with Champlain Housing Trust, one of the oldest and largest CLT programs in the nation, as the largest contributor. In terms of the resale formula, 85 percent of units are subject to the appraisal- based formula with the seller recouping a share of the appreciation (e.g., 25 percent in the case of Champlain Housing Trust ). Another 10 percent rely on indexed formulas, most commonly based on the CPI.

Table 1:

Breakdown of Unit Characteristics in SEH Sample

Program Type
Community Land Trust	73.0%
Limited Equity Cooperative	9.1%
Deed Restriction	15.7%
Other	2.2%
Region
Midwest	27.8%
Northeast	40.3%
South	6.7%
West	25.2%
Resale Formula Type
Appraisal-based	84.9%
Indexed	10.1%
Other	5.0%
N	1,177

Source: HomeKeeper National Data Hub

The PSID data is directed by the Institute for Social Research at the University of Michigan and began with a nationally representative sample in 1968 that has been supplemented since 1997 to include more recent immigrants, particularly Hispanic and Latinx households. The survey has been conducted every two years since 1999. The PSID has been one of the main sources of panel data for wealth studies to examine the relationship between tenure type and wealth accumulation and is recognized for including detailed questions about a household’s financial situation (Di et al., 2007; Herbert et al., 2013; Reid, 2004; Turner and Luea, 2009). Our unit of analysis consists of tenure periods, which are defined as the period of time between each move reported by individuals listed as head of household at both the beginning and end of the period.⁷ Therefore, if a household moved multiple times within the study period, each tenure period is treated as a separate unit. For owners, we limit our sample to households who purchased and resold a home between 1997 and 2017 for which wealth information is available, with at least one intervening survey wave between moves. For renters, we limit our sample to households who moved between 1997 and 2017 for which wealth information is available, with at least one intervening survey wave between moves in order to have information about starting and ending wealth. We also remove anomalous cases from both groups, such as cases where mortgage principal increases from zero to non-zero during the tenure period, and cases where zero non-equity wealth is reported at both the beginning and end of the tenure period.

For owners, our main measure of housing equity is estimated based on self-reported home value minus outstanding mortgage balance in the waves immediately succeeding the home purchase and immediately preceding the resale, annualized based on the time between these two surveys.⁸ For owners and renters, we also estimate a measure of non-housing equity wealth by adding up savings, pensions, and financial investments (but not wealth associated with businesses, farms, or additional real estate investments), less outstanding non-housing debts. For both owners and renters, we eliminate observations for which reported home value, mortgage, and other wealth components are missing in either time period.

The SEH programs are designed to serve low- and moderate-income households earning less than 80 percent of the AMI, primarily serving first-time homebuyers earning between 51–80 percent of AMI. They have therefore substantially lower income than the general population. Owners and renters participating in the PSID also differ substantially in terms of characteristics like age, income, household structure, and race/ethnicity.

In order to compare the three groups – SEH owners, PSID owners, and PSID renters – we implement a propensity score matching (PSM) approach using nearest neighbor matching.⁹ This means that in the matched sample, some of the PSID households are used several times if they are the best match for multiple SEH households. Tenure periods are matched based on age of household head, household income, household size, race/ethnicity of household head, and purchase period, as well as state-level or MSA-level (where applicable) median home value and income. PSID observations with income above $150,000 are excluded to improve the degree of matching based on income, given that among SEH owners only one observation had income above $145,000. The race/ethnicity variable is broken down into non-Hispanic white, non-Hispanic Black or African American, Hispanic, and Other.¹⁰ The locational controls at the metropolitan level (or state-level for units outside metropolitan areas) ensure that matched households face similar market environments in terms of house value and household income. All dollar amounts are expressed in 2019 real dollars.

Using this technique, we match each combination of groups: 1) PSID owners and SEH owners; 2) PSID renters and SEH owners; and 3) PSID renters and PSID owners. We focus on the first two models in the analysis and present the results for M odel 3 in Appendix A. For M odels 1 and 2, we match with replacement so that an individual in the control group (PSID owners or renters) can be matched to more than one individual in the treatment group (SEH owners) in order to improve the quality of the nearest match and ensure the treatment and control groups are as similar as possible. Model 3 matches PSID owners and PSID renters that were successfully matched to SEH households in the first two models, in order to assess the differences between renters and owners that are demographically similar to the SEH sample.

Our analysis focuses on the differences in median annual changes in home equity for SEH and PSID owners and annual changes in non-equity wealth for renters, calculated by dividing the total change in these wealth measures by the duration of tenure.¹¹ We would have preferred to observe changes in non-equity wealth for SEH owners as well but do not have access to that information. We rely on the assumption that SEH owners will not experience absolute declines in non-housing wealth. Based on this assumption, our comparison between equity wealth growth for owners and non-equity wealth growth for renters demonstrates the extent to which the wealth of owners increases over and above the growth in non-equity wealth. By comparing PSID owners and renters in Appendix A, we are able to show that this assumption holds for PSID owners and that the comparisons of changes in non-equity wealth for renters and home equity of overall wealth for owners show similar patterns.

We estimate quantile regressions incorporating both matching and survey weights (Conley and Galenson 1994; Koenker and Hallock 2001). Matching weights account for the matching with replacement in the models comparing SEH and PSID households so that PSID households that are used as a match for several SEH households are properly accounted for. Survey weights for PSID households account for the longitudinal structure of the survey that has led to a differential representation of different socio-demographic groups based on income, race and ethnicity, and immigration status in particular.

The PSM approach we adopt does not allow us to estimate the causal impact of participating in SEH programs on wealth accumulation since participating households are not assigned randomly and the PSM does not account for differences in propensity to save or in familial resources between SEH owners and PSID owners and renters. Rather, the PSM allows us to compare actual wealth changes for SEH owners and PSID owners and renters with similar sociodemographic characteristics and locational contexts.

Results

Figures 1a-1b display the change in absolute standardized mean difference (SMD) from before (right) to after (left) the matching process for each model, showing that our matching strategy effectively balances the treatment and control groups and that the SMD falls below the 0.1 threshold commonly used to define adequate matches for all variables in all three models except for household size in Model 1.¹² In particular, the matched households have similar age, income and move- in year as well as similar regional median income and home value, factors that have been shown to be key in impacting wealth accumulation through homeownership.

Figure 1: Absolute Standardized Mean Difference.

a- SEH Owners vs. PSID Owners

b- SEH Owners vs. PSID Renters

Table 2 presents descriptive statistics for the three matched samples used as part of the comparison between SEH owners and PSID households. Overall, the distributions across samples confirm that the three post-matching samples are well balanced across key variables that have been associated with wealth. Degree of balance is calculated via the absolute standardized mean difference (SMD)¹³ between the control and treatment groups for each of the matching variables. A well-matched variable generally has an absolute SMD less than 0.1, indicating that the mean difference between the treatment and control groups is one-tenth the size of the population standard deviation. The median household income of the post-matching samples is between $43,000 and $44,000, substantially lower than the area median income ($57,000 to $58,000), reflecting the fact that SEH programs serve primarily low- and moderate-income households earning between 60 and 80 percent of their area median income. The average household head is between 35 and 36 years old, since SEH participants are often first-time homebuyers. In terms of racial/ethnic composition, the sample is disproportionately white Non-Hispanic/Latinx (79 percent of SEH households) with only about 5 percent Black or African American households, and 5 percent Hispanic or Latinx households.¹⁴ The average household size is 2.1 to 2.2 and the average length of tenure is slightly longer for SEH households (6.2 years) than for matched PSID Owners (5.9) and Renters (5.1). This difference is substantially smaller than in the unmatched data where the average length of tenure is 5.4 years for owners and 3.1 years for renters, illustrating the stability of SEH homeowners discussed above.

Table 2:

Descriptive Statistics for SEH Homeowners and PSID Owners and Renters, Post-Match

	SEH	PSID Owners	PSID Renters
Median Housing Wealth	$2,802	$27,045	NA
Median Other Wealth	NA	$7,477	$2,934
Household Income	$43,135	$43,665	$43,135
Age	35.2	36.1	35.8
Black or African American	4.9%	6.1%	5.3%
Hispanic or Latinx	5.3%	4.6%	4.2%
Other Race or Ethnicity	10.4%	12.4%	12.5%
White Non-Hispanic/Latinx	79.4%	76.9%	78.1%
Household Size	2.2	2.5	2.2
Length of Tenure	6.2	5.9	5.1
Move-In Period 1997–2007	57.4%	49.1%	53.8%
Move-In Period 2008–2012	31.9%	50.2%	44.7%
Move-In Period 2013–2019	10.7%	0.7%	1.5%
Area Median Household Income	$57,921	$57,791	$58,007
Area Median House Value	$330,573	$341,973	$328,356
N	1,177	302	504
N Weighted	1,177	1,177	1,177

Source: HomeKeeper National Data Hub and PSID

Note: All dollar figures are in 2019 dollars and represent individual and location characteristics as of the move.

In terms of purchase period, most observations are for households that purchased or moved in between 1997 and 2007, prior to the Great Recession (49 to 57 percent). A substantial share purchased or moved between 2008 and 2019, during the Great Recession or its aftermath (32 to 50 percent), and while 10.7 percent of SEH households purchased and sold during the recovery period of 2013 to 2019, only 0.7 percent of PSID owners and 1.5 percent of PSID renters moved during this period, in part due to the right censoring of post-2017 moves. Previous work (Wang et al. 2019) has shown broadly similar levels of wealth building across these three periods, although slightly lower for those who purchased and sold during the recovery period. The higher share of SEH households who purchased and sold during the recovery period relative to the control groups, would therefore be expected to have a limited downward impact on the SEH estimates.

Our analysis focuses on differences in median rather than differences in means, given the relatively small sample size and the presence of observations with large annual changes in wealth that can substantially affect group averages. Differences in median, t-statistics, and 95 percent confidence intervals are estimated using quantile regressions with a dummy for whether matched pairs are in the treatment or control group (SEH owners relative to PSID owners and renters in M odels 1 and 2).

Table 3-a reports the difference in median annual change in housing wealth between the SEH and PSID owners. The median SEH owner experienced an annual increase in housing wealth of $1,657 in real terms compared to $2,079 for the median PSID owner. The difference of $423 is not statistically significant, but at about 20 percent of the estimated wealth gains for PSID owners, it is substantial. Over the average six year holding period, this means that a SEH household would have accumulated about $10,000 in housing wealth, compared to $12,500 for a similar PSID owner. These numbers are somewhat smaller than what has been found in the literature on overall owners, both because these households have characteristics (such as lower household incomes) associated with lower wealth accumulation, and because our study period includes a period of substantial housing wealth loss during the Great Recession.

Table 3:

Differences in median annual wealth change for matched comparison groups

a- SEH Owners vs. PSID Owners
	SEH	PSID Owners	Difference	t	95% Confidence Interval
Median Annual Change in Housing Wealth	1,657.7	2,079.5	−421.9	−0.7	−1,686.0 – 851.9
b- SEH Owners vs. PSID Renters
	SEH	PSID Renters	Difference	t	95% Confidence Interval
Median Annual Change in Housing Wealth vs Other Wealth	1,657.7	15.5	1,642.1***	25.9	1,518.1 – 1,766.2

Note:

^*p<0.1;

^**p<0.05;

^***p<0.01.

Table 3-b shows that overall wealth for matched renters in the PSID only increased by $16 a year in real terms, which is both substantially and statistically significantly lower than the median annual wealth change reported for SEH owners. This suggests that as long as the SEH homeowners do not lose substantial non-equity wealth during their tenure period, they were able to accumulate substantially more wealth than similar renters.

Appendix A compares the median annual change in housing wealth and other wealth by PSID owners and renters. The results show that the median PSID owner also added significantly more housing wealth than the median increase in overall wealth experienced by renters. While PSID owners also experienced a larger increase in non-equity wealth than renters, this only represented $224 per year; far less than the growth in home equity wealth. These results are supportive of the fact that the comparison of changes in housing wealth for SEH owners relative to the change in other wealth for PSID renters may understate rather than overstate the extent to which SEH owners experience higher overall wealth growth than PSID renters.

Discussion

Homeownership has been associated with household wealth building through the accumulation of equity generated by the forced saving mechanism associated with paying back one’s mortgage principal as well as long-term home value appreciation. At the same time, barriers to accessing homeownership, particularly in the form of borrowing constraints, can prevent households from entry for which it might be optimal to switch from renting to owning. Parental wealth, including wealth generated from homeownership, has been found to be an important factor in young households becoming homeowners. In addition, rising housing costs (for both types of tenure) have further complicated access to homeownership for low- and moderate-income renters with limited access to capital.

In this context, SEH programs have the potential to facilitate access to homeownership for low- and moderate-income households, for whom owning may be the desired tenure choice but who would not be able to afford a house that meets their needs without the SEH option. SEH programs are designed to maintain long-term affordability by limiting appreciation at resale. By design, this limits the amount of wealth building SEH homeowners will experience through appreciation but does not affect the forced saving mechanism channel.

Our results indicate that SEH owners experienced real median net home equity gains of $1,658 per year for owners who bought and resold their home since 1997. This means that a household who stayed six years in their house would have accumulated about $10,000. This is less than the $2,080 in median annual housing wealth gain or about $12,500 over six years experienced by similar owners in the PSID but still represents about 80 percent of that amount (and the difference is not statistically significant). By comparison, similar renters in the PSID sample only accumulated about $100 in real terms over that period and the difference in wealth gain between owners and renters is statistically significant.

The estimated amount of housing wealth built by SEH homeowners is more than 1.5 times the median net worth of renters ($6,000 as of 2019) and represents an amount sufficient to put a five percent down payment on a $200,000 house. For first-time homeowners, SEH programs can therefore help households start accumulating wealth and enable them to build their credit and downpayment to become homeowners of unrestricted houses, if they desire to do so in the future.

Overall, these results indicate that SEH homeowners are able to build real wealth through home equity gains (both from house price appreciation and repaying their mortgage principal) over their holding period. The level of housing wealth- building by SEH owners may be slightly lower than the appreciation experienced by similar owners of unrestricted properties (although not statistically significant), but it is substantially higher than the overall non-equity wealth built by renters. These conclusions indicate that SEH can be an effective wealth-building tool – particularly among low and moderate-income households that may have limited access to other homeownership opportunities – and can therefore serve as an effective substitute for unrestricted homeownership options with respect to home equity accumulation.

Further work is needed to examine whether these results hold for people of color. The historical lack of access to homeownership by people of color has been identified as one of the drivers of the persistent racial-ethnic wealth inequality gap, and increasing access to homeownership options is particularly important for these households. Our small sample size limits our ability to provide group-specific estimates, but future research should examine whether SEH owners of color experience similar levels of housing wealth building to white owners (which is not the case for owners of unrestricted units).

In addition, these results aggregate all organizations and program types. Future work is needed to examine potential variations across SEH program types and resale formulas. Finding whether certain program designs are most effective in supporting wealth building is particularly important for local governments and non-profit organizations that are considering expanding support for these programs.

Finally, differences in non-financial outcomes, such as neighborhood locations between SEH participants and similar owners and renters is also an area worth exploring further to examine whether such programs support access to different types of neighborhoods in addition to supporting wealth building.

Appendix A:

Differences in median annual wealth change for PSID Owners and PSID Renters that were Matched to SEH Owners

	PSID Owners	PSID Renters	Difference	t	95% Confidence Interval
Median Annual Change in Housing Wealth vs Other Wealth	2,079.5	32.5	2,047.1***	46.1	1,989.5 – 2,104.6
Median Annual Change in Other Wealth	223.5	32.5	191.0***	23.5	144.1 – 237.9

Note:

^*p<0.1;

^**p<0.05;

^***p<0.01.