“All roads lead to Rome?” Performance evaluation across different types of community land trusts based on a large-scale survey

ABSTRACT

Community land trusts (CLTs) seek to keep homes and other urban spaces permanently affordable and community controlled. As their number across the United States has increased, different iterations of the CLT model seem to have proliferated in practice, stoking scholarly debate as to their varying outcomes and benefits. There has, however, been little attempt to empirically measure the relationship between this institutional diversity and outcomes. Applying institutional theory to a national CLT dataset, we identify five main organizational sub-types of CLT: traditional CLTs, start-up CLTs, government-housed CLTs, nonprofits with a CLT/shared equity (SE) program, and adapted CLTs. Statistical tests confirm a high degree of similarity in operational scope, organizational capacity, and performance outcomes across the most prominent sub-types. The limited statistical differences which can be identified are consistent with known CLT and urban institutional development processes. Further studies might seek to determine how consequential such limited differences may be.

KEYWORDS:

• Community land trust
• permanently affordable housing
• historical institutionalism
• community control
• shared equity housing

Introduction

As communities in the United States (U.S.) face increasingly entrenched housing affordability crises, community land trusts (CLTs), which aim to keep homes and other types of space permanently affordable and community controlled (Davis, 2010; Ehlenz & Taylor, 2019), have attracted heightened interest, with an increase in the number of CLTs and CLT-like programs in operation (Wang et al., 2023). With this heightened activity, scholars have not only empirically examined the effects of CLTs on populations and communities, but they have also chronicled the evolutionary institutional development of CLTs themselves, documenting the emergence of significant internal heterogeneity in their structure as organizations over time, generating debate regarding the ability of CLTs to achieve different aspects of their mission. Specifically, scholars have suggested different CLT iterations may be more or less effective in accomplishing the model’s core mission of procuring diverse types of permanently affordable housing that is also community controlled. Nonetheless, there has been little attempt to comprehensively, empirically measure how performance, i.e., the ability to accomplish core housing and community control-related missions, might vary across different types of CLTs.

To intervene in this scholarly debate and address this critical gap in the literature, this study asks: are there significant differences with respect to scope, capacity, and performance across major “families” or types of CLT models? To answer this question, using the results of questions incorporated into a comprehensive survey of all U.S. CLTs and CLT-like, related Shared Equity (SE) entities, we first inductively identify five main organizational types of entities employing CLT models in operation today in the U.S.: traditional CLTs, start-up CLTs, government-housed CLTs, nonprofits with a CLT/SE program, and, critically, what we call an “adapted” CLT, which deviates from the traditional CLT with respect to its governance, with potential implications for community control.¹ For the three most widespread types, we also conduct a statistical analysis to determine if significant performance and outcome differences, as measured by other questions from this same survey, can be observed between them.

In general, we find far more in common across major sub-types than we find substantively statistically different between them, suggesting that perhaps many, if not all, institutional pathways of CLT development may lead to some fairly similar outcomes: all roads may, in fact, lead to Rome. These commonalities include similar residential portfolio size and tenure types, stewardship activities, affordability term, staff capacity, financial health indicators, and Diversity, Equity, and Inclusion (DEI) practices. Differences across the models appear to reflect the institutional development paths which have produced these varying types, consistent with existing literature on CLTs’ evolution, and as consistent with known urban institutional development processes.

Literature review

As CLTs have received increasing attention as a possible solution to the housing crisis, garnering high-profile coverage in outlets such as the New York Times (Fahy, 2023), scholars have continued to identify and debate their benefits and limits. Below, we apply existing theory on urban historical institutional development to CLTs’ emergence and evolution, before offering a detailed review on variations of CLT models, and empirical evidence on their effects.

The emergence and evolution of CLTs’ definitions in the U.S.

In recent years, scholarship on institutionalism has been applied to distinctly urban phenomena, to understand how the organization of urban space evolves (Marwell & Morrissey, 2020; Sorensen, 2015, 2018; Spicer & Casper-Futterman, 2023). For our purposes, a key tenet of this scholarship is that urban organizational models (like the CLT) may evolve over time through processes such as “institutional layering” (Sorensen, 2015; Thelen, 1999), whereby new or updated “rules” are developed for an organizational model, even as the old rules persist in operating in parallel, as well, and may not be supplanted within that model’s operating “organizational field” (Fligstein & McAdam, 2011). Such work usefully frames any review of the development and persistence of different versions of the CLT, within what scholars variably call the “CLT field” (DeFilippis et al., 2019), “CLT movement” (DeFilippis et al., 2018), and “CLT industry” (DeFilippis et al., 2018). Below, we trace the organizational evolution of CLTs informed by such scholarship, as to explain how a variety of organizational structures came to operate with the CLT label in the U.S., which today is home to more than 300 CLTs and related SE entities, which steward more than 40,000 housing units (Wang et al., 2023).

The first major publication outlining an organizational model using the CLT label (Swann et al., 1972) was published by the International Independence Institute (III). It defined the CLT as a “quasi-public” body (Swann et al., 1972, p. 1) to hold land in stewardship for the community’s use and control. But when it came to the model’s specifics, while it drew on the experiences of what is now generally agreed to be the first CLT—New Communities, Inc., founded by Black U.S. civil rights movement organizers in 1969 just outside of Albany, Georgia, to secure community control over rural land (Davis, 2010, 2014; Swann et al., 1972)—the III publication noted many CLT organizational configurations might be possible. It noted, for example, that a CLT’s board and voting corporate membership might be one and the same for some period of time (Swann et al., 1972, p. 36); that the board might have varying numbers of classes/types of representatives (Swann et al., 1972, p. 40); that some contexts might warrant a regional, rather than neighborhood, CLT scale (Swann et al., 1972, p. 28); that a Community Development Corporation (CDC) could sponsor a CLT, with varying degrees of overlap and autonomy (Swann et al., 1972, pp. 31–32); and that government, not just “community,” could act as land owner, while ceding governance to the community (Swann et al., 1972, p. 3).

By the early 1980s, as the CLT model began to spread, the III had been superseded by the Institute for Community Economics (ICE), which came to act as a key CLT organizational developer and institutional actor in shaping the CLT model’s “rules.” By 1982, ICE had published The Community Land Trust Handbook, which stated CLTs were democratically structured nonprofit corporations with boards elected from the membership, and, specifically, an equally tri-partite board of residents, community residents, and at-large public representatives (Institute for Community Economics, 1982, p. 18). Some of these features, however, were “happened upon” (Davis, 2014, p. 27), rather than derived by some formal process. This tri-partite definition today corresponds to what CLT development organizations refer to as the “traditional” CLT model (Lowery et al., 2021). This definition also typically specifies a 99-year lease with resale restrictions today (Lowery et al., 2021).

ICE and the National Community Land Trust Network (now Grounded Solutions Network, or GSN) continued to play an active role in documenting and developing CLT best practices, with the former creating a legal manual in 1991 and a technical academy in 2006, and the latter publishing a technical manual in 2011 (Davis, 2014). GSN, a national nonprofit organization promoting the CLT model, and sometimes critiqued by some CLT scholars for acting as a “trade association” (DeFilippis et al., 2018) to advance a more “technocratic” approach to CLTs (DeFilippis et al., 2018), nonetheless claims to embrace an expansive and inclusive definition of CLT. Its definition specifies three types of representatives in the board and the provision of lasting community assets, but not tri-partite board structure or membership (Grounded Solutions Network [GSN], 2023).

These CLT definitions, however, did not necessarily or always align with those created in legislation. By the early 1990s, as more CLTs continued to emerge,² a federal legislative definition was created, so CLTs could access U.S. Department of Housing and Urban Development (HUD) funds. Legislation of this era, into which Rep. Sanders inserted a definition provided to him by a small group of CLT advocates (Davis, 2014), defined the CLT in a way that was largely consistent with the CLT movement of the time, a definition which today is sometimes called the “classic” CLT model by practitioners. Through HUD’s HOME Investment Partnership Program (HOME), CLTs can qualify as a special type of “community housing development organization” (CHDO, a legal category with its own definition and requirements for funding eligibility purposes, and which requires that one third of the board be represented by low-income persons or areas) that is eligible for CHDO technical assistance funds. Through the HOME program, HUD states that a CLT (a) is not sponsored by a for-profit organization, (b) acquires and holds land with formula-based restrictions on the transfer price of the improvements by lessees, as to assure perpetual affordability, and (c) has a board that consists of CLT residents, members residing in the CLT area, and members representing public interest (HUD, 2001).

Other federal regulatory agencies, however, may operate with a slightly different definition. The Department of Agriculture’s Rural Housing program (7 CFR § 3555.10, n.d.; 42 U.S. Code § 1472, n.d.) defines the CLT as a private (i.e., explicitly excludes government ownership) entity and states slightly different board requirements (i.e., does not specify that a majority of board members must be elected). Notably, neither definition precisely matches the one sometimes put forth by CLT organizations of the “traditional” model, with one-third/one-third/one-third board-seat allocations. Furthermore, these federal definitions have focused primarily on housing, even though the “original” definitions from the CLT movement also encompassed other types of public uses and commercial space.

State governments, meanwhile, have also offered varying definitions of a CLT, which may be utilized for state legal incorporation purposes, state funding eligibility, or both. As of 2018, 15 states had adopted some form of CLT statute (Decker, 2018). Five were comprehensive CLT statutes, while 10 were only partially specified with respect to the definitional rules of a CLT (Decker, 2018). These state laws vary with respect to how the CLT is defined, with differences usually relating to the nature of income level restrictions, ground lease renewability, and tax exemption eligibility. The consequences of failing to comply with these legal definitions by self-described CLT remains unclear (Decker, 2018, p. 500). But some states, aware of the disconnect between their laws and CLTs’ actual practices, have moved to update their legislation so that any disconnect between legislation and practice does not become an operational issue. California, for example, moved to pass legislation (Bill Text AB-430, 2023) to address this concern, so that the state’s CLT definition can encompass the range of actual organizational practices in use.

Varying definitions in practice: Varieties of CLTs and related SE entities

Reflecting this institutional evolution and lack of private or public regulatory definitional consistency, today a range of CLTs and closely related CLT-like models have emerged and remain in operation, generating scholarly debate over how the organizational varieties of CLT may relate to difference in their mission-related outcomes.

The range of CLT entities is sometimes argued to include some “CLT-like” entities, which might not always meet a strict definition of a “classic” CLT, but nonetheless are inspired by or seek to caucus, organizationally and politically, with the CLT movement, and perhaps even adopt the CLT label in their name. Earlier this century, CLT advocates began to conceive of both CLTs and some of these CLT-like entities as part of a broader class of Shared Equity (SE) entities. The SE concept,³ developed by a central figure in the CLT “movement,” John Emmeus Davis (Davis, 2006), encompasses all non-market homeownership forms which preserve affordability by limiting equity, so that initial public and social investments can be “recycled” (Axel-Lute, 2021) over time to benefit future generations of low-income households. The SE concept thus includes other collective tenure ownership models, such as co-operatives (Axel-Lute, 2021; Crabtree et al., 2012; Ehlenz & Taylor, 2019).⁴ But by 2011, when the last prior national CLT survey was conducted (Thaden, 2012), a number of nonprofit organizations had begun to adopt SE homeownership models which were not co-operatives and might not meet a strict definition of a CLT, but variably referred to themselves as “CLTs,” “SE programs,” or “below market-rate programs.”

These latter CLT-like entities may deviate from some of the features associated with the “classic” CLT model, but nonetheless seek to steward permanently affordable and community-controlled housing and/or commercial space (Crabtree et al., 2012; Davis, 2006; Ehlenz & Taylor, 2019). As such, some practitioners now refer to CLTs and closely related shared equity entities together, even though some of the latter may not operate with “classic” tri-partite governance arrangements, and may not offer the full range of tenure options (e.g., both rental and ownership opportunities for different types of households and life stages) (Wang et al., 2023).

Even among those organizations which may call themselves CLTs, some may not operate formally as a nonprofit at all, but rather, may technically be owned or sponsored by the government, with some degree of community control embedded into the governance model (Spicer et al., 2022). Though such configurations have been critiqued for potentially reduced community control (DeFilippis et al., 2018), it has also been noted that early CLT organizers (Swann et al., 1972) envisioned that it might be possible to combine public ownership with a degree of community control (Spicer et al., 2022). Community control is a mechanism through which CLTs advance social justice and racial equity, with varying effect due to differing CLT capacities and local political environments (Lowe, 2022).

Some CLTs may also operate at starkly different scales, both in terms of the number of units they steward, as well as their intended service area (Fujii, 2016), with larger service areas perhaps resulting in a less of a focus on community control-related goals (DeFilippis et al., 2018). Other CLTs, meanwhile, may be structured as nonprofit corporations, but may not adhere to a strict tri-partite governance structure, nor have a majority of its board as elected members. Some may also be called CLTs but may be subsidiaries or partners, with varying degrees of autonomy and dependence, of CDCs or other similar nonprofit and community programs, as was originally envisioned by early CLT movement builders, as noted earlier (Lowe et al., 2022; Swann et al., 1972; Thaden, 2012).

Some of this research has explicitly addressed the way in which certain ideal-type variations in CLT structures may affect the ability of CLTs to achieve different types of missions, and might thereby be conceived of as occupying different product (e.g., property management development vs. community organizing/advocacy) or institutional (e.g., nonprofit vs. governmental) niches within the community development system (Ferguson & Stoutland, 1999). Scholars have specifically note that the “classic” or “traditional” version of a CLT, which had definitionally emerged by the early 1980s, as noted above, called for tri-partite governance of a nonprofit organization, with three classes of corporate membership in the nonprofit organization of the CLT: residents (i.e., lessees of the CLT), nonresident local community members, and the public at large. Early iterations also focused on the establishment of new organizations which sought to achieve community control over urban space (DeFilippis, 2003), not only by securing housing, but by hosting a wide range of commercial and social use activities in spaces maintained at cost, rather than to maximize profit.

These same scholars have argued, however, as CLTs increasingly were created by existing community organizations funded by large philanthropic institutions in the community development “industry” (DeFilippis et al., 2019), and increasingly focused on permanently affordable housing at scale (Thaden & Pickett, 2019), they began to shift to focus primarily on housing provision, while streamlining or reducing tri-partite governance with corporate membership. As a result, while some scholars do note tri-partite governance is not sufficient to assure either community control or the development of a meaningful sense of community among participants (Kruger et al., 2020), they have nonetheless argued that governance and ownership configurations which deviate from the “classic” tri-partite model may be associated with a reduction in community control, and a reduction in broader aims beyond securing housing that might be more concordant with conceptual notions of a CLT “movement” (vis-à-vis a CLT “industry”), as referenced earlier.

Empirical evidence on CLTs’ benefits and impact

Notwithstanding this substantial corpus of scholarship detailing the potential drawbacks associated with different types of CLTs, there has been little comprehensive effort to broadly measure the universe of CLTs and categorize them into different ideal-types based on their organizational differences, nor has there been much effort to attempt to measure whether or how these variations actually relate to performance and/or outcomes, that is, CLTs’ ability to meet their goals of securing community-controlled, permanently affordable housing and commercial space, particularly for lower-income populations. Most studies cited above, which focus on specific types of CLTs and their potential drawbacks or benefits, utilize small-N case study designs, and use qualitative data to make inferences from few cases, often also drawn from few regions. This is not to say that such studies have no value. Rather, these studies do not attempt to empirically operationalize or generalize their claims that some model versions may be systematically associated with diminished performance or outcomes. They also do not attempt to comprehensively define what major types of CLTs might exist, either.

This is not to say there has not been substantial CLT research of a comprehensive, statistical nature. CLTs have been empirically shown to improve access to homeownership (Wang et al., 2019), increase residential and neighborhood stability (Nelson et al., 2020; Wang et al., 2019), reduce gentrification (Choi et al., 2018), enhance wealth creation (Acolin et al., 2021), reduce mortgage debt (Theodos et al., 2019), and advance racial equity (Wang & Thaden, 2022). CLT unit owners are, as compared to market-rate homeowners with comparable levels of household income, more likely to be Black and female-headed households (Schneider et al., 2023). CLTs may also produce heightened resident engagement (Lowe & Thaden, 2016), and can also provide permanently affordable rental housing as well (Decker, 2018). Yet these mostly quantitative, larger-N studies typically do not specify which CLT sub-types their statistical analyses address, in part because such data has not previously been systematically collected and analyzed.

Despite this extensive literature documenting the benefits of CLTs, as well as the previously reviewed literature on their institutional variety, several inter-related questions remain empirically unanswered: what are the main varieties of CLTs and CLT-like SE models in operation today, and what are the demonstrable, performance-related consequences of the development of such variation? Are certain CLT types more likely to enable certain types of households to secure permanently affordable housing? Do some offer fewer support services for residents, or operate at fundamentally different scales, with consequences for their ability to achieve CLT goals such as stewarding permanently affordable housing and enabling broader community control? If such differences exist, can we definitively state some types of CLTs clearly occupy different product niches (Ferguson & Stoutland, 1999) in the community development system? For example, might “traditional” CLT models, which might be of an older vintage given the institutional development reviewed above, be more effective in achieving a greater degree of community control? These questions have not been systematically examined.

Data and method

Study scope and data collection

To answer these questions, we leverage data derived from the new national CLT/SE entity survey (Wang et al., 2023). The survey employed a broad definitional scope to capture all possible variations of the CLT model. The scope included self-identified CLTs (except for those found to be exclusively conservation land trusts) and nonprofits with SE programs (except for nonprofits exclusively providing co-operative housing). Wang et al. (2023) specified the study scope in detail.

To include SE programs operating with CLT-like characteristics, the survey included those with SE homes that were single-family units as defined by the conventional mortgage market (i.e., detached units, attached units, and condominiums). The survey distinguished single-family SE homes from other residential unit types such as rental units, lease-to-purchase units, cooperative housing units, manufactured homes, or homeownership units that did not have resale restrictions.

As to enable the questions above to be empirically tested with data, we developed a survey questionnaire that included three main sections: (1) organizational structure and characteristics, such as organizational type, service area, staff, board, and membership; (2) SE home portfolios, such as resale control mechanisms, SE home characteristics, home acquisition, financial characteristics, property and resale counts, and homeowner characteristics; and (3) size and characteristics of other residential portfolio types, and nonresidential development activities. The CLT/SE entity survey questionnaire was developed using a secure third-party online data collection platform to enhance the completeness and consistency of survey responses. The survey instrument was beta tested by staff from eight CLTs, which were selected to represent geographic, programmatic, and residential portfolio diversity. Feedback from the beta testers was incorporated in finalizing the survey instrument.

Data collection for this study started in May 2021 and concluded in October 2022. It consisted of three phases. The first phase involved the development of an up-to-date CLT/SE directory with contact information for the purpose of survey administration in later phases. The second phase focused on administration of an online survey. And the third phase included additional outreach and data verification. Wang et al. (2023) described detailed information about data collection in each phase.

Table 1 summarizes data collection effort. A total of 314 CLT/SE entities were identified, with a 62% survey response rate. The total count included 17 CLT/SE entities (most are newly established organizations) that were added to the final list after the survey administration. After excluding organizations with incomplete survey response, 144 CLT/SE entities were retained for the typology analysis, representing 46% of the final list.

Table 1. Summary of data collection effort.

Item Name	Entity Count/Percent
Final list of CLT/SE entities	314
CLT/SE entities added to the final list after survey	17
CLT/SE entities included in the survey	297
Non-respondents	114
Respondents	183
Complete survey	144
Incomplete survey	39
Survey response rate	62%
Complete survey as percentage of final list	46%

While the data collection effort was carefully designed and executed, survey biases likely existed. There may, for example, be missing CLT/SE entities from the final population list, as smaller and newer organizations that had not yet documented or updated their activities online were likely to be excluded during the website validation process. In addition, the survey response rate was notably higher in areas where regional partners helped verify the CLT directory and reach out their member CLT/SE entities. In areas lacking regional coalitions, entities may, again, have been systematically under-reported. Despite these potential biases, the dataset constituted the most comprehensive effort to date to understand the prevalence, practices, and impact of the CLT/SE field in the U.S. Figure 1 shows the locations of 144 CLTs included in this study and all 314 CLTs from the original dataset.

Figure 1. Map of sample and all community land trusts in the United States.

Typology development

Based on the literature reviewed earlier, as well as the survey responses to specific questions on organizational characteristics, we deployed a replicable waterfall process model to develop a CLT/SE entity typology that separated “traditional CLTs” from other variations, and then further subdivided other variations based on their organizational differences with respect to five characteristics, all suggested as germane by the literature as reviewed above: entity type, membership, board, governance, and asset base.

Specifically, based on the data collected, we partitioned and assigned the respondents to typology categories based on whether the entity is a nonprofit organizations or a government entity, whether the entity possesses a corporate “community” membership (i.e., anyone within a stated community area, as per the CLT/SE’s by-laws or articles of incorporation, can become a corporate member of the organization), whether the entity has a distinct board of directors (as separate from the corporate membership), whether the board has a tri-partite or lessee-centered governance structure, and whether the entity actually holds any land in trust.

We applied a modified waterfall model, which is a type of process model, to identify major ideal types of CLTs and assign respondents to membership in a given type. Figure 2 presents the sequential development process of our partitioning and assigning process model.⁵

Figure 2. Typology development process.

We first identified “start-up” CLTs from survey respondents, which were required to meet two criteria to be assigned to this category: (1) the entity must have the intent to create either a standalone CLT or to develop a CLT program; and (2) there was no land in trust or it was not ground leased at the time of survey, i.e., its asset base did not include any land in trust which would reasonably meet a CLT definition. Given the nature of these criteria, we did not need to consider entity type, membership existence, and board existence and structure at this step, given that start-ups may not yet reached a stage in their development to have yet addressed these features. It is also important to clarify that the “start-up” type is not based on year of founding, but instead reflects the entity’s development phase. Hence, a CLT which was formed or established decades ago, but which operates today without land in trust, could be classified as belonging to this category.

The next category is the “government-housed CLT.” If a CLT is operated as a program operating under or by a public agency, it was assigned to this category. Using this as a sole exclusionary criterion, we did not consider membership existence and board structural traits in assigning respondents to this category. We did require, however, that there be land in trust or ground leased, as to mutually exclude and distinguish it from the start-up category.

The third type, as shown in Figure 2, is also based on entity type, and separately categorizes CLT/SE programs which are not government-housed, but also not independently housed as a stand-alone organization, either, i.e., it operates as part of a larger nonprofit with a CLT/SE program. Many of the organizations belonging to this type were Habitat for Humanity affiliates or CHDOs that established a CLT program or a CLT-like SE program—in addition to their existing residential portfolios—to develop permanently affordable housing units. Similar to government-housed CLTs, membership and board structure are not required under this type, while land in trust or ground leased is required.

The last step, as shown in Figure 2, enables us to distinguish what we call “adapted CLTs” from traditional CLTs. We define adapted CLTs as nonprofit organizations that met two out of three organizational structural criteria relating to corporate membership, the board, or asset base. Most cases involved one of two scenarios: (1) they either had (or were in the process of including) at least one lessee on board and there is land in trust or ground leased (but no membership), or (2) they had corporate community membership and at least one lessee on board (but no land in trust or ground leased, despite that there are permanently affordable housing units being operated by the entity).

Finally, we then operationalized a definition of traditional CLTs by restricting this category to nonprofit organizations (i.e., government ownership excluded). Entities assigned to this category were also required to have a corporate community membership. Ideally, we would also expect traditional CLTs to have a tri-partite board. In practice, we relaxed this criterion by including organizations that had at least one third of lessees or residents living in housing provided by the organization (hereinafter, lessees) on the board. In addition, traditional CLTs must have land in trust or that is ground leased. Note for this criterion, these two conditions need not be exclusive; a traditional CLT may have deed-restricted units whose land is not in trust, in addition to ground leased units. Under traditional CLTs, meeting all four criteria indicates that the underlying entity is structured to deliver community control and stewardship—which are the heart of a CLT’s mission.

Critically, there is a relationship between an entity’s assignment to a specific typology category, and both the existence and type of housing unit that might be associated with an entity. Start-up CLTs, for example, do not have any permanently affordable housing units or land. For nonprofits with a CLT/SE program and government-housed CLTs, permanently affordable housing must exist to demonstrate such program is active. For SE programs, however, there must be at least one SE homeownership unit in the program. For both traditional CLTs and scenario 1 of adapted CLTs (as reviewed above), the presence of housing units was not required because they could have nonresidential developments or just land; however, if there was any housing unit, those under the CLT structure must be permanently affordable. For scenario 2 of adapted CLTs reviewed above, permanently affordable housing units are required.

Analytical approach

Figure 2 summarizes the sample size under each typology category type. Due to the small sample sizes for start-up CLTs (n = 14) and government-housed CLTs (n = 3), the following analyses focus on comparing nonprofits with a CLT/SE program (n = 36) and adapted CLTs (n = 52) with traditional CLTs (n = 39). About half of the entities under the “non-profits with a CLT/SE program” are nonprofits with a CLT program, which were self-reported as a CLT program and were validated in the original CLT Census study according to Wang et al. (2023). The rest half of the entities under this category are nonprofits with a CLT-like, SE program.⁶ Excluding start-up CLTs and government-house CLTs is not to say that the first two categories are inconsequential, but rather, are difficult to further study analytically due to their low frequency in our survey. Notwithstanding the above-referenced sample bias, this low frequency in our survey is likely reflective of their comparative occurrence in reality.

Because the CLT survey revealed a wide variation of board composition in terms of the existence and proportion of lessees and community residents on board, we also conducted a supplemental typology validation analysis by further examining the board composition and membership existence of traditional CLTs and adapted CLTs. The objective of this part of analysis was to examine to what extent community control may be said to be effectively operationalized in CLT models’ governance and organizational structures: based on the literature, the presence of corporate community membership and a tri-partite board arguably represents a fuller exercise of community control, while the lack of corporate community membership, as well as no lessee or community resident may reflect diminished or limited community control. We did not include nonprofits with a CLT/SE program in this part of analysis because, in most cases, their responses were often about their parent nonprofits as opposed to their CLT-level governance.

Finally, we examined CLT/SE outcome variables from the survey that can be broadly grouped into three categories: scope, capacity, and performance.⁷ Table 2 describes each variable in this study. These variables, while comprehensive, may not capture nuances due to the nature of survey data collected at the national scope. For example, we included a variable measuring the number of residential activities, but the dataset does not contain further information about the scope and effectiveness of those activities. Consequently, the commonalities and differences across CLT types revealed in this study commemorate the level of detail from the CLT/SE dataset. We compared the three largest categories in our typology: traditional CLTs, adapted CLTs, and nonprofits with a CLT/SE program, separately, across these variables. For nominal variables, chi-square tests were conducted. For continuous variables, Mann Whitney U-tests were applied. The purpose of conducting these statistical tests was to make quantitative decisions about whether differences exist between variables of the population based on the sample data. We chose different statistical testing methods based on the type of sample data. The next section presents the analysis results.

Table 2. Description of variables.

Variable Name	Description
Scope:
Year of Establishment	Year in which organization was established.
No. of Unit Types	Count of residential unit types (i.e., shared equity homeownership, rental, lease-to-purchase, cooperative, manufactured homes, non-shared equity homeownership).
SE Resales in 2021	Number of resales for shared equity homes in 2021.
SE Prog. Staff Ratio	Ratio of shared equity program staff to organizational staff.
No. of Residential Activities	Count of residential activities (i.e., property management, HOA support, home repairs/weatherization, HUD-certified homebuyer counseling, non-HUD homebuyer counseling, loss mitigation counseling).
No. of Residential Activities—All	Count of nonresidential activities an organization ever had (e.g., commercial development, agricultural development); up to 7 types of activities.
No. of Community Support Types—All	Count of community support types an organization ever had (e.g., job training, community organization, policy advocacy); up to 7 types of support.
No. of Nonresidential Activities—Since 2020	Count of nonresidential activities an organization had since 2020 (e.g., commercial development, agricultural development, community organization, policy advocacy); up to 7 types of activities.
No. of Community Support Types—Since 2020	Count of community support types an organization had since 2020 (e.g., job training, community organization, policy advocacy); up to 7 types of support.
No. of Post-Purchase Support Types	Count of post-purchase homeowner support types (i.e., home maintenance support, home repair & replacement support, improvement support, financial issue support).
Racial Equity Priority?	Whether organization prioritizes racial equity/justice (yes-1, no-0).
Racial Equity Policy?	Whether organization addresses racial equity/justice in its bylaw or formal policies (yes-1, no-0).
SE Prog. Supported by Internal Revenue?	Whether internally generated revenue subsidizes operating budget for the shared equity homeownership portfolio (yes-1, no-0).
SE Unit Presence?	Whether there is any shared equity unit (yes-1, no-0).
Rental Unit Presence?	Whether there is any rental unit (yes-1, no-0).
Income Level	Income requirement for shared equity homes (i.e., below 80% AMI, 80% AMI, 120% AMI, above 120% AMI).
Region	Region (i.e., Northeast, Midwest, South, West, Puerto Rico).
Service Area	Organization’s service area (i.e., one or more neighborhoods, city, one county, multiple counties, metropolitan statistical area, state).
Capacity:
Avg. SE Resales/Year	Average number of resales per year for shared equity homes.
SE Type Diversity Index	Simpson’s diversity index of development types for shared equity homes (i.e., single-family detached house and townhomes, duplexes or SF with ADU, condominiums).
No. of Org. Staff	Number of staff in the organization.
No. of SE Prog. Staff	Number of staff in the shared equity program.
No. of SE Acquisition Types	Count of acquisition types for shared equity homes (e.g., new construction, rehabilitation, donated properties); up to 9 types.
Performance:
Years between Org. and SE Prog. Establishments	Difference between shared equity program establishment year and organization establishment year.
Years between Org. and Rental Prog. Establishments	Difference between rental program establishment year and organization establishment year.
No. of SE Units—2022	Number of new and projected shared equity homes in 2022.
No. of Rental Units—2022	Number of new and projected rental units in 2022.
No. of Other Units	Number of units other than shared equity homes or rental units.
No. of Lost SE Units	Difference in shared equity unit count between currently in portfolio (by the end of 2021) and ever created.
No. of Projected SE Units/Year	Average number of projected shared equity homes per year.
No. of Projected Rental Units/Year	Average number of projected rental units per year.
No. of External Revenue Sources for SE Units—All	Count of external revenue sources to fund shared equity homes since the first shared equity home (e.g., community development block grant, state housing trust fund, individual donation); up to 26 source types.
No. of External Revenue Sources for SE Units—Since 2019	Count of external revenue sources to fund shared equity homes since 2019 (e.g., community development block grant, state housing trust fund, individual donation); up to 26 source types.
Years between SE Prog. Establishment and 1st Unit Acquisition	Difference in years between shared equity program establishment and property acquisition of the first shared equity home.
Years between SE Prog. Establishment and 1st Unit Sell	Difference in years between shared equity program establishment and sell of the first shared equity home.
% of 1st-Time Homebuyer	Percentage of first-time homebuyer in shared equity homes.
% of Families with Children	Percentage of families with child(ren) in shared equity homes.
% of Single-Mother Households	Percentage of single-mother households in shared equity homes.
% of POC-Headed Households	Percentage of households headed by person of color in shared equity homes.
% of Black-Headed Households	Percentage of households headed by Black in shared equity homes.
Affordability <30 Years?	Whether the affordability term for shared equity homes is less than 30 years (yes-1, no-0).
% of Internal Revenue in SE Operating Budget	Internally generated revenue (or earned income) as the percentage of the operating budget for shared equity homes (i.e., 0%, 1–20%, 21–40%, 41–60%, 61–80%, 81–100%).

Findings

Outcomes analysis

Overall, when compared to adapted CLTs and nonprofits with a CLT/SE program, traditional CLTs had more similarities than differences with respect to scope, capacity, and performance measures. Please see Table 3, which presents the results of comparative analysis, for comprehensive metrics. In terms of similarities for organizational scope measures, on average, traditional CLTs were statistically similar to the other two types in the number of residential activities (1.5–2.1 per organization) and the number of post-purchase support activities (1.5–1.9 per organization). They were similar in Diversity Equity and Inclusion (DEI) focus: 51–64% of surveyed organizations reported prioritizing racial equity and justice, and 39–43% addressed racial equity and justice in their bylaws or other formal policies. These organizations were generally evenly distributed across four regions in the United States, with higher share in the west region for adapted CLTs and nonprofits with a CLT/SE program (but not statistically different). Most organizations across CLT types (80–84%) serve households earning between 80–120% area median income (AMI). In addition, organizations in three CLT types were similar in presence of SE units (72–86%) and the share of internal revenue to support operating cost of SE program (59–72%).

Table 3. Comparative analysis results.

Variable	Traditional CLT				Adapted CLT					Nonprofit with CLT/SE
	n	Mean	Median	SD	n	Mean	Median	SD	p	n	Mean	Median	SD	p
Scope:
Year of Establishment	38	2005	2006	11	53	2006	2009	13		36	1992	1990	9	***
No. of Unit Types	38	1.8	2.0	1.0	53	1.5	1.0	0.9		36	1.3	1.0	0.7	**
SE Resales in 2021	38	2.8	0	5.7	53	1.2	0	2.2		36	1.3	0.5	2.0
SE Prog. Staff Ratio	28	0.736	1.000	0.364	39	0.790	1.000	0.312		35	0.346	0.100	0.362	***
No. of Residential Activities	38	2.1	1.5	1.6	53	1.5	1.0	1.4		36	2.1	2.0	1.4
No. of Residential Activities—All	38	1.5	1.0	1.7	53	0.9	0.0	1.3		36	0.8	0.0	1.1
No. of Community Support Types—All	38	3.2	3.0	1.7	53	2.0	2.0	1.9	***	36	2.6	3.0	2.0
No. of Nonresidential Activities —Since 2020	38	0.3	0.0	0.9	53	0.2	0.0	0.4		36	0.0	0.0	0.0	**
No. of Community Support Types —Since 2020	38	0.2	0.0	0.7	53	0.4	0.0	1.0		36	0.2	0.0	0.7
No. of Post-Purchase Support Types	30	1.9	2.0	1.4	39	1.7	2.0	1.2		31	1.5	1.0	1.1
Racial Equity Priority?	36	0.639			53	0.509				35	0.543
Racial Equity Policy?	36	0.417			53	0.396				36	0.389
SE Prog. Supported by Internal Revenue?	32	0.594			43	0.721				32	0.656
SE Unit Presence?	38	0.789			53	0.717				36	0.861
Retail Unit Presence?	38	0.579			53	0.491				36	0.278			***
Income Level	32				44					34
Below 80% AMI	3	0.094			3	0.068				6	0.176
80% AMI	16	0.500			14	0.318				21	0.618
120% AMI	11	0.344			22	0.500				6	0.176
Above 120% AMI	2	0.062			5	0.114				1	0.029
Region	38				53					36
Midwest	7	0.184			10	0.189				6	0.167
Northeast	10	0.263			13	0.245				7	0.194
PR	1	0.026			0	0.000				0	0.000
South	9	0.237			9	0.170				5	0.139
West	11	0.289			21	0.396				18	0.500
Service Area	38				53					36				**
Neighborhood(s)	6	0.158			9	0.170				4	0.111
City	12	0.316			13	0.245				2	0.056
County	12	0.316			18	0.340				15	0.417
Multiple counties	4	0.105			7	0.132				9	0.250
MSA	2	0.053			2	0.038				5	0.139
State	2	0.053			4	0.075				1	0.028
Capacity:
Avg. SE Resales/Year	37	1.7	0.1	3.5	53	0.4	0.0	0.9		35	0.5	0.0	1.4
SE Type Diversity Index	32	0.203	0.100	0.231	46	0.088	0.000	0.163	*	35	0.145	0.000	0.211
No. of Org. Staff	38	8.1	3.3	20.5	53	4.3	2.5	6.1		36	25.1	14.5	26.5	***
No. of SE Prog. Staff	34	2.8	2.0	2.6	46	2.5	2.0	2.2		35	3.9	2.0	5.0
No. of SE Acquisition Types	30	3.3	3.0	2.2	39	2.6	2.0	1.6		31	2.7	2.0	2.0
Performance:
Years between Org. and SE Prog. Establishments	29	0.8	0.0	3.5	39	2.2	0.0	4.6	*	31	12.7	9.0	13.7	***
Years between Org. and Rental Prog. Establishments	22	7.1	5.5	7.7	26	5.4	4.0	5.0		10	11.6	8.0	12.6
No. of SE Units—2022	38	94	33	151	53	45	25	70		36	140	27	360
No. of Rental Units—2022	38	87	2	357	53	45	0	108		36	51	0	133
No. of Other Units	38	14	0	60	53	11	0	68		36	26	0	77
No. of Lost SE Units	38	2.4	0.0	9.0	53	1.0	0.0	5.2		36	14.1	0.0	51.7
No. of Projected SE Units/Year	38	19.1	5.5	27.0	52	12.6	4.0	32.6		36	16.5	6.0	22.7
No. of Projected Rental Units/Year	38	7.3	0.0	15.5	53	12.2	0.0	43.5		35	2.7	0.0	10.6
No. of External Revenue Sources for SE Units—All	35	5.8	5.0	3.7	46	5.2	4.0	3.6		35	6.1	5.0	4.3
No. of External Revenue Sources for SE Units—Since 2019	35	3.7	3.0	4.2	46	2.0	1.0	2.3		35	2.3	0.0	3.8
Years between SE Prog. Establishment and 1st Unit Acquisition	30	2.0	1.5	2.7	40	1.1	1.0	1.3		30	0.2	0.0	1.2	***
Years between SE Prog. Establishment and 1st Unit Sell	27	3.2	3.0	3.0	40	2.6	2.0	2.8		30	1.6	1.0	3.9	***
% of 1st-Time Homebuyer	26	0.814	0.980	0.324	30	0.847	1.000	0.290		24	0.955	1.000	0.146	**
% of Families with Children	24	0.427	0.425	0.326	26	0.448	0.400	0.323		22	0.747	0.850	0.257	***
% of Single-Mother Households	24	0.240	0.140	0.284	24	0.187	0.185	0.165		20	0.369	0.358	0.271	*
% of POC-Headed Households	26	0.880	0.935	0.200	38	0.849	0.945	0.244		22	0.837	0.880	0.195
% of Black-Headed Households	28	0.358	0.195	0.398	28	0.254	0.105	0.301		23	0.282	0.140	0.321
Affordability <30 Years?	34	0.029			44	0.000				34	0.029
% of Internal Revenue in Operating Budget	34				46					35				*
0%	4	0.118			8	0.174				10	0.286
1–20%	13	0.382			17	0.370				18	0.514
21–40%	5	0.147			7	0.152				2	0.057
41–60%	4	0.118			6	0.130				0	0.000
61–80%	4	0.118			3	0.065				3	0.086
81–100%	4	0.118			5	0.109				2	0.057

CLT: community land trust; SE: shared equity; Mann Whitney U-tests were conducted for all numerical variables and Chi-square tests for nominal variables; *<0.1, **< 0.05, ***<0.01.

In terms of similarities for organizational capacity measures, on average, traditional CLTs were similar to the other two types in the number of staff for SE program (2.5–3.9 staff per organization) and the number of SE property acquisition type (2.6–3.3 per organization). On average, traditional CLTs had 1.7 resales per year, higher than that for adapted CLTs (0.4 resales) and nonprofits with a CLT/SE program (0.5 resales). However, the discrepancies were not statistically significant.

Regarding similarities for performance measures, traditional CLTs were similar to the other two types in the number of years to establish a rental program after organizational establishment (7.1 years vs. 5.4 to 11.6 years). In addition, traditional CLTs were similar to the other two types in median unit counts: a typical organization had 25 to 33 SE units, zero to two rental units, and zero other type of unit. Although the average unit count by tenure type and by CLT type varied substantially, it was driven by a few organizations with a large number of units, and the variation was not statistically significant. A typical organization did not lose any of its SE units, although the average number of lost SE units varied across CLT types (1.0 to 14.1, despite not statistically significant). In addition, there was no statistical difference in terms of projections of unit productivity across CLT types: on average, organizations were self-projecting to develop 12.6 to 19.1 SE units and 2.7 to 12.2 rental units per year. Across CLT types, there was no statistical difference in terms of the average number of external funding sources per organization (5.2 to 6.1 since the first SE unit, and 2.0 to 3.7 since 2019). In terms of affordability term/length, nearly all organizations across CLT types had their SE units with at least 30 years of affordability. Last but not least, traditional CLTs were similar to the other two types in the average shares of households of color (84–88%) and Black households (25–36%).

As for differences, compared to adapted CLTs and nonprofits with a CLT/SE program, on average traditional CLTs spent significantly less time acquiring their first SE property after organizational establishment (0.8 years vs. 2.2 years and 12.7 years, respectively). On average adapted CLTs had a lower number of residential community supports (2.0 vs. 3.2, p < .05) and lower diversity of SE development type (0.09 vs. 0.20, p < .10) compared to traditional CLTs.

In addition, we find that nonprofits with a CLT/SE program diverge in some statistically significant ways from traditional CLTs. Specifically, compared to traditional CLTs, nonprofits with a CLT/SE program had a lower share of rental units (28% vs. 58%, p < .01), more organization staff (25.1 vs. 8.1, p < .01), lower share of SE staff (35% vs. 74%, p < .01), and slightly lower number of nonresidential development activities since 2020 (0.2 vs. 0.3, p < .05). Also, nonprofits with a CLT/SE program were older with respect to their year of organizational establishment (on average 1992 vs. 2005, p < .01), took longer to establish a CLT/SE program after organizational establishment (on average 12.7 years vs. 0.8 years, p < .01), and took less time to sell their first SE property after SE program establishment (on average 1.6 years vs. 3.2 years, p < .01). More nonprofits with a CLT/SE program than traditional CLTs (80% vs. 50%) allocated 20% or lower internally generated revenue to the operating budget for SE homes. Nonprofits with a CLT/SE program also had a higher share of first-time homebuyers (96% vs. 81%, p < .05), families with child(ren) (75% vs. 43%, p < .01), and single-mother families (37% vs. 24%, p < .1) living in SE units. In terms of service areas, compared to traditional CLTs, a lower share of nonprofits with a CLT/SE program served at the city level (6% vs. 32%), but they had higher shares at the county (42% vs. 32%), multiple-county (25% vs. 11%), and metropolitan statistical area levels (14% vs. 5%).

Supplemental typology analysis

Our supplemental typology analysis affirms that there is a wide range of board-membership configurations. Organizational distribution by membership presence and board composition is shown in Figure 3 for traditional CLTs and adapted CLTs. In terms of membership existence, by definition all traditional CLTs had a corporate membership, whereas the majority of adapted CLTs (45 out of 53, or 85%) did not have corporate membership. In terms of board composition, we first analyzed the response data structure with respect to the number of lessees on board. If at least one third of board seats designated by the by-laws were lessees, the CLT was assigned to “Tier One.” “Tier Two” included CLTs with at least some lessees on board but less than one third of board seats. “Tier Three” had no lessees on the board. Suggestive of the possibility of an underlying bi-modal distribution in the overall population, in which organizations pursued either full community and lessee governance or none at all, the majority of adapted CLTs were identified as Tier One and Tier Three (25 and 22 CLTs, respectively), with only five adapted CLTs in Tier Two.

Figure 3. Number of community land trusts by membership presence and board composition (Black dots: traditional CLT; gray dots: adapted CLT).

Within each Tier, we then identified further sub-variation by sub-stratifying based on the presence of other community members on the board. For Tier One, a tri-partite board represents the “best” or most idealized board structure, as per the literature review, followed by a board with at least one third of board seats being lessees. More traditional CLTs employed an equally tri-partite board structure (n = 22) than not (n = 17), whereas a slightly higher number of adapted CLTs had at least one third of lessees (n = 14) than those with a tri-partite board (n = 11). For Tier Two, though there is a small sample size which limits our ability to make generalizable analyses, adapted CLTs with some lessees (but less than one third) on the board were more likely to have no community resident presence (n = 3) than having some community residents on board (n = 2). A similar pattern was found in Tier Three: adapted CLTs with no lessees on the board were more likely to have no community resident presence (n = 13) than they were to have at least one third of the board be community residents (n = 7). There were two adapted CLTs in Tier Three that did not have any requirements with respect to the presence of a community resident on the board, but had a requirement about having a corporate member presence on the board.

Discussion

As CLTs in the U.S. have proliferated, so has their institutional variety, as have the number of studies examining their benefits. However, to date, the major substantive varieties of CLTs in operation, at least with respect to their institutional and/or organizational structure, had not been comprehensively delineated using a comprehensive or empirical process. As a result, prior studies had not compared practice and performance across the range of CLT varieties in operation today. By addressing these literature gaps, we have sought not only to inform policymakers, CLT providers, and community leaders when allocating resources for affordable housing and designing CLT programs, but also to inform the CLT field’s growth and direction.

Using a new national CLT/SE entity dataset (Wang et al., 2023), this study first sought to advance the literature through the development of a five-type CLT typology, offering data on the prevalence of each sub-type, as well. Beyond this key first finding, which we discuss below, we also offer two other broad sets of findings, related to outcomes and governance regimes across CLT sub-types. Second, in comparing the scope, capacity, and performance of traditional CLTs with adapted CLTs and nonprofits with a CLT/SE program, we found more in common than we found different across the sub-types. Third, in examining nuances in the micro-governance structures in terms of board composition and membership presence for traditional and adapted CLTs, we similarly found commonalities suggesting adherence to shared principles across varied governance instantiations, despite the existence of some notable differences. Below, we review each of these three major sets of findings in greater detail, before discussing some limitations and the future research implications of our findings.

Prevalence of CLT sub-types

With respect to the five major CLT sub-types we have identified, acknowledging potential sample biases as a limitation of our findings, we suggest that the frequency of each of the sub-types reflects the definitional evolution of the CLT in practice over the last half century. Nonetheless, despite the dominance of the traditional CLT in the historical CLT literature, there are fewer traditional CLTs than adapted CLTs in the survey sample (n = 39 vs. 52). This suggests that while traditional CLTs are a primary and well-developed sub-type, they are not the dominant or “typical” form of CLTs in the U.S. today.

Another primary category is the nonprofit organization with an embedded CLT/SE program, and we observe a trend that increasing numbers of nonprofit community development organizations have adopted the CLT model as an internal program. One in four organizations in our sample belongs to this category, higher than 19% documented in the first national survey of CLTs in 2007 (Sungu-Eryilmaz & Greenstein, 2007). The development of this sub-type is consistent with literature which suggests that CLT development has been affected by its interaction with the community development “industry” in recent decades (DeFilippis et al., 2019), resulting in the proliferation and growth of these instantiations of the CLT.

Finally, while governments can deploy the CLT model and act as owner (Spicer et al., 2022), with only three government-housed CLTs identified in the survey, our typology composition affirms that the nonprofit corporation structure is the backbone of CLT development, and likely reflects the fact that for much of the CLT’s history, direct state ownership of enterprise has been out of favor in the U.S. as an institutional model, reducing uptake and use of this model alongside the others.

Nonetheless, the persistence of these multiple ideal-type CLT models is consistent with notions of institutional layering, through which multiple iterations of what is initially a single institutional model or organizational form may persist as social actors modify its rules to apply to new contexts and circumstances. This common urban institutional development phenomenon results in different types of models which co-exist in parallel, each developing in response to broader forces at given moments in time. For our purposes, a key point is that such layering is a distinctly different evolutionary pathway than, for example, institutional displacement, where a new model sweeps aside another one entirely (as has arguably occurred with other urban organizational forms in the past, perhaps with “settlement houses,” for example). It is possible, therefore, that different “versions” of the same “model” may persist and come to serve different “product” niches in the community development system, as noted earlier, and as originally envisioned in the first CLT model documentation published more than half-century ago (Swann et al., 1972).

Commonalities and differences across CLT sub-types

Our second set of findings pertain to differences in outcomes across the three most prevalent sub-types. We found that traditional CLTs in large part share similar traits and outcomes with adapted CLTs and nonprofits with a CLT/SE program. Specifically, we did not discern any statistical difference with respect to the size of residential portfolio, residential tenure types, stewardship activities, affordability term, commitment to DEI practices, staff capacity, or financial health indicators.

The fact that there are more commonalities than discrepancies across these metrics suggests that deviations in their governance and operational structure from the traditional CLT model do not necessarily or substantively affect mission or practice in statistically identifiable ways, as previous studies had suggested on an anecdotal or small-N basis (Crabtree et al., 2012; Davis, 2006; Ehlenz & Taylor, 2019; Thaden & Pickett, 2019). We therefore cannot definitively conclude that different sub-types of CLTs fundamentally occupy different product niches in the community development system, nor can we conclude that CLT sub-types whose development might reflect the presence of “industry” forces yield fundamentally different outcomes than those aligned with “movement”-based, community organizing approaches. This is not to say that no such differences exist, but rather, if they exist, they do not register in a manner which can be readily captured from descriptive survey data and a resulting statistical analysis. It is also possible that such differences are still emergent, and that sufficient time has not yet passed for them to register on a statistically significant level.

Meanwhile, those statistical differences between the three major sub-types which we were able to identify and measure generally seemed to merely reflect the inherent nature of each of the major categorizations, i.e., they were tautological in nature. For instance, CLT/SE programs created in nonprofits tend to be newer in vintage than traditional CLTs, yet these programs took less time to produce SE units, served larger areas, and tend to serve great proportion of first-time homebuyers and families with children. These findings indicate that nonprofits with established capacity, resources, and other programmatic activities may help the newly created CLT/SE program “hit the ground running” in more effectively delivering SE units and purposefully serving needed families than traditional CLTs, which may take longer in forming “from scratch” to be able to build the necessary organizational capacity to a sufficient scale to accomplish these goals. Traditional CLTs and, more broadly, CLT startups, could more effectively achieve their goals within a shorter timeframe by conducting market assessments and partnering with other community-based and real estate entities. These efforts can help them integrate into the local affordable housing ecosystem and identify families in need.

In addition, we find that compared to traditional CLTs, adapted CLTs are less involved in community activities and less diverse in SE development type. The reduction of community engagement and residential development type may be associated with membership absence or a lack of community presence in board composition of the adapted CLTs, again likely reflecting the nature of the categorization. Though some scholars have treated community engagement activities as a proxy for community control of CLTs (DeFilippis et al., 2018), it is not clear from findings of this study whether adapted CLTs are associated with a lesser degree of community control, as the response measures we derived from the survey were too coarse to meaningfully assess community control.

Nuances in CLT micro-governance structures: Board-membership configurations

Finally and relatedly, further examination of CLT governance with respect to the frequency of certain types of board compositions and membership presence also indicates significant overlap and commonalities in practices across a varying range of board-membership configurations. There are three notable sub-findings in this regard.

First, the vast majority of traditional and adapted CLTs have either a tri-partite governance model, or at least one third of board directors designated to lessees, suggesting that this key element of the “classic” CLT model, which is often believed to help foster substantive community control (Kruger et al., 2020), is in fact dominant in practice. Second, we find that substantially more adapted CLTs have no lessee requirement in their bylaws with respect to their board composition, than those requiring some, but less than one third, of board directors designated to lessees. While lacking statistical evidence to support this argument, we suspect that many of these adapted CLTs may simply be in the early stages of their development, with none or too few lessees to make the lessee requirement into bylaw, which, as noted in the literature review, is wholly consistent with the evolutionary pathways first envisioned by CLT practitioners a half-century ago (Swann et al., 1972). Third, it is also a fairly common practice for adapted CLTs not to have a membership, the significance of which remains to be determined.

Future research implications and limitations

Based on these findings, future studies could examine why some CLTs choose not to follow the “classic” board-membership configuration, and might also examine this variation’s relationship to differences in organization mission and programmatic activity. Further study might also empirically examine the efficacy of community control among CLTs with different board-membership configurations, and effectiveness of alternative governance practices, as well.⁸ Finally, future studies might also seek to identify which CLT sub-type(s) might be best to deploy in specific contexts, as to further maximize each sub-type’s effective use.

While the new CLT dataset presented a unique opportunity to address the research question of this study, it also had clear limitations. First, the small sample size restricted our ability to conduct a robust regression analysis. Hence, we were not able to control for the effect of other variables while examining the difference of any given trait across CLT categories. Second, the data derived from a comprehensive national survey lack depth and nuance in terms of organizational structure and programmatic characteristics that may help interpret the commonalities and differences across CLT types. Third, because the data do not contain board or membership information for the CLT/SE program under a parent nonprofit organization, we are not able to examine to what extent membership-board configuration of the CLT/SE program deviates from traditional or adapted CLTs. Lastly, we do not have data to examine the evolution of governance structure, if any, within each CLT in our sample, therefore we do not know how a change in the governance structure of a single entity may affect that CLT’s mission, scope of work, institutional capacity, and performance.

Conclusion

Findings of this research support the premise that CLT evolution has generally followed the posited contours of field or institutional layering as an urban development process (Sorensen, 2015, 2018), as variations in CLT models have emerged without seeming to supplant the traditional CLT model, and in ways which reflect broader forces shaping government housing and community development policy over time. In conjunction, however, we also observe that variations in CLT sub-type and/or governance structure do not necessarily impede stewardship or delivery of lasting affordable housing or other community assets. Beyond the CLT, this more broadly suggests that it might be possible to use substantively different implementation variations of an ostensibly shared underlying organizational model or organizational form (here, the CLT, though this could apply to other models), to achieve the same goals. Empirically, it perhaps provides some support for David Harvey’s (2012) critique that the “left as a whole is bedeviled by an all-consuming ‘fetishism of organizational form’” (p. 125), in which proponents of urban change become so concerned with a dogmatic adherence to only using certain versions of these models that it “stands in the way of exploring appropriate and effective solutions.” (p. 70) The details of the organizational form, indeed, might therefore not be the most important factor in shaping the content of what different collectives and groups do to affect urban change.

These findings are not intended to undermine the need for clear CLT definitions that align with the CLT field’s evolution, however, for the purposes of accessing resources and making a greater impact as the field grows. If certain organizational models like the CLT are to receive access to government and philanthropic support, it is important to identify and define the range of ways CLTs are organized to realize their missions in practice, as to assure that the intended targets of resource support are, in fact, the ones receiving it. If multiple sub-types of CLTs, however, can readily achieve many of the CLT movement’s shared goals—if most, if not all, roads lead to Rome—it also might make sense to assure they are all well identified for the benefit of practitioners, funders, and scholars alike, as to enable the broadest and most effective use of CLT models today.

Acknowledgments

We extend our gratitude to every community land trust practitioner who contributed to the 2022 Census study. We also thank Grounded Solutions Network for supporting this research. We are particularly grateful to Emily Thaden for her insights during the early development stage of this study. Additionally, we thank Cissi Xu for her research assistance. Our appreciation goes to Dr. Pfeiffer for editing this article and the three anonymous reviewers for their constructive comments. All errors are our own.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Notes on contributors

Ruoniu Wang

Ruoniu Wang is an assistant professor in the Runstad Department of Real Estate in the College of Built Environments at the University of Washington. He studies spatial justice and inclusive communities, including their impacts reflected in the built environment, human behaviors, and policy interventions. He has designed and conducted a U.S. Census of inclusionary housing policies, a U.S. census of community land trusts, and a national performance evaluation of shared equity homeownership programs. He grounds his research with applied tools to democratize data for low-income communities.

Jason Spicer

Jason Spicer is an assistant professor at the Marxe School of Public and International Affairs at Baruch College at the City University of New York. His research and teaching focuses on alternative economic models most commonly associated with community economic development and the social economy. He is the author of a new book on cooperatives with Oxford University Press, and his research has appeared in a wide range of leading social science journals.

Notes

1. This was the first such study undertaken in a decade, the 2022 Census of CLTs and Shared Equity Entities (see Wang et al., 2023), As explained in the Census documentation, the purpose was to comprehensively capture various types of CLT models. Many nonprofits have adopted CLT-like shared equity homeownership models using ground lease and tri-partite board structure and may refer to themselves as “shared equity programs,” or “permanently affordable housing programs.” Although they may not self-report as CLTs, they have effectively employed the CLT model, and, as noted later in this study, sometimes may politically caucus with the CLT “movement” and “industry.” The CLT/shared equity dataset does not include co-operative housing, which departs from the traditional CLT model.

2. By the mid-1990s, the number of CLTs in operation in the U.S. exceeded 100. https://cltweb.org/timeline/

3. Shared equity homes facilitate affordable homeownership preservation through a resale restriction, often administered by a CLT or other nonprofit organization. These homes typically utilize a legal mechanism‚usually a ground lease or a deed restriction, that: (1) includes provisions ensuring that the homes shall be kept affordable for subsequent families below a certain income level for an affordability term of at least 30 years after recordation; (2) includes a resale formula that limits the homeowner’s proceeds upon resale; and (3) has a preemptive purchase option provided to the program administrator or such administrator’s assignee (e.g., an eligible homebuyer) who may purchase the homeownership unit from the homeowner at resale.

4. Though some CLTs may contain housing co-operatives, the co-operative movement itself is organizationally and analytically separate from the CLT movement. Housing co-operatives involve joint ownership of the building, with shareholders owning shares in the co-operative corporation. CLTs and related SE models typically operate with a different logic with the land and improvement value owned separately, and without a shareholder model. Both are nonetheless part of a broader suite of collective tenure models (Ehlenz, 2018; Wegmann et al., 2017), but U.S. housing cooperatives emerged through a separate process and are beyond the scope of this study.

5. Though, to our knowledge, they are less commonly used in social science-based research, waterfall models are common in computer science, as well as in engineering, and have long been used in the development of new software and in project implementation. They have also been utilized in finance, notably real estate finance (Altshuler & Schneiderman, 2011; Staiger, 2018). First developed and proposed by a California Institute of Technology professor (Royce, 1970), the traditional waterfall model involved a linear, sequential assignment process, in which tasks, phases, or data are sequentially partitioned, in step-wise and, if possible, mutually exclusive fashion (Alshamrani & Bahattab, 2015; Balaji & Murugaiyan, 2012; Kramer, 2018), with remaining tasks, phases, or data further classified by subsequent step-wise partitioning.

6. Non-profits with a CLT-like, SE program were defined as those entities operated by a nonprofit, holding land in trust, and using resale-formula to ensure permanent affordability for their homeownership units.

7. Some variables such as the share of internal revenue to support operating cost of SE program can be grouped under two or more categories. The main purpose of the grouping is to help interpret comparison results for a long list of variables.

8. For example, instead of employing membership, some CLTs chose to establish a homeowner committee, which may take on some of the responsibilities typically performed by corporate members.