Where Subsidized Housing Lands, and Why It Matters: A Review of LIHTC Siting, Neighborhood Opportunity, and Fair Housing

1. Introduction

The Low-Income Housing Tax Credit (LIHTC) is the federal government's primary tool for encouraging the development and rehabilitation of affordable rental housing (Keightley, 2023). Created by the Tax Reform Act of 1986, it became the de facto federal rental production program within its first decade (Cummings and DiPasquale, 1999). Unlike housing vouchers, which let a household choose its neighborhood, LIHTC attaches the subsidy to a building that must remain affordable for decades. A single allocation decision therefore fixes the neighborhood context for every household that lives in the building over its compliance period, and often beyond it (Lens and Reina, 2016).

That fact carries more weight now than when the program began. A large body of research in sociology and economics argues that neighborhoods shape life chances (Sampson, 2012; Sharkey, 2013), and experimental and quasi-experimental work now shows that the neighborhood a child grows up in affects adult earnings, college attendance, and other long-run outcomes (Chetty, Hendren, and Katz, 2016; Chetty and Hendren, 2018; Chyn, 2018). Tract-level estimates of these effects are publicly available (Chetty et al., 2018), so the opportunity profile of a proposed LIHTC site can be measured before an allocation is made.

Location is also a legal question. The Fair Housing Act obliges federal housing programs to affirmatively further fair housing, and HUD's 2015 rule set out a planning process to that end (U.S. Department of Housing and Urban Development, 2015). In Texas Department of Housing and Community Affairs v. Inclusive Communities Project (2015), a case that arose from a challenge to the state's LIHTC allocations, the Supreme Court held that disparate-impact claims are cognizable under the Fair Housing Act. Siting is therefore an efficiency question, an equity question, and a civil-rights question at once.

Despite this, the relevant evidence sits in literatures that rarely speak to one another. Studies of LIHTC location are mostly descriptive and focus on poverty and race. Studies of neighborhood effects focus on households that move, mostly with vouchers, and rarely on LIHTC tenants. Studies of LIHTC spillovers focus on property values and crime in receiving neighborhoods, and rarely ask what those results imply for where projects should go. This review brings the three together.

We proceed as follows. Section 2 describes the scope of the review. Section 3 explains how program design builds location choices into each allocation. Section 4, Section 5, Section 6 and Section 7 take up the four core questions: where LIHTC developments go, why they go there, what location means for tenants, and what it means for receiving neighborhoods. Section 8 weighs these findings against one another. Section 9 maps policy levers to each part of the siting decision, and Section 10 sets out a research agenda.

2. Scope and Approach

This is a narrative review rather than a systematic review or meta-analysis. We focus on peer-reviewed empirical work on the United States, supplemented by policy reports and legal sources where they document program rules or policy changes. We include studies that address at least one of four questions: where LIHTC developments are located relative to neighborhood poverty, race, and opportunity; what explains those locations; how neighborhood context affects the households that live in subsidized housing; and how LIHTC developments affect the neighborhoods around them.

Two boundaries are worth stating. First, we draw on the neighborhood-effects literature even where it does not study LIHTC directly, because it supplies the main rationale for siting in high-opportunity areas. Second, we draw on studies of other place-based subsidized housing, such as public housing and city-financed housing, where they illuminate mechanisms that also apply to LIHTC. We note where a finding comes from a program other than LIHTC.

Sources were identified through searches of economics, planning, sociology, and housing policy journals and working-paper series, using terms that combined the program name with siting, location, neighborhood, spillover, property values, segregation, and opportunity, and through the reference lists of the studies found. Because the review's purpose is to connect literatures rather than to estimate a pooled effect, we prioritize studies that are widely cited, use credible designs, or document institutional features that the other studies take for granted. The selection is necessarily partial, and the conclusions should be read as a synthesis of the main lines of evidence rather than a complete census of the literature.

We weight evidence by research design. Where available, we emphasize randomized experiments, natural experiments, and quasi-experimental designs that exploit program rules, such as discontinuities in credit formulas or changes in tract designations. Descriptive studies are essential for documenting patterns, but we are careful not to read causal explanations into them. Table 1 summarizes the empirical studies that carry the most weight in the sections that follow.

3. How Program Design Builds in Location

LIHTC is administered through the tax code but allocated by states. Each year, state housing finance agencies receive a per-capita allocation of credits, which they award to developers through a competitive process governed by a Qualified Allocation Plan (QAP). Developers sell the credits to investors, usually through syndicators, and the equity raised reduces the debt the project must carry, allowing it to charge restricted rents (Keightley, 2023; Desai, Dharmapala, and Singhal, 2010). The credit is claimed over ten years, while properties must meet income and rent restrictions for a 15-year compliance period, followed by an extended-use period that brings the total commitment to at least 30 years for most properties.

Two credit types matter for siting. The competitive 9% credit is designed to deliver a subsidy worth about 70% of a project's qualified basis in present value, and the 4% credit, used mainly with tax-exempt bond financing, about 30% (Keightley, 2023). Because 9% credits are scarce, the QAP's scoring rules directly shape which sites win. The 4% credit is not subject to the same competitive allocation, so its location depends more on where bond-financed projects pencil out.

Location enters program rules in several explicit ways. Projects located in Qualified Census Tracts (QCTs), which are defined by high poverty or a high share of low-income households, or in Difficult Development Areas, which have high construction, land, and utility costs relative to incomes, may claim up to 130% of eligible basis. The federal statute also directs states to give preference to projects in QCTs whose development contributes to a concerted community revitalization plan. Since the Housing and Economic Recovery Act of 2008, states may also designate additional buildings for the basis boost, giving them a way to favor sites outside QCTs.

These rules embed a tension. The QCT boost and revitalization preference reward investment in low-income neighborhoods, consistent with a place-based view of the program as a tool for community development. The fair-housing obligation and the neighborhood-effects evidence point toward lower-poverty, higher-opportunity locations, consistent with a people-based view of the program as a route to mobility. The siting evidence reviewed below should be read with this tension in mind.

Program design also shapes who lives in LIHTC housing. Using tenant-level data from 18 states, O'Regan and Horn (2013) find that LIHTC serves many households with extremely low incomes, most of whom also receive rental assistance, and that developments range from economically mixed to heavily concentrated among the poorest tenants. This matters for siting: the neighborhood a building occupies affects not only moderate-income working households but also some of the lowest-income households in subsidized housing.

The time limits on affordability add a further dimension. After the initial 15-year compliance period, investors typically exit and properties are recapitalized, resyndicated, or, where owners can obtain release from extended-use commitments, converted toward market rents (Keightley, 2023). The incentive to convert is strongest in neighborhoods where market rents have risen, which are often the same neighborhoods that have become higher-opportunity places. As a result, the opportunity profile of the LIHTC stock reflects not only where properties were originally sited but also which ones remain affordable, and the program can lose its best-located units first unless preservation policy intervenes (Lens and Reina, 2016). Siting and preservation are therefore two stages of the same decision.

4. Where LIHTC Developments Go

Spatial analyses consistently find that LIHTC properties are clustered. Dawkins (2013) applies point-pattern methods to the largest U.S. metropolitan areas and finds that in most of them, LIHTC properties are more clustered than the multifamily housing stock as a whole. Clustered properties tend to sit in dense central-city locations with higher poverty rates and larger minority populations, although the degree of clustering varies considerably across metropolitan areas. This variation is itself informative: it suggests that local institutions and markets, not only federal rules, shape where projects land.

The early program record points in the same direction. In their analysis of the first decade of LIHTC, Cummings and DiPasquale (1999) document the program's growth into the main federal production vehicle and the changing economics of credit pricing, which affected which projects were feasible. Across studies, central cities have hosted a disproportionate share of LIHTC units, even as some states and metropolitan areas have seen growth in suburban placements.

Whether clustering raises or lowers segregation depends on the scale of analysis and the comparison group. Owens (2015) examines the spread of assisted housing across U.S. metropolitan neighborhoods between 1980 and 2005–2009. She finds that deconcentration produced only modest gains in economic integration for very low-income families, in part because high-income families became more segregated over the same period. Even a program that spreads its units more widely may therefore have limited effects on segregation if other forces move in the opposite direction.

City-level analyses can diverge from metropolitan averages. Within a single city, the relevant comparison is not between central city and suburb but between neighborhoods that differ in race, income, land costs, and zoning. In New York City, for example, tract-level data for 2003–2023 show LIHTC units concentrated in predominantly Black and Hispanic tracts (Zhang, 2026), a pattern that metropolitan-scale summaries would not reveal. Within-city studies are still relatively rare, and they matter because many of the decisions that shape siting, including zoning, land disposition, and local subsidy layering, are made at the city level.

Claims about where LIHTC goes also depend on the benchmark. Comparing LIHTC tracts with the population as a whole answers a different question from comparing them with the rental stock, with the existing multifamily stock, or with the parcels on which multifamily housing could legally be built. A program that simply followed the existing rental stock would look concentrated relative to the population, and a program that located wherever land was zoned for multifamily use would look concentrated relative to both. Dawkins's comparison with the multifamily stock is informative precisely because it holds this constant. Studies that use different benchmarks are not in conflict so much as answering different questions, and the choice of benchmark should follow from the policy question at hand: whether allocators are choosing among feasible sites in a biased way, or whether the set of feasible sites is itself the problem.

Finally, the location of the stock changes over time not only through new construction but also through exits. As properties reach the end of their affordability periods, owners in appreciating neighborhoods have the strongest incentive to convert to market rents. Lens and Reina (2016) show that the opportunity profile of subsidized properties approaching expiration varies, so decisions about which properties to preserve are themselves siting decisions. A complete picture of where LIHTC housing is located must account for both entry and exit.

5. Why They Go There

Several explanations recur in the literature. They are difficult to separate empirically because they push in the same direction, toward lower-cost, higher-poverty neighborhoods.

Program incentives. The QCT basis boost increases the credit available for projects in high-poverty tracts, improving their financial feasibility relative to otherwise similar projects elsewhere. Evidence that this incentive changes behavior comes from studies that exploit the credit formula. Baum-Snow and Marion (2009) use a discontinuity created by the boost and find that tracts where projects receive 30% more credit receive substantially more low-income units. Freedman and Owens (2011) use changes in the federal rules that determine QCT status as a quasi-experiment. These studies establish that the boost moves development at the margin. State QAPs add a second layer of incentives. Ellen and Horn (2018) find that awarding points for high-opportunity locations is associated with modest shifts toward lower-poverty siting, and Khadduri (2013), reviewing QAPs from 36 states, finds that only a minority offered incentives for projects in high-opportunity neighborhoods at the time, and recommends limiting priorities that favor low-income areas.

Land and construction markets. LIHTC developers face the same land and construction costs as market-rate developers, and land is cheapest in low-income neighborhoods. Evidence on crowd-out shows how closely LIHTC follows the market. Eriksen and Rosenthal (2010) find substantial crowd-out of unsubsidized rental construction, with instrumental-variables estimates larger than OLS because LIHTC development is drawn to areas already ripe for new construction. They conclude that the program affects where low- and moderate-income rental housing is built more than how much is built. Sinai and Waldfogel (2005) estimate that roughly two of every three government-subsidized units displace units the private market would otherwise have provided, with less crowd-out in more populous markets. Malpezzi and Vandell (2002) likewise find evidence consistent with a high rate of substitution, though their test cannot rule out lower rates. Freedman and McGavock (2015) estimate that 30–50% of LIHTC-induced rental development is offset by fewer unsubsidized units in the same neighborhood. Taken together, these results imply that LIHTC siting is shaped by the same development economics that drive private construction.

Zoning and land-use regulation. Multifamily housing cannot be built where it is not allowed. Restrictive local regulation is widespread and is most intense in coastal metropolitan areas (Gyourko, Hartley, and Krimmel, 2021), and it raises housing costs where it binds (Glaeser and Gyourko, 2018). Rothwell and Massey (2009) find a strong association between low-density zoning and racial segregation across metropolitan areas, and evidence that anti-density zoning slows desegregation over time. Because LIHTC projects are almost always multifamily, low-density zoning effectively removes many higher-income neighborhoods from the set of feasible sites. Historical and political accounts show that such regulation has long been used, explicitly or implicitly, to maintain racial and economic boundaries (Rothstein, 2017; Trounstine, 2018).

Local opposition. Even where zoning permits multifamily housing, projects often require discretionary approvals that give neighbors leverage. Attitudes toward affordable housing are shaped by ideology and by racial and class stereotyping, and neighborhood opposition is an effective barrier to siting (Tighe, 2010). Using a national survey, Scally and Tighe (2015) show that the participation of wealthier, more organized groups in local review can crowd out marginalized groups and push projects toward places where opposition is weaker. Einstein, Glick, and Palmer (2019) document that participants in local land-use meetings are disproportionately opposed to new housing and unrepresentative of their communities, so participatory institutions can reproduce inequality. Developers anticipate this, and some sites are never proposed at all.

Development capacity. A fifth, less studied factor is where the organizations that build LIHTC housing are based and what kinds of projects they know how to deliver. Mission-driven nonprofit developers and community development corporations are often rooted in the low-income neighborhoods they serve, and their pipelines, relationships with local officials, and access to city-owned land reflect that base. The federal revitalization preference for projects in QCTs reinforces this pattern. Whether developer capacity is an independent cause of clustering or a response to the other constraints has, to our knowledge, received little direct study.

These explanations interact. The QCT boost makes low-income sites more profitable, land markets make them cheaper, zoning makes alternatives illegal, and opposition makes alternatives slow and uncertain. Because they reinforce one another, relaxing any one constraint may have limited effects unless others are addressed at the same time.

6. What Location Means for Tenants

The case for siting in higher-opportunity areas rests on the neighborhood-effects literature. Sociological work has long argued that neighborhoods shape residents' lives through social organization, institutions, and exposure to violence, and that these effects can persist across generations (Sampson, 2012; Sharkey, 2013). A central lesson of more recent reviews is that the question is not whether neighborhoods matter but which dimensions matter, for whom, and over what time frame (Sharkey and Faber, 2014).

Experimental evidence has sharpened this picture. The Moving to Opportunity (MTO) experiment offered families in high-poverty public housing vouchers to move to lower-poverty neighborhoods. Early evaluations found no significant effects on adult economic self-sufficiency but improvements in mental health for adults and female youth (Kling, Liebman, and Katz, 2007), and long-term follow-up found improvements in adults' subjective well-being (Ludwig et al., 2012). Later analysis showed that children whose families moved before age 13 had higher adult earnings and college attendance (Chetty, Hendren, and Katz, 2016). Using families that moved across the country, Chetty and Hendren (2018) find that outcomes improve with each additional year of childhood spent in a better neighborhood. Chyn (2018), studying children displaced by public housing demolitions in Chicago, finds higher employment and earnings and fewer violent-crime arrests in adulthood relative to children whose buildings were not demolished.

Why neighborhoods matter is less settled than whether they do. Proposed mechanisms include the quality of local schools, exposure to violence, the density of local institutions and social networks, environmental hazards, and access to jobs (Sampson, 2012; Sharkey and Faber, 2014). The MTO results for adults, with gains in mental health and well-being but not in earnings (Kling, Liebman, and Katz, 2007; Ludwig et al., 2012), suggest that safety and stress may be as important as economic opportunity for some outcomes. For LIHTC siting, the distinction matters because opportunity indices weight these mechanisms differently, and a site that ranks highly on one dimension, such as school quality, may rank poorly on another, such as transit access to jobs.

These findings have a direct implication for LIHTC. If childhood exposure to a neighborhood affects adult outcomes, then the location of family LIHTC housing shapes the long-run prospects of the children who grow up in it. The Opportunity Atlas (Chetty et al., 2018) translates these estimates to the tract level and has made it possible for allocating agencies to measure neighborhood opportunity directly.

Supply-side placement also complements mobility policy. Bergman et al. (2024) show that customized search assistance raised the share of voucher families moving to high-upward-mobility areas from 15% to 53%, which suggests that barriers to neighborhood choice, not only preferences, keep low-income families out of high-opportunity areas. Siting LIHTC housing in those areas removes one such barrier by placing affordable units where few would otherwise exist.

Evidence linking neighborhood effects to LIHTC tenants specifically is more limited. Ellen, Horn, and Kuai (2018) examine the neighborhoods LIHTC tenants actually live in and find that poor and minority tenants live in significantly more disadvantaged neighborhoods than other LIHTC tenants, so averages across all LIHTC properties can overstate the access that the most disadvantaged tenants receive. They also suggest separating access to transit from broader opportunity indices, since the two may conflict. Drawing on residents' own experiences in California, Reid (2019) documents the benefits of stable, affordable housing as well as the barriers tenants face in improving their economic situations, and suggests that residents' own views of opportunity may differ from those embedded in common opportunity measures.

Two cautions follow. First, most causal evidence on neighborhood effects comes from voucher holders and displaced public housing residents, not from LIHTC tenants, and the effects of moving into a new LIHTC building may differ. Second, the benefits of a high-opportunity neighborhood may depend on whether residents can access its institutions, such as schools, which in turn depends on design, management, and local attendance policies.

7. What Location Means for Neighbors

The other side of the trade-off is the receiving neighborhood. Early studies of affordable housing and nearby property values produced mixed results. Reviewing seventeen such studies, Nguyen (2005) concludes that the effect depends on the design and management of the housing, its compatibility with the host neighborhood, and the degree of concentration, rather than on affordable housing as a category.

Studies using stronger designs have generally found neutral or positive effects in low-income neighborhoods. In New York City, Schwartz et al. (2006) find sizable positive spillovers from city-supported housing investments that persist over time, are larger for bigger projects, and appear partly to reflect the removal of existing nuisances such as vacant lots and abandoned buildings. Ellen et al. (2007) find little evidence that federally subsidized rental housing, including LIHTC, depressed nearby property values in the city. Using a discontinuity in the credit formula, Baum-Snow and Marion (2009) find that LIHTC developments raise property values in declining areas and reduce incomes in gentrifying ones. Diamond and McQuade (2019) estimate that a LIHTC development in a low-income neighborhood raises nearby house prices and generates sizeable welfare gains for local homeowners and renters, while effects in higher-income areas are negative.

Effects also differ across cities and with concentration. Woo, Joh, and Van Zandt (2016) find negative price effects near LIHTC developments in Charlotte but upgrading effects in Cleveland, varying with neighborhood income mix. In Cook County, Voith et al. (2022) study 508 developments built between 1997 and 2016 and find positive effects on surrounding property values overall, particularly in lower-income communities, which speaks to concerns that concentrating projects erodes their benefits.

Evidence on other outcomes points in a similar direction. Freedman and Owens (2011) find that LIHTC development in the poorest neighborhoods reduces violent crime, measurable at the county level, with no detectable effect on property crime. Di and Murdoch (2013) find no robust evidence that LIHTC openings reduce the performance of nearby elementary schools in Texas. Freedman and McGavock (2015) find that LIHTC raises the share of poor residents in recipient neighborhoods and lowers median income, but the effects are small.

Property-value effects require careful interpretation. Rising prices benefit homeowners but raise costs for renters, so the same spillover can be a gain for some existing residents and a loss for others (Diamond and McQuade, 2019). Price effects also capitalize expectations about future neighborhood change, and they say little directly about the well-being of LIHTC tenants themselves. Studies that track renters' outcomes and residential moves in receiving neighborhoods, alongside prices, would help distinguish revitalization that benefits existing residents from change that displaces them (Baum-Snow and Marion, 2009; Freedman and McGavock, 2015).

Reviewing this body of work, Dillman, Horn, and Verrilli (2017) conclude that neighborhood context determines whether LIHTC production brings measurable gains, with the clearest benefits in distressed neighborhoods, where projects tend to replace blight and add investment. In stronger neighborhoods, effects are smaller and occasionally negative.

8. Reconciling the Evidence: The Siting Trade-Off

Read together, Section 6 and Section 7 present a genuine trade-off rather than a simple story. From the tenant's perspective, siting in low-opportunity neighborhoods limits access to the environments that most improve children's long-run outcomes. From the neighborhood's perspective, siting in distressed areas can bring investment, raise property values, and reduce crime. Both effects appear to be real, and both are largest in the same places: the distressed neighborhoods where LIHTC has historically clustered.

Three considerations help weigh these effects. First, they accrue to different people. The neighborhood benefits documented in Section 7 accrue largely to existing homeowners and residents, while the opportunity costs accrue to LIHTC tenants and especially their children. A siting pattern that maximizes one can sacrifice the other. Second, they operate on different time scales. Property-value effects can appear within a few years, while the effects of childhood exposure accumulate over decades. Third, both depend on context. Not every low-income neighborhood is a low-opportunity neighborhood, and not every high-income neighborhood offers the institutions that drive mobility (Reid, 2019).

A further consideration is that the trade-off is not purely a matter of aggregate welfare. Fair-housing law treats the exclusion of protected groups from parts of a housing market as a harm in itself, independent of whether the neighborhoods they are steered toward improve (Texas Department of Housing and Community Affairs v. Inclusive Communities Project, 2015). From this perspective, an allocation pattern that systematically places family housing in minority neighborhoods raises concerns even if each individual project generates local benefits. Conversely, a pattern that withdraws investment from minority neighborhoods in the name of mobility raises its own equity concerns for the residents who remain. Welfare estimates of the kind produced by Diamond and McQuade (2019) are an essential input to these judgments, but they do not settle them.

This suggests that the useful policy question is not whether to site in low-income or high-opportunity neighborhoods, but what balance across the two a jurisdiction's portfolio should strike, and which specific sites deliver the most value on each dimension. A portfolio that includes both revitalization projects in distressed neighborhoods and family housing in high-opportunity neighborhoods can capture benefits on both sides. The evidence reviewed in Section 5 suggests that most jurisdictions are far from such a balance, because the constraints on siting in high-opportunity areas are binding.

9. Policy Levers

Each explanation in Section 5 points to a different lever, and each lever acts on a different part of the siting decision.

Allocation rules. QAP scoring can reward high-opportunity locations directly (Ellen and Horn, 2018), limit the weight given to low-income-neighborhood priorities, and remove provisions that block such projects, such as requirements for local government support (Khadduri, 2013). California's adoption of opportunity maps in its allocation process is one prominent example, and resident perspectives suggest that the definition of opportunity used in such maps deserves scrutiny (Reid, 2019).

Basis boosts. Since 2008, states have been able to extend the basis boost to additional buildings they designate, which lets them offset the higher land costs of high-opportunity sites rather than rewarding only Qualified Census Tracts. How states use this discretion, and with what effect, is an open empirical question.

Land use. Zoning reform that permits multifamily housing in more neighborhoods addresses the constraint documented by Rothwell and Massey (2009) and Gyourko, Hartley, and Krimmel (2021). Because LIHTC follows development economics closely (Eriksen and Rosenthal, 2010), expanding the set of legal and feasible sites may be a precondition for any allocation reform to work.

Local review. Reforms that reduce discretionary review for projects that meet objective standards limit the leverage of organized opposition (Scally and Tighe, 2015; Einstein, Glick, and Palmer, 2019).

Preservation. Prioritizing the preservation of subsidized properties in high-opportunity neighborhoods protects access that would otherwise be lost as affordability periods end (Lens and Reina, 2016).

Complementary mobility support. Search assistance and landlord outreach for voucher holders (Bergman et al., 2024) address the demand side of the same problem and can complement supply-side siting policy.

Monitoring and evaluation. Allocating agencies can make siting outcomes visible by reporting the neighborhood characteristics of each year's awards against stated portfolio goals, using a consistent opportunity measure and a consistent benchmark. Because QAPs are revised regularly, building evaluation into those revisions, for example by phasing in new scoring criteria or comparing similar applications on either side of a scoring threshold, would produce the kind of evidence on policy effects that the literature currently lacks (Ellen and Horn, 2018).

Table 2 summarizes these levers, the constraint each addresses, and the main open questions.

Fair-housing law provides the frame for all of these levers. The disparate-impact standard confirmed in Texas Department of Housing and Community Affairs v. Inclusive Communities Project (2015) and the planning obligations in HUD's 2015 rule (U.S. Department of Housing and Urban Development, 2015) give allocating agencies both reasons and tools to examine the cumulative effects of their siting decisions, although the rule's status has changed since it was issued.

10. Gaps and Research Agenda

Five gaps stand out.

Separating mechanisms. Few studies separate the effect of program incentives from the land-market, regulatory, and political factors that push siting in the same direction. Designs that exploit changes in allocation rules, QAP scoring, or tract designations offer one path, as do zoning reforms that change the set of feasible sites without changing program incentives. Studies that combine allocation records with parcel-level land and zoning data would make it possible to observe the choice set developers actually face.

Tenant outcomes. The causal neighborhood-effects evidence comes mostly from vouchers and household moves, and has not been linked directly to the long-run outcomes of LIHTC tenants. Linked administrative records on tenants, combined with tract-level opportunity measures, would make it possible to estimate whether the neighborhood in which a LIHTC property sits affects its residents' earnings, health, and children's education.

Within-city variation. Most siting evidence is national or metropolitan, while allocation, zoning, and land markets operate locally. Within-city studies can identify where policy constraints bind and can evaluate local reforms (Zhang, 2026).

Measuring opportunity. Opportunity indices differ in what they measure, and residents may value different features than those indices capture (Ellen, Horn, and Kuai, 2018; Reid, 2019). Research that tests which measures predict tenant outcomes would help agencies choose among them.

The 4% credit and preservation. Much of the siting literature focuses on competitive 9% allocations, yet bond-financed 4% projects account for a large share of units, and their location is governed less by QAP scoring than by bond availability and project economics (Keightley, 2023). Little is known about whether their siting differs systematically from 9% projects, or about how preservation and recapitalization decisions shift the location of the stock over time (Lens and Reina, 2016).

Methodologically, the strongest designs in this literature exploit features of program rules that generate quasi-random variation in where credits flow, such as the credit-formula discontinuity used by Baum-Snow and Marion (2009) and the QCT rule changes used by Freedman and Owens (2011). Similar opportunities arise whenever allocation rules change: QAP revisions that introduce or remove opportunity points, state-designated basis boosts, and zoning reforms that alter the set of feasible sites. Pairing such variation with tenant-level records and with measures of neighborhood change on both sides of a boundary would allow researchers to estimate the full trade-off described in Section 8 within a single design, rather than assembling it from studies of different places and periods.

11. Conclusions

Where LIHTC housing is built matters for the households who live in it and for the neighborhoods that receive it. The evidence reviewed here shows persistent clustering in higher-poverty, higher-minority areas; explanations for that clustering that include program incentives, land markets, zoning, and local opposition; large effects of neighborhood exposure on children's outcomes; and real, context-dependent benefits of LIHTC investment for distressed neighborhoods.

These findings do not resolve into a single rule. They suggest instead that jurisdictions should manage their LIHTC portfolios deliberately, combining revitalization projects with family housing in high-opportunity areas, and that doing so requires acting on several constraints at once. Better evidence on which levers actually move siting, and on how siting affects LIHTC tenants themselves, would put that balance on a firmer footing.