Do Homeowners Move When a Different-Race Neighbor Moves in Next Door?

The study asks whether homeowners are more likely to move when a new next-door neighbor is of a different race. The authors use national housing transaction data matched with mortgage records, allowing them to observe both household race and the exact timing and location of moves. They compare move rates for homeowners who receive a different-race neighbor immediately next door versus two or three doors away on the same block. Both Black and White homeowners are more likely to move when the different-race neighbor is next door. The size effects are modest but meaningful: about a 6% increase for Black homeowners and 3% for White homeowners.

The Policy Scientist's Perspective

This article offers timely evidence on a policy issue with broad societal implications: how neighborhood change continues to reflect racial dynamics even in a period of declining explicit prejudice. Understanding these mechanisms matters because residential sorting shapes access to schools, safety, health, and long-term mobility. The data set used in the study is unusually large, precise, and well-matched, allowing credible identification. The nearest-neighbor method provides a compelling quasi-experimental strategy with clear causal logic. The findings appear generalizable to other metropolitan areas, and the study makes a meaningful contribution to the modern segregation literature.

Full Citation and Link to Article

Bayer, P., Casey, M., McCartney, W. B., Orellana-Li, J., & Zhang, C. (forthcoming). Distinguishing causes of neighborhood racial change: A nearest neighbor design. American Economic Review.

Central Research Question

The article asks whether homeowners respond directly to the racial identity of a new next-door neighbor, independent of broader neighborhood change, shifting amenities, or expectations about future demographic transitions. Specifically, the study investigates whether Black and White incumbent homeowners are more likely to move when a new homeowner of a different race arrives immediately next door, as compared to a similar arrival two or three houses away on the same residential block. The authors frame the question as a test of whether hyper-local racial preferences—distinct from preferences over neighborhood-level composition—continue to shape residential mobility in contemporary U.S. housing markets.

Previous Literature

The article builds on longstanding theories of racial sorting and neighborhood transition, including Schelling’s classic tipping model, which argues that small changes in racial composition can trigger broader demographic shifts. Earlier empirical work has documented stable patterns of racial stratification, yet researchers have struggled to determine whether households respond directly to neighbors’ race or indirectly to correlated changes in amenities, expectations, or socioeconomic indicators.

Survey-based studies from Farley, Frey, Logan, Charles, and others show that racial preferences remain present in stated residential choices, but these methods are limited because respondents must imagine hypothetical neighborhoods with all other attributes held constant—an unrealistic cognitive task. More recent quasi-experimental research attempts to separate preferences over racial composition from preferences over neighborhood amenities, but challenges remain because demographic change typically coincides with shifts in safety, schooling, commercial services, and perceived future neighborhood trajectories.

The authors contribute to this literature by proposing a research design that isolates the racial identity of the nearest neighbor from other contextual changes. They also extend more recent methodologies used in related work on political sorting, social interactions, and residential spillovers. The study’s contribution lies in its ability to provide a micro-level test of whether the race of the immediately adjacent neighbor independently affects mobility decisions, thus informing broader debates about segregation, integration, and the persistence of racial stratification in U.S. cities.

Data

The study uses two primary data sources. First, the authors draw on CoreLogic housing transaction records, which provide detailed information on property characteristics, transaction timing, home sales, and precise geolocation. These records allow the construction of a longitudinal panel of homes and their ownership transitions.

Second, these transaction data are matched to Home Mortgage Disclosure Act (HMDA) loan application files, which offer demographic information about buyers, including race, ethnicity, income, loan purpose, and lender characteristics. The matched dataset thus combines precise spatial information with homeowner demographics and financial details. The authors impose a series of restrictions to ensure comparability across units and to minimize heterogeneity in the housing stock. They restrict the analysis to owner-occupied units built after 1900, exclude very large or unusually old homes, and limit the sample to census blocks dominated by single-family or small multifamily homes.

The study focuses exclusively on Black and White homeowners because these groups represent the central axis of historical residential stratification in U.S. housing markets. Hispanic and Asian households are excluded primarily because racial categorization is more complex for these groups and because sample sizes for Asian households are small.

The resulting dataset spans nearly two decades and includes tens of millions of homeowner-quarter observations. It is sufficiently large to identify rare but meaningful events—such as the arrival of a new different-race neighbor immediately next door—and to support high-precision estimation. The dataset’s granularity and national scope allow examination of neighborhood racial change in a wide range of contemporary metropolitan areas.

Methods

The authors employ a nearest-neighbor quasi-experimental design, comparing incumbent homeowners who receive a new different-race neighbor immediately next door (one door away) with those who receive such a neighbor two or three doors away on the same side of the same block during the same quarter. Because these homeowners are geographically proximate, they share the same exposure to block-level amenities, school zones, public services, and neighborhood-level demographic change. The primary identifying assumption is that within such fine geographic distances, the exact location of the newly arrived neighbor is quasi-random with respect to unobserved incumbent attributes.

The logic of the design is straightforward: if changes in amenities or expectations drive mobility, then residents one, two, or three doors down should respond similarly. If, however, the identity of the immediate neighbor has independent causal force, the incumbent directly adjacent to the new homeowner should exhibit a greater mobility response than those slightly farther away.

The authors estimate regression models with block-by-quarter fixed effects, ensuring comparisons occur only within the same micro-neighborhood and time period. They also conduct balance tests to show that the treatment (next-door) and control (two to three doors away) groups are virtually identical in observable characteristics, including home size, age, homeowner income, loan type, and tenure. Additional robustness tests address alternative explanations, such as responses to the arrival of any new neighbor, income differences, or potential within-block amenity variation. Across all specifications, the estimated effect remains stable in magnitude and highly statistically significant.

Findings/Size Effects

The results indicate that both Black and White incumbent homeowners exhibit higher move propensities when a new neighbor of a different race moves in immediately next door compared to when the new neighbor moves in two or three houses away. For Black homeowners, the effect corresponds to roughly a 6 percent increase in the likelihood of moving within two years. For White homeowners, the effect is approximately a 3 percent increase. These estimates represent lower bounds because the control group—those two to three doors down—may also respond, albeit less strongly, to the same demographic change.

The study also documents substantial heterogeneity. The effects are strongest in high-density settings where homes are closest together; for White incumbents receiving a new Black neighbor, the effect diminishes sharply as physical distance increases, and it essentially disappears beyond approximately 25 meters. This implies that hyper-local cues matter significantly more than broader neighborhood composition. Income-related heterogeneity is also present: incumbents with higher incomes show stronger responses, as do cases where the new different-race neighbor has a lower income. The authors also find stronger responses in Northern cities than in Southern ones, consistent with longstanding regional differences in segregation patterns.

Importantly, the authors show that proximity to a new different-race neighbor has no detectable effect on subsequent housing prices within the block, suggesting that changes in expected amenities or future neighborhood trajectories are unlikely to drive the observed mobility patterns. Instead, the evidence supports the conclusion that the racial identity of the next-door neighbor plays a direct role in residential decision-making.

Conclusion

The study demonstrates that racial preferences at the level of immediate neighbors remain a measurable and meaningful force in contemporary residential mobility decisions. The findings suggest that even in an era of declining overt prejudice and increasing suburban integration, hyper-local racial dynamics continue to shape patterns of neighborhood change. Because residential mobility drives long-run segregation patterns and influences access to opportunity, the study provides an important contribution to research on housing markets and urban inequality.

Methodologically, the article advances efforts to identify causal drivers of neighborhood sorting by introducing a design that effectively isolates neighbor-specific racial preferences from correlated amenities and expectations. The results indicate that the persistence of racial stratification cannot be understood solely through broad neighborhood-level processes; micro-level interactions matter as well. The study’s national scope and rigorous identification strategy allow for credible generalization to a wide array of U.S. metropolitan areas, underscoring its relevance for ongoing research into the mechanisms behind residential segregation and the sustainability of integrated neighborhoods.