Whether new rail transit stations displace low-income residents from their immediate vicinity remains unresolved at the spatial scale where theory predicts the effect operates. The only national, individual-level US test (Delmelle and Nilsson 2020, using the Panel Study of Income Dynamics linked to census tracts, station openings 1970–2014) found no evidence that low-income residents disproportionately leave new transit neighbourhoods — but tract-level geography, uniform treatment of all stations, and unobserved within-tract moves leave two coupled questions open: (i) does highly localised, sub-tract (block-face or parcel-level) displacement, including short-distance moves that begin and end inside the same census tract, occur immediately adjacent to new stations; and (ii) how do out-migration effects vary across station types (accessibility, ridership, transit-oriented versus transit-adjacent development, land availability) and neighbourhood/market contexts? Later fine-grained studies give context-dependent answers — demolition-driven direct displacement near densified Zurich stations, renter displacement within 1.8 km of new Shanghai stations, only modest churn signals across 19 UK schemes, and no disproportionate out-migration in Los Angeles tax-record data — while US household-level datasets of the needed spatial precision (e.g., consumer reference data — commercial databases of household addresses) have so far been applied to gentrification generally or to single corridors. The problem is to deliver household-level, sub-tract-granularity evidence that supports a causal reading of station-adjacent out-migration of low-income residents, differentiated by station and market type, with claims scoped to the station population and contexts actually studied.

Background

Rail transit investment is often accompanied by the fear of 'transit-induced gentrification': a new station raises the accessibility and amenity value of surrounding land, housing values and rents nearby tend to rise, and higher-income households may outbid lower-income households for these newly accessible locations. If the residents most dependent on transit are priced out of the locations best served by it, transit investment produces an equity paradox (for a review, see Delmelle, 2021). 'Displacement' here means the departure of lower-income residents from their neighbourhood under housing-market pressure — as distinct from ordinary residential mobility — and it is empirically hard to isolate because low-income households move frequently even where nothing is changing.

The most direct national-scale test of the displacement half of this hypothesis in the United States is Delmelle and Nilsson (2020). They linked the Panel Study of Income Dynamics (PSID), a long-running US household panel survey, to census-tract locations and traced the out-migration of individuals from tracts within five years of the opening of a new fixed-guideway rail station (a stop on rail running on its own dedicated track, such as light rail, metro, or commuter rail) anywhere in the United States between 1970 and 2014. A census tract is a moderately small statistical unit (roughly 1,600–8,000 residents) but is much larger than a station's immediate catchment. They found that low-income individuals are more mobile everywhere, but are not significantly more likely to move out of new transit neighbourhoods than comparable residents elsewhere — a null result robust across decades, for renters alone, for different time windows, and for alternative definitions of 'transit neighbourhood'.

What this design cannot see follows from its geography. Because the tract is the finest spatial unit of the PSID linkage, displacement occurring at scales finer than the tract — on the block face (the properties fronting one side of a street) or on parcels immediately adjacent to a station — or moves that begin and end inside the same tract are unobservable by construction, so the national null could reflect genuine absence of displacement or spatial aggregation that dilutes a highly localised effect. All stations were also treated uniformly, whereas theory and the price-capitalisation literature (studies of how transit access is reflected in nearby property values) suggest effects should differ with station accessibility, ridership performance, surrounding land availability, and whether development is genuinely transit-oriented (compact, walkable station-area planning, usually abbreviated TOD) rather than merely transit-adjacent. The paper's limitations and future-research discussion calls, in paraphrase, for finer-grained geocoded administrative or tax-record data to determine whether highly localised displacement exists, and for quantifying heterogeneous impacts of specific station and market types; the first author's subsequent review reiterates that census tracts 'may be too large to capture changes that may only occur in the area immediately surrounding a station' and that these limitations 'could be addressed by the use of alternate datasets' (Delmelle, 2021).

The intervening literature has moved on both fronts, but outside the original setting. Companion studies by the same group examined eviction filings (a court-record proxy for involuntary moves) near new stations in four US cities, with very limited evidence of elevated filings (Delmelle, Nilsson and Bryant, 2021); the destinations of movers, finding that low-income movers from transit neighbourhoods with limited station coverage tend to relocate to more disadvantaged tracts (Nilsson and Delmelle, 2020); and the characteristics of in-movers, using Home Mortgage Disclosure Act (HMDA) loan-application records, which are individual-level but observe buyers rather than leavers, around a new Charlotte light rail line (Delmelle, Nilsson and Schuch, 2021), complemented by a resident survey finding no elevated stated moving propensity near stations (Nilsson et al., 2020). Household-level datasets of the required spatial precision now exist — Swiss and English linked registry or consumer data, Shanghai individual rental records, Toronto tax-filer panels, and US consumer reference data (commercial databases of household addresses compiled from consumer-marketing records) such as the Data Axle historical residential file — and have produced mixed verdicts: demolition-driven direct displacement of incumbent low-income residents near densified train stations in Zurich (Lutz, Wicki and Kaufmann, 2024); rent increases and out-migration of renters replaced by wealthier newcomers within 1.8 km of new stations in Shanghai (Deng, Li and Chen, 2025); only modest signals of churn across 19 UK schemes, heterogeneous by local economic context (Kandt, van Dijk and Todd, 2025); a strong gentrification–displacement link measured with household-level data in Columbus, Ohio, though not transit-specific (Khaleel and Hanlon, 2026); and household-level evidence of gentrification displacement in New York and San Francisco (Song and Chapple, 2024), while a Los Angeles study using local tax records, as reported in Delmelle (2021), found no disproportionate out-migration. Heterogeneity by neighbourhood type and by housing tenure has been documented for single corridors in Charlotte (Schmidt and Bardaka, 2024) and Helsinki (Meriläinen et al., 2024), and survival-analysis designs (time-to-event statistical models of how long residents remain in place) for migration near corridors have now appeared (Schmidt and Bardaka, 2026). What no study has yet delivered is the combination the open question calls for: household-level, sub-tract-granularity evidence on out-migration immediately around new rail stations, differentiated by station and market type — and coverage of the national or systematically multi-city station population that would settle whether the US national null is an artifact of spatial aggregation, which remains open.

Problem Statement

The problem, as left open by Delmelle and Nilsson (2020) and not resolved by later work, is empirical and twofold, the two parts being linked aspects of one objective — measuring the displacement effect of new rail stations with adequate spatial resolution and differentiation:

(1) Localised displacement. Using geocoded household-level longitudinal data at finer spatial resolution than the census tract (e.g., administrative, tax-record, or consumer-reference data locating households at address, block or parcel level), determine whether low-income residents are disproportionately likely to move out of the immediate vicinity of newly opened rail transit stations — including short-distance displacement that begins and ends within the same census tract, which tract-level designs cannot observe — relative to matched locations without new stations. This settles whether the national null result reflects genuine absence of transit-induced displacement or spatial aggregation that dilutes a block-face-scale effect.

(2) Station- and context-heterogeneity. Quantify how any out-migration (and related displacement-pressure) effects vary across station types — accessibility, ridership performance, surrounding land availability, transit-oriented versus transit-adjacent development — and across neighbourhood and housing-market contexts, rather than treating all fixed-guideway stations uniformly.

The comparison against matched locations without new stations must be built so that its estimates support a causal reading of out-migration at this sub-tract resolution — with outcomes tied to the timing of station announcement, construction and opening — distinguishing displacement under station-induced housing-market pressure from ordinary residential mobility. Claims must be scoped to the population of stations and contexts actually studied: evidence covering the national or systematically multi-city station population is required specifically to settle whether the national null result of Delmelle and Nilsson (2020) is a spatial-aggregation artifact, whereas sub-tract evidence for a single city or corridor answers the immediate-vicinity displacement question for the population studied.

References

  1. Delmelle, E. C., & Nilsson, I. (2020). New rail transit stations and the out-migration of low-income residents. Urban Studies, 57(1), 134–151. https://doi.org/10.1177/0042098019836631
  2. Delmelle, E. C. (2021). Transit-induced gentrification and displacement: The state of the debate. In R. H. M. Pereira & G. Boisjoly (Eds.), Social Issues in Transport Planning (Advances in Transport Policy and Planning, Vol. 8, pp. 173–190). Elsevier. https://doi.org/10.1016/bs.atpp.2021.06.005 (preprint: https://doi.org/10.31235/osf.io/5ka2g)
  3. Nilsson, I., & Delmelle, E. C. (2020). Impact of new rail transit stations on neighborhood destination choices and income segregation. Cities, 102, 102737. https://doi.org/10.1016/j.cities.2020.102737
  4. Delmelle, E. C., Nilsson, I., & Bryant, A. (2021). Investigating transit-induced displacement using eviction data. Housing Policy Debate, 31(2), 326–341. https://doi.org/10.1080/10511482.2020.1815071
  5. Delmelle, E. C., Nilsson, I., & Schuch, J. C. (2021). Who's moving in? A longitudinal analysis of home purchase loan borrowers in new transit neighborhoods. Geographical Analysis, 53(2), 237–258. https://doi.org/10.1111/gean.12234
  6. Nilsson, I., Schuch, J. C., Delmelle, E. C., & Canales, K. L. (2020). Should I stay or should I go? A survey analysis of neighborhood change and residential mobility concerns around new light rail stations in Charlotte, NC. Journal of Transport Geography, 86, 102790. https://doi.org/10.1016/j.jtrangeo.2020.102790
  7. Allen, J., Higgins, C. D., Silver, D., & Farber, S. (2023). Are low-income residents disproportionately moving away from transit? Journal of Transport Geography, 110, 103635. https://doi.org/10.1016/j.jtrangeo.2023.103635
  8. Schmidt, A., & Bardaka, E. (2024). The role of transit investments and initial neighborhood attributes on gentrification outcomes. Transportation Research Part D: Transport and Environment, 137, 104463. https://doi.org/10.1016/j.trd.2024.104463
  9. Schmidt, A., & Bardaka, E. (2026). Transit-induced household migration: A survival analysis of heterogeneous income effects near urban rail corridors. Transportation Research Part A: Policy and Practice, 212, 105140. https://doi.org/10.1016/j.tra.2026.105140
  10. Meriläinen, M., Karhula, A., Kurvinen, A., Falkenbach, H., & Ala-Mantila, S. (2024). Transit-induced socioeconomic ascent and new metro stations in Helsinki Metropolitan Area: Distinct effects on renters, homeowners, and pre-existing housing dwellers. Journal of Transport Geography, 114, 103758. https://doi.org/10.1016/j.jtrangeo.2023.103758
  11. Lutz, E., Wicki, M., & Kaufmann, D. (2024). Creating inequality in access to public transit? Densification, gentrification, and displacement. Environment and Planning B: Urban Analytics and City Science, 51(9), 2196–2212. https://doi.org/10.1177/23998083241242883
  12. Kandt, J., van Dijk, J., & Todd, J. (2025). How do major public transport investments induce gentrification? Evidence from consumer data. Transportation Research Part D: Transport and Environment, 142, 104684. https://doi.org/10.1016/j.trd.2025.104684
  13. Deng, K. K., Li, L., & Chen, J. (2025). Impact of public investment on residential gentrification: Individual-level evidence from Shanghai's rail transit expansion. Urban Studies, 62(11), 2328–2347. https://doi.org/10.1177/00420980241310366
  14. Khaleel, S., & Hanlon, B. (2026). Gentrification-induced displacement, destination neighborhoods, and destination houses for low-income households in Columbus, Ohio. Journal of the American Planning Association, 92(2), 163–178. https://doi.org/10.1080/01944363.2025.2593359
  15. Song, T., & Chapple, K. (2024). Does gentrification constrain housing markets for low-income households? Evidence from household residential mobility in the New York and San Francisco metropolitan areas. Urban Studies, 61(13), 2603–2622. https://doi.org/10.1177/00420980241244699
  16. Deng, K. K., Li, L., & Zong, Q. (2025). Investigating transit-induced displacement with housing renter relocation data from urban China. Transportation Research Part D: Transport and Environment, 141, 104674. https://doi.org/10.1016/j.trd.2025.104674