What Happens to Neighborhoods When States Override Local Zoning Laws?

Blanco and Sportiche (2025) ask how residents are affected by and respond to denser housing built under Massachusetts Chapter 40B, which allows developers to override restrictive local zoning. They analyze geocoded 40B developments linked to property sales, residential address histories, and voting records. They find that the average development does not significantly affect nearby property values; the 95% confidence interval ranges from a 0.4% decline to a 2.4% increase. Larger developments reduce nearby values by about 9%, but induce only 1.6% of long-term incumbent residents within 0.2 miles to move. They find no meaningful increase in residents’ political participation.

Why This Article Was Selected for The Policy Scientist

This article addresses a timely policy question as states increasingly reconsider the balance between local zoning authority, housing supply, affordability, and access to high-opportunity communities. Its importance extends beyond Massachusetts because restrictions on residential construction shape regional housing markets, household mobility, and patterns of economic opportunity throughout the United States. Blanco and Sportiche have already contributed substantially to research on housing policy and Chapter 40B, and this study advances that work by examining what happens to incumbent residents after development occurs. The linked property, migration, and voting data are unusually strong, and the stacked difference-in-differences design provides a credible causal framework. Generalizability remains somewhat constrained by Massachusetts’ distinctive statewide policy structure, but the findings should inform evaluation of comparable zoning reforms elsewhere. The article is forthcoming in American Economic Journal: Applied Economics, one of the leading applied economics journals published by the American Economic Association.

Full Citation and Link to Article

Blanco, H., & Sportiche, N. (2025). There goes the neighborhood? The local impacts of state policies that override municipal zoning. American Economic Journal: Applied Economics. Advance online publication. https://www.aeaweb.org/articles?id=10.1257/app.20250044

Central Research Question

The study examines what happens when state policy overrides restrictive municipal zoning and permits substantially denser housing in affluent, low-density communities. The authors focus on Massachusetts Chapter 40B, which allows developers to bypass some local zoning restrictions in municipalities that fall below specified affordable-housing thresholds. The central question is not simply whether the policy produces additional housing, but whether nearby incumbent residents experience measurable economic consequences and subsequently alter their behavior. The authors therefore investigate three principal outcomes: changes in nearby property values, residential mobility among existing residents, and political responses to the new development. The broader concern is whether adverse reactions by incumbent residents could undermine the objectives of statewide efforts to expand housing opportunities in municipalities where local land-use rules otherwise constrain construction.

Previous Literature

The study builds on several established strands of housing and urban economics research. One body of literature finds that restrictive land-use regulation reduces housing supply and contributes to higher prices. Important examples include Glaeser and Gyourko (2002), Glaeser and Ward (2009), Kok, Monkkonen, and Quigley (2014), and Gyourko and Krimmel (2021). Related research links restrictive zoning with residential segregation, including Trounstine (2020) and Resseger (2022), while Hsieh and Moretti (2019) connect housing constraints to slower aggregate economic growth.

A second literature examines how new construction affects nearby housing markets. Li (2021), Pennington (2021), and Asquith, Mast, and Reed (2023) generally find that additional market-rate construction can reduce nearby housing prices through increased supply. Schwartz et al. (2006), however, show that subsidized development can increase surrounding property values when it replaces existing neighborhood disamenities. Particularly relevant is Diamond and McQuade (2019), who find that Low-Income Housing Tax Credit developments reduce property values in comparatively high-income neighborhoods.

A third strand concerns residential sorting and political responses. Tiebout (1956) conceptualized household relocation as “voting with one’s feet,” while Oates (1972), Epple and Romer (1991), Banzhaf and Walsh (2008), and Bayer and McMillan (2012) developed related theories and evidence concerning residential sorting. The authors also draw on studies showing that high-income households often prefer neighborhoods containing similarly situated households, including Bayer, Ferreira, and McMillan (2007), Boustan (2013), Guerrieri, Hartley, and Hurst (2013), and Diamond (2016). The present study extends these literatures by jointly examining property values, migration, and political behavior following a statewide zoning override.

Data

The analysis relies on an unusually detailed combination of geocoded housing, property, individual, and electoral data. The researchers begin with validated geographic information for Chapter 40B developments and focus primarily on 899 developments that entered the permitting process between 1995 and 2019. These data contain exact locations, development size, subsidized-unit counts, rental or ownership status, and permit dates.

Housing prices come from Zillow’s Transaction and Assessment Database, which contains the universe of residential property sales in Massachusetts with consistent coverage beginning in 1987. The dataset includes sale prices, dates, addresses, mortgage information, and property characteristics such as lot size, bedrooms, bathrooms, structure type, and year built. The analysis is restricted to arms-length transactions involving single-family homes and condominiums.

Residential mobility is measured using Infutor address-history data covering most U.S. adults since the 1980s. These records allow the researchers to determine where individuals lived and when they moved. Voting behavior is measured using Massachusetts voter files containing participation histories for federal, state, and local elections. Census data provide additional neighborhood characteristics. Taken together, the linked datasets permit the authors to follow housing markets and incumbent residents for as long as twelve years after development approval.

Methods

The principal identification strategy is a stacked difference-in-differences design organized around geographic “rings” surrounding each Chapter 40B development. Homes and residents located very near a new development serve as the treatment group, while homes and residents slightly farther away provide the comparison group. The basic logic is that properties separated by only a few tenths of a mile should share most underlying neighborhood characteristics but differ considerably in their exposure to the new development.

The empirical models compare outcomes across rings before and after permit approval using event-study specifications. Development-by-calendar-year fixed effects absorb common temporal changes around each project, while development-by-ring fixed effects account for persistent differences across distances. The authors also control for observable property characteristics and cluster standard errors at the development level. Corrective sample weights are used so that the stacked difference-in-differences estimates can be interpreted as average treatment effects across developments.

Several robustness tests strengthen the causal interpretation. The authors use alternative comparison-ring distances, restrict the analysis to developments with complete permit information, remove observations duplicated across overlapping development areas, and employ an alternative design comparing currently treated locations with locations surrounding developments approved later. A placebo test that artificially shifts the treatment date nine years earlier produces no corresponding post-treatment price effects. The alternative identification strategy generates broadly similar results, including long-run property-price declines of approximately 13 percent near large developments.

Findings/Size Effects

The average Chapter 40B development substantially increases local housing density without producing a statistically significant decline in nearby property values. Population density rises approximately 50 percent within 0.1 miles of a development and 16 percent within 0.2 miles. For properties within 0.1 miles, the estimates are sufficiently precise to reject price reductions greater than 0.4 percent and increases greater than 2.4 percent relative to homes located 0.5 to 0.6 miles away.

Development size, however, produces substantial heterogeneity. Developments containing fewer than 50 units increase nearby property values by approximately 5 percent, whereas developments containing at least 50 units reduce nearby values by roughly 9 percent. These negative effects emerge about eight years after permit approval and are concentrated within 0.1 miles of the development. The estimated losses imposed on surrounding properties are relatively small compared with the housing value created: larger developments generated approximately $9.31 billion in housing value between 1995 and 2019, while estimated losses among nearby single-family properties totaled about $290 million, or approximately 3 percent of the value created.

Residents living within 0.1 miles of large developments become approximately 83 percent more likely to move, but the large percentage change reflects a low baseline probability and therefore translates into a small number of actual departures. The average large development causes only about 2.28 of 140 long-term incumbent residents living within 0.2 miles to leave during the first twelve years following approval. This represents approximately 1.6 percent of incumbent residents. Effects decline rapidly with distance and disappear beyond 0.2 miles. Moreover, the migration response is concentrated among residents not identified as homeowners; incumbent homeowners show essentially no measurable response.

Political responses are similarly limited. The researchers find no statistically significant increase in participation in general elections and no compelling evidence of increased participation in local elections. Residents who had already experienced Chapter 40B development were also no more likely to support repealing the law in Massachusetts’ 2010 referendum.

Conclusion

The evidence indicates that statewide zoning overrides can produce substantial increases in housing density without generating large average economic or behavioral reactions among incumbent residents. Although large developments cause meaningful property-value declines for immediately adjacent properties, those effects are highly localized. Even where price declines occur, residential departures are modest in absolute terms and political mobilization is essentially absent.

The findings therefore distinguish between strong opposition to proposed development and residents’ behavior after development actually occurs. Existing residents may object intensely during the permitting process, yet completed projects do not necessarily generate widespread property losses, population turnover, or subsequent electoral opposition. The study also demonstrates that development scale matters: average effects conceal a sharp contrast between smaller projects, which increase nearby property values, and larger projects, which reduce them. More broadly, the results suggest that evaluating zoning reform requires attention not only to housing production but also to the magnitude, geographic concentration, and durability of effects on existing residents.