The Race to Exclude: Residential Growth Controls in California Cities, 1970–1992

Abstract

Local regulations that restrict residential growth are a key driver of California’s affordable housing crisis. Scholars have argued these growth controls were implemented in the late 20th century by cities intending to exclude Black households. However, growth controls may also have plausibly been driven by a desire to exclude growing Hispanic, Asian, and foreign-born populations; by increased concern about the negative environmental consequences of population growth; or by homeowners’ or cities’ fiscal motivations. I jointly test these competing explanations using time-varying data on the adoption of a variety of residential growth controls covering California cities from 1970 to 1992. I find that, all else equal, cities with a lower share of Black residents—both in absolute terms, and relative to their metropolitan area—were more likely to pass residential growth controls. I also find some evidence that growth controls were more likely to be passed in areas experiencing greater Black population growth and in cities more supportive of White-Black segregation. Finally, I find strong evidence that, net of other factors, cities in areas more supportive of policies to protect the environment were more likely to pass residential growth controls.

Keywords:

• Land use/zoning
• growth management
• regulation
• demographics
• urban environment

Acknowledgements

Thanks to Kip Jackson for providing the land use regulation data used in this paper. Additional thanks to H. Jacob Carlson, Prentiss Dantzler, Frank Donnelly, Neil Fligstein, Andrew Foster, Christopher B. Goodman, David Harding, Margot Jackson, Dov Kadin, Jacob Krimmel, David Lindstrom, John Logan, Kevin Mwenda, Carolina Reid, Daniel Schneider, Caleb Scoville, Mary Shi, Susan Short, David Weil, Michael White, and anonymous reviewers for helpful feedback at different stages of this project. Previous versions of this paper were presented at the 2022 ASA and PAA Annual Meetings, and to audiences at the University of Toronto, Brown University, and the University of Michigan.

Disclosure Statement

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

Notes

1 In California, municipalities may carry the title of “city” or “town,” with no legal distinction between the two. In this paper, I use “city” to refer to all Californian municipalities.

2 Notably, more contemporary research still finds that White households are averse to Black neighbors (Cutler et al., 1999; Emerson et al., 2001; Mummolo & Nall, 2017).

3 Other studies of White racial attitudes in the late 20th century contrast with this picture, suggesting that Whites held more hostile attitudes towards Blacks than Hispanic or Asians, and more hostile attitudes towards Hispanics than Asians (& Oldendick, 1996).

4 It is important to note that homeowners’ desire to protect their property value is not necessarily discordant with homeowners’ desire to exclude racial minorities; indeed, given the prevalence of White households who did not want to live near racial minorities, declines in racial segregation would lead to reduced demand for homes, and thus reduced property values.

5 These regulations include: adequate public facilities ordinances for commercial/industrial development, rezoning commercial/industrial land to less intense use, reducing permitted commercial/industrial building height, limiting the square feet of commercial development, and limiting the square feet of industrial development.

6 Online Appendix Table 1 lists the 29 propositions selected, alongside brief summaries of their text.

7 Online Appendix Table 3 contains full model results with coefficients for population pressures, spatial diffusion controls, and a cubic function of year included. Of note, the first two models contain fewer observations than the second two models. This is because there are metropolitan areas where no city passes an explicit population limit or lower-density regulation, and thus the metropolitan fixed-effect perfectly predicts the dependent variable. Stata drops these observations when estimating the model.

8 In Online Appendix A, I also find generally similar results from negative binomial regression models where the dependent variable is the number of residential growth controls in a given category that a city adopts by 1992, and the independent variables are measured in 1970 or in the decennial Census after incorporation, whichever is later.

Additional information

Funding

Funding for this research was provided by the National Science Foundation Graduate Research Fellowship Program [1752814], the National Institute of Child Health and Human Development [T32-HD007275], the UC Berkeley Opportunity Lab, the Horowitz Foundation for Social Policy, the UC Berkeley Institute of Governmental Studies, and the Brown University Population Studies and Training Center (PSTC), which receives funding from the National Institutes of Health [P2C HD041020]. I acknowledge support from the James M. and Cathleen D. Stone Center for Inequality Dynamics at the University of Michigan.

Notes on contributors

Joe LaBriola

Joe LaBriola is a Research Assistant Professor in the Survey Research Center at the University of Michigan's Institute for Social Research. His research uses survey and administrative data to examine racial and socioeconomic inequalities in the contemporary United States, with a particular focus on racial inequalities in housing and wealth.