Abstract
This paper presents causal evidence of a significant positive effect of rent control status on eviction filing rates in San Francisco, CA. Two publicly available data sets, of eviction notices (n = 21,806) and property tax records (n = 1,978,687), are combined using a regression discontinuity design to estimate a local average treatment effect of ∼1.3% of evictions per residential unit per year conditioned on positive rent control status. Compared to the baseline rate of eviction notices over this same time period, the findings suggest that for a given tenant, positive rent control status (i.e., living in a rent-controlled unit) increases the likelihood of eviction by approximately 240% per year. This finding is best understood not as an inherent characteristic of rent control policy in general, but rather as the result of specific state-wide laws, passed in the years following the adoption of rent control in San Francisco, which granted rent-controlled property owners an economic incentive to evict and the legal means to do so.
Keywords:
Acknowledgments
I thank Paul Waddell, Carolina Reid, and Timothy Thomas for their guidance and encouragement, as well as Professor Erin McElroy and Dan Sakaguchi at AEMP for their help in assembling the eviction filing data.
Disclosure Statement
No potential conflict of interest was reported by the author.
Notes
1 San Francisco Administrative Code Chapter 37.
2 Owner-occupied buildings of four units or fewer were excluded until 1995. As this change predates any of the eviction data used in this study, this fact is largely inconsequential to the results of the study.
3 Notable exceptions include Early and Olsen (Citation1998); Glaeser and Luttmer (Citation2003); Gyourko and Linneman (Citation1990); Heskin et al. (Citation2000); Krol and Svorny (Citation2005); Moon and Stotsky (Citation1993); Murray et al. (Citation1991); Nagy (Citation1995, Citation1997); and Sims (Citation2007).
4 Early and Olsen (Citation1998), Krol and Svorny (Citation2005), and Ambrosius et al. (Citation2015) could be considered exceptions, but these rely on census data rather than disaggregated tenant observations.
5 In this case, the eviction records were graciously provided to the author, unaltered, courtesy of the Anti-Eviction Mapping Project (Maharawal & McElroy, Citation2018).
6 In keeping with Asquith (Citation2019), this paper will use both “eviction” and “eviction notice” in reference to the Rent Board data.
7 Properties with a tenancy-in-common (TIC) use code are excluded from the analysis.
8 The eviction data used in this study include no tenant-specific data, so building-level aggregation does not result in the loss of any information.
9 Because the assessor records only provide the property built year, and not month or day, I consider all properties built in 1979 to be rent controlled. By potentially including uncontrolled properties in the sample of controlled properties, it is possible that the estimated treatment effect is conservative in magnitude. However, repeated tests of the models with and without properties built in 1979 did not significantly alter the results.
10 Here I use “no-fault” to describe any of the following nine eviction types: owner move-in (OMI), capital improvement, Ellis Act, condo conversion, substantial rehabilitation, lead remediation, Good Samaritan tenancy ends, development agreement, and demolition. All other evictions, except for those where the eviction type was not indicated, are considered “at-fault” or “breach of lease” evictions. See the Appendix for a full enumeration of eviction types, categories, and observations.
11 A 2002 survey found that 32% of San Francisco renters were unaware of the rent control status of their units, whereas many others were misinformed: https://sfrb.org/sites/default/files/FileCenter/Documents/1885-tenantreportfinal.pdf
12 Nationwide data suggest that the proportion of filings that result in an actual eviction could be as low as one third (Leung et al., Citation2021).
13 See California Civil Code 1954.50-1954.535.
Additional information
Notes on contributors
Max Gardner
Max Gardner is a PhD candidate in civil systems engineering in the Department of Civil and Environmental Engineering at UC Berkeley. His research interests include housing affordability, behavioral modelling of location choice processes in the context of intraurban migration, and microsimulation for regional transportation and land-use forecasting.