Who Banishes? City Power and Anti-homeless Policy in San Francisco

Abstract

Despite cross-disciplinary attention to laws targeting homeless behavior in cities, systematic analysis of the power dynamics behind the adoption and implementation of such laws is surprisingly scarce. This article addresses that oversight by investigating the politics of anti-homeless policies in San Francisco, a critical and revealing case. Using a mixed-methods approach that joins qualitative analysis of public records with spatial and statistical analysis of precinct-level election results, census data, and geocoded police and 311 records, it evaluates previously unmeasured claims concerning the relative influence of social and economic forces in determining policy adoption and assesses whether enforcement patterns betray preferential treatment. Findings suggest that in the face of mobilized opposition, an anti-homeless regime composed of business and neighborhood merchants, elected officials, conservatives, and homeowners each contribute resources required to pass anti-homeless laws. Contrary to past claims, enforcement practices do not appear to privilege only the downtown business district.

Keywords

city power homelessness local politics public space mixed-methods

Since Dahl’s (1961) landmark Who Governs?, enduring debate over the sources of power in city politics has yielded little consensus. Those in the pluralist camp, following Dahl’s lead, maintain that power is dispersed and that no one group controls the outcomes of all political debates. Others, who claim that economic competition constrains the alternatives available to elected officials, emphasize the privileged role business plays in shaping city policy. Still others bridge the divide and argue that city politics is guided by governing coalitions composed of economic elites and government leaders forced to collaborate with the voting public.

If, as Jessica Trounstine (2018) has recently claimed, “Battles over the control of urban space have always been the primary driver of city politics,” then debates over exclusionary homeless policies offer an especially potent and revealing realm for investigating city power dynamics (p. 3). Anti-homeless policies—including laws prohibiting sleeping in public, sitting on sidewalks, or panhandling—incite battles between use values and exchange values of urban space (Logan and Molotch 1987). Indeed, existing explanations of anti-homeless law adoption largely reflect theories of city power. Numerous accounts claim that cities adopt such policies in response to the demands of residents, pressure from local business and industry, or a combination of both. With urban homelessness a pressing concern in cities across the United States (Thrush 2018), it is important to understand the motivations behind the particular policies cities adopt to address homelessness. As of yet, however, empirical assessments of the relative influence and resources contributed by political actors to bring about the adoption of anti-homeless laws have been surprisingly scarce. Despite extensive and cross-disciplinary analysis of these laws,

there are in fact remarkably few detailed studies of precisely how such policies come in to being, of the make-up of the coalitions (of private businesses, the local state and police) that lie behind them, or of the practices through which they are enacted on the ground. (May and Cloke 2013, p. 896)

The research reported here ventures into this neglected terrain. It seeks to analyze the political forces behind exclusionary laws intended to banish (Beckett and Herbert 2010) unhoused populations from select public spaces. In doing so, it also seeks to contribute a more nuanced account of city power than recent ambitious efforts that, following Peterson (1981), focus on city budgets and spending patterns for evidence of influence over policy decisions (e.g., de Benedictis-Kessner and Warshaw 2016; Einstein and Kogan 2016; Hajnal and Trounstine 2010). The budget is not the only policy through which power is expressed. Furthermore, when attention is directed toward the implementation of city policy, “the conceptual neatness of Peterson’s developmental, redistributive, and allocational typology of policies breaks down empirically” (Stone 2017, p. 14). This is certainly the case with anti-homeless policies. Even when such laws are formally drafted as allocational policies to be implemented by police or public works departments to serve the needs of the city as a whole, in practice, they may be enforced to appease the interests of a narrow sector of urban society (Goetz 2009).

In what follows, a mixed-methods research strategy is employed to assess the relative influence of political actors in promoting the adoption and shaping enforcement patterns of two anti-homeless policies in San Francisco, a city in which homeless concerns have long shaped local politics. Analysis focuses on the 2010 “sit-lie” ordinance banning sitting or lying on sidewalks and the 2016 law prohibiting tents on sidewalks. The research involves three complimentary empirical steps: (1) qualitative analysis of public records to establish the cast of political actors engaged in the debates and identify their policy preferences, (2) geographic information system (GIS) and spatial regression analysis of election results to estimate neighborhood characteristics associated with electoral support for anti-homeless ballot measures, and (3) spatial analysis of police records and 311 service requests to investigate enforcement patterns for signs of more subtle forms of political power.

Findings indicate that while big business and select neighborhood merchants played a major role in getting anti-homeless proposals on the agenda and shaping the public debate, pressure from these economic actors proved insufficient to ensure passage of the proposed legislation. In each case, city officials proved unwilling or unable to adopt the anti-homeless proposal. Instead, officials supportive of the measures strategically put the decision directly to city voters. Ultimately, the ballot proved to be the key resource required to adopt the proposals. Neighborhoods with higher proportions of conservatives and homeowners played critical roles in supplying the votes needed to pass each measure. Once adopted, anti-homeless policy enforcement patterns contradict previous claims that such laws are implemented chiefly to benefit downtown businesses. Instead, enforcement of the sit-lie ordinance has been especially strict in low-density residential areas and near the neighborhood merchants who played the biggest role in bringing the issue to the attention of local government, while tent ban–related enforcement reveals little evidence of bias and appears to be mostly shaped by bureaucratic capacity. Taken together, the findings provide evidence of an anti-homeless regime in San Francisco, a loose-knit coalition composed of big business and neighborhood merchants, strategic public officials, conservatives and homeowners, collectively capable of crafting punitive homeless policy even in the paradigmatic progressive city.

Theory and Expectations

The debate over city power is waged primarily between two camps: the pluralists who highlight the power of organized groups and elections, versus those who claim that economic imperatives constrain the policy options available to cities (Hajnal and Trounstine 2010; Stone 2017). Both traditions, as well as urban regime theory (Stone 1989), which draws together elements from each, provide valuable expectations concerning the political forces behind the adoption of anti-homeless policy in cities.

The pluralist tradition—of which Dahl’s (1961) Who Governs? is emblematic—understands political resources as dispersed through society, allowing no single group to monopolize decision-making across all policy areas. This perspective assumes that citizens will engage in politics when issues directly impact their lives, but otherwise, will exert indirect influence over officials who remain wary of the latent power of the citizenry to hold them accountable in the next election (Dahl 1961). Although some policies may be especially shaped by pressure from business, mobilized neighborhood groups and city voters also influence city policy making. Much research supporting the pluralist perspective involves comparison between cities. Donovan and Neiman (1992) find that politically active populations—those with high turnout rates and legacies of pushing for ballot measures—are most successful in efforts to constrain local growth policies. Spending preferences of local leaders are found to reflect those of local populations (Connolly and Mason 2016), while local policy choices have been linked to local political ideology (Choi et al. 2010), public opinion (Hajnal and Trounstine 2010), and political culture (Goetz 1994). Other work supporting pluralistic expectations assesses variation in political activity and influence within cities. Urban social geography—the distribution of social groups across a city space—impacts the priorities and risks that different groups face, how they engage in politics, and what they demand from local government when they do (Jun and Musso 2013). City policies and projects impact neighborhoods unevenly, which differentiates political activity across city space (Kinsey et al. 2010). A sense of investment in place may also shape residents’ levels and forms of political activity (Kearns 1995). Homeowners—who possess both sentimental attachments to place and economic incentives to protect property values—are frequently seen as one of the most potent forces in local politics (Einstein, Palmer, and Glick 2019; Fischel 2001; Verba, Schlozman, and Brady 1995).

Peterson’s (1981) City Limits frequently serves as the figurehead for those who claim that the imperatives of economic development are what really shape city policy making. From this perspective, local politics is overshadowed by competition between cities seeking to attract both investment and a footloose middle-class more likely to “vote with their feet” by seeking out cities that adopt policies they prefer, especially regarding taxing and spending practices (Minkoff 2009; Tiebout 1956). Elected officials are seen as constrained to an agenda that can reliably promote local economic development. Elections, group activism, and ideology matter little compared to the economic pressures that cities face (Dye 1979). City politics is viewed as “groupless,” and low levels of political participation are understood as a rational response to the limited impact such involvement can achieve (Peterson 1981, p. 128). The only groups that really matter for city politics are “growth machine” coalitions including representatives from business, real estate, and government who all share an interest in local development and land values (Logan and Molotch 1987). Cities also behave much like businesses themselves. “Entrepreneurial cities” (Harvey 1989) invest in urban place-making by developing “new urban glamour zones” (MacLeod 2002, p. 605) intended to attract tourists and wealthy residents to commercial centers, while concurrently adopting increasingly stringent surveillance practices for regulating urban space, all in the aim of supporting the local economy (Coleman, Tombs, and Whyte 2005; MacLeod 2002; Mitchell 1997).

According to the regime theory of city politics, power depends upon a collaborative governing coalition incorporating elected officials, business interests, and portions of the voting public, each contributing unique and necessary resources to the bargain (Stone 1989). While officials possess the authority to enact policy, they depend on both business investment and electoral support to maintain power and achieve collective goals. Regime theory “rejects both pluralist assumptions that governmental authority is adequate to make and carry out policies, as well as structuralist assumptions that economic forces determine policy” (Mossberger and Stoker 2001, p. 812). Instead, governing decisions depend on cooperation and compromise among coalition members who, as members of the regime, achieve far more together than they would have through solitary efforts (Stone 1989). DeLeon’s (1992) study of San Francisco reconceives regime theory and documents an “antiregime”: a progressive coalition capable blocking business demands, stalling development, and to striving to “protect community from capital” (p. 98).

Within the literature on the criminalization of homelessness, it is common to find claims about the political impact of organized groups or economic pressures, but far less common to find evidence regarding the distinct political impact of the various forces. In fact, within several seminal works on the topic one may find claims that would seem to support any of the competing theories of city power. For example, Beckett and Herbert (2010) emphasize the role played by mobilized anti-crime activists, neighborhood groups, middle-class residents, and city voters in pressuring a divided city council to adopt new exclusionary laws, much as the pluralists would predict. Yet elsewhere, the authors describe these laws “as a kind of public subsidy on behalf of downtown commercial interests” (p. 30) adopted “in significant part, as a response to concerns from the business community about the effects of disorderly people on consumption patterns” (p. 64), as the economic imperative perspective would predict. Similarly, Mitchell (1997, p. 308) notes that “vocal opposition” to proposed anti-homeless legislation can force a city council to turn the decision over to city voters, though he ultimately casts anti-homeless laws as strategic efforts by local officials to attract and appease local investment (p. 313). Again, both social and economic pressures seem to be at play, but their relative influence is not distinguished. Perhaps most relevantly, Gowan’s (2010) thorough analysis of homelessness in San Francisco highlights both the pluralistic political pressure from merchants and city residents alongside public–private partnerships through which city government and real estate developers jointly pursue homeless-displacement strategies neither partner is capable of pursuing independently, a finding more in line with a regime conception of city power. The point is not to criticize these important works, but rather to show that understanding the relative influence of social and economic pressures on anti-homeless policies demands further empirical scrutiny.

Some research on the politics of homelessness leads us to expect that mobilized groups of residents should impact the anti-homeless policy process. Sylvestre (2010b) finds that the residents most threatened by the homeless are most likely to get involved. Perhaps most prominent are cases of NIMBY-oriented groups threatened by proposals to site homeless service-providing facilities in their neighborhoods (Oakley 2002; Takahashi 1997). Politically conservative residents are especially likely to perceive the homeless as threatening (Farrell 2005; Lee, Farrell, and Link 2004), and so may also be more supportive of anti-homeless laws. On the other hand, more politically progressive groups and homeless advocacy organizations often mobilize against proposed anti-homeless laws (Beckett and Herbert 2010; Oakley 2002). Findings concerning public opinion among racial or ethnic groups toward homelessness are mixed and uncertain (Beckett and Herbert 2010; Lee, Farrell, and Link 2004). Evidence that anti-homeless policy making is shaped primarily by the actions of mobilized groups will buoy a pluralist view of homeless politics in San Francisco.

Other work attributes the “punitive turn” in city homeless policy to the pressures faced by city officials competing to “attract and retain highly mobile capital and the footloose middle-class” (DeVerteuil 2006, p. 110). In this work, anti-homeless laws are adopted primarily in response to concerns that visible signs of homelessness will deter shoppers and tourists (Goetz 2009; Mitchell 2011). Steffen (2012) points to business elites as the chief architects of anti-homeless proposals in Atlanta.¹ Local business also seeks to influence how exclusionary laws are implemented, leading to spatially uneven patterns of enforcement that privilege business and commercial zones while neglecting other parts of the city (Herring 2014; Goetz 2009). If homeless politics in San Francisco aligns neatly with expectations drawn from the economic imperative perspective, findings will indicate that unified government and business leaders craft exclusionary policies with little concern for or involvement by city residents. Enforcement patterns prioritizing downtown tourist and shopping destinations will further support this perspective.

Regime theory encourages us to look for the unique resources contributed by government, business, and residents in the process through which anti-homeless policies are adopted. While business and certain public officials may be primarily motivated by citywide economic concerns, residents and other officials may be brought into the coalition when campaigns activate concerns over public safety, property values, or perceived disorder. Leaders may strategically frame homeless concerns in an effort to galvanize public support for punitive laws (Amster 2003). As a result, the “exclusionary tactics of urban redevelopment are thus intertwined and legitimated by a growing cultural intolerance of our society’s most truly disadvantaged” (Stuart 2015, p. 944). A regime view of homeless politics in San Francisco will be supported by findings indicating that the adoption of anti-homeless policies relies upon the interrelated participation of city residents, business, and government.

Data and Method

Homelessness has been a salient political issue in San Francisco for decades and a top concern for both residents and recent mayors who, once in office, adopt policy approaches that have swung pendulum-like from prioritizing social service provision to adopting aggressive “military campaign” (MacDonald 1995) styles of homeless policy (Gowan 2010; Green 2019). It is also a city long characterized by a large and politically active progressive population (DeLeon 1992)—a group likely to oppose proposed punitive exclusionary policies—as well as an active business community encouraging business-friendly local governance.

In studying the adoption of homeless policy in San Francisco, I follow Murphy (2009) in considering San Francisco “a revelatory and noteworthy case” (p. 308). In it exist all potential political forces that dominant theories of city power point to as influential in determining policy outputs. For Gowan (2010, p. 239), San Francisco represents “a particularly important case of the criminalization of homelessness” specifically because such punitive policies have been adopted despite opposition from more progressive political forces. “If it could happen there,” she tells us, “it could happen anywhere.” In addition, as a result of the city’s high-density and constrained geography, areas of concentrated poverty and homelessness brush up against San Francisco’s high-end retail areas and tourist attractions. Unlike more sprawling or low-density cities, in San Francisco, homelessness can be neither hidden nor ignored.

The policies I focus on fit neatly within what Mitchell (1998a, 1998b, 2011) categorizes as “anti-homeless” laws. These laws criminalize behaviors or activities that individuals experiencing homelessness, because they lack private property or reliable access to shelter, are forced to do in public. Common examples of these laws include bans on vagrancy, trespassing, loitering, sleeping in public, camping in public space, and panhandling. Since enforcement depends on the subjective decisions of police officers, these laws “essentially criminalize status for some urban residents; individuals may be targeted for how they look and what they symbolize rather than specific behaviors” (Beckett and Herbert 2010, p. 15). Anti-homeless laws “banish” the visibly poor from prime public urban spaces (Beckett and Herbert 2010; Duncan 1978).

The adoption and enforcement of two anti-homeless laws are analyzed below. The first, a so-called “sit-lie” ordinance prohibiting sitting or lying on city sidewalks, was adopted in 2010. The second, a law banning tents on city sidewalks, was adopted in 2016. These policies were chosen because they are the two most recent anti-homeless laws adopted by the city and because they apply to all urban spaces in San Francisco, whereas bans on panhandling, sleeping in parks, or loitering near ATMs can only be enforced in select parts of the city.²

A mixed-methods research strategy designed to assess the relationship between city power and anti-homeless policy proceeds in three analytical steps. The first step aims to set the stage for the political contest by identifying the political actors involved in the public debate and to establish their policy preferences. To do so, I analyze three forms of public records: (1) video of public comment made during meetings of the Public Safety Committee of the San Francisco Board of Supervisors (BOS), (2) arguments for and against the propositions printed in Voter Information Pamphlets published by the city’s Department of Elections, and (3) data on campaign contributions.³ Public comments before the San Francisco BOS were coded and categorized according to three key characteristics: whether the commenter supported or opposed the proposed measure, how the commenter identified themselves (i.e., as a resident, homeless advocate/activist, or business owner/representative), and the commenter’s issue of primary concern (i.e., public safety/harassment, impact on business, civil liberties, discriminatory enforcement practices, preserving public space, or encouraging the prioritization of service provision over criminalization). Analysis of public comments before the BOS is complemented by qualitative content analysis of statements published in the Voter Information Pamphlets as well as campaign contribution data.

Since each proposed measure was eventually put to voters as a local ballot proposition, the second analytical step involves analyzing election results to determine whether particular social groups or neighborhoods played particularly impactful roles in passing the ballot measures. Existing research has yet to sufficiently investigate (and empirically determine) which urban residents bear chief responsibility for the adoption of anti-homeless policies. Homeowners, motivated by preserving property values, may be more likely than renters to support such laws (Fischel 2001). More conservative voters have been found to be more concerned by the threat of individuals perceived as homeless (Farrell 2005; Lee, Farrell, and Link 2004). Residents who reside in areas with high concentrations of homeless-related concerns may also perceive higher levels of threat and may, in turn, be especially supportive of punitive legislative efforts. This portion of the analysis provides a first empirical examination of these hypothesized relationships between the voting urban public and anti-homeless proposals. Although available data do not allow direct measurement of individual-level voting preferences, precinct-level election results joined with census data and geocoded 311 “homeless concerns” service requests allow us to discern characteristics of the neighborhoods that were most likely to support the anti-homeless measures.

For both the 2010 and 2016 elections (the years in which Measure L, the proposed sit-lie ordinance, and Measure Q, the proposed tent ban, were on the ballot, respectively), precinct-level election results were joined to GIS shapefiles—spatial data sets representing the geographic boundaries of each electoral precinct within the city.⁴ Election results provide information on three variables of interest: the percent of voters in each precinct who supported the anti-homeless ballot measure (the dependent variable), the percent of precinct voters registered as Republicans (a proxy for conservative voters),⁵ and precinct-level voter turnout. I also include a dummy variable indicating whether the city Supervisor representing each precinct supported the ballot measure or not. Census data (from the 2010 decennial census and the 2016 American Community Survey) were aggregated from block groups to each electoral precinct to establish the rates of homeownership as well as the racial and ethnic composition of each precinct.⁶ While there is no theoretically motivated reason to expect that racial and ethnic characteristics will influence precinct-level support for anti-homeless measures, it is important to control for neighborhood demographics since social geography can powerfully influence political behavior (Enos 2017) and may bias perceptions of neighborhood blight or disorganization (Sampson 2012).⁷

Finally, I utilize data on 311 service requests to assess whether voters living in neighborhoods with more reported homeless-related concerns were more likely to support the ballot measures. In many big cities, residents can make a variety of service requests either by phone (by dialing 311) or, more likely, via online or mobile portals. In San Francisco, residents can make a number of service requests categorized as “homeless concerns,” though in 2010, far fewer request types were available than are today. For the 2010 election, I gather all “Human Waste” and “Medical Waste” service requests for the two years leading up to the election. Human waste (generally feces, though occasionally other items) and medical waste (generally used syringes) are commonly attributed to individuals experiencing homelessness and were the concerns most often cited by citizens who voiced support for the sit-lie law.⁸ Service requests were geocoded, and the number falling within each precinct, plus those within 100 feet of precinct borders, were summed and serve as the measure of the variable I term concentrated concern.⁹ By 2016, more types of “homeless concerns” service requests were available to San Francisco residents. For this year, I collect the number of “Encampment” requests falling within each precinct plus those within 100 feet of the precinct boundary for the two years prior to the election date.¹⁰ The use of 311 service requests is by no means a perfect measure of any of the relevant concerns. Neighborhood service request counts likely capture not only varying levels of the particular concern, but also variation in resident activism (Minkoff 2016; White and Trump 2018). While it is likely that the number of 311 requests and the issue being reported will generally trend together, human and medical waste or encampments will surely stimulate varying levels of response in different parts of the city.¹¹ So, the “concentrated concern” variable should be understood as capturing not just an objective count of the concern of interest, but also the intensity of local response to that concern in voters’ residential neighborhoods.¹²

Following Anselin (2005), election data are analyzed using spatial regression. Preliminary ordinary least squares (OLS) regression indicated significant levels of spatial autocorrelation, a common symptom of spatial data and one that can bias results and undercut their reliability. Diagnostic tests determined that a spatial error model of maximum likelihood estimation best attenuated such concerns and yielded the most reliable estimates. Spatial error weights are generated for each precinct utilizing second-order Rook contiguity matrices incorporating an average of 15.63 “neighbors” per precinct (inclusive of first-order neighbors). Election results for both 2010 and 2016 are modeled using this technique.

The third and final step of this analysis investigates whether enforcement patterns of anti-homeless policies betray more subtle, place-based signs of urban power. According to Sylvestre (2010a) “by showing how the law actually operates…we obtain a more accurate understanding of state power and structure” (p. 804). Anti-homeless policies, enforced by “street-level bureaucrats” (Lipsky 1980) afforded wide discretion to regulate particular behaviors and uses of public space, are vulnerable to enforcement practices that effectively privilege particular neighborhoods and the residents or businesses within them. Evidence that enforcement prioritizes certain city regions or land uses may reveal a “geography of power” (Mitchell 2011, p. 942) that should inform our understanding of the power dynamics shaping anti-homeless policies.

Assessing bias in enforcement of anti-homeless laws is not a straightforward matter. It is difficult to disentangle variation in enforcement from variation in the behavior being regulated. Some neighborhoods will have higher concentrations of the concerns targeted by the anti-homeless laws (i.e., sitting on sidewalks or tent encampments) and we should expect them to be enforced more frequently in those neighborhoods. So, we can’t simply equate higher frequency of enforcement with preferential treatment. Instead, I focus on variation in quality of enforcement. Assessing the severity of the punishment (for the sit-lie ban) and the time taken for public agencies to complete service requests (for the tent ban) helps control for spatial variation in homeless-related concerns and focus in on variation in enforcement practices.

Enforcement patterns related to each anti-homeless law were analyzed using a GIS shapefile provided by the San Francisco Planning Department which characterizes 435,603 parcels of land within the city by zoning type (e.g., Residential, Downtown Retail, Community Business). For the sit-lie ban, police records of all instances of enforcement in the five years following the law’s adoption (2011–2015) were collected. Each record is characterized as one of three levels of enforcement severity: warning, citation, or booking. Each sit-lie enforcement record was geocoded and joined to any city parcel within 100 feet. This allows me to aggregate the number of sit-lie warnings, citations, and bookings that occurred within 100 feet of each land use type. The assumption motivating this technique is that land uses or neighborhoods with higher rates of citations or bookings relative to warnings (i.e., stricter implementation) receive privileged enforcement of the law compared to land uses and neighborhoods in which warnings (i.e., lax implementation) dominate. Since two neighborhoods—Haight Ashbury and the Tenderloin—were frequently referenced in public comments before the BOS (and both have reputations for concentrated homelessness), I also aggregate the sit-lie records to the police districts representing each neighborhood. Doing so allows me to assess whether public comments and press coverage predicting preferential enforcement in the Haight bore out.

Assessing enforcement practices of the tent ban proves somewhat more difficult, particularly because it was generally agreed that the city already possessed the legal authority to clear encampments prior to the 2016 law’s adoption. Both the mayor at the time of adoption (Ed Lee) and the director of the city’s Department of Homelessness and Supportive Housing viewed the new law as one among “a whole host of tools” for addressing encampments (quoted in Fagan and Green (2017)). So, rather than analyzing only instances in which the 2016 tent ban was enforced, I cast a wider net and assess city response to encampments more broadly. This approach allows me to assess preferential treatment involved in the family of laws that collectively authorize encampment clearance.¹³ Since the city’s Department of Public Works is primarily responsible for encampment complaints, a modified measure of privileged enforcement was required. A total of 26,543 “Encampment” 311 service requests from 2017, the year following the policy’s adoption, were collected and geocoded.¹⁴ The time between service request initiation and completion (time-to-completion) was calculated for each request. Again, all requests were joined to city parcels within 100 feet, allowing me to calculate average time-to-completion for various land use types. The assumption guiding this method is that faster time-to-completion for encampment requests (which equates to faster tent clearance) indicates preferential enforcement.

Results

Actors and Preferences

Initially, Mayor Gavin Newsom—whose early political success is frequently attributed to “his tough-love approach to the homeless problem” (D. E. Murphy 2003)—partnered with second district Supervisor Alioto-Piers to introduce the sit-lie ordinance for consideration by the San Francisco BOS. The city had debated similar laws before. One such law, adopted in the 1970s, was eventually deemed unconstitutional. Another proposal, pushed by police chief-turned-mayor Frank Jordan, was narrowly defeated by voters in 1994 (Gowan 2010). These past experiences informed and inspired high levels of public involvement in the debate over the 2010 proposal. Opening one BOS Public Safety Committee meeting, the committee chairman noted that more individuals had signed up to make public comment than had at any other committee meeting he had attended. A total of 155 public comments were made during two hearings on the proposed sit-lie law (on March 1 and May 10, 2010). Data on commenter identification and the key issues emphasized are summarized in Table 1.¹⁵

Table 1.

Public Comment on Sit-Lie Ban Proposal.

	Proponents	Opponents
Percent of all public comments	26%	60%
Speaker identification
Business/Merchant	34%	1%
Resident	37%	25%
Homeless activist/advocate	—	26%
Issue of concern
Safety/Harassment	51%	—
Business concerns	24%	—
Discriminatory enforcement	—	41%
Homeless services/facilities	—	10%
Use value of public space	—	11%
Total comments: 155

Note. Classifications are not exclusive. Only speakers who could be identified or concerns that could be categorized are included.

About one in four speakers expressed support for the proposed ordinance. Of those, 37% identified themselves as city residents, while 34% identified as merchants or business owners. These sit-lie proponents emphasized two key issues in their arguments: 51% focused on concerns of public safety, harassment, and intimidation, while 24% argued that allowing individuals to block sidewalks was bad for business. One speaker, “Tracy,” recalled her boyfriend being attacked on Haight Street by an individual she often saw sitting on the sidewalk nearby. The vice president of South of Market Business Association claimed “this city doesn’t appreciate that the economy is the ultimate citizen, and that’s what it’s all about, because there’s no money for the youth homeless without taxes.” The Haight Ashbury neighborhood was mentioned more than any other part of the city, and several members of the Haight Ashbury Improvement Association (an organization of neighborhood merchants) spoke in favor of the bill. Interestingly, one speaker voiced support for the law but was concerned that all the attention to the Haight would overshadow similar concerns in the Tenderloin neighborhood, where she lived.

The ordinance was opposed by 60% of individuals making public comment. Of these speakers, 25% identified as residents, and 26% as advocates and activists for the homeless. Several religious leaders and self-identified homeless individuals spoke in opposition to the bill as well. By far, the most frequent concern was that the sit-lie law would lead to discriminatory enforcement by police. One member of the Haight Ashbury Neighborhood Council (an organization of residents) argued that the law would result in “criminalizing appearance rather than conduct,” a concern shared by 41% of opponents. Several speakers mentioned that the original sit-lie ban in the 1970s had been used to target homosexuals and gay culture. The Executive Director of the city’s Coalition on Homelessness claimed that “once again, destitute people are being used as political scapegoats,” and others (including Supervisor Chris Daly) agreed the proposal was a purely political maneuver. Still others expressed concern that the law would infringe on civil liberties, disproportionately affect day laborers, or inhibit the public’s use of public space.

The BOS ultimately voted down the measure by a margin of eight votes to three. In response, the Mayor used his authority to include the proposal as Measure L—termed “Civil Sidewalks”—on the November 2010 ballot, giving city voters a direct say on the matter. Arguments printed in the Voter Information Pamphlet for that election provide further information on which political actors favored or opposed the measure. In addition to the Mayor and three supportive city Supervisors, the Haight Ashbury Improvement Association, the chamber of commerce, and the city’s Republican Party all printed statements supporting of the measure. Other supporters included the Asian Pacific Democratic Party and the Police Officers Association. Again, supporters emphasized concerns about safety and harassment and claimed that the sit-lie law would bolster business. Opponents of the measure included more politically progressive politicians and party organizations, homeless advocates, and the Haight Ashbury Neighborhood Council, who echoed concerns about discriminatory enforcement practices and potential infringements on civil rights.

While neighborhood merchants may have played a central role in public comment sessions, it was the power of big business that dominated contributions to the campaign in favor of the sit-lie ordinance. Donations to the “Coalition for Civil Sidewalks” came predominantly from members of the city’s “growth machine” (Logan and Molotch 1987) as well as from the city’s downtown Union Square Business Improvement District (BID), which contributed revenue gathered through property assessments to support the campaign (Selbin et al. 2018). The chamber of commerce, real estate and investment firms along with their presidents, partners, and supportive committees contributed a majority of the $397,903 raised by the campaign in favor of the law. Opponents of the ballot measure raised just $9,251 for their “Sidewalks are for People” campaign to defeat Measure L.

Unlike the sit-lie law, debate over the so-called “Housing not Tents” proposal in 2016 involved drastically less public comment directed toward city officials. Second district Supervisor Farrell, the author and chief champion of the tent ban, made no attempt to pass his proposal through the BOS. Instead, with the support of three other supervisors, he put the issue directly on the 2016 ballot as Measure Q, “Prohibiting Tents on Public Sidewalks.” The only reason public comment was heard on the matter is that Supervisor Kim—who ultimately opposed the proposal—raised a competing measure in a Public Safety Committee meeting on September 8, 2016. Public comments in this hearing focused predominantly on Ferrell’s proposed tent ban, which by then had already been put on the ballot.

Only 12 public comments were made before the BOS regarding the proposed tent ban. While one member of a neighborhood business association spoke in favor of the proposal, most supporters represented big business and industry associations. A vice president from the chamber of commerce claimed the increased prevalence of tent encampments was due primarily to changes in public policy, while an executive from the San Francisco Travel Association argued that current procedures did not permit authorities to clear encampments quickly enough. Nearly all opponents of the tent ban were homeless advocates who, like the Executive Director of the Coalition on Homelessness, argued that efforts should be directed toward providing housing and services, not on tent clearance.

Arguments published in the 2016 Voter Information Pamphlet and contributions to the campaigns for and against the tent ban tell a story similar to 2010. In addition to the four city Supervisors who sponsored the bill, the chamber of commerce, merchants associations, and the Police Officers Association all expressed support for the tent ban. Supporters emphasized the dangers and health risks of encampments. Opponents, including members of the BOS and other elected officials, homeless advocates, religious leaders, and the Haight Ashbury Neighborhood Council, urged voters to reject the measure, and argued that efforts should focus on finding housing for individuals experiencing homelessness, not criminalizing them. Opponents shared the sentiments expressed in the statement contributed by the Faithful Fools Street Ministry: “This measure is an expression of frustration, not a meaningful response to the growing crisis of homelessness.” “San Franciscans for Housing not Tents,” the political committee funding the campaign for the tent ban, raised $822,047 in contributions primarily from the chamber of commerce and other associations representing real estate and tourism in the city, along with several large contributions from the city’s tech elite (Wong 2016). “People for True Housing and Homeless Solutions 2016,” the committee funding the campaign against measure, raised only $8,777, a mere 1% of funds raised by supporters of the tent ban.

The Votes

Measure L (sit-lie ban) and Measure Q (tent ban) passed with support from 54.3% and 51.8% of voters, respectively. However, support for these measures varied considerably across city space. Figure 1 maps precinct-level support for the sit-lie ballot measure along with spatial variation of key explanatory variables.¹⁶ Regression results for both elections are reported in Table 2. Findings are largely consistent across the years and can be discussed concurrently. For the sit-lie ordinance, model 1 measures concentrated concern in terms of “human waste” 311 service requests, while “medical waste” is the concentrated concern in model 2.

Figure 1.

Precinct-level support for Sit-Lie Ordinance in 2010 and spatial variation of explanatory variables.

Table 2.

Precinct-Level Support for Anti-Homeless Ballot Measures.

	Sit-Lie Ordinance (2010)		Tent Ban (2016)
	(1)	(2)	(3)
% Republican	1.249*** (0.067)	1.246*** (0.066)	0.849*** (0.067)
% Homeownership	0.074*** (0.017)	0.072*** (0.016)	0.064*** (0.012)
Concentrated concern	0.002 (0.002)	0.010** (0.004)	0.003* (0.001)
Turnout	0.142*** (0.038)	0.131*** (0.037)	0.038 (0.032)
Supervisor support	0.000 (0.008)	0.000 (0.008)	0.022*** (0.005)
% Black	–0.083* (0.042)	–0.090* (0.041)	0.051 (0.032)
% Asian	0.038^† (0.021)	0.040^† (0.021)	0.095*** (0.017)
% Latino	–0.156*** (0.034)	–0.162*** (0.034)	0.007 (0.023)
Spatial weight	0.786*** (0.044)	0.774*** (0.046)	0.879*** (0.031)
Constant	0.283*** (0.029)	0.293*** (0.028)	0.296*** (0.032)
N	572	572	579
R ²	.85	.86	.79

Note. Spatial error models of maximum likelihood estimation. Standard errors in parentheses. Summary statistics in the appendix.

^† p < .10. *p < .05. **p < .01. ***p < .001.

In both the 2010 and 2016 elections, a higher percentage of registered Republicans was associated with a statistically significant increase in support for the anti-homeless ballot measures. Substantively, moving from one standard deviation below the mean precinct percent Republican to one stand deviation above it (while holding all other variables at their means) is associated with a 13% increase in support for the sit-lie ballot measure, and a 7% increase in support for the tent ban. Similarly, precincts with higher rates of homeownership tended to provide higher levels of electoral support for the measures. A shift from one standard deviation below to one standard deviation above the mean precinct homeownership level was associated with a 3% increase in support for each anti-homeless policy.

In both cases, higher levels of concentrated concern were associated with increased support for the anti-homeless measures, though for only one such measure in 2010. Voters living in precincts with more homeless-related 311 service requests (concerning “medical waste” in 2010 and “encampments” in 2016) were more likely to support the anti-homeless policies than voters in precincts with fewer such complaints. However, while statistically significant, the substantive association between concentrated concern and support for anti-homeless laws was rather modest. Holding all other variables at their means, the models estimate that moving from one standard deviation below the concentrated concern means to one standard deviation above them is associated with just under a 1% increase in precinct-level support for the anti-homeless proposal each year. Precinct counts of “Human waste” 311 requests were not found to have any statistically significant relationship with support for the sit-lie ban.

Support for the anti-homeless ballot measures was also shaped in part by the racial or ethnic compositions of precinct populations. Precincts with higher proportions of Black and Latino residents tended to be less supportive of the sit-lie proposal. Asian and Asian-American neighborhoods, on the other hand, tended to be somewhat more supportive of anti-homeless ballot measures, though the relationship between percent-Asian and support for the sit-lie measure drops just below standard levels of statistical significance in models 1 (p = .071) and 2 (p = .058). There were 118 Asian and Asian-American-majority precincts in 2010, and 111 in 2016. Average support in these precincts for the ballot measures was 54% in 2010 and 52% in 2016, compared to 49% and 46% in all other precincts, respectively.¹⁷

Results concerning political variables are inconsistent across the years. In 2010, higher rates of turnout were associated with support for the sit-lie ban, though there was no comparable relationship in 2016. Conversely, in 2016, precincts represented by Supervisors supportive of the tent ban contributed higher levels of support for Proposition Q, though Supervisor support was not a statistically significant predictor of precinct support for the sit-lie ban in 2010.

Enforcement

Table 3 reports enforcement patterns of both the sit-lie ordinance and the tent ban for select land use types and the two police districts of interest.¹⁸ For the sit-lie ban, citywide, 38% of all enforcement records were warnings, 49% were citations, and 14% were bookings. However, enforcement patterns vary considerably by land use and neighborhood. Enforcement in the Haight Ashbury Neighborhood Commercial zone—ground zero for the debate over the adoption of the law—was considerably stricter than in most other parts of the city. Fully 72% of enforcement instances in the Haight were either citations or bookings. It is also worth noting that over a third of all sit-lie enforcement records (1,665 of 4,813 total records) occurred within 100 feet of Haight Neighborhood Commercial parcels, leading to 5.2 enforcement instances per parcel, compared to just .01 instances per parcel citywide. Parcels zoned for lower-density residential buildings and houses also had stricter enforcement patterns than the city average. In contrast, in the two downtown zoning types with the highest levels of sit-lie enforcement (Downtown Retail and Downtown Office Special Development), warnings were far more common.¹⁹ The contrast between enforcement in the Tenderloin Police District and the Park District (which includes the Haight) is particularly striking. In the Park district, 73% of all enforcement instances were citations or bookings, and only 27% resulted in warnings. In the Tenderloin, the numbers are nearly flipped: only 34% of enforcement instances resulted in a citation or booking, while 66% were warnings.

Table 3.

Enforcement Patterns by Parcel Zoning Type.

Zoning District	Sit-Lie Enforcement (2011–2015)					Tent Ban Enforcement (2017)
Zoning District	Total Incidents	Incidents per Parcel	% Warning	% Citation	% Booking	Total Requests	Request per Parcel	Mean time- to-completion
Citywide	4,813	0.01	0.38	0.49	0.14	26,543	0.06	68.72
Haight Street Neighborhood Commercial	1,665	5.20	0.28	0.59	0.13	31	0.10	48.97
Castro Street Neighborhood Commercial	381	3.28	0.41	0.41	0.17	303	2.61	68.45
Upper Market Neighborhood Comm. Transit	233	0.99	0.24	0.48	0.27	1,192	5.07	68.24
Public	944	0.48	0.40	0.47	0.13	7,410	3.76	70.54
Service/Arts/Light Industrial	106	0.16	0.42	0.44	0.14	769	1.20	69.4
Downtown Retail	147	0.16	0.56	0.28	0.16	112	0.12	61.4
Downtown Office (Special Development)	185	0.11	0.56	0.36	0.08	369	0.22	80.25
Mission Street Neighborhood Comm. Transit	96	0.10	0.58	0.28	0.14	630	0.68	73.43
Downtown Support	23	0.09	0.30	0.43	0.26	111	0.43	79.76
Regional Commercial	12	0.07	0.67	0.33	0.00	624	3.69	70.51
Heavy Industrial	6	0.06	0.17	0.67	0.17	311	2.88	76.52
Residential-House, Three Family	1,295	0.05	0.28	0.58	0.14	2,521	0.09	64.57
Residential—Commercial, High Density	405	0.04	0.59	0.35	0.06	2,068	0.22	64.68
Chinatown Community Business	12	0.03	0.67	0.25	0.08	75	0.20	62.47
South of Market Residential Enclave	21	0.03	0.81	0.05	0.14	1,137	1.53	71.06
Residential-Mixed, Moderate Density	163	0.02	0.21	0.66	0.12	559	0.07	63.63
Neighborhood Commercial Moderate Scale	81	0.02	0.41	0.53	0.06	757	0.19	62.3
Mixed Use Office	22	0.02	0.68	0.18	0.14	636	0.56	70.75
Residential-Mixed, Low Density	236	0.01	0.23	0.62	0.15	1,220	0.07	69.67
Downtown Office	10	0.01	0.20	0.80	0.00	185	0.10	57.26
Police District
Park District (incl. the Haight)	2,051	0.14	0.27	0.60	0.13
Tenderloin District	262	0.22	0.66	0.29	0.06

Note. Incidents and requests are counted under multiple categories if they occur within 100 feet of multiple parcel types.

Analysis of encampment-related enforcement yields more ambiguous findings. Citywide, the average time-to-completion for encampment service requests was 68.7 hours, or nearly three days. There is some variation around the mean, but there is little evidence of a discernible pattern of enforcement privileging particular neighborhoods or land uses. Response time is a bit quicker in some residential or neighborhood commercial zones, but these areas also tend to receive fewer requests. For the 52 zoning types included in the full analysis (which include 92% of all parcels), time-to-completion and requests per parcel were correlated at 0.34, significant to the .05 level.²⁰ Volume of requests, not land use or city region, appears to be a better predictor of the time it takes city agencies to respond to encampment service requests.

Discussion

Jointly, analysis of public records and election data reveals evidence of an ensemble of actors who all contribute resources critical to the adoption of anti-homeless policies in San Francisco. Large business and industry groups sought to shape public discourse by contributing substantial financial resources to support both ballot measures, allowing the campaigns in favor of the sit-lie ordinance and the tent ban to dramatically outspend opponents of the proposals. However, analysis of public comments before the BOS reveals that neighborhood merchants (especially those representing the Haight Ashbury neighborhood) played a particularly dominant role in pressuring city officials to adopt the sit-lie ban.

Although powerful economic actors in the city supported both the sit-lie ordinance and the tent ban, elected officials were unwilling, unable, or perhaps never intended to use their authority to turn these proposals into legislation. In both instances, the BOS was divided between supporters and opponents of the proposed anti-homeless laws. The data analyzed offer no way to conclusively determine what motivated city officials to propose, support, or oppose the measures. We cannot be certain whether Mayor Newsom and Supervisor Farrell proposed their respective measures for purely political purposes as critics commonly claim (Fagan and Green 2017; Walsh 2002), or that Supervisors’ stances on the proposed measures were primarily informed by the opinions of their constituents, though in 2016 there does appear to have been an association between Supervisor representation and voter support for the tent ban. Nor can we conclusively determine that Supervisors who opposed the sit-lie proposal were especially susceptible to the arguments presented by the mobilized and vocal opposition, although it seems likely that, absent such opposition and attention, officials would have been more willing to respond to the demands of the business community. But what appears more important than the forces that caused the divide among city officials is that the divide itself existed at all. Evidence that city officials were not united in support for the anti-homeless measure and refused to directly legislate the will of the business community presents problems for perspectives of city power that acknowledge only economic pressure in their accounts of city policy making. City officials did, however, contribute a crucial resource required for the ultimate adoption of each anti-homeless law: the authority to place the proposals on the ballot. Without this formal power delegated to the mayor or a coalition of four Supervisors, it is far less likely either proposal would have become law.

Election results indicate that the politics of anti-homeless policies in San Francisco is by no means “groupless” (Peterson 1981, p. 128). Two key groups—homeowners and registered Republicans—ultimately contributed the final resource required to secure the adoption of each anti-homeless measure: votes. Precincts with larger shares of homeowners and Republicans provided consistently higher levels of electoral support for the anti-homeless ballot measures. In considering these results, it is important to reiterate that the electoral precinct is the unit of analysis, not the individual. It should not be uncritically inferred that Republicans and homeowners were more supportive of the measures—such inferences would be vulnerable to the ecological inference problem (King 1997). What we can confidently say is that support was higher for the measures in neighborhoods with more Republicans and homeowners. With that in mind, two interpretations of these results seem most plausible. First, an increased density of conservative or home-owning residents may foster a neighborhood culture especially concerned with blight or disorganization associated with homelessness. More homeowners might lead to more active homeowners associations that work to inform neighborhood residents of the benefits of banning sitting and lying and tents on sidewalks. The other interpretation is that Republicans and homeowners were in fact more likely to support anti-homeless proposals, as past research would predict, even if methodological limitations inhibit direct evidence of this individual-level relationship. Either way, we can be fairly confident that both groups—whether directly or indirectly—played an important role in passing both ballot measures.

The racial/ethnic composition of neighborhoods also appears to have influenced electoral support for ballot measures, and while ecological inference concerns should again encourage cautious interpretation, qualitative findings bolster the statistical relationships revealed. The finding that precincts with greater proportions of Latino residents contributed lower levels of support for the sit-lie proposal in 2010 makes sense in light of concerns expressed during BOS hearings by residents of the Mission neighborhood (home to a large Latino population) and Supervisor Campos who represented the Mission, that the law would be used to target day laborers. Analysis also indicates that predominantly Asian and Asian-American precincts supported both ballot measures at higher rates than city averages. Among those in 2010 who spoke in favor of the sit-lie ordinance at BOS hearings were a number of parents of students attending a new Chinese Immersion school (located on Haight Street). The Asian Pacific Democratic Club also expressed support for Measure L in the 2010 Voter Information Pamphlet, claiming “Our school children are threatened by people who encamp on public sidewalks and use aggressive behavior to intimidate pedestrians” (Department of Elections 2010, p. 147). Furthermore, city Supervisors representing the Sunset neighborhood in the fourth district (home to many Asian and Asian-American residents) supported both proposed measures. Joining findings from qualitative and statistical analyses affords greater confidence that neighborhood ethnic composition played a role in shaping electoral support of the anti-homeless legislation.

Cumulatively, analysis of the adoption process for these two anti-homeless laws indicates that members of an informal coalition composed of business interests, public officials, and groups of city residents each contributed unique resources required to turn proposal into law. To follow pluralists in emphasizing only elections and the capacity of mobilized groups to shape city policy, or to adopt the economic imperative perspective and attend only to the influence of financial pressures and business elites over policy makers, is to paint too simplistic a picture of what is actually a nuanced, multilayered relationship. Instead, regime theory, by acknowledging the critical interrelated political influence of capital, public authority, and elections, more accurately captures the dynamics behind the anti-homeless policies studied here. Anti-homeless policies in San Francisco appear to be the result of what amounts to an anti-homeless regime in the city.

What, then, does analysis of enforcement patterns contribute to our understanding of anti-homeless politics in the city? Despite frequent claims that anti-homeless policies are enforced to appease downtown business interests (Goetz 1994; Mitchell 2011), findings reveal little evidence of such narrow preferential treatment. Instead, the sit-lie ordinance is enforced most strictly in the streets surrounding the merchants who mobilized for its adoption, as well as in lower-density residential areas. Enforcement also varies dramatically even between two neighborhoods—Haight Ashbury and the Tenderloin—both known to have high concentrations of individuals experiencing homelessness. Police in the Haight were far more likely to issue citations and bookings, while in the Tenderloin warnings were more prevalent. The Tenderloin resident who testified in favor of the law but worried that enforcement would privilege the Haight over other neighborhoods appears to have been justified in her concerns.²¹ Had the analysis indicated that downtown business interests are the chief beneficiaries of anti-homeless laws, regardless of who contributes to the adoption process, it would have undercut claims that power is shared among members of the anti-homeless regime. But this was not the case. Instead, findings indicate that the Haight neighborhood merchants (and homeowners in low-density neighborhoods) are not just members of the coalition needed to adopt these laws, but also benefit from how the laws are implemented, revealing an additional, subtle level of power over city policy.

Evidence concerning encampment clearance, on the other hand, betrays little evidence of biased enforcement. Although results are not definitive, what appears to determine response time to encampment service requests is the volume of requests submitted in a particular region, indicating that, at least in this case, service provision seems to be determined primarily by agency capacity, not political power. Several limitations make analyzing the tent ban more difficult and problematic than analysis of sit-lie enforcement. As noted above, the city possessed the legal authority to clear encampments before voters passed Proposition Q, so there is less of a direct connection between the tent ban adopted and the enforcement practices being assessed. Furthermore, while service request “time to completion” is the only data available for assessing enforcement bias, it is likely shaped by a host of factors that cannot be included in the analysis. Despite limitations, this exploratory investigation into the implementation of anti-homeless laws should encourage further empirical inquiry into the enforcement patterns of policies intended to regulate urban public space.

Given the history, severity, and political salience of homelessness in San Francisco, can these findings speak to homeless politics in other cities? The San Francisco case is a perfect storm, possessing all the actors and forces that we would expect to shape debates over anti-homeless policy, and this limits the extent to which findings here may be generalized to other cities. However, this study of San Francisco reveals political dynamics that will likely be present in debates over anti-homeless policies elsewhere. The unique resources possessed by business, local officials and city residents—capital, authority, and votes—will surely interact (though in various configurations) to shape the adoption process. Groups who feel threatened—like homeowners concerned about property values, or the Latino residents of the Mission concerned the sit-lie ordinance would discriminate against day laborers—will likely mobilize to influence the debate. In cities lacking the political “countervailing forces” present in San Francisco, business elites and city officials will more easily influence the policy-making process (Steffen 2012). Like San Francisco, the debate will not conclude when policies are adopted. Enforcement practices of anti-homeless laws in other cities, too, will likely be shaped by local power dynamics, and may continue defining winners and losers long after proposals are enshrined in law.

Anti-homeless Policy and City Power

Homelessness remains an issue of major public concern and political debate in cities throughout the United States, so it is important to better understand the political forces behind the particular homeless policies that cities pursue. In recent years, anti-homeless policies have been adopted in San Francisco because of the collective efforts and resources contributed by a varied cast of political actors. Business and neighborhood merchants played a major role getting policy proposals on the agenda and shaped public debate through testimony and campaign contributions. But business and merchant pressure was not sufficient to convince elected officials to legislate anti-homeless policies themselves.

In both 2010 and 2016, officials responded strategically by using their authority to place proposals on the ballot, leaving the ultimate decision over anti-homeless policies to the voters, as San Francisco politicians had done for decades before. Residents in more conservative neighborhoods with higher percentages of homeowners provided the votes necessary to pass both measures. Electoral support also came from those neighborhoods submitting higher levels of homeless-related 311 service requests. Once the laws were adopted, they have at times been enforced in a biased manner, but not to the exclusive benefit of downtown businesses as is frequently assumed. Instead, the sit-lie ordinance has privileged the neighborhood merchants who played such a substantial role in bringing the proposal to the attention of city officials. Encampment clearance, on the other hand, appears more shaped by administrative capacity than preferential treatment.

So then, who banishes? What political power dynamics lead to the adoption and shape enforcement of punitive anti-homeless laws in the paradigmatic progressive city? To attribute sole responsibility to any one political actor distorts the nuanced and dynamic nature of homeless politics in San Francisco. Business associations and neighborhood merchants, elected officials, homeowners, and conservatives all play an important part. The anti-homeless policies studied here were achieved by the city’s anti-homeless regime: an informal coalition of actors all contributing unique resources necessary to adopt anti-homeless policies. Without the participation of merchants and big business, elected officials, and select social groups, it is far less likely the anti-homeless policies analyzed would have been adopted. Once on the books, multiple members of the anti-homeless regime—and not just downtown business—benefit from biased enforcement of anti-homeless laws. Power is at play not only in the adoption process, but in implementation as well.

Beyond empirical findings, this study makes a compelling case for applying a mixed-methods approach to the study of city power. Evidence of power possessed by assorted political actors must be discovered in different sources and by utilizing different techniques. Qualitative, statistical, and spatial analyses were jointly required to identify the political actors involved and the resources they contributed toward shaping the anti-homeless policy debate. Lacking theoretical guidance, electoral opposition to anti-homeless policy from Latino neighborhoods and support from Asian neighborhoods only make sense when joined with qualitative analysis of statements made before elected officials and printed in voter guides. Similarly, claims that region-specific merchants powerfully influenced the policy adoption process are more convincing when joined by evidence that the laws are also enforced more strictly in their neighborhoods.

Conceptions of city power can and should be enriched by incorporating analysis of both policy adoption and the subtle ways power is expressed through policy enforcement. As Stone (2017) has noted, the distinction between policies that distribute selective benefits and those providing collective benefits blurs when implementation practices are taken into account. Assessing both the political forces that lead to policy adoption as well as the power relations betrayed by patterns of policy enforcement yields a more detailed and dynamic depiction of political influence. If city power fundamentally involves struggles over urban space, focusing more attention on spatial variation in the implementation of public policy is likely to enrich our understanding of who governs cities today.

Footnotes

Appendix.

Summary Statistics for Spatial Regression Models.

	2010				2016
	Minimum	Maximum	M	SD	Minimum	Maximum	M	SD
Support for anti-homeless ballot measures	0.16	0.78	0.50	0.11	0.28	0.68	0.47	0.07
% Republican	0.00	0.31	0.09	0.05	0.00	0.24	0.07	0.04
% Homeownership	0.00	0.92	0.41	0.23	0.00	0.98	0.42	0.24
Human Waste 311s (model 1)	0	1,003	35.72	117.6	—	—	—	—
Medical Waste 311s (model 2)	0	30	1.46	3.40	—	—	—	—
Encampment 311s (model 3)	—	—	—	—	0	2,893	80.74	235.26
Turnout	0.25	0.86	0.61	0.09	0.45	0.99	0.81	0.08
Supervisor support	0	1	0.28	0.45	0	1	0.39	0.49
% Black	0.00	0.57	0.06	0.08	0.00	0.60	0.05	0.08
% Asian	0.04	0.93	0.31	0.19	0.03	0.91	0.31	0.19
% Latino	0.01	0.63	0.14	0.12	0.00	0.68	0.14	0.12

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) received no financial support for the research, authorship, and/or publication of this article.

ORCID iD

David J. Amaral

Notes

References

Amos

Brian

P. McDonald

Michael

Watkins

Russell

. 2017. “When Boundaries Collide: Constructing a National Database of Demographic and Voting Statistics.” Public Opinion Quarterly 81 (S1): 385–400. doi:10.1093/poq/nfx001.

Amster

Randall

. 2003. “Patterns of Exclusion: Sanitizing Space, Criminalizing Homelessness.” Social Justice 30 (1): 195–221.

Anselin

Luc

. 2005. Exploring Spatial Data with GeoDa: A Work Book. Spatial Analysis Laboratory, Univ. of Illinois. Accessed May 1, 2019. http://www.csiss.org/clearinghouse/GeoDa/geodaworkbook.pdf.

Beckett

Katherine

Herbert

Steven Kelly

. 2010. Banished: The New Social Control in Urban America. Oxford: Oxford Univ. Press.

Benedictis-Kessner

Justin de

Warshaw

Christopher

. 2016. “Mayoral Partisanship and Municipal Fiscal Policy.” The Journal of Politics 78 (4): 1124–38. doi:10.1086/686308.

Bonds

Eric

Martin

Leslie

. 2016. “Treating People Like Pollution: Homelessness and Environmental Injustice.” Environmental Justice 9 (September). doi:10.1089/env.2016.0021.

Campbell

David E.

2006. Why We Vote: How Schools and Communities Shape Our Civic Life. Princeton: Princeton Univ. Press.

Choi

Sang Ok

Bae

Sang-Seok

Kwon

Sung-Wook

Feiock

Richard

. 2010. “County Limits: Policy Types and Expenditure Priorities.” The American Review of Public Administration 40 (1): 29–45. doi:10.1177/0275074008328171.

Coleman

Roy

Tombs

Steve

Whyte

Dave

. 2005. “Capital, Crime Control and Statecraft in the Entrepreneurial City.” Urban Studies 42 (13): 2511–30. doi:10.1080/00420980500380428.

10.

Connolly

Jennifer M.

Mason

Dyana P.

. 2016. “Ideology and Local Public Expenditure Priorities.” Political Research Quarterly 69 (4): 830–41. doi:10.1177/1065912916665702.

11.

Dahl

Robert A.

1961. Who Governs? Democracy and Power in an American City. New Haven: Yale Univ. Press.

12.

DeLeon

Richard Edward

. 1992. Left Coast City: Progressive Politics in San Francisco, 1975-1991. Lawrence: University Press of Kansas.

13.

Department of Elections. 2010. “Voter Information Pamphlet.” City and County of San Francisco. Accessed May 1, 2019. https://sfpl.org/pdf/main/gic/elections/November2_2010.pdf.

14.

DeVerteuil

Geoffrey

. 2006. “The Local State and Homeless Shelters: Beyond Revanchism?” Cities 23 (2): 109–20. doi:10.1016/j.cities.2005.08.004.

15.

Donovan

Todd

Neiman

Max

. 1992. “Citizen Mobilization and the Adoption of Local Growth Control.” The Western Political Quarterly 45 (3): 651–75. doi:10.2307/448686.

16.

Duncan

James

. 1978. “Men Without Property: The Tramp’s Classification and Use of Urban Space.” Antipode 10 (1): 24–34. doi:10.1111/j.1467-8330.1978.tb00292.x.

17.

Dye

Thomas R.

1979. “Politics Versus Economics: The Development of the Literature on Policy Determination.” Policy Studies Journal 7 (4): 652–62. doi:10.1111/j.1541-0072.1979.tb01360.x.

18.

Einstein

Katherine Levine

Kogan

Vladimir

. 2016. “Pushing the City Limits: Policy Responsiveness in Municipal Government.” Urban Affairs Review 52 (1): 3–32. doi:10.1177/1078087414568027.

19.

Einstein

Katherine Levine

Palmer

Maxwell

Glick

David M.

. 2019. “Who Participates in Local Government? Evidence from Meeting Minutes.” Perspectives on Politics 17 (1): 28–46.

20.

Enos

Ryan D.

2017. The Space between Us: Social Geography and Politics. Cambridge: Cambridge Univ. Press.

21.

Fagan

Kevin

Green

Emily

. 2017. “SF’s Voter-Approved Camp-Sweep Measure More Symbol than Substance.” San Francisco Chronicle, March 26. Accessed December 18, 2019. https://www.sfchronicle.com/bayarea/article/SF-s-voter-approved-camp-sweep-measure-more-11028060.php.

22.

Farrell

Chad R.

2005. “Sharing Neighborhoods: Order and Disorder in Homeless-Domiciled Encounters.” American Behavioral Scientist 48 (8): 1033–54. doi:10.1177/0002764204274198.

23.

Fischel

William A.

2001. The Homevoter Hypothesis: How Home Values Influence Local Government Taxation, School Finance, and Land-use Policies. Cambridge: Harvard Univ. Press.

24.

Goetz

Edward G.

1994. “Expanding Possibilities in Local Development Policy: An Examination of U.S. Cities.” Political Research Quarterly 47 (1): 85–109. doi:10.1177/106591299404700105.

25.

Goetz

Edward G.

2009. “Land Use and Homeless Policy in Los Angeles.” International Journal of Urban and Regional Research 16 (4): 540–54. doi:10.1111/j.1468-2427.1992.tb00195.x.

26.

Gowan

Teresa

. 2010. Hobos, Hustlers, and Backsliders: Homeless in San Francisco. Minneapolis: Univ. of Minnesota Press.

27.

Green

Matthew

. 2019. “TIMELINE: The Frustrating Political History of Homelessness in San Francisco.” KQED News, August 2. Accessed December 15, 2019. https://www.kqed.org/news/11765010/timeline-the-frustrating-political-history-of-homelessness-in-san-francisco.

28.

Hajnal

Zoltan

Trounstine

Jessica

. 2010. “Who or What Governs? The Effects of Economics, Politics, Institutions, and Needs on Local Spending.” American Politics Research 38 (6): 1130–63. doi:10.1177/1532673X10362870.

29.

Harvey

1989. “From Managerialism to Entrepreneurialism—The Transformation in Urban Governance in Late Capitalism.” Geografiska Annaler, Series B: Human Geography 71 (1): 3–17. doi:10.2307/490503.

30.

Herring

Chris

. 2014. “The New Logics of Homeless Seclusion: Homeless Encampments in America’s West Coast Cities.” City & Community 13 (4): 285–309. doi:10.1111/cico.12086.

31.

Jun

Kyu-Nahm

Musso

Juliet

. 2013. “Participatory Governance and the Spatial Representation of Neighborhood Issues.” Urban Affairs Review 49 (1): 71–110. doi:10.1177/1078087412453704.

32.

Kearns

Ade

. 1995. “Active Citizenship and Local Governance: Political and Geographical Dimensions.” Political Geography, Spaces of Citizenship 14 (2): 155–75. doi:10.1016/0962-6298(95)91662-N.

33.

King

Gary

. 1997. A Solution to the Ecological Inference Problem: Reconstructing Individual Behavior from Aggregate Data. Princeton, NJ: Princeton Univ. Press.

34.

Kinsey

Barbara Sgouraki

Bartling

Hugh

Peterson

Anne F.

Baybeck

Brady P.

. 2010. “Location of Public Goods and the Calculus of Voting: The Seattle Monorail Referendum.” Social Science Quarterly 91 (3): 741–61. doi:10.1111/j.1540-6237.2010.00717.x.

35.

Lee

Barrett A.

Farrell

Chad R.

Link

Bruce G.

. 2004. “Revisiting the Contact Hypothesis: The Case of Public Exposure to Homelessness.” American Sociological Review 69 (1): 40–63. doi:10.1177/000312240406900104.

36.

Lipsky

Michael

. 1980. Street-Level Bureaucracy: Dilemmas of the Individual in Public Services. New York: Russell Sage Foundation.

37.

Logan

John R.

Molotch

Harvey Luskin

. 1987. Urban Fortunes: The Political Economy of Place. Berkeley, CA: Univ. of California Press.

38.

MacDonald

Heather

. 1995. “Commentary: San Francisco’s Matrix Program for the Homeless.” Criminal Justice Ethics 14 (1): 2–80. doi:10.1080/0731129X.1995.9991983.

39.

MacLeod

Gordon

. 2002. “From Urban Entrepreneurialism to a ‘Revanchist City’? On the Spatial Injustices of Glasgow’s Renaissance.” Antipode 34 (3): 602–24. doi:10.1111/1467-8330.00256.

40.

May

Jon

Cloke

Paul

. 2013. “Modes of Attentiveness: Reading for Difference in Geographies of Homelessness.” Antipode 46:894–920. doi:10.1111/anti.12043.

41.

Minkoff

Scott L.

2009. “Minding Your Neighborhood: The Spatial Context of Local Redistribution.” Social Science Quarterly 90 (3): 516–37. doi:10.1111/j.1540-6237.2009.00629.x.

42.

Minkoff

Scott L.

2016. “NYC 311: A Tract-Level Analysis of Citizen–Government Contacting in New York City.” Urban Affairs Review 52 (2): 211–46. doi:10.1177/1078087415577796.

43.

Mitchell

Don

. 1997. “The Annihilation of Space by Law: The Roots and Implications of Anti-Homeless Laws in the United States.” Antipode 29 (3): 303–335. doi:10.1111/1467-8330.00048.

44.

Mitchell

Don

. 1998a. “Anti-Homeless Laws and Public Space: I. Begging and the First Amendment.” Urban Geography 19 (1): 6–11.

45.

Mitchell

Don

. 1998b. “Anti-Homeless Laws and Public Space: II. Further Constitutional Issues.” Urban Geography 19 (2): 98–104.

46.

Mitchell

Don

. 2011. “Homelessness, American Style.” Urban Geography 32 (7): 933–56. doi:10.2747/0272-3638.32.7.933.

47.

Mossberger

Karen

Stoker

Gerry

. 2001. “The Evolution of Urban Regime Theory: The Challenge of Conceptualization.” Urban Affairs Review 36 (6): 810–35.

48.

Murphy

Dean E.

2003. “Left Faces Left in San Francisco Runoff Vote for Mayor.” The New York Times, December 7. Accessed December 23, 2019. https://www.nytimes.com/2003/12/07/us/left-faces-left-in-san-francisco-runoff-vote-for-mayor.html.

49.

Murphy

Stacey

. 2009. “‘Compassionate’ Strategies of Managing Homelessness: Post-Revanchist Geographies in San Francisco.” Antipode 41 (2): 305–25. doi:10.1111/j.1467-8330.2009.00674.x.

50.

Oakley

Deirdre

. 2002. “Housing Homeless People: Local Mobilization of Federal Resources to Fight NIMBYism.” Journal of Urban Affairs 24 (December): 97–116. doi:10.1111/1467-9906.00116.

51.

O’Brien

Daniel Tumminelli

. 2015. “Custodians and Custodianship in Urban Neighborhoods: A Methodology Using Reports of Public Issues Received by a City’s 311 Hotline.” Environment and Behavior 47 (3): 304–27. doi:10.1177/0013916513499585.

52.

Openshaw

Stan

. 1983. The Modifiable Areal Unit Problem. Norwich: Geo Books.

53.

Peterson

Paul E.

1981. City Limits. Chicago: Univ. of Chicago Press.

54.

Sampson

Robert J.

2012. Great American City: Chicago and the Enduring Neighborhood Effect. Chicago: Univ. of Chicago Press.

55.

Selbin

Jeffrey

Campos-Bui

Stephanie

Epstein

Joshua

Lim

Laura

Nacino

Shelby

Wilhlem

Paula

Stommel

Hannah

. 2018. “Homeless Exclusion Districts: How California Business Improvement Districts Use Policy Advocacy and Policing Practices to Exclude Homeless People from Public Space.” SSRN Scholarly Paper ID 3221446, Social Science Research Network, Rochester, NY. https://papers.ssrn.com/abstract=3221446.

56.

Selbin

Jeffrey

Campos-Bui

Stephanie

Feldstein

Sasha

Fisher

Marina

Miller

Nathaniel

. 2016. “California’s New Vagrancy Laws: The Growing Enactment and Enforcement of Anti-Homeless Laws in the Golden State (2016 Update).” SSRN Scholarly Paper ID 2794386, Social Science Research Network, Rochester, NY. https://papers.ssrn.com/abstract=2794386.

57.

Steffen

Charles G.

2012. “The Corporate Campaign against Homelessness: Class Power and Urban Governance in Neoliberal Atlanta, 1973–1988.” Journal of Social History 46 (1): 170–96. doi:10.1093/jsh/shs031.

58.

Stone

Clarence N.

1989. Regime Politics: Governing Atlanta, 1946-1988. Lawrence: University Press of Kansas.

59.

Stone

Clarence N.

2017. “Trends in the Study of Urban Politics: A Paradigmatic View.” Urban Affairs Review 53 (1): 3–39. doi:10.1177/1078087416674328.

60.

Stuart

Forrest

. 2015. “On the Streets, Under Arrest: Policing Homelessness in the 21st Century.” Sociology Compass 9 (11): 940–50. doi:10.1111/soc4.12324.

61.

Sylvestre

Marie-Eve

. 2010a. “Disorder and Public Spaces in Montreal: Repression (and Resistance) Through Law, Politics, and Police Discretion.” Urban Geography 31 (6): 803–24. doi:10.2747/0272-3638.31.6.803.

62.

Sylvestre

Marie-Eve

. 2010b. “Policing the Homeless in Montreal: Is This Really What the Population Wants?” Policing and Society 20 (4): 432–58. doi:10.1080/10439463.2010.523114.

63.

Takahashi

Lois M.

1997. “The Socio-Spatial Stigmatization of Homelessness and HIV/AIDS: Toward an Explanation of the NIMBY Syndrome.” Social Science & Medicine 45 (6): 903–14. doi:10.1016/S0277-9536(96)00432-7.

64.

Thrush

Glenn

. 2018. “Homelessness Rises Slightly Despite Strong Economy, Federal Report Finds.” The New York Times, December 18. Accessed May 17, 2019. https://www.nytimes.com/2018/12/17/us/politics/homelessness-increases-seattle-new-york-.html.

65.

Tiebout

Charles M.

1956. “A Pure Theory of Local Expenditures.” Journal of Political Economy 64 (5): 416–24. doi:10.1086/257839.

66.

Trounstine

Jessica

. 2018. Segregation by Design: Local Politics and Inequality in American Cities. Cambridge: Cambridge Univ. Press. doi:10.1017/9781108555722.

67.

Verba

Sidney

Schlozman

Kay Lehman

Brady

Henry E.

. 1995. Voice and Equality: Civic Voluntarism in American Politics. Cambridge: Harvard Univ. Press.

68.

Walsh

Joan

. 2002. “Bummed Out.” Salon, November 4. Accessed January 10, 2020. https://www.salon.com/2002/11/04/homeless_2/.

69.

White

Ariel

Trump

Kris-Stella

. 2018. “The Promises and Pitfalls of 311 Data.” Urban Affairs Review 54 (4): 794–823. doi:10.1177/1078087416673202.

70.

Wong

Julia Carrie

. 2016. “Wealthy San Francisco Tech Investors Bankroll Bid to Ban Homeless Camps.” The Guardian, October 12. Accessed January 9, 2020. https://www.theguardian.com/us-news/2016/oct/12/san-francisco-homeless-proposition-q-tech-investors.