[Recommended] - Abstract - Academic Writers Bay

FIND A SOLUTION AT Academic Writers Bay

Resisting Broken Windows:The Effect of Neighborhood Disorder on Political BehaviorSeptember 18, 2019AbstractConcurrent housing and opioid crises have increased exposure to street-crime, homelessnessand addiction in American cities. What are the political consequences of this increased neighborhood disorder? We examine a change in social context following the relocation of homelessnessand drug treatment services in Boston. In 2014, an unexpected bridge closing forced nearly1,000 people receiving emergency shelter or addiction treatment to relocate from an island inthe Boston Harbor to mainland Boston, causing sustained increases in drug-use, loitering, andother features of neighborhood disorder. Residents near the relocation facilities mobilized tomaintain order in their community. In the subsequent Mayoral election, their turnout grew 9percentage points while participation in state and national elections was unchanged. Moreover,increased turnout favored the incumbent Mayor, consistent with voter learning about candidatequality following local shocks. Voters responded to neighborhood changes at the relevant electoral scale and rewarded responsive politicians.Keywords: Neighborhood disorder, local politics, political behavior, voter attributionConcurrent homelessness and drug crises have raged across many North American cities, sparking contentious debates about the appropriate political response (Sperance 2018, Whelan 2018, Wing 2019). Thespatially concentrated nature of these crises – reified by tent encampments, shelters, drug clinics, and syringes – portend particular worry because these visible elements of disorder encroach on the spaces livedand traveled on by citizens. Consequently, the governmental response to homelessness and addiction hasbeen a central question in dozens of recent local elections (Appendix Table A1) and defined the tenures ofincumbent mayors across the country (Halverstadt 2018, Malas & Lazo 2018). This anxiety around exposureto homelessness and addiction reflects the longstanding concern “that the street is disorderly, a source ofdistasteful, worrisome encounters” (Wilson 1975, 65). Neighborhood disorder – often metaphorically described as “broken-windows” – has for decades motivated aggressive “quality-of-life” policing and dominatedmayoral elections (Vitale 2008).1 Despite the first-order importance of order maintenance to local politics,political science has been largely silent about its effect at the ballot box.2Yet a strong theoretical basis suggests that exposure to neighborhood disorder will increase politicalparticipation and influence vote choice, particularly in local elections. While political opinion in the UnitedStates is increasingly nationalized – and hence decoupled from spatial variation in issue impacts – proximityto crime remains an important determinant of attitudes towards crime and policing (Hjorth 2017, Hopkins2018). Street-level encounters with homeless and drug-using populations heighten fear of crime (Wilson1968, Vitale 2008) and provoke discomfort and disgust among many individuals (Clifford & Piston 2017).Fear and discomfort stemming from repeated exposure to disorder likely impels residents to lobby policeand local officials to address disorder and to organize new forms of social capital to amplify their collectivevoice. Thus, we hypothesize that residents will respond to neighborhood disorder by mobilizing to protecttheir neighborhood, resulting in higher levels of political participation, and that the intensity of the politicalresponse will increase with proximity to disorder (Tobler 1970).To test these hypotheses, we examine an abrupt relocation of homeless and drug-using populations inBoston to identify how disorder affects voting behavior. We combine rich contextual analysis of the casewith large-scale quantitative analyses of (1) everyday political participation through non-emergency andemergency requests for government services, (2) individual-level voting data measuring participation inlocal, state, and national elections, and (3) electoral choices in Mayoral and Presidential elections. We findthat requests for government services increased, turnout increased in local elections (but not in state or1For examples, concerns about rampant disorder in the early 1990s propelled “law-and-order” mayoral candidates like New YorkCity’s Rudy Guliani and San Francisco’s Frank Jordan to unseat progressive incumbents (Vitale 2008).2“Quality of life and disorder continue to be among the most urgent issues local politicians address, regardless of party affiliation”(Kelling & Coles 1997, 14). Michener (2013) uses the Chicago Neighborhood Survey to examine the relationship between disorder,attending community meetings, and contacting local officials.1national ones), and, contrary to our expectations, affected precincts swung towards the incumbent Mayor,but showed no changes in the next Presidential election.These results provide the first systematic evidence that increased disorder alters voter behavior, increasingturnout and changing vote choice. We further show that changes in neighborhood disorder cause individuallevel political responses at local, but not state or national, levels. In this case, voters responded solely atthe relevant electoral scale – the local government, responsible for policing, public works, and siting sheltersand drug treatment facilities. Because the relocation of homeless populations is a pervasive political issue(McKellar 2017, Smith & Smith 2018, Steinbuch 2019, Billings 2019), our finding that increased disorderprovokes a substantial local political response generalizes to a large set of cases involving the construction ofnew facilities (Innes 2019) or displacing existing drug or homeless encampments (Orange 2013, Kim 2017).Moreover, these results suggest that “man-made” shocks do not always produce anti-incumbent electoraleffects. The abrupt relocation we study is neither a natural disaster nor an economic downturn but shareswith these events a complex causal origin, high salience, and an opportunity to learn about the incumbent’squality through her response (Ashworth et al. 2018). Consistent with studies of natural disasters (Healy &Malhotra 2009, 2010), we find the incumbent benefited from adeptly managing the fallout.The Politics of DisorderDisorder comprises the physical (e.g. litter, graffiti, vacant buildings) and behavioral (e.g. public drug use,loitering, panhandling) features of neighborhoods that suggest the deterioration of social control and diminish quality of life (Gracia 2014, 4325). More informally, disorder tracks the“readily observable” featuresof “bad” neighborhoods (Michener 2013). Neighborhood residents report highly similar levels of disorderregardless of race, class, homeownership, and age (Skogan 1992, 54-77). A rich literature in sociology andpublic health suggests that disorder forms an important part of neighborhood context, associated with fearof crime (Hinkle & Weisburd 2008), distrust of neighbors (Ross & Jang 2000), stress (Latkin & Curry 2003),and reduced outdoor exercise (Molnar et al. 2004).Disorder is linked to the political process through the government’s role in providing safety and sanitationon the streets and sidewalks and in providing shelter and social services to disadvantaged people. Politicalparticipation can pressure governments to provide additional services to manage disorder, such as policing.A community leader in Ridgewood, Queens explains, “When you hear of a drug location, you have to screambloody murder to the police captain until you get results” (Kelling & Coles 1997, 56). Indeed, maintainingorder is a central demand that citizens place on their police, and police foot patrols are a common responseto these demands (Wilson 1968, 1975). Moreover, political action can pressure governments to devote2additional resources towards managing public streets and sidewalks to reduce physical features of disorderlike discarded needles (Melendez 2018) and human feces (Tyler 2018). Lastly, political action can dispersesources of disorder like tent encampments (Holland 2018) and halt or delay the construction of new facilitiesperceived to herald disorder (Bittle 2017) or push for the construction of such facilities elsewhere.Because people are more likely to participate in politics when the benefits exceed the costs (Riker & Ordeshook 1968), we expect that disorder will lead to greater political participation. Our argument proceedsas follows: disorder diminishes quality-of-life and is linked to the political process through the provisionof government services like assistance to the poor, policing, and sanitation. Consequently, citizens will become involved in politics to impel government officials to manage disorder. In other words, the instrumentalbenefits of participation increase as disorder increases. Dahl (1961) succinctly captures the essence of ourargument in his account of neighborhood resistance to building metal-roof apartments in mid-century NewHaven. Faced with what they perceived to be the existential threat of “slum-housing” development, residents of the Hill Neighborhood saw their fate linked with the political approval of the development’s zoningpermit. Organizing to pressure the city to deny a permit for the development, they formed a neighborhoodassociation, raised money to mount a legal defense, and protested at city meetings (Dahl 1961, 193-197).Living in a neighborhood marked by features of disorder such as street-crime, vagrancy, littered syringes,and public drug use is a source of aggravation, fear, and disgust for many residents (Skogan 1992, Michener2013). As one Seattle resident explains, “There’s a level of anxiety having a homeless camp on the otherside of my hedge. It gets higher as the garbage grows” (Wing 2019). Disorder portends a more generalthreat of neighborhood decay (Skogan 1992), and neighborhoods with high-levels of disorder like Philadelphia’s Kensington (Whelan 2018) or Portland’s Bayside (Billings 2018) are often described as “under siege.”Feelings of threat may increase the benefits of political participation to the extent that residents believe thatrelief comes from government intervention.3Threats to the neighborhood can also mobilize residents to organize in defense of the community as theysee it, creating new forms of social capital that amplify the collective voice (Orbell & Uno 1972, Hansen1985). For example, people who perceive more disorder in their community are also more likely to attendcommunity meetings (Michener 2013). The development or renewal of civic organizations, in turn, altersthe calculus of political behavior in favor of more participation: group membership subsidizes information,increases mobilization, and provides opportunities for honing civic skills (Leighley 1996).In addition to these instrumental reasons, disorder may induce expressive political participation (Schuessler2000) in the same way that crime victimization induces participation, providing an outlet for victims of crime3Political scientists have long argued that the distribution of racial groups across space shapes inter-group conflict and politicalbehavior through similar threat mechanisms (Key 1949, Enos 2014, 2015, Anoll 2018).3to “move beyond fear and submission” and reinvent themselves as organizers and activists (Bateson 2012,572). A San Francisco resident who became an activist on local street homelessness described her persistentreporting of unseemly street behavior as providing a sense of control over an unsettling environment. “I’mso tired of the complacency, of people saying, ‘There’s nothing we can do’… The one thing we can do is makenoise” to say “This isn’t OK” (Knight 2015).This discussion indicates that increased disorder should provoke a sharp political response. Disorderportends both fear of crime and a more general neighborhood deterioration (Skogan 1992), and resolutionof these fears are linked to the political process through government provision of public safety, sanitation,and social services to manage disorder. Consequently, political action provides both the means for residentsto impel government to better manage disorder and a sense of control amid the chaos. These theoreticalconsiderations lead to the following hypothesis:H1: Increased neighborhood disorder will increase political participation.Changes in neighborhood disorder may also affect vote choice. Local voters may blame the elected officials who presided during the changes and vote against those incumbents in subsequent elections (Hopkins &Pettingill 2018). Anecdotal evidence suggests that voters punished incumbent Mayors who failed to managegrowing disorder in the 1990s (Vitale 2008). More generally, a robust literature documents retrospectivevoting in American elections, as voters consider changes in their own lives during an incumbent’s tenure toassess candidate performance and inform their vote choices (Fiorina 1981, Healy & Malhotra 2013, Malhotra& Margalit 2014). As a Pomona, California City Councilor explains, “If [the new shelter] doesn’t work” toease concerns about homelessness, “we’re all going to get voted out” (Walker 2017). More recent literatureassesses whether natural disasters and other “shocks” influence incumbent electoral fortunes (Achen & Bartels 2016, Fowler & Hall 2018). Events outside the direct control of a politician still offer an opportunityfor rational voters to learn about the competence of their elected official through her response, potentiallyshaping re-election prospects (Ashworth et al. 2018). When neighborhood quality-of-life declines duringthe tenure of the incumbent, voters may learn that the incumbent is not equipped to deal with the problemand instead support a potentially higher-quality challenger. Consequently, both simple and sophisticatedconceptions of retrospective voting generate our second hypothesis:H2: Increased neighborhood disorder will increase anti-incumbent voting.Hypotheses 1 and 2 predict that changes in neighborhood disorder will a provoke political response. Butthe mechanisms through which it does so are a function of individual geographic proximity to neighborhoodchanges. Those living closest to sources of neighborhood disorder will be those most affected by it; for4example, the effect of homeless shelter openings on property crimes are concentrated very close to the shelterand decay beyond 400 meters (Faraji et al. 2018). Thus, threat perception, and ensuing political response,should intensify with proximity to the sources of neighborhood disorder (Enos 2017). This expectation isconsistent with the literature on geography and group threat, and is derived from Tobler’s (1970) first law ofgeography: “everything is related to everything else, but near things are more related than distant things.”Testing this hypothesis speaks to whether residents are reacting to real changes in their everyday experience(“objective” disorder) or their perceptions of changes from local media and politicians (“subjective” disorder)(Michener 2013). We formalize this intuition as hypothesis three:H3: Treatment effects will increase with proximity to the affected sites.We test these hypotheses by measuring changes in political behavior in national, statewide, and localelections before-and-after a change in neighborhood disorder. Estimating effects at different electoral scalesprovides insight into whether the political response to disorder is targeted at the level of government closestto management of the problem (Larsen 2018) or spread across all levels of government (Sances 2017, Larsenet al. 2019) with implications for voter rationality and retrospective voting.Research DesignTo test the effect of neighborhood disorder on political participation, we analyze an abrupt change in thegeographic distribution of homeless and drug-using populations in the city of Boston, detailed in Figure 1.For decades, Long Island, an island located in the Boston Harbor, was the hub of the city’s social-services,including the city’s largest homeless shelter and most of its addiction treatment services (BHCHP 2014,Ramirez 2015b). That changed when an inspector unexpectedly condemned the bridge connecting theisland with the mainland in October 2014.4 Following the condemnation of the bridge, the city of Bostonabruptly terminated the social services provided on the island, including 57% of the city’s total substance usetreatment beds (BHCHP 2014). About 1,000 people receiving daily services on the island were displaced,including 450 homeless people previously staying in the Long Island Homeless Shelter and another 300people in recovery living in the various residential detoxification facilities and transitional housing programs4Closure of the island was a shock to service providers and residents; Jonathan Scott, president of a recovery program for womenon the island, said “When we were first told we had to leave, we didn’t know the reason” (Irons & Dungca 2014). Press reports notethat “medications, journals, and plates of food that had just been warmed were all left behind” (Irons & Dungca 2014). The city hadpreviously announced plans to begin repairing the 63-year old bridge in September. An engineer involved with the reconstructionproject told local media that reconstruction would only require partial lane closures, and “under no circumstance will the whole bridgebe closed” (Burrell 2014).5Figure 1: Key Events in the Long Island Shock1951 City builds a bridge connecting Long Island in the Boston Harbor to the mainland.1983 City opens its largest homeless shelter and addiction treatment services on island.Oct. 8, 2014 (3PM) Long Island Bridge deemed imminently unsafe.Oct. 8, 2014 (3:50 PM) City declares immediate evacuation of island. Police block island access.Oct. 8, 2014 (7PM) Hundreds of residents relocated to emergency sites. Island services are terminated.Jan. 2015 Newly built Southampton Street Shelter opens near emergency sites.April 2016 Supportive Place for Observation and Treatment (SPOT) opens in South End.Oct. 2016 Opioid Urgent Care Center opens in South End.Nov. 7. 2017 Walsh defeats challenger Tito Jackson in Mayoral race.January 1, 2018 Mayor Walsh promises to re-build Long Island Bridge in Inauguration Speech.Sept. 27, 2018 State environmental board approves permit for bridge reconstructionon the island (Ramirez 2015b). Displaced facilities and associated populations were primarily relocated tothe Boston Medical Center area at the boundaries of the South End, Roxbury and Dorchester neighborhoods,as shown in Figures 2 and 3. This area was the site of the immediate re-location because it contained severalfacilities operated by the Boston Public Health Commission, which had just three hours to find new placesto host the so-called “Long Island Refugees” (BHCHP 2014, Ramirez 2015a). In the immediate aftermath ofthe Long Island closure, 250 men slept on the basketball court at the Public Health Commission’s South EndFitness Center, 45 women slept on cots in the atrium of the Commission’s homeless clinic in the South End,and 20 others stayed in emergency cots at the Commission’s Wood-Mullen Shelter (Burrell 2014, Becker &Jolicoeur 2014). These temporary relocation facilities were eventually replaced by newly built shelter andrenovated sites in the relocation area (Pressley 2015).According to the city’s public health agency, closure of Long Island facilities resulted in “rapid changes…throughout Boston, particularly in the South End neighborhoods and the already busy intersection of AlbanyStreet and Massachusetts Avenue” (BHCHP 2014). This part of Boston – along a stretch of MassachusettsAvenue containing the Boston Medical Center, neighborhoods of the South End and Lower Roxbury, andthe mostly industrial Newmarket Square district – has long been home to a concentration of social serviceproviders and methadone clinics, giving it the reputation of “Methadone Mile” (Daniel 2016). But existingresidents and local employees reported that the quality-of-life in the neighborhood declined sharply afterthe Long Island Bridge closure. Following the opening of the Southhampton Street Shelter, crime increased.Between January and October of 2015, violent crimes increased 30% and drug violations by 55% near the6Figure 2: Map of Long Island and Methadone MileL o n g I s l a n dSouth EndRoxburyDorchesterSouth BostonMMAA SS SS AA CC HH UU SS EE TT TT SS AAVV EE NN UU EEEsri, HERE, Garmin, © OpenStreetMap contributors, and the GIS user communityThe approximate stretch of Massachusetts Avenue called Methadone Mile, the site of Long Island services relocation, is colored in red.Long Island is outlined in black.new shelter, while these crimes declined citywide (Abel 2015). These quantitative trends match qualitativeaccounts of the neighborhood change. One resident posted that “I’ve been periodically walking from myhome in the South End to my business at Newmarket since 1988. the situation in 2015 is far worse thanI’ve ever seen. I now walk up Albany St. rather than Mass Ave. because of the crowds… to avoid any drugtrafficking. Also, the level of theft and break-ins in the Newmarket neighborhood is sky-rocketing” (Reilly2016). Years later, residents, politicians, and the media have continued to blame the closure of the LongIsland Bridge for exacerbating public safety and quality-of-life issues in the neighborhood (Karedes 2018,Jonas 2019, Singh 2019), and neighborhood associations have eyed re-opening the Long Island campus asintegral to solving the neighborhood’s seemingly intractable homelessness and drug problems (Seth 2018a).Ground zero of the crisis is the busy intersection of Melnea Cass Boulevard and Massachusetts Avenue,situated between Boston Health Care for the Homeless and the Wood Mullens Shelter to the north of the7Figure 3: Map of Voters Near SitesDots represent individual voters. Circles mark areas within 500, 1000, and 1600 meters of Wood-Mullens Shelter.intersection and the Southampton Street Shelter and two methadone clinics south of the intersection (Ramos& Allen 2016). To illustrate this, and the broader consequences of the neighborhood change, we downloadedimages from Google Street View of the same plots of land at the intersection before and after the Long IslandShock. As Figure 4 shows, dozens of people loitering, and pervasive litter, are apparent after the closureof the Long Island Bridge. Other before-and-after images captured in Google Street View (Appendix FigureA2) show the removal of benches at nearby bus stops and the fencing off of a patch of grass on Melnea CassBoulevard where people previously congregated, a move lampooned in local media with the headline “News25 Investigates: Mayor’s controversial fence moves addiction problem across the street” (Rasmussen & Smith2017). While differences in the photos may also be attributable to seasonal differences5, these images reflectthe neighborhood changes induced by the Long Island shock.5The 2017 images are from July, while the 2013 images are from September. Google did not offer images in 2017 and 2013 fromthe same month.8Figure 4: Neighborhood Disorder: Evidence from Google Street ViewMass Ave. & Melnea Cass, 2013 (Before Long Island Shock)2017 (After Long Island Shock)Mass Ave. & Melnea Cass, 2014 (Before Long Island Shock)2017 (After Long Island Shock)9The closing of the Long Island Bridge, and the consequent termination of the island’s facilities, increasedthe local density of homeless individuals and people who use drugs. The re-location of the displaced populations was not random, as they relocated to an area where the Boston Public Health Commission operatedexisting facilities. But that the re-location occurred at all, and the timing of the increased neighborhood disorder, is plausibly unrelated to the attitudes and behaviors of voters living near the re-location sites. As such,we argue that this event—the Long Island Shock—constitutes an as-if random increase in neighborhooddisorder.The Long Island Shock provides analytic leverage into the effect of neighborhood disorder on politicalbehavior. Our primary measure of political behavior is geolocated, individual-level administrative votingrecords for every registered voter in Massachusetts. These data, which include gender, race and ethnicity,place of residence, and voting history, were organized and provided to the researchers by the political analytics firm L2. We supplement these data with citywide calls for services records, precinct-level electoral results,and home-ownership parcel data downloaded from the city assessor. We collected 311 non-emergency calldata from the city of Boston’s website and 911 call data from the Boston Police Department via public recordsrequest.For each registered voter in Boston, we calculate the Crow’s distances (straight line) in meters fromtheir place of residence to the sites that saw an influx in homeless persons and drug users following the LongIsland Shock. A respondent’s distance to the nearest site – the minimum of the distance to any of the sites – isour measure of exposure. With these data, we measure differences in political participation before and afterthe bridge closing as a function of proximity to the sources of neighborhood changes. Defining treatment(living close to the re-location sites) and control (living far from the re-location sites) by increasing distances,we measure differences in participation across a spectrum of definitions of proximity, avoiding problems ofaggregation and scale common to geographic analysis (Openshaw 1983, White 1983). To test the effect ofincreased neighborhood disorder on vote choice, we examine changes in incumbent vote share with precinctlevel electoral data from the city of Boston that we digitized and geocoded. We calculate the Crow’s distanceof each voting precinct centroid to the relocation sites.Following a Boston City Council report on the Long Island Bridge Closure (Pressley 2015), we define thescope of relocated sites as those shown in Table 1. While we measure distance to these specific sites, theyare all within a few blocks of one another, and collectively serve as a measure of proximity to the increase inneighborhood disorder.We test whether increased neighborhood disorder caused voters living close to the sites to vote at higherrates (compared to voters living farther away) in the immediate next statewide, national, and mayoralelections, in 2014, 2016, and 2017, respectively. For each type of election we compared turnout rates to the10Table 1: New Locations for Long Island FacilitiesProgram Beds Bed Type New AddressSOAR 50 Transitional Housing 112 Southampton StreetSafe Harbor 20 Housing for HIV patients 112 Southampton StreetLong Island Shelter 368 Emergency Shelter 112 Southampton StreetWoods Mullen Shelter 200 Emergency Shelter 794 Massachusetts AvenueSouth End Fitness Center 200 Emergency Shelter 35 Northampton Streetimmediately prior election of the same type, the 2010 midterm election, the 2012 presidential election, andthe 2013 mayoral election. Our analysis is restricted to voters who lived in the same residence and wereregistered to vote in each election cycle. We identify these voters by restricting the analysis to voters in theL2 file whose registration date is prior to date of the first election in a given election pair. Voter registrationdates are updated when voters change their registration, generally when they move residences. Limiting ouranalysis to non-movers omits people who moved residences in between elections. Moving is a more costlybehavior than voting, so individuals who are motivated to move in response to increased neighborhooddisorder are probably likely to respond politically as well. Therefore, limiting our results to non-moversdownward biases our results toward null turnout effects.To test our hypotheses, we employ a difference-in-differences design, estimating the change in turnoutand vote share between like-elections before and after the shock to neighborhood disorder. Our unit of analysis for turnout is individual registered voters, and the unit of analysis for vote share is electoral precincts.We estimate the following average treatment effects:ATE = [P(Votet=2jd < d∗) – P(Votet=1jd < d∗)] – [P(Votet=2jd > d∗) – P(Votet=1jd > d∗)]In a linear regression, we estimate this ATE for turnout with the following equation, for voter i in election t:Turnoutit = αi + γt + β[I(di < d∗) · γt] + itHere, αi represents individual-level fixed effects, γt is a dummy for the treated period (or the change inturnout, on average, between election t = 1 and t = 2 for the untreated individuals). A voter’s distance tothe closest relocation facility is denoted di. I(di < d∗) is an indicator variable equal to one for individualswho live within the designated distance of any facility, thus delineating treated units. We estimate thisequation at increasing distances of d∗ (100; 200; 300; : : : ; 1600 meters). Finally, it is the idiosyncratic modelerror. The estimate for β – the change in turnout for treated units, relative to control – is our difference-indifferences estimator.We also measure treatment as a continuous function of logged distance from the sites. We log distance11because we expect that the effect of proximity decays exponentially; i.e., marginal increases in distance fromthe sites have less influence on the treatment the further away from the sites. That is, we estimate thefollowing:Turnoutit = αi + γt + β[log(di) · γt] + itWe use a similar difference-in-differences approach to estimate the effects on vote share. In precinct p,election t, the vote share received by the incumbent is modeled as follows:Incumbent Vote Sharept = αp + γp + β[I(dp < 1 mile) · γt] + Ω[Λi · γt] + where αp represents a precinct fixed effect, γp represents an election fixed effect (i.e., the swing to MayorWalsh from 2013 to 2017 when d is zero), and β captures the extent to which the change in incumbent voteshare varies between the 24 precincts that lie within one-mile of the vector of facilities and the remainingprecincts in Boston (di is the minimum distance of every pth precinct centroid to any of the facilities). Λp is avector of precinct-level controls for representing average quantities from the voter file or relevant indicatorsfor particular districts/neighborhoods (more details below) where Ω is a vector of coefficients correspondingto each control.We estimate equations for 3 election pairs: the 2010 and 2014 midterm elections, the 2012 and 2016general elections, and the 2013 and 2017 Boston mayoral elections.6 For all models, we first estimatethe specification without co-variate controls. Second, we estimate specifications that use exact matchingto compare treated and control individuals on the following characteristics: race, age, gender, income,home-ownership, and party identification. Age, gender and party identification are recorded directly in theMassachusetts voter file. Race is imputed by L2 using an imputation method based on name, neighborhood,socio-economic variables, and other demographic information. Income matching uses the average censusblock group income category of the voter’s residence. We combined the voterfile data with property datafrom the city of Boston to identify if a voter is homeowner. The exact-matching algorithm matches eachtreated unit to all possible control units with exactly equal co-variate values (age is coarsened into 10-yearbins), creating subclasses in which all treated and control units have the same covariate values. We thenuse the size of each subclass as weights in a Weighted Least Squares regression analysis while includingthe matching variables as covariates. Standard errors are clustered at the individual level in all turnout6We do not analyze changes in voting between city council local elections in non-mayoral election years due to variation across citycouncil districts in challenges to incumbent city councillors. We also do not estimate changes in vote share between the 2010 and 2014midterm elections because Boston’s electoral precincts were changed in 2011 following the decennial census.12specifications.7We estimate a third specification to account for potential confounding from different municipal candidates running for election in 2013 compared to 2017. In 2013, Marty Walsh’s opponent, John Connolly, wasan at-large City Councillor from West Roxbury, and turnout in that neighborhood was much higher in 2013than 2017. In 2017, Mayor Walsh ran against Tito Jackson, the city councillor from District 7, which coversparts of the Roxbury, Dorchester, South End, and Fenway neighborhoods of Boston. Changes in voter turnoutbetween 2013 and 2017 may be a function of different candidates running, and since the candidates receivedisproportionate shares of their home district’s votes, a Jackson or Connolly effect may appear as an effectof proximity to the Long Island re-location sites. Therefore, we estimate a specification that matches on thesame covariates as in the previous specification and includes “fixed-effects” for Boston’s city council districtsand the West Roxbury neighborhood (which is not coterminous with any city council district). These fixedeffects represent the average trend in turnout for these places and account for changes in voting behaviordriven by the presence of neighborhood locals on the mayoral ballot in one election but not the other, aswell as competitiveness of ward city council races concurrent with the mayoral election. This makes thecomparison for these places “within-neighborhood,” a conservative approach that rules out the possibilitythat any of the changes in turnout between council districts and between West Roxbury and the rest of thecity of Boston are due to distance from the relocation sites.8Causal inference in the difference-in-differences setup hinges on the parallel trends assumption thatchanges in political behavior between residents located near the relocation sites, and residents of matchedgender, income, age, party identification, homeownership, and race who live further away, would have beensimilar in expectation had the Long Island Shock not occurred. In the Appendix (Figure A1), we examinepre-trends in turnout, finding that turnout rates moved in parallel between treatment and control areas priorto the shock.ResultsTo further buttress our claim that the Long Island Shock caused an abrupt increase in disorder, and to providesome initial evidence that the change in disorder altered daily interactions with government, we first present7Our pre-analysis plan specified bootstrap standard errors. Results are not substantially or substantively different using eithermethod, but the clustered standard errors tend to be larger than the bootstrapped ones, so we chose to report the more conservativeclustered standard errors. We include bootstrapped standard errors in the Appendix Table A8.8Including fixed effects for all city council districts, instead of a single dummy for Tito Jackson’s district, deviates from our preregistration. After receiving feedback about differential levels of competitiveness in ward city council races concurrent with the mayoralrace, we were persuaded to use council district fixed effects. Results are substantively similar in sign, magnitude, and significance.13changes in requests for government services related to disorder. In the areas surrounding the relocationfacilities, there was an uptick in 311 non-emergency requests for government services plausibly related tohomelessness.9 As Figure 5 illustrates, the frequency of weekly service requests within one-mile of therelocation sites increased both relative to the rest of Boston and in absolute terms in the months followingthe October 2014 bridge closure. As these requests are “by definition a measure of the service demandsthat neighborhoods place on city governments” (White & Trump 2018), their increase is consistent withburdensome neighborhood changes. But 311 calls have at least two kinds of measurement error: conceptual(311 calls in these categories may not exactly map onto to disorder) and false-positive (311 calls can bemade without basis). To mollify these concerns, we performed a parallel analysis of 911 calls, focusingon 911 calls for overdoses and “emotionally disturbed persons” (police terminology for calls responsive toerratic behavior and/or mental health emergencies) which we obtained by public records request. Theseare less susceptible to measurement error because they are more specific and because fraudulent calls areillegal. Weekly 911 calls for drug poisonings and emotionally disturbed persons in the vicinity of relocationsites increased in absolute terms and relative to the rest of Boston (Figure 5). A negative binomial regressionfinds significant evidence of a larger discontinuous jump within a mile of relocation sites for both 311 calls(p < 0:01) and 911 calls (p < 0:01) after the Long Island Shock (Appendix Table A3).These changes may reflect both changes in neighborhood conditions and deliberate political mobilization.Consistent with this analysis, various neighborhood association minutes encourage residents to “call 911when the need arises. . . it will increase statistics” (CSNA 2015) and “get dots on the map” (BFSNA 2016).Boston formed a special unit to respond to the high frequency of 911 overdose calls in the area (Atkinson2017).Having shown these changes in immediate political behavior, we next present results for the effect ofthe Long Island Bridge closing on turnout for voters living close to the re-location sites. Figure 6 presentsthe effect of the closing on the likelihood of voting in the next mayoral election with the treatment groupdefined as living within increasing distances from the re-locations sites. The number above each pointestimate indicates the number of registered voters living within that distance – the treated sample. The estimates demonstrate that the neighborhood change following the Long Island Bridge closing caused increasedturnout among proximate residents in the 2017 mayoral election relative to residents living farther away9Plausibly homelessness related 311 calls were identified by string searching the open-ended CLOSURE REASON field in the datafor the string “homeless” and using only calls in the set of call types corresponding to these explicitly designated homelessness calls.This process captured 8 call types out of the data’s total 216 call types, and subsets the data from about 1.3 million service requests to160,000 service requests. Subsetting the data in this way helps deal with clearly unrelated shocks to 311 calls, such as periods of heavysnowfall. The 8 call types are “Requests for Street Cleaning”, “Ground Maintenance”, “Poor Conditions of Property”, “CE Collection”,“Illegal Dumping”, “General Comments For a Program or Policy”, “Housing Discrimination Intake Form”, and “Highway Maintenance”.14Figure 5: Spike in weekly 311 Calls and 911 Calls following Long Island ClosurePlots show predicted weekly counts of 311 calls (top) and 911 calls (bottom), with 95% confidence intervals, fromnegative binomial regressions with month fixed effects and linear and quadratic trends interacted with the date of thebridge closure. Points are connected by a Loess fit.from the sites. The first plot displays the point estimates and 95% confidence intervals for the models without covariates. The second plot presents the estimates for our matching specifications. The third plot showsthe point estimates for the specifications with both matching and separate trends by city council district andin West Roxbury. Except for the results with treatment defined as living within 300 meters of the sites, wherethe treated sample size is lowest, we find consistently significant (p < :05) positive effects on turnout acrosstreatment definitions. Effect sizes range from approximately 5 to 9 percentage points, initially decreasing astreatment is defined by greater distances, stabilizing by about 800 meters.Decreasing effects on turnout further away from the relocation sites suggest that the effect of neighborhood disorder increases with proximity. To formally test this hypothesis, we estimate models measuring theeffect of treatment as a continuous function of logged distance from the sites. Table 2 reports the resultsfrom 3 specifications: without covariates, with covariates, and with covariates and the time trend interacted15with council district fixed effects and the West Roxbury neighborhood. Across models, voters are less likelyto vote in 2017 compared to 2013 the greater the distance they live from the sites.10 The effect sizes aresubstantially similar across models.11 These results provide evidence in support of Hypothesis 3, that theeffect on turnout is a function of distance to the sites.While we observe effects of increased neighborhood disorder on mayoral election voting, we do notsee such effects in statewide and national elections. Figure 7 shows the effect of proximity on changes inpresidential turnout between 2016 and 2012 (top) and midterm voting between 2014 and 2010 (bottom)without any controls. The presidential election results are generally indistinguishable from zero, negativeat lower distances but positive at greater distances. None of the midterm election estimates are statisticallysignificant and hover around zero. The midterm and presidential specifications measuring the effect oftreatment as a continuous function of logged distance from the sites also show small (no demographiccontrols) or null (with demographic controls) effects of distance on turnout. These results are shown in theOnline Appendix.10The 2017 coefficient is positive, but this is conditional on the distance being at zero (and the covariates being at their base categories,where applicable). Generally, mayoral turnout was lower in 2017 than 2013.11These estimates and the estimates in the tests of hypothesis 1 are robust to a Bonferroni multiple comparisons test16Figure 6: Mayoral Turnout, 2017 vs 2013Raw AssociationDistrict FEs and West Roxbury trendPoint estimates represent change in turnout (for treatment relative to control) under the definition of treatment asliving within the distance to the relocation sites indicated on the x-axis. Error bars are 95% confidence intervals.Numbers above each estimate indicate the number of treated voters.Taken together, these results show that voters mobilized politically in response to increased neighborhooddisorder, but that this response was limited to local elections. The 5 to 9 percentage point effect on mayoralturnout is striking against baseline levels in local elections. In 2013, 38% of registered voters in the city ofBoston voted in the mayoral election, and just 27.8% voted in the 2017 election.The political effects were not limited to participation. Contrary to Hypothesis 2, proximate voters werealso more likely to support the incumbent Mayor, other things equal. Across various council-district fixed17Table 2: Turnout Change as function of log distanceDependent variable:Mayoral Turnout, 2017 v 2013(1) (2) (3)2017 –0.126∗∗∗ 0.318∗∗ 0.338∗∗(0.003) (0.139) (0.133)2017 · log(Distance) –0.036∗∗∗ –0.026∗∗∗ –0.032∗∗∗(0.002) (0.002) (0.004)2017 · West Roxbury –0.052∗∗∗(0.009)Covariates N Y YCouncil District Trends N N YObservations 430,444 430,444 430,444R2 0.747 0.752 0.753Adjusted R2 0.494 0.503 0.505Residual Std. Error 0.355 (df = 215220) 0.352 (df = 215180) 0.351 (df = 215171)Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01Table shows the coefficient estimates for the change in turnout as a function of the logged distance voters live from therelocation sites. Standard errors clustered at the individual level are shown in parentheses. Matching covariates inmodels 2 and 3 include voter race, age, gender, party identification, homeownership status, and the median income ofthe census block in which they live.18Figure 7: Effect of Proximity on Turnout at other scales (raw associations)Presidential Turnout, 2016 vs 2012Midterm election turnout, 2014 vs 2010Point estimates represent change in turnout (for treatment relative to control) under the definition of treatment asliving within the distance indicated on the x-axis to the relocation sites. Error bars are 95% confidence intervals.Numbers above each estimate indicate the number of treated voters.effect specifications12 (Table 3), the swing towards incumbent Marty Walsh was more than 5 percentagepoints larger in precincts within one-mile of the relocation sites, including when adjusting for demographiccovariates (Model 2) and adding an indicator for the West Roxbury neighborhood, home of Walsh’s previous12The main specifications for vote choice differ from those in our pre-registration. A detailed explanation is provided in the Appendix.19Figure 8: Model 3: Swing to Walsh in Nearby PrecinctsBar plots represent the change in Walsh vote share from 2013 to 2017 for Boston electoral precincts within 1 mile ofthe relocation sites (bottom) and beyond 1 mile (top). Error bars are 95% confidence intervals.(2013) opponent (Model 3). To give a sense of the effect-size, Figure 8 illustrates the change in Walsh Voteshare at precincts near and far away from the relocation sites and the uncertainty of these estimates usingthe parameters from Model 3; precincts whose geographic centers fell within one-mile of the relocation sitesexperienced about a 15 percentage point swing to Walsh, while those further away experienced about a 10percentage point swing (his citywide swing was 13.8%). No such incumbent swing was observed betweenthe 2016 and 2012 Presidential races (Appendix Table A15).1313Precinct boundaries changed before the 2012 election, precluding a similar analysis of 2014 and 2010 vote share.20Table 3: Mayoral Vote Share, 2017 vs 2013Dependent variable:Walsh Swing(1) (2) (3)< 1 Mile 8.698∗∗ 6.771∗∗∗ 5.587∗∗∗(3.558) (2.322) (1.864)White % –0.057 –0.150∗(0.106) (0.086)Hispanic % –0.456∗∗∗ –0.361∗∗∗(0.116) (0.094)Mean Age 1.414∗∗∗ 0.724∗∗∗(0.176) (0.153)Low Income % 0.112 0.049(0.164) (0.131)High Income % 1.888∗∗∗ 2.683∗∗∗(0.334) (0.277)Black % –0.366∗∗∗ –0.400∗∗∗(0.093) (0.075)West Roxbury Dummy 31.428∗∗∗(2.728)Council FE Yes Yes YesObservations 253 253 253R2 0.519 0.814 0.881Adjusted R2 0.501 0.803 0.873Residual Std. Error 14.349 (df = 243) 9.027 (df = 237) 7.237 (df = 236)F Statistic 29.120∗∗∗ (df = 9; 243) 69.275∗∗∗ (df = 15; 237) 109.349∗∗∗ (df = 16; 236)Note: ∗p<0.1; ∗∗p<0.05; ∗∗∗p<0.01Table shows coefficient estimates for the change in support for Mayor Walsh (for treatment relative to control) from2013 to 2017 for Boston electoral precincts. Treatment is defined as precincts within 1 mile of the relocation sites.Covariates are calculated at the precinct level from the voterfile data. Standard errors clustered at the precinct level areshown in parentheses.21DiscussionNeighborhood Disorder & Local Political ParticipationThese results indicate that a shock to neighborhood disorder leads to a substantively large increase in political participation. Residents living in neighborhoods proximate to the relocation sites responded to increasedneighborhood disorder by voting at a higher rate in the subsequent Mayoral election than voters living further from the sites (Hypothesis 1).14 Moreover, the effect sizes were largest for voters who lived closestto the sources of neighborhood disorder (Hypothesis 3), consistent with geographic proximity capturing achange in everyday life on the street for residents that modulates the strength of the political response. Building on important work establishing the association between self-reported neighborhood disorder, attendingcommunity-meetings, and contacting public officials (Michener 2013), our study uses a compelling researchdesign (Dunning 2010) to show how disorder affects voting behavior and everyday daily participation. Assuch, we have demonstrated how an important and salient dimension of social geography drives politicalbehavior. While our quantitative analyses illustrate the impact on electoral politics, the political response tothe Long Island Bridge closure was not limited to voting or requesting government services. Exposure to theaffected areas has led residents and others to introduce legislation, hold protests, oppose the construction ofnew social services, attend city council hearings, and contact the mayor and other local officials to demandthat the city government curb disorder, as detailed in the Appendix. In 2018, 900 people wrote letters insupport of rebuilding the Long Island Bridge in an effort coordinated by the South End Forum, a group ofneighborhood associations (Seth 2018a). This multifaceted political response shows how neighborhood disorder galvanizes voters to participate in politics on both intensive margins (more intense participation) aswell as extensive margins (more people participating).Another important finding is that increased neighborhood disorder influences participation and votechoice only in local elections. Proximity to the relocation sites did not structure turnout or vote choice innational and statewide elections. Thus, voters responded to local events – the decision to relocate city facilities for homelessness and addiction treatment – by participating in local politics and holding local politiciansaccountable. This contrasts with voter “attribution errors” posited in other contexts (Sances 2017, Achen &Bartels 2016). Instead of a mechanical anti-incumbent backlash across levels of government, the electoralresponse was nuanced, consistent with voters of sufficient rationality to identify the proper scale at which todirect action and the proper politicians to hold accountable (Larsen 2018). The power to curb neighborhood14In the Appendix, we detail analyses of heterogeneous treatment effects and placebo tests which explore and ultimately rule outracial threat as an alternative explanation. We also present results from a robustness check using randomization inference (comparingthe observed result with thousands of hypothetical alternative relocation sites).22disorder in these Boston neighborhoods lies with the municipal government, which is responsible for localpolicing, litter control, public works and park maintenance and operates the Boston Public Health Commission and Boston Emergency Shelter Commission, which control the city’s public homeless shelters, syringedistribution program, and drug treatment services.In response to neighborhood problems such as increased disorder, existing residents can respond by passive acceptance, political action, or leaving the area (Orbell & Uno 1972). What are the conditions underwhich they choose political action? Information provision may be an important mediator between neighborhood change and voter decision-making (Hopkins 2010, de Benedictis-Kessner 2018). While neighborhoodchange often provokes NIMBY attitudes (Hankinson 2018), voters require sufficient knowledge to infer thatundesirable neighborhood changes can be mitigated through increased political participation (Arceneaux& Stein 2006, Gasper & Reeves 2011). Our empirical design cannot disentangle the precise mechanismsbehind the increased participation but, following this logic and our review of the case, we posit that localmedia, place-based neighborhood associations and business groups, and politicians informed voters aboutthe connections between increased neighborhood disorder and the closure of the Long Island Bridge, on theone hand, and the role of government in ameliorating disorder on the other hand. Indeed, the creation ofsocial capital to coordinate responses to disorder may be an important intermediate step in the effects onvoting behavior. Consistent with this mechanism, analyses detailed in the Appendix use the Boston Neighborhood Survey to show that respondents who perceive more disorder in their communities are more likelyto participate in community meetings.When do Negative Shocks Benefit the Incumbent?Why did voters closer to the re-location sites choose to reward Mayor Walsh? And how do these proincumbent results cohere with the increases in political participation? We argue that the Long Island Shockraised the salience of addiction and homelessness for voters near the relocation sites, who learned the roletheir government played in managing these issues in part as a consequence of this shock (de BenedictisKessner 2018). This increased salience resulted in higher levels of local participation that advantaged candidates like Walsh who were perceived to be qualified on these issues. Negative shocks provide insight into anincumbent’s quality, which they can contrast with challenger quality (Ashworth et al. 2018). In the naturaldisaster literature, an effective incumbent response has been shown to produce electoral benefits (Healy &Malhotra 2009) in part by mobilizing supporters to turnout to vote (Chen 2013). And Walsh appears to havemanaged the Long Island Shock adequately, replacing treatment beds and responding to local complaintswith additional police patrols, needle pickups (City of Boston 2015), street outreach to drug users (Becker232017a), beautification efforts (Allen 2016, Kaeslin 2017), and providing an air-conditioned “day-center”(Marcelo 2017) to keep “some of the chaos off the street” (Rasmussen & Alulema 2018). On the broaderopioid epidemic, Walsh added treatment referral services to the city’s 311 hotline, boosted appropriationsfor addiction treatment, expanded housing services to address chronic homelessness, and formed the onlymunicipal Office of Recovery Services in the US (Becker 2016, Marcelo 2017).Some anecdotal evidence reinforces the view that voters rewarded Walsh for effectively managing thebad hand he was dealt (Ashworth et al. 2018). Jonathan Scott, head of a drug rehabilitation organization,said the Mayor “has my full support and endorsement… the fact that we [replaced the beds lost on LongIsland] in 2 and a half years was nothing short of a miracle” (Freyer 2017). Non-experts affected by theLong Island Shock emphasize that the Mayor has done a good job, even though their situation is bad. ACEO of a local trucking company explained that while conditions have worsened, she did not blame Walsh:“The mayor has tried very, very hard, and I think he has done a good job but I think what they need to dois open the island again…” (Becker 2017b). Moreover, Walsh may have benefited from prior credibility onthe topics of addiction and homelessness that colored how voters perceived his response (van Zuylen-Wood2016, Becker 2017b). Walsh had himself received treatment on Long Island and spoke about recovery fromaddiction in his 2013 Mayoral bid and the 2016 Democratic National Convention. This credibility has helpedbuffer Walsh against criticism of his management of the Long Island Shock. John McGahan, president of alocal addiction treatment organization, said “When I hear people throwing rocks at him on this issue, itdisturbs me, because I know how much he cares” (Becker 2017b). In contrast, the City Councillor whochallenged Walsh previously worked as a representative for an opioid drug, undermining his campaign’semphasis on the opioid crisis (O’Sullivan 2017). Therefore, voters likely saw Walsh as better equipped toaddress drug problems, a more important issue for voters in neighborhoods closer to the relocation.15 Thislogic comports with partisan issue ownership and issue salience (B´elanger & Meguid 2008), though at playhere is candidate issue ownership.LimitationsThere are several limitations to this study. One challenge to this paper and to studying disorder more generally is difficulty in measurement. Because our research design requires demonstrating that closure of theLong Island Bridge substantially increased disorder at the relocation sites, we have presented multiple formsof evidence consistent with an increase in disorder, including descriptions from local media, neighborhood15In an interesting contrasting case, an emergency room doctor at Boston Medical Center unseated the area’s longtime state representative in a 2018 primary campaign focusing on his expertise in managing the district’s drug problems (Capelouto 2018, Boston GlobeEditorial Board 2018).24association meeting minutes, reports from the city’s public health agency and city council, official crimestatistics, images from Google Street View, and original statistical analyses of non-emergency 311 calls andemergency 911 calls. Any one of these changes provides modest evidence of changes in disorder. But triangulating between these various forms of evidence suggests that closure of the Long Island Bridge causedsignificant neighborhood changes and was a salient event for the neighborhood’s existing residents. However, what is less clear is how to quantify this increase in disorder to make comparisons between this caseand others, an important limitation. A second limitation concerns the case we study, an abrupt but persistent“shock” to disorder. Such shocks may operate differently from a more gradual neighborhood change (e.g.they be more salient), limiting the external validity of our inferences. That said, abrupt shocks to disorderoccur with some frequency, usually tracking the dislocation of encampments of people who are homeless oruse drugs (Orange 2013, Whelan 2018). A third limitation involves disentangling the political effects of achange in disorder from the the government’s response to disorder (the presence of the Mayor and other cityofficials in the area and increased investment in government services like policing or needle pick-ups). This isnot a unique problem for our study, as the political effects of other kinds of shocks (such as natural disasters)similarly depend on the government’s response (Healy & Malhotra 2009, 2010, Chen 2013). This issue ismost acute for the incumbent vote share effects–it seems implausible that incumbent vote share would haveincreased near the relocation sites absent substantial efforts by the Mayor’s office to mitigate disorder. Withregards to participation, the presence of increases in non-voting forms of mobilization (neighborhood association organizing, protests, requests for government services) suggest that at least some of the mobilizationresponse is due to changes in disorder itself, not solely to the actions of government officials, because theseactions were temporally prior to any government response. We suspect that this extends partially to the voterturnout results, but admit that this is a tentative extrapolation, and we cannot fully decompose the turnouteffect in this case. Still, our study provides a compelling glimpse into the political implications of disorder,an important and understudied phenomenon connected to ongoing drug and homelessness crises in NorthAmerica. We hope future research advances our understanding of disorder by addressing these limitations.ConclusionThis study shows that increased neighborhood disorder causes increased participation across a range of political behaviors. Voters living near the re-location sites were more likely to request government services,vote in the next mayoral election, and more likely to vote for the incumbent Mayor. They also engaged inless systematic behaviors including contacting elected officials, protesting (Irons 2017, Maniscalco 2017),organizing to restrict the construction of new social services (CSNA 2016, Seth 2018b, Daniel 2018), sub-25mitting letters in support of rebuilding the Long Island Bridge (Seth 2018a), and attending neighborhoodassociation meetings (Daniel 2016). But the level of government mattered. While we observed widespreadincreases in local political participation, we did not find similar effects in national and state elections. Theresponse was limited to the relevant, local scale. This contrast with “attribution errors” posited in othercontexts motivates inquiry into the mechanisms by which neighborhood change provokes political action,including the informational roles of local elites and neighborhood associations.Closure of the Long Island Bridge – and consequent relocation of homeless individuals, people with addiction, and social services to mainland Boston – provides insight into the political consequences of abruptincreases in neighborhood disorder. Boston is one of many cities dealing with highly concentrated – andpoliticized – homelessness and drug use (Kim 2017), and thus these results speak to a broader phenomenonof growing disorder in the shadow of addiction. Sue Sullivan of Boston’s Newmarket Square Business Association pointed out that “It’s not just here. Look at San Francisco. I have my interns looking at everymajor city, and nobody knows how to deal with this” (Sperance 2018).16 Indeed, growing problems withunsheltered homelessness have prompted declarations of “homelessness states of emergency” in major citieslike Los Angeles, Portland, Sacramento, San Diego, and Seattle, along with the state of Hawaii and dozensof other municipal and county governments. Our results suggest that exposure to these crises will increaseparticipation in local politics, but the electoral effect on incumbents will depend on the quality of their responses and of their challengers. The normative implications of these positive claims are ambiguous; bywidening the participatory gap, increased participation among the less marginalized group could produceless equitable policymaking (Sances 2016).Our discussion has focused on the political response of stably-housed residents near the re-location sites.But it seems appropriate to give the final words to the Long Island Refugees, the already-vulnerable peopledisrupted by the termination of the island’s facilities when its long-neglected bridge failed. We hope ourresearch invites scholarship into the geographic factors that structure the services available for vulnerablepeople, such as proximity to existing neighborhoods and whether treatment services concentrate or diffusevulnerable people. This has clear policy relevance. Federal law requires patients treated with methadone—the gold standard therapy for opioid addiction—make frequent (typically daily) visits to a specialized clinic,a unique burden that both clusters vulnerable people (potentially provoking NIMBY opposition) and limitsaccess (Samet et al. 2018, de Benedictis-Kessner & Hankinson 2019). Despite these barriers, efforts to sitenew services can be successful (Lupick 2018), and future research could assess the conditions under whichthey succeed. The continued suffering on Boston’s “Methadone Mile” and beyond adds urgency to such work.16In a 2016 survey of candidates in the San Francisco Bay region, the overwhelming majority believed easing homelessness was theregion’s top priority (Villa & Noori 2016).26ReferencesAbel, D. (2015), ‘Crime rises after shelter arrives in newmarket area’, Boston Globe . Accessed: 2018-08-21.Achen, C. & Bartels, L. (2016), Democracy for Realists: Why Elections Do Not Produce Responsive Government,Princeton Studies in Political Behavior, Princeton University Press.Allen, E. (2016), ‘Public art, cleaner streets coming to boston thanks to partnership’, Boston Globe . Accessed:2018-12-17.Anoll, A. P. (2018), ‘What makes a good neighbor? race, place, and norms of political participation’, AmericanPolitical Science Review 112(3), 494–508.Arceneaux, K. & Stein, R. M. (2006), ‘Who is held responsible when disaster strikes? the attribution ofresponsibility for a natural disaster in an urban election’, Journal of Urban Affairs 28(1), 43–53.Ashworth, S., Bueno de Mesquita, E. & Friedenberg, A. (2018), ‘Learning about voter rationality’, AmericanJournal of Political Science 62(1), 37–54.Atkinson, D. (2017), ‘Team will target ods, minor emergencies’, Boston Herald . Accessed: 2018-08-21.Bateson, R. (2012), ‘Crime victimization and political participation’, American Political Science Review106(3), 570–587.Becker, D. (2016), ‘State’s opioid epidemic is vividly seen on boston’s ’methadone mile”, WBUR . Accessed:2018-08-30.Becker, D. (2017a), ‘How walsh’s opioid outreach program is working, 6 months in’, WBUR . Accessed:2018-08-28.Becker, D. (2017b), ‘In mayoral race, disagreement over boston’s approach to tackling opioid crisis’, WBUR .Accessed: 2018-08-21.Becker, D. & Jolicoeur, L. (2014), ‘Homeless displaced by long island bridge closure remain in flux’, WBUR .Accessed: 2018-08-15.B´elanger, E. & Meguid, B. M. (2008), ‘Issue salience, issue ownership, and issue-based vote choice’, ´ ElectoralStudies 27(3), 477–491.BFSNA (2016), ‘General meeting minutes’. Accessed: 2018-09-28.URL: https://static1.squarespace.com/static/57b884e0d2b857c2cf2dac69/t/57bd8c528419c2e71779a3be/1472040020912/03-15+General+Meeting+Minutes.pdf27BHCHP (2014), ‘Long island bridge closure: Impact and recommendations’.URL: https://www.bhchp.org/sites/default/files/field/image/LIS%20Fact%20Sheet 0.pdfBillings, R. (2018), ‘Bayside at rock bottom: Portland neighborhood is under siege’, Portland Press Herald .Accessed: 2019-05-28.Billings, R. (2019), ‘Hostile crowd criticizes portland proposal to build emergency shelter at barron center’,Press Herald . Accessed: 2019-04-30.Bittle, J. (2017), ‘No room at the holiday inn’, Village Voice . Accessed: 2019-05-27.Boston Globe Editorial Board (2018), ‘Endorsement: Liz malia, jeffrey sanchez, and jon santiago for mass.state representatives’. Accessed: 2018-09-13.Burrell, C. (2014), ‘Long island bridge closed; homeless forced to leave shelter’, Patriot Ledger .Capelouto, J. (2018), ‘Two longtime state representatives ousted in primary upsets’, Boston Globe . Accessed:2018-09-13.Chen, J. (2013), ‘Voter partisanship and the effect of distributive spending on political participation’, American Journal of Political Science 57(1), 200–217.City of Boston (2015), ‘Mayor walsh deploys crews to clear boston’s public spaces of hazardous, discardedneedles’. Accessed: 2018-08-28.URL: https://www.cityofboston.gov/news/default.aspx?id=20186Clifford, S. & Piston, S. (2017), ‘Explaining public support for counterproductive homelessness policy: Therole of disgust’, Political Behavior 39(2), 503–525.CSNA (2015), ‘Report: Homeless methadone clients’. Accessed: 2018-09-25.
URL:minutes/
https://web.archive.org/web/20180915224647/https://chestersquareneighbors.org/meeting
CSNA (2016), ‘Notes from march 2, 2016 meeting’. Accessed: 2018-09-29.URL: https://web.archive.org/web/20180929212823/https://chestersquareneighbors.org/2016/04/17/notesfrom-march-2-2016-meeting/Dahl, R. A. (1961), Who governs?: Democracy and power in an American city, Yale University Press.Daniel, S. (2016), ‘Methadone mile neighbors look for security’, Boston Sun . Accessed: 2018-08-16.28Daniel, S. (2018), ‘Wsana talks about sifs, state legislation and marijuana’, Boston Sun . Accessed: 2018-09-29.de Benedictis-Kessner, J. (2018), ‘How attribution inhibits accountability: Evidence from train delays’, TheJournal of Politics 80(4), 1417–1422.de Benedictis-Kessner, J. & Hankinson, M. (2019), ‘Concentrated burdens: How self-interest and partisanshipshape opinion on opioid treatment policy’, American Political Science Review pp. 1–7.Dunning, T. (2010), ‘Design-based inference: beyond the pitfalls of regression analysis?’, Rethinking socialinquiry: Diverse tools, shared standards pp. 273–311.Enos, R. D. (2014), ‘The causal effect of intergroup contact on exclusionary attitudes’, Proceedings of theNational Academy of Sciences of the United States of America 111(10), 3699–3704.Enos, R. D. (2015), ‘What the demolition of public housing teaches us about the impact of racial threat onpolitical behavior’, American Journal of Political Science .Enos, R. D. (2017), The space between us: Social geography and politics, Cambridge University Press.Faraji, S.-L., Ridgeway, G. & Wu, Y. (2018), ‘Effect of emergency winter homeless shelters on property crime’,Journal of Experimental Criminology 14(2), 129–140.Fiorina, M. (1981), Retrospective Voting in American National Elections, Yale University Press.Fowler, A. & Hall, A. B. (2018), ‘Do shark attacks influence presidential elections? reassessing a prominentfinding on voter competence’, The Journal of Politics 0(0), 000–000.Freyer, F. (2017), ‘Mixed reviews for walsh on handling of opioid crisis’, Boston Globe . Accessed: 2018-08-21.Gasper, J. T. & Reeves, A. (2011), ‘Make it rain? retrospection and the attentive electorate in the context ofnatural disasters’, American Journal of Political Science 55(2), 340–355.Gracia, E. (2014), Neighborhood Disorder, Springer Netherlands, Dordrecht, pp. 4325–4328.URL: https://doi.org/10.1007/978-94-007-0753-52751Halverstadt, L. (2018), ‘Faulconer’s homelessness record marked by reluctance, struggles and a few successes’, Voice of San Diego . Accessed: 2019-05-27.Hankinson, M. (2018), ‘When do renters behave like homeowners? high rent, price anxiety, and nimbyism’,American Political Science Review 112(3), 473–493.29Hansen, J. M. (1985), ‘The political economy of group membership’, American Political Science Review79(1), 79–96.Healy, A. & Malhotra, N. (2009), ‘Myopic voters and natural disaster policy’, American Political Science Review103(3), 387–406.Healy, A. & Malhotra, N. (2010), ‘Random events, economic losses, and retrospective voting: Implicationsfor democratic competence’, Quarterly Journal of Political Science 5(2), 193–208.Healy, A. & Malhotra, N. (2013), ‘Retrospective voting reconsidered’, Annual Review of Political Science16(1), 285–306.Hinkle, J. C. & Weisburd, D. (2008), ‘The irony of broken windows policing: A micro-place study of therelationship between disorder, focused police crackdowns and fear of crime’, Journal of Criminal justice36(6), 503–512.Hjorth, F. (2017), ‘The influence of local ethnic diversity on group-centric crime attitudes’, British Journal ofPolitical Science p. 1–23.Holland, G. (2018), ‘Although l.a. mayor calls latest crackdown on homeless camps a success, legal issuescloud city’s plans’, Los Angeles Times . Accessed: 2019-05-27.Hopkins, D. (2018), The Increasingly United States: How and Why American Political Behavior Nationalized,Chicago Studies in American Politics, University of Chicago Press.Hopkins, D. J. (2010), ‘Politicized places: Explaining where and when immigrants provoke local opposition’,American Political Science Review 104(1), 40–60.Hopkins, D. J. & Pettingill, L. M. (2018), ‘Retrospective voting in big-city us mayoral elections’, PoliticalScience Research and Methods 6(4), 697–714.Innes, S. (2019), ‘Desperate to give up opioids, patients flock to 24/7 methadone clinic in phoenix’, ArizonaRepublic . Accessed: 2019-04-30.Irons, M. (2017), ‘Students keep finding needles near their school. they want action from the city.’, BostonGlobe . Accessed: 2018-09-20.Irons, M. & Dungca, N. (2014), ‘Abrupt closing of long island bridge was slow to arrive’, Boston Globe .Accessed: 2018-09-16.30Jonas, M. (2019), ‘Breaking down “operation clean sweep”’. Accessed: 2019-09-02.URL: https://commonwealthmagazine.org/housing/breaking-down-operation-clean-sweep/Kaeslin, T. (2017), ‘Coalition makes progress revitalizing medical campus neighborhood’, Daily Free Press .Accessed: 2018-12-17.Karedes, D. (2018), ‘Safety concerns mount after woman in her 70s is stabbed in south end’, Boston 25 News. Accessed: 2018-09-17.Kelling, G. L. & Coles, C. M. (1997), Fixing broken windows: Restoring order and reducing crime in ourcommunities, Simon and Schuster.Key, V. (1949), Southern Politics in State and Nation, Knopf, New York.Kim, V. (2017), ‘The struggle to clean up methadone mile, camp heroin, skid row’, The Fix . Accessed:2018-10-9.Knight, H. (2015), ‘Time for non-homeless people to rant and rave’, San Francisco Chronicle . Accessed:2019-05-27.Larsen, M. V. (2018), ‘Is the relationship between political responsibility and electoral accountability causal,adaptive and policy-specific?’, Political Behavior pp. 1–28.Larsen, M. V., Hjorth, F., Dinesen, P. T. & Sønderskov, K. M. (2019), ‘When do citizens respond politicallyto the local economy? evidence from registry data on local housing markets’, American Political ScienceReview pp. 1–18.Latkin, C. A. & Curry, A. D. (2003), ‘Stressful neighborhoods and depression: a prospective study of theimpact of neighborhood disorder’, Journal of health and social behavior pp. 34–44.Leighley, J. (1996), ‘Group membership and the mobilization of political participation’, The Journal of Politics58(2), 447–463.Lupick, T. (2018), Fighting for Space: How a Group of Drug Users Transformed One City’s Struggle withAddiction, Arsenal Pulp Press.Malas, N. & Lazo, A. (2018), ‘Los angeles mayor’s political future tied to plan to solve city’s homeless crisis’,Wall Street Journal . Accessed: 2019-05-27.Malhotra, N. & Margalit, Y. (2014), ‘Expectation setting and retrospective voting’, The Journal of Politics76(4), 1000–1016.31Maniscalco, M. (2017), ‘Joys and worries’. Accessed: 2018-09-22.URL: http://www.mysouthend.com/news///218552Marcelo, P. (2017), ‘Cleaning up ‘methadone mile’ and other drug havens’, Boston.com . Accessed: 2018-08-21.McKellar, K. (2017), ‘All 3 new homeless shelters now under construction after slow start in south salt lake’,Desert News Utah . Accessed: 2019-04-30.Melendez, L. (2018), ‘Mayor london breed says she’ll be ’aggressive’ in cleaning san francisco streets’, ABC7News . Accessed: 2018-12-17.Michener, J. (2013), ‘Neighborhood disorder and local participation: Examining the political relevance of“broken windows”’, Political Behavior 35(4), 777–806.Molnar, B. E., Gortmaker, S. L., Bull, F. C. & Buka, S. L. (2004), ‘Unsafe to play? neighborhood disorder andlack of safety predict reduced physical activity among urban children and adolescents’, American journalof health promotion 18(5), 378–386.Openshaw, S. (1983), Concepts and Techniques in Modern Geography, Geo Books, Norwich, UK, chapter TheModifiable Areal Unit Problem.Orange, R. (2013), ‘Bergen to close norway’s biggest heroin park’, TheLocal.no . Accessed: 2018-09-22.Orbell, J. M. & Uno, T. (1972), ‘A theory of neighborhood problem solving: political action vs. residentialmobility’, American Political Science Review 66(2), 471–489.O’Sullivan, J. (2017), ‘Tito jackson once worked for drug maker’, Boston Globe . Accessed: 2018-09-14.Pressley, A. (2015), ‘Responding to the impact of the abrupt closing of the long island bridge on boston’shomeless and recovery communities’. Accessed: 2018-10-12.URL: https://urlzs.com/et1XZRamirez, A. (2015a), ‘Homelessness in boston, 1 year after losing the long island shelter’, WBUR . Accessed:2018-08-15.Ramirez, A. (2015b), ‘Long island: Closure and fallout timeline’, Spare Change News . Accessed: 2018-08-21.Ramos, N. & Allen, E. (2016), ‘Life and loss on methadone mile’, Boston Globe . Accessed: 2018-09-18.Rasmussen, E. & Alulema, P. (2018), ‘25 investigates: Undercover video raises questions about city’s tent’,Boston 25 News . Accessed: 2018-12-27.32Rasmussen, E. & Smith, E. (2017), ‘25 investigates: Mayor’s fence moves addiction problem across thestreet’, Boston 25 News . Accessed: 2018-09-16.Reilly, A. (2016), ‘In boston’s newmarket square, an uneasy coexistence for businesses and the homeless’,WGBH Boston . Accessed: 2018-08-16.Riker, W. H. & Ordeshook, P. C. (1968), ‘A theory of the calculus of voting’, American political science review62(01), 25–42.Ross, C. E. & Jang, S. J. (2000), ‘Neighborhood disorder, fear, and mistrust: The buffering role of social tieswith neighbors’, American journal of community psychology 28(4), 401–420.Samet, J. H., Botticelli, M. & Bharel, M. (2018), ‘Methadone in primary care—one small step for congress,one giant leap for addiction treatment’, New England Journal of Medicine 379(1), 7–8.Sances, M. W. (2016), ‘The distributional impact of greater responsiveness: Evidence from new york towns’,The Journal of Politics 78(1), 105–119.Sances, M. W. (2017), ‘Attribution errors in federalist systems: When voters punish the president for localtax increases’, The Journal of Politics 79(4), 1286–1301.Schuessler, A. A. (2000), A logic of expressive choice, Princeton University Press.Seth, D. (2018a), ‘Mayor says long island could fill need for supportive housing’, Boston Sun . Accessed:2018-09-17.Seth, D. (2018b), ‘Medical marijuana company applies to zba for tremont street’, Boston Sun . Accessed:2018-09-17.Singh, A. (2019), ‘Campbell calls for new focus on addiction, homelessness after ’operation clean sweep”.Accessed: 2019-09-02.URL: https://www.wgbh.org/news/local-news/2019/08/20/campbell-calls-for-new-focus-on-addictionhomelessness-after-operation-clean-sweepSkogan, W. G. (1992), Disorder and decline: Crime and the spiral of decay in American neighborhoods, Univ ofCalifornia Press.Smith, D. & Smith, D. (2018), ‘Residents erupt in anger at garcetti’s town hall for planned homeless shelterin venice’, Los Angeles Times . Accessed: 2019-04-30.33Sperance, C. (2018), ‘The opioid epidemic is turning commercial buildings into deadly hazmat zones, andno one knows what to do about it’, Bisnow.com . Accessed: 2018-09-17.Steinbuch, Y. (2019), ‘San francisco residents launch gofundme to oppose homeless shelter’, NY POST .Accessed: 2019-04-30.Tobler, W. R. (1970), ‘A computer movie simulating urban growth in the detroit region’, Economic Geography46, 234–240.Tyler, C. (2018), ‘San francisco’s new ’poop patrol’ will try to find waste before you do’, KGO . Accessed:2019-05-27.van Zuylen-Wood, S. (2016), ‘Marty walsh is not tom menino’, Boston Magazine . Accessed: 2018-09-14.Villa, L. & Noori, C. (2016), ‘How the candidates view homelessness’. Accessed: 2019-05-29.URL: https://projects.sfchronicle.com/2016/candidates-homeless-survey/Vitale, A. S. (2008), City of disorder: How the quality of life campaign transformed New York politics, NYUPress.Walker, T. (2017), ‘Is compassion giving way to anger over homelessness in southern california?’, OrangeCounty Register . Accessed: 2019-05-27.Whelan, A. (2018), ‘Philadelphia’s kensington ‘under siege’ as opioid-linked homelessness soars’, Philly.com. Accessed: 2018-09-22.White, A. & Trump, K.-S. (2018), ‘The promises and pitfalls of 311 data’, Urban Affairs Review 54(4), 794–823.White, M. J. (1983), ‘The measurement of spatial segregation’, American journal of sociology 88(5), 1008–1018.Wilson, J. (1975), Thinking about crime, Basic Books (AZ).Wilson, J. Q. (1968), ‘Varieties of police behavior’, New York: Atheneum .Wing, J. (2019), ‘Tensions flare as illegal homeless camps spread in seattle’, KNKX . Accessed: 2019-05-27.34

WE OFFER THE BEST CUSTOM PAPER WRITING SERVICES. WE HAVE DONE THIS QUESTION BEFORE, WE CAN ALSO DO IT FOR YOU.
Assignment status: Already Solved By Our Experts
(USA, AUS, UK & CA PhD. Writers)
CLICK HERE TO GET A PROFESSIONAL WRITER TO WORK ON THIS PAPER AND OTHER SIMILAR PAPERS, GET A NON PLAGIARIZED PAPER FROM OUR EXPERTS

Order Your Custom Paper — Best Custom Essay Writing Services

QUALITY: 100% ORIGINAL PAPER – NO PLAGIARISM – CUSTOM PAPER

Why Choose Us?

100% non-plagiarized Papers
24/7 /365 Service Available
Affordable Prices
Any Paper, Urgency, and Subject
Will complete your papers in 6 hours
On-time Delivery
Money-back and Privacy guarantees
Unlimited Amendments upon request
Satisfaction guarantee

How It Works

Click on the “Place Your Order” tab at the top menu or “Order Now” icon at the bottom and a new page will appear with an order form to be filled.
Fill in your paper’s requirements in the “PAPER DETAILS” section.
Fill in your paper’s academic level, deadline, and the required number of pages from the drop-down menus.
Click “CREATE ACCOUNT & SIGN IN” to enter your registration details and get an account with us for record-keeping and then, click on “PROCEED TO CHECKOUT” at the bottom of the page.
From there, the payment sections will show, follow the guided payment process and your order will be available for our writing team to work on it.

About AcademicWritersBay.com

AcademicWritersBay.com is an easy-to-use and reliable service that is ready to assist you with your papers 24/7/ 365days a year. 99% of our customers are happy with their papers. Our team is efficient and will always tackle your essay needs comprehensively assuring you of excellent results. Feel free to ask them anything concerning your essay demands or Order.

AcademicWritersBay.com is a private company that offers academic support and assistance to students at all levels. Our mission is to provide proficient and high quality academic services to our highly esteemed clients. AcademicWritersBay.com is equipped with competent and proficient writers to tackle all types of your academic needs, and provide you with excellent results. Most of our writers are holders of master’s degrees or PhDs, which is an surety of excellent results to our clients. We provide assistance to students all over the world.
We provide high quality term papers, research papers, essays, proposals, theses and many others. At AcademicWritersBay.com, you can be sure of excellent grades in your assignments and final exams.