a report from Councilor E. Denise Simmons, Chair and Councilor Sumbul Siddiqui of the Housing Committee for a public hearing held on September 27, 2018 to discuss Affordable Housing Overlay District

Allachment A Opening Remarks of Councilor Simmons for Housing Committee Hearing Sept 27, 2018, 5:30 pm in Sullivan Chamber Call of the Meeting: The Housing Committee shall meet to receive an update from the Community Development Department on the Affordable Housing Overlay District, and an update on the annual Inclusionary Zoning report. The Housing Committee shall also meet the City Manager's candidate for the open Cambridge Housing Authority Board position. Good Evening, Tonight, we have a packed agenda and therefore I'II make my opening remarks brief - and I will ask that we ALL work to keep our remarks concise and to the point We have three main topics I am looking to cover at this hearing. First, we are going to meet the City Manager's nominee for the open CHA Board position. Second, the Community Development Department will brief us on the first of the new annual Inclusionary Zoning reports. Third, the CDD will provide an update on progress being made towards establishing an Affordable Housing Overlay for the city. Any of these topics could easily fill a two hour hearing, and so I will urge us all to be mindful of the fact that we may not get to cover every last detail and every last question this evening - and if that's the case, I assure you that we will carry on the discussion in future hearings if need be. So please do not feel that this is your "one bite at the apple." I also want us to be mindful of the larger time constraints we are operating under - specifically, we are on the cusp of October, and as we move deeper into Autumn, we know it will become more challenging to schedule hearings around the holiday season and around our unpredictable weather. So I want to be very intentional in how we use our time with the Housing Committee hearings in the next couple of months, and I want to put us in a position where we can send the Affordable Housing Overlay to the full Council for deliberation and voting by the end of February or early March. In order to meet this goal, ! am going to again urge my colleagues to be mindful of what we send to the CDD and the Solicitor's office for review in the coming months. If we send dozens of new initiatives to these departments to vet during the balance of the year, we are setting ourselves up to clog up the system - and we are then setting ourselves up for nothing of substance to be achieved this term. I don't think any of us wants to needlessly spin our wheels, and therefore we are all going to have to practice a little restrain and be strategic in how we are using our resources. With that being said, I would like to turn the floor over to the City Manager and the Director of the Cambridge Housing Authority to introduce Elaine DeRosa, the candidate for the vacant CHA Board Position. Gentlemen, the floor is yours. QUESTIONS TO ASK OF CHA BOARD NOMINEE: What strategies would you enact to try to promote a fairer sense of distribution of affordable housing in our Community? What actions might you take to fight against NIMBYism?

What would you say are the greatest challenges facing the CHA as an organization, and the CHA Board in particular? QUESTIONS TO ASK CDD: How is it that Utile ended up with purview over the Affordable Housing Overlay portion of this agenda? Can you summarize how this came to be? For the purposes of planning our future Committtee hearings: can you provide a general time of when the CDD's work on the Overlay legislation will be completed, and what specific steps are needed to pass it to a final vote on the Council? This will help us better understand how much time we have to work with, and what our deadlines must be. That the City Manager be and hereby is requested to direct the Community ORDERED: Development Department to provide a written timeline of what specific steps must take place in order to take a final vote on the Affordable Housing Overlay legislation, and to report back on this matter to the City Council within 14 business days. What is the impact on the Affordable Housing Overlay District from the various other Overlay Districts (such as the Flood Plains Overlay District and the Central Square Arts Overlay District) that the CDD is being asked to review? Can work on all of these items proceed without impeding the work toward completing the Affordable Housing Overlay District? What is the potential impact on the City's ability to create new housing if all of these overlay districts are enacted? Would there be a cumulative effect that would essentially create impossible conditions for new development? Or can these various overlay districts co-exist without imposing undue and untenable constraints on the City's ability to consider new development?

Miles 0.5 Fresh Pond 1096 1 - 15 0 units 16 - 50 51 - 100 101 - 200 201 - 318 Number of Affordable Units Neighborhood Boundaries No residential units in block Cambridge, Massachusetts by Census Block Number of Affordable Units into a single polygon. about the location of housing units. areas from uninhabited open space, Notes on Census block geography: Attachment B Charles River • A large Census block at Fresh Pond Reservation was divided into three • Two Census blocks that included portions of Danehy Park and adjacent Census block boundaries were altered at four locations to increase clarity • Two Census blocks comprising Washington Elms were merged together polygons to separate the open space from the housing enclaves at Neville • The Census block at Russell Field was divided into two polygons in order to put the adjacent residential properties along Clifton St into a separate polygon. Map prepared by Brendan Monroe on September 13, 2018. CDD GIS C:\Projects|Housing/AffordableNumberByBlock11x17.mxd

Charles River Memorial Do areas from uninhabited open space. properties were divided into seven polygons in order to separate developed put the adjacent residential properties along Clifton St into a separate polygon. polygons to separate the open space from the housing enclaves at Neville into a single polygon. • The Census block at Russell Field was divided into two polygons in order to Notes on Census block geography: about the location of housing units. Census block boundaries were altered at four locations to increase clarity • A large Census block at Fresh Pond Reservation was divided into three • Two Census blocks that included portions of Danchy Park and adjacent • Two Census blocks comprising Washington Elms were merged together Map prepared by Brendan Monroe on September 13, 2018. CDD GIS C:lProjectslHousing/AffordableRatioByBlock11x17.mxd Cambridge, Massachusetts by Census Block to All Housing Units Ratio of Affordable Units No residential units 0.19 - 10.0% 25.0 - 50.0% 50.0 - 75.0% 0% 75.0 - 100.0% 10.0 - 25.0% •- Neighborhood Boundaries Ratio of Affordable Housing to All Housing Fresh Pond PAy Fresh Pond

City of Cambridge September 27, 2018 utile CAMBRIDGE ENVISION Envision Cambridge City Council Housing Committee DRAFT envison.cambridgema.gov Attachment C

City of Cambridge Agenda utile Envision Cambridge • 100% Affordable Citywide Overlay Analysis DRAFT • Zoning Ideas Generated from the Envision Cambridge Process envison.cambridgema.gov

City of Cambridge • Focus groups • Walking tours • Youth workshops • Public workshops • Street team events utile • Digital engagement • Online and paper surveys Envision Cambridge • Advisory committee and working groups DRAFT a wide-reaching engagement process envison.cambridgema.gov Hundreds of new policy and program ideas were generated from

City of Cambridge impacts. utile Ideas to test include: across all planning topics. Envision Cambridge - 100% atfordable housing overlay - Environmental performance incentive - "Super-inclusionary" housing program • Some recommendations have a large-scale, measurable ettect on development and its outcomes • Additional analysis is needed to understand range of DRAFT envison.cambridgema.gov We want to test the effectiveness of priority zoning-related ideas Members of the public discussed proposed recommendations at a public meeting in July.

Goals: City of Cambridge district) utile permitting process Analysis of Working Groups New Policy Ideas create new affordable units Envision Cambridge DRAFT community input instead of discretionary permitting approvals • Allow for as-of-right approvals with required design review and Zoning policy for 100% affordable housing developments only: • Offer density bonuses and relief from dimensional standards (height, • Makes it easier and quicker to permit 100% affordable developments setbacks, open space) and parking where necessary (varies by zoning • Help reduce cost of producing affordable housing through expediated affordable housing's viability in areas where it's been more ditticult to • Foster equitable distribution of affordable housing citywide by expanding What is a 100% Affordable Housing Citywide Overlay? envison.cambridgema.gov

City of Cambridge housing. utile residential space Analysis of Working Groups' New Policy Ideas - Gives a 30% density bonus Envision Cambridge beyond the city's current requirements • Super-Inclusionary would be a voluntary apply to new residential buildings or - Requires that 20% of new residential • Current inclusionary zoning requirements: - Like the existing policy, it would only development is dedicated to affordable program that provides a larger density bonus units or more than 10,000 square feet of in exchange for additional affordable housing conversions which create 10 or more new Floor Ratio DRAFT 150% 250% · 200% 100% 50% 0% districts) and overlay Base Zoning (including base • Density under current zoning Zoning (existing, What is a "super-inclusionary" housing program? mandatory) Inclusionary voluntary) Affordable Tier A: 30% (hypothetical, voluntary) Affordable Hypothetical Super-Inclusionary Program Tier B: 35% hypothetical, envison.cambridgema.gov Density added under Super-Inclusionary voluntary) Affordable Tier C: 40% (hypothetical,

Year: Type: Target City of Cambridge 2020 Municipal utile - District energy performance, such as: Analysis of Working Groups New Policy Ideas - Resiliency measures construction is required 2022 Small (1-4 units) Residential Envision Cambridge set forth by the Net Zero Action Plan - Net positive construction after net zero 2025 Current Net Zero Action Plan requirements by use: Multifamily • A voluntary program that provides a density bonus in exchange for improved environmental building - Net zero construction ahead of the requirements DRAFT 2025 Commercial 2025 zero GHG emissions by mid-century. Institutional 2030 What is an environmental performance incentive policy? THE PATH TO A Laboratory DELZERO envison.cambridgema.gov Cambridge's 2015 Getting to Net Zero Framework is an action plan to get all buildings to net CAMBRIDyE

City of Cambridge utile Envision Cambridge Overlay Analysis DRAFT 100% Affordable Citywide envison.cambridgema.gov

needs. equity. City of Cambridge of differences. utile income community. Analysis of Working Groups' New Policy Ideas Envision Cambridge distribution of attordable housing citywide. • Expand affordable rental and homeownership • Provide a variety of housing options for people of • Foster neighborhoods of opportunity and equitable preservation and complementary infill development. different socioeconomic levels, life stages, and physical • Provide access to opportunities for all people regardless opportunities to enable Cambridge to thrive as a mixed- DRAFT Work toward addressing race-based disparities and racial Advances Envision Cambridge shared community goals: Maintain the existing patterns of the city through a mix of West Cambridg more Why study a 100% Affordable Housing Citywide Overlay? affordable envison.cambridgema.gov

City of Cambridge units utile Analysis of Working Groups' New Policy Ideas Envision Cambridge affordable housing developers. complete 100% affordable housing • Challenges to building affordable housing: • 100% Attordable Housing Citywide Overlay would: DRAFT - High land costs and competition from market-rate developers Enable affordable housing developers to better compete with market-rate developers Why study a 100% Affordable Housing Citywide Overlay? - Appeal of discretionary approvals can add significant cost, long delays, and significant risk to envison.cambridgema.gov - More difficult to build affordable housing in some areas of the city given zoning limitations (i.e. Help reduce costs for affordable housing development by balancing land costs with the number of for affordable housing developments without discretionary approvals and significant zoning reliet) Expand affordable housing's viability in areas where it's been more difficult to create new affordable - Remove permitting uncertainty and subsequent additional costs, delays, and risks to make it easier to density is limited in Residence A and B districts, where high land costs make development infeasible units needed to make projects feasible and lowering soft costs through an expedited approval process 10

City of Cambridge utile Fresy Pond Pony Analysis of Working Groups' New Policy Ideas Envision Cambridge DRAFT Number of Affordable Units by Census Block Charles River 1096 1 - 15 0 units 16 - 50 51 - 100 envison.cambridgema.gov 201 - 318 101 - 200 Number of Affordable Units -- Neighborhood Boundaries No residential units in block

City of Cambridge nonprofits). utile affordable units) • Main objectives are to: Analysis of Working Groups' New Policy Ideas developers in growth areas create new affordable units Envision Cambridge family housing and new construction low-, moderate, and middle-income people DRAFT housing production is determined by the availability of funding. - Development is subsidized by federal, state, and city government. • The City already partners with affordable housing developers (usually local - Two or three projects are typically completed each year (approx. 50-60 - Enable affordable housing developers to better compete with market-rate • The overlay does not change the number of units produced because affordable - FY19 funding for Affordable Housing Trust totals $13.6 million in City funds - Expand affordable housing's viability in areas where it's been more difficult to - City-funded affordable housing is created through acquisition of existing multi- - "Affordable" housing is income-restricted where housing costs are affordable to 100% Affordable Housing Citywide Overlay: Background envison.cambridgema.gov nonprofit developer in Cambridge. affordable housing development built by a 625 Putnam Ave is an example of a 100% 12

City of Cambridge utile fundamental viability Analysis of Working Groups New Policy Ideas affordable multi-family housing Envision Cambridge standards (height, setbacks, open space) and allowed use in Residence A and B districts); without creating uncertainty about a project's • Allow for as-of-right approvals with required • Allow affordable multi-family and townhouse Planning Board design review and community • Allow for conversion of any existing structure to developments in all districts (e.g. not currently an For 100% affordable housing developments only: • Offer density bonuses and relief from dimensional input instead of discretionary permitting approvals • A design review process would ensure good design parking where necessary (varies by zoning district) DRAFT ERE DE densities together in close proximity. Huron Ave. at Vassal Ln. has different What would the 100% Affordable Housing Citywide Overlay entail? envison.cambridgema.gov

City of Cambridge locations utile apply to the entire city Analysis of Working Groups' New Policy Ideas Envision Cambridge zoning to be competitive in the market of 2.0 to be financially feasible in most Cambridge Street, and areas of potential • The 100% Affordable Housing Overlay could - In higher-density areas such as Mass Ave, 2.5 times the density allowed by current base - In residential neighborhoods and other areas, affordable developments would need an FAR change, affordable developments would need DRAFT 100% Affordable Housing Overlay Ideas for Discussion envison.cambridgema.gov Residential Neighborhoods Potential Change Corridors and Areas of 14

change. City of Cambridge utile Envision Cambridge competitive with market-rate development. Cambridge Street, and the areas of potential affordable housing feasible in most locations. A 2.5x density bonus above base zoning made • Based on the land value, HR&A calculated the needed to make 100% affordable development feet that occurred within 500-feet of Mass Ave, • HR&A analyzed recent property transactions for properties with land area of at least 7,500 square number of affordable units and FAR that would be DRAFT envison.cambridgema.gov Commercial Corridors and Areas of Potential Change Analysis Potential Change 15 Corridors and Areas of

6.1+ 4.1 - 6.0 2.1 - 4.0 0.0 - 2.0 City of Cambridge Floor Area Ratio utile Analysis of Working Groups New Policy Ideas will be redeveloped. Envision Cambridge Not all parcels in these areas DRAFT Urban Form Comparison - Existing Built Density 1.6 5.0 3.9 envison.cambridgema.gov the form of a building. preservation, and design rules. FAR and rules about 7.9 4.3 2.4 height, lot coverage, historic guidelines combine to manage FAR is one of many urban form Floor Area Ratio Precedents

6.1 + 4.1 - 6.0 2.1 - 4.0 0.0 - 2.0 City of Cambridge Floor Area Ratio utile Analysis of Working Groups' New Policy Ideas will be redeveloped. Envision Cambridge Not all parcels in these areas by Current Zoning (Residential) DRAFT Urban Form Comparison - Density Allowed 1.6 5.0 3.9 envison.cambridgema.gov the form of a building. preservation, and design rules. FAR and rules about 2.4 7.9 4.3 height, lot coverage, historic guidelines combine to manage FAR is one of many urban form Floor Area Ratio Precedents

6.1 + 4.1 - 6.0 2.1 - 4.0 0.0 - 2.0 City of Cambridge Floor Area Ratio utile Analysis of Working Groups' New Policy Ideas will be redeveloped. Envision Cambridge Not all parcels in these areas 100% Affordable Overlay DRAFT Urban Form Comparison - Max. Density under 1.6 5.0 3.9 envison.cambridgema.gov the form of a building. preservation, and design rules. FAR and rules about 7.9 4.3 2.4 height, lot coverage, historic guidelines combine to manage FAR is one of many urban form Floor Area Ratio Precedents

City of Cambridge most locations. utile buildings to be built today. Envision Cambridge development feasible in ditterent areas. current zoning does not allow most existing Residential Overlay Analysis building meet that density requirement while • CDD analyzed recent property transactions in complementing the surrounding neighborhood? would be required to make an affordable housing the number of units and floor area ratio (FAR) that Residential A & B zoning districts, and calculated * • An FAR of 2.0 made attordable housing teasible in The present analysis asks: How can a contemporary • These neighborhoods' historic urban fabric includes some parcels with relatively dense development, but DRAFT other constraints • Methodology: "test-fit" analysis condition and quickly test how a real envison.cambridgema.gov building would fit on that parcel, given - Test-fits take a real or hypothetical parcel Residential Neighborhoods

classes City of Cambridge utile - B: 5,000-10,000 st - A: Less than 5,000 st - C: Greater than 10,000 sf • Representative Shape: Rectangular Envision Cambridge • Representative Context: Mixed Density • Representative Size: Three land area size- #Parcels [phone removed] 600 1200 400 1-1,000 DRAFT 1,001-2,000 2,001-3,000 3,001-4,000 4,001-5,000 5,001-6,000 6,001-7,000 7,001-8,000 8,001-9,000 9,001-10,000 10,001-11,000 Finding a generic parcel condition in residential areas Class 11,001-12,000 Land Area (sf) 12,001-13,000 13,001-14,000 14,001-15,000 Sizes for Each Size 15,001-16,000 Representative Parcel Residential A & B Parcels by Land Area 16,001-17,000 envison.cambridgema.gov 17,001-18,000 18,001-19,000 19,001-20,000 20

feet). City of Cambridge fit the development Other Test-Fit Assumptions utile insin Canese units and smaller units for smaller households). DRAFT Include at-grade ground-floor unit entrances and/or ramps to • Maintain the front setback typical of the surrounding context. unit (this represents an average point between larger family sized • Test development at both three stories (33 feet) and four stories (44 • Vary setbacks, open space, and parking requirements as necessary to • Approximately 1,000 square feet of gross floor area for each housing envison.cambridgema.gov

City of Cambridge 2 units, 2 stories utile Analysis of Working Groups' New Policy Ideas Maximum allowed under current zoning (many existing buildings are non-compliant) Envision Cambridge 7 units, 3 stories Affordable Citywide Overlay like in residential neighborhoods? Example of buildout under 100% DRAFT buildings, consistent with the historic mix of densities found throughout Cambridge. 8 units, 4 stories Affordable Citywide Overlay What would the 100% Affordable Citywide Overlay look Example of buildout under 100% Examples of hypothetical buildout on small parcels (<5,000sf): Res. A or B Prototypical Neighborhood 100% Affordable Housing developments in residential zoning districts would take the form of mid-rise multifamily envison.cambridgema.gov to height, residential buildings in are not zoning neighborhoods compliant, due 22 density, or both 69% of existing

) City of Cambridge 2 units, 2 stories utile Analysis of Working Groups' New Policy Ideas Maximum allowed under current zoning (many existing buildings are non-compliant) Envision Cambridge 10 units, 3 stories Affordable Citywide Overlay like in residential neighborhoods? Example of buildout under 100% DRAFT consistent with the historic mix of densities found throughout Cambridge. 12 units, 4 stories Affordable Citywide Overlay What would the 100% Affordable Citywide Overlay look Example of buildout under 100% envison.cambridgema.gov 100% Affordable Housing developments in residential zoning districts would take the form of mid-rise multifamily buildings, to height, residential buildings in are not zoning compliant, due neighborhoods 23 Examples of hypothetical buildout on medium-sized parcels (5,000sf - 10,000sf): Res. A or B Prototypical Neighborhood density, or both 69% of existing

1 unit, 2 stories City of Cambridge utile Analysis of Working Groups New Policy Ideas Maximum allowed under current zoning (many existing buildings are non-compliant) Envision Cambridge 17 units, 3 stories Affordable Citywide Overlay like in residential neighborhoods? Example of buildout under 100% DRAFT buildings, consistent with the historic mix of densities found throughout Cambridge. 23 units, 4 stories Affordable Citywide Overlay What would the 100% Affordable Citywide Overlay look Example of buildout under 100% Examples of hypothetical buildout on large parcels (>10,000sf): Res. A Prototypical Neighborhood 100% Affordable Housing developments in residential zoning districts would take the form of mid-rise multifamily envison.cambridgema.gov to height, residential buildings in are not zoning 24 compliant, due neighborhoods density, or both 69% of existing

City of Cambridge utile Analysis of Working Groups New Policy Ideas Envision Cambridge DRAFT Examples of recent infill development in Cambridge In these examples, the new developments are at a higher density than the surrounding buildings. envison.cambridgema.gov 25

Size Class 5,000- >10,000 <5,000 st 10,000 st 3 City of Cambridge 4 floors = 44 ft. 3 floors = 33 ft. Area 3,971 6,025 11,650 Parcel utile Floors* Test-fits Comparison 2,500 1,980 Floor 3,580 5,790 3,000 5,790 Area per *typical older triple decker is 35 - 40 ft. tall Envision Cambridge GFA 7,920 7,500 17,370 12,000 10,740 23,160 Area 9,140 6,080 6,100 9,875 15,660 20,905 Leasable DRAFT • 2 1.89 1.49 1.78 FAR ** Average Unit GFA = 1,000 sf. 17 12 10 23 Units** building space is used for open space 50% 41% 50% 50% 37% 50% Open Space** Unit count = Building GFA / Average Unit GFA 11 ft 11 ft 20 ft 20 ft 20 ft 20 ft Front Side envison.cambridgema.gov Setbacks 24' 12' 12' 24' Side 13' 13' Rear 26

A-1 A-2 District City of Cambridge Max 0.50 0.50 FAR 0.50 utile 6,000 2,500 4,500 Min Lot Area / DU Zoning Constraints 15 20 25 Envision Cambridge Setback Min Front to 25) to 35) to 20) 10 (sum 15 (sum Setback 7.5 (sum Min Side DRAFT 25 25 25 Setback Min Rear 35 35 35 Max Height Zoning rules in red would need relief for 100% affordable housing developments. 50% 50% 40% Ratio Min OS Ratio Parking 1 per DU 1 per DU 1 per DU No No No envison.cambridgema.gov Allowed? Multifamily 27

• City of Cambridge utile • Height considerations: • Parking considerations: Other Considerations Envision Cambridge protection, and/or tuck-under parking. parking compared to other parts of the city. also a possibility, but raises cost of development. - Tested at 3 and 4 stories, 33 and 44 feet respectively DRAFT - Existing homes in some districts have driveways, putting less pressure onto on-street - Parking ratios like 0-0.5 spaces per unit are more typical in affordable developments. - An additional 5-7 feet would allow for pitched roofs, a raised first floor for stoop, flood optimal density), as well as parking requirements. Green roofs and stormwater management - At-grade spaces require compromise with open space requirements; tuck-under parking is • Open space requirements require compromise with height requirements (in order to achieve the envison.cambridgema.gov 28

City of Cambridge utile complementary designs. achieve their housing goals. Design Considerations Envision Cambridge • Cambridge can build on the design innovation currently DRAFT • Though an affordable overlay should provide for an easier underway in similar cities that manage infill development to should still be subject to design review, to ensure the best viable process of development approval, 100% affordable development outcomes for open space siting, maintenance of existing trees, and envison.cambridgema.gov 29

Attachment D Who Participates in Local Government? Evidence from Meeting Minutes* Katherine Levine Einstein' Maxwell Palmert David Glick® June 29, 2018 Forthcoming, Perspectives on Politics Abstract Scholars and policymakers have highlighted institutions that enable community partic- ipation as a potential buffer against existing political inequalities. Yet, these venues may be biasing policy discussions in favor of an unrepresentative group of individu- als. To explore who participates, we compile a novel data set by coding thousands of instances of citizens speaking at planning and zoning board meetings concerning housing development. We match individuals to a voter file to investigate local political participation in housing and development policy. We find that individuals who are older, male, longtime residents, voters in local elections, and homeowners are significantly more likely to participate in these meetings. These individuals overwhelmingly (and to a much greater degree than the general public) oppose new housing construction. These participatory inequalities have important policy implications and may be contributing to rising housing costs. *This research was funded by Boston University's Initiative on Cities. Many thanks to Mirya Holman, Spencer Piston, Jessica Troustine, and participants at the Vanderbilt Local Political Economy Conference, University of Wisconsin-Milwaukee Political Science Research Workshop, American Political Science Asso- ciation "New Faces of Urban Politics" Mini-Conference, and Boston Area Research Initiative Spring 2018 Conference for their helpful comments. We gratefully acknowledge our outstanding research assistants Luisa Godinez Puig and Sarah Sklar. All errors are our own. *Assistant Professor, Department of Political Science, Boston University. kleinst@buedu. #Assistant Professor, Department of Political Science, Boston University. [email removed]. §Associate Professor, Department of Political Science, Boston University. [email removed].

Many local leaders view neighborhood activism and participation as a key source of policy information and a critical form of civic engagement. Almost half of mayors selected neighborhood meetings as one of the top two ways they learn about their constituents' views (Einstein, Glick, and LeBlanc 2017), and the National League of Cities highlighted neighborhood meetings as a critical component of community engagement (Hoene, Kingsley, and Leighninger 2013). The celebration of neighborhood participation is not new. President Lyndon B. Johnson's Community Action Program—part of the 1964 Economic Opportunity Act—institutionalized neighborhood involvement in the allocation of federal urban spending. President Jimmy Carter made neighborhoods a cornerstone of his administration's housing programs (Carter 1980). Scholars of local government and normative theorists more broadly contend that institutions that spur neighborhood-based political participation help provide voice to underrepresented groups, enhance citizen efficacy, and are integral to a thriving democracy (Berry, Portney, and Thomson 1993; Fung 2006; Michels and Graat 2010; Stone and Stoker 2015). Moreover, such institutions may offer opportunities for compromise via deliberative democracy (Gutmann and Thompson 2012). Indeed, the local level may offer the most potential to benefit from such institutions as participation and efficacy are greater in smaller jurisdictions (Oliver 2001; Lassen and Serritzlew 2011; Oliver, Ha, and Callen 2012). In some ways, local institutions that enable direct citizen involvement echo national efforts to increase political participation among socioeconomically disadvantaged voters. In response to participatory inequalities, some policymakers and advocates have pursued a variety of initiatives designed to facilitate registration, offer more early voting, and shorten lines at polling places, for example. These policies may, however, have unanticipated consequences. In some cases, they may exacerbate the very inequities they attempt to solve. Berinsky (2005) finds that reforms designed to facilitate voting actually increase sociocconomic inequalitics in turnout; de Kadt (2017) uncovers a similar phenomenon in South Africa. Burden et al. (2013) discover that, while Election Day registration has a positive effect on overall turnout, early voting appears to decrease turnout in isolation. 2

Institutions designed to encourage and empower neighborhood participation in local politics could also have distorting consequences for the distribution of influence. We examine this possibility using the substantively important case of housing policy. In the wake of the excesses of urban renewal (Rae 2004; Schleicher 2013) and the dominance of pro- growth, developer-oriented urban politics (Logan and Molotch 1987), local governments have promulgated institutions designed to constrain developers and empower neighborhood- level and environmental interests (Logan and Rabrenovic 1990; Gerber and Phillips 2003; Glaeser and Ward 2009; Schleicher 2013). One example is a movement in many localities to allow and encourage neighborhood participation in zoning and planning board meetings. Such participation gives neighbors an opportunity to inform appointed board members and local elected officials of their views on projects ranging from large developments to modest renovations. It also offers the potential to extract concessions from developers (sometimes directly (Hankinson 2013)). However, greater participation may amplify some voices more than others. The concen- trated costs of development projects in particular may create strong incentives for neigh- borhood groups that are highly affected by a proposal to mobilize against development. In contrast, the diffuse benefits of an increased housing supply are less likely to motivate participation from the broader population of a city/region that might benefit from more housing. Land use regulations may provide these highly motivated individuals the tools with which to restrict higher density projects. This failure to construct an adequate supply of housing has important policy consequences. The Obama White House identified national housing affordability as a critical policy challenge, arguing that "the growing severity of undersupplied housing markets is jeopardizing housing affordability for working families, increasing income inequality by reducing less-skilled workers' access to high-wage labor markets, and stifling GDP growth by driving labor migration away from the most productive regions" (White House 2016). The lack of affordable housing in areas with high mobility could have a profound negative impact on many children's life

opportimities (Chetty, Herdren, and Katv 2016). While housing crises in some of the nation's coastal cities has been the focus of media attention, a lack of affordable housing is a national crisis. There is not a single county in the country in which a minimum-wage earner can afford an average two-bedroom rental (National Low Income Housing Coalition 2017). Housing affordability and supply are inextricably linked. Economists have attributed the current affordability crisis in large part to insufficient supply (Quigley and Rosenthal 2005; Glaeser, Gyourko, and Saks 2005; Gyourko, Saiz, and Summers 2008; Glaeser and Ward 2009; Glaeser 2011; Gyourko and Molloy 2014; Hsieh and Moretti 2015). Moreover, insufficient housing supply may hamper efforts at environmental sustainability. Greater housing density helps reduce sprawl (Glaeser 2011) and is a cornerstone of local efforts to mitigate climate change (Barro 2017). We ask how participation may play a role in restricting development. To assess local political participation, previous studies have relied primarily on surveys (Hankinson 2018; Marble and Nall 2017; Wong 2018), voting (Fischel 2001; Gerber and Phillips 2003; Wong 2018), case studies of meetings (Mansbridge 1980; Fiorina 1998), and aggregate-level analyses of meeting participation (Fung 2006). In contrast, we rely on directly observing both who participates in policy discussions about housing development, and how they participate. We do so across a range of communities by compiling and coding new data on all citizen participants in planning Board and zoning board meetings dealing with the construction of multiple housing umits in 97 Massachusetts cities and towns. We match thousands of individual participants to the Massachusetts voter file to explore who participates in local political meetings. This data set is the first comprehensive effort to measure the behavior of community meeting participants. Moreover, we juxtapose the opinions of meetings attendees with the vote on a statewide housing ballot referendum to provide a novel comparison of attendee views with those of the broader public. This allows us to learn two separate attributes of meeting attendees: (1) whether they are demographically representative of their broader communities, and; (2) whether they are attitudinally representative of their broader 4

communities. We find that meeting participants are unrepresentative of the broader public in a variety of ways. They are more likely to be older, male, longtime residents, voters in local elections, and homeowners. Moreover, these individuals overwhelmingly oppose the construction of new housing: almost two-thirds of these participants speak out in opposition to new housing development. A sizable minority of meeting participants—especially housing opponents— are repeat participators who attend multiple meetings to speak out about local housing projects. Meeting attendees generally raise a wide variety of issues, from concerns about. local trees to traffic. These results suggest that the structure of public meetings surrounding housing development likely contributes to a failure in many locations to produce a sufficient housing supply. More broadly, they reveal that institutions designed to enhance democratic responsiveness may have perverse consequences on participation, the views that policymakers hear, and/or outcomes. This article makes two important contributions. First, while a multitude of political science studies have identified inequalities in political participation, this article is the first able to document inequalities in who shows up to salient public meetings. Rather than using surveys or vote returns, this study is the only one to our knowledge that directly observes participants in politics to precisely measure inequalities. Second, it makes a novel theoretical argument about the nature of participation in housing policy. We argue, that even in areas where public opinion broadly favors redressing housing shortages with increased supply, specific housing development proposals will disproportionately garner opposition that In the housing policy arena, institutions and behavior is empowered by local institutions. align in a way that enable non-majoritarian outcomes with tangible implications for housing availability. 5

1 Who Participates At the heart of all of the predictions outlined below are general and fundamental questions about grass-roots democracy. Throughout our analysis, we consider two competing views about neighborhood-level civic engagement on housing policy. The first is that these meetings are an opportunity for efficacious civic engagement, mediation of competing interests (Dahl 1961; Berry, Portney, and Thomson 1993), and deliberative democracy (Gutmann and Thompson 2012). The second, in contrast, views neighborhood activism as captured by a small, unrepresentative group with strong views (Mansbridge 1980; Fiorina 1998; Kain 2012). A wide body of scholarship in American politics suggests that more socioeconomically advantaged individuals are more likely to participate and to have their voices amplified in key policy discussions (Schlozman, Verba, and Brady 2012; Gilens 2014; Hajnal and Trounstine 2016). Political science research also generally finds higher levels of political participation among the elderly, who have the time, resources, and policy interest that allow for and encourage involvement in politics (Campbell 2005; Schlozman, Verba, and Brady 2012). Those that have lived in the same place for a greater duration (Kang and Kwak 2003; Gay 2012), and own their homes (Fischel 2001) also participate in politics at higher rates. Men-especially white men--are more likely to engage in direct contact and collective action relative to women (Mansbridge 1980; Kittilson 2016). We suspect these broad findings will also apply to participants in neighborhood meetings. This would fit with research on participatory small-group decision processes that contends that such institutions are unrepresentative in similar ways to other forms of political participation (Mansbridge 1980; Sanders 1997). While participatory inequalities have been widely studied—though not precisely empirically measured —it is less obvious whether meeting attendees will be predisposed to hold particular attitudes. Accounts of anti-development, NIMBY (not in my backyard) sentiments among homeowners generally predominate urban politics research (Fischel 2001; Hankinson 2018). Many contemporary commentaries on housing, however, point to the influence of a new

housing coalition featuring poverty and affordable housing advocates, developers, and urban planners as a potential check on NIMBY sentiments from entrenched homeowners (Semuels 2017; Yglesias 2017). Indeed, recent evidence from ballot initiatives and surveys suggests that, in liberal communities, mixed-income developments may generate at least some public support, with individuals basing their preferences for housing on ideology rather than pure economic self-interest (Wong 2018). We argue that the development of new housing may disproportionately induce participa- tion from individuals with unrepresentative opinions. The potential externalities of housing proposals are spatially concentrated while the benefits are diffuse. Proposed housing de- velopments have potentially profound effects on neighborhood property values, amenities, and quality of life (Fischel 2001). Increasing the housing supply reduces housing prices (Quigley and Rosenthal 2005; Glacser, Gyourko, and Saks 2005; Gyourko, Saiz, and Summers 2008; Glaeser and Ward 2009; Glaeser 2011; Gyourko and Molloy 2014; Hsieh and Moretti 2015). This reduction in housing prices would adversely impact the economic interests of local homeowners. Interestingly, renters may also feel that-new housing developments are detrimental to their economic interests. Recent experimental evidence suggests that renters in high-cost housing markets believe that new developments will raise their rents (Hankinson 2018). Moreover, housing developments frequently represent stark changes in neighborhood environments and composition. Studies of racial and ethnic politics have found such rapid changes to be strong motivators for attitudes and behavior (Green, Strolovitch, and Wong 1998; Hopkins 2010; Enos 2016). In contrast, we anticipate that proponents of new housing development will be compar- atively less likely to attend meetings on proposed projects. The economic benefits of new housing supply are diffuse. Any change in housing affordability from a single project is likely to be barely perceptible, particularly when weighed against the visible costs experienced by a narrower subset of the neighborhood. Indeed, even if the benefits were comparable, prospect, 7

theory suggests that losses have a greater impact on behavior than equivalently sized gains (Kahneman and Tversky 1979). Moreover, at least some of the individuals most likely to benefit from a new housing development (potential new residents) live outside the jurisdiction in which the development is proposed. In contrast, virtually all of those experiencing the costs of new housing already reside in that jurisdiction. Relative to supporters, then, housing development opponents are more likely to: (1) be informed about developments happening in their community and (2) be able to target their own appointed/elected officials in voicing their views about housing. Both information (Lassen 2005) and efficacy (Shingles 1981; Finkel 1985) are positively associated with political participation. Interestingly, this bias towards opposition to specific projects may differentiate housing from some other areas. We contend that even those individuals who are predisposed to support the construction of affordable housing in the abstract will inclined to oppose specific housing project proposals in their communities. This sharply differs from immigration policy, for example. Iyengar et al. (2013) find that citizens in advanced industrialized democracies support the admission of individual immigrants, while generally opposing more open immigration policies. In addition, we also anticipate that those who participate will do so with high intensity and frequency. The factors listed above that should disproportionately spur opposition to local housing development will likely also foment strong public opinions. Intense viewpoints are linked with a greater propensity for political participation (Fiorina 1998; Kain 2012; Pew Research Center 2014). Therefore, we expect meeting attendees in general-and, in • particular, opponents of new housing development—to attend repeat meetings. Finally, we expect participants to exhibit high levels of expertise. In previous predictions, we suggested that participants are likely to be socioeconomically advantaged and perceive significant costs of proposed housing developments. We might expect a highly educated group that views changes to the housing stock as a major threat to learn about and cite local zoning laws and land use regulations. They may also solicit the views of experts-such 8

as lawyers, engineers, architects, and other real estate professionals—to provide strong and well-sourced arguments about a potential housing development. This expertise might lead to well-informed neighborhood dialogue, but, it could also exacerbate political inequalities. Lupia and Norton (2017) suggest that deliberative democracy may not work as intended if participating interlocutors use sophisticated language as a form of political power to drown out other policy discussion. Perhaps strikingly in the context of rising national partisan polarization (Abramowitz 2010), we do not expect partisanship to predict participation in housing meetings of to affect the issues that individuals raise. While partisanship certainly impacts local politics (Tausanovitch and Warshaw 2014; Einstein and Kogan 2016), we anticipate that the immediacy of neighborhood-level concerns will swamp partisan leanings on housing issues. Indeed, Marble and Nall (2017) use survey experiments to show that liberal homeowners-—while generally favorable towards redistributive programs-prioritize their home values over their ideological preference for affordable housing (though see (Wong 2018)). 2 Data and Methods To evaluate who participates, we assembled a novel data set of all citizen participants in planning board and zoning board meetings between 2015-2017 in 97 cities and towns in metropolitan Boston. One reason we focused on Massachusetts is data availability. As a consequence of the Commonwealth's open meeting law, Massachusetts localities are required to provide detailed meeting minutes for all public bodies. These minutes must include "a summary of the discussions on each subject." A majority of cities/towns in metropolitan Boston have interpreted this to mean including the names and addresses of all members of the public who spoke at the meeting. In addition to the data availability, the Boston metro region has other advantagcous traits for studying participation in the hyper local politics of housing development. While compact,

the Boston metro area includes an unusual mumber of independent cities and towns. Indeed, there are dozens of autonomous local communities with their own demographics, politics, and local regulations within 50 miles of Boston. Boston's surrounding communities range from small, leafy, "bedroom" towns to more diverse small cities. The housing stock in the area includes estates, modest starter homes, three family "triple deckers," and taller modern apartment buildings. While the eastern Massachusetts economy and housing marking are doing quite well relative to other parts of the country, there is still great variation across municipalities in terns of housing demand, availability, and cost. Moreover, the strength of the overall housing market is an asset for this study because it means there is demand for housing, and a market for new development, almost everywhere. Lastly, the fact that Eastern Massachusetts is generally liberal makes it a difficult test for some of the hypotheses. It is disproportionately populated by people who, in the abstract, would tend to support more housing and efforts at improving access to affordable housing. ; In Table 1 we provide summary statistics about a variety of traits (mean, minimum and maximum) for the 97 cities and towns for which we have coded meeting comments. As the data show, our sample is, as would be expected in eastern Massachusetts, relatively white (86% on average) and affluent. More important than the means are the ranges of these variables, many of which are directly pertinent to the theoretical expectations. For example, the sample has tremendous variation in terms of residential density (237 to nearly 17,000 people per square mile), housing prices ($200K to $1.2MM median), population growth (0% to 11% from 2010-2015), and age (9% to 28% over 65). To assemble our dataset, we downloaded all available public hearing minutes for local planning and zoning boards. In all cities and towns, these are the two bodies responsible for reviewing any housing developments not permitted by right under local zoning code. Such housing projects were publicly reviewed by one or both bodies in such cases. In many of these meetings, owners or developers are petitioning for variances (exceptions) to the underlying regulations. Under Chapter 40A in Massachusetts, all public hearings for such bodies are 10

Table 1: Traits of cities and towns for which we have participation data min mean max 183382 25772 4427 Population 1976 237 16880 Population Density -0 11 5 Population Growth 2010-2015 53 24 42 Median Age 28 9 15 Percent Over 65 98 86 17 Percent White 15 2 Percent Black 0 76 0 5 Percent Hispanic 34852 199519 Median Household Income 97650 Median House Price 205200 1170400 431844 24 4 43 Distance from Boston (miles) 97 Observations published in "a newspaper of general circulation in the city or town once in each of the two successive weeks, the first publication to be not less than fourteen days before the day of the hearing." Cities/towns also are required to post a notice "in a conspicuous place in the city or town hall" with similar advanced notice. Moreover, the city/town also must mail a notice of a public hearing to "parties of interest," which are defined as "the petitioner, abutters, owners of land directly opposition on any public or private street or way, and abutters to the abutters within three hundred feet of the property line of the petitioner as they appear on the most recent applicable tax list" (Commonwealth of Massachusetts 2017). We utilized all minutes that were posted on cities' and towns' websites. The public hearings captured in our database covered a wide range of policy areas, ranging from the construction of large multifamily or mixed use housing developments with hundreds of rental units to the addition of wireless communication towers. We focus on all hearings concerning housing developments featuring the construction of more than one unit of housing.! This focus reflects our interest in the polities of increasing housing supply via 'By definition, all meetings are those in which a developer or homeowner is asking for an exemption to the local zoning code. Projects approved by right do not go through the local zoning process. It is possible that only those projects that require an exemption generate public opposition. Indeed, the drawing of these maps is in and of itself intensely political (Rothstein 2017). The meeting minutes fcature many citizen opponents 11

densifying communities with high demand. Even within this more limited policy category, public meeting minutes exhibit enormous variation. Some of these projects are relatively small (e.g. a family seeking to add an accessory apartment), while others are expansive proposals from large professional development companies. Some meetings feature comments from one neighbor who shows up to support a friend in obtaining a variance from local zoning regulations. Others, in contrast, are filled with dozens of comments from residents with deep concerns about a proposed project. Using these minutes, we created a database of all public comments surrounding the development of more than one housing unit. Each observation—-which is at the comment level—includes the name and address of the meeting participant.? We also code whether the individual supports, opposes, or is neutral about a proposed housing project. Finally, when available, we also include a code describing the reason(s) the participant expressed along with her support/opposition/neutrality. These reasons encompassed a wide variety of topics, including parking, environmental concerns, traffic, density, affordability, noise, aesthetics/history, property values and septic systems, among others. 3 A full codebook describing these categories and criteria for inclusion is included in the appendix. Because some of the meeting minutes provide extraordinary detail-including in some cases exact transcripts of proceedings—we are also able to also analyze valuable qualitative data. Even without merging these data with any other information, we can make valuable observations. Because each publie comment is an observation, we can calculate the proportion of meeting attendees who are repeat participants and how many meetings these individuals attend). Moreover, we can learn the proportion of individuals who support/oppose the development of additional housing and the reasons they typically cite. to changes in the zoning code as well (these individuals are not included in the data set analyzed here). "If an individual speaks multiple times at a meeting about different housing developments, she receives one observation per housing project. If participant makes multiple comments about the same project at the same meeting, her comments are concatenated into one observation. Finally, if the same individual attends multiple meetings to comment about the same project, she is coded as one observation per meeting. ' Intercoder reliability checks showed that coders agreed 100% of the time about whether a comment should be labeled support/oppose/neutral. They selected the same set of 19 total topic categories 85% of the tine. 12

What's more, because we have the names and addresses of these individuals, we can merge them with data from the Massachusetts voter file to learn more about their demographics. Using a fuzzy matching algorithm, we link meeting commenters with registered MA voters.4 We were able to match 2,580 of the 3,123 people in the set of participants (82.6%). As many people commented more than once, we were able to match the speakers of 85.4% of the comments to the voter file. The voter file offers some important demographic data about these meeting participants, and allows us to compare these individuals to city/town-level demographics. In particular, the voter file provides data on individuals' age, gender, partisanship, history of voter turnout in elections at all levels, and registration date at current address (which we use as a rough proxy for duration of residence). While this analysis obviously will not convey a complete picture of (un)representativeness-it does not include income or race, most notably—it offers unprecedented insight into the individuals who participate in local democratic proceedings. 3 Results We begin by using the voter file to compare those who participated in local meetings to those in their towns who did not. Table 2 presents the difference in means between commenters and non-commenters. On average, meeting participants are older, have lived at their residence for longer (proxied by the length of their voter registration at that location), and are more likely to be men. Women constitute 51.3% of the voter file, but only 43.3% of the commenters at development meetings. As expected, we find no differences in partisanship. Democrats, Republicans, and Independent/Unaffiliated voters participate at similar rates. There are significant differences based on vote history. The individuals who participated in development 4We matched on name and address, the only data on participants available. Due to a large number of typos and misspellings, we used a fuzzy string matching algorithm and manual review of the matches. A majority of the people who we were unable to match are likely in the voter file, but could not be matched due to name duplication and missing addresses. 13

meetings voted at roughly twice the frequency of those who did not." Table 2: Difference in Means Between Commenters and All Voters Non-Commenters Commenters N Mean Mean Difference N Variable 7.818** 50.893 58.711 2,566 Age 1,535,520 5.549** 17.377 11.828 2,580 1,618,375 Reg. Length -0.080** Female 0.433 0.513 1,618,375 2,580 0.002 0.317 0.320 Reg: Democrat 1,618,375 2,580 0.001 0.112 0.111 2,580 1,618,375 Reg. Republican 0.566 0.002 2,580 1,618,375 0.563 Reg: Independent % Elections Voted 0.502 2,580 1,618,375 0.272 0.230** Table 3 presents logit models using the full voter file, where the dependent variable is an indicator of whether or not the resident participated in a development meeting. 'The first specification includes only individual-level variables, the second includes town-level controls (town averages for each individual variable), and the third includes town-level fixed effects. The results are consistent across all three specifications." Voters are more likely to participate when they are older, have lived in the same address for longer, and vote more frequently. Female voters are less likely to participate, and we observe no partisan differences. These results broadly confirm that meeting participants are demographically unrepresentative of their towns in ways consistent with our theoretical predictions. One key independent variable that we cannot assess using the voter file is homeownership. While we are unable to collect homeownership data for the thousands of commenters in the data, we did match the 85 individuals who participated in the Town of Arlington's Zoning and Planning Board meetings with data from the Registry of Deeds. We selected the Town of Arlington because: (1) the relatively high number of comments (122 comments from 85 individuals) in the town allowed us to make reliable comparisons with town-level 5% Elections Voted is calculated as the share of elections between 2010 and 2016 in which in individual voted. The total number of possible elections varies by town. «We also examined various subsample models, including restricting the data to towns with at least 15 commenters. Such restrictions do not have any meaningful effect on the results. 14

Table 3: Logit Models of Commenters Relative to Full Voter File (1) (3) (2) 0.003* 0.005** 0.004** Age (0.001) (0.001) (0.001) 0.017** 0.019** 0.012** Reg. Length (0.002) (0.002) (0.002) Female -0.391** -0.404** -0.408** (0.040) (0.040) (0.040) 0.039 0.100 0.109 Reg. Democrat (0.070) (0.070) (0.068) 0.149* 0.113 0.158* Reg: Independent (0.064) (0.064) (0.064) 2.088** % Elections Voted 2.218** 2.052** (0.076) (0.075) (0.076) Observations 1,538,086 1,538,086 1,538,086 97 97 97 Towns X Town Controls Town FEs Standard errors in parentheses P<0.05 **p<0.01, * demographics, and (2) the town has a good mix of homeowners and renters (39% of the population are renters and 61% homeowners). We find that, consistent with our predictions, homeowners are significantly overrepresented as meeting participants; while 39% of the town rents, renters only comprise 22% of participants. Next, we assess the proportion of meeting attendees in our full data set who participated in multiple meetings. Somewhat in contrast with our predictions, most participants only attended a single meeting: Eight-three percent of the commenters in our sample spoke at only one meeting. The average person made 1.3 comments, and 45 people made five or more comments. Among the participants that we matched to the voter file, the only significant predictor of the number of comments made is political party. Democrats were less likely to make multiple comments, and Republicans were more likely to do so. 15

3.1 Predicting Commenter Positions Turning to the positions expressed by meeting participants, the overwhelming majority of attendees spoke out in opposition to proposed new housing. Sixty-three percent of all comments were in opposition to proposed housing projects, while only 14.6% expressed support; the remaining 22.8% of comments were neutral. These results strongly suggest that, as predicted, the incentives to show up and oppose new housing are far stronger than those to participate in support. We also use individual-level variables to predict which participants oppose new housing. Table 4 presents the results of this analysis. Consistent with theoretical predictions, all else equal, those who appeared at multiple meetings are more likely to speak in opposition. Women and infrequent voters are also more inclined, on average, towards opposition. Democrats, in contrast, are more likely to support projects and less likely to be neutral or oppose them than independent or Republican participants. This last finding is consistent with Democrats having more progressive views on housing (Marble and Nall 2017), but contrasts with much of the media coverage on the NIMBY movement, which suggests that NIMBYism is particularly prevalent among progressives (Capps 2015; Paul 2015). Our results suggest that, within the progressive places facing housing crises likely to engender NIMBYism, Republicans are more likely show up to meetings in opposition to new housing. This finding suggests that. liberal homeowners and renters may, in some instances, overcome a neighborhood-based opposition to new housing (Hankinson 2018; Marble and Nall 2017) to support more dense housing consistent with their ideological preferences. Overall, though, support for new housing remains low among both affiliates of both parties: only 19.4% of Democrats and 12.8% of Republicans expressed support for proposed projects at public meetings. The failure of individual-level demographics like age and gender to predict opposition to housing construction is methodologically important. We theorized that meeting participants would be weighted towards opposition because of a combination of the concentrated costs of new housing, prospect theory, and residence in the jurisdiction where housing is proposed. 16

Table 4: Logit Models of Commenter Positions (2) (1) (3) DV=Neutral or Oppose DV=Support DV=Neutral DV=Oppose 0.005 0.004 -0.005 -0.005 Age (0.005) (0.005) (0.003) (0.004) 0.004 -0.004 -0.013** 0.007 Reg. Length (0.005) (0.004) (0.005) (0.005) 0.253** -0.253** 0.123 0.013 Female (0.080) (0.098) (0.098) (0.070) -0.479** 0.113 -0.360** 0.466** Reg: Democrat (0.119) (0.163) (0.163) (0.139) 0.027 -0.041 0.214 -0.153 Reg. Independent (0.158) (0.158) (0.129) (0.111) -0.653** -0.460** 0.664** 0.159 % Elections Voted (0.155) (0.155) (0.111) (0.127) 0.038 0.058** -0.038 -0.052* Number of comments (0.029) (0.029) (0.025) (0.021) 3,629 3,629 3,629 3,629 Observations Standard errors in parentheses p<0.05 **p<0.01,* A propensity towards opposition, however, could also simply be a consequence of the un- representative demographics of meeting participants; perhaps older men, for example, are both more likely to participate in planning and zoning meetings and more likely to oppose the construction of new housing: Instead, we find that the predictors of participation in meetings are completely different from those that explain positions in meetings. Older and male individuals are more likely to participate in meetings, but, conditional on participation, age and gender do not predict opposition to new housing. 17

3.2 Support for Housing in the Voting Booth, Opposition at Meet- ings We have demonstrated that a large majority of individuals who attend zoning and planning board meetings express opposition to the projects under consideration. To assess if such opposition is disproportionate, we compare meeting participation to the results of an important ballot referendum concerning housing policy. In 2010, Massachusetts held a referendum to repeal Chapter 40B, a law promoting affordable housing that permits developers to bypass local zoning regulations if: (1) the town's housing stock is less than 10% affordable and (2) at least 20-25% of the proposed units have long-term affordability restrictions. Across the state, a majority of voters favored keeping the law, and the referendum to repeal Chapter 40B failed with only 42% of the vote. Figure 1 shows the distribution of the vote supporting 40B by town. Across the cities in our sample, 56% of voters in the referendum adopted the pro-affordable housing position and opposed repeal of Chapter 40B, and there was majority support against repeal in 61 of the 96 towns.? This comports with state-level figures, where 58% of voters opposed the repeal. This is a significantly greater level of support than evinced by the mere 15% of meeting commenters who spoke in support of the construction of new housing: This is especially striking given that Chapter 40B deals exclusively with affordable housing. We would expect opposition to affordable housing to be greater than opposition to market-rate housing based on prior scholarship on public opinion surrounding housing (Tighe 2010).& If anything, then, our measure of general public opinion is biased towards opposition, and should be more similar to the opinions evinced in our meetings minutes. The relative toughness of this particular test makes the 40-percentage point difference between 40B support and support for housing projects at public meetings all the more striking. We do not have 40B repeal results for Boylston, MA. &Only 3% of negative comments cited affordability. Thus, there is little evidence that our commenter data would be biased towards opposition because it featured market-rate, rather than affordable, housing developments. 18

Figure 2 shows the relationship between town-level vote against repealing Chapter 40B and the percentage of comments in each town that were supportive of multifamily housing developments.' While there is a positive correlation between opposition to the 40B repeal and positive comments, in every town, fewer than half of the meeting comments were positive. For example, in Cambridge, the town with the highest support for 40B (80% of voters opposed repeal), only 40% of comments at development meetings supported multifamily housing: Indeed, almost every town in Massachusetts exhibited higher support for Chapter 40B than for the development of specific multifamily housing projects. While voters in tilese towns supported affordable housing construction in the abstract, a significant majority of those who attended development meetings opposed the development of specific project proposals. 3.3 Reasons Expressed for Supporting and Opposing Development Finally, we also investigate the reasons individuals cited when expressing their support/opposition on housing projects. While many meeting minutes simply noted whether participating indi- viduals supported or opposed a project, some provided greater detail—in some cases exact transcripts of individuals' comments. Figure 3 shows the frequency of each reason given by the position taken by commenters. Perhaps the most striking result is the variety of reasons offered, including flood sus- ceptibility, septic systems, environmental concerns, neighborhood character, and parking; among other things. Moreover, there are notable differences in the reasons provided by supporters and opponents. Supporters of new housing were significantly more likely to mention affordability concerns. Opponents, in contrast, were more likely to raise traffic, environmental, flooding, and safety concerns. The reasons cited suggest that, unsurprisingly, commenters raise issues that reflect the contexts in which their communities are situated. Almost one-quarter of opposing comments cited traffic, and most of these highlighted specific instances of congestion. A Manchester- We restrict the sample to the 70 towns where there were at least 10 comments. 19

30 20 Towns 10 0 10% 20% 50% 60% 70% 80% 90% 100% 40% 30% 0% Vote for Affordable Housing in 40B Referendum Figure 1: Support for 40B Referendum 100% 90% % Comments Supporting Multifamily Housing Developments 0% 0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100% Vote for Affordable Housing in 40B Referendum Figure 2: Support for 40B Referendum 20

oppose neutral support Aesthetics Affordability Building Foundation Corruption Density Diversity Environment Flooding Height/Shadows Home Values/City Finances Neighborhood Character Noise Non-Compliance Parking Pedestrian Impact Privacy Safety Schools Septic/Water Traffic 0% 5% 10% 15% 20% 0% 5% 10% 15% 20% 0% 5% 10% 15% 20% % of Group Naming Reason Figure 3: Reasons Given by Commenters, Grouped by Position Taken 21

by-the-Sea woman observed that "traffic has increased at a fast rate even without the new building" in her community. One Foxborough man "commissioned his own traffic study as he feels the impact of cars and children on the area have not been adequately addressed. He has lived in the area for a few years and compares the peak traffic periods to a demolition derby." (This commenter's ability to commission his own traffic study also illustrates the unrepresentative resources that many of these participants have available to them. Traffic studies typically cost thousands of dollars.) Similarly, almost 15% of comments opposing new housing mention flooding concerns, and many of these cited specific instances of water in basements, yards, or nearby streets. A Newburyport woman noted that "Boyd Drive already experienced flooding. The impact on existing homes was not assessed." A Reading man "explained that a couple of homes on Dustin Road have a lot of water and flooding problems, and opined that rain gardens will not work." Given the historically exclusionary aims of many zoning and land use regulations (Troun- stine 2016; Rothstein 2017), the comments may also provide a means of evaluating the extent to which race and racial bias drive opposition to the construction of new housing. In particular, the nearly 11 percent of commenters who cited "neighborhood character" in opposition to a housing project may be using raciálly coded language. '° Indeed, many activists and media observers view such concerns in this light. Jacobus (2017) notes: "If you are like me, when someone says they want to 'preserve the character' of a community, what you hear is that they want to exclude poor people and people of color." A few of the comments that fell under the neighborhood character umbrella appear to be racialized. One man in Beverly—a town that is 83% white—critiqued the design of a building as "ridiculous" and said "Beverly is going to look like Chelsea." 62% of Chelsea's population is Hispanic (and Chelsea is six towns away). He went on to ask if "there is a restriction put on the building that there is to be no Section 8 housing in the building." Several other comments in the 10 Publie safety may, on its face, also seem like it includes concerns evincing underlying racial biases. In most cases, however, these comments had to do with emergency vehicle access and pedestrian safety in heavy traffic. 22

database similarly argued that their homogenous communities would resemble much more diverse ones if a project were approved. Most of the comments referencing neighborhood character, however, are not explicitly linked with race. A Dighton woman opposed a project because she felt it was "not consistent with the neighborhood. A multi-family home built on a slab is going to negatively impact the values of homes in the neighborhood. The other homes in the neighborhood are single family homes that are owner occupied." There may very well be racial undertones to this woman's opposition--and there are alinost certainly some class concerns. But, there is nothing explicitly in her comments that clearly ties her opposition to racial bias. Many of the comments that referenced neighborhood character across a variety of towns were remarkably similar to hers; a Concord man "spoke in opposition to the project and the change in the neighborhood character." A woman in Hudson "was worried about the character of the neighborhood and how this doesn't fit in." The content of these comments also allows us to qualitatively capture the knowledge and expertise of these commenters. Many commenters cited their professional backgrounds in law, design, engineering, architecture, and real estate in making assessments of housing projects that personally affected their communities. In addition, the content of many of their comments suggested an extraordinary familiarity with highly complex local land use regulations. Commenters would frequently cite specific statutes in arguing that a particular project was not in compliance with local zoning regulations. One commenter in Arlington "inquired about setbacks, the parking reduction bylaw, and whether the project would go before the Commission." An engineer in the town of Andover critiqued a developer's traffic study and stormwater analysis: "He stated that as an engineer he knows what kinds of games can be played with numbers. He gives no credibility to these counts. He added that Merrimack College traffic is not de minimus....He asked for a written report from the DPW on the impacts of proceeding with the facility." Participants in these meetings frequently displayed a high level of knowledge often derived from their own professional backgrounds—-that they 23

used when engaging in local political proceedings, consistent with our predictions. 4 Policy Impact Given the affordability and sustainability crises facing many American cities and towns, the participatory bias outlined above presents a potentially serious obstacle to change. Perhaps most importantly, our results reveal that zoning board and planning board officials are overwhelmingly hearing opposition to the construction of new housing. Often, the only voice that these public officials (and meeting attendees) hear speaking in support of new housing is the developer, whose financial stake in the project makes him poorly suited to make the case that new construction is publicly beneficial. This opposition can be persuasive. One local affordable housing lawyer we interviewed critiqued the Massachusetts system's emphasis on transparency as propagating exclusion: the towns are "controlled by older and richer people than the town as a whole, and it's bad! Under the guise of making things more transparent, we end up creating a much more exclusive system than would otherwise exist." A housing consultant recounted that, in her experience, neighbors' opposition typically resulted in money for neighbors, delay, and/or changes to the project—all of which render the project more expensive. A planning board member in a suburban MA town similarly highlighted delay as a frequent outcome of neighborly opposition: she "typically wouldn't deny a project because of public opposition, but would slow it down a lot." Another planning board official from a different town described a recent project delayed by months as a consequence of "older" opponents "concerned about parking." These delays are consequential. As another housing lawyer put it: "delay is the biggest enemy of development.... the ability of anyone to delay development is the ability to kill it." This corroborates academic work that implicates public opposition to new development as an important driver of rising housing costs (Fischel 2001). To more concretely illustrate the persuasive impact public comments have on plan- 24

ning /zoning decisions, we explore the meeting minutes of two cities in depth: Cambridge and Worcester, MA. We select these cities for several reasons. First, their meeting minutes were unusually detailed (indeed, Cambridge's minutes were exact transcriptions). Second, they are both locations where we would not necessarily expect NIMBY attitudes to prevail. Cambridge is one of the most liberal cities in the country and facing a massive affordable housing crisis; since we found a strong association between Democratic affiliation and support for new housing in our analysis of meeting minutes, we might expect local officials in Cambridge to similarly prefer a greater supply of dense housing: Worcester is one of the poorest cities in our data set; in less affluent cities, concerns about diminished tax base should, in theory, generate more official support for new residential developments (Peterson 1981). Both cities thus represent tough tests for observing a significant policy impact." Obviously, these case studies tracing the evolution of a couple of proposals cannot perfectly measure the policy impact of these land use regulations. In an ideal world, we would be able to randomize the implementation of measures encouraging public input in the zoning process, or at least observe variation in these institutions. Unfortunately, because all MA towns operate under the same zoning law mandating public input in the zoning process—and, indeed, these regulations are widespread nationally—we do not have the cross-sectional variation to measure policy impact in this way. Moreover, national-level data on land use regulations-including longitudinal data—are extraordinarily difficult to generate. The most detailed available data on land use regulations are cross-sectional MA regulations from the Housing Regulation Database, and required several years of painstaking work to assemble (Pioneer Institute for Public Policy Research and Rappaport Institute for Greater Boston 2005). Finally, any study that did look for impact via changes in projects through the meeting process would also have to account for the fact that initial proposals may reflect existing institutional contexts. We believe, however, that these case studies —while imperfect—strongly suggest that these meeting comments shape important policy outcomes. 1 In the 2010 referendum, 80% of Cambridge voters and 65% of Worcester voters opposed repealing the Chapter 40B law promoting affordable housing development. 25

In 2016, a group of neighbors attended a Cambridge Planning Board meeting in staunch opposition to a proposal seeking to convert an abandoned commercial warehouse into four residential units. Neighbors worried, among other things about "density," "insufficient parking;." "demolition," "building foundations," and that "the development is very non- compliant." Members of the planning board took these concerns very seriously, and cited them in making multiple additional demands of the developer. Planning board member Tom Sienieicz observed: This board member would find it very, very difficult tonight....in light of the input we've gotten from abutters and my review of the documents, to make findings in affirmative...It seems like there is the potential to engage in a more detailed conversation with the community to see whether....the developer can assuage the primary concerns of parking; of density, and the issue of settlement....I would also include the potential...for the Board to ask for a parking analysis or a traffic analysis. Fellow board member Ahmed Bur built on Sienieicz's concerns: "In addition to what Tom said, I would also request some sort of geotech engineering study done. More than one person mentioned houses sinking based on water." Other members of the Cambridge Planning Board largely echoed these concerns, similarly rooting them in neighbors' stated objections at the meeting. In one of the most liberal cities of the country, a group of neighbors uniformly opposed the development of new housing. The Cambridge Planning Board agreed that these concerns were valid, and suggested a variety of measures imposing significant new costs on the developer, including additional parking and geotech studies. The developer returned to the planning board three months later in January 2017, having completed both the parking and geotech study and altering his proposal in a number of ways to suit neighbors' concerns: "A number of the neighbors thought that four units was too many and whether we could actually consider having a successful project with only three, and we've come to a resolution that we are going to do that." The developer also agreed to increase the number of parking spaces per unit from one to two. Neighbors thus imposed multiple costs on the developer; geotech 26

and parking studies cost thousands of dollars. Additional months of delay similarly impose significant carrying costs. Finally, and most importantly, the developer has lost the value of an additional unit and use of space now occupied by the additional parking spots. This reduction is not only costly to the developer—it also reduces the overall housing supply in a city desperate for more housing (and likely made each of the three remaining units larger and more expensive). While one unit is obviously not going to have a significant impact on a city's overall housing supply, this process repeating itself hundreds of times starts to have a marked influence on housing availability. Moreover, anticipation of this process might deter meritorious projects from even being proposed and/or push the proposals that are made in the direction of more expensive, higher end, units to make the economics work. This policy impact is also evident in less affluent cities, where concerns about diminished tax base should, in theory, generate more official support for new residential developments. A proposed 36-unit condominium in Worcester, MA met steep neighborhood opposition at at a 2015 Worcester Zoning Board meeting. One man cited his status as a representative of the Brown Square Neighborhood Group and former zoning board member to question the legality of the proposal. The meeting minutes describe his views: "He stated that he does not believe the proposal meets the statute regulations to be considered hardship. He believes that the petition should be denied and that the developer is only looking to maximize for profitability. This does not fit in with the character of the neighborhood." Another man similarly worried about negative impacts on "neighborhood character and social structures" as well as "property values." As in our example in Cambridge, Worcester Zoning Board members were deeply concerned about neighborhood opposition. Meeting minutes described one board member's response to neighborhood opposition in the 2015 meeting concerning the development of low-rise condominiums: Mr. Abramoff [Worcester Zoning Board Chair stated that he believes that the design looks like this is an institution. The project needs to have a lot of landscaping to be more appealing. He is concerned the density is very high and 27

also about the amount of impervious area. He would like to see the applicant meet with the neighborhood again because right now there is a big gap from what is proposed to what the neighbors want. Other board members concurred in a unanimous vote. This meant that discussion of the proposal would be continued through the next meeting six weeks later, and that construction approval was delayed by a further two months. At the subsequent meeting, neighborhood opposition to the proposed low-rise condominium development remained intense, despite the developer having reduced the number of housing units from 36 to 24. This neighbor's comments perhaps most succinctly described his community's concerns: "NAME) stated that there was no compromise or agreement at the neighborhood meeting. They do not want to this type of project in the neighborhood." The board agreed: "Mi. Wanat [Worcester Zoning Board member| stated that the applicant addressed some of his concerns, but that he is concerned with this development not quite fitting in to the neighborhood and the traffic that will be due to the density. Mr. Haddon concurred." The developer opted to withdraw his proposal at this point; neighborhood opposition successfully killed the project. The fact that neighborhood opposition had such a potent impact is striking, and speaks to the generalizability of the political inequality we have document in this article. Worcester is not the sort of advantaged city frequently featured in media and academic accounts of NIMBYism. As a former industrial city 40 miles outside of Boston, Worcester has considerably lagged the Greater Boston region's explosive economic growth. It nonetheless features housing policy dynamics that would not be out of place in San Francisco or Palo Alto. These case illustrate the potential of citizens to persuade local officials; commenters have other means, however, of effecting policy. Frequent attendance at meetings also in some instances indicates citizens' willingness to pursue legal challenges against developers and/or the city/town. Multiple individuals in our data set attended meetings with lawyers of identified themselves as lawyers opposing projects in a personal capacity. In a few cases, we were able to match individuals in our data set with lawsuits filed in the Massachusetts Land Court on the development in question. Given the importance of lawsuits as a key avenue for 28

stymying development (Glaeser and Ward 2009), such implied threats (or actual lawsuits) can have a potent impact. Finally, prior research using these data shows that the most highly regulated places in MA permit the least multifamily housing (Glaeser and Ward 2009). This fact is consistent with public meetings constraining the supply of housing. In the absence of stringent land use regulations, housing developments can be constructed "by right," without necessitating any planning or zoning board meetings. In contrast, review of variance requests by these boards—in concert with public meetings—is associated with production of significantly less multifamily housing 4.1 Generalizing Beyond Massachusetts One potential limitation of our analyses is that all of our data are from one state: Mas- sachusetts. It is possible that Massachusetts' town meeting tradition and strong local zoning control (1) lead to a particularly unrepresentative set of citizens who oppose new housing development and/or (2) make housing opponents particularly impactful. While we are unable to rigorously quantify meeting participation in other states, suggestive evidence indicates that these trends hold, at least to some extent, elsewhere. First, we conducted detailed case studies of the zoning codes in six cities with widely varying institutional and regional contexts: Charleston, SC, Charlotte, NC, Los Angeles, CA, Milwaukee, WI, Phoenix, AZ, and San Francisco, CA. The zoning codes in all six cities mandate the solicitation of public input at multiple stages in the development process, confirming that analogous procedures to those in the Boston area are present elsewhere. In addition, we surveyed 115 mayors of cities over 75,000 (a response rate of 25%).!? Among other topics, we asked mayors whether they believed housing development was more influenced by "majority public opinion" or a "small group with strong views." 60% of mayors 12 We recruited mayors of all cities over 75,000 with a combination of personalized emails and phone calls. All interviews were conducted over the phone, ensuring that we spoke directly with mayors. The survey covered a wide array of topics, including climate change, federalism, and race. 29

selected "small group with strong views," and, in more qualitative elaborations, described opposition remarkably similar to that captured in our Massachusetts data. Multiple mayors mentioned dominant elderly groups, while others highlighted the impact of well-organized oppositional neighborhood associations. Interestingly, in all cases, mayors who elaborated on the "small groups" in their cities mentioned individuals/groups who opposed the construction of new housing-consistent with our finding that meeting attendees overwhelmingly oppose housing development. Finally, we highlight one case with a differing institutional and socioeconomic context: Milwaukee, WI. While NIMBYism has been well-documented in coastal cities like Boston and San Francisco, comparatively less media and scholarly attention has focused on whether opposition to higher density holds in less affluent communities with lower housing prices like deindustrializing Milwaukee which, unlike many of the Massachusetts cities/towns, is governed by a strong mayor system rather than a town meeting. Nonetheless, at least in pockets of the city, media accounts and comments from local officials suggest that an unrepresentative group of neighbors dominate public hearings in similar ways that we observe in eastern Massachusetts. On multiple occasions, after attending hearings concerning housing developments in gentrifying parts of the city, Milwaukée Mayor Tom Barrett has remarked, "I didn't realize everyone on the East Side was an architect" (Jannene 2014). An interview with a Milwaukee alderman confirmed that the mayor used this comment repeatedly and was struck by "well-informed design critiques from professors" at local community meetings. The alderman noted at his community meetings that there were "a lot of regulars" and that he "know|s| who I'm going to run into..architects and lawyers. Lawyers show up in lawyerly manner." He also believed—as we found in our limited quantitative data analysis-that a disproportionate share of meeting attendees were homeowners, not renters. Perhaps more importantly, the Milwaukee alderman—like the individuals interviewed in Massachusetts—believed that the individuals who attended these meetings had important policy impact. He noted that "the voices of abutters carry a lot of weight," in how he voted 30

on a development project and that, in some cases it "only takes one voice" to influence a project. Local political bloggers similarly highlighted cases of neighborhood opposition delaying projects by months (Jannene 2012, 2014). 5 Prescriptions for Local Democracy This paper has uncovered two related forms of bias. The first is that an unrepresentative group disproportionately participates in public meetings concerning housing development. The second is that the concentrated costs and diffuse benefits of housing development spur a group of highly affected individuals to both participate and oppose new housing. The first can potentially be addressed with measures that help to mitigate disparities in participation. In particular, policymakers could do more to include renters in the housing development process. While there is some evidence that renters express hostility towards housing development (Hankinson 2018), Marble and Nall (2017) find that renters exhibit more progressive attitudes towards new housing compared with homeowners. One way to enhance renter participation is to ensure that they are aware of developments in their community. In many Massachusetts communities, notices are mailed to property-owning abutters. In other words, notices are sent to landlords, not their tenants who actually reside in the abutting properties (e.g. Town of Arlington 2016). In many cases, then, individuals who live nearby may not even be aware of proposed housing developments. Fung (2006) notes that, for institutions of empowered participation to operate effectively, they must be structured in ways that encourage participation by all. The bias towards opposition is harder to address, in part because it is normatively murkier whether it is problematic that the most affected individuals are the most likely to participate and oppose projects. While there are broader negative societal consequences of failing to increase the supply of housing, the era of developer-dominated politics suggests that ignoring (or even not privileging) abutters' concerns is also normatively problematic. Policymakers 31

might consider restructuring public hearings to encourage greater deliberation and genuine responsiveness to participating interlocutors (Fung 2006; Gutmann and Thompson 2012). Of course, genuine deliberation requires the representation of all sides of a debate. With only 15 percent of comments in support of new housing, it is difficult to imagine a well-informed back-and-forth policy discussion surrounding many of the housing developments in many of these meeting minutes. Finally, these meetings raise important questions about the level of expertise needed to participate in public deliberation (Fung 2006). Many of the commenters exhibit a high level of specialized knowledge about local land use and zoning: On the one hand, this bias towards high knowledge could dissuade some underrepresented voices from speaking up at meetings. On the other, as a society, we may want individuals to have a base level of knowledge about local land use prior to participating in important policy debates surrounding housing: While this paper has uncovered some troubling participatory biases in public meetings, these issues do not necessarily mean that neighborhood-level politics are inherently unrep- resentative. Scholars have identified other policy arenas where these meetings do appear to significantly enhance the participation of socioeconomically disadvantaged groups (Fung 2006). Moreover, a developer-dominated system like the one that existed prior to the move- ment towards neighborhood participation is unlikely to yield significantly better outcomes in terms of affordability. Similarly, moving towards a system in which elites on zoning and planning boards wield the greatest influence may not necessarily yield greater democratic accountability; indeed, the demographic and attitudinal composition of zoning and planning board members may not be so different than that of meeting attendees. We hope that future research can build upon our findings to improve the functionality of these public meetings and that political scientists and policymakers alike can learn important lessons about implementing higher quality democracies from these meeting minutes. 32

References Abramowitz, Alan. 2010. The Disappearing Center: Engaged Citizens, Polarization, and American Democracy. Yale University Press. Barro, Josh. 2017. "Here's what mayors and governors should really do to support the Paris Agreement." Business Insider. Available at http://www.businessinsider.com/what-mayors- and-governors-should-do-to-support-paris-agreement-2017-6. Accessed on August 14, 2017. Berinsky, Adam. 2005. "The Perverse Consequences of Electoral Reform in the United States." American Politics Research 33 (4): 471-491. Berry, Jeffrey M., Kent E. Portney, and Ken Thomson. 1993. The Rebirth of Urban Democracy. Washington, D.C.: Brookings Institution Press. Burden, Barry C., David T. Canon, Kenneth R. Mayer, and Donald P. Moynihan. 2013 "Election Laws, Mobilization, and Turnout: The Unanticipated Consequences of Election Reform." American Journal of Political Science 58 (1): 95-109. Campbell, Andrea L. 2005. How Policies Make Citizens: Senior Political Activism and the American Welfare State. 2005: Princeton University Press. Capps, Kriston. 2015. "How Fair Housing Will Turn Liberal Cities Conservative." CityLab. Available at https://www.citylab.com/equity/2015/07/how-fair-housing-will-turn-liberal- cities-conservative/398642/. Accessed on August 24, 2017. Carter, Jimmy. 1980. "Housing and Community Development Act of 1980. Remarks on Signing S. 2719 Into Law.". Available at http://www.presidency.ucsb.edu/ws/?pid=45238. Accessed on August 15, 2017. Chetty, Raj, Nathaniel Herdren, and Lawrence F. Katz. 2016. "The Etfects of Exposure to Better Neighborhoods on Children: New Evidence from the Moving to Opportunity Experiment." American Economic Review 106 (4): 855-902. Commonwealth of Massachusetts. 2017. "Chapter 40A, Section 11.". Available at https: //malegislature.gov/Laws/GeneralLaws/PartI/TitleVII/Chapter40A/Section11. Accessed on August 18, 2017. Dahl, Robert. 1961. Who Governs? Democracy and Power in the American City. New Haven: Yale University Press. de Kadt, Daniel. 2017. "Bringing the polls to the people: How increasing electoral access encourages turnout but exacerbates political inequality". Available at https://www.dropbox. com/s/g9787v81pq8mu53/dekadt_polls_to_the_people.pdf?dl=0. Accessed on August 15, 2017. Einstein, Katherine Levine, David M. Glick, and Conor LeBlanc. 2017. '2016 Menino Survey of Mayors.". Available at https://www.bu.edu/ioc/files/2017/01/2016-Menino-Survey-of- Mayors-Final-Report.pdf. Accessed on August 14, 2017. 33

Einstein, Katherine Levine, and Vladimir Kogan. 2016. "Pushing the City Limits: Policy Responsiveness in Municipal Government." Urban Affairs Review 1 (52): 3-32. Enos, Ryan D. 2016. "What the Demolition of Public Housing Teaches Us About the Impact of Racial Threat on Political Behavior." American Journal of Political Science 60 (1): 123-142. Finkel, Steven E. 1985. "Reciprocal Effects of Participation and Political Efficacy: A Panel Analysis." American Journal of Political Science 20 (4): 891-913. Fiorina, Morris P. 1998. "Extreme Voices: A Dark Side of Civic Engagement." In Civic Engage- ment in American Democracy, ed. Theda Skocpol and Morris P. Fiorina. Washington,D.C: Brookings Institution Press. Fischel, William A. 2001. The Homevoter Hypothesis: How Home Values Influence Local Government Taxation, School Finance, and Land-Use Policies. Cambridge, MA: Harvard University Press. Fung, Archon. 2006. Empowered Participation: Reinventing Urban Democracy. Princeton, NJ: Princeton University Press. Gay, Claudine. 2012. "Moving to Opportunity: The Political Effects of a Housing Mobility Experiment." Urban Affairs Review 48 (2): 147-179. Gerber, Elizabeth R., and Justin H. Phillips. 2003. "Development Ballot Measures, Interest Group Endorsements, and the Political Geography of Growth Preferences." American Journal of Political Science 46 (4): 625-639. Gilens, Martin. 2014. Affluence and Influence: Economic Inequality and Political Power in America. Princeton, NJ: Princeton University Press. Caesar a Vare net, Hea thie, and tapi: New Out: Prenter Invention Makes Us Richer, Glaser, Edward L., and Bryce Ward. 2009. "The Causes and Consequences of Land Use Regulation: Evidence from Greater Boston." Journal of Urban Economics 65 (3): 265-278. Glaeser, Edward L., Joseph Gyourko, and Raven E. Saks. 2005. "Why is Manhattan So Expensive? Regulation and the Rise in Housing Prices." Journal of Law and Economics 6: Green, Donald P., Dara Z. Strolovitch, and Janelle S. Wong. 1998. "Defended Neighborhoods, Integration, and Racially Motivated Crime." American Journal of Sociology 104 (2): 372- 403. Gutmann, Amy, and Dennis Thompson. 2012. The Spirit of Comprise: Why Governing Demands it and Campaigning Undermines It. Princeton, NJ: Princeton University Press. 34

Gyourko, Joseph, Albert Saiz, and Anita A. Summers. 2008. "A New Measure fo the Lo- cal Regulatory Environment for Housing Markets: The Wharton Residential Land Use Regulatory Index." Urban Studies 45 (3): 693-721. Gyourko, Joseph, and Raven Molloy. 2014. "Regulation and Housing Supply.". NBER Working Paper 20536. Hajnal, Zoltan L., and Jessica L. Trounstine. 2016. "Race and Class Inequality in Local Politics." In The Double Bind: The Politics of Racial and Class Inequalities in the Americas, ed. Juliet Hooker and Jr. Alvin B. Tillery. Washington,D.C: American Political Science Association. Hankinson, Michael. 2013. "Externalities or Extortion? Privatizing Social Policy through Community Benefits Agreements." Harvard Journal of Real Estate: G-13. High Hankinson, Michael. 2018. "When Do Renters Behave Like Homeowners? and NIMBYism." American Political Science Review. Rent, Price Anxiety, Available at https://www.cambridge.org/core/journals/american-political-science- review/article/when-do-renters-behave-like-homeowners-high-rent-price-anxiety-and- nimbyism/72B5F7B4CFEC099E9EC4A7A1EAD8611D Accessed on April 20, 2018. Hoene, Christopher, Christopher Kingsley, and Matthew Leighninger. 2013. "Bright Spots in Community Engagement.". Available at https://www.knightfoundation.org/media/uploads/ publication_pdfs/BrightSpots-final.pdf. Accessed on August 14, 2017. Hopkins, Daniel J. 2010. "Politicized Places: Explaining Where and When Immigrants Provoke Local Opposition." American Political Science Review 104 (01): 40-60. Hsieh, Chang-Tai, and Enrico Moretti. 2015. "Why Do Cities Matter? Local Growth and Aggregate Growth.". Technical report. National Burean of Economic Research Working Paper No. 21154. Iyengar, Shanto, Simon Jackman, Solomon Messing, Nicholas Valentino, Toril Aalberg; Raymond Duch, Kyu S. Hahn, Stuart Soroka, Allison Harell, and Tetsuro Kobayashi. 2013. "Do Attitudes About Immigration Predict Willingness to Admit Individual Immigrants." Public Opinion Quarterly 77 (3): 641-665. Jacobus, Rick. 2017. "We Are All NIMBYs..Sometimes." Shelterforce: The Voice of Com- munity Development. Available at https://shelterforce.org/2017/09/12/we-are-all-nimbys- sometimes/. Accessed on September 20, 2017. Jannene, Jeramey. 2012. "East Library Replacement Project on Track." Urban Milwaukee. Available at http://urbanmilwaukee.com/2012/12/13/eyes/. Accessed on August 21, 2017. Jannene, Jeramey. 2014. "Developers Slay NIMBY Objectors." Urban Milwaukee. Avail- able at http://urbammilwaukee.com/2014/09/01/eyes-on-milwankee-developer-slays-nimby- objectors/. Accessed on August 21, 2017. 35

Kahneman, Daniel, and Amos Tversky, 1979. "Prospect Theory: An Analysis of Decision Under Risk." Econometrica 47 (2): 263-291. Kain, Bruce. 2012. Democracy More or Less: America's Political Reform Quandary. New York: Cambridge University Press. Kang; Naewon, and Nojin Kwak. 2003. "A Multilevel Approach to Civic Participation: Individual Length of Residence, Neighborhood Residential Stability, and Their Interactive Effects with Media Use." Communication Research 30 (1): 80-106. Kittilson, Miki Caul. 2016. "Gender and Political Behavior.". Oxford Researel Encyclopedia of Politics. Oxford University Press. Available at http://politics.oxfordre.com/view/10. 1093/acrefore/9780190228[phone removed]/acrefore-9780190228637-e-71?print=pdf. Accessed on August 22, 2017. Lassen, David Dreyer. 2005. "The Effect of Information on Voter Turnout: Evidence from a Natural Experiment." American Journal of Political Science 49 (1): 103-118. Lassen, David Dreyer, and Soren Serritzlew. 2011. "Jurisdiction Size and Local Democracy: Evidence on Internal Political Efficacy from Large-scale Municipal Reform." American Political Science Reviews 105 (2): 238-258. Logan, John R., and Gordana Rabrenovic. 1990. "Neighborhood Associations: Their Issues, Their Allies, and Their Opponents." Urban Affairs Review 26 (1): 68-94. Logan, John R., and Harvey L. Molotch. 1987. Urban Fortunes: The Political Economy of Place. Berkeley, CA: The University of California Press. Lupia, Arthur, and Anne Norton. 2017. 'Inequality is Always in the Room: Language and Power in Deliberative Democracy." Daedalus 156 (3): 64-76. Mansbridge, Jane. 1980. Beyond Adversary Democracy. New York: Basic Books. Marble, William, and Clayton Nall. 2017. "Beyond 'NIMBYism': Why Americans Sup- port Affordable Housing But Oppose Local Housing Development.". Available at https: //www.dropbox.com/s/bdo3riekpnywgot/MarbleNall_NIMBYPaper.pdf?dl=0. Accessed on August 15, 2017. Michels, Ank, and Laurens De Graaf. 2010. "Examining Citizen Participation: Local Partici- patory Policy Making and Democracy." Local Government Studies 36 (4): 477-191. National Low Income Housing Coalition. 2017. "Out of Reach.". Available at http://nlihc. org/sites/default/files/ooi/OOR_2017.pdf. Accessed on August 14, 2017. Oliver, J. Eric. 2001. Democracy in Suburbia. Princeton, NJ: Princeton University Press. Oliver, J. Eric, Shang E. Ha, and Zachary Callen. 2012. Local Elections and the Politics of Small-Scale Democracy. Princeton, NJ: Princeton University Press. 36

Paul, Lauren Gibbons. 2015. "Not in Newton's Back Yard: Affordable Housing in Newton." Boston Magazine. Available at http://www.bostonmagazine.com/property/article/2015/ 02/24/affordable-housing-in-newton/. Accessed on August 24, 2017. Peterson, Paul E. 1981. City Limits. University of Chicago Press. Pew Research Center: 2014. 'The 'U-Shape of Political Activism: Higher at Ideological Ex- tremes, Lower in Center.". Available at http:// www.people-press.org/2014/06/12/section-5- political-engagement-and-activism/pp-2014-06-12-polarization-5-01/. Accessed on August 15, 2017. Pioneer Institute for Public Policy Research, and Rappaport Institute for Greater Boston. 2005. "Massachusetts Housing Regulation Database.". Prepared by Amy Dain and Jenny Schuetz. Available at http://www.masshousingregulations.com/dataandreports.asp. Accessed on March 2, 2017. Quigley, John M., and Larry Rosenthal. 2005. "The Effects of Land Use Regulation on the Price of Housing: What Do We Know? What Can We Learn?" Cityscape 81 (1): 69-137. Rae, Douglas W. 2004. City: Urbanism and Its End. New Haven, CT: Yale University Press. Rothstein, Richard. 2017. The Color of Law. New York: W.W. Norton. Sanders, Lyn M. 1997. "Against Deliberation." Political Theory 25: 347-76. Schleicher, David. 2013. "City Unplanning." Yale Law Journal 122: 1672-1736. Schlozman, Kay, Sidney Verba, and Heury E. Brady. 2012. The Unheavenly Chorus: Unequal Political Voice and the Broken Promise of American Democracy. Princeton, NJ: Princeton University Press. Semuels, Alana. 2017. "From 'Not in My Backyard' to 'Yes in My Backyard'." The Atlantic. Available at https://www.theatlantic.com/business/archive/2017/07/yimby-groups-pro- development/532437/. Accessed on April 22, 2018. Shingles, Richard D. 1981. "Black Consciousness and Political Participation: The Missing Link." American Political Science Review 75 (1): 76-91. Stone, Clarence N., and Robert P. Stoker: 2015. Urban Neighborhoods in a New Era. Chicago: University of Chicago Press. Tausanovitch, Chris, and Christopher Warshaw. 2014. "Representation in Municipal Govern- ment." American Political Science Review 108: 605-641. Tighe, J. Rosie. 2010. "Public Opinion and Affordable Housing: A Review of the Literature." Journal of Planning Literature 25 (1): 3-17. Town of Arlington. 2016. "19R Park Avenue Abutters' List - 300 Feet From Site.". Available at http://www.arlingtonma.gov/home/showdocument?id 33197. Accessed on August 21, 2017. 37

Trounstine, Jessica. 2016. "Segregation and Inequality in Public Goods." American Journal of Political Science 60 (3): 709-725. White House. 2016. "Housing Development Toolkit.". Available at https://www.whitehouse. gov/sites/whitehouse.gov/files/images/Housing_Development_Toolkit/20f.2.pdf. Ac- cessed on March 2, 2017. Wong, Weihuang. 2018. "Our Town: Support for Housing Growth When Localism Meets Liberalism.". MIT Political Science Department Research Paper No. 2018-11. Available at https://papers.ssrn.com/sol3/papers.cf?abstract_id=3173813. Accessed on June 19, 2018. Yglesias, Matthew. 2017. "A promising new coalition looks to rewrite the politics of urban housing." Vor. Available at https://www.vox.com/policy-and-politics/2017/7/24/16010720/ de-comprehensive-plan. Accessed on April 22, 2018. 38

Appendix Comment Coding Every time a public participant at a zoning or planning meeting was identified by name and address, and spoke about a project that implicated multiple housing units, we coded a) their information, b) information about the address of the project they spoke about, c) whether they were supportive, neutral, or opposed and, when they gave reasons or asked questions about topics that fit into one of our 20 categories. The two major coding decisions were a) how to code the participant's tone and b) how to code their reasons. Tone The support/neutral/oppose variable is coded support or oppose if the coder can detect any hint in either direction. Most supportive comments were quite explicit and included phrases such as "I support this project," and "this is good for the town" Oppose comments fell into two categories. Some explicitly expressed opposition in general: "this is bad for the town," "I'm opposed to this project." Other comments coded "oppose" focused on specific reasons (see below) with a negative tone or valence: "I'm worried about traffic," "it will make the street more dangerous," or "it doesn't fit the neighborhood." Comments coded neutral were generally sincere, or at least neutrally phrased questions. Asking "How will this affect the wildlife" would be coded neutral. Many of these neutral comments likely came from skeptical or even opposed residents who couched their views in a formally neutral question. We coded these as neutral rather than try to guess or assume why they were asking about things with a negative valence. This should make the coding reasonably conservative. Content When possible, we coded the substance of each commenter using the scheme depicted in Table 5. We allowed for multiple content areas per commenter such that a person who raised both traffic and environmental concerns would get both comment codes. Data Matching We matched the commenter data to a Massachusetts voter file from the voter data firm Nation Builder. For each comment, the only available fields to identify the commenter were their town, name, and address. We used probabilistic string matching on names and addresses using the Stata reclink2 package. We manually reviewed each match to eliminate false matches. We matched commenters to the voter file using three different combinations of the available fields. In all combinations, we required that the voter's mailing address town corresponded to the town of the meeting. 1. First name, last name, address, town: 94% of matches 2. First name, last name, town: 5% of matches. Each match reviewed to verify that first name differences were due to plausible nicknames or middle names. 3. First name from commenters to Middle name from voter file, address, town: 1% of matches. A small number of matches (10) were rejected because the commenter matched to multiple people in the voter file. Most often, one commenter matched to a father and son with the 39

same name and address. Without suffixes or middle initials, we were unable to ditterentiate between these pairs. 40

Table 5: Comment issue coding scheme Density Arguments that the new development will make the population too dense in the area The building will be too tall/short and will cast unacceptable shadows. Height/Shadows Includes arguments about wind from the building (often a result of the height) Too much strain on parking, proposal doesn't account for enough parking: Parking Traffic Vehicular traffic only (not pedestrian) Arguments that the development will harm improve/influence the quality Schools of the local public schools Arguments about the development increasing housing prices, including Affordability affordable housing, etc. includes income diversity Arguments about impact on diversity. Includes disabilities handicap Diversity accessible) Construction may lead to flooding either during or after. Project may Flooding affect drainage Building Foundation Construction will damage the foundation of neighboring buildings Noise Construction causing noise or the development making the area noisier New housing too close with views into property and other related concerns Privacy Trees/Green Arguments about trees, parks, green space, wildlite, and environmental impact, includes air pollution concerns Space/Environment Aesthetics "It's ugly" "it doesn't match the other buildings" "building doesn't fit" Includes arguments about visual and historic character of area. Not compliant with Complaining the development does not comply with zoning laws (often argue that zoning laws are agreed to after a collective participatory process, zoning therefore should not be ignored) Raises safety concerns about children, snow removal, intersections etc. Safety Pedestrian Includes pedestrian/bicycle traffic. Also sidewalk issues To show difference between density and explicit fears of socioeco- Neighborhood Charac- ter nomic/racial diversity, arguments about preserving history and questions of "fit" that are not about the building itself. Concerns about who will be moving into the neighborhood and using neighborhood resources; ai- guments that this is a "great addition to the neighborhood." Arguments about "changing" the neighborhood Includes arguments about a development decreasing property values and Home value/city rev- "hurting my property values" or questions about enues reducing city revenues, whether a property will be a "net financial gain for the city'" Septic/ water system Only applies to suburbs without sewer systems. Comnents about unethical dealings, corrupt officials, developers cheating Corruption residents. Requires more than saying that developers have not listened to residents 41

Table 6: Top 10 Reasons Given by Position Taken Neutral Oppose Support Traffic (23.1%) Aesthetics (11.1%) Environment (14.3%) Environment (18.6%) Density (9.7%) Septic/Water (8.2%) Flooding (7.0%) Flooding (11.9%) Affordability (9.5%) Safety (14.8%) Environment (9.3%) Density (11.9%) Neighborhood Character (6.9%) Aesthetics (5.6%) Aesthetics (11.9%) Parking (5.6%) Parking (4.2%) Septic/Water (10.9%) Pedestrian Impact (3.5%). Traffic (5.3%) Neighborhood Character (10.5%) Home Values/City Finances (5.3%) Safety (3.4%) Pedestrian Impact (5.0%) Non-Compliance (3.3%) Parking (9.9%) Non-Compliance (7.1%) Home Values/City Finances (3.2%) Diversity (5.0%) 42

AttaChment E "Living in inclusionary housing has allowed me to live in the city that I grew up in. This in turn has allowed me to raise my children in a place that I hold dear to my heart. If it weren't for this program I would have never been able to afford to stay in my city." - Inclusionary Housing Renter 2018 Inclusionary Housing Report Community Development Department City of Cambridge September 2018 IRAM FAROOQ Assistant City Manager for 344 Broadway Community Development Cambridge, MA 02139 Voice: [phone removed] SANDRA CLARKE Fax: [phone removed] Deputy Director TTY: [phone removed] Chief of Administration www.cambridgema.gov KHALIL MOGASSABI Deputy Director Chief of Planning

TABLE OF CONTENTS 3 Introduction 4 Part I: Production Report 4 Affordable Housing Stock 6 Inclusionary Projects Complete and In Development Activity Since 2017: Complete and Approved Inclusionary Projects 8 Units by Bedroom Size - Current and Changes in Past Year. 10 Middle Income (80%-120% of AMI). 11 Development Pipeline.. 11 Conclusion. 12 Map of Inclusionary Units. 13 Part II: Inclusionary Rental Program 14 Overview of Complete Rental Units as of September 1, 2018 15 Characteristics of Households Living in Inclusionary Housing. 23 Characteristics of Recently Housed Tenants. 26 Summary of Property Management Review. 28 Middle Income Program 30 Part III: Homeowners.. Attachment: Summary of the Rental Applicant Pool, 5/1/18

INTRODUCTION This report presents the current status of the City of Cambridge's Inclusionary Housing Program. Since the beginning of the Inclusionary Housing Program in 1998, 1102 affordable units have been successfully created through the inclusionary zoning ordinance or through similar zoning provisions. This includes 891 rental units and 211 homeownership units. In April 2017, the City Council adopted revisions to the inclusionary provisions in the zoning ordinance. The new provisions call for an annual review and report on the Inclusionary Housing Program. This is the first annual report since the enactment of the new provisions. The report provides information on inclusionary housing production, the status of the inclusionary rental housing program, and information on the homeownership program. The report is divided into three parts: Part I - Production: This section details the status of all inclusionary units - both rental and ownership. It describes the current status of inclusionary housing production and describes the activity in the program since the adoption of the revisions to the Ordinance in April 2017. Part II - Rental Program: A report on the Rental Program was provided to the Housing Committee in April 2017. This section provides updates to the 2017 report and includes information on the status of applicants to the rental program and on the current tenants in inclusionary units. Part III - Homeownership: This section provides information on the resales and tenure of inclusionary ownership units. The changes to the inclusionary zoning ordinance approved in April 2017 were the first changes in nearly 20 years. These changes, effective as of December 1, 2016, include: • The set aside for affordable units is now based on a percentage of residential net floor area rather than a percentage of units. The set aside for affordable units is 20% of total dwelling unit net floor area after the application of the inclusionary zoning density bonus (for a transition period from 12/1/16 to 6/30/17, the set aside for affordable floor area was 15% of total net residential floor area). Previously the set aside was calculated at 15 % of the units before the application of the bonus, which typically resulted in 11-12% of the units as affordable. A monetary contribution is required if the dwelling unit net floor area designated for affordable units is less than the required 20% of floor area and the remaining floor area is not enough for a unit. • The monetary contribution is based on the amount of subsidy needed to create equivalent floor area in an affordable building funded by the Cambridge Affordable Housing Trust - currently $397 per square foot. • The inclusion of three-bedroom units is required in larger projects (over 30,000 square feet). For smaller projects, in which there are any three-bedroom units, an affordable three-bedroom unit is required. • Income eligibility for rental units remains at 80% of Area Median Income (AMI). The income eligibility for homeownership is increased to 100% of AMI. Ownership units will be priced at 30% of the income of a household at 90% AMI. Rents for all units were based on 30% of a household's gross income. The rent for studio units under the revised ordinance is based on 25% of a household's gross income. Rents for all other units will continue to be calculated based on 30% of a household's income. Tenants' incomes continue to be recertified annually; however, households remain eligible so long as their income is less than 100% AMI. • If a household's income falls below 40% of AMI when they are recertified, they will pay a minimum rent which is equivalent to the affordable rent for a tenant at 40% AMI. 3

PART I: PRODUCTION REPORT Production of new affordable housing through the inclusionary housing program has continued at a rapid pace over the last year. This new production includes developments approved both under the revised inclusionary provisions and those in effect prior to the 2017 amendments. Although the new provisions were adopted in April 2017, with an effective date of December 1, 2016, there continues to be a pipeline of developments that were permitted prior to the effective date of the new requirements and have yet to get approvals for their inclusionary units or sought a building permit. It is expected that these developments will continue to be added to the inclusionary housing portfolio over the next two or more years along with those subject to the new provisions. The information on inclusionary housing production which follows looks at inclusionary housing developments in two time frames: analysis of the totality of inclusionary developments to date (and those in the pipeline) and data on the changes in production since April 2017 when the revised ordinance was adopted. Data is provided on: the status of the affordable housing stock, specifically the inclusionary housing developments; the bedroom size of inclusionary housing units; and the pipeline of inclusionary developments. HOUSING STOCK Affordable Housing Stock: Inclusionary housing is one of the ways that affordable units are created in Cambridge. Below is a chart showing the different types of affordable housing in Cambridge as of June 30, 2018. In the past few years, most of the new affordable housing units in the City have been created through inclusionary housing. However, affordable housing owned and operated by the Cambridge Housing Authority and non-profit and private affordable housing providers makes up most of thé City's affordable housing stock. Affordable units make up 14.84% of the 54,713 units of housing in the City. This inventory is updated at the end of each fiscal year. Cambridge Affordable Housing Stock (as of 6/30/2018) Total=8,117 Scattered Site FTHB, 99, 1% Privately-Owned Housing, 1,264, 16% Non-Profit Sponsored Inclusionary Housing, 2,937, Housing, 1102, 36% 14% Public Housing, 2,715, 33% • Public Housing • Non-Profit Sponsored Housing - Privately-Owned Housing • Inclusionary Housing • Scattered Site FTHB Figure 1: Cambridge Affordable Housing Stock, as of 6/30/17).

Housing Starts: The following table shows the number of units for which building permits were issued for new housing starts in all housing from 2015 to the first half of 2018. This includes building permits for projects of all sizes-from single-family homes to multi-family residential projects with hundreds of units. Larger market rate developments with inclusionary units are included in these total as are new units built in all affordable developments. Calendar Year # of Permits 485 2[phone removed] 2017 [phone removed] - First six months Table 1: Housing starts by year 5

INCLUSIONARY PROJECTS COMPLETE AND IN DEVELOPMENT All Inclusionary Units: As of September 1, 2018, there are 1,102 affordable units approved under inclusionary zoning or similar zoning requirements. Of these, currently 935 units are complete and an additional 167 units are under construction. Fourteen of the units under construction will be ready for occupancy within the next two months. Approved Units - Complete and In Development 9/1/2018 1000 935 900 196 800 700 600 500 400 739 300 200 100 152 Complete In Development • Rental • Homeownership Figure 2: Approved Units, completed and in develop: ment as of 9/1/18 Growth: The chart below shows the growth of the inclusionary housing program since its inception in 1998 and the projected growth for the next three years based on developments currently in the pipeline. The majority of the growth in the program since 2011 is due to an increase in the development of rental units. Inclusionary Housing Stock Growth 1800 1660 1600 150g,' 1337 , 1400 1425 1102, 1200 Rental 1265 1126 1000 Ownership 800 891 600 -Total 235 400 211 235 211 200 0 FY17 FY02 FY05 FY07 FY06 FY09 FY10 FY19 Figure 3: Inclusionary Housing Stock Growth, current and projected. 6

ACTIVITY SINCE 2017: COMPLETE AND APPROVED INCLUSIONARY PROJECTS Developments and Units: Inclusionary housing units are counted when they are approved, just prior to the issuance of a building permit. From April 1, 2017, when the revised inclusionary ordinance was approved, through September 1, 2018 , seven new inclusionary housing projects have been approved and are currently in development. These seven projects include a total of 127 affordable units, 123 rental units and 4 homeownership units. During the same time period, 56 new rental and 3 new homeownership units were complete. Units Complete and In Development 1000 891 900 152 768 800 700 143 600 500 400 739 625 300 211 207 200 14 100 196 193 0 Rental Rental Homeownership Homeownership As of 9/1/2018 As of 4/1/2017 • Complete = In Development Figure 4: All units complete and in development, 2017 and 2018 Of the approved projects, the revised inclusionary provisions apply to three projects - 47 Bishop Allen Drive, 195 -211 Concord Turnpike, and 1699 Massachusetts Avenue. These are all rental projects which were permitted between December 1, 2016 and June 30, 2017. Therefore, 15% of the net residential square feet is devoted to affordable units. These 3 projects include a total of 49 new affordable units. The total size of the approved projects ranged from 17 units to 320 units. As the inclusionary requirement is now calculated as a percentage of net residential square feet in order to increase the production of family size units, the resulting percentage of affordable units is less than the percentage of affordable square feet in a development. In the three developments cited above there are 49 affordable units are out of a total of 360 units, or 13.6% of the units. It is expected that the percentage of units in developments which will be subject to the 20% net residential square foot requirement will be between 17%- 18% of all units. Monetary Contribution with New Ordinance: The net residential square feet of affordable units will not always add up to the required percentage of floor area. The aggregate amount of inclusionary floor area will be as close as possible to the required amount, while also meeting other requirements for the inclusionary units, such as the provision of three-bedroom units in certain developments, location of units, and bedroom size distribution. To make up this difference, a monetary contribution to the Affordable Housing Trust is required for the difference between the required residential floor area and the residential floor area of the inclusionary units. The contribution rate is currently $397 per square foot. The contribution for the three approved projects will amount to a combined total of $32,633. When the project is nearing completion, this amount will be confirmed and due prior to the issuance of a Certificate of Occupancy for each building. 7

Homeownership Increase: Although new rental housing development has continued over the last several years and, as a result, the most significant increase in the number of inclusionary units is through rental developments, we have seen a few homeownership developments in the last two years. For three years, from March 2013 to April 2016, no new homeownership projects were produced. Since April 2016, 4 new developments with a total of 17 affordable homeownership units have been approved. UNITS BY BEDROOM SIZE-CURRENT AND CHANGES IN PAST YEAR Family Sized Units-Changes Under New Ordinance: While the prior inclusionary housing provisions required affordable units to be provided in proportion to the bedroom size of all the units in the project, a goal of the 2017 amendments was to increase the proportion of family sized units and reduce the proportion of studio units. The revised inclusionary ordinance contains provisions to promote family sized units, which are defined as units with three or more bedrooms and at least 1,100 square feet of floor area. For buildings with more than 30,000 square feet of net residential floor area, a family sized unit is required for every 6,000 square feet of inclusionary housing floor area. Applying this requirement at 195-211 Concord Turnpike, resulted in six of the twelve, or 50%, of the three-bedroom units in the project, being designated as affordable. Although 23% of all the units in the property are studio units, only 8% of the inclusionary units are studios. For smaller projects, the proportion of affordable family sized units to all affordable housing units must be equal to or greater than the ratio of all family sized market rate units to all market rate units. For example, one of the inclusionary developments to which the new inclusionary provisions apply has a total of 17 units with two family sized units (three or more bedrooms). The result is that at least one affordable unit must be a family sized unit, while the previous inclusionary provisions would not have required any affordable family sized units for this project. The following chart compares the percentage of rental units approved by bedroom size under the prior ordinance and under the revised ordinance. Bedroom Sizes of Rental Units Created Before and After Ordinance Revisions 50% 45% 45% 43% 40% 34% 29% 25% 20% 16% 16% 15% 12% 10% 5% 5% 0% 3+ BR Studio 2BR 1 BR • Units Under New Ordinance (49 units) •i Units Under Prior Ordinance (842 units) Figure 5: Bedroom sizes of rental units created before and after the ordinance revisions. 8

Changes in Number and Percentage of Units by Bedroom Size Since April 1, 2017: The chart below illustrates the change in the number and percentage of approved rental and ownership units by bedroom size, since April 2017. This includes developments which are complete and those in construction. It also includes middle income units. Distribution of Rental and Homeownership Unit Sizes, 2017 and 2018 891 769 207 211 100% 6%, 53 5%, 36 17%, 35 17%, 35 90% 80% 35%, 271 34%, 300 70% 48%, 99 48%, 101 60% 50% 46%, 357 45%, 400 40% 30% 33%, 68 33%, 70 20% 10% 15%, 14%, 138 105 2%, 5 2%, 5 0% Rental 2017 Rental 2018 Home Ownership 2017 Home Ownership 2018 • 3+ Bedroom 2 Bedroom • Studio 1 1 Bedroom Figure 6: Rental Homeownership Units, 2017 and 2018. Three-bedroom units: Significantly. there are currently 30 completed rental units with three or more bedrooms. This number will significantly increase to 53, as 23 three-bedroom rental units are now under construction. This includes units in developments that are subject to the old or the new inclusionary housing provisions. Homeownership: There have not yet been new homeownership developments approved under the 2017 amendments to the inclusionary zoning provisions. Homeownership projects have always had more three- bedroom units than rental projects. Three-bedroom units account for 17% of all inclusionary homeownership units. Studio Units: The impact of the change to rents for studio units in the revised ordinance, which reduces the rent from 30% to 25% of a household 's income, will not be evidenced until the first studio units built under the revised ordinance are complete in approximately 18-24 months. 9

MIDDLE INCOME (80-120% OF AMI) Middle income units have been required to be included in some rental projects. There is one occupied project with middle income units at 270 Third Street. There are additional projects which require middle income units in the pipeline. Two buildings are currently under construction: 249 Third Street with two middle income units and Mass & Main with nine middle income rental units. In addition, the second residential building to be developed by Alexandria also requires middle income units. The unit distribution is currently under review and is estimated to have 25 to 29 middle income units. Finally, the special permit for 40 Thorndike Street (the Sullivan Courthouse redevelopment) requires 8 of the 24 residential units be middle income (along with 8 low and moderate-income units). Middle Income Rental Units 70 58 60 • Anticipated 50 32 40 • Approved 30 11 20 • Existining 10 15 0 Number of Units Figure 7: Middle Income Rental Units 10

DEVELOPMENT PIPELINE The pipeline of inclusionary housing developments has remained active over the past year. This is a summary of the projects that are currently permitted and expected to be developed in the future. Several developments in the pipeline were permitted prior to the enactment of the new inclusionary provisions and therefore remain subject to the former inclusionary housing provisions in the ordinance. Developments Under Inclusionary Review: Four inclusionary developments are currently under review by CDD staff and are expected to be approved in the coming months providing 72 affordable rental units. Of these, one inclusionary development, 605 Concord Avenue, was granted a special permit in December 2016 and is subject to the new inclusionary provisions in the ordinance at 15% requirement. Permitted Pipeline Projects: There are several projects with approved special permits, including multi- phased planned unit development (PUD) projects, that are expected to result in applications for inclusionary housing approval over the coming months. Three developments with approved special permits will be subject to the new inclusionary provisions with a 20% affordability requirement. These are 50 Cambridge Park Drive with 294 approved rental units; 1043-1059 Cambridge Street with 18 approved rental units; and 55 Wheeler Street with 525 approved units. Discussions have begun on the inclusionary unit requirements for 50 Cambridge Park Drive. The development at 55 Wheeler Street would become the biggest producer of new inclusionary housing. Other projects which are permitted and in the pipeline are a mid-sized building in North Cambridge on Elmwood Street, the next residential building in North Point (Parcel I) and MIT's residential building. Pre-approval Projects: In addition, there continue to be proposals for new residential projects submitted for approval to the Planning Board. Currently, the projects seeking approval include a small homeownership project and a mid-sized rental project. There are additional developments which are preliminary and have not yet submitted applications to the Planning Board. CONCLUSION Over the last year, residential projects have moved forward under both the old and new inclusionary housing provisions and new residential projects are continuing to be approved and proposed. While three projects have been permitted under the new inclusionary provisions, it will take some time to fully assess the impact of the revised ordinance. One early finding is that the new ordinance is resulting in a greater proportion of larger, family size units. In addition, there continues to be a pipeline of residential projects that will create new affordable inclusionary units throughout the city. 11

Fresh Pond BL Complete 1-4 5 - 15 31 - 58 Figure 8: Map of Inclusionary Unit Locations by status, September, 2018 1-4 5 - 15 31 - 58 16 - 30 In Development by Location and Status Number of Inclusionary Units 31 - 58 16 - 30 Under Review Cambridge, Massachusetts Inclusionary Unit Locations 0.25 Miles Charles River 12

PART II: INCLUSIONARY RENTAL PROGRAM The last overview of the Inclusionary Rental Program was completed in May 2017 and provided data on tenants housed during the period of April 1, 2016 to March 31, 2017. This report summarizes the program's data on recently housed tenants up to April 1, 2018 and compares tenant characteristics as of that date to April 1, 2017. The report also looks at approval and denial rates of applications from 2010 to April 1, 2018, and data on the middle-income portfolio. As for applicants, a detailed summary report of the applicant pool as of May 1, 2018 is attached. The waiting pool consists of 3,075 applicants. Of these, 767 are Cambridge residents, 361 non-residents who work in Cambridge and 1,947 non-resident applicants. Of the767 resident applicants, 270 hold a Section 8 Mobile Voucher through the Cambridge Housing Authority. In addition to working with applicants and tenants, the Housing Division maintains a continuous relationship with the property managers of inclusionary buildings. Staff works with the property management staff on all lease-ups, annual recertifications, and other issues as may arise. CDD meets with property managers regularly to review the goals and requirements of the program, train new staff, and to respond to concerns and resolve identified issues. 13

OVERVIEW OF COMPLETE RENTAL UNITS Since April 2017 many developments which were in construction have been completed. As of September 1st 2018, the City's Inclusionary Rental Program consists of 736 completed rental units, an increase of 114 units. This includes 15 units which are affordable to middle income households. (There are three rental units which are included in the Production section that are not included in this section, because those three units are owned and managed by the Cambridge Housing Authority, not CDD). As the tables below illustrate, more than half of the low-moderate completed units are studios and one-bedrooms. About 40% of the completed units are two bedrooms or larger. Though there has not been an increase in completed three-bedrooms over the last year, we expect to see an increase in family sized units in the future because of the new inclusionary provisions in the zoning ordinance. Completed Units by Bedroom Size, 2018 OBR 2BRs 3BRs 4BRs Total 1BR 1 100 721 341 250 29 Low-Mod Program 100% 4% 0.1% 35% 14% 47% 15 0 0 8 7 Middle Income Pro- 0% 0% gram 100% 0% 47% 53% 736 29 1 100 Total 258 348 Completed Units by Bedroom Size, 2017 OBR 3BRS 4BRs 1BR 2BRs Total 221 25 287 73 1 607 Low-Mod Program 0% 100% 36% 47% 4% 12% 8 Middle Income 15 7 Program 0% 100% 53% 0% 47% 0% 1 73 25 622 Total 229 294 Tables 2 and 3: Total Completed Units by Bedroom Size, 2017 and 2018 14

CHARACTERISTICS OF HOUSEHOLDS LIVING IN INCLUSIONARY HOUSING The data on households living in inclusionary units is analyzed as of April 1, 2017 and April 1, 2018. During this time 15 new tenants were added. Household Size: Since 60% of complete inclusionary units are studio and one bedrooms, it follows that most households (58%) in the program are single person households. Nineteen percent of households are two-person households and the remaining 23% are three or more-person households. There is little change since last year. HH Sizes - Occupied Units, 2018 Bedroom Total 3pp 4pp 1pp 5pp брр 2pp Size Studio 72 1*** 73 0 1 270 20 0 290 1 2 12* 71 37 90 6 217 3 1*** 0 3 10** 24 5 5 4 0 0 1 1 Total 354 42 112 5 81 605 11 Percent 2% 19% 59% 1% 13% 7% 100% * This includes residents with live-in aides or reasonable accommodations **This includes families with children of the opposite sex or with ages 10 years or more between them ***This includes families who have grown or shrunk in size while in the program Data does not include middle income units HH Sizes - Occupied Units, 2017 Bedroom Total 4pp 5pp 3pp 1pp 2pp 6pр Size 0 Studio 67 67 1 265 18 283 13* 93 3 • 36 70 215 3 0 0 11** 3 24 4 6 4 0 0 1 1 81 111 345 Total 4 590 40 9 7% 19% 2% 58% 14% Percent 100% 1% *this includes residents with live-in aides or reasonable accommodation **this includes families with children of the opposite sex or with ages 10 years or more between Data does not include middle income units Tables 4 and 5: Occupied Units by Household Size, 2017 and 2018. Does not include Middle Income Program 15

Families with Children: The tables below show the number of households with children. One hundred and ninety-four families in the portfolio have at least one child in their household. Of these families, 103, or 53%, have at least one child under 6. Ninety-one, or 47%, of families have children who are between the ages of 6 and 18. Eighty-seven percent of families with children reside in a two-bedroom apartment. The program's preference system prioritizes households with children under 6 followed by a preference for households with children under 18. Families with Children by Bedroom size, Occupied Units, 2018 Percent Total Bedroom Size Children Under 6 Children Under 18 0 1 Studio 1% 1 0 1 0 0% 2 169 76 93 87% 9 12% 14 23 0 4 1 1% 1 Total 103 100% 91 194 Percent 47% 53% 100% Data does not include middle income units Families with Children by Bedroom Size, Occupied Units, 2017 Bedroom Size Children Under 6 Children under 18 Percent Total 0 0 0 0% Studio 1 0 1 1 0% 120 176 88% 56 3 14 11% 23 9 4 0.5% 1 1 Total 100% 201 135 66 Percent 67% 33% 100% Data does not include middle income units Tables 6 and 7: Families with Children by Bedroom Size, 2017 and 2018. Does not include Middle Income Program 16

Children in the Portfolio: These tables show the total number of children in the portfolio. In 2018, there are 304 children under 18 in the rental portfolio. Forty-one percent are under the age of 6 and 59% are between 6 and 18 years old. Eighty-one percent of children live in two-bedroom units. In 2017 there were 310 children under 18 in the rental portfolio. There has been a slight decrease in the number of children in the portfolio since 2017 (310). This is in part due to children aging in the portfolio. Total Number of Children by Bedroom Size, Occupied Units, 2018 Total Bedroom Size Children Under 6 Children Under 18 Percent 1 1 0% Studio 0 1 0% 109 81% 137 246 3 18% 14 55 41 4 1% 2 2 Total 124 100% 304 180 Percent 100% 59% 41% Data does not include middle income units Total Number of Children by Bedroom Size, Occupied Units, 2017 Bedroom Size Total • Children Under 6 Children under 18 Percent 0 Studio 0% 1 1 0 1 0% 80% 147 101 248 20 18% 57. 37 4 1 4 1% 3 Total 168 100% 310 142 Percent 54% 46% 100% Data does not include middle income units Tables 8 and 9: Total Number of Children by Bedroom Size 2017 and 2018. Does not include Middle Income Program 17

Resident Income Levels (not including voucher holders): When a household initially leases an inclusionary unit their gross income must be between 50% and 80% of the AMI or they must have a housing voucher. (A tenant with a mobile housing voucher may have an income lower the 50% AMI). Upon annual recertification of non-voucher tenants, incomes change and are sometimes below 50% or above 80% of AMI. Tenants whose income falls below 50% of AMI remain eligible for their unit. Tenants whose income rises above 80% AMI remain eligible so long as their income does not exceed 100% of AMI. For these reasons, there are variations in the income levels of tenants who do not have vouchers. Most non- voucher residents still earn between 50% to 80% of AMI. There has not been a significant change to the proportion of tenants in this income range from 2017 to 2018. The percent of residents below 40% AMI also has not significantly changed. The significant changes are in the 40% to 50% AMI range, with an increase in households and a corresponding decrease in tenants with income over 80% AMI. Please note this chart does not include data from the middle-income program. AMI Levels, All Non-Voucher Holders 8% 12% 1% 2018 10% 69% 12% 3% 2017 8% 8% 68% • 80%-100% • > 100% 50%-80% • 40%-50% • <40% Figure 9: AMI Levels for all non-voucher holders, 2017 and 2018., 18

Voucher Utilization in Inclusionary Housing: The level of voucher utilization, as shown in the charts below, has remained the same between 2017 and 2018. Households with vouchers account for more than half of the portfolio. Both 2017 and 2018 voucher utilization rates represent an increase from previous years. The percentage of residents with vouchers in inclusionary units was 45%. In 2016 that percentage rose to 53%, followed by 56% in 2017 and 2018. For the last two years, 87% of all voucher holders in the portfolio have received their voucher from the Cambridge Housing Authority. Voucher Utilization, 2018 295, 87% 338, 267,44% 56% 43, 13% • Non-Voucher Tenants • CHA Other Voucher Voucher Utilization, 2017 287, 87% 259, 331, 44% 56% 44, 13% = Other Voucher • Non-Voucher Tenants • CHA Figures 10 and 11: Voucher Utilization rate, 2017 and 2018. 19

Length of Tenancy: Length of Tenancy: The chart below demonstrates that there has been a decrease in the households who have moved into their units in the last 24 months (49% in 2017 and 30% in 2018). It follows that there has been an increase in the length of time that households are staying in their units. The percentage of households in their units for two to five years has increased from 24% to 45 %. The percentage of tenants staying over five years has remained steady. Length of Tenancy 50% 45% 40% 35% 29% 30% 26% 25% 20% 20% 19% 19% 17% 17% 14% 15% 13% 10% 9% 10% 8% 5% 0% 2017 2018 • <12 Months • 12-24 Months 24-36 Months • 36-60 Months • 60-120 Months • 120+ Months Figure 12: Length of Tenancy, 2017 and 2018. Does not include the Middle Income Program Rental Turnover Rate: The chart below illustrates the number and percentage of occupied units which became vacant and leased to new households during each fiscal year during the past 5 years. In FY 2014, 11.2% of units turned over, followed by 10.8% in FY 2015 and 12.9% in FY 2016. In FY 2017, the turnover rate dropped to 6.8%. Inclusionary Rental Turnover Rate, by Fiscal Year 70 12.9% 60 11.2% 10.8% 50 40 6.8% 30 59 42 20 37 33 10 FY2015 FY2014 FY2016 FY2017 Number and Percent of Units Turning Over Figure 13: Inclusionary rental program turnover rate, FY2014—FY 2017. 20

Race of Head of Household: When applicants complete their initial applications, they are asked, but not required, to indicate their race and ethnicity. As of April 1, 2018, 344, or 66%, tenants have provided this infor- mation. 27% identified as Black/African American followed by 22% who selected White, 5% who selected Asian, 1% who selected Native Hawaiian/Other Pacific Islander or American Indian/Alaska Native. Race of Heads of Household 1% 5% 27% 44% - 1% 22% • American Indian/Alaska Native • Asian • Native Hawaiian/Other Pacific Islander • Black/African American • White • No data* Figure 15: Race of Head of House Hold. Includes Middle Income Program Ethnicity of Head of Household: Ethnicity was voluntarily provided by 73% of current residents. Ten per- cent identified as Hispanic or Latino. Ethnicity of Heads of Household 10% 27% 63% • Hispanic or Latino • Not Hispanic or Latino • No Data Figure 16: Ethnicity of Head of House Hold. Includes Middle Income Program 21

Age of Head of Household: The number of households in the age ranges of 30-39, 40-54, and 55-69 are evenly distributed and have remained largely the same since 2017, except the 55-69 age range, which saw an increase of 14 heads of household. Age Ranges of Heads of Household 2018 7% 11% 26% 28% 28% 2017 12% 29% 26% 7% 26% Age in Years • 70+ 140-54 • 55-69 •I < 30 • 30-39 Figure 14: Age Ranges of Heads of Household, 2017 and 2018. Does not include the Middle Income Program 22

CHARACTERISTICS OF RECENTLY HOUSED TENANTS Recent Lease-ups: Between April 1, 2017 and March 31, 2018, 77 new households were housed in rental units, 89% had Cambridge residency, which is the same percentage as the last overview (April 1, 2016 to March 31, 2017). Studio units were sometimes filled with applicants who did not receive Cambridge residency but who work in Cambridge. Cambridge Residency Status of New Tenants 120 101 100 11 77 80 8 60 90 40 69 20 2018 2017 • Non-Cambridge Resident • Cambridge Resident Figure 17: Recent Lease-ups by residency status, 2017 and 2018. Families with Children: There were 22 families with children who entered leases during this time, representing 29% of all recent lease-ups. The percentage of families with children is driven by the size of units that are available. Most of these families had children under 6. This is a decrease from last year, when 39 families with children were newly housed. 45 New Tenants with Children 39 40 35 30 25 22 36 20 15 18 10 5 0 2018 2017 • With Children under 6 • With Children under 18 Figure 18: Recent Lease-ups by families with children, 2017 and 2018. 23

Emergency Need: During this time period there were 26 newly housed tenants, 34% of the total, who had an emergency need. Homelessness, overcrowding, paying more than 50% of income in rent were the most prevalent emergency needs. Emergency Need, New Tenants 45 • No Fault 39 40 Eviction 4 35 6 • Paying over 50% 30 26 25 • Code Violation 20 16 10 15 • Over- 2 crowded 10 6 12 5 • Homeless 5 2018 2017 Residents with Emergency Need Figure 19: New tenants by emergency need, 2017 and 2018. Voucher Utilization in New Tenants: Participation of voucher holders is lower than in previous years. Between April 1, 2017 and March 31, 2018, 25 voucher holders were successfully housed, representing only 32% of new lease-ups. This represents a decrease from last year, when voucher holders comprised 59% of new lease- ups. Of those 25 voucher holders, 21 (80%) obtained their voucher through the Cambridge Housing Authority. Most of the new units were one- and two-bedroom units. Voucher Utilization, New Tenants 21 2018 52 56 2017 41 • Non Voucher • Other Voucher • CHA Figure 20: Voucher Utilization among recent lease-ups, 2017 and 2018 24

Type of Housing Prior to Entering Program: In both 2017 and 2018, prior to moving into inclusionary housing, most households new to the program were living in market rate housing. In 2018, 24 (31%) of new tenants lived in some type of affordable housing, were in a shelter, or had been homeless prior to moving to their inclusionary unit. In 2017, 30, (30%) of recent lease-ups were living in affordable housing, were in a shelter, or had been homeless prior to moving to their inclusionary unit. Prior Residency - New Tenants 9 3 3 9 53 77 2018 8 5 101 16 71 2017 • Market • Non-profit/Other Affordable Public Housing • Transitional Housing/Shelter • Other Inclusionary Figure 21: Prior residency among recent lease-ups, 2017 and 2018 25

SUMMARY OF PROPERTY MANAGEMENT REVIEW The Community Development Department approves applicants for eligibility for the inclusionary program, based on their income and assets. When applicants are approved, their applications are sent to a property with an available appropriately sized unit. Property managers review applicants for credit, CORI, and, in some cases, landlord references. Each property has its own standards which are reviewed and approved by CDD prior to the leasing of the units in the building. The standards for tenant approval must be consistent with those of market rate tenants. For instance, landlord references can only be used as criteria for approval if they are also used for market rate tenants. Applicants' incomes are not used to determine creditworthiness. Based on the property management review, applicants are approved or denied for the unit. If they are denied, applicants have five days to appeal the denial to the management company, who may or may not overturn it. This section provides information on the results of reviews by properties between 2010 to April 1, 2018. Total Reviews, 2010-April 1, 2018: Between the beginning of 2010 and April 1, 2018, 829 applicants were reviewed by a property management company in at least one building. One hundred and fifty-seven applicants (18%) were denied and 705 (82%) were approved. Applicants Processed, Total =862 157,18% 705, 82% = Approved Denied Figure 22: Approvals and denials, 2010 to 4/1/18 Reviews by Voucher Status, 2010- April 1, 2018: Between 2010 and April 1, 2018, 289 applicants were voucher holders and 573 were non-voucher holders. The overall approval rate for voucher holders was 81% and for non-voucher holders was 83%. Approvals and Denials, 2010-4/1/2018 600 573 103 500 18.0% 400 289 300 54 82.0% 18.7% 470 200 81.3% 235 100 Non-Voucher Holders Voucher Holders • Approved • Denied Figure 23: Approvals and denials, total, all years. 26

Applicant Appeals for Tenancy Denials, 2010-2018: If an applicant is denied by a property, they can appeal the denial to discuss reasons for the denial with the property management. There have been 57 appeals since 2010, of which 24 (40%) were by voucher holders and 33 (60%) were by non-voucher holders. Non- voucher holders have a slightly higher appeal success rate, at 41%, than voucher holders, at 33%. Appeals by Voucher Status 33 35 30 24 25 19 20 15 15 10 14 5 0 Non-Voucher Voucher Holders Holders • Denied Approved Figure 24: Appeal results, 2010-4/1/2018 Reasons for Denial: Between 2010 and April 1, 2018, the most common reason for denial was an applicant's credit, followed by eviction history, landlord reference, criminal history, and civil court. Denial of Non Voucher Holders Denial of Voucher Holders 3% 5% 5% 3% 6% 8% 6% 8% 76% 80% • Credit • Eviction = Landlord • Criminal • Civil Court • Credit • Eviction • Landlord • Criminal • Civil Court Figure 25 and 26: Reasons for Denial by Voucher Status, 2010-4/1/2018 27

MIDDLE INCOME PROGRAM There are currently 15 middle income units in the rental portfolio for households whose incomes are between 80% to 120% AMI. The middle-income units are at one property in Kendall Square. They were first leased up in 2015. This is a small number of units from which to identify trends; however, a few characteristics can be noted. The information below, unless otherwise noted, is based on households in these units on April 1, 2018. On that date, 14 of the 15 units were occupied with one vacancy. Since the units were first leased up until April 1, 2018, there has been a total of four vacancies here. Tenants by Household Size: There are 7 one-bedroom units and 8 two-bedroom units in the portfolio. The following chart shows the household size of all tenants. The seven one-bedroom units have one or two occupants and the two-bedroom units have two to five occupants. Middle Income Tenants by Household Size 2 5 2 • 1pp # 2pp 3pp 4pp • 5pp Figure 27: Middle Income Tenants by Household Size Number of children: As of April 1, 2018, there were nine households with children under 18. The total number of children is 15 with 6 under 6 years old and 9 between 6-18 years old. Children in the Middle Income Program 10 9 8 6 3 2 Children Under 18 Children Under 6 Figure 28: Middle Income Tenants by Household Size 28

Tenant Income: The following chart shows the percentage of AMI for the 14 occupied units. It should be noted that 8 of the 15 units are designated for households at 80%-100% AMI and 7 units are designated for households at 100%-120% AMI. The range of incomes are governed by this requirement. The vacant unit is a two-bedroom unit designated for a household in the 80%-100% AMI income range. The one household who is over 120% AMI moved out of the unit after April 1, 2018. Middle Income Tenant AMI Percentages 140% 124% 118% 116% 120% 110% 107% 107% 104% 97% 100% 90% 88% 88% 86% 83% 83% 80% 60% 40% 20% 0% Middle Income Inclusionary Households Figure 29: Middle Income Tenant AMI Percentages 29

PART 3: HOMEOWNERS The Housing Division oversees sales and resales of affordable homeownership units which are developed through a variety of programs, including non-profit development, financial assistance, Homebridge, and inclusionary housing. Inclusionary homeownership units account for approximately 40% of all affordable ownership units. The data below provides information on the frequency of resales and the length of time that owners remain in inclusionary ownership units. Turnover of Ownership Units: The first resale of inclusionary homeownership units was in 2008. There have been between one to five inclusionary homeownership units resold per year since that time. The chart below shows the number of units resold per year, the bedroom size, and the total inclusionary homeownership stock occupied each year. The rate of resale of inclusionary homeownership units is low. The highest rate of resales was in 2014 and 2016 when 2.4% of the units were resold. Inclusionary Sales Annual Turnover Total IZ HO - % Occupied Stock by Date Units Sold FY IZ Total 2 Bedroom 1 Bedroom 3 Bedroom 1 0 1 0 123 2008 0.8% 0 1 0 0.6% 158 2009 2 0 182 1.1% 2010 2.1% 2 2011 187 2 1 1.6% 187 2012 2 0 189 1.1% 2013 2.6% 2 3 0 194 2014 1 1.0% 194 1 2015 1 194 4 2016 2.6% 4 1 2 1 2017 195 2.1% 2 0 2018 195 1.0% 2 Table 10: Inclusionary Sales 2010 to present Inclusionary Homeownership Tenure: Most inclusionary homeownership units were first occupied between FY2006 to FY2010. 152 of the 195 total occupied units, or 78%, were made available to buyers for the first time during that period. To date, there is not a lot of difference between the age of the unit and the length of current owner tenure. The chart below shows the median age of inclusionary units compared to the median length of time they have been occupied by bedroom size. Unit Size by Median Age Median Length of Number of of Unit in Ownership in Years Bedrooms Years 0/1 9.5 8.8 2 11.1 10.6 3 10.6 12.1 Table 11: Homeownership Tenure There are many reasons that owners sell an affordable unit. For units where the seller's motivation is documented, the most common reason is a change in family size, such as a marriage, birth of a child, or addition of elderly parents to the household. This was the reason given in 36.2% of sales. Other common reasons were job relocation (18.8%); the desire to own an unrestricted home (15.9%); or the death of the owner (8.7%). 30

N/A N/A July 2017 May 2010 April 2017 May 2010 May 2010 May 2010 May 2010 May 2010 May 2010 May 2010 May 2010 May 2014 May 2010 May 2010 May 2010 April 2016 May 2010 June 2011 June 2014 March 2018 March 2016 October 2016 August 2017 January 2013 October 2013 yet considered** December 2016 November 2014 September 2017 oldest application not 26 26 23 23 98 61 72 72 184 123 123 voucher* "Cambridge" 19 24 18 63 non- 91 91 95 15 48 48 15 voucher applicants Cambridge 13 13 13 101 891 101 813 633 633 180 790 790 Non- 180 Voucher applicants 0% 1% 4% 0% 4% 0% 2% 2% 9% 3% 0% 0% 2% 1% 5% 0% 0% 67% 12% 70% 13% 10% 12% 14% 13% 61% 12% 15% 100% 100% 100% the pool the pool the pool % of total % of total % of total number of number of number of applicants in applicants in applicants in 52 19 14 30 17 14 48 94 35 11 20 2-br 158 742 1-br 123 0-br 150 210 302 126 131 138 128 318 169 288 660 [phone removed] 1060 applicants applicants applicants points points points tier tier total total total resident subtotal resident subtotal resident subtotal non-resident subtotal non-resident subtotal non-resident subtotal Summary of the Rental Applicant Pool by Preference Group on 5.1.18 Cambridge resident, emergency need Non-resident, work in Cambridge, emergency need Non-resident, no preference Non-resident, work in Cambridge Non-resident, work in Cambridge, emergency need Non-resident, work in Cambridge, child under 18, emergency need Non-resident, work in Cambridge, emergency need Non-resident, work in Cambridge, child under 18 Non-resident, child under 18, emergency need Non-resident, child under 6, emergency need Cambridge resident Non-resident, child under 18 Non-resident, emergency need Non-resident, work in Cambridge Non-resident, child under 6 Cambridge resident Non-resident, work in Cambridge, child under 6, emergency need 2 Bedroom Cambridge resident, child under 6, emergency need Cambridge resident, child under 18, emergency need Cambridge resident, child under 18 Non-resident, work in Cambridge, child under 6 Non-resident, no preference Non-resident, emergency need Non-resident, no preference Cambridge resident Studio 1 Bedroom Cambridge resident, emergency need Cambridge resident, child under 6 Non-resident, work in Cambridge Cambridge resident, emergency need Non-resident, emergency need

N/A N/A N/A N/A N/A N/A N/A N/A May 2014 May 2010 May 2014 May 2010 May 2010 June 2010 May 2010 May 2010 May 2010 May 2010 April 2012 June 2011 April 2013 April 2015 June 2104 March 2013 March 2016 August 2015 August 2012 October 2015 October 2012 October 2017 February 2012 December 2015 yet considered** December 2013 December 2015 113 113 voucher* "Cambridge" oldest application not 17 42 17 42 voucher applicants 250 160 160 250 Voucher applicants 1% 0% 0% 0% 0% 0% 0% 2% 5% 0% 0% 1% 2% 2% 1% 4% 4% 1% 7% 9% 0% 2% 0% 0% 0% 5% 20% 15% 17% 13% 10% 25% 21% 12% 10% 10% 100% 100% the pool the pool % of total % of total number of number of applicants in applicants in m IN N 15 14 23 23 319 609 290 130 103 4-br applicants applicants in points points N tier total total resident subtotal non-resident subtotal non-resident subtotal Cambridge resident, child under 18, emergency need Cambridge resident, child under 6, emergency need Cambridge resident, child under 6 Cambridge resident, emergency need 4 Bedroom Non-resident, child under 6 Non-resident, work in Cambridge, child under 6, emergency need Non-resident, work in Cambridge, emergency need 3 Bedroom Cambridge resident, child under 18, emergency need Non-resident, work in Cambridge, child under 6, emergency need Non-resident, work in Cambridge, child under 6 Non-resident, child under 6, emergency need Non-resident, work in Cambridge, child under 18 Non-resident, work in Cambridge, child under 18 Non-resident, work in Cambridge Non-resident, work in Cambridge, child under 6 Non-resident, work in Cambridge Non-resident, child under 18, emergency need Non-resident, child under 6, emergency need Non-resident, emergency need Cambridge resident Cambridge resident, child under 6, emergency need Cambridge resident, emergency need Non-resident, work in Cambridge, emergency need Non-resident Non-resident Non-resident, child under 18, emergency need Cambridge resident, child under 18 Non-resident, child under 6 Non-resident, emergency need Non-resident, child under 18 Non-resident, work in Cambridge, child under 18, emergency need Non-resident, work in Cambridge, child under 18, emergency need Cambridge resident, child under 6 Non-resident, child under 18 Cambridge resident Cambridge resident, child under 18

457 270 187 " voucher* "Cambridge 37 237 206 voucher applicants* 460 Non- 1674 2134 Voucher applicants 25% 63% 12% 100% [phone removed] 3075 applicants a non-Cambridge Voucher. Non-residents are identified as having a "Cambridge" voucher and non-Cambridge voucher. **oldest application not yet considered" are applicants who have not been sent to available property. All Resident Applicants Applicants may be eligible for more than one bedroom size. All Applicants *This is an accurate account at the time of application of Cambridge residents which have a "Cambridge" voucher and includes Cambridge residents who have Non-resident, work in Cambridge Non-resident NOTES TOTAL ALL POOLS The data in this report is from the Rental Applicant Pool on 5/1/2018

Atlachment F Thoughts on the Affordable Housing Overlay concepts, 9/27/18 Overall I support the idea of a citywide affordable housing overlay. I appreciate the work of the consultant and city staff to develop this concept. I do have a number of concerns, which I hope will be addressed. My final position on this Affordable Housing Overlay will depend on how my concerns are addressed. A strong argument for the Affordable Housing Overlay is that it enables affordable units to be located in Res A and B areas, where the units could be in a different size and type of affordable building than currently happens, and the Overlay could make it easier than now to offer affordable homeownership, which Cambridge needs much more of. The maps and the breakout by neighborhood in the Housing Distribution Analysis (esp. the breakout on the upper right corner of p. 2) are good arguments for the affordable housing overlay. Another argument is that an Overlay will bring more economic and racial diversity to neighborhoods that use to be diverse, but where rising housing prices have pushed out people who are not well-off. I do not want the Overlay result in tearing down older buildings. Instead, adding to and modifying existing buildings should be prioritized and shown in examples. Under current zoning, a big old house may be re-developed as luxury condos that most current Cambridge residents can not afford. With the Affordable Housing Overlay, the big old house could be re-developed as a somewhat higher # of affordable condos that current residents can afford. Residents of affordable tend to reside longer term in their homes, and to have children in the public schools, so they will contribute to the strength of the neighborhood. Just like I do not want to see older houses destroyed, I do not want to see mature trees destroyed. The proposal should state some prioritization of trees relative to housing, and it should protect trees 1 year before the sale of a property. The Overlay should include requirements for some % of family sized units- such as at least 50% 2BR or larger. We should require what we need. I have concerns about the small setbacks, such as only 5 feet from the property line. In Res. C, the minimum setback is 7.5, so 1 do not think as-of right the setback should be smaller than that. Possibly change the setbacks, so that required setbacks are measured from existing buildings, not just from property lines. If I were the immediate abutter, the setback from my building is what would matter to me. Another idea would be to have flexible setbacks that relate to the existing setbacks on the block. If the typical front setback were, for example, 15 feet, the front setback for the affordable property could be 20% less, or 12 feet. "Super-Inclusionary" proposal Using the 2.5 times existing density approach means that areas that are already more dense will become even more dense. For example, Central Sq., already home of some of the greatest allowed density (see

p. 17), and an area that already has the highest % of affordable housing in the city (see Housing Distribution) and the greatest # of people of color, would have much higher density than Porter Sq. We just upzoned Central Sq., and there has not been time for that upzoning to take effect yet. I would like there to be an upper limit on the proposed increased density. An FAR of 6 is already very dense. I would like clarification on whether all of the increased density would have to be used in order to be economically viable. If that is the case, then there is no room for upper limits on height or other elements of urban form, and I would likely oppose many of the resulting developments. I look forward to continuing to work on both of these proposals and hope to be able to fully support them. Sincerely, Lee Farris 269 Norfolk St 02139 :

Atlachment 6 CITY OF CAMBRIDGE Community Development Department Iram Farooq Assistant City Manager for MEMORANDUM Community Development To: Iram Faroog, Assistant City Manager for Community Sandra Clarke Development Deputy Director From: Chief of Administration Christopher Cotter, Housing Director W Date: September 12, 2018 Ke: Housing Distribution by Neighborhood CDD staff have completed the attached analysis detailing the City's stock of tordable housing, along with the total number of housing units, in each of th ity's thirteen neighborhoods as of Tune 30, 201 Affordable units on the attached analysis include all rental and homeownership units that are subject to long-term rent ot sales price restrictions and income- eligibility limits of any level, as well as some market units in predominantly affordable developments. Units that were under construction or permitted as of June 30, 2018 are included in this affordable housing count. Information on overall housing stock was provided by Cliff Cook, as described in the attached memo. Affordable units are categorized as follows: • Non-Profit Sponsored Housing: includes all housing sponsored and/or owned by non-profit developers; • Public Housing; includes all state and federally subsidized public housing and former public housing (e.g. RAD conversion) properties controlled and managed by the Cambridge Housing Authority; • Inclusionary Housing: includes all privately-owned affordable housing created under the Inclusionary Housing Ordinance and similar zoning provisions • Privately-Owned Affordable Housing: includes all privately-owned and/or sponsored affordable housing properties, including limited equity cooperatives, and excluding Inclusionary Housing; and, • Scattered-Site Homeownership: includes attordable homes purchased by first-time homebuyers, excluding those created through Inclusionary Housing and other developers. 344 Broadway Cambridge, MA 02139 Voice: [phone removed] Fax: [phone removed] TTY: [phone removed] www.cambridgema.gov

Cambridge Affordable Housing Stock Distribution Summary (as of June 30, 2018) Non-Profit Privately- Total Scattered- Public Inclusionary All Housing Owned Sponsored Units Site FTHB Affordable Housing Housing Housing Housing 54,713 Units: 1,264 2,715 2,937 8,117 1,102 35.96% 100.00% 14.84% 12.52% 15.67% Percentage: 34.64% 1.20% NOTES AND DEFINITIONS: Affordable Housing: all rental and homeownership units subject to long-term rent or sale price restrictions and income-eligibility restrictions; includes units affordable to low, moderate and middle-income households; includes market units in predominantly affordable developments. Non-Profit Sponsored Housing sponsored and/or owned by non-profit developers Housing: Public Housing: State and federally subsidized public housing and former public housing (eg RAD conversion) properties controlled and managed by the Cambridge Housing Authority Inclusionary Housing: Provisions ored carabidge using eated un dard sintu zonary Housing programs; includes buildings under construction Privately-Owned Affordable Housing: rely und ander sponsored ar are noting raries riding Scattered-Site Homeownership: ferdable ugh pursday by it ane fore buyers excluding those CDD Analysis of City housing stock as of 6/30/2018 (Cliff Cook) All Housing Units: Prepared August 21, 2018

Affordable Housing Distribution by Neighborhood" (as of June 30, 2018) All Privately Non-Profit % of Scattered- Inclusionary Total % of Total Housing Owned Sponsored neigh. Site FTHB Housing Housing Units Housing Housing 932 467 12.3% 7,553 11.5% 69 17 1 - East Cambridge 389 2.3% 36 0 0.4% 1,542 36 20.6% 3,015 4 11 64 620 7.6% 215 326 - Wellington-Harrington 34.5% 13.6% 98 1,107 3,206 14 653 70 4 - The Port 372 65 6,491 20.0% 12 1,296 151 16.0% 593 475 5 - Cambridgeport 6,736 455 32 6.8% 6 5.6% 351 58 6 - Mid-Cambridge 16.7% 8.2% 43 669 4,016 128 7 - Riverside 316 178 5.3% 111 2,086 2 1.4% 40 3 - Agassiz 49 16 8,004 6 586 9.8% 7.2% 33 240 9 - Neighborhood 9 82 215 1.3% 53 6 0.7% 4,210 8 37 10 - West Cambridge 7,612 24.5% 14 1,862 511 305 454 578 22.9% 11 - North Cambridge 21.0% 2.9% 5 1,108 0 73 233 155 12 - Cambridge Highlands 13.8% 4 1.9% 0 1,134 157 153 [13 - Strawberry Hill 54,713 99 14,84% 100.0% 1,102 8,117 1,264 2,715 2,937 1.22% 15.57% 13.58% 33.45% 36.18% PUBLIC HOUSING UNITS NON-PROFIT HOUSING 14.3% 389 69 [1 - East Cambridge 1- East Cambridge 2.3% 2 - MIT 0.0% 2 - MIT 0.0% 7.9% 215 3 - Wellington-Harrington 11.1% 326 3 - Wellington-Harrington 4 - The Port 663 20.4% 4 - The Por 12.7% 372 17.5% 593 15 - Cambridgeport 475 20.2% 5 - Cambridgeport 2.1% 58 16 - Mid-Cambridge 351 12.0% 6 - Mid-Cambridge 6.8% 178 7 - Riverside 7 - Riverside 316 10.8% 0.6% 16 8 - Agassiz 49 1.7% 8 - Agassiz 3.4% 92 9 - Neighborhood 9 7.3% 215 9 - Neighborhood 9 0.3% 10 - West Cambridge 1.3% 37 10 - West Cambridge 578 21.3% 11 - North Cambridge 11 - North Cambridge 454 15.5% 0.0% 0 155 12 - Camb Highlands (12 - Cambridge Highlands 5.3% 153 5.6% 0.0% [13 - Strawberry Hill [13 - Strawberry Hill 100.0% 2,715 100,0% 2,937 INCLUSIONARY HOUSING UNITS PRIVATE HOUSING UNITS 0 0.0% 1 - East Cambridge 457 1 - East Cambridge 41.5% 0.0% 2-MIT 2 - MIT 3.3% 36 5.1% 64 3 - Wellington-Harrington 4 3 - Wellington-Harrington 0.4% 98 7.8% 4 - The Port 70 6.4% 14 - The Port 11.9% 151 5 - Cambridgeport 65 5.9% 5 - Cambridgeport 32 6 2.5% 8 - Mid-Cambridge 0.5% 6 - Mid-Cambridge 7 - Riverside 10.1% 128 43 7 - Riverside 3.9% 40 3.2% 8 - Agassiz 4 0.4% 8 - Agassiz 19.0% 9 - Neighborhood 9 33 240 3.0% 9 - Neighborhood 9 0 0.0% 6 0.5% 10 - West Cambridge 10 - West Cambridge 40.4% 11 - North Cambridge 511 305 11 - North Cambridge 27.7% 0.0% 0 12 - Cambridge Highlands 6.6% 73 12 - Cambridge Highlands 0.0% 0 13 - Strawberry Hill 0.0% [13 - Strawberry Fill 100.0% 1,284 100.0% 1,102 NOTES AND DEFINITIONS: SCATTERED-SITE HOMEOWNERSHIP all rental and homeownership units subject to long-term rent or Affordable Housing: 17.2% - East Cambridge sale price restrictions and income-eligibility restrictions; 0.0% 2 - MIT 11.1% 3- Wellingion Harrington" developments 14.1% 4 - The Port 12 12.1% 5 - Cambridgeport 8 8.1% 6 - Mid-Cambridge Housing sponsored and/or owned by non-profit developers 4 Non-Profit Sponsored Housing: 4.0% 17 - Riverside 2 2.0% 8 - Agassiz 6.1% 9- Neighborhood 9 Stale and federally subsidized public housing (including R 2.0% Pubilc Housing: 10 - West Cambridge enversion) propertles controlled and managed by ( 14.1% 11 - North Cambridge Cambridge Housing Authorily 12 - Cambridge Highlands 5.1% 13 - Strawberry Hill 4.0% Inclusionary Housing: Privaley-ownod affordable housing created under the 99 100.0% Inclusionary Housing Provisions of the Cambridge Zoning Ordinance and simllar zoning-based programs; includes buddings under construction Privately-owned and/or sponsored affordable housing Privately-Owned Affordable Housing: properties, Including limited equity cooperatives and excluding Inclusionary Housing Propared August 21, 2018 Affordable homes purchased by first-time homebuyers, Scattered-Site Homeownership: excluding those created through Inclusionary Housing and other devolopars All Housing Units: COD Analysis of Cily housing stock as of 6/3012018 (Clif Cook)

DRAFT MEMORANDUM TO: CHRIS COTTER, CASSIE ARNAUD CLIFF COOK PROM: Cc: MELISSA PETERS, IRAM FAROOQ, SANDRA CLARKE, KHALIL MOGASSABI, JEFF ROBERTS SUBJECT: HOUSING STOCK COUNT AS OF JUNE 30, 2018 (END OF FY 2087). DATE: AUGUST 10, 2018 Buildout Database Analysis Having just updated the Buildout Database with development data from the past fiscal yeat, the Database generated the following housing totals for Carabridge as of June 30, 2018, the end of Fiscal year 2018. Figures ate compared to neighbothood totals provided by the 2010 Census, adjusted to cottect fot ertors and omissions: Difference • 2010 US Census 2018 2018-2010 City Analysis Neighborhood (Adjusted) 1,753 5,800 7,553 East Cambridge 726 816 1,542 Area 2/MIT 19 3,015 2,996 Wellington Harrington 415 3,206 The Port/Area Four 2,791 117 6,491 6,374 Cambridgeport 18 Mid-Cambridge 6,736 6,718 (323) 4,339 4,016 Riverside (30) 2,086 2,116 Agassiz 360 6,004 5,644 Nelghborhood Nine West Cambridge 4,210 15 4,195 5,618 7,612 1,994 North Cambridge 757 351 1,108 Cambridge Highlands (81) Strawberry Hill 1,215 1,134 5,740 48,973 54,713 City Total The end of IFY 2018 figure is 1,423 units larger than that fox Jute 30, 2017, when the prior fiscal the most recent annual change reflects a tripling of the older figure.

Since the 2010 Census was conducted on April 1, 2010 the Cambridge housing stock has gtown by 11.7%, after adjustments to the base figute for errors and omissions principally related to graduate student apartment style housing. This works out to a 1.36% annual compound rate of growth over the 99 month period. The following chart summarizes change in the housing stock since the 2010 Census. Total Housing Units 60,000 55,000 50,00D ////////# 45,000 40,000 35,000 30,000 30-Jun-13 30-Jun-l 30-Jun-15 30-Jun-16 30-Jun-17 30-Jun-18 2010 Census Adjusted Adjusted Adjusted (1-Apr-10) Definition of Housing Included in June 30, 2018 Buildout Database Analysis The Buildout Database figures derive ftom analysis of data from the VISION Assessing data systern, Planning Board decisions, and ISD bullding permits, as supplemented by infotmation frotn the 2012 Street Listing published by the Elections Commission, affordable housing listings, VISION systena notes, and doorways, mailboxes, and street numbers visible in VISION photos and Google StreetView. Reported figutes comprise all housing wits, as defined by the US. Census Bureau, which generally define a housing unit to be one oi more rooms that open onto a common area ot butiding exterlot and where the resident(s) have their own bathing and cooking facilities. The Census Bureau treats lodging houses, SROs and assisted living residences as housing units. What is and Is Not Included in the Figures Reported Here 1. The base for analysis is taken from the FY 13 version of the Carbridge Assessing Department VISION database, which include data about all parcels through December 31, 2011. 2. Reported housing figures include: a. All existing non-institutional housing units, as defined by the Census Bureau, including units completed during the stated fiscal year. 2

b. Housing units for which a building permit has been issued as of June 30, 2018. (Counts of units under construction come from the CDD-maintained Development Log ot Energov. Reported Energor numbers are subject to revision to correct etrots entered into permits by applicants.) c. Units found in mixed use properties after evaluation on a case by case basis to adjust the numbet of units reported by Assessing to include all residential units located on each property. (Corrected data comes from the 2012 Street Listing published by the Elections Commission, notes included in the VISION record, and doorways, mailboxes, and steet numbers visible in VISION photos and Google StreetView.) d. All graduate student apartment-style housing owned by the universities, whether subject to the property tax of exempt and whether ot not those units are counted as such by VISION ot the Census Bureau. e. Lodging House/SRO units where services are not provided as patt of an organized progtam and tesidency is not predicated upon participation in that ptograin. Such units are counted as housing on a 1 unit fot 1 room basis. (For the purposes of the housing tally, Lodging House/SRO buildings are those either licensed as such by the Licensing Commission, which have the appropriate land use code assigned by the Assessing Depattment, ot ate known to have this use based upon affordable housing information,) f. Lodging House/SRO buildings that are non-institutional in form, whete services are provided as pait of an organized program, and tesidency is predicated upon participation in that program ate treated as housing when the building would otherwise likely have a residential use (e. g., a triple-decker is counted as three units regardless of the number of SRO style rooms). g: SRO units located in the YMCA and YWCA buildings, including those that come with services, counted on 1 unit for 1 room basis. • 3. Reported housing figures exclude: 2. Housing units fot: which a demolition permit was issued prior to July 1, 2018. b. All undergraduate housing owned by the universities whether taxable of tax exempt and whethet or not those units are counted as such by VISION or the Census Bureau, including apartment style utits such as the DeWolfe Street units owned by Harvard. c. Dormitory style graduate student housing, such as the Hatvard Law School Gropius dormitories and the Cronkite Center: d. Lodging House/ SRO rooms or beds where services are provided as part of an otganized progtai and residency is predicated upon participation in that program, with the exception of YMCA and YWCA building units. (Lodging House/SRO buildings wight: be counted as housing, depending upot the style of the building.) є. All other forms of group quarters housing, including dormitories, nursing hornes, homeless shelters, and jails. 3

Ferors and Omissions Due to the dispersed nature of the data and the need to recompile figures collected in a variety of databases, there is always the possibility of ertots being introduced into the housing unit counts. The following corrections resulted in significant changes to the inventory: 1. Correction of two query errors between generation of the FY 14 and the FY 15 teports had a significatt effect on the size of the inventory. The error affected parcels with tore than ore small residential building, specifically those buildings with one, two or three units. In a number of cases the count of housing units was overstated. In some cases the number was doubled, and in others the figure quadrupled. Correcting the error lowered the total housing stock by about 1,000 units compared to prior figures. 2. One notable change occurted between the FY 13 and FY 14 calculations. There was a reduction it units in Riverside due to the correction of errors that led to the ovet counting of Harvard- owned undergraduate-occupied units. ( In light of these etrots and omissions, the FY13 and FY14 figutes included in the chart found earlier in this memo have been adjusted downward by 1,200 to approximate their effect on the overall total. 3. A correction made to the initial FY 2017 total lowered that figure by 98 units, from 53,388 reported in some places to the 52,290 used for this analysis