Boswell Proposal April2010!11!1

Dissertation Proposal

Linking Neighborhood Housing Market Typologies to Revitalization Strategies: Spatial Analysis of Residential Investments on Property Values in Baltimore, Maryland (tentative)

April 2010

Lynette Boswell

PhD Candidate

Department of Urban Planning and Design

University of Maryland, College Park

ABSTRACT. This study will examine the link between neighborhood housing market typologies, revitalization strategies, and the effectiveness of these strategies in neighborhoods in Baltimore, Maryland. It will analyze the impacts of public residential investments on surrounding property values across neighborhood housing markets, while considering other mitigating factors, which include: magnitude and concentration of investments, proximity to neighborhood assets, and external neighborhood conditions. This study will develop a neighborhood housing market typology for Baltimore, Maryland, using the cluster statistical method to evaluate neighborhood change theories and models. Traditional hedonic price regression models will estimate the effects of residential investments on nearby property values (using neighborhood housing market typology indicators), and a comparison spatial hedonic regression model will be used to address spatial dependency and errors.

1

INTRODUCTION

Neighborhood revitalization strategies have evolved since the Model City program in the 1970’s,

downtown reinvestment and Enterprise (EC) programs in the 1980’s, the Empowerment (EZ)

and Homeownership programs in the 1990’s, to various strategies of targeted Community

Development Block Grant (CDBG) and HOME partnership funds since 2000. Throughout the

late 1970s to the early 1990s, federal policymakers have demonstrated interests in geographically

targeted urban development strategies: however, these past programs have presented mixed or

few results in distressed cities (Oakley and Tsao, 2007, and 2006, and Ladd, 1997). Despite the

presence of targeted redevelopment and investment policies, many cities remain blighted and

economically unstable. Based on these challenges, in the last decade, scholars (e.g. Mallach,

2006, and Galster, Tatian, Accordino, 2006) have encouraged cities to think more strategically of

their place-based methods, and understand the patterns of neighborhood change as well as

regional housing market dynamics to guide city-wide decisions and policies.

Today, weak market cities – “characterized as economically distressed areas with little market

demand for an excessive supply of housing stock (Mallach, 2006)” -- have reconsidered their

efforts, and instead of focusing on massive redevelopment in depopulated areas, they are moving

towards stabilization strategies. Stabilization is defined as gradual neighborhood improvement

and market correction. It is about rebuilding confidence and fostering market recovery. Market

recovery happens when stabilizing actions (vacant homes rehabilitated and new homes

developed) exceed destabilizing events (new foreclosures and homes abandoned) (Mallach,

2006, 2008). The goal of stabilization strategies is to retain population (Mallach, 2006, Galster

et al., 2006) and improve the existing fabric of affected neighborhoods. The general principles

2

of stabilization efforts include: 1) understanding the local housing markets; 2) targeting few key

areas; 3) targeting based on community resources and capacity; 4) matching strategies to market

conditions; 5) integrating strategies with other neighborhood strategies; and, 6) leveraging

private investment in more stable areas than heavily distressed neighborhoods (Mallach, 2008;

Mallach, 2006; Galster et al., 2006; Burnett and Brophy, 2003).

Cities such as Richmond, Virginia; Baltimore, Maryland; and Cleveland, Ohio, have begun to

employ these strategies, and spatially concentrate investments in areas exhibiting active housing

markets. A key tool used by these cities to address neighborhood challenges is neighborhood

housing market typologies. Typologies are used to classify neighborhoods into distinctive

categories through quantitative analysis based on housing and socio-economic characteristics.

Once categorized, neighborhood categories are matched with applicable stabilization strategies.

The concept of linking strategies to neighborhood and housing market conditions implies only

certain strategies are successful in different neighborhoods based on neighborhood conditions. It

further suggests neighborhoods exist along a continuum of decline and renewal; therefore, it is

assumed that less stable blocks can be improved through aggressive block level targeting of infill

development and housing rehabilitation, and massive redevelopment, such as demolition and

new construction, will revitalize more distress blocks.

Scholars have proven that strategically targeted investments at the block level present positive

measureable impacts (Galster, et Al., 2006; Galster, et Al., 2004; and Ding, et al. 2000).

However, as a policy tool, this model has not been explicitly tested for accuracy across different

neighborhood housing markets. It is important to question: Is the effectiveness of stabilization

strategies linked to housing market typologies? Under what neighborhood conditions will

strategies more likely stem or reverse neighborhood decline, as indicated by changes in nearby

3

property values? Moreover, how should these strategies be applied – based on geographic scale,

proximity to neighborhood resources, or in response to other economic forces- within respective

neighborhood markets?

The purpose of this study is to evaluate the effects of new construction and rehabilitation

residential investments on surrounding properties in varying neighborhood housing markets

through market cluster analyses. This study will use both a traditional ordinary least squares

(OLS) and spatial hedonic regression models to estimate the impacts of residential investments

on nearby property values. These models will be used to determine which neighborhood housing

market indicators influence the impacts of investments on surrounding property values. The

empirical analysis will be applied to Baltimore, Maryland.

BACKGROUND

As summarized in table 1, Baltimore’s population declined by 17% from 1990 to 2000, and

suffered from slight increases in poverty and unemployment rates. Median household income

also declined, though the Metropolitan Statistical Areas (MSA) average income increased

slightly. Rapid changes among the population, between 1970 and 2000, and declining jobs in

regions surrounding Baltimore negatively impacted housing markets in the City prior to 1998, as

households moved to growing regions with job opportunities. Between 1970 and 2000, a

significant percentage of the population and job growth occurred in counties south and west of

Baltimore City, in Washington DC counties: Montgomery, Fredrick, and Prince Georges

Counties. Jobs in the D.C. metropolitan area increased from 450,000 to 1.1 million,

demonstrating 147% job growth, followed by significant population gains. Baltimore City and

4

County contain more jobs with 1.5 million jobs in 2000, but did not grow as fast as surrounding

regions.

Table 1: Baltimore Socio-Economic Change from 1970 - 2000

Source: Data adapted from the City of Baltimore Housing Plan, 2005

Massive out-migration from the Baltimore City slightly reversed in the beginning of 2000 due to

expensive housing in and around the Washington DC metropolitan areas. Transportation

linkages to DC jobs, through the Maryland Area Regional Commuter (MARC) train, and more

affordable housing options in Baltimore made the City more attractive to young professionals

(National City of Leagues, 2005). Over a five year period, between 2000 and 2005, Baltimore

experienced an economic resurgence in its housing market due to changing demographics, an

increase in regional housing demand, and small job growth in the Baltimore metropolitan area

(City of Baltimore Housing Plan (BHP), 2005). This led to an increase in market activity among

home sale and new construction. In 2000, median home values were $82,228, and in 2005, there

was an approximate 60% increase to $132,021. Between 2002 and 2005, a total of 17,215

permits for property investments above $5,000 were issued by the City (BHP, 2005). New

construction in the City occurred along the City’s waterfront and edge neighborhoods – Clipper

Mill and Cylburn Meadowns -- and increased over 4,000 units from 2000 to 2005. Development

efforts continued past 2005 but were stalled in 2007 due to high foreclosures in new and

redeveloped areas of the City.

5

1. Housing and Economic Development in Baltimore Neighborhoods

Starting in 1975 through 1995, the City of Baltimore relied on HUD’s Community Development

Block Grants (CDBG), HOME partnership, Empowerment (EZ) and Enterprise Zones (EC),

Hope VI, and housing tax credit programs to address older neighborhoods (Cohen, 2000). The

CDBG is a HUD categorical program allocated, at the discretion of the City, to provide housing

rehabilitation loans, public service and facilities, infrastructure, business finance and commercial

revitalization assistance to non-profit economic and housing development organizations, and

private sector groups and households. The HOME Partnership program is also a HUD sponsored

block grant program designed to create affordable housing for low-income households whose

income does not exceed more than 50 percent of HUD- adjusted median income. This program

is the City’s only public assistance grant used exclusively for homeowners and rental, new

construction and rehabilitation investments. The EZ and EC grants, which terminated in 2009,

were used by the City to provide companies located in designated zones with income tax credits,

deduction on business property and tax-exempt facility bonds.

Between 1975 and 2000, the City received $500 million in CDBG funds, (BHP, 2005) $100

million in Empowerment Zones (EZ), and an additional $350 million in Enterprise Zones (EC)

through federal grants and tax incentives (Davis and Brocht, 2002). Prior to 1980, a significant

proportion of this funding was directed toward downtown revitalization efforts (Wong and

Peterson, 1986: 299; and Polese and Stern, 2000: 420) based on the re-development focus of the

CDBG program, and local support from Baltimore’s private sector and financial industries. By

1990, these funds were concentrated in areas with high percentages of vacant and abandoned

6

properties to support Baltimore’s neighborhood organizations, NGOs and Community

Development Corporations (CDC’s), on a first-come basis. Since 1992, the City has received

$131 million in HOME partnership funds, representing approximately 15 percent of overall

federal spending in Baltimore, while CDBG dollars represents approximately 55 percent of

federal funds disbursed to the City.

2. Baltimore Neighborhood Housing Market Typology

Federal programs are applied in Baltimore to target disadvantaged groups based on socio-

economic characteristics and not market conditions and/or neighborhood resources. With limited

funding and growing disinvestments in neighborhoods, in 2002, the City Planning Department

categorized its 255 neighborhoods into 55 neighborhood statistical areas1 (NSAs) using the

City’s Real Property Database. Duplicating a real estate analysis of neighborhood markets

developed for the city of Philadelphia by the proprietary group, the Reinvestment Fund; the City

categorized neighborhoods based on housing indicators: abandonment rate, median assessed

home value, and percent homeownership. Data was aggregated to the NSA to represent

neighborhoods. Areas were then categorized into relatively homogenous groups, or ‘clusters”

based on neighborhood conditions. The statistical method used to create discrete categories is

called cluster analysis (Miller and Whicker, 1999). This housing market classification scheme is

labeled: a neighborhood housing market typology. The City used the typology to link each

1 Baltimore neighborhood statistical areas were developed by the Baltimore City Planning Department and the Family League of Baltimore City, and used by Baltimore City Data Collaborative. Boundaries are based on City’s neighborhood and community organizations with further consideration for existing census tract boundaries to create statistical profiles. The concept of NSA’s was initiated in 1978 by HUD, for cities wanting to access Section 8 substantial rehabilitation subsidies. Cities were to develop detailed revitalization plans for neighborhoods, with public and private housing and community development resources to meet neighborhood revitalization needs (Varady, 1986).

7

cluster to revitalization strategies, identified as initiatives to stabilize market conditions.

Following are Baltimore’s 2005 neighborhood housing market typology categories:

1. Preservation: Healthy, attractive areas with high owner-occupancy rates and high property values. Housing stock which is well maintained and vacancy and abandonment rates are very low.

2. Stabilization: High home ownership rate, low abandonment, relatively low median home values, signs of deferred maintenance.

3. Reinvestment: Moderate real estate values, average home ownership rates and substantial vacancy rates.

4. Redevelopment: Deteriorated housing stock, abandoned buildings and vacant lots, traditional market forces not working.

The typology was revised in 2008. For finer detail, census block groups replaced NSAs as a

smaller unit of analysis, and the pool of variables was expanded to represent a broader range of

market activities. Additional housing variables included: assessed home values, percent vacant

homes and lots, percent foreclosure, percent homeownership, percent single-family homes,

percent commercial land, and percent residential rental subsidies. The cluster market typology

expanded from four to five categories: competitive and emerging, stable, transitional, and

distressed (Baltimore’s 2008 definitions, maps, and strategies are provided in Appendix 2).

The cluster analysis of neighborhoods provides an empirical framework for the City to

strategically match public resources to neighborhood housing market conditions, identifying

areas with relatively strong and weak markets. Baltimore along with many other cities, such as

Cleveland, Ohio; Memphis, Tennessee; and Philadelphia, Pennsylvania, are employing

typologies to allocate and target federal government programs and funding. As more cities use

neighborhood typologies to link neighborhood conditions to revitalization strategies and target

resources, it is important to understand the impacts of these interventions in different

neighborhood markets.

8

LITERATURE REVIEW

The literature on neighborhood dynamics and change theories, cities investment strategies and

impacts, and neighborhood housing market typologies provide a context to make inferences of

whether and to what extent the effectiveness of revitalization strategies are influenced by

neighborhood housing markets. It also provides a framework to understand what types of

investments are more suitable at different stages in a neighborhood’s course of change.

1. Neighborhood Change Theories and Models

Theories of neighborhood change provide a foundation to explain the process of how and why

neighborhoods change or improve over time. Specific neighborhoods are defined by three

dimensions: population, neighborhood conditions, and neighborhood expectations about the

future of the neighborhood (Kolodny, 1983).

Scholars remain inconclusive about which models of neighborhood change theories accurately

portrait cycles of neighborhood dynamics. Common perspectives are bonded in the following

schools: demographic/ecological, socio-cultural/organizational, political economy and social

movements. Classic models of neighborhood change are based in the demographic/ecological

disciplines: (1) the life-cycle, (2) the arbitrage model, and (3) the composition model. The life-

cycle model is commonly used in the context of neighborhood planning.

In general, neighborhood change theories consist of two assumptions. First, all neighborhoods

go through a predictable ‘life-cycle’, though there is significant contention among scholars

whether decline is an inevitable part of the neighborhood life-cycle. Second, changes in

neighborhoods are influenced by factors which impact individual household behaviors based on

residential preferences. Kolodny (1980) and additional studies (Bruch, 2006; Galster, 1998; and

9

Rappaport, 1997) identifies internal and external neighborhood factors which negatively affect

household decisions to leave neighborhoods as push factors, such as: physical decline in housing

stock, commercial, and institutions; out movement of higher income households; and, voluntary

residential preferences based on race, ethnicity or class. Models which reflect the evolution and

changes in the direction of neighborhoods are helpful for cities to plan for and attempt to arrest

decline in areas which experience negative impacts or push factors.

1.1. Life-Cycle Model

The life-cycle model originated among the urban ecologists of the Chicago School, spearheaded

by David Birch (1971). Birch suggests all neighborhoods exist on an evolutionary continuum

which begins as a stable community and ultimately declines at the end of the cycle. In 1975, the

Public Affairs Counseling (PAC 1975)real estate research corporation, developed under HUD’s

Office of Policy Development and Research identified five stages of neighborhood development

based on Birch’s2 model of city development: 1) Healthy, 2) Incipient Decline, 2) Clearly

Declining, 4) Accelerating Decline, 5) Abandoned. PAC explicitly indicated abandonment as a

final stage in the neighborhood life-cycle, and suggested it was inevitable. This work was

followed with Downs’ 1980 model of neighborhood stages with specific descriptions of

neighborhoods at each stage (Downs, 1980: pg. 63-64):

Stage 1: stable and viable. Relatively new and thriving or relatively old and stable neighborhoods with rising property values with desirable amenities and housing stock that continue to attract residents to maintain them.

Stage 2: minor decline. Older areas where structural maintenance is beginning to decline. Minor physical deficiencies are visible and density is higher in area. Property

2 Birch’s evolutionary processes include: low density rural areas (stage 1), wave of development (stage 2), fully developed high quality residential (stage 3), packing (stage 4: overcrowding, low income immigration, aging structures), and thinning (stage 5) (Kodolony, 1980). He later tested these factors against New Haven,

Connecticut’s 1967 census data, and included stage 6: recapture.

10

values are stable or increasing slightly. Public services and social status are below stage 1.

Stage 3: clear decline. Rental properties are increasing or dominant in the housing market. Minor physical deficiencies are visible and overall confidence in the neighborhood is weak wit increasing abandoned housing. Social status is lower than stages 1 and 2 because lower income groups predominates the housing market

Stage 4: heavily deteriorated. Housing is very deteriorated and dilapidated, and most structures are in need of repair. Profitability of rental units is poor and housing is only being marketed to lower income groups. Abandonment is prolific in the neighborhood and a pessimistic view of the future of the neighborhood.

Stage 5: unhealthy and nonviable. Neighborhoods are at the terminal point with massive disinvestment and abandonment. Residents represent the lowest social status in the City and region. Area is not considered marketable with increasing crime and depleting pubic services.

Downs’ analysis placed neighborhoods on a continuum, and suggested change could occur in

either direction along the continuum, and at any stage neighborhoods could be stable, improving,

or declining. In this conjecture, Downs’ argument differs from PAC’s conception of

neighborhood change. PAC places neighborhoods and cities in static points along the life-cycle,

all moving towards decline and abandonment; while Downs argues an oscillating process.

Downs claimed specific factors increased neighborhoods’ susceptibility to decline or revitalize

(Downs, p. 65), and “possibilities to reverse decline hinges on the stage of the neighborhood

along the life-cycle continuum.” Downs (1981:66) further noted that the effectiveness of

specific policies is influenced by what stage a neighborhood is in when the strategy is applied, all

other factors equal. As a result, neighborhoods in the latter stages of decline present more

difficulty in gradual improvement than those in the earlier stages.

1.2. Neighborhood Dynamics and Housing Markets

The life-cycle model captures physical variations in the housing market and the socio-economic

characteristics of neighborhoods at different points on the continuum; however, this model

11

provides little explanation for how and why households move along the continuum or impetus of

change in neighborhood conditions. Economists Maclennan (1982) and Grisby et al. (1987),

expand this literature by adding another dimensionality to the discussion, and propose individual

households move up and down markets based on their preferences, decisions of neighbors

regarding property maintenance, and location and spatial elements of the neighborhood.

Macleanan and Grisby scrutinized models of neighborhood change in their observations of

neighborhoods through the lens of segmented housing markets. Their contributions to

neighborhood change theories suggest that neighborhoods contain housing sub-markets with

varying economic, social, institutional and demographic forces which impact neighborhood

change. Housing sub-markets are defined as a collection of dwelling units, which offer similar

housing structures, public services, and location amenities (Grisby et al. 1987) within a

metropolitan area or city. Substitutable differences in the mix of housing type, quality of

housing stock, and price of housing units, and in some cases socio-economic characteristic of

residents or location attributes create unique sub-markets (Grisby et al. 1987). Market conditions

of the sub-market impacts the demand for the neighborhood based on household preferences.

Grisby suggests a filtering model3 concept. He links neighborhood conditions to residents’

socio-economic resources, their choices and ability to maintain properties, and decisions to

remain in neighborhoods. Grisby assertions are supported by literature on neighborhood change

and household preferences (Bruch, 2006; Rapaport, 1997). This literature combines

3 Grisby et al. (1987) suggest that households based on their preferences and socio-economic status move up and down sub-markets through a filtering process. Filtering occurs if upwardly mobile households filter up sub-markets

and are replaced by households in a lower socio-economic group.

12

This diagram exhibits the linear process of the impacts of housing market indicators on neighborhood conditions, which lead to neighborhood stage models. Source: Diagram adapted from neighborhood stage and housing market frameworks by Downs, 1980; and Grisby et al., 1987.

concepts of the life-cycle stage model and neighborhood dynamics, and housing sub-markets to

provide a framework to understand the link between housing market conditions, factors which

impact household decisions, and ultimately market forces at the neighborhood level. These

elements can influence how and when cities invest in neighborhoods.

2. Neighborhood Revitalization Strategies and Program Impacts

The successes of neighborhood redevelopment activities are closely linked to theories of

neighborhood change, changing housing markets, and the dynamics socio-economic

characteristics of neighborhoods. Studies which evaluate the impacts of public investments in

13

Neighborhood Life Stages

Stage 1 Stage 2 Stage 3 Stage 4 Stage 5

Stable Minor Clear Heavily Unhealthy

Decline Decline Decline

De

Neighborhood Submarkets

Housing Characteristics Housing Quality

Socio-Economic Characteristics Location Variable

Neighborhood Assets/Services

Changes in Socio-Economic Characteristics Changes in Households’ and Institutions’ Behavior

Changes in Level of Maintenance

Figure 1: Neighborhood Submarket Indicators and Downs’ (1980) Life Cycle Model

cities (Ellen, 2006; Galster, 2006; Popkin, 2004; Zielenbach, 2003; Ding, et al., 2000; and

Salaama, 1999) suggest that the aim of the intervention, target groups, and scale of the strategy

are important factors. Since the New Deal program in the 1930’s federal programs and funding

have taken different forms to alleviate urban poverty and provide adequate housing to

disadvantage groups (Keating, 1999). Two common approaches of investing in neighborhoods

are city-wide and target-neighborhood revitalization strategies (Mallach, 2009); differences

among these approaches are based on their spatial level of targeting resources.

2.1. Neighborhood Revitalization Strategies

City-wide strategies are initiatives in which neighborhood revitalization is carried out in all

neighborhoods, or in all neighborhoods eligible for federal resources based on meeting threshold

criteria (Mallach, 2006). Criteria are determined by income or other socio-economic

characteristics. The Model City program and city-wide allocations CDBG funds are examples of

these strategies. Target-neighborhood revitalization strategies differs from the city-wide strategy

which attempts to address all neighborhoods –or all neighborhoods meeting a criteria threshold –

in that these initiatives target few, select neighborhoods based on priority criteria (Mallach,

2006). Due to limited city resources and substantial city-wide distress, a small number of

neighborhoods are selected and targeted for public investments. Target-neighborhood strategies

may include EZ and EC zones based on selection process or city/ non-profit initiatives which

aggressively target initiatives at the block-level. City-wide and neighborhood-target approaches

are both people- and place-based strategies; however, city-wide strategies tend to be more scatter

site and politically driven. These strategies are explored below.

14

2.1.1. City-Wide Revitalization Strategies

The Model City program, a categorical grant program, enacted in 1966 by HUD was the first

city-wide, place-based initiative (Keating, 1999), in the U.S.A. This federal program initially

targeted 75 cities, and later doubled that number for a total of 150 cities, to alleviate inner-city

disinvestment and improve neighborhood conditions. The Model City program concentrated

comprehensive social services, inclusive of “bricks and mortar” in lower-income communities to

combat urban unrest in mostly minority neighborhoods faced with disproportionately high rates

of poverty and unemployment, racial discrimination in housing and employment, and poor

housing and public services (Keating and Smith, 1996). Initially the program’s purpose was to

target a few neighborhoods (Keating and Smith, 1996: 52); however, the implementation of the

program made funding accessible to all neighborhoods which demonstrated negative socio-

economic conditions (Scott, 1969). Criticized for little impact, in 1972, Congress appropriated

$150 million dollars to close out the program and fold it into the CDBG program due to lack of

funding and political support (Keating ed., 1996).

The CDBG program, which stemmed from the conversion of several HUD categorical grant

programs (including Model Cities and Urban Renewal) into a single federal grant, allocated

funding to entitlement communities and urban counties based on annual funding formulas

(Keating, 1999). Still active today, this program is available to communities which demonstrate

significant blight; however, leverage and discretion is given to cities as to where and how they

spend CDBG funds. This flexibility makes the program focus broad and less geographically

targeted towards poor neighborhoods. Minneapolis, Minnesota is an example of a city-wide

strategy which uses CDBG funds as a large scale effort to engage in planning and housing

15

improvement. CDBG resources are available to all eighty-one neighborhoods in the City,

through allocations vary based on neighborhoods’ conditions and needs.

EZ and EC zones developed and administered by HUD in the late 1980s and early 1990s, are

similar in scope to CDBG funding initiatives. Job creation and economic development were

specific focuses of these initiatives. Under the Clinton Administration, the federal government

provided at least $100 million to designated EZs and $3 million to ECs annually over 10 years to

provide tax incentives to businesses located in zones. Six urban areas received EZ funding,

including: Baltimore, Atlanta, Chicago, Detroit, New York, and Philadelphia-Camden, while 65

urban areas received EC funding. Though spatially located, these zones vary based on

populations assisted, and scale, which vary in many cities, from small highly distressed areas to

areas located in or adjacent to distressed neighborhoods (Turnham and Bonjourni, 2004).

2.1.2. Neighborhood- Targeted Revitalization Strategies

Within the last decade cities, have begun to be more strategic in their allocation of CDBG funds

by aggressively concentrating resources in fewer neighborhoods, at the block level for greater

visible impact. The revitalization efforts of cities such as Baltimore, Maryland, and Richmond,

Virginia are examples of target-neighborhood planning. These cities incorporate market-

oriented and data-driven processes that focus on neighborhoods with active real estate markets,

neighborhood assets, and strong neighborhood capacity.

In 1999, the City of Richmond, Virginia, initiated the Neighborhood in Bloom (NIB) program

which targeted eighty percent of CDBG, HOME, and Local Housing Initiative Corporation

(LISC)4, funds in six of 49 neighborhoods representing low to moderate income areas and other

4 LISC is a national organization, which mobilizes corporate, government and philanthropic support to provide local community development organizations with loans, grants and equity investments, and technical and management

16

neighborhood assessments using a city-wide neighborhood typology. The city committed

aggressive code enforcement, public safety, community empowerment efforts, and housing

investment resources in neighborhoods with extensive problems (high concentration of vacant

structures and poverty rates) and neighborhoods showing significant signs of decline (low

revitalization capacity and significant number of vacant structures) for two years, and

sequentially expanded the program to include two additional neighborhoods (Mallach,

2006:255). Over a ten year period, crime rates decreased, and property values increased along

with homeownership rates in select low to moderate income neighborhoods (Mallach, 2006).

In 2000, the City of Baltimore partnered with the Healthy Neighborhood Initiative (HNI) to

develop a program similar to NIB. Unlike the NIB program, the HNI is a non-profit led

initiative, which targets approximately one million of the City’s $44 million allocation of CDBG,

HOME, and local bank financial pools, in 15 of 55 Baltimore’s NSA. Resources are used for

rehabilitation, acquisition, homeownership assistance and neighborhood empowerment efforts,

on selected blocks in selected NSAs. The selection process of neighborhoods was similar to the

NIB initiative. At the census block level, between 2000 and 2006, select neighborhoods

presented positive impacts, such as; a rise in home sale prices and values, fewer vacant units,

decreases in homes remaining on the market for more than 120 days, and increases in the number

of homes sold in target blocks (Boehlke, 2002).

2.2. Impact of Public Investments on Property Values

assistance.

17

Scholars and policy maker’s present mixed results in the analysis of the impacts of revitalization

programs (see Table 2). Empirical studies (Garvin, 2002; Keating, 1999; and Washnis, 1973)

which evaluate the Model City and Minneapolis program focus on qualitative measures and

participatory processes. Studies which examine the impacts of ECs and EZs commonly use

difference-in-difference statistical methods, and present some successes in the reduction in

poverty and unemployment within the program zones, and spillover effects in adjacent

neighborhoods (Oakley and Tsao, 2007; and 2006). Empirical studies claim EC and EZ

programs’ net impacts are minimal due to the large areas targeted, and overwhelming negative

physical conditions and poor socio-economic characteristics of selected neighborhoods (Boarnet,

2001 Papke, 1994; Papke, 1993, Erickson and Friedman, 1990;). Studies which analyze CDBG

spending in cities are also mixed. Galster et al., (2000) analysis of CDBG spending between

1994 and1996 in 17 cities, is not conclusive of whether investments have measureable impacts

due to data limitations, the inability to standardize across cities and account for all factors which

affect community development investments and neighborhood outcomes.

18

Ding, et al. (2000) examination of public residential investments in Cleveland, Ohio, presents

more robust findings with the use geographic information systems (GIS) and a hedonic

19

Table 2: Neighborhood Revitalization Strategies

Source: Table is adapted from ABT Associates (2006), “Neighborhood Revitalization” and additional literature which analyze the impacts of neighborhood revitalization strategies from 1960 to present.

regression price models. They analyze the effects of investments on surrounding property values,

and conclude impacts are based on distance to investments, scale of investments, type of

investments (new construction or rehabilitation), and socio-economic and demographic

characteristics of the neighborhood. Using spatially lagged variables and interacting income

variables with housing investments, this study demonstrates that positive impacts occurred when

public investments were new construction, large scale, and concentrated (within 150 feet of

rehabilitated investments, and 300 feet of new construction investments), or located in low

income areas with few minorities.

20

results demonstrate that block level targeted investments have positive impacts on surrounding

properties. Using adjusted interrupted time series’ statistical method, their study suggests homes

that were concentrated in large numbers on few blocks produced a greater appreciation in market

values than comparable homes in similarly distressed neighborhoods, controlling for structural

elements and neighborhood conditions.

In summary, studies which evaluate the impacts of CDBG and other federal programs on

property values indicate a need for resources to be concentrated at the block level, and infer that

neighborhood conditions and external factors impact the investments’ effect. However,

conclusions cannot be drawn on whether strategies’ impacts have less influence in a relatively

more distressed neighborhood than in a neighborhood that is not as negatively affected. Few

studies explicitly test the impacts of housing revitalization efforts on varying housing

submarkets, and many fail to address additional variables which impact housing markets such as:

located within target reinvestment zones (EZ, EC and Business redevelopment zones), presence

of active neighborhood associations, and proximity variables to neighborhood assets (e.g. parks

and high performing schools).

3. Neighborhood Housing Market Typologies

A thorough understanding of how neighborhood’s change over time and the impact of city-wide

or target-neighborhood strategies allow cities to plan for and address changing conditions in

neighborhoods. Neighborhood housing market typologies is a critical tool in this analysis.

Typologies allow cities to capture and categories neighborhood conditions, and link applicable

5 Triage policy is defined as “the exclusion of severely declining urban areas from services and programs on the grounds that the intensity of their needs cannot be met and the provision of services is therefore inefficient (Marcuse, 1982)”.

21

strategies to address challenges. As a policy tool, neighborhood housing market typologies are

not new. In the 1980’s Goetze and Colton examined housing dynamics, census data, and market

trends in Boston, and identified a link between housing conditions, neighborhood stages, and

public interventions. They used housing conditions as key elements to capture neighborhood

dynamics and change, and developed a two-dimensional conceptual framework for

neighborhood classification.

3.1. Initial Development of the Neighborhood Typology (1980)

Based on life-cycle theories and filtration concepts of neighborhood change, Goetze and Colton

(1980) argued that similar structures built in dissimilar neighborhoods of the same city at the

same time could experience various housing dynamics, thus discounting the life-cycle theory

which focuses on the aging of housing structures and corresponding neighborhood conditions

(1980: 33). Condition and obsolesce, according to Goetze and Colton, did not go hand in hand.

To demonstrate this, Goetze and Colton’s created a neighborhood classification system of

neighborhoods similar to life-stage models, analyzing the following indicators: rising sales price,

turn-over rates, vacant units, foreclosures and tenure. Based on housing indicators, Goetz and

Colton developed a market dynamic matrix to identify three market types: rising (more

applicants than vacancies), stable (supply and demand balance out), and declining (fewer

households seeking to remain than available dwellings). Each market type was coupled with

corrective measures to address neighborhood conditions. Measures or strategies to counteract

market conditions included: assisting residents in upgrading efforts and constructing additional

22

housing for good areas; providing code enforcement, tax incentives and repairs for fair areas;

and, relocation, demolition, and homesteading for poor areas.

By monitoring and analyzing market changes through neighborhood classification, Goetze and

Colton inferred that cities could intervene and create environments to build household

confidence and ensure households remain in the neighborhood; or, improve neighborhood

conditions to attract new households. In this framework, city interventions impact householders’

decisions, and indirectly impact neighborhood conditions and markets. Whereas Downs argue

the effectiveness is based on neighborhood stages, Goetze and Colton suggest effectiveness of

the intervention relies on the city’s ability to counteract neighborhood dynamics, and increase

demand where housing markets are weak (1979:34).

3.2. Neighborhood Housing Market Typology (2000)

More than 20 years later, numerous cities such as Richmond Virginia; Baltimore, Maryland;

Philadelphia, Pennsylvania; Cleveland, Ohio; Chicago, Illinois; Flint, Michigan; and,

Indianapolis, Indiana, are employing Goetze and Colton’s neighborhood market classification

technique through the development of neighborhood housing market typologies. Methods and

statistical analysis, and purposes for typologies differ among these cities, but the general concept

remains the same: provide an understanding of neighborhoods’ strengths and weaknesses, and

prescribe reuse strategies to effect change or stimulate market forces (Mallach, 2006).

3.2.1. Typology Methods

Cities’ typologies consist of several methods to categorize city neighborhoods into distinct

housing market typologies. Simple methods consist of z-scores (Cleveland, Ohio Typology

(2003)) or descriptive data aggregated to census blocks into categories based on levels of distress

23

or market criteria (Mallach, 2003). A more common, robust method is the statistical cluster

analysis, which provides several methods to categorize neighborhoods based on selected

neighborhood indicators.

Cluster analysis is an empirical technique developed in the 1930’s by social scientists and further

defined by biologists in the 1960’s (Julnes, 1999; Johnson and Wichern, 1996) to explore

patterns and relationships among various phenomenons. Cluster analysis refers to the

classification or ordering of entities into similar groups or classes to explain a phenomenon.

Similarities among groups or classes are presented in meaningful categories which involve,

“clusters of objects that display small within-cluster variations relative to the between-cluster

variation” (Julnes, 1999: 295). Meaning, the object represents the entity, and they are

differentiated by clustering the objects into categories based on measures of similarities, and

each category presents distinct characteristics. Cluster analysis differs from principle component

analysis and factor analysis, which attempt to classify objects and variables based on similarities;

in that this processes does not attempt to identify continuous dimensions. Cluster analyses

identify separate categories to place occurrences or phenomenon.

3.2.2. Defining Neighborhood Clusters

The process of defining a cluster is simple. In clustering analysis, data are not standardized

because the purpose is to develop meaningful variations among variables. A method is used to

create categories based on Euclidean distances or correlation coefficients. Methods include three

solutions: hierarchical methods, iterative partitioning methods, and factor analytic methods. The

hierarchical method is for entities that belong to higher-order clusters, and are used for smaller

sample sizes, while the iterative portioning or K-means methods calculates the multi-dimensional

24

center of the clusters and then iteratively reassigns the entities into clusters based on whose

center is closest (Julnes, 1999:582). The factor analytic method is a categorical version of factor

analysis in which entities are reassigned in this method as the analysis process occurs, and

clustering of objects is based on the characteristics of measured variables. The analysis of

variance (ANOVA) is used to ensure statistically significant differences exist among different

clusters.

Hierarchical and k-means are commonly used methods in the literature to explore submarkets

and other real estate classifications (Bourassa, et al., 1999; and Gershoff, et al., 2009) and to

develop robust neighborhood housing market analyses. In many studies, these methods are

coupled with principal component analysis to initially identify factors or variables, which are

subsequently inputted in the cluster method to develop the typology (Bourassa et al., 1999 and

Weisbourd, 2009). Factor analytic methods are used by practitioners to develop clusters based on

access to data and limited capacity to run more robust models.

3.2.3. Neighborhood Cluster Indicators and Linked Strategies

Cities’ indicators, housing data, and strategies used to develop neighborhood typologies differ

based on the uniqueness of the city and neighborhood conditions. Following are examples of

neighborhood health variables used in housing market typologies through the cluster analysis

(Based on Philadelphia housing market typology):

Housing tenure Age of housing Percentage of units demolished Percentage of vacant properties Percentage of properties deemed dangerous or imminently dangerous Percentage of properties categorized as commercial Year, value, type, and price of most recent sales Total count of residential units

25

Percentage of households surveyed with high- and very high-risk credit scores

Commonly used typologies categories with definitions and corresponding strategies include:

Stable (Protection): areas that are well-established and successful, where the goal is to maintain the existing housing stock and improve services.

Transitional (Revitalization): areas with aging but still viable housing stock, often close to declining industrial areas. The goal here is to invest strategically to enhance desirability of housing and encourage stability.

Distressed (Redirection): areas with declining incomes and major social service needs, or whose housing stock cannot compete with housing elsewhere in the city or metro area. The goal here is to work toward major landscape transformation and attract much-needed new investment. (Mallach, 2006: Minneapolis Neighborhood Revitalization program).

Table 3 provides market typologies which employ both robust cluster analysis and simpler z-

score methodology, from Philadelphia, PA; Baltimore, MD; Cleveland, OH; Memphis, TN;

Kansas, MO; and Indianapolis, IN.

Similar to Goetze and Colton work, which identify matching strategies to address market

conditions, strategies used by cities in table 3 are based on assumptions of neighborhood change

models about local conditions, and mitigating factors to forestall neighborhood decline and

stabilize the supply of housing units and demand. It is assumed that interventions applied in

different neighborhoods based on typologies will influence existing and potential residents’

attitudes and long-term expectations about the future of the neighborhood. For example,

neighborhoods which exhibit high vacancy rates, low homeownership, and substantially low

household incomes may be classified as distressed neighborhoods, and the corresponding

strategy to address the distress is to redevelop neighborhood blocks or redirect investments

around existing neighborhood assets. Massive demolition of vacant and abandoned properties

may be more appropriate for this market type based on neighborhood stage models because

vacancy rates are high and units have deteriorated.

26

Table 3: Neighborhood Housing Market Typologies

City Market Typologies Unit of Analysis Indicators Strategy Methodology Purpose

Baltimore, MDCompetitive, Emergining,

Stable, Transitional, Distressed

Census Block Group

Median Home Value Sales, Percent Commercial Land, Percent Rental Subsidies,

Percent Forclosure, Percent Vacant Homes, Percent Home

Ownership, Percent Single Family Homes, and Percent

Vacant Lots

Preventative measures for competitive/emerging typologies; stabilization efforts for stable and

transitional markets; and demolition and substantial

redevelopment in distressed markets.

Factor analytic cluster analysis

Guide city's investment strategies for both daily

operations and long-term planning; use of data to

target HUD Neighborhood

Stabilization program dollars.

Philadelphia, PA

High Value/Appreciating, Steady, Transitional,

Distressed, and Reclamation

Census Block Group

Median Sale price, Variance in sales price, Percent Rental Subsidies, Vacancy factor,

Percent Foreclosure, Percent Commercial use, Percent Owner Occupied, Housing

Units per Acre, New Construction

Invest in areas with the greatest potential, close to strong markets or with stable real estate market;

and extensive redevelopment in distessed areas.

Factor analytic cluster analysis

Program was designed to renew and strengthen

Philadelphia’s urban neighborhoods through specific public action.

Kansas, MO

Developing Areas, Conservation Areas,

Stabilization Areas, and Redeveloping Areas

Census Tract

Population change, median household income,

unemployment, household on public assistance, persons below poverty level, persons with high school education, #

crimes, single headed households, median housing value, housing unit change, vacant housing, residential building/demolition permits

ratio, owner occupied housing, household income to housing

payment ratio

Preserve what is valuable and prevent decline by

addressing problems and their root causes. Undertake many

“standard actions” simultaneously to address housing

maintenance/rehabilitation, code enforcement

z-score; final score calculated as the average of all z-scores for each

variable within each block group

Provide neighborhood assessments for long

term planning intiatives, which Strategically apply

public and private resources in a way that

is based on theexisting conditions,

trends, opportunities, strengths and needs of

diverse areas.

Cleveland, OHRegional Choice, Stable,

Transitional, Fragile, Distressed

Census Tract

Median Assessed Value, Percent change in Median

Value, Net Change in number of single family, Foreclosure rate, homeownership rate, boarded up/condemnation rate, housing

rate below fair, vacant and distressed structure rate,

demolition rate

Encourage rehabilitation in strong and stable

markets; use NSP funds in conjunction with

HOME, CDBG and LIHTC resources to rebuild

areas; create sustainable homeownership

rehabilitation markets

z-score; final score calculated as the average of all z-scores for each

variable within each block group

Provide information to assist the Department of

Community Development, City of Cleveland, and other stakeholders in the

development of program strategies that promote

urban revitalization.

Indianapolis, IN A, B, C, D, Census Block Group

Housing Vacancy (90-day postal vacancy), total assessed

housing value, and owner occupied (certified property

ownership)

A: Market regionally; B: active code and nuiance

reinforcement; C: selective demolition,

target resources, and limit concentrated subsidized

housing; and D: site acquisition and land

assembly

Factor analytic cluster analysis

Guide more efficient decision making by

matching resources, policies, and strategies

with neighborhood conditions

Memphis, TN

Zone 1: Classic distressed

neighborhoods; Zone 2: Vulnerable "swing"

neighborhoods; Zone 3: Stable neighborhoods of

choice; and Zone 4: Uptrending traditional

neighborhoods

Census Tract

Socio-economic variables, demography population data,

variables related to community development essentials,

housing and neighborhoods, crime, school quality, and

health indicators

Zone 1: Classic distressed

neighborhoods; Zone 2: Vulnerable "swing"

neighborhoods; Zone 3: Stable neighborhoods of

choice; and Zone 4: Uptrending traditional

neighborhoods

Zones geographically categorized based on

market criterion.

Stabilize neighborhoods and develop intervention for foreclosures; overlay zones with foreclosure

data to guide investment decisions

27

Table 3: Neighborhood Housing Market Typologies

3.3. Gaps in the Current Literature

Two significant gaps in the current literature which address the link between neighborhood

market typologies and revitalization efforts are limitations in the development of neighborhood

housing market typologies and the lack of robust, spatial models which identify investments’

impacts on surrounding property values.

3.3.2. Development of housing market typologies

Julnes (1999) notes the cluster analysis process is simple, and straightforward, however, the

cluster analysis is still an emerging statistical method and is not extensively supported by a body

of statistical reasoning literature. In most typologies, the variables are observed over a one to

two year period which provides very little indication of neighborhood change demonstrated by

the turn-over in housing units, population change, and other variables, or capture neighborhood

change or dynamics. The focus on real estate activities ignores other neighborhood quality

elements, which influence neighborhood choice, residential mobility, and impact market forces.

Examples of these elements include: crime, quality of neighborhood schools and business

establishments (Galster, et al., 2006; Galster et al. 2000; and Sampson, et al., 2002).

Additionally, little information is provided on threshold levels of how strategies should change

over time based on neighborhood improvement or movement along the continuum. Studies have

used cluster analyses to develop typologies to define housing submarkets (Bourassa, et al.,

1999), but the literature does not apply these methods to evaluate where or within which housing

market the intervention is more likely to be effective. Therefore, the use of typologies to

evaluate and prescribe program interventions must be re-examined to address neighborhood

change indicators, expand to include more socio-economic variables which define housing

28

markets, and create a clearer understanding of transitions of neighborhoods across market type

when interventions are applied.

3.3.3. Exploring Spatial Hedonic Regression Models

Hedonic regressions and difference-in-difference variations of hedonic models are commonly

used in studies to evaluate the impact of housing investments, and these studies attempt to

address spatial dependence and capture non-fixed spatial variables and neighborhood effects

through the phenomenon of spatial drift of coefficients6 (Can 1990 and 1992; Ding, 2000).

Though methods are robust, OLS regression model may be insufficient to capture spatial

dependence in housing markets through spatial interaction and diffusion effects, hierarchies of

place and spatial spillover. In the last ten years, studies have begun to employ the geographic or

locally weighted regression (GWR) models to analyze spatial impacts, through spatial hedonic

regression models.

Spatial hedonic regression models, developed by Anselin (1988 and 1990), are used to

investigate spatial non-stationary estimates, and capture variability in the quality of amenities,

which is limited in the conventional hedonic model using OLS((Cho et, al., 2006; and

Weissbourd, et al., 2009). These models’ ability to capture housing markets spatial dependence

are generally found in studies valuing the quality of the environment (Carruthers and Clark,

2009; and Cho et, al., 2006), which apply GWR models to estimate the marginal implicit price

and series of implicit demand functions to describe the relationship between the price of distance

from environmental amenities and dis-amenities. Comparing the classic hedonic regression

model with spatial model, studies show improvement in the predictive ability of the model and

6 Spatial drift of coefficients attempt to capture neighborhood effects by using interaction terms, a concept built upon by the interaction of Cartesian coordinates with housing attributes to generate a unique location values (Fik et al.,

2003 in Carruthers et al., 2009).

29

ability to address highly correlated variables due to spatial dependence. GWR models are not

constrained by rigid boundaries and may inform policy makers on the spatial impact of

investments.

The next section provides a detailed development of the research question to guide this project.

RESEARCH QUESTION

This study will address the relationship between neighborhood market typologies and

stabilization efforts with one specific question:

1. To what extent do varying market conditions, as a measure of neighborhood housing

market typologies, impact the effectiveness of residential housing investments in

neighborhoods? Effectiveness will be measured as positive or negative impacts on

surrounding residential sales prices. There are four mitigating factors to consider. (1)

Magnitude; (2) Concentration; (3) Proximity; and (4) External Neighborhood conditions

A. Is the effect based on the amount of public investment put into the neighborhood?

B. Is the effect based on spatial concentration or density of investments?

C. Is the effect based on whether investments are located within targeted

reinvestment zones?

D. Is the effect based on whether investments are in proximity to

neighborhood/community assets?

E. Do other neighborhood conditions or spatial aspects such as surrounding

redevelopment projects, or other unique neighborhood indicators (organizational

structure of neighborhood groups) influence stabilization efforts?

30

These questions will be examined within each market typology, and addressed through

quantitative and qualitative analysis based on five hypotheses.

Hypothesis 1:

The first hypothesis states the effects of residential investment on nearby property values is

based on where neighborhoods exist within a market continuum. To test this hypothesis,

Baltimore block groups will be divided into neighborhood housing market typologies. Block

groups will represent the neighborhood unit of analysis. This step will be completed using the

statistical cluster technique to isolate four to five typologies. Typologies will include stable,

minor decline, clear decline, heavily deteriorated, and declined. This hypothesis assumes single-

family homes in fewer distressed markets will have a greater appreciation than homes in more

distressed markets.

Hypothesis 2:

This hypothesis addresses public investment. Based on the amount of public investment in each

market typology, the value of single-family homes in minor decline markets will experience

increases in values and require less public investment, while more distressed markets will require

higher investment for positive change (Galster, 2005). All things equal, this assumption is based

on the City’s ability to leverage private and existing homeowners’ investment in neighborhoods.

It assumes public investment in minor decline markets will result in a greater increase in

property values than the same investment in distressed markets. Alternatively, more public

investment will be needed in distressed neighborhoods than in minor decline markets to achieve

the same increase in property values.

31

Hypothesis 3:

This hypothesis addresses the spatial concentration of investments. Current literature states

targeted and highly concentrated investments achieve greater impact because investments are

sustained and highly visible (Galster, 2006, Ding, 2000, and Ellen, 2001). Therefore, nearby

property values will be positively impacted if they are located in areas where public residential

investments are highly concentrated. These effects may differ by market typology. This

hypothesis assumes that the value of nearby single-family homes located in areas with high

concentrations of city investments will experience increases in housing values in comparison to

single-family homes in areas with low concentrations of city investments.

Hypotheses 4 through 5:

The last two hypotheses address proximity and neighborhood externalities. The first hypothesis

addresses and controls for proximity to neighborhood assets, such as community institutions,

parks, or public transit options. The second hypothesis attempts to control for major

redevelopment projects, such as HOPE VI projects or housing and commercial developments in

the City based on spill-over effects. Both hypotheses assume proximity to these elements will

positively or negatively impact property values (Mallach, 2006; Accordino and Johnson, 2000;

and Rapaport, 1997). Additional neighborhood variables, such as active neighborhood groups,

will be included in the model and qualitative analysis will be completed.

METHODOLOGY

1. Cluster Analysis

A hybrid cluster analysis using the hierarchical and alternate version of K-means cluster method

will be used to develop the neighborhood housing market typology for Baltimore City. A

neighborhood housing market typology currently exists for the City; however, there are

32

significant limitations in the variables selected to define neighborhood change and neighborhood

health indicators. This study will identify four to five housing market types to be developed

through the cluster analysis, and assumes that there is a statistical relationship between increases

in property values across neighborhood market types. Two cluster models for 1990 and 2000

will be developed to complete descriptive analysis of neighborhood transition between

neighborhood types, and impacts of investments across market types. A cluster analysis will also

be completed for 2004/2005 data.

2. Empirical Model

The hedonic price regression method is a common technique used to determine the effects of

neighborhood attributes on housing prices. This method is used to determine the estimated

demand for housing attributes and the impact of public services and interventions. Hedonic

models estimates property values based on structural variables, such as housing type, number of

bathrooms, heating units, and presences of a garage, or basement (Ding, 2000). Neighborhood

variables include: institutions, services and public safety; and distance measures such as distance

to the central business district, employment centers and transportation nodes (Knapp and Ding,

2003; and Ding, 2000). Additional, socio-economic variables or indicators which impact

households’ opportunities and neighborhood services and conditions, based on John Powell

seminal work7 will be included.

For this model, this study assumes there is a relationship between neighborhood market

conditions and the effect of residential investment on nearby property. Sales in 2004 and 2005,

within 150 and 300 feet of the residential investments, will be used to determine impact. Studies

identify 150 to 300 feet as measurable distances to analyze new and rehabilitation residential

7 John Powell (1999) work focus on the effects of race, poverty, and urban sprawl in the U.S., and highlights lack of access to regional opportunities create poverty, and significant barriers for residents.

33

investments (Ding, 2000, and Ellen, 2001). These steps will require the use of GIS to measure

distance variables and capture investments around home sales. The basic model is expressed as:

R = f(S, N, NR, IR)

Where

R= rent (sale price);

S = Structural characteristics;

N = Vector of socio-economic neighborhood characteristics (including location variables);

IN = Vector of investment of new construction; and

IR = Vector of rehabilitation; and

e = error term

Thus the model is:

ln (R) = (β0 + β1S+ β2N + β3NR + β4IR +e) (Model 1)

Due to the non-linear form of prices and values, the dependent variable, sales price will be used

as a common form semi-log of home sales price for this model. The structure attributes and

neighborhood characteristic variables will represent control variables.

To address the research question of public investments and level of concentration or scale, a

second model will extend the basic model to include public investment expenditures and density

of investment. These investments will be measured at the block level, and will include dummy

variables for total cumulative city residential investment below $20,000 and above $20,000

(Galster, 2001). Concentration (number of HOME investments within 150 and 300 feet) and

scale measures (number of housing units in a building) will be based on the number of

investments and number of housing units in the investment, respectively. This impact will be

tested using the spatial drift of coefficients phenomenon. This approach is used in Ding (2000)

34

study, and is a common approach to capture spatial lagged variables and neighborhood effect in

the hedonic model. Additionally, other variables which demonstrate neighborhood market

conditions such as: household income, foreclosure rates, and homeownership rates will also be

included as interaction variables.

Each variable will be interacted with HOME Investments, and form the following model:

ln (R) = (β0 + β1S+ β2N + β30NR + β40IR+ β31NR*N+ β411R*N +e) (Model 2)

Targeted reinvestment zones, such as Baltimore’s Healthy Neighborhood Initiative blocks, Main

Street Reinvestment area and other neighborhood initiatives, and areas with strong neighborhood

groups, will be included as a third model to control for other neighborhood effects and

externalities. Strong neighborhood organizations will be represented in the model with groups

which have operating support above $20,000. Lastly, proximity measures using geographic

information systems will be used to identify sales in proximity to neighborhood assets such as

parks, high achieving schools, and other large public investments.

OLS models will be followed by a second stage demand function and geographic weighted

regression model to address spatial dependences and auto-correlation specifications.

3. Definitions and Measures: House Price, Stabilization and Neighborhood Unit

3.1. House Price

Housing prices are extensively used in empirical studies to examine home investment impacts on

neighborhood conditions, quality, and spill-over effects (Ellen, et al., 2001; Ding, 2000; Ding

and Knaap, 2003; Galster et al., 2005). This variable is an imperfect measure of quality, and is

influenced by residents’ socio-economic status, neighborhood amenities, and dis-amenities, and

public services (Ding and Knaap, 2003). House price is commonly used to capture

35

neighborhood effects and included in studies as a common form semi-log of home sales price,

due to its non-linear form.

3.2. Neighborhood Unit

There is little consensus among scholars in defining the urban neighborhood (Galster, 2003;

Grisby et al., 1987). Neighborhoods are defined based on natural or political boundaries,

characteristics of householders, or even by sub-markets and housing structural types. For the

purpose of this study, the neighborhood unit will be defined based on Census block groups. A

Census block is the smallest level of geography designated by Census Bureau, and generally

approximate actual city street blocks in urban areas (US Census). Units in close proximity will

share common attributes and tend to be classified in the same sub-market (Galster, 2003);

therefore, the Census block group unit of analysis was selected to represent the neighborhood

unit.

3.3. Data Set

The data set for this study will include three sources: 1990 and 2000 census, Maryland Property

View, and Baltimore Department of Housing and Community Development (BDHCD) (see table

4-7). The study will focus on the City HOME funds. HOME partnership is administered by

HUD and provided to cities to create affordable housing for low-income households. A local

government can use HOME funds for grants, direct loans, loan guarantees and other forms of

credit enhancement, or rental assistance. Eligible households for HOME assistance vary with the

nature of the funded activity. For rental housing and rental assistance, at least 90 percent of

benefiting families must have incomes that are no more than 60 percent of the HUD-adjusted

median family income for the area. For rental units, at least 20% of the units must be occupied

36

by families with incomes that do not exceed 50% of the HUD-adjusted median. For households,

incomes must not exceed 80 percent of the area median. These requirements may impact which

market housing investments are located.

For the City of Baltimore, HOME Partnership funds are used for rehabilitation and new

construction of owner and rental residential properties. HOME housing investment data

includes: new construction and rehabilitation collected from 1998 through 2003. The (BDHCD)

provides information on the location, amount, type (new construction or rehabilitation), and year

of HOME Partnership investments. This data will be linked with Maryland Property View data

for additional information on the structural elements of housing units and socio-economic data

compiled from the US census. Data will include sales prices of single-family, owner-occupied,

arm-length transactions in 2004 and 2005. This timeframe was chosen to capture city sales

activity before the significant, negative impact of the economic and housing market crisis which

occurred in 2006 through 2009.

The 2000 Census will be used to obtain data on the block level, including socio-economic and

housing characteristic neighborhood indicators. Socio-economic data will include income,

education, and employment rates. Housing variables will include homeownership, occupancy,

age, structural variables, and number of units. These data will be complemented with Maryland

Property View data which includes structural elements on properties, sales data and assessed

values. Additionally, Maryland Property View data provides verification on occupancy data and

type of transaction.

37

TIMELINE

I plan to finalize and defend this proposal in April 2010. During the months of May, June, and

July of 2010 I will complete data collection. During June and July of 2010, interviews will be

conducted with neighborhood groups, organization and City departments to discuss

neighborhood activities. From August to November of 2010, I will complete data analysis, run

model, and begin to write up the analysis findings. I anticipate defending the dissertation in the

fall semester of 2011.

38

Table 4: Data Variables

Variables Description Source Year

Sales PriceMaryland

Property View2004-2005

%PERMIT

Percent of residential properties that undergo rehabilitation investment

above $5,000

Baltimore DHCD

1998/2003

%VACSTRUC

Percent of residential properties that are vacant and abandoned at year's

end

Baltimore DHCD

1998/2003

%VACLOTPercent of residential lots

that are vacant and abandoned at year's end

Baltimore DHCD

1998/2003

%VIOL

Percent of residential properties with other types

of housing violations (excluding vacant) at year

end

Baltimore DHCD

1998/2003

AVGSALE Median Sale Price Property View 1998-2003THHSOLD Total Housing Units Sold Property View 1998-2003

AVG#DAYSMedian number of days on

the marketBaltimore

DHCD1998/2003

%HHOWN

Percent of housing units owner-occupied (single family homes and condo

units only)

Property View 1998/2003

%MORTGRate of properties under

mortgage foreclosureProperty View 1998/2003

AVGHHVALUE Average housing values Property View 1998/2003

AVGAGEAverage age of housing

unitsProperty View 1998/2003

Grade Property Grade Property View 1998/2003

HOUSING

39

Table 5: Data Set Variables (Continued)

Variables Description Source Year

HSSCOREMaryland High School Assessments (HSA)

BCPSS 1998/2003

ELEMSSCORE

Maryland School Assessment (MSA) Test

scores for 3rd, 5th, and 8th grades

BCPSS 1998/2003

Public TransitAccess to public transit

stopsBaltimore

DHCD

Toxic Waste Proximity to toxic wasteBaltimore

DHCD

ParksAccess to parks and open

spaceBaltimore

DHCD

Commercial Access to commercialBaltimore

DHCD%Commercial Percent of commercial

Water Access to waterBaltimore

DHCD

Unemployment Unemployment rate BLS 1998/2003

Change UnemployChange in unemployment

rate

% PopulationPercent of population age 25-64 that have some college

and aboveCensus 1990/2000

HH Income Median Household Income Census 1990/2000

% HH IncomeChange in median hh

incomeCensus

Pov. Rate Poverty Rate 1990/2000Change Poverty Change in poverty rate Census

EDUCATION

MOBILITY/TRANSPORTATION

HEALTH/ENVIRONMENT

SOCIO/ECONOMIC

40

Table 6: Data Set Variables (Continued)

Variables Description Source Year

% Comm. Permit

Percent of commercial properties that undergo rehabilitation investment

above $5,000

Baltimore DHCD

1998/2003

% Comm. Vacant

Percent of commercial properties that are vacant and abandoned at year's

end

Baltimore DHCD

1998/2003

CrimeRt Crime RateBaltimore

DHCD1998/2003

VioCrim Violent crime rateBaltimore

DHCD1998/2003

HRMort High Risk Mortgages HMDA 1998/2003

Associa.Active Neighborhood

associations and block clubs

Baltimore DHCD

CHAP CHAP ZoneBaltimore

DHCD2004-2005

Herit. Zone Heritage ZoneBaltimore

DHCD2004-2005

HNI Heathy NeighborhoodHealthy

Neighborhood Intiative

2004-2005

EC Enterprise ZoneBaltimore

DHCD2004-2005

EZ Empowerment ZoneBaltimore

DHCD2004-2005

BDCZBaltimore Development

Corp ZoneBaltimore

DHCD2004-2005

Hope VIHUD HOPE VI Project

investmentsBaltimore

DHCD2004-2005

EDBDI East Baltimore

Development Inc. InitiativeBaltimore

DHCD2004-2005

<20,000 City Investment < $20,000Baltimore

DHCD2004-2005

>20,000 City Investment > $20,000Baltimore

DHCD2004-2005

TARGET REINVESTMENT ZONE

MAJOR REDEVELOPMENT

PUBLIC INVESTMENT EXPENDITURE

NEIGHBORHOOD STABILITY/QUALITY

41

1. APPENDIX 1.1. Bibliography References 1.2. Baltimore Neighborhood Market Typology Map A1, 2

1.1. BIBLIOGRAPHY REFERENCES

“Baltimore, Maryland: A Neighborhood Investment Strategy Paying Off” (National City of Leagues, October 2005)

A. C. Helms. "Understanding Gentrification: An Empirical Analysis of the Determinants of Urban Housing Renovation," Journal of Urban Economics Vol. 54 (2003): 474-498.

Abelson, P. “Properties Prices and the Value of Amenities,” Journal of Environmental Economics and Management, Vol. 6: 11-28

Accordino, John and Gary Johnson. “Addressing the Vacant and Abandonment Property Problem.” Journal of Urban Affairs, Vol. 22, No. 3 (2000): 301-315.

Alan Mallach, “Bringing Buildings Back: From Abandoned Properties to Community Assets,” National Housing Institute and Rutgers University Press (2006): 75-85.

Alan Mallach, “Managing Neighborhood Change,” National Housing Institute (2008) Arnott, R., R. Davidson, and D. Pines, “Housing Quality, Maintenance and Rehabilitation,”

Review of Economic Studies L (1983): 467-494available from the National Housing Institute at: http://www.nhi.org/research/521/managing_neighborhood_change/

Ayse Can “Specification and Estimation of Hedonic Housing Price Models,” Regional Science and Urban Economics Vol. 22 (1992): 453-474.

Baltimore City Consolidated Plan (July 2005). available at: http://www.baltimorehousing.org/inside_plansandreports.asp (accessed: October 9, 2009)

Baxter, Vern and Mikey Lauria. Residential Mortgage Foreclosure and Neighborhood Change. Housing Policy Debate, Vol. 11, No. 3 (2000): 675-699.

Brasington, David and Diane Hite. Demand for environmental quality: A spatial Hedonic Analysis. Regional Science and Urban Economics. Vol. 35, No. 1 (2005): 57-82.

Bourassa, Steven, Foort Hamelink, Martin Hoesli, and Bryan D. MacGregor. Defining Housing Submarkets. Journal of Housing Economics. Vol 8 (1999): 160-183.

Brown, Prudence. “Comprehensive Neighborhood-based Initiatives,” Cityscape Vol. 2(May 1996).

Bruce, Grady D. and Robert E. Witt. “Developing Empirically Derived City Typologies: An Application of Cluster Analysis” The Sociological Quarterly Vol. 12 (Spring 1971): 238-246

Bruch, Elizabeth E. Neighborhood Choice and Neighborhood Change. American Journal of Sociology. Vol. 112, No. 3 (November 2006): 667–709.

C. Rapaport, "Housing Demand and Community Choice: An Empirical Analysis,” Journal of Urban Economics Vol. 42, No. 2 (September 1997): 243-260.

Carmon, Naomi. “Neighborhood Regeneration: The State of the Art” Journal of Planning Education and Research 17, No. 2 (1997): pp. 131-144

Carruthers, John and David Clark. Valuing Environmental Quality: A Space-based Strategy. U.S. Department of Housing and Urban Development working paper # REP 07-01; submitted in final form to the Journal of Regional Science May 2009.

42

Christine M. E. Whitehead, "Urban Housing Markets: Theory and Policy," in Paul Cheshire and Edwin S. Mills, eds., Handbook of Urban and Regional Economics, 3 North Holland, 1999: 1560-1594.

Cho, Seong-Hoon Cho, J. M. Bowker, and William Park. Measuring the Contribution of Water and Green Space Amentities to Housing Values: An Application and Comparision of Spatially-weighted Hedonic Models. Selected Paper prepared for presentation at the American Agricultural Economics QAssociatioin Annual Meeting, Long Beach, California, July 23-26, 2006.

Clay, Philip and Robert Hollister. Neighborhood Policy and Planning. Lexington, Massachusetts: D.C. Heath and Company, 1983.

Clay, Phillip. Neighborhood Renewal. Lexington, Massachusetts: D.C. Heath and Company, 1979.

Cohen, James. “Abandoned Housing: Exploring Lessons from Baltimore” Housing Policy Debate Vol. 12, No. 3 (2001): 415-448.

Cortes, Alvaro. “Estimating the Impacts of Urban Universities on Neighborhood Housing Markets: An Empirical Analysis” Urban Affairs Review Vol. 39, No. 3(2004): 342-375

Ding et al. “The Effect of Residential Investment on Nearby Property Values: Evidence from Cleveland, Ohio” Journal of Real Estate Research ½, Vol. 19 (2000): pp. 25-48

Ding, Chengri, Knaap, Gerrit-Jan. “Property Values in Inner City Neighborhoods: The effects of Home ownership, Housing Investment, and Economic Development.” Housing Policy Debate Vol. 13(2003):701-727.

Ding, Chengri, Simons, Robert, Baku, Esmail. “The Effect of Residential Investment on Nearby Property Values: Evidence from Cleveland, Ohio”. JRER Vol. 19 (2000): 23-48.

Downs, Anthony. Neighborhoods and Urban Development. Washington: The Brookings Institution, 1981.

E.O. Olsen, "The Demand and Supply of Housing Services: A Critical Survey of the Empirical Literature," in HURE-II.Vol. 2 (1987): 989-1022

Ellen et al. “Building Homes, Reviving Neighborhoods: Spillovers from Subsidized Construction of Owner-Occupied Housing in New York City” Journal of Housing Research, Vol. 12, No. 2(2001): 185-216.

Engberg, Rodney A. and Susan W. Friedman. 1999. “State Enterprise Zones and Local Housing Markets.” Journal of Housing Research Vol.10:163-187

F., Redburn, Buss, Terry. Public Policies for Distressed Communities. Lexington, Massachusetts: D.C. Heath and Company, 1982.

Frankel, David M. and Ady Pauzner. Expectations and the Timing of Neighborhood Change. Journal of Urban Economics. Vol. 51, No. 2(March 2002): 295-314.

Fulbright-Anderson, K., Anne C Kubisch and James P. Connell (Eds.). New Approaches to Evaluating Community Initiatives, Vol. 2: Theory, Measurement, and Analysis. Washington, D.C.: The Aspen Insitute, 1998.

Galster, et al. “Measuring the Impacts of Community Development Initiatives: A New Application of the Adjusted Interrupted Time-Series Method” Evaluation Review Vol. 28 (2004): 502-538.

Galster, George and Ronald B. Mincy. Understanding the Changing Fortunes of Metropolitan Neighborhoods, 1980 to 1990. Housing Policy Debate, Vol. 3, No. 4(1998): 303-352.

43

Galster, George C. An Econometric Model of the Urban Opportunity Structure:Cumulative Causation among City Markets, Social Problems, and Underserved Areas.Washington, DC: Fannie Mae Foundation. 1998.

Galster, George, Chris Walker, Chris Hayes and Patrick Boxall. “Measuring the Impact of CDBG Spending on Urban Neighborhoods,” Housing Policy Debate Vol. 15, No. 4(2004): 903-934

Galster, George, Ken Temkin, Chris Walker and Noah Sawyer. “Measuring the Impacts of Community Development Initaitives,” Evaluation Review Vol. 28, No. 6(2004): 1-38.

Galster, George, Roberto G. Queria, and Alvaro Cortes. Identifying Neighborhood Thresholds: An Empirical Exploration. Housing Policy Debate, Vol. 3, No. 11 (2000): 701-732.

Galster, George, Tatian, Peter, Accordino, John. “ Targeting Investments for Neighborhood Revitalization”. Journal of the American Planning Association Vol. 72 (2006): 457-467.

Galster, George. Are Neighborhoods Self-Stablizing? Exploring Enogenous Dynamics. Urban Studies, Vol. 44, No. 1, (2007) :167-185.

Galster,George. Nature of Neighborhoods. Urban Studies, Vol. 38, No. 12 (2001): 2111-2124Gibb, Kenneth, O’Sullivan, Tony. Housing Economics and Public Policy. Malden,

Massachusetts: Blackwell Science Ltd, 2003.Glaeser, Edward L., and Joseph Gyourko. “Urban Decline and Durable Housing.” Working

Paper no. 8598 NBER, Cambridge, MA. (November, 2001).Gershoff, Elizabeth, Sara Pedersen, and J. Lawrence Aber. Creating Neighborhood Typologies

of GIS-based Data in the Absence of Neighborhood-Based Sampling: A Factor and Cluster Anlaytic Strategy. Journal of Prevention and Intervention in the Community. Vol. 37 (2009) 35-47

Greenbaum, Robert and John Engberg.. “An Evaluation of State Enterprise Zone Policies.” Policy Studies Review 17 (2000):29-46.

Grigsby,William and M. Baratz, G. Galster, and D. Maclennan. “Dynamics of Neighborhood and Decline.” Progress in Planning Vol. 28, No.1(1987): 1-76

Grogan, Paul s. and Tony Proscio. Comback Cities. Boulder, Colorado: Westview Press, 2000.Gyourko, Joseph E. and Saiz, Albert, "Urban Decline and Housing Reinvestment: The Role of

Construction Costs and the Supply Side" (May 2003). Federal Reserve Bank of Philadelphia Working Paper No. 03-9. Available at SSRN: http://ssrn.com/abstract=570323

Hays, R. Allen.Chapter 3: Housing Markets and Sub-markets. P. 57-81 in The Federal Government and Urban Housing: Ideology and Change in Public Policy. Second edition. Albany: State University of New York Press, 1995.

Hiller, Amy and Dennis Cullhane. “Abandoned Housing: Exploring Lessons from Baltimore.” Housing Policy Debate Vol. 12 (2001):449-448.

J. Gyourko and A. Saiz, "Reinvestment in the Housing Stock: The Role of Construction and Supply Side Costs," Journal of Urban Economics, (2004): 238-256.

Jerome Rothenberg, George C. Galster, Richard V. Butler and John Pitkin, The Maze Of Urban Housing Markets: Theory, Evidence, and Policy. Chicago: The University of Chicago Press, 1991

John F. Kain and John M. Quigley, "Measuring the Value of Housing Quality," Journal of the American Statistical Association, Vol. 65 (1970): 532-548.

Jud, G. and J. Watts. “Schools and Housing Values,” Land Economics, Vol. 57, No. 3: 459-470.

44

Kasarda, John D. Inner-City Concentrated Poverty and Neighborhood Distress: 1970 to 1990. Housing Policy Debate Vol. 4, No. 3 (1993):253–302.

Keating, W, Krumholz, Norman, Star, Philip. Revitalizing Urban Neighborhoods. Lawrence, Kansas: University Press of Kansas, 1996.

Keating, W., Krumholz, Norman. Rebuilding Urban Neighborhoods. Thousand Oaks, California: Sage Publications, 1999.

Klain, John F., Apgar, William. Housing and Neighborhood Dynamics: A Simulation Study. Cambridge, Massachusetts: Harvard University, 1985.

Koebel, C. Theodore, et al. “Urban Redevelopment, Displacement and the Future of the American City” Community Affairs Office Federal Reserve Bank of Richmond (April 1996).

Kolodny, Robert. Some Policy Implications of Theories of Neighborhood Change, p. 93–110.In Neighborhood Policy and Planning, ed. Phillip L. Clay and Robert M. HollisterLexington, MA: Heath, 1983

Krupka, Douglas J. & Noonan, Douglas S."Neighborhood Dynamics and the Housing Price Effects of Spatially Targeted Economic Development Policy," IZA Discussion Papers 4308, Institute for the Study of Labor (IZA), 2009

Kuhn, Randall and Dennis P. Culhane. “Applying Cluster Analysis to Test a Typology of Homelessness by Pattern of Shelter Utilization: Results from the Analysis of Administrative Data.” American Journal of Community Psychology, Vol. 26, No. 2(1998): 207-232

Lang, Robert. “Did Neighborhood Life-Cycle Theory Cause Urban Decline?” Housing Policy Debate Vol. 11(2000):1-6.

Lee, Chang et al., “The Differential Impacts of Federally Assisted Housing Programs on Nearby Property Values: A Philadelphia Case Study” Departmental Papers (SPP), University of Pennsylvania, 1999.

Little, James T. 1976. Residential preferences, neighborhood filtering and neighborhood change. Journal of Urban Economics Vol. 3: 68-81.

Lynch, A. and D. Rasmussen. “Measuring the Impact of Crime on House Prices,” Applied Economics, Vol. 33 (2001): 1981-1989.

Mallach, Alan. Bringing Buildings Back. Montclair, New Jersey: National Housing Institute, 2006.

Mallach, Allan. “From Abandonment to Reuse: Issues and Policies in Urban Abandonment.” Prepared for seminar hosted by Fannie Mae Foundation, 5 November 2001.

Megbolugbe, Isaac, Linneman, Peter D., Hoek-Smit, Marja. “Understanding Neighborhood Dynamics: A Review of the Contributions of William G. Grisby.” Urban Studies Vol.33 (1996): 1779-1795.

Megbolugbe, Isaac, Marja C. Hoek-Smit, and Peter D. Linneman. Understanding Neighborhood Dynamics: A Review of the Contributions of William G. Grigsby. Retrieved from http://knowledge.wharton.upenn.edu/papers/700.pdf

Metzger, John T. Planned Abandonment: The Neighborhood Life-cycle theory and National Urban Policy. Housing Policy Debate, Vol. 11, No. 1 (2000): 7-40.

Metzger, John T.. “Planned Abandonment: The Neighborhood Life-Cycle Theory and National Urban Policy.” Housing Policy Debate Vol. 11(2000): 7-40.

45

Oakley, Deirdre, Hui-Shien, TSAO. “A New Way Of Revitalizing Distressed Urban Communities? Assessing The Impact Of The Federal Empowerment Zone Program.” Journal of Urban Affairs Vol. 28 (2006):443-471.

Oakley, Deirdre, Tsao, Hui-Shein. “Socioeconomic gains and slipover effects of geographical targeted initiatives to combat economic distress: An examination of Chicago’s Empowerment Zone.” Cities Vol. 24(2007):43-59.

Ottensmann, John R. “Positive Discrimination, Spatial Targeting, and Urban Policy” Journal of Planning Literature Vol. 2 (1987): 243-253

Palmist, Raymond B. 1992. “Valuing Localized Externalities.” Journal of Urban Economics Vol. 31: 59-67

Pollock, Marcus, Rutkowski, Ed. The Urban Transition Zone: A Place Worth A Fight. J.H Furst Co. ,Inc, 1998

Quercia, Robert G. and George Galster. Threshold Effects and Neighborhood Change. Journal of Planning Education and Research, Vol. 20, No. 2, 146-162 (2000).

Quercia,Roberto,and George Galster. “Threshold Effects and Neighborhood Change.” Journal of Planning Education and Research Vol. 20(2000):146-435.

Reid, Carolina. “Neighborhoods in Bloom: Measuring the Impact of Targeted Community Investments” Community Investments (Winter 2006): 24

Rohe, William, Stewart, Leslie. “Home ownership and Neighborhood Stability”. Housing Policy Debate Vol. 7(1996): 37-81.

Roni, Peter. 1999. “Evaluating Community Development Interventions.” In Urban Problems and Community Development, edited by Ronald Ferguson and William Dickens. Washington, D.C.: Brooking Insitution Press: 521-559; 565-567

Rosenthal, Staurt. "Old homes, externalities, and poor neighborhoods. A model of urban decline and renewal," Journal of Urban Economics, Elsevier, Vol. 63(3), pages 816-840

Rosenthal, Stuart. Old homes, externalities, and poor neighborhoods. A model of urban decline and renewal. Journal of Urban Economics. Vol. 63, Iss. 3,(2008): 816-840.

Rusk, David. Baltimore Unbound: A strategy for Regional Renewal. The Abell Foundation, Baltimore, Maryland,1996.

Sawickim David S. and Patrice Flynn. 1996. “Neighborhood Indicators: A Review of the Literature and an Assessment of Conceptual and Methodological Issues.” Journal of the American Planning Association Vol. 62:165-183

Schill et al. “Revitalizing Inner-City Neighborhoods: New York City’s Ten-Year Plan” Housing Policy Debate Vol. 13, No. 3 (2002): 529-566.

Schilling, Joseph and Naomi Friedman. “The Revitalization of Vacant Properties: Richmond Virginia Case Study” Prepared for the International City/County Management Association (Washington DC).

Schwab, William A. The Predictive Value of Three Ecological Models: A Test of the Life-Cycle, Arbitrage, and Composition Models of Neighborhood Change. Urban Affairs Review, Vol. 23, No. 2 (1987):295-308

Schwartz, Alex. “New York City and Subsidized Housing: Impacts and Lessons of the City’s $5 Billion Capital Budget Housing Plan” Housing Policy Debate Vol. 10, No. 4 (1999): 839-877.

Schwirian, Kent. “Models of Neighborhood Change.” Annual Review of Sociology Vol. 19(1983):83-102.

46

Simons et al. “The Value Impact of New Residential Construction and Neighborhood Disinvestment on Residential Sales Price” Journal of Real Estate Research ½, Vol. 15 (1998): 147-161.

Stephen Shephard, “Hedonic Analysis of Housing Markets,” in Handbook of Regional and Urban Economics, Cheshire and Mills (eds.), Vol. 3(1999): 1595-1635.

Temkin, Kenneth and William Rohe. Neighborhood Change and Urban Policy. Journal of Planning Education and Research, Vol. 15(Apr 1996): 159 - 170.

Turnham, Jennifer, Bonjorni, Jessica. “Review of Neighborhood Revitalization Initiatives.” Prepared for the Neighborhood Reinvestment Corporation by ABT Associates (Washington DC), 2002.

Van Ryzin, Gregg G. and Andrew Genn. “Neighborhood Change and the City of New York’s Ten-Year Housing Plan” Housing Policy Debate Vol. 10, No. 4 (1999): 799-838.

Vaughan, Roger J. The Urban Impacts of Federal Policies: Vol. 2. Economic Development. Santa Monica, CA: RAND Corporation, 1977

Wachter, Susan and Kevin Gillen. “Public Investment Strategies: How They Matter for Neighborhoods in Philadelphia: Identification and Analysis” For the William Penn Foundation (October 2006).

Weicher, John and Thomas Thibodeau, “Filtering and Housing Markets: An Empirical Analysis,” Journal of Urban Economics, Vol. 23 (1988): 21-40.

Whalley, Diane. “Neighborhood Variations in Moderate Housing Rehabilitation Program Impacts: An Accounting Model of Housing Quality Change.” Economic Geography Vol. 64, No. 1 (1988): 45-61.

Wu, J. et al. “Amenities in an Urban Equilibrium Model: Residential Development in Portland, Oregon,” Land Economics, Vol. 80, No. 1: 19-22

www.nlc.org/ASSETS/8E57CCBAB45046208EB67726EF27EE78/Chap%202%20Baltimore.pdf – (November 5, 2007).

Yannis M. Iaonnides and Jeffrey E. Zabel. "Neighborhood Effects and Housing Demand." Journal of Applied Econometrics, Vol. 18 (2002):563-584.

47

Competitive             

Neighborhoods in this category, like Federal Hill, Canton and Homeland, have robust housing markets with high owner-occupancy rates and high property values. Foreclosure, vacancy and abandonment rates are all very low.  Most direct market interventions are not necessary in the Competitive market  Basic municipal services such as street maintenance are essential to maintaining these markets.  While densities do vary single-familydetached homes predominate and these areas typically don’t have a mix of housing types. 

Emerging            

48

Neighborhoods in the “Emerging” category, such as Abell, Hampden and Mt. Vernon, have robust housing markets but with home ownership rates slightly below the citywide average; this category appeals to property owners interested in tapping into a strong rental market  Median sales price is above $244,000.  Additional incentives for development and investment in the Emerging market would recognize its potential for growth.  There is more variety in housing types and more commercial areas than in the competitive cluster.

Stable         

This cluster includes neighborhoods such as Reservoir Hill, Lauraville and Violetville.  Median sale price is around $160,000 and the rate of foreclosure is just below the City average of 5%.  In Stable markets, the City should consider stabilizing and marketing any vacant houses.  Traditional housing code enforcement is also essential to maintain the existing housing stock.  Home ownership is still significant at 55%.

Transitional             

Neighborhoods in the "Transitional" category, such as Allendale, Belair Edison and Kenilworth Park, are found typically at the inner edge of the stable neighborhoods. These neighborhoods have moderate real estate values with median sale prices between $80,000-$100,000, with higher median sales in areas with commercial land uses.  Foreclosure rates are slightly higher than average, but occupancy rates are still higher than average.  This cluster also has the highest rate of rental subsidy.  The City should support homeowners who may be facing economic hardships due to the national economy.

Distressed                 

These neighborhoods, which include Middle East, Penn North and Westport, have nearly 4 times the levels of vacant homes and vacant lots as found in other categories.   Sale prices typically range from $36,000 - $40,000.  Distressed markets tend to rely on comprehensive housing market inventions, such as site assembly and tax increment financing.   One of the six criteria for identifying the Growth Promotion Areas includes neighborhoods located in distressed markets.  Demolitions in the Distressed markets should be clustered to create potential for greater public safety as well as marketability.  The housing type here is predominately rowhouse.

49