abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 1
CONTENTS
INTRODUCTION ............................................................................................................................................................. 2
I. NEIGHBORHOOD CHANGE IN BALTIMORE: AN OVERVIEW ................................................................................... 5
A. Citywide trends ......................................................................................................................................................... 6
B. Neighborhood trajectories ...................................................................................................................................... 11
II. THE DIMENSIONS OF NEIGHBORHOOD CHANGE IN BALTIMORE ....................................................................... 16
A. Demographic change .............................................................................................................................................. 16
1. Black population change ..................................................................................................................................... 16
2. White and Latinx population change .................................................................................................................. 19
B. Housing market change ........................................................................................................................................... 20
1. Real estate market dynamics .............................................................................................................................. 22
2. Homeownership ................................................................................................................................................. 30
III. KEY NEIGHBORHOOD CLUSTERS ........................................................................................................................... 33
A. Predominantly black moderate-income neighborhoods ........................................................................................... 34
B. The Northeast Triangle ............................................................................................................................................ 36
C. Gentrifying neighborhoods ...................................................................................................................................... 40
IV. CLOSING COMMENTS .............................................................................................................................................. 46
APPENDIX I: METHODOLOGY ....................................................................................................................................... 50
APPENDIX 2: CORRELATIONS BETWEEN ECONOMIC, DEMOGRAPHIC,
AND HOUSING-MARKET FACTORS ............................................................................................................................... 52
Drilling Down in Baltimore’s Neighborhoods:
Changes in racial/ethnic composition and
income from 2000 to 2017
by Alan Mallach
Prepared for the Abell Foundation
Baltimore, Maryland
April 2020
The opinions in this report are those of the author and do not represent the position of either the Abell Foundation
or the Center for Community Progress.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 2
INTRODUCTION
Urban neighborhoods—their dynamics, their eects, and their transformation in the 21st century—have
become one of the most heatedly contested issues in the ongoing discussion about the present and future of
American cities. Few discussions about urban neighborhoods fail to bring in the issue of gentrication, which
has become one of the most widely used (and, I would suggest, misused) words in the urban lexicon. At the
same time, research has increasingly shown the power of neighborhood eects, and how devastating living
or growing up in a distressed, concentrated poverty area can be on people’s health and well-being, their life
expectancy, and their prospects for better earnings and upward mobility later in life.
All of these issues are powerfully present in Baltimore, a city that in many respects is experiencing strong
revival, and in others, continuing decline. Baltimore is a city intensely polarized by race and economic status.
What happens in Baltimore’s neighborhoods, whether they are gentrifying or declining, continuing to
struggle or growing in strength, not only is critical for framing public policy, but also denes what kind of a
city Baltimore is, and what kind of city its residents want it to be. And yet, little is known about what is actually
going on. In recent news posts, Baltimore has been portrayed by President Trump as a “rat and rodent infested
mess,”
1
and by a highly publicized national study as one of the ve most rapidly gentrifying cities in the United
States.
2
Neither is true, of course, although Baltimore undoubtedly has some neighborhoods that are in deep
distress, and some neighborhoods that are gentrifying.
The purpose of this report is to ll in part of that knowledge gap by providing an initial picture of what has
actually been happening in Baltimore’s neighborhoods since the beginning of the 21st century—that is,
to what extent have neighborhoods moved upward economically, moved downward, or stayed largely the
same, and what does that mean in terms of population change, economic condition, and housing markets.
Specically, this report looks at Baltimore’s roughly 200 census tracts, breaking them down into categories
by race and by economic level in 2000 (as described in the next section) and presenting how they’ve changed
since. For example, have neighborhoods that were similar economically but dierent racially in 2000 followed
similar or dierent trajectories from then until now? And if their trajectories were dierent, in what ways?
I look at how neighborhoods have changed; whether they moved upward or downward economically; whether
they gained or lost population; how their racial composition did or did not shift; and how their housing market
conditions, including home ownership rates and sales prices, changed. I tried to get a sense of how many and
which neighborhoods were gentrifying versus declining, and how those trends relate to population change,
particularly in the city’s black population.
Although I look closely at gentrication in Baltimore, this report is not about gentrication as such. It is
about the larger picture of neighborhood change. Gentrication is one part of that picture—a signicant part,
but one that aects only a small minority of Baltimore’s neighborhoods. Most of Baltimore’s neighborhoods
are changing, but in dierent ways. Those changes are being driven by major demographic shifts in the city’s
population, which are in turn driving major changes in Baltimore’s housing market. This local change
parallels the hollowing of the middle class and the increasing polarization of wealth and poverty seen at
the national level.
The largest single factor driving change in Baltimore is that Baltimore is losing its working- and middle-class
families. That factor plays out very dierently across the city’s racial divide. While Baltimore is losing white as
well as black middle- and working-class families, it is gaining a young, highly skilled and high-earning white—
but not black—population through in-migration. As a result, the white population is becoming more auent,
1 Tweet on July 27, 2019; the exact wording was “[Rep. Elijah] Cumming District is a disgusting, rat and rodent infested mess.”
2 National Community Reinvestment Coalition, Shifting Neighborhoods: Gentrication and Cultural Displacement in American Cities (March 2019);
e.g., “Baltimore and Philadelphia metro areas are in the top 10 list, with the fourth and fth largest number of gentried tracts in the study” (p.15).
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 3
and the black population is becoming poorer. That shift reverberates through the housing market. In those
areas to which young, auent white households are moving, housing demand is strong and prices are rising.
In those areas from which working- and middle-class black families are leaving, housing demand is weak,
prices are largely at, and abandonment is distressingly common.
All of those who are engaged in working to make Baltimore a healthier, stronger city need to ask the question:
Why is Baltimore losing its working- and middle-class families, particularly its African American ones? I hope
this report will encourage conversation around that question.
The relationship between demographic change and market change underlies most of what is happening to
Baltimore’s neighborhoods. Markets may not be fair, but they are powerful, and they tend to work in ways
little aected by political decisions and community aspirations. They can be inuenced, but only if they are
thoroughly understood. This is particularly relevant to the subject of gentrication. In light of the role that
this issue plays in many discussions of neighborhood change in Baltimore, I will address it briey in this
introduction, and then in more detail later.
While gentrication may have dierent meanings for dierent people, I dene it here, as do almost all
researchers who study and write about it, as a combination of signicant increases in both house prices and
household incomes in a given area.
3
This reects the understanding that gentrication is about both the inux
of more auent households into an area and the increase in that area’s house prices above some citywide
or regional benchmark. It is not the same as displacement. Displacement is a dicult term because as with
gentrication itself, it can mean dierent things to dierent people in dierent contexts. Most precisely,
displacement refers to an involuntary process—that is, people being forced to leave their homes, for any
number of reasons—as opposed to people voluntarily moving, again for any number of reasons.
The data cannot tell us whether displacement, in the sense given above, is happening in Baltimore’s relatively
small gentrifying area, but the data suggest, in the words of progressive journalist Jarrett Murphy, that “the
issue isn’t displacement of the poor, it’s replacement.”
4
The one available statistical measure of displacement,
the rate of evictions, shows no correlation with any indicator of gentrication. It is a product, above all, of
poverty and high rental cost burden. I am not arguing that there is no displacement, as dened above,
connected to gentrication in Baltimore. As philosophers and scientists have long pointed out, proving
the absence of something is often impossible. I nd, however, that the changes taking place in those
neighborhoods are fully explained by replacement, not displacement. In the course of that process, far more
lower-income white households have been replaced than black households, while in many cases out-moving
larger black households have been replaced by smaller black households.
This report is not a complete picture of Baltimore’s neighborhoods. Neighborhoods are complicated things.
A neighborhood is more than its economic trajectory; it is a product of the commitment of its residents and
property owners, or the absence of that commitment; the presence or absence of neighborhood organizations
and institutions; the levels and character of the interactions among its residents, and much more. At the
same time, understanding economic and demographic trends and how they then drive housing markets
is fundamental to understanding neighborhoods, while many market-related factors, such as trends in
homeownership rates or population movements, powerfully aect the social as well as the economic dynamics
of the neighborhood.
3 In some cases, researchers add signicant increase in educational attainment (particularly the percentage of adults with a bachelor’s or higher
degree) to the rst two factors. This is the basic framework that was adopted by the National Community Reinvestment Coalition in the study cited
earlier; I do not disagree with the framework, although I question how it was applied.
4 Jarrett Murphy, “The Complicated Research on how Gentrication Aects the Poor,” CityLimits, November 20, 2015. https://citylimits.
org/2015/11/20/the-complicated-research-on-how-gentrication-aects-the-poor/
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 4
Finally, this report is not about pointing ngers. The picture painted here is in large part the reection of
powerful and long-term changes in our nation’s demographic and economic character and reects historical
patterns of discrimination, segregation, redlining, and white ight. The ability of the city’s current community
leaders and advocates to rapidly undo the city’s underlying social, economic, and physical challenges is
severely limited. That said, there are many things that can be done to redress inequity, and the strategic
framework recently adopted by the city’s Department of Housing and Community Development represents a
serious, thoughtful eort to begin grappling with many of them.
5
I look rst at the larger city picture, and the patterns of variation in neighborhood change by race and
economic condition, followed by a discussion of the implications of change for population change, as well
as change in house values, home ownership rates, and other key neighborhood indicators. The next section
looks at particular patterns of neighborhood change, including gentrication and the decline of middle-
income neighborhoods. The following section looks at three neighborhood clusters of particular signicance in
Baltimore, and a nal section oers some key takeaways with particular implications for public policy.
Some of the ndings presented in this report may be surprising, and some may be upsetting. That said, it
is important to lay out the facts as dispassionately as possible, so that they can be understood, and so they
can help further the discussion among people who care about the city and its neighborhoods to bring about
change for the better to Baltimore’s neighborhoods. The thrust of this report, however, is not to recommend
what those changes should be, but to lay out, as best I can, the picture of neighborhood change in a dynamic,
beautiful, but deeply challenged city.
To study Baltimore’s neighborhoods, I used census tracts, the unit created and used by the U.S. Census Bureau
for small area analysis. With the city divided into nearly 200 tracts, they are small enough to be meaningful
and relatively homogenous, and they have the advantage that nearly all datasets are available by census tract.
Census tracts are not the same as the Neighborhood Statistical Areas (NSAs) used by the city, but the two are
often roughly comparable. Thus, when I refer to a neighborhood by name in this report, the reader should
understand that I am referring to areas that are similar but not identical to that named neighborhood.
I then segmented the city’s census tracts into categories based on race and economic condition. With respect
to race, I used the percentage of black population, and with respect to economic level, I used the median
6
tract household income. I looked at data for 2000, 2010, and 2017. The breakdown in race and income is
shown in the matrix in Table 1 (income ranges are relative to the citywide median household income). I use
the descriptive terms for the economic and racial composition of the city’s neighborhoods shown in the matrix
frequently in the report.
5 Baltimore City Department of Housing and Community Development, A New Era of Neighborhood Investment: A Framework for Community
Development, November 2018.
6 Median refers to the midpoint of a range of numbers (i.e., that number where half of the numbers are lower and half are higher). It is dierent
from average, which is the sum of the numbers divided by the number in the range.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 5
Table 1: Neighborhood Category Matrix By Economic Level and Racial Composition
ECONOMIC
COMPOSITION
RACIAL
COMPOSITION
Neighborhood Type Range
0-29.9% Black 30-69.9% Black 70-100% Black
Predominantly
White
Mixed
Predominantly
Black
Low Income 0-59.9% X
Moderate Income 60-99.9% X X X
Middle Income 100-149.9% X X X
Upper-Middle Income 150-199.9% X X X
Upper Income 200%+ X
The matrix oers a total of 15 possible neighborhood categories. The actual number of categories is 11, as
shown by “X” in the table. There are no census tracts (e.g., predominantly white low-income tracts) in the other
categories. The income ranges—those within which the tract median falls—for the three time periods I looked
at are shown in Table 2. A tract that had a median income of $40,000 in 2000 would be considered middle
income, and if its median fell to $35,000 in 2010, it would be considered moderate income at that point. A more
detailed description of my methodology is provided in Appendix 1.
Table 2: Income Ranges by Neighborhood Type for 2000, 2010, and 2017
NEIGHBORHOOD TYPE RANGE 2000 2010 2017
Low Income 0-59.9% $0-$18,046 $0-23,631 $0-27,984
Moderate Income 60-99.9% $18,047-30,078 $23,632-39,386 $27,985-46,641
Middle Income 100-149.9% $30,079-45,117 $39,387-59,079 $46,642-69,961
Upper-Middle Income 150-199.9% $45,118-60,156 $59,082-78,772 $69,962-93,282
Upper Income 200%+ $60,157+ $78,773+ $93,283+
Citywide Median $30,078 $39,386 $46,641
I. NEIGHBORHOOD CHANGE IN BALTIMORE: AN OVERVIEW
This section will look at how each of the 11 neighborhood types has shifted since 2000—which types of
neighborhood have moved upward, which have moved downward, and which have stayed largely the same.
When I talk about upward and downward, I am talking about movement from one income range to another;
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 6
that is, when a census tract that was middle income in 2000 becomes an upper-income tract in 2017, or vice
versa. Upward movement is an indicator of potential gentrication, but does not in itself demonstrate that an
area is gentrifying. I explore additional factors that can point to gentrication in a later section of the report.
A. Citywide trends
Before looking at neighborhood categories, though, it is useful to take a quick look at the overall pattern of
change in the city of Baltimore since 2000, as shown in Table 3. Baltimore is a majority-minority city, with 62.1%
identifying as black or African American, 30.6% white, 0.26% American Indian and Alaskan Native, 3% Asian,
1.7% other races and 2.5% two or more races. Five and a quarter percent of the population identies as Latinx
of any race.
7
The city lost 30,000 people between 2000 and 2010, but its population has been roughly stable
since then. Specically, since 2000, the city has lost roughly 30,000 black residents and 30,000 white residents,
while gaining nearly 20,000 Latinx and 5,000 Asian residents. Since 2010, however, Baltimore has lost nearly
10,000 black residents, while its white population has stabilized, largely as a result of in-migration. Figure 1
shows the cumulative change by year from 2010 to 2017 for the city’s white and black populations.
Figure 1: Cumulative Population Change by Race 2010-2017
Baltimore has seen solid economic growth in recent years. Since 2010, the city has added nearly 20,000 jobs.
Household incomes in Baltimore have grown at a rate nearly 50% greater than the national rate over that
period; as a result, the median Baltimore household’s income has risen from 72% to 81% of the national
median. Income growth, however, has been greater among white households, whose incomes have grown at
more than double the rate of black households. As a result, the city’s income distribution has become more
polarized. In 2000, the median black household income was 71% of the median non-Latinx white household
income; by 2017, it had dropped to 49% of the median non-Latinx white household income.
7 Data from the 2014-2018 American Community Survey.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 7
Table 3: Citywide Trends
2000 2010
CHANGE
2000-2010
2017
CHANGE
2010-2017
CHANGE
2000-2017
Total Population (1) 651,154 620,538 -4.7% 619,796 -0.1% -4.9%
Black Population 418,951 399,121 -4.7% 389,222 -2.5% -7.1%
Latinx Population 11,061 22,821 +106.2% 30,729 +34.7% +177.8%
White Non-Latinx
Population
201,566 173,972 -13.7% 170,916 -1.8% -15.2%
Median Household
Income
$30,078 $39,386 +30.9% $46,641 +18.4% +55.1%
Average Annual Rate
of Change
+2.7% +2.4%
Black Median
Household Income
$26,202 $33,260 +26.9% $36,428 +9.5% +39.0%
White Median
Household Income
$37,113 $55,249 +48.9% $72,085 +30.5% +94.2%
% in Poverty 22.9% 21.3% -1.6% 22.4 +1.1% -0.5%
Homeowners 129,079 118,655 -8.1% 113,558 -4.3% -12.0%
Renters 128,117 119,737 -6.5% 126,233 +5.4% -1.5%
Homeownership
Rate
50.2% 49.2% 47.4%
Median Sales Price
(2)
$60,000 $91,000 +51.7% $106,000 +16.5 +76.7%
Sales Volume 10,211 6,647 -34.9% 10,433 +57.0% +2.2%
Median Monthly
Gross Rent
$498 $859 +72.5% 1,009 +17.5% +102.6%
Average Annual
Change
+$36 +5.6% +$21 +2.3%
SOURCE: 2000 Decennial Census, 2006-2010 and 2013-2017 American Community Survey, CoreLogic (sales price and volume)
(1) Breakdown of population by race/ethnicity does not include other racial groups and people indicating two or more racial
group membership.
(2) Median of sales by census tract
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 8
Sales prices in Baltimore, after collapsing with the foreclosure crisis and the Great Recession, have been slowly
recovering, but in most parts of the city, they are still well below national levels
8
or their 2007 peak. Since 2000,
rents have increased more rapidly than sales prices, and are now slightly above the national median rent. One
byproduct of this is that the percentage of Baltimore renters who are cost-burdened (spending over 30% of
their income for rent) went from 43% to 53% between 2000 and 2017. The most rapid rent rise and growth in
cost-burdened households took place between 2000 and 2010, and both have been largely stable since then.
Despite rising rents and aordable sales prices, Baltimore is losing homeowners. Since 2000, the city has lost
15,000 homeowners, and the homeownership rate has dropped from 50% to 47%, falling below 50% for the
rst time since 1930.
This short description makes clear that there is not one Baltimore. Baltimore is a large city, within which
many inconsistent, even conicting trends exist side by side. As a result, its neighborhoods are moving in
many dierent directions: some upward, some downward, and some staying much the same. This is to be
expected, but it reects an important overarching point about neighborhood change that is often overlooked.
Change is the norm. The majority of urban neighborhoods are engaged in an ongoing process of change.
Change can be upward, downward, or back again, driven by a complex mix of local, regional, and national
economic, demographic, and social trends and consumer preferences.
9
There is nothing inevitable either about
gentrication or neighborhood decline.
Table 4 shows how the distribution of the city’s neighborhoods has shifted since 2000 by race and economic
level. Reecting the national trends both with respect to a diminishing middle class and an increasing process
of “economic sorting” by income group, the number of neighborhoods in the middle in Baltimore has declined,
while the number of those at either end has risen. The number of upper-income areas has nearly tripled. This
shift can be seen vividly in Figure 2.
Figure 2: Change in Number of Baltimore Neighborhoods by Economic Level 2000 TO 2017
Low Moderate Middle Upper Middle Upper
15
10
5
0
-5
-10
-15
-20
8 The median price for existing homes in the United States during 2017 was approximately $246,000, or more than double the median price in
Baltimore.
9 For a comprehensive discussion of the dynamic processes of neighborhood change, and the factors involved, see my paper What Drives Neigh-
borhood Trajectories in Legacy Cities? Understanding the Dynamics of Change (2015), published by the Lincoln Institute of Land Policy, and
available at https://www.lincolninst.edu/publications/working-papers/what-drives-neighborhood-trajectories-legacy-cities.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 9
Table 4: Change in Number of Neighborhoods Over Time by Race and Economic Level
ALL NEIGHBORHOODS
2000 2010 2017
Low Income 22 28 32
Moderate Income 88 84 84
Middle Income 72 59 55
Upper-Middle Income 12 17 10
Upper Income 6 12 17
TOTAL
200 200 198
PREDOMINANTLY BLACK NEIGHBORHOODS
2000 2010 2017
Low Income 20 22 26
Moderate Income 59 63 62
Middle Income 31 27 16
Upper-Middle Income 2 1 0
Upper Income 0 0 0
TOTAL
112 113 104
PREDOMINANTLY WHITE NEIGHBORHOODS
2000 2010 2017
Low Income 0 1 0
Moderate Income 11 9 9
Middle Income 29 17 13
Upper-Middle Income 5 9 8
Upper Income 6 12 17
TOTAL
51 48 47
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 10
MIXED NEIGHBORHOODS
2000 2010 2017
Low Income 2 5 6
Moderate Income 18 12 13
Middle Income 12 15 26
Upper-Middle Income 5 7 2
Upper Income 0 0 0
TOTAL
37 39 47
When one looks at how neighborhoods have shifted by both economic level and race, however, a startling
contrast appears. Predominantly white neighborhoods tend to move upward in their trajectories, while
predominantly black neighborhoods tend to move downward. In 2000, there were 31 predominantly black
middle-income census tracts in Baltimore, or not quite 1 out of every 6 tracts. By 2017, there were only 16, or
half as many. As noted later, these neighborhoods did not gentrify, and many are in decline.
While the number of predominantly white middle-income census tracts also went down, the ones that changed
mostly moved upward. Indeed, the great majority of Baltimore’s gentrifying neighborhoods come from the
ranks of largely white formerly moderate- and middle-income neighborhoods. Indeed, a close look at the shift
in the distribution of predominantly white census tracts shows dramatic change: In 2000, only 1 out of 5 of
these tracts was upper-middle or upper income, but by 2017, over half fell into those categories.
This shift reects a major change in the makeup of Baltimore’s white population. While that population
was historically distributed fairly evenly across the full income spectrum from rich to poor, it is increasingly
becoming one of auent households. At the same time, the income distribution of the city’s black population
is moving in the opposite direction, reecting the out-migration of much of the city’s black middle class. Both
are shown in Figure 3.
Figure 3: Citywide Change in Income Distribution of Households by Race 2000 TO 2017
Low Moderate Middle Upper Middle Upper
50%
40%
30%
20%
10%
0%
-10%
-20%
-30%
Black
White
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 11
The only economic segment of the city’s black population that is growing (either in share or numbers) is the
low-income population, while the only segment of Baltimore’s white population that is growing is the upper-
income population, reecting the dierential in- and out-migration taking place. Relative to their share of the
existing population, white in-migration to Baltimore is signicantly greater than black in-migration,
10
and as I
discuss later, white in-migration is disproportionately made up of well-educated (bachelor’s or higher degree)
and high-earning people in their 20s and 30s, who are clustering in a small number of areas in the city.
As I discuss in a later part of this report, a large part of the loss of white low-income households in Baltimore
is associated with gentrication. Between 2000 and 2017, the city lost 8,200 low-income white households,
of which 3,800 were in gentrifying census tracts. While some of this loss may potentially be considered
displacement, with these households moving elsewhere,
11
a signicant part of the loss may also be
attributable to mortality; over the same period, the number of white Baltimoreans over 65 dropped by
10,000, or nearly 30%.
B. Neighborhood trajectories
Given these citywide trends, what does it mean for individual neighborhoods? For an initial answer, I asked a
simple question: For each neighborhood in each of the 11 categories in 2000, where did they stand, in terms
of both race and economic level, in 2010 and 2017? Once I had the data, in order to present answers to that
question in a visually meaningful form, I color-coded each neighborhood based on its 2000 economic level and
racial distribution, and where it stood in both respects in 2010 and 2017.
Table 5: Economic Trajectories of Middle-Income Neighborhoods by Racial Category
NUMBER YEAR UP DOWN ALL CHANGE SAME
PREDOMINANTLY BLACK
31
2010 0 12 12 19
2017 0 18 18 13
MIXED
12
2010 2 1 3 9
2017 0 1 1 11
PREDOMINANTLY WHITE
28
2010 12 3 15 13
2017 15 4 19 9
10 For the period 2013 through 2017, the average annual number of white in-migrants was equal to 10% of the underlying white population base,
while the average annual number of black in-migrants was equal only to 4% of the underlying black population base. Average annual Latinx in-mi-
gration was nearly 12% of their population base. Latinxs were the only group with a net positive migration balance, although the white net loss
was far smaller than the black net loss.
11 Between 2000 and 2017, although the white population of Baltimore County dropped by over 50,000, the number of white residents over 65
increased by 3,200, suggesting the possibility of some migratory eects.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 12
Table 6 and Table 7 show the trajectories for predominantly white and predominantly black middle-income
neighborhoods (with median incomes between 100% and 150% of the citywide median income in 2000). The
color-coding key appears at the bottom of Table 6. By scanning across the three sets of columns, one can
quickly get a sense of the extent and direction of change in the cluster of census tracts from 2000 to 2010, and
from 2010 to 2017. Table 8 summarizes the data from the two tables below and also shows the trajectories for
the smaller number of racially mixed middle-income neighborhoods.
Table 6: Trajectories of Predominantly White Middle-Income Neighborhoods
by Black Population Share and Economic Level
2000
2010 2017
CATEGORY
% BLACK CATEGORY % BLACK CATEGORY % BLACK
24510010100 4% 24510010100 4% 24510010100 7%
24510010200 18% 24510010200 11% 24510010200 5%
24510010300 1% 24510010300 5% 24510010300 2%
24510010400 3% 24510010400 5% 24510010400 2%
24510010500 7% 24510010500 6% 24510010500 4%
24510020100 14% 24510020100 11% 24510020100 7%
24510030200 21% 24510030200 35% 24510030200 31%
24510120100 10% 24510120100 8% 24510120100 9%
24510130600 3% 24510130600 4% 24510130600 7%
24510130700 6% 24510130700 9% 24510130700 7%
24510130803 11% 24510130803 14% 24510130803 10%
24510130806 18% 24510130806 26% 24510130806 19%
24510230200 3% 24510230200 5% 24510230200 1%
24510230300 2% 24510230300 7% 24510230300 4%
24510240100 0% 24510240100 2% 24510240100 3%
24510240400 2% 24510240400 3% 24510240400 1%
24510250206 3% 24510250206 14% 24510250206 13%
24510250401 11% 24510250401 26% 24510250401 29%
24510260404 26% 24510260404 31% 24510260404 31%
24510260605 14% 24510260605 19% 24510260605 17%
24510260900 3% 24510260900 5% 24510260900 8%
24510270402 25% 24510270402 53% 24510270402 58%
24510270501 13% 24510270501 32% 24510270501 42%
24510270502 22% 24510270502 51% 24510270502 58%
24510270703 19% 24510270703 36% 24510270703 36%
24510271101 20% 24510271101 36% 24510271101 34%
24510272004 10% 24510272004 15% 24510272004 15%
24510272005 10% 24510272005 16% 24510272005 10%
0-29% African American Very Low Income (0-59% citywide median)
30-69% African American Low Income (60-99% citywide median)
70-100% African American Middle Income (100-149% citywide median)
Upper-Middle Income (150-199% citywide median)
Upper Income (200%+ citywide median)
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 13
The majority of Baltimore middle-income neighborhoods in 2000 were no longer middle-income
neighborhoods by 2017 (39 out of 71). But the trajectories of predominantly white and predominantly black
neighborhoods were very dierent. The great majority of predominantly white neighborhoods that changed
moved upward economically, while all of the predominantly black neighborhoods that changed moved
downward economically. Mixed neighborhoods showed much less change. By 2017, nine of the predominantly
white middle-income neighborhoods had become upper-income (200% or more of the city median income),
and another six had become upper-middle income. These are the neighborhoods in which the great majority
of Baltimore’s gentrication has taken place. The geographic distribution of predominantly black and
predominantly white middle neighborhoods, and their trajectories from 2000 to 2017, are shown in Map 1.
Table 7: Trajectories of Predominantly Black Middle-Income Neighborhoods
by Black Population Share and Economic Level
2000 2010 2017
24510080102 94% 24510080102 97% 24510080102 98%
24510090100 89% 24510090100 88% 24510090100 84%
24510090300 79% 24510090300 77% 24510090300 68%
24510090600 98% 24510090600 97% 24510090600 89%
24510130805 76% 24510130805 78% 24510130805 86%
24510150701 98% 24510150701 98% 24510150701 92%
24510150702 98% 24510150702 98% 24510150702 96%
24510150900 97% 24510150900 97% 24510150900 95%
24510151100 100% 24510151100 98% 24510151100 99%
24510160801 100% 24510160801 99% 24510160801 97%
24510160802 99% 24510160802 98% 24510160802 97%
24510200702 97% 24510200702 97% 24510200702 97%
24510250101 85% 24510250101 90% 24510250101 81%
24510250102 84% 24510250102 89% 24510250102 96%
24510260203 90% 24510260203 90% 24510260203 95%
24510260301 80% 24510260301 92% 24510260301 89%
24510260302 81% 24510260302 92% 24510260302 91%
24510260403 97% 24510260403 94% 24510260403 92%
24510270802 86% 24510270802 94% 24510270802 91%
24510270803 82% 24510270803 89% 24510270803 89%
24510270805 84% 24510270805 87% 24510270805 84%
24510270901 94% 24510270901 97% 24510270901 96%
24510270902 92% 24510270902 93% 24510270902 91%
24510270903 90% 24510270903 89% 24510270903 74%
24510271002 97% 24510271002 96% 24510271002 95%
24510271900 72% 24510271900 76% 24510271900 73%
24510280101 89% 24510280101 92% 24510280101 88%
24510280102 97% 24510280102 97% 24510280102 97%
24510280200 96% 24510280200 95% 24510280200 91%
24510280401 82% 24510280401 85% 24510280401 79%
24510280402 93% 24510280402 98% 24510280402 97%
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 14
Table 8: Economic Trajectories of All Other Neighborhoods by Racial Category
NUMBER YEAR UP DOWN ALL CHANGE SAME
LOW PREDOMINANTLY BLACK
19
2010 6 (note 1) NA (note 2) 6 13
2017 6 (note 1) NA (note 2) 6 13
MODERATE PREDOMINANTLY BLACK
59
2010 4 8 12 46
2017 3 14 17 42
MODERATE MIXED
18
2010 6 2 8 10
2017 6 3 9 9
MODERATE PREDOMINANTLY WHITE
11
2010 5 0 5 6
2017 5 0 5 6
UPPER-MIDDLE PREDOMINANTLY BLACK
3
2010 0 1 1 2
2017 0 3 3 0
UPPER-MIDDLE MIXED
4
2010 0 2 2 2
2017 0 4 4 0
UPPER-MIDDLE PREDOMINANTLY WHITE
5
2010 2 1 3 2
2017 2 1 3 2
UPPER PREDOMINANTLY WHITE
6
2010 NA (note 2) 0 0 6
2017 NA (note 2) 1 1 5
NOTES TO TABLE 8
(1) This “upward” movement, with the exception of one census tract, represents movement from low to moderate, which is not
necessarily a meaningful change in neighborhood conditions, but is likely to represent little more than the statistical phenomenon
known as regression to the mean. One tract did, however, move from low to middle over the study period; this census tract, which
roughly corresponds to the Greenmount West neighborhood, is unique in that respect, and will be discussed further later.
(2) Because the low-income and upper-income categories occupy the bottom and top of the category scale, and because the
methodology used to dene upward and downward movement is movement between categories, no downward movement for the
former, or upward movement for the latter, can take place.
During the same period, the black population living in what had been predominantly white middle-income
neighborhoods increased by 6,300, or roughly 72%, as seven of the census tracts in this category moved from
being predominantly white to racially mixed. Most of these tracts were in the northeastern part of the city, an
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 15
area to which large numbers of African American families have moved since 2000. By comparison, only one of
the 31 predominantly black middle-income census tracts saw any change in its racial distribution.
Table 8 shows the same information for the other neighborhood categories. The pattern that can be seen in
the middle-income neighborhoods generally holds true across economic levels. The majority of predominantly
black low- and moderate-income census tracts showed little or no change, but far more moderate-income
tracts moved downward, falling below 60% of the citywide median income between 2000 and 2017, than
moved upward. By contrast, nearly half of the predominantly white moderate-income tracts moved upward.
Similarly, all three of the largely black upper-middle income tracts in 2000 had moved downward by 2017.
Figure 4 summarizes and compares the overall neighborhood trend pattern for predominantly white and
predominantly black moderate, middle, and upper-middle neighborhoods, which could go either up or down,
from 2000 to 2017.
MAP 1: TRAJECTORIES OF MIDDLE NEIGHBORHOODS FROM 2000 TO 2017
Predominantly Black Neighborhoods Predominantly White Neighborhoods
272004
272003
271900
271801
271700
271501
130805
271600
151300
151200
151100
151000
280200
280302
150900
150800
150701
150702
150500
130400
130600
120700
130100
130200
090800
080101
270101
260303
260302
080102
080200
080302
260301
260202
260201
270102
260102
260203
260402
260403
130300
150400
150600
160802
280402
160700
280401
280403
280404
250101
250102
250103
200800
200702
200600
250206
210200
200500
200400
200200
190100
180100
180300
200100
180200
020200
030100
030200
020100
010500
261100
261000
080301
080400
070300
070200
070100
060200
060100
060300
010100
010200
190300
200300
040200
160600
160500
150300
150200
150100
140100
140200
280500
110100
120600
120300
120201
170200
170100
170300
160100
160200
160300
160400
140300
120500
090900
080600
080500
090700
090200
270903
090600
090500
090300
090100
090400
080700
080800
070400
040100
100100
100200
100300
120400
100200
250303
250301
250205
250207
250203
250204
250401
250402
250600
240100
010400
010300
060400
240400
230300
210100
220100
240200
020300
240300
230200
230100
260605
260604
260501
260700
260404
260401
260900
260800
250500
200701
160801
280301
271503
271400
271300
271200
271101
271102
271001
271002
270805
270804
270801
270802
270901
270902
270803
270600
270301
270302
270402
270502
270501
270703
270702
270701
270401
260101
270200
120100
120202
130700
130803
130806
130804
271802
272007
272006
280101
280102
272005
272004
272003
271900
271801
271700
271501
130805
271600
151300
151200
151100
151000
280200
280302
150900
150800
150701
150702
150500
130400
130600
120700
130100
130200
090800
080101
270101
260303
260302
080102
080200
080302
260301
260202
260201
270102
260102
260203
260402
260403
130300
150400
150600
160802
280402
160700
280401
280403
280404
250101
250102
250103
200800
200702
200600
250206
210200
200500
200400
200200
190100
180100
180300
200100
180200
020200
030100
030200
020100
010500
261100
261000
080301
080400
070300
070200
070100
060200
060100
060300
010100
010200
190300
200300
040200
160600
160500
150300
150200
150100
140100
140200
280500
110100
120600
120300
120201
170200
170100
170300
160100
160200
160300
160400
140300
120500
090900
080600
080500
090700
090200
270903
090600
090500
090300
090100
090400
080700
080800
070400
040100
100100
100200
100300
120400
100200
250303
250301
250205
250207
250203
250204
250401
250402
250600
240100
010400
010300
060400
240400
230300
210100
220100
240200
020300
240300
230200
230100
260605
260604
260501
260700
260404
260401
260900
260800
250500
200701
160801
280301
271503
271400
271300
271200
271101
271102
271001
271002
270805
270804
270801
270802
270901
270902
270803
270600
270301
270302
270402
270502
270501
270703
270702
270701
270401
260101
270200
120100
120202
130700
130803
130806
130804
271802
272007
272006
280101
280102
272005
Remained middle-income neighborhoods in 2017
Moved downward to low or moderate income
Moved upward to upper-middle or upper income
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 16
Figure 4: Summary of Neighborhood Trajectories From 2000 to 2017 by Race
To summarize, neighborhood change is a fundamental reality in Baltimore. Large numbers of neighborhoods
are moving both upward and downward economically, a few moved in one direction from 2000 to 2010,
and then reversed direction after 2010. Within the three middle-income categories—which represent the
great majority of Baltimore neighborhoods—nearly half changed category, either up or down, between 2000
and 2017.
In the next section, I will explore the signicance of these changes, and how they aect such factors as
population change, house values, and homeownership rates.
II. THE DIMENSIONS OF NEIGHBORHOOD CHANGE IN BALTIMORE
Economic change has powerful implications for other dimensions of neighborhood change. In this section, I
will explore how change aects two critical dimensions of Baltimore’s neighborhoods:
• Demographic change, including gains, losses, and population shifts by race, ethnicity,
and income; and
• Housing market change, including sales prices and sales volumes, and changes in
homeownership rates.
A. Demographic change
In this section, I will look at population shifts by race and by income, for the city’s African American, white, and
Latinx populations.
1. Black population change
Understanding trends aecting Baltimore’s African American population is particularly important for a number
of reasons. First, it is by far the city’s largest racial or ethnic group. Second, as I have shown, Baltimore’s
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 17
African American population has beneted proportionately less from the city’s strong but uneven revival
than Baltimore’s white population, exacerbating the racial wealth gap. Third, despite progress over the years,
African Americans face racialized disadvantage owing to structural racial and socioeconomic segregation, and
they are still more likely to experience discrimination and uneven treatment than others. Fourth, at least some
documents have treated population loss and gentrication-driven displacement as being eectively one and
the same, an assumption that needs to be critically examined.
As shown in Table 3, despite a natural increase (excess of births over deaths) of more than 20,000,
12
Baltimore’s
black population declined by nearly 20,000 from 2000 to 2010, and by an additional 10,000 from 2010 to 2017,
an overall decline of 7% since 2000. While I may not be able to answer the question of “why” with any precision,
the relationship between neighborhood trends and population change may suggest some answers.
Table 9: Change in Black Population 2000 to 2017 by 2000 Neighborhood Type
NEIGHBORHOOD
TYPE IN 2000
PREDOMINANTLY
BLACK
MIXED
PREDOMINANTLY
WHITE
NUMBER % NUMBER % NUMBER %
Low Income -10,716 -22.7%
Moderate Income -34,408 -18.8% -1,781 -7.1% +2,273 +52.8%
Middle Income -231 -0.2% +5,055 +21.7% +6,301 +72.4%
Upper-Middle
Income
+195 +3.5% +1,378 +20.9% -293 -21.4%
Upper Income +1,372 +84.7%
TOTAL CHANGE
45,160 -13.1% +4,652 +8.5% +9,653 +60.3%
Table 9 shows the gain or loss in black population by 2000 neighborhood type. Neighborhoods that were low-
income, predominantly black neighborhoods in 2000 lost 10,716 black residents, or nearly 1 out of 4 of those
living in those neighborhoods.
With minor exceptions, black population decline in Baltimore is the result of black households moving out of
predominantly black low- and moderate-income neighborhoods. That does not mean that there are no other
census tracts that saw declines in black population. There are a small number of such tracts, including some
that are gentrifying. But taken as a whole, movement out of predominantly black low- and moderate-income
neighborhoods unrelated to gentrication is driving black population decline in Baltimore. As I noted earlier,
this loss is disproportionately made up of middle- and upper-income households.
The table also shows a strong pattern of black households moving into neighborhoods that were either racially
mixed or predominantly white in 2000. These areas saw an increase of more than 14,000 in black residents
from 2000 to 2017. That inow, however, oset less than one-third of the outow from predominantly black
12 The positive natural increase balance is shrinking, however, as the number of births to black parents in Baltimore has dropped sharply since 2000
from 7,034 to 4,828 in 2017. This is a decline of 31%, far greater than the proportional decline in the overall population.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 18
low- and moderate-income neighborhoods. The others can be assumed to have left Baltimore City.
13
While
their destinations are not known, it is notable that over the same period, the black population of Baltimore
County grew by nearly 82,000 and that of Anne Arundel County by 26,000, growth that was likely to have been
fueled in part by out-migration from Baltimore City.
Map 2: Spatial Distribution of Black Population Gains and Losses 2000 to 2017
272004
272003
271900
271801
271700
271501
130805
271600
151300
151200
151100
151000
280200
280302
150900
150800
150701
150702
150500
130400
130600
120700
130100
130200
090800
080101
270101
260303
260302
080102
080200
080302
260301
260202
260201
270102
260102
260203
260402
260403
130300
150400
150600
160802
280402
160700
280401
280403
280404
250101
250102
250103
200800
200702
200600
250206
210200
200500
200400
200200
190100
180100
180300
200100
180200
020200
030100
030200
020100
010500
261100
261000
080301
080400
070300
070200
070100
060200
060100
060300
010100
010200
190300
200300
040200
160600
160500
150300
150200
150100
140100
140200
280500
110100
120600
120300
120201
170200
170100
170300
160100
160200
160300
160400
140300
120500
090900
080600
080500
090700
090200
270903
090600
090500
090300
090100
090400
080700
080800
070400
040100
100100
100200
100300
120400
100200
250303
250301
250205
250207
250203
250204
250401
250402
250600
240100
010400
010300
060400
240400
230300
210100
220100
240200
020300
240300
230200
230100
260605
260604
260501
260700
260404
260401
260900
260800
250500
200701
160801
280301
271503
271400
271300
271200
Northeast Triangle
271101
271102
271001
271002
270805
270804
270801
270802
270901
270902
270803
270600
270301
270302
270402
270502
270501
270703
270702
270701
270401
260101
270200
120100
120202
130700
130803
130806
130804
271802
272007
272006
280101
280102
272005
Loss of 1,500 or more
Loss of 1,000 to 1,499
Loss of 500 to 999
Gain of 1,500 or more
Gain of 1,000 to 1,499
Gain of 500 to 999
These trends have led to a dramatic shift in the spatial distribution of the city’s black population, as shown
in Map 2. Most tracts in East and West Baltimore are losing population, while most of the gain in black
population is taking place in an area that I call the “Northeast Triangle” shown on the map, including Loch
Raven, Overlea, Glenham-Belhar, Cedonia, and Frankford. That area saw its black population grow by 17,500
from 2000 to 2017.
13 The net out-migration was actually substantially larger than the reported population loss, since Baltimore’s black population maintained a positive
although gradually shrinking birth/death ratio throughout the period, resulting in a natural increase in the city’s black population since 2000 of
20,000 to 25,000. Notably, however, the number of births to black parents in Baltimore has dropped sharply since 2000 from 7,034 to 4,828 in
2017, a decline of 31%, far greater than the proportional decline in the overall population.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 19
The broad trend in Baltimore’s black population is the movement out of low- and moderate-income
neighborhoods, largely in East and West Baltimore, and movement toward the Northeast Triangle, but even
more so, outside Baltimore City entirely.
2. White and Latinx population change
As noted earlier, both black and white populations in Baltimore have declined by roughly 30,000 each since
2000. Since the city’s white population is much smaller than its black population, however, the proportionate
decline has been far greater, 15% compared to 7%. In contrast to the population decline in the city’s black
population, which is concentrated in predominantly black low- and moderate-income neighborhoods, white
population decline is widely distributed across most neighborhood types, except for small increases in
predominantly white upper-middle and upper-income tracts, as shown in Table 10. The increase in white
population in predominantly black moderate-income tracts is largely attributable to white population growth in
one census tract containing part of the Charles Village and Harwood NSAs near the Johns Hopkins campus.
Table 10: Change in White Population 2000 to 2017 by 2000 Neighborhood Type
NEIGHBORHOOD
TYPE IN 2000
PREDOMINANTLY
BLACK
MIXED
PREDOMINANTLY
WHITE
NUMBER % NUMBER % NUMBER %
Low Income -67 -2.3%
Moderate Income +475 +5.3% -4,231 -15.5% -7,667 -27.2%
Middle Income -1,746 -17.6% -6,921 -35.4% -7,947 -4.7%
Upper-Middle
Income
-661 -39.2% -2,616 -24.6% +305 +3.4%
Upper Income +387 +3.3%
TOTAL CHANGE
-1,999 -13,868 -14,922
Census tracts where the white population declined by more than 500 people greatly outnumbered those that
gained by a similar amount, as shown in Map 3. The greatest white population loss was in South Baltimore and
in the Northeast Triangle, while a handful of tracts showed signicant gain, principally downtown and around
the Inner Harbor.
As noted earlier, Baltimore’s white population has stabilized in recent years as a result of strong white
in-migration, visible both in census data and in homebuying activity. Ongoing replacement of the city’s
white population appears to be taking place, as long-time residents, many moderate- and middle-income,
leave or pass away, and are replaced by generally younger and more auent new arrivals, who are largely
concentrated in a smaller part of the city.
Baltimore’s Latinx population is the fastest growing segment of the city’s population, although only 5% of the
total. The area of greatest Latinx concentration is in Highlandtown and eastward, where they make up 30% to
45% of the total population of six census tracts. Overall, however, the city’s Latinx population is fairly dispersed;
only 25% of the city’s Latinx population lives in that area of Latinx concentration.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 20
Map 3: Spatial Distribution of White Population Gains and Losses 2000 to 2017
272004
272003
271900
271801
271700
271501
130805
271600
151300
151200
151100
151000
280200
280302
150900
150800
150701
150702
150500
130400
130600
120700
130100
130200
090800
080101
270101
260303
260302
080102
080200
080302
260301
260202
260201
270102
260102
260203
260402
260403
130300
150400
150600
160802
280402
160700
280401
280403
280404
250101
250102
250103
200800
200702
200600
250206
210200
200500
200400
200200
190100
180100
180300
200100
180200
020200
030100
030200
020100
010500
261100
261000
080301
080400
070300
070200
070100
060200
060100
060300
010100
010200
190300
190200
200300
040200
160600
160500
150300
150200
150100
140100
140200
280500
110100
120600
120300
120201
170200
170100
170300
160100
160200
160300
160400
140300
120500
090900
080600
080500
090700
090200
270903
090600
090500
090300
090100
090400
080700
080800
070400
040100
100100
100200
100300
120400
100200
250303
250301
250205
250207
250203
250204
250401
250402
250600
240100
010400
010300
060400
240400
230300
210100
220100
240200
020300
240300
230200
230100
260605
260604
260501
260700
260404
260401
260900
260800
250500
200701
160801
280301
271503
271400
271300
271200
Northeast Triangle
271101
271102
271001
271002
270805
270804
270801
270802
270901
270902
270803
270600
270301
270302
270402
270502
270501
270703
270702
270701
270401
260101
270200
120100
120202
130700
130803
130806
130804
271802
272007
272006
280101
280102
272005
B. Housing market change
Up to this point, I have focused on the population of Baltimore’s neighborhoods: how they are distributed
by income and race, and how populations have shifted from one part of the city to others since 2000. In this
section, I explore changes in Baltimore’s neighborhood housing markets—the dynamics of buying, selling,
and renting homes and apartments in the city’s neighborhoods. Before digging into the numbers, a brief
discussion of why housing markets—particularly the patterns of homebuying and selling—are so important to
understanding the dynamics of Baltimore’s neighborhoods may be useful.
While the housing market is far from the only thing that determines whether or not a neighborhood is a vital,
thriving community, it powerfully aects neighborhood outcomes. The demand for housing in a neighborhood
reects the extent to which people choose to live there rather than elsewhere, given their means and their
locational preferences. When people choose to buy a home in a particular neighborhood, they are making
a longer-term commitment to that neighborhood that often leads to behaviors that enhance neighborhood
vitality. Conversely, if people only live in a neighborhood because they lack other locational choices, and leave
if they can, their behavior is likely to reect that perspective and the neighborhood is likely to suer as a result.
Housing markets are a critical underpinning of neighborhood strength and vitality.
Loss of 1,000 or mor
e
Loss of 500 to 999
Gain of 1,000 or mor
e
Gain of 500 to 999
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 21
Where market demand is weak, prices are low and sales are few. Houses sit empty for a long time, and
those that sell are more likely to attract investors than owner-occupant buyers. Homeowners make fewer
improvements because they are unlikely to get their money back if they sell, while property owners are more
likely to fall behind on mortgage or property tax payments and let their houses go into tax sales or mortgage
foreclosure. All of those forces, in turn, often lead to houses eventually becoming abandoned, and in many
cases, economically unfeasible to rehabilitate and restore to use. Conversely, too rapid growth in demand
and prices can destabilize a neighborhood, encouraging speculation and undermining neighborhood stability
and cohesion.
This study is not a neighborhood-level market analysis of Baltimore. Such an analysis already exists, as for a
number of years, Baltimore City has commissioned regular analyses of small-area market conditions from the
Philadelphia-based Reinvestment Fund, and used those analyses to create neighborhood market typologies.
14
That information has been used to help design a number of city strategies, including the Vacants to Value
program. The purpose of this report is to look at the market dimensions of neighborhood change, and to
relate them to the shifts in household incomes and population movements described earlier.
To measure market strength and weakness, I look at three factors:
• Sales price
The price at which houses sell is the single most powerful measure of market strength or weakness.
This is particularly true in Baltimore, where most neighborhoods are dominated by row houses and
where many dierent neighborhoods contain houses of largely similar size, vintage, and construction.
A three- story row house can sell for over $500,000 in Bolton Hill, and a physically all-but-identical
15
one
may sell for less than one-tenth that price less than a mile to the west.
• Sales volume
For a housing market to be healthy, there have to be enough buyers to absorb the supply. If there are
too few buyers, properties may sit empty and ultimately be abandoned. Conversely, too many buyers
can overheat the market, or be a sign of speculation and ipping.
• Percentage of investor buyers
A high share of investor buyers in a neighborhood made up largely of single-family homes is a warning
sign. Not only is a reasonably high level of owner-occupants important for a stable neighborhood,
but the absence of owner-occupant buyers is also a sign that there are few people willing to make a
personal commitment to the neighborhood, as distinct from buyers who see the neighborhood purely
as a prot opportunity.
Finally, I look at change in the number of homeowners and the homeownership rate, which is an important
indicator of market conditions and neighborhood strength. Not only is there evidence that homeownership
may be an important factor in fostering neighborhood stability and community engagement, but
there is also evidence that declines—particularly if rapid—in homeownership can have a destabilizing
eect on neighborhoods.
16
14 These analyses are available at https://planning.baltimorecity.gov/maps-data/housing-market-typology.
15 Identical in the sense of having a similar architectural appearance, structural quality, square footage, and interior conguration. The Bolton Hill
row house is likely, however, to be in substantially better condition.
16 There is a substantial body of research on the impacts of homeownership. Much of the research is summarized in Lawrence Yun and Nadia Evan-
gelou, “Social Benets of Homeownership and Stable Housing,” published by the National Association of Realtors (2016) and available at https://
realtoru.edu/wp-content/uploads/2014/06/Homeownership-Stable-Housing.pdf. The evidence for family and behavior eects of homeowner-
ship is much stronger than the direct evidence of neighborhood eects, which to some extent must be inferred from the former. The eects of
declines in homeownership have been studied less, but one solid study is Chengri Ding and Gerrit-Jan Knapp, “Property Values in Inner-City Neigh-
borhoods: The Eects of Homeownership, Housing Investment and Economic Development,” Housing Policy Debate 13:4 (2003) 701–727. Clearly,
however, there is no generalizable “magic number” as to what a homeownership rate should be.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 22
As I will discuss later, these factors—particularly increases in sales price—are relevant to evaluating whether
and where gentrication may be taking place.
1. Real estate market dynamics
Baltimore shared in the housing bubble that consumed the United States from 2000 to 2007 and in the
subsequent bust. As Figure 5 shows, home sales prices in Baltimore more than doubled from 2000 to 2007,
going from $60,000 to $132,000, and then plummeted, falling to $75,000 by 2011. After at prices for a few
years, prices have started to recover, reaching a median of $106,000 in 2017.
17
As with other trends, the change
in prices was not experienced evenly across the city. Indeed, the most dramatic price phenomenon since 2000
has been the extraordinary variation in price change from one part of the city to another.
Figure 5: Citywide Median Sales Price by Year 2000 to 2017
Figure 6: Distribution of Sales Prices by Census Tract Relative to the Citywide Median
2000, 2010, and 2017
17 Preliminary 2018 data provided by CoreLogic shows a solid increase from the 2017 gures used in this report.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 23
What is striking about house prices in 2000 by census tract is how little price variation there was in Baltimore’s
housing market at the time. The lowest-priced census tract had a median that was more than half the citywide
median, and sales prices more than double the citywide median were found in only 15 tracts (7.5% of the total).
There were few tracts where the market was not functioning, and few upscale tracts where prices were high.
Prices in over 80% of the city’s census tracts fell in the relatively narrow range between 50% and 150% of the
citywide median price, as shown in Figure 6.
By 2010, this had already changed dramatically. The number of census tracts in the middle-income range had
dropped from 168 to 107, and the number continued to drop—to 84—from 2010 to 2017. More and more
tracts were at the bottom and the top of the home price range, and fewer and fewer were in the middle.
18
Another perspective on home sales price change comes by looking at the gain or loss in value in constant
dollars; that is, prices adjusted for ination.
19
From 2000 to 2017, the median sales price in Baltimore increased
by 24% in constant dollars, a respectable performance in light of the city’s boom-bust price cycle. Overall,
58% of the city’s tracts gained value, and 42% lost value. The change, however, was not evenly distributed.
Ordinarily, one would expect the gains and losses to be distributed along a bell-shaped curve,
20
with most of
the gains and losses clustered close to the middle. In actuality, as Figure 7 shows, the distribution was the
opposite. Few tracts changed only a little in house value. Far more gained a lot or lost a lot. Nearly a quarter of
the city’s tracts lost 30% or more in median house value in constant dollars, while over a quarter gained more
than 50% in value, and 1 out of 7 tracts saw their median house value in constant dollars more than double.
This is perhaps the single most vivid illustration of the economic polarization that has taken place in Baltimore
over the past two decades.
Figure 7: Distribution of Sales Price Change in Constant Dollars
2000 to 2017 (Number of Census Tracts)
18 For the statistically minded, the standard deviation of sales prices went from $42,417 in 2000 to $85,844 in 2010, and $103,056 in 2017.
19 For purposes of calculating ination, I used the change in the Consumer Price Index, which increased by 42.1% from June 2000 to
June 2017.
20 This distribution is so common that it is also referred to as a “normal distribution.”
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 24
This polarization relates closely to the racial composition of the neighborhood. As Figure 8 shows, 3 out of 5
predominantly white census tracts saw house prices increase by more than 50% in constant dollars, compared
to 2 out of 5 racially mixed, and only 1 out of 10 largely black tracts. Conversely, house prices declined by 20%
or more in nearly half of all predominantly black tracts, compared to less than 1 out of 5 mixed and 1 out of 12
predominantly white tracts. It is also strongly linked to the distribution of population losses and gains.
This has two powerful implications for the future of Baltimore’s predominantly black neighborhoods. First, it
has led to massive loss of wealth for many of the city’s black homeowners, for whom home equity typically
represents the greater part of their overall household wealth. Second, because stability and potential
appreciation in house value are an important consideration in homebuying decisions, it discourages
homebuyers, whether black or white, from buying homes in these neighborhoods. There is strong evidence,
with a handful of exceptions, that this is currently taking place.
Figure 8: Percentage Distribution of Sales Price Change from 2000 to 2017
by Racial Composition of Census Tract
The pattern is similar, but somewhat more complicated, for sales prices and sales volumes by both the
racial and income category of the census tract. Starting with the proposition that the average turnover in
an existing pool of residential properties typically runs 6% to 7% per year,
21
and allowing room for annual
uctuations, suggests that an annual volume of home sales in the range of 5% to 8% of the existing houses
in a given neighborhood (the “sales ratio”) can be considered the “Goldilocks” range—not too cold, and not
too hot. Ratios signicantly below that level are likely to lead to property deterioration, and in many cases
abandonment, as movers are unable to nd buyers or renters to replace them.
22
21 See, e.g., F.J.Fabozzi, The Handbook of Mortgage-Backed Securities, New York, NY: McGraw-Hill (2005); M. Piazzesi and M. Schneider, “Housing and
macroeconomics,” Handbook of Macroeconomics, 2, Elsevier, 2016: 1547-1640.
22 In a neighborhood where a signicant amount of new construction or substantial rehabilitation of houses for sale is taking place, the Goldilocks
range will be potentially signicantly higher, as the optimal number of buyers is the sum of those buying in the existing stock (i.e., 5% to 8% of
that stock) and the buyers of the new units coming on the market.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 25
Overall, the sales ratio for Baltimore in 2017 was 5.3%, meaning that residential sales in 2017 were equal to
5.3% of the city’s single-family housing inventory. While on the low side, it is within the Goldilocks range and
suggests that the housing market, on the whole, is functioning fairly well. As with other measures, however,
the citywide statistics mask considerable variation by neighborhood. Figure 9 shows the picture by census tract
racial and income category, while Table 11 shows detail for both sales prices and sales volumes for each of the
neighborhood categories.
Figure 9: Sales Ratios for Census Tracts by Racial and Income Category, 2017
The greatest market weakness is concentrated in predominantly black neighborhoods where the median
household income is at or below the citywide median—that is, low- and moderate-income neighborhoods.
Those neighborhoods have also seen signicant loss in house value in constant dollars. Black middle-
income neighborhoods are “hanging in” in terms of sales volumes, although barely, and seeing only nominal
appreciation in sales price. Adjusted for ination, house values in predominantly black middle-income
neighborhoods have been increasing at well under 1% per year since 2000, while predominantly white middle-
income neighborhoods saw an average increase of 4% per year over that period. Sales volumes in the city’s
mixed and predominantly white neighborhoods are consistently within the Goldilocks range, except for the ve
largely white upper-middle census tracts. On closer look, however, this reects unusually high volumes in two
areas: One is the Village of Cross Keys, and the second is to the west and south of the Inner Harbor, where a
great deal of new construction is taking place, pushing optimal sales volumes upward.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 26
Table 11: Sales Prices and Sales Volumes by Neighborhood Category
NUMBER 2000 2010 2017
2017 SALES
RATIO
CHANGE
2000-2017
CHANGE IN
CONSTANT $$
LOW PREDOMINANTLY BLACK
Median price $52,450 $51,250 $69,125 +33%
-7%
Number of sales 445
191 297 2.2%
MODERATE PREDOMINANTLY BLACK
Median price $49,950 $61,000 $50,625
+1% -29%
Number of sales 1,934 1,621 1,831 3.2%
MODERATE MIXED
Median price $53,950 $84,225 $94,759 +76% 24%
Number of sales 914 746 1,006 6.3%
MODERATE PREDOMINANTLY WHITE
Median price $56,000 $145,000 $148,600 +165% 87%
Number of sales 591 474 751 7.5%
MIDDLE PREDOMINANTLY BLACK
Median price $62,000 $97,000 $101,550 +64% 12%
Number of sales 1,621 1,066 1,826 4.7%
MIDDLE MIXED
Median price $78,750 $120,375 $137,436 +75% 23%
Number of sales 982 590 890 5.9%
MIDDLE PREDOMINANTLY WHITE
Median price $86,375 $187,250 $237,250 +175% 93%
Number of sales 2,301 1,550 2,517 9.3%
UPPER-MIDDLE PREDOMINANTLY BLACK
Median price $75,975 $130,200 $146,950 +93% 36%
Number of sales 124 78 149 5.6%
UPPER-MIDDLE MIXED
Median price $76,000 $145,000 $156,000 +105% 41%
Number of sales 310 179 324 6.4%
UPPER-MIDDLE PREDOMINANTLY WHITE
Median price $124,000 $206,606 $296,675 +139% 68%
Number of sales 408 266 375 11.8%
UPPER PREDOMINANTLY WHITE
Median price $201,500 $426,250 $424,298 +111% 48%
Number of sales 545 322 442 6.8%
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 27
Sales prices and sales volumes are powerfully correlated. As Table 12 shows, the median sales price in tracts
that had a sales ratio of 2% or less was barely $44,000, or roughly 40% of the citywide median; by contrast,
prices in tracts with sales ratios of 10% or more were $268,500, or roughly 2.5 times the citywide median,
which also reects the fact that these are the areas where most new construction is taking place.
Table 12: Relationship of Median Sales Price to Sales Ratio
<2% 2%-2.99% 3%-4.99% 5%-7.99% 8%-9.99% 10%+
Median
sales price
$44,081 $53,512 $71,500 $129,500 $184,700 $268,500
Turning to the third measure, the percentage of investor buyers in the market, the patterns are similar to those
described above, but with a key dierence. The long-term trend since 2000 reects the eects of the housing
bubble and bust, and the slow recovery. Figure 10 shows that the investor share peaked at over 40% in 2005,
remained generally over 30% through 2013, and has declined as the Baltimore housing market has recovered
since then.
23
Figure 10: Share of Investor Buyers in Baltimore Market 2000 to 2017
As Figure 11 shows, the share of investor buyers is higher in predominantly black census tracts. Roughly 1
out of 3 buyers in predominantly black census tracts is an investor buyer, compared to less than 1 out of 5 in
predominantly white tracts.
Investor buyers, however, still make up more than 40% of all buyers in 1 out of every 5 census tracts in the city,
and more than 50% in nearly 1 out of 10, as shown in Map 5. While this is a signicant improvement since the
2005 peak, when that was true of over half of the city’s tracts, it is still a serious concern. These tracts tend to
23 Preliminary 2018 data shows a further decline in the investor buyer share of the market.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 28
be areas with very low house sales prices. While some investors may be buying substandard or vacant houses
to rehabilitate and then sell or rent them, others may be buying houses at very low prices in order to “milk”
them—that is, rent them as is, make minimal if any repairs, perhaps not even pay property taxes, and plan to
walk away after a few years, having made a decent return from cash ow alone.
24
Figure 11: Percentage of Investor Buyers in 2017 by Tract Racial and Income Category
While milking, a predatory form of landlord behavior, is far from the norm in Baltimore, the likelihood of a
given building being treated in that fashion is a straightforward reection of the economics of owning and
operating rental housing, and is typically found only where rents are very high relative to house values—in
other words, locations where it is possible to make a return entirely from the cash ow from the dwelling
unit, with little concern for the ultimate value of the property. This is measured by a simple equation, the ratio
between the annual gross rent and the value of the property, which can be approximated by comparing the
annual gross rent for 2017 as reported in the American Community Survey with the median 2017 sales price
for homes in that census tract.
Using as a rule of thumb that tracts where the median house value is less than ve times the median annual
rent are at risk of milking, I found that 54, or slightly more than one-quarter of the census tracts in Baltimore,
meet that criterion, as shown in Map 5. Of these, the great majority are predominantly black census tracts,
with the majority of these being moderate-income census tracts. I am not suggesting that all or most of the
landlords in these areas are in fact predatory in their behavior, but that the market characteristics of those
areas place the areas at risk of drawing potentially predatory investors. It is notable, however, that there is
considerable overlap between Map 4 and Map 5, which shows those parts of the city where investors have the
largest share of the housing market.
24 I discuss investor strategies and the conditions under which milking is likely to take place in detail in my paper “Lessons from Las Vegas: Housing
markets, neighborhoods, and distressed single-family property investors,” Housing Policy Debate 24, no. 4 (2014): 769-801.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 29
Map 4: Areas with High Investor Buyer Share in 2017
272004
272003
271900
271801
271700
271501
130805
271600
151300
151200
151100
151000
280200
280302
150900
150800
150701
150702
150500
130400
130600
120700
130100
130200
090800
080101
270101
260303
260302
080102
080200
080302
260301
260202
260201
270102
260102
260203
260402
260403
130300
150400
150600
160802
280402
160700
280401
280403
280404
250101
250102
250103
200800
200702
200600
250206
210200
200500
200400
200200
190100
180100
180300
200100
180200
020200
030100
030200
020100
010500
261100
261000
080301
080400
070300
070200
070100
060200
060100
060300
010100
010200
190300
200300
040200
160600
160500
150300
150200
150100
140100
140200
280500
110100
120600
120300
120201
170200
170100
170300
160100
160200
160300
160400
140300
120500
090900
080600
080500
090700
090200
270903
090600
090500
090300
090100
090400
080700
080800
070400
040100
100100
100200
100300
120400
100200
250303
250301
250205
250207
250203
250204
250401
250402
250600
240100
010400
010300
060400
240400
230300
210100
220100
240200
020300
240300
230200
230100
260605
260604
260501
260700
260404
260401
260900
260800
250500
200701
160801
280301
271503
271400
271300
271200
271101
271102
271001
271002
270805
270804
270801
270802
270901
270902
270803
270600
270301
270302
270402
270502
270501
270703
270702
270701
270401
260101
270200
120100
120202
130700
130803
130806
130804
271802
272007
272006
280101
280102
272005
Map 5: Areas at Greater Risk of Predatory Milking Practices
272004
272003
271900
271801
271700
271501
130805
271600
151300
151200
151100
151000
280200
280302
150900
150800
150701
150702
150500
130400
130600
120700
130100
130200
090800
080101
270101
260303
260302
080102
080200
080302
260301
260202
260201
270102
260102
260203
260402
260403
130300
150400
150600
160802
280402
160700
280401
280403
280404
250101
250102
250103
200800
200702
200600
250206
210200
200500
200400
200200
190100
180100
180300
200100
180200
020200
030100
030200
020100
010500
261100
261000
080301
080400
070300
070200
070100
060200
060100
060300
010100
010200
190300
200300
040200
160600
160500
150300
150200
150100
140100
140200
280500
110100
120600
120300
120201
170200
170100
170300
160100
160200
160300
160400
140300
120500
090900
080600
080500
090700
090200
270903
090600
090500
090300
090100
090400
080700
080800
070400
040100
100100
100200
100300
120400
100200
250303
250301
250205
250207
250203
250204
250401
250402
250600
240100
010400
010300
060400
240400
230300
210100
220100
240200
020300
240300
230200
230100
260605
260604
260501
260700
260404
260401
260900
260800
250500
200701
160801
280301
271503
271400
271300
271200
271101
271102
271001
271002
270805
270804
270801
270802
270901
270902
270803
270600
270301
270302
270402
270502
270501
270703
270702
270701
270401
260101
270200
120100
120202
130700
130803
130806
130804
271802
272007
272006
280101
280102
272005
40% to 49.9% investor-buyer shar
e
50% or higher investor-buyer shar
e
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 30
2. Homeownership
As briey noted earlier, the homeownership rate has been dropping steadily in Baltimore in recent years,
falling below 50% for the rst time since 1930. Since 2000, the number of homeowners in the city has dropped
by roughly 15,500 (12%), potentially contributing to the destabilization of many city neighborhoods. The
number of homeowners has dropped much more than the number of renters, which also reects in part the
extent to which much of the new housing being created in Baltimore is rental, rather than owner-occupied
housing. This is true in high-density, upscale areas like Harbor East as well as in residential neighborhoods
where vacant houses are being rehabilitated. The great majority of homes rehabilitated through the city’s
Vacants to Value program are reused as rental rather than owner-occupied housing. This reects the
signicantly more favorable returns available from rental housing in many parts of the city.
25
It also reects a shortfall in homebuying in many parts of the city. To sustain homeownership, one must have a
steady ow of new homebuyers to replace those who move or pass away. Knowing that the average length of
stay of homeowners in Baltimore is 15 years, one can infer that 6%-7% will move or pass away each year, which
in turn dictates that the number of new homebuyers will be roughly equal to 6%-7% of the number of existing
owners. The actual number is substantially less than that, particularly with respect to black homeowners. In
2010, there were 64,242 black homeowners in Baltimore, but in 2018, there were only 1,873 black homebuyers,
or 2.9% of the number of owners.
26
Although a signicant improvement from the post-recession low point
of 2011, when only 715 mortgages to black homebuyers were made in Baltimore, it is far too few to sustain
current black homeownership rates. Those buyers, moreover, as I will discuss further in a later section, are
concentrating in a few areas, particularly the area I have called the Northeast Triangle.
The shortfall is not a function of lack of access to mortgages. Indeed, the dramatic increase in buyers since
2011 reects growing access to mortgages for large numbers of African American homebuyers.
27
At the same
time, the data also reect the growing movement toward the suburbs, as more mortgages to black buyers—
particularly relative to the existing number of black homeowners—are being made in Baltimore and Anne
Arundel counties. Over 3,000 black homebuyers received mortgages to buy homes in those two counties in
2018. In both of those counties, homebuying is close to or exceeds what might be considered the minimum
replacement rate to sustain or grow the existing homeowner pool.
As Map 6 shows, however, the loss of homeownership has been uneven. While some areas (particularly East
and West Baltimore) have lost 30% or more of their homeowners since 2000, others have remained relatively
stable and a few have even seen an increase, mainly in neighborhoods around the Inner Harbor
and downtown.
25 For a more detailed discussion of this point, see pages 35-37 of Tackling the Challenge of Blight in Baltimore: An Evaluation of Baltimore’s Vacants
to Value Program, prepared by the Center for Community Progress (2017), available at https://www.communityprogress.net/lebin/Baltimore_Va-
cant_to_Value_Report_Final.pdf.
26 This is the number of purchase mortgages made to black owner-occupant buyers, not the total number of buyers, which include cash buyers and
people nancing houses through means not reected in HMDA data such as seller nancing. Those numbers, however, are likely to be fairly small.
27 This does not mean that all race-based mortgage disabilities have been eliminated. They have been reduced, however, and growing numbers of
black homebuyers have been able to navigate the mortgage system. That said, a signicant constraint still exists that disproportionately aects
low property value areas with respect to appraisal procedures generally and lenders’ reluctance to make small mortgage loans, typically for
amounts below $50,000.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 31
Table 13: Black Homebuyers and Homeowners in Baltimore and Surrounding Counties
Number of
homebuyer
mortgages
in 2018
Number of
homeowners in 2010
Mortgages as % of
homeowners
Baltimore City
1,873 64,242 2.9%
Baltimore County
2,058 37,302 5.5%
Anne Arundel County
1,421 15,123 9.4%
% in Baltimore City
35% 55%
Map 6: Percentage Change in Number of Homeowners 2000 to 2017
Insucient Data
-30.00% or less
-29.99% – -15.00%
14.99% – 0.00%
0.01% – 14.99%
15.00% or more
SOURCE: Map created by PolicyMap
As Figure 12 shows, the greatest losses were in predominantly black low- and moderate-income areas, where
the number of homeowners dropped 20% and 25%, respectively. Largely black middle-income neighborhoods
also saw signicant losses; all in all, roughly 80% of the net loss in homeownership in Baltimore came from the
city’s predominantly black neighborhoods, much more than their share of the city’s homeowners.
This does not mean, however, that the loss of the city’s homeowners was mostly made up of black households;
indeed, the decline in the number of white households over the same period was greater. It highlights the
spatial shift among black households in Baltimore. As large numbers of black families have bought homes in
other parts of the city, particularly in the Northeast Triangle, their gains have oset losses in the number of
white homeowners in those areas. Map 7 shows the areas where the number of black homeowners increased
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 32
by 10% or more and the number of white homeowners declined by 10% or more from 2000 to 2017, with the
Northeast Triangle highlighted. In addition to that area, similar trends can be seen in the southwest part of the
city, in neighborhoods such as Violetville and Morrell Park.
Figure 12: Change in Number of Homeowners from 2000 to 2017 by Tract Category
Map 7: Areas with Both Black Homeownership Growth and White Homeownership Decline 2000 to 2017
SOURCE: Map created by PolicyMap
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 33
Table 14 provides a summary of the neighborhood data presented in this section of the report. A discussion
of relationships and correlations between the various data elements presented in this section appears as
Appendix 2.
Table 14: Summary Characteristics By Neighborhood Category
*Other vacant is a category used by the U.S. Census Bureau to denote those vacant properties that are neither oered for rent or
sale, held pending occupancy by tenants or buyers, or used for seasonal or temporary occupancy. While it is not limited to vacant
and abandoned properties, it can be seen as a rough surrogate for long-term vacant, abandoned properties.
III. KEY NEIGHBORHOOD CLUSTERS
In the preceding section, I looked at the trends in each of the dierent categories that make up Baltimore’s
neighborhoods, segmented by household income and race. In the course of that analysis, three dierent
neighborhood clusters stand out as representing the most signicant trends in neighborhood change in
Baltimore since 2000:
• Predominantly Black moderate-income neighborhoods. These neighborhoods, which make up 30% of
the city’s census tracts, account for the greatest part of the city’s black population loss, as well as the
sharpest declines in property values and homeownership in the city.
• The Northeast Triangle. These neighborhoods, which make up 11% of the city’s census tracts, were
mostly predominantly white in 2000. They are undergoing signicant change with an inux of black
homebuyers as well as an exodus of white households.
CATEGORY
Number
of Tracts
(2000)
Black
Population
Change
2000-2017
White
Population
Change
2000-2017
Sales
Price
Change
2000-
2017
Investor
Buyer
Change
2017
Homeowner
Change
2000-2017
Home
Ownership
Rate 2017
Poverty
Rate
2017
Other
Vacant*
Share
2017
Low
Income
Black 19 -10,716 -67 31.8% 27.1% -20.0% 20.8% 41.6% 16.9%
Moderate
Income
Black 59 -34,408 +475 1.4% 34.4% -24.8% 37.8% 29.1% 20.3%
Mixed 18 -1,781 -4,231 75.6% 26.1% -12.4% 29.4% 25.2% 13.1%
White 11 +2,273 -7,667 165.4% 20.7% -18.1% 47.8% 20.0% 6.4%
Middle
Income
Black 31 -231 -1,746 63.8% 31.0% -10.9% 58.9% 19.2% 8.1%
Mixed 12 +5,055 -6,921 74.5% 27.6% -8.1% 60.1% 14.3% 5.4%
White 28 +6,305 -7.947 174.7% 13.9% +0.4% 58.1% 17.2% 5.5%
Upper-
Middle
Income
Black 2 +195 -661 93.4% 16.4% -5.0% 78.0% 11.3% 4.9%
Mixed 5 +1,378 -2,616 105.0% 15.8% -3.3% 73.7% 12.4% 4.3%
White 5 -293 +305 139.3% 11.8% No 43.3% 10.4% 3.7%
Upper
Income
White 6 +1,372 +587 110.6% 7.2% +0.3% 71.5% 4.5% 4.8%
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 34
• Gentrifying neighborhoods. These neighborhoods, which are estimated to represent 14% of the
city’s census tracts, are clustered around the Inner Harbor, Johns Hopkins University, and downtown.
They constitute the most visible and highly publicized manifestation of neighborhood change in
Baltimore today.
These are not the only neighborhood clusters in Baltimore, although collectively they make up more than half
of the city’s neighborhoods. Each represents, however, a dierent and distinct challenge facing the city—the
challenge of decline, the challenge of maintaining stability, and the challenge of managing growth. This section
will drill down into the dynamics of each of these three neighborhood clusters.
A. Predominantly black moderate-income neighborhoods
This cluster is the largest single neighborhood cluster in Baltimore, making up both in 2000 and 2017 roughly
30% of the city’s census tracts, and roughly half of all predominantly African American tracts. In 2000, they
were not the city’s poorest areas, but nonetheless had median incomes between 60% and 100% of the citywide
median. As shown in Map 8, 42 (71%) of these tracts remained moderate income in 2017, 14 (24%) moved
downward into the low-income category, and three moved upward into the middle-income category. Four,
including the three that moved upward, moved from predominantly black to racially mixed; the overwhelming
majority, however, remained predominantly black.
Map 8: Trajectories of Predominantly Black Moderate-Income Neighborhoods 2000 to 2017
272004
272003
271900
271801
271700
271501
130805
271600
151300
151200
151100
151000
280200
280302
150900
150800
150701
150702
150500
130400
130600
120700
130100
130200
090800
080101
270101
260303
260302
080102
080200
080302
260301
260202
260201
270102
260102
260203
260402
260403
130300
150400
150600
160802
280402
160700
280401
280403
280404
250101
250102
250103
200800
200702
200600
250206
210200
200500
200400
200200
190100
180100
180300
200100
180200
020200
030100
030200
020100
010500
261100
261000
080301
080400
070300
070200
070100
060200
060100
060300
010100
010200
190300
200300
040200
160600
160500
150300
150200
150100
140100
140200
280500
110100
120600
120300
120201
170200
170100
170300
160100
160200
160300
160400
140300
120500
090900
080600
080500
090700
090200
270903
090600
090500
090300
090100
090400
080700
080800
070400
040100
100100
100200
100300
120400
100200
250303
250301
250205
250207
250203
250204
250401
250402
250600
240100
010400
010300
060400
240400
230300
210100
220100
240200
020300
240300
230200
230100
260605
260604
260501
260700
260404
260401
260900
260800
250500
200701
160801
280301
271503
271400
271300
271200
271101
271102
271001
271002
270805
270804
270801
270802
270901
270902
270803
270600
270301
270302
270402
270502
270501
270703
270702
270701
270401
260101
270200
120100
120202
130700
130803
130806
130804
271802
272007
272006
280101
280102
272005
*
*
*
*
Remained moderate-income neighborhoods in 2017
Moved downward to low-income
Moved upward to middle-income
“Hanging in” census tracts (census tracts that are keeping pace with the
overall citywide trajectory of growth)
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 35
On its face, that description suggests relative stability. The actual picture, however, is quite dierent:
• These neighborhoods account for roughly three-quarters of the total loss in black
households in Baltimore.
• These neighborhoods account for half of the loss in homeowners in Baltimore.
• The typical home in these neighborhoods has lost 30% of its value since 2000 in constant
(ination-adjusted) dollars.
The three upwardly moving census tracts are outliers in this category; two are on the edge of the increasingly
upscale Harbor East area, and may be potentially headed toward gentrication, and the third falls within
the orbit of the Johns Hopkins University campus. The characteristics of tract 604 have been aected by the
demolition of the Broadway Homes housing project and the construction of new housing on part of the
site since 2000. These three areas are seeing increases in household income and house prices, as well as an
increase in homeownership in the tracts close to Harbor East resulting from rehabilitation of vacant properties
and some inll construction. They have also seen a decline in their black population. While there has been
some growth in white residents, it has been far less than the decline in the black population. These tracts,
however, contained only 4% of the population of this neighborhood cluster. The trajectory of the remaining
96% is the story of this cluster.
Table 15: Selected Indicators for Predominantly Black Moderate-Income Neighborhoods
TRACT REMAINED
MODERATE INCOME
TRACT MOVED DOWNWARD
TO LOW INCOME
Median income change in constant $$
2000-2017
-8.3% -26.3%
Change in black population 2000-2017 -23,042 -8,980
% change in black population 2000-2017 -17.0% -12.1%
Median sales price in 2017 $50,000 $50,980
% change in sales price in constant $$
2000-2017
-30.8% -30.7%
Sales ratio 2017 2.9% 3.2%
Investor buyer share in 2017 35.7% 43.1%
Change in number of homeowners
2000-2017
-6,165 -1,547
% change in number of homeowners
2000-2017
-25.9% -26.9%
Homeownership rate in 2017 39.2% 33.0%
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 36
It is striking that there is little dierence in the key indicators except for income change between the tracts
that remained moderate income and those that moved downward. Given the greater dierence, however,
between the two subclusters in Table 15 in their homeownership rate and investor buyer share, it is possible
that either or both of those factors may have aected the relative income stability of those tracts that
remained moderate income.
I stress relative stability because the broad income ranges obscure a key issue: The great majority of the
“stable” tracts actually declined relative to the city of Baltimore as a whole. While in this cluster, the 2000
median income was 86% of the citywide median; by 2017, it was only 75% of the citywide median. Although
precise statistics on migration by income range are not available, it can reasonably be assumed that the black
families that moved out of these areas were of higher incomes than those who remained behind.
One question remains: Of these 42 census tracts, did all decline, or did some hang in, in the sense of keeping
pace with the overall citywide trajectory of growth? To answer that question, I looked to see which tracts, if
any, saw both income change and sales price change from 2000 to 2017 at least on par with citywide change
over the same period—that is, tracts that did not fall behind citywide levels on either measure. Only three
tracts met those criteria, which are shown in Map 8.
28
One is the Penn North neighborhood (tract 1303), south
of Druid Hill Park; a second is the Barclay area (tract 1204), immediately north of Greenmount West; and
the third is tract 2604.02, a small part of the large Frankford area near the city’s eastern border. The Barclay
neighborhood has been the focus of major investment supported by the city, state, and federal government
and the community, as has Penn North although to a lesser extent, and these investments both appear to have
had some impact. The circumstances aecting the other area is unclear. It is important to note, however, with
respect to the “hanging in” portion of Frankford, that the entire growth, in both income and sales price, took
place between 2000 and 2010. Since 2010, that neighborhood has been losing ground at levels paralleling
its peers.
The upshot is that almost all of Baltimore’s largely black moderate-income neighborhoods, many of which
were relatively healthy in 2000, are losing ground, and many are in crisis. Families continue to leave, and
household incomes are in sharp decline, while the housing market is on the edge of market failure. While the
median house value in these neighborhoods in 2000 was over 80% of the citywide median, it is now below
50%. The number of new buyers is far too low to absorb the supply of housing, the share of investor buyers is
far above the citywide average, and vacant housing is becoming endemic in some areas. The future of these
neighborhoods is one of the most dicult challenges faced by the city of Baltimore.
B. The Northeast Triangle
The role of what I have called the Northeast Triangle as a focus of black homebuying, and as the area of the
greatest black population growth in Baltimore, has been noted earlier. This area is also notable because of its
relatively high level of racial integration compared to the rest of the city, and because it represents the major
remaining reservoir of stable middle-class housing in the city. Its continued stability, therefore, has important
implications for the city’s future.
To drill down into this area, I looked at 20 census tracts containing roughly 13% of the city’s population, as
shown in Map 9. The map also shows which tracts were predominantly black, and which racially integrated in
2017. The signicance of that distinction will be discussed below.
The course of change in the Triangle has been uneven. During the decade from 2000 to 2010, this area saw
both white ight and signicant black in-migration. The area lost 40% of its white population or 13,000 people,
28 Tract 2707.01, which corresponds to the Idlewood NSA, and which by virtue of being entirely renter-occupied could not meet the criteria by deni-
tion, showed signicant income increase, particularly between 2010 and 2017.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 37
while gaining almost the same number of black residents. Along with a small Latinx increase, the overall
population of the area remained stable. Since 2010, however, the decline in the area’s white population has
slowed signicantly from an average of 1,300/year to 200/year; the black population has continued to grow,
resulting in an overall increase of population in the area of nearly 4,000 since 2010. The number of households,
however, has stayed the same, suggesting that many larger families with children are most probably replacing
older empty-nester families.
29
Map 9: Northeast Triangle Census Tracts
272004
272003
271900
271801
271700
271501
130805
271600
151300
151200
151100
151000
280200
280302
150900
150800
150701
150702
150500
130400
130600
120700
130100
130200
090800
080101
270101
260303
260302
080102
080200
080302
260301
260202
260201
270102
260102
260203
260402
260403
130300
150400
150600
160802
280402
160700
280401
280403
280404
250101
250102
250103
200800
200702
200600
250206
210200
200500
200400
200200
190100
180100
180300
200100
180200
020200
030100
030200
020100
010500
261100
261000
080301
080400
070300
070200
070100
060200
060100
060300
010100
010200
190300
200300
040200
160600
160500
150300
150200
150100
140100
140200
280500
110100
120600
120300
120201
170200
170100
170300
160100
160200
160300
160400
140300
120500
090900
080600
080500
090700
090200
270903
090600
090500
090300
090100
090400
080700
080800
070400
040100
100100
100200
100300
120400
100200
250303
250301
250205
250207
250203
250204
250401
250402
250600
240100
010400
010300
060400
240400
230300
210100
220100
240200
020300
240300
230200
230100
260605
260604
260501
260700
260404
260401
260900
260800
250500
200701
160801
280301
271503
271400
271300
271200
271101
271102
271001
271002
270805
270804
270801
270802
270901
270902
270803
270600
270301
270302
270402
270502
270501
270703
270702
270701
270401
260101
270200
120100
120202
130700
130803
130806
130804
271802
272007
272006
280101
280102
272005
Racially integrated (>30% white and <70% African American)
Predominantly (70%+) African American
Table 16 shows some key market indicators for the Triangle for 2000, 2010, and 2017. Sales prices in the area
saw a more pronounced bubble eect than in most of the rest of the city, perhaps driven not only by the rapid
inux of black buyers, but also by a sharp increase in the share of investor buyers, which more than doubled
from 2000 to 2010. Price increases substantially outpaced income increases during that period.
Price increases have continued—but more slowly—since 2010, and the share of investor buyers has dropped
from 29% to 22%. Some neighborhoods have recovered fairly well from the collapse that followed the
mortgage bubble, such as the two census tracts that correspond to the Hamilton Hills neighborhood, as shown
in Figure 13. Others, such as Frankford, have failed to regain much of the value lost at that time.
29 An alternative hypothesis is that couples that had previously moved to the area as childless couples are now having children, but that is less likely.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 38
Table 16: Key Market Indicators for Northeast Triangle
2000 2010 2017
Median sales price
$77,250 $134,700 $137,450
% of citywide median
128.8% 149.7% 130.9%
Investor buyer share
13.5% 29.2% 21.8%
Median income
$38,925 $50,099 $55,936
$ of citywide median
129.4% 127.2% 119.9%
Homeownership rate
63.7% 62.2% 61.0%
The sales ratio (the ratio between the number of sales and the number of single-family units) in the Triangle as
a whole in 2017 was a healthy 6%. Although the decline in household incomes and in the home ownership rate
are of some concern, the declines are modest. Overall, the areawide data suggest that the Triangle is at a point
of relative stability.
Figure 13: Median Sales Price Trends in Hamilton Hills 2000 to 2017
There are signicant dierences, however, between those parts of the Triangle whose populations are racially
mixed, and those that are predominantly black, raising potential future concerns. On a series of key indicators,
the seven predominantly black census tracts
30
are doing signicantly less well than the 12 racially mixed tracts.
Median income growth since 2000 was 24.9%, corresponding to a decline of 15% in constant ination-adjusted
30 There are actually eight such tracts in the Triangle, but one 2707.01 (Idlewood) has no owner-occupied housing, and thus no real estate market
data. As noted earlier, in terms of household income growth, this all-renter neighborhood has done well, particularly since 2010.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 39
dollars, compared to 52.5% in the racially mixed tracts. Sales ratios in 2017 were 4.7% in the predominantly
black tracts, just below the bottom of the Goldilocks range, compared to 7% in the mixed tracts, squarely in
the middle of the range. The median sales price in six of the seven predominantly black tracts was below the
median for the Triangle as a whole.
All three indicators relate closely to a fourth, which is the share of homebuyer mortgages going to black
homebuyers. As Figure 14 shows, by superimposing a trend line on the graph, the correlation between the
black share of new homebuyers and the black share of the existing population in 2017 was so close as to be
nearly absolute.
31
Put dierently, the number of white homebuyers declines in direct proportion to the growth
in the black population.
Figure 14: Share of Homebuyer Mortgages to Black Homebuyers
and Black Population Share by Census Tract 2017
This pattern is deeply distressing but should not come as much of a surprise. As stated earlier, neighborhoods
in Baltimore have been strongly aected by historical patterns of discrimination, segregation, redlining, and
white ight. The eect of racial perceptions and prejudices on homebuying choices is well known and has
been widely documented.
32
It says nothing about the quality of the homebuyers, their desire to sink roots in
the neighborhood, or any other attribute that may be associated with homeownership. It says a great deal,
however, about the quantity of homebuyers—that is, the size of the homebuyer pool likely to consider buying
in a given neighborhood.
Black homebuyers make up between 25% and 30% of the total pool of homebuyers in the city of Baltimore,
and about 15% of the total pool in the Baltimore metropolitan area.
33
While not insignicant, that is only a
31 The actual correlation is 0.864545, as close to a perfect 1.0 as any correlation is likely to be in the real world.
32 The research of Maria Krysan, a sociologist at the University of Illinois at Chicago, is particularly notable on this subject, e.g., “Does race matter in
the search for housing? An exploratory study of search strategies, experiences, and locations,” Social Science Research 37, no. 2 (2008): 581-603,
and many other papers.
33 This is based on the black share of total mortgages made and reported under the Home Mortgage Disclosure Act from 2014 through 2016. While
a small share of homes are bought by owner-occupants through all-cash deals, the number is small, and in all likelihood, given black/white wealth
disparities, the share of black home purchase mortgage borrowers is likely to be slightly higher than their share of all homebuyers.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 40
small part of the total market. With neighborhoods competing with one another throughout the city and
metro area, a neighborhood that has access to few of the remaining 85% of potential regional buyers is at an
inherent disadvantage. This is compounded by the fact that the trend among black homebuyers is increasingly
to buy homes in suburban areas or in largely white or mixed areas in the city, thus further reducing the
buyer pool for homes in predominantly black neighborhoods. With less demand for those homes, their price
will be less than that of similar homes in neighborhoods treated more favorably by the market, while these
neighborhoods will also have more diculty recovering from shocks like the foreclosure crisis and the Great
Recession, a phenomenon that has been referred to as a “discrimination tax” or a “segregation tax.”
Thus, over and above the ethical and social imperatives in fostering integration, the city of Baltimore has a
strong utilitarian argument in favor of eorts to sustain and, if possible, grow strong integrated communities
in the Northeast Triangle, ensuring that those neighborhoodsremain attractive to middle-class homebuyers of
all racial and ethnic backgrounds.
C. Gentrifying neighborhoods
The subject of gentrication is highly contentious because there is no real consensus about what
“gentrication” means. The term has come to be used in many dierent ways, many of which go well beyond
British sociologist Ruth Glass’ meaning when she coined the term over 50 years ago.
34
That said, for purposes
of this analysis, treating gentrication solely as a process of measurable neighborhood change, a denition is
needed that can measure the process of change in the demographic and housing market characteristics of a
neighborhood triggered by the in-migration of the more auent.
Three measures are widely used in research on the subject: (1) increase in household incomes; (2) increase
in sales prices; and (3) increase in educational attainment, namely the percentage of adults with a bachelor’s
degree or higher.
35
A neighborhood could show change in one of these measures for reasons unrelated to
gentrication. Because, for example, elderly households tend to have lower incomes than socioeconomically
similar families in their peak earning years, the replacement of elderly homeowners by young families
into a residential neighborhood could trigger a signicant increase in incomes without any change in
social or economic character.
36
Similarly, construction of even a small cluster of new homes in a low-value
neighborhood—even if subsidized for low- or moderate-income buyers—could appear in the numbers as a
dramatic increase in sales prices in that neighborhood from one year to the next.
For the purposes of this report and in the interest of simplicity, I will use the rst two measures, increases
in incomes and sales prices. I will discuss educational attainment as well as age distribution as additional
factors below. For incomes, I identify the neighborhoods that changed category in the following ways from
2000 to 2017:
Low or moderate ➜ middle, upper middle, or upper
Middle ➜ upper middle or upper
34 Ruth Glass coined the term in her 1964 book, London: Aspects of Change, where she describes in language worth quoting in full how: one by one,
many of the working class quarters of London have been invaded by the middle classes – upper and lower. Shabby modest mews and cottages […]
have been taken over when their leases have expired and have become elegant expensive residences. Larger Victorian homes, downgraded in an
earlier or recent period […] have been upgraded once again. The current social status and value of such dwellings are […] enormously inated by
comparison with previous levels in these neighborhoods. Once this process of ‘gentrication’ starts in a district, it goes on rapidly until all or most
of the original working class occupiers are displaced, and the whole social character of the district is changed (Glass, 1964). Glass was far from
the rst to notice or give a name to the process she described, but it was her coinage that stuck, and has become the term that dominates the
discourse about neighborhood change today.
35 This reects the reality that in today’s American economy, a four-year university degree is arguably the single most powerful proxy for high eco-
nomic and social status. Incomes and higher education correlate very strongly.
36 This may be taking place in the Glenham-Belhar neighborhood in the Northeast Triangle.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 41
If an upper-middle income neighborhood has moved up to the upper-income category, however, I do not
consider that to be gentrication.
For sales prices, the most meaningful measure is price appreciation relative to the citywide rate, rather than
the absolute dollar amount. I treat gentrifying tracts as being those where the median sales price rose over
that period by 50% more than the citywide rate. Since the median sales price in Baltimore went from $60,000
to $106,000 from 2000 to 2017, for an increase of not quite 77%, tracts that increased by 115% or more meet
that criterion. In actuality, in all but four of the 27 census tracts that meet this criterion, sales prices increased
at double or more the citywide rate—that is, by 153% or more over the study period. Those 27 census tracts—
with approximate neighborhood equivalents, and their changes in income category, black population share,
and sales price from 2000 to 2017—are shown in Table 17. The tracts are shown in Map 10.
A number of salient points ow from Table 17 and Map 10:
• There are no major surprises. These neighborhoods tend to be the ones most frequently characterized
as gentrifying in the local media. They are all incremental expansion of already strong areas, as in the
case of Harbor East, or they extend from major nodes of activity, such as the Inner Harbor, downtown,
or the Johns Hopkins Homewood campus.
• Over half of the gentrifying tracts were middle-income tracts in 2000—not wealthy, but not low-income
either. Only one gentrifying tract—Greenmount West—was a low-income neighborhood (median
income under 60% of the citywide median) in 2000.
• 18, or two-thirds, of the 27 gentrifying tracts were predominantly white in 2000, while only four were
predominantly black. Five were mixed.
Map 10: Gentrifying Neighborhoods 2000 to 2017
272004
272003
271900
271801
271700
271501
130805
271600
151300
151200
151100
151000
280200
280302
150900
150800
150701
150702
150500
130400
130600
120700
130100
130200
090800
080101
270101
260303
260302
080102
080200
080302
260301
260202
260201
270102
260102
260203
260402
260403
130300
150400
150600
160802
280402
160700
280401
280403
280404
250101
250102
250103
200800
200702
200600
250206
210200
200500
200400
200200
190100
180100
180300
200100
180200
020200
030100
030200
020100
010500
261100
261000
080301
080400
070300
070200
070100
060200
060100
060300
010100
010200
190300
200300
040200
160600
160500
150300
150200
150100
140100
140200
280500
110100
120600
120300
120201
170200
170100
170300
160100
160200
160300
160400
140300
120500
090900
080600
080500
090700
090200
270903
090600
090500
090300
090100
090400
080700
080800
070400
040100
100100
100200
100300
120400
100200
250303
250301
250205
250207
250203
250204
250401
250402
250600
240100
010400
010300
060400
240400
230300
210100
220100
240200
020300
240300
230200
230100
260605
260604
260501
260700
260404
260401
260900
260800
250500
200701
160801
280301
271503
271400
271300
271200
271101
271102
271001
271002
270805
270804
270801
270802
270901
270902
270803
270600
270301
270302
270402
270502
270501
270703
270702
270701
270401
260101
270200
120100
120202
130700
130803
130806
130804
271802
272007
272006
280101
280102
272005
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 42
Table 17: Gentrifying Neighborhoods in Baltimore
Put dierently, as shown in Table 18, of 110 “gentriable”
37
predominantly black census tracts in 2000, four had
gentried by 2017. Of 40 “gentriable” predominantly white tracts, 18 had gentried. Outcomes for racially
mixed tracts fell in between.
This pattern is consistent with the experience in other cities, where areas most likely to gentrify are not
only those that are predominantly white to begin with, but also those that are not deeply impoverished or
disinvested areas. This pattern has been found in studies of Chicago, New York City, Philadelphia, and St.
37 For our purposes, a “gentriable” tract is one that was neither upper-middle nor upper income in 2000.
TRACT NEIGHBORHOOD NEIGHBORHOOD CATEGORY BLACK POP SHARE MEDIAN SALES PRICE CHANGE
2000 2010 2017 2000 2017 2000 2017
1205 Greenmount West Low Mod Middle 94.3% 57.1% $82,250 $257,300 2.594
1308.04 Hampden Mod Middle Middle 5.9% 7.6% $56,000 $210,000 4.136
2607 Highlandtown Mod Mod Middle 8.6% 2.3% $41,200 $182,950 5.285
401 Downtown Mod Middle Middle 42.9% 14.1% $90,000 $207,300 2.261
602
Patterson Park
Neighbors
Mod Middle Middle 69.0% 42.1% $54,900 $184,700 3.641
1207 Remington Mod Middle Middle 33.5% 22.5% $48,900 $142,500 3.170
1401 Bolton Hill Mod Mod Middle 37.3% 34.2% $145,000 $371,500 2.793
2610
Highlandtown/Pat
Park North
Mod Mod Middle 43.6% 27.8% $73,000 $325,000 4.603
603 Butchers Hill Mod Middle Middle 74.1% 43.7% $59,750 $189,000 3.121
604 Washington Hill Mod Mod Middle 85.7% 64.8% $62,500 $175,000 2.736
1203
Charles Village/
Harwood
Mod Mod Middle 70.3% 49.4% $64,900 $124,000 2.353
1202.01 Abell Mod Middle Upper Mid 13.6% 18.0% $80,000 $268,500 3.156
2611 Canton Mod Upper Upper 3.6% 7.0% $73,000 $325,000 4.603
105
Upper Fells Pt/
Butchers Hill
Middle Upper Upper Mid 7.3% 4.3% $95,000 $280,000 3.211
201 Upper Fells Point Middle Upper Mid Upper Mid 1.5% 6.6% $68,500 $280,000 3.394
302
Little Italy/
Jonestown
Middle Upper Mid Upper Mid 20.8% 31.2% $125,000 $273,750 2.160
1306 Hampden Middle Middle Upper Mid 3.4% 7.0% $57,450 $225,000 4.352
1308.06 Woodberry Middle Middle Upper Mid 18.1% 18.9% $47,250 $212,000 3.701
101 Canton Middle Upper Mid Upper 3.7% 6.8% $129,900 $345,000 2.533
102 Patterson Park Middle Middle Upper 1.8% 5.2% $63,500 $259,900 4.457
103 Patterson Park Middle Upper Mid Upper 0.9% 2.0% $67,000 $275,000 3.955
104 Canton Middle Upper Mid Upper 3.2% 2.3% $133,000 $305,000 2.297
2302
Federal Hill/So
Baltimore
Middle Upper Upper 3.0% 13.3% $90,000 $289,950 2.989
2303
So Baltimore/Pt
Covington
Middle Middle Upper 1.9% 4.4% $52,000 $265,000 5.173
2401 Locust Point Middle Upper Upper 0.2% 3.1% $63,500 $371,250 6.614
2404 Locust Point Middle Upper Mid Upper 1.7% 0.6% $96,715 $331,567 3.081
2609 Brewers Hill Middle Upper Mid Upper 3.3% 7.7% $65,000 $293,750 4.538
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 43
Louis.
38
This does not mean that there is no gentrication in predominantly black neighborhoods, but it is the
exception rather than the rule. Gentrication tends to take place when such neighborhoods are exceptionally
well situated relative to strong neighborhoods or to economic centers, such as Greenmount West, or in cities
like Washington D.C. that are experiencing intense gentrication pressures, and where there are few, if any,
historically white working-class neighborhoods available to be gentried.
Table 18: Frequency of Gentrication by Neighborhood Category
“Gentriable”
tracts in 2000
Tracts that
gentried by 2017
Percentage of
“gentriable” tracts
Predominantly white
40 18 45.0%
Mixed
32 5 15.6%
Predominantly black
110 4 3.6%
All tracts
182 27 14.8%
Gentrication in Baltimore, as in many other cities, is largely driven by a very specic group: young
predominantly white people, so-called millennials, with university degrees. While they are not the only
people moving into the city’s gentrifying areas, they are the driving force of change. As Table 19 shows,
the transformation of these areas has been paralleled by a massive increase in their number of university
graduates, and in the share of their population aged 25 to 34. The share of university graduates in these
tracts more than tripled, pushing the citywide share close to the national level. In 2017, 80% of the residents
of downtown Baltimore had a bachelor’s or higher degree, and 55% were aged 25 to 34. Map 11 shows
the areas of greatest millennial concentrations, clustered around the Inner Harbor, downtown, and Johns
Hopkins University.
Table 19: Change in Share of Adults with a Bachelor’s or Higher Degree
and Population Aged 25 to 34, 2000 to 2017
% of Adults with Bachelor’s
or Higher Degree
% of Population Aged
25 to 34
2000 2017 CHANGE 2000 2017 CHANGE
Gentrifying census tracts
19.9% 67.0% 237.2% 20.2% 36.3% 79.7%
Baltimore City
19.1% 30.4% 59.2% 14.4% 18.6% 29.2%
United States
24.4% 30.9% 26.6% 14.0% 13.7% -2.1%
38 For Chicago, see Hwang, Jackelyn, and Robert J. Sampson. “Divergent pathways of gentrication: Racial inequality and the social order of renewal
in Chicago neighborhoods.” American Sociological Review 79:4 (2014): 726-751.For Philadelphia, see Moselle, Aaron and Annette John-Hall. “The
surprising truth behind the racial dynamics of gentrication in Philly” Philadelphia, PA: WHYY, March 13, 2018. https://whyy.org/articles/surpris-
ing-truth-behind-racial-dynamics-gentrication-philly/ For New York and Chicago, see Timberlake, Jerey M., and Elaina Johns-Wolfe. “Neighbor-
hood ethno-racial composition and gentrication in Chicago and New York, 1980 to 2010.” Urban Aairs Review 53:2 (2017): 236-272;. For St. Louis,
see Swanstrom, Todd, Henry S. Webber and Molly W. Metzger. “Rebound neighborhoods in older industrial cities: The case of St. Louis” in Brown,
Alexandra, David Buchholz, Daniel Davis and Arturo Gonzalez, ed. Economic Mobility: Research & ideas on Strengthening Families, Communities
and the Economy. St. Louis, MO: Federal Reserve Bank of St. Louis and Board of Governors of the Federal Reserve System (2016) and Mallach, Alan
and Karen Beck Pooley. “What drives neighborhood revival? Qualitative research ndings from Baltimore and St. Louis.” Cambridge, MA: Lincoln
Center of Land Policy, Working Paper WP18AM12018 (2018).
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 44
Gentrication in Baltimore has resulted in two separate population declines. The number of black households
living in these tracts has dropped slightly, while the lower-income white population in the area has dropped by
far more, as both have been increasingly replaced by more auent white residents.
The number of black residents has dropped from its historic peak in most, although not all, of the gentrifying
tracts. To provide a perspective on this, the trends from 1980 to the present for total population and black
population, and the average annual change in each by decade, are shown in Table 20. The table shows that the
black population in these tracts increased from 1980 to 2000, and has since declined, at a slightly faster but still
moderate rate, to where their black population share is now roughly the same as it was in 1980. In contrast,
total population (largely non-Latinx white) declined from 1980 to 2000, but it has been growing back since
then, adding 6,300 people since 2000.
Map 11: Millennial Concentrations in Baltimore
Insucient Data
29.99% or less
30.00% – 39.99%
40.00% or more
SOURCE: Map created by PolicyMap
The loss in black population in gentrifying tracts represents a very small part of total black population loss in
Baltimore. Since 2000, over 60% of Baltimore’s census tracts have seen a decline in black population, losing a
total of 64,396 black residents, oset roughly in half by smaller gains in other parts of the city, most notably
the Northeast Triangle. The 15 gentrifying tracts that lost black population lost a total of 5,657 black residents,
slightly less than 9% of the citywide total, oset by gains of roughly 1,500 black residents in the 12 gentrifying
tracts that gained black population.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 45
Table 20: Change in Total Population and Black Population in Gentrifying Census Tracts 1980 to 2017
POPULATION TOTALS 1980 1990 2000 2010 2017
Total population
79,418 74,550 63,821 67,765 70,157
Black population
14,689 16,473 17,382 15,115 13,201
% Black
18.5% 22.1% 27.2% 22.3% 18.8%
AVERAGE ANNUAL CHANGE
1980-1990 1990-2000 2000-2010 2010-2017
Change in total population
-487 -1073 394 341
Change in black population
178 91 -227 -273
Looking at the change in the number of black households in these tracts, however, a somewhat dierent
picture emerges. Although the black population dropped by over 4,000 from 2000 to 2017, the number of
households dropped by less than 700. The number of black households increased in half of the gentrifying
tracts. This reects a drop in the size of the average black household in these areas from 2.7 to 2.3 people.
In other words, much of the change was attributable to the larger black households moving out and being
replaced in large part by smaller ones or, alternatively, to existing households becoming smaller, either as a
result of children growing up and moving out, or elderly household members passing.
During the same period, however, these areas were losing lower-income white residents at a much more
rapid pace. The number of lower-income white households, those earning $25,000 or less in 2000 adjusted
for ination in 2017,
39
dropped by nearly 3,800 households over that period, or 52%. The respective trends
are compared in Table 21. Because the population of most of these areas was largely white in 2000, it logically
follows that most of the impact of population change and economic change would be felt by the lower-income
white population. Some market shifts are colorblind.
Table 21: Change in Number of Households in Gentrifying Census Tracts 2000 to 2017
Households in 2000 Households in 2017 N change % change
Black
6,430 5,740 -690 -10.7%
Lower-Income White
7,258 3,463 -3,795 -52.3%
All of this reinforces the point that “what’s going on isn’t displacement of the poor—it’s replacement.” In the
absence of strong intentional action to counteract the trend, replacement becomes all but inevitable even
with no overt displacement.
40
As an area attracts more auent buyers or renters, those buyers or renters in
Baltimore are more likely to be white than black. As a result, as units turn over, a progressively larger share of
the new buyers or renters will be white compared to their share of the pre-existing owners or tenants, or they
will be auent white households rather than lower-income ones, white or black. The change is consistent with
the premise that that is indeed what is taking place.
39 That income level roughly corresponds to 80% of the citywide median. I compared the number of households earning that level in 2000 with the
number earning $35,000 in 2017, a dierence that approximately tracks the rate of ination over that period.
40 As noted earlier, the term “displacement” with its inference that the families involved were involuntarily displaced, e.g., compelled to leave
because of rehabilitation, condominium conversion, or some similar reason associated with price appreciation. While I have no information on
whether or not this has taken place in any of these neighborhoods, the central point here is that the outcomes visible in the data can easily be
explained by turnover processes without any involuntary displacement taking place.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 46
Tenants are more strongly aected by market change than homeowners. Turnover among tenants is far higher
than among homeowners. The duration of the average tenant’s stay in the same unit in Baltimore is two years,
compared to 15 years for homeowners. Moreover, while homeowners may nd themselves under pressure to
sell, they nonetheless are likely to benet economically from the appreciation taking place in the neighborhood
and have the option to stay in their home. Tenants lack these options.
As a result, these neighborhoods are seeing not only a gradual but steady decline in the number of black
tenants as they are replaced by white tenants or buyers, but also a highly uneven and variable pattern in
the number of black homeowners. In some neighborhoods, including Greenmount West and Bolton Hill,
the number of black homeowners has increased.
41
This is not taking place in other gentrifying tracts, but
the data suggest that has less to do with homeowners being more likely to move out and more to do with
the fact that there are few black homebuyers moving into these neighborhoods. In 2017, 1,287 homebuyers
received mortgages to buy homes in the city’s gentrifying neighborhoods; of this total, only 83 (6%) were black
homebuyers. Almost half of all white homebuyers in Baltimore that year bought their homes in one or another
of these 27 tracts, compared to 1 out of 20 black homebuyers.
Finally, the data presented here describes where gentrication, as dened earlier, has already led to
measurable change, although in many areas it remains a work in progress. Over the coming decades,
depending on many dierent factors, gentrication may extend beyond the existing tracts. If so, based on past
experience not only in Baltimore but also elsewhere, it will move slowly and incrementally into areas adjacent
to those already gentrifying. Whether and to what extent this happens, however, will depend not only on what
happens inside Baltimore, but on regional, national and even global economic and demographic trends.
While prediction is inherently dicult, careful analysis of year-by-year data on sales prices, building permits,
and other measures should allow city ocials and others concerned with this issue to identify emerging
trends, and begin to frame intentional strategies not to prevent change from happening, but to the extent
feasible, mitigate the harms and maximize the benets of change for existing residents and other lower-
income Baltimoreans.
42
IV. CLOSING COMMENTS
The most powerful single conclusion that ows from the preceding pages is that, when it comes to
neighborhood trajectories in Baltimore, race trumps income. If one looked at two Baltimore neighborhoods in
2000 that were all but identical in their social and economic characteristics, but one was largely white and the
other largely black, that one piece of information would be enough to predict with high probability where each
would stand relative to the other in 2017 or 2020. As a matter of equity and social policy, this is undesirable
and unacceptable; in order to address it constructively, however, we must acknowledge its reality.
The second conclusion is the magnitude of the loss of black households from out-migration. Although
Baltimore had a net loss of some 30,000 black residents from 2000 to 2017, that is the tip of a much larger
iceberg. Looking at roughly 60% of neighborhoods that lost black population, these neighborhoods lost over
60,000 black residents; when one adds that the excess of births over deaths in city’s black population was
over 20,000 during that period, the sheer magnitude of the population shift becomes apparent. As black
households have either moved from inner neighborhoods to outer ones, particularly to the Northeast Triangle,
inside the city, or left the city entirely, large numbers of traditionally black neighborhoods have lost population.
They have become poorer as their middle-class residents have left, and are seeing the changes that typically
follow sustained population loss—dropping house values, declining homeownership rates, higher vacancy
rates, and rising abandonment.
41 It is worth noting that, according to the most recent available HUD data, nearly 30% of the households in Greenmount West and slightly more
than 50% of the renter households live in subsidized housing, either in aordable housing developments or as holders of Housing Choice Vouch-
ers.
42 Although beyond the scope of this project, careful analysis of some of the datasets used in this report, particularly in terms of recent sales price
trends, might potentially provide clues to emerging areas of gentrication.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 47
These shifts are most pronounced in the roughly 30% of the city’s census tracts that are predominantly black
and where the median household income is between 60% and 100% of the citywide median. Many middle-
income largely black areas are also struggling, but by and large are not seeing the extent of decline visible in
the moderate-income neighborhoods. Many are showing signs of decline, however, and strategies to arrest
their decline are likely to be important in order to help them maintain or regain stability.
43
At the same time, the city’s white population is moving in the opposite direction. While Baltimore is still
losing white working-class and middle-class families, it is gaining auent ones, in particular young so-called
millennials with university degrees and highly marketable skills. White households also represent a much
larger part of the homebuyer market than black ones. As noted earlier, the rate of white in-migration into
Baltimore is signicantly higher than that of black in-migration, while in recent years the number of white
homebuyers in the city has been much greater than the number of black homebuyers.
Add to this picture the painful reality that white homebuyers are far less likely to buy homes in predominantly
African American neighborhoods than in neighborhoods that are either predominantly white or racially
mixed, and the full extent of the market disparity becomes clear. The continued reality of racial discrimination,
whether with respect to real estate sales, mortgage lending or any other factor, simply adds further pressure
to an already severely imbalanced situation.
The upshot is that a relatively small number of neighborhoods are seeing increased investment and
homebuying activity, and a much larger number are either treading water or declining. The great majority of
white homebuying activity is going into those areas like Roland Park that were stable upper-middle or upper-
income areas in 2000 (about 9% of Baltimore’s census tracts) and the roughly 14% of the city’s tracts that have
signicantly gentried since then.
From the standpoint of public and social policy, the city of Baltimore faces three distinct neighborhood
challenges. First and foremost among these, I believe, is the challenge of stabilizing and reversing the
decline of as many as possible of the city’s struggling predominantly black moderate- and middle-income
neighborhoods. This is both a physical and an economic problem, and as such, poses a classic conundrum.
Without greater income and wealth-building opportunities for their residents, it is unlikely that the
neighborhoods can truly become stabilized. At the same time, if their residents gain new skills and better jobs,
open successful businesses, and then leave their neighborhoods, they benet, but their neighborhoods, and
for the most part the city of Baltimore, do not.
Along with economic opportunity strategies, this calls for a determined eort to improve the quality of
life in these neighborhoods—a term that encompasses their physical environment, public safety, quality
education, and more—to make them better places for everyone regardless of their income and education,
and to make them places where people who have the ability to choose among neighborhoods, and can aord
to move either to other parts of the city or its suburban surroundings, will choose to stay or move into. The
strategic framework recently released by the city’s Department of Housing and Community Development is an
important step in this direction.
This discussion underscores the central role that the loss of working- and middle-class families plays in
fostering neighborhood decline in Baltimore. If that loss is to be stemmed, and in time reversed, all public,
private, and nonprot stakeholders must ask the two-part question: Why are they leaving, and what can be
done to change the conditions that are prompting them to leave?
43 Many of this group of neighborhoods t into the neighborhood cluster that the City’s Department of Housing and Community Development refers
to as SCENs, or Strategic Code Enforcement Neighborhoods, and as such, have been a major part of the City’s Vacants to Value strategy. While
that strategy has been highly eective, in the narrow sense of it having led to rehabilitation and reuse of large numbers of vacant properties in
those areas, it has not in most cases changed their trajectory, making clear that the emergence of vacant properties in these neighborhoods is
more a symptom than a cause of their underlying problems.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 48
The second challenge is to ensure the long-term stability of the areas where black homebuyers have been
moving in recent years. The largest cluster of these areas is in the Northeast Triangle, but a smaller number
of other neighborhoods, such as Ashburton and Howard Park in the northwestern part of the city, are also
drawing black homebuyers. These neighborhoods are important to the city of Baltimore. Not only are they
attractive neighborhoods, in most cases with detached single-family homes very dierent from the iconic
Baltimore row house, but also they are the areas where, to the extent that black middle-class families are
staying in the city, most are moving. These areas represent a key reservoir of stable property values, tax
revenues, and engaged citizens. Maintaining their vitality, by making sure that they continue to oer a high
level of amenity value, including not only good schools and public safety, but also attractive, well-maintained
public open space, strong commercial hubs or corridors, and strong neighborhood institutions, should be an
ongoing eort. No one should assume that they will “take care of themselves.”
Finally, the issue of present and future gentrication represents both an opportunity and a challenge for
Baltimore. From many perspectives, the market transformation of the areas around the Inner Harbor,
downtown, and the Johns Hopkins campus is a positive trend for the city. Compared to the nation, and even
more markedly when compared to its surrounding counties, Baltimore, despite its progress in recent years,
remains a very poor city. Household incomes and property values are far lower than in the surrounding area,
and the city lacks the resources to address its daunting challenges. For the city’s scal and economic survival,
it needs to draw and hold an economically diverse population, and attract investment in homes, multifamily
housing, and commercial properties.
Moreover, it is important to remember a basic principle: neighborhoods change. With the arguable exception
of perhaps the most stable high-income and the most distressed areas, neighborhoods are in constant ux.
They change economically, they change culturally, and their racial or ethnic mix changes. To hope to freeze any
neighborhood in its economic, social, and racial conguration of a particular moment in time is an exercise,
whatever one’s intentions, that is bound to fail.
Baltimore is arguably fortunate in at least one respect, in that the modest scale and gradual pace of
gentrication in Baltimore compared to magnet cities like Washington D.C. or Seattle mean that any household
that is priced out of one area may still nd housing in other parts of the city, often nearby. That may be poor
consolation for a family that has deep roots in a particular neighborhood, but it does make an economic
dierence. Moreover, it is important to acknowledge the dierence, as noted earlier, between displacement
and replacement. The process of population change in gentrifying neighborhoods may not involve any
overt action to push people out; in a separate 2016 analysis, I found no relationship between the volume of
evictions and the rate of increase in rent levels, and a negative relationship between evictions and increases
in household income as a proxy for gentrication.
44
Recent research from New York City, a city with far more
intense market pressures than Baltimore, found that gentrication did not aect the household mobility rates
of low-income families.
45
Cities cannot freeze neighborhoods or tenants in place, nor is it likely to be a sound strategy even if it were
possible. At the same time, the city should encourage production of long-term aordable housing in areas
undergoing gentrication, in order to create a pool of units that are not driven by the market pressures on
those areas, and work with both tenants and landlords to encourage increased use of housing vouchers in
those areas. At present, few of the city’s gentrifying neighborhoods have more than a handful of subsidized
units. Only in Greenmount West and Bolton Hill do subsidized units, including vouchers, make up more than
25% of the neighborhood’s rental housing stock.
44 I realize that eviction is far from the only form of action that leads to displacement. It is, however, the one form of action that can be directly mea-
sured, and logic would suggest that if involuntary displacement were taking place to any signicant degree, it would be reected in the eviction
statistics.
45 Kacie Dragan, Ingrid Ellen, and Sherry A. Glied. “Does Gentrication Displace Poor Children? New Evidence from New York City Medicaid Data.”
NBER Working Paper No. 25809, May 2019.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 49
In closing, I recognize the magnitude of the city’s task in addressing its neighborhood challenges. Although
the city and its partners have accomplished a great deal in recent years, far more needs to be done. I hope
that this analysis, which I believe is the rst detailed, factually grounded analysis of the city’s neighborhood
conditions and trends, will be a valuable resource in that eort.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 50
APPENDIX I: METHODOLOGY
Whenever one conducts a study of neighborhood conditions and trends, particularly in a city as large as
Baltimore, one must begin by making a series of decisions of how to divide the city into neighborhoods, and
how to organize those neighborhoods for purposes of the study. There is no simple right or wrong way to do
this, as one can oer some rationale for a variety of dierent approaches. Generally speaking, though, the
choices must be reasonable, and they must end up with the data being organized into a small enough set of
categories so that it is manageable and comprehensible.
The rst step is dividing the city into neighborhoods. The Baltimore Planning Department has divided the city
into over 200 Neighborhood Statistical Areas (NSAs). Although at times perhaps arbitrary, this breakdown
probably represents the closest parallel to how Baltimoreans dene the neighborhoods they live in. However,
the NSA boundaries do not correspond to census tracts, the standard unit by which almost all small-area
data are compiled. As a result, obtaining and assembling data on neighborhood change by NSA would be an
extraordinarily time-consuming and dicult process, which would also require many compromises with data
quality and availability. Conversely, the Baltimore Neighborhood Indicators Project has divided the city into
55 Community Statistical Areas, for which it provides an awesome body of datasets. Those areas, however
useful they are for many purposes, are too large—and often contain within them subareas of widely varying
character—to be most useful for the purposes of this study. In the end, I decided to use census tracts, the
unit created and used by the U.S. Census Bureau for small-area analysis. With the city divided into nearly
200 tracts, they are small enough to be meaningful and relatively homogenous and have the advantage that
nearly all datasets are available by census tract.
46
While census tracts are not the same as the NSAs, they are
often roughly comparable to the neighborhoods designated by the city; thus, when I refer to a neighborhood
by name in this report, the reader should understand that I am referring to areas that are approximately the
same as that named neighborhood, not identical.
To try to show separately how each of the 200 census tracts did or did not change between 2000 and 2017
would not only be unwieldy and unduly time consuming, but also result in a report that would be so detailed
it would be meaningless except as a reference document. In order to provide meaningful results, I segmented
the city’s census tracts into categories based on race and economic condition. With respect to race, I used
the percentage of black population, and with respect to economic level, I used the median
47
tract household
income. I looked at data for 2000, 2010, and 2017. After exploring a number of alternatives, I arrived at
the breakdown shown in the matrix that follows, based on income ranges relative to the citywide median
household income. I will use the descriptive terms for the economic and racial composition of the city’s
neighborhoods shown in the matrix frequently in the report.
46 Another small advantage of using census tracts is that it allows one to make comparisons with the data provided in the NCRC report cited earlier.
In order to be able to compare census tract data over time, I have utilized the Neighborhood Change Data Base created by Geolytics, Inc., which
normalizes data by census tract boundaries for each decade from 1970 to those used since 2010.
47 Median refers to the midpoint of a range of numbers, dened as that number where half of the numbers are lower and half are higher. It is dier-
ent from average, which is the sum of the numbers divided by the number in the range.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 51
Table 1: Neighborhood Category Matrix by Economic Level and Racial Composition
Economic Composition Racial Composition
Neighborhood Type
Range
48
0-29.9% Black 30-69.9% Black 70-100% Black
Predominantly White Mixed Predominantly Black
Low Income
0-59.9% X
Moderate Income
60-99.9% X X X
Middle Income
100-149.9% X X X
Upper-Middle Income
150-199.9% X X X
Upper Income
200%+ X
The matrix oers a total of 15 possible neighborhood categories. The actual number of categories is 11, as
shown by “X” in the table.
49
There are no census tracts (e.g., predominantly white low-income tracts) in the
other categories. The income ranges—that is, the range within which the tract median falls—for the three
time periods looked at are shown in Table 2. A tract that had a median income of $40,000 in 2000 would be
considered middle income, and if its median fell to $35,000 in 2010, it would be considered moderate income
at that point. The greater part of the data that I used comes either from the decennial (every 10 years) census,
or from the American Community Survey (ACS), an annual survey of a sample of households conducted by the
U.S. Census Bureau.
50
The ACS provides data on demographics and economic condition of residents, as well
as housing data, such as homeownership rates or vacancy data. To supplement these data, I used housing
market data acquired from CoreLogic, including sales prices and volumes, and the split between owner-
occupant and investor buyers, by census tract.
Table 2: Income Ranges by Neighborhood Type for 2000, 2010, and 2017
NEIGHBORHOOD TYPE RANGE 2000 2010 2017
Low Income
0-59% $0-$18,046 $0-23,631 $0-27,984
Moderate Income
60-99% $18,047-30,078 $23,632-39,386 $27,985-46,641
Middle Income
100-149% $30,079-45,117 $39,387-59,079 $46,642-69,961
Upper-Middle Income
150-199% $45,118-60,156 $59,082-78,772 $69,962-93,282
Upper Income
200%+ $60,157+ $78,773+ $93,283+
Citywide Median
$30,078 $39,386 $46,641
SOURCE: 2000 Decennial Census, 2006-2010 and 2013-2017 American Community Survey
48 I explored using dierent ranges, particularly ranges that were distributed more evenly relative to the citywide median (such as 0-49%, 50-79%,
80-119%, 120-149%, and 150%+), but found that because the citywide median is so low relative to the metro-area median as well as the nation-
al median, that would result in ranges that were lower than what it reasonably means to be “middle” or “upper” income in that larger context.
Baltimore’s median income is roughly 80% of the national median, so 100-149% of the Baltimore median is equivalent to 80-120% of the national
median.
49 There was a single census tract in the low-income/mixed category. It did not show signicant change in either economic level or racial mix over
the study period.
50 Because of the small size of the ACS samples, annual data is not provided at the census tract level. Instead, the U.S. Census Bureau combines
data for ve-year periods (referred to as the “Five-Year American Community Survey”), which it provides for census tracts. With respect to income
data, the data for earlier years are inated to the last year in the series; therefore, the data shown in Table 2, even though they are for a ve-year
period, measure the income for the years shown in the table.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 52
APPENDIX 2 : CORRELATIONS BETWEEN ECONOMIC, DEMOGRAPHIC,
AND HOUSING-MARKET FACTORS
Over the preceding pages, we’ve described how change in a variety of areas—population change, change
in house values, homeownership, and more—varies depending on the type of neighborhood, based on
neighborhood income and racial distribution.
But the important thing about all of these dierent factors is that they are not independent of one another—
they are related. How one changes aects, to varying degrees and in varying directions, what happens to the
others. Homeownership rates, sales prices, incomes, vacant properties, and the share of investor buyers all
aect each other in dierent ways. In order to tease out these relationships, statisticians use a simple measure
known as correlation, which measures the extent to which the relationship between the distributions of two
factors, or variables, is likely to be the product of chance, or reects some actual relationship or connection
between the factors. A correlation can be positive, when both variables move in the same direction, such as the
proximity of a place to the equator and the average temperature in that place; or it can be negative, when both
variables move in the opposite direction, such as the proximity of a place to the equator and its average annual
snowfall. A correlation of 0 reects pure chance, or no relationship; a correlation of 1 reects absolute identity
between the distribution of two factors. In real life, of course, nothing is likely to have a correlation of either 0
or 1, but the higher the number, the stronger the relationship between the two variables.
51
As is well known, correlation does not mean causality. Just because there is a relationship between two factors
does not prove that Factor A caused Factor B, or vice versa. Figuring out what factors caused other factors is
far more complicated and uncertain than simply showing that they are related to each other. That said, nding
a strong correlation between two factors tells us that something meaningful is going on here, and that it
may be worthwhile to look more closely at them, and to think about what that relationship may indeed mean
regarding the future of the city’s neighborhoods.
To illustrate this point, I have calculated correlations for a cluster of factors in Baltimore’s middle-income
neighborhoods, which I show in Table 1 for each of the subcategories of middle-income neighborhoods, with
correlations color-coded in three categories: weak (more than 5% probability of chance), moderately strong
(1%-5% probability of chance), and very strong (less than 1% probability of chance).
Looking rst at the cluster of predominantly white middle-income neighborhoods, I see that the correlations
between the factors are, with few exceptions, very strong. As incomes go up, white populations go up, sales
prices go up, homeownership levels go up, and the share of investor buyers goes down. These are all quite
consistent with what is known about the dynamics of change, and the conditions in Baltimore. That sales
prices should go up with higher incomes is clearly to be expected, while, because I’ve shown the city’s white
population is becoming increasingly auent, the further correlation between increase in white population,
incomes, and sales prices is logical. As sales prices go up, the share of investor buyers goes down, perhaps
because if prices are too great a multiple of the rent roll, the investor’s return on an investment will become
too small. Investors typically look for markets where the ratio between the sales price and the annual rent roll
is low enough to ensure a decent return on equity.
51 The way the strength of a relationship is measured in correlation is by determining the probability or likelihood that the relationship is the product
of chance from 100% to, say, 0.00001%. So, when I say that a relationship is signicant at the 0.01 (or 1%) level, the probability that the relation-
ship is the result of pure chance is only 1%; in other words, there is a 99% probability, or likelihood, that there is a relationship between the two
variables and only a 1% likelihood that it is pure chance. For practical purposes, statisticians generally consider anything with a probability of the
relationship being a product of chance of 10% or more as not being statistically meaningful.
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 53
Table 1: Correlations Between Selected Variables for Middle-Income Neighborhoods
PREDOMINANTLY WHITE MIDDLE-INCOME NEIGHBORHOODS
INCOME Δ
BLACK
POP Δ
WHITE
POP Δ
SALES
PRICE Δ
SALES
VOLUME
2017
INVESTOR
BUYERS
2017
BLACK POP Δ
-.4257
WHITE POP Δ
.6506 -.5018
SALES PRICE Δ
.8303 -.4540 .6199
SALES VOLUME 2017
.3158 -.2453 .7129 .3735
INVESTOR
BUYERS 2017
-.6370 .3942 -.5072 -.6642 -.3758
HOMEOWNER Δ
. 4475 -.2360 .6772 .5501 .6537 -.3492
MIXED MIDDLE-INCOME NEIGHBORHOODS
INCOME Δ
BLACK
POP Δ
WHITE
POP Δ
SALES
PRICE Δ
SALES
VOLUME
2017
INVESTOR
BUYERS
2017
BLACK POP Δ -.1708
WHITE POP Δ .3718 -.8005
SALES PRICE Δ .7525 -.4224 .4217
SALES VOLUME 2017 .5611 .2158 -.2452 .5064
INVESTOR
BUYERS 2017
-.4176 .1229 . 2712 -.5223 .0841
HOMEOWNER Δ .4871 .6018 -.3230 .2780 .4098 -.4394
PREDOMINANTLY BLACK MIDDLE-INCOME NEIGHBORHOODS
INCOME Δ
BLACK
POP Δ
WHITE
POP Δ
SALES
PRICE Δ
SALES
VOLUME
2017
INVESTOR
BUYERS
2017
BLACK POP Δ
.2201
WHITE POP Δ
.1882 -.5686
SALES PRICE Δ
.3402 . 3346 .1043
SALES VOLUME 2017
.2001 -.3142 .2452 . 1559
INVESTOR
BUYERS 2017
-.3202 -.2965 -.0452 -.7333 -.0203
HOMEOWNER Δ
.4596 -.1410 .6954 . 3322 . 0738 -.3196
Relationship statistically signicant at the .01 level
Relationship statistically signicant at the .05 level
Relationship not signicant at the .05 or stronger level
abell foundation
Drilling Down in Baltimore’s Neighborhoods | THE ABELL FOUNDATION 54
The correlations in the mixed and predominantly black middle-income neighborhoods are far less strong
than in the predominantly white middle-income neighborhoods. So, while there is a similar relationship
between rising incomes, higher sales prices, and fewer investor buyers in predominantly black middle-income
neighborhoods, it is weaker—particularly with respect to sales price—than in the white middle-income
neighborhoods. That reects the painful reality that predominantly black neighborhoods throughout the
United States pay what could be called a “discrimination tax” or “segregation tax,” indicating how the real
estate market perceives predominantly black neighborhoods, in which homes in largely black neighborhoods
will carry lower market values than those in predominantly white neighborhoods of comparable social and
economic character.
52
Along similar lines, it is worth noting that greater sales activity measured by sales volume in largely white
tracts is strongly linked to higher prices, along with signicant declines in the share of investor buyers
and increases in the share of homeowners, but the same is not true in predominantly black tracts. This
suggests that the housing markets are working quite dierently in the two categories of neighborhood, and
indeed, may raise a question about the ecacy of some possible strategies to increase market activity in the
latter areas.
While the relationships between these and other variables could be explored in far greater detail, that
would be beyond the scope of this report. The key point I want to make here, though, is that there are many
relationships between the dierent elements that go into neighborhood change, that they are complex,
and above all, that they do not work the same way in all situations or conditions. This, in turn, reinforces the
important policy point that a given neighborhood stabilization or revitalization strategy will not work the same
way in dierent neighborhoods, because its outcomes are not simply a product of the strategy itself, but a
product of the interaction between the strategy and the particular conditions of the neighborhood.
52 For a recent report documenting this reality nationally, see Andre Perry, Jonathan Rothwell and David Harshbarger, “The Devaluation of Assets
in Black Neighborhoods,” Washington DC: Brookings Institution (2018), available at https://www.brookings.edu/research/devaluation-of-as-
sets-in-black-neighborhoods/.
