Legacy City Revitalization:
The Role of Federal Historic
Tax Credit Projects
Kelly L. Kinahan, AICP
Doctoral Candidate
Levin College Research Conference
August 20, 2015
Research Questions
(1) What is the distribution of RTC activity across legacy city
neighborhood types?
(2) What is the relationship between historic tax credit activity
and changes neighborhood racial, socioeconomic, and
housing characteristics?
2
Legacy cities:
 shrinking, post-industrial, right-sizing, etc.
 Long-term population loss
 Economic restructuring
 Vacancy, abandonment, foreclosures
 Baltimore, Cleveland, Philadelphia, Richmond, & St. Louis
Federal historic rehabilitation tax credits (RTCs):
 20% federal income tax credit
 National Register of Historic Places
 Income-producing (not owner-occupied)
 “the largest federal program specifically supporting historic
preservation” (NPS, 2015, p.1)
 $73 billion of investment for more than 40,000 preservation
projects
3
Literature Framework: Legacy
Cities
 Demolition:
 Blight remediation and neighborhood stabilization
 Not a question of if, but where and to what extent
 Unknowns: effect on surrounding property values, balance
protection of cultural and historic resources, long-term impacts
 Historic buildings:
 Core assets for neighborhood revitalization and stabilization
 Preservation is largely absent from both broader policy
discussions and the implemented approaches
 Recent revitalization trends, particularly downtowns
 Unknowns: how is preservation activity contributing to
revitalization?
4
Literature Framework: Urban
Preservation
 Analysis of program effects: standard economic impact
analyses at state and federal levels
 Historic preservation and gentrification
 “preservation efforts are more prone to cause displacement than
redevelopment projects involving new construction” because
“property values and rents begin to increase even before the real
estate experiences much improvement” (Werwath, 1998, p. 489)
 Limited empirical evidence
 All types of neighborhood investment?
 Weak market context of legacy cities?
5
Data
Historic Tax Credit data:
 Technical Preservation Services Division of the National Parks
Service
 Federal RTC projects from 1998-2007
Geolytics’ Neighborhood Change Database:
 Normalized census tract data to 2010 boundaries
 Pre-intervention: Census 2000
 Post-intervention: American Community Survey 2006-2010 (5year estimates)
6
Dependent Variables
Race/Ethnicity
Expected Sign
Share of Non-Hispanic White
+/-
Share of Non-Hispanic Black
+/-
Share of Hispanics
+/-
References
Bures, (2001); Deng, (2012); Mallach (2015);
Podagrosi & Vojnovic, (2008); Swanstrom &
Webber, (2014)
Bures, (2001); Hollander (2010); Mallach
(2015); Podagrosi & Vojnovic, (2008);
Swanstrom & Webber, (2014)
Bures, (2001); Swanstrom & Webber, (2014)
Socio-economic
Bachelor's Degree or greater
+
Professional/Technical workers
+
Median Household Income
+
Share of low-, middle-, and upper income persons
+/-
Poverty Rate
-
Households
+
Allison, (2005); Coulson & Leichenko,
(2004); Deng, (2012); Mallach (2015);
Montgomery, (2004)
Allison, (2005); Blakely, (2001); Coulson &
Leichenko, (2004); Filion, (2010); Florida,
(2002); Swanstrom & Webber, (2014)
Allison, (2005); Coulson & Leichenko,
(2004); Smith, (1998); Mallach (2015);
Swanstrom & Webber, (2014); Werwath,
(1998)
Allison, (2005); Coulson & Leichenko,
(2004); Smith, (1998); Mallach (2015);
Swanstrom & Webber, (2014); Werwath,
(1998)
Allison, (2005); Coulson & Leichenko,
(2004); Deng (2012); Hollander (2010);
Smith, (1998); Swanstrom & Webber,
(2014); Werwath, (1998)
Birch, (2005); NPS, (2014); Stern & Seifert,
(2010); Ryberg-Webster, (2014a, 2014b)
Housing
Median Housing Value
+
Median Rent
+
Allison, (2005); Bures, (2001); Coulson &
Leichenko, (2004); Deng (2012); Smith,
(1998); Swanstrom & Webber, (2014);
Werwath, (1998)
Allison, (2005); Bures, (2001); Coulson &
Leichenko, (2004); Deng (2012); Smith,
(1998); Swanstrom & Webber, (2014);
Werwath, (1998)
7
Methods
(1) what is the distribution of RTC activity across legacy city
neighborhood types and transition patterns?
Descriptive statistics, maps
8
Neigbborhood Neigbborhood
Type
Category
Black, Stressed,
&
Disadvantaged
Highly
Distressed
Collapsed
Urban Core
Highly
Distressed
Competitive &
Educated some
Distress
Stable
Declining &
Black
Highly
Distressed
Educated
Newcomers
Stable
Established &
Stable
Homeowners
Stable
Highly
Bifurcated:
Success &
Distress
Stable
White
Immigrants
Highly
Distressed
Description
low-value housing;low-income families with
little educational attainment; residents are
primarily black, unemployed, and living
below the poverty line
long-term renters, high rates of poverty and
public assistance among renters; high
vacancy rates, weak housing values, and
low educational attainment; large share of
the population is under 18, black, and
unemployed; higher-than average rents;
some white residents
high-value housing, well-educated singles,
and higher-than-average income; white
residents, rents higher than city and MSA
averages; high rates of poverty and public
assistance among renters
weak housing values, paired with low
educational attainment, high rates of public
assistance, among low-income black
families
high housing values with well-educated, highincome singles
high levels of homeownership, low levels of
poverty and public assistance, people that
have lived in their homes and the
neighborhood for an extended period of
time, low vacancy, and little multifamily
housing
high rates of poverty and public assistance,
transient renters; high-value housing
occupied by well-educated singles; lowincome renters
very old housing stock; primarily white
residents; some foreign born and Hispanic
residents
Share of
Distribution
Highest
Neighborhood
Concentration
Ratio
City Census Year
Gaining/Losing
Share, 1970-2010*
Most Common
Transition Pattern
19%
STL
(1.20)
2010
(1.21)
+7.7%
Collapsed Urban
Core; Declining &
Black (all years)
7%
BAL
(1.24)
1990
(1.10)
-0.1%
Black, Stressed &
Disadvantaged
Bifurcated: Success &
Distress;
Educated Newcomers
(all years)
8%
PHI
(1.32)
2010
(1.11)
+1.4%
19%
RVA
(1.19)
2000
(1.21)
+4.1%
Black, Stressed, &
Disadvantaged
-8.1%
Established & Stable
Homeowners (19702010); Declining &
Black (1970-2000)
13%
RVA
(1.49)
1970
(1.30)
16%
STL
(1.30)
1970
(1.14)
-2.9%
Educated
Newcomers(1970-80);
White Immigrant
(1980-2010)
7%
RVA
(2.11)
2000
(1.21)
+2.1%
Competitive &
Educated some
Distress
11%
PHI
(1.28)
1970
(1.20)
-4.2%
Black, Stressed &
Disadvantaged
*Based on the share of all 1970 neighborhoods in the neighborhood type compared to the share of 2010 neighborhood in the type.
9
Tracts
Highly Distressed
Black, Distressed, & Disadvantaged
Collapsed Urban Core
Declining and Black
White Immigrants
Subtotal
Stable
Competitive & Educated, some Distress
Educated Newcomers
Established and Stable Homeowners
Highly Bifurcated
Subtotal
Total Tracts with RTC Activity
Total Tracts
Total RTC Projects
Total RTC Investment
Baltimore
RTC
Projects
RTC
Tracts
28%
8%
18%
5%
59%
29%
5%
7%
0%
41%
11%
10%
14%
6%
41%
37%
2%
2%
17%
59%
RTC
Investment
Tracts
Cleveland
RTC
Projects
RTC
Tracts
17%
2%
3%
0%
22%
$ 50,306,545
27%
$ 13,313,764
8%
$ 32,476,179
19%
$
10%
15%
64%
32%
12%
16%
4%
64%
43%
1%
1%
34%
78%
21%
198
194
$620,770,963
$ 305,166,688
5%
$ 2,636,555
5%
$ 6,483,810
18%
$ 210,387,420
9%
85%
36%
12%
0%
0%
24%
36%
Richmond
RTC
Projects
RTC
Tracts Tracts
Highly Distressed
Black, Distressed, & Disadvantaged
Collapsed Urban Core
Declining and Black
White Immigrants
Subtotal
Stable
Competitive & Educated, some Distress
Educated Newcomers
Established and Stable Homeowners
Highly Bifurcated
Subtotal
Total Tracts with RTC Activity
Total Tracts
Total RTC Projects
Total RTC Investment
27%
3%
21%
2%
53%
36%
4%
11%
0%
50%
0%
14%
15%
18%
47%
0%
14%
0%
36%
50%
RTC
Investment
Tracts
RTC
Tracts
Philadelphia
RTC
Projects
10%
6%
4%
1%
21%
$ 55,023,634
$ 20,383,484
$ 6,599,639
$
34,201
13%
18%
5%
23%
13%
59%
8%
2%
22%
0%
31%
48%
0%
0%
31%
79%
15%
172
90
$625,580,807
$ 480,377,334
$
$
$ 63,162,515
87%
13%
9%
13%
6%
41%
43%
8%
2%
16%
69%
RTC
Investment
RTC
Tracts Tracts
38%
0%
1%
0%
39%
$ 179,310,603
$
4,196,399
$
4,778,212
$
30%
26%
15%
9%
15%
17%
12%
7%
9%
59%
50%
0%
5%
0%
56%
61%
42%
66
369
$624,961,613
$
$ 40,854,688
$
$ 395,821,709
70%
8%
24%
5%
0%
18%
3%
9%
24%
41%
50%
St. Louis
RTC
Projects
3%
23%
18%
4%
48%
RTC
Investment
2%
2%
16%
0%
20%
$ 13,028,492
$
882,086
$ 112,600,212
$
10%
64%
3%
1%
13%
80%
14%
375
188
$1,299,874,545
$ 677,656,517
$ 35,045,262
$
6,688,452
$ 453,973,524
90%
RTC
Investment
$
$
$
$
20,650,408
148,329,832
26,205,470
11,902,295
16%
26% $ 782,145,803
0% $
0% $
391,795
26% $ 316,499,079
52%
84%
32%
106
388
$1,306,124,683
10
Methods
(1) what is the distribution of RTC activity across legacy city
neighborhood types and transition patterns?
Descriptive statistics, maps
(2) what is the relationship between historic tax credit activity
and neighborhood racial, socioeconomic, and housing
characteristics?
Difference-in-difference regression model
 understand whether rehabilitation activity accelerated
neighborhood changes or if they continued at the same rate
 baseline differences between RTC and non-RTC tracts are accounted
for
12
Non-Hispanic Whites
Non-Hispanic Blacks
Hispanic
Households
Bachelor's Degree or Greater
Professional/Technical Workers
Poverty
Median Household Income
Very low-income (30% of city MHI or less)
Low-income (31-50% of city MHI)
Moderate income (51-80% of city MHI)
Middle income (81-120% of city MHI)
Upper income (120% of city MHI or greater)
Share of Housing Units 50 years or older
Median Rent
Median Housing Value
RTC Non-RTC RTC Non-RTC
Tracts
Tracts
Tracts
Tracts
2000
2010
43.4%
37.8% 44.1%
33.3%
46.1%
52.2% 42.2%
54.0%
4.3%
5.6%
5.9%
7.6%
1,341
1,386
1329
1308
27.0%
15.9% 36.9%
19.2%
24.2%
17.8% 28.0%
19.0%
28.4%
23.8% 28.7%
25.7%
$36,009 $39,507 $38,319 $36,322
23.5%
19.4% 17.9%
15.8%
9.2%
8.8% 12.6%
13.9%
12.9%
13.7% 11.0%
12.4%
13.9%
14.5% 15.2%
16.4%
40.8%
43.6% 43.3%
41.2%
63.8%
57.6% 73.5%
74.5%
$680
$656
$847
$774
$121,406 $88,471 $211,250 $137,811
13
Statistical Model
Yit=α+β1(RTCi)+β2(Postt)+β3(RTC ⋅ Post)it+ β4(City)+ϵit
where:
α = intercept
β = coefficient
RTC = dummy variable for neighborhoods with historic tax credit
investment between 1998 and 2007
Post = dummy variable for the post treatment period of 2010
City= Citywide fixed effects
Y = Value (in log form) for revitalization indicators in census tract i in
year t
ϵ = a random error term with the usual assumed statistical properties
14
Model 1
Model 2
Model 3
Model 4
Model 5
Model 6
All RTC
tracts
Above
median RTC
tracts
Below
median RTC
tracts
Stable (2000)
RTC tracts
All RTC tracts
Above
median RTC
tracts
178
89
89
95
178
89
All non-RTC
tracts
All non-RTC
tracts and
below
median RTC
tracts
All non-RTC
tracts and
above
median RTC
tracts
All non-RTC
tracts and all
non-Stable
RTC tracts
Matched
comparison
tracts
Matched
comparison
tracts and
below
median RTC
tracts
n
742
831
831
835
152
241
Total
observations
920
920
920
920
330
330
Treatment
Group
n
Comparison
Group
15
Model 1
Dependent Variables
Model 6
Above
RTC and
Median
Above
Below
Stable RTC
Matched
RTC and
Full Model Median Median RTC
Tracts
Comp.
Matched
RTC Tracts
Tracts
(2000)
Model
Comparison
Group
Race/Ethnicity
Percent Hispanic
-0.050
Percent Non-Hispanic Black
-0.069 **
Percent Non-Hispanic White
-0.001
Socio-economic
Income Groups
Very-low income
-0.001
Low-income
-0.065 **
Moderate-income
-0.056 *
Middle-income
0.002
Upper-income
0.060 *
Median Houshold Income
0.078 **
Percent Bachelor's or more
0.040
Percent Professional/Technical workers 0.059 *
Poverty Rate
-0.052
Housing
Households
0.038
Median Housing Value
0.086 ***
Median Rent
0.024
Observations
920
All Dependent variables are in natural log form.
***p<0.01, **p<0.05, *p<0.1
Model 2
Model 3
Model 4
Model 5
-0.053
-0.060 *
0.011
-0.012
-0.030
-0.013
-0.056 *
-0.062 *
-0.017
0.280 -0.078
0.317 -0.066
0.292 0.063
0.033
-0.046
-0.020
0.065 **
0.019
0.066 **
0.029
0.034
-0.030
-0.034
-0.038
-0.052 *
-0.062 *
0.058 *
0.036
0.023
0.043
-0.038
0.013
-0.087 ***
-0.072 **
0.012
0.024
0.083 ***
0.012
0.001
-0.031
0.414
0.269
0.602
0.826
0.631
0.331
0.292
0.181
0.781
0.066 **
0.060 **
0.026
920
-0.017
0.052 *
0.005
920
0.051
0.092 ***
0.037
920
0.088
-0.044
0.012
0.127 **
-0.024
0.060
0.038
0.033
-0.001
0.304 0.117 **
0.117 0.053
0.367 0.047
330
330
20
Key Findings & Policy
Implications
 RTC incentivizes investment in weak market conditions
 RTC projects are contributing to important LC neighborhood
revitalization goals- attracting more upper-income earners
and professional/technical workers as well as increased
median household income and median housing values
 Negative change effects- loss of non-Hispanic black and lowand moderate-income residents. Most evident in Stable
neighborhoods
 These differences vary, however, based on the scale of RTC
investment, the status of the neighborhood prior to RTC
activity, and the composition group of comparison tracts
21
Key Findings & Policy
Implications
Legacy city policy: Including RTCs in strategic targeting
frameworks for neighborhood stabilization
Federal/state policy: Requiring greater coordination with local
planning efforts; training for real estate developers on
program use
Evidence that RTC activity does play a significant role in the
complex neighborhood change processes of legacy cities
22
Legacy City Revitalization:
The Role of Federal Historic
Tax Credit Projects
Kelly L. Kinahan, AICP
Doctoral Candidate
Levin College Research Conference
August 20, 2015