ArticlePDF Available

Abstract and Figures

The objective of this paper was to examine the profit levels, energy use and environmental impacts of two residential development scenarios in a watershed in the Philadelphia region under two zoning assumptions. The two scenarios were based on economic suitability and environmental suitability. A key question was whether these occurred together in the Pennypack Creek Watershed. Suitability analyses in ArcGIS using criteria for profit and for local sustainability parsed out two sets of developable areas. Buildouts to satisfy 2035 population projections in these areas using CommunityViz software were based on actual municipal zoning ordinances. In a unified zoning scheme created by the authors, a density-adjusted number of housing units are placed watershed-wide without municipal restrictions. Profit data for buildings in each zip code were used to compute a Weighted Profit per Square Meter. Household units were associated with a particular type of automobile and average Vehicle Kilometers Traveled in the relevant census tracts. The GREET program was used to compute energy use, air pollution emissions and greenhouse gas emissions. A Weighted Water Quality Index and Index of Biological Integrity were used to assess water-related impacts based on recent monitoring data supplied by the Philadelphia Water Department. It was no surprise that ECON-UNI and ECON-MUNI generated higher profit than ENV-MUNI and ENV-UNI. ENV-UNI had lower energy use and environmental impacts than all others. That ECON-MUNI had the second lowest energy use and environmental impacts, and the highest water quality, was unexpected. Some policy proposals and conclusions end the paper.
Content may be subject to copyright.
This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
http://www.elsevier.com/authorsrights
Author's personal copy
Landscape
and
Urban
Planning
125
(2014)
188–206
Contents
lists
available
at
ScienceDirect
Landscape
and
Urban
Planning
j
o
ur
na
l
ho
me
pag
e:
www.elsevier.com/locate/landurbplan
Research
Paper
Housing
location
in
a
Philadelphia
metro
watershed:
Can
profitable
be
green?
John
A.
Sorrentinoa,,
Mahbubur
R.
Meenarb,
Alice
J.
Lambertc,
Donald
T.
Wargod
aDepartment
of
Economics,
Center
for
Sustainable
Communities,
Temple
University,
580
Meetinghouse
Road,
Ambler,
PA
19002,
United
States
bCenter
for
Sustainable
Communities,
Department
of
Community
&
Regional
Planning,
Temple
University,
United
States
cFormerly
at
the
Bucks
County
(PA)
Planning
Commission,
United
States
dDepartment
of
Economics,
Temple
University,
United
States
h
i
g
h
l
i
g
h
t
s
Housing
is
placed
in
a
Philadelphia-area
watershed
according
to
profitability
and
sustainability
under
two
different
zoning
schemes.
Profit,
energy
use,
air
pollution,
greenhouse
gases,
water
quality
and
biological
integrity
are
assessed
for
each
scenario-zoning
combination
and
compared.
Implications
of
the
results
are
used
for
policy
recommendations.
a
r
t
i
c
l
e
i
n
f
o
Article
history:
Received
12
March
2013
Received
in
revised
form
31
January
2014
Accepted
1
February
2014
Keywords:
Housing
Location
Profitability
Sustainability
Watershed
planning
a
b
s
t
r
a
c
t
The
objective
of
this
paper
was
to
examine
the
profit
levels,
energy
use
and
environmental
impacts
of
two
residential
development
scenarios
in
a
watershed
in
the
Philadelphia
region
under
two
zoning
assumptions.
The
two
scenarios
were
based
on
economic
suitability
and
environmental
suitability.
A
key
question
was
whether
these
occurred
together
in
the
Pennypack
Creek
Watershed.
Suitability
analyses
in
ArcGIS
using
criteria
for
profit
and
for
local
sustainability
parsed
out
two
sets
of
developable
areas.
Buildouts
to
satisfy
2035
population
projections
in
these
areas
using
CommunityViz
software
were
based
on
actual
municipal
zoning
ordinances.
In
a
unified
zoning
scheme
created
by
the
authors,
a
density-
adjusted
number
of
housing
units
are
placed
watershed-wide
without
municipal
restrictions.
Profit
data
for
buildings
in
each
zip
code
were
used
to
compute
a
Weighted
Profit
per
Square
Meter.
Household
units
were
associated
with
a
particular
type
of
automobile
and
average
Vehicle
Kilometers
Traveled
in
the
relevant
census
tracts.
The
GREET
program
was
used
to
compute
energy
use,
air
pollution
emissions
and
greenhouse
gas
emissions.
A
Weighted
Water
Quality
Index
and
Index
of
Biological
Integrity
were
used
to
assess
water-related
impacts
based
on
recent
monitoring
data
supplied
by
the
Philadelphia
Water
Department.
It
was
no
surprise
that
ECON-UNI
and
ECON-MUNI
generated
higher
profit
than
ENV-MUNI
and
ENV-UNI.
ENV-UNI
had
lower
energy
use
and
environmental
impacts
than
all
others.
That
ECON-
MUNI
had
the
second
lowest
energy
use
and
environmental
impacts,
and
the
highest
water
quality,
was
unexpected.
Some
policy
proposals
and
conclusions
end
the
paper.
©
2014
Elsevier
B.V.
All
rights
reserved.
1.
Introduction
Land
use
change
is
considered
by
some
analysts
to
be
the
most
important
human-induced
environmental
transformation
(Wolman
&
Fournier,
1987).
Suburbanization
in
the
US
has
been
a
predominant
form
of
land
use
change
that
has
become
increasingly
automobile-dependent
and
has
lost
ties
with
central
cities.
New
Corresponding
author.
Tel.:
+1
267
468
8370.
E-mail
addresses:
jsorrent@temple.edu
(J.A.
Sorrentino),
meenar@temple.edu
(M.R.
Meenar),
alicejlambert@gmail.com
(A.J.
Lambert),
docwargo@temple.edu
(D.T.
Wargo).
development
has
extended
into
prime
agricultural
and
wooded
lands,
and
other
environmentally
sensitive
areas
(Batty
&
Xie,
2005;
Cervero,
2003;
Cullingworth
&
Caves,
2003;
Galster
et
al.,
2001;
Walker,
2004).
Environmental
degradation
at
the
suburban
fringe
includes
an
increase
in
the
release
of
greenhouse
gases,
degradation
of
lakes
and
streams
and
loss
of
biodiversity
(Walker,
2004).
Such
growth
is
also
thought
to
cause
many
socioeconomic
ills
(Adams,
Bartelt,
Elesh,
&
Goldstein,
2008).
That
some
of
this
can
be
avoided
is
the
thrust
of
the
present
work.
Watershed
planning,
conducted
within
watershed
boundaries,
and
land
use
planning,
usually
focused
on
municipal
boundaries,
are
often
two
different
planning
processes.
A
watershed-based
planning
approach
is
a
“coordinating
framework
for
environmental
0169-2046/$
see
front
matter
©
2014
Elsevier
B.V.
All
rights
reserved.
http://dx.doi.org/10.1016/j.landurbplan.2014.02.005
Author's personal copy
J.A.
Sorrentino
et
al.
/
Landscape
and
Urban
Planning
125
(2014)
188–206
189
management
that
focuses
public
and
private
sector
efforts
to
address
the
highest
priority
problems
within
hydrologically-
defined
geographic
areas”
(Browner,
1996).
A
multi-municipal
program,
with
governments
and
watershed
associations
working
together
to
establish
and
apply
regulations,
offers
a
comprehen-
sive
way
to
manage
the
natural
and
built
environment.
This
study
examines
such
an
approach,
despite
the
difficulties
that
exist
in
states
such
as
Pennsylvania
where
fundamental
land
use
decisions
are
made
by
municipalities
(Hershberg,
2003;
Kenney,
1997).
Although
the
integration
of
local
planning
processes
is
increas-
ingly
being
seen
as
a
critical
public
policy
challenge
(Carter,
Kreutzwiser,
&
de
Loë,
2005;
Mitchell,
2005;
Plummer,
de
Grosbois,
de
Loe,
&
Velaniskis,
2011),
few
studies
have
analyzed
the
process
of
using
a
watershed-based
planning
approach
to
locate
residen-
tial
development.
Steiner,
McSherry,
and
Cohen
(2000)
performed
a
suitability
analysis
for
four
land
uses,
including
housing
devel-
opment,
within
a
large,
rural
watershed
in
the
western
US.
The
authors
found
areas
suitable
for
housing,
but
did
not
focus
on
the
impacts
of
development
in
the
areas
they
found
suitable.
Tang,
Engel,
Pijanowski,
and
Lim
(2005)
also
studied
a
large
watershed
in
an
already
urbanized
and
industrialized
area
in
the
US
Mid-
west.
They
concentrated
on
the
environmental
impacts
of
previous
development,
but
made
very
general
recommendations
for
future
watershed
decision
making.
Brown
(2000)
proposed
using
housing
density
as
a
water
quality
indicator
in
another
large
US
Midwest
watershed,
but
did
not
estimate
the
impacts
of
new
residential
development.
The
present
study
uses
suitability
analyses
to
locate
areas
for
housing
buildouts
in
a
small
watershed
in
the
eastern
US,
examines
the
energy
use
and
environmental
impacts
of
four
scenario-zoning
combinations
and
makes
some
policy
recommen-
dations
based
on
the
results.
By
comparing
economic
impacts
and
energy/environmental
impacts
of
housing
location
schemes,
the
present
work
aims
to
determine
whether
profitable
and
green
development
can
happen
together.
It
is
thought
that
this
approach
can
add
a
new
dimension
to
planning
at
the
watershed
level.
Using
a
mix
of
regulations
and
incentives,
municipalities
can
implement
watershed-level
plans
within
their
boundaries
by
channeling
development
in
agreed-upon
directions.
As
Daniels
and
Daniels
(2003,
p.
3)
write,
“Land
use
planning
.
.
.
needs
to
emphasize
redevelopment
and
infill
within
cities
and
sub-
urbs,
maintaining
quality
built
environments,
preserving
valuable
natural
areas
and
working
landscapes,
and
carefully
designing
greenfield
developments.”
While
the
work
for
this
paper
is
focused
on
the
former
objectives,
there
was
no
attempt
to
implement
design
at
the
subdivision
level.
The
next
section
describes
the
study
area
and
discusses
how
the
data
were
used
to
generate
suitable
areas
for
development,
to
locate
buildings
in
those
areas,
and
to
measure
the
impacts
of
such
loca-
tion.
The
third
section
presents
and
discusses
the
empirical
results.
Policy
implications
and
conclusions
follow.
2.
Data
and
methods
The
logical
sequence
by
which
the
analysis
proceeds
is
as
follows:
After
setting
the
geo-political
context,
the
elements
of
residential
development
that
comprise
economic
suitability
and
environmental
suitability
are
discussed.
These
elements
were
operationalized
in
Geographic
Information
Systems
(GIS)
software,
ArcGIS
from
the
Environmental
Science
and
Research
Institute
(ESRI),
to
create
suitable
areas
for
the
economically
suitable
and
environmentally
suitable
scenarios.
CommunityViz
was
used
to
perform
buildouts
for
the
two
scenarios
under
actual
and
“unified”
zoning.
The
means
are
described
by
which
profit,
energy
use,
air
pollution,
greenhouse
gases,
water
quality
and
biological
integrity
resulting
from
the
housing
location
patterns
were
computed.
2.1.
The
study
area
Located
in
the
Delaware
River
Basin
(Kaufman,
Homsey,
Belden,
&
Ritter-Sanchez,
2011),
the
90
km2Pennypack
Creek
Watershed
(PCW)
crosses
through
12
municipalities
within
three
counties
in
southeastern
Pennsylvania
on
its
way
to
the
Delaware
River.
About
328,000
people
live
within
its
boundaries
according
to
the
2010
US
Census.
The
Creek
is
a
public
amenity,
contributing
to
the
public
water
supply
and
used
extensively
for
recreation.
From
1950
to
1980
the
watershed
outside
the
City
of
Philadel-
phia
limits
experienced
significant
development.
As
of
2012,
single-family
homes
made
up
38%
and
multi-family
homes
12%
of
the
watershed
(Fromuth,
2012).
Over
the
last
decade,
the
PCW
has
been
estimated
to
be
about
30%
impervious
(PWD,
2003,
2009).
The
PCW
is
serviced
by
public
water
suppliers,
and
it
is
estimated
that
stormwater
collection
systems
are
installed
in
65%
of
it.
Many
of
these
systems
were
designed
only
to
collect
runoff
and
discharge
it
offsite.
This
has
resulted
in
increased
flooding,
destabilized
stream
channels,
severe
erosion
and
sedimentation.
A
municipal
treatment
plant
contributes
a
large
portion
of
the
base
flow
in
the
Creek,
resulting
in
additional
nutrients.
In
the
midst
of
the
urbanization
within
the
PCW,
there
has
been
significant
effort
by
conservation
groups
and
government
agencies
to
preserve
land
as
open
space
(Fig.
1)
(DVRPC,
2010),
Figs.
2
and
3
are
intended
to
give
the
reader
spatial
views
of
the
watershed.
Local
political
independence
is
evident
in
Pennsylvania,
and
plays
into
land
use
considerations
significantly.
In
the
mid-1970s,
the
state
legislature
adopted
the
Home
Rule
Charter
stating,
“A
municipality
.
.
.
may
exercise
any
function
not
denied
by
this
Con-
stitution,
by
its
home
rule
charter
or
by
the
General
Assembly
at
any
time”
(PA
DCED,
2003).
The
Pennsylvania
Municipalities
Planning
Code
Act
of
1968
permits
municipalities
to
make
land
use
decisions.
In
the
PCW,
there
are
12
different
zoning
codes
regulating
land
use
and
(in
most
codes)
dwelling
density.
Though
most
municipal-
ities
have
their
own
protective
measures
to
conserve
floodplains
and
preserve
open
space,
the
end
result
is
often
a
disconnected
set
of
preserved
parcels
throughout
the
watershed.
The
potential
synergies
of
joint
measures
across
municipalities
are
thereby
not
realized.
Section
303(d)
of
the
US
Clean
Water
Act
describes
the
nature
of
an
impaired
stream.
The
Act
requires
that:
“The
states
identify
all
waters
where
required
pollution
controls
are
not
sufficient
to
attain
or
maintain
applicable
water
quality
standards,
and
estab-
lish
priorities
for
development
of
Total
Maximum
Daily
Loads
based
on
the
severity
of
the
pollution
and
the
sensitivity
of
the
uses
to
be
made
of
the
waters,
among
other
factors”
(US
EPA,
2014).
The
Pennypack
Creek
is
listed
by
the
Pennsylvania
Department
of
Environmental
Protection
(PA
DEP)
as
an
impaired
stream
for
two
designated
uses,
aquatic
life
and
recreation.
Total
Maximum
Daily
Loads
have
been
assigned
by
the
PA
DEP
for
a
number
of
pollutants.
They
include
quantifiable
reductions
for
trichoroethylene,
fecal
col-
iform,
dissolved
oxygen-consuming
pollutants,
phosphorous,
and
suspended
solids.
The
responsibility
to
implement
the
reductions
lies
with
either
wastewater
treatment
operators
or
local
municipal-
ities.
Both
must
apply
to
the
US
Environmental
Protection
Agency
(EPA)
for
a
National
Pollution
Discharge
Elimination
System
Permit
for
permission
to
discharge
pollutants.
Wastewater
operators
and
municipalities
have
to
pay
for
and
administer
additional
controls
to
reduce
the
amount
of
pollutants
in
their
wastewater
or
storm
sewer
systems.
These
efforts
are
not
done
in
coordination,
reducing
pollutants
in
a
piecemeal
fashion.
In
1978,
the
Pennsylvania
legislature
recognized
that
flooding
and
water
quality
problems
existed
because
regulations
were
not
standardized
throughout
watersheds.
It
enacted
the
Stormwater
Management
Act
(SMA),
requiring
the
PA
DEP
to
designate
water-
sheds,
and
establish
guidelines
for
the
preparation
of
Stormwater
Author's personal copy
190
J.A.
Sorrentino
et
al.
/
Landscape
and
Urban
Planning
125
(2014)
188–206
Fig.
1.
Land
uses
in
the
PCW.
Management
Plans
(SMP).
These
plans
encourage
comprehensive
planning
and
management
of
stormwater
by
requiring
counties
to
prepare
the
plans
and
to
develop
ordinance
language
for
munici-
palities.
Within
the
SMP
planning
process,
counties
are
responsible
for
establishing
a
Watershed
Planning
Advisory
Committee.
Each
municipality
is
required
to
adopt
stormwater
management
ordi-
nances
that
are
consistent
with
the
standards
and
criteria
of
the
plan.
Although
these
stormwater
runoff
performance
standards
are
required
to
be
consistent
throughout
the
watershed,
municipalities
still
have
the
freedom
to
structure
and
enforce
ordinances
as
they
see
fit.
The
only
legislation
which
requires
multi-municipal
plan-
ning
for
watersheds
is
the
SMA.
Municipal
officials
have
limited
resources.
They
are
generally
not
eager
to
commit
to
voluntary
activities.
Often
municipal
officials
want
to
know
how
combined
planning
efforts
will
benefit
their
jurisdictions
directly
before
they
commit.
Multi-municipal
planning
for
watersheds
tends
to
be
most
successful
when
incentives
are
offered,
or
local
government
needs
besides
watershed
conservation
are
met
(Barletta,
Dahme,
&
Maimone,
2007).
The
present
study
developed
a
unified
zoning
scheme
for
the
PCW
to
investigate
whether
consistent
classification
would
result
in
more
profit
and/or
less
environmental
degradation.
To
arrive
at
the
unified
residential
zoning
scheme,
the
residential
zoning
cate-
gories
and
codes
of
the
suburban
municipalities
with
portions
that
lie
in
the
PCW,
and
those
of
the
City
of
Philadelphia,
were
gathered
and
recorded.
Using
key
words
from
the
actual
categories,
the
eight
generic
categories
listed
vertically
in
Table
1
were
created.
The
key
zoning
parameters
used
in
the
CommunityViz
buildouts
are
listed
horizontally
in
Table
1.
The
homogenized
values
were
derived
by
different
methods
based
on
the
amount
of
variation
in
the
actual
codes.
Dwelling
Units
per
Building
and
Floors
per
Building
were
chosen
by
straight
frequency.
A
spatially-weighted
average
was
cal-
culated
for
Building
Separation,
Building
Density
and
Road
Setback.
The
weights
contained
the
percentages
of
the
relevant
municipal-
ities
that
were
zoned
with
each
code,
and
the
percentages
of
the
PCW
land
area
that
the
portions
of
these
municipalities
occupied.
2.2.
Residential
development:
economic
suitability
In
general,
housing
demand
is
largely
a
function
of
the
price
of
housing
as
altered
by
the
fact
that
mortgage
interest
is
federal
tax-
deductible
in
the
US.
The
effective
price
of
a
housing
unit
(H)
is
the
after-tax
user-cost
(ATUCH)
as
shown
in
Eq.
(1).
ATUCH=
PH
(L
MC
+
M
+
RET)
+
U
PH
(L
MR
+
RET)
ATR
(1)
The
variables
are
designated
as
follows:
PH
the
price
of
a
housing
unit
L
loan-to-value
ratio
Author's personal copy
J.A.
Sorrentino
et
al.
/
Landscape
and
Urban
Planning
125
(2014)
188–206
191
Fig.
2.
Counties
and
municipalities
in
the
PCW.
MC
mortgage
constant
that
multiplies
the
loan
amount
to
get
the
annual
principal
plus
interest
payment
M
maintenance
costs
as
a
percent
of
PH
U
utility
costs
MR
mortgage
interest
rate
RET
real
estate
taxes
as
a
percent
of
PH
ATR
average
income
tax
rate
Studies
of
US
housing
markets
broadly
agree
on
the
price
of
housing
as
ATUC
(e.g.,
Gill
&
Haurin,
1991;
Gillen,
2005).
These
studies
are
also
in
agreement
on
the
percentages
used
for
main-
tenance
cost,
real
estate
taxes
and
average
income
tax
rate.
The
loan-to-value
ratio
is
the
amount
of
the
purchase
price
that
a
bank
will
finance.
The
mortgage
constant
is
a
statistic
that,
when
multiplied
by
the
loan
amount,
yields
the
annual
payment
of
prin-
cipal
and
interest
on
a
home
loan.
Utility
costs
have
been
estimated
from
actual
values
for
a
sample
of
75
homes
randomly
chosen
within
the
PCW
zip
codes
on
www.trulia.com
(Trulia.com,
2011).
Finally,
deducted
from
the
ATUC
is
tax
savings.
Current
and
anticipated
household
income/wealth,
and
the
hedonic
characteristics
of
a
particular
home
and
its
neighborhood,
are
the
other
major
determinants
of
housing
choice.
Since
the
object
of
the
present
study
is
location,
the
characteristics
of
a
buyer’s
target
house
are
not
directly
relevant.
Galster
et
al.
(2001)
posit
a
taxonomy
of
neighborhood
characteristics
that
depict
an
inter-
active
relationship
between
the
individual
households
and
the
natural
(environmental)
and
built
(structural,
infrastructural,
prox-
imity)
environments,
the
community
(demographic,
class
status,
Table
1
Unified
zoning
classifications
and
requirements.
District
Zoning
code
Dwelling
units
per
building
Floors
per
building
Building
separation
(m)
Building
density
(per
ha)
Road
setback
(m)
Low
L
1
3
7.62
2.69
15.24
Low–medium
LM
1
3
7.62
5.66
12.19
Medium
M
1
3
4.57
10.13
9.14
High
H
2
3
4.72
14.43
9.14
Multi-family
MF
8
3
7.62
19.62
9.14
High
urban
HU
1
3
2.44
43.32
3.51
High
urban
duplex
HUD
2
3
2.44
63.01
3.84
Multi-family
urban
MFU
25
10
0
7.17
0
Author's personal copy
192
J.A.
Sorrentino
et
al.
/
Landscape
and
Urban
Planning
125
(2014)
188–206
Fig.
3.
Zip
codes
in
the
PCW.
social-interactive,
sentimental),
and
institutions
(tax/public
service
package,
political)
that
exist
at
the
neighborhood
scale
and
above.
The
relationship
between
the
economic
variables
and
the
char-
acteristics
of
a
particular
home
is
easier
to
track
than
their
relationship
to
the
local
neighborhood.
Despite
the
added
difficul-
ties,
hedonic
pricing
has
also
been
applied
to
neighborhood
and
environmental
characteristics
(e.g.,
Albouy
&
Lue,
2011;
Gibbons
&
Machin,
2008;
Jim
&
Chen,
2006;
Poor,
Pessagno,
&
Paul,
2007;
The
Reinvestment
Fund,
2009).
Infrastructure,
tax/public
service
package
and
proximity
to
local
institutions
seem
immediately
relevant
to
location
choice
with
respect
to
impacts
on
the
cost-of-
living.
Available
transportation
modes
and
distances
to
frequent
destinations
directly
impact
the
household
budget.
Home
and
neighborhood
security
directly
improve
quality
of
life.
Good
pub-
lic
schools
enable
loftier
prospects
for
higher
education
and
career
paths
for
children,
without
the
higher
expense
of
private
educa-
tion.
Convenient
and
pleasant
open
space
offsets
the
expense
of
obtaining
similar
amenities
in
the
private
sector.
As
a
local
exam-
ple
of
the
importance
of
these
variables,
The
Reinvestment
Fund
(2009)
found
in
their
analysis
of
the
City
of
Philadelphia
that
a
1%
increase
in
a
Structural
Decline
Score
(based
on
building
vacancies
and
demolitions,
lien
sales
for
unpaid
taxes,
and
water
shut-offs
due
to
unpaid
bills)
in
a
neighborhood
reduced
the
sales
price
of
homes
in
that
area
by
$16.15
per
m2;
a
1%
increase
in
the
Crime
Score
(based
on
higher
levels
of
drug
use
and
possession,
a
range
of
aggravated
and
weapons
related
offenses,
and
arson)
reduced
the
sales
price
of
homes
in
that
block
by
$10.76
per
m2;
and
each
1%
increase
in
the
school
catchment
area
Pennsylvania
System
of
School
Assessment
%-Proficiency
Score
(based
on
state-wide
math-
ematics
and
reading
tests)
increased
the
sales
price
of
homes
in
that
block
by
$5.60
per
m2.
Households
juxtapose
information
on
the
home-specific
and
neighborhood
characteristics
with
their
economic
variables
and
attitudes
toward
the
environment,
undertake
tradeoffs,
and
make
decisions.
Though
there
is
evidence
to
the
contrary
(e.g.,
Levine
&
Inam,
2004),
it
is
assumed
in
what
follows
that
housing
demand
drives
housing
supply
(Glaeser,
Gyourko,
&
Saiz,
2008).
This
assumption
is
reflected
by
the
presence
of
demand-side
criteria
in
the
economic
suitability
analysis.
Housing
supply
is
determined
by
the
existing
stock
minus
the
removal
of
deteriorated
units
plus
the
addition
of
newly-built
units.
However,
only
the
sales
of
those
houses
in
the
stock
in
a
specific
geographic
area
determine
the
market
price
in
that
area.
The
devel-
oper
decides
whether
s/he
can
make
a
profit
building
and
selling
a
house
at
that
price.
S/he
simply
makes
the
product,
sells
it
in
a
one-off
transaction
and
reaps
the
profit
as
the
difference
between
the
sales
price
and
the
cost
of
production.
Specifically,
developer
profit
for
a
single
unit
is
the
settlement
price
minus
the
salesper-
son’s
commission,
the
settlement
costs,
land
costs,
construction
costs,
infrastructure
costs,
“soft”
costs
and
overhead.
The
latter
five
Author's personal copy
J.A.
Sorrentino
et
al.
/
Landscape
and
Urban
Planning
125
(2014)
188–206
193
Table
2
Housing
unit
projections.*
Municipality
2010
census
household
size
(persons
per
occupied
housing
unit)
Vacancy
rate
(2%
of
occupancy
rate)
Occupancy
rate
+
vacancy
rate
(persons
per
unit)
Change
in
population
in
watershed
2010–2035
2035
housing
units
needed
Upper
Southampton
Township
2.569
0.051
2.620
410
156
Warminster
Township
2.539
0.051
2.590
2465
952
Abington
Township
2.587
0.052
2.639
457
173
Bryn
Athyn
Borough
3.213
0.064
3.277
58
18
Hatboro
Borough
2.419
0.048
2.467
283
115
Horsham
Township 2.732 0.055 2.787 1709
613
Jenkintown
Borough 2.196 0.044 2.240
4
2
Rockledge
Borough
2.420
0.048
2.468
31
13
Upper
Dublin
Township
2.721
0.054
2.775
164
59
Upper
Moreland
Township
2.417
0.048
2.465
1492
605
Total
2706
*Lower
Moreland
and
Philadelphia
were
excluded
due
to
non-positive
population
projections.
costs
sum
to
the
overhead-adjusted
total
cost
of
providing
the
unit.
The
soft
costs
include
items
such
as
appraisal,
permit
application,
review
and
inspection
fees.
Overhead
is
generally
computed
as
a
standard
percentage
of
the
sum
of
the
other
costs.
Well-accepted
formulas
and
sources
for
housing
supply
and
construction
costs
have
been
used.
Whereas
most
models
estimate
a
national
or
metro-region-wide
housing
supply
(Blackley,
1999;
Glaeser,
Gyourko,
&
Saiz,
2008;
Kinsey,
1992;
Mayer
&
Somerville,
2000;
Somerville,
1999),
prices
and
costs
by
zip
codes
within
the
PCW
have
been
examined
for
this
analysis.
This
disaggregated
approach
is
thought
to
be
somewhat
unique
in
the
housing
literature.
2.3.
Residential
development:
environmental
suitability
The
goals
of
simultaneously
seeking
positive
economic,
envi-
ronmental
and
social
justice
outcomes
from
development
have
been
pursued
at
the
international,
national
and
local
levels
for
decades.
From
Maclaren
(1996),
Berke
and
Manta-Conroy
(2000)
and
Wheeler
(2000),
the
following
goals
of
local
sustainability
were
chosen:
(1)
protection
of
the
natural
environment,
(2)
minimal
use
of
nonrenewable
resources
and
(3)
responsible
regional
cooper-
ation
by
local
governments.
In
this
paper,
goals
(1)
and
(2)
are
directly
sought
via
criteria
that
are
imposed
in
the
environmen-
tal
suitability
analysis
below.
Regional
cooperation
in
our
study
means
multi-municipal
collaboration
within
the
watershed
region.
The
unified
zoning
scheme
described
above
represents
such
col-
laboration
in
the
analysis.
The
PCW
municipalities
are
thought
to
comply
with
the
hypothetical
unified
scheme.
It
is
incorporated
in
two
suitability
analyses
and
two
buildouts
below.
Some
policy
rec-
ommendations
based
on
the
housing
location
are
also
made
in
the
sequel
to
encourage
such
regional
cooperation.
The
notion
of
environmentally
sustainable
land
development
is
not
alien
to
the
land
development
industry.
Of
particular
inter-
est
to
this
study,
the
US
Green
Building
Council
(2011)
promotes
standards
and
sponsors
education
and
training
programs
to
foster
sustainability.1A
rather
spirited
enunciation
of
the
need
for
devel-
opment
to
involve
people,
the
planet,
and
profit
is
given
as
part
of
the
Sustainable
Land
Development
Initiative
(SLDI):
Today’s
reality
is
that
the
‘people’
are
driving
demand
for
prac-
tices
that
steward
the
‘planet.’
There
are
many
sound
land
development
practices
.
.
.
which
not
only
reduce
development
costs,
but
add
sales
premium
potential
and
provide
additional
1The
Council
has
headquarters
in
Washington,
DC,
and
a
Certification
Institute
in
Philadelphia.
Its
mission
is
“To
transform
the
way
buildings
and
communities
are
designed,
built
and
operated,
enabling
an
environmentally
and
socially
responsible,
healthy,
and
prosperous
environment
that
improves
the
quality
of
life.”
environmental
benefits.
Such
practices
require
a
more
holis-
tic
and
sophisticated
approach
than
that
which
is
typically
employed
today,
but
nevertheless,
offer
significant
cost
saving
opportunities
(TriplePundit,
2010).
An
implication
of
the
three-pronged
approach
is
that
the
environ-
ment
and
less-advantaged
citizens
need
not
suffer
when
there
are
positive
economic
outcomes
for
developers.
This
scenario
will
con-
centrate
only
on
environmental
criteria.
2.4.
Two
hypothetical
scenarios
and
their
impacts
The
Delaware
Regional
Planning
Commission
(DVRPC)
is
the
metropolitan
planning
organization
for
the
nine-county
Philadel-
phia
region.
As
part
of
its
portfolio
of
projects,
it
makes
projections
about
population
and
employment
growth
for
counties
and
munic-
ipalities
within
the
region
with
their
collaboration.
The
PCW
contains
all
or
part
of
12
municipalities
in
Bucks,
Montgomery
and
Philadelphia
Counties.
The
2035
projected
percentage
popu-
lation
growth
for
Bucks
County
is
21%.
It
is
15%
for
Montgomery
County,
and
0%
for
Philadelphia
County
(DVRPC,
2007).
The
actual-
zoning
scenarios
presented
below
hypothetically
placed
in
the
zip
codes
the
buildings
required
to
house
the
DVRPC
population
fore-
casts,
given
zip
code-specific
household
sizes.
These
projections
appear
in
Table
2,
with
those
municipalities
(Lower
Moreland
and
Philadelphia)
with
negative
or
zero
projected
growth
omitted.
As
noted
by
Church
(1999,
chap.
20)
and
Murray
(2010),
mod-
ern
GIS
have
evolved
to
include
sophisticated
analytical
tools
that
help
organize
location
problems.
Two
suitability
models
were
developed
using
the
ArcGIS
10.2
(ESRI,
2013),
one
for
economic
suitability
and
one
for
environmental
suitability.
The
GIS
base
lay-
ers
contained
data
reflecting
development
desirability.
Table
A.1
lists
the
data
sets
needed
to
produce
the
suitability
layers.
To
build
the
models,
each
vector
layer
was
converted
to
a
grid-formatted
raster
layer
using
values
for
a
single
suitability
criterion.
Each
raster
layer
was
then
reclassified
with
an
assigned
scale
value
ranging
from
2
to
10
(most
suitable).
Additionally,
the
model
required
layer-
influence
percentages
for
each
layer
that
sum
to
100.
Housing
data
representative
of
the
zip
codes
in
the
PCW
are
listed
in
Table
3.
Because
the
housing
data
representative
of
the
zip
codes
in
the
PCW
show
that
single
detached
homes
make
up
the
majority
of
new
housing
supply,
prices
and
costs
of
sin-
gle
homes
have
been
listed.
Adjustments
made
in
the
buildouts
reflect
the
presence
of
multi-household
supply.
Together
with
the
values
in
Table
4,
these
data
were
used
to
provide
econom-
ically
suitable
areas
for
development.
While
most
criteria
reflect
demand-side
desirability,
it
is
thought
that
developers
benefit
from
higher
asset
prices
for
homes
sold
as
demand
factors
make
them
more
desirable.
The
median
house
price
data
were
taken
Author's personal copy
194
J.A.
Sorrentino
et
al.
/
Landscape
and
Urban
Planning
125
(2014)
188–206
Table
3
Housing
price-related
data
(2012
$).
Zip
code
Median
house
price
($)
ATUC
($)
Common
Level
Ratio
Price
per
m2($)
Build
cost
per
m2($)
Profit
per
m2($)
18966
278,800
23,796
0.1273
1571.53
1140.22
431.31
18974
271,000
23,231
0.1273
1528.48
1140.22
388.25
19001
203,200
18,320
0.7512
1485.42
1221.17
264.25
19002
392,500
32,033
0.8153
1722.23
1221.17
501.06
19006
237,000
20,768
0.7354
1614.59
1221.17
393.42
19009
290,000
24,608
0.7108
1334.73
1221.17
113.56
19025
326,000
27,215
0.8153
1506.95
1221.17
285.78
19040
223,000
19,754
0.7188
2174.31
1221.17
953.14
19044
225,000
19,899
0.7188 1302.43 1221.17
81.27
19046
247,500
21,529
0.7401
1194.79
1221.17
26.37
19090
182,000
16,784
0.7401
1474.66
1221.17
253.49
19111
151,500
14,575
0.2846
1259.38
1002.01
257.37
19114
131,500
13,126
0.2846
1173.27
1002.01
171.25
19115
179,000
16,567
0.2846
1334.73
1002.01
332.71
19116
205,000
18,450
0.2846
1442.36
1002.01
440.35
19136
98,950
10,768
0.2846
968.75
1002.01
33.26
19149
105,000
11,206
0.2846
893.40
1002.01
108.61
19152
155,000
14,828
0.2846 1227.09
1002.01
225.07
Table
4
Neighborhood
criteria
for
economic
suitability.
Zip
code
Educational
attainment
Index
of
household
income
Crime
index
School
district
name
School
district
rating
18966
14.051
1.176
2.33
Centennial
0.893
18974
13.19
0.931
3.77
Centennial
0.893
19001
13.704
0.939
4.99
Upper
Moreland
0.767
19002
14.609
1.4
1.76
Upper
Dublin
0.878
19006
14.4
1.413
3.5
Lower
Moreland
0.985
19009
16.122
1.757
0.15
Bryn
Athyn
0.985
19025
15.629
1.623
3.08
Upper
Dublin
0.878
19040
13.374 0.833
2.38
Hatboro/Horsham
0.851
19044
13.731
0.912
3.79
Hatboro/Horsham
0.851
19046
14.497
1.255
4.99
Abington
0.697
19090
13.37
0.858
4.48
Abington
0.697
19111
12.571
0.643
5.27
Philadelphia
0.178
19114
12.661 0.699 5.79
Philadelphia
0.178
19115
12.879
0.698
4.08
Philadelphia
0.178
19116
13.115
0.737
3.85
Philadelphia
0.178
19136
11.954
0.599
5.7
Philadelphia
0.178
19149
12.281
0.618
7.17
Philadelphia
0.178
19152
12.369
0.645
6.7
Philadelphia
0.178
from
actual
house
sales
in
the
“Real
Estate
Trends”
section
of
www.trulia.com
(Trulia.com,
2011)
for
houses
sold
between
the
dates
1
October
and
31
December
2012.
It
is
important
to
note
that
sales
prices
for
settled
homes
are
the
only
appropriate
data
to
use
for
the
asset
price
of
a
house.
There
was
a
wide
range
of
housing
prices,
with
the
non-Philadelphia
zip
codes
commanding
much
higher
prices.
Therefore,
those
zip
codes
had
much
higher
user
costs
of
housing.
As
a
suitability
criterion,
a
higher
median
price
was
valued
more
highly
from
the
supply
side.
The
ATUC
as
given
in
Eq.
(1)
is
the
annual
cost
of
carrying
the
house.
For
suit-
ability,
a
lower
ATUC
was
given
a
higher
value
from
the
demand
side.
In
general,
housing
quality
is
important
to
potential
home
buy-
ers.
A
reasonable
proxy
for
the
quality
of
a
house
is
average
price
per
square
meter
(m2).
The
average
price
per
m2came
from
Trulia.com
(2011).
The
costs
of
construction
per
m2came
from
R.S.
Means
(2011)
and
were
updated
using
the
2%
increase
in
the
Producer
Price
Index
for
Materials
and
Supply
Inputs
to
Residential
Construction
for
December
2012
(US
BLS,
2012).
The
land
component
cost
was
set
at
25%
of
the
total
sales
price.
The
overhead
and
profit
figures
came
from
interviews
with
local
Philadelphia
metro
area
housing
developers.
The
overhead
was
set
at
10%
and
the
profit
17%
of
con-
struction
costs.
In
the
suitability
analysis,
the
higher
the
average
profit,
the
more
suitable
is
the
area
from
the
supply-side.
From
among
the
elements
in
Galster
et
al.
(2001),
those
in
Table
4
have
been
chosen
to
represent
neighborhood.
Each
zip
code
in
the
PCW
is
associated
with
a
public
school
district.
SchoolDigger.com
(2011)
ranks
schools
at
elementary,
middle,
and
high
school
levels
by
percentile
rank.
It
ranks
districts
using
the
arithmetic
average
of
the
percentile
ranks
of
the
schools
in
the
dis-
trict.
Column
six
in
Table
4
lists
these
scores.
A
higher
school
district
rating
implies
higher
suitability
from
the
demand
side.
Educational
attainment
was
found
using
American
FactFinder
(US
Census
Bureau,
2011).
Years
of
school
attained
were
multiplied
by
the
frequency
of
each
category
in
the
zip
code
population
to
get
the
frequency-weighted
average
years
of
school.
Column
two
in
Table
4
displays
a
range
from
slightly
below
12
to
slightly
above
16.
The
higher
the
educational
attainment,
the
higher
the
suit-
ability
of
that
zip
code
from
the
demand
and
supply
sides.
Mean
household
income
was
also
gotten
from
American
FactFinder
(US
Census
Bureau,
2011).
The
arithmetic
average
of
the
mean
incomes
of
the
zip
code
populations
was
computed.
The
index
of
house-
hold
income
was
calculated
for
each
zip
code
by
dividing
the
mean
income
of
that
zip
code
by
the
average
over
all
zip
codes.
Column
three
in
Table
4
shows
these
values.
The
higher
the
index
of
house-
hold
income,
the
higher
the
suitability
from
the
supply
side,
but
perhaps
also
the
demand
side.
Computing
the
crime
index
was
a
bit
more
involved.
Data
on
the
number
of
robberies,
aggravated
assaults,
burglaries
and
thefts
were
found
and
divided
by
thousands
of
people.
The
four
types
of
crimes
were
given
“severity
weights”
of
0.3,
0.3,
0.2
and
0.2
and
a
weighted
sum
of
the
four
crime
rates
was
calculated.
These
numbers
populate
the
crime
index
column
in
Table
4.
The
lower
the
crime
index,
the
higher
the
suitability
from
the
demand
and
supply
sides.
Author's personal copy
J.A.
Sorrentino
et
al.
/
Landscape
and
Urban
Planning
125
(2014)
188–206
195
Fig.
4.
Economically
suitable
areas.
Table
A.2
shows
five
(all
but
profit
based
on
natural
breaks)
reclassified
values
for
each
economic
suitability
criterion
raster
layer
ranging
from
2
to
10
(most
suitable).
The
land
use
classes
were
assigned
values
based
on
estimated
likelihood
of
conversion
into
residential
development
with
modern
construction
technol-
ogy.
The
layer
influences
were
ascertained
through
interviews
with
three
developers,
one
of
whom
is
Mr.
Thomas
Bentley,
President
of
Bentley
Homes,
Paoli,
PA
(personal
communication,
8
January
2014),
a
highly
involved
member
of
the
American
Association
of
Home
Builders.
Mr.
Bentley
shared
national
data
on
home
builders
and
the
determinants
of
housing
supply.
Table
A.3
shows
the
val-
ues
chosen.
Fig.
4
shows
the
most
economically
suitable
areas
for
residential
development
in
the
PCW.
The
criteria
used
to
determine
the
areas
for
residential
devel-
opment
consistent
with
local
sustainability
are
listed
vertically
in
Table
A.4.
Some
of
the
layers
were
merged
for
processing
conve-
nience.
The
assigned
scale
values
for
each
suitability
layer
ranging
from
2
to
10
(most
suitable).
The
layer
influences
were
solicited
in
an
online
survey
developed
in
Qualtrics
(2012)
of
municipal
Envi-
ronmental
Advisory
Council
members
in
the
watershed
throughout
the
months
of
March
and
April
in
2011.
The
six
question
survey
asked
for
inputs
in
choosing
suitability
layers
and
their
influences
in
the
model,
and
was
distributed
to
the
municipal
managers
and
chairpersons
of
the
Environmental
Advisory
Councils
of
each
municipality
with
land
area
in
the
PCW.
Overall,
the
survey
reached
52
people.
Of
the
43
responses
recorded,
36
were
complete
and
used.2Fig.
5
contains
the
environmentally
suitable
areas.
Each
municipality
was
clipped
to
the
watershed
boundaries.
As
noted,
the
zoning
classifications
for
residential
development
within
the
watershed
vary
among
municipalities.
Each
zoning
class
includes
information
on
lot
size
requirements,
height
restrictions,
and
setback
distances.
The
zoning
classifications
were
used
as
the
base
layer
to
run
the
buildout
analyses.3
The
software
chosen
to
conduct
the
buildouts
for
this
study
was
CommunityViz
4.1
Scenario
360
(Placeways,
2009).
The
Buildout
Wizard
requires
a
GIS
base
layer,
such
as
a
zoning
layer,
to
provide
density
allowances
for
each
zoning
class.
The
software
prompts
user
input
on
zoning
classifications,
existing
buildings,
minimum
lot
sizes,
minimum
setback
distances
for
front,
back
and
side
yards,
height
restrictions,
and
other
data.
Using
this
information,
the
Wiz-
ard
then
calculates
and
spatially
allocates
the
maximum
number
of
buildings
that
could
be
built
on
the
developable
land.
The
choice
was
made
in
the
Wizard
to
“follow
roads.”
2This
survey
of
public-spirited
individuals
was
performed
as
part
of
another
study
by
one
of
the
present
authors.
It
contrasts
with
the
small
number
of
developers
asked
to
divulge
proprietary
information
on
commercial
activities.
3More
detailed
information
on
the
areas
that
the
zoning
classes
cover
is
available
from
the
authors.
Author's personal copy
196
J.A.
Sorrentino
et
al.
/