ArticlePDF Available

Impacts of Sea Level Rise on Real Estate Prices in Coastal Georgia *

  • Georgia Southern University, Savannah, US

Abstract and Figures

This paper adopts a hedonic pricing model to study the impact of vulnerability to inundation from sea level rise on home prices in Savannah, Georgia. We find that homes most at risk from sea level rise are associated with an approximate 3.1 percent price discount. The results are consistent with prior studies, which uses data from different locations in U.S. coastal areas. We also find that the discount is more significant in our later sample period, indicating that house buyers may be becoming more aware of the climate risk. This paper contributes to the understanding of house pricing factors in the study area regarding the sea level rise effects.
Content may be subject to copyright.
(2020) 49, 43–52
Impacts of Sea Level Rise on Real Estate Prices in
Coastal Georgia
Jason Beckaand Meimei Linb
aDepartment of Economics, Georgia Southern University, USA
bDepartment of Geology and Geography, Georgia Southern University, USA
Abstract: This paper adopts a hedonic pricing model to study the impact of vulnerability to inundation
from sea level rise on home prices in Savannah, Georgia. We find that homes most at risk from sea level
rise are associated with an approximate 3.1 percent price discount. The results are consistent with prior
studies, which uses data from different locations in U.S. coastal areas. We also find that the discount is
more significant in our later sample period, indicating that house buyers may be becoming more aware of
the climate risk. This paper contributes to the understanding of house pricing factors in the study area
regarding the sea level rise effects.
Keywords: coastal Georgia, house prices, hedonic model, sea level rise
JEL Codes: R30, R31, R11
It is widely accepted that human beings have altered our natural environments in ways
that could trigger irreversible consequences for global climate systems. Since the mid 20th
century, global average temperatures have risen and the resultant retreat of land ice and
thermal expansion of water molecules has caused sea level rise (SLR) worldwide. While
the exact rate at which temperature and sea level will continue to rise is unknown, SLR is
projected to be in the range of one to six feet by the end of the century and is expected to
increase at a faster rate beyond 2100 (Meyer and Pachauri, 2014).
The impact of SLR is uneven across regions with the most severe effects concentrated
in low-lying coastal areas. Due to SLR, coastal areas will likely experience flooding, sub-
mergence, and erosion during and beyond the 21st century (Meyer and Pachauri, 2014).
Exacerbating this situation further is the fact that coastal regions are often densely popu-
lated. According to “National Coastal Population Report: Population Trends from 1970 to
Jason Beck is an Associate Professor of Economics at Georgial Southern University, Savannah, GA. Meimei
Lin is an Assistant Professor of Geography at Georgia Souther University, Svannah, GA. Corresponding
Author: Jason Beck E-mail:
(c) Southern Regional Science Association 2020
ISSN 1553-0892, 0048-49X (online)
44 The Review of Regional Studies 50(1)
2020,” approximately 164 of 313 million people in the United States (approximately 52 per-
cent) live in coastal watershed counties and the population density of coastal areas is roughly
three times that of the U.S. in general (National Oceanic and Atmospheric Administration,
While the effects of SLR will likely be far reaching, one area where its effects will be
particularly felt is in the real estate market. Given that real estate is typically the largest
investment for the median U.S. household (Campbell, 2006), a sharp decline in the value of a
significant fraction of the housing stock would represent a large decrease in wealth for many
households in affected areas. Yale Climate Opinions map data (Howe et al., 2015) suggest
that 61 percent of adults in Chatham County, GA (where Savannah is located and the focus
of this study) “are worried about global warming,” which is among the highest in coastal
Georgian counties. If buyers recognize this threat, it is possible that the risk of SLR may be
reflected in current home values for vulnerable properties.
A recent literature has emerged examining the effects of expected SLR on property values.
Bernstein et al. (2019) use a national database of home transactions from 2007-2016 and find
that vulnerable properties were associated with a 7 percent discount. They suggest that this
discount may be driven in part by buyer sophistication in that non-owner occupied properties
sell for a larger discount than owner-occupied units. They also note that this discount has
increased over time. Furthermore, they find that among owner-occupiers, areas that express
greater concern for (and belief in) SLR are associated with larger penalties than areas that
are less worried about the effects of climate change.
Baldauf et al. (2018) further capitalize on the heterogeneity of beliefs regarding the
negative effects of climate change to see how people in different areas have responded. They
connect survey data on beliefs across the U.S. regarding climate change with a comprehensive
dataset of housing transactions and find that the property value discount is 7 percent larger
for “believer” areas compared to “non-believer” areas. Bakkensen and Barrage (2017) find
that the heterogeneity of beliefs regarding SLR contributes to the selection into coastal
homes. In particular, people living in the flood zone tend to underestimate their flood risks
compared to those of inundation models, whereas people who live outside the flood zone are
more concerned with flood risks.
Using data from the Chesapeake Bay area, Walsh et al. (2017) also find a discount for
properties vulnerable to SLR but that the discount is offset in the presence of adaptive
structures such as bulkheads and ripraps. They show that homes in the high risk inundation
areas sell for an average of 19-23 percent less if there is a lack of protective structures.
However, if certain protection structures are in place, property values of the high risk zones
can increase up to 21 percent.
Based on hedonic analyses of property transaction data in Galveston County, TX, Atreya
and Czajkowski (2014) argue that the impact of flood risk on housing value is dependent
upon its water-related amenities such as distance to the coast. Houses within a quarter mile
of the nearest coastline sell at a price premium despite having higher flood risks. Ortega and
Taspinar (2018) discover that housing prices significantly dropped after Hurricane Sandy in
New York City. Overall, they find a 9 percent reduction in property values in storm-affected
areas compared with similar houses outside the storm.
Southern Regional Science Association 2020.
This project uses Multiple Listing Service (MLS) data in a hedonic model to see if such
properties vulnerable to SLR sell for less in Savannah, Georgia, a coastal region of the U.S.
We find homes that are at high risk (i.e. ones that would be inundated with SLR of 0-3
feet) sell at a 3.1 percent discount relative to those that would only be inundated at SLR
greater than 6 feet. Homes that would only be inundated with a 4-6 feet rise in sea level see
no difference in price relative to safe1homes.
Our results suggest that climate change may already be having an impact on housing
values. Such information could be valuable for policy makers when considering options to
mitigate the impact of climate change, especially in the most vulnerable areas. As any policy
response will weigh the costs and benefit of action, getting as clear of a picture as possible
regarding the costs of SLR becomes important. Home devaluation from SLR would be a
part of those costs.
Real estate transaction data were obtained via the MLS of Savannah, GA over the period
2007-2016. Over this time, there were roughly 42,000 single-family homes sold through the
MLS. The data are detailed enough to allow for a number of hedonic controls related to
physical home characteristics, property location, and timing of sale.
To generate the variables associated with vulnerability to inundation, the GIS shapefile
for Georgia was downloaded from the National Oceanic and Atmospheric Administration
(NOAA) SLR Calculator. This online data portal provides access to an online SLR calculator
viewer as well as the underlying shapefiles for all 48 lower U.S. States. This GIS shapefile of
Georgia includes polygons of potentially impacted areas under 0-6 feet SLR and was further
clipped to include just Savannah, GA. All home sales of Savannah were geocoded from the
longitude and latitude data provided in the MLS and overlaid with 0-6 feet SLR zones to
generate vulnerability measures to coastal inundation.
The housing data were restricted to include only existing (i.e. not newly built) homes
with a reported age of 2 years or greater and selling for more than $50,000 and less than $2
million. Entries with blank or illogical values were likewise removed. Analysis begins with
34,807 sold homes.
To allow home characteristics to affect sales outcomes nonlinearly, we follow Levitt
and Syverson (2008) by operationalizing all independent variables as binary regres-
sors. The variables, ON E BEDROOM,T W OB EEDROO M,T H REE BE DROOM ,
mous variables equaling one if the observation has one, two, three, four, five, or six
plus bedrooms respectively, and zero otherwise. Full and half baths are controlled
for in a similar fashion with the variables ON EF U LLBAT H ,T W OF U LLBAT H ,
1Here, the term “safe” is used in a relative context. Since Meyer and Pachauri (2014) estimates SLR to be
in the range of one to six feet by 2100, we consider homes that would only be inundated with a SLR of over
six feet to be “safe.”
Southern Regional Science Association 2020.
46 The Review of Regional Studies 50(1)
Table 1: Definitions of Variables in the Analysis
P RI CE : real sales price of house, base year 2012
LnP RI CE : the natural log of the real sales price
SQF T 1S QF T 5: a series of binary variables to indicate if the house falls in the 1st square footage
quintile (0-1311 sqft), 2nd sqft quintile (1312-1580), 3rd sqft quintile (1581-1930), 4th sqft quintile
(1931-2541) or 5th sqft quintile (2542+)
T W OB EDROO MS : a binary variable to indicate if the house has two bedrooms
T H REBE DROOM S : a binary variable to indicate if the house has three bedrooms
F OU RBE DROOM S : a binary variable to indicate if the house has four bedrooms
F I V E BED ROOM S: a binary variable to indicate if the house has five bedrooms
SI X P LUS BE DROOM S: a binary variable to indicate if the house has six or more bedrooms
T W OF U LLBAT H : a binary variable to indicate if the house has two full bathrooms
T H REEF U LLB AT H: a binary variable to indicate if the house has three full bathrooms
F OU RF U LLBAT H : a binary variable to indicate if the house has four full bathrooms
F I V E P LUS F U LLBAT H: a binary variable to indicate if the house has five or more full bathrooms
ON EH ALF BAT H : a binary variable to indicate if the house has one half bathrooms
T W OH ALF BAT H : a binary variable to indicate if the house has two half bathrooms
T H REEH ALF BAT H : a binary variable to indicate if the house has three half bathrooms
F OU RP LU SH ALF BAT H : a binary variable to indicate if the house has four or more half bathrooms
F I REP LAC E: a binary variable to indicate if the house has a fireplace
ON EGARAGE: a binary variable to indicate if the house has one garage space
T W OGARAGE : a binary variable to indicate if the house has two garage spaces
T H REEP LU S GARAGE: a binary variable to indicate if the house has three or more garage spaces
Y2007 Y2016: a series of binary variables to indicate if the house was sold in year 2007-2016
25Y EARS : a binary variable to indicate if the house was 2-5 years old at sale
610Y EARS : a binary variable to indicate if the house was 6-10 years old at sale
11 25Y EARS : a binary variable to indicate if the house was 11-25 years old at sale
26 50Y EARS : a binary variable to indicate if the house was 26-50 years old at sale
51 100Y EARS : a binary variable to indicate if the house was 51-100 years old at sale
101 + Y EARS : a binary variable to indicate if the house was 101 or more years old at sale
P OOL: a binary variable to indicate if the house has a swimming pool
W AT ERF RO NT : a binary variable to indicate if the house was marketed as a waterfront property
SLR1: a binary variable to indicate if the house would be inundated with a 0-1 foot rise in sea level
SLR2: a binary variable to indicate if the house would be inundated with a 1.01-2 foot rise in sea level
SLR3: a binary variable to indicate if the house would be inundated with a 2.01-3 foot rise in sea level
SLR4: a binary variable to indicate if the house would be inundated with a 3.01-4 foot rise in sea level
SLR5: a binary variable to indicate if the house would be inundated with a 4.01-5 foot rise in sea level
SLR6: a binary variable to indicate if the house would be inundated with a 5.01-6 foot rise in sea level
HIRI SK: a binary variable to indicate if the house would be inundated with a 0-3 foot rise in sea level
LOW RI SK : a binary variable to indicate if the house would be inundated with a 3.01-6 ft rise in sea level
NORI SK: a binary variable to indicate if the house would only be inundated with a >6 ft rise in sea level
ZI P 1Z IP 167: binary variables =1 when house was sold in the associated 6-digit ZIP code, 0 otherwise
is controlled for with dummy variables for square footage quintiles (see Table 1 for specific
ranges). Yearly fixed effects are represented by the variables Y2007-Y2016. Age of the
house is controlled for via five dummy variables for different age ranges, specifically
25Y EARS, 6 10Y EARS, 11 25Y EARS, 26 50Y EARS, 51 100Y EARS, and
101 + Y EARS.2Additional physical characteristics, such as the presence of a fireplace,
presence of a swimming pool, and the number of garage spaces are controlled for, as well as
2Recall that new homes are excluded from the sample.
Southern Regional Science Association 2020.
Table 2: Descriptive Statistics
Variable Mean Std. Deviation
P RI CE 222153.4 171802.4
LnP RI CE 12.116 0.595
SQF T 1 0.200 0.400
SQF T 2 0.200 0.400
SQF T 3 0.200 0.400
SQF T 4 0.200 0.400
SQF T 5 0.200 0.400
ON EBEDROOM 0.017 0.131
T W OB EDROO MS 0.104 0.305
T H REBE DROOM S 0.584 0.492
F OU RBE DROOM S 0.244 0.429
F I V E BED ROOM S 0.044 0.205
SI X P LUS BE DROOM S 0.006 0.077
T W OF U LLBAT H 0.693 0.046
T H REEF U LLB AT H 0.157 0.363
F OU RF U LLBAT H 0.030 0.171
F I V E P LUS F U LLBAT H 0.008 0.090
ZE ROH ALF BAT H 0.688 0.463
ON EH ALF BAT H 0.299 0.457
T W OH ALF BAT H 0.012 0.109
T H REEH ALF BAT H 0.0006 0.023
F OU RP LU SH ALF BAT H 0.0001 0.009
F I REP LAC E 0.620 0.485
ZE ROGARAGE 0.345 0.475
ON EGARAGE 0.130 0.336
T W OGARAGE 0.483 0.499
T H REEP LU S GARAGE 0.041 0.199
25Y EARS 0.152 0.359
610Y EARS 0.193 0.394
11 25Y EARS 0.275 0.446
26 50Y EARS 0.190 0.392
51 100Y EARS 0.150 0.357
101 + Y EARS 0.037 0.188
P OOL 0.053 0.224
W AT ERF RO NT 0.116 0.320
CONDO 0.133 0.340
SLR1 0.018 0.134
SLR2 0.004 0.063
SLR3 0.014 0.117
SLR4 0.033 0.178
SLR5 0.034 0.181
SLR6 0.040 0.195
HIRI SK 0.036 0.187
LOW RI SK 0.107 0.308
NORI SK 0.857 0.350
whether the home was a single-family dwelling or a condominium/townhouse.
Location is accounted for in the hedonic model via a series of dummy variables each
representing a 6-digit zip code similar to Beck et al. (2018). These zip codes were constructed
Southern Regional Science Association 2020.
48 The Review of Regional Studies 50(1)
by truncating available 9-digit zip codes thereby creating 167 different locational controls.
The largest represents 4.8 percent of the sample. Note that not all 5-digit zips will necessarily
have ten distinct sub areas. By controlling for location with these 6-digit zip codes, an added
degree of granularity is achieved.
In addition to being close to the coast, the Savannah area has many inland waterways
and thus many homes adjacent to water. Since vulnerability to SLR will often be correlated
with proximity to water, it is difficult to separate the potential price effects of SLR risk
from the other effects associated with water adjacent homes. These effects could have a
negative component, in that these properties may have a higher risk of non-SLR related
flooding, and/or a positive component related to the amenity value of waterfront property.
We control for a home being directly on the water with a dummy variable equaling one if the
home is marketed as “waterfront.” Unfortunately, since low elevation areas are susceptible
to both SLR-related inundation and non-SLR flooding, it is difficult to disentangle the two
effects. Thus, our estimates might be considered an upper bound for price effects due to
As is common in the literature, the dependent is the natural log of sales price. The
nominal sales prices provided by the MLS were adjusted into year 2012 dollars via the
Implicit Price Deflator.3
The average home in the sample has three bedrooms, two full baths, is 28 years old,
and sold for $222,153. Tables 1 and 2 provide descriptions of each variable and summary
statistics, respectively.
Following Haag et al. (2000), Levitt and Syverson (2008), Cebula (2009), Beck et al.
(2018), and others, we employ a hedonic (logged) pricing model to estimate the effects of
the independent variables on (logged) home prices. We estimate this model via OLS with
robust standard errors:
ln(P rice)j=f(Ij, Ej, Oj) (1)
where ln(P rice)jis the natural log of the price of house j,Ija vector of interior and exterior
physical characteristics for house j,Ejis a vector of time of sale characteristics for house j,
and Ojis a vector of characteristic controlling for SLR risk associated with house j.
Results for the hedonic model are presented in Table 3. Coefficients on the control variables
largely match expectations and previous research. Larger homes sell for progressively more,
as do homes with more full and half baths. Newer homes tend to sell for more, except relative
to homes over one hundred years old. Given the presence of a well-known and desirable
historical district in Savannah, it is unsurprising that historic homes are not associated with
a sale price discount (see Cebula (2009)). Amenities such as a fireplace, a swimming pool,
more garage spaces, and the property being waterfront are associated with higher selling
prices. The negative coefficient on the bedroom variables can be explained by the inclusion
Southern Regional Science Association 2020.
Table 3: OLS Estimates of the Hedonic Pricing Model for the Dependent
Variable: ln(price)
Model 1: 6 SLR Model 2: 2 SLR Control
Control Variables (SLR1-SLR6) Variables (HIRISK & LOWRISK)
Variable Coef t-stat p-val Coef t-stat p-val
SQF T 2 0.154 28.700 0.000 0.154 28.750 0.000
SQF T 3 0.284 47.610 0.000 0.285 47.700 0.000
SQF T 4 0.449 63.840 0.000 0.449 63.930 0.000
SQF T 5 0.666 74.790 0.000 0.666 74.900 0.000
ON EB EDRO OM -0.149 -8.060 0.000 -0.149 -8.080 0.000
T W OBE DROOM S 0.020 2.670 0.008 0.020 2.650 0.008
F OU RBE DROOM S -0.025 -5.720 0.000 -0.025 -5.710 0.000
F IV EBE DROOM S -0.064 -6.670 0.000 -0.064 -6.680 0.000
SI XP LU SB EDRO OMS -0.105 -3.260 0.001 -0.105 -3.250 0.001
T W OF U LLBAT H 0.160 19.750 0.000 0.159 19.730 0.000
T HRE EF U LLBAT H 0.298 29.340 0.000 0.298 29.320 0.000
F OU RF U LLBAT H 0.516 34.150 0.000 0.516 34.140 0.000
F IV EP LU SF U LLBAT H 0.739 24.970 0.000 0.739 24.970 0.000
ON EH ALF BAT H 0.089 21.890 0.000 0.089 21.880 0.000
T W OH ALF BAT H 0.180 9.470 0.000 0.180 9.440 0.000
T HRE EH ALF BAT H 0.425 3.650 0.000 0.425 3.650 0.000
F OU RP LU SHALF B AT H 0.164 2.830 0.005 0.163 2.830 0.005
F IRE P LACE 0.099 25.420 0.000 0.099 25.420 0.000
ON EGARAGE 0.073 13.160 0.000 0.073 13.190 0.000
T W OGARAGE 0.163 31.660 0.000 0.163 31.700 0.000
T HRE EP LU SGARAGE 0.316 31.740 0.000 0.316 31.770 0.000
Y2008 -0.037 -5.570 0.000 -0.037 -5.570 0.000
Y2009 -0.117 -16.830 0.000 -0.117 -16.830 0.000
Y2010 -0.178 -25.290 0.000 -0.178 -25.280 0.000
Y2011 -0.250 -35.450 0.000 -0.250 -35.450 0.000
Y2012 -0.217 -31.300 0.000 -0.217 -31.320 0.000
Y2013 -0.134 -20.180 0.000 -0.134 -20.200 0.000
Y2014 -0.063 -9.800 0.000 -0.063 -9.800 0.000
Y2015 0.015 2.470 0.014 0.015 2.470 0.014
Y2016 0.074 12.010 0.000 0.074 12.010 0.000
610Y EARS -0.020 -4.530 0.000 -0.020 -4.520 0.000
11 25Y EARS -0.048 -9.670 0.000 -0.048 -9.700 0.000
26 50Y EARS -0.108 -16.060 0.000 -0.108 -16.060 0.000
51 100Y EARS -0.059 -5.490 0.000 -0.059 -5.500 0.000
101 + Y EARS 0.047 2.450 0.014 0.047 2.450 0.014
P OOL 0.089 12.440 0.000 0.089 12.470 0.000
W AT ERF RO NT 0.198 34.600 0.000 0.198 34.550 0.000
CONDO -0.092 -13.380 0.000 -0.091 -13.490 0.000
SLR1 -0.025 -1.860 0.062
SLR2 -0.077 -3.450 0.001
SLR3 -0.022 -1.420 0.156
SLR4 0.000 0.040 0.965
SLR5 0.006 0.690 0.487
SLR6 0.005 0.630 0.528
HI RI SK -0.031 -3.170 0.002
LOW RI SK 0.004 0.690 0.493
R20.788 0.788
F 57.230 56.340
N 34807 34807
Note: 167 6-digit zip code fixed effects are present but not reported
of the square footage variables and is consistent with other research (see Salter et al. (2012),
Waller and Jubran (2012), Seagraves and Gallimore (2013) Allen et al. (2015), among others).
Both the national decline in home values and the subsequent rebound over the last
decade are visible in these results. Relative to 2007, Savannah housing prices appear to have
bottomed out in 2011 before trending back upwards and approaching their 2007 levels by
Southern Regional Science Association 2020.
50 The Review of Regional Studies 50(1)
Turning now to our variables of interest, we see that the coefficients on SLR1 and SLR2
are negative and statistically significant. They roughly exhibit a tapering effect, mostly
decreasing in magnitude and statistical significance as the level of risk decreases. Properties
that are the most vulnerable to sea level rise (i.e. those that would be inundated with a rise
in sea level of one foot or less) are associated with a 2.5 percent penalty relative to those
that are under little/no risk of being inundated. This result is statistically significant at the
10 percent level. Properties that would be inundated with a two-foot or three-foot sea level
rise are associated with a statistically significant 7.7 percent discount. Homes inundated
with a three to six foot rise were not associated with a statistically significant different sales
price than safe homes. It is unclear why the associated coefficient on SLR2 is larger than
the coefficient on SLR1. Despite the fairly rich set of control variables that are provided
in the MLS data, unobservable heterogeneity is always a possibility and it could be that
the homes that are most directly threatened by SLR are also associated with unobservable
positive home characteristics such as beach access or an ocean view. Examined as a whole, a
general picture emerges of an association between degree of property vulnerability and lower
sales price.
A similar model (Model 2) was also estimated with the seven sea level rise variables
(SLR1-SLR6 and “no risk”) collapsed into three categories: no risk (as previously defined),
low risk (homes that would be inundated with a four to six foot increase in sea level), and
high risk (homes that would be inundated with a SLR of three feet or fewer). Relative to the
reference group (“no risk”), homes designated as high risk were associated with a statistically
significant 3.1 percent lower price. Low-risk homes saw no noticeable impact on sales price.
Over time, information levels regarding the risk of SLR may increase. Additionally, the
time before the most direct consequences of SLR will be felt grows shorter. It is thus possible
that the effects of SLR risk on housing prices may change over time. To examine this, we
divide the ten years of available data into two five-year periods: 2007-2011 and 2012-2016.
In estimating Model 2 on each of the two five-year sub-samples, we find that the discount
associated with the highest risk properties increases from 3.4 percent to 4 percent across the
two periods. This difference is statistically significant.
Since any public policy action will consider the costs and benefit of action, information on the
costs of SLR is important when formulating a policy response. One of the many costs of SLR
will be home devaluation in vulnerable areas. This paper combines MLS home transaction
data from 2007-2016 for Savannah, Georgia with GIS mapping to explore the connection
between properties at risk of inundation as a result of SLR and their observed price. The
results suggest that properties that would be inundated with a sea level rise of three feet
or less were associated with a 3.1 percent discount. Properties that would be inundated
with a 4-6 feet SLR saw no discount relative to safe homes. Additionally, we find that the
discount was statistically larger in the period of 2012-2016 than it was in 2007-2011. These
findings suggest that buyers may be aware of the risk of vulnerable properties, are bidding
accordingly, and this effect has grown over time. Our results are in line with other research,
Southern Regional Science Association 2020.
such as Bernstein et al. (2019), which reports a 7 percent discount for at risk homes.
This study is not without limitations. For example, it focuses on only one narrow ge-
ographic location and one should be cautious about extrapolating these results to other
coastal areas where inundation risk and median home values may be different. Furthermore,
work by Walsh et al. (2017) suggests that adaptive structures may reduce the impact of
SLR on home values. This study does not take into account adaptations home owners may
undertake as their understanding of SLR changes over time.
Future research could focus on other coastal areas and utilize newer data as they become
available. Our results suggest the impact of inundation risk on home prices was greater
in more recent years, but available data only extended to 2016. As new transaction data
become available, it would be interesting to see if the effect continues to increase.
Allen, Marcus T., Anjelita Cadena, Jessica Rutherford, and Ronald Rutherford. (2015) “Ef-
fects of Real Estate Brokers’ Marketing Strategies: Public Open Houses, Broker Open
Houses, MLS Virtual Tours, and MLS Photographs,” Journal of Real Estate Research,
37(3), 343–369.
Atreya, Ajita and Jeffrey Czajkowski. (2014) “Is Flood Risk Universally Sufficient to Offset
the Strong Desire to Live Near the Water,” Risk Management and Decision Processes
Center Working Paper No. 2014-09.
Bakkensen, Laura A. and Lint Barrage. (2017) “Flood Risk Belief Heterogeneity and Coastal
Home Price Dynamic Going Under Water,” National Bureau of Economic Research Work-
ing Paper No. 23854.
Baldauf, Markus, Lorenzo Garlappi, and Constantine Yannelis. (2018) “Does Climate
Change Affect Real Estate Prices? Only If You Believe in It,” Working Paper. Available
online in January 2020 at id=3240200.
Beck, Jason, Stephen Bray, and Aaron Trapani. (2018) “Using Agent Remarks to Explore
the Principal-Agent Relationship in Residential Real Estate Brokerage,” The Empirical
Economic Letters, 17(5), 569–577.
Bernstein, Asaf, Matthew Gustafson, and Ryan Lewis. (2019) “Disaster on the Horizon: The
Price Effects of Seal Level Rise,” Journal of Financial Economics, 134(2), 253–272.
Campbell, John Y. (2006) “Household Finance,” The Journal of Finance, 61(4), 1554–1604.
Cebula, Richard J. (2009) “The Hedonic Pricing Model Applied to the Housing Market of the
City of Savannah and Its Savannah Historic Landmark District,” The Review of Regional
Studies, 39(1), 9–22.
Haag, Jerry T., Ronald C. Rutherford, and Thomas A. Thomson. (2000) “Real Estate Agent
Remarks: Help or Hype?,” Journal of Real Estate Research, 20(1–2), 205–215.
Howe, Peter D., Matto Mildenberger, and Leiserowitz Anthony Marlon, Jennifer. R.. (2015)
“Geographic Variation in Opinions on Climate Change at State and Local Scales in the
USA,” Nature Climate Change, 5, 596–603.
Levitt, Steven D. and Chad Syverson. (2008) “Market Distortions when Agents Are Better
Informed: The Value of Information in Real Estate Transactions,” National Bureau of
Economic Research Working Paper No. 11053.
Southern Regional Science Association 2020.
52 The Review of Regional Studies 50(1)
Meyer, Leo and Rajendra K. Pachauri. (2014) “Climate Change 2014: Synthesis Report,”
Intergovernmental Panel on Climate Change.
National Oceanic and Atmospheric Administration. (2013) “National Coastal Population
Report: Population Trends from 1970 to 2020,” US Department of Commerce.
Ortega, Francesc and S¨uleyman Taspinar. (2018) “Rising Sea Levels and Sinking Property
Values: Hurricane Sandy and New York’s Housing Market,” Journal of Urban Economics,
106, 81–100.
Salter, Sean P., Franklin G. Mixon Jr., and Ernest W. King. (2012) “Broker Beauty and
Boon: A Study of Physical Attractiveness and its Effects on Real Estate Brokers Income
and Productivity,” Applied Financial Economics, 21(1), 811–825.
Seagraves, Philip and Paul Gallimore. (2013) “The Gender Gap in Real Estate Sales: Nego-
tiation Skill or Agent Selection?,” Real Estate Economics, 41(3), 600–603.
Waller, Bennie D. and Ali M. Jubran. (2012) “The Impact of Agent Experience on the Real
Estate Transaction,” Journal of Housing Research, 21(1), 67–82.
Walsh, Patrick, Charles Griffiths, Dennis Guignet, and Heather Klemick. (2017) “Modeling
the Property Price Impact of Water Quality in 14 Chesapeake Bay Counties,” Ecological
Economics, 135, 103–113.
Southern Regional Science Association 2020.
... Alternatively, there is a fair amount of literature on the effects of other natural disasters such as floods, storms, wildfires, rising sea levels, hurricanes and earthquakes on housing prices. Most of these studies found that natural disasters have discounting effects on housing prices (Murdoch et al., 1993;Pompe and Rinehart, 1995;Mueller et al., 2009;Beltrán et al., 2018;Beck and Lin, 2020;Graff Zivin et al., 2020). Hedonic models are also more common in this field; these models control for the impacts of housing characteristics such as square footage and the number of bathrooms on the price. ...
... For example, several studies used OLS ©Southern Regional Science Association 2021. models with dummy variables that accounted for the potential impact of wildfires (Mueller et al., 2009), floods (Beltrán et al., 2018), sea level rise (Beck and Lin, 2020), and hurricanes (Graff Zivin et al., 2020) on local housing values. Notably, Graff Zivin et al. (2020) -whose methodology we follow in this paper -finds a positive impact on housing prices for houses sold in locations impacted by Florida hurricanes in the following 1-3 years. ...
Full-text available
Abstract: This paper examines the effects of four significant tornadoes on housing prices in Moore, Oklahoma. We use a hedonic difference-in-difference approach by considering transactions made preceding and following major tornadic events in Moore (occurring in 1999, 2003, 2010, and 2013). Our dataset spans nearly 30,000 housing transactions between 1990 and 2020. The length of tornado impacts is evaluated using a set of time indicators for the years leading up to (and after) the tornadoes. We find a 2-5% decrease in housing prices during the first year after a tornado for houses in the destructive path. However, no such impacts exist in years 2-7. We also employ three different specifications (OLS, spatial lag, and spatial error) to find the most appropriate model for considering the potential spatial processes at work. The results are largely similar across specifications.
... Further analyses of economic value and ecosystem services could incorporate changes in social welfare that stem from improvements or degradation in recreational experiences (e.g., fishing, bird watching, boating), commercial fishing, and cultural associations within the Nisqually River Delta. When combined with other services such as coastal protection, fish production, recreation, cultural resources, and aesthetic value, the total economic value of marshes is orders of magnitude higher than when considering SC-CO 2 alone (Lynne et al. 1981;Barbier 2019;Beck and Lin 2020;Good 2020). ...
Full-text available
Sea-level rise (SLR) and obstructions to sediment delivery pose challenges to the persistence of estuarine habitats and the ecosystem services they provide. Restoration actions and sediment management strategies may help mitigate such challenges by encouraging the vertical accretion of sediment in and horizontal migration of tidal forests and marshes. We used a process-based soil accretion model (Coastal Wetland Equilibrium Model) combined with a habitat classification model (MOSAICS) to estimate the effects of SLR, suspended sediment, and inland habitat migration on estuarine habitats, soil carbon accumulation, and economic value of climate change mitigation of carbon accumulation (social cost of carbon dioxide) in a macrotidal estuary in the northwest USA over 100 years (2011 to 2110). Under present-day sediment levels, we projected that after 100 years, most high salt marsh would remain with < 100 cm SLR, but substantial area converted to transitional (low) salt marsh and mudflat with ≥ 100 cm SLR. Increasing sediment availability increased the projected resilience of transitional salt marsh to SLR but did not prevent declines in high marsh area. Projected total carbon accumulation plateaued or declined with ≥ 100 cm SLR, yet the economic value of carbon accumulation continued to rise over time, suggesting that the value of this ecosystem service was resilient to SLR. Doubling or tripling sediment availability increased projected carbon accumulation up to 7.69 and 14.2 kg m ⁻² and increased total economic value up to $373,000 and $710,000, respectively. Allowing marsh migration supported conversion of upland to freshwater marsh, with slight increases in carbon accumulation. These results inform climate adaptation planning for wetland managers seeking to understand the resilience of estuarine habitats and ecosystem services to SLR under multiple management strategies.
... However, coastal properties in Savannah, Georgia, showed small price discounts of only a few percent from 2007 to 2016 despite being exposed to several feet of sea level rise (J. Beck & Lin, 2020). Similar price discounts were found for high-risk properties in Miami-Dade County, FL, but interestingly these were lower for highly priced properties purchased as non-primary residences, suggesting a greater risk tolerance by wealthy buyers (Fu & Nijman, 2020). ...
Full-text available
Increasing coastal flooding threatens urban centers worldwide. Projections of physical damages to structures and their contents can characterize the monetary scale of risk, but they lack relevant socioeconomic context. The impact of coastal flooding on communities hinges not only on the cost, but on the ability of households to pay for the damages. Here, we repurpose probabilistic risk assessment to analyze the monetary and social risk associated with coastal flooding in the San Francisco Bay Area for 2020–2060. We show that future coastal flooding could financially ruin a substantial number of households by burdening them with flood damage costs that exceed discretionary household income. We quantify these impacts at the census block group scale by computing the percentage of households without discretionary income, before and after coastal flooding costs. We find that for several coastal communities in San Mateo County more than 50% of households will be facing financial instability, highlighting the need for immediate policy interventions that target existing, socially produced risk rather than waiting for potentially elusive certainty in sea level rise projections. We emphasize that the percentage of financially unstable households is particularly high in racially diverse and historically disadvantaged communities, highlighting the connection between financial instability and inequity. While our estimates are specific to the San Francisco Bay Area, our granular, household‐level perspective is transferable to other urban centers and can help identify the specific challenges that different communities face and inform appropriate adaptation interventions.
Full-text available
A potential principal-agent problem between real estate agents and their clients is well recognized in the literature. Several studies compare the sales outcomes of agent-owned homes with those sold for clients and find evidence that agents convince clients to sell too quickly and too cheaply. This paper uses the non-public agent-to-agent remarks in the MLS to not only compare sales outcomes of agent-owned homes with client-owned homes in Savannah, GA, but also to explore other agent/seller relationships such as a stated agent/client material relationship and agents representing other agents. Contrary to other studies, we find no evidence of a principal-agent problem between agents and sellers.
Full-text available
Addressing climate change in the United States requires enactment of national, state and local mitigation and adaptation policies. The success of these initiatives depends on public opinion, policy support and behaviours at appropriate scales. Public opinion, however, is typically measured with national surveys that obscure geographic variability across regions, states and localities. Here we present independently validated high-resolution opinion estimates using a multilevel regression and poststratification model. The model accurately predicts climate change beliefs, risk perceptions and policy preferences at the state, congressional district, metropolitan and county levels, using a concise set of demographic and geographic predictors. The analysis finds substantial variation in public opinion across the nation. Nationally, 63% of Americans believe global warming is happening, but county-level estimates range from 43 to 80%, leading to a diversity of political environments for climate policy. These estimates provide an important new source of information for policymakers, educators and scientists to more effectively address the challenges of climate change.
How do climate risk beliefs affect coastal housing markets? This paper provides theoretical and empirical evidence. First, we build a dynamic housing market model and show that belief heterogeneity can reconcile prior mixed evidence on flood risk capitalization. Second, we implement a door-to-door survey in Rhode Island, finding significant flood risk underestimation and sorting based on risk perceptions and amenity values. Third, we estimate that coastal prices exceed fundamentals by 6%-13% in our benchmark area, with potentially higher overvaluation in other locations. Finally, we quantify both allocative inefficiency and distributional consequences arising from flood risk misperceptions and insurance policy reform.
This paper analyzes the effects of hurricane Sandy on the New York City housing market using a large parcel-level dataset that contains all housing sales for 2003–2017. The dataset also contains geo-coded FEMA data on which building structures were damaged by the hurricane and to what degree. Our estimates provide robust evidence of a persistent negative impact on flood zone housing values. We show the gradual emergence of a price penalty among flood zone properties that were not damaged by Sandy, reaching 8% in year 2017 and showing no signs of recovery. In contrast, damaged properties suffered a large immediate drop in value following the storm (17–22%), followed by a partial recovery and convergence toward a similar penalty as non-damaged properties. The partial recovery in the prices of damaged properties likely reflects their gradual restoration. However, the persistent price reduction affecting all flood-zone properties is more consistent with a learning mechanism. Hurricane Sandy may have increased the perceived risk of large-scale flooding episodes in that area.
Homes exposed to sea level rise (SLR) sell for approximately 7% less than observably equivalent unexposed properties equidistant from the beach. This discount has grown over time and is driven by sophisticated buyers and communities worried about global warming. Consistent with causal identification of long-horizon SLR costs, we find no relation between SLR exposure and rental rates and a 4% discount among properties not projected to be flooded for almost a century. Our findings contribute to the literature on the pricing of long-run risky cash flows and provide insights for optimal climate change policy.
The existence of the real estate brokerage industry is generally attributed to high transaction costs in real estate markets. Brokers are typically expected to market sellers' properties, assist in contract negotiations, and coordinate the post-contract tasks necessary to close transactions. Presumably, brokers can perform these duties at lower cost than sellers. In addition to cost efficiencies, brokers may also impact market outcomes. Numerous researchers have investigated whether or not the use of brokers as well as various broker actions, broker characteristics, and broker/seller legal relationships affect market outcomes in the form of price and/or, time-on-the-market effects. We extend this line of research by considering price, time-on-market, and probability of sale effects in relation to four specific broker strategies: public open houses, broker open houses, MLS virtual tours, and MLS photographs. The results indicate positive relationships between these strategies and house prices and mixed relationships between these strategies and probability of sale and time-on-market.
The Chesapeake Bay and its tributaries provide a range of recreational and aesthetic amenities, such as swimming, fishing, boating, wildlife viewing, and scenic vistas. Living in close proximity to the Bay improves access to these amenities and should be capitalized into local housing markets. We investigate these impacts in the largest hedonic analysis of water quality ever completed, with over 200,000 property sales across 14 Maryland counties. We use a spatially explicit water quality dataset, along with a wealth of landscape, economic, geographic, and demographic variables. These data allow a comprehensive exploration of the value of water quality, while controlling for a multitude of other influences. We also estimate several variants of the models most popular in current literature, with a focus on the temporal average of water quality. In comparing 1 year and 3 year averages, the 3 year averages generally have a larger implicit price. Overall, results indicate that water quality improvements in the Bay, such as those required by EPA's Total Maximum Daily Load, could yield significant benefits to waterfront and near-waterfront homeowners.
This research examines the productivity of real estate agents who acquire and maintain their real estate salesperson’s license for two years or less (ROOKIE) relative to more experienced agents who have been licensed agents for 10 years or more (VETERAN). Many individuals pursue a real estate license in search of riches rather than a career and, as such, may not have fully contemplated the relationship between the economy, the housing market, and agent experience. The findings of this study show that properties listed by ROOKIE agents will sell for approximately 10% less than those listed by more experienced agents. This finding is compounded by the fact that properties listed by these agents also endure a significantly longer marketing duration than those of more experienced agents. This is in contrast to properties listed by VETERAN agents which sell for approximately 2% more than those of their less experienced counterparts and did so 32% faster. Finally, while the inexperienced agent does not significantly influence the probability of a sale, the more experienced agent does significantly increase the probability of a successful transaction. These findings may prove to be quite useful to sellers, buyers, and brokers, as well as those seeking a career in real estate.
This study examines differences in net selling price for residential real estate across male and female agents. A sample of 2,020 home sales transactions from Fulton County, Georgia are analyzed in a two-stage least squares, geospatial autoregressive corrected, semi-log hedonic model to test for gender and gender selection effects. Although agent gender seems to play a role in naïve models, its role becomes inconclusive as variables controlling for possible price and time on market expectations of the buyers and sellers are introduced to the models. Clear differences in real estate sales prices, time on market, and agent incomes across genders are unlikely due to differences in negotiation performance between genders or the mix of genders in a two-agent negotiation. The evidence suggests an interesting alternative to agent performance: that buyers and sellers with different reservation price and time on market expectations, such as those selling foreclosure homes, tend to select agents along gender lines.