# RECORD eeedb48b06f3660101615bf3de6b27e1951880ee
# TITLE: Impact of public water supply unreliability on residential property prices in Marion County, West Virginia
# YEAR: 2021.0  VENUE: Agricultural and Resource Economics Review  DOI: 10.1017/age.2021.24
# AUTHORS: Fahad Alzahrani; A. Collins
# PAGES: 25  CANDIDATE TABLE PAGES (1-based): [8, 11, 12, 15, 21, 22, 23, 24]
# ABSTRACT: Abstract Water supply unreliability in many public water systems stems from aging infrastructure. We measure unreliability by the issuance of boil water notices (BWNs) within one year prior to single-family residential sale observations. Using a spatial quantile regression framework on transactions between 2012 and 2017, we find statistically significant, negative relationships between BWNs and residential properties. The estimated impacts of unreliability on residential housing prices, however, are not uniform across the distribution of prices. Specifically, we find that BWNs have a larger impact on medium- to low-priced houses (at or below the 60 percent quantile) compared with high-priced houses. An aggregate marginal willingness-to-pay value of $4.2 million was computed for a one-day reduction in annual BWN throughout Marion County.


===== PAGE 1 (score 8, numeric-paren 0, table-words 1, var-words 6) =====

RESEARCH ARTICLE
Impact of public water supply unreliability on
residential property prices in Marion County,
West Virginia
Fahad Alzahrani1
and Alan R. Collins2*
1King Faisal University, Al-Ahsa, Eastern Province, Saudi Arabia and 2West Virginia University,
Morgantown, WV, USA
*Corresponding author. Email: alan.collins@mail.wvu.edu
(Received 1 February 2021; revised 28 September 2021; accepted 14 October 2021; first published online 
Abstract
Water supply unreliability in many public water systems stems from aging infrastructure.
We measure unreliability by the issuance of boil water notices (BWNs) within one year
prior to single-family residential sale observations. Using a spatial quantile regression
framework on transactions between 2012 and 2017, we find statistically significant, nega-
tive relationships between BWNs and residential properties. The estimated impacts of
unreliability on residential housing prices, however, are not uniform across the distribu-
tion of prices. Specifically, we find that BWNs have a larger impact on medium- to
low-priced houses (at or below the 60 percent quantile) compared with high-priced
houses. An aggregate marginal willingness-to-pay value of $4.2 million was computed
for a one-day reduction in annual BWN throughout Marion County.
Keywords: boil water notices; hedonic property price model; spatial quantile regression; water
infrastructure
Introduction
Studies have shown that even developed countries, like the U.S., face water challenges due to
problems associated with affordability, demographic changes, high environmental quality
expectations, and aging infrastructure (Wescoat, Headington, and Theobald 2007; OECD
2016; Allaire, Wu, and Lall 2018; Allen et al. 2018). Examining affordability concerns,
Mack and Wrase (2017) project increasing water affordability issues with raising water
rates, particularly in the state of West Virginia, where the highest percentage of at-risk cen-
sus tracts (46 percent) for households unable to afford water bills of any state in the nation.
Aging water infrastructure, coupled with major funding shortfalls, is threatening the
nation’s water security (Alfredo et al. 2016).1 Many of the pipes and main lines that
© The Author(s), 2021. Published by Cambridge University Press on behalf of the Northeastern Agricultural and Resource
Economics Association. This is an Open Access article, distributed under the terms of the Creative Commons Attribution
licence (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted re-use, distribution, and reproduction in
any medium, provided the original work is properly cited.
1Water infrastructure involves what is constructed to pump, divert, transport, treat, store, and deliver
water, as well as to collect, treat, and discharge storm and wastewater.
doi:10.1017/age.2021.24
Agricultural and Resource Economics Review (2022), 51, 105–129
15 November 2021)
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press


===== PAGE 8 (score 27, numeric-paren 9, table-words 6, var-words 4) =====

Unlike OLS, quantile regression specifies changes in the conditional quantile, which
allows us to observe how housing characteristics are valued differently by consumers of
housing at different price ranges. Similar to equation 8, the spatial quantile lag model is
expressed as
ln(P) = r(t)W ln(P) + ZB(t) + 1(t)
(9)
which will result in estimates based on the conditional quantile, taking account of the
spatial effect at the same time.
To estimate the spatial quantile model, the two-stage quantile regression approach
proposed by Kim and Muller (2004) is used to account for the endogeneity in the spa-
tially lagged variable. In the first stage, the spatially lagged exogenous variables WZ and
Z are used to predict the spatially lagged endogenous variable W ln(P) at an individual
quantile. The predicted

W ln(P) is substituted for W ln(P) in the spatial lag model to
eliminate the correlation between the spatially lagged endogenous variable and the
error term. Then, the second-stage regression for that quantile is performed to obtain
ρ(τ) and β(τ). This two-stage procedure is then repeated for each quantile.
Data
Our analysis is based on two main sources of data: (1) West Virginia Office of
Environmental Health Services (OEHS) and (2) Marion County Assessor. The next
two subsections provide descriptions of these data.
Boil water notices
To measure water supplier unreliability, we use BWNs issued by either public water sys-
tems or the Marion County Health Department. The OEHS provides information on all
BWNs in West Virginia starting from 2012. Between January 2012 and December 2017,
there were a total of 440 BWNs in Marion County, varying annually between 64 and 80
with no discernible trend over time. For each notice, the provided information includes
the following: when the notice is issued and lifted, the name of the public water system
plus its PWSID, the public health sanitation district where the notice is issued, the
Table 2. Boil Water Notices (BWNs) means by price quantile
Price quantile
Mean number of
BWNs days per year
Mean of BWN experienced
one year prior to sale
$31,928 (0.1)
4.694
0.811
$59,864 (0.2)
1.907
0.471
$81,286 (0.3)
0.890
0.313
$102,267 (0.4)
1.314
0.319
$124,881 (0.5)
0.995
0.301
$144,611 (0.6)
1.185
0.291
$167,371 (0.7)
0.540
0.268
$210,602 (0.8)
0.549
0.228
$301,113 (0.9)
0.505
0.162
Alzahrani and Collins
112
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press


===== PAGE 9 (score 25, numeric-paren 2, table-words 1, var-words 21) =====

reason for the notice, and details about the affected areas or street names.4 Using infor-
mation from the SDWIS/Fed database and details from each notice, we obtained city
and county affected by each notice.5
Housing data
The Marion County Assessor provides data for all residential property sales between
January 2012 and December 2017. This data set contains 7,972 residential property
transactions. By dropping all property sales without structures, the number of transac-
tions declined to 5,525. To ensure only arm’s length sales in our analysis, a process fol-
lowing Herriges, Secchi, and Babcock (2005) is used to drop all properties that were less
than 50 percent of their assessed values and/or sold for less than $5,000. Therefore, after
eliminating these sales, the total number of transactions included in our analysis is
1,985 single-family residential properties.6
To capture housing attributes previously found in the literature to influence housing
prices, we include variables for housing characteristics in the hedonic model such as
size, story height, age, number of bedrooms, total number of bathrooms, and existence
of house amenities (e.g., air-conditioning and fireplace). In addition, we control for two
variables that have been consistently neglected in hedonic property price models: (1)
house physical condition and (2) material and workmanship quality. Based upon asses-
sor data for physical condition, we rank houses from Unsound = 1 to Excellent = 6. For
quality, we use the quality grade factor provided by the county’s assessor ranging from
E = Poor (0.5) to X = Excellent (2.5). According to the assessor, these two variables are
unrelated, as the first one measures only the physical condition, but the second mea-
sures the quality of the building regardless of the physical condition. We compute a
simple correlation coefficient, and we find that it is equal to 0.60, indicating a moderate
correlation between the physical condition and the quality of the building.
All property sales are geocoded using house’s street address with ArcGIS to measure
neighborhood and proximity characteristics such as crime, school quality, distance to
Fairmont State University, and to the nearest highway interchange.7 For crime, we
use the total crime index that compares the average local crime level to that of the entire
U.S. (an index of 100 is average). This index includes both property and violent crimes
for 2010 at the block group level, and it is obtained from the ArcGIS database. For
school quality, we follow the literature and use proficiency tests as a proxy
(Brasington 1999). Specifically, we use the percentage of students in elementary school
who scored at or above proficiency levels in math tests. We obtain these shapefiles for
attendance zones in the county from the National Center for Education Statistics
(NCES) and the data for math tests from the GreatSchools website. For distance
4PWSID is an identification code that starts with the state’s initials and followed by seven numbers.
These codes can be used to find more information about the water system (e.g., city and county served
or the source of the primary water) in the Safe Drinking Water Information System/Federal Warehouse
(SDWIS/Fed).
5Some water systems were not available on SDWIS/Fed, so we used the WV Drinking Water Watch to
obtain the necessary information. Available at: http://129.71.204.189:1977/DWWpublic/index.jsp.
6County Assessor data did not include information about the source of potable water for each property.
Given the high percentage of the county population being served by public water systems, all properties in
this data set are assumed to be supplied by public water for potable water needs as opposed to private wells.
7This variable was included since Fairmont serves a “bedroom” community for the more prosperous
communities of Morgantown to the north and Clarksburg to the south.
Agricultural and Resource Economics Review
113
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press


===== PAGE 11 (score 27, numeric-paren 19, table-words 2, var-words 4) =====

Results
First, a single market assumption for residential properties in Marion County is con-
firmed from our data using a Chow test on models comparing two data subsamples:
(1) the city of Fairmont and (2) the rest of Marion County. To examine the changing
impact of water supply unreliability over the distribution of housing prices, the spatial
quantile regression model results are reported in Appendix Tables A1 and A2. All esti-
mates for the spatial autoregressive coefficient (ρ) are positive and statistically signifi-
cant, meaning that a positive spatial similarity exists between residential property sale
prices. Focusing on the BWN variables with coefficients statistically different from
Table 3. Property transactions and BWNs by tax districts in Marion County, West Virginia, between 2012
and 2017
Number
District
Property
transactions
BWNs
Properties affected by
BWNs (%)
1
Barrackville
53
4
30 (56.6)
2
Fairmont District
78
20
23 (29.5)
3
Fairmont Citya
485
51
138 (28.5)
4
Grant Annexa
136
20
45 (33.1)
5
Union Citya
177
11
34 (19.2)
6
Winfield Citya
28
4
21 (75)
7
Fairview
9
3
1 (11.1)
8
Farmington
14
5
10 (71.4)
9
Grant District
236
20
32 (13.6)
10
Grant Town
19
2
12 (63.2)
11
Lincoln District
133
37
69 (51.9)
12
Mannington District
17
3
11 (64.7)
13
Mannington City
75
7
59 (78.7)
14
Monongah-Grantb
24
0
14 (58.3)
15
Monongah-Lincolnb
19
0
12 (63.2)
16
Paw Paw District
59
31
41 (69.5)
17
Rivesville
50
24
37 (74)
18
Union District
113
12
29 (25.7)
19
Winfield District
86
11
22 (25.6)
20
Worthington
5
1
4 (80)
21
White Hall
25
2
11 (44)
22
Pleasant Valley
144
17
45 (31.3)
Multiple Districts
—
65
—
Total
1,985
350
700 (35.3)
aPart of Fairmont City.
bBWNs that affect districts 14 and 15 usually affect other districts as well.
Agricultural and Resource Economics Review
115
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press


===== PAGE 12 (score 22, numeric-paren 0, table-words 2, var-words 16) =====

Table 4. Summary statistics and variable descriptions
Variable
Description
Unit
Average
St. Dev.
Min.
Max.
Dependent variable
House price
House’s sale price
$
135,338
81,082
6,000
650,000
Independent variables
BWN variables
Days BWNs_year
Number of days the house was under BWNs a year
before the sale
Days
1.40
5.52
0
116
BWN
1 = if the house was affected by a BWN, 0 = otherwise
0/1
0.35
0.48
0
1
Structural variables
SQFT
Building area in square feet
ft2
1,645.45
658.89
480
5,488
Story height
Story height
Number
1.27
0.42
1
3
Construction material
Material used in construction, 1 = if the material used is
Alum/Vinyl, 0 = otherwise
0/1
0.63
0.48
0
1
Physical condition
Physical condition of the house (1–6, 1 = unsound, 6 = excellent)
Number
4.75
0.82
1
6
Quality grade factor
Quality of the material and workmanship (0.5–2.5, 0.5 = E, 2.5 = X+)
Number
1.05
0.24
0.5
2.5
Age
Age of the house
Years
63
36
0
167
Bedrooms
Number of bedrooms
Number
2.80
0.76
1
12
Total bathrooms
Total number of bathrooms (full + half)
Number
1.88
0.81
1
6
AC
Heating system with AC, 1 = central with A/C, 0 = otherwise
0/1
0.77
0.42
0
1
Fireplace
1 = if there is a wood-burning fireplace, 0 = otherwise
0/1
0.31
0.46
0
1
Basement garage
1 = if there is basement garage, 0 = otherwise
0/1
0.29
0.45
0
1
Alzahrani and Collins
116
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press


===== PAGE 13 (score 13, numeric-paren 0, table-words 1, var-words 9) =====

Property variables
Lot size
Lot size
Acres
0.96
5.30
0.01
130.28
Sidewalk
1 = if there is a sidewalk, 0 = otherwise
0/1
0.06
0.24
0
1
Fence
1 = if there is a fence around the property, 0 = otherwise
0/1
0.49
0.50
0
1
Neighborhood and proximity variables
Crime
Total crime index
Number
54.05
24.28
19
101
School quality
Percentage of students scoring at or above proficiency levels in
the math test
%
48.91
14.26
29
71
Distance to highway exit
Distance to the nearest highway interchange (I-79)
Miles
2.63
2.68
0.09
17.47
Distance to FSU
Distance to Fairmont State University (FSU)
Miles
3.19
2.57
0.14
17.30
Note: Observations = 1,985.
Agricultural and Resource Economics Review
117
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press


===== PAGE 15 (score 47, numeric-paren 33, table-words 3, var-words 4) =====

Table 5. Two stage least square (2SLS) estimates for the spatial quantile models (dependent variable is ln (P))
2SLS
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Days
BWNs_year
−0.0065***
(0.0016)
−0.0664***
(0.0173)
−0.0271**
(0.0121)
−0.0155
(0.0095)
−0.0075
(0.0048)
−0.0064**
(0.0033)
−0.0049**
(0.0025)
−0.0048
(0.0034)
−0.0015
(0.0033)
0.0009
(0.0030)
BWN
−0.1853***
(0.0186)
−0.5392***
(0.0653)
−0.2971***
(0.0483)
−0.1750***
(0.0265)
−0.1017***
(0.0209)
−0.0676***
(0.0171)
−0.0509***
(0.0185)
−0.0449**
−(0.0449)
−0.0074
(0.0205)
0.0077
(0.0172)
Note: Bootstrapped standard errors in parentheses.
***p<0.01, **p<0.05.
Table 6. MWTP ($) for a change in water supply reliability from the estimated spatial models
2SLS
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Days BWNs_year
1,160
2,679
2,093
982
868
BWN
35,116
28,690
24,746
18,304
12,875
10,421
9,090
9,129
Notes: The values represent coefficients that are statistically significant at 5 percent; empty cells are for insignificant coefficients. Since BWN is a binary variable, its marginal impact is equal to:
Mean P in quantile * (exp(b) −1) * (1/1 −ρ).
Agricultural and Resource Economics Review
119
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press


===== PAGE 16 (score 8, numeric-paren 0, table-words 3, var-words 2) =====

spatial correlation impacts between houses impacted by the BWN that are embedded in
the MWTP estimates, (b) enhanced risk perceptions of households impacted by the
BWN as explained below, and (c) the capitalized defensive expenditure value may rep-
resent the cost of leaving the home temporarily while the BWN is in effect rather than
the economic cost of in-house responses.
Furthermore, we see that the MWTP values for consumers of low-priced houses are
greater than for consumers of high-priced houses. That is, a reduction of one day under
BWN per year will increase housing prices at the medium to lower end of the housing
price distribution by $2,100 to $2,700, whereas houses at the mid-range of the distribu-
tion will increase by $850 to $1,000 (Table 6). These impacts range from 0.60 percent to
8.4 percent of average sale prices across these quantiles. Houses at the 0.7 quantile and
above of the price distribution show no statistically significant impacts from the days of
the BWN variable. Finally, given the mean distribution of BWNs by quantile in Table 2,
a threshold value (π) of approximately 1.0 BWNs annually is indicated by the statisti-
cally insignificant results at the 0.7 quantile and above.
One interpretation of these results is that the greater prevalence of BWN for
lower-priced houses leads to greater risk perceptions and awareness about public water
supply problems among these housing consumers. Previous studies have shown that
consumers who experience more problems with their water supply perceive the associ-
ated health risks differently and were willing to pay more to fix water supply problems
than those consumers who experience water supply problems infrequently (Anadu and
Harding 2000; Genius and Tsagarakis 2006). As shown in Table 2, experience with and
duration of BWNs consistently declines with housing price quantiles.
Overall, water supply unreliability has a substantial impact on the value of residential
properties in Marion County. To calculate the aggregate MWTP across the entire
county for a one-day reduction of BWN, three factors are multiplied together: (1)
the average value of statistically significant coefficients from the spatial quantile regres-
sion ($1,694), (2) the estimated annual average number of owner-occupied housing
units that would be expected to be affected by the BWN in Marion County between
2013 and 2017 (6,150), and (3) finally, from the quantile regression results, a factor
of 0.4 as the percentage of houses would incur a price impact from these notices.
Following this procedure, we obtain an aggregate MWTP value of $4.2 million for a
one-day reduction of the annual BWN throughout Marion County.
Conclusions
Aging water infrastructure, combined with funding shortfalls, poses serious challenges
to local governments and water systems across the U.S. The number of water main
breaks in the U.S. has been increasing every year resulting in more interruptions in pub-
lic water service (Folkman 2018). Since achieving full reliability in water service is most
likely not an optimal policy, decision makers at local water utilities must attempt to bal-
ance between cost and risk associated with their policies (Howe et al. 1994). However,
water customers’ perceptions of the risk associated with supply disruptions are usually
unknown to decision makers. Therefore, studies have attempted to examine the WTP
for
water
supply
reliability
using
stated
preference
approaches
(Genius
and
Tsagarakis 2006; Vásquez and Espaillat 2016).
In this research, we examine the impact of water supply unreliability on single-
family, residential property sales in Marion County, West Virginia, using spatial quan-
tile hedonic property price models. To measure the impact of water supply unreliability,
Alzahrani and Collins
120
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press


===== PAGE 21 (score 198, numeric-paren 163, table-words 2, var-words 13) =====

Appendix
Spatial quantile regression models—dependent variable is ln (P)11
Table A1. Spatial quantile regression results for model (1)—days BWNs_year
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Days BWNs_year
−0.0664***
(0.0173)
−0.0271**
(0.0121)
−0.0155
(0.0095)
−0.0075
(0.0048)
−0.0064**
(0.0033)
−0.0049**
(0.0025)
−0.0048
(0.0034)
−0.0015
(0.0033)
0.0009
(0.0030)
SQFT
0.0002***
(0.0000)
0.0002***
(0.0000)
0.0002***
(0.0000)
0.0002***
(0.0000)
0.0002***
(0.0000)
0.0003***
(0.0000)
0.0002***
(0.0000)
0.0002***
(0.0000)
0.0002***
(0.0000)
Story height
−0.0575*
(0.0306)
−0.0303
(0.0283)
−0.046*
(0.0244)
−0.0413**
(0.0192)
−0.0323*
(0.0173)
−0.0266
(0.0198)
−0.0092
(0.0243)
0.0052
(0.0225)
−0.0063
(0.0222)
Construction material
−0.0672**
(0.0286)
−0.0668***
(0.0231)
−0.0636***
(0.0185)
−0.0525***
(0.0154)
−0.0297**
(0.0141)
−0.0298*
(0.0160)
−0.0285*
(0.0168)
−0.0256
(0.0174)
−0.0311*
(0.0159)
Physical condition
0.2683***
(0.0281)
0.2637***
(0.0217)
0.2266***
(0.0187)
0.1998***
(0.0146)
0.1758***
(0.0156)
0.1634***
(0.0158)
0.1472***
(0.0181)
0.1373***
(0.0167)
0.1473***
(0.0159)
Quality grade factor
0.4243***
(0.1245)
0.3806***
(0.0792)
0.3937***
(0.0677)
0.3875***
(0.0544)
0.3909***
(0.0488)
0.3472***
(0.0507)
0.364***
(0.0668)
0.4398***
(0.0532)
0.4374***
(0.0657)
Age
−0.0025***
(0.0007)
−0.0025***
(0.0005)
−0.0026***
(0.0005)
−0.0024***
(0.0003)
−0.0025***
(0.0004)
−0.0026***
(0.0004)
−0.0025***
(0.0004)
−0.0021***
(0.0004)
−0.0016***
(0.0003)
Bedrooms
0.0235
(0.0264)
0.0198
(0.0212)
0.0411**
(0.0164)
0.0487***
(0.0150)
0.0518***
(0.0126)
0.0463***
(0.0158)
0.0319**
(0.0152)
0.0448***
(0.0140)
0.0474***
(0.0133)
Total bathrooms
0.0059
(0.0320)
0.0258
(0.0192)
0.0372**
(0.0153)
0.0572***
(0.0134)
0.034***
(0.0123)
0.028**
(0.0142)
0.0279*
(0.0145)
0.0152
(0.0149)
0.0152
(0.0178)
AC
0.2074***
(0.0540)
0.2328***
(0.0335)
0.2202***
(0.0309)
0.2217***
(0.0250)
0.2169***
(0.0244)
0.2099***
(0.0259)
0.1711***
(0.0233)
0.1615***
(0.0271)
0.0909**
(0.0371)
(Continued)
Agricultural and Resource Economics Review
125
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press


===== PAGE 22 (score 185, numeric-paren 154, table-words 1, var-words 9) =====

Table A1. (Continued.)
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Fire place
0.1527***
(0.0362)
0.1059***
(0.0220)
0.113***
(0.0191)
0.1052***
(0.0138)
0.0905***
(0.0158)
0.0867***
(0.0150)
0.0799***
(0.0163)
0.0667***
(0.0176)
0.0553***
(0.0153)
Basement garage
0.0573**
(0.0263)
0.0169
(0.0203)
0.0298*
(0.0168)
0.0159
(0.0140)
0.0113
(0.0144)
0.0142
(0.0144)
0.0085
(0.0185)
0.0032
(0.0172)
−0.0058
(0.0179)
Lot size
−0.0009
(0.0075)
0.0088
(0.0065)
0.0112**
(0.0044)
0.0105**
(0.0050)
0.0117*
(0.0065)
0.0108
(0.0073)
0.0244**
(0.0102)
0.0212**
(0.0100)
0.0381***
(0.0070)
Sidewalk
−0.0753
(0.0730)
−0.0562
(0.0529)
−0.0216
(0.0462)
−0.035
(0.0325)
−0.0415
(0.0291)
−0.0376
(0.0298)
−0.0559*
(0.0315)
−0.0885***
(0.0298)
−0.0933**
(0.0410)
Fence
0.1351***
(0.0308)
0.1127***
(0.0249)
0.0967***
(0.0160)
0.0944***
(0.0155)
0.0903***
(0.0130)
0.0912***
(0.0163)
0.0764***
(0.0161)
0.0659***
(0.0156)
0.0434***
(0.0161)
Crime
−0.0004
(0.0010)
−0.0004
(0.0005)
−0.0007
(0.0005)
−0.0005
(0.0005)
−0.0002
(0.0004)
−0.0005
(0.0004)
−0.0008
(0.0005)
−0.0005
(0.0004)
−0.0006
(0.0005)
School quality
0.0014
(0.0014)
0.0001
(0.0011)
−0.0006
(0.0008)
−0.0014**
(0.0007)
−0.001
(0.0007)
−0.001
(0.0008)
−0.0009
(0.0010)
−0.0005
(0.0007)
−0.0007
(0.0007)
Distance to highway exit
−0.0046
(0.0116)
−0.007
(0.0085)
−0.0156**
(0.0072)
−0.0204***
(0.0061)
−0.0139**
(0.0058)
−0.0103*
(0.0062)
−0.0154**
(0.0064)
−0.0128**
(0.0064)
−0.014**
(0.0067)
Distance to FSU
−0.0093
(0.0106)
−0.0115
(0.0078)
−0.0072
(0.0060)
0.0005
(0.0054)
−0.0016
(0.0052)
−0.0044
(0.0052)
−0.0024
(0.0056)
0.0015
(0.0061)
0.0055
(0.0067)
ρ
0.2085***
(0.0439)
0.2248***
(0.0341)
0.1709***
(0.0302)
0.1705***
(0.0245)
0.1862***
(0.0248)
0.1841***
(0.0266)
0.1633***
(0.0288)
0.1628***
(0.0267)
0.1694***
(0.0264)
Constant
6.7475***
(0.5188)
6.7145***
(0.4221)
7.569***
(0.3615)
7.7436***
(0.2913)
7.6613***
(0.2727)
7.8642***
(0.3232)
8.3129***
(0.3527)
8.2941***
(0.3355)
8.2796***
(0.3264)
Notes: Bootstrapped standard errors in parentheses. The statistical software used for the estimation of spatial quantile regression, and the qregspiv command of the McSpatial package in the
software R. This command does not return goodness-of-fit measures for the regression results. Year-fixed effects are included in this model.
***p<0.01, **p<0.05, *p<0.1.
Alzahrani and Collins
126
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press


===== PAGE 23 (score 231, numeric-paren 192, table-words 1, var-words 16) =====

Table A2. Spatial quantile regression results for model (2)—BWN
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
BWN
−0.5392***
(0.0653)
−0.2971***
(0.0483)
−0.175***
(0.0265)
−0.1017***
(0.0209)
−0.0676***
(0.0171)
−0.0509***
(0.0185)
−0.0449**
−(0.0449)
−0.0074
(0.0205)
0.0077
(0.0172)
SQFT
0.0003***
(0.0000)
0.0002***
(0.0000)
0.0002***
(0.0000)
0.0002***
(0.0000)
0.0002***
(0.0000)
0.0003***
(0.0000)
0.0002***
(0.0002)
0.0002***
(0.0000)
0.0002***
(0.0000)
Story height
−0.0357
(0.0372)
−0.0317
(0.0285)
−0.0466**
(0.0205)
−0.0374*
(0.0197)
−0.0238
(0.0217)
−0.0326*
(0.0194)
−0.0065
−(0.0065)
0.0089
(0.0211)
−0.0044
(0.0208)
Construction
material
−0.078***
(0.0262)
−0.0762***
(0.0235)
−0.0519**
(0.0202)
−0.0567***
(0.0159)
−0.0407***
(0.0152)
−0.0311**
(0.0146)
−0.0275*
−(0.0275)
−0.0197
(0.0170)
−0.0308**
(0.0151)
Physical
condition
0.2086***
(0.0238)
0.2265***
(0.0227)
0.2058***
(0.0173)
0.2014***
(0.0173)
0.1733***
(0.0176)
0.1622***
(0.0153)
0.1478***
(0.1478)
0.1407***
(0.0166)
0.1474***
(0.0181)
Quality grade
factor
0.5079***
(0.0963)
0.4762***
(0.0734)
0.4752***
(0.0509)
0.4117***
(0.0578)
0.4025***
(0.0609)
0.3566***
(0.0585)
0.3477***
(0.3477)
0.4225***
(0.0632)
0.4505***
(0.0692)
Age
−0.0021***
(0.0006)
−0.0025***
(0.0005)
−0.0024***
(0.0004)
−0.0025***
(0.0005)
−0.0026***
(0.0004)
−0.0026***
(0.0003)
−0.0023***
−(0.0023)
−0.0023***
(0.0004)
−0.0017***
(0.0003)
Bedrooms
0.0165
(0.0161)
0.024
(0.0192)
0.0406**
(0.0166)
0.0463***
(0.0165)
0.0447***
(0.0154)
0.0481***
(0.0130)
0.0366**
(0.0366)
0.0433***
(0.0128)
0.047***
(0.0141)
Total bathrooms
−0.0178
(0.0211)
0.021
(0.0223)
0.0406***
(0.0148)
0.0454***
(0.0168)
0.0335***
(0.0118)
0.0273*
(0.0141)
0.037***
(0.0370)
0.0148
(0.0172)
0.0143
(0.0153)
AC
0.1847***
(0.0347)
0.2161***
(0.0372)
0.2305***
(0.0272)
0.2244***
(0.0271)
0.2281***
(0.0268)
0.2081***
(0.0295)
0.1684***
(0.1684)
0.1468***
(0.0287)
0.0936***
(0.0334)
Fire place
0.1368***
(0.0260)
0.1306***
(0.0234)
0.1319***
(0.0204)
0.1094***
(0.0159)
0.0972***
(0.0149)
0.0896***
(0.0159)
0.0836***
(0.0836)
0.077***
(0.0165)
0.0604***
(0.0171)
Basement garage
0.0209
(0.0253)
0.0326
(0.0214)
0.0232
(0.0184)
0.017
(0.0179)
0.0153
(0.0151)
0.0081
(0.0153)
−0.0006
−(0.0006)
0.0019
(0.0165)
−0.0104
(0.0181)
(Continued)
Agricultural and Resource Economics Review
127
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press


===== PAGE 24 (score 150, numeric-paren 126, table-words 1, var-words 6) =====

Table A2. (Continued.)
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
Lot size
0.0022
(0.0049)
0.0088
(0.0066)
0.0122**
(0.0050)
0.0111**
(0.0047)
0.01*
(0.0057)
0.0112
(0.0078)
0.022***
(0.0220)
0.0218**
(0.0099)
0.0372***
(0.0074)
Sidewalk
−0.0141
(0.0515)
−0.0541
(0.0345)
−0.0527
(0.0443)
−0.0385
(0.0334)
−0.0523
(0.0400)
−0.0374
(0.0312)
−0.0647**
−(0.0647)
−0.0876**
(0.0357)
−0.0889**
(0.0445)
Fence
0.0783***
(0.0277)
0.104***
(0.0226)
0.1003***
(0.0176)
0.0936***
(0.0182)
0.0993***
(0.0156)
0.0938***
(0.0170)
0.0729***
(0.0729)
0.066***
(0.0144)
0.0514***
(0.0155)
Crime
−0.0001
(0.0008)
−0.0009
(0.0007)
−0.0007
(0.0005)
−0.0004
(0.0004)
−0.0001
(0.0005)
−0.0004
(0.0004)
−0.0007*
−(0.0007)
−0.0006
(0.0005)
−0.0007
(0.0005)
School quality
0.0018
(0.0012)
0.00002
(0.0011)
−0.0008
(0.0009)
−0.0012
(0.0008)
−0.0005
(0.0007)
−0.0009
(0.0007)
−0.0008
−(0.0008)
−0.0005
(0.0007)
−0.0005
(0.0007)
Distance to
highway exit
−0.0158
(0.0104)
−0.0151*
(0.0089)
−0.02***
(0.0067)
−0.022***
(0.0065)
−0.0123*
(0.0071)
−0.0124**
(0.0058)
−0.0146***
−(0.0146)
−0.0163***
(0.0061)
−0.0141*
(0.0076)
Distance to FSU
−0.0024
(0.0094)
−0.0072
(0.0079)
−0.0032
(0.0064)
0.0014
(0.0061)
−0.0033
(0.0061)
−0.0006
(0.0048)
−0.0017
−(0.0017)
0.0012
(0.0056)
0.0043
(0.0065)
ρ
0.2047***
(0.0326)
0.1631***
(0.0298)
0.1507***
(0.0284)
0.1497***
(0.0236)
0.1619***
(0.0281)
0.1693***
(0.0274)
0.158***
(0.1580)
0.1442***
(0.0255)
0.1539***
(0.0263)
Constant
7.0339***
(0.3612)
7.5819***
(0.3441)
7.8347***
(0.3136)
7.9704***
(0.2879)
7.9343***
(0.3276)
8.0273***
(0.3113)
8.3479***
(8.3479)
8.5324***
(0.3105)
8.4602***
(0.3152)
Notes: Bootstrapped standard errors in parentheses. The statistical software used for the estimation of spatial quantile regression, and the qregspiv command of the McSpatial package in the
software R. This command does not return goodness-of-fit measures for the regression results. Year-fixed effects are included in this model.
***p<0.01, **p<0.05, *p<0.1.
Alzahrani and Collins
128
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press


===== PAGE 25 (score 3, numeric-paren 0, table-words 0, var-words 3) =====

Figure A1. Distance Histogram Between Neighbors in the Weight Matrix.
Cite this article: Alzahrani F, Collins AR (2022). Impact of public water supply unreliability on residential
property prices in Marion County, West Virginia. Agricultural and Resource Economics Review 51, 105–129.
https://doi.org/10.1017/age.2021.24
Agricultural and Resource Economics Review
129
https://doi.org/10.1017/age.2021.24 Published online by Cambridge University Press