# RECORD 63065c2cad027c9578aace5fb84b3bd4ba000e2b # TITLE: Hurricane-Induced Discharges from Superfund Sites: A Hedonic Price Analysis of Real Estate Market Responses # YEAR: 2024.0 VENUE: Land Economics DOI: 10.3368/le.101.1.030423-0019r1 # AUTHORS: Mehrnoosh Asadi; Jean Dubé; Pallab Mozumder; Eric Strobl # PAGES: 36 CANDIDATE TABLE PAGES (1-based): [29, 30, 31, 32, 33, 34] # ABSTRACT: Abstract This study investigates the impact of hurricane-induced discharges from Superfund sites on residents’ risk perception and housing values in Harris County, Texas. Employing a difference-in-differences hedonic approach with property sales data from 2013 to 2019, we analyze how residents adjust their risk perception based on environmental hazard information and whether this affects the real estate market. Examining proximity to Superfund sites, we find a 27% reduction in price for houses located within half a mile of impacted sites, diminishing with distance. The findings provide valuable insights for enhancing hazardous waste management through targeted interventions, risk assessments, and response plans. ===== PAGE 1 (score 8, numeric-paren 0, table-words 0, var-words 8) ===== 1 Hurricane-Induced Discharges from Superfund Sites: A Hedonic Price Analysis of Real Estate Market Responses Mehrnoosh Asadia,*, Jean Dubéb, Pallab Mozumderc, Eric Strobld,e a Assistant Professor, Department of Economics, College of Saint Benedict/Saint John’s University, 37 S College Avenue, Saint Joseph, MN, 56374, USA, email: masadi001@csbsju.edu b Professeur, Université Laval, École Supérieure d'Aménagement et de Développement (ÉSAD), Québec, QC G1V 0A6, Canada. Phone number: 418 656-2131, email: jean.dube@esad.ulaval.ca c Professor, Department of Earth & Environment, Department of Economics, and Institute of Environment, Florida International University (FIU), AHC5-370, University Park Campus, FIU, 11200 SW 8th Street, Miami, FL 33199, USA. Phone number: 305-348-7146, email: mozumder@fiu.edu d Professor, Department of Economics, University of Bern, Hochschulstrasse 6, 3012 Bern, Switzerland. Phone number: +41 31 684 56 35, email: eric.strobl@unibe.ch e Research Professor, Department of Economics, University of Birmingham, Birmingham, B15 2TT United Kingdom. email: e.strobl@bham.ac.uk Abstract: This study investigates the impact of Hurricane-induced discharges from Superfund sites on residents' risk perception and housing values in Harris County, Texas. Employing a difference-in- difference Hedonic approach with property sales data from 2013 to 2019, we analyze how residents adjust their risk perception based on environmental hazard information and whether this affects the real estate market. Examining proximity to Superfund sites, we find a 27% reduction in price for houses located within half a mile of impacted sites, diminishing with distance. The findings provide valuable insights for enhancing hazardous waste management through targeted interventions, risk assessments, and response plans. Keywords: Superfund Sites, hurricane-induced discharges from Superfund Sites, difference-in-difference Hedonic approach, housing prices, Harris County, Texas JEL Code: Q51, Q54 * Corresponding author at Department of Economics, College of Saint Benedict/Saint John’s University, 37 S College Avenue, Saint Joseph, MN, 56374, USA, email: masadi001@csbsju.edu. doi:10.3368/le.101.1.030423-0019R1 by guest on October 20, 2026. Copyright 2024 Downloaded from ===== PAGE 29 (score 16, numeric-paren 0, table-words 2, var-words 8) ===== 29 TABLES Table 1: Descriptive statistics and definition of variables used in the analysis. Variables Mean SD Min Max Sale price (in nominal $US) 243123.20 178913.90 38700 1249900 Lot size (foot square) 8487.66 9550.11 719 216998 Age (year) 24.4738 22.7355 0 133 Number of bedroom 3.3349 0.5842 2 4 Number of bathroom 1.0440 0.2051 1 2 Quality index 2.5419 0.7001 1 4 Distance to closest park (in miles) 1.2621 0.8712 0.0121 9.8120 Distance to closest school (in miles) 0.5553 0.3630 0.0141 4.8438 Distance to closest rail (in miles) 1.5292 1.1434 0.0004 7.1191 Distance to CBD (in miles) 16.3517 7.5865 0.4693 43.2705 Distance to closest highway (in miles) 0.7221 0.6115 0.0002 4.3443 Distance to nearest Superfund Site flooded (in miles) 13.4039 7.7813 0.1746 42.3446 Distance to nearest Superfund Site non-flooded (in miles) 6.5460 3.6611 0.0200 24.9993 Entire sample consisting of 5162 observations. by guest on October 20, 2026. Copyright 2024 Downloaded from ===== PAGE 30 (score 30, numeric-paren 0, table-words 3, var-words 7) ===== 30 Table 2: Descriptive statistics and definition of variables used in the analysis (treatment group vs control group) Before matching Treatment Control Variables Mean SD Min Max Mean SD Min Max Sale price (in nominal $US) 135568.40 95174.08 39000.00 1112500.00 241020.80 184057.60 38746.00 1249900.00 Lot size (foot square 8002.68 4935.67 1416.00 38956.00 6545.92 4410.80 1403.00 41060.00 Age (year) 50.7685 27.9000 0.0000 119.0000 39.3402 32.8095 0.0000 128.0000 Number of bedroom 2.9512 0.6273 2.0000 4.0000 3.0099 0.6216 2.0000 4.0000 Number of bathroom 1.0173 0.1306 1.0000 2.0000 1.0276 0.1639 1.0000 2.0000 Quality index 3.2693 0.5342 1.0000 4.0000 2.7135 0.7800 1.0000 4.0000 Distance to closest park (mile) 0.8399 0.6046 0.0348 2.7063 0.7954 0.4880 0.0347 2.6766 Distance to closest school (mile) 0.5078 0.3823 0.0255 2.0326 0.4687 0.3517 0.0148 2.5410 Distance to closest rail (mile) 0.6077 0.4445 0.0105 2.0218 0.7762 0.6957 0.0007 3.7410 Distance to CBD (mile) 9.2985 6.0508 1.9181 20.7789 7.9820 6.0022 0.4693 22.0822 Distance to closest highway (mile) 0.3237 0.2481 0.0061 1.7462 0.4423 0.3664 0.0045 2.1187 Note: Total sample size is 635 observations for treatment area and 4527 observations for control area SD: standard deviation After Matching Treatment Control Variables Mean SD Min Max Mean SD Min Max Sale price (in nominal $US) 136418.80 97184.71 39000.00 1112500.00 249555.60 193201.50 38746.00 1249900.00 Lot size (foot square) 7320.46 3761.80 1416.00 38956.00 6281.09 4069.38 1403.00 41060.00 Age (year) 51.2719 28.0661 0.0000 119.0000 42.3081 33.3166 0.0000 128.0000 Number of bedroom 2.9484 0.6331 2.0000 4.0000 2.9824 0.6292 2.0000 4.0000 Number of bathroom 1.0172 0.1302 1.0000 2.0000 1.0290 0.1678 1.0000 2.0000 Quality index 3.2806 0.5434 1.0000 4.0000 2.7008 0.8164 1.0000 4.0000 Distance to closest park (mile) 0.8603 0.6156 0.0348 2.7063 0.7671 0.4971 0.0347 2.6766 Distance to closest school (mile) 0.5008 0.3903 0.0255 2.0326 0.4139 0.2969 0.0148 2.5410 Distance to closest rail (mile) 0.5795 0.4068 0.0105 1.9979 0.8051 0.7329 0.0007 3.7410 Distance to CBD (mile 8.3676 5.4514 1.9181 20.6712 6.9476 5.4109 0.4693 21.2468 Distance to closest highway (mile) 0.3158 0.2391 0.0061 1.7462 0.4449 0.3656 0.0045 2.1187 by guest on October 20, 2026. Copyright 2024 Downloaded from ===== PAGE 31 (score 28, numeric-paren 17, table-words 4, var-words 1) ===== 31 Table 3: Estimation results (Without matching). transactions within 1.5-mile from Superfund site ATE Zones Exponential Gaussian Independent variables Treatment (home sales within 1.5 miles) -0.6223 *** -0.4958 *** -0.5678 *** -0.5686 *** (0.0946) (0.1071) (0.0888) (0.0887) Treatment × After Harvey 0.1130 (0.1100) Treatment × After Harvey × (<0.5 mile) -0.2772 ** (0.1268) Treatment × After Harvey × (<1 mile) -0.1333 * (0.0718) Treatment × After Harvey × Distance to Superfund Sites † -0.4415 * -0.1699 ** (0.2296) (0.0742) Structural Characteristics Yes Yes Yes Yes Spatial Characteristics Yes Yes Yes Yes Year Fixed Effects Yes Yes Yes Yes Parallel Trend Yes Yes Yes Yes Spatial fixed effect (census tract) ‡ Yes Yes Yes Yes Number of observations 2409 2409 2409 2409 R-squared 0.6394 0.6420 0.6393 0.6394 Note: Standard errors are clustered within the census tract and reported in parentheses. *p < 0.1. **p < 0.05. ***p < 0.01 † Two mathematical transformations are considered to extrapolate the impact locally. ‡ Information is derived from the first 3-digit or 4-digit census tract. When there are more than 500 observations available within a 3-digit tract, the census tract is subdivided into a 4-digit tract. by guest on October 20, 2026. Copyright 2024 Downloaded from ===== PAGE 32 (score 28, numeric-paren 17, table-words 4, var-words 1) ===== 32 Table 4: Estimation results (With matching). transactions within 1.5-mile from Superfund site ATE Zones Exponential Gaussian Treatment (home sales within 1.5 miles) -0.6067 *** -0.4861 *** -0.5566 *** -0.5580 *** (0.1013) (0.1117) (0.0907) (0.0906) Treatment × After Harvey 0.1029 (0.1226) Treatment × After Harvey × (<0.5 mile) -0.2876 ** (0.1279) Treatment × After Harvey × (<1 mile) -0.1262 * (0.0682) Treatment × After Harvey × Distance to Superfund Sites † -0.5238 ** -0.1945 *** (0.2123) (0.0688) Structural Characteristics Yes Yes Yes Yes Spatial Characteristics Yes Yes Yes Yes Year Fixed Effects Yes Yes Yes Yes Parallel Trend Yes Yes Yes Yes Spatial fixed effect (census tract) ‡ Yes Yes Yes Yes Number of observations 1963 1963 1963 1963 R-squared 0.6500 0.6527 0.6500 0.6502 Note: Standard errors are clustered within the census tract and reported in parentheses. *p < 0.1. **p < 0.05. ***p < 0.01 † Two mathematical transformations are considered to extrapolate the impact locally. ‡ Information is derived from the first 3-digit or 4-digit census tract. When there are more than 500 observations available within a 3- digit tract, the census tract is subdivided into a 4-digit tract. by guest on October 20, 2026. Copyright 2024 Downloaded from ===== PAGE 33 (score 42, numeric-paren 29, table-words 4, var-words 1) ===== 33 Table 5: Estimation results with DDD (without matching) ATE Zones Exponential Gaussian Treatment -0.1283 ** -0.0691 -0.1027 * -0.1020 * (0.0520) (0.0510) (0.0527) (0.0524) Control 0.2878 ** 0.2474 * 0.2794 * 0.2789 * (0.1435) (0.1428) (0.1413) (0.1421) Treatment × After Harvey 0.0678 (0.1105) Treatment × After Harvey × (<.5 mile) -0.3371 *** (0.1013) Treatment × After Harvey × (<1 mile) -0.1181 * (0.0604) Treatment × After × Distance to Superfund Site ‡ -0.6172 ** -0.1885 *** (0.2429) (0.0703) Control × After Harvey -0.0657 (0.0520) Control × After Harvey × (<.5 mile) 0.0899 (0.0582) Control × After Harvey × (<1 mile) 0.0022 (0.0467) Control × After Harvey × Distance to Superfund Sites † -0.1889 -0.0250 (0.1902) (0.0841) Structural Characteristics Yes Yes Yes Yes Spatial Characteristics Yes Yes Yes Yes Year Fixed Effects Yes Yes Yes Yes Parallel Trend Yes Yes Yes Yes Spatial fixed effect (census tract) ‡ Yes Yes Yes Yes Number of observations 5162 5162 5162 5162 R-squared 0.6832 0.6846 0.6832 0.6831 Note: Standard errors are clustered within the census tract and reported in parentheses. *p < 0.1. **p < 0.05. ***p < 0.01 † Two mathematical transformations are considered to extrapolate the impact locally. ‡ Information is derived from the first 3-digit or 4-digit census tract. When there are more than 500 observations available within a 3-digit tract, the census tract is subdivided into a 4-digit tract. by guest on October 20, 2026. Copyright 2024 Downloaded from ===== PAGE 34 (score 42, numeric-paren 29, table-words 4, var-words 1) ===== 34 Table 6: Estimation results with DDD (with matching) ATE Zones Exponential Gaussian Treatment -0.1727 *** -0.1164 * -0.1483 ** -0.1476 ** (0.0653) (0.0614) (0.0633) (0.0628) Control 0.2995 ** 0.2656 * 0.2917 ** 0.2907 ** (0.1424) (0.1401) (0.1398) (0.1408) Treatment × After Harvey 0.0629 (0.1193) Treatment × After Harvey × (<.5 mile) -0.3454 *** (0.1054) Treatment × After Harvey × (<1 mile) -0.0949 (0.0592) Treatment × Distance to Superfond Site ‡ -0.6705 ** -0.2055 *** (0.2602) (0.0728) Control × After Harvey -0.0567 (0.0559) Control × After Harvey × (<.5 mile) 0.1155 (0.0896) Control × After Harvey × (<1 mile) -0.0182 (0.0445) Control × After Harvey × Distance to Superfund Sites † -0.1927 -0.0128 (0.2050) (0.0914) Structural Characteristics Yes Yes Yes Yes Spatial Characteristics Yes Yes Yes Yes Year Fixed Effects Yes Yes Yes Yes Parallel Trend Yes Yes Yes Yes Spatial fixed effect (census tract) ‡ Yes Yes Yes Yes Number of observations 4444 4444 4444 4444 R-squared 0.6944 0.6958 0.6944 0.6944 Note: Standard errors are clustered within the census tract and reported in parentheses. *p < 0.1. **p < 0.05. ***p < 0.01 † Two mathematical transformations are considered to extrapolate the impact locally. ‡ Information is derived from the first 3-digit or 4-digit census tract. When there are more than 500 observations available within a 3-digit tract, the census tract is subdivided into a 4-digit tract. by guest on October 20, 2026. Copyright 2024 Downloaded from ===== PAGE 35 (score 1, numeric-paren 0, table-words 0, var-words 1) ===== 35 FIGURES Figure 1: Marginal effect of the Impacted Superfund sites on home values Figure 1.a: Estimation with matching Figure 1.b: Estimation without matching by guest on October 20, 2026. Copyright 2024 Downloaded from