# RECORD b4d90ac556c12775841d148b92b32c2089b062dd
# TITLE: The Effects of Landfills on Rural Residential Property Values: Some Empirical Evidence
# YEAR: 2000.0  VENUE: Journal of Regional Analysis & Policy  DOI: 10.68479/001c.9345
# AUTHORS: R. Bouvier; J. M. Halstead; K. Conway; A. Manalo
# PAGES: 15  CANDIDATE TABLE PAGES (1-based): [9]
# ABSTRACT: The question of whether solid waste landfills affect residential property values has long been a subject of debate. Past research has resulted in mixed conclusions. The current study examines six landfills, which differ in size, operating status, and history of contamination. The effect of each landfill is estimated by the use of multiple regression. In five of the landfills, no statistically significant evidence of an effect was found. In the remaining case, evidence of an effect was found, indicating that houses in close proximity to this landfill suffered an average loss of about six percent in value.


===== PAGE 1 (score 5, numeric-paren 0, table-words 0, var-words 5) =====

 
The Effect of Landfills on Rural Residential 
Property Values: Some Empirical Evidence 
 
Rachel A. Bouvier, John M. Halstead, Karen S. Conway, and 
Alberto B. Manalo* 
 
Abstract.  The question of whether solid waste landfills affect residential 
property values has long been a subject of debate.  Past research has re-
sulted in mixed conclusions.  The current study examines six landfills, 
which differ in size, operating status, and history of contamination.  The 
effect of each landfill is estimated by the use of multiple regression.  In 
five of the landfills, no statistically significant evidence of an effect was 
found.  In the remaining case, evidence of an effect was found, indicating 
that houses in close proximity to this landfill suffered an average loss of 
about six percent in value. 
 
1. Introduction 
 
The United States is one of the leading solid waste generators in the 
world.  Estimates from the Environmental Protection Agency predict that by 
the year 2010, the United States will be generating municipal solid wastes at 
the unprecedented rate of 250 million tons per year (US EPA, 1990).  Despite 
recent improvements in recycling and incinerating techniques, the sanitary 
landfill remains the most widely used method of disposal.   
The question of whether solid waste landfills affect residential property 
values has long been a subject of debate.  The hedonic technique, which at-
tempts to infer people’s preferences by the way they behave in the market, 
has been used in the past for this purpose, with widely differing results.  This  
 
* Rachel Bouvier is graduate research assistant, Department of Economics, University of Massa-
chusetts and Lecturer, Mount Holyoke College; John M. Halstead is Professor, Department of 
Resource Economics and Development, University of New Hampshire; Karen Conway is Asso-
ciate Professor of Economics, University of New Hampshire; and Alberto Manalo is Associate 
Professor, Department of Resource Economics and Development, University of New Hamp-
shire. 
 
 
JRAP (2000)30:2                                                                      
 


===== PAGE 9 (score 169, numeric-paren 136, table-words 5, var-words 2) =====

Effects of Landfills on Rural Residential Property Values                                                                           31 
 
rate, but if the structure of the housing markets differs, the intercept would 
not necessarily capture the difference.   
 
 
Table 3.  Results of Landfill Property Value Model 
TOWN 
PEPPERELL 
HUDSON 
BELCHERTOWN 
CLINTON 
LEICESTER 
WARE 
Adj. R2 
0.5506 
0.7335 
0.7034 
0.5466 
0.4127 
0.6215 
F stat: 
5.737 
11.275 
16.881 
5.662 
3.115 
8.225 
Prob> F: 
0.0001 
0.0001 
0.0001 
0.0001 
0.0033 
0.0001 
Number 
of obs. 
59 
47 
101 
59 
46 
67 
 
Variable 
 
Coefficient Variable (Standard Error)a 
Intercept 
10.0756*** 
(0.3842) 
9.8012*** 
(0.3795) 
10.4123*** 
(0.2798) 
11.0511* 
(0.5332) 
10.8524*** 
(0.4942) 
9.9469*** 
(0.3218) 
MILF1 
-0.0627* 
(0.0354) 
-0.0280 
(0.0486) 
-0.0065 
(0.0271) 
-0.0099 
(0.0207) 
0.2558 
(0.1588) 
0.0073 
(0.0309) 
MICBD1 
0.0386 
(0.0367) 
0.0522*** 
(0.0193) 
-0.0085*** 
(0.0265) 
0.0298*** 
(0.0167) 
-0.0082 
(0.0238) 
-0.0139 
(0.0249) 
MIHWY1 
0.0001 
(0.0008) 
-0.0017 
(0.0016) 
-0.0008 
(0.0010) 
-0.0012 
(0.0022) 
-0.0038** 
(0.0018) 
-0.0012 
(0.0022) 
AGE 
-0.0021 
(0.0050) 
-0.0057*** 
(0.0015) 
-0.0045*** 
(0.0014) 
-0.0083*** 
(0.0047) 
-0.0086** 
(0.0032) 
-0.0028 
(0.0018) 
AGE2 
0.0000 
(0.0002) 
0.00002** 
(0.00001) 
2.16e-04*** 
(7.15e-05) 
0.000053 
(0.000039) 
0.0004* 
(0.0002) 
4.86e-05 
(1.00e-04) 
BDRM 
0.0650** 
(0.0315) 
0.1063** 
(0.0403) 
0.0730*** 
(0.0246) 
0.0576*** 
(0.0323) 
-0.0178 
(0.0234) 
0.0920*** 
(0.0326) 
OTHRM 
0.0169 
(0.0246) 
0.0213 
(0.0180) 
0.0843*** 
(0.0191) 
0.1069* 
(0.0348) 
0.0078 
(0.0314) 
0.0445* 
(0.0244) 
BTHRM 
0.1236*** 
(0.0345) 
0.0041 
(0.0471) 
0.1191*** 
(0.0381) 
-0.036 
(0.0611) 
0.1610 
(0.0819) 
0.0235 
(0.0406) 
LNLTSZ 
0.1171*** 
(0.0316) 
0.1639*** 
(0.0370) 
0.0601** 
(0.0265) 
0.0255 
(0.05122) 
0.0056 
(0.0419) 
0.01066*** 
(0.0305) 
GRACE 
0.1107*** 
(0.0316) 
0.0397 
(0.0427) 
0.0183 
(0.0351) 
0.1298** 
(0.0552) 
0.0629 
(0.0679) 
0.1602*** 
(0.0547) 
FPLACE 
0.0896** 
(0.0414) 
0.0624** 
(0.0284) 
0.0008 
(0.0330) 
0.0842 
(0.0625) 
0.0935* 
(0.0534) 
0.0633 
(0.0477) 
POOL 
0.0194 
(0.0661) 
0.0901* 
(0.0522) 
0.0660 
(0.0599) 
-0.0160 
(0.1040) 
0.2151* 
(0.1118) 
0.0564 
(0.0781) 
SOLD92 
-0.0086 
(0.0427) 
0.1888** 
(0.0767) 
0.1039** 
(0.0478) 
0.1018 
(0.0955) 
-0.0268 
(0.2175) 
0.0681** 
(0.0787) 
SOLD93 
0.0262 
(0.0526) 
0.0759 
(0.0549) 
0.0260 
(0.0453) 
-0.0284 
(0.0756) 
-0.0078 
(0.2233) 
0.0336 
(0.0843) 
SOLD94 
0.0528 
(0.0430) 
0.0756 
(0.0527) 
0.0344 
(0.0445) 
-0.0992 
(0.0775) 
-0.0243 
(0.2156) 
-0.1101 
(0.0795) 
aWhite corrected standard errors 
   ***statistically significant at the 99% level 
   **statistically significant at the 95% level 
   *statistically significant at the 90% level 
 
 


===== PAGE 10 (score 26, numeric-paren 0, table-words 12, var-words 2) =====

32 
                                                            Bouvier, Halstead, Conway & Manalo 
7. Results of Individual Regressions 
 
Results of each regression are summarized in Table 3.  F-statistics for all 
six models indicate that all were significant at the one percent level.  The 
more salient features of the regressions are highlighted here.  Three of the 
data sets (Pepperell, Clinton and Hudson) showed evidence of model mis-
specification when Ramsey ReSET tests were performed.  Some ex post tests 
were then performed on the functional form of the model to determine if the 
Ramsey ReSET tests would yield different results under different functional 
forms.  However, the results did not change significantly.  Therefore, the 
model remains in its current form, with a cautionary word as to the presence 
of misspecification in three of the models. 
The Pepperell data set generated an adjusted R2 of 0.5506 and an F statis-
tic of 5.737, indicating that the null hypothesis (the model explains none of 
the variation in housing price) can be decisively rejected.  As noted below, 
since Pepperell was the only model where heteroskedasticity was present, 
corrected (White) standard errors are reported with the results in Table 3.  
All of the variable coefficients had the hypothesized sign.  The landfill vari-
able coefficient is negative, as hypothesized, and it is statistically significant 
at the ten percent level.    
  In Hudson, although the adjusted R2 of 0.7335 attests to the predictive 
power of the model, the distance to the landfill variable coefficient was not 
statistically significant at the five or ten percent level (although it did have 
the hypothesized sign).  There is nothing unexpected in the model; most of 
the estimated coefficients were statistically significant and had the hypothe-
sized sign.   
The Belchertown data set yielded an adjusted R2 of 0.7034.  Of the nine 
structural characteristics, six were significant at the five percent level and 
had the hypothesized sign.  Although the estimated coefficient on the landfill 
variable is the hypothesized sign, a 95 percent confidence interval around the 
estimated coefficient does include zero.  Thus, there is no statistically signifi-
cant evidence of an effect of the Belchertown landfill on the surrounding 
property values. 
In the Clinton data set, only two variable coefficients emerged statisti-
cally significant at the five percent level and three more variables were sig-
nificant at the ten percent level.  The Clinton model’s adjusted R2 is 0.5466.   
Again, the distance from the landfill variable coefficient is of the hypothe-
sized sign, but statistically insignificant. 
Leicester’s regression resulted in an adjusted R2 of only 0.4125.  This 
means that the model has the weakest explanatory power of the six towns;  
however, it is still statistically significant at the five percent level.  The coeffi-
cient on the landfill variable is positive, which is contrary to expectations 
(although it is statistically insignificant at the five percent level).  Two of the 
variable coefficients in Leicester's regression are statistically significant at the 
five percent levels;  four are statistically significant at the ten percent level.