THE EFFECT OF ETHANOL PLANTS ON RESIDENTIAL
PROPERTY VALUES: EVIDENCE FROM MICHIGAN
By
Timothy R. Hodge

A THESIS
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
MASTER OF SCIENCE
Agricultural, Food, and Resource Economics
2011

ABSTRACT
THE EFFECT OF ETHANOL PLANTS ON RESIDENTIAL
PROPERTY VALUES: EVIDENCE FROM MICHIGAN
By
Timothy R. Hodge
Since the mid 1990s, bio-fuel producers have built more than 130 ethanol plants across
the United States, the majority of which have been placed in the upper Midwest. While
politicians and the industry have praised the positive effects of ethanol facilities, it is important
to explore the potential negative impacts. This study examines one negative effect that is not yet
fully understood: the impact ethanol plants have on the value of residential property located near
a new ethanol facility.
To meet this objective, sales data for residential properties sold between 1999 and 2009
from two ethanol communities in Michigan and the hedonic method are used to evaluate the
impact on property values over time and across homes in each community. Use of sales data
over this ten-year period provides a unique analysis as it enables a comparison of properties preand post- plants coming on line. Furthermore, use of pre- and post-plant sales data provides
greater confidence that any observed negative effect is truly the result of the ethanol plant and
not some pre-existing, unobserved factor. Conclusions confirm that ethanol plants may have
large negative effects, depressing the value of homes as much as 18% and as far as two miles
away. However, these results may not be universal as conditions, tastes, and preferences differ
across space and time.

ACKNOWLEDGEMENTS

I am very grateful to my major professor and advisor, Dr. Mark Skidmore. His guidance
and encouragement throughout my years at Michigan State have made this and other projects not
only possible, but also extremely enjoyable. I am also grateful to my committee members, Dr.
Scott Loveridge and Dr. Pat Norris, whose efforts and suggestions have greatly improved the
quality of this research. I also thank Dr. Scott Swinton for providing insightful feedback during
initial presentations, and Nancy Corey for providing data from Caro.
I am also very grateful to my parents, without whom this project would not have been
possible. In addition to assisting with data collection and idea generation for this project, my
mother and father have provided invaluable support and direction throughout my life.
Finally, I thank my best friend and wife, Drea, for her constant love, support, and ability
to keep things in perspective.

iii

TABLE OF CONTENTS

LIST OF TABLES ............................................................................................................. v
LIST OF FIGURES .......................................................................................................... vi
1. INTRODUCTION .......................................................................................................... 1
2. LITERATURE REVIEW ............................................................................................... 6
2.1 Negative Effects of the Ethanol Industry ................................................................... 6
2.2 Measuring the Effect of Negative Externalities on Property Values ......................... 8
2.3 Contribution .............................................................................................................10
3. METHODS ................................................................................................................... 12
3.1 Hedonic Approach ................................................................................................... 12
3.2 Model ....................................................................................................................... 13
3.3 Structural Attributes (X) .......................................................................................... 17
3.4 Neighborhood Attributes (L) ................................................................................... 18
3.5 Time ......................................................................................................................... 18
3.6 Functional Form ....................................................................................................... 19
4. DATA ........................................................................................................................... 20
4.1 Areas of Study.......................................................................................................... 20
4.2 Data on House Prices and Desired Attributes .......................................................... 21
4.3 Additional Concerns ................................................................................................ 22
5. RESULTS ..................................................................................................................... 26
5.1 House Characteristics: Caro ..................................................................................... 26
5.2 Neighborhood Attributes: Caro................................................................................ 29
5.3 House Characteristics: Marysville ........................................................................... 30
5.4 Neighborhood Attributes: Marysville ...................................................................... 32
6. CONCLUSION ............................................................................................................. 36
APPENDICES ................................................................................................................. 38
APPENDIX A ............................................................................................................... 39
APPENDIX B ................................................................................................................ 40
APPENDIX C ................................................................................................................ 41
APPENDIX D ................................................................................................................ 42
APPENDIX E ............................................................................................................... 43
APPENDIX F ................................................................................................................ 45
REFERENCES ................................................................................................................ 47
iv

LIST OF TABLES

Table 1. St Clair River summary statistics.......................................................................33
Table 2. Further examination on the effect of the St Clair River.....................................33
Table 3. Description of variables…………......................................................................40
Table 4. Caro summary statistics......................................................................................41
Table 5. Marysville summary statistics............................................................................42
Table 6. Caro regression results………….......................................................................43
Table 7. Marysville regression results.............................................................................45

v

LIST OF FIGURES

Figure 1. Historical ethanol production (millions of gallons) ............................................. 1
Figure 2. Michigan ethanol plants location, name, start date, and nameplate production
capacity............................................................................................................................... 20
Figure 3. Location of select properties in Caro and Marysville......................................... 25
Figure 4. Average house prices for the Caro area............................................................... 28
Figure 5. Average house prices for the Marysville area..................................................... 32
Figure 6. Property sales less than $20,000.……………..................................................... 39

vi

1. INTRODUCTION
Ethanol production in the United States has steadily been on the rise since the mid 1990s
(Figure 1). As of January 2010, there were 189 ethanol plants in operation in the United States
with capacity totaling over 13 billion gallons per year. This capacity is expected to exceed 14.4
1

billion gallons per year once current projects are complete.

Production (Millions of Gallons)
12,000
11,000
10,000
9,000
8,000
7,000
6,000
5,000
4,000
3,000
2,000
1,000
0

Source: Renewable Fuels Association
Figure 1. Historical ethanol production (millions of gallons)

As noted by Hahn & Cecot (2009), the continued growth of the ethanol industry is
1

See the Renewable Fuels Association website for production details, http://ethanolrfa.org.
1

primarily the result of politicians and scientists seeing ethanol as a way to promote
environmental and energy security goals. To stimulate the production and use of ethanol to meet
these goals, significant strides in production levels have been spurred by generous subsidies and
government mandates at both national and state levels (Cotti & Skidmore, 2010). Recent
estimates highlight that incentives are accomplishing their intended effect as the industry
displaced the need for 364 million barrels of oil (approximately five percent) in 2009
(Urbanchuk, 2010).
In addition to the environmental and political benefits associated with the ethanol boom,
it is often cited that this growth is benefiting rural America by reshaping its economic base. As
Don Cumpton, the director of economic development in Hereford, Texas, stated in an interview
with the New York Times, “These projects are bringing 100 new jobs to our town. It's not as if
Dell computer's going to be setting up shop here. We'd be nuts to turn something like this down"
(Barrionuevo, 2006). Recent calculations have validated this claim, highlighting numerous
benefits to both local and national sectors of the economy. In a report prepared for the
Renewable Fuels Association, Urbanchuk (2010) estimated that the 2009 production levels
supported nearly 400,000 jobs in all sectors of the economy, contributed $53.5 billion to GDP,
and added $16 billion to household income.
Despite the allure of ethanol, the industry is not without skeptics and researchers in a
number of fields are analyzing the industry to fully understand its impact. Four general
conclusions highlighting the negative effects of ethanol production include: increasing grain and
food prices in local and world markets (McNew & Griffith, 2005; Runge & Senauer, 2007),
environmental degradation (Pimental, 1991; Pimental 2003; Niven, 2005; Searchinger et al.
2008), inefficient production (Pimental, 2003), and costs exceeding benefits (Gardner, 2007;

2

Hahn & Cecot, 2009). Specifically, Gardner (2007) concluded that the federal subsidies and
mandates in the ethanol industry generate long-run deadweight losses between three and four
billion dollars annually. Likewise, Hahn & Cecot (2009) estimated that the costs of increasing
ethanol production to ten billion gallons per year (e.g. production costs, distribution costs, and
increased emission costs) would exceed benefits (e.g. oil displacement, greenhouse gas emission
reductions, and air toxic emission reductions) by three billion dollars annually in 2012 (holding
policies fixed).
Although research highlighting the industry’s negative impact is growing, one subject
with minimum exposure is the effect of ethanol plants on surrounding residential property
values, a topic worth our attention if we are to fully capture the economic costs of the industry.
While Turnquist, Fortenbery, & Foltz (2008) have considered this issue at the municipal level,
providing a good first approximation of anticipated effects and insight into what might be
observed, a detailed analysis at the property level is required to fully understand and capture the
industry’s effect on the value of residential properties.
It is reasonable to anticipate that ethanol plants will have a negative effect on neighboring
property values. The odor emitted by a plant is considered offensive, even nauseating, to some
(Meersman, 2001). On this basis alone, ethanol plants may depress property values as
2

homebuyers may require a discounted price to be willing to live with the smell. Beyond this,
two additional negative externalities that may depress property values include: 1) toxic emissions
could escape the plant requiring evacuation for a large radius surrounding the plant, and 2)
increased truck traffic with associated noise and safety concerns.
2

Studies that have examined the negative effect of smell and air quality on property values
include: Harrison & Rubinfeld (1978), Nelson (1978), Zabel & Kiel (2000), and Saphores &
Aguilar-Benitez (2005). Although the variables used, areas of study, and methods employed
differ among these studies, the results were generally negative and statistically significant.
3

This study recognizes the externalities ethanol plants may impose on residents and
provides the first in-depth analysis examining the change in residential property values
associated with the presence of a new ethanol plant. To meet this objective, property level data
has been collected from two Michigan communities that currently have operating facilities. The
two communities in this study represent the range of communities in which plants are locating
across the country: one is a small farming town with a low amount of vegetation (other than
crops) and little pre-existing industry (Caro), while the other is more populated with a high
amount of vegetation and other already established industrial sites (Marysville).
The data compiled includes more than 600 residential sales in each community over a
ten-year period (1999-2009), and the analysis relies on the well-established hedonic method to
evaluate the impact on property values over time and across homes in each community. Hedonic
analysis has a long history and has been used to measure numerous externalities on property
values, ranging from the positive effect of community gardens (Been & Voicu, 2006) to the
negative impact of hazardous waste sites (McCluskey & Rausser, 2003). Using a ten-year
period, this study provides a unique hedonic analysis by directly comparing properties sold prior
to each plant’s production with those sold after operations begin. In addition, use of data over a
ten-year period provides greater confidence that any observed negative effect is truly the result of
the ethanol plant and not some pre-existing, unobserved factor. Conclusions confirm that
ethanol plants may have a large impact on property values, depressing the value of homes as
much as 18% and as far as two miles away. However, this conclusion may not be universal as
consumer tastes and preferences differ across space and time. In addition, conditions
surrounding residential properties may limit the impact of an ethanol plant. These conditions

4

include, but are not limited to, the inability to see the plant and the existence of pre-existing
industry in the general proximity of the new ethanol plant.
The following section provides an overview of all studies that consider the effects of
ethanol plants on neighboring property values. A review of these studies highlights the need for
an in-depth, property level analysis. In addition, this section includes a survey of previous
literature examining the impact of local negative externalities. Among these, special attention is
given to the techniques used and the wide range of findings. Section 3 focuses on the hedonic
method and how it allows researchers to monetize different property attributes, including
negative externalities. Details regarding the specifications used in the present study are also
discussed in Section 3. Details regarding the two cities and the data used in this study are
discussed in Section 4. The results are presented in Section 5, and Section 6 offers concluding
remarks.

5

2. LITERATURE REVIEW
2.1 Negative Effects of the Ethanol Industry
The negative effects of the ethanol industry that have been examined by researchers
include: costs exceeding benefits, increasing food and grain prices in both local and world
markets, environmental degradation, production inefficiency, and effects on neighboring
property values. Of key interest in this study is the effect of an ethanol plant on neighboring
property values. To date, four studies have examined the effect of an ethanol plant on
neighboring property values and the effect on residential properties is still not yet well
understood. Of these four studies, only two are directly relevant to the issue of residential
properties; two studies examine the impact of ethanol plants on agricultural land (Henderson &
Gloy, 2008; Blomendahl & Johnson, 2009). The two studies examining farmland are less
relevant to the present study as the value of residential property is based on its housing attributes
while agricultural property is valued on its anticipated future earnings. Stemming from this
difference, the opposite effect of an ethanol facility on farm properties is expected because
higher expected returns are anticipated as demand for local corn may increase with the
introduction of a plant (McNew & Griffith, 2005).
Turning our attention to those studies highlighting the impact of an ethanol plant on
residential property values, the first study was conducted in 2007 by a consulting firm hired by
the city of Portsmouth, Virginia. City officials were concerned about the potential impact that a
216 million gallon per year (MGY) ethanol plant would have on local property values. Using a
few Texas communities as a baseline, the report concluded that housing values could decline
between eight and forty-six percent for homes within a two-mile radius of the plant (Hoyer and
Saewitz, 2007). Three major shortcomings of this study are highlighted by Turnquist,

6

Fortenbery, & Foltz (2008). First, the methodology of the consulting firm was not described in
the report and thus the conclusions are difficult to evaluate. Second, the communities the firm
used were not named. Third, there exists no evidence of any public, peer-reviewed assessment
of ethanol plant effects in Texas or anywhere else. Beyond these issues, there is an additional
concern with regards to the validity of this report. According to the 2008 Energy Report of
Texas, the first operational ethanol plant in Texas went online in early 2008.

3

Therefore, a 2007

report estimating the impact of an ethanol plant in Texas cannot measure the effects of an
operational plant. It is possible that the authors extrapolated from other types of refineries found
in Texas, but even so, the effect of an ethanol refinery may be quite different than other types of
refineries.
Given the shortcomings of the 2007 report, the remaining investigation by Turnquist,
Fortenbery, & Foltz (2008) is the only public source document that offers an examination of the
effect of ethanol plants on residential properties. In their study, the authors’ evaluation of the
effect ethanol plants have on neighboring property is twofold. First, the authors analyze the
impact of a plant on the rate of agricultural land conversion. Their expectation is that as the
value of agricultural land increases because of expected commodity price increases, the rate of
agricultural land conversion to other uses will diminish (relative to other communities). Second,
the authors investigate the impact of an ethanol plant on residential property values. To
undertake their analysis, the authors examined four ethanol facilities in Wisconsin (all
operational by 2006) and collected municipal level tax assessment data from 2000 to 2006. To
capture the effect, the four ethanol plants were geographically located and zones of two, ten,
twenty-five, and fifty miles around each plant were created. These zones acted as representative

3

See http://www.window.state.tx.us/specialrpt/energy/
7

distances from the plant to each municipality. The authors note that municipalities closest to
ethanol facilities experienced continued growth in residential land values after the plant began
operations; however, the growth appeared to be less than municipalities further away.
Specifically, municipalities closest to the plant experienced growth rates of fifty percent while
municipalities in the rest of the state experienced growth rates of eighty percent. Testing the
differences between the rates of growth among municipalities, the authors were unable to
confirm that the ethanol communities experienced less growth because the differences were not
statistically significant. In conclusion, the authors suggest that some properties within the closest
municipality may have experienced adverse effects from the plant's existence; however, any
potential negative effect was offset at the municipal level. The authors deduce that if any effect
on residential land is to be determined, a detailed analysis at the sub-municipal level is needed.
2.2 Measuring the Effect of Negative Externalities on Property Values
Beyond the ethanol industry, a number of studies have investigated the effect of positive
and negative externalities on surrounding property values. Among such studies a wide range of
methods have been employed, from surveys to hedonic models. Also, numerous types of
externalities have been examined, from the negative impact of incinerators (Kiel & McClain,
1995) to the positive effect of open space (Bolitzer & Netusil, 2000). For reasons previously
discussed, it is anticipated that the impact of an ethanol plant on residential properties will be
negative, and thus this section provides a brief survey of previous studies focused on the effects
of localized negative externalities. Special attention is given to the various techniques that have
been implemented to measure negative effects on neighboring property and the range of results

8

(even among studies examining the same type of externality).

4

In a foundational paper, Blomquist (1974) measured the effect of an electrical power
plant on neighboring property values. Employing the hedonic technique, Blomquist’s model
used the average value of houses within a census block as the dependent variable and the
“effective distance” from the power plant as the key independent variable to capture the power
plant’s effect. The effective distance equaled the actual distance from the plant for those
properties less than 11,500 feet away and equaled 11,500 feet for properties farther away. His
report concluded that properties within 11,500 feet increased in value by nine-tenths of a percent
for each additional ten percent increase in distance from the plant.
Nelson (1981) and Gamble & Downing (1982) examined the impact of the Three Mile
Island incident on neighboring property values. In addition, Gamble & Downing studied the
impact of nuclear reactors without accidents. Rather than follow the approach of Blomquist and
use average prices at the block level, both studies set a new standard by employing individual
house sales as their unit of observation. Among the independent variables, Nelson employed two
variables to capture the impact of the plant on neighboring properties. The first was a dummy
variable indicating whether or not the house was sold before or after the accident. The second
was an interaction of the indicator variable with the date of sale, to capture the changing impact
of the plant over time. Gamble & Downing also used two different variables to examine the
effect of the nuclear plant. The first was an indicator variable representing properties from
which one could see the plant and the second was distance from the plant. Both studies
concluded that nuclear reactors did not have the anticipated negative impact.
4

For an extensive discussion on studies measuring local externalities and their results, see
Sirmans, Macpherson, & Zietz (2005). Also, see Boyle & Kiel (2001) or Jackson (2001) for an
extensive survey of studies examining the impact of environmental externalities (e.g. air quality,
water quality, and undesirable land use).
9

Since Gamble & Downing’s study, a broad range of research measuring the impact of
negative externalities on neighboring property values has been completed. Although not an
exhaustive list, these studies examined the effects of the following negative externalities:
hazardous waste sites (McClelland, Schulze & Hurd, 1990; Michaels & Smith, 1990; Smolen,
Moore, & Conway, 1992; Thayer, Albers & Rahmatian, 1992; McCluskey & Rausser, 2003);
landfills (Nelson, Genereux, & Genereux, 1992; Reichert, Small, & Mohanty, 1992); natural gas
facilities (Flower & Ragas, 1994; Boxall, Chan, & McMillan, 2005); and superfund sites
(Kohlhase, 1991; Kiel, 1995; Deaton & Hoehn, 2004). Although the externalities being studied
were different, conclusions were generally the same with most studies indicating a statistically
5

significant negative effect on property values. However, the spatial ranges and monetary
impacts varied greatly. As an example of the variation in spatial ranges, Nelson, Genereux, &
Genereux’s (1992) study estimated that the impact of landfills reached homes within two miles
of the site, whereas Kohlhase’s (1991) study on toxic sites estimated that the impact was felt up
to six miles away. As an example of the price differentials examined, Michaels & Smith’s
(1990) study on hazardous sites estimated that the benefit for each additional mile from the site
was $115 while Smolen, Moore, & Conway’s (1992) study on hazardous sites estimated a
benefit of $9,000-14,000 for each additional mile. Observing large differences in monetary and
spatial impacts, even within similar site studies, is not surprising. Each area being studied at
different time periods is unique. The price consumers are willing to pay for a given bundle of
attributes (or even a single attribute) will differ as tastes and preferences change across time and
space (Sirmans et al, 2005).

5

The only study that did not have the value of the property increase as the distance from the site
increased was that of Flower & Ragas (1994). The authors concluded that more prestigious
neighborhoods, as well as other positive influences, caused the unexpected result.
10

2.3 Contribution
The present study adds to both lines of research discussed above. In terms of research
examining the ethanol industry, little is known about how ethanol plants affect residential
property values. This is an important consideration if we are to fully understand the potential
economic costs imposed by the industry. With regard to the general research implementing
hedonic methods to examine the impact of negative externalities, this work provides a simple
and unique method to evaluate the impact by allowing within-community comparisons before
and after the plant begins production. Although including data before an externality exists is not
required to implement the hedonic approach, doing so allows the researcher to make withincommunity comparisons across time and reduces potential concerns that the negative effect is
caused by some unknown, pre-existing factor at (or near) the site.

11

3. METHODS
3.1 Hedonic Approach
Over the past forty years, there has been considerable discussion concerning theoretical
and empirical methods to estimate the price of housing attributes (including externalities). The
approach that has been widely used is hedonic price analysis. To determine the impact of an
ethanol plant on neighboring residential property values, I follow the general theoretical
approach outlined by Rosen (1974).

6

As with any good, a housing unit may be described as a vector of n objectively measured
attributes (H = h1, h2,..., hn). Such attributes encompassed in a housing unit include: the
characteristics of the structure (i.e. square footage, number of bedrooms, number of bathrooms,
age, etc.), the land the structure is on, and the location in which it exists. Beyond forming the
housing unit, each attribute also has its own implicit price and it is the sum of these prices that

n
determine how much a house is worth ( P   p ) (Brasington & Hite, 2005; Kashian,
H
i
i 1
Eiswerth, & Skidmore, 2006). However, the price of each attribute is not readily observed as a


house cannot be disaggregated and sold in separate markets. Therefore to obtain the price of a
given attribute, one can use data on the final price of the house and variables that characterize the
attributes embodied in the unit to derive the hedonic price function [P(H)=F(h1, h2,..., hn)]. This
function, in turn, allows empirical estimation of the implicit marginal price of a given attribute
(Palmquist, 1984). Quoting Rosen (1974, pg. 34): “Econometrically, implicit prices are
estimated by the first-step regression analysis (product prices regressed on characteristics) in the
6

Although Rosen was not the first to employ hedonic pricing techniques to estimate implicit
prices of goods (Haas, 1922a; Wallace, 1926; Court, 1939; Ridker & Henning, 1967), he was the
first to support interpretation and estimation through a well-defined theoretical model.
12

ˆ
construction of hedonic price indexes” [ P(H) ]. Using the first-step regression, the implicit
ˆ
ˆ
marginal price of the ith component is defined as p  P(H) /h (Goodman, 1978).
i
i

In addition to estimating the implicit marginal price of each attribute, hedonic analysis

may include a second stage to recover the inverse demand functions for individual attributes. To

recover the demand functions, the implicit prices calculated in the first stage are regressed
against observed quantities and socioeconomic characteristics of the consumers (Freeman, 1979).
However, as highlighted by Malpezzi (2003), reliable second stage estimation is difficult as it is
plagued with identification problems, imperfect specifications, and non-robust coefficient
estimates. For these reasons, the second stage estimation is rarely used. Like previous studies
examining the impact of negative externalities on house prices, this study focuses solely on the
hedonic method’s first stage to estimate the marginal impact of the ethanol plant on the price of
neighboring properties.
3.2 Model
The model in this study will analyze the impact of an ethanol plant using two approaches.
The first approach measures distance of each property from the ethanol plant as a continuous
variable. This approach is illustrated by the following equation:

m
n
PRICE      X    L   preD   DISTpre 
i
j ij
l il 1
i 2
i
j 1
l 1

k

 DISTpost   DISTpost2    Time  
3



i

4

i



13

t 1

t

it

i

[1]

7

where PRICE represents the real sales price of each house (i), X represents chosen structural
ij
i
attributes, L represents chosen neighborhood attributes, preD is an indicator variable
il
i


representing properties sold before a plant began production, DISTpre represents the Euclidean
i


distance to the plant’s address prior to production, DISTpost represents the Euclidean distance
i

2 represents distance squared, Time represents
to the operational plant’s address, DISTpost
it
i

year indicator variables to capture the sale date of each house, and  is the error term.
i


The objective of equation [1] is to provide a clear examination on whether neighboring


properties experience adverse effects from the ethanol plant. To accomplish this, three particular
variables in equation [1] are of interest: preDi, DISTprei, and DISTposti. The coefficient for the
first variable indicates whether or not the intercept differs between properties sold before and
after the plant began production. The coefficients for the latter two variables provide a clear
estimate of the ethanol plant’s impact on property values as the distance between the plant and a
given property increases. It is anticipated that the coefficient for DISTprei will be statistically
insignificant, providing evidence that there were no negative externalities at (or near) the
location prior to the operational plant. The coefficient for DISTposti is expected to be
statistically significant and positive, indicating that the closer a house is to the plant the lower its
price. Finally, the coefficient for DISTpost2 will be included to measure non-linear effects of
i
the ethanol plant. It is anticipated that the coefficient for DISTpost2 will be negative, implying
i

that there is a diminishing effect as the distance between the plant and the property increases.


7

Prices will be converted to 2009 dollars using the Consumer Price Index (CPI).
14

In addition to the anticipated results of the two key distance variables (DISTprei and
DISTposti), including distance to the site prior to the operational plant is a unique feature of this
study. The majority of previous hedonic work has focused on measuring the impact of an
externality exclusively during the years the plant is in operation. A potential downfall of
excluding the distance prior to the facility’s existence is attributing a negative effect to the
facility when the true cause may be some unobserved factor that prevailed prior to the plant.
Including properties and their distance to the site before the plant’s operations allows for direct
comparisons within the same community to ensure there was no unforeseen, pre-existing factor.
Three hedonic studies could be found that explicitly measure the impact of projects at
different stages. First, Smolen, Moore, & Conway (1992) examined the impact of an existing
hazardous waste site and the impact of a proposed hazardous waste site. The authors conclude
that properties in the community with the proposed plant were not impacted while properties in
the community having an existing waste site were impacted between $9,000 and $14,000 for
each additional mile up to 2.6 miles. This is different from the present study, however, as
different communities were used as comparisons. As noted above, differences between
communities may exist and direct comparisons may not be made unless all differences are
accounted for.
Second, McMillen & Thorsnes (2003) examine the impact of a smelter on property
values at different phases in Tacoma, Washington. Specifically, the authors examined the impact
of the smelter during the following four stages: 1) operational, 2) closing, 3) Superfund site
designation, and 4) cleanup. McMillen & Thorsnes concluded that the discount associated with
proximity to the smelter converted to a premium once the site was closed and designated as a
Superfund site. The authors state that this outcome was observed since the smelter was located

15

in an otherwise attractive location. McMillen & Thornses use a similar technique as employed in
the present study; however, they examine the removal of a negative externality rather than the
introduction of a negative externality.
Finally, Kiel & McClain (1995) examined the impact of an incinerator on home sales
during different time periods in North Andover, Massachusetts. The stages that were examined
include: pre-rumor, rumor, construction, beginning operations, and continued operations. The
authors found that the incinerator had no impact until the construction phase, at which time the
property values increased approximately $2,300 for each additional mile from the plant. Once
the incinerator began operations, the impact was much larger ($8,100 per mile) and over time the
impact slightly decreased ($6,600 per mile). This is directly comparable with the current study;
however, the current study takes Kiel & McClain’s general results into consideration and
implements two time periods (pre- and post-production).
Although equation [1] is an adequate first approximation, the negative effect of the
ethanol plant measured in equation [1] may be larger (i.e. more negative) for properties closer to
the plant since the coefficient for DISTposti captures the effect on all properties (including those
miles away). With this consideration, a second regression will be used to further appraise the
impact on properties closest to the plant. This model is represented by the following equation:

m
n
PRICE      X    L   preD   DISTpre 
i
j ij
l il 1
i 2
i
j 1
l 1



d
k
  Ring    Time  
r
ir
t
it i
r 1
t 1

[2]

where PRICE , X , L , preD , DISTpre , Time , and  are the same as equation [1]. Rather
i ij il
i
i
it
i

than measure the Euclidean distance between the plant and each residential property, Ring
ir




  
16



measures incremental, half-mile rings around the ethanol plant. It is expected that relatively
large negative impacts will be observed for the first few rings but the effects will dissipate as the
distance increases.
3.3 Structural Attributes (X)
While the list of structural attributes that could be included is extensive, only those
variables suggested in the literature as consistently having a significant impact on the value of
residential properties are included. The structural characteristics collected include: size of the
house, size of the lot, age of the house, number of bedrooms, number of bathrooms, number of
stories, the existence of a basement, the existence of an attached garage, and the existence of
central air conditioning (Palmquist, 1984; Pollakowski, 1995; Brasington & Hite, 2005). It is
expected that as most of these housing attributes increase (or are present as in the case of a
garage, a basement, and central air), they will have a positive impact on the price of property.
Three exceptions to this expectation include: age of the house, number of bedrooms, and number
of stories. While it is obvious that the age of the house will have a negative impact, the expected
impact that the number of bedrooms and number of stories will have on the price is ambiguous.
Once size is controlled for, additional rooms translate to smaller rooms. It is therefore unclear
whether more small rooms are preferred to fewer large rooms. Similarly, an additional story
translates into more a divided house. In their review, Sirmans et al. (2005) observed bedrooms
to be positive in twenty-one out of forty studies including the variable and stories to be positive
in only four out of thirteen studies. Some studies exclude the number of bedrooms or stories
altogether, citing that they are not reliable when the size of the structure is included due to
collinearity (McClelland, Schulze, & Hurd, 1990). The ethanol plant’s effect remains the same
whether or not bedrooms and stories are included.

17

In addition to including the linear measurements of each structural attribute, it may be
beneficial to use the squares of the lot size, house size, and age of the house since these are
expected to influence the value of a house in a nonlinear fashion (Brasington & Hite, 2005).
Nonlinearities are expected due to diminishing returns in consumption (Witte, Sumka, &
Erekson, 1979).
3.4 Neighborhood Attributes (L)
The list of potential neighborhood attributes is extensive. Focusing on the neighborhood
variables that are most likely to have an impact, two key variables that are often cited include:
distance to the town center and an indicator variable equal to one if the property is located within
8

the city limit and zero otherwise. It is anticipated that each variable will be positive as a town
may offer a variety of amenities. As with distance to the ethanol plant, it may be beneficial to
include the square of distance to the town center to measure diminishing effects.
3.5 Time
As stated above, the sale date of each house will be captured by a year indicator variable.
Since prices used are in real terms (inflated to 2009 values), inclusion of the sale date captures
the effect of general market trends over time (as well as other fixed effects). Although the effect
of each year indicator variable depends on local housing market trends, a negative effect each
year post 2006 is anticipated as the housing market collapsed nationwide.
3.6 Functional Form
While hedonic price models have been used to routinely analyze the market price of
multiple housing attributes, a common challenge for all hedonic studies is selecting the
8

Direction of the property from the plant was also considered, but Abeles-Allison & Connor
(1990) have shown this to not be statistically significant in their study examining hog farms.
Furthermore, there exist a small number of observations downwind from the ethanol plant (see
page 25) and inference based on only a few observations would be unreliable.
18

appropriate functional form (Cropper, Deck, & McConnell, 1988). Since theory provides no a
priori guidance regarding functional form, it is common to empirically determine the functional
form that best fits the data (Palmquist, Roka, & Vukina, 1997). Following previously cited
literature, two functional forms were considered: linear and semi-log (natural logarithm of the
dependent variable). To determine which model best fit the data, each specification’s sum of
squared residuals were compared once the observed prices were normalized by their geometric
means. Palmquist & Danielson (1989) show that this procedure is equivalent to the Box-Cox
criterion.
Based on the sum of squared residuals, the two specifications are not statistically
different from each other and both the semi-log and linear results are presented in the
appendices. Although the results of the two functional forms are similar, discussion of the
results in Section 5 is based on the semi-log form. Historically, implementing the semi-log
specification is preferred and the semi-log regressions presented below are generally more
conservative (Sirmans et al., 2005).

19

4. DATA
4.1 Areas of Study
Figure 2 presents a map detailing all operating ethanol plants in Michigan. Of the five
operational plants, only two are examined in this study: POET Biorefining in Caro (operational
October 2002) and Marysville Ethanol, LLC in Marysville (operational September 2007). These
locations were chosen for two key reasons: 1) both plants have operated for a significant period
of time allowing for more pre- and post- plant operations sales data; and 2) both communities
have had a substantial number of sales during the time period of interest.

9

Study Area:
Caro
POET Biorefining
October 2002
53 mmgy
Study Area:
Marysville
Marysville Ethanol, LLC
September 2007
50 mmgy

Lake Odessa
Carbon Green Bioenergy
September 2006
50 mmgy

Riga
Green Plains Renewable Energy
February 2007
57 mmgy

Albion
The Andersons Albion Ethanol, LLC
August 2006
55 mmgy

Figure 2. Michigan ethanol plants location, name, start date, and nameplate
10
production capacity

9

Other plant locations are more rural and a very limited number of sales were available for those
communities since the plant went online (30 or less for each community). Inference drawn from
regressions with a small number of observations would be unreliable.
10
There are currently no plants in Michigan’s Upper Peninsula.
20

4.2 Data on House Prices and Desired Attributes
To conduct the hedonic analysis described above, detailed information on the price, the
structural attributes, and the neighborhood characteristics of houses surrounding the ethanol
plants was assembled. Two Michigan multiple listing services (MLS) provided data:
MiRealSource and RMLS.

11

All available sales data between 1999 and 2009 were collected for

each community. A total of 1,956 home sales were obtained (909 from Caro and 1,046 from
Marysville).
Upon examining the data, two remaining issues required attention: 1) the neighborhood
characteristics of each property still needed to be determined since multiple listing services do
not provide these features; and 2) a large number of Caro properties required structural
characteristics to be added as information was missing in both MLS databases. To handle the
first issue, all included properties were geo-coded and mapped using a Geographic Information
Systems (GIS) database. Mapping the sales resulted in 129 dropped observations (60 in Caro
and 69 in Marysville), as the addresses were not in the GIS database. Furthermore, upon
examination of the mapped properties, a large number of properties were apparently miscoded
and were located large distances from the areas of interest (often appearing in cities or towns
miles away). These observations were therefore excluded. The final number of mapped data
included 824 sales in Caro and 887 sales in Marysville. Using these 1,712 sales, the
neighborhood characteristics of interest were determined using the GIS database.
With the aid of the local assessor, the issue of missing structural data of Caro properties
was resolved. Of the properties collected, the assessor was unable to locate the following: age

11

Both MiRealSource and RMLS are online member services.
21

for 123 properties, square footage for an additional three properties, and lot size for an additional
27 properties. Thus, the final Caro dataset that included all attributes totaled 671 sales.
4.3 Additional Concerns
Beyond data completeness, there are three additional concerns with the dataset. The first
issue stems from the rural nature of the data. As highlighted in Henderson & Gloy (2008), the
effect of an ethanol plant on farmland property may be positive, resulting from increased demand
for neighboring farmland commodities. Beyond the positive effect of the ethanol plant on
farmland properties, it is reasonable to expect valuation differences to arise between residential
and farmland properties as the latter are marketed and sold primarily based on their agricultural
production capacities rather than their housing attributes. Therefore, to be sure that only
residential properties are considered, all properties with more than ten acres have been excluded
from the dataset. This restriction is a common practice for hedonic analyses of residential
properties in rural communities (Abeles-Allison & Connor, 1990; Palmquist, Roka, & Vukina,
1997; Herriges, Secchi, & Babcock, 2005). An additional 43 observations were excluded for
Caro, and zero for Marysville.
The second concern stems from potential human error. In obtaining records from the
MLS databases, there is no procedure to ensure the data reported are completely accurate. It is
possible that some sales were mistyped and not enough zeros were included. This could result in
biased estimates of the ethanol plant’s impact, but only if the errors are somehow systematically
related to the proximity of the home to the plant. Nevertheless, several homes were well below
the typical price range. In particular, there were 19 transactions indicating a selling price of less
than $20,000. Thus, I chose to exclude 17 observations for Caro and two observations for

22

12

Marysville.

Although Herriges et al. (2005) expressed a concern that dropping such

observations might exclude the properties most affected by the ethanol plant, this concern is
outweighed by the potential for human error. Unreported regressions including these
observations yield results that are similar to those presented in this paper and maps detailing the
location of the dropped properties in relation to the ethanol plant are provided in Appendix 1.
The final concern relates to whether the omission of additional structural and
neighborhood characteristics might lead to omitted variable bias, especially with regard to the
coefficients of interest. Beyond the structural characteristics cited above, one additional
structural characteristic in Marysville that requires attention is the fact that some residential units
are condominiums. To handle this issue, an indicator variable equal to one if the unit is a
condominium and zero if it is a house is included: a negative sign is expected (indicating
condominiums sell for lower prices). Beyond the neighborhood characteristics cited above, there
is one additional neighborhood characteristic in Caro and two in Marysville that should be
included. Of particular concern in Caro is the potential effect of the sugar plant. The sugar plant
is located within the city limits and may adversely affect properties as the plant emits a pungent
odor. To handle this, I include a variable that represents the Euclidean distance (and distancesquared) between the sugar plant’s address and each residential property. Of particular concern
in Marysville are the potential effects of the St. Clair River and the Detroit Edison power plant.
An indicator variable representing those properties within one half-mile of the river will be
included to estimate the impact of the river. It is anticipated that the river will have a positive
effect on these properties. As with the sugar plant in Caro, the Euclidean distance (and distance-

12

These transactions could also reflect family-to-family sales, in which socio-economic
attachment influence selling price more than the physical attributes of the residence (Robison &
Ritchie, 2010).
23

squared) from each property to the Detroit Edison power plant’s address will be used to measure
the effect of the power plant on neighboring properties. As highlighted in the literature review,
the impact of the power plant is anticipated to be negative (Blomquist, 1974).
Figure 3 shows the location of a select number of collected properties within each
community, as well as the location of the ethanol plants, sugar plant (Caro), and power plant
(Marysville). Furthermore, the rings used in the analysis are represented with the distance (in
miles) of each ring from the ethanol plant. The full list of variables included and their definitions
can be viewed in Appendix 2. Appendix 3 and 4 provide the summary statistics for Caro and
Marysville, respectively.

24

2.5
2.0
1.5
1.0

3.0
2.5
2.0
1.5

Figure 3. Location of select properties in Caro (above) and Marysville (below)

25

5. RESULTS
Given that housing markets are highly localized and spatially segmented (Sirmans et al,
2005), regressions for the two locations are estimated separately to obtain the implicit marginal
price for the housing attributes in each community. The results for Caro are presented in
Appendix 5 and the results for Marysville are presented in Appendix 6. The results presented in
columns (1) and (2) reflect equation [1], using DISTpost to capture the effect of the ethanol
i
plant on property values and DISTpost2 to capture non-linear effects. To further examine the
i

impact, and to better capture non-linearities, columns (3) and (4) present the non-linear effects of
the ethanol plant in
Caro using the approach reflected in equation [2]. Furthermore, columns (1)
and (3) of each table contain the results for the semi-log estimation and columns (2) and (4)
contain the linear estimation results.
5.1 House Characteristics: Caro
First, consider the results for the standard variables included in the analysis. The
structural variable coefficients have the correct signs, are statistically significant, and appear to
be reasonable estimates when transformed to dollar values. Examining the variables concerning
size of the house and size of the lot, the results indicate that an additional square foot of living
space adds approximately $36 of value, on average, while an additional acre of land adds
13

$7,800.

The impact of age on the house is the only negative effect among the structural

13

Special attention must be given when interpreting coefficients including a squared term since
the two coefficients may not simply be combined to get the estimated impact of each variable in
percentage terms. Rather, the first derivative of the combined coefficients must be calculated.
For example, to estimate the impact of an additional square foot on the value of the average
house from column 1 in Appendix 5, the following calculation must be made: % change =
2
[(coefficient from LivingArea) + (2*coefficient from squared LivingArea *LivingArea)]*100.
Using this equation yields: % change = [(0.0006) + (2*-0.0000000878*1440)]*100 = 3.47%,
26

characteristics, as expected. The results indicate that there is a decrease of approximately $530
for each additional year since construction of the house. Since the coefficients for Basement,
Garage, and AC represent the impacts of dummy variables, one cannot simply multiply each by
100 to get the corresponding percent change (as is done for continuous variables). Following the
procedure provided Halvorsen & Palmquist (1980), the presence of a basement increases the
value of the average house by $23,700 and the presence of an attached garage increases the
average house price by approximately $22,450. The presence of central air conditioning also
generates a large premium, increasing the value of the average house by $14,200. Finally, an
additional full bathroom increases the value of the average house by approximately $8,600.

14

As shown in Appendix 5, two structural characteristics that are not statistically significant
include: Bedrooms and Stories. This result is not surprising from a theoretical perspective as
mulicollinearity is anticipated with the inclusion of LivingArea. This result is also not surprising
from an empirical perspective as previous hedonic studies indicate statistical insignificance when
the size of the house is included (Kashian, Eiswerth, & Skidmore, 2006; McClelland, Schulze, &
Hurd, 1990).

15

Turning attention to the year indicator variables which capture market trends, the only
years that are statistically different than the base year (1999) include: 2002 and 2006-2009. Each
coefficient is negative, and from 2006 to 2009 a larger negative impact is observed for each
where 1440 is the average living area from the summary statistics provided in Appendix 3.
Multiplying this by the average house price ($103,380) gives an increase of $36 per square foot,
on average.
14
Although the estimates for garage, central air, and bathroom seem high, each is within the
typical range cited by Sirmans et al. (2005). The presence of central air conditioning may also
represent a proxy for updating a house. This would further support the large premium observed
from central air conditioning.
15
Furthermore, correlation tests between these variables were significant at the one percent
level.
27

additional year. In 2006, house values decreased by 19.8% (or approximately $20,500) since
1999. By 2009, properties in the Caro community experienced large net decreases as a result of
the recession and the coefficient for 2009 indicates an estimated decline of approximately 49.5%
in the average house sale since 1999! While this may seem drastic, Figure 4 highlights annual
average house prices in the Caro area and supports this result.

Real House Price ($2009)
$180,000
$160,000
$140,000
$120,000
$100,000
$80,000
$60,000
$40,000
$20,000
$0
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Source: Michigan Association of Realtors
Figure 4. Average house prices for the Caro area

16

16

While the large (and seemingly significant) decline from 1999 to 2000 is not consistent with
the regression results, the data collected for 1999 was from the Lapeer Association of
REALTORS while the data for 2000-2009 was from the Lapeer and Upper Thumb Association
of REALTORS (a group which includes Caro). Using the Lapeer Association of REALTORS for
1999 is the only way of consistently representing all years since it is not apparent what group of
realtors the Caro area had belonged to at that time (if any).
28

5.2 Neighborhood Attributes: Caro
Examination of the included neighborhood characteristics validates the expectation of the
ethanol plant’s impact and provides insight concerning other neighborhood effects. As expected,
being located within the city limits appears to have a positive effect on the house (although
statistically insignificant). The coefficient for preD is statistically insignificant, indicating there
is no change in the intercept between properties sold prior to plant operations. Unexpected is the
statistically insignificant effect of distance from the town center. Another unanticipated result is
the statistically insignificant effect of the sugar plant, indicating that it has no effect on the price
of nearby residential properties.
Finally, the estimated effect of the ethanol plant is negative. This result should be
considered in two parts. In column (1), the coefficient DISTpre represents the distance to the
location of the plant before the plant was online. This coefficient is statistically insignificant and
verifies that there was no externality (positive or negative) at, or near, the site prior to plant
production. The second coefficient, DISTpost, shows the negative impact of the plant once
operations began. One interpretation of the combined effect of DISTpost and DISTpost2 is that
property values increase as their distance from the plant increases, but at a decreasing rate. The
average property experienced an increase in value of approximately 
3.4% for each mile from the
plant. In terms of dollar values, this translates into an increase of $3,600 per mile. While this
effect is as expected, the range in distance that houses are from the plant varies greatly (from
two-tenths of a mile to ten miles). Since DISTpost measures the average impact of all houses
within this range, the estimated impact of the plant on the closest properties may be
underestimated.

29

To more completely examine the effect of the ethanol plant on houses closer to the plant
and to examine nonlinear effects with easier interpretation, several rings representing interval
distances from the plant have been added to the regression. Ring1 represents all properties
within one mile from the plant, Ring1.5 represents properties between 1-1.5 miles, Ring2
represents properties between 1.5-2 miles, Ring2.5 represents properties between 2-2.5 miles,
and Ring3 represents properties between 2.5-3 miles. All other properties outside these rings
serve as the baseline since the effect is not anticipated to reach farther than three miles. Note,
however, that Ring1 includes all properties within one mile while the others are measured in
half-mile increments. The first ring extends to one mile because there are a small number of
observations within the first half-mile and imprecise estimates would likely result due to the
small number of observations. Interpreting the results shown in columns 3 and 4 of Appendix 5,
the ethanol plant has a negative effect on properties as far as two miles away. As anticipated,
those properties closer to the plant experience a larger negative effect than was measured in the
Euclidean distance regression. Interpreting the results, those within two miles sold for 15-18%
less after the plant began operations.
5.3 House Characteristics: Marysville
Examining the Marysville estimates and comparing them with Caro confirms that tastes
and preferences differ among consumers as the price consumers are willing to pay for a given
bundle of attributes (or even a single attribute) differ. The most striking differences are those
that change statistical significance, including: LotSize is no longer significant, Bedrooms and
Stories are no longer insignificant, InTown is significant and negative, toTC is significant and
positive, and the ethanol plant has no impact.

30

Examining the structural characteristics, most have their anticipated sign, are statistically
significant, and appear to be reasonable estimates when transformed to dollar values. The
coefficient for LivingArea indicates that an additional square foot of living space adds
approximately $54 of value, on average. For each year older the house becomes, the value
decreases by approximately $430. The presence of a basement generates a smaller premium than
was estimated in Caro, increasing the value of the average house by $18,400; while the presence
of an attached garage generates a larger premium than was estimated in Caro, increasing the
value of the house by $29,000. Having central air conditioning also generates a slightly larger
premium in Marysville, increasing the value of the average house by $15,600. Finally, an
additional full bathroom increases the value of the average house by approximately $14,000.

17

As mentioned, Bedrooms and Stories are statistically significant. An additional bedroom
increases the value of the average house by $4,400 and an additional story decreases the value by
$11,700. Finally, the indicator representing condominiums shows that the average condo sells
for approximately $21,500 less than the average house (ceteris paribus).
Examining the year indicator variables highlights that the first seven years are statistically
no different than the base year (1999) and the coefficients for 2007-2009 are each statistically
significant and negative. Comparing these results with Caro, the Marysville housing market
appears to have experienced a smaller negative impact in 2007 and 2008; however, by 2009 the
Marysville market was in the same situation as Caro with the average house sale approximately
47% less than in 1999. As before, these are not unreasonable estimates as Figure 5 shows the
yearly trends in the Marysville housing market.

17

Again, the estimates are within the ranges presented by Sirmans et al. (2005).
31

Real House Price ($2009)
$200,000

$180,000
$160,000
$140,000
$120,000
$100,000
$80,000
$60,000
$40,000
$20,000
$0
1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009
Source: Michigan Association of Realtors
Figure 5. Average house prices for the Marysville area

5.4 Neighborhood Attributes: Marysville
The results of the included neighborhood variables are surprising. The variables
representing in town, distance to town, distance to the river, and distance to the power plant are
statistically significant and have signs opposite of what was anticipated. Being located in town
(InTown) has a large negative effect and being located farther from the center of town (toTC)
appears beneficial. However, once the squared term for toTC is included the effect validates
expectations: the average house decreases approximately 2.8% for each additional mile from the
town center.
The coefficient for River, indicating the properties within one-half mile of the River, is
negative. Observing a negative effect from the river seems counterintuitive; however, the

32

negative effect may be the result of a large number of properties within a half-mile not having
access to or a view of the river. If a large number of the properties within one-half mile are
without access or a view, there is no reason to anticipate a positive effect. In fact, a negative
effect could be anticipated as the properties could be in the floodplain. To examine this issue
further, five new variables have been created: River1 representing properties within 0.1 miles of
the river, River2 representing properties within 0.1-0.2 miles of the river, River3 representing
properties within 0.2-0.3 miles of the river, River4 representing properties within 0.3-0.4 miles of
the river, and River5 representing properties within 0.4-0.5 miles of the river. The summary
statistics for these variables are shown in Table 1 and the majority of properties within one-half
mile of the river are farther than two-tenths of a mile. These properties may cause the observed
negative effect. To confirm these suspicions, Table 2 shows the coefficients of each new
variable by including them into the regressions from Appendix 6. As the results show, the river
has a positive effect on properties closest to the river.

Variable
River1
River2
River3
River4
River5
# of Obs.

Table 1. St Clair River summary statistics
Mean
Std Dev
Min
0.030
0.172
0
0.024
0.152
0
0.042
0.200
0
0.089
0.285
0
0.023
0.149
0
885

Max
1
1
1
1
1

Table 2. Further examination on the effect of the St Clair River
Independent
Dependent Variable
Variables
Ln(RealPrice)
0.1446***
River1
(0.0527)
0.2810***
River2
(0.0760)
-0.1141**
River3
(0.0508)
-0.0515*
River4
(0.0312)

33

River5

Table 2 (cont’d)
-0.1023***
(0.0371)

As with the sugar plant in Caro, the distance to the power plant has the opposite sign of
what was anticipated. Unlike the sugar plant in Caro, it is statistically significant; however, the
effect is minimal as the average property value decreases by approximately 0.7% percent for
each mile from the plant.
Finally, the estimated effect of the ethanol plant is statistically insignificant. To ensure
that these results are robust and that the conclusion is not a result of downward bias from the
furthest properties, several rings representing interval distances from the plant have been added
to the regression. One key difference exists between the rings created for Marysville and Caro:
the first ring for Marysville (Ring1.5 in Appendix 6) extends to 1.5 miles from the plant. This
was done because there are a small number of observations within the first mile of the plant and
imprecise estimates would likely result due to the small number of observations. Beyond
Ring1.5, the remaining rings mirror those created for Caro. Again, all properties farther than
three miles serve as the baseline because the effect is not anticipated to reach farther than three
miles. Examining the regressions in Column 3 of Appendix 6, the coefficient representing all
properties within 1.5 miles of the plant is statistically insignificant. This provides additional
evidence that properties closest to the ethanol plant have not experienced any depreciation in
value from the ethanol plant. Finally, it is worth noting that Ring2 and Ring2.5 are positive and
highly significant, perhaps indicating some positive externality that has not been considered.
To provide insight into why there was no measurable impact on neighboring property
values in Marysville when Caro was clearly affected, three hypotheses have been considered.
The first hypothesis relates to data issues. Although a sufficient number of observations have

34

been accumulated to present satisfactory hedonic results, there are very few properties within a
one-mile radius of the Marysville plant (less than one percent of all observations). Although the
impact was felt up to two miles in Caro and the number of properties within two miles in
Marysville is significant (approximately 34% of all observations), as observed with other
attributes there is no reason to expect the impact in Marysville to reach the same distance as was
experienced Caro. Therefore, the few observations within one mile of the ethanol plant may not
have been enough to observe an adverse effect. The second hypothesis centers on visibility.
Trees in the Marysville community are abundant. Perhaps the adage, “Out of sight, out of
mind,” applies to the Marysville ethanol plant. That is, visibility may be a requirement for some
ethanol plants to adversely impact the surrounding community. The third hypothesis stems from
pre-existing conditions at the plant’s location. The Marysville ethanol plant was placed in an
already developed industrial area, whereas Caro did not have pre-existing industrial facilities.
Perhaps failure to observe an impact stems from the plant not adding any perceived negative
externalities. All three of explanations are reasonable; unfortunately the data are not suitable to
determine which is the best explanation.

35

6. CONCLUSION
This study contributes to the existing literature in three ways. First, an in-depth analysis
concerning the impact of an ethanol plant on residential properties has been provided. Two
communities with ethanol plants were examined in this study to determine whether ethanol
plants have adverse effects on nearby property values. Each community examined offers a
different landscape and are in many ways representative of ethanol plant communities across the
country. Marysville is a larger community with more industry, whereas Caro is a smaller
farming community with little pre-existing industry. As highlighted, the location of an ethanol
plant may adversely affect neighboring property values; depressing the value of homes as much
as 18% and as far as two miles away. However, these impacts have been shown to not be
universal, probably because consumer tastes and preferences differ across time and space. In
addition, inconsistent impacts may stem from the physical surroundings of ethanol plants such as
vegetation and the existence of other industries across communities.
These findings have practical significance for community planners considering whether
to allow a plant to locate in their community and determine a suitable location. These results
suggest that community planners should direct ethanol plants to be built in areas where they are
not seen or are among pre-existing industrial buildings to minimize the impact. However, this is
not always possible as the ethanol industry is filling the landscape of rural America, where
vegetation (other than fields) and pre-existing industry are minimal.
The second contribution of this study is a potential upper limit on the effect of ethanol
plants on residential property values that should be added to future cost-benefit studies
examining the ethanol industry. Although the current study has highlighted the impacts of this
cost on property owners, future cost-benefit studies may need to include this cost in terms of

36

property taxes. If property values are decreasing as a result, homeowners and the community
may experience negative effects as property values and property taxes decline.
Finally, this study contributes to the general research implementing the hedonic method
by illustrating a technique to evaluate negative externality impacts in a way that allows within
community comparisons before and after the plant begins production. Using this approach helps
to assure researchers, decision makers, and others that any observed negative impact is not the
result of other pre-existing conditions in the community.

37

APPENDICES

38

APPENDIX A

Figure 6. Property sales less than $20,000
39

APPENDIX B
Table 3. Description of variables
Variable
REALPRICE
LivingArea
LotSize
Age
Bedrooms
Baths
Stories
Basement
Garage
AC
InTown
toTC
RIVDum

preD
DISTpre
DISTpost
toSugarPlant
toPowerPlant

Description
Sales price of the residential property (2009 dollars)
Size of the residential structure (square feet)
Size of the property associated with the residential structure (acres)
Age of the residential structure, estimated as continuous numbers with
each number representing an additional decade.
Number of bedrooms
Number of bathrooms
Number of stories
Indicator variable to distinguish whether the residential structure has a
basement (1 = structure has a basement, and 0 otherwise).
Indicator variable to distinguish whether the residential structure has a
garage (1 = structure has a garage, and 0 otherwise).
Indicator variable to distinguish whether the residential structure has
central air conditioning (1 = structure has central air, and 0 otherwise).
Indicator variable to distinguish whether the property is located within
the city limits (1 = property within the limits, and 0 otherwise).
Distance to the town center, measured in miles (to the nearest
hundredth).
Indicator variable to distinguish whether the property is located within
a half mile from the St. Clair River (1 = property within this range,
and 0 otherwise). This variable only applies to Marysville properties.
Indicator variable to distinguish whether the property was sold prior to
the start date of the ethanol plant (1 = property sold prior to the
operational plant, and 0 otherwise).
Distance to the ethanol plant prior to production, measured in miles
(to the nearest hundredth).
Distance to the operating ethanol plant, measured in miles (to the
nearest hundredth).
Distance to the sugar plant, measured in miles (to the nearest
hundredth). This variable only applies to Caro properties.
Distance to the power plant,measured in miles (to the nearest
hundredth). This variable only applies to Marysville properties.

40

APPENDIX C
Table 4. Caro summary statistics
Variable
REALPRICE
LivingArea
LotSize
Age
Bedrooms
Baths
Stories
Basement
Garage
AC
InTown
toTC
preD
DISTpre
DISTpost
Ring1
Ring1.5
Ring2
Ring2.5
Ring3
toSugarPlant
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
# of Obs.

Mean
103,380
1440
1.49
4.68
2.99
1.60
1.26
0.71
0.75
0.33
0.52
2.06
0.22
3.00
2.73
0.09
0.18
0.26
0.08
0.05
2.21
0.01
0.11
0.11
0.12
0.11
0.15
0.13
0.11
0.09
0.05

Std Dev
47,888
533
2.10
2.99
0.77
0.59
0.40
2.14
0.42
2.16
1.85
2.02
-

Min
20,000
540
0.07
0
1
1
1
0
0
0
0
0
0
0.76
0.26
0
0
0
0
0
0.2
0
0
0
0
0
0
0
0
0
0
611

41

Max
414,496
5000
10
15
6
4
3
1
1
1
1
10.1
1
10.85
10.85
1
1
1
1
1
10
1
1
1
1
1
1
1
1
1
1

APPENDIX D
Table 5. Marysville summary statistics
Variable
REALPRICE
LivingArea
LotSize
Age
Bedrooms
Baths
Stories
Basement
Garage
AC
Condo
InTown
toTC
River
preD
DISTpre
DISTpost
Ring1
Ring1.5
Ring2
Ring2.5
Ring3
toPowerPlant
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
# of Obs.

Mean
152,125
1380
0.25
3.39
2.94
1.79
1.31
0.78
0.87
0.55
0.12
0.98
0.88
0.20
0.82
2.43
2.48
0.01
0.14
0.20
0.14
0.29
1.55
0.03
0.13
0.12
0.14
0.09
0.13
0.11
0.09
0.08
0.06

Std Dev
58,938
459
0.23
2.36
0.66
0.68
0.42
0.63
0.40
0.38
0.82
0.87
0.78
-

Min
20,000
550
0
0
1
1
1
0
0
0
0
0
0.06
0
0
0.63
0.63
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
885

42

Max
451,167
4492
3.86
10
6
4
2.5
1
1
1
1
1
4.85
1
1
5.56
5.56
1
1
1
1
1
5.53
1
1
1
1
1
1
1
1
1
1

APPENDIX E
Table 6. Caro regression results
Independent
Variables
Intercept
LivingArea
LivingArea2
LotSize
2

LotSize
Age
2

Age

Bedrooms
Baths
Stories
Basement
Garage
AC
InTown
toTC
toTC

2

preD
DISTpre
DISTpost
DISTpost

2

Ln(RealPrice)
10.468***
(0.1678)
0.0006***
(0.0001)
-8.78e-08***
(2.03e-08)
0.0918***
(0.0244)
-0.0056**
(0.0025)
-0.0767***
(0.0152)
0.0027**
(0.0011)
0.0285
(0.0247)
0.0834**
(0.0334)
-0.0026
(0.0407)
0.2061***
(0.0326)
0.1965***
(0.0312)
0.1288***
(0.0265)
0.0136
(0.0523)
-0.1157
(0.0909)
0.0169
(0.0125)
0.0982
(0.0925)
0.0244
(0.0280)
0.0738*
(0.0392)
-0.0072
(0.0048)

Dependent Variable
RealPrice
Ln(RealPrice)
23187
10.758***
(17812)
(0.1927)
42.390***
0.0006***
(10.986)
(0.0001)
-0.0019
-8.58e-08***
(0.0031)
(1.99e-08)
9524.7***
0.0945***
(2608.6)
(0.0243)
-552.86**
-0.0058**
(278.18)
(0.0025)
-8951.1***
-0.0747***
(1461.5)
(0.0154)
370.69***
0.0027**
(107.38)
(0.0011)
1619.8
0.0276
(2204.9)
(0.0246)
9063.2***
0.0853**
(2948.8)
(0.0335)
-2769.3
-0.0056
(4335.9)
(0.0412)
21387***
0.2090***
(2868.4)
(0.0326)
16320***
0.1968***
(2514.3)
(0.0311)
13189***
0.1247***
(2455.8)
(0.0270)
-599.28
0.0363
(5705.3)
(0.0550)
1353.0
-0.0957
(8130.1)
(0.1111)
161.14
0.0165
(1052.8)
(0.0134)
15207*
0.0121
(9020.3)
(0.0773)
4726.5
0.0120
(2907.0)
(0.0135)
12126***
(3826.6)
-1062.2**
(445.02)
43

RealPrice
59136***
(20658)
43.845***
(10.917)
-0.0021
(0.0031)
9586.5***
(2592.1)
-564.04**
(273.82)
-8859.1***
(1503.0)
372.70***
(111.21)
1142.90
(2208.5)
9404.8***
(2975.4)
-3148.5
(4457.8)
21626***
(2873.8)
16455***
(2525.1)
12679***
(2473.8)
5058.1
(5611.7)
10621
(10389)
-818.31
(1218.7)
1926.7
(7046.8)
1613.4
(1304.1)
-

Table 6 (cont’d)
Independent
Variables

Ln(RealPrice)

Ring1

-

Ring1.5

-

Ring2

-

Ring2.5

-

Ring3

-

toSugarPlant
toSugarPlant
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
R-squared

2

0.0661
(0.0964)
-0.0126
(0.0132)
-0.0976
(0.0856)
-0.0794
(0.0753)
-0.1346*
(0.0788)
-0.1237
(0.1017)
-0.0158
(0.1011)
-0.0978
(0.1001)
-0.2212**
(0.1002)
-0.4019***
(0.1042)
-0.5156***
(0.1112)
-0.6839***
(0.1245)
0.6679

Dependent Variable
RealPrice
Ln(RealPrice)
-0.1776**
(0.0920)
-0.1971**
(0.0914)
-0.1637**
(0.0738)
-0.0446
(0.0650)
-0.0910
(0.0726)
-12695
0.0134
(9583.7)
(0.1244)
741.05
-0.0099
(1187.1)
(0.01379)
-13703*
-0.0920
(8166.0)
(0.0876)
-1230.3
-0.1170
(7998.9)
(0.0733)
-4629.2
-0.1646**
(8309.9)
(0.0765)
-1445.2
-0.1541
(9932.8)
(0.1044)
11222
-0.0526
(10119)
(0.1024)
-3372.9
-0.1294
(9827.9)
(0.1023)
-10329
-0.2520**
(9922.9)
(0.1024)
-26667***
-0.4283***
(10199)
(0.1070)
-34202***
-0.5391***
(10397)
(0.1130)
-50396***
-0.7259***
(10897)
(0.1236)
0.7028

# of Obs.

0.6703

RealPrice
-28152***
(8376.4)
-24195***
(8246.5)
-16750**
(7174.2)
-8517.8
(6073.0)
-4970.2
(7443.5)
-23022*
(12340)
1844.0
(1321.1)
-10968
(8066.2)
-4987.0
(7607.6)
-7819.3
(7933.9)
-4516.3
(9776.3)
6785.5
(9889.8)
-6353.2
(9672.6)
-13849
(9721.9)
-29582***
(10132)
-37168***
(10168)
-54734***
(10740)
0.7036

611

Notes: All regression results are corrected for heteroskedasticity. Asterisks denote
significance at the 1% (***), 5% (**), and 10% (*) levels.

44

APPENDIX F
Table 7. Marysville regression results
Independent
Variables
Intercept
LivingArea
LivingArea2
LotSize
2

LotSize
Age
2

Age

Bedrooms
Baths
Stories
Basement
Garage
AC
Condo
InTown
toTC
toTC

2

River
preD
DISTpre

Ln(RealPrice)
11.098***
(0.2104)
0.0005***
(0.0001)
-5.18e-08***
(1.69e-08)
0.0454
(0.0591)
-0.0294*
(0.0160)
-0.0417***
(0.0103)
0.0020
(0.0013)
0.0291*
(0.0157)
0.0921***
(0.0146)
-0.0767***
(0.0204)
0.1142***
(0.0198)
0.1741***
(0.0248)
0.0976***
(0.0140)
-0.1526***
(0.0326)
-0.1378*
(0.0707)
0.0971**
(0.0475)
-0.0669**
(0.0336)
-0.1017***
(0.0309)
0.1845
(0.1700)
0.0176
(0.0344)

Dependent Variable
RealPrice
Ln(RealPrice)
111190***
11.003***
(20034)
(0.1426)
26.084**
0.0005***
(10.969)
(0.0001)
0.0101***
-5.43e-08***
(0.0031)
(1.69e-08)
5230.0
0.0988
(8346.0)
(0.0617)
-4660.2**
-0.0401**
(2358.4)
(0.0165)
-7815.2***
-0.0416***
(1496.0)
(0.0100)
534.25***
0.0016
(174.84)
(0.0013)
714.79
0.0318**
(2316.4)
(0.0152)
14631***
0.0978***
(2303.7)
(0.0147)
-8758.1***
-0.0692***
(2768.4)
(0.0202)
13070***
0.1118***
(2255.9)
(0.0194)
19205***
0.1728***
(2880.1)
(0.0239)
11140***
0.0946***
(1879.1)
(0.0139)
-22411***
-0.1365***
(4708.9)
(0.0318)
-15821*
-0.1668**
(8146.4)
(0.0729)
24540***
0.2710***
(7758.6)
(0.0652)
-13972***
-0.1157***
(4481.3)
(0.0334)
-9546.4**
-0.0883***
(4258.4)
(0.0297)
9310.5
0.0861
(12102)
(0.0954)
3949.5
0.0381
(4070.2)
(0.0323)
45

RealPrice
103185***
(16218)
26.124**
(10.644)
0.0099***
(0.0031)
13631
(8497.4)
-6754.7***
(2418.1)
-8495.9***
(1430.2)
535.15***
(172.84)
-189.95
(2244.0)
14219***
(2316.9)
-8656.7***
(2753.7)
13446***
(2224.0)
18337***
(2779.2)
11252***
(1869.4)
-20444***
(4708.7)
-21346**
(8596.4)
29352***
(10503)
-18979***
(4713.2)
-6961.7*
(4062.5)
-8104.8
(8277.1)
10413***
(2545.6)

Table 7 (cont’d)
Independent
Variables

Ln(RealPrice)
0.0599
(0.1270)
-0.0160
(0.0236)

DISTpost
DISTpost

2

Ring1.5

-

Ring2

-

Ring2.5

-

Ring3

-

toPowerPlant
toPowerPlant
2000
2001
2002
2003
2004
2005
2006
2007
2008
2009
R-squared

2

-0.2188***
(0.0701)
0.0681***
(0.0263)
0.0518
(0.0505)
0.0120
(0.0439)
-0.00004
(0.0434)
-0.0207
(0.0434)
-0.0140
(0.0435)
-0.0095
(0.0429)
-0.0354
(0.0440)
-0.1190**
(0.0491)
-0.2706***
(0.0785)
-0.6334***
(0.0879)
0.8182

Dependent Variable
RealPrice
Ln(RealPrice)
-8169.9
(8963.2)
411.82
(1436.9)
0.0315
(0.0655)
0.1662***
(0.0584)
0.1387***
(0.0499)
0.0475
(0.0329)
-41448***
-0.2894***
(11844)
(0.0752)
13396***
0.0901***
(3766.0)
(0.0271)
13518*
0.0568
(7616.9)
(0.0476)
7679.8
0.0157
(6164.6)
(0.0401)
5901.7
0.0055
(6080.6)
(0.0397)
2969.3
-0.0076
(6032.7)
(0.0395)
3430.9
-0.0058
(6092.9)
(0.0398)
4947.2
0.0012
(6118.0)
(0.0391)
-918.19
-0.0303
(6408.2)
(0.0402)
-10844
-0.1134**
(7554.2)
(0.0448)
-32619***
-0.2586***
(9342.9)
(0.0758)
-61327***
-0.6463***
(9495.6)
(0.0856)
0.8250

# of Obs.

0.8229

RealPrice
2315.7
(7800.9)
8756.5
(7883.5)
1399.7
(7091.7)
-5975.3
(4240.3)
-50413***
(13548)
17177***
(4087.5)
13513*
(7471.1)
7035.7
(5975.1)
6251.7
(5898.4)
2465.1
(5854.8)
2548.0
(5914.2)
4884.8
(5955.6)
-1622.8
(6241.6)
-11397
(7249.9)
-32988***
(9150.5)
-62431***
(9250.2)
0.8286

885

Notes: All regression results are corrected for heteroskedasticity. Asterisks denote
significance at the 1% (***), 5% (**), and 10% (*) levels.

46

REFERENCES

47

REFERENCES

Abeles-Allison, Mark, and Larry J. Connor. "An Analysis of Local Benefits and Costs of
Michigan Hog Operations Experiencing Environmental Conflicts." Michigan State
University Extension. June 1990.
Barrionuevo, Alexei. "Boom in Ethanol Reshapes Economy of Heartland." The New York
Times. 25 June 2006.
Been , Vicki and Voicu, Ioan, "The Effect of Community Gardens on Neighboring Property
Values" (2006). New York University Law and Economics Working Papers. Paper 46.
Blomendahl, Ben, and Bruce Johnson. "Has the Ethanol Industry Impacted Nebraska Land
Values?” Cornhusker Economics 8 April 2009.
Blomquist, Glenn. "The Effect of Electric Utility Power Plant Location on Area Property Value."
Land Economics 50.1 (1974): 97-100.
Bolitzer, B. and N.R. Netusil. “The Impact of Open Spaces On Property Values In Portland,
Oregon.” Journal of Environmental Management 59 (2000): 185–193.
Boxall, Peter C., Wing H. Chan, and Melville L. McMillan. "The Impact of Oil and Natural Gas
Facilities on Rural Residential Property Values: A Spatial Hedonic Analysis." Resource
and Energy Economics 27 (2005): 248-69.
Boyle, Melissa A., and Katherine A. Kiel. "A Survey of House Price Hedonic Studies of the
Impact of Environmental Externalities." Journal of Real Estate Literature 9.2 (2001): 11744.
Brasington, David M., and Diane Hite. "Demand for Environmental Quality: A Spatial Hedonic
Analysis." Regional Science and Urban Economics 35.1 (2005): 57-82.
Court, A. T. Hedonic Price Indexes with Automotive Examples, Dynamics of Automoblie
Demand, New York, NY: General Motors, 1939.
Cropper, Maureen L., Leland B. Deck, and Kenneth E. McConnell. "On the Choice of Functional
Form for Hedonic Price Functions." The Review of Economics and Statistics 70.4 (1988):
668-75.
Cotti, Chad, and Mark Skidmore. "The Impact of State Government Subsidies and Tax Credits in
an Emerging Industry: Ethanol Production 1980–2007." Southern Economic Journal 76.4
(2010): 1076-93.
Deaton, B J., and John P. Hoehn. "Hedonic analysis of hazardous waste sites in the presence of
other urban disamenities." Environmental Science & Policy 7.6 Dec. (2004): 499-508.
48

Flower, Patrick C., and Wade R. Ragas. "The Effects of Refineries on Neighborhood Property
Values." Journal of Real Estate Research 9.3 (1994): 319-38.
Freeman, III, A. Myrick. "Hedonic Prices, Property Values and Measuring Environmental
Benefits: A Survey of the Issues." The Scandinavian Journal of Economics 81.2 (1979):
154-73.
Gamble, H. B., and R. H. Downing. "Effects of Nuclear Power Plants on Residential Property
Values." Journal of Regional Science 22.4 (1982): 457-78.
Gardner, Bruce. "Fuel Ethanol Subsidies and Farm Price Support." Journal of Agricultural &
Food Industrial Organization 5.2 (2007).
Goodman, Allen C. "Hedonic Prices, Price Indices and Housing Markets." Journal of Urban
Economics 5 (1978): 471-84.
Haas, G. C. A Statistical Analysis of Farm Sales in Blue Earth County, Minnesota, as Basis for
Farm Land Appraisal, Masters Thesis, the University of Minnesota, 1922a.
Hahn, Robert, and Caroline Cecot. "The Benefits and Costs of Ethanol: An Evaluation of the
Government’s Analysis." Journal of Regulatory Economics 35.3 (2009): 275-95.
Halvorsen, Robert, and Raymond Palmquist. "The Interpretation of Dummy Variables in
Semilogarithmic Equations." The American Economic Review 70.3 (1980): 474-75.
Harrison, Jr, David, and Daniel L. Rubinfeld. "Hedonic Housing Prices and the Demand for
Clean Air." Journal of Environmental Economics and Management 5 (1978): 81-102.
Henderson, J. and B. A. Gloy. "The Impact of Ethanol Plants on Cropland Values in the Great
Plains.” Regional Research Working Paper (2008), Federal Reserve Bank of Kansas City.
Herriges, Joseph A., Silvia Secchi, and Bruce A. Babcock. "Living with Hogs in Iowa: The
Impact of Livestock Facilities on Rural Residential Property Values." Land Economics
81.4 (2005): 530-45.
Hoyer, Meghan, and Mike Saewitz. "Study Denounces Project's Impact." The Virginian-Pilot 20
Nov. 2007.
Jackson, Thomas O. “The Effects of Environmental Contamination on Real Estate: A Literature
Review” Journal of Real Estate Literature 9.2 (2001): 93-116.
Kashian, Russ, Mark E. Eiswerth, and Mark Skidmore. "Lake Rehabilitation and the Value of
Shoreline Real Estate: Evidence from Delevan, Wisconsin." The Review of Regional
Studies 36.2 (2006): 221-38.

49

Kiel, Katherine A. "Measuring the Impact of the Discovery and Cleaning of Identified
Hazardous Waste Sites on House Values." Land Economics 71.4 (1995): 428-35.
Kiel, Katherine A., and Katherine T. McClain. "House Prices During Siting Decision Stages:
The Case of an Incinerator From Rumor Through Operation." Journal of Environmental
Economics and Management 28 (1995): 241-55.
Kohlhase, Janet E. “The Impact of Toxic Waste Sites on Housing Values.” Journal of Urban
Economics 30 (1991): 1-26.
Malpezzi, Stephen. “Hedonic Pricing Models: A Selective and Applied Review,” in Housing
Economics and Public Policy: Essays in Honor of Duncan MacLennan, Anthony
O’Sullivan and Kenneth Gibb (Eds.), Blackwell, 2003.
McClelland, Gary H., William D. Schulze, and Brian Hurd. "The Effect of Risk Beliefs on
Property Values: A Case Study of a Hazardous Waste Site." Risk Analysis 10 (1990): 48597.
McCluskey, Jill J., and Gordon C. Rausser. "Hazardous Waste Sites and Housing Appreciation
Rates." Journal of Environmental Economics and Management 45 (2003): 166-76.
McMillen, Daniel P., and Paul Thorsnes. "The Aroma of Tacoma: Time-Varying Average
Derivatives and the Effect of a Superfund Site on House Prices." Journal of Business and
Economic Statistics 21.2 (2003): 237-46.
McNew, Kevin, and Duane Griffith. "Measuring the Impact of Ethanol Plants on Local Grain
Prices." Review of Agricultural Economics 27.2 (2005): 164-80.
Meersman, Tom. "Ethanol plant pose pollution challenge." Star Tribune 8 July 2001
[Minneapolis-St. Paul].
Michaels, R. Gregory, and V. Kerry Smith. "Market Segmentation and Valuing Amenities with
Hedonic Models: The Case of Hazardous Waste Sites." Journal of Urban Economics 28.2
(1990): 223-42.
Nelson, Arthur C., John Genereux, and Michelle Genereux. “Price Effects of Landfills on House
Values.” Land Economics 68 (1992): 359-65.
Nelson, Jon P. "Residential Choice, Hedonic Prices, and the Demand for Urban Air Quality."
Journal of Urban Economics 5.3 (1978): 357-69.
Nelson, Jon P. "Three Mile Island and Residential Property Values: Empirical Analysis and
Policy Implications." Land Economics 57.3 (1981): 363-72.
Niven, Robert K. "Ethanol in Gasoline: Environmental Impacts and Sustainability Review
Article." Renewable and Sustainable Energy Reviews 9.6 (2005): 535-55.

50

Palmquist, Raymond B. "Estimating the Demand for the Characteristics of Housing." The
Review of Economics and Statistics 66.3 (1984): 394-404.
Palmquist, Raymond B., and Leon E. Danielson. "A Hedonic Study of the Effects of Erosion
Control and Drainage on Farmland Values." American Journal of Agricultural Economics
71 (1989): 55-62.
Palmquist, Raymond B., Fritz M. Roka, and Tomislav Vukina. "Hog Operations, Environmental
Effect, and Residential Property Values." Land Economics 73.1 (1997): 114-24.
Pimentel, David. "Ethanol Fuels: Energy Security, Economics, and the Environment." Journal of
Agricultural and Environmental Ethics 4.1 (1991): 1-13.
Pimentel, David. "Ethanol Fuels: Energy Balance, Economics, and Environmental Impacts Are
Negative." Natural Resources Research 12.2 (2003): 127-34.
Pollakowski, H. O. "Data Sources for Measuring Housing Price Changes." Journal of Housing
Research 6.3 (1995): 377-87.
Ridker, Ronald G., and John A. Henning. "The Determinants of Residential Property Values
with Special Reference to Air Pollution." Journal of Political Economy 82 (1967): 34-55.
Reichert, Alan K., Michael Small, and Sunil Mohanty. "The Impact of Landfills on Residential
Property Values." Journal of Real Estate Research 7.3 (1992): 297-314.
Robison, Lindon J., and Bryan K. Ritchie. Relationship economics: the social capital paradigm
and it's application to business, politics and other transactions. N.p.: Ashgate Publishing
Co., 2010.
Rosen, Sherwin. "Hedonic Prices and Implicit Markets: Product Differentiation in Pure
Competition." Journal of Political Economy 82.1 (1974): 34-55.
Runge, C. Ford, and Benjamin Senauer. "How Biofuels Could Starve the Poor." Foreign Affairs
86.3 (2007): 41-53.
Saphores, Jean-Daniel, and Ismael Aguilar-Benitez. "Smelly Local Polluters and Residential
Property Values: A Hedonic Analysis of Four Orange County (California) Cities." Estudios
Economicos 20.2 (2005): 197-218
Searchinger, T. R. Heimlich, R.A. Houghton, F. Dong, A. Elobeid, J. Fabiosa, S. Tokgoz, D.
Hayes, and T. Yu. “Use of U.S. Croplands for Biofuels Increases Greenhouse Gases
Through Emissions from Land-Use Change.” Science 319 (2008): 1238-1240.

Sirmans, G. Stacy, David A. Macpherson, and Emily N. Zietz. "The Composition of Hedonic
Pricing Models." Journal of Real Estate Literature 13.1 (2005): 1-44.

51

Smolen, Gerald E., Gary Moore, and Lawrence V. Conway. "Economic Effects of Hazardous
Chemical and Proposed Radioactive Waste Landfills on Surrounding Real Estate Values."
Journal of Real Estate Research 7.3 (1992): 283-95.
Thayer, M., H. Albers and M. Rahmatian. “The Benefits of Reducing Exposure to Waste
Disposal Sites: A Hedonic Housing Value Approach.” Journal of Real Estate Research 7.3
(1992): 265-82.
Turnquist, Alan & Fortenbery, T. Randall & Foltz, Jeremy. "Progress or Devastation? The
Effects of Ethanol Plant Location on Local Land Use," 2008 Annual Meeting, July 27-29,
Orlando, Florida, American Agricultural Economics Association.
Urbanchuk, John M. "Contribution of the Ethanol Industry to the Economy of the United States."
Renewable Fuels Association. N.p., 12 Feb. 2010.
Wallace, Henry A. "Comparative Farm-Land Values in Iowa." Journal of Land and Public Utility
Economics 2 (1926): 92-385.
Witte, Ann D., Howard J. Sumka, and Homer Erekson. "An Estimate of a Structural Hedonic
Price Model of the Housing Market: An Application of Rosen's Theory of Implicit
Markets." Econometrica 47.5 (1979): 1151-73.
Zabel, Jeffrey E., and Katherine A. Kiel. "Estimating the Demand for Air Quality in Four U.S.
Cities." Land Economics 76.2 (2000): 174-94.

52