Land Use Policy 19 (2002) 91–98

The value of open spaces in residential land use

Jacqueline Geoghegan*
Department of Economics, Clark University, 950 Main Street, Worcester MA 01760, USA
Received 30 January 2001; received in revised form 11 July 2001; accepted 16 August 2001

Abstract

The preservation of open spaces has become an important policy topic in many regions. Policy tools that have been used include:
cluster zoning; transferable development rights; proposed land taxes to fund purchases of remaining open spaces; and private
organizations that buy land. This paper develops a theoretical model of how different types of open spaces are valued by residential
land owners living near these open spaces, and then, using a hedonic pricing model, tests hypotheses concerning the extent to which
these different types of open spaces are capitalized into housing prices. The empirical results from Howard County, a rapidly
developing county in Maryland, USA, show that ‘‘permanent’’ open space increases near-by residential land values over three times
as much as an equivalent amount of ‘‘developable’’ open space. This methodology can be used to help inform policy decisions
concerning open space preservation, such as effectively targeting certain areas for preservation, or as a means of creative financing of
the purchase of conservation easements, through the increase in property taxes, resulting from the associated increase in property
values. r 2002 Elsevier Science Ltd. All rights reserved.

Keywords: Open space; Land preservation; Hedonic models

Introduction space preservation: in 2000, 174 out of 208 ballot issues

passed generating $7.4 billion in conservation funding;
The preservation of open space is an important policy in 1999, $1.8 billion was authorized; and in 1998, $8.3
issue in the United States and elsewhere, as land in billion (Land Trust Alliance, http:\\www.lta.org).
forestry and agricultural uses have been converted into As part of the policy discussion and to better inform
residential and commercial uses. From 1982 to 1992, policy decisions concerning how these monies can be
approximately 6.2 million acres of agricultural land and efficiently spent, further research into the benefits of
5.1 million acres of forested land were converted to open spaces is needed. While there are a myriad of
urban and other developed uses in the United States potential benefits associated with preserving open
(Vesterby et al., 1997). Open space has associated with it spaces, this paper focuses specifically on the benefits of
many potential public goods, such as aesthetic, recrea- local open spaces to residential consumers. This paper
tion, and biodiversity values as well as other associated hopes to contribute some insights into this issue by
ecosystem services, for example flood control and water developing a theoretical model of the economic value of
purification. In response to the market failure associated open spaces in individual residential location choice,
with public goods, responses have been varied: cluster and subsequently estimating a hedonic model of
zoning; transferable development rights; proposed land residential land values that includes as explanatory
taxes to fund purchases of remaining open spaces; and variables measures of the different types of open spaces
private organizations that buy land such as The Nature around each residential location. Specifically, this
Conservancy and The Trust for Public Lands. Indeed, current research focuses on the differential value of
high public interest in the United States has been ‘‘permanent’’ and ‘‘developable’’ open spaces have on
demonstrated in the results from the last three elections near-by residential land prices.
from ballot initiatives to generate funding for open The paper is structured as follows: the next section
gives a brief background on relevant policies on
*Tel.: +1-508-793-7226; fax: +1-508-793-8849.
preserving open space in the United States, followed
E-mail address: [email protected] (J. Geoghegan). by a review of the economic literature. Next a theoretical

0264-8377/02/$ - see front matter r 2002 Elsevier Science Ltd. All rights reserved.
PII: S 0 2 6 4 - 8 3 7 7 ( 0 1 ) 0 0 0 4 0 - 0
92 J. Geoghegan / Land Use Policy 19 (2002) 91–98

model of residential location choice is developed, lands (Maryland Office of Planning, hptt:\\www.op.sta-
followed by the empirical specification of a case study te.md.us).
in a county in suburban Maryland with tests of the
theoretical hypotheses, with econometric results. The
paper concludes with a brief discussion and suggestions Review of economic literature
for further research.
Questions concerning the economic value of open
spaces have been addressed by economists using two
Policies in the United States broad methodological approaches for estimating eco-
nomic values of market and nonmarket goods and
In the United States, a myriad of local, state, and services: stated preference and revealed preference. The
federal regulations have the potential to affect land use former approach relies on survey techniques to elicit
patterns (Platt, 1996). Land use planning and regulation individual preferences and values for economic goods
is usually a function of state and local governments, and services, while the latter approach uses observed
traditionally performed through zoning regulations and market choices that individuals make to reveal their
subdivision ordinances. In addition, federal government underlying preferences and to estimate their values for
activities can indirectly affect land use patterns. These goods and services (Freeman, 1993). In the survey
include the allocation of resources for the transportation approach, for example, respondents are asked how
network and the deductions of mortgage interest and much they are willing to pay to preserve land in open
property taxes in the federal income tax code. uses. The value of preserving agricultural land uses was
Zoning remains the most prevalent land use control estimated using this approach by Halstead (1984),
tool in affecting the location of different land uses, but Beasley et al. (1986), and Kline and Wichelns (1994,
in general zoning can only affect the density of 1996a, b, 1998). In this latter series of papers, the
residential development and it is not prohibited out- authors find empirical evidence that in a region of the
right. In addition, there are a few policy ‘‘levers’’ United States, high population and property values lead
available to local governments that can be used to affect to greater support for agricultural land use preservation
the incentives for development, either through manda- programs. This approach has also been used to estimate
tory or voluntary regulations. Examples of mandatory the value of preserving open space in an urban setting in
regulations include urban growth boundaries, adequate the US (Breffle et al., 1998).
public facilities ordinances, differential development The revealed preference approach, which relies on
fees, down-zoning, and public provision of water and market transactions and the hedonic pricing methodol-
sewer. Governments can also use voluntary market ogy, has also been used to estimate the value of open
mechanisms such as agricultural and other conservation spaces. While further development of the hedonic
easements, transferable development rights to create pricing model approach is given below, as this is the
incentives for development to occur in specific regions, modeling framework for the empirical study in this
or they can explicitly buy parcels of remaining open paper, the basic concept is that a residential property is a
space. heterogeneous good made up of a bundle of character-
Results from a recent survey show that 46 of the 50 istics, each of which contributes to the sales price of the
US states have some form of open space preservation good, including environmental attributes of the residen-
program in place (National Governors’ Association, tial parcel, such as the amount of open space in the
http:\\www.nga.org). These include using different neighborhood of the parcel.
funding sources such as real estate transfer fees and The early literature in this area focused on the value
hunting licenses to generate monies to purchase land or of urban parks and greenbelts in residential property
conservation easements as well as grants for habitat values, with statistical results that showed that the
protection and restoration. In 1997, the General farther away a residential property was from the
Assembly in the state of Maryland (the location of the greenbelt or park, the less its selling price, controlling
empirical case study below) adopted a number of for all other attributes (Kitchen and Hendon, 1967;
programs to implement the ‘‘Smart Growth’’ initiative Weicher and Zerbst, 1973; Hammer et al., 1974;
of Governor Parris Glendening. The goals of Smart McMillan, 1974; Correll et al., 1978; Peiser and
Growth in Maryland are to revitalize previously Schwann, 1993). Later research analyzed the value of
developed areas, preserve open space, and to discourage British forest land in residential prices in a series of
sprawl development in rural areas. The state directs papers by Garrod and Willis (1992a, b, c), similarly in
support and infrastructure funding to locally designated Tyrvainen and Meittinen (2000) as well as the effects of
‘‘Priority Funding Areas’’ and in addition, a ‘‘Rural a golf course on residential land values in Do and
Legacy Program’’ provides funding for the purchase of Grudnitski (1995). Ready et al. (1997) compare stated
conservation easements on agricultural and forested preference to hedonic models for measuring the amenity
 J. Geoghegan / Land Use Policy 19 (2002) 91–98 93

benefits from farmland and find that the two methodol- model. There is a rich history of spatial economic
ogies generate estimates within twenty percent of each models of land use, which can be grouped into two basic
other. categories: monocentric city models and public finance/
In a series of papers by Bockstael and colleagues (Bell spatial amenities models (Mieszkowski and Mills, 1993).
and Bockstael, 2000; Geoghegan et al., 1997; Leggett The modern monocentric city model of urban spatial
and Bockstael, 2000) the authors find empirical effects structure, where the distribution of land uses on a
of different kinds of open space, measured at different featureless plain around a central business district
scales, on residential land values. For example, in (CBD) is a result of an equilibrium between the
Geoghegan et al. (1997), the authors find that within a declining land rent gradient and increasing transporta-
tenth of a kilometer radius, the proportion of open tion costs, was developed by Alonso (1964), Mills
space positively impacts land values, but within a 1-km (1967), and Muth (1969). In more recent times, these
buffer this variable negatively influences land prices. models have become more realistic and complicated by
They interpret this result to suggest that individuals incorporating polycentric cities, different expectations
value open space as a view from their house, but that concerning the future, etc. (for a complete review of this
individuals prefer more diverse land uses at the larger literature, see Anas et al., 1998).
scale. For example, this could be interpreted that The public finance/spatial amenities models are
individuals prefer to be able to walk to other important derived from the Tiebout (1956) model, where indivi-
land uses from their houses, such as commercial land duals choose their residences based on location-specific
uses for local shopping. public goods. That is, individuals ‘‘vote with their feet’’
For a case study in England, Cheshire and Sheppard for a package of housing services, as well as local fiscal
(1995) use a hedonic model to analyze the difference and nonfiscal amenities in their choice of residential
between the effects of publicly and privately owned open location (Hoyt and Rosenthal, 1997). The applied
space amenities on residential property in two towns. literature, largely based on aggregate spatial data such
They include many location variables to capture local as census data, has incorporated aspects of both strands
spatial amenities. However, given data constraints, these of theoretical models, for example, by including as
authors can only measure their variables of interest at an explanatory variables distance measures and neighbor-
aggregate level of 1 km2. Variables are measured by hood measures, such as Dubin (1988, 1992), Can (1990,
placing each residential observation in a 1-km grid, then 1992) and McMillen (1992, 1995). The model developed
extracting the percent of the relevant land uses in that in this paper similarly incorporates both strands of the
1 km2. This results in one aggregate variable for all literature by including both aspects of location. That is,
observations in 1 km2. They use variables measuring the it is assumed that individuals have preferences on their
‘‘percent of land in accessible open space’’, that is land locations: specifically, as the location relates to com-
with public access and ‘‘percent of land in inaccessible muting distances as well as the spatial amenities
open space’’, land with no provision of public access to surrounding their location.
capture the difference between use and nonuse values of The theoretical model is an extension of Brueckner
the open space. They find the impact depends on the (1983). It is assumed that there are identical consumers
amount of open space amenities in two towns. with the utility function:
While these two variables are similar in spirit to the Uðc; q; y; spermanent ; sdevelopable Þ; ð1Þ
‘‘permanent open space’’ and ‘‘developable open space’’
variables in this paper, there are important differences. where c is the composite nonhousing numeraire good, q
The data for this current paper are spatially explicit and the consumption of housing floor space, y the consump-
disaggregate, so the variables included in this paper vary tion of private yard space, spermanent the permanent open
for each observation. Rather than examining the space surrounding parcel and sdevelopable the developable
influence of access, the variables in this paper are used open space surrounding parcel.
to examine the willingness to pay higher prices for the There are two potential types of open space around a
availability of open space around a residential location. residential parcel: permanent open space, such as parks
Because the data contain permanently preserved as well and lands that have conservation easements (i.e. have
as developable agricultural and forestland, this paper sold their development rights); and developable open
measures the effect of both current land uses and future space, such as privately owned forested land and
expectations of surrounding land uses. agricultural land. In the model, the open space goods
are public goods, not individually owned by a residential
homeowner. But it is hypothesized that consumers
The theoretical model prefer more open space to less and this will be
capitalized into the selling price of the home. Consumers
The theoretical model developed in this paper follows are assumed to be myopic, that is, they do not expect
in the tradition of the basic residential location choice that the farm next to their neighborhood will get
94 J. Geoghegan / Land Use Policy 19 (2002) 91–98

converted to residential use at any time soon, although and the recreational and aesthetic value of water
we do expect individuals to value ‘‘permanent’’ open (Lansford and Jones, 1995).
space higher. This differentiation between types of open In general, hedonic pricing models assume that the
space is the addition to the Brueckner (1983) model as heterogeneous housing good is comprised of a bundle of
well as the empirical test of the hypotheses developed. characteristics, such as
It is assumed that all consumers receive the same
income m at their place of employment in the CBD, so R ¼ a þ Sb þ Lg þ Gt þ e; ð4Þ
that in equilibrium, consumers have a uniform level of
utility. All consumption bundles must satisfy where R is an (n  1) vector of housing prices, S is an
Uðc; q; y; spermanent ; sdevelopable Þ ¼ U; where U is the uni- (n  k) matrix of parcel/structure characteristics, L is an
form utility level. For each consumer % to %reach U; the (n  l) matrix of neighborhood characteristics, G is an
rental payment R for a house located x miles from % the (n  m) matrix of spatial and locational variables, a; b; g;
CBD with a given q; y; spermanent ; and sdevelopable ; must be t are the associated parameter vectors and e is an (n  1)
vector of random error terms.
Uðm  tðxÞ  R; q; y; spermanent ; sdevelopable Þ ¼ U; ð2Þ A specific version of this general hedonic pricing
%
where tðxÞ is the commuting cost from distance x: This model is estimated to test the hypotheses that different
equation implicitly defines the consumer bid-rent func- types of open space around a residence contributes
tion positively to housing values. The empirical application is
focused on one suburban county of Washington, DC
R ¼ Rðq; y; spermanent ; sdevelopable ; x; m; UÞ: ð3Þ and Baltimore, MD that is undergoing some of the most
%
rapid growth in Maryland: Howard County. As tax
To test this theory, a hedonic model of land values is rates, public services, school quality, and crime rates
developed to test the hypotheses that individuals are vary over counties, but not within counties in this area,
willing to pay more for locations with greater private the model includes a single county in order to focus
open space and public open space, controlling for the specifically on the land use issue and to abstract away
other variables that affect residential land values as from these other issues.
suggested by the above theoretical model, such as The data for this paper are from the Maryland Office
distance to CBD, other spatial amenities, and char- of Planning’s (MOP) geocoded data of land parcels and
acteristics of the parcel and housing structure. associated sales transactions including descriptions of
the parcel and associated housing structure. The sales
transactions used to estimate the hedonic model
The empirical model included all geocoded and consistent transactions
between September 1993 and June 1996 of residential
Hedonic models are reduced form statistical models properties, resulting in a dataset of 5599 observations.
that seek to trace out, at a point in time, the locus of The selling prices were deflated annually using the urban
equilibrium transactions prices as a function of the Consumer Price Index. Data from the 1990 US
characteristics of the heterogeneous real estate trans- Population are used to control for neighborhood effects.
acted. Innumerable applications exist in the literature, Data on land use are also from the Maryland Office of
dating from the 1950s. Because real estate is a Planning, which is used as a ‘‘base map’’ for land use,
complicated good with many dimensions, differences with further information on conservation easements
in selling prices of houses will be dictated by a number from the Maryland Department of Natural Resources
of factors, including the quality of the housing structure added to this base map, for the creation of the
on the property, neighborhood characteristics, the ‘‘permanent open space’’ land use category. The data
accessibility to the CBD, as well as the environmental on conservation easements were not in a map format;
amenities associated with the property. In the past the dataset only included land area of the easement and
several decades, environmental economists have used the location of the centroid of the easement parcel.
these types of models to reflect the contribution of Therefore, in order to add these data to the land use
environmental amenities to the selling price of a house. map, all of these parcels were assumed to be circular in
The first application in environmental economics shape, with an associated radius so that the area of each
(Ridker and Henning, 1967) tested the relationship circle was the area of the conservation easement parcel
between property values and air quality and did indeed in the original dataset. See Fig. 1 for this final land use
find a statistically significant relationship (for further map and associated transaction points for the hedonic
information on hedonic models, see Freeman (1993)). model. This approach will add some amount of error to
Other environmental applications include: the value of the calculations of land use areas, but as long as these
agricultural erosion (Dorfman et al., 1996); the impact errors are not systematic in any way, the estimation
of hazardous waste sites (Michaels and Smith, 1990); results will not be biased.
 J. Geoghegan / Land Use Policy 19 (2002) 91–98 95

Fig. 1. Howard County Land use.

In explaining the variation in the transaction price of principal employment centers in the region, as well as
the parcel, four types of explanatory variables are used, the distance to nearest town (DTOWN). The theoretical
which are the specific variables for the general variables model developed in this paper predicts that the
in Eq. (4). All of the variables, with brief descriptions, coefficients on the distance variables to be negative,
can be found in Table 1. The parcel characteristic’s set reflecting the trade-off between housing prices and
includes the lot size (LOTSIZE), the year the house was commuting costs. To account for neighborhood
built (YRBLT), and square footage of the house characteristics, the third set of variables include
(SQFT). As with many previous hedonic models, the information on the block group that each parcel is in,
price of a parcel is assumed to be nonlinearly related to which were extracted from the 1990 US Census of
these variables. Thus, the natural log of these variables Population. These include: population density (POP-
are used. Binary variables are included to control for DEN); percent of population with college education
the assessor’s perceived quality of the house (FAIR, or higher (PEDUCB); and the natural log of the
AVER, GOOD, VGOOD, with the omitted category median household income (LNINC). It is expected that
of FAIR in the regression below), and the number higher population density would decrease housing
of stories (NOSTORY) in the house. It is expected prices, while education level and median household
that lot size, size of the house, and quality of income would increase housing prices, holding all else
construction will all contribute positively to the selling equal.
price of the house. As the county does not have a Finally, indices were constructed to calculate the
significant amount of ‘antique’ houses, which could amount of ‘‘developable open space’’ (DEVPOPEN)
command a premium price as elsewhere in Maryland, it and ‘‘permanent open space’’ (PERMOPEN) that
is expected that newer houses will also have higher surrounds each parcel within a 1600-km radius. This
prices, all else equal. distance was used to capture the open spaces that could
The second set of explanatory variables relates to the be seen in a 20-min walk from each parcel. The land
location attributes of the parcels. This includes the uses that were included in ‘‘developable open space’’
Euclidean distance from each parcel to Washington, DC included agricultural cropland, pasture, and forest,
(DISTDC) and Baltimore, MD, (DISTBALT) the two and ‘‘permanent open space’’ included parks and lands
96 J. Geoghegan / Land Use Policy 19 (2002) 91–98

Table 1
Variable names and definitions (n ¼ 5599)

Variable name Definition Units Mean value

PRICE Deflated transaction price \$ (1996=\$1) 241,000

LOTSIZE Size of parcel Acres 3.5
YRBLT The year the house was built yr 1982
SQFT Size of house Square feet 2300
NOSTORY Number of stories in house Count 1.8
FAIR Quality of house 0, 1 0.03
AVER Quality of house 0, 1 0.57
GOOD Quality of house 0, 1 0.36
VGOOD Quality of house 0, 1 0.04
DISTDC Euclidean distance to Washington, DC km 39
DISTBALT Euclidean distance to Baltimore, MD km 15
DISTTOWN Euclidean distance to nearest town km 5
POPDEN Population density in block group Individual/km2 0.0006
PEDUCB Percent of individuals in block group with bachelor’s degree Percent 17
INCOME Median income of block group \$ (1990) 63,000
DEVPOPEN % of land in 1600-m buffer in developable open space Percent 35
PERMOPEN % of land in 1600-m buffer in permanent open space Percent 12

with conservation easements. The main theoretical higher quality all contribute positively to the selling
hypotheses to be tested in the empirical model are that price of a house. As is predicted by the theoretical
each of the estimated coefficients on these open space model, the farther a house is from the major employ-
variables is positive, while the coefficient on the ment centers of Washington, DC and Baltimore, MD,
permanent open space variable is of larger magnitude the lower the selling price; both of these estimated
than the estimated coefficient on the developable open coefficients are negative and statistically significant. The
space variable. distance to nearest town variable is positive and
As hedonic models are reduced form models of an statistically significant. While towns can also be employ-
equilibrium locus of offers and bids for residential land, ment centers as well as locations of stores and other
economic theory cannot inform the empirical specifica- amenities, the positive coefficient suggests that there
tion (see Cropper et al. (1988) for further discussion). might be congestion effects occurring near these towns,
Therefore, following previous research, (Bockstael, so that individuals are not willing to pay as much for a
1996; Geoghegan et al., 1997) flexible function forms house there, all else equal. The census neighborhood
were used, with a semi-log specification as the best fit. variables are also all statistically significant and of the
The equation that is estimated is expected sign. Higher population density in a neighbor-
ln Price ¼ b1 þ b2 ln LOTSIZE þ b3 ln YRBLT hood decreases housing values, while higher incomes
and education levels increases housing values, again,
þ b4 ln SQFT þ b5 NOSTORY
holding all else equal.
þ b6 AVER þ b7 GOOD þ b8 VGOOD The main hypothesis to be tested from the theoretical
þ b9 DISTDC þ b10 DISTBALT model and the policy question raised in this paper,
þ b11 DISTTOWN þ b12 POPDEN concern the estimated coefficients on the open space
variables. The estimated coefficients on both permanent
þ b13 PEDUCB þ b14 ln INCOME
open space and the developable open space are positive,
þ b15 DEVPOPEN þ b16 PERMOPEN: ð5Þ with the estimated coefficient on permanent open space
being over three times the magnitude as the estimated
coefficient on developable open space, as was hypothe-
Estimation results and discussion sized. However, only the estimated coefficient on
permanent open space is statistically significant at the
The results of this estimation are presented in Table 2. 5% level, while the estimated coefficient on the
A Cook–Weisberg test for heteroskedasticity, using the developable coefficient is statistically significant at
fitted values of the dependent variable, fails to reject the slightly less than the 10% level, with a p-value of 0.08.
null hypothesis of homoskedasticity. The estimated Nonetheless, this result does suggest that individuals
coefficients on all the structural characteristics are value permanent open space more than developable
positive and statistically significant, as expected. Larger open space, as they are willing to pay more to live near
lot sizes, larger houses, newer houses, and houses of permanent open space, all else being equal.
 J. Geoghegan / Land Use Policy 19 (2002) 91–98 97

Table 2 a return on the land preservation investment, which

Regression resultsa could be targeted to finance further conservation
Variable name Estimated coefficient T-statistic easements or more simply as general local government
b revenue. Future research, therefore, includes using
CONSTANT 8.55 11.42
LNLOTSIZE 0.081b 11.42 the current residential land tax for each parcel to
LNYRBLT 0.066b 5.66 compute the expected increase in county tax revenue
LNSQFT 0.211b 15.34 generated over 20 yr by the current land preservation
NOSTORY 0.109b 7.76 programs in Howard County, and then to compare this
AVER 0.276b 9.68
revenue to the cost of preserving the lands in question.
GOOD 0.464b 15.09
VGOOD 0.627b 15.77 Then simulations can be performed to investigate
DISTDC –4.45e–06b 3.99 where the program could be spatially targeted in the
DISTBALT 6.38e–06b 5.65 county to generate the largest net local benefits of the
DISTTOWN 1.29e–05b 6.04 program.
POPDEN 24.88b 2.15
PEDUCB 0.341b 3.16
LNINCOME 0.133b 5.25
DEVPOPEN 0.074 1.74
PERMOPEN 0.257b 4.84 Acknowledgements
a 2
Dependent variable: LNPRICE, N ¼ 5599; R ¼ 0:34:
b
Signifies statistical significance at the 5% level. The research is this paper was supported by US EPA
Grant #R825309 and NASA New Investigator Program
in Earth Sciences Grant # NAG5-8559. I would like to
thank Laura Schneider, Nancy Bockstael, Lori Lynch
Conclusions and Shawn Bucholtz for assistance with the spatial data
as well as two anonymous reviewers for their very
As a greater number of local governments develop helpful comments.
policy initiatives to preserve open space, research, such
as presented in this paper, can help contribute to the
analysis and design of such policies, particularly in
assisting in target economic-based programs, such as the References
purchase of development rights, conservation ease-
ments, or complete parcels. Using a spatially explicit Alonso, W., 1964. Location and Land Use. Harvard University Press,
approach, as was done here, can also help in targeting Cambridge.
specific locations for the spending of public monies, and Anas, A., Arnott, R., Small, K., 1998. Urban spatial structure. Journal
of Economic Literature 36 (3), 1426–1464.
can potentially demonstrate the trade-off between Beasley, S.D., Workman, W.G., Williams, N.A., 1986. Estimating
buying more area in rural areas, where land values are amenity values of urban fringe farmland: a contingent valuation
generally cheaper, versus buying land in more densely approach: note. Growth and Change 17 (4), 70–78.
populated regions, where land is more expensive, but Bell, K.P., Bockstael, N.E., 2000. Applying the generalized method of
moments estimation approach to spatial problems involving
could potentially contribute further to local housing
microlevel data. Review of Economics and Statistics 82 (1), 72–82.
values. Or, given further data on environmental values, Bockstael, N.E., 1996. Modeling economics and ecology: the
such as habitat and watershed protection benefits, this importance of a spatial perspective. American Journal of Agri-
modeling approach could be used to rank parcels cultural Economics 78 (5), 1168–1180.
according to these values as well. Breffle, W.S., Morey, E.R., Lodder, T.S., 1998. Using contingent
The definition and calculation of the ‘‘permanent valuation to estimate a neighbourhood’s willingness to pay to
preserve undeveloped urban land. Urban Studies 35 (4), 715–727.
open space’’ variable includes farms that have sold their Brueckner, J.K., 1983. The economics of urban yard space: an
development rights. The positive and statistically ‘‘implicit-market’’ model for housing attributes. Journal of Urban
significant coefficient on this variable in the hedonic Economics 13 (2), 216–234.
regression demonstrates that individuals value this type Can, A., 1990. The measurement of neighborhood dynamics in urban
of near-by farmland in their residential location choice, house prices. Economic Geography 66 (3), 254–272.
Can, A., 1992. Specification and estimation of hedonic housing price
so these values are not only potentially capitalized into models. Regional Science and Urban Economics 22 (3), 491–508.
the individual farmland prices (Nickerson and Lynch, Cheshire, P., Sheppard, S., 1995. On the prices of land and the value of
2001), but also into the surrounding residential land amenities. Economica 62 (246), 247–267.
prices. Further research will focus on the relationship Correll, M.R., Lillydahl, J.H., Singell, L.D., 1978. The effects of
greenbelts on residential property values: some findings on the
between how increases in nearby residential parcels from
political economy of open space. Land Economics 54 (2), 207–217.
conservation easements increase the assessed value of a Cropper, M.L., Deck, L.B., McConnell, K.E., 1988. On the choice of
property. In turn, this could potentially lead to an functional form for hedonic price functions. Review of Economics
increase in the land tax revenue from the parcel, yielding and Statistics 70 (4), 668–675.
98 J. Geoghegan / Land Use Policy 19 (2002) 91–98

Do, A.Q., Grudnitski, G., 1995. Golf courses and residential house Lansford Jr., N.H., Jones, L.L., 1995. Marginal price of lake
prices: an empirical examination. Journal of Real Estate Finance recreation and aesthetics: an hedonic approach. Journal of
and Economics 10 (3), 261–270. Agricultural and Applied Economics 27 (1), 212–223.
Dorfman, J.H., Keeler, A.G., Kriesel, W., 1996. Valuing risk-reducing Leggett, C.G., Bockstael, N.E., 2000. Evidence of the effects of water
interventions with hedonic models: the case of erosion protection. quality on residential land prices. Journal of Environmental
Journal of Agricultural and Resource Economics 21 (1), 109–119. Economics and Management 39 (2), 121–144.
Dubin, R., 1988. Estimation of regression coefficients in the presence McMillan, M., 1974. Open space preservation in developing areas: an
of spatially autocorrelated error terms. Review of Economics and alternative policy. Land Economics 50 (4), 410–417.
Statistics 70 (3), 466–474. McMillen, D.P., 1992. Probit with spatial autocorrelation. Journal of
Dubin, R., 1992. Spatial autocorrelation and neighborhood quality. Regional Science 32 (3), 335–348.
Regional Science and Urban Economics 22 (3), 433–452. McMillen, D.P., 1995. Selection bias in spatial econometric models.
Freeman, A.M., 1993. The Measurement of Environmental and Journal of Regional Science 35 (3), 417–436.
Resource Values: theory and Methods. Resources for the Future, Michaels, R.G., Smith, V.K., 1990. Market segmentation and valuing
Washington, DC. amenities with hedonic models: the case of hazardous waste sites.
Garrod, G., Willis, K., 1992a. The amenity value of woodland in Great Journal of Urban Economics 28 (20), 223–242.
Britain: a comparison of economic estimates. Environmental and Mieszkowski, P., Mills, E.S., 1993. The causes of metropolitan
Resource Economics 2 (4), 415–434. suburbanization. The Journal of Economic Perspectives 7 (3),
Garrod, G., Willis, K., 1992b. The environmental economic impact of 135–147.
woodland: a two stage hedonic price model of the amenity value of Mills, E.S., 1967. An aggregative model of resource allocation in a
forestry in Britain. Applied Economics 24 (7), 715–728. metropolitan area. American Economic Review 57 (2), 197–210.
Garrod, G., Willis, K., 1992c. Valuing goods’ characteristics: an Muth, R., 1969. Cities and Housing. University of Chicago Press,
application of the hedonic price method to environmental Chicago.
attributes. Journal of Environmental Management 34 (1), 59–76. Nickerson, C., Lynch, L., 2001. The effect of farmland preservation
Geoghegan, J., Wainger, L., Bockstael, N.E., 1997. Spatial landscape programs on farmland prices. American Journal of Agricultural
indices in a hedonic framework: an ecological economics analysis Economics 83 (2), 341–351.
using GIS. Ecological Economics 23 (3), 251–264. Peiser, R.B., Schwann, G.M., 1993. The private value of public open
Halstead, J.M., 1984. Measuring the nonmarket value of Massachu- space within subdivisions. Journal of Architectural and Planning
setts agricultural land. Journal of the Northeastern Agricultural Research 10 (2), 91–104.
Economics Council 13 (1), 12–19. Platt, R.H., 1996. Land Use and Society: geography, law and public
Hammer, T.R., Coughlin, R.E., Horn IV, E.T., 1974. The effect of a policy. Island Press, Washington DC.
large urban park on real estate values. American Institute of Ready, R.C., Berger, M.C., Blomquist, G., 1997. Measuring benefits
Planners Journal, 40(4), 274–277. from farmland: hedonic pricing vs. contingent valuation. Growth
Hoyt, W.H., Rosenthal, S.S., 1997. Household location and Tiebout: and Change 28 (4), 438–458.
do families sort according to preferences for locational amenities? Ridker, R.G., Henning, J.A., 1967. The determinants of residential
Journal of Urban Economics 42 (2), 159–178. property values with special reference to air pollution. Review of
Kitchen, J.W., Hendon, W.S., 1967. Land values adjacent to an urban Economics and Statistics 49 (2), 246–257.
neighborhood park. Land Economics 43 (3), 357–361. Tiebout, C.M., 1956. A pure theory of local expenditure. Journal of
Kline, J., Wichelns, D., 1994. Using referendum data to characterize Political Economy 64 (5), 416–424.
public support for purchasing development rights to farmland Tyrvainen, L., Meittinen, A., 2000. Property prices and urban forest
programs. Land Economics 70 (2), 223–233. amenities. Journal of Environmental Economics and Management
Kline, J., Wichelns, D., 1996a. Measuring public preferences for the 39 (2), 205–223.
environmental amenities provided by farmland. European Review Vesterby, M., Daugherty, A., Heimlich, R., Claassen, R., 1997. Major
of Agricultural Economics 23 (4), 421–436. land use changes in the contiguous 48 states. AREI Updates. No. 3.
Kline, J., Wichelns, D., 1996b. Public preferences and farmland USDA, ERS, NRED, June.
preservation programs. Land Economics 72 (4), 538–549. Weicher, J.C., Zerbst, R.H., 1973. The externalities of neighborhood
Kline, J., Wichelns, D., 1998. Measuring heterogeneous preferences for parks: an empirical investigation. Land Economics 49 (1),
preserving farmland and open space. Ecological Economics 26 (2), 99–105.
211–224.