Estimating the effect of a land parcel index using hedonic price analysis

Abstract

The purpose of this study was to statistically test the effect of a parcel index – consisting of a combination of a soil index, a fertility index and a location index – intended to be used as a price-determining indicator for the sale of agricultural land at farmland markets. In the hedonic price model, the coefficients of the variables representing parcel index, population, gross return and parcel irrigation investment status were positive and statistically significant at a significance level of 0.01. There was a negative relationship between parcel size and sale price, which implied that the selling price per decare tends to decrease as the parcel size increases. In the study area, the prices of farmland with large parcel sizes and irrigation efficiency investments were higher. The population density in the region and gross income from farmlands were the major factors that generated demand for the land. The hedonic price model establishes an important link between the parcel index and the sale price of farmland. Based on this link, parcel index-based pricing can contribute significantly to the creation of a farmland market in Turkey.

Full text

427
Agricultural Economics – Czech, 68, 2022 (11): 427–433 Original Paper
https://doi.org/10.17221/72/2022-AGRICECON
Supported bythe Scientic Research Projects Coordination Unit ofAkdeniz University inTurkey (Project No.3592).
Estimating the eect ofaland parcel index
using hedonic price analysis
B A, S K*
Department ofAgricultural Economics, Agriculture Faculty, Akdeniz University, Antalya, Turkey
*Corresponding author: skaraman@akdeniz.edu.tr
Citation: Aksu B., Karaman S. (2022): Estimating the eect ofaland parcel index using hedonic price analysis. Agric. Econ.
– Czech., 68: 427–433.
Abstract: epurpose ofthis study was tostatistically test the eect ofaparcel index –consisting ofacombination
ofasoil index, a fertility index and alocation index –intended tobe used asa price-determining indicator for the
saleofagricultural land atfarmland markets. Inthe hedonic price model, the coecients ofthe variables representing
parcel index, population, gross return and parcel irrigation investment status were positive and statistically signicant
atasignicance level of0.01. ere was anegative relationship between parcel size and sale price, which implied that
the selling price per decare tends todecrease asthe parcel size increases. Inthe study area, the prices offarmland with
large parcel sizes and irrigation eciency investments were higher. epopulation density inthe region and gross in-
come from farmlands were the major factors that generated demand for the land. ehedonic price model establishes
animportant link between the parcel index and the sale price offarmland. Based onthis link, parcel index-based pricing
can contribute signicantly tothe creation ofafarmland market inTurkey.
Keywords: asymmetric information; double-log model; farmland market; soil index; Vuong test
e largest share inthe active capital ofagricultural
enterprises island. eland isafactor inproduction,
but unlike other factors of production, it is immov-
able, has a xed supply and is not subject to depre-
ciation (Raup 2003). It is used in both livestock and
plant production. evalue ofland isdirectly aected
bytheproceeds from agricultural production. Estimat-
ing the value ofland isnot easy because ithas numer-
ous variable characteristics, even invery small parcels.
Proceeds from parcels within the immediate vicinity
ofeach other can vary widely.
e market for farmland diers from other mar-
kets inthat the supply offarmland isxed, unlike the
case in many other transactions. eir characteristic
features are that they can only besold where they are
located, that they have their own individual character-
istics, that they can bepurchased and sold atland of-
ces and that their sale istaxable. Inaddition tothese
characteristics, they can also feature aclose relation-
ship tohumans and multi-stakeholder ownership.
e farmland market does not meet the requirements
ofafully competitive market. Since land isaninherently
heterogeneous resource, there can be a limited num-
ber ofbuyers and sellers inthe market, and itisthere-
fore considered an imperfectly competitive market.
e farmland market largely depends on local sup-
ply and demand. Due to the restrictions of the farm-
land market, there are numerous small local markets
in which buyers and sellers operate. A price level
isformed ineach local market. is price level reects
the local forces ofsupply and demand that represent the
utility value for buyers and sellers.

428
Original Paper Agricultural Economics – Czech, 68, 2022 (11): 427–433
https://doi.org/10.17221/72/2022-AGRICECON
Since farmland is an immovable property, trading
activities are limited toownership. Asamatter offact,
what is sold is not the physical land but the right
topossess itor, alternatively, the rent obtained from it.
ecapitalized rent orland price isformed bythe con-
version offarmland rent into money. eprice offarm-
land isdetermined bynatural (all factors including the
soil and the climate), demographic (population), social
(location, access) and economic (capital investments
inland) factors (Drescher etal. 2001; Huang etal. 2006).
Interms ofagricultural use, soil quality, water supply,
farmland yields, parcel size, proximity tomarkets, land
rent and agricultural subsidies are the factors that are
most frequently cited as major determinants of farm-
land prices (Lloyd etal. 1991; Awasthi 2009). Inaddi-
tion, location-specic features that do not reect the
agricultural characteristics of the land are capitalized
inland prices (Spinneyetal.2011).
e main purpose of this study was to statistically
test aparcel index, which was created toserve asthe
rst trading indicator to determine the market price
for farmland. Forthis purpose, ahedonic price model
was estimated for farmland sales inthe study area, and
statistically signicant characteristics were identied
inthe model. e null hypothesis (H0) that farmland
prices are not aected bythe parcel index was tested
against asample ofregional data onactual land market
transactions. is isthe rst study inthe published liter-
ature totest the eect ofaparcel index, which isacom-
bination ofsoil index, fertility indexand location index,
onfarmland prices. Inthis regard, itsuggests how nec-
essary itistouse aparcel index asanindicator ofthe
sale prices offarmland.
MATERIAL AND METHODS
e chosen study area was the rural area of Kirka-
gac, located inthe west ofthe Aegean region ofTurkey.
First, the location, block and parcel details of farm-
lands sold in2015 were obtained from the real estate
oce inthe Kirkagac municipality. Out ofthe target
population list ofpeople engaged inagricultural pro-
duction and those selling farmland in the Kirkagac
region, 164farmers were selected byapurposive sam-
pling method. In November 2016, face-to-face inter-
views were conducted with the farmers who accepted
the invitation, and aquestionnaire was lled in. Some
of the information obtained from the questionnaires
was conrmed by a parcel query application to the
General Directorate of Land Registry and Cadastre.
eland assets, soil classes and parcel characteristics
ofthe rural area of Kirkagac were obtained from the
Manisa land asset publication prepared bythe General
Directorate ofRural Services (MLA 1998).
e hedonic price model. Itisimpossible for buyers
and sellers touse asingle market price for farmland be-
cause each parcel offarmland exhibits aunique com-
bination ofattributes, and therefore, its valuation must
beafunction ofthe quantity and value ofa combina-
tion of the dierent attributes present. Rosen (1974)
denes hedonic prices as'the implicit prices ofattrib-
utes'. ese prices are calculable and implied because
there isno direct market equivalent for them. Equa-
tion(1) indicates the basic hedonic price model:
1
i ij
m
j
j
P X
=
= β
∑
(1)
where: βj –marginal implied price for the characteris-
ticj; Xij–set of explanatory variables.
By adding anerror term toEquation(1), regression
analysis can beused totest the hypotheses ofthe model
and ofβj, and toobtain estimates forβj. Since there are
noguidelines about the functional form ofthe hedonic
price model in terms of economic theory, the Box-
-Cox transformation was applied totest the functional
forms. In this approach, the nonlinear parameter λ
isadded to the dependent and independent variables
(Box and Cox 1964). egeneral hedonic regression
model istherefore asfollows:
[ ] [ ]
12
0
11
2
0, Var
nm
j k ik i
j
k
ii
i i
j
P XZ
E
λλ
= =
=β+β +β +ε
ε= ε=σ
∑∑
(2)
where: β0 –constant term; m–number oftransform-
able variables; n–number ofnon-transformable discrete
variables; Zik–discrete independent variable (irrigation
investment status); εi–residuals that eliminate the homo-
scedasticity restriction; λ1,λ2 –Box-Cox transformations.
e hedonic regression model consists of the de-
pendent variable farmland sales price (Pi), the continu-
ous independent variableXij (parcel index, parcel size,
population ofthe settlement where the land was sold,
and gross return from the land) and the discrete inde-
pendent variableZik (the irrigation investment status),
which is a dummy variable [FigureS1 in electronic
supplementary material (ESM); for the ESM see the
electronic version]. emaximum likelihood ratio test

429
Agricultural Economics – Czech, 68, 2022 (11): 427–433 Original Paper
https://doi.org/10.17221/72/2022-AGRICECON
can be used to determine the functional form of the
hedonic price model. Individual and combined tests
ofBox-Cox parameters give unexpected results. ere-
fore, the Vuong (1989) test can be used as acomple-
mentary test tochoose among four functional forms.
Wecan dene the likelihood ratio for each individual
observation i using the following formula:
( )
( )
12 12
1
2
1
,
1
1
i ii
jj kk j k
n
i
i
n
i
i
i
LR ll ll
n LR
n
Vuong
LR LR
n
=
=
λλ λλ = −




=
−
∑
∑
(3)
where: n – number ofobservations; LRi –likelihood
ratio between the modelsj andk (LRi = llj − llk); llj,llk
–likelihood ratio for the jand kmodels.
e Vuong test statistic isasymptotically distributed
asa standard normal distribution. While the positive
values higher than the critical value Nα/2 (atasigni-
cance level of α) conrm model j, the negative val-
ues lower than Nα/2 conrm model k. Acccordinly,
|Vuong|≤Nα/2 indicates that there isno signicant dif-
ference between the kand jmodels.
RESULTS AND DISCUSSION
Sample characteristics. e buyer evaluates the
utility of the property for their future purposes.
eseller evaluates the utility ofthe proceeds fromthe
sale (or what they will buy with the proceeds from
thesale) inrelation tothe benet ofthe existing prop-
erty tothem. Itis the property's marginal utility that
determines the economic importance ofthe property
tothe potential buyer. When the buyer's and seller's ac-
ceptance prices are equal, itbecomes the sale price.
In the study area, the average sale price for farmland
has been USD 2 493.9. Farmland in this region has
been sold ataminimum ofUSD607.1 and atamaxi-
mum ofUSD5714.3. ecoecient ofvariation (CV)
offarmland sale prices was 52%. erefore, farmland
was oered for sale athighly varying prices (Table1).
e parcel index is calculated by adding together
the soil index, fertility index and parcel location index
scores. Soil index consists ofasoil prole, topsoil tex-
ture, slope ofthe land and other features. Itisavalue
ranging from 0to100(Arici and Akkaya Aslan 2014).
e soil index value was obtained from soil surveys
and maps prepared previously for Manisa. efertility
index was determined according to the fertility indi-
cators ofthe land. Itis avalue ranging from 0 to10.
elocation index iscreated bytaking into considera-
tion the proximity ofthe property toresidential areas,
the geometric shape ofthe parcel, the available trans-
port facilities, the current irrigation status, etc., and
ranges from 1to20.
When calculating the parcel index, 70% of the soil
index (TE) isadded tothe index scores determined for
fertility and location. is index can beused asanin-
dicator for buyers, asitcontains three major elements
ofthe sale offarmland. Previous studies indicated that
the sale price of farmland depends on their fertility
and distance tothe market. ecurrent index was dis-
cussed and tested using alternative weights byTezcan
etal. (2020). Asaresult oftheir study, itwas stated that
when calculating the parcel index, each plot ofagricul-
tural land should add additional criteria toreect its
own characteristics and location. Inother words, itwas
shown that the current index weights cannot bestand-
ard but can be changed for each plot of agricultural
land by experts working on this subject. However,
itshould not beforgotten that the land for which the
parcel index iscalculated isused for agricultural pur-
poses. Minimum and maximum parcel indices of the
farmland inthe study area were 12.0and 83.4, respec-
tively. eaverage parcel index value inthis region was
Table 1. Descriptive statistics
Variable Name ofvariable Unit Expected sign Min. Max. Mean SD CV (%)
Piparcel sale price USD/decare dependent variable 607.1 5 714.3 2 493.9 1 294.6 51.9
x1parcel index – + 12.0 83.4 37.2 15.1 40.6
x2parcel size m2+/– 219.6 50 631.0 7 051.6 6 002.5 85.1
x4population person +130.0 3 181.0 1 346.3 929.0 69.0
x5gross return USD +39.4 4 017.9 751.9 455.9 60.6
x6irrigation investment – + 0 1 0.390 0.489 125.4
CV –coefficient ofvariation
Source: Authors' own elaboration

430
Original Paper Agricultural Economics – Czech, 68, 2022 (11): 427–433
https://doi.org/10.17221/72/2022-AGRICECON
37.2. eCV for the parcel index was 40.6%. is value
indicated that farmland grades vary inthat region (Ta-
ble1). Ahigh parcel index isexpected tohave apositive
eect onthe sale price ofthe parcel.
Parcel size isabasic physical characteristic that isex-
pected toaect the selling price offarmlands because
alarger size offarmland means ahigher overall value
than that of a smaller piece of land. However, their
value per decare decreases ata decreasing rate since
a larger plot of land will attract fewer potential buy-
ers. is situation reects a curvilinear relationship
between the two variables. erefore, the size offarm-
lands isexpected tohave aninverse relationship with
the selling price per decare and isincorporated into the
hedonic equation inanon-linear fashion. eland size
and sales price ofland per decare are expected tohave
astatistically signicant relationship because the ma-
jority offarmlands are small due totheir fragmentation
through inheritance, but the direction ofthis relation-
ship isuncertain. Also, Ritter etal. (2020) stated intheir
study that the size–price relationship may change over
time and may dier for sub-samples.eaverage par-
cel size in the region was 7 051.6 m2. e smallest
of the parcels sold was 219.6 m2 and the largest one
was 50631m2. eCV ofparcel size was 85%. Inother
words, parcel sizes varied greatly and were diverse
innature (Table1).
Demographic factors are addressed in many he-
donic studies either in the form of local population
density or annual population growth rate. Accord-
ing to Palmquist and Danielson (1989), the popula-
tion density ofthe district where the parcel islocated
can be used to measure current population pressure,
while the population growth rate can represent popula-
tiongrowth expectations. Depending onthe population
ofthe settlement, the selling price offarmland varies.
erefore, apositive relationship isexpected tobeob-
served between the population of the settlement and
the selling price of farmland. e average population
ofthe settlements inthe study area was 1346.3people.
epopulation ofthe smallest settlement was 130peo-
ple, whereas the population ofthe largest settlement was
3181people. at makes the CV ofsettlementpopula-
tion 69%. Settlement population data isdiverse.
Gross return iscalculated for the parcel sold. epro-
duction pattern of the parcel, and the income gener-
ated, can beanindicator for buyers because the farmers
who want tobuy aparcel assign aprice for theagricul-
tural land taking into account the return onthe capi-
tal they invest. egreater the expected future returns
onapiece ofland, the higher the present value ofthat
land is expected to be. e gross return per decare
in the study area was USD 751.9. It was found that
farmers generated anincome equal toone-third ofthe
average sales price ofparcels. eCV ofthe gross re-
turn ofparcels was 60.6%. erefore, the gross return
onthe parcels vary. Itisexpected that the income gen-
erated from a parcel will have a positive eect on its
selling price (Table1).
To alarge extent, parcel-specic characteristics de-
termine the productive function and income generating
capacity ofthe parcel. Both the natural and man-made
conditions of the parcel complement each other for
the specic production target, and investments in ir-
rigation eciency increase the value of farmland.
e availability of modern technologies for drainage
and water saving in the parcel increases the income
of the buyer, thereby aecting the selling price. is
variable was included in the model as adummy vari-
able. epresence and absence ofwater eciency and
drainage investments inthe parcels were assigned the
values of1and0, respectively (Table1).
Estimation results. ree basic steps were followed
inthe estimation ofthe hedonic price model. First, Box-
-Cox transforms and Vuong tests were applied to se-
lect the functional form used in the estimation ofthe
hedonic price model. As a result of Box-Cox trans-
formations and Vuong tests, alog–log function form
was chosen from alternative functions (TablesS1, S2
inESM; for the ESM see the electronic version).
Second, statistical tests were performed to conrm
that the hedonic price model assumption was made
using results from ordinary least squares (OLS) re-
gression testing. Inaddition, goodness oft measures
were estimated for the OLS regression. eestimated
hedonic price model ispresented inTable2. ecoef-
cients estimated inTable2 are the implied prices for
each of the characteristics or attributes considered.
In the hedonic price model, the coecients of parcel
index, parcel size, population, gross return and par-
celirrigation investment were positive and statistically
signicant atasignicance level of1%. esettlement
size ofthe region where the land is located is usually
a factor that aects the parcel price. e population
isanindicator ofthe size ofthe settlement. epositive
sign ofthe population indicated that there was aposi-
tive relationship between the price of the parcel and
this variable. epopulation coecient indicated that
when the population inthe settlement increased by1%,
the parcel price increased by0.46%. In other words,
the higher the population was, the wider the non-ag-
ricultural use of the land was. e increased demand

431
Agricultural Economics – Czech, 68, 2022 (11): 427–433 Original Paper
https://doi.org/10.17221/72/2022-AGRICECON
for land due toanincreasing population has apositive
eect on farmland prices. In some areas where little
farmland isoered for sale and there isastrong demand
for land for non-agricultural use, farmland prices may
increase signicantly. Currently, farmland in Turkey
isdistributed and fragmented, with many small parcels
owned by multiple owners through inheritance. Ac-
cording tothe information onthe farmer registration
system inrelation tocultivated farmland, the number
ofparcels owned byagricultural enterprises is6and the
parcel size is13decares (FRS 2016).
Large parcel size is regarded as desirable by farm-
ers for agricultural production with modern and large
equipment. Small parcel size increases production cost s
duetoalack ofeconomies ofscale. Duetothe regional
production pattern, the size of farmland in demand
may vary. Previous research has found that parcel size
has anegative eect onthe sales price ofland. Asthe
size ofthe parcel offarmland oered for sale increases,
the value ofthat farmland perdecare decreases atade-
creasing rate (Brorsen et al. 2015). is means that
alarger parcel isexpected to besold atalower price
per decare than a smaller parcel. e reason is that
fewer buyers compete for larger parcels inthe market.
To check for a quadratic (decreasing) eect of parcel
size on price, the parcel size squared was included
inthe hedonic model. ecoecient ofthe parcel size
(squared) had anegative relationship onthe sale price,
reecting the fact that the sale prices per decare tended
to decrease as the parcel size increased (Figure S2
inESM; for the ESM see the electronic version).
Gross return refers to the gross income obtained
from the market value of products cultivated on the
farmland sold. e expected income from farmland
isgenerally considered tobethe main factor that de-
termines its value. A rational farmland buyer prefers
tobuy the parcel he thinks will benet him the most
asregards the return onthe price he will pay byconsid-
ering the balance between land prices and the income
from the 'use ofthe land'. egross return had aposi-
tive eect onthe parcel sales price and was statistically
signicant atasignicance level of1%. When the gross
return increased by 1%, the price of the farmland in-
creased by0.15%. erefore, buyers take into consider-
ation the gross return ofthe farmland when evaluating
its sale price. Since the demand for farmland in the
study area isoften associated with agricultural produc-
tion rather than non-agricultural use, one can say that
the gross income obtained from the farmland isade-
terminative factor informing the price ofthe farmland.
Itisknown that farmland prices increase when farm-
ing income ishigh (Xuetal. 1993). is relationship
is based, in part, on the income eect experienced
bypotential buyers offarmland.
Table 2. Estimation results for the hedonic price function
Variables (xi)Coecients (βi) SE t-statistic Marginal probability (P)
x10.20 0.06 3.24 0.00
x21.01 0.34 3.00 0.00
x3–0.05 0.02 –2.53 0.01
x40.46 0.03 15.29 0.00
x50.15 0.04 4.00 0.00
x60.22 0.06 3.69 0.00
Constant –1.20 1.45 –0.83 0.41
Diagnostics Te s t Value Level ofsignicance
Heteroscedasticity F-statistic 1.184 0.265***
Multicollinearity VIF 1.420 less than 5
Normality test Jarque-Bera X24.797 0.112
Determination
coecients
R20.756 –
adjusted R20.746 –
Ramsey RESET F(1.157)-statistic 0.090 0.767*
Residual sum ofsquares RSS 13.700 –
F-statistic –81.098 0.000
*, ***Significance atthe 10% and 1% levels, respectively; RESET –regression equation specification error test; VIF –vari-
ance inflation factor
Source: Authors' own elaboration

432
Original Paper Agricultural Economics – Czech, 68, 2022 (11): 427–433
https://doi.org/10.17221/72/2022-AGRICECON
Previous studies included the variables ofsoil quality,
location and fertility individually inthe hedonic price
model (Vasquez et al. 2002). ey found that these
variables generally had a positive eect on the price
offarmland. Inthis present study, lands were classied
bycreating aparcel index combining the soil index and
fertility index, which denes the permanent and vari-
able characteristics ofthe soil, and the location index,
which denes the distance tothe settlements orenter-
prise headquarters. e parcel index is an important
indicator for buyers, asitincludes soil quality, fertility
and location. Ithad apositive eect onthe farmland
price asexpected and was signicant atasignicance
level of1%. When the parcel index increased by1%, the
farmland price increased by0.20%. ehedonic price
model establishes animportant link between the par-
cel index and the sale price offarmland. rough this
link, parcel index-based pricing can make asignicant
contribution to the development of a farmland mar-
ketinTurkey.
Farmlands can beimproved for specic uses, thereby
creating appreciation. Forexample, investment inwater
ponds for agricultural production, modern technolo-
gies for irrigation orerosion, and drainage requirements
all add value tothe farmland. Palmquist and Danielson
(1989) found that farmland values were signicantly af-
fected byboth potential erosion and drainage require-
ments. Installation of modern irrigation systems that
will save water, and soil improvements for drainage, af-
fect the sale price ofthe parcel because they increase the
potential income ofthe buyer. Itwas found that the sales
price offarmland in which irrigation eciency invest-
ments had been made was 22% higher onaverage than
the sales price ofother farmland. Investments infarm-
lands for irrigation eciency in the region contribute
tothe creation ofahigher demand for farmland.
Diagnostics tests. Heteroscedasticity and multicol-
linearity are two common problems that arise when
working with cross-sectional data in econometric
analyses. Multicollinearity islikely toreduce the accu-
racy ofthe estimated parameters. erefore, avariance
ination factor (VIF) was used to reveal any possible
multicollinearity among independent variables and
needs to be dened. e mean VIF of all variables
inboth models was 1.42(ranging from 1.05to1.65).
Anumber less than5 indicated that multicollinearity
was not an issue in these models. Furthermore, the
presence ofheteroscedastic error terms inthe hedonic
price model was tested. Forthis purpose, the White test
result was F-statistic=1.18 (P=0.26), meaning that H0,
the homoscedasticity hypothesis, cannot be rejected.
Since the Jarque and Bera test isP= 0.11, H0, the hy-
pothesis that the error terms are normally distributed,
cannot be rejected. e Ramsey regression equation
specication error test (RESET) test (F-statistic) was
found tobe0.54(P=0.46). is value indicated that
the hypothesisH0, which states that there isnospeci-
cation error inthe model atasignicance level of1%,
should not be rejected. An F-statistic of 81.098 indi-
cated that the model was highly signicant. ecoef-
cient ofdetermination(R2) indicated that 76% ofthe
changes in price, the dependent variable, were ex-
plained by the land characteristic variables included
inthe hedonic price model (FigureS3 inESM; for the
ESM see the electronic version).
Discussion. Most of the studies on the agricultural
land market are not aimed atincreasing the eciency
of the agricultural land market. Studies in published
literature are mostly designed todetermine the factors
aecting the price of agricultural land (Dacko et al.
2021). It is known that the agricultural land market
does not meet the requirements of a perfectly com-
petitive market because ofthe diversity ofagricultural
land. Inthis market, the dierence between the knowl-
edge ofthe buyers and sellers about the characteristics
ofthe agricultural land provides market power tothe
seller and causes the market to be ineective. With
the agricultural land parcel enquiry, basic information
such asarea, quality and location are presented openly
toeveryone without any restrictions, while much in-
formation –price inparticular –isnot presented (Lin
and Zhang 2021). erefore, the parcel enquiry in-
formation isnot sucient when making the decision
topurchase agricultural land. Forthis, it isclear that
there isaneed for aparcel index that buyers can benet
from when making such apurchase.
With this study aparcel index was created, which
is expected to be used as an important indicator for
both buyers and sellers in the agricultural land mar-
ket, and its eect on the agricultural land sales price
was tested. endings showed that the parcel index
for the agricultural land market had asignicant eect
onthe formation ofsale prices. With the parcel index,
asymmetric information conditions were eliminated,
and itisexpected that the agricultural land market will
work more eectively. Agricultural land market trans-
parency improves market eciency (Seifert etal. 2021).
ere are some limitations in this study. First, the
parcel index was calculated for apredetermined area.
erefore, the limitation ofour study is that it could
not determine the role of the parcel index at a na-
tional level. Forthis purpose, the parcel index should

433
Agricultural Economics – Czech, 68, 2022 (11): 427–433 Original Paper
https://doi.org/10.17221/72/2022-AGRICECON
be published in agricultural land consolidation pro-
jects. Second, the rates taken inthe parcel index were
kept constant. ese rates can be changed according
tothe intended use of the agricultural land. Forthis,
while calculating the parcel index ofagricultural lands,
changes can bemade to the rate calculations at are-
gional level.
CONCLUSION
In this study, the parcel index, which contributes
to the formation of the sales price of farmland, was
developed for the rst time in published literature,
thereby contributing tothe farmland market. Knowl-
edge ofthe classication offarmlands byparcel index
score isessential for landowners aswell asland buyers,
developers and land policymakers. e results ofthe
hedonic price model indicated that the parcel index
has astrong eect on the price offarmland. e im-
portance of using all available information on farm-
lands isproven. Furthermore, hedonic results can also
beuseful inthe policy-making decisions ofagricultural
public agency representatives for the management and
marketing offarmland.
REFERENCES
Arici I., Akkaya Aslan S.T. (2014): Land Consolidation Plan-
ning and Project. Bursa, Turkey, Dora: 237.
Awasthi M.K. (2009): Dynamics and resource use eciency
ofagricultural land sales and rental market inIndia. Land
Use Policy, 26:736–743.
Box G., Cox D.R. (1964): Ananalysis oftransformations. Jour-
nal ofthe Royal Statistical Society SeriesB, 26:211–252.
Brorsen B.W., Doye D., Neal K.B. (2015): Agricultural land
and the small parcel size premium puzzle. Land Econom-
ics, 91:572–585.
Drescher K., Henderson J., McNamara K.T. (2001): Farmland
prices determinants. In:American Agricultural Econom-
ics Association Annual Meeting, Chicago, Illinois, US,
Aug5–8,2001: 1–14.
Dacko M., Wojewodzic T., Pijanowski J., Taszakowski J.,
Dacko A., Janus J. (2021): Increase in the value of agri-
cultural parcels –Modelling and simulation ofthe eects
ofland consolidation project. Agriculture, 11:388.
FRS (2016): Farmer Registration System Records. Farmer
Registration System (FRS), Ministry of Agriculture and
Forestry. Available at https://www.tarimorman.gov.tr/
TRGM/Menu/49/E-Hizmetler (accessed Aug26,2016).
Huang H., Miller G.Y., Sherrick B.J., Gómez M.I. (2006):
Factors inuencing Illinois farmland values. American
Journal ofAgricultural Economics, 88:58–470.
Lloyd T.A., Rayner A.J., Orme C.D. (1991): Present-value
models ofland prices in England and Wales. European
Review ofAgricultural Economics, 18:141–166.
Lin L., Zhang C. (2021): Land parcel identication. In:DiL.,
Üstündağ B.(eds.): Agro-geoinformatics: eory and Prac-
tice. Cham, Switzerland, Springer: 163–174.
MLA (1998): Manisa Land Asset s (MLA). Manisa, Turkey, Pub-
lications ofthe General Directorate ofRural Services:149.
Palmquist R.B., Danielson L.E. (1989): Ahedonic study ofthe
eects oferosion control and drainage onfarmland values.
American Journal ofAgricultural Economics, 71:55–62.
Raup P.M. (2003): Disaggregating farmland markets. In:
MossC.B., Schmitz A. (eds.): Government Policy and
Farmland Markets: eMaintenance ofFarmer Wealth.
Hoboken, US, John Wiley &Sons: 15–25.
Ritter M., Hüttel S., Odening M., Seifert S. (2020): Revisiting
the relationship between land price and parcel size inag-
riculture. Land Use Policy, 97:104771.
Rosen S. (1974): Hedonic prices and implicit markets: Prod-
uct dierentiation inpure competition. Journal ofPolitical
Economy, 82:34–55.
Spinney J., Kanaroglou P., Scott D. (2011): Exploring spa-
tial dynamics with land price indexes. Urban Studies,
48:719–735.
Seifert S., Kahle C., Hüttel S. (2021): Price dispersion
infarmland markets: What isthe role ofasymmetric in-
formation? American Journal ofAgricultural Economics,
103:1545–1568.
Tezcan A., Buyuktas K., Akkaya Aslan S.T. (2020): Amulti-
-criteria model for land valuation inthe land consolidation.
Land Use Policy, 95:104572.
Vasquez O., Nelson J.R., Hamilton J.R. (2002): Regression
analysis todetermine the eects of land characteristics
onfarmland values in South-Central Idaho. American
Society ofFarm Managers and Rural Appraisers, 65:69–77.
Vuong Q.H. (1989): Likelihood ratio test for model selection
and non-nested hypothesis. Econometrica, 57:307–333.
Xu F., Mittelhammer R.C., Barkley P.W. (1993): Measuring
the contributions ofsite characteristics tothe value ofag-
ricultural land. Land Economics, 69:356–369.
Received: March 14, 2022
Accepted: October 31, 2022
Published online: November 15, 2022