- Email: [email protected]
Factors affecting farmland prices in the Czech Republic
Land Use Policy 30 (2013) 130–136
Contents lists available at SciVerse ScienceDirect
Land Use Policy journal homepage: www.elsevier.com/locate/landusepol
Factors affecting farmland prices in the Czech Republic Petr Sklenicka a,∗ , Kristina Molnarova a , Katerina C. Pixova a , Miroslav E. Salek b a b
Department of Land Use and Improvement, Faculty of Environmental Sciences, Czech University of Life Sciences, Prague, Czech Republic Department of Ecology, Faculty of Environmental Sciences, Czech University of Life Sciences, Prague, Czech Republic
a r t i c l e
i n f o
Article history: Received 10 September 2011 Received in revised form 31 January 2012 Accepted 9 March 2012 Keywords: Land market Agricultural land conservation Land value Land management Soil quality
a b s t r a c t The spatial variability of farmland prices is determined by factors reflecting agricultural use, and also by location-specific characteristics, which are crucial to the conversion of farmland to non-farming uses. In co-operation with experienced real-estate brokers, we collected data from 286 transactions carried out in 2008. We identified factors to be analyzed at the parcel scale and tested their effect on the variability of farmland prices in the Czech Republic using general linear modeling. Our results indicate that the most powerful factor in explaining the sale price per square metre was proximity to a settlement, and significantly higher prices were found close to existing built-up areas. The next most powerful factors were: municipality population, travel time to the capital city, accessibility of the parcel, and natural soil fertility. The results have been interpreted to determine the threshold values for significant factors that support future non-agricultural use of farmland and significantly raise current farmland prices. The values supporting non-agricultural use of farmland are proximity to a settlement (up to 100 m), proximity to a larger municipality (above 5000 inhabitants), short travel time to the capital city (up to 1 h) and accessibility to the parcel via the transportation network. © 2012 Elsevier Ltd. All rights reserved.
Introduction The price of farmland is determined by many agronomical, economic, demographic and geographic factors (Drescher et al., 2001; Huang et al., 2006; Skaloˇs et al., 2011). The determinants of farmland price volatility vary according to future land development (Plantinga and Miller, 2001). From the standpoint of agricultural use, the factors most frequently quoted as significant determinants of farmland price are soil quality, water supply, land rents, farm returns, farm size, location in relation to markets, various land lease arrangements, and agricultural subsidies (Palmquist and Danielson, 1989; Lloyd et al., 1991; Sogaard, 1993; Bastian et al., 2002; Awasthi, 2009). The value of non-agricultural characteristics of farmland has been noted in many previous studies that describe the frequently speculative character of business transactions where the buyer intends to develop the land, usually for residential, commercial or recreational purposes. Buyers with a special motivation often pay a premium to obtain agricultural land (Drozd and Johnson, 2004). In comparison with these forms of motivation, non-agricultural use of farmland for habitats or for open spaces is usually a less significant driver of increases in farmland prices (Skaloˇs and
∗ Corresponding author at: Czech University of Life Sciences, Faculty of Environmental Sciences, Department of Land Use and Improvement, Prague 6, Suchdol, 165 21, Czech Republic. Tel.: +420 224 384 352; fax: +420 234 381 848. E-mail address: [email protected] (P. Sklenicka). 0264-8377/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.landusepol.2012.03.005
Engstová, 2010). However, these forms of use also increase the competition for farmland (Bastian et al., 2002). Barnard (2000) estimates that non-agricultural influences account for about onequarter of the average market value of US farm real estate. In such cases, location-specific characteristics, which do not reflect the agricultural characteristics of the land, are capitalized into land prices (Cho and Newman, 2005; Tan et al., 2009; Spinney et al., 2011). The conversion of farmland to non-agricultural use is, according to many authors, supported especially by proximity to a settlement, i.e. the distance to the edge or the centre of the nearest municipality (e.g. Cheshire, 1995; Guiling et al., 2009). Naydenov (2009) confirmed a significant negative relationship between distance to the capital city and land prices in one part of Bulgaria, whereas in another part of the country the effect of the distance of the parcels from the seaside was prevalent. Besides proximity to a metropolitan area, Stewart and Libby (1998) emphasize the role of the quality of the infrastructure and of accessibility, especially proximity to a highway or to a state road. Drescher et al. (2001) and Lisec and Drobne (2009) also describe the influence of the natural amenities of an area on the farmland market, noting that the presence of natural amenities increases recreational activities and retirement activities, which are further capitalized into land prices. Palmquist and Danielson (1989) and Guiling et al. (2009) found a positive influence of the size of the adjacent settlement, or of the local population, on land prices. Forster (2006) reports a higher level of conversion of farmland to residential and commercial use in areas where population growth is occurring.
P. Sklenicka et al. / Land Use Policy 30 (2013) 130–136
Similarly as in other post-socialist countries (Deininger and Jin, 2003; Drobne et al., 2009), the real estate market of the Czech Republic and the methods for its evaluation are still in the development phase. To date, no comprehensive studies focusing on the sales of agricultural land have been published. However, partial results (e.g. Sklenicka and Salek, 2008) show that since the transition to democracy in 1989 and the restitution of agricultural land in the following years, there has been a clear prevalence of the rental market over the sales market. The results further show that factors affecting landowners’ decisions to sell or rent their land include low current prices of agricultural land and lack of credit policies enabling potential farmers to buy land. This situation leads to renting subjects farming large blocs of land, a phenomenon which many studies have shown to be potentially disadvantageous both from the environmental standpoint (Petit and Usher, 1998; Lovett-Doust et al., 2003) and from the standpoint of the efficiency of agricultural production (Deininger and Jin, 2003). Where agricultural land is sold, the transactions usually involve individual parcels or groups of parcels, rather than the whole farm. The two main acts used within the Czech legal system to control the transformation of agricultural land to buildable land are the Act no. 183/2006 on Spatial Planning and Building Rules (“Building Act”) and Act No. 334/1992 on Protection of Agricultural Land Resources. While spatial planning, institutionalized by the Building Act, protects agricultural land by zoning, the Act on Protection of Agricultural Land Resources enables the authorities to charge relatively high one-time fees for the transformation of agricultural land to buildable land within this zoning. However, the current system does not utilize some of the measures proposed by recent research (e.g. Deininger and Jin, 2003; Deininger et al., 2003) to promote efficient land utilization and access to the land by farmers by poor but efficient farmers, such as a realistic level of land taxation and efficient credit policies to support farming subjects and prevent distress sales. Studies of farmland prices can be divided into two broad categories, according to their focus on agricultural or non-agricultural factors. A comprehensive overview of both these categories of studies and of the variables that are used has been presented by Shi et al. (1997). Our study cites some more recently published works. Apart from these two categories, there are models that combine the approaches of the two types of studies. One of the earliest of these models was a study by Scharlach and Schuh (1962). More recent examples include the works of Bernischka and Binkley (1994), and Plantinga and Miller (2001). Some of these studies use time-series data, some use cross-sectional data, and some combine these two types. Both parcel level data and aggregated data from secondary sources are used for model estimation (Shi et al., 1997). The goal of this study is, in co-operation with experienced realestate brokers, to identify the factors which can act as determinants of land prices in the Czech Republic. We then test the significance of these factors, related both to agricultural and non-agricultural use of farmland, for the spatial variability of land prices at the parcel scale, and verify the hypothesized relationships between the most significant predictors and the price of agricultural land (Table 1).
Methods Data collection In this study, the term “farmland” is used as a synonym for the word “agricultural land”, meaning “land including arable land, land under permanent crops and land under permanent meadows and pastures” (OECD, 1997). The data was collected in the course of 2008, and samples were chosen throughout the Czech Republic (Fig. 1) to represent the
131
Table 1 Factors potentially influencing farmland prices, identified by 17 real estate brokers. The table lists the frequency of repeatedly identified factors and the influence of each factor on farmland prices predicted by the brokers. Only variables identified as potentially significant by more than one respondent are listed. Where the brokers agreed in their predictions, a positive influence of the higher value of a factor on land prices is marked (+), and a negative influence is marked (−). Where the opinion was not uniform, the influence was marked (?). Based on further discussion with the participating brokers, the factor Travel Time to a City was subsequently divided into 3 separate factors: Travel Time to the Capital City (over 1 million inhabitants), Travel Time to the Regional Capital (hundreds of thousands of inhabitants) and Travel Time to a District Town Capital (tens of thousands of inhabitants), to describe the influence of the size of the town combined with the travel time to this town. Factor potentially influencing the prices of farmland
Number of identifications
Hypothesized influence
Parcel size Travel time to a city Proximity to a settlement Soil fertility Parcel accessibility Municipality population Distance from a recreationally used waterbody Risk of soil and crop contamination Inundated area Highly eroded soil Scenic vistas Nature conservation Steep slope Real income Land rents Systematic drainage Irrigation Infrastructure Angling opportunities Shape of the plot
17 16 16 14 14 12 4
+ − − + + + −
4 4 4 3 3 3 3 3 2 2 2 2 2
− − − + ? − + + ? ? + + ?
entire range of the country’s natural and socio-geographic heterogeneity. We co-operated with 17 real-estate agencies, whose scope covers all 14 regional administrative units of the Czech Republic. The sample used in this study includes all transactions carried out by these 17 agencies in 2008 in which only one parcel or a group of adjoining parcels was sold. In other transactions, the price reflected the variable characteristics of all the sold parcels, and it would therefore be impossible to determine the influence of individual factors. All transactions included in the sample took place between a willing buyer and a willing seller, there were no distress sales or transactions between co-owners, as all these circumstances could influence the price in manners which would be difficult or impossible to assess. To objectivize the initial choice of potential determining factors, brokers (n = 17) with a minimum of 5 years of experience in real estate were asked to list the 10 most significant factors that influence land prices in the Czech Republic. The number of brokers who identified each factor as important determined the importance attached to the factor. Six predictors were chosen on the basis of this simple questionnaire. As shown in Table 1, these 6 factors reached a very high level of correspondence (at least 12 out of 17 brokers identified them as important; i.e. more than 70% of the brokers), whereas the level of correspondence in other factors was relatively low (they were identified as important by 4 or fewer brokers; i.e. fewer than 25%). Based on further discussion with the participating brokers, the factor Travel Time to a Large Town was divided into 3 separate factors: Travel Time to the Capital City (over 1 million inhabitants), Travel Time to the Regional Capital (hundreds of thousands of inhabitants) and Travel Time to the District Town (tens of thousands of inhabitants), to describe the influence of the size of the town combined with the travel time to this town. The study therefore tested 8 predictors, 3 of which can be described as agricultural and 5 as non-agricultural explanatory
132
P. Sklenicka et al. / Land Use Policy 30 (2013) 130–136
Table 2 Description of the explanatory variables used in the study. Variables
Abbr.
Data type
Data source
Data mean; range (min.–max.)
Agricultural variables Parcel size
AREA
Natural soil fertility
FERT
Total area of purchased parcel or group of adjoining parcels [ha] Weighted mean of administrative prices of farmland [CZK m−2 ]
3.7523 ha; (0.0485–18.4511) 5.69 CZK m−2 ; (1.10–16.58)
Parcel accessibility
ACCESS
Access to parcel via public transportation network Accessible/Inaccessible
Database of office for surveying, mapping and cadastre Soil quality maps (vector, 1:500–1:5000), database of office for surveying, mapping and cadastre Orthophotographs, maps of the office for surveying, mapping and cadastre
Non-agricultural variables Proximity to a settlement
DIST
Municipality population
INHAB
Shortest distance from the edge of the parcel to the edge of a built-up area [m] Population size [No]
Travel time to capital city
CAPIT
Travel time to regional capital
REGIO
Travel time to a district town
DISTR
Travel time by car from the location of the parcel to the Capital City [min] Travel time by car from the location of the parcel to the regional capital [min] Travel Time by car from the location of the parcel to the district town [min]
variables of farmland prices. These factors, their data type, data sources, means and ranges are listed in Table 2. Parcel Size was determined directly from the Database of the Office for Surveying, Mapping and Cadastre. Where more than one adjacent parcel was sold at the same time, the sum of their sizes was entered into the evaluation. Natural Soil Fertility, comprising a complex of soil properties without soil improvement measures, is expressed for individual soil quality units as administrative prices. These administrative prices were calculated for all farmland based on long-term testing of all 2199 soil quality units according to the yields of the main field crops (Becvarova et al., 1988), and are used in the legislative and tax system of the Czech Republic. Municipality Population was determined from the database of the Czech Statistical Office as the number of inhabitants of the nearest village or town, as settlement in the Czech Republic is almost entirely nucleated. The values for the rest of the independent variables were determined in the GIS environment (Arc GIS 9.2), by overlaying the
Orthophotographs, maps of the office for surveying, mapping and cadastre Database of the Czech Statistical Office GIS of Czech Ministry of Transport, Maps of Office for Surveying, Mapping and Cadastre GIS of Czech Ministry of Transport, Maps of Office for Surveying, Mapping and Cadastre GIS of Czech Ministry of Transport, Maps of the Office for Surveying, Mapping and Cadastre
Access – 74%; no access – 26%
167.5 m; (0–1620)
2998.2; (1–67,543) 139.8 min; (35–289) 61.9 min; (5–113)
27.0 min; (3–56)
cadastral map with current orthophotomaps. Parcel Accessibility was evaluated visually. The parcel was evaluated as accessible if at least one public road or public field road provided access to its edge. Settlement Proximity was measured as the shortest distance between the edge of the parcel and the edge of the nearest built-up area. The values of the predictors Travel Time to the Capital City, Travel Time to the Regional Capital, and Travel Time to the District Town were calculated as the travel time by car to the centres of these cities. These calculations were based on vector data by the Czech Ministry of Transport, which contains roads of all categories with speed limits. The dependent variable in our study was Farmland Price [CZK m−2 ]. The sale prices were determined from 286 transactions which took place in 2008 throughout the Czech Republic. Fig. 1 shows that the collected data was evenly distributed across the country (not locally concentrated). Data on the prices and identification of the sold parcels was provided by the 17 real-estate agencies participating in the study. The average sale
Fig. 1. Locations of the 286 farmland sales in the Czech Republic analyzed here.
P. Sklenicka et al. / Land Use Policy 30 (2013) 130–136
DIST INHAB CAPIT ACCESS CAPIT:DIST FERT:CAPIT DIST:ACCESS AREA:DIST FERT
−0.022 0.126 −0.841 0.170 −0.034 0.826 −0.019 0.007 0.145
0.056 0.005 0.245 0.036 0.010 0.315 0.006 0.003 0.067
0.90 2.32 0.57 1.06 0.57 0.33 0.53 0.31 0.23
1 1 1 1 1 1 1 1 1
<0.0001 <0.0001 0.0006 0.0001 0.0006 0.0088 0.0009 0.011 0.03
price was 38.58 CZK m−2 , and prices ranged from 4.00 CZK m−2 to 200 CZK m−2 . All 8 explanatory variables were further analyzed for these parcels. All transactions included in this study were carried out in Czech crowns (CZK). The exchange rate (both in 2008 and currently) is approximately 1 EUR to 25 CZK. Data analysis We applied general linear modeling to reveal the driving factors affecting farmland prices across the Czech Republic. As we do not have sufficiently detailed data to justify the use of hedonic models with exactly specified and reliable variables, we used the most general form to analyze possible predictors of farmland price. We use general linear modeling, rather than the specific hedonic approach, which would require solutions of complicated partial differential equations to fully characterize market conditions/equilibrium. Knowledge of the general effects of selected predictors provides an opportunity for a further detailed economic evaluation of the selected factors using hedonic models, and an opportunity to decompose the price of the items into separate components that determine the price. First, we checked the normality of all continuous variables to be included in the model. While REGIO and DISTR were normally distributed (Kolmogorov–Smirnov test, both d < 0.08, P > 0.1), logarithmic transformation was applied to Farmland Prices, INHAB, FERT and CAPIT to normalize the data (Kolmogorov–Smirnov test, all d < 0.07, P > 0.1). Area and DIST were square-root transformed (Kolmogorov–Smirnov test, both d < 0.08, P > 0.05 after transformation). The effects of eight fixed predictors and their first-order interactions were included in the null model and, afterwards, nonsignificant variables (P > 0.05) were eliminated step-by-step, using the backward selection procedure to achieve a minimum adequate model (Crawley, 2007). F-tests were applied to assess the contributions of particular terms to the model deviances and to the calculation of statistical significances (alpha stated as 0.05). All procedures were performed using R. Results Five out of the eight fixed variables proved to be significant predictors of farmland prices (P < 0.05). Two of them were agricultural variables (ACCESS, FERT) and three were non-agricultural variables (DIST, INHAB and CAPIT). Moreover, the variability in land prices was significantly influenced by four interactions – CAPIT:DIST, FERT:CAPIT, DIST:ACCESS, AREA:DIST. The significant terms explain 65.6% of the total variation. Table 3 lists the variables that contribute significantly to the explanation of the sale prices of farmland. The linear relationship of the remaining three fixed predictors and 24 interactions was not statistically significant. Settlement Proximity (DIST) is the most powerful predictor of the variability of farmland prices. As shown in Fig. 2a, the price of a parcel directly adjoining a built-up area (DIST = 0;
-2
Farmland Price (CZKm )
P
150
100
50
0 0
200
400
600
800
1000
1200
1400
1600
1800
DIST (m)
(b)
200
-2
df
Farmland Price (CZKm )
2
150
100
50
0 0
10
20
30
40
50
60
70
80
INHAB (thousands)
(c)
200
-2
SE
Farmland Price (CZKm )
Estimate
200
150
100
50
0 0
60
120
180
240
300
CAPIT (min)
(d)
200
-2
Predictor
(a)
Farmland Price (CZKm )
Table 3 Results of the model presenting the predictors and their interactions which contributed significantly (P < 0.05) to the variability in farmland prices.
133
150
100
50
0 0
5
10
15
20
-2
FERT (CZKm ) Fig. 2. Relationships between significant fixed effects and farmland prices on real (non-transformed) data visualising the real courses of these relationships: (a) Settlement Proximity (DIST); (b) Municipality Population (INHAB); (c) Travel Time to the Capital (CAPIT); and (d) Natural Soil Fertility (FERT).
mean = 88.72 CZK m−2 ) is more than four times higher than the price of parcels situated more than 100 m from a built-up area. In parcels situated up to 100 m from the edge of a builtup area (DIST = 1–100; mean = 46.20 CZK m−2 ), the price is still more than twice higher than the price for more distant parcels, whereas at distances over 100 m from a built-up area (DIST > 100; mean = 19.04 CZK m−2 ), the price is relatively constant. The population size of the adjacent municipality (INHAB) is also a strong predictor of farmland prices. The results show a strong trend toward increasing farmland prices with increasing population of the adjacent municipality (Fig. 2b). In municipalities with more
134
P. Sklenicka et al. / Land Use Policy 30 (2013) 130–136
there is a mild positive influence of Parcel Size on farmland prices at distances over 100 m from a built-up area, the influence of Parcel Size is strongly negative in parcels closer to a built-up area.
-2
Farmland Price (CZKm )
50
40
Discussion 30
20
10
0 No access
Access
Fig. 3. Effect of Parcel Accessibility (ACCESS) on farmland prices. The mean scores (bars) and standard errors (whiskers) are presented using non-transformed data.
than 5000 inhabitants, the average land price is approximately 2.5 times higher (INHAB > 5000; mean = 71.07 CZK m−2 ) than in municipalities with fewer than 1000 inhabitants (INHAB < 1000; mean = 28.86 CZK m−2 ). Of the variables expressing travel times to different levels of statutory towns (Capital City, Regional Capital, District Town), only Travel time to the Capital City (CAPIT, Fig. 2c) proved statistically significant. In land situated within 60 min from the capital city, the price of farmland was more than twice higher (CAPIT < 60; mean = 74.44 CZK m−2 ) than in land with longer travel times. Where the travel time was more than 60 min, the influence of the capital city on land prices was minimal (CAPIT > 60; mean = 35.44 CZK m−2 ). Another significant predictor was Parcel Accessibility (ACCESS). The results (Fig. 3) show that the price of accessible plots was more than twice as high (mean = 45.18 CZK m−2 ) as the price of inaccessible plots (mean = 18.97 CZK m−2 ). The last single factor significantly influencing land prices was Natural Soil Fertility (FERT). However, the role of this factor is ambivalent, as prices tend to increase both in land with the highest Natural Soil Fertility and in land with the lowest Natural Soil Fertility (Fig. 2d). The strongest of the four significant interactions is between Travel Time to the Capital City and Proximity to a Settlement (CAPIT:DIST). The influence of proximity to the edge of the nearest built-up area is stronger in parcels with a longer travel time to the capital city. In areas with travel time to the capital over 1 h, the price of farmland within 100 m from a built-up area is on an average 218% higher than the price of farmland further from a built-up area. However, in areas within 1 h of travel time from the capital city, the difference is only 42%. The interaction between Travel time to the Capital City and Natural Soil Fertility (CAPIT:FERT) shows a significantly diminished influence of soil fertility on farmland prices in areas within 60 min of travel time from the capital city. In these areas, the average price of the most fertile land (over 8.00 CZK m−2 ) is 47% lower than the price of less fertile land (less than 8.00 CZK m−2 ), whereas in more remote areas the price of the most fertile land is 6% higher. The interaction between Settlement Proximity and Parcel Accessibility (DIST:ACCESS) shows that the negative influence of lack of accessibility on farmland prices is strongest in parcels within 100 m from a built-up area, where the average price of inaccessible parcels is 52% lower than the price of accessible parcels. At distances over 100 m, the difference in price is only 36%. The interaction between Parcel Size and Settlement Proximity (AREA:DIST) shows contrasting trends in parcels within 100 m from built-up areas and parcels that are farther away. While
The most significant factor behind the spatial volatility of farmland prices is Proximity to a Settlement. There is a significant increase in farmland prices up to approximately 100 m from the edge of a built-up area, but distinctly the highest prices are paid for farmland directly adjoining a built-up area. This phenomenon could theoretically be attributed to willingness to pay higher prices where the buyer owns farmland adjoining a village and bordering on his/her recent parcel (there is usually higher land fragmentation near to villages). This would lead to consolidation of parcels owned by one owner, and to mitigation of ownership fragmentation, a major factor reducing farm profitability (Di Falco et al., 2010). However, an examination of the ownership of parcels bordering with sold land has proved that this explanation is incorrect, as only 2% of the land was bought by owners of neighbouring parcels. We therefore conclude that the main reason behind this sharp increase in farmland prices is speculative interest motivated by conversion of farmland and subsequent sale of the land for non-farm uses. This is in accordance with findings of Drozd and Johnson (2004), Tan et al. (2009) and Spinney et al. (2011). In the Czech Republic, buildable land is usually delimited by the Master plan, and development permits are nearly always granted only in direct continuity with the current built-up area. Our interpretation of this factor is supported by previous studies, e.g. by Livanis et al. (2006) and Guiling et al. (2009). Some authors measure the distance of the plot from the centre of the village (Cheshire, 1995). In our opinion, the distance of the edge of the parcel from the edge of the current built-up area is a more accurate variable. This approach takes into account relevant conditions, such as the connection to paved roads and to the municipal infrastructure, as well as the probability of conversion of the land into non-agricultural use in the master plan. Moreover, problems with eccentrically placed village centres are eliminated (Delbecq and Florax, 2010). The second most important factor behind the spatial variability of farmland prices proved to be Municipality Population. Our results have confirmed a positive influence of population size on an increase in farmland prices. This result can also be interpreted with regard to speculation on the conversion of farmland mainly to buildable land, as the higher prices of farmland adjoining larger municipalities imitate the higher prices of building land and family homes (Livanis et al., 2006). However, in this case, the variability in farmland prices can also be explained by higher competition between potential local buyers in larger municipalities and by the pressure of this competition toward an increase in prices (Palmquist and Danielson, 1989; Guiling et al., 2009). Speculative purchases of farmland for future development can also explain the high volatility of prices connected with another factor: Travel Time to the Capital City. In areas within 1 h of travel time from the capital city, farmland prices are significantly higher, and within this interval prices increase further with increasing proximity to the capital. This phenomenon cannot be explained by a higher concentration of farmers in and around the capital city. Travel time around 1 h is still acceptable for commuting to the capital city and its outskirts. This is an area traditionally known to have lower unemployment rates, the highest average salary, the highest number of education opportunities, and other benefits (Henley, 1998; McDonald and McMillen, 1998). Therefore, as with the previously discussed factors, the higher prices in areas within a shorter travel time to the capital city can be explained by speculation on high demand for building plots or family homes. A similar influence of
P. Sklenicka et al. / Land Use Policy 30 (2013) 130–136
regional capitals (with a population size of hundreds of thousands) or district towns (with a population size of tens of thousands) has not been proved, contrary to our expectations. In contrast to the conclusions of Naydenov (2009) and along with e.g. Henley (1998), we conclude that length units are not suitable for explaining the lucrativeness of some areas in relation to the capital city, or in relation to large cities in general. The travel time needed to reach the city by car or by some other means of transport is a more suitable descriptive unit. Places which are easily accessible by a highway or, for example, by a high-speed train, can be as much as twice as far away as areas with low accessibility, and still require the same travel time (Haurin and Rosenthal, 2004). The accessibility of a parcel is generally the main condition for fulfilling ownership rights. In the Czech Republic, many parcels have lost their accessibility due to land-use consolidation and closure of field roads during the socialist period (Sklenicka, 2006). The owners of parcels that are currently not accessible by the transportation network are discriminated against in the agricultural use of their land, and in renting their land for farming purposes. Both of these activities are dependent on cooperation with the owners or tenants of neighbouring plots. Our results show that there is a similar discrimination in farmland prices, which reflect the interest in converting farmland to other uses (Zhong et al., 2011). There is apparent unwillingness to pay higher prices for inaccessible land and to risk failure in future negotiation of access with the neighbouring owners. In our opinion, this handicap affects both sales of farmland for agricultural purposes and speculative purchases of land for future non-farm use. This result of our study supports the significance of land consolidation programmes in consolidating land ownership and providing accessibility to each plot by a system of field roads (Sklenicka et al., 2009). The last single factor significantly influencing the variability of farmland prices is Natural Soil Fertility. A considerable increase in farmland prices occurs only in the most fertile land, which can be explained by higher demand from farmers. Higher soil fertility and its influence on land prices mainly reflect interest in using the land for agriculture, as the prices of building plots, recreational plots and other plots are usually not determined by soil fertility (Sklenicka et al., 2002). The only exception may be the least fertile farmland, where speculative purchases for non-farming purposes are supported by the ease and the higher probability of obtaining permission to convert the land to residential use, as well as by the significantly lower financial cost of obtaining this permission than in the case of more fertile land (Forster, 2006; Spinney et al., 2011). This phenomenon is manifested in our results, where higher farmland prices were observed both in the most fertile land and in the least fertile land. This result is partly in contradiction with our original hypothesis, in which we expected a distinctly positive relationship between Natural Soil Fertility and farmland prices. Proximity to a Settlement was also the predictor that occurred most frequently in statistically significant interactions with other predictors. The influence of proximity to the edge of a built-up area was much more pronounced in parcels with a longer travel time to the capital city. This result can be explained by the expectations of faster and more extensive development of residential areas in the proximity of the capital city, resulting in higher demand for land, including plots further from the edges of the currently builtup areas of the villages in this zone. Settlement Proximity in interaction with Parcel Accessibility confirms the importance of the accessibility of a parcel, especially at shorter distances from current built-up areas. This result indicates the high importance of providing access to more expensive parcels, where conversion to residential use is expected. Potential buyers apparently regard poor accessibility of more expensive parcels as a much higher risk than in the case of cheaper parcels. At the same time, inaccessible farmland can easily be consolidated
135
with other parcels owned by the same owner in the process of land consolidation (Sklenicka, 2006), whereas in residential parcels this consolidation is much more complicated and much less likely to happen. The third significant interaction of Proximity to a Settlement is with Parcel Size. Parcel Size was a significant factor only in this interaction, despite the fact that it was the only factor that was identified as important by all the real-estate brokers. Some previous studies have reached different conclusions (e.g. Maddison, 2000). We had expected this factor to be very powerful, especially due to the extreme fragmentation of farmland in the Czech Republic (Sklenicka and Salek, 2008). The results indicate that a positive influence of Parcel Size on farmland prices applies only at longer distances from currently built-up areas. In parcels closer to builtup areas, this influence is negative. This result can be explained by the higher cost of large parcels in more expensive zones closer to a built-up area, where speculation on converting farmland to residential use is frequent. Another possible explanation is the demand for larger parcels for agricultural use (further from current builtup areas), as farming larger blocks of land is more efficient, due to mechanization and the agronomic methods that are used (Gonzales et al., 2004). In the interaction between Travel Time to the Capital City and Natural Soil Fertility, the role of Natural Soil Fertility is distinctly diminished in areas with a shorter travel time. This factor only begins to influence land prices in areas with a longer travel time. Similarly to Henley (1998) and to the variables discussed above, the key reason behind this trend is apparently the buyer’s intention to use the land. In areas with shorter travel time to the capital city, speculation on the conversion of farmland to residential use is more frequent (McDonald and McMillen, 1998). Buyers therefore demand less fertile land with lower taxation levels and a much lower fee for conversion to residential use. In areas with a longer travel time to the capital city, there is a long-term preference for agricultural use, with farmland prices positively reflecting higher soil fertility.
Conclusions Our study has investigated the key factors influencing spatial variability in farmland prices in the Czech Republic, and has analyzed the effects of these factors. In cooperation with 17 experienced real estate brokers, we chose 8 factors that the brokers regarded as the main determinants of farmland prices. Subsequent linear modeling confirmed the significance of 5 single factors and 4 interactions, and showed that, at some values of these significant predictors, farmland prices increased by as much as several hundred percent. The interpretation of the results indicates that specific values of these factors increase the demand for farmland for future non-agricultural use. This non-agricultural use is supported mainly by proximity to currently built-up areas (up to 100 m), proximity to a large municipality (over 5000 inhabitants), short travel time to the capital city (up to 1 h), and access to the parcel via the transportation network. However, our study has for the first time proved that travel time to the regional capital or to the district town is not a significant determinant of farmland prices. Our results do not show relevant differences in the role of factors affecting land markets in comparison with other European or U.S. open-market states. The results further indicate the significant influence of land consolidation projects and of master planning on farmland prices, and the important role of these two forms of landscape planning in protecting farmland from urban sprawl and in supporting the land market (Lisec et al., 2008). Knowledge of factors influencing sale prices of farmland is important for landowners, and also for land purchasers, developers and land policy makers.
136
P. Sklenicka et al. / Land Use Policy 30 (2013) 130–136
Acknowledgments This study was supported by grant No. QH82162 from the Ministry of Agriculture of the Czech Republic “Possibilities of defragmentation of agricultural land”. The authors owe special thanks to Robin Healey for editing the manuscript, and to 17 real estate brokers for their help with obtaining the data. References Awasthi, M.K., 2009. Dynamics and resource use efficiency of agricultural land sales and rental market in India. Land Use Policy 26, 736–743. Barnard, C.H., 2000. Urbanization affects a large share of farmland. Rural Cond. Trends 10, 57–63. Bastian, C.T., McLeod, D.M., Germino, M.J., Reiners, W.A., Blasko, B.J., 2002. Environmental amenities and agricultural land values: a hedonic model using geographic information systems data. Ecol. Econ. 40, 337–349. Becvarova, V., Vasek, P., Vanicek, F., 1988. Bonitace Zemedelskych Pud a Smery Jejich Vyuziti. 4. dil. FMZVZ, Prague. Bernischka, M., Binkley, J.K., 1994. Land price volatility in a geographically dispersed market. Am. J. Agr. Econ. 76, 185–195. Cheshire, P., 1995. On the price of land and the value of amenities. Economica 62, 247–267. Cho, S.H., Newman, D.H., 2005. Spatial analysis of rural land development. Forest Policy Econ. 7, 732–744. Crawley, M.J., 2007. The R Book. Wiley & Sons, Chichester. Delbecq, B.A., Florax, R.J.G.M., 2010. Farmland allocation along the rural–urban gradient: the impact of urbanization and urban sprawl. In: Proceedings of Annual Meeting of Agricultural and Applied Economics Associations, Denver, CO, pp. 1–17. Deininger, K., Jin, S., 2003. Land Sales and Rental Markets in Transition: Evidence from Rural Vietnam. World Bank Policy Research Working Paper 3013, Washington DC. Deininger, K., Gonzalez, M.A., Castagnini, R., 2003. Comparing Land Reform and Land Markets in Colombia: Impacts on Equity and Efficiency. World Bank Policy Research Working Paper, Washington DC. Di Falco, S., Penov, I., Aleksiev, A., van Rensburg, T., 2010. Agrobiodiversity, farm profits and land fragmentation: evidence from Bulgaria. Land Use Policy 27, 763–771. Drescher, K., Henderson, J., McNamara, K., 2001. Farmland prices determinants. In: Proceedings of American Agricultural Economics Association Annual Meeting, Chicago, IL. Drobne, S., Grilj, T., Lisec, A., 2009. Real estate market activity in Slovenia in 2000–2006. Geodetski vestnik 53, 561–579. Drozd, D.J., Johnson, B.B., 2004. Dynamics of a rural land market experiencing farmland conversion to acreages: the case of Saunders County, Nebraska. Land Econ. 80, 294–311. Forster, D.L., 2006. An overview of U.S. farm real estate markets. In: Proceedings of CICA International Confederation of Agricultural Credit, Quebec, Canada, pp. 1–28. Gonzales, X.P., Alvarez, C.J., Crecente, R., 2004. Evaluation of land distributions with joint regard to plot size and shape. Agr. Syst. 82, 31–43. Guiling, P., Brorsen, B.W., Doye, D., 2009. Effect of urban proximity on agricultural land values. Land Econ. 85, 252–264. Haurin, D.R., Rosenthal, S.S., 2004. The Sustainability of Homeownership: Factors Affecting the Duration of Homeownership and Rental Spells. Report for U.S. Department of Housing and Urban Development, Washington, DC. Henley, A., 1998. Residential mobility housing equity, and the labour market. Econ. J. 108, 414–427.
Huang, H., Miller, Y., Sherrick, B.J., Gómez, M.I., 2006. Factors influencing Illinois farmland values. Am. Agric. Econ. Assoc. 88, 458–470. ˇ R., 2008. Modelling the rural land transaction Lisec, A., Ferlan, M., Lobnik, F., Sumrada, procedure. Land Use Policy 25, 286–297. Lisec, A., Drobne, S., 2009. The influence of protected natural and cultural heritage on land management/market: the case of Slovenian natural protected areas. Spatium 20, 41–48. Livanis, G., Moss, C., Breneman, V.E., Nehring, R.F., 2006. Urban sprawl and farmland prices. Am. J. Agr. Econ. 88, 915–929. Lloyd, T.A., Rayner, A.J., Orme, C.D., 1991. Present value model of land prices in England and Wales. Eur. Rev. Agric. Econ. 18, 141–166. Lovett-Doust, J., Biernacki, M., Page, R., Chan, M., Natgunarajah, R., Timis, G., 2003. Effects of land ownership and other landscape-level factors on rare-species richness. Landsc. Ecol. 18, 621–633. Maddison, D., 2000. A hedonic analysis of agricultural land prices in England and Wales. Eur. Rev. Agric. Econ. 27, 519–532. McDonald, J.F., McMillen, D.P., 1998. Land values, land use, and the first Chicago zoning ordinance. J. Real Estate Financ. 16, 135–150. Naydenov, I., 2009. Factors Influencing Farmland Prices in Bulgaria, Master Thesis, Maastricht University. OECD, 1997. Glossary of Environment Statistics, Studies in Methods, Series F, No. 67, United Nations, New York. Palmquist, R.B., Danielson, L.E., 1989. A hedonic study of the effects of erosion control and drainage on farmland values. Am. J. Agr. Econ. 71, 55–62. Petit, S., Usher, M.B., 1998. Biodiversity in agricultural landscapes: the ground beetle communities of woody uncultivated habitats. Biodivers. Conserv. 7, 1549–1561. Plantinga, A.J., Miller, D.J., 2001. Agricultural land values and the value of rights to future land development. Land Econ. 77, 56–67. Scharlach, W.C., Schuh, G.E., 1962. The land market as a link between rural and urban sectors of the economy. J. Farm Econ. 44, 1406–1411. Shi, Y.J., Phips, T.T., Colyer, D., 1997. Agricultural land value under urbanizing influence. Land Econ. 73, 90–100. Skaloˇs, J., Engstová, B., 2010. Methodology for mapping non-forest wood elements using historic cadastral maps and aerial photographs as a basis for management. Journal of Environment Management 91, 831–843. ˇ ´ Z., Trpáková, I., Santr ˚ cková, M., Uhlíˇrová, L., Kukla, Skaloˇs, J., Weber, M., Lipsky, uˇ P., 2011. Using old military survey maps and orthophotograph maps to analyse long-term land cover changes – case study (Czech Republic). Appl. Geogr. 32, 426–438. Sklenicka, P., Lhota, T., Cecetka, J., 2002. Soil porosity along a gradient from forest to field. Die Bodenk 53, 181–187. Sklenicka, P., 2006. Applying evaluation criteria for the land consolidation effect to three contrasting study areas in the Czech Republic. Land Use Policy 23, 502–510. Sklenicka, P., Salek, M., 2008. Ownership and soil quality as sources of agricultural land fragmentation in highly fragmented ownership patterns. Landscape Ecol. 23, 299–311. Sklenicka, P., Hladik, J., Strelecek, F., Kottova, B., Lososova, J., Cihal, L., Salek, M., 2009. Historical, environmental and socio-economic driving forces on land ownership fragmentation, the land consolidation effect and project costs. Agric. Econ. – Czech. 55, 571–582. Sogaard, V., 1993. The land market cycle. Eur. Rev. Agric. Econ. 20, 65–76. Spinney, J., Kanaroglou, P., Scott, D., 2011. Exploring spatial dynamics with land price indexes. Urban Stud. 48, 719–735. Stewart, P.A., Libby, L.W., 1998. Determinants of farmland value: the case of DeKalb County, Illinois. Appl. Econ. Persp. Policy 20, 80–95. Tan, R., Beckmann, V., van den Berg, L., Qu, F., 2009. Governing farmland conversion: comparing China with the Netherlands and Germany. Land Use Policy 26, 961–974. Zhong, T.Y., Huang, X.J., Zhang, X.Y., Wang, K., 2011. Temporal and spatial variability of agricultural land loss in relation to policy and accessibility in a low hilly region of southeast China. Land Use Policy 28, 762–769.
Contents lists available at SciVerse ScienceDirect
Land Use Policy journal homepage: www.elsevier.com/locate/landusepol
Factors affecting farmland prices in the Czech Republic Petr Sklenicka a,∗ , Kristina Molnarova a , Katerina C. Pixova a , Miroslav E. Salek b a b
Department of Land Use and Improvement, Faculty of Environmental Sciences, Czech University of Life Sciences, Prague, Czech Republic Department of Ecology, Faculty of Environmental Sciences, Czech University of Life Sciences, Prague, Czech Republic
a r t i c l e
i n f o
Article history: Received 10 September 2011 Received in revised form 31 January 2012 Accepted 9 March 2012 Keywords: Land market Agricultural land conservation Land value Land management Soil quality
a b s t r a c t The spatial variability of farmland prices is determined by factors reflecting agricultural use, and also by location-specific characteristics, which are crucial to the conversion of farmland to non-farming uses. In co-operation with experienced real-estate brokers, we collected data from 286 transactions carried out in 2008. We identified factors to be analyzed at the parcel scale and tested their effect on the variability of farmland prices in the Czech Republic using general linear modeling. Our results indicate that the most powerful factor in explaining the sale price per square metre was proximity to a settlement, and significantly higher prices were found close to existing built-up areas. The next most powerful factors were: municipality population, travel time to the capital city, accessibility of the parcel, and natural soil fertility. The results have been interpreted to determine the threshold values for significant factors that support future non-agricultural use of farmland and significantly raise current farmland prices. The values supporting non-agricultural use of farmland are proximity to a settlement (up to 100 m), proximity to a larger municipality (above 5000 inhabitants), short travel time to the capital city (up to 1 h) and accessibility to the parcel via the transportation network. © 2012 Elsevier Ltd. All rights reserved.
Introduction The price of farmland is determined by many agronomical, economic, demographic and geographic factors (Drescher et al., 2001; Huang et al., 2006; Skaloˇs et al., 2011). The determinants of farmland price volatility vary according to future land development (Plantinga and Miller, 2001). From the standpoint of agricultural use, the factors most frequently quoted as significant determinants of farmland price are soil quality, water supply, land rents, farm returns, farm size, location in relation to markets, various land lease arrangements, and agricultural subsidies (Palmquist and Danielson, 1989; Lloyd et al., 1991; Sogaard, 1993; Bastian et al., 2002; Awasthi, 2009). The value of non-agricultural characteristics of farmland has been noted in many previous studies that describe the frequently speculative character of business transactions where the buyer intends to develop the land, usually for residential, commercial or recreational purposes. Buyers with a special motivation often pay a premium to obtain agricultural land (Drozd and Johnson, 2004). In comparison with these forms of motivation, non-agricultural use of farmland for habitats or for open spaces is usually a less significant driver of increases in farmland prices (Skaloˇs and
∗ Corresponding author at: Czech University of Life Sciences, Faculty of Environmental Sciences, Department of Land Use and Improvement, Prague 6, Suchdol, 165 21, Czech Republic. Tel.: +420 224 384 352; fax: +420 234 381 848. E-mail address: [email protected] (P. Sklenicka). 0264-8377/$ – see front matter © 2012 Elsevier Ltd. All rights reserved. doi:10.1016/j.landusepol.2012.03.005
Engstová, 2010). However, these forms of use also increase the competition for farmland (Bastian et al., 2002). Barnard (2000) estimates that non-agricultural influences account for about onequarter of the average market value of US farm real estate. In such cases, location-specific characteristics, which do not reflect the agricultural characteristics of the land, are capitalized into land prices (Cho and Newman, 2005; Tan et al., 2009; Spinney et al., 2011). The conversion of farmland to non-agricultural use is, according to many authors, supported especially by proximity to a settlement, i.e. the distance to the edge or the centre of the nearest municipality (e.g. Cheshire, 1995; Guiling et al., 2009). Naydenov (2009) confirmed a significant negative relationship between distance to the capital city and land prices in one part of Bulgaria, whereas in another part of the country the effect of the distance of the parcels from the seaside was prevalent. Besides proximity to a metropolitan area, Stewart and Libby (1998) emphasize the role of the quality of the infrastructure and of accessibility, especially proximity to a highway or to a state road. Drescher et al. (2001) and Lisec and Drobne (2009) also describe the influence of the natural amenities of an area on the farmland market, noting that the presence of natural amenities increases recreational activities and retirement activities, which are further capitalized into land prices. Palmquist and Danielson (1989) and Guiling et al. (2009) found a positive influence of the size of the adjacent settlement, or of the local population, on land prices. Forster (2006) reports a higher level of conversion of farmland to residential and commercial use in areas where population growth is occurring.
P. Sklenicka et al. / Land Use Policy 30 (2013) 130–136
Similarly as in other post-socialist countries (Deininger and Jin, 2003; Drobne et al., 2009), the real estate market of the Czech Republic and the methods for its evaluation are still in the development phase. To date, no comprehensive studies focusing on the sales of agricultural land have been published. However, partial results (e.g. Sklenicka and Salek, 2008) show that since the transition to democracy in 1989 and the restitution of agricultural land in the following years, there has been a clear prevalence of the rental market over the sales market. The results further show that factors affecting landowners’ decisions to sell or rent their land include low current prices of agricultural land and lack of credit policies enabling potential farmers to buy land. This situation leads to renting subjects farming large blocs of land, a phenomenon which many studies have shown to be potentially disadvantageous both from the environmental standpoint (Petit and Usher, 1998; Lovett-Doust et al., 2003) and from the standpoint of the efficiency of agricultural production (Deininger and Jin, 2003). Where agricultural land is sold, the transactions usually involve individual parcels or groups of parcels, rather than the whole farm. The two main acts used within the Czech legal system to control the transformation of agricultural land to buildable land are the Act no. 183/2006 on Spatial Planning and Building Rules (“Building Act”) and Act No. 334/1992 on Protection of Agricultural Land Resources. While spatial planning, institutionalized by the Building Act, protects agricultural land by zoning, the Act on Protection of Agricultural Land Resources enables the authorities to charge relatively high one-time fees for the transformation of agricultural land to buildable land within this zoning. However, the current system does not utilize some of the measures proposed by recent research (e.g. Deininger and Jin, 2003; Deininger et al., 2003) to promote efficient land utilization and access to the land by farmers by poor but efficient farmers, such as a realistic level of land taxation and efficient credit policies to support farming subjects and prevent distress sales. Studies of farmland prices can be divided into two broad categories, according to their focus on agricultural or non-agricultural factors. A comprehensive overview of both these categories of studies and of the variables that are used has been presented by Shi et al. (1997). Our study cites some more recently published works. Apart from these two categories, there are models that combine the approaches of the two types of studies. One of the earliest of these models was a study by Scharlach and Schuh (1962). More recent examples include the works of Bernischka and Binkley (1994), and Plantinga and Miller (2001). Some of these studies use time-series data, some use cross-sectional data, and some combine these two types. Both parcel level data and aggregated data from secondary sources are used for model estimation (Shi et al., 1997). The goal of this study is, in co-operation with experienced realestate brokers, to identify the factors which can act as determinants of land prices in the Czech Republic. We then test the significance of these factors, related both to agricultural and non-agricultural use of farmland, for the spatial variability of land prices at the parcel scale, and verify the hypothesized relationships between the most significant predictors and the price of agricultural land (Table 1).
Methods Data collection In this study, the term “farmland” is used as a synonym for the word “agricultural land”, meaning “land including arable land, land under permanent crops and land under permanent meadows and pastures” (OECD, 1997). The data was collected in the course of 2008, and samples were chosen throughout the Czech Republic (Fig. 1) to represent the
131
Table 1 Factors potentially influencing farmland prices, identified by 17 real estate brokers. The table lists the frequency of repeatedly identified factors and the influence of each factor on farmland prices predicted by the brokers. Only variables identified as potentially significant by more than one respondent are listed. Where the brokers agreed in their predictions, a positive influence of the higher value of a factor on land prices is marked (+), and a negative influence is marked (−). Where the opinion was not uniform, the influence was marked (?). Based on further discussion with the participating brokers, the factor Travel Time to a City was subsequently divided into 3 separate factors: Travel Time to the Capital City (over 1 million inhabitants), Travel Time to the Regional Capital (hundreds of thousands of inhabitants) and Travel Time to a District Town Capital (tens of thousands of inhabitants), to describe the influence of the size of the town combined with the travel time to this town. Factor potentially influencing the prices of farmland
Number of identifications
Hypothesized influence
Parcel size Travel time to a city Proximity to a settlement Soil fertility Parcel accessibility Municipality population Distance from a recreationally used waterbody Risk of soil and crop contamination Inundated area Highly eroded soil Scenic vistas Nature conservation Steep slope Real income Land rents Systematic drainage Irrigation Infrastructure Angling opportunities Shape of the plot
17 16 16 14 14 12 4
+ − − + + + −
4 4 4 3 3 3 3 3 2 2 2 2 2
− − − + ? − + + ? ? + + ?
entire range of the country’s natural and socio-geographic heterogeneity. We co-operated with 17 real-estate agencies, whose scope covers all 14 regional administrative units of the Czech Republic. The sample used in this study includes all transactions carried out by these 17 agencies in 2008 in which only one parcel or a group of adjoining parcels was sold. In other transactions, the price reflected the variable characteristics of all the sold parcels, and it would therefore be impossible to determine the influence of individual factors. All transactions included in the sample took place between a willing buyer and a willing seller, there were no distress sales or transactions between co-owners, as all these circumstances could influence the price in manners which would be difficult or impossible to assess. To objectivize the initial choice of potential determining factors, brokers (n = 17) with a minimum of 5 years of experience in real estate were asked to list the 10 most significant factors that influence land prices in the Czech Republic. The number of brokers who identified each factor as important determined the importance attached to the factor. Six predictors were chosen on the basis of this simple questionnaire. As shown in Table 1, these 6 factors reached a very high level of correspondence (at least 12 out of 17 brokers identified them as important; i.e. more than 70% of the brokers), whereas the level of correspondence in other factors was relatively low (they were identified as important by 4 or fewer brokers; i.e. fewer than 25%). Based on further discussion with the participating brokers, the factor Travel Time to a Large Town was divided into 3 separate factors: Travel Time to the Capital City (over 1 million inhabitants), Travel Time to the Regional Capital (hundreds of thousands of inhabitants) and Travel Time to the District Town (tens of thousands of inhabitants), to describe the influence of the size of the town combined with the travel time to this town. The study therefore tested 8 predictors, 3 of which can be described as agricultural and 5 as non-agricultural explanatory
132
P. Sklenicka et al. / Land Use Policy 30 (2013) 130–136
Table 2 Description of the explanatory variables used in the study. Variables
Abbr.
Data type
Data source
Data mean; range (min.–max.)
Agricultural variables Parcel size
AREA
Natural soil fertility
FERT
Total area of purchased parcel or group of adjoining parcels [ha] Weighted mean of administrative prices of farmland [CZK m−2 ]
3.7523 ha; (0.0485–18.4511) 5.69 CZK m−2 ; (1.10–16.58)
Parcel accessibility
ACCESS
Access to parcel via public transportation network Accessible/Inaccessible
Database of office for surveying, mapping and cadastre Soil quality maps (vector, 1:500–1:5000), database of office for surveying, mapping and cadastre Orthophotographs, maps of the office for surveying, mapping and cadastre
Non-agricultural variables Proximity to a settlement
DIST
Municipality population
INHAB
Shortest distance from the edge of the parcel to the edge of a built-up area [m] Population size [No]
Travel time to capital city
CAPIT
Travel time to regional capital
REGIO
Travel time to a district town
DISTR
Travel time by car from the location of the parcel to the Capital City [min] Travel time by car from the location of the parcel to the regional capital [min] Travel Time by car from the location of the parcel to the district town [min]
variables of farmland prices. These factors, their data type, data sources, means and ranges are listed in Table 2. Parcel Size was determined directly from the Database of the Office for Surveying, Mapping and Cadastre. Where more than one adjacent parcel was sold at the same time, the sum of their sizes was entered into the evaluation. Natural Soil Fertility, comprising a complex of soil properties without soil improvement measures, is expressed for individual soil quality units as administrative prices. These administrative prices were calculated for all farmland based on long-term testing of all 2199 soil quality units according to the yields of the main field crops (Becvarova et al., 1988), and are used in the legislative and tax system of the Czech Republic. Municipality Population was determined from the database of the Czech Statistical Office as the number of inhabitants of the nearest village or town, as settlement in the Czech Republic is almost entirely nucleated. The values for the rest of the independent variables were determined in the GIS environment (Arc GIS 9.2), by overlaying the
Orthophotographs, maps of the office for surveying, mapping and cadastre Database of the Czech Statistical Office GIS of Czech Ministry of Transport, Maps of Office for Surveying, Mapping and Cadastre GIS of Czech Ministry of Transport, Maps of Office for Surveying, Mapping and Cadastre GIS of Czech Ministry of Transport, Maps of the Office for Surveying, Mapping and Cadastre
Access – 74%; no access – 26%
167.5 m; (0–1620)
2998.2; (1–67,543) 139.8 min; (35–289) 61.9 min; (5–113)
27.0 min; (3–56)
cadastral map with current orthophotomaps. Parcel Accessibility was evaluated visually. The parcel was evaluated as accessible if at least one public road or public field road provided access to its edge. Settlement Proximity was measured as the shortest distance between the edge of the parcel and the edge of the nearest built-up area. The values of the predictors Travel Time to the Capital City, Travel Time to the Regional Capital, and Travel Time to the District Town were calculated as the travel time by car to the centres of these cities. These calculations were based on vector data by the Czech Ministry of Transport, which contains roads of all categories with speed limits. The dependent variable in our study was Farmland Price [CZK m−2 ]. The sale prices were determined from 286 transactions which took place in 2008 throughout the Czech Republic. Fig. 1 shows that the collected data was evenly distributed across the country (not locally concentrated). Data on the prices and identification of the sold parcels was provided by the 17 real-estate agencies participating in the study. The average sale
Fig. 1. Locations of the 286 farmland sales in the Czech Republic analyzed here.
P. Sklenicka et al. / Land Use Policy 30 (2013) 130–136
DIST INHAB CAPIT ACCESS CAPIT:DIST FERT:CAPIT DIST:ACCESS AREA:DIST FERT
−0.022 0.126 −0.841 0.170 −0.034 0.826 −0.019 0.007 0.145
0.056 0.005 0.245 0.036 0.010 0.315 0.006 0.003 0.067
0.90 2.32 0.57 1.06 0.57 0.33 0.53 0.31 0.23
1 1 1 1 1 1 1 1 1
<0.0001 <0.0001 0.0006 0.0001 0.0006 0.0088 0.0009 0.011 0.03
price was 38.58 CZK m−2 , and prices ranged from 4.00 CZK m−2 to 200 CZK m−2 . All 8 explanatory variables were further analyzed for these parcels. All transactions included in this study were carried out in Czech crowns (CZK). The exchange rate (both in 2008 and currently) is approximately 1 EUR to 25 CZK. Data analysis We applied general linear modeling to reveal the driving factors affecting farmland prices across the Czech Republic. As we do not have sufficiently detailed data to justify the use of hedonic models with exactly specified and reliable variables, we used the most general form to analyze possible predictors of farmland price. We use general linear modeling, rather than the specific hedonic approach, which would require solutions of complicated partial differential equations to fully characterize market conditions/equilibrium. Knowledge of the general effects of selected predictors provides an opportunity for a further detailed economic evaluation of the selected factors using hedonic models, and an opportunity to decompose the price of the items into separate components that determine the price. First, we checked the normality of all continuous variables to be included in the model. While REGIO and DISTR were normally distributed (Kolmogorov–Smirnov test, both d < 0.08, P > 0.1), logarithmic transformation was applied to Farmland Prices, INHAB, FERT and CAPIT to normalize the data (Kolmogorov–Smirnov test, all d < 0.07, P > 0.1). Area and DIST were square-root transformed (Kolmogorov–Smirnov test, both d < 0.08, P > 0.05 after transformation). The effects of eight fixed predictors and their first-order interactions were included in the null model and, afterwards, nonsignificant variables (P > 0.05) were eliminated step-by-step, using the backward selection procedure to achieve a minimum adequate model (Crawley, 2007). F-tests were applied to assess the contributions of particular terms to the model deviances and to the calculation of statistical significances (alpha stated as 0.05). All procedures were performed using R. Results Five out of the eight fixed variables proved to be significant predictors of farmland prices (P < 0.05). Two of them were agricultural variables (ACCESS, FERT) and three were non-agricultural variables (DIST, INHAB and CAPIT). Moreover, the variability in land prices was significantly influenced by four interactions – CAPIT:DIST, FERT:CAPIT, DIST:ACCESS, AREA:DIST. The significant terms explain 65.6% of the total variation. Table 3 lists the variables that contribute significantly to the explanation of the sale prices of farmland. The linear relationship of the remaining three fixed predictors and 24 interactions was not statistically significant. Settlement Proximity (DIST) is the most powerful predictor of the variability of farmland prices. As shown in Fig. 2a, the price of a parcel directly adjoining a built-up area (DIST = 0;
-2
Farmland Price (CZKm )
P
150
100
50
0 0
200
400
600
800
1000
1200
1400
1600
1800
DIST (m)
(b)
200
-2
df
Farmland Price (CZKm )
2
150
100
50
0 0
10
20
30
40
50
60
70
80
INHAB (thousands)
(c)
200
-2
SE
Farmland Price (CZKm )
Estimate
200
150
100
50
0 0
60
120
180
240
300
CAPIT (min)
(d)
200
-2
Predictor
(a)
Farmland Price (CZKm )
Table 3 Results of the model presenting the predictors and their interactions which contributed significantly (P < 0.05) to the variability in farmland prices.
133
150
100
50
0 0
5
10
15
20
-2
FERT (CZKm ) Fig. 2. Relationships between significant fixed effects and farmland prices on real (non-transformed) data visualising the real courses of these relationships: (a) Settlement Proximity (DIST); (b) Municipality Population (INHAB); (c) Travel Time to the Capital (CAPIT); and (d) Natural Soil Fertility (FERT).
mean = 88.72 CZK m−2 ) is more than four times higher than the price of parcels situated more than 100 m from a built-up area. In parcels situated up to 100 m from the edge of a builtup area (DIST = 1–100; mean = 46.20 CZK m−2 ), the price is still more than twice higher than the price for more distant parcels, whereas at distances over 100 m from a built-up area (DIST > 100; mean = 19.04 CZK m−2 ), the price is relatively constant. The population size of the adjacent municipality (INHAB) is also a strong predictor of farmland prices. The results show a strong trend toward increasing farmland prices with increasing population of the adjacent municipality (Fig. 2b). In municipalities with more
134
P. Sklenicka et al. / Land Use Policy 30 (2013) 130–136
there is a mild positive influence of Parcel Size on farmland prices at distances over 100 m from a built-up area, the influence of Parcel Size is strongly negative in parcels closer to a built-up area.
-2
Farmland Price (CZKm )
50
40
Discussion 30
20
10
0 No access
Access
Fig. 3. Effect of Parcel Accessibility (ACCESS) on farmland prices. The mean scores (bars) and standard errors (whiskers) are presented using non-transformed data.
than 5000 inhabitants, the average land price is approximately 2.5 times higher (INHAB > 5000; mean = 71.07 CZK m−2 ) than in municipalities with fewer than 1000 inhabitants (INHAB < 1000; mean = 28.86 CZK m−2 ). Of the variables expressing travel times to different levels of statutory towns (Capital City, Regional Capital, District Town), only Travel time to the Capital City (CAPIT, Fig. 2c) proved statistically significant. In land situated within 60 min from the capital city, the price of farmland was more than twice higher (CAPIT < 60; mean = 74.44 CZK m−2 ) than in land with longer travel times. Where the travel time was more than 60 min, the influence of the capital city on land prices was minimal (CAPIT > 60; mean = 35.44 CZK m−2 ). Another significant predictor was Parcel Accessibility (ACCESS). The results (Fig. 3) show that the price of accessible plots was more than twice as high (mean = 45.18 CZK m−2 ) as the price of inaccessible plots (mean = 18.97 CZK m−2 ). The last single factor significantly influencing land prices was Natural Soil Fertility (FERT). However, the role of this factor is ambivalent, as prices tend to increase both in land with the highest Natural Soil Fertility and in land with the lowest Natural Soil Fertility (Fig. 2d). The strongest of the four significant interactions is between Travel Time to the Capital City and Proximity to a Settlement (CAPIT:DIST). The influence of proximity to the edge of the nearest built-up area is stronger in parcels with a longer travel time to the capital city. In areas with travel time to the capital over 1 h, the price of farmland within 100 m from a built-up area is on an average 218% higher than the price of farmland further from a built-up area. However, in areas within 1 h of travel time from the capital city, the difference is only 42%. The interaction between Travel time to the Capital City and Natural Soil Fertility (CAPIT:FERT) shows a significantly diminished influence of soil fertility on farmland prices in areas within 60 min of travel time from the capital city. In these areas, the average price of the most fertile land (over 8.00 CZK m−2 ) is 47% lower than the price of less fertile land (less than 8.00 CZK m−2 ), whereas in more remote areas the price of the most fertile land is 6% higher. The interaction between Settlement Proximity and Parcel Accessibility (DIST:ACCESS) shows that the negative influence of lack of accessibility on farmland prices is strongest in parcels within 100 m from a built-up area, where the average price of inaccessible parcels is 52% lower than the price of accessible parcels. At distances over 100 m, the difference in price is only 36%. The interaction between Parcel Size and Settlement Proximity (AREA:DIST) shows contrasting trends in parcels within 100 m from built-up areas and parcels that are farther away. While
The most significant factor behind the spatial volatility of farmland prices is Proximity to a Settlement. There is a significant increase in farmland prices up to approximately 100 m from the edge of a built-up area, but distinctly the highest prices are paid for farmland directly adjoining a built-up area. This phenomenon could theoretically be attributed to willingness to pay higher prices where the buyer owns farmland adjoining a village and bordering on his/her recent parcel (there is usually higher land fragmentation near to villages). This would lead to consolidation of parcels owned by one owner, and to mitigation of ownership fragmentation, a major factor reducing farm profitability (Di Falco et al., 2010). However, an examination of the ownership of parcels bordering with sold land has proved that this explanation is incorrect, as only 2% of the land was bought by owners of neighbouring parcels. We therefore conclude that the main reason behind this sharp increase in farmland prices is speculative interest motivated by conversion of farmland and subsequent sale of the land for non-farm uses. This is in accordance with findings of Drozd and Johnson (2004), Tan et al. (2009) and Spinney et al. (2011). In the Czech Republic, buildable land is usually delimited by the Master plan, and development permits are nearly always granted only in direct continuity with the current built-up area. Our interpretation of this factor is supported by previous studies, e.g. by Livanis et al. (2006) and Guiling et al. (2009). Some authors measure the distance of the plot from the centre of the village (Cheshire, 1995). In our opinion, the distance of the edge of the parcel from the edge of the current built-up area is a more accurate variable. This approach takes into account relevant conditions, such as the connection to paved roads and to the municipal infrastructure, as well as the probability of conversion of the land into non-agricultural use in the master plan. Moreover, problems with eccentrically placed village centres are eliminated (Delbecq and Florax, 2010). The second most important factor behind the spatial variability of farmland prices proved to be Municipality Population. Our results have confirmed a positive influence of population size on an increase in farmland prices. This result can also be interpreted with regard to speculation on the conversion of farmland mainly to buildable land, as the higher prices of farmland adjoining larger municipalities imitate the higher prices of building land and family homes (Livanis et al., 2006). However, in this case, the variability in farmland prices can also be explained by higher competition between potential local buyers in larger municipalities and by the pressure of this competition toward an increase in prices (Palmquist and Danielson, 1989; Guiling et al., 2009). Speculative purchases of farmland for future development can also explain the high volatility of prices connected with another factor: Travel Time to the Capital City. In areas within 1 h of travel time from the capital city, farmland prices are significantly higher, and within this interval prices increase further with increasing proximity to the capital. This phenomenon cannot be explained by a higher concentration of farmers in and around the capital city. Travel time around 1 h is still acceptable for commuting to the capital city and its outskirts. This is an area traditionally known to have lower unemployment rates, the highest average salary, the highest number of education opportunities, and other benefits (Henley, 1998; McDonald and McMillen, 1998). Therefore, as with the previously discussed factors, the higher prices in areas within a shorter travel time to the capital city can be explained by speculation on high demand for building plots or family homes. A similar influence of
P. Sklenicka et al. / Land Use Policy 30 (2013) 130–136
regional capitals (with a population size of hundreds of thousands) or district towns (with a population size of tens of thousands) has not been proved, contrary to our expectations. In contrast to the conclusions of Naydenov (2009) and along with e.g. Henley (1998), we conclude that length units are not suitable for explaining the lucrativeness of some areas in relation to the capital city, or in relation to large cities in general. The travel time needed to reach the city by car or by some other means of transport is a more suitable descriptive unit. Places which are easily accessible by a highway or, for example, by a high-speed train, can be as much as twice as far away as areas with low accessibility, and still require the same travel time (Haurin and Rosenthal, 2004). The accessibility of a parcel is generally the main condition for fulfilling ownership rights. In the Czech Republic, many parcels have lost their accessibility due to land-use consolidation and closure of field roads during the socialist period (Sklenicka, 2006). The owners of parcels that are currently not accessible by the transportation network are discriminated against in the agricultural use of their land, and in renting their land for farming purposes. Both of these activities are dependent on cooperation with the owners or tenants of neighbouring plots. Our results show that there is a similar discrimination in farmland prices, which reflect the interest in converting farmland to other uses (Zhong et al., 2011). There is apparent unwillingness to pay higher prices for inaccessible land and to risk failure in future negotiation of access with the neighbouring owners. In our opinion, this handicap affects both sales of farmland for agricultural purposes and speculative purchases of land for future non-farm use. This result of our study supports the significance of land consolidation programmes in consolidating land ownership and providing accessibility to each plot by a system of field roads (Sklenicka et al., 2009). The last single factor significantly influencing the variability of farmland prices is Natural Soil Fertility. A considerable increase in farmland prices occurs only in the most fertile land, which can be explained by higher demand from farmers. Higher soil fertility and its influence on land prices mainly reflect interest in using the land for agriculture, as the prices of building plots, recreational plots and other plots are usually not determined by soil fertility (Sklenicka et al., 2002). The only exception may be the least fertile farmland, where speculative purchases for non-farming purposes are supported by the ease and the higher probability of obtaining permission to convert the land to residential use, as well as by the significantly lower financial cost of obtaining this permission than in the case of more fertile land (Forster, 2006; Spinney et al., 2011). This phenomenon is manifested in our results, where higher farmland prices were observed both in the most fertile land and in the least fertile land. This result is partly in contradiction with our original hypothesis, in which we expected a distinctly positive relationship between Natural Soil Fertility and farmland prices. Proximity to a Settlement was also the predictor that occurred most frequently in statistically significant interactions with other predictors. The influence of proximity to the edge of a built-up area was much more pronounced in parcels with a longer travel time to the capital city. This result can be explained by the expectations of faster and more extensive development of residential areas in the proximity of the capital city, resulting in higher demand for land, including plots further from the edges of the currently builtup areas of the villages in this zone. Settlement Proximity in interaction with Parcel Accessibility confirms the importance of the accessibility of a parcel, especially at shorter distances from current built-up areas. This result indicates the high importance of providing access to more expensive parcels, where conversion to residential use is expected. Potential buyers apparently regard poor accessibility of more expensive parcels as a much higher risk than in the case of cheaper parcels. At the same time, inaccessible farmland can easily be consolidated
135
with other parcels owned by the same owner in the process of land consolidation (Sklenicka, 2006), whereas in residential parcels this consolidation is much more complicated and much less likely to happen. The third significant interaction of Proximity to a Settlement is with Parcel Size. Parcel Size was a significant factor only in this interaction, despite the fact that it was the only factor that was identified as important by all the real-estate brokers. Some previous studies have reached different conclusions (e.g. Maddison, 2000). We had expected this factor to be very powerful, especially due to the extreme fragmentation of farmland in the Czech Republic (Sklenicka and Salek, 2008). The results indicate that a positive influence of Parcel Size on farmland prices applies only at longer distances from currently built-up areas. In parcels closer to builtup areas, this influence is negative. This result can be explained by the higher cost of large parcels in more expensive zones closer to a built-up area, where speculation on converting farmland to residential use is frequent. Another possible explanation is the demand for larger parcels for agricultural use (further from current builtup areas), as farming larger blocks of land is more efficient, due to mechanization and the agronomic methods that are used (Gonzales et al., 2004). In the interaction between Travel Time to the Capital City and Natural Soil Fertility, the role of Natural Soil Fertility is distinctly diminished in areas with a shorter travel time. This factor only begins to influence land prices in areas with a longer travel time. Similarly to Henley (1998) and to the variables discussed above, the key reason behind this trend is apparently the buyer’s intention to use the land. In areas with shorter travel time to the capital city, speculation on the conversion of farmland to residential use is more frequent (McDonald and McMillen, 1998). Buyers therefore demand less fertile land with lower taxation levels and a much lower fee for conversion to residential use. In areas with a longer travel time to the capital city, there is a long-term preference for agricultural use, with farmland prices positively reflecting higher soil fertility.
Conclusions Our study has investigated the key factors influencing spatial variability in farmland prices in the Czech Republic, and has analyzed the effects of these factors. In cooperation with 17 experienced real estate brokers, we chose 8 factors that the brokers regarded as the main determinants of farmland prices. Subsequent linear modeling confirmed the significance of 5 single factors and 4 interactions, and showed that, at some values of these significant predictors, farmland prices increased by as much as several hundred percent. The interpretation of the results indicates that specific values of these factors increase the demand for farmland for future non-agricultural use. This non-agricultural use is supported mainly by proximity to currently built-up areas (up to 100 m), proximity to a large municipality (over 5000 inhabitants), short travel time to the capital city (up to 1 h), and access to the parcel via the transportation network. However, our study has for the first time proved that travel time to the regional capital or to the district town is not a significant determinant of farmland prices. Our results do not show relevant differences in the role of factors affecting land markets in comparison with other European or U.S. open-market states. The results further indicate the significant influence of land consolidation projects and of master planning on farmland prices, and the important role of these two forms of landscape planning in protecting farmland from urban sprawl and in supporting the land market (Lisec et al., 2008). Knowledge of factors influencing sale prices of farmland is important for landowners, and also for land purchasers, developers and land policy makers.
136
P. Sklenicka et al. / Land Use Policy 30 (2013) 130–136
Acknowledgments This study was supported by grant No. QH82162 from the Ministry of Agriculture of the Czech Republic “Possibilities of defragmentation of agricultural land”. The authors owe special thanks to Robin Healey for editing the manuscript, and to 17 real estate brokers for their help with obtaining the data. References Awasthi, M.K., 2009. Dynamics and resource use efficiency of agricultural land sales and rental market in India. Land Use Policy 26, 736–743. Barnard, C.H., 2000. Urbanization affects a large share of farmland. Rural Cond. Trends 10, 57–63. Bastian, C.T., McLeod, D.M., Germino, M.J., Reiners, W.A., Blasko, B.J., 2002. Environmental amenities and agricultural land values: a hedonic model using geographic information systems data. Ecol. Econ. 40, 337–349. Becvarova, V., Vasek, P., Vanicek, F., 1988. Bonitace Zemedelskych Pud a Smery Jejich Vyuziti. 4. dil. FMZVZ, Prague. Bernischka, M., Binkley, J.K., 1994. Land price volatility in a geographically dispersed market. Am. J. Agr. Econ. 76, 185–195. Cheshire, P., 1995. On the price of land and the value of amenities. Economica 62, 247–267. Cho, S.H., Newman, D.H., 2005. Spatial analysis of rural land development. Forest Policy Econ. 7, 732–744. Crawley, M.J., 2007. The R Book. Wiley & Sons, Chichester. Delbecq, B.A., Florax, R.J.G.M., 2010. Farmland allocation along the rural–urban gradient: the impact of urbanization and urban sprawl. In: Proceedings of Annual Meeting of Agricultural and Applied Economics Associations, Denver, CO, pp. 1–17. Deininger, K., Jin, S., 2003. Land Sales and Rental Markets in Transition: Evidence from Rural Vietnam. World Bank Policy Research Working Paper 3013, Washington DC. Deininger, K., Gonzalez, M.A., Castagnini, R., 2003. Comparing Land Reform and Land Markets in Colombia: Impacts on Equity and Efficiency. World Bank Policy Research Working Paper, Washington DC. Di Falco, S., Penov, I., Aleksiev, A., van Rensburg, T., 2010. Agrobiodiversity, farm profits and land fragmentation: evidence from Bulgaria. Land Use Policy 27, 763–771. Drescher, K., Henderson, J., McNamara, K., 2001. Farmland prices determinants. In: Proceedings of American Agricultural Economics Association Annual Meeting, Chicago, IL. Drobne, S., Grilj, T., Lisec, A., 2009. Real estate market activity in Slovenia in 2000–2006. Geodetski vestnik 53, 561–579. Drozd, D.J., Johnson, B.B., 2004. Dynamics of a rural land market experiencing farmland conversion to acreages: the case of Saunders County, Nebraska. Land Econ. 80, 294–311. Forster, D.L., 2006. An overview of U.S. farm real estate markets. In: Proceedings of CICA International Confederation of Agricultural Credit, Quebec, Canada, pp. 1–28. Gonzales, X.P., Alvarez, C.J., Crecente, R., 2004. Evaluation of land distributions with joint regard to plot size and shape. Agr. Syst. 82, 31–43. Guiling, P., Brorsen, B.W., Doye, D., 2009. Effect of urban proximity on agricultural land values. Land Econ. 85, 252–264. Haurin, D.R., Rosenthal, S.S., 2004. The Sustainability of Homeownership: Factors Affecting the Duration of Homeownership and Rental Spells. Report for U.S. Department of Housing and Urban Development, Washington, DC. Henley, A., 1998. Residential mobility housing equity, and the labour market. Econ. J. 108, 414–427.
Huang, H., Miller, Y., Sherrick, B.J., Gómez, M.I., 2006. Factors influencing Illinois farmland values. Am. Agric. Econ. Assoc. 88, 458–470. ˇ R., 2008. Modelling the rural land transaction Lisec, A., Ferlan, M., Lobnik, F., Sumrada, procedure. Land Use Policy 25, 286–297. Lisec, A., Drobne, S., 2009. The influence of protected natural and cultural heritage on land management/market: the case of Slovenian natural protected areas. Spatium 20, 41–48. Livanis, G., Moss, C., Breneman, V.E., Nehring, R.F., 2006. Urban sprawl and farmland prices. Am. J. Agr. Econ. 88, 915–929. Lloyd, T.A., Rayner, A.J., Orme, C.D., 1991. Present value model of land prices in England and Wales. Eur. Rev. Agric. Econ. 18, 141–166. Lovett-Doust, J., Biernacki, M., Page, R., Chan, M., Natgunarajah, R., Timis, G., 2003. Effects of land ownership and other landscape-level factors on rare-species richness. Landsc. Ecol. 18, 621–633. Maddison, D., 2000. A hedonic analysis of agricultural land prices in England and Wales. Eur. Rev. Agric. Econ. 27, 519–532. McDonald, J.F., McMillen, D.P., 1998. Land values, land use, and the first Chicago zoning ordinance. J. Real Estate Financ. 16, 135–150. Naydenov, I., 2009. Factors Influencing Farmland Prices in Bulgaria, Master Thesis, Maastricht University. OECD, 1997. Glossary of Environment Statistics, Studies in Methods, Series F, No. 67, United Nations, New York. Palmquist, R.B., Danielson, L.E., 1989. A hedonic study of the effects of erosion control and drainage on farmland values. Am. J. Agr. Econ. 71, 55–62. Petit, S., Usher, M.B., 1998. Biodiversity in agricultural landscapes: the ground beetle communities of woody uncultivated habitats. Biodivers. Conserv. 7, 1549–1561. Plantinga, A.J., Miller, D.J., 2001. Agricultural land values and the value of rights to future land development. Land Econ. 77, 56–67. Scharlach, W.C., Schuh, G.E., 1962. The land market as a link between rural and urban sectors of the economy. J. Farm Econ. 44, 1406–1411. Shi, Y.J., Phips, T.T., Colyer, D., 1997. Agricultural land value under urbanizing influence. Land Econ. 73, 90–100. Skaloˇs, J., Engstová, B., 2010. Methodology for mapping non-forest wood elements using historic cadastral maps and aerial photographs as a basis for management. Journal of Environment Management 91, 831–843. ˇ ´ Z., Trpáková, I., Santr ˚ cková, M., Uhlíˇrová, L., Kukla, Skaloˇs, J., Weber, M., Lipsky, uˇ P., 2011. Using old military survey maps and orthophotograph maps to analyse long-term land cover changes – case study (Czech Republic). Appl. Geogr. 32, 426–438. Sklenicka, P., Lhota, T., Cecetka, J., 2002. Soil porosity along a gradient from forest to field. Die Bodenk 53, 181–187. Sklenicka, P., 2006. Applying evaluation criteria for the land consolidation effect to three contrasting study areas in the Czech Republic. Land Use Policy 23, 502–510. Sklenicka, P., Salek, M., 2008. Ownership and soil quality as sources of agricultural land fragmentation in highly fragmented ownership patterns. Landscape Ecol. 23, 299–311. Sklenicka, P., Hladik, J., Strelecek, F., Kottova, B., Lososova, J., Cihal, L., Salek, M., 2009. Historical, environmental and socio-economic driving forces on land ownership fragmentation, the land consolidation effect and project costs. Agric. Econ. – Czech. 55, 571–582. Sogaard, V., 1993. The land market cycle. Eur. Rev. Agric. Econ. 20, 65–76. Spinney, J., Kanaroglou, P., Scott, D., 2011. Exploring spatial dynamics with land price indexes. Urban Stud. 48, 719–735. Stewart, P.A., Libby, L.W., 1998. Determinants of farmland value: the case of DeKalb County, Illinois. Appl. Econ. Persp. Policy 20, 80–95. Tan, R., Beckmann, V., van den Berg, L., Qu, F., 2009. Governing farmland conversion: comparing China with the Netherlands and Germany. Land Use Policy 26, 961–974. Zhong, T.Y., Huang, X.J., Zhang, X.Y., Wang, K., 2011. Temporal and spatial variability of agricultural land loss in relation to policy and accessibility in a low hilly region of southeast China. Land Use Policy 28, 762–769.