WTO agreement on agriculture: Potential consequences for agricultural production and land-

WTO agreement on agriculture: Potential
consequences for agricultural production and landuse patterns in the Swiss lowlands
Robert Huber & Bernard Lehmann
Robert Huber (Corresponding author)
Bernard Lehmann
Agri-food and Agri-environmental Economics Group, Institute for
Environmental Decisions, ETH Zurich, Switzerland
E-mail: [email protected]
A strong link exists between agricultural production and landscape.
Globalisation (more open agricultural markets) will change agricultural production and thus landscape will change as well. In this
paper, we address the following questions: (a) what economic effects
can be expected with respect to agricultural production structures
in a high cost production region such as the Swiss lowlands given
a substantial development in the WTO; and (b) how does this structural change influence land-use patterns? We discuss the expected
economic effects from a theo-retical point of view and implement
these findings in a spatially explicit normative programming model
The long-term objective of the WTO members is “to establish a fair and market-oriented agriculture trading system”.
In the final act of the Uruguay Round of Multilateral Trade
Negotiations, the Agreement on Agriculture (AoA) constituted a first step in this reform process (Mosoti & Gobena,
2007). The AoA distinguishes three main pillars in the negotiation process: export subsidies, agricultural market access,
and domestic support. In the long run, all export subsidies
must be eliminated, significant reductions in tariffs and
expansion in tariff rate quotas should increase access to
agricultural markets and domestic support must be reduced
to green box measures with a minimal impact on trade. For
high cost production regions with an elevated level of support such as Switzerland or Norway, the outcome of this
process and the associated change in agricultural policies
will alter agricultural production structures significantly.
In Switzerland, farmers manage more than a third of the
for a case study region in the Swiss lowlands. The results show a
wide range of possible economically efficient outcomes depending
on production costs and farmers’ preferences. Our results imply
that, if production costs were to sink sufficiently, income maximizing farmers would focus on grassland based milk production. This
would only lead to a modest change in the existing land-use patterns
since our case study region is currently dominated by dairy farms. If
production costs remain high, agricultural production would shift to
more extensive production activities in order to maximize the sectoral
income. In this case, the local landscape would change noticeably.
WTO, land-use change, mathematical programming model, GIS
Geografisk Tidsskrift-Danish Journal of Geography 109(2): 131145, 2009
country’s surface area. Consequently, changes in the agricultural production system as a result of WTO negotiations
will also have an impact on agricultural landscapes and
thus the landscape as a whole. This long-term perspective
raises a lot of scepticism. Agriculture in these high income
countries is perceived to be more than just a provider of
food and fibre while more open markets are associated with
an undermining of the environmental and social functions
of agriculture. Consequently, questions arise regarding the
economic, ecological and social consequences which can be
expected from the implementation of world market prices
and WTO compatible domestic support in high cost / high
support regions. The construction of such a (counterfactual)
WTO scenario represents a reference point for the evaluation of new agricultural policies (Hodge, 2008). Without
such a hypothetical (but nevertheless realistic) scenario, the
discussion of the possible outcome of freer trade will always
be biased by the status quo and lend strength to protectionist
and conservative forces.
Geografisk Tidsskrift-Danish Journal of Geography 109(2) 131
The goal of this article is to discuss the expected outcome of such a scenario in a high cost production region.
More specifically, we look at expected economic, ecological and social effects for a local landscape in the Swiss
lowlands given a green box compatible direct payment
system and world market prices. Our approach consists of:
a) Discussion of the expected effects from a (theoretical)
economic point of view;
b) An application of these effects to a case study region in
the Swiss lowlands. This entails a modelling approach
since no data for such a scenario is available. Thereby,
we use a normative mathematical programming model;
c) Discussion of the outcome in a broader (ecological and
social) context.
It is important to note that the result of our normative
modelling approach is not a precise prediction of what
the local landscape will look like in the future. However,
we are able to identify economic driving forces, resource
limitations and relevant decision variables in a specific
scenario. The advantage of this approach is that it allows
us to ‘think outside of the box’ and to discuss, in complete
freedom, the consequences of such a scenario (Happe &
Balmann, 2006). As a result, transparency in the discussion on the effects of freer trade on local landscapes can
be enhanced.
The article is organized as follows. The next section
will focus on expected outcomes of WTO negotiations
and the existing adaption strategies of Swiss agricultural
policy. Then, a general overview of the linkages between
agricultural production and agricultural landscape is given.
Thereafter, we discuss the economic consequences of world
market prices for agricultural production and the corresponding landscape. Our methodology is presented and a
model application to the Swiss lowlands is provided. In the
last two sections, we discuss results and present conclusions.
Swiss national policy and the WTO
In the 1990’s, high public spending, overproduction and
environmental problems led to a change in Swiss agricultural policy. In line with the AoA, Switzerland changed
its Federal Constitution in 1996 (BLW, 2004). Since then,
there has been an ongoing reduction of market support
and agricultural markets have been liberalized. Domestic
support shifted to decoupled direct payments. At the same
132 Geografisk Tidsskrift-Danish Journal of Geography 109(2)
time, a cross compliance approach was introduced in order
to guarantee a certain level of environmental goods and
services (Joerin et al., 2006). However, overall support
remained high. In the last twenty years, the Producer Support Estimate (PSE), an indicator for agricultural support,
only sank from 78 to 66% (OECD, 2007).
In the WTO negotiations, Switzerland leads the G-10
group, which represents net food importing countries.
These countries oppose the idea of completely free agricultural markets and refer to the multifunctional role of
agriculture. From a purely economic perspective, multifunctionality describes the fact that agriculture produces
both commodity and non-commodity outputs (OECD,
2001). If the latter involves some kind of market failure,
social welfare may be enhanced by supporting agriculture
to provide these non-commodity outputs (e.g. landscape).
The idea of multifunctional agriculture meets criticism.
For instance, Anderson (2000) argues that only a minor
trade-off is required to meet domestic policy objectives
on the one hand and agricultural protection reform objectives as envisaged in the WTO rules on the other. In this
respect, the OECD argues that support should pursue the
different objectives with a minimum of economic distortion
both domestically and internationally (OECD, 2003). In a
broader perspective, however, queries do arise concerning this rather neoliberal point of view. (For a discussion
of multifunctionality in the context of WTO negotiations
see: Dibden et al., 2009; Potter & Tilzey, 2007; Potter &
Burney, 2002).
As a final agreement in the multilateral WTO Doha
round has been rescheduled (2009), Switzerland aims for
a bilateral agreement in the agri-food sector with its largest trading partner, the European Union. In either case, the
producer prices received by farmers would sink considerably.
In conclusion, Swiss farmers will face substantial producer price reductions in the medium term. This applies
even without a concluding agreement in the Doha round.
The green box compatible direct payment system, however,
will remain the central pillar in Swiss agricultural policy.
Upcoming modifications will even reinforce the green box
compatibility of this direct payment system (Vogel et al.,
Agricultural production and agricultural
Historical perspective
In the last centuries, landscape rarely was a designed good
but most frequently the result of different production systems. Historically speaking, changing agricultural production processes have always left their mark on the landscape.
Thus, these transformations in agricultural landscapes reflect the changing needs for food, fibre and energy and the
associated policy measures. Until 1850, for instance, Swiss
agriculture was dominated by subsistence farming, which
mainly involved crop production. With the emergence of
the steam engine and the corresponding transport facilities, wheat and other crops were imported to Switzerland.
Price relationships between crops and livestock changed
significantly. Agricultural production switched from crops
to milk and meat production and thus grassland became the
dominant land-use. Economic crises and two world wars
led to the emergence of a highly regulated agricultural sector. Crop production was supported in order to secure food
availability in times of need. This led to a re-emergence of
the production of a wide variety of crops in Switzerland.
Landscape is an amalgam of natural, economic and cultural
aspects and can be defined in different ways. Hence it is
important to delimit the use of the term landscape for this
article. We will focus on agricultural land-use e.g., the
extent of crop production, grassland and agricultural nature
conservation areas. Since we concentrate on the effects
on agricultural land-use, we do not consider agricultural
landscape elements (trees, hedgerows), other forms of land
cover (forest, building areas) or invisible landscape functions such as biodiversity conservation and the preservation
of natural resources (nutrient runoff and soil conservation).
The link between agricultural production and agricultural landscape is straightforward (Vanslembrouck & Van
Huylenbroeck, 2005). Farmers use natural resources (e.g.
land) in their production process which leads inevitably to
a change in the natural state of these resources. However,
the analysis of the interactions between agricultural production and the provision of landscape is complex. On the
one hand, farmers’ choices are influenced by a complex
set of factors such as input and output prices, technological
innovations and policy measures to name but a few. On the
other hand, the link between different production activities
and the natural resources is often a non-linear and complex
process which can exhibit threshold effects. The level of
intensity, for instance, can have a significant impact on
the provision of landscape and its functions. This twofold
complexity is the source of a wide range of literature on the
interactions between agriculture and landscape (Ferrari &
Rambonilaza, 2008; Mander et al., 2007; Brouwer, 2004).
Microeconomic perspective
Given our research goal, we focus on a microeconomic
interpretation of the connection between agriculture and
agricultural landscape. According to Boisvert (2001), jointness can originate from three causes: technological interdependencies, non-allocable inputs or fixed inputs on farm
level. Abler (2004) summarises non-allocable inputs and
fixed inputs as economic interdependencies.
• Technological interdependencies refer to the fact that
the agricultural production process, such as the use of
fertilizer or pesticide, can have inseparable (typically
negative) effects on landscape amenities.
• In the case of non-allocable inputs, the outputs of two
products depend on each other. The provision of open
space, for instance, is related to some form of agriculture which produces a certain output (e.g. milk, meat or
fibre). The input factor land cannot be allocated solely
to either the production of the commodity or to the
provision of open space.
• Fixed inputs on farm level lead to a competitive relationship between the commodity and non-commodity
outputs. For example, if the amount of land belonging
to a single farm is fixed, the allocation of a certain
amount of land to biodiversity conservation will lead
to a reduction in the amount of land available for milk
The causes of jointness vary widely when it comes to
landscape provision. Obviously, both technical and economic interdependencies contribute to the provision of
landscapes. On an aggregate level, a combination of these
causes contributes to the provision of landscape benefits.
Consequently, changes in the agricultural production process (e.g. the variation in inputs) will alter the landscape.
Expected economic effects of world market prices
for the agricultural sector in Switzerland
The theoretical economic aspects which could be expected
from our scenario are manifold and vary on different scales
(farms, regions, countries). It would be impossible to inGeografisk Tidsskrift-Danish Journal of Geography 109(2) 133
clude all effects in a model approach. Therefore, we focus
on three central aspects of the discussion on free trade and
their consequences for agricultural production:
• Specialisation of the agricultural sector on activities
with a comparative cost advantage;
• Structural change in order to realize economies of scale;
• Importance of natural conditions due to the sinking
value of the marginal product.
The concept of comparative (cost) advantages refers to
the fact that one country (or person / farm) can produce a
certain product at lower opportunity costs. Given free trade
markets, this increases the incomes, and thus welfare, in
the trading countries. The idea of comparative advantage
goes back to David Ricardo (1821) who showed that if
two countries produce two goods, a country can benefit by
specialising in one good and trading the other, even if it has
lower productivity for both goods. This model is clearly a
simplification compared to the real world. Nevertheless, it
illustrates an economic driving force if market access (and
thus trade) is increased.
Economies of scale represent lower production costs
through an expansion of production. This refers to structural
change in the agricultural sector. Structural change, however, has become a catchword in the economic assessment
of the agricultural sector all over the world. Irrespective of
the actual size of a farm in any country, structural change
always has the potential to increase the competitiveness
of the sector. Switzerland, however, has small structures
compared to other industrialized countries. Therefore, the
potential for improvement is large even within the concept
of family based / smallholder farms.
The gain in the importance of natural conditions can be
inferred from the lower value of the marginal product due
to lower output prices. The corresponding economic rule
states that the factor price must correspond to the value of
the marginal product. If the output price decreases (and the
factor price remains constant), the marginal product must
increase. This can have two effects: if the agronomic production suitability of farmland is high, production becomes
more intensive. If this suitability is low, however, farmland
is abandoned. The abolition of price support increases the
relevance of natural production conditions. This effect is
not unique to high cost regions. The same processes can
also be observed in different countries (Primdahl, 2010).
However, we are aware that the direct payment system will
reduce this effect.
134 Geografisk Tidsskrift-Danish Journal of Geography 109(2)
As a result of the combination of these three effects, our
scenario (world market prices and the existing green box
compatible direct payment system) would imply that in the
long-term, there will be larger farms practicing intensive
agricultural production on good soils while marginal land is
abandoned. This qualitative interpretation also reveals the
limits of our approach. Given these economic tendencies, it
is obvious that a diversification strategy could also ensure
economic viability. Added value products (geographical
indications e.g. AOC), product differentiation (organic agriculture, tourist services) and vertical integration (products
processed directly on the farm) are important issues in the
context of decreasing producer prices. However, our model
approach is not suitable for the illustration of such effects.
In the next section, we apply the three above-mentioned
economic concepts to a case study region in the Swiss
lowlands using a normative programming model.
Model application to the Swiss lowlands
Mathematical programming models (MPM) are based on
the principles of neoclassical economics. Thus, economic
agents are profit optimizers and in combination with limited
resources represented by model restrictions, these models
incorporate the fundamental economic problem: making
the most of limited resources (Buysse et al., 2007). MPMs
are meaningful when analysing the environmental impacts
of agriculture because the basic linkage between agricultural production and environmental indicators can conveniently be modelled. Differences in agricultural production
technologies can be attributed with different coefficients
for environmental outcomes (Buysse et al., 2007). This
basic approach applies also to the implementation of the
microeconomic causes of jointness between agricultural
production and landscape. Technical interdependencies
are implemented through coefficients. Different amounts
of nitrogen loss, for instance, are associated with different
levels of bovine grazing intensity. More importantly, economic interdependencies can be represented through model
results. Given a specific economic environment, a farmer’s
decisions lead to a certain land-use pattern and its associated
ecological effects. Thus, our approach allows a quantification of joint economic and ecological effects on a landscape
level. However, we are aware that the representation of rural
landscapes in MPMs is a difficult task (Janssen & van Ittersum, 2007). There is a lack of knowledge on the interaction
between agricultural practices and the ecological outcomes
on a landscape level (Rossing et al., 2007; Tscharntke et al.,
2005). Moreover, the implementation of agri-environmental
schemes as an activity in our model does not necessarily
mean that the intended environmental outcome is attained
since there is a difference between the performance and the
outcome effect of an agri-environmental scheme (Primdahl
et al., 2003; Oñate et al., 2000).
We use a spatially explicit sector supply model to assess
the effects of world market prices on the land-use patterns
in the Swiss lowlands. This is a linear optimization model
which maximizes the aggregate annual income (labour
income plus land rents) of a specific region giving consideration to cropping constraints, plant nutrient requirements,
manure production, forage and fertilizer balances, as well
as structural constraints and the natural production conditions (Peter, 2008; Huber, 2009a; Hartmann et al., 2009).
The specification of the policy environment in the model
depicts Swiss agricultural policies in detail. This entails
the regulations of the cross compliance approach in the
Swiss agricultural law (balanced use of fertilizers, appropriate proportion of ecological compensation areas, crop
rotation, suitable soil protection measures, selection and
specific application of plant treatment products, animalfriendly conditions for livestock). However, we integrated
only WTO compatible forms of direct payments, i.e., only
payments based on acreage and no payments based on the
number of animals.
The model includes all important activities relating to
income generation, land-use, livestock as well as ecological
indicators. In addition, extensive agricultural activities such
as sheep or goat husbandry as well as nature conservation
activities (without the production of food or feed) are explicitly part of the model (Table 1).
Economic concepts are implemented in the programming model in various ways. The implementation of the
comparative cost advantage and the corresponding specialisation is inherent to a normative programming model.
It is the purpose of such a model to choose the best option
under a given set of resource constraints. In fact, these
models even tend towards ‘overspecialization’. Linear
mathematical programming solutions have a tendency to
produce extremely specialised solutions since the number
of production possibilities employed is influenced by the
amplitude of the constraint set and the associated production possibilities (Wiborg et al., 2005).
Economies of scale are implemented through the introduction of benchmark farms and full working load of
the machinery. With respect to benchmark farms, German
construction data is used instead of Swiss. This is necessary because existing production structures in Switzerland
are very small and thus production costs are very high.
The implementation of such a benchmark farm reduces the
construction costs (amortisation, interest) per cow by 35%.
Without such a reduction in production costs, the imple-
Table 1: Model activities and specification.
Model activities
Root crops (sugar beet, potatoes)
Cereals (wheat, barley, triticale)
Oil seeds (sunflowers, rape)
Grassland (permanent, rotational)
Yields per parcel (soil and climatic suitability);
intensity levels (intensive, mid-intensive,
extensive); size of the parcel
Milk (dairy cattle, rearing cattle, goats)
Beef cattle (sucklers, calves, bulls)
Meat (pigs, lamb, broilers)
Eggs (pullets, laying hens)
Animal type; housing system and size; livestock
efficiency; feeding system; free range management
Extensive grassland
Meadows with minimum size of 0.05 ha, no
fertilizer or phytosanitary measures allowed,
restrictions on mowing (late cut – after June 15)
Rotational fallows
Areas sown with indigenous wildflowers, forbs
and legumes, integrated in the crop rotation (at the
same location for one to three vegetation periods),
fertilization and treatment with insecticides are not
Geografisk Tidsskrift-Danish Journal of Geography 109(2) 135
mentation of our chosen scenario with world market prices
would result in a complete abandonment of agriculture. In
fact, this is just what interest groups in favour of border
protection claim. However, the purpose of our article is to
bypass such foregone conclusions and to discuss a possible
scenario with agriculture still in place. The basic idea is
that if there is no border protection, Swiss production costs
(analogous to the producer prices) approach those of its
closest neighbours.
In order to represent the increased importance of natural
conditions, our normative programming model is linked
to a Geographical Information System (GIS) model. The
latter provides detailed information on soil and climatic
suitability for agricultural production. Based on the existing
land characteristics, the GIS model forms continuous land
units which are homogenous in their agricultural production suitability. In addition, these land units contain information on the climatic suitability, average slope and the
suitability for biodiversity conservation. The latter assesses
the proximity to natural habitats and represents a better
accessibility from nature conservation activities to bodies
of water (lakes, streams) or woods (forests, trees, hedgerows) The assumption is that agricultural land units with
a direct connection to natural habits are more suitable for
biodiversity conservation than those which are separated
or isolated.
This information enters into the optimization model.
Production and climatic suitability influence yields, while
size and average slope influence production costs on each
specific parcel. The information on the suitability for biodiversity conservation determines the allocation of the nature
conservation activities in the model (Huber, 2009a).
Methodological limitations of our MPM have some
well-known limitations. Firstly, the one-dimensional objective function (income maximisation) does not represent
diverging preferences, values and risk behaviour. Secondly,
its linearity and the lack of feedback effects which would
be expected, for instance, from input and output markets.
Thirdly, a calibration to real world data is not possible
due to the integration of German construction data and the
fact that existing structural conditions are intentionally not
taken into consideration. However, the original model fits
to real world data in an appropriate manner (Hartmann et
al., 2009). Thus, we assume that the model specification
represents the agricultural production process quite adequately. Furthermore, the GIS model disregards property
rights. In reality, landscape fragmentation would be higher
than in our modelling approach. Nevertheless, our model is
suitable for the purpose of this article since we do not make
136 Geografisk Tidsskrift-Danish Journal of Geography 109(2)
any predictions about the future but try to identify driving
forces, resource limitations and relevant decision variables
in a specific scenario. However, the above-mentioned disadvantages must be taken into account if the scenario is
discussed in the broader context.
Case study region
The case study region (District Muri) is situated in the
central part of the Swiss lowlands and has an area of approximately 10,000 ha. The region can be characterized as
a peri-urban rural area. This means, it is neither part of an
agglomeration nor a city. The distance to agglomerations
(Zurich), however, is small. Agricultural structures are
dominated by mixed farms (50%). 304 out of 536 farmers are milk producers. The number of cows amounts to
slightly more than 7,000 (13 cows per farmer). In addition,
there is considerable pig and poultry production. Livestock units per ha (all animals) amounts to 1.9. 20% of the
farms have an agricultural area of less than 10 ha, 46%
have 10-20 ha available for production and 34% cultivate
an area bigger than 20 ha, whereby only 3% of these have
over 40 ha of land. Average farm size is approximately 18
ha which is slightly above the Swiss average of 16.7 ha
per farm. Still, average farm size must be characterised as
very small. Land-use is dominated by grassland (57% of
total area), cereals (21%) and maize production (16%). The
district lies in a valley in the bottom of which climatic and
soil conditions are good. The hillsides are less suitable for
agricultural production.
Producer prices for agricultural commodities in Switzerland are significantly higher than the EU or world market
level. On average (2003 to 2006), wheat prices in France
and Germany, for instance, were 70% lower than in Switzerland. With regard to the price of milk, differences in
Europe were smaller (70% of Swiss prices). However, the
milk price in New Zealand amounted to only 40% of the
Swiss price.
When constructing the scenario, we assume that Swiss
producer prices approximate world market prices. As a
proxy for world market prices, we apply US producer price
averages from 2003 to 2006 (Table 2). In order to avoid
the problem of exchange rate fluctuations, we multiplied
existing model parameters (in Swiss francs CHF) by the
price level in the US (percentage).
Furthermore, we assume that production costs also
decrease. Lower fixed costs are modelled by introducing
benchmark farms and full working load of the machinery.
Table 2: Implemented producer price reduction.
Unit: USD
Milk price
Beef cattle
Sugar beet
reduction in
the model
Source: FAO (2009)
Benchmark farms have lower investment costs because
they are at least three times bigger than the average Swiss
farm today. We introduce German data in order to represent
these scale effects. Small production units also lead to high
machinery costs since farmers do not utilize their machines
to the full. Again, the assumption that bigger farms will
make more use of their machinery leads to the assumption
of lower fixed cost in the production process (Table 3).
Table 3: Fixed cost reductions.
In addition, we introduced a cost level parameter. This
allows us to show changes in the production pattern if production costs (fixed and variable) are further decreased.
The idea behind reduced production costs is that open markets will also influence factor markets. This is certainly
valid for the cost of pesticides since these products are
indirectly protected by the existing policy environment.
However, the cost of services (e.g. for the veterinarian)
is influenced more strongly by local economic conditions
(purchasing power, salary etc.) than by global markets.
Thus, it is difficult to assess the overall dimension of this
cost reduction. In the model calculations, we apply a sensitivity analysis in order to assess the different cost levels.
In addition to the different cost levels, we calculate
the optimal solutions for different levels of opportunity
costs for labour. In our model, opportunity costs represent a
minimal factor compensation for labour. If the farmer earns
less than the level of opportunity costs, the corresponding
economic activity does not enter the optimal solution. As
a result, the income per hour and farmer has a lower limit
and thus represents an exit threshold. Low opportunity
costs imply that farmers have preferences to stay in the
agricultural sector that are not solely dictated by financial
motives. High opportunity costs, on the other hand, imply
that the farmers have the possibility to work outside the
sector and earn at least these opportunity costs.
Opportunity costs can certainly also be seen as production costs. The difference is that the production cost level is
an exogenous variable for the individual farmer whereas the
level of opportunity costs can vary significantly between
the farmers due to, for example, differences in preferences.
In reality, it can be observed that farmers tend to produce
even with very low hourly wages. Thus, results from our
two calculation setups (decreasing production costs and
decreasing level of opportunity costs) differ with respect
to their interpretation.
(selected) Machinery
Seed drill
Trailer (grass)
Source: KTBL (2006); ART (2006)
Our results show long-term, static effects on land-use and
farm characteristics. Figure 1 illustrates the results with a
decreasing production cost level (1A) and a reduction in
the opportunity cost for labour (1B). In 1A, opportunity
costs are set at CHF 10; in 1B, the cost level is set at 85%.
Thus, the vertical line in each figure indicates the identical
solution in the two calculation sets.
If production costs remain high in our scenario, extensive forms of agricultural activities dominate the solutions.
Geografisk Tidsskrift-Danish Journal of Geography 109(2) 137
1. A) Decreasing production cost level (Opportunity cost for labour: 10 CHF)
Animal production
Extensive grassland
Rotational fallows
Milking cows
1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65
1.00 0.95 0.90 0.85 0.80 0.75 0.70 0.65
Production cost level (%)
Production cost level (% )
1. B) Decreasing opportunity costs for labour (Production cost level 0.85)
Animal production
Extensive grassland
Rotational fallows
Milking cows
Opportunity cost for labour (CHF)
Opportunity cost for labour (CHF)
Figure 1: Land-use and livestock production with decreasing production costs and opportunity costs for labour.
Sheep grazing (production of lamb) would dominate landuse. As a result, most of the agricultural area is used for
nature conservation activities. Almost two thirds of the
agricultural area is used as extensive permanent grassland
(64%). Rotational fallows account for 11% of the total area.
Cash crops, maize, rape (oil) and sugar beet are cultivated
on 14% of the area. The rest of the area (11%) is used as
more intensive grassland in order to produce fodder stocks
for the winter and to feed the small number of cows in the
optimal solution. This number, however, increases gradually with a decreasing production cost level. If the production cost level is lower than 85% of the existing level,
grassland based milk production emerges as the dominant
138 Geografisk Tidsskrift-Danish Journal of Geography 109(2)
production activity in the case study region. The amount of
extensive grass dairy cows may consume is restricted due
to dietary reasons. Thus, land-use changes with a switch
from extensive to more intensive forms. Extensive grassland dwindles to a level of 1,900 ha, corresponding to 20%
of the total area. Rotational fallows make way for more
productive activities. Cash and fodder crops increase to a
level of 2,000 ha, which corresponds to 22% of the total
area. The production of lamb disappears completely in the
optimal solution if the cost level sinks below the 80% level
(right figure in 1A).
The same effects can be observed if the opportunity
costs for labour, representing a minimal factor compen-
Table 4: Existing land-use compared to model output for different levels of agricultural production.
Land-use 2005 (ha)
Model output
Nature conservation
Opportunity cost
% of 2005
Extensive agriculture
% of 2005
Intensive agriculture
% of 2005
Oil seeds
Nature conservation schemes
Root crops
Production cost level
Oil seeds
Nature conservation schemes
Root crops
sation for the farmer, are varied between 28 and 0 CHF
(Figure 1B). With high opportunity costs, agricultural
production would be virtually non-existent. Due to the
high direct payments for cultivated land, however, nature
conservation schemes continue to be pursued in the optimal
solution (rotational fallows 80%, extensive grassland 10%
of the total area). Practically speaking, the region would
be a flowering field. If opportunity costs are decreased,
a two step process can be observed. Firstly, nature conservation schemes are replaced by extensive agricultural
production activities. In our model approach this is represented by grazing sheep. In analogy to the situation with
high production costs, the amount of extensive permanent
grassland increases. Secondly, if opportunity costs sink
even further, more intensive agricultural activities enter the
optimal solution. Again, grassland based milk production
emerges as the dominant agricultural activity. However,
with decreasing production costs, there is a further decline
in the amount of crops compared to the calculations. This
can be deduced from the fact that lower opportunity costs
favour labour-intensive activities such as milk production.
Apart from this difference, the results of the two calculation setups show the same effects. If production costs remain high, there would be a shift in agricultural production
and land-use would be dominated by nature conservation. If
production costs sink, either through lower production costs
or lower opportunity costs, extensive agricultural activities
such as sheep grazing are the most profitable. A further
decline in costs implies that intensive grassland based milk
production would become the most profitable agricultural
activity in the case study region. To sum up, there are three
possible stages of agricultural production that can be deduced from our scenario: (a) nature conservation with little
agricultural production, (b) extensive agriculture, and (c)
intensive agriculture. Tables 4 and 5 compare the model
output associated with these three stages to the existing
agricultural structures in 2005. Table 4 focuses on land-use,
whereas Table 5 shows differences in the number of farmers, income per farmer, animal production and livestock
Geografisk Tidsskrift-Danish Journal of Geography 109(2) 139
Table 5: Existing number of farms, income per farmer, animals and livestock intensity compared to model output for different levels of
agricultural production.
Data 2005
Model output
Nature conservation
Opportunity cost
Number of farms
% of 2005
Extensive agriculture
% of 2005
Intensive agriculture
% of 2005
Dairy cows
Income per farmer (CHF)
Livestock intensity per ha
Production cost level
Number of farms
Dairy cows
Income per farmer (CHF)
Livestock intensity per ha
Overall, the comparison shows large differences between existing agricultural production and the model results. This is hardly surprising since our assumptions are
somewhat extreme and no consideration is given to any
adjustment process. Nevertheless, results reveal the economic driving forces and the consequences for land-use
under our world market scenario. As was to be expected,
the greatest differences are to be found between existing agriculture and the nature conservation state, while intensive
agriculture exhibits the smallest differences. Table 4 and 5
also illustrate clearly the main conclusions of Figure 1: low
production costs favour a grassland based milk production;
high production costs encourage extensive sheep production; high opportunity costs make a nature conservation
state more profitable.
With respect to land-use, our results show that the share
of abandoned land is very low. Even without productive
agriculture, the share of abandoned land in the model is
less than 5%. On the one hand, this reflects rather small
differences in the agricultural production suitability in our
case study region. On the other hand, it emphasizes the role
of high direct area payments in Swiss agricultural policy as
a main economic driving force. Obviously, direct payments
can compensate the reduction in price support with respect
140 Geografisk Tidsskrift-Danish Journal of Geography 109(2)
to the input of land. Agricultural land would be cultivated
even without the production of a commodity. Moreover,
our model results imply that crop production declines and
grassland increases (except for the nature conservation
state in which there is only a small share of productive
agriculture). In the case of crop production, we can observe
a shift from cereals to oil seeds due to the changes in relative output prices. The area dedicated to crop production
in absolute terms, however, is clearly less than the existing
share. In the case of zero opportunity costs, farmers would
focus on the production of sugar beet on land units with
high production suitability and stop cultivating cereals and
oil seeds. In the case of reduced costs, however, oil seeds
(rape) remain in the optimal solution. The reduction in the
area for cash crops is due to the comparative cost advantage
of lamb and milk production in the case of an extensive and
intensive agricultural production respectively. The differences shown in Table 5 underline this conclusion. Livestock numbers change drastically depending on the state of
agricultural production. Extensive agricultural production
would lead to a tenfold increase in the number of sheep.
However, the number of sheep currently in the region is
rather low. Intensive agricultural production would lead to
a 40% increase the number of dairy cows. Nevertheless,
livestock intensity would be lower than in the actual situation. This can be inferred from the fact that in any case, pig
and poultry production does not enter the optimal solution.
However, in reality, the production of pigs and poultry
accounts for 36% of the livestock units. In addition to the
unfavourable production parameters, the drop out of these
model activities can be ascribed to the animal protection
law in Switzerland which sets a maximum stock level of
1,500 and 18,000 for pig and poultry stables respectively.
Thus, economies of scale are limited.
The number of farms also varies widely in the different
states of production. In the case of intensive agriculture,
over 80% of the farms will remain in business. In this
case, the income per farmer would be comparable to the
existing income. Given the structural development in our
modelling approach, a reduction in production costs to the
70% level would even result in a 19% increase agricultural
income. In contrast, nature conservation would be possible
with 90% less farms. In this case, the income per farmer
increases more than threefold. With extensive agricultural
sheep production, the increase in income would almost be
twofold. Yet again, this is basically a consequence of high
direct payments in the Swiss policy scheme.
The results of our modelling approach can be represented in a GIS map (Figure 2). The graphical representation of the model solution illustrates the consequences for
the land-use patterns under the three states of agricultural
production. Given high opportunity costs, nature conservation schemes dominate the landscape (a). If the level of
opportunity costs sinks, extensive grassland emerges as
the dominant form of land-use (b). The same land-use can
be observed if opportunity costs are moderate (CHF 10)
but production costs remain high (d). If costs sink even
further, grassland emerges as the dominant land-use (e).
Extensive land-use shifts to the less productive land units
on the hillsides of the case study region. If opportunity costs
are assumed to be 0, sugar beet is the only crop cultivated
and the rest of the area would be used as grassland for milk
production (c).
In conclusion, our model shows that the economic outcome of our WTO scenario has different potential outcomes
for the land-use patterns. The structural change and spe-
Extensive agriculture
Intensive agriculture
Opportunity cost CHF 20 (b)
Opportunity cost CHF 0 (c)
Cost level 100% (d)
Cost level 70% (e)
4.5 km
Nature conservation
Opportunity cost CHF 28 (a)
Figure 2: Selected GIS representation of model results.
Geografisk Tidsskrift-Danish Journal of Geography 109(2) 141
cialisation implemented in our model approach, together
with the WTO compatible direct payment system, lead
to a more productive agricultural sector which is able to
produce food even with world market prices. However, if
farmers are unable to further reduce their production costs
or if they claim a high wage for their working hours, the
outcome may be completely different. From a productive
perspective, grassland based milk production is the most
likely agricultural activity to be pursued in our case study
region. This outcome would have the lowest effect on landuse patterns because land-use would be comparable to the
current state. In contrast, the existing direct payment system
with high area payments and additional green payments can
be seen as an economic driving force which favours a more
extensive agricultural production. Thus, the emergence of
a ‘nature conservation agriculture’ must be considered as
an equally plausible outcome of our scenario.
From a historical point of view, agricultural landscapes
have always reflected the need for food and, to a lesser
extent, fibre and energy. Since the change in agricultural
production structures is an inevitable process, local agricultural landscapes in Switzerland are going to change
irrespectively of more globalized markets (El Benni &
Lehmann, 2010). However, a new agreement in the WTO
negotiations would change the market system for food considerably and thus the characteristics of land-use patterns
will undergo even more distinctive changes.
The results of our calculations show that, under world
market prices, agriculture in the Swiss lowlands would,
to a certain extent, be economically viable. However, an
increase in productivity and an adequate reduction in production costs is a prerequisite for a productive agricultural
sector. In this case, farmers would focus on milk production. Since the existing landscape is already dominated
by dairy farms, the change in land-use and thus landscape
would be moderate. In contrast, high production costs or
high opportunity costs for farmers’ labour would entail
more extensive forms of agricultural land-use. This would
change the land-use patterns significantly. Moreover, our
model results show an improvement in ecological indicators. In the world market price scenario, farmers have to
provide the same environmental services as they do today.
In addition to model inherent environmental services, farmers increase extensive forms of land-use (extensive grassland) and cease intensive (independent of land-use) farming
142 Geografisk Tidsskrift-Danish Journal of Geography 109(2)
activities such as pig and poultry production in the optimal
solution. As a result, livestock intensity in the case study region sinks irrespective of production and opportunity costs.
This leads to a reduction in environmentally harmful emissions from agriculture. Moreover, our calculations show
that the whole agricultural area is still cultivated. Thus,
open space amenities would be provided in any case. This
is due to high direct payments for any use of agricultural
land. However, the level of agricultural income in the case
of ‘nature conservation agriculture’ indicates that the direct
payments must be adapted to sinking commodity prices
otherwise the extent of the support may be disproportionate. Moreover, regional differences must be emphasised.
In Switzerland, for example, the transfer of our results to
a mountainous region would certainly lead to the wrong
conclusions. In marginal areas (in terms of agricultural
production suitability and geographical situation), the need
for support is much more pronounced. Our model suggests
that marginal areas are used as extensive grassland and that
production intensity on good farmland remains high. This
will be even more pronounced in regions with greater differences in natural conditions or with greater susceptibility
to climate change.
The question remains how more extensive land-use and
the corresponding change in landscape would be perceived
by the public and whether it would be accepted even if the
associated ecological consequences were favourable. In
this respect, Anderson (2000) argues that it is not possible
to predict whether new uses of land (for different farm
activities, or for golf courses, recreation parks or similar)
would be any less aesthetically pleasing than the current
uses. In addition, Bromley (2000) advocates the idea that
if agriculture is not competitive, farmers should focus on
the provision of landscape and habitat management. In
this case, the production of food and fibre would be a secondary product of landscape and habitats. In support of
such arguments, an economic valuation study in the Swiss
lowlands shows that the public has a positive willingness
to pay for more extensive farmland at the expense of high
intensity grassland (Schmitt et al., 2005), and Schüpbach
et al. (2009) as well as Junge et al. (2009) demonstrate that
Swiss citizens perceive landscapes with a high share of
extensive farmland as prettier than those with a lower level.
As our results imply, lower commodity prices associated
with high direct payments could push agriculture in this
However, the amount of extensive grassland in our scenario would be far greater than the demand reported in these
studies. Moreover, a choice experiment involving politi-
cians from the case study region reveals that they would not
accept such a scenario (Huber, 2009b). The conservative
political right will not accept certain land-use scenarios if
they impede agricultural production of food and fibre too
much. This is in line with opinion surveys of Swiss citizens
which report that food production is still perceived as one of
the main functions of agriculture (e.g. Tutkun et al., 2007).
Furthermore, this also reflects the image most farmers have
of themselves. For example, a choice experiment carried
out by Lips & Gazzarin (2008) shows that dairy farmers
in the Swiss lowlands exhibit a strong preference for milk
production even if they have lower salaries and poorer
working conditions compared to other sectors. To sum
up, more extensive forms of agricultural land-use would
be opposed even if they were economically efficient.
Food security is an important issue in the context of opposition to more extensive forms of land-use. Indeed, food
security arguments served to justify most of the protectionist measures observed in the last century. Due to the food
crisis in 2008, this issue has re-emerged as a topic for the
attention of agricultural policy makers. Even though there
are major reservations concerning such policies in industrialized countries (Mann, 2008), they are nevertheless a
political reality. As a consequence, the change in perspective supported by Bromley and the corresponding radical
new forms of agriculture would not meet with public or
political acceptance in Switzerland. In this connection, Wilson (2007) argues that productivist and post-productivist
action and thought will occur simultaneously anyway. This
underlines the existence of other economic driving forces
not implemented in our model, such as product differentiation or the possibility to work part-time on the farm. Thus,
our results show the range of an economically efficient
spectrum encompassing productivist to post-productivist
activities rather than socially optimal solutions.
While acknowledging the methodological limits of
our approach, the overall advantage of our line of thought
would be the possibility to break the vicious circle of the
so-called ‘subsidy trap’ (Happe & Balmann, 2007): In
Switzerland, small-scale agriculture is reliant on financial
assistance in order to be able to produce at all. Yet it is precisely the agricultural policy of the past which has pushed
farms into this situation. Certain compensation payments
are certainly necessary and can be justified – but the reason
for them must be made transparent. From an economic
perspective, the most promising policy measures are the
implementation of a direct payment system which distorts
trade as little as possible (OECD, 2003) and the adoption of
policy measures that support structural change by allowing
farmers to reduce their production costs. This would enable
farmers to produce commodities in a competitive environment and to efficiently provide agricultural landscapes at
the same time.
We analysed the consequences of world market prices for
agricultural production and the land-use patterns in the
Swiss lowlands using a mathematical programming model.
Given a sufficient reduction in production costs, our results
imply that income maximizing farmers would focus on
grassland based milk production. This would only lead to
a modest change in the existing landscape since our case
study region is currently dominated by dairy farms. If production costs remain high, agricultural production would
shift to more extensive production activities in order to
maximize the sectoral income. However, if a certain level
is exceeded, farmers would merely cease production and
cultivate their land in order to get direct payments. This
would change the land-use patterns considerably. The main
driving forces behind this development are the implementation of the direct payment system and the farmers’ possibility to reduce their production costs, in particular, by means
of structural change which would result in more productive
farms. Moreover, different preferences of the individual
farmers influence the outcome of our scenario significantly.
Low opportunity costs, representing noneconomic preferences to remain in the agricultural sector, reinforce the
trend towards grassland based milk production in the Swiss
lowlands. The results of our scenario show that freer trade
and the corresponding lower prices for commodities do not
necessarily mean that agricultural production vanishes in
the Swiss lowlands and that the existing agricultural landscape disappears. In fact, the model showed a large range
of possible, economically efficient outcomes. Whether or
not these outcomes are sustainable and how they could be
implemented remain open questions. The answers must be
sought beyond the confines of a purely economic analysis.
Abler, D. (2004): Multifunctionality, Agricultural Policy,
and Environmental Policy. Agriculture and Resource
Economics Review 33(1): 8-18.
Geografisk Tidsskrift-Danish Journal of Geography 109(2) 143
Anderson, K. (2000): Agriculture’s multifunctionality and
the WTO. Australian Journal of Agricultural and Resource Economics 44(3): 475-494.
ART (2006): Maschinenkosten 2007. ART-Bericht 664.
Tänikon, Forschungsanstalt Agroscope ReckenholzTänikon (ART).
BLW (2004): Swiss Agricultural Policy: Objectives, tools,
prospects. Bern, Swiss Federal Office for Agriculture.
Boisvert, R.N. (2001): The Production Relationship Underlying Multifunctionality. Pp. 27-57 in: OECD (ed.):
Multifunctionality: Towards an analytical framework.
Paris, OECD Publications.
Bromley, D. (2000): Can Agriculture Become an Environmental Asset? World Economics 1(3): 127-139.
Brouwer, F. (2004): Sustaining Agriculture and the Rural
Environment. Governance, Policy and Multifunctionality. Advances in Ecological Economics. Cheltenham,
UK, Edward Elgar Publishing Ldt.
Buysse, J., Van Huylenbroeck, G. & Lauwers, L. (2007):
Normative, positive and econometric mathematical
programming as tools for incorporation of multifunctionality in agricultural policy modelling. Agriculture,
Ecosystems & Environment 120(1): 70-81.
Dibden, J., Potter, C. & Cocklin, C. (2009): Contesting
the neoliberal project for agriculture: Productivist and
multifunctional trajectories in the European Union and
Australia. Journal of Rural Studies 25(3): 299-308.
El Benni, N. & Lehmann, B. (2010): Swiss agricultural
policy reform: Landscape changes in consequence of
national agricultural policy and international competition pressure. Pp. Chapter 5 in: Primdahl, J. & Swaffield,
S. (eds.): Globalisation and Agricultural Landscapes.
Change Patterns and Policy trends in Developed Countries. Cambridge, Cambridge University Press.
FAO (2009): FAOSTAT Statistics [on-line]. Food and Agriculture Organization of the United Nations. Available
from: http://faostat.fao.org/site/339/default.aspx [Accessed April 2009].
Ferrari, S. & Rambonilaza, M. (2008): Agricultural multifunctionality promoting policies and the safeguarding of
rural landscapes: How to evaluate the link? Landscape
Research 33(3): 297-309.
Happe, K. & Balmann, A. (2006): Survival without subsidies. Agrifuture 2/06: 14-16.
Happe, K. & Balmann, A. (2007): Does it also work without subsidies? The possible consequences for European
agriculture [on-line] Pp. 25-30. IAMO. Available from:
http://www.iamo.de/dok/iamo2007_en.pdf [Accessed
March 2009].
144 Geografisk Tidsskrift-Danish Journal of Geography 109(2)
Hartmann, M., Huber, R. & Peter, S. (2009): Strategies
to mitigate greenhouse gas and nitrogen emissions in
Swiss agriculture. IED Working Paper, Agri-food and
Agri-environmental Economics Group, ETH Zürich.
Hodge, I. (2008): To What Extent are Environmental Externalities a Joint product of Agriculture? Overview
and Policy Implications. Pp. 85-118 in OECD (ed.):
Multifunctionality in Agriculture: Evaluating the Degree of Jointness, Policy Implications. Paris, OECD
Huber, R. (2009a): Economies of scope in the agricultural
provision of ecosystem services: An application to a
high cost production region. Agrarwirtschaft (accepted).
Huber, R. (2009b): Valuation of agricultural land-use scenarios with choice experiments: a political market share
approach. Contributed paper. 1st International conference on landscape economics. Vienna July 2-4.
Janssen, S. & van Ittersum, M.K. (2007): Assessing farm
innovations and responses to policies: A review of bioeconomic farm models. Agricultural Systems 94(3):
Joerin, R., Campo, I.S., Maier, T. & Flury, C. (2006): Market liberalization and the role of direct payments in Switzerland. Paper presented at the Annual meeting of Japanese Association of Regional Agriculture and Forestry
Economics [on-line]. Osaka, St. Andrew’s University.
Available from: http://www.afee.ethz.ch/people/Associated/joerin/Publications/Swiss_Report_Oct_13_2006.
pdf [Accessed April 2009].
Junge, X., Jacot, K.A., Bosshard, A. & Lindemann-Matthies, P. (2009): Swiss people’s attitudes towards field
margins for biodiversity conservation. Journal for Nature Conservation 17(3): 150-159.
KTBL (2006): Betriebsplanung Landwirtschaft 06/07.
KTBL Datensammlung. Darmstadt, Kuratorium für
Technik und Bauwesen in der Landwirtschaft e.V.
Lips, M. & Gazzarin, C. (2008): What are the preferences
of Dairy Farmers regarding their Work? A Discrete
Choice Experiment in the Eastern Part of Switzerland. Contributed Paper 44132 [on-line]. Ghent, XIIth
Congress of the European Association of Agricultural
Economists (EAAE). Available from: http://purl.umn.
edu/44132 [Accessed April 2009].
Mander, Ü., Wiggering, H. & Helming, K. (2007): Multifunctional Land Use. Meeting Future Demands for
Landscape Goods and Services. Berlin Heidelberg,
Mann, S. (2008): Degrees of Jointness for Food Security
and Agriculture. Pp. 159-170 in OECD (ed.): Multifunctionality in Agriculture: Evaluating the Degree of Jointness, Policy Implications. Paris, OECD Publications.
Mosoti, V. & Gobena, A. (2007): International trade rules
and the agriculture sector. Selected implementation issues. FAO Legislative Study. Rome, FAO, Communication Division.
OECD (2001): Multifunctionality towards an analytical
framework. Paris, OECD Publications.
OECD (2003): Multifunctionality: The Policy Implications.
Paris, OECD Publications.
OECD (2007): Agricultural Policies in OECD Countries:
Monitoring and Evaluation 2007. Chapter 13, Switzerland. OECD, Paris.
Oñate, J.J., Andersen, E., Peco, B. & Primdahl, J. (2000):
Agri-environmental schemes and the European agricultural landscapes: the role of indicators as valuing tools
for evaluation. Landscape Ecology 15(3): 271-280.
Peter, S. (2008): Modellierung agrarökologischer Fragestellungen unter Berücksichtigung struktureller Veränderungen in der Schweizer Landwirtschaft. PhD Thesis,
Agri-food and Agri-environmental Economics Group.
Zurich, ETH.
Potter, C. & Burney, J. (2002): Agricultural multifunctionality in the WTO--legitimate non-trade concern or
disguised protectionism? Journal of Rural Studies 18(1):
Potter, C. & Tilzey, M. (2007): Agricultural multifunctionality, environmental sustainability and the WTO:
Resistance or accommodation to the neoliberal project
for agriculture? Geoforum 38(6): 1290-1303.
Primdahl, J. (2010): Globalisation and the local agricultural
landscape: current change patterns and public policy interventions. Pp. Chapter 8 in: Primdahl, J. & Swaffield,
S. (eds.): Globalisation and Agricultural Landscapes.
Change Patterns and Policy trends in Developed Countries. Cambridge, Cambridge University Press.
Primdahl, J., Peco, B., Schramek, J., Andersen, E. & Oñate,
J.J. (2003): Environmental effects of agri-environmental
schemes in Western Europe. Journal of Environmental
Management 67(2): 129-138.
Ricardo, D. (1821): On the Principles of Political Economy
and Taxation. London: John Murray. Library of Economics and Liberty [on-line]. Available from: http://
www.econlib.org/library/Ricardo/ricP2a.html [Accessed April 2009].
Rossing, W.A.H., Zander, P., Josien, E., Groot, J.C.J.,
Meyer, B.C. & Knierim, A. (2007): Integrative model-
ling approaches for analysis of impact of multifunctional agriculture: A review for France, Germany and
The Netherlands. Agriculture, Ecosystems & Environment 120(1): 41-57.
Schmitt, M., Schläpfer, F. & Roschewitz, A. (2005): Bewertung von Landschaftsveränderungen im Schweizer
Mittelland aus Sicht der Bevölkerung eine Anwendung
der Choice-Experiment-Methode. Birmensdorf, Eidgenössische Forschungsanstalt für Wald Schnee und
Landschaft WSL.
Schüpbach, B., Junge, X., Briegel, R., Lindemann-Matties,
P. & Walter, T. (2009): Ästhetische Bewertung landwirtschaftlicher Kulturen durch die Bevölkerung. ARTSchriftenreihe 10. Reckenholz.
Tscharntke, T., Klein, A.M., Kruess, A., Steffan-Dewenter,
I. & Thies, C. (2005): Landscape perspectives on agricultural intensification and biodiversity – ecosystem
service management. Ecology Letters 8(8): 857-874.
Tutkun, A., Haller, T. & Lehmann, B. (2007): Ungebrochene
Befürwortung einer produzierenden Landwirtschaft
– sofern sie tier- und umweltgerecht ist. Trendbericht
UNIVOX Teil III A Landwirtschaft 2006. Institute for
Environmental Decisions, Zürich, ETH.
Vanslembrouck, I. & Van Huylenbroeck, G. (2005): Landscape Amenities. Economic Assessment of Agricultural
Landscapes. Landscape Series. Dordrecht, SpringerVerlag.
Vogel, S., Lanz, S., Barth, L. & Böbner, C. (2008): Ziele für
eine multifunktionale Landwirtschaft. Agrarforschung
15(8): 390-395.
Wiborg, T., McCarl, B.A., Rasmussen, S. & Schneider,
U.A. (2005): Aggregation and Calibration of Agricultural Sector Models Through Crop Mix Restrictions
and Marginal Profit Adjustments. Paper presented at the
XIth EAAE Congress, Copenhagen [on-line]. Available
from: http://purl.umn.edu/24567 [Accessed April 2009].
Wilson, G.A. (2007): Multifunctional agriculture a transition theory perspective. Wallingford, CABI.
Geografisk Tidsskrift-Danish Journal of Geography 109(2) 145