Full text of DOI:10.1080/090647100750014385

This article was downloaded by: [FU Berlin]
On: 31 October 2014, At: 02:53
Publisher: Taylor & Francis
Informa Ltd Registered in England and Wales Registered Number: 1072954 Registered office: Mortimer House,
37-41 Mortimer Street, London W1T 3JH, UK
Acta Agriculturae Scandinavica, Section B — Soil &
Plant Science
Publication details, including instructions for authors and subscription information:
http://www.tandfonline.com/loi/sagb20
Farm-related Factors Influencing the Adoption of Short-
rotation Willow Coppice Production Among Swedish
Farmers
Anders Roos , Håkan Rosenqvist , Erik Ling & Bo Hektor
Published online: 05 Nov 2010.
To cite this article: Anders Roos , Håkan Rosenqvist , Erik Ling & Bo Hektor (2000) Farm-related Factors Influencing the
Adoption of Short-rotation Willow Coppice Production Among Swedish Farmers, Acta Agriculturae Scandinavica, Section B —
Soil & Plant Science, 50:1, 28-34, DOI: 10.1080/090647100750014385
To link to this article: http://dx.doi.org/10.1080/090647100750014385
PLEASE SCROLL DOWN FOR ARTICLE
Taylor & Francis makes every effort to ensure the accuracy of all the information (the “Content”) contained
in the publications on our platform. However, Taylor & Francis, our agents, and our licensors make no
representations or warranties whatsoever as to the accuracy, completeness, or suitability for any purpose of the
Content. Any opinions and views expressed in this publication are the opinions and views of the authors, and
are not the views of or endorsed by Taylor & Francis. The accuracy of the Content should not be relied upon and
should be independently verified with primary sources of information. Taylor and Francis shall not be liable for
any losses, actions, claims, proceedings, demands, costs, expenses, damages, and other liabilities whatsoever
or howsoever caused arising directly or indirectly in connection with, in relation to or arising out of the use of
the Content.
This article may be used for research, teaching, and private study purposes. Any substantial or systematic
reproduction, redistribution, reselling, loan, sub-licensing, systematic supply, or distribution in any
form to anyone is expressly forbidden. Terms & Conditions of access and use can be found at http://
www.tandfonline.com/page/terms-and-conditions

Acta Agric. Scand., Sect. B, Soil and Plant Sci. 2000: 50, 28–34
Printed in Ireland. All rights reser6ed
ACTA AGRICULTURÆ
SCANDINAVICA
ISSN 0906–4710
Farm-related Factors Influencing the
Adoption of Short-rotation Willow
Coppice Production Among Swedish
Farmers
Roos, A., Rosenqvist, H., Ling, E. and Hektor, B. (Department of
Forest Management and Products, Swedish University of Agricultural
Sciences, P.O. Box 7060, SE-750 07 Uppsala, Sweden). Farm-related
factors influencing the adoption of short-rotation willow coppice produc-
tion among Swedish farmers. Accepted February 24, 2000. Acta Agric.
Scand., Sect. B, Soil and Plant Sci. 50: 28–34, 2000.
Anders Roos*,
H a˚ kan Rosenqvist, Erik Ling
and Bo Hektor
Department of Forest Management
and Products, Swedish University of
Agricultural Sciences, P.O. Box 7060,
SE-750 07 Uppsala, Sweden
©
2000 Taylor & Francis.
The area of short-rotation willow coppice energy plantations on
Swedish farmland increased quickly in the early 1990s because of
subsidies for energy crop production, an increased CO tax on fossil
2
fuel and an already existing biofuel market in the country. In this
study farm-related determinants for the adoption of short-rotation
willow coppice production among Swedish farmers are identified and
estimated. A Tobit model is applied to cross-sectional data on
Swedish farmers in 1995. The results show that the decision to plant
willow and the areas planted depend positively on arable land area,
forest land area, the area of other land types, leasing out of arable
land and tractor ownership. Negative factors are pasture area, ten-
ancy and animal production. Differences in willow growing between
ownership types, age groups and geographical regions are also impor-
tant. The policy implications of the results are discussed.
Key words: adoption, bioenergy,
energy crops, Tobit model.
Introduction
ropean Union have set targets for raising bioenergy
production (European Commission, 1997).
Short-rotation energy crop production is considered
in many countries as one possible means of living up
to the commitments in the 1997 Kyoto Protocol for
countering global climate change. Energy crops could
also become alternative non-food cash crops for
farmers across Europe if support for foodcrop pro-
duction were reduced. According to the White Paper
by the European Commission, the share of bioenergy
of the total energy production of the European
Energy crops are also being investigated in the
USA, where 8–16 Mha of arable land can be used for
biomass production in the future (Hohenstein &
Wright, 1994). The Energy Information Administra-
tion of the Department of Energy in the USA expects
an annual growth of 0.8% for renewables until 2020,
where biomass from energy crops and wood accounts
for more than half the energy production. Several
programmes support private bioenergy investments
(
3
EU15) countries has the potential to increase from
% to 8.5% by 2010. Half of this quantity would
derive from energy crops. Most countries in the Eu-
(
EIA, 1998).
Commercial willow plantations in Sweden ex-
panded in the early 1990s, for several reasons. First,
the implementation phase of a deregulated food pol-
icy between 1991 and 1996 implied subsidies for
*
Corresponding author.
2
8

Short-rotation willow coppice production
−
1
planting willows amounting to 1200 ECU ha
and
for fencing (Proposition 1989/90,
46). Secondly, taxes on sulphur and CO for fossil
and fertilizers (Mbata, 1997). Limited dependent-
variable models have often been used in these studies.
The inspection of the adoption literature led to the
conclusion that there is a proven methodology avail-
able for analysing the adoption by Swedish farmers
of short-rotation willow production.
−
1
4
1
80 ECU ha
2
fuels in heat production increased considerably in
991, improving the competitiveness of biofuels in the
energy market (Larsson Rosenqvist, 1997).
1
&
Thirdly, a large biofuel market already exists in Swe-
den, based on the demand from the district heating
plants of wood fuel from the forest sector. The fourth
positive factor for the expansion of commercial wil-
low plantations is the extensive research and plant
breeding programmes of willow since the 1970s
The aim of this study was to identify the farm-re-
lated factors that influence the adoption of willow
production by Swedish farmers. The study was geo-
graphically restricted to farm enterprises in southern
and central Sweden (G o¨ taland and Svealand) that
were growing willow in 1995 (see map in Fig. 1).
Northern Sweden is excluded from the study since a
harsh climate for willow production makes dedicated
energy plantations less suitable in this part of the
country.
(
Christersson & Senneby-Forsse, 1994).
Since the two issues of how to increase the share of
renewable energy and how to reduce surplus food
production are relevant in many countries, realistic
predictions about prospective energy-crop growers
are needed. Better knowledge about the adoption
patterns of willow production in Sweden during the
early 1990s could help policy makers and decision
makers in Sweden, and in other countries where
short-rotation energy forests are discussed, to design
effective energy and agricultural policies and identify
clear-cut target groups for information campaigns.
National estimates of land availability and of the
technical and economic status of short-rotation en-
ergy crop production are found in several studies for
the USA (e.g. Wright & Hohenstein, 1994) and for
Sweden (SOU, 1992). Several studies in the USA and
in Europe use Geographical Information Systems
Materials and methods
Theoretical model
An adapted theoretical model used by Saha et al.
(1994) was applied. It was assumed that the farmer,
or farm enterprise, can choose to combine willow
cultivation with alternative profit-maximizing activi-
ties. Let the farm enterprise be described by a profit
function that includes the profits of both willow
growing and other activities on the farm:
e
s−
y=by+p yf(x)−c(p, y, x)+y (p, A−y, x)
(1)
(
GIS) to describe possible spatial aspects of future
bioenergy systems (e.g. Liu et al., 1993; Dagnell,
995; Downing & Graham, 1996). Most studies to
Total profits, p, depend on: per hectare subsidies
discounted to present value, b, times the willow area,
y; the present value of woodfuel prices, p , multiplied
1
e
date have, however, been unable to include knowl-
edge about adoption patterns for energy crops, spe-
cifically among the farmers, simply because such
empirical data do not exist.
Studies of adoption patterns by farmers are numer-
ous, ranging from analyses of tillage practices (Gould
et al., 1989; Lexmon & Andersson, 1998) to innova-
tions (Saha et al., 1994) and the use of crop varieties
by the area under willow production, y, times the
wood fuel produced on 1 ha of willow land, f(.),
which is a function of a vector of farm (and farmer)
attributes, x; discounted costs for planting, managing
and harvesting the willow plantations, c(.), which is a
function of discounted netput prices, p, willow area,
y, and farm attributes, x; and finally, on the profits
s−
from non-willow production on the farm, y . The
latter expression is a function of the discounted price
vector for outputs and inputs, p, total crop area
minus the willow area, A−y, and attributes of the
farm, x. The first-order condition for profit maxi-
mization with regard to y will then be:
e
s−
y
b+p f(x)−c (p, y, x)+y (p, A−y, x)=0
(2)
y
The expression describes how the marginal benefits
from an increasing willow production consist of rev-
enues from subsidies and present values of future
willow woodchip sales. One cost is production costs
for planning, planting, management and harvesting,
where c (.)\0. Another cost is the ‘‘cost’’ for non-
willow profits foregone, where yy B0.
Fig. 1. Area of the study showing the regions. NE, North-east;
NW, north-west; SE, south-east; SW, south-west; S, south.
y
s−
2
9

A. Roos et al.
The optimal choice of willow area can then be
described as:
area, conditional on being over the ‘‘limit’’, i.e. that
more than 0 ha of willows are being grown.
Differentiations of the estimated results will allow
ꢀ
ꢀ
e
y =y (b, p , p, x, A)
(3)
us to obtain for different values of x : (1) the ex-
i
The optimal solution of willow cultivation is a
function of subsidies offered, woodfuel prices, prices
for inputs and non-biofuel outputs, farm and farmer
attributes, and total arable land area on the farm.
This study focuses on how size and farm characteris-
tics influence the willingness to grow willow and the
area planted. The effects of different economic fac-
tors, e.g. prices, subsidies and other macroeconomic
factors, could not be analysed since the planting year
was not known.
The above analysis assumes interior solutions to
the optimization problem, but it is clear that the
non-interior solution y*=0 is feasible, i.e. that no
willows are grown. Theoretically, the opposite solu-
tion is also possible, i.e. that 100% of the arable land
is used for willow plantations. This situation is, how-
ever, rare and not a primary issue for this analysis.
pected energy willow area, (2) the willow area condi-
tional to being above the limit, (3) and the
probability of being above the limit (McDonald &
Moffitt, 1980).
The model implies the assumption that the deci-
sions to grow willow and how many hectares to plant
are made simultaneously.
Data
Based on the general conclusions of previous adop-
tion studies (Saha et al., 1994; Mbata, 1997; Lexmon
&
Andersson, 1998), the variables in the present
model describe the farmer, the farm, farm size, land
use, production and location.
Data on all 1158 farms in southern Sweden with a
farming area of more than 2 ha growing more than
0
.1 ha of willows were obtained from the 1995 Farm
Register (FR) compiled by Statistics Sweden (1995).
For comparison, a stratified sample from 1995 of 535
non-willow farmers from the same geographical re-
gion was also used.
The definitions of the variables are presented in
Table 1.
The Tobit model
The great majority of Swedish farmers does not grow
willows. In technical terms this means that the depen-
dent variable in this study, willow area on the farm,
is left-censored at zero since the value cannot be
negative. The Tobit model is a common method for
analysing censored samples of this type. It has been
used in several agricultural adoption studies (e.g.
Gould et al., 1989; Goodwin & Schroeder, 1994;
Mbata, 1997).
It is reasonable for two reasons to assume that the
independent variables included in the model are ex-
ogenous. First, the decision to grow willows was
made after 1990, generally after decisions about other
traditional farm production orientations had already
been made. Secondly, the areas of arable land, pas-
ture, forests and other uses depend on fundamental
conditions which are likely to change very little over
the space of a few years. However, variables describ-
ing uses of arable land other than for growing wil-
lows have not been included owing to potential
endogenity problems that would otherwise yield bi-
ased results.
A positive sign for the untransformed arable land
area variable is expected, simply because the range of
different crops and varieties is expected to be wider
on a large farm than on a small farm. Few additional
hypotheses can be formulated about the sign of the
model coefficients.
The Tobit model (Tobin, 1958), adapted for analy-
sis of the factors influencing willow growing, can be
defined as:
y =ix +m if y \0
i
i
i
i
yi=0 otherwise
(4)
where y is the willow plantation area on farm i, x is
i
i
the vector of explanatory variables, b is a coefficient
vector and o is the error term.
i
Maximum-likelihood procedures are available to
produce consistent and asymptotically normal
parameter estimates for Tobit models (Amemiya,
1973).
2
If the error terms are independent and N(0, s ),
the expected woodfuel use can be written as:
Results and discussion
E[y ]=F(z )E(y ꢀy \0)
i
i
i
i
The estimation results are presented in Table 2. To
avoid the ‘‘dummy-variable trap’’ of perfect collinear-
ity, the dummies representing zero areas of forest
land, pasture land and other land were removed, as
were the variables for owner ages 35–50 years and
the geographical dummy symbolizing farms in the
ixi
where z = |
(5)
i
In Eqn (5), F(z ) denotes the probability that farm i is
growing willows for biofuels, and the second term on
the right-hand side describes the expected willow
i
3
0

Short-rotation willow coppice production
Table 1. Definitions of variables
1% level, one at the 5% level and two at the 10%
level. Seven coefficients were insignificant.
Variable
Definition
As expected, the sign of the arable land area coeffi-
cient was positive, although the marginal influence of
the variable decreased, as indicated by the negative
sign of the squared arable-land coefficient. The influ-
ence of forest area on willow growing was equally
positive, suggesting that forest owners have a better
general knowledge than other farmers about growing
long-rotation varieties, such as trees. They may even
already have experience of selling wood fuel in the
form of logging residues. For pasture areas of more
than 7 ha a negative relationship with willow growing
was detected. This may be due to the fact that
farmers with animals usually try to keep land avail-
able for both pasture and fodder production, which is
not compatible with willow growing. Medium-sized
Willows
Arable
land
Arable
land2
For0
For1
For2
For3
Past0
Past1
Past2
Oth0
Willow area (ha)
Arable land area (ha)
Arable land area squared
1 if forest area=0, 0 otherwise
1 if forest area 0–25 ha, 0 otherwise
1 if forest area 25–80 ha, 0 otherwise
1 if forest area 80 ha–, 0 otherwise
1 if pasture area=0, 0 otherwise
1 if pasture area 0–7 ha, 0 otherwise
1 if pasture area 7 ha–, 0 otherwise
1 if other land area=0, 0 otherwise
1 if other land area 0–7 ha, 0
otherwise
Oth1
Oth2
Leasout
1 if other land area 7 ha–, 0 otherwise
1 if arable land is leased out; 0
otherwise
Table 2. Tobit estimations
Tenancy 1 if arable land is leased; 0 otherwise
Inst/Ltd
1 if owner is a limited company or
institution; 0 otherwise (if the owner is
a person)
1 if owner age B35 years; 0 otherwise
1 if owner age 35–50 years; 0 otherwise
1 if owner age 50–65 years; 0 otherwise
1 if owner age E65 years; 0 otherwise
1 if horses; 0 otherwise
1 if milk cows; 0 otherwise
1 if other cattle; 0 otherwise
1 if pigs; 0 otherwise
1 if sheep; 0 otherwise
Number of tractors
Variable
Coefficient
SD
P-value
Intercept
Arable land
Arable land2
For1
For2
For3
Past1
Past2
Oth1
Oth2
Leasout
Tenancy
Inst/Ltd
–35 y
50–65 y
65 y–
Horses
Cows
Cattle
Pigs
Sheep
Harvester
Tractors
Irrigation
Ecological
Hours
NE
−700.55***
0.0235***
−2.15E-7***
2.8673
98.5222*** 12.6762
106.661***
−6.8463
−68.8454*** 11.8247
34.0404***
−3.3320
32.4583*** 12.4959
−33.3419***
26.4386*
−36.5587*
35.0230***
−80.4911*** 12.6901
−20.9750**
−151.513***
−73.9290*** 10.0900
−35.5978*** 12.4298
−13.0963
−9.9022
17.67
0.0011
1.359E-8 0.0001
11.3542
0.0001
0.0001
B35 y
35–50 y
50–65 y
65 y+
0.8006
0.0001
0.0001
0.46
0.0001
0.0003
0.79
0.0094
0.0001
0.058
Horses
Cows
Cattle
Pigs
Sheep
Tractor
14.7483
9.3379
9.1939
12.2648
Harvester 1 if harvester; 0 otherwise
Irrigation 1 if irrigation; 0 otherwise
Ecological 1 if ecological production; 0 otherwise
Hours
Northeast 1 if in north-east; 0 otherwise
Northwest 1 if in north-west; 0 otherwise
Southeast 1 if in south-east; 0 otherwise
West
South
8.6504
13.9672
20.8936
9.4312
0.080
Estimated work hours
0.0002
0.0001
0.036
0.0001
0.0001
0.0042
0.230
9.9912
17.3181
1 if in south-west; 0 otherwise
1 if in south; 0 otherwise
11.6193
9.7626
0.2608
0.310
southernmost part of the area. Likelihood ratio (LR)
tests were conducted to check the relevance of the
variables symbolizing owner class (institutional and
limited owners)/age groups, geographical situation.
The whole model in Table 2 was also tested against
the alternative with all coefficients, except the inter-
cept, restricted at zero. All LR tests described above
provide support for the model presented in Table 2.
The McFadden pseudo-R statistic (Maddala, 1983),
0.8229***
53.6132*** 13.7265
0.0016
0.0001
0.467
13.0593
0.000044
134.136***
−61.1990*** 17.8603
−59.0333*** 15.6666
−67.9900*** 13.6976
17.9349
0.000218 0.841
12.8786
0.0001
0.0006
0.0002
0.0001
NW
SE
W
s
278.0836
6.8881
0
.110, is not notably different from what has been
Estimated mean values for untransformed variables.
Base case: No forest, pasture or other land, age 35–50
y, South.
recorded in similar investigations.
The model results include coefficients for 29 inde-
pendent variables, 19 of which were significant at the
Significant values: *PB0.1; **PB0.05; ***PB0.01.
3
1

A. Roos et al.
areas of other land types (e.g. housing land, unpro-
ductive land) were positively correlated with willow
cultivation, whereas the same relationship was nega-
tive and insignificant for larger areas of other land
types. Leasing out of land was positively associated
with willow growing. The explanation may be that
both land uses are complementary reactions to a
situation of surplus land ownership relative to other
production factors. Tenancy had a negative relation-
ship with willows for just the opposite reason: that
land is primarily needed for other, less land-intensive
production than willow growing.
The results show that the institutional/limited own-
ership category, together with the age span of 50–65
years, was the most adoption-prone owner group,
even when other variables had been accounted for.
Institutional owners may be interested in willow pro-
duction because they have adopted a more aggressive
management style. The fact that the work on these
farms is often carried out by employees may ease the
move towards more extensive crop production (wil-
lows) since the owner more easily than the family
farmer can compensate for the reduced input of
labour by reducing the number of employees. The
low willow-planting activity in the oldest owner age
group was similar to the observations of other adop-
tion studies (e.g. Mbata, 1997; Lexmon & Andersson,
positive attitude towards new technologies in general.
It would be interesting to measure the effect of total
work hours by family members and employees on
willow growing on the farm. This variable was, how-
ever, too unreliable and was therefore not included in
the analysis.
The differences in willow-growing intensity be-
tween the different geographical regions probably
reflect variations in several aspects, the biofuel mar-
ket, soils, infrastructure and climate. In the north-
eastern region of the studied area, a high population
density and large demand for biomass fuels are com-
bined with a well-developed woodfuel sector based on
residues from the forest sector and a high woodfuel
consumption in the surrounding district heating
plants. This may have induced farmers to have more
confidence in a future market for willow chips. The
north-eastern region also includes large areas of
medium-quality agricultural land which is probably
suitable for willows.
In the western parts of the investigated area, a less
developed woodfuel market has probably reduced the
interest in growing willows. A high supply of wood
fuels from the forest sector has a similar effect in the
south-east. Farmers in these areas are probably less
confident about the future demand for wood chips
from willow coppice plantations. Low rainfall during
the growth season and coarse-textured soils in this
region could also have reduced willow production. In
the extreme south however, a somewhat higher wil-
low-planting activity could again be due to the high
population density and the potential for a high wood-
fuel demand. Willow growing in the north-east and
the south could finally be due partly to a good
infrastructure in these areas of advisors and special-
ized companies who can manage the willow planta-
tions. The role of local advisors has been shown
previously by Ling (1996).
1998). The negative sign for the coefficient represent-
ing owner ages under 35 years was more surprising,
as earlier adoption studies found that young farmers
are more eager to adopt new methods than older
farmers. This result could reflect a situation where
limited financial resources among the youngest group
makes it risk averse. In the intermediate age groups,
however, both a stable economic situation and a
positive attitude towards new crops may contribute
to a high adoption rate.
All variables symbolizing animal production had
negative correlations with willow production. Crop-
producing farms may have more easily selected par-
cels that qualified to receive subsidies for
willow-planting activities. Furthermore, on farms
that concentrate on animal or milk production, land
is needed for fodder production and manure deposi-
tion. Although willows are being tested by re-
searchers as vegetation filters for sewage plants, few
applicable and cost-effective methods exist today for
applying manure on land under willow cultivation.
The positive sign of the irrigation coefficient might
be confusing. The use of irrigation might be a sign of
low rainfall and coarse-textured soils, which are less
suitable for willow. However, our results suggest that
the relative suitability of willow on these soils when
compared with grain crops is still good. Furthermore,
irrigation and tractor ownership could also be associ-
ated with an active management style with a more
An illustration of the effect of age, geographical
location, milk cows and forest land on expected
willow area, given that the farmer is a willow grower,
and on the probability of adopting willow cultivation,
is shown in Figs 2 and 3.
An increased arable land area, from 30 to 150 ha,
leads to a higher probability of willow production
(Fig 2) and to willow production on a greater area
(Fig 3). Figs 2 and 3 also show that changes in the
independent variables have a stronger effect on the
probability of adopting willow than on the area of
willow cultivated.
Conclusions
This study shows that several variables describing the
farmer, the farm type, land use and geographical
3
2

Short-rotation willow coppice production
Fig. 2. Probability of growing willows. NE, North-east; NW, north-west; SE, south-east; SW, south-west; S, south. Base case: no forest,
pasture or other land, age group 35–50, S. Dark bars: 150 ha; lighter bars: 30 ha.
Fig. 3. Expected willow area for willow growers. NE, north-east; NW, north-west; SE, south-east; SW, south-west; S, south. Base case: no
forest, pasture or other land, age group 35–50, S. Dark bars: 150 ha; lighter bars: 30 ha.
3
3

A. Roos et al.
situation influence the decisions of Swedish farmers
to adopt willow cultivation.
Renewable Energy Laboratory, Golden CO. NREL/CP-200-
098, pp. 140–150.
Downing, M. & Graham, R. L. 1996. The potential supply and
cost of biomass energy crops in the Tennessee Valley Authority
Region. Biomass Bioenergy 11, 283–303.
EIA, 1998. Annual Energy Outlook 1999 With Projections to 2020.
Energy Information Administration, Dept of Energy, Washing-
ton DC.
European Commission 1997. Energy for the Future: Renewable
Sources of Energy. White Paper for a Community Strategy and
Action Plan, Communication from the Commission
COM(97)599, Brussels.
Goodwin, B. K. & Schroeder, T. C. 1994. Human capital, pro-
ducer education programs, and the adoption of forward-pricing
methods. Am. J. Agric. Econ. 76, 936–947.
Gould, B. W., Saupe, W. E. & Klemme, R. M. 1989. Conservation
tillage: the role of farm and operator characteristics and the
perception of soil erosion. Land Econ. 65, 167–182.
Hohenstein, W. G. & Wright, L. L. 1994. Biomass energy produc-
tion in the United States: an overview. Biomass Bioenergy 6,
8
The results of the study have policy implication
both for Sweden and for other countries that are
planning to increase energy crop production. Adop-
tion patterns of energy crops can be predicted. This is
particularly true for the introduction phase. The re-
sults also help policy makers and players on the
market to focus on the right farmers in information
and marketing efforts. Furthermore, if there is a
policy goal to increase energy crop adoption among
farmers with low adoption rates, e.g. small-scale
farmers, the results suggest that the barriers to adop-
tion first must be identified. Tailor-made incentive
programmes will probably have to be introduced in
such cases. The results can also be useful for assessing
biofuel production in different regions and for locat-
ing conversion plants based on energy coppice. A
heating plant based on willow may, for instance, be
more appropriate in a region with many large farms
than in a region with small-scale diary farms. Caution
should, however, accompany every translation of
these results to other countries. It is still likely that
the adoption pattern of energy crops in most regions
and countries will vary between different types of
farm.
1
61–173.
Larsson, S. & Rosenqvist, H. 1997. Willow production in Swe-
den—politics, cropping development and economy. Paper pre-
sented at the European Energy Crops Conference 30
September–1 October 1996, Enschede, The Netherlands. Orga-
nized by BTG Biomass Technology Group BV.
,
Lexmon, A. & Andersson, H. 1998. Adoption of minimum tillage
practices. Swed. J. Agric. Res. 28, 29–38.
Ling, E. 1996. Integration av skogsbr a¨ nslesortimentet p a˚ distrikt-
sniv a˚ —en n a¨ tverksstudie av tv a˚ lokala skogsbruksomr a˚ den i
s o¨ dra Sverige. In: Bioenergins nuvarande och framtida konkur-
renskraft—tre djupstudier inom f o¨ r a¨ dlingskedjan, Frankel, Ling
Future studies could investigate further the energy
crop adoption process, especially concerning the im-
portance of locational aspects (e.g. soil type, local
biofuel market volume). A technique to produce re-
gional biofuel supply projections based on empirical
data could also be developed.
&
Lundgren, IIIEE Research Report 96:1. Lund University. (In
Swedish.)
Liu, W., Merriam, R. A., Phillips, V. D. & Singh, D. 1993.
Estimating short-rotation Eucalyptus saligna production in
Hawaii: an integrated yields and economic model. Biores. Tech-
nol. 45, 167–176.
Maddala, G. S. 1983. Limited-Dependent Variables in Economet-
rics. Econometric Society Monographs 3. Cambridge University
Press, Cambridge.
Mbata, J. N. 1997. Factors influencing fertilizer adoption and rates
of use among small-scale food crop farmers in the rift valley area
of Kenya. Q. J. Int. Agric. 36, 285–302.
Acknowledgements
The authors received helpful comments from two
referees. The project was financed by the Swedish
Environmental Protection Agency.
McDonald, J. & Moffitt, R. 1980. The uses of Tobit analysis. Rev.
Econ. Statist. 62, 318–321.
Saha, A., Love, H. A. & Schwart, R. 1994. Adoption of Emerging
technologies under output uncertainty. Am. J. Agric. Econ. 76,
References
8
36–846.
Amemiya, T. 1973. Regression analysis when the dependent vari-
able is truncated normal. Econometrica 41, 997–1016.
SOU, 1992. Bioenergikommissionens slutbet a¨ nkande. SOU
1992:90. Allm a¨ nna F o¨ rlaget, Stockholm. (In Swedish.)
Christersson, L. & Senneby-Forsse, L. 1994. The Swedish pro-
gramme for intensive short-rotation forests. Biomass Bioenergy
Statistics Sweden 1995. Information om Lantbruksregistret 1995
8
(Information about the farm register). Orebro. (In Swedish.)
Tobin, J. 1958. Estimation of relationships for limited dependent
variables. Econometrica 26, 24–36.
Wright, L. L. & Hohenstein, W. G. (eds) 1994. Dedicated feed-
stock supply systems: their current status in the U.S.A. Biomass
Bioenergy 6, No. 3 (Special issue), 159–241.
6
, 145–149.
Dagnell, S. A. P. 1995. Resource mapping and analysis of UK
farm livestock manures. Proceedings of the Second Biomass
Conference of the Americas: Energy, Environment, Agriculture,
and Industry, Aug. 21–Aug. 24, 1995. Portland, OR. National
3
4