Author’s Accepted Manuscript
Sheep production systems in semi-arid zone:Changes and simulated bio-economic performancesin a case study in Central Chile
Paula Toro-Mujica, Claudio Aguilar, Raúl Vera,José Rivas, Antón García
PII: S1871-1413(15)00317-0DOI: http://dx.doi.org/10.1016/j.livsci.2015.07.001Reference: LIVSCI2785
To appear in: Livestock Science
Received date: 7 November 2014Revised date: 22 June 2015Accepted date: 2 July 2015
Cite this article as: Paula Toro-Mujica, Claudio Aguilar, Raúl Vera, José Rivasand Antón García, Sheep production systems in semi-arid zone: Changes andsimulated bio-economic performances in a case study in Central Chile, LivestockScience, http://dx.doi.org/10.1016/j.livsci.2015.07.001
This is a PDF file of an unedited manuscript that has been accepted forpublication. As a service to our customers we are providing this early version ofthe manuscript. The manuscript will undergo copyediting, typesetting, andreview of the resulting galley proof before it is published in its final citable form.Please note that during the production process errors may be discovered whichcould affect the content, and all legal disclaimers that apply to the journal pertain.
www.elsevier.com/locate/livsci
1
Sheep production systems in semi-arid zone: Changes and simulated bio-economic
performances in a case study in Central Chile
Paula Toro-Mujicaa*
, Claudio Aguilara, Raúl Vera
a, José Rivas
b and Antón García
b
aDepartamento de Ciencias Animales, Facultad de Agronomía e Ingeniería Forestal, Pontificia Universidad Católica
de Chile. Av. Vicuña Mackenna 4860, Santiago, Chile.
bDepartamento de Producción Animal, Facultad de Veterinaria, Universidad de Córdoba, Campus de Rabanales,
14071 Córdoba, España
ABSTRACT
A sheep farm typology was developed to describe the evolution of sheep farming between the
censuses of 1997 and 2007 in the semiarid zone of Central Chile. The typology yielded three
groups (I to III) that accounted for 81, 17 and 2.5% of the farms respectively, differing in farm
size and in the ratio of sheep to cattle and other agricultural activities. Sheep represented 80-86%
of the livestock units in small farms, as opposed to 53% in the larger, more diversified, ranches.
Farm-based technical and economic parameters were not available. Stochastic mathematical
simulation of bioeconomic performance for prototype farms representative of each of the three
groups showed differences accounted for by farm size, farm diversification and animal breed.
Large between-farms within group variation in performance suggest the existence of room for
incorporation of technology. Smaller flock sizes in I and II were associated with larger
production costs and less income per lamb and per kg live weight. Larger farms carrying Merino
produced more lambs per ewe and had lower unitary costs. Unaccounted for costs of family labor
in the smaller farms, together with some evidence of their gradual decapitalization, explain the
continued existence of the small sheep farm sector. Implications for the future development of
these farms are discussed.
2
Keywords: small farms, land use, typology, simulation modeling
1. INTRODUCTION
The Mediterranean region of Chile extends for about 1,000 km between 30° and 37° S,
ranging from arid areas in the N to humid in the S (Le Houerou, 2004), and covers some of the
most important agricultural regions of the country. Within this area, the central, most populated,
semi-arid portion includes a variety of land uses, and has a long history of agricultural occupation
(Cáceres, 2010). The valleys and flatlands are dedicated to intensive, irrigated horticulture,
viticulture and pomology, whereas hilly lands are typically dominated by a woody savanna
(Moreira-Muñoz, 2011), and are largely in the hands of family farms whose main source of
income is meat and wool sheep raised in extensive low input-low output systems. These dry
lands, degraded by decades of cereal monocropping (Brunel et al., 2011; Kapur and Ersahin,
2014), are covered by low yielding naturalized annual pastures (Ovalle and Squella, 1996) thus
resulting in low animal and per ha performance (Vera et al., 2013). Grassland-based sheep
production systems are of historical, social, environmental and economic importance throughout
the world’s Mediterranean regions (Cosentino et al., 2014). Sheep systems are frequently based
on sub humid to semiarid degraded rangelands that limit physical productivity and are highly
seasonal (Specht et al., 1988), but may also be associated with multifunctionality characteristics
(Hadgigeorgiou et al., 2005), including conservation of the landscape and biodiversity (Baumont
et al., 2014; Casasús et al., 2012; García-Martínez et al., 2011; Winkler, 1999). The forage
resources of Mediterranean environments and the environmental services that they provide have
been the subject of numerous reviews (e.g. Arroyo and Zedler, 1995; Baumont et al., 2014), but
their future is uncertain (de Rancourt et al., 2006).
3
Not surprisingly, between-farm variability is large in Mediterranean regions, compounded by
changes in policies, national and international market opportunities, and variable institutional
support over time (Carmona et al., 2010).
Numerous authors have advanced the view that government policies should ideally consider
this wide variation so as to skew policies towards the weaker sectors (Boyazoglu and Morand-
Fehr, 2001; Boyazoglu et al., 2002; Daskalopoulou and Petrou, 2002), but neither the EU
subsidies (Caballero, 2001) nor the credits and policies available in Chile differentiate between
production systems. In consequence, the importance of sheep production systems in the
Mediterranean continues to decrease (Bernués et al., 2011), a trend replicated in Chile where
sheep farms decreased 17% in the 1997-2007 period (INE, 2007, 1997). Sheep flocks in Central
Chile are mostly based on Suffolk and Merino, and are owned mostly by small and medium
farmers. Reduction in wool prices, particularly for medium wool, have forced farmers to favor
meat production via the gradual introduction of new breeds that give rise to larger, easier to cut
carcasses, and with larger added value. This process further increased between-farms variability,
and modified their production and profitability. In this context the development of farm
typologies constitutes a tool to identify structural features of production systems, generate a
framework within which policies may address the needs of specific farm categories, and identify
farms with a need or potential to adopt new technologies.
Targeting policies, technical assistance, and innovative technologies for fragile environments
can be helped through the identification of farms clusters and recommendation domains as
carried out by typological studies (Barrantes et al., 2009; Cortez-Arriola et al., 2015;
Daskalopoulou and Petrou, 2002; Girard et al., 2008; Madry et al., 2013; Ripoll-Bosch et al.,
2014) together with a variety of information technologies (Blackmore and Apostolate, 2011;
Carmona et al., 2010; Dallies et al., 2009) to support farmer decision making. For example,
4
Meyland et al. (2013) recognizing the ecosystem importance of soils, combined a farm typology
with a conceptual cropping system with the purpose of identifying system-appropriate
alternatives to promote soil conservation. Similarly, Bohnet et al., (2011a) developed a farm
typology of cattle systems in the Bowen-Broken watershed, to understand the relationships
between ranchers, land, and the capacity to incorporate sustainable practices in order to decrease
the rate of sedimentation and nutrient loading of the Great Barrier Reef.
Mathematical modeling techniques that integrate economic and agro ecological information
allow the early assessment of possible policies and technical interventions, due to their ability to
rapidly focus on specific farm types and production systems at low cost, as exemplified among
others by Bohnet et al. (2011b), Buysse et al. (2007), Cacho et al. (1995), Catrileo et al. (2009)
Dalgliesh et al. (2009), Ruben et al. (1998), Toro et al. (2009) and Zimmermann et al. (2009).
Thus, the joint use of these tools allows prioritization of support policies for farm groups
based on technical and economic indicators. They also permit to value environmental services
and societal values, including the generation of opportunities for farm labor (Olaizola et al.,
2015), so that marginal farms may reach profitability levels compatible with their continuing
existence.
The objective of the present paper was to develop a typology of sheep production systems in
the Mediterranean area of Central Chile to identify and quantify changes that have taken place in
the interval between the last two agricultural surveys, in terms of structural, and aggregated
economic and technological variables. Since surveys do not identify technical and economic
parameters for production systems, nor was data available elsewhere, the bio-economic
performance of prototype farms representative of each typological group was simulated to
identify groups of farms marginally profitable, and to assess interventions that would increase
their performance and continued livelihood.
5
2. MATERIALS AND METHODS
2.2. Study area
The study was carried out in the central, semi-arid zone of Chile (33° to 35° 01’ S, 70° 02' W
to the Pacific Ocean). The surface area is 16,387 km2, 29.5% of which is under native forests,
24% in rangelands, 17% croplands and 9% forestry (INE, 2007). The topography is varied, with
altitudes varying between 0 to 5135 masl, and encompasses four agro ecological regions: Coastal
areas, a Central Valley with ocean influences, an interior Valley, and the Andean foothills. The
average yearly rainfall varies between 518 mm in the interior Valley to 580 mm in the foothills
(Santibañez, 1993) but annual rainfall has been decreasing over the last 20 years (IPPC, 2014).
Winter rainfall predominates, and June and July are the wettest months but with a highly erratic
rainfall ranging between 100 and 300 mm per month (Pizarro, 2007).
2.2. Data selection and population
Farms that raise ruminant animals cover 1.6 million ha in the study region, and 18% of them
carry sheep (INE, 2007). Data available in the 1997 and 2007 surveys for sheep farms include
dimensional and social variables, land tenure, and farmer-related characteristics. Survey data
were then processed to obtain derived quantitative and qualitative variables (Ruiz et al., 2008).
The number of farms carrying at least some sheep was 3,466 in 1997, and 2,793 in 2007 (INE,
2007, 1997). Some farms had nominal numbers of sheep and therefore only sheep farms with a
minimum of 5 ha, stocking rates of at least 0.1 sheep LU/ ha and a minimum of 40% of the LU as
sheep were kept for further analysis. Initially, farms were grouped into four classes, according to
the number of sheep animal units in each, as shown in Table 1. Subsequently, noncommercial
farms with less than 5 LU were excluded from the analysis. These farms represented a large
6
proportion of the total, but with a trend towards decreasing absolute numbers. The number of
farms classified as T1 increased over the period 1997-2007, whereas that of T0 decreased,
possibly due to increases in flock size of the latter that allowed them to reclassify as T1 farms in
2007. Excluding category T0 from the study population left a total (T1+T2+T3) of 352 farms for
1997 and 395 for 2007 (Table 1).
To characterize sheep production systems, quantitative variables extracted and/or generated from
the 2007 Census were classified into three categories as follows: dimensional, intensity and
diversification, and social variables. The dimensional variables showed a wide range of variation
in terms of crop area and flock size (Table 2). Variables related to intensification and
diversification were relatively less variable. Sheep numbers as proportion of total livestock, and
the area covered by natural grasslands, showed coefficients of variation less than 50%. Although
with a high coefficient of variation, 95% of the farm population had sheep stocking rates ranging
between 0.25 and 0.32 sheep LU. The persistence of extensive sheep systems in the region is
evidenced by variable, but low, sheep stocking rates, large proportion of native pastures, and low
areas of sown annual and perennial forages. Social variables showed a high percentage of tenured
areas, advanced age of farmers, and low permanent and total labor use, all of which coincide with
the extensive nature of these systems.
The geographical location of sheep farms (Figure 1) showed larger concentrations in the
Coastal area and Central Valley. The municipalities of Litueche, La Estrella and, Marchigüe had
65 and 67% of the commercial farms reported for 1997 and 2007 respectively, and showed an
increasing number of farms owning 30-100 LU.
7
2.2.1. Typological classification and Inter-census evolution
The method proposed by Escobar and Berdegué (1990) and used by Guillem et al., (2012),
and Toro-Mujica et al., (2012) was applied to farm classification. It includes three stages, as
follows: review and selection of variables, factor analysis, and cluster analysis. This approach
was applied to the 2007 data yielding a number of farm types as well as equations to estimate
factors for each farm. Quantitative variables included dimension, land use and tenure, and the
composition of the farm crop and livestock portfolio. An initial descriptive analysis led to
elimination of variables with a coefficient of variation less than 60%. This was followed by
calculation of the correlation matrix between variables, to eliminate those highly correlated
among each other (r>90%) and those totally uncorrelated. Bartlett’s chi-square test was used to
ensure adequate correlations, and the Kaiser-Meyer-Olgin index was calculated to determine
sampling adequacy (Uriel and Aldás, 2005). The remaining standardized variables were subject
to a factor analysis using principal components to extract the factors (Hair et al., 2009). Factors
with eigenvalues larger than 1 were considered significant (Uriel and Aldas, 2005). The varimax
rotation was applied to these factors to maximize simplification (Aggelopoulos et al., 2009; Hair
et al., 2009).
A hierarchical cluster analysis was used to separate farms into distinct groups, using the Ward
and centroid methods to delimit them (Álvarez-López et al., 2008; Riveiro et al., 2013). The
Euclidean, square Euclidean and Manhattan distances were calculated for each of these methods.
Selection of the number of groups was based on observation of the respective dendograms and
variation in the cluster coefficients in successive stages (Caballero and Fernandez-Santos, 2009).
8
The resolution of the groups was tested with analysis of variance and discriminant analyses, and
the final solution accepted was that for which the discriminant analysis classified the majority of
the farms, and in addition generated significant differences in the largest number of original,
observed, variables.
Contingency and chi-square tests were used to determine the dependence of the typological
groups on flock size prior to further comparisons. When significant associations were found, the
adjusted residues allowed determination of combinations of typological groups with flock size
that differed from expectations (Caballero, 2001). Comparisons between typological groups
identified for the 2007 Census were carried through analysis of variance or Kruskal-Wallis test,
of the original quantitative variables and by contingency and chi-square tests of the qualitative
variables. The variables loadings or weights for each factor in the analysis of the 2007 survey
data were then applied to the 1997 farm data to make the two datasets compatible with each
other. The 1997 data thus transformed, were subjected to a discriminant analysis and grouped in
the typological groups defined for 2007. Changes between the two censuses were assessed by
statistical comparison of the selected variables.
All statistical analyses were performed with SPSS 11.5 (Pérez, 2005).
2.3. Bio-economic performance
Due to lack of on-farm technical and economic parameters for sheep systems in the study
region, a mathematical, dynamic, simulation model was use to estimate the above parameters.
Aguilar et al., (2006) developed a well tested and validated model to estimate lamb production
under grazing conditions with and without supplementation of ewes and lambs, parameterized for
the target region. The model is energy-driven, and all energy transactions use the CSIRO (2007,
9
1990) equations with appropriate corrections for the effect of breed. Body condition score (CS)
was used to calculate reproductive performance as in CSIRO (2007), but CS was estimated from
body weight using the regression proposed by Frutos et al., (1997) for Churra ewes in Spain,
corrected based on the standard reference weight (CSIRO, 2007) of the local breeds. CS thus
calculated were input into the logistical regression of Molina et al., (1994) developed for
Manchega ewes in Mediterranean environments, to estimate ewe fertility at mating. Lamb birth
weights were estimated as proposed by Donald and Russell (1970), corrected for the breed effect,
and modified to include the effect of feeding level as in Kelly et al., (2006). The latter authors
also showed a “U” type of relationship between lamb birth weight and lamb death rate, and data
for Merino were derived from Oldham et al., (2011). All other animal parameters were calculated
as in CSIRO (2007). Lambs were sold when they reached 25 kg liveweight and 120 d of age.
The model included three stochastic variables, namely, the voluntary forage intake by grazing
sheep with a normal distribution for random factors as proposed by Naylor et al., (1966), and
ewe and lamb death rates with respective means of 3 and 5% yearly, in a geometric distribution.
The model also calculates costs associated with labor use, feeds, and other minor inputs, and
simulates lamb births and deaths and their weight gain, thus providing estimates of farm outputs
and profitability. Input data consisted of the dimensional and other variables supplemented with
average climate data for the region. Following detailed examination of the large number of output
variables, some of them were selected and summarized to provide data indicative of the scale of
production (Yearly lamb production per farm), the most important farmer-controlled management
variable (Stocking rate), data on reproductive performance as influenced by feeding strategy and
ewe condition, the degree of farm diversification (Sheep-derived income; Sheep LU), and lastly,
estimates of economic farm performance (Net income, Operational income, Average cost, Mean
operational cost). Sheep-derived income is the ratio between income originating in sheep
10
production and total farm income. All economic results are expressed per unit lamb produced.
Sheep livestock units (LU) is intended to represent the importance of the sheep flock relative to
all farm livestock (Gaspar et al., 2008; Milan et al., 2006). Net income is the total farm income
minus the market value of all fixed and variable costs. The mean operational cost equals net
income plus the opportunity cost of all animals and the cost of forage consumed. The average
cost is the ratio of total costs divided into the number of lambs sold. The average operational cost
is the ratio between the total operational cost and the number of lambs sold (Aguilar et al., 2006).
The model was employed to simulate a random sample of farms. The sample consisted of 8, 5
and 3 farms for Groups I, II, and III respectively, representing a minimum of 2.5% of the
respective totals. For each farm 20 replicates were performed to accommodate stochastic
variables previously described.
3. RESULTS AND DISCUSSION
3.2. Farm typology, 2007 Census
3.2.1. Characterization variables
Ten of the 35 census quantitative variables were discarded due to low CVs. The correlation
matrix was used to eliminate 11 additional variables, thus leaving 14 for further analysis. The
KMO test was 0.7 suggesting an adequate sample, and Bartlett’s chi-square test was highly
significant (P<0.001) indicating also the adequacy of the correlation matrix for the subsequent
factor analysis (Uriel and Aldas, 2005). Four factors were selected that accounted for 71.7% of
the variance (Table 2).
11
The first factor accounted for 36.5% of the variance and was related to the majority of the
dimensional variables examined, including farm size, irrigated area, cropped area, flock size and
size of the labor force. These results partially resembled those of García-Martínez et al. (2009)
who found that the first factor accounting for 23% of the variance was closely correlated with
flock size (r=0.74) in Spain’s mountain cattle systems. In the present study, the second factor
accounted for 13.5% of the variance and was related to the relative size of the sheep component
in the farming systems examined, where low factor values indicated the large proportion of LU as
sheep. The third factor accounted for 11.3% of the variance and was related to high value orchard
and vineyard crops; high values for this factor are indicative of non-sheep related agricultural
activities. Lastly, the fourth factor explains 10.3% of the variance and was related to degree of
farm intensification. High values for this factor suggest larger than average sheep stocking rates,
associated with a larger relative area of sown perennial forages.
3.2.2. Typology
The definitive cluster analysis yielded three groups or clusters using Ward’s method and
estimation of the Euclidean distance. A discriminant analysis using the cluster groups as
classification variable, correctly classified 97% of the farms. Most of the erroneously assigned
farms corresponded to farms of group I that were classified as group II. Groups I, II, and III
accounted for 81, 17, and 2.5% of the farms respectively. Figure 2 shows farm distribution based
on the two first clusters. Differences between Groups I and II are mostly accounted for by the
weights of factor 2, indicative of Group I farms specializing in sheep production. Group III on the
other hand shows high weights for factor 1, indicative of larger farms and higher sheep stocks.
Groups I and II are more homogenous for factor 2, relative to factor 1.
12
Table 3 shows between group differences and similitude for the 34 variables related to farm size,
degrees of specialization and intensification, and social variables. Farms in Group III were
significantly larger than in Groups I and II. Differences between farm sizes gave rise to
significant differences in areas of annual and perennial forages, forest plantations, orchards and
vineyards. Group I had smaller percentages of LUs as cattle, croplands, forested areas, and less
diversity in crops and livestock, whereas the sheep stocking rates were significantly larger than
for Group II.
In general, the three groups showed significant differences in 26 of the 34 variables considered
(P<0.05). Using these differences and the geographical locations of the farms it was possible to
characterize each group as follows.
Group I accounts for 80.8% of the region’s sheep farms. The majority are located in the
Coastal and Interior dry land regions, with two municipalities, Litueche and La Estrella,
accounting for 55.8% of the sheep farms (Figure 3). In general they are small to medium-sized
farms, with small areas of cereals (less than 2.3% of the area). 72% of the farms are family
owned and include very limited use of contract labor, thus indicating that they are essentially
family farms operated by their individual owners. This type of farm applies the survival strategy
I-1 defined by Bowler (1992) and Ilbery (2001), in which the challenge is to maintain a viable
agricultural enterprise, without recourse to diversification. Primary education is predominant
(61%), and only 10% of the owners have technical or university degrees. Sheep flocks are small,
and the main feed resource is native pastures. Forest plantations, orchards, perennial pastures and
annual forages areas cover 3.7, 0.7, 0.7 and 1.1% of the average arm area respectively. Lack of
irrigation (0.6% of the farm area) and very limited infrastructure were associated with little
diversity in crops and animals.
13
Group II includes 17% of the sheep farms, with similar geographical location to Group I. The
municipality of Marchigue contains 29% of this group’s farms (Figure 3). Farm and flock size
are also similar to Group I, and the main differences are that cattle represent 34.1% of the LU,
and have somewhat larger areas of wheat croplands (3.4%) and forested areas (8.6%).
With regard to qualitative variables, farms in this group have better infrastructure, including
stockyards and water reservoirs. As for Group I, farms are owned by individual farmers, who also
have similar educational levels.
Group III is the smallest group, with only 2.5% of the sheep farms, and are mostly located in
the municipalities of Peralillo and Marchigue (Figure 3). Farm sizes tend to be large and have
correspondingly larger herds, including 70.4% of the LU as sheep and 25.8% as cattle. In this
sense, they are intermediate between the two previous groups. They also include larger areas of
sown pastures and annual forages, reforested areas, and irrigated sectors. Nevertheless, the
percentage areas of these land uses differ from the other groups only in the relative importance of
native pastures (41% of the total area), and larger surface areas of corn, orchards, vineyards and
sown forages, associated with larger irrigated areas.
These farms generally have better infrastructure including facilities for forage conservation
and better stockyards. 60% of the farms belong to various types of corporations, have a paid
administration (90%), and 66% of the managers have technical or university degrees.
3.2.3. Inter-census changes
The comparison between the two censuses showed significant differences (P<0.01) in 21 of
the 35 variables considered (Table 4). There are absolute and relative decreases in livestock
numbers and in cereal crops (mostly wheat crops), and native pastures (Table 4). The average
14
number of crop and livestock enterprises per farm decreased between the two censuses,
suggesting an increased specialization of these farms.
Contrary to expectations, the increase in the age of farmers and managers was less than the 10
years elapsed between the two censuses but given the initial advanced age of both, the change
may have been due changes in management and decreases in number of farmers. Regarding
qualitative variables, there was a decrease in areas allocated to stockyards and water reservoirs,
with the exception of an increase in feedlot areas. There were decreases in the areas managed by
estates and de facto corporations, and an increase in farms managed by paid managers.
The classification of the 1997 sheep farms according to the groups defined for 2007
subdivided the farm population in 67.9, 30.4 and 1.7% for groups I to III respectively, showing
that Group II experienced the largest changes. Two trends were evidenced, firstly that a large
proportion of farms increased the sheep stock and decreased that of cattle, thus moving to Group
I, and secondly, a small group of farms increased farm and stock size, moving to Group III.
The largest differences between the two census data was observed in Group I. Contrary to the
overall population, Group I showed significant decreases in farm, corn, and perennial pastures
areas, with a trend towards sheep specialization relying mostly on native pastures and decreasing
areas of cereals. The limited irrigated areas were mostly dedicated to orchards. There was also a
decrease in the numbers of farms with stockyards, and little change in the remaining
infrastructure. Trends in ownership were similar to that of the whole farm population, with
increases in individual ownership and decreases in other legal modalities.
There were no changes in Group II in the relative proportions of cattle to sheep units, or in the
ratios of perennial crops and forages, and in forested areas, but there were decreases in the
15
relative areas of native pastures and wheat. The social structure remained unchanged, but the
number of appointed managers decreased. There was an increase in irrigated infrastructure,
possibly due to the application of Law 18450 (MINAGRI, 1985), that stimulated private
investments in irrigation and drainage.
Group III showed the least changes over the 1997-2007 period, with decreases in wheat crops
and in the number of crops and livestock enterprises per farm. Land ownership increased and
permanent paid labor decreased, whereas the number of paid managers increased, associated with
the surface area and ownership type of these farms. The average sheep stocking rate of the region
and of the groups did not change over the 1997-2007 period, indicating the lack of intensification
and the continued dominance of extensive sheep meat production systems. The frequency of
stockyards and water reservoirs in Group III (30% each) was considerably larger than in Groups I
and II (barns 4.1 and 16.7%; water reservoirs 3.1 and 15.2% respectively). Furthermore, the
majority of farms in Group III were managed by a paid administrator.
3.3. Bio-economic performance
The simulated bio-economic performance showed between-group differences that are
essentially the consequence of three variables: farm size, farm diversification, and animal breed
respectively (Table 5). The smaller flock size of Groups I and II led to larger production costs
and less income per lamb and per kg live weight. The degree of diversification is reflected in the
relative income from sheep, the relative number of sheep LU, and of total LU/ha. Similar to data
reported by Cecchi et al., (2010) Group I farms that had little or no access to irrigation practiced
extensive grazing management, with little or no inclusion of crops so that the large majority of
16
income was derived from sheep. Differences associated with sheep breed can be observed
between Groups I and II relative to Group III since Merino predominate in the latter and it is
associated with larger flocks, lighter birth and weaning weights, and a larger number of lambs
and kg sold per ewe due to higher reproductive performance. On the other hand, reliance on
native pastures, limited sown pastures and scarce supplementation led to similar production per
hectare. Nevertheless, Groups I and II tended to have slightly larger stocking rates if expressed as
kg/ha.
The large (60%) difference between average financial and operational costs, suggests the
existence of opportunity costs ignored by small farmers and that explain the continued existence
of these systems. This finding is similar to that observed for ecological sheep farms in Castilla-La
Mancha, where unaccounted for costs of family labor explain the subsistence of family farms
(Toro-Mujica et al., 2012).
The present results indicate that sheep farms are profit-seeking, but profitability is variable. Some
costs, particularly in the smaller properties, are underestimated because farmers use family labor
exclusively and possibly also by hidden decapitalization of farm infrastructure. Simulation
modeling of prototypes farms showed that economic variables were highly variable in response
to modest variations in the stochastic variables, and to variations in the farmer-controlled
variables of stocking rate and sheep breed. This is a positive outcome since it suggests that there
are opportunities for improved farm management in Groups I and II probably, but not
exclusively, through horizontal transference of technology and knowledge. The alternative
hypothesis, that is, that differences in performance between farms within groups are due to
differences in resources endowment could not be tested with the available data. Regardless, other
forms of economic incentives, such as directed credit, and technical assistance should be
17
implemented if society values the social, cultural and environmental contribution of small
farmers.
Small and medium farms are likely to be more susceptible to climate change than larger
properties in which the scale and variability of the landscape affords more alternatives to alleviate
climate stress. It is predicted that in the next two to three decades temperatures in Chile will
increase 2-3ºC, with more pronounced increases of 2.7-3ºC in the study region
(Neuenschwander, 2010). Australian simulation studies for regions with comparable climate
indicate that a 3ºC temperature increase might decrease forage production for up to 30%
(McKeon et al., 2009). Equivalent simulation models and research approaches are available for
Chile (Castellaro et al, 2012; Vera et al, 2013), but the possible impacts of climate change have
not yet been explored. The study region has experienced a sustained rate of decrease in annual
rainfall since 1960, whose effect would be further aggravated if temperatures increase as
anticipated, with possible negative consequences on plant growth. In this context the continued
existence of small farms, such as those of Groups I and II, highly dependent on extensive grazing
of native pastures, raises a number of issues (IFPRI, 2005). Similar to the findings of Dubeuf
(2011), sheep systems in Mediterranean Chile face significant economic, social and
environmental challenges. Achieving balance between them would require government support in
terms of infrastructure and training, as well as a larger and more systematic integration with crops
in mixed-farming systems, all of which would contribute to larger profitability, and conservation
of the landscape and its biodiversity (Franco et al., 2012).
It has been repeatedly stated (e.g. Devendra, 2010) that technologies for these fragile systems
need to be socially acceptable, allow for the conservation of rural culture and traditions, enhance
systems’ resilience when faced with imminent climate changes, and conserve biodiversity.
18
European institutions and policies consider that small farms have an important role in rural areas.
Evidence from the European small farm sector indicates that they constitute a means to deal with
rural poverty, that it is advisable to promote their diversification, and that they can provide
environmental and other non-trade public goods (ENRD, 2011, 2010). As Altieri and Toledo
(2011) indicate, traditional family farms such as those in Groups I and II have the potential to
implement agro-ecological management practices based on maintenance of biodiversity,
provision of ecosystem services, allow the existence of regulatory socio-cultural institutions, and
may generate products with local identity. In that case, they may therefore contribute to local,
regional and national food security, and additionally, they may reduce local input and products
transportation costs, and reduce the agricultural ecological footprint. In this context and to the
extent possible, agricultural policies should support sheep systems of maximum resilience, self-
sufficient, economically viable, energetically efficient, socially acceptable, and able to conserve
natural resources. Regarding diversification, the European evidence also suggests that impact
may be related to household age and education, and farm location (distance from markets), as
well as agricultural policy.
The continued existence of small farms managed by an older population with little formal
education is a matter of concern. Strong sector- and group-specific, long-term policies would be
required in Chile, as well as elsewhere, if their continued existence is considered desirable from
the point of view of maintaining a rural population that can conserve typical lifestyles and
traditions, and maintain cultural identity and landscape diversity of the region. Examples of such
policies include the development of a strong story line (“relato” in Spanish), enhanced access to
market and technical information, development of growth poles, diminishing transaction costs,
promotion of continued rural education for the decreasing young to medium-age remaining
19
population, opportunities for off-farm employment, stimulus to the diversification of sheep
systems beyond just simply undifferentiated meat production, promotion and valorization of
land and water conservation and promotion of artisanal leather, wool and other byproducts of
sheep production, strengthening technical and financial assistance, and linking small farms to
supply and marketing chains, among others but it is clear that no single policy or action would be
sufficient to achieve these aims (Ilbery, 1991, McNally, 2001, Morgan-Davies et al., 2012). The
role of women, rural schools, and other local institutions in this scenario remains to be
established but it would appear to be a major priority (Kazakopoulos and Gidarakou, 2003).
4. CONCLUSIONS
The area of native pastures in the study region decreased 30% over 10 years, reflecting a
partial abandonment of sheep raising in exchange for higher value agricultural activities. The
smaller number of paid labor, the larger surface area per laborer and larger own areas over time
indicate modest changes in the farm portfolio, as well as rural migration to urban centers. In
general, sheep farms did not show improvements in irrigation or farmer education, indicating the
limited dynamism of these farms. The number of sheep-only farms in Group I increased from
1997 to 2007, and was accompanied by increases in flock size and area sown to improved
pastures. These changes were interpreted as a need to diminish cost and increase farm outputs,
but with continued importance of sheep production as the main source of income for family
farms. On the other hand the sheep-cattle farms of group II decreased in number. These farms
were better endowed, including limited irrigation, which allowed for more profitable land uses
that represented 50% of farm income.
20
The much larger surface areas of Group III farms allowed the coexistence of sheep farming
with other agricultural activities, where sheep graze cereal stubbles, forested areas, and irrigated
crops. Simulated economic performance of these farms showed lower costs per lamb produced
due to dilution of labor costs over a large number of agricultural activities. The use of Merino in
these farms resulted in a larger number of lambs produced per ewe, leading to larger farm
profitability. Lastly, large simulated bioeconomic differences in performance between farms
within groups suggests that there exists opportunities for improvements in technical and
economic performance among the least profitable small farms.
5. ACKNOLEDGEMENTS
The senior author acknowledges support of Comisión Nacional de Ciencia y Tecnología
(CONICYT), project FONDECYT 3130346 for her postdoctoral studies. The comments of two
anonymous reviewers are appreciated.
6. REFERENCES
Aggelopoulos, S., Soutsas, K., Pavloudi, A., Sinapis, E., Petrou, D., 2009. Suggestions for
reformation in sheep farms based on production cost. J. Food. Agric. Environ. 3, 561-566.
Aguilar, C., Vera, R., Allende, R., Toro, P., 2006. Supplementation, stocking rates, and economic
performance of lamb production systems in the Mediterranean-type region of Chile. Small
Ruminant Res. 66, 108-115.
21
Altieri, M., Toledo, V.M., 2011. The agroecological revolution of Latin America: rescuing
nature, securing food sovereignity and empowering peasants. J. Peasant Stud. 38 (3), 587-
612.
Álvarez-López, C.J., Riveiro, J.A., Marey-Pérez, M.F., 2008. Typology, classification and
characterization of farms for agricultural production planning. Span. J. Agric. Res. 6, 125-
136.
Arroyo, M.T.K., Zedler, P., Fox, M., 1995. Ecology and biogeography of Mediterranean
ecosystems in Chile, California and Australia. Springer-Verlag, New York.
Barrantes, O., Ferrer, C., Reine, R., Broca, A., 2009. Categorization of grazing systems to aid the
development of land use policy in Aragon, Spain. Grass Forage Sci. 64, 26-41.
Baumont, R., Carrère, P., Jouven, M., Lombardi, G., López-Francos, A., Peeters, A., Porqueddu,
C., Martin, B., 2014. Forage resources and ecosystem services provided by Mountain and
Mediterranean grasslands and rangelands. Zaragoza: CIHEAM/INRA/FAO/VetAgro Sup
Clermont-Ferrand/Montpellier SupAgro. 843 p.
Bernués, A., Ruiz, R., Olaizola, A., Villalba, D., Casasús, I., 2011. Sustainability of pasture-
based livestock farming systems in the European Mediterranean context: Synergies and
trade-offs. Livest. Sci. 139, 44-57.
Blackmore, S., Apostolidi, K., 2011. FUTUREFARM- integration of farm management
informaton systems to support real-time management decisions and compliance of
standards. Project Final Report. URL:
http://www.futurefarm.eu/system/files/FFD8.9_Final_Report_4.1_Final.pdf
Bohnet, I. C., Roberts, B., Harding, E., Haug, K. J., 2011a. A typology of graziners to inform a
more targeted approach for developing natuarl resource management policies and
agricultural extension programs. Land Use Policy 28, 629-637.
22
Bohnet, I.C., Roebeling, P.C., Willaims, K.J., Hozworth, D., va Grieken, M.E., Pert, P.L., Kroon,
F.J., Westcott, D.A., Brodie, J., 2011b. Landscapes Toolkit: an integrated modelling
frameworkd to assis stakeholders in exploring options for sustainable landscape
development. Landscape Ecol. 26, 1179-1198
Bowler, I., 1992. Sustainable Agriculture as an Alternative Path of Farm Business Development,
in: Bowler, I.R., Bryant, C.R., Nellis, M.D. (Eds.), Contemporary Rural Systems in
Transition, Vol. 1: Agriculture and Environment. CAB International, Wallingford, pp.
237–253.
Boyazoglu, J., Morand-Fehr, P., 2001. Mediterranean dairy sheep and goat products and their
quality. A critical review. Small Ruminant Res. 40 (1), 1-11.
Boyazoglu, J., 2002. Livestock research and environmental sustainability with special reference
to the Mediterranean basin, Small Ruminant Res. 45(2), 193-200.
Brunel, M., Meza, F., Ros, R., Santibañez, F., 2011.Effects of topsoil loss on wheat productivity
in dryland zones of Chile. J. Soil Sci. Plant Nutr. 11, 129-137.
Buysse, J., VanHuylenbroeck, G., Lauwers, L., 2007. Normative, positive and econometric
mathematical programming as tools for incorporation of multifunctionality in agricultural
policy modelling. Agric. Ecosyst. Environ. 120, 70-81.
Caballero, R., 2001. Typology of cereal-sheep farming systems in Castile-La Mancha (south-
central Spain). Agric. Syst. 68, 215-232.
Caballero, R., Fernández-Santos. X., 2009. Grazing institutions in Castilla-La Mancha, dynamic
or downward trend in the Spanish cereal-sheep system. Agric. Syst. 101, 69-79.
Cáceres, J. 2010. Una vieja y olvidada relación económica: Perú y Chile en torno al trigo, in:
Encuentro de Historia Económica. México pp. 1-11. UNAM.
23
Cacho, O.J., Finlayson, J.D., Bywater, A.C., 1995. A simulation model of grazing sheep: II.
Whole farm model. Agric. Syst. 48, 27-50.
Carmona, A., Nahuelhual, L., Echeverría, C., Báez, A. 2010., Linking farming systems to
landscape change: An empirical and spatially explicit study in southern Chile. Agric.
Ecosyst. Environ. 139, 40-50.
Casasús, I, Rogosic, J., Rosati., A., Stokovic, I., Gabila, D. 2012. Animal farming and
environmental interactions in the Mediterranean Region. EAAP publication Nº131.
Wageningen Academic. Netherlands. 276 pp.
Castellaro, G., Aguilar, C., Vera, R., Morales, L., 2012. A simulation model of mesophytic
perennial grasslands. Chil. J. Agric. Res. 72, 388-396.
Catrileo, A., Toro, P., Aguilar, C., Vera, R., 2009. Use of supplement and variation in the
stocking rate in cow-calf systems on temperate pastures in Chile: a simulation approach.
Anim. Prod. Sci. 49, 1059-1067.
Cortez-Arriola, J., Rossing, W.A., Massiotti, R.D.A., Scholberg, J.M., Groot, JC, Tittonell, P.,
2015. Leverages for on-farm innovation from farm typologies? An illustration for family-
based dairy farms in north-west Michoacán, Mexico. Agric. Syst. 135, 66-76.
Cosentino, S.L., Porqueddu, C., Copani, V., Patanè, C., Testa, G., Scordia, D., Melis, R., 2014.
European grasslands overview: Mediterranean region, in: EGF at 50: The Future of
European Grasslands. Hopkins, A., et al. (Eds). Aberystwyth University. European
Grassland Federation 19, 41-56.
CSIRO, 1990. Feeding Standards for Australian Livestock. Ruminants. Commonwealth
Scientific and Industrial Research Organization. CSIRO Publications, East Melbourne,
Australia, 266 pp.
24
CSIRO, 2007. Nutrient requirements of domesticated ruminants. Commonwealth Scientific and
Industrial Research Organization. CSIRO Publishing: Melbourne. 270 pp.
Dalgliesh, N.P., Foale, M.A., McCown, R.L., 2009. Re-inventing model-based decision support
with Australian dryland farmers. 2. Pragmatic provision of soil information for paddock-
specific simulation and farmer decision making. Crops and Pasture Science 60, 1031-
1043.
Daskalopoulou, I., Petrou, A., 2002. Utilising a farm typology to identify potential adopters of
alternative farming activities in Greek agriculture. Journal of Rural Studies 18, 95-103.
de Rancourt, M., Fois, N., Lavín, M.P., Tchakérian, E., Vallerand, F., 2006. Mediterranean
sheep and goats production: An uncertain future. Small Ruminant Res. 62, 167-179.
Devendra, C., 2010. Concluding synthesis and the future for sustainable goat production. Small
Ruminant Res. 89, 125-130.
Donald, H.P., Russell, W.S., 1970. The relationship between weight of ewes at mating and
weight of new born lambs. Anim. Prod. 12, 273-280
Dubeuf, J.P., Ruiz Morales, F.de A., Castel Genis, J.M., 2010. Initiatives and projects to promote
the Mediterranean local cheeses and their relations to the development of livestock
systems and activities. Small Ruminant Res. 93, 67-75.
ENRD. 2011. Draft Conference Summary. The present and the future of small farms in the
European Union. Krakow, Poland: European Network for Rural Development.
ENRD.2010. Semi-subsistence farming in Europe: Concepts and key issues. Sibui, Romania:
European Network for Rural Development.
Escobar, G., Berdegué, J., 1990. Metodología de tipificación de la producción campesina de la
provincia de Ñuble, Chile. RIMISP. Chile, Santiago de Chile.
25
Franco, J.A., Gaspar, P., Mesias, F.J., 2012. Economic analysis of scenarios for the sustainability
of extensive livestock farming in Spain under the CAP. Ecol. Econ. 74, 120-129.
Frutos, P., Mantecón, A., Ciráldez, F., 1997. Relationship of body condition score and live
weight with body composition in mature Churra ewes. Anim. Sci. 64: 447-452
García-Martínez, A., Bernués, A., Olaizola, A.M., 2011. Simulation of mountain cattle farming
system changes under diverse agricultural policies and off-farm labour scenarios. Livest.
Sci 137, 73-86.
García-Martínez, A., Olaizola, A., Bernués, A., 2009. Trajectories of evolution and drivers of
change in European mountain cattle farming systems. Animal 3, 152-165.
Gaspar, P., Escribano, A.J., Mesías, F.J., Escribano, M., Pulido, F., 2011. Goat systems of
Villuercas-Ibores area in SW Spain: Problems and perspectives of traditional farming
systems. Small Ruminant Res. 97, 1–11.
Gaspar, P., Escribano, M., Mesías, F.J., Ledesma, A., Pulido, F., 2008. Sheep farms in the
Spanish rangelands (dehesas): Typologies according to livestock management and
economic indicators. Small Ruminant Res. 74, 52-63.
Girard, N., Duru, M., Hazard, L., Magda, D., 2008. Categorising farming practices to design
sustainable land-use management in mountain areas. Agron. Sustain. Dev. 28, 333-343.
Guillem, E.E., Barnes, A.P., Rounsevell, M., Renwick, A. 2012. Refining perception-based
farmer typologies with the analysis of past census data. J. Environ. Manage. 110, 226-
235.
Hadgigeorgiou, I., Osoro, K., de Almeida, J.P.F, Molle, G. 2005. Southern European grazing
lands: Production, environmental and landscape management aspects. Livest. Prod. Sci.
96, 51-59.
26
Hair, J.F., Black, W.C, Babin, B.J., Anderson, R.E., 2009 Multivariate Data Analysis, seventh ed.
Prentice Hall. 816 pp.
IFPRI. 2005. The future of small farms. Proceedings of a Research Workshop.Washington, DC
Ilbery, B., 1991. Farm diversification as an adjustment strategy on the urban fringe of the West
Midlands. J. Rural Stud. 7, 207-218.
Ilbery, B., 2001. Alternative farm enterprises on the urban fringe: evidence from the UK. Paper
Presented on the Symposium “Open space functions under urban pressure” Gent, 19–21
September 2001.
INE, 1997. Censo Agropecuario 1997. Base de microdatos. Instituto Nacional de Estadísticas.
Santiago, Chile. URL: www.ine.cl
INE, 2007. Censo Agropecuario 2007. Base de microdatos. Instituto Nacional de Estadísticas.
Santiago, Chile. URL: www.ine.cl
IPPC, 2014. Climate change 2014: Impacts, Adaptation, and Vulnerability. Technical
Summary.Chapter 27, Central and South America.
Kapur, S., Ersahin, S., 2014. Soil security for ecosystem management. Mediterranean Soil
Ecosystems 1. Springer International Publishing. 167 pp.
Kazakopoulos, L., Gidarakou, I., 2003. Young women farm heads in Greek agriculture: entering
farming through policy incentives. J. Rural Stud. 19, 397-410.
Kelly, R.W., Greeff, J.C., Macleod, I., 2006. Lifetime changes in wool production of Merino
sheep following differential feeding in fetal and early life. Aust. J. Agr. Res. 57, 867-876.
Le Houerou, H.N. 2004. An agro-bioclimatic classification of arid and semiarid lands in the
isoclimatic Mediterranean zones. Arid. Land Res. Manag. 18, 301-346.
27
Madry, W., Mena, Y., Roszkowska-Madra, B., Gozdowski, D., Hryniewski, R., Castel, J.M.,
2013. An overview of farming system typology methodologies and its use in the study of
pasture-based farming sysems: a review. Span. J. Agric. Res 11, 316-326.
McKeon, G., Stone, G.S., Syktus, J.I., Carter, J.O., Flood, N.R., Ahrens, D.G. Bruget, D.N.,
Chilcott, C.R., Cobon, D.H., Cowley, R.A., Crimp, S.J., Fraser, G.W., Howden,
S.M., Johnston, P.W., Ryan, J.G., Stokes, C.J., Day, K.A., 2009. Climate change
impacts on northern Australian rangeland livestock carrying capacity: a review of issues.
Rangeland J. 31, 1-29.
McNally, S., 2001. Farm diversification in England and Wales - what can we learn from the farm
business survey?. J. Rural Stud 17, 247-257.
Meylan, L., Merot, A., Gary, C., Rapidel, B., 2013. Combining a typology and a conceptual
model of cropping system to explore the diversity of relationships between ecosystem
services: The case of erosion control in coffee-based agroforestry systems in Costa Rica.
Agric. Syst. 118, 52-64.
Milán, M.J., Bartolomé, J., Quintanilla, R., García-Cachán, M.D., Espejo, M., Herráiz, P.L.,
Sánchez-Recio, J.M., Piedrafita, J., 2006. Structural characterisation and typology of beef
cattle farms of Spanish wooded rangelands (dehesas). Livest. Sci. 99, 197-209.
MINAGRI, 1985. Ley Nº 18.450, que aprueba Normas para el Fomento de la Inversión Privada
en obras de Riego y Drenaje y sus modificaciones
Moreira-Muñoz, A., 2011. Plant geography of Chile. Springer Netherlands. 346 pp.
Morgan-Davies, C., Waterhouse, T., Wilson, R., 2012. Characterization of farmers’ responses to
policy reforms in Scottish hill farming areas. Small Ruminant Res. 102, 96-107.
Naylor, T.H., Balintfly, J.L., Burdick, D., Chu, K., 1966. Computer simulation techniques. New
York: Wiley. 352 pp.
28
Neuenschwander, A., 2010. El cambio climático en el sector silvoagropecuario de Chile.
Fundación para la Innovación Agraria (FIA). Santiago
Olaizola, A.M., Ameen, F., Manrique, E., 2015. Potential strategies of adaptation of mixed
sheep-crop systems to changes in the economic environment in a Mediterranean mountain
area. Livest. Sci. 176, 166-180.
Olaizola, A.M., Chertouh, T., Manrique, E., 2008. Adoption of a new feeding technology in
Mediterranean sheep farming systems: Implications and economic evaluation. Small
Ruminant Res. 79, 137-145.
Oldham, C.M., Thompson, A.N., Ferguson, M.B., Gordon, D.J., Kearney, G.A., Paganoni, B.L.,
2011. The birthweight and survival of Merino lambs can be predicted from the profile of
liveweight change of their mothers during pregnancy. Anim. Prod. Sci. 51, 776-783.
Ovalle, M.C., Squella, F., 1996. Terrenos de pastoreo con pastizales anuales en el área de
influencia climática mediterránea, in: Ruiz, N.I. (Ed.), Praderas para Chile. Instituto de
Investigación Agropecuaria. Santiago, pp. 429-466.
Pérez, C., 2005. Técnicas estadísticas con SPSS. Editorial Pearson Educación, S.A. Madrid,
España.
Pizarro, J., 2007. Atlas climatológico de Chile. Campos térmicos y pluviométricos. Subdirección
de Climatología y Meteorología Aplicada. Santiago. Chile.
Ripoll-Bosch, R., Joy, M., Bernués, A., 2014. Role of self-sufficiency, productivity and
diversification on the economic sustainability of farming systems with autochthonous
sheep breeds in less favoured areas in Southern Europe. Animal 8,1229-1237.
Riveiro, J.A., Mantecón, A.R., Álvarez, C.J., Lavín, P., 2013. A typological characterization of
dairy Assaf breed sheep farms at NW of Spain based on structural factor. Agric. Syst.
120, 27-37.
29
Ruben, R., Moll, H., Kuyvenhoven, A., 1998. Integrating agricultural research and policy
analysis: analytical framework and policy applications for bio-economic modelling.
Agric. Syst. 58, 331-349.
Ruiz, F.A., Castel, J.M., Mena, Y., Camúñez, J., González-Redondo, P., 2008. Application of the
technico-economic analysis for characterizing, making diagnoses and improving pastoral
dairy goat systems in Andalusia (Spain). Small Ruminant Res. 77, 208-220.
Santibañez, F., Uribe, J.M., 1993. Atlas agroclimático de Chile: Regiones sexta, séptima, octava
y novena. Ministerio de Agricultura. Fondo de Investigación Agropecuaria. Santiago.
Chile.
Specht, R.L., Rundel, P., Westman, W.E., Catting, P.C., Mjer, J., Greenslade, P., 1988.
Mediterranean-type ecosystems. A data source book. Kluwer Academic Publishers. 248
pp.
Toro, P., Catrileo, A., Aguilar, C., Vera, R., 2009. Modeling supplementation strategies for beef
steer rearing and fattening systems in southern Chile. Chil. J. Agric. Res. 69(2), 207-213.
Toro-Mujica, P., García, A., Gómez-Castro, G., Perea, J., Rodríguez-Estévez, V., Angón, E.,
Barba, C., 2012. Organic dairy sheep farms in south-central Spain: Typologies according
to livestock management and economic variables. Small Ruminant Res. 104, 28-36.
Uriel, E., Aldas, J., 2005. Análisis multivariante aplicado. Thomson Editores, Madrid. España.
Vera, R., Aguilar, C., Toro, P., Squella, F., Perez, P., 2013. Performance of lambs grazing an
annual Mediterranean pasture or fed supplements based on olive oil cake or maize and its
influence on system outputs. Anim. Prod. Sci. 53, 516-522.
Winkler, S., 1999. Biological diversity of drylands, arid, semi-arid, savanna, grassland and
Mediterranean ecosystems. Background paper. Fourth Meeting of the Subsidiary Body on
Scientific, Technical and Technological Advice, Montreal, Canada, 21–25 June 1999.
30
Zimmermann, A., Heckelei, T., Pérez Domínguez, I., 2009. Modelling farm structural change for
integrated ex-ante assessment: review of methods and determinants. Environ. Sci. Policy.
12(5), 601-618.
Figure captions:
Figure 1: Geographical location of sheep farms
Figure 2: Distribution of the sheep farms according to the scores obtained for Factor 1 and
Factor 2
Figure 3: Geographical location of sheep farms of Groups I, II and III in 2007
31
Table 1: Distribution of sheep farms according to flock size reported in the 1997 and 2007
agricultural surveys, expressed in livestock units (LU).
Class No of LU
1997 2007 Change, %
Nº % Nº %
T0 Less than 5 LU 612 63,5 471 42,4 -23,0
T1 5-30 LU 277 28,7 333 48,5 20,2
T2 30-100 LU 40 4,1 39 5,7 -2,5
T3 >100 LU 35 3,6 23 3,4 -34,3
Commercial sheep farms Over 5 LU 352 36,5 395 57,6 12,2
Total no. of farms
964 100 866 100 -10,2
32
Table 2: Factors selected in the factor analysis, eigenvalues, partial and cumulative variances and
correlation coefficients of the variables with the corresponding factor (as per Hairs et al., 2009 ).
Factor
Nº.
Eigenvalue
Observed Variables Correlation of variables
with the respective factor
% explained variance
(% cumulative
variance)
1
5,1
36,5
(36,5)
Surface area, ha 0,90
Sheep LU 0,86
Cereals, ha 0,72
Cattle LU 0,75
Irrigated area, ha 0,79
Total workers, nº 0,86
Forested area, ha 0,90
2
1,9 Cattle, %LU 0,90
13,5 Sheep, %LU -0,88
(50,0) No. of crops 0,45
3
1,6 Orchards, ha
Vineyards, ha
0,84
0,86 11,3
(61,3)
4
1,4 Sheep stocking rate, LU/ha
Perennial forages , % area
0,84
0,85 10,3
(71,6)
33
33
Table 3: Statistical descriptors of three farm groups identified through hierarchical cluster
analysis of the 2007 survey data, based on sheep livestock units.
Total
Groups
p Dimensional variables I II III
Number of farms --- 319 66 10 --
Surface area, ha 160.1±444.5 137.3±345.2ª 100.6±87.8ª 1281.6±1715b 0.01
Livestock units, sheep 26.6±58.4 24.4±49.2ª 15.7±13.2ª 165.9±199.2b 0.01
Livestock units, cattle 4.5±20.6 1.1±3.9ª 9.4±7.1b 78.2±105.3c 0.01
Total LU 33.9±73.8 28.3±53.0ª 28.6±20.8ª 249.0±288.4b 0.01
Cereals, ha 2.7±9.3 1.8±6.7ª 4.3±6.3ª 19.2±39.4b 0.01
Maize, ha 0.6±6.5 0.1±0.7ª 0.5±1.6ª 16.7±38.9b 0.01
Wheat, ha 1.9±6.6 1.5±6.6ª 3.6±6.2b 2.5±5.4ªb 0.05
Native pastures, ha 64.4±30.4 102.5±308.3ª 64.4±73.7b 604.1±1097.3ªb 0.01
Improved pastures, ha 1.5±19.9 1.8±22.2 0.4±3.0 1.7±5.4 0.88
Sown perennial forages, ha 0.3±2.2 0.1±0.7ª 0.5±2.6ª 5.3±10.3b 0.01
Sown anual forages, ha 2.9±38.4 0.7±3.0ª 1.0±1.6ª 85.5±237.3b 0.01
Forested area, ha 12.2±107.4 3.6±14.0ª 11.1±27.2ª 294.8±633.2b 0.01
Orchards, ha 7.2±0.1 0.3±1.3ª 0.2±0.5ª 13.8±11.7b 0.01
Vineyards, ha 0.5±4.0 0.1±1.0ª 0.05±0.2ª 16.2±19.1b 0.01
Irrigated area, ha 1.1±11.9 0.2±1.9ª 0.4±1.6ª 34.8±69.3b 0.01
Variables related to intensification
Sheep stocking rate, LU/ha 0.28±0.33 0.31±0.36b 0.17±0.07ª 0.15±0.04ª 0.01
Sheep, %LU 80.4±19.7 86.1±16.3c 53.8±9.9ª 70.4±18.4b 0.01
Cattle, %LU 9.1±15.1 3.4±8.5ª 34.1±12.3c 25.8±18.7b 0.01
Native pastures, % surface area 64.4±30.4 65.5±30.7b 63.2±28.0b 41.0±28.9ª 0.04
Improved pastures, % surface area 1.6±10.2 1.8±11.1 0.6±4.5 0.5±1.5 0.63
Sown perennial forages, % surface
area 0.7±4.3 0.7±4.7 0.4±1.3 1.3±2.1 0.48
Sown annual forages, % surface area 1.2±3.8 1.1±3.7 1.3±2.7 2.9±8.7 0.32
Forested area, % surface area 4.6±10.3 3.7±9.4ª 8.6±12.7b 8.2±15.0b 0.01
Orchards, % surface area 0.7±2.6 0.7±2.7ª 0.3±1.1ª 3.5±4.2b 0.01
Cereals, % surface area 2.7±7.2 2.3±7.2ª 4.6±6.3b 3.5±10.6ª 0.05
Maize grain, % surface area 0.7±5.4 0.6±5.7 0.8±3.1 3.0±8.7 0.39
Wheat, % surface area 1.6±4.2 1.3±3.8ª 3.4±5.7b 0.7±1.9ª 0.01
Irrigated area, % surface area 0.7±5.5 0.6±5.7ª 0.7±3.2ª 4.6±10.6b 0.08
Livestock classes, nº 2.9±1.3 2.6±1.2ª 4.0±1.0b 2.9±0.9ª 0.01
Crops, nº 1.2±1.5 1.0±1.3ª 2.4±1.8b 2.2±2.1b 0.01
Social variables
Age, years 61.9±13.3 61.4±13.2 63.9±13.7 70.6±10.1 0.19
Ownland, % 87.2±30.3 71.5±42.7ª 77.2±38.3ªb 99.1±2.9b 0.08
Total workers, nº 1.7±2.7 1.3±1.4ª 1.5±1.1ª 13.0±9.81b 0.01
Permanent workers, nº 0.8±2.6 0.47±0.7ª 0.42±0.7ª 12.8±10.0b 0.01
34
34
Table 4: Farm structural changes between 1997 and 2007.
Groups
Total I II III
Dimensional Variables 1997 2007 p 1997 2007 p 1997 2007 p 1997 2007 p
Number of farms 239 319 --- 107 66 --- 6 10 --- 352 395 ---
Surface area, ha 204.6b 137.3a 0.04 166.9 100.6 0.96 1281.4 1281.6 0.99 211.5 160.1 0.11 Animal units, sheep 37.6 24.5 0.39 23.4 15.8 0.86 211.8 165.9 0.66 36.2 26.6 0.59
Animal units, cattle 3.2b 1.1a 0.001 16.3 9.4 0.42 130.6 78.2 0.42 9.3b 4.5a 0.01
Total animal units 45.3 28.3 0.25 44.5 28.6 0.59 474.2 249 0.22 52.3b 34a 0.02 Cereals, ha 11.5b 1.8a 0.001 18.9b 4.3a 0.001 82.7 19.2 0.11 17.6b 2.7a 0.01
Wheat, ha 5.5b 1.5a 0.001 8.5 3.6 0.001 36.1b 2.5a 0.04 6.9b 1.9a 0.01
Native pastures, ha 163.4b 102.5a 0.03 115.2b 64.4a 0.05 443 604.1 0.74 153.5b 108.8a 0.06 Improved pastures, ha 0.7 1.8 0.48 11.8 0.4 0.19 317.4 1.7 0.15 9.5 1.5 0.11
Sown perennial forages, ha 1.2b 0.1a 0.02 3 0.5 0.25 282 5.3 0.13 6.5 0.3 0.11
Forested area, ha 8.0 3.6 0.31 7.4 11.1 0.39 267.2 294.8 0.93 12.3 12.2 0.99 Orchards, ha 0.1a 0.3b 0.05 0.4 0.2 0.47 13.3 13.8 0.95 0.4 0.6 0.48
Irrigated area, ha 1.3b 0.2a 0.001 5.6 0.4 0.22 25.8 34.8 0.79 3.0 1.1 0.13
Variables related to intensification Sheep stocking rate, LU/ha 0.29 0.31 0.46 0.19 0.17 0.15 0.16 0.15 0.59 0.3 0.3 0.27
Sheep, %LU 84.4 86.2 0.23 51.7 53.8 0.22 60.8 70.4 0.33 74.1ª 80.4b 0.01
Cattle, %LU 5.0b 3.4a 0.03 35.2 34.1 0.70 21.9 25.8 0.66 14.4b 9.1b 0.01 Native pastures, % surface area 74.7b 65.4a 0.001 73.8a 63.2b 0.001 58.8 41 0.31 74.1b 64.4a 0.01
Improved pastures, % surface area 0.5a 1.8b 0.07 2.0 0.6 0.34 15 0.5 0.17 1.2 1.6 0.56
Sown perennial forages, % surface area 1.5 0.7 0.17 0.9 0.4 0.21 10.9 2.1 0.21 1.5b 0.7a 0.08
Forested area, % surface area 2.1 3.7 0.97 3.6a 8.6b 0.00 10.1 8.2 0.82 2.7 4.6 0.79
Orchards, % surface area 0.2a 0.7b 0.02 0.3 0.3 0.82 0.8 3.5 0.15 0.2a 0.7b 0.01 Cereals, % surface area 10.7b 2.3a 0.01 14.4b 4.6a 0.00 5.4 3.7 0.63 14.2b 2.7a 0.01
Wheat, % surface area 5.0 1.3 0.001 7b 3.4a 0.01 2.5 0.7 0.13 5.5b 1.6a 0.01
Irrigated area, % surface area 2.2 0.6 0.001 1.2 0.7 0.54 1.4 4.6 0.49 1.9b 0.7a 0.06 Livestock classes, nº 3.2 2.6 0.001 4.8 4.0 0.001 4.2 2.9 0.01 3.7b 2.9a 0.01
Crops, nº 2.0 1.0 0.001 3.4 2.4 0.001 5.3 2.2 0.04 2.5b 1.2a 0.01
Social variables
Age, years 59.2a 61.4b 0.06 60.4a 64b 0.10 60 70.7 0.4 59.5a 61.9b 0.02
Land own, % 62.9a 71.5b 0.02 61.3a 77.2b 0.05 50a 99.1b 0.05 62.2a 73.2b 0.01
Total workers, nº 2.4b 0.4a 0.001 2.8 0.3 0.001 30 12.6 0.22 3.0b 0.7a 0.01 Permanent workers, nº 1.8b 0.1a 0.001 1.9 0.2 0.001 22.3b 2.5a 0.06 2.2b 0.2a 0.01
Surface area per paid worker, ha 233.0b 92.4a 0.08 83.5a 190.1b 0.02 108.6a 303.2b 0.04 90a 229.2b 0.01
35
35
Table 5: Bio-economic results of simulation runs for prototypes of three farm groups (monetary
values in Chilean $ of 2013 (1 US=$ 610)).
Variable Group I Group II Group III p
Yearly lamb production per farm, nº 59.4±26.8a 101±58
a 2,820±2,917.5b 0.01
Lambs per ewe per year, nº 0.8±0.1a 0.8±0.1
a 1±0.1b 0.05
Sheep-derived income, % of total 79.5±23.4b 54.3±20.8
ab 44.9±11.4a 0.07
Stocking rate, LU/ha pasture 0.35±0.13 0.39±0.31 0.29±0.09 0.85
Stocking rate, LU/ha total 0.22±0.05b 0.19±0.07
ab 0.11±0.01a 0.10
Net income, $/lamb 10,300±8,009a 8,721±6,055
a 17,985±3,224b 0.07
Operational income*, $/lamb 27,724±3,811a 30,003±1,808
ab 33,598±5,015b 0.09
Average cost, $/lamb 29,311±7,413b 33,062±6,154
b 23,964±3,169a 0.08
Mean operational cost, $/lamb 11,887±2,724b 11,779±1,819
ab 8,351±4,961a 0.10
Sheep LU, % of total 95.5±8.7b 59.1±5.6
a 70.2±10.1a ≤0.01
*Does not consider the opportunity costs of pastures and animals.
Highlights
Sheep systems in the semiarid zone of Central Chile are classified into three farms types
Meat- and- wool sheep are mainly run by family farms at close to subsistence level.
Economic subsistence of family farms occurs at the expense of not valuing some production
resources
Social and environmental sustainability require support of public policies
36
36
Figure 2: Geographical location of sheep farms
37
37
Figure 2: Distribution of the sheep farms according to the scores obtained for Factor 1 and
Factor 2
-4
-3
-2
-1
0
1
2
3
4
5
6
-4 -2 0 2 4 6 8 10
Fact
or
2
Factor 1
Group I Group II Group III
38
38
Figure 3: Geographical location of sheep farms of Groups I, II and III in 2007