Sustainable Livestock Farming in the European Union: A Study on Beef Farms in NUTS 2 Regions

4.1. Principal Component Analysis

The PCA allowed the identification of four factorial dimensions, rotated with Varimax rotation, for a total explained variance of 73.1%. The four principal components (PCs) collect and synthesize information from the original correlated variables and represent new uncorrelated factorial dimensions. These new variables make it possible to identify and describe the traits and characteristics of EU beef farming patterns.

The first PC accounts for about 23% of the variance and describes a “Professional large livestock production” whose farming management patterns are highly externally labor-intensive. This component includes farms based on a non-family labor force (0.825) and managed by legal persons (0.734). A legal person pertains to a legally recognized entity, such as a corporation or organization, created by law, while a natural person refers to an individual human being that can own property, enter into contracts, be subject to laws, and assume personal responsibilities. This legal person management system is further confirmed by the presence of a negative correlation with a family labor force (−0.815) and with management by physicalnatural persons (−0.792). On the other hand, this pattern shows a positive correlation with a large farm surface, of more than 50 ha (0.569), and with a high density of livestock units per farm (0.478). Furthermore, considering the negative coefficient of small to medium farm size (−0.506), large farms strongly characterize this factorial dimension.

The second component explains about 20% of the total variance and is characterized by a high presence of permanent grasslands (0.936), which refers to grass-covered land intended for continuous agricultural use without periodic conversion, meadows, and pastures (0.703), which instead have no implication for permanent land use, and a fair presence of livestock farms (0.437). On the contrary, the strongly negative correlation with arable (−0.940) and cereal-crop lands (−0.871) suggests that it is non-specialized extensive livestock farming with some marginal characteristics. Considering that the variables expressed in this PC refer to extensive and marginal agriculture with a good presence of zootechnical activity, this component was named “Marginal livestock farming”.

The third PC, which accounts for about 16% of the total variance, includes medium- (0.602) and large-sized farms (0.512), with a significant presence of pastures and meadows (0.439) and a certain prevalence of cattle farms in the territory (0.578). The negatively correlated variables of irrigated agricultural area (−0.625) and small-sized farms (−0.890) improve the characterization of this component. The variables included in this component suggest an extensive management profile with sustainable features, considering, for instance, a reduced water consumption. In fact, the variables expressed in this case suggest areas with large-sized farms associated with pastures and low irrigation use. These are therefore elements of environmental sustainability, as the high presence of pastures (typically polyphytic, rich in plant species) and the reduced use of irrigation can preserve biodiversity and prevent soil erosion, respectively; for these reasons, this PC was defined as “Extensive and sustainable livestock farming”.

The fourth and last component, describing about 13% of the variability, depicts the most market-oriented model with the highest economic performance, which is why this PC was named “Business-based cattle farming”. Indeed, it is positively correlated with an economic indicator such as the high standard output (0.860). The business orientation is evidenced by rational management, as shown by the high percentage of irrigated areas (0.492) and to the high livestock density coefficient. The significance of the variables Livestock units per unit of utilized agricultural area (LSU/UAA) (0.716) and livestock units per livestock farm (LSU = 0.715) describes highly specialized beef farming. The latter component can also be considered intensive cattle farm management because it uses a low percentage of agricultural land.

4.2. Cluster Analysis

A cluster analysis was carried out on the obtained factorial dimensions and, in order to assess the optimal number of clusters, a two-step cluster analysis was employed using the silhouette index.

Cluster 1 strongly expresses the fourth component, “Business-based cattle farming”, while the other components are all negatively expressed, in particular the PC related to “Extensive and sustainable livestock farming”. This cattle farming pattern, which is widespread in a small number of NUTS 2 regions, is therefore characterized by farms with a high degree of specialization, livestock intensity, and economic performance, which can be referred to as “High-yield specialist livestock farms”.

Cluster 2 expresses positively the PC called “Marginal livestock farming”, where the component “Extensive and sustainable livestock farming” is strongly negative. These characteristics help to outline a territorial model where the presence of cattle farms is scarce and, if present, has marginal characteristics. In view of the fact that the remaining PCs are negatively related to the cluster, this group can be identified as “Non-specialized extensive cattle-farms”, thus falling into areas with low beef-producing activity.

Cluster 3 is strongly characterized by marginal livestock farming, with a high presence of meadows, pastures, and permanent grasslands. These conditions lead to a certain propensity for sustainability, given that the second and third PCs are strongly expressed. It is therefore possible to consider this cluster as “Marginal and sustainable cattle farms”.

Cluster 4 is distinguished from the others by the significant presence of intensive livestock farming, organized in corporate form, as observed in the first PC. The other components have coefficients close to 0; so, they are barely, or not at all, relevant in its characterization. Considering this profile, it is possible to name the cluster “Highly intensive cattle farms”.

The last cluster (Cluster 5), which includes the third PC, contains elements of extensiveness and sustainability, unlike the third cluster, where NUTS 2 regions are more characterized by attributes of marginality. Furthermore, considering that the expression of the second PC is negative, it is possible to suppose that these territories are mainly represented by extensive and sustainable farms. For these reasons, Cluster 5 was named “Low environmental impact and large-sized farms”.

It should be noted that not all macro-areas identified by the NUTS 2 classification could be used for PC analysis and subsequent clustering, mainly due to the lack of data about the variables used in the multivariate analysis. For this reason, about 16% of the areas were not analyzed and, together with other EU areas, are represented in crossed-out white.

Cluster 2 is the second smallest cluster by diffusion and is mainly concentrated in the Balkan, Aegean, and Eastern European areas. In addition, some Italian regions (Trentino-Alto Adige and some others in southern Italy) and about two-thirds of Portugal are included. Cattle farming in this group is not very widespread and is characterized by a low level of organization, probably due to other livestock farming of more profitable species or to a different land use.

Cluster 3 is mainly present in the central-northern area of Great Britain, the whole of Ireland, and most of the Alpine region of central-southern Europe, plus some areas of northern Spain and central-eastern France. The breeding most represented here is consistent with a large presence of meadows and pastures, aiming at the maintenance of a marginal, but more widespread, type of livestock farming and with more marked aspects of environmental sustainability.

Cluster 4 includes most of the Spanish and French territories, as well as Central and Eastern Europe, with parts of Germany and Poland, the Czech Republic, and Slovakia. Again, the nature of the cluster is in line with the spatial location of the NUTS 2, which includes part of areas with strong livestock farming intensity. This is particularly the case of the French and Eastern European territories, which are already large-scale exporters of cattle for breeding and slaughter.

Cluster 5 is by far the group with the widest distribution, as it is present in most of the Italian peninsula, almost all of Germany, Poland, the Baltic States, Finland, and Sweden, as well as in southern Great Britain and parts of northern France. These areas are characterized by large-scale farms. The widespread incidence of this cluster throughout the EU territory, and in particular in Eastern European territories, suggests that this kind of beef production, extensive and with a low environmental impact because of pastures and low irrigation use, could represent a substantial share of production within the EU. However, extensive farming does not always translate into broad production and economic capacity, which is why these farms could be mainly oriented toward satisfying the local or national demand.