Live bovine animal export dynamics and structural heterogeneity across European member states, 2014-2024: A cluster -based analysis Authors: Evija Skujina, Irina Pilvere Published: 2026 ABSTRACT The European Union (EU) cattle sector is facing major structural reforms related to the implementation of the European Green Deal Farm to Fork Strategy and the Common Agricultural Policy (CAP) 2023–2027. This study examines changes in the export of live bovine animals and cattle population dynamics across the EU–23 Member States over the period 2014–2024. Several EU Member States experienced notable year-to-year fluctuations in live bovine animal exports relative to cattle populations. The export and live bovine animal population long-term growth trends were applied to a time series. Export Intensity and the Compound Annual Growth Rate (CAGR) indicator were applied to a time series. In addition, hierarchical and K – means cluster analysis were used to identify differences between EU-23 Member States.The analysis revealed a decoupled dynamic: while live bovine animal export increased by 58% (from EUR 2.26 billion to EUR 3.58 billion), the total cattle population declined by 8.6% (from 76.2 to 69.7 million head). Three distinct clusters of EU Member States were identified: (1) countries with high export intensity and declining cattle population; (2) countries with rapid export growth of live bovine animals from low baseline cattle population (growth-driven cluster); and (3) countries with low live bovine animal export intensity and domestic market orientation.These findings suggest the presence of heterogeneity within the EU–23 Member States in the EU cattle sector and raise concerns regarding long-term competition in live bovine animal exports. Implications The results indicate a within EU–23 Member States cattle sector, where export growth occurs concurrently with declining cattle population. The identification of three distinct export intensity clusters highlights the absence across EU Member States. From a policy perspective, a uniform approach under the CAP may not adequately address the diversity of EU Member States national conditions. Therefore, future policy development, particularly beginning in 2028, Member States with high export intensity may need targeted support to maintain or rebuild cattle populations, while countries with lower export intensity could focus on enhancing export capacity. Further live bovine animal export and cattle population research is necessary to better understand the drivers behind the observed divergence, including economic incentives, cattle genetic potential, and evolving market conditions. Such analysis is essential for evaluating the long-term implications for food security, rural development, and environmental sustainability within the EU regions. Introduction The European Green Deal, through initiatives such as the Farm to Fork Strategy and the CAP 2023–2027, outlines a transition toward more sustainable agricultural systems (European Commission, 2025). These policy initiatives place particular emphasis on public health, climate resilience, biodiversity, and reducing greenhouse gas (GHG) emissions. They also have significant implications for the livestock sector, particularly through the stricter animal welfare requirements and market regulation affecting live bovine animals (Krzyzanowski, 2021; Molitorisová & Burke, 2023). Limiting livestock density is one measure proposed to reduce the negative environmental impact of agricultural production (del Campo et al., 2025). The cattle sector in the EU is characterised by substantial heterogeneity in geographical distribution, farm size and structure, production volume, and economic profitability. Differences between the Western and Eastern, as well as Northern and Southern EU regions, reflect variation in natural conditions, economic situation, and historical development (Vinci, 2022). To reduce the environmental impact of the livestock sector, EU regions need specific approaches and strategies to address these differences (Hocquette et al., 2018; Sydd, 2025). The CAP 2023–2027 provides financial and policy support to the livestock sector while simultaneously establishing ambitious targets for climate neutrality, biodiversity, and sustainable rural development (European Commission, 2024). However, these policy objectives also introduce additional requirements for livestock farming (European Commission, 2025). Stricter environmental standards and changing consumer demand for sustainably produced food are increasing pressure on cattle producers to adapt production and trade practise. At the same time, the livestock sector operates within a global market environment, where trade flows are influenced by economic, geopolitical, and regulatory factors (del Campo et al., 2025). Live bovine animal exports affect the development of the cattle sector. On the one hand, the export of live cattle leads to higher profits for domestic businesses and enhances the competitiveness of the cattle sector. On the other hand, they may reduce the availability of livestock for domestic production and processing (Cowley, 2021; Hocquette et al., 2018). The export dynamics of live bovine animals are influenced by multiple factors, including high cost of transport, animal welfare regulations, disease control, environmental and ethical considerations, and political and geopolitical instability (Adams et al., 2025). These factors may negatively affect the balance between export competitiveness, compliance with animal welfare standards, and environmental sustainability objectives within EU livestock policy (Kearney et al., 2024). European Commission forecasts indicate that EU animal exports will gradually decrease by 3.2% per year between 2022 and 2035, mostly due to animal shortages and restrictive access to certain markets due to geopolitical factors (European Commission, 2025). Production growth will be driven by higher livestock productivity rather than by increases in agricultural area or livestock numbers. Additional environmental requirements pose new challenges for market competition in the EU, particularly in South American countries, which are shifting to other global markets (OECD/FAO, 2024). In Europe, cattle production is one of the most important agricultural and food sectors (Vinci, 2022). In recent years, the EU CAP has increasingly focused on sustainable agricultural production, circular economy principles, and the reduction of food losses among food supply chain actors (Sydd, 2025). Despite the growing policy relevance of the cattle sector within the framework of the European Union Green Deal and Common Agricultural Policy 2023–2027, existing studies have primarily focused on cattle production system environmental impacts or farm-level sustainability indicators. Few studies analyse the relationship between live bovine animal export dynamics and cattle population changes across the EU Member States. There is a lack of comparative studies examining how the intensity of live bovine animal export differs across EU Member States. The gap limits of the analysis of relationship between export dynamics, cattle population trends, and competitiveness in EU Member States. Therefore, this study aims to analyse changes in live bovine animal exports and live bovine animal population across EU Member States from 2014 to 2024, and identify distinct patterns of export intensity through cluster analysis and assess implications for the cattle sector. This study provides several approaches: First, the study provides a comparative analysis of live bovine animal exports and population dynamics across the EU–23 Member States over the period 2014–2024. Second, the study applies a multi-method statistical framework combining time-series analysis, CAGR, cluster analysis, and ANOVA to identify patterns in export performance. Third, the study analyses differences in live cattle exports among the EU–23 Member States, reflecting varying levels of market orientation and structural development in the cattle sector, highlighting the implications associated with export growth and declining bovine populations. The hypothesis of the study is that EU–23 Member States form statistically distinct clusters with respect to live bovine animal export intensity, reflecting different levels of market orientation and development in the cattle sector. Materials and methods Research data sources The study uses the following Eurostat datasets the bovine population and intra EU trade live bovine animal (Intra-EU27) for EU Member States, presented in Table S1. The availability of data from each EU Member State for the period from 2014 to 2024 was verified at the data processing stage. After filtering the data, the final sample included EU–23 Member States, resulting in datasets of n = 253 observations for the period 2014–2024. The EU Member States with significant data gaps (Sweden, Finland, Malta, Cyprus) were excluded from further analysis, because Finland, Cyprus, and Malta data were not reported for the full analyse period, while Sweden only published data for selected years. This data's incomplete coverage prevents the development of consistent and comparable time series, which are necessary for the calculation of research indicators. Incomplete data uses could introduce bias and significantly affect the reliability of statistical estimates. Cluster analysis is particularly sensitive to data distribution, scaling, and completeness of observations; integration of such incomplete data could distort the cluster structure, reduce its stability, and make interpretation difficult. To ensure robustness, methodological consistency, and cross-country analysis comparability, Finland, Cyprus, Malta, and Sweden were excluded from further analysis. The missing 2023 data for Bulgaria were marked “:” in the Eurostat database, indicating unavailable data rather than zero. Linear interpolation was applied to estimate the missing 2023 value, as data were available for both 2022 and 2024, with no structural breaks in the export value of live bovine animals (EUR) and the weight of live bovine animal exports. This approach ensured the continuity of the time series while avoiding artificial volatility. The data interpolation was calculated according to the following formula:(1)y2023=y2022+y20242 The arrangement and filtering of the data ensured the continuity of the research and the preservation of the time series trend. Eurostat data on the population of live bovine animals for several EU Member States included official Eurostat observation flags indicating specific data quality and methodological issues in the annual data. Data on live bovine animals for Bulgaria (2017), Portugal (2021), and France (2021–2023) were classified as provisional (marked as “p”), indicating that the data might be more accurate in future data updates. Data for France in 2019 and 2020 were marked as “d” (definition differences), indicating methodological differences relative to other EU Member States or years. Several EU Member States (Croatia in 2018 and 2022; Hungary in 2020; Italy in 2022 and 2024) received a mark “b” (break in time series), indicating a break in the time series or a change in methodology. Eurostat dataset notes concerning the live bovine animal population were retained and included in the analysis. The research interpreted the live bovine animal population and live bovine animal exports data with caution, given their specific nature. Research period The research covers the period from 2014 to 2024. Research area The region examined by the research covered EU–23 Member States: Austria (AT), Belgium (BE), Bulgaria (BG), Croatia (HR), Czechia (CZ), Denmark (DK), Estonia (EE), France (FR), Germany (DE), Greece (EL), Hungary (HU), Ireland (IE), Italy (IT), Latvia (LV), Lithuania (LT), Luxembourg (LU), the Netherlands (NL), Poland (PL), Portugal (PT), Romania (RO), Slovakia (SK), Slovenia (SI) and Spain (ES). Data analyses For data analysis and interpretation, the monetary values of live bovine animal exports were recalculated to millions of EUR. The weight of live bovine animals was recalculated from 100 kg to tonnes, and the live bovine animal population from 1000 to millions of heads to ensure data comparability. Such scaling made it possible to compare export data and identify the positions of the EU–23 Member States in the EU trade market, and, indirectly, to identify structural trends in live bovine animal exports by the EU–23 Member States. The research into live bovine animal exports for the EU–23 Member States used three main indicators: (1) the value of live bovine animal exports (million EUR); 2) the weight of live bovine animal exports (tonnes); 3) the live bovine population (million heads). Data for the EU-23 Member States was analysed using IBM SPSS Statistics Version 29.0.2.0, yielding the average live bovine animal population and value of live bovine animal exported, and the ratio of live bovine animal exports to the population of live bovine animal at the EU–23 Member States level. Figure. S1 graphically shows the methodology for preparing and analysing the data. Several indicators were used to analyse the data. Export intensity indicator. To identify structural differences and export orientation in the live bovine animal exports sector across the EU–23 Member States, an export intensity indicator (EI) was constructed. The indicator represents the ratio of the average live bovine animal exports to the average live bovine animal population in each country. This approach links export performance to the structural size of the cattle sector. Export intensity was calculated by the following formula:(2)EIi=ViPiwhere EIi is the export intensity indicator (million EUR/million heads); Vi is the average value of live bovine animal exports in 2014–2024 (million EUR); Pi is the average live bovine animal population in 2014–2024 (million heads). This formulation action allowed export activity to be evaluated and assessed relative to the available livestock base rather than in absolute terms. Such an approach is particularly relevant for the live bovine animal sector, where population size constitutes a key structural constraint on export capacity. The approach builds on the previous study, which used export intensity as an indicator of international competitiveness and market specialization (Bojnec & Fertő, 2019; Tłuczak, 2015). However, unlike earlier studies, which defined export intensity as the ratio of export value to output value, the present study relates export intensity to the live bovine animal population. This allows export intensity to be interpreted as an indicator of the structure and market orientation at the national level, capturing the average value per million heads. Consequently, the indicator facilitates cross-country comparability of export performance across the EU–23 Member States. Calculation of absolute annual and base-year changes. To establish yearly changes in the export of live bovine animals by the EU–23 Member States, the research calculated absolute yearly values. This enabled identification of changes in exports in value terms (million EUR), physical volume (tonnes), and the live bovine animal population (million heads), thereby allowing comparison with the previous year.(3)δmch=ym−ym−1where δmch is the absolute change in live bovine exports or the live bovine animal population in the EU–23 Member States; Ym is the value of the respective indicator for each EU–23 Member States for the reporting year m (export value of live bovine animal in million EUR; export volume of live bovine animal in tonnes; live bovine animal populations in million heads); Yₘ₋₁ is the value of the respective indicator for each of the EU–23 Member States for the previous year. The following formula was used to calculate absolute base-year changes:(4)δmb=ym−y2014where δmmb is a change in live bovine animal exports and in the live bovine animal population in the EU–23 Member States from the base year; Ym is the value of the respective indicator for each EU–23 Member States for the reporting year (value of live bovine animals exported in million EUR; volume of live bovine animals exported in tonnes; live bovine animal populations in million heads); Y2014 is the value of the respective indicator for each EU–23 Member States for 2014. Note: Annual absolute changes were calculated based on year-to-year comparisons. For the first year, where no prior-year data were available, the value was denoted as “.” or “0″. In the analysis, several formulas were used to calculate the percentage change in annual exports of live bovine animals and in the live bovine animal population relative to the previous year. The change rate was calculated according to formula 5.(5)Tmch=(ym/ym−1)x100where Tmch is the rate of annual change in live bovine exports or the bovine animal population for the EU–23 Member States; Ym is the value of the respective indicator for each EU–23 Member States for the reporting year m (value of live bovine animal exported in million EUR; volume of live bovine animal exported in tonnes; live bovine animal population in million heads); Ym₋₁is value of the respective indicator for each EU–23 Member State for the previous year, multiplied by 100 to get the percentage. Changes from the base year were calculated according to formula 6.(6)Tmb=(ym/yy−1)x100where Tmb is a percentage increase in live bovine animal exports or the population of live bovine animals from the base year for each EU–23 Member State; Ym is the value of the respective indicator for each EU–23 Member State for the reporting year m (value of live bovine animal exported in million EUR; volume of live bovine animal exported in tonnes; live bovine animal population in million heads); Yy-1 is the value of the respective indicator for each EU–23 Member State for 2014 multiplied by 100 to get the percentage. Calculation of the average level for a time series. To analyse long-term trends in live bovine animal exports and the live bovine animal populations in the EU–23 Member States, the research calculated the average level of the time series. The indicator enabled the determination of the average level of live bovine animal exports over the entire period analysed. The average level of the time series was calculated according to formula 7.(7)ymean=∑m=1nymnwhere Ymean is the average level of the time series showing the volume of live bovine exports or the live bovine animal population in the EU–23 Member States over eleven years (2014–2024); Ym is the value of the respective indicator for each EU–23 Member State for the reporting year m (value of live bovine animal exported in million EUR; volume of live bovine animal exported in tonnes; live bovine animal population in million heads); n is the number of years in the analysis period (2014–2024, n= 11). Compound Annual Growth Rate (CAGR). During the analysis period in several EU –23 Member States, exhibited substantial year-to-year fluctuations in live bovine animal exports (both in value and weight) relative to the bovine animal populations. Because simple base-year comparisons may be distorted by short-term volatility and low initial export values, the CAGR was used to assess long-term trends. CAGR provides a standardised measure of average annual change between 2014 and 2024 and improves comparability across countries by smoothing interannual fluctuations. The compound annual growth rate was calculated according to formula 8.(8)CAGRm.%=[(Yn.2024Y1.2014)1n−1−1]×100where CAGR m. % is the compound annual growth rate for the respective indicator for EU-23 Member States expressed as a percentage; n is the number of years in the period 2014–2024 (n=11), (n-1). corresponds to the number of growth periods (10 intervals between observations) Yn is the value of the indicator for the last year of the period (2024); Y1 is the value of the indicator for the first (base) year of the period (2014). The study period covers 11 years (2014–2024), which corresponds to 10 growth intervals between observation. The exponent is expressed as 1/(n-1) CAGR was calculated for the following indicators: 1) the value of live bovine animal exported (million EUR); 2) the volume of live bovine animal exported (in tonnes); 3) the population of live bovine animal (million heads). Multiplying the result by 100, the CAGR was expressed as a percentage for long-term growth. Analysis of correlations and variable selection for cluster analysis. To identify long-term changes in live bovine animal exports and structural differences in exports among EU–23 Member States, correlation analysis and a multicollinearity test were performed on an initially constructed dataset. The dataset included live bovine animal exports in EUR and kilograms; absolute and recalculated values in millions of EUR and tonnes; average live bovine animal population; annual and base-year absolute changes; the average level and change for the time series; export intensity; and long-term growth rates (CAGR). The calculations were performed using IBM SPSS Statistics Version 29.0.2.0, Analyse-Correlate-Bivariate menu. Seven variables were initially included in the analysis: (1) the average value of live bovine animal exported. in million EUR; 2) the average population of bovine animal in million heads; 3) the average value of live bovine animal exported in tonnes; 4) average export intensity; 5) CAGR for export value %; 6) CAGR for export volume in tonnes %; 7) CAGR for the live bovine animal population %. Since the export growth indicators (in tonnes) are linearly dependent on export value (EUR), they were excluded from further cluster analysis, thereby retaining only independent, economically interpretable indicators. In the subsequent cluster analysis, the following variables were selected: (1) average export value in million EUR; 2) average export intensity; 3) a CAGR for export value; 4) a CAGR for the population of live bovine animals. The correlation analysis confirmed that there was no multicollinearity between the four variables (see the Results section) making them suitable for use in a single cluster analysis model. Before performing the cluster analysis all quantitative indicators were standardised using Z-scores. Standardisation was necessary because clustering algorithms are sensitive to variable scaling, and without normalisation larger-scale variables would dominate distance calculations. The Z-score transformation ensures comparability across all variables (mean = 0; SD = 1) and enables the derivation of methodologically correct cluster results. Hierarchical cluster analysis. Hierarchical cluster analysis was employed to classify the EU–23 Member States according to structural similarities in live bovine animal export performance. The clustering was based on the following standardised variables: average export value (in million EUR), average export intensity, CAGRs for live bovine animal exports, and the live bovine animal population. The agglomeration method and a distance measure. Given the relatively small sample size of data (n = 23), hierarchical clustering using Ward’s method with the squared Euclidean distance was applied to identify homogeneous groups of EU Member States. Ward’s method was selected because it minimizes within cluster variance at each group of the agglomeration process, resulting in compact and well separated clusters compared with other hierarchical linkage methods. The Ward’s method is less prone to chaining effects and is therefore considered suitable for small datasets, as it ensures more stable clustering and reduces intra-cluster dispersion (Dinh et al., 2025; Hill-Sullins et al., 2025; Ikotun et al., 2023). A dendrogram analysis enabled the identification of the optimal number of clusters and naturally formed groups. Determining the optimal number of clusters. A visual analysis of the dendrogram and the agglomeration graph (Elbow method) indicated a potentially optimal number of 3 or 4 clusters. However, in both cases, France consistently formed a separate cluster of a single EU–23 Member States that was statistically distinct from the other groups of EU–23 Member States. This systematic separation identified France as a structurally unique outlier (see the Results section), which could distort the cluster structure and complicate the interpretation of the results. Therefore, France was excluded from further analysis, and the final clustering was performed for the EU–22 Member States. After determining the optimal number of clusters, a K-means clustering algorithm was applied to identify cluster centroids and refine cluster composition. Enhancement of the datasets and stability of the clusters. The final clustering was performed for the EU–22 Member States using the K-means method with a predetermined number of clusters (K = 3). However, the cluster affiliation of Germany identified by hierarchical clustering (hereinafter, HK; Ward's method) differed from that determined by K-means. These differences could be explained by the fundamentally different optimisation approaches applied by the two methods. The HK method forms clusters sequentially combining items with the least mutual distance at each step. The agglomeration table shows that Germany (item 9) was merged with Belgium and the Netherlands in the ninth step, with a coefficient of 2.104, thereby forming the first cluster together with other EU Member States. The dendrogram (see Figure. S5) visualises this hierarchical structure with Germany falling within the first cluster. In contrast, the K-means method is based on the principle of global optimisation, iteratively minimising the sum of squared distances between data points and their cluster centroids. The analysis showed that for Germany, the Euclidean distance to the centroid of the third cluster (1.855) was less than to the centroid of the first cluster; therefore, the K-means algorithm optimally reclassified Germany into the third cluster. This difference cannot be attributed to an algorithmic error; rather, it reflects the complementary nature of clustering methods: HK reveals the structural similarity and sequential formation of data (an agglomerative process), whereas K-means optimizes intra-cluster homogeneity more accurately. As a result, the assignment of Germany to the third cluster by the K-means method was methodologically justified and consistent with its relative proximity to the corresponding centroid. The data calculations were performed using IBM SPSS Statistics Version 29.0.2.0. Analyse–Classify–Hierarchical Cluster Analysis menu and K-means Cluster Analysis. K-means clustering and validation of the results. K-means clustering enabled the identification of cluster affiliations and the computation of final cluster centroids based on standardised Z-scores. The final 3-cluster distribution was used as an independent variable in a one-way ANOVA and Tukey HSD post hoc tests to statistically assess differences among the identified EU Member State clusters and to ensure the validity of the results. The analysis was performed using IBM SPSS Statistics Version 29.0.2.0. Analyse–Compare Means and Proportions menu. Justification of Statistical Approach. The analysis of live bovine animal export across the EU–23 Member States required a methodological approach that captures both temporal dynamics and structural differences between EU–23 Member States. Descriptive statistics provided an overview of export volumes and trends in the live bovine animal population. However, they were insufficient to identify underlying structural patterns or statistically significant differences between the EU–23 Member States. The analysis of both temporal dynamics and structural differences, the study combines time series analysis, hierarchical and K-means cluster analysis, and ANOVA. Time series allowed the identification of export and live bovine animal population trends over the period 2014–2024, highlighting dynamic changes in export performance across the EU–23 Member States. Cluster analysis was used to group the EU–23 Member States according to similarities in export intensity, growth rates, and changes in live bovine animal population, thereby revealing structural patterns within the EU–23 Member States. Finally, ANOVA was applied to test whether the observed differences between clusters are statistically significant, providing robust validation of the identified groupings. The combination of these methods enabled a more comprehensive understanding of export live bovine animal dynamics than descriptive statistics alone. It simultaneously captured temporal developments, structural heterogeneity, and statistically significant differences across the 23 EU Member States. Limitations of the research One of the main methodological limitations involved the heterogeneity of data and their structural differences across data sources. For the dataset, information was selected and aggregated from several Eurostat databases with different metadata levels and units of measurement. The EU trade statistical office collects information on the value of live bovine animal exports in EUR and on weight in 100 kg, while data on the live bovine animal population are measured in thousands of livestock units. These differences in units of measurement limited the direct comparability of data and made difficulties to estimate economic values accurately. Due to differences in data (e.g., number value/weight), the data were used as complementary sources of information for the analysis. In addition, a limitation of the study is the availability of data. The EU Member States Finland, Cyprus, and Malta data were not reported for the full analysis period, while Sweden only published data for selected years. This incomplete coverage prevented the development of consistent and comparable time series, which were necessary for the calculation of research indicators. The interpretation of the results should therefore account for this limitation when comparing EU Member States. Results This section presents the live bovine animal population and export intensity of live bovine animals in the EU–23 Member States, describing the overall economic potential of the market for live bovine animal exports in the EU–23 Member States. Characteristics of live bovine animal exports in EU member states Diverging trends in live bovine animal exports and live bovine animal population dynamics The live bovine animal export trends in the EU–23 Member States during 2014–2024 revealed that live bovine animal exports increased in both value and volume (see Table S2). The value of live bovine animal exported increased from EUR 2262.59 million to EUR 3575.13 million by 58%, while the export volume in tonnes increased by 15.77%. At the same time, the total population of live bovine animals in the EU–23 Member States decreased from 76.2 million to 69.70 million (8.6%). The divergence between export value and export volume suggests an increase in unit export price or changes in the composition of exported animals, such as higher value livestock categories. The linear trend analysis (R2=0.8075) indicates a structural pattern in the EU–23 Member States, where the export value increased despite a decline in the overall live bovine animal population (see Figure. S2). These changes in the cattle sector indicate a shift from production expansion toward efficiency, improvements in productivity and modernization, with greater emphasis on gains in biodiversity, considerations, and herd management practises aimed at reducing costs while improving animal welfare and limiting antibiotics (European Commission, 2023; Smeets Kristkova et al., 2015). Export intensity in the EU – 23 member states In the EU market, large numbers of animals are transported and exported to third countries every year (European Union, 2023). The average export intensity of live bovine animals (see Table 1) reveals significant differences across the EU–23 Member States. In the period 2014–2024, in the Czech Republic, export intensity showed a high average value (EUR 168.60 million at a volume of 72.395.78 tonnes), which was higher than in Luxembourg (EUR 23.66 million at a volume of 11.870.96 tonnes). The difference in export value between the two countries was primarily attributable to differences in the size of their live bovine populations: the Czech Republic had 1.37 million, whereas Luxembourg had 0.19 million. However, after recalculating the average export intensity per million live bovine animals, the values were similar in the Czech Republic (EUR 123.36 million) and Luxembourg (EUR 122.94 million). In Slovenia (EUR 112.00 million) and Slovakia (EUR 111.72 million), the figures were identical. This suggests a high export orientation in the cattle market relative to the average live bovine animal population in these EU–23 Member States during 2014–2024. Medium-high export intensity was observed for France (EUR 74.24 million), Hungary (EUR 72.52 million), the Netherlands (EUR 61.81 million), and Belgium (EUR 60.35 million). The results showed that live bovine animal exports are an important component of livestock farming; however, the live cattle market accounts for a relatively small share of the total cattle sector. During this period, the Baltic States exhibited varying export intensities for live bovine animals. The export intensity was slightly higher in Estonia (EUR 82.32 million) than in Latvia (EUR 77.34 million), thereby indicating a relatively small live bovine animal population in both Member States. In Lithuania, the live bovine animal population was the highest at 0.66 million, and export intensity was EUR 55.11 million per million live bovine animals.Table 1Average export intensity, export value, and export volume of live bovine animals in the EU–23 Member States, 2014–2024.Table 1 dummy alt textCountriesAverage export value, million EURAverage export, tonnesAverage live bovine animal population, million headExport intensity, million EUR per million headCzechia168.6072 395.781.37123.36Luxembourg23.6611 870.960.19122.94Slovenia53.5924 737.080.47112.86Slovakia49.1123 719.150.44111.72Estonia20.619 669.920.2582.32Latvia30.6112 896.660.4077.34Croatia32.2213 595.100.4375.07France1344.194 34 880.9418.1174.24Hungary63.3923 272.050.8772.52Netherlands239.761 01 160.893.8861.81Belgium142.4945 803.392.3660.35Lithuania36.2814 693.320.6655.11Romania60.7019 730.341.9331.37Austria53.1520 239.531.9028.04Denmark40.0214 392.261.5126.59Portugal28.6111 888.871.6117.78Germany200.0469 387.2511.6717.14Spain68.0321 655.056.3910.65Ireland63.7020 131.726.529.78Poland28.0710 506.186.134.58Italy9.133 505.546.191.47Greece0.28120.400.590.48Bulgaria0.2075.360.570.36Source: Author’s calculations based on Eurostat,2014 –2024 (CN8;apro_mt_lscatl). Table 1Average export intensity, export value, and export volume of live bovine animals in the EU–23 Member States, 2014–2024.Table 1 dummy alt textCountriesAverage export value, million EURAverage export, tonnesAverage live bovine animal population, million headExport intensity, million EUR per million headCzechia168.6072 395.781.37123.36Luxembourg23.6611 870.960.19122.94Slovenia53.5924 737.080.47112.86Slovakia49.1123 719.150.44111.72Estonia20.619 669.920.2582.32Latvia30.6112 896.660.4077.34Croatia32.2213 595.100.4375.07France1344.194 34 880.9418.1174.24Hungary63.3923 272.050.8772.52Netherlands239.761 01 160.893.8861.81Belgium142.4945 803.392.3660.35Lithuania36.2814 693.320.6655.11Romania60.7019 730.341.9331.37Austria53.1520 239.531.9028.04Denmark40.0214 392.261.5126.59Portugal28.6111 888.871.6117.78Germany200.0469 387.2511.6717.14Spain68.0321 655.056.3910.65Ireland63.7020 131.726.529.78Poland28.0710 506.186.134.58Italy9.133 505.546.191.47Greece0.28120.400.590.48Bulgaria0.2075.360.570.36Source: Author’s calculations based on Eurostat,2014 –2024 (CN8;apro_mt_lscatl). Average export intensity, export value, and export volume of live bovine animals in the EU–23 Member States, 2014–2024. Source: Author’s calculations based on Eurostat,2014 –2024 (CN8;apro_mt_lscatl). This indicated that all three Baltic States had relatively high export intensity of live bovine animals. The coupled income support implemented under the EU CAP over two funding periods: 2014–2022 and 2023–2027, provides direct payments to cattle farmers, thereby contributing to the cattle sector in the Baltic States. Although the support mechanisms differ between the Baltic States. In Latvia, for example, the support scheme involves intensifying production, raising animal welfare standards, and limiting antimicrobial resistance, as well as supporting investments in infrastructure, farm buildings, and biosecurity measures; consequently, animal exports have increased (Ministry of Agriculture of the Republic of Latvia, 2025). In contrast, lower export intensity was observed in Poland (EUR 4.58 million per million bovine), Greece (EUR 0.48 million), and Bulgaria (EUR 0.36 million), indicating that live bovine animal exports were less significant in these countries. Regional patterns for the detailed country analyses The average export value and the average live bovine animal population in the relevant EU–23 Member States (see Table 1, Figure. S3) revealed that France and Germany had the highest export potential during 2014–2024. In France, the average export value was EUR 1344.19 million, with an average of 18 million head. In contrast, in Germany, the average export value was EUR 200.04 million, with an average of 11.67 million head. The results show that France had the largest population of live bovine animals. France and Germany had the highest export potential among the EU–23 Member States. The Baltic States exhibited a different export pattern from that of Germany and France. In 2014–2024, the average export value in Latvia was EUR 30.61 million, with an average of 0.40 million cattle; in Estonia, it was EUR 20.61 million (0.25 million head); in Lithuania, it was EUR 36.28 million (0.66 million head). The ratio of exports to live bovine animal population in the Baltic States was significantly higher than in the large EU–23 Member States. Cattle production in the Baltic States is more oriented towards export markets than in other EU–23 Member States, given the relatively small domestic market. During 2014–2024, the opposite trend was observed in live bovine animal exports in Spain, Italy, Greece, Portugal, Bulgaria, and Romania (see Table 1, Figure. S3), with low export volumes indicating the dominance of domestic consumption or restrictions on the production system. Overall, the average values of live bovine animal exports and the average live bovine animal population show differences in export intensity across the EU–23 Member States. The next section analyses in live bovine animal exports, in more detail, changes for the EU–23 Member States. Evolution of live bovine animal exports in the EU–23 member states Competitiveness in live bovine animal exports for 2014–2024 revealed differences among EU Member States. The differences may be related to the impact of the CAP 2023–2027, particularly with respect to climate change and the implementation of environmental measures in the cattle sector (European Commission, 2025). The CAP recognises and supports the role of organic farming through a cattle market (Nowak & Kobiałka, 2024). Cattle farming, part of a global food supply chain, is influenced by global trade conditions, disrupting production, and increasing environmental pressure. Therefore, a growing need for a globally coordinated and environmentally friendly trade policy (Terpou et al., 2025). Changes in live cattle exports between the EU–23 Member States reveal structural differences over the years. These differences reflect variations in market orientation and production structure (Terpou et al., 2025). The annual and base-year changes, as well as the live bovine animal population in the EU–23 Member States, did not follow a single linear pattern, but rather reflected different groups of Member States: 1. EU–23 Member States and years experiencing a strong increase in exports; 2. EU–23 Member States and years with stable or proportionate annual and base-year growth in live bovine animal exports. A comparison of the average base-year growth (%) of the live bovine animal population with the average annual growth of live bovine exports in tonnes (%) in the EU – 23 Member States in 2014–2024 (see Figure. S4) revealed significant changes in live bovine animal exports, particularly in 2018 when the base-year growth was almost 6.8 times higher than in 2017 and in 2023, 8.8 times higher compared with the previous year. In 2024, compared with 2017, the live bovine animal population in the EU–23 Member States decreased by 2%, accounting for 92.95% of the total live bovine animals recorded in 2014. An increase in exports was not achieved through an expansion, but rather through production optimisation, increased animal weight, and responses to market demand and trade conditions. A similar pattern is evident in the evolution of export value (Figure. S5), where the average annual growth of EU–23 Member States exports increased significantly in 2018 (2.99 times that of 2017) and 2023 (5.16 times that of 2022). The overall base-year growth indicated long-term structural development with live bovine animal exports in 2024 exceeding the 2014 level by 3.4 times. The increase in live bovine animal exports reflects changes in market demand and production systems. This may be partly related to external factors such as COVID-19 restrictions and the CAP 2023–2027 requirements for animal welfare and the environmental standards (Adams et al., 2025; European Commission, 2025). Case studies: Greece and Croatia To illustrate the structural patterns identified in the analysis, Greece and Croatia were selected due to their high base growth rates in exports of live bovine animals. In both countries, the average base growth rates increased five to six times compared to the initial period, distinguishing them from the majority of EU–23 Member States and justifying closer examination. In Greece, live bovine exports fluctuated, decreasing in 2017 (EUR 1.8 thousand) and increasing substantially in 2018 (EUR 83.2 thousand). In 2022, exports decreased to EUR 10.2 thousand, followed by a significant increase to EUR 969.9 thousand in 2023 and EUR 1.48 million in 2024. During the same period, the live bovine animal population remained relatively stable, with a slight decrease from 659 thousand in 2014 to 595.2 thousand in 2024. A similar pattern was observed in Croatia, where the value of live bovine animal exports increased from EUR 5.25 million in 2014 to EUR 66.33 million in 2024, while the live bovine animal population remained relatively stable from 441 thousand to 415 thousand. The export growth in both countries occurred without a corresponding increase in the live bovine animal population, compared to other EU–23 Member States, where live bovine animal population and export figures were different. The decrease in the live bovine population in Greece can be explained by the spread of Lumpy skin disease (LSD) in 2015, which is an infectious disease of cattle caused by the LSD virus genus Capripoxvirus family Poxviridae. When LSD is detected in a Member State, strict livestock export restrictions are introduced on live cattle and related products, significantly affecting the cattle sector (Tasioudi et al., 2016). Similarly, fluctuation in Croatia’s live bovine animal exports reflected broader structural challenges in the livestock sector. Since 2020, production costs have increased in the livestock sector, influenced by the COVID-19 pandemic and the war in Ukraine. In 2022, foreign trade in livestock products was also in deficit: EUR 3.7 million for live cattle. EUR 435 million for meat products. In live cattle imports, the largest share consisted of fattening cattle weighing 80–160 kg, whereas in live cattle exports, slaughter cattle, whose weight exceeded 300 kg, accounted for the largest share in 2022. It is important to note that the cattle sector was constrained by high costs and a shortage of calves, which reduced its profitability. In 2023, an additional problem was African swine fever, which reduced trade in live animals and meat products in Croatia (Mikuš et al., 2024). The average base-year growth rates for live bovine animal exports (2014–2024) were as follows: Greece, 516%; Croatia, 613%. Due to the high volatility observed in annual export values, the compound annual growth rate (CAGR) was used to analyse changes, thereby smoothing short-term fluctuations, and identifying trends in the development of live bovine animal export. These characteristics explain the classification of Greece and Croatia into a separate cluster in the K-means analysis. Trends in live bovine animal exports (CAGR analysis) As shown in Table 2 and Figure. S6, the CAGR for the live bovine animal population in almost all EU–23 Member States was low, close to zero, or slightly negative. This suggests a relatively stable or slightly decreasing live bovine animal population during the analysis period. In contrast, the CAGR for live bovine animal exports showed significant dispersion across the EU–23 Member States (see Table 2 and Figure. S6), in most of them, live bovine animal exports increased. The largest increase was reported in Greece (38.97%), Croatia (28.86%), Estonia (18.49%), and Latvia (16.41%). In some countries, such as Ireland and Spain, a strong increase in exports was observed (CAGR > 10%); however, the increase in live bovine population was minimal at 0.10% in Ireland and 0.82% in Spain.Table 2Compound Annual Growth Rate (CAGR) of live bovine animal export value and volume, population in EU–23 Member States, 2014–2024.Table 2 dummy alt textNr.EU Member StateExport value CAGR, %Export volume CAGR, %Population CAGR, %1Greece38.9735.60−1.012Croatia28.8622.50−0.443Estonia18.4911.07−1.294Latvia16.418.35−1.805Ireland10.297.310.106Hungary10.246.020.827Slovenia9.993.70−0.318Lithuania8.606.90−2.039Czechia6.542.770.1810Belgium6.481.40−1.3911Netherlands6.446.45−1.5612Portugal5.951.41−0.4113Slovakia4.890.55−1.1014France3.650.27−1.5415Germany3.46−0.39−1.9516Spain3.04−1.190.1617Denmark3.020.67−0.9318Italy2.47−2.73−0.6019Luxembourg1.260.59−1.0820Austria−1.30−5.19−0.7421Romania−3.57−6.18−1.3322Poland−4.16−7.900.9023Bulgaria−15.13−20.48−0.06Note. CAGR -compound annual growth rate. Values represent average growth rates over the period 2014–2024.Source: Authors’ calculations based on Eurostat, 2014–2024. Table 2Compound Annual Growth Rate (CAGR) of live bovine animal export value and volume, population in EU–23 Member States, 2014–2024.Table 2 dummy alt textNr.EU Member StateExport value CAGR, %Export volume CAGR, %Population CAGR, %1Greece38.9735.60−1.012Croatia28.8622.50−0.443Estonia18.4911.07−1.294Latvia16.418.35−1.805Ireland10.297.310.106Hungary10.246.020.827Slovenia9.993.70−0.318Lithuania8.606.90−2.039Czechia6.542.770.1810Belgium6.481.40−1.3911Netherlands6.446.45−1.5612Portugal5.951.41−0.4113Slovakia4.890.55−1.1014France3.650.27−1.5415Germany3.46−0.39−1.9516Spain3.04−1.190.1617Denmark3.020.67−0.9318Italy2.47−2.73−0.6019Luxembourg1.260.59−1.0820Austria−1.30−5.19−0.7421Romania−3.57−6.18−1.3322Poland−4.16−7.900.9023Bulgaria−15.13−20.48−0.06Note. CAGR -compound annual growth rate. Values represent average growth rates over the period 2014–2024.Source: Authors’ calculations based on Eurostat, 2014–2024. Compound Annual Growth Rate (CAGR) of live bovine animal export value and volume, population in EU–23 Member States, 2014–2024. Note. CAGR -compound annual growth rate. Values represent average growth rates over the period 2014–2024. Source: Authors’ calculations based on Eurostat, 2014–2024. Live bovine animal exports decreased in Bulgaria (−15.13%) and Austria (−1.30%), where the CAGR for the live bovine animal populations was low: Bulgaria (−0.06%) and Austria (−0.74%), which may indicate a recession in the cattle sector or a market reorientation (see Table 2). In Poland, the total value of live bovine animal exports decreased by 4.16% per year on average during the analysis period, whereas the live bovine animal population increased by 0.90% per year (see Table 2). Poland may have shifted towards the domestic market or processing, and exports consisted of live bovine animals with lower unit values. Accordingly, the cattle industry increasingly focused on the domestic market and product processing rather than on the export of live bovine animals (Hocquette & Chatellier, 2011; Terpou et al., 2025). The Baltic States demonstrated a similar structural model in the development of live bovine animal exports (see Table 2 and Figure. S6). In all three countries, the live bovine animal population decreased, while export values and volumes increased, indicating an orientation towards foreign markets. Lithuania was characterised by stable export growth with exports increasing at a CAGR of 8.60% in value and 6.90% in volume despite the fact that the live bovine animal population decreased by 2.03%. Latvia showed a more dynamic trend, with export growing at 16.41% in value, while the live bovine animal population decreased by 1.80% per year. Estonia, although the smallest in absolute terms, exhibited the highest export intensity, with export values increasing at a CAGR of 18.49%. volumes at 11.07% while the live bovine animal population decline remained relatively moderate (1.29%). Overall, the Baltic States demonstrate a consistent pattern in which exports of bovine animals expand despite decreasing live bovine animal population, suggesting increasing export specialization and competitiveness in external markets. After processing EU–23 Member States data on the growth of live bovine animal exports in both value and volume terms, and comparing these with changes in the total live bovine animal population, the figures revealed that live bovine animal exports continued to grow despite stagnation or declines in the total population. This provides a basis for further cluster analysis. Formation of groups of the EU–23 member states for growth in the live bovine animal sector Correlation analysis The correlation analysis (Table 3) revealed a strong relationship between the export value and volume of live bovine animal exports in tonnes (r = 0.998, p < 0.001). Live bovine animal exports in the EU–23 Member States were predominantly driven by volume. A correlation was also observed between export value and the live bovine animal population (r = 0.790; p < 0.001), suggesting that higher export volume and a larger live bovine animal population are associated with higher export value. Similarly, the export volume of live bovine animals (in tonnes) was correlated with the live bovine animal population (r = 0.777; p < 0.001), indicating that higher numbers of live bovine animals in EU–23 Member States are associated with higher volumes of live bovine animal exports.Table 3Pearson correlation matrix of key export variables in the EU–23 Member States, 2014–2024.Table 3 dummy alt textVariableStatisticsExport value, EUR millionLive bovine animal population. million headExport volume, tonnesExport intensity, EUR million per million headExport value growth rate (CAGR)Export volume growth rate (CAGR)Live bovine animal population growth rate (CAGR)Export value, EUR millionPearson Correlation10.790⁎⁎0.998⁎⁎0.160−0.094−0.069−0.251Sig. (2-tailed)0.0000.0000.4660.6700.7550.249N23232323232323Live bovine animal population, million headPearson Correlation0.790⁎⁎10.777⁎⁎−0.263−0.224−0.201−0.116Sig. (2-tailed)0.000-0.0000.2250.3030.3580.599N23232323232323Export volume, tonnesPearson Correlation0.998⁎⁎0.777⁎⁎10.192−0.093−0.064−0.255Sig. (2-tailed)0.0000.000-0.3800.6740.7720.240N23232323232323Export intensity, EUR million per million headPearson Correlation0.160−0.2630.19210.1640.156−0.163Sig. (2-tailed)0.4660.2250.380-0.4550.4780.459N23232323232323Export value growth rate (CAGR)Pearson Correlation−0.094−0.224−0.0930.16410.984⁎⁎−0.159Sig. (2-tailed)0.6700.3030.6740.455-0.0000.470N23232323232323Export volume growth rate (CAGR)Pearson Correlation−0.069−0.201−0.0640.1560.984⁎⁎1−0.186Sig. (2-tailed)0.7550.3580.7720.4780.000-0.396N23232323232323Live bovine animal population growth rate (CAGR)Pearson Correlation−0.251−0.116−0.255−0.163−0.159−0.1861Sig. (2-tailed)0.2490.5990.2400.4590.4700.396-N23232323232323** Correlation is significant at the 0.01 level (2-tailed).Note: Calculations performed using IBM SPSS Statistics Version 29.0.2.0.Source: Authors’ calculations based on Eurostat, 2014–2024. Table 3Pearson correlation matrix of key export variables in the EU–23 Member States, 2014–2024.Table 3 dummy alt textVariableStatisticsExport value, EUR millionLive bovine animal population. million headExport volume, tonnesExport intensity, EUR million per million headExport value growth rate (CAGR)Export volume growth rate (CAGR)Live bovine animal population growth rate (CAGR)Export value, EUR millionPearson Correlation10.790⁎⁎0.998⁎⁎0.160−0.094−0.069−0.251Sig. (2-tailed)0.0000.0000.4660.6700.7550.249N23232323232323Live bovine animal population, million headPearson Correlation0.790⁎⁎10.777⁎⁎−0.263−0.224−0.201−0.116Sig. (2-tailed)0.000-0.0000.2250.3030.3580.599N23232323232323Export volume, tonnesPearson Correlation0.998⁎⁎0.777⁎⁎10.192−0.093−0.064−0.255Sig. (2-tailed)0.0000.000-0.3800.6740.7720.240N23232323232323Export intensity, EUR million per million headPearson Correlation0.160−0.2630.19210.1640.156−0.163Sig. (2-tailed)0.4660.2250.380-0.4550.4780.459N23232323232323Export value growth rate (CAGR)Pearson Correlation−0.094−0.224−0.0930.16410.984⁎⁎−0.159Sig. (2-tailed)0.6700.3030.6740.455-0.0000.470N23232323232323Export volume growth rate (CAGR)Pearson Correlation−0.069−0.201−0.0640.1560.984⁎⁎1−0.186Sig. (2-tailed)0.7550.3580.7720.4780.000-0.396N23232323232323Live bovine animal population growth rate (CAGR)Pearson Correlation−0.251−0.116−0.255−0.163−0.159−0.1861Sig. (2-tailed)0.2490.5990.2400.4590.4700.396-N23232323232323** Correlation is significant at the 0.01 level (2-tailed).Note: Calculations performed using IBM SPSS Statistics Version 29.0.2.0.Source: Authors’ calculations based on Eurostat, 2014–2024. Pearson correlation matrix of key export variables in the EU–23 Member States, 2014–2024. ** Correlation is significant at the 0.01 level (2-tailed). Note: Calculations performed using IBM SPSS Statistics Version 29.0.2.0. Source: Authors’ calculations based on Eurostat, 2014–2024. A strong relationship was found between the annual growth rates of exports over the analysis period. The growth rate of export value (CAGR) was almost identical to the growth rate of export volume in tonnes (CAGR) (r = 0.984; p < 0.001). Export value and export volume increased at a similar rate. In contrast, export intensity (million EUR per million live bovine animals) showed a weak negative relationship with the live bovine animal population (r = −0.263; p = 0.225). This relationship was not statistically significant. EU–23 Member States with larger live bovine animal population tended to have lower export intensity, indicating a greater role of live bovine animal consumption. Whereas, smaller EU–23 Member States tend to be more export-oriented. The export volume growth rate (CAGR) for the live bovine animal population showed a weak, negative correlation with export volume (r = −0.251; p = 0.249). The EU–23 Member States with higher export volumes of live bovine animals tend to have lower live bovine animal population growth. Overall, the correlation analysis shows relationships between live bovine animal export value and export volume as well as between their growth rates, while export intensity reflects different structural dimensions of the export system. Hierarchical clustering Hierarchical cluster analysis was performed using Ward’s linkage method and squared Euclidean distance based on four standardized indicators: average live bovine animal export value, export intensity, CAGR of export value, and CAGR of the live bovine animal population. The number of clusters was determined based on dendrogram structure and changes in the agglomeration coefficient. The hierarchical analysis identified three groups. A significant increase in the agglomeration coefficient suggested that clusters with different internal structures. High heterogeneity in live bovine animal exports across the EU–23 Member States was combined at this stage. Among the 23 Member States of the EU, France demonstrated the highest indicator value with respect to live bovine animal exports Z = 4.46 (see Table S3). In contrast, export intensity (Z = 0.55), export CAGR (Z = −0.32), and live bovine animal population CAGR (Z = −0.94) remained within a comparable range relative to other EU Member States. The data indicate that export value was the main factor influencing the distance between countries in the cluster analysis. During the cluster formation process, several clearly distinguishable groups of the EU–23 Member States were identified, with France consistently separated from the other EU Member States and differing statistically from them. The cluster analysis showed that France formed an isolated cluster, which provided justification for its exclusion from subsequent analysis. The final clustering was performed on the EU–22 Member States to improve cluster stability. The agglomeration schedule of hierarchical cluster analysis for EU–22 Member States (see Table S4) shows that up to step 14 of the cluster formation process, a relatively small increase in the distance coefficient was observed, thereby smoothly combining similar EU Member States into clusters. From steps 16–18, the distance coefficient increased significantly from 14,414 to 23,285, then to 31,153 (step 19) and 44,448 (step 20), and finally to 65,818 (step 21), which was used to select the number of clusters (see Table S4). The sequence of cluster formation and the distances between them are visualised in a dendrogram (see Figure. S5) with the horizontal axis representing quadratic Euclidean distances between the clusters (scale 0–25). The optimal number of clusters was determined based on the dendrogram structure and an increase in the agglomeration coefficient, which indicated separation into three clusters. The three-cluster solution was also consistent with the elbow criterion based on changes in the agglomeration coefficient. The dendrogram shows three main clusters, which were used as the basis for the K-means clustering procedure. The first cluster, characterised by short quadratic Euclidean distances (∼2), includes Belgium, the Netherlands, Lithuania, Germany, Estonia, Latvia, Luxembourg, Slovakia, the Czech Republic, Slovenia, and Hungary (n = 11). The cluster includes pairs of EU Member States, such as Belgium and the Netherlands, Estonia, and Latvia, which could be combined at the initial clustering stage, reflecting similar structural changes. These EU Member States were relatively close, thereby forming a homogeneous regional sub-cluster and a high degree of mutual similarity in live bovine animal export characteristics compared with other EU Member States. The second cluster combined Croatia and Greece (n = 2). Both Member States were characterised by internal similarity and formed a compact cluster that was visually separated from the other Member States (see Figure. S5). This export profile for live bovine animals was similar but significantly different from that of the other EU Member States. The third cluster combined Austria, Denmark, Italy, Spain, Portugal, Poland, Romania, Ireland, and Bulgaria (n = 9) (Figure. S5). The internal structure of this cluster was highly heterogeneous, comprising both highly developed Western EU Member States and fast-growing Eastern EU Member States. Despite significant internal differences in export volumes and composition, the overall cluster profile of the EU Member States was sufficiently similar to warrant a single cluster. Based on a dendrogram and changes in the agglomeration coefficient, for further analyse the optimal three-cluster solution was selected and used in the subsequent K-means analysis. Clustering and validation of live bovine export dynamics in EU member states K-means clustering was performed using four standardised indicators: average export value, export intensity, CAGR of export value, and CAGR of live bovine animal population, based on the sample of EU–22 Member States, selected to capture both export performance and structural characteristics of the cattle sector. The results identified three structurally distinct patterns in live bovine animal exports across EU–22 Member States, reflecting statistically significant differences in key clustering variables. The comparative characteristics of the identified clusters were expressed as standardised Z-scores of export intensity, export growth (CAGR), and live bovine animal population growth, and provide a basis for structural disparity between clusters (see Table 4).Table 4Comparative characteristics of clusters of EU-22 Member States, 2014–2024.Table 4 dummy alt textClusternEU Member StatesZ-score Export IntensityZ-scoreCAGRExport valueZ-score Live bovine animal population CAGRCluster 1High Intensity10SK, BE, EE, LU, SI, LV, NL, CZ, LT, and HU0.880.16−0.24Cluster 2Rapid Growth2HR and GR−0.322.43−0.61Cluster 3Low intensity10AT, BG, DK, DE, IT, IE, PL, PT, RO, and ES−0.87−0.610.32Note: Values represent standardized Z-scores. Differences between clusters were evaluated using the Tukey HSD post-hoc test at α = 0.05.Note: Cluster membership was determined using the K-means algorithm based on four standardized indicators: average export value, export intensity, CAGR of export value, and CAGR of the live bovine animal population. IBM SPSS Statistics Version 29.0.2.0.Source: Author’s calculations based on Eurostat data, 2014– 2024. Table 4Comparative characteristics of clusters of EU-22 Member States, 2014–2024.Table 4 dummy alt textClusternEU Member StatesZ-score Export IntensityZ-scoreCAGRExport valueZ-score Live bovine animal population CAGRCluster 1High Intensity10SK, BE, EE, LU, SI, LV, NL, CZ, LT, and HU0.880.16−0.24Cluster 2Rapid Growth2HR and GR−0.322.43−0.61Cluster 3Low intensity10AT, BG, DK, DE, IT, IE, PL, PT, RO, and ES−0.87−0.610.32Note: Values represent standardized Z-scores. Differences between clusters were evaluated using the Tukey HSD post-hoc test at α = 0.05.Note: Cluster membership was determined using the K-means algorithm based on four standardized indicators: average export value, export intensity, CAGR of export value, and CAGR of the live bovine animal population. IBM SPSS Statistics Version 29.0.2.0.Source: Author’s calculations based on Eurostat data, 2014– 2024. Comparative characteristics of clusters of EU-22 Member States, 2014–2024. Note: Values represent standardized Z-scores. Differences between clusters were evaluated using the Tukey HSD post-hoc test at α = 0.05. Note: Cluster membership was determined using the K-means algorithm based on four standardized indicators: average export value, export intensity, CAGR of export value, and CAGR of the live bovine animal population. IBM SPSS Statistics Version 29.0.2.0. Source: Author’s calculations based on Eurostat data, 2014– 2024. As shown in Table 4, the clusters differ mainly in terms of export intensity and export growth rates. The ANOVA results indicate that these two variables differ statistically between clusters, whereas the remaining variables do not. Cluster 1 is characterised by high export intensity (Z = 0.88) and moderate export growth (CAGR Z = 0.16), reflecting a strong export orientation compared to other clusters. Cluster 2 shows very rapid export growth (CAGR Z = 2.43) despite relatively low export intensity (Z = −0.32), that this cluster is driven mainly by growth dynamics rather than export intensity. This cluster is primarily distinguished by export growth (CAGR) rather than export intensity. This is consistent with the Tukey results, which show no statistically significant difference in export intensity between clusters 2 and 3. In contrast, Cluster 3 is characterised by low export intensity (Z = −0.87) and negative export growth (CAGR Z = −0.61), indicating a more domestically oriented bovine sector with lower dependence on live bovine animal exports across EU –22 Member States. The distances between cluster centroids (Table 5) indicate differences between clusters. The smallest distance was observed between clusters 1 and 3 (2.004), while the largest distance was found between clusters 2 and 3 (3.112), reflecting more pronounced export differences in development patterns between these groups of EU Member States. Before performing ANOVA, the assumptions of normality and homogeneity of variance were assessed using the Shapiro–Wilk and Levene tests. The Shapiro–Wilk test indicated minor deviations from normality for the export value indicator (p = 0.005–0.016), whereas the remaining variables did not deviate significantly from normality (p > 0.05). The Levene test showed homogeneity of variances across clusters for all variables (export value: F(2.19) = 1.658. p = 0.217; export CAGR: F(2.19) = 0.314. p = 0.734; population CAGR: F(2.19) = 0.640. p = 0.538).Table 5Cluster centroids and distances for EU–22 Member States, in 2014–2024.Table 5 dummy alt textVariablesCluster123Final Cluster CentresZ-scoreAverage live bovine animal export value, million EUR−0.13503−0.37772−0.23584Z-scoreAverage Export Intensity0.88350−0.32240−0.87427Z-score CAGRExport value, million EUR0.160442.43102−0.61464Z-scoreCAGRlive bovine animal population−0.238140.038560.32428Distances Between Cluster CentresCluster 1n/a2.5972.004Cluster 22.597n/a3.112Cluster 32.0043.112n/aNote: K-means clustering using four standardized indicators: average export value, export intensity, industry CAGR, million EUR, and CAGR, number of heads. IBM SPSS Statistics Version 29.0.2.0.Source: Author’s calculations based on Eurostat, 2014–2024. Table 5Cluster centroids and distances for EU–22 Member States, in 2014–2024.Table 5 dummy alt textVariablesCluster123Final Cluster CentresZ-scoreAverage live bovine animal export value, million EUR−0.13503−0.37772−0.23584Z-scoreAverage Export Intensity0.88350−0.32240−0.87427Z-score CAGRExport value, million EUR0.160442.43102−0.61464Z-scoreCAGRlive bovine animal population−0.238140.038560.32428Distances Between Cluster CentresCluster 1n/a2.5972.004Cluster 22.597n/a3.112Cluster 32.0043.112n/aNote: K-means clustering using four standardized indicators: average export value, export intensity, industry CAGR, million EUR, and CAGR, number of heads. IBM SPSS Statistics Version 29.0.2.0.Source: Author’s calculations based on Eurostat, 2014–2024. Cluster centroids and distances for EU–22 Member States, in 2014–2024. Note: K-means clustering using four standardized indicators: average export value, export intensity, industry CAGR, million EUR, and CAGR, number of heads. IBM SPSS Statistics Version 29.0.2.0. Source: Author’s calculations based on Eurostat, 2014–2024. The results of ANOVA (see Table 6) showed that two variables: export intensity (F = 24.614; p < 0.001) and CAGR (million EUR) (F = 24.919; p < 0.001) differed significantly between clusters.Table 6ANOVA of EU–22 Member States live bovine animal export, 2014–2024.Table 6 dummy alt textANOVAVariablesSum of SquaresdfMean SquareFSig.η²(Effect size)Z-score Average live bovine animals export value, millionBetween Groups0.11820.0591.0710.3620.101Within Groups1.047190.055---Total1.16521----Z-score Average Export IntensityBetween Groups15.64327.82224.614<0.0010.724Within Groups6.038190.318---Total21.68121----Z-score CAGR Export, millionBetween Groups15.85027.92524.919<0.0010.724Within Groups6.043190.318---Total21.89321----Z-score CAGR live bovine animal populationBetween Groups1.58220.7910.7710.4770.075Within Groups19.497191.026---Total21.07921----Note: K-means clustering using four standardized indicators: average export value, export intensity, industry CAGR, million EUR, and CAGR of live bovine animal population, number of heads. IBM SPSS Statistics Version 29.0.2.0. Statistical significance was evaluated at α = 0.05.Source: Author’s calculations based on Eurostat data, 2014–2024. Table 6ANOVA of EU–22 Member States live bovine animal export, 2014–2024.Table 6 dummy alt textANOVAVariablesSum of SquaresdfMean SquareFSig.η²(Effect size)Z-score Average live bovine animals export value, millionBetween Groups0.11820.0591.0710.3620.101Within Groups1.047190.055---Total1.16521----Z-score Average Export IntensityBetween Groups15.64327.82224.614<0.0010.724Within Groups6.038190.318---Total21.68121----Z-score CAGR Export, millionBetween Groups15.85027.92524.919<0.0010.724Within Groups6.043190.318---Total21.89321----Z-score CAGR live bovine animal populationBetween Groups1.58220.7910.7710.4770.075Within Groups19.497191.026---Total21.07921----Note: K-means clustering using four standardized indicators: average export value, export intensity, industry CAGR, million EUR, and CAGR of live bovine animal population, number of heads. IBM SPSS Statistics Version 29.0.2.0. Statistical significance was evaluated at α = 0.05.Source: Author’s calculations based on Eurostat data, 2014–2024. ANOVA of EU–22 Member States live bovine animal export, 2014–2024. Note: K-means clustering using four standardized indicators: average export value, export intensity, industry CAGR, million EUR, and CAGR of live bovine animal population, number of heads. IBM SPSS Statistics Version 29.0.2.0. Statistical significance was evaluated at α = 0.05. Source: Author’s calculations based on Eurostat data, 2014–2024. These results export intensity and export growth were the key variables differentiating the identified clusters. In contrast, the average volume of exports (F = 1.071; p = 0.362) and the change in the live bovine animal population (F = 0.771; p = 0.477) were not statistically significant. Large effect sizes were observed for export intensity and export growth (η² = 0.724), with strong differentiation between clusters in export intensity and the CAGR of export value growth. The stability of the cluster solution was additionally assessed using the silhouette coefficient. The four-variable standardized indicators (Z-scores): average export value, export intensity, CAGR of export value, and CAGR of live bovine animal population model produced a silhouette coefficient of approximately 0.52, reflecting a moderate cluster structure, with acceptable cohesion within clusters and separation between clusters. The distribution of EU–22 Member States across clusters was uneven. The largest cluster includes 45.5%, while the smallest cluster consists 13.6%. The ratio between the largest and smallest cluster was 3.33, indicating a relatively small cluster. To assess the robustness of the clustering results, an additional Two Steps cluster analysis was performed using a reduced set of two key z-scores standardised variables, as export intensity and export value growth CAGR. The silhouette coefficient increased to approximately 0.60, suggesting a stronger cluster structure compared to the four-variable model. The largest cluster includes 49.5%, while the smallest cluster consists 27.3%. The ratio between the largest and smallest cluster was 1.50. Both models identified three clusters; the distribution of EU–22 Member States across clusters was slightly different. Indicates that export intensity and export growth are the primary drivers of cluster formation. The two-variable model improved statistical separation, while the four-variable model was retained for further analysis as it provides a more comprehensive presentation of structural characteristics of live bovine animal exports, including population dynamics. The two-variable model resulted in more balanced clusters; however, the four-variable hierarchical and K-means solutions retained a smaller, more distinct cluster. The results of the hierarchical cluster analysis showed that Germany (item No 9) (Figure. S5) was grouped with Belgium and the Netherlands at an early stage of cluster formation, with the distance coefficient of 2.104 (Table 5). The dendrogram (see Figure. S5) revealed this hierarchical structure with Germany grouped with the 11 EU Member States in the first cluster. However, the K-means clustering method, which optimises distances to cluster centroids, resulted in a different classification. For Germany the Euclidean distance (1.855) to the centroid of Cluster 3 was smaller than to Cluster 1, which led to its reassignment to Cluster 3 (see Table S4). Overall, revealed that the growth of live bovine animal exports in the EU Member States was structurally differentiated and follows different patterns. These differences support the use of cluster analysis for testing the hypothesis. Sustainability implications of live bovine animal export structures in the EU-22 member states The cluster analysis results provided the basis for discussing potential sustainability risks associated with different export structures in the EU–22 Member States. The first cluster (n = 10) was characterised (see Table 4) by a high export intensity (Z = 0.883), moderate positive growth in export value (CAGR million EUR. Z = 0.160) and a decline in the live bovine animal population (CAGR for live bovine animal population. Z = −0.238). This group included Belgium, the Czech Republic, Estonia, Hungary, Latvia, Lithuania, Luxembourg, the Netherlands, Slovakia, and Slovenia. The second cluster (n = 2) included Croatia and Greece (see Table 4). These EU Member States were characterised by a rapid increased in the value of live cattle exports (CAGR, million EUR, Z = 2.43), while maintaining a relatively stable live bovine animal population (CAGR for live bovine population, Z = 0.038), and a low export intensity (Z = −0.322). Export growth in these Member States appears high, but is largely due to a low initial base, characterised by relatively small export volumes and a low level of the live bovine animal population in the base year. In this case, even modest increases result in high growth rates. This structural profile places Croatia and Greece in a separate – growth-driven cluster. The third cluster (n = 10) included Austria, Bulgaria, Denmark, Germany, Italy, Ireland, Poland, Portugal, Romania, and Spain (see Table 4) characterised a low average export intensity (Z = −0.874), a stable or moderately increasing live bovine animal population (CAGR for live bovine population, Z = 0.324), and a negative growth rate of export value (CAGR, million EUR, Z = −0.615). Suggests a stronger orientation towards the domestic market and processing activities rather than live bovine animal exports. The K-means clustering showed that the EU –22 could be divided into three distinct clusters according to the intensity of live bovine animal exports and export growth rates (see Fig. 1).Fig. 1Classification of EU–22 Member States by K-means clusters and distance-to-centroid values, 2014–2024.Fig 1 dummy alt text Classification of EU–22 Member States by K-means clusters and distance-to-centroid values, 2014–2024. Note: Grey-shaded countries indicate EU Member States not included in the analysis. K-means clustering performed using IBM SPSS Statistics Version 29.0.2.0. based on four standardised indicators: average export volume, export intensity, compound annual growth rate (CAGR) of export value, million EUR, and CAGR of live bovine animal populations, number of heads. Distance values indicate proximity to cluster centroids. Source: Author's calculations based on Eurostat, 2014–2024. The ANOVA results showed that export intensity differed significantly across clusters (F(2.19) = 24.614, p < 0.001). The Tukey HSD post hoc test shows that Cluster 1 differed significantly from Cluster 2 (p = 0.032) and Cluster 3 (p < 0.001). No statistically significant differences were found between Clusters 2 and Cluster 3 in terms of export intensity (p = 0.432). Suggests that Cluster 2 is distinguished mainly by export growth rather than export intensity. Showed that export intensity emerges as a key factor in differentiating the clusters. The first cluster is the only one with a higher export intensity than the other clusters. At the same time, this division may have several advantages, including market prices, liquidity, and demand elasticity. However, the structural profile of the cluster also allows the identification of potential sustainability risks associated with the long-term resilience of live bovine animal exports. Particularly where high export intensity is combined with a declining live bovine animal population. In the cattle sector, with a perennial biological cycle and time-consuming herd rebuilding, such a development model may be considered structurally sensitive, as exports may compete with domestic consumption and production capacity. The cluster's high export intensity is not necessarily associated with a simultaneous increase in production capacity or the live bovine animal population. May indicate increased pressure on the sustainability and resilience of the cattle sector. In the long term, it could lead to several potential risks that require further empirical investigation: (1) possible decrease in organic production; (2) increased dependence on imported genetic and feed resources, and (3) reduced flexibility in export markets or in the event of policy shocks (Cowley, 2021; Greenwood, 2021; Rich, 2024), that the cattle sector could become particularly sensitive to external shocks, including fluctuations in international prices, restrictions on export markets, or changes in veterinary and trade regulations. Discussion A structurally differentiated pattern of live bovine animal exports across EU Member States from 2014 to 2024 highlights both changes in exports and differences among EU Member States, as well as the absence of a uniform development trajectory within the sector. Live bovine animal exports from the EU-23 Member States increased in terms of both value and volume over the period 2014 to 2024, while cattle export value increased despite a decline in the live bovine animal population, consistent with (OECD/FAO, 2024) forecasts, that agricultural growth in advanced economies is increasingly driven by higher productivity rather than by output growth. A consideration of methodology France is exclusion from the final cluster analysis. Hierarchical clustering consistently classified France as a statistical outlier, exceptionally high export value and cattle population. France including would have affected the homogeneity of the clusters and reduced the interpretability of the results. Nevertheless, France is structurally significant in the EU cattle farming sector and deserves separate consideration in policy discussions, given its scale and central role in European livestock production. Four Member States: Sweden, Finland, Malta, and Cyprus were excluded due to incomplete data coverage over the study period. While this decision was necessary to ensure consistency and comparability of the time series indicators, it introduces a limitation in terms of fully capturing the diversity of the EU. Based on their structural characteristics, these countries would likely align with low-intensity or domestically oriented clusters; however, this assumption requires further empirical validation. The correlation analysis is based on a small sample size (n = 23), and therefore the results are interpreted cautiously, providing only indicative rather than definitive relationships. The cluster export performance is not explained by cattle population, but rather by a combination of export intensity and export growth dynamics. The results of the study reveal three distinct structural profiles among EU Member States. The first group is characterised by high export intensity alongside declining cattle populations. This points to a focus on external cattle markets, even in situations where the renewal of live bovine populations may be limited. The second group, which includes Croatia and Greece, stands out for particularly high export growth rates (CAGR). However, growth was mainly reflected by a low initial export base, rather than by consistently high export intensity. The third group consists of EU Member States with low export intensity and relatively stable or moderately increasing cattle populations. This suggests a stronger orientation towards domestic markets and processing activities, rather than export expansion. The export growth in the EU cattle sector is not necessarily linked to an increase in the live bovine animal population. This pattern, where export values rise while cattle populations remain stable or decrease, may result from structural changes in the cattle sector, such as improvements in live bovine animal productivity, shifts in herd composition, or higher unit values of exported animals. This study did not directly analyse these factors. From a sustainability perspective, the cluster structure allows discussion of possible implications without implying causal relationships. Clusters with high export intensity, the combination of strong export orientation and decreased cattle populations, may raise questions about long-term sustainability and production capacity. High-growth clusters may represent more or less stable system, where rapid export intensity increases are influenced by a low initial live bovine animal population base, rather than stable product capacity. Low-intensity clusters may be less integrated into international markets, but could benefit from more resilient domestic supply chains. As several studies suggest, the sustainability and resilience of the cattle sector in the long term could pose risks requiring further empirical investigation, such as a possible decrease in organic production. Increased dependence on imported genetic and feed resources or reduced flexibility in export markets or in the event of policy shocks. The cattle sector could become particularly sensitive to external shocks, including fluctuations in international prices, restrictions on export markets, or changes in veterinary and trade regulations (Cowley, 2021; Greenwood, 2021; Hocquette et al., 2018; Hocquette & Chatellier, 2011; Terpou et al., 2025; Vinci, 2022). Study limitations and future research. The analysis is primarily based on aggregated live bovine animal exports and live bovine animal population data, which may mask structural differences at the EU Member State level. At the same time, variations in data quality and reporting across the EU Member States may affect the comparability of some indicators. The study focuses on export dynamics and does not directly analyse farm-level productivity, production costs, or environmental performance. Future research could take this analysis further by applying panel econometric models (e.g., fixed-effects models) to quantify the causal relationships between EU Member States country-specific economic variables and export intensity. Subsequent studies could also explore the link between export-oriented live bovine animal (beef breed cattle) farming and biodiversity protection, as well as animal welfare standards across different regions of the EU. Such research would provide a deeper understanding of how trade dynamics interact with sustainability goals in the changing European agricultural policy framework. Conclusion This study provides an overview of the structural dynamics of live bovine animal exports in the EU-23 Member States over the period from 2014 to 2024. The results show that export growth occurred despite a decline in the live cattle population. The cluster model of live bovine animal exports across three clusters highlights the structural heterogeneity of the EU live cattle sector and demonstrates the need for a differentiated policy approach to Member States national production systems, market orientation, and sustainability goals. These structural differences are essential for rethinking future policies that support both competitiveness and the long-term sustainability of the cattle sector in the EU without creating additional competitiveness risk for EU Member States. Ethics approval Not applicable. Data and model availability statement Data on the live bovine animal population and live bovine animal export are available on the Eurostat website: Database Eurostat. The calculations are based on datasets reported directly to Eurostat by each country: EU trade since 1988 by HS2-4–6 and CN8. EU policies present datasets organised according to specific EU policies. The calculations used to conduct the study are available from the corresponding author upon reasonable request. The analyses and models were implemented using IBM SPSS Statistics. Declaration of generative AI and AI-assisted technologies in the writing process During the preparation of this work, the authors used ChatGPT (OpenAI) to assist with language editing and improving the clarity of the manuscript. After using this tool. The authors reviewed and edited the content as needed and take full responsibility for the published article's content. Financial support statement None. Ethical statement This study is based exclusively on secondary data analysis and does not involve any direct experimentation with humans or animals. Therefore, ethical approval was not required CRediT authorship contribution statement Evija Skujina: Writing – original draft, Methodology, Investigation, Formal analysis, Data curation. Irina Pilvere: Writing – review & editing, Methodology, Conceptualization. Declaration of competing interest The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.