Unveiling Determinants of Successful Dairy Farm Performance from Dairy Exporting EU Countries

Abstract

The dairy sector is the second largest agricultural sector in the EU and Lithuania. It faces economic challenges (price volatility, farm consolidation and downsizing, etc.), but its importance outstrips other agricultural sectors (combining agro-systems and providing valuable food products for people). The aim of the study is to identify the vulnerabilities of dairy farms and to consider how to improve their performance after analysis of dairy farms in dairy exporting EU countries. As the problem of the study is complex, a set of indicators was analysed, including farm size, milk yield per cow, number of cows per annual work unit (AWU), milk production per fodder area, feed autonomy, milk price, total operating costs, depreciation, farm net value added per AWU, milk production per capita, and GDP at current prices per capita. The analysis was carried out using data from the Farm Accountancy Data Network (FADN) for 2017–2019. As Lithuania’s dairy sector is export-oriented, EU countries with more than 100% milk self–sufficiency (18 countries in total) were chosen for the comparison. The multi-criteria decision-making methods (MCDM) were used for the study. The multi-criteria evaluation revealed that countries that are leaders in dairy exports obtained the best values of dairy farm performance. These countries (Denmark, Netherlands, Ireland, and Belgium) have the highest farm size, production scale, productivity, and income indicators. While Slovenian, Latvian, and Lithuanian dairy farms performed poorly in terms of productivity and economic indicators, these countries achieve competitiveness in the dairy sector through lower milk prices, higher utilisation of own resources, and higher levels of public support.

1. Introduction

The dairy sector holds an important role in the economies of many European Union (EU) countries. The EU is the second-largest milk producer in the world (after India) and the largest exporter of dairy products [1]. Dairy production is the second-largest agricultural sector in the EU, accounting for 12% of all agricultural output [2]. The EU’s Common Agricultural Policy (CAP), oriented towards the market (abolition of quotas, trade liberalisation), facilitated the expansion of milk production and foreign trade in dairy products.

In 2023, compared to 2014 which was the last year of milk quota system, EU milk production increased by 5.6% [3], and the quantity of exported dairy products increased by 10.5% [4]. The EU dairy sector became an equal player in the global dairy market, competing with the dairy production from the United States of America (USA) where dairy farms are very large and intensive, as well as with the dairy production from low-cost milk production countries such as New Zealand [5]. This drives EU dairy farms and the dairy industry to increase their competitiveness and production efficiency. From 2014 to 2020, milk yield per cow in the EU increased by 14%, and the average number of cows per farm rose from 35 to 43 [6], indicating that milk production is concentrating in fewer but larger farms. Despite the growth in farm sizes, the economic situation of dairy farms remains challenging. Dairy farms face volatility in milk prices and input costs [7,8], issues with generational succession [9], all of which pose additional challenges to the economic viability of dairy farms. It is noteworthy that dairy farms receive insufficient market returns, with almost 40% of all dairy farm income consisting of income support payments [10].

The growing global population, urbanisation, and increasing average income per capita create favourable conditions for the growth in demand for dairy products [1]. However, dairy farms face sustainability challenges. On one hand, livestock farming, including dairy farming, is considered a crucial part of the terrestrial ecosystem, helping to maintain biodiversity and natural grasslands, and plays a vital role in the food chain. Dairy cows can process fibrous materials such as grass and by-products of the food industry, converting them into milk and meat, and provide people with essential valuable nutrients. Milk and dairy products offer energy and a significant amount of protein and micronutrients, including calcium, magnesium, selenium, riboflavin, and vitamins B5 and B12. They are the fifth largest source of energy and the third largest source of protein and fat in the human diet [11]. Additionally, dairy farming helps combat poverty by providing income to many farmers. According to FAO [5], a quarter of all farms worldwide have dairy animals.

On the other hand, dairy farms are identified as major emitters of greenhouse gases, contributing to climate change. It is estimated that globally, dairy cattle emit about 4.4% of anthropogenic greenhouse gases [12]. Furthermore, intensive farming leads to environmental degradation and biodiversity loss [11]. Consumer concerns about the environmental impact of dairy production and demands for higher animal welfare standards are driving investments in production technologies and changes in production methods, which require additional costs, thereby reducing the efficiency of milk production.

Favourable natural conditions, the availability of grass fodder, a strong tradition of milk production and consumption of dairy products, and an advanced and technologically modern dairy industry make the dairy sector in Lithuania one of the largest and most important agricultural sectors. Dairy production is the second biggest agricultural sector (after cereals), accounting for about 17% (2022), and dairy cows are kept by 21.9% of all registered farms [13]. The Lithuanian dairy industry is an important exporter—exports of dairy products account for about 60% of the total income of the Lithuanian dairy industry, with dairy products exported to more than 60 countries [14]. The Lithuanian dairy sector, as in other EU countries, shows trends of intensification and concentration of production, but no expansion of milk production is recorded. Data show that the number of dairy farms is shrinking every year. On average, around 11% of farms are withdrawing from milk production each year between 2004 and 2023. The number of dairy cows fell by 54% over 20 years [14].

This study aims to assess the economic situation of dairy farms in dairy exporting EU countries, to identify the weaknesses of dairy farms, and to consider the opportunities for improving their performance. The study uses three MCDM methods, namely, Simple Additive Weighting (SAW), Technique for Order of Preference by Similarity to an Ideal Solution (TOPSIS) and Evaluation based on Distance from Average Solution (EDAS) to assess the impact of various factors on the performance of dairy farms. These methods allow for evaluation of the impact of various factors on the performance of dairy farms. The novelty of the study lies not only in supplementing knowledge about the economic situation of dairy farms in EU countries, but also in evaluating dairy farms in the context of sustainability and dairy exporting EU countries (the study examines those EU countries whose dairy sectors are more export oriented, with milk production exceeding domestic needs). The research provides deeper insights by identifying which factors interact and complement each other, enabling the achievement of the best results. Therefore, the results of this study could be useful in designing policy support measures to increase the competitiveness of dairy farms and the national dairy sector.

The article is organised as follows: Section 2 provides a literature review. Section 3 describes the methods, starting with a description of the three MCDM methods (SAW, TOPSIS, and EDAS) used, followed by a presentation of the indicators used in the research to assess the situation of dairy farms. The results are discussed in Section 4. Finally, the last section draws conclusions on our empirical results.

2. Literature Review

Climate change and biodiversity loss are forcing EU farms to adapt dairy production to rising environmental standards, while maintaining the competitive position of the dairy sector in the face of growing global demand for dairy products. The development of dairy production in the context of sustainability was one of the most important research topics in the last decade. A scientometric analysis using VOSviewer (version 1.6.20) software and analysing scientific publications (scientific articles and reviews) extracted from the Web of Science (Core Collection) platform over the period from 2004 to 2024 was carried out, yielding 470 publications selected on the basis of the following key words: “milk farm*” OR “dairy farm*” OR “milk production” AND “economic*” OR “sustainab*” shows that research on dairy farming is widely interpreted and covers a wide range of aspects (environmental, ecological, veterinary medicine, etc.). The analysis identified four clusters (each with a different colour) that are prevalent in economic research on dairy farms (Figure 1). The cluster in red highlights issues related to research on dairy farming as an economic microsystem—economic evaluation of the farm, evaluation of policies, and evaluation of innovation. The green cluster includes a large group of studies related to environmental impact assessment, reflecting research on the environmental dimension of sustainability (greenhouse gas (GHG), N, P emissions, manure management, life cycle assessment) and the relevance of these studies. The third cluster (shown in blue) includes studies related to dairy herd management (feeding, treatment, lactation, and reproduction management). The fourth cluster, which is the smallest, shown in yellow, covers animal welfare research, bringing together research on changing consumption patterns, the dairy industry, and the impact on consumer health. These issues are part of the research questions on the social dimension of sustainability, which are often linked to public concerns about the impact of agriculture on animal and human welfare, food quality, and consumer attitudes [15].

As, this study assesses the economic situation of EU dairy farms in the context of the concept of sustainability, we mainly focus on economic sustainability in this literature review. Dairy farms are profit-oriented businesses, and the economic sustainability of farms is probably the most important issue for most farmers [16]. For farms or enterprises to be able to conduct and expand their agricultural activities, income must exceed costs, at least in the long term, leading to the survival of production [17]. Neglecting economic sustainability may lead to resistance or reduce the interest of farms to take the necessary actions to improve environmental performance and social conditions.

There is no single definition of economic sustainability in the scientific literature. From a theoretical point of view, economic sustainability could be viewed in two ways: in the first case, it focuses on the sustainable use of natural resources within a defined economic framework, i.e., economic sustainability is achieved when economic activity is not at the expense of natural resources; in the second case, it is achieved when there is economic growth, meaning that in order to be economically sustainable, there has to be a return on the capital invested in the company [18]. From these concepts comes the concept of economic sustainability, which is generally understood as an economic activity that helps to sustain long-term economic growth while at the same time not harming the environment and taking into account the interests of the community, i.e., striking a balance between economic growth and the needs of the environment and community.

Different methods and indicators are used to measure economic sustainability, which are selected according to the purpose of the sustainability measurement (research, monitoring, certification, policy making, farm advisory services, farm self-assessment, consumer information, and regional planning), the level of the measurement (sectoral level, regional level, farm level, product level, supply chain level, and benchmark level), the geographical coverage, the scope of the business activity, and the level of the dimension of sustainability [19].

Profitability is one of the key factors used to assess economic sustainability [20,21,22,23]. Profitability is calculated by comparing revenues and costs by the difference between revenues and costs, or by the ratio between profits and sales revenue/equity/costs. From a data availability point of view, financial profit is the simplest (easiest) to calculate for farms, as it can be extracted from accounting records [24]. However, extending the financial profit indicator from an economic point of view to include the opportunity cost of the farm’s own resources (labour, land, and capital) allows for a broader view of the farm’s economic profitability and a longer perspective [25]. In this case, indicator of farm viability is obtained, which is widely used to assess both sustainability and the economic situation of farms [23,26,27,28,29,30,31].

Competitiveness is another important factor in the economic sustainability of farms [21]. The cost of production is one of the most important indicators for assessing the competitiveness of a farm; in the long run, only farms that manage to cover all economic costs (including opportunity costs) are viable [21]. As dairy farms are mostly price takers [7], cost analysis and cost management are one of the key factors in maintaining the competitiveness of a farm or sector. There is no single definition of competitiveness in economic theory [32], so researchers use a variety of indicators to assess competitiveness: Parzonko and Bórawski in 2020 [33] used the family labour remuneration as an indicator in their research, and Irz and Jansik in 2015 [34] assessed the competitiveness of dairy farms using partial productiveness and total factor productivity measures. The following sub-topics were analysed in the sustainability assessment methodologies to assess competitiveness: value added, innovation, investment, market orientation, local economy, quality, and farm diversification. It should be noted that indicators for competitiveness sub-indicators are often qualitative, and therefore often appear as qualitative case studies [35,36,37].

The concept of sustainability Implies the desire to produce more output with fewer resources, so productivity and efficiency indicators are an important part of measuring economic sustainability [38,39,40]. Productivity indicators measure the ability of factors of production to produce output and are expressed as the ratio between output and the factor of production (land, labour, capital, and livestock) [32]. Efficiency indicators show how efficiently a farm uses technology and measure the output produced from a given set of inputs.

An analysis of the scientific literature showed that researchers often focus on a narrower field of research compared to sustainability, choosing to analyse one or more economic aspects of dairy farming. Kołoszycz in 2020 [27], Spicka et al. in 2019 [25], and Wilczyński and Kołoszycz in 2021 [31] conducted on-farm economic viability studies in order to determine the viability of dairy farms in different countries, taking into account the size of the farm. The results reveal that dairy farm viability results are highly sensitive to the used estimates of unpaid farm wages and cost of equity [41], while there were no common definitions for these estimates in the economic literature [25].

Parzonko and Bórawski in 2020 [33], Stoychev and Ivanov in 2022 [42], Viira et al. in 2014 [43], and Ziętara and Mirkowska in 2023 [44] examined the competitiveness level of dairy farms and the dairy sector, finding that farm size is one of the main factors in determining the competitiveness of dairy farms and has a significant impact on economic efficiency, including labour productivity. A study on dairy farm income was conducted by Średzińska et al. in 2021 [45], assessing the profitability of EU dairy farms using an aggregated indicator. Syrůček, Bartoň and Burdych in 2022 [46] analysed the break-even points of dairy farms in EU countries to determine the minimum milk yield and milk price required to achieve a certain level of profitability.

3. Materials and Methods

In order to achieve the objective of the study—to assess the economic situation of EU dairy farms in the context of EU dairy exporting countries, to identify the weaknesses of dairy farms, and to consider the potential for improvement of their performance—an analysis of dairy farms was carried out using three MCDM methods: SAW, TOPSIS, and EDAS. As the Lithuanian dairy sector is export oriented, the objects of the research—countries being compared—were selected from the EU countries where milk supply exceeds the country’s domestic needs by more than 100% (a total of 18 countries). The data from the Farm Accountancy Data Network (FADN) for the years 2017–2019 were used for the analysis. The period from 2017 to 2019 was chosen for analysis because of the relatively stable period of activity between two crisis periods (after the price collapse starting in 2014 and the abolition of EU quotas in 2015 to the pandemic of the Covid-19 in 2020, and the war in Ukraine caused by Russia in 2022). Typically, in studies that do not seek to analyse the rate of variability, a 3-year period is sufficient to remove the random effect of variability due to natural or other events [47].

MCDM methods are useful when dealing with multidimensional phenomena. In order to achieve the accuracy of the results, three MCDM methods—SAW, TOPSIS, and EDAS—were applied. A search of the Web of Science (WoS) Core Collection at https://www.webofscience.com/ (accessed 18 April 2024), with the following query string processed in the search engine: TOPIC (search title, abstract, author keywords, plus keywords) = “agriculture” AND “TOPSIS” OR “SAW” OR “EDAS”) revealed that MCDM methods are most commonly found in environmental studies in agricultural research [48,49,50,51], sustainability assessments [52,53,54,55], and evaluations of competitiveness, efficiency, or public economic support measures [56,57,58,59,60,61], in terms of assessing the impact of the factors analysed on farm performance or for comparative analysis between countries (Figure 2).

These methods are based on different aggregation principles (utility functions and reference points), which may lead to different results. The SAW method incorporates indicator values and their weights into a single indicator, the so-called method criterion [62]. The TOPSIS method searches for the ideal solution that has the smallest distance from the best values and the largest distance from the worst values of the indicators [62]. The EDAS method solution is based on a compatible distance from the average value of the alternative solution as a reference point [63]. The choice of the three methods allows the disadvantages of one method to be compensated by the advantages of the other methods.

The SAW method is an MCDM method used to calculate the composite scores of the alternatives and is based on the linear utility function for multi-criteria evaluation [62]. The criterion of the method $S_j$ combines values of the indicators and their weights into one index. To obtain the $S_j$ values, firstly, the values of the selected indicators were normalised. Maximizing the indicator, values were normalised as follows:

$${\overline{r}}_{ij}=\frac{r_{ij}}{\underset{j}{\mathit{max}}{r}_{ij}},$$

(1)

Minimizing the indicator, the linear normalisation ratios are computed as follows:

$${\overline{r}}_{ij}=\frac{\underset{j}{\mathit{min}}{r}_{ij}}{r_{ij}},$$

(2)

where $r_{ij}$ is the value of the i-th indicator for the j-th alternative (in this case, countries), $\underset{j}{\mathit{max}}{r}_{ij}$—the highest value of i-th indicators for the j-th alternative, $\underset{j}{\mathit{min}}{r}_{ij}$—is the lowest value among all alternative i-th indicators for the j-th alternative.

The sum of the normalised values ${\tilde{r}}_{ij}$ of all indicators $S_j$ for each j-th alternative is calculated using formula:

$$S_j=\sum _{i=1}^m{\omega }_i{\tilde{r}}_{ij},$$

(3)

where ${\omega }_i$ is the weight of the i-th indicator, ${\tilde{r}}_{ij}$ is the normalised value of the i-th indicator for the j-th alternative. The values of the $S_j$ criterion take between 0 and 1. The highest value of the $S_j$ criterion means the best performance.

The basic principle of the TOPSIS method is to find the best available alternative (in our case, countries) according to ideal and anti-ideal solutions: approaching the ideal solution with the shortest distance from the best values of indicators and with the longest distance from the worst values of indicators [62]. The TOPSIS method normalizes the initial data and relies on the distance between the best and the worst points. The vector normalisation is performed as follows:

$${\tilde{r}}_{ij}=\frac{r_{ij}}{\sqrt{{\sum }_{j=1}^n{r_{ij}}^2}},(i=1, \dots ,m; j=1, \dots ,n)$$

(4)

where ${\tilde{r}}_{ij}$ is the normalised value of the j-th alternative according to the i-th indicator.

Then ideal V* and anti-ideal TOPSIS solutions $V^{-}$ are identified:

$$V^{\mathrm{*}}=\left\{V_1^{\mathrm{*}},V_2^{\mathrm{*}},\dots ,V_m^{\mathrm{*}}\right\}=\left\{\left(\underset{j}{\max }\frac{{\omega }_i{\tilde{r}}_{ij}}{i}\in I_1\right),\left(\underset{j}{\min }\frac{{\omega }_i{\tilde{r}}_{ij}}{i}\in I_2\right)\right\},$$

(5)

$$V^{-}=\left\{V_1^{-},V_2^{-},\dots ,V_m^{-}\right\}=\left\{\left(\underset{j}{\min }\frac{{\omega }_i{\tilde{r}}_{ij}}{i}\in I_1\right),\left(\underset{j}{\max }\frac{{\omega }_i{\tilde{r}}_{ij}}{i}\in I_2\right)\right\},$$

(6)

where I₁ is the set of indexes for maximised indicators, I₂ is the set of indexes for minimised indicators, and ${\omega }_i$ is the weight of the i-th indicator.

The total distance $D_j^{\mathrm{*}}$ between each option and an ideal solution V*, and the total distance $D_j^{-}$ between each option and an anti-ideal solution $V^{-}$ are calculated as follows:

$$D_j^{\mathrm{*}}=\sqrt{\sum _{i=1}^m{({\omega }_i{\tilde{r}}_{ij}-V_i^{\mathrm{*}})}^2,}$$

(7)

$$D_j^{-}=\sqrt{\sum _{i=1}^m{({\omega }_i{\tilde{r}}_{ij}-V_i^{-})}^2,}$$

(8)

The best alternative according to the TOPSIS method corresponds to the highest value of the main criterion $C_j^{\mathrm{*}}$, which is calculated using formula:

$$C_j^{\mathrm{*}}=\frac{D_j^{-}}{D_j^{\mathrm{*}}+D_j^{-}}, (j=1,...,n) ({0\le C}_j^{\mathrm{*}}\le 1).$$

(9)

The EDAS method was proposed by Ghorabaee in 2015 [63,64]. The aim of EDAS is to identify the best alternative using a normalisation technique based on the average decision. Compared to TOPSIS, EDAS lets us evaluate more realistic solutions, because in real life, achieving the minimum distance to the ideal solution or the maximum distance to the anti-ideal solution does not guarantee the best solution. EDAS utilizes two measures called PDA (positive distance from the average value) and NDA (negative distance from the average value) to determine the score of each alternative and their relative ranking order. The highest PDA and the lowest NDA values describe the best alternative.

For each indicator the average solution is calculated:

$$AV={\left[{AV}_j\right]}_{1\times m},$$

(10)

where each element of AV is computed as follows:

$${AV}_j=\frac{{\sum }_{i=1}^n{r}_{ij}}{n},$$

(11)

$PDA$ and $NDA$ matrices are constructed:

$$PDA={\left[{PDA}_{ij}\right]}_{n\times m},$$

(12)

$$NDA={\left[{NDA}_{ij}\right]}_{n\times m}.$$

(13)

The elements of the PDA and NDA matrices are calculated with respect to the criterion $j$. If the j-th criterion must be maximised, then elements of the matrices ${PDA}_{ij}$ and ${NDA}_{ij}$ are calculated as follows:

$${PDA}_{ij}=\frac{\mathrm{m}\mathrm{a}\mathrm{x}(0,\left(r_{ij}-{AV}_j\right))}{{AV}_j},$$

(14)

$${NDA}_{ij}=\frac{\mathrm{m}\mathrm{a}\mathrm{x}(0,\left({AV}_j-r_{ij}\right))}{{AV}_j}.$$

(15)

If the j-th criterion must be minimised, then elements of the matrices ${PDA}_{ij}$ and ${NDA}_{ij}$ are calculated according to the formulas below:

$${PDA}_{ij}=\frac{\mathrm{m}\mathrm{a}\mathrm{x}(0,\left({AV}_j-r_{ij}\right))}{{AV}_j},$$

(16)

$${NDA}_{ij}=\frac{\mathrm{m}\mathrm{a}\mathrm{x}(0,\left(r_{ij}-{AV}_j\right))}{{AV}_j}.$$

(17)

Thus, ${PDA}_{ij}$ and ${NDA}_{ij}$ represent positive and negative distances for alternative $i$ with regards to criterion $j$.

Then, positive and negative distances of each alternative are aggregated by applying weighted summation. Weighted sums ${SP}_i$ and ${NP}_i$ are calculated as follows:

$${SP}_i=\sum _{j=1}^m{w}_j{PDA}_{ij},$$

(18)

$${SN}_i=\sum _{j=1}^m{w}_j{NDA}_{ij},$$

(19)

where $w_j$ is the weight of the criterion $j$.

Aggregated indicators ${SP}_i$ and ${NP}_i$ are then normalised:

$${NSP}_i=\frac{{SP}_i}{{max}_i\left({SP}_i\right)},$$

(20)

$${NSN}_i=1-\frac{{SN}_i}{{max}_i\left({SN}_i\right)}.$$

(21)

The composite score is calculated for each alternative as an average of two normalised aggregates:

$${AS}_i=\frac{1}{2}\left({NSP}_i+{NSN}_i\right),$$

(22)

where $0\le {AS}_i\le 1$. Alternatives are ranked in descending order according to the composite score.

Based on the literature review and taking into account the complexity of the research problem, a set of indicators (

Table 1. Selected dairy farming indicators for the multicriteria evaluation.

Indicator	Description	Source of Literature
Productivity indicators:
Farm size	Average number of dairy cows (livestock unit (LU)) per holding. Allows to estimate herd size and production scale.	Gonzalez-Mejia et al., 2018 [65]; Kołoszycz and Świtłyk, 2019 [66]; Syrůček et al., 2019 [67]; Wilczyński and Kołoszycz, 2021 [31]
Milk yield per cow	Average milk production per cow per year (kg). Measure of production efficiency.	Syrůček et al., 2019 [67]; Wilczyński and Kołoszycz, 2021 [31]
Number of cows per AWU	Number of dairy cows per AWU 1 Indirect measure of technology. Useful for comparing farm productivity, and for socio-economic characterisation.	Gonzalez-Mejia et al., 2018 [65]; Van Der Meulen et al., 2014 [15]
Milk production per fodder area	Milk volume produced per ha fodder area (milk in tons/ ha fodder area). Measure of land use intensity for dairy cows. Useful for characterising farms and comparing management practices.	Cele et al., 2022 [68]; Kołoszycz and Świtłyk, 2019 [66]; Spicka et al., 2019 [25]
Feed autonomy	Share of own-produced feed in total feed used in farm (%). Measure of self-sufficiency in feed.	Entrena-Barbero et al., 2024 [69]
	Economic indicators:
Milk price	Annual milk price (Eur/t).	Entrena-Barbero et al., 2024 [69]; Galnaitytė et al., 2017 [56]
Total operating costs	Total operating costs include specific costs for milk production (purchased concentrates, purchased coarse fodder, farm use of non-fodder crops, specific forage costs, milk herd renewal costs, and the milk levy and other specific livestock costs (veterinary, etc.))+ non-specific costs: upkeep of machinery and buildings, power (fuel and electricity), contract work, taxes and other dues, taxes on land and buildings, insurance for farm buildings, and other direct costs (Eur/t) [70].	Galnaitytė et al., 2017 [56]; Hanrahan et al., 2018 [71]
Depreciation	Annual depreciation of buildings, equipment (Eur/t milk).	Key et al., 2018 [72]
Farm net value added/AWU	Farm net value added (FNVA) = total output − intermediate consumptions + balance subsidies and taxes − depreciation divided by AWU (FNVA/AWU, Eur).	Madau et al., 2017 [73]
	External economic environment indicators:
Milk production per capita	Indicates the level of self-sufficiency in milk and the export orientation of the country’s dairy sector, calculated as the amount of milk produced in the country divided by the country’s population (due to the different methodologies for calculating the self-sufficiency indicator and the varying values of the indicator reported in different sources, we did not use the self-sufficiency indicator in the study to assess the sector’s dependence on exports) (kg milk per capita). In EU countries, the average consumption of milk and dairy products per capita was 251 kg per year [74], suggesting that in countries where per capita production is well above this level of consumption, the self-sufficiency in milk and dairy products is high.	Stoychev and Ivanov, 2022 [42]
Gross domestic product (GDP) at market prices	GDP at current prices, Eur per capita.	Bórawski et al., 2020 [75]; Nowak and Kaminska, 2016 [58]

¹ Annual work unit (AWU) is the full-time equivalent employment, i.e., the total hours worked divided by the average annual hours worked in full-time jobs in the country.

) was selected for the quantitative assessment, reflecting the main objectives of EU agriculture: strengthening sustainability, increasing the extensiveness of production, and moving towards a circular economy: productivity indicators—farm size, milk yield per cow, number of cows per AWU, milk production per fodder area, and feed autonomy; economic indicators—milk price, total operating costs, depreciation, and farm net value added per AWU; and indicators of economic environment—milk production per capita and GDP at current prices.

The indicators chosen to reflect the main characteristics of dairy farms and allow the measurement of the performance of dairy farms in the context of economic results, the environment, and animal welfare. Farm size is considered to determine the technical efficiency and profitability of a farm [76], and large farms are able to achieve economies of scale, which leads to a profitable farm operation [77]. Milk yield per dairy cow is a key factor in achieving production scale [78,79], higher yield generally means less inputs per production unit. The relative number of workers on a farm indicates the level of farm mechanisation and allows for the examination of the potential replacement of labour with technology [80].

Milk production per forage area and feed autonomy allows for an assessment of the environmental sustainability of the farm. On one hand, the forage area indicates the farm’s ability to provide its own feed (self-sufficiency); on the other hand, it is an indicator of the environmental dimension of sustainability. A decreasing forage area per cow indicates a trend towards production intensification, which suggests greater exploitation of natural resources [81]. The area per cow also identifies compliance with animal welfare requirements on farms (whether the animals have the opportunity to go outside or graze). Milk production per forage area also indicates the productivity of the use of the area [79]. A higher share of feed produced on farms compared to the share of feed brought in means more self-sufficiency for farmers. This indicator can also be considered as an indicator of circularity, as it reflects one of the principles of circularity: minimise external resource input [69].

One possible way to reduce environmental impacts is to move towards more extensive farming, so the indicators assessing environmental aspects and production extensiveness are the same in our study. The researchers use different indicators to assess the extensiveness of milk production: Latruffe et al. [82], used three indicators: livestock density (stocking rate), calculated as the number of LSU per hectare of forage area, the ratio of fodder area to utilised agricultural area (UAA), and the share of the rented area to UAA. Kellemann and Salhofer in 2014 [83] used two indicators: livestock density and milk yield per cow per year.

The economic indicators chosen for the study relate to milk price and milk production costs, which are key indicators of a sustainable dairy farm as well as indicators for assessing the overall economic competitiveness of farms in the dairy market at local and international levels [84]. The value of the capital invested in the farm was measured in terms of the annual depreciation rate to determine the average annual utilisation of the capital stock [73].

The other indicator used in the analysis, farm net value added (FNVA), is one of the indicators of profitability and represents the remuneration of the fixed factors of production (labour, land, and capital), regardless of whether they are external or on-farm factors [70]. This indicator makes it possible to compare farms irrespective of the dependence of the factors of production they use (whether they are family farms or external), which is relevant when farms are not homogeneous in terms of the dependence of the factors of production used.

The assessment criteria include two indicators of the country’s economic environment, milk production per capita and GDP per capita, which reflect the country’s availability of milk and dairy products and the dependence of the dairy sector on exports and the potential for domestic consumption.

The weights of the selected indicators were weighted on the basis of the experts’ assessments and are presented in

Table 2. Weights of indicators and directions for optimisation.

	Farm Size (Dairy Cows/Farm)	Cows/AWU	Milk Yield per Cow (kg/cow)	Milk/Fodder Area (t/ha)	Feed Autonomy (Share of Own Produced Feed %)	Milk Price (Eur/t)	Total Operating Costs (Eur/t)	Depreciation (Eur/t)	FNVA/AWU (Eur/AWU)	Milk Production per Capita (kg/capita)	GDP at Market Prices (Eur/capita)
Weight	0.1	0.1	0.1	0.05	0.05	0.15	0.15	0.05	0.15	0.05	0.05
Direction	max	max	max	max	max	max	min	min	max	max	max

. For the purpose of the study, 10 experts were interviewed and asked to assess the weights of the selected indicators. The experts were selected from different fields of activity (six from the field of science and policy making and four from the field of dairy farm practice).

4. Results and Discussion

Since the EU abolished milk quotas in April 2015, dairy farmers are free to decide how to increase milk production. Market liberalisation had different effects on milk production in the analysed EU countries: Ireland, Luxembourg, Czech Republic, Belgium, and Poland expanding their milk production by 23–50%, while in other countries, milk production even decreased (with the largest decreases in Finland, France, Lithuania, Sweden, and Slovakia) [85]. In Lithuania, milk production in 2023 was lower by 6% compared to 2014 (Figure 3).

Despite the different physical characteristics of dairy farms in the selected EU countries, an analysis of data from the Dairy Farm Structure Survey 2010–2020 (Eurostat survey, the most recent relevant survey) [86] shows the same trends in all the countries analysed (Figure 4): the number of dairy farms is declining, while the size of the herd (number of cows per farm) is increasing. In the EU countries analysed, the number of farms with cows in 2020 decreased by 51.0% compared to 2010, with the largest decreases in the Baltic countries: Lithuania by 65.9%, Estonia by 69.3% and Latvia by 60%. The smallest decreases were observed in Ireland and the Netherlands (decreases of 5.2% and 20.6%, respectively). The average number of cows per farm in the EU countries analysed almost doubled from 19.4 cows in 2010 to 38.7 cows in 2020. The highest increases were observed in the countries with the smallest farm size in 2010: In Estonia, the average farm size increased by 2.9 times, while in Poland, Latvia, Lithuania, and Hungary it increased by 2.2–2.0 times. This means that these countries had the greatest potential for farm expansion due to their evolving farm structures. The evolution of cow numbers over the period analysed varied between the countries concerned: in several EU countries, cow numbers increased (in Ireland (+46.4%), Luxembourg (+21.2%), the Netherlands (+7.7%), Belgium (+3.3%), and Austria (+1.3%)), while in others, cow numbers decreased (with the largest decreases in Lithuania (−31%), Slovakia (−22.2%), and Latvia (−16.3%). In the EU countries considered, the number of cows in 2020 will decrease by 2.6% compared to 2010.

This study focuses on specialised dairy farms that derive 2/3 of their total standard production from dairy farming activities (FADN code 45). The survey covered farms that were included in the FADN survey as a representative sample according to FADN methodology in each analysed country out of a total of 11,020 farms [6].

The physical and productivity characteristics of specialised dairy farms in the EU countries examined vary considerably (

Table 3. Selected specialised dairy farms physical and productivity indicators in 2017–2019.

Countries 2	Sample Farms	Farm Size (LU) 1	Forage Area (ha)	Milk Production (t)	Milk Yield per Cow (kg/cow) 1	Livestock Density (LU/ha)	Milk per ha Fodder Area (t/ha) 1	Feed Autonomy (%) 1	Cows/AWU 1	AWU	Share of Family Work Unit (%)
Ireland	317	80	65	467	5865	1.23	7.2	31%	47.3	1.7	84%
Denmark	407	188	126	1854	9841	1.49	14.3	33%	54.0	3.5	35%
Netherlands	367	102	57	901	8815	1.80	15.8	22%	55.1	1.9	88%
Luxembourg	248	77	88	601	7809	0.87	6.8	27%	42.6	1.8	87%
Estonia	120	114	199	1085	9554	0.57	5.3	38%	18.9	6.0	15%
Lithuania	324	13	25	76	5980	0.51	3.0	55%	7.2	1.8	80%
Latvia	341	20	43	134	6786	0.46	3.1	44%	9.1	2.2	60%
Austria	660	22	25	156	7228	0.86	6.2	31%	12.8	1.7	99%
Finland	239	41	54	381	9233	0.77	7.1	29%	19.5	2.1	81%
Germany	2887	73	62	590	8047	1.18	9.5	28%	31.9	2.3	64%
France	1350	64	80	456	7086	0.81	5.7	30%	32.1	2.0	86%
Belgium	305	75	54	593	7875	1.40	11.0	24%	38.6	2.0	96%
Poland	2562	19	15	117	6169	1.27	7.8	47%	10.2	1.9	94%
Slovenia	138	19	17	110	5709	1.15	6.6	47%	11.2	1.7	100%
Czech Republic	220	178	305	1471	8259	0.58	4.8	56%	9.9	17.9	7%
Sweden	364	88	128	805	9103	0.69	6.3	26%	32.1	2.8	63%
Hungary	108	56	44	448	7955	1.28	11.6	38%	13.3	4.2	22%
Slovakia	63	232	670	1701	7326	0.35	2.5	69%	8.4	27.6	2%
Mean	x	81	114	664	7702	0.96	7.5	38%	25	4.7	65%
Minimum	x	13	15	76	5709	0.35	2.5	22%	7.25	1.7	2%
Maximum	x	232	670	1854	9841	1.80	15.8	69%	55.1	27.6	100%
Std. deviation	x	61	152	533	1255	0.39	3.6	13%	16.2	6.7	33%
Median	x	74	60	529	7842	0.87	6.7	32%	19.2	2.1	80%
Coefficient of variation	x	75%	133%	80%	16%	41%	48%	33%	64%	141%	50%

¹ indicators used in the MCDM assessment; ² countries ranked according to the indicator—milk production per capita.

). The most variable indicators are farm size (number of dairy cows on the farm, forage area, and milk production on the farm) and number of dairy cows per AWU. Slovakia, Czech Republic, Denmark, Estonia, and the Netherlands have the largest dairy farms in terms of number of cows, while the smallest farms are in Lithuania (13 cows), Poland (19 cows), Slovenia (19 cows), and Latvia (20 cows). The average specialised dairy farm size in the EU countries considered is 81 cows, with a coefficient of variation of 75%; the average of forage area is 86 ha with a coefficient of variation of 133%; and the average of milk production on the farm is 664 t with a coefficient of variation of 80%.

The countries with the smallest specialised dairy farms (Lithuania, Latvia, and Slovenia) also have the lowest productivity indicators (milk yield per cow, milk per hectare of fodder area, and cow number per AWU), but they have a high level of self-sufficiency in feed. This indicates that extensive milk production prevails in these countries and that farms rely mainly on their own resources.

The best productivity rates are found in countries with a predominance of large farms. The Netherlands, Denmark, Ireland, and Belgium have the most intensive specialised dairy farms, with the highest milk production per ha of fodder area and the highest cow number per AWU. However, these countries face difficulties in implementing environmental requirements (in particular N and P limits per ha) [87]. Slovakia stands out in this group of countries, where farms are characterised by large herds, and extensive milk production dominates, with the lowest livestock densities and the highest level of self-sufficiency in animal feed.

Denmark, Estonia, Sweden, and Finland have the highest cow productivity rates. In these northern EU countries, the natural conditions lead to a higher investment requirement per cow for the construction and equipping of barns, and therefore, in order to make the farm profitable, efforts are made to increase the milk yield per cow [88], thus reducing the fixed costs per unit of production.

The economic performance of specialised dairy farms (

Table 4. Selected specialised dairy farms economic indicators in 2017–2019.

Countries 2	Milk Price (Eur/t) 1	Coupled Payment (Eur/t)	Total Operating Costs (Eur/t) 1	Gross Margin (Eur/t)	Depreciation Eur/t milk 1	FNVA/AWU Eur 1	Balance Subsidies and Taxes/AWU	Share of Subsidies in FNVA	Milk Production per Capita kg 1	GDP at Market Prices (Eur/capita) 1
Ireland	335	0.0	201	134	24.3	50,776	12,670	25%	1624	67,283
Denmark	390	0.0	274	116	37.4	94,312	21,153	22%	965	52,177
Netherlands	394	0.0	261	133	48.7	72,573	10,404	14%	837	44,963
Luxembourg	341	0.0	235	106	92.4	55,157	47,430	86%	673	98,880
Estonia	312	6.4	230	88	41.5	27,079	13,413	50%	608	19,627
Lithuania	288	17.2	186	119	74.6	8,067	5,359	66%	554	16,240
Latvia	292	23.7	217	99	45.1	10,622	7,192	68%	509	15,003
Austria	378	0.6	235	143	94.2	22,002	13,968	63%	428	43,443
Finland	375	83.5	357	101	91.8	26,800	38,566	144%	428	42,280
Germany	363	0.0	244	119	47.7	46,764	18,177	39%	398	40,643
France	360	5.2	245	119	64.5	35,488	18,395	52%	374	35,083
Belgium	349	0.9	205	145	41.9	51,916	12,451	24%	367	40,350
Poland	307	11.2	166	152	40.7	12,896	4,747	37%	314	12,993
Slovenia	318	11.5	234	95	98.8	9,088	6,643	73%	298	22,073
Czech Republic	334	16.9	284	67	45.1	24,980	17,786	71%	295	19,777
Sweden	376	11.4	303	85	54.0	51,599	36,012	70%	273	46,793
Hungary	307	37.4	270	74	26.5	24,199	13,809	57%	200	13,967
Slovakia	334	37.6	329	43	51.9	17,799	14,013	79%	167	16,463
Mean	342	15	249	108	56.7	35,673	17,344	58%	517	36,002
Minimum	288	0	166	43	24.3	8,067	4,747	14%	167	12,993
Maximum	394	83	357	152	98.8	94,312	47,430	144%	1624	98,880
Std. deviation	32	20	47	28	23.0	23,056	11,510	30%	338	21,705
Median	338	9	239	111	48.2	26,940	13,888	60%	413	37,717
Coefficient of variation	9%	140%	19%	26%	41%	65%	66%	51%	65%	60%

¹ indicators used in the MCDM assessment; ² countries ranked according to the indicator—milk production per capita.

) also shows that there was considerable variation between farms. In 2017–2019, the average milk price in the countries analysed was 342 Eur/t, the lowest price was recorded in Lithuania (288 Eur/t) and the highest in the Netherlands (394 Eur/t). The analysis showed that the share of exported milk was not the decisive factor in determining the milk price (Ireland, Denmark, the Netherlands, Luxembourg, Estonia, and Lithuania were the most export-dependent countries, but this group had both the lowest and the highest prices), but that the milk price was determined by the country’s economic capacity. It should be noted that only some of the countries analysed provided coupled support to farmers. The highest level of coupled support was in Finland, followed by Slovakia, Hungary, Latvia, Lithuania, and Czech Republic.

The indicator, FNVA/AWU, allows a comparison between farms regardless of the dependence of the fixed factors of production (land, labour, and capital) being either owned by the farm or external. This is relevant for this study on EU dairy farms because farms are not homogeneous in terms of labour, e.g., among the countries analysed, Slovakia, Czech Republic, Estonia, Hungary, and Denmark stand out, where the labour of family members accounts, respectively, for 2, 7, 11, 15, and 35% of AWU, while in the other countries, family work unit accounted for between 60 and 100% of AWU. In the period 2017–2019, the average value of the FNVA/AWU amounted to 35,673 Eur. The seven countries with the largest cow herds and the highest productivity rates, Denmark, the Netherlands, Sweden, Belgium, Ireland, Luxembourg, and Germany, had higher-than-average FNVA/AWU. The farms with the lowest values for the productivity indicators (Lithuania, Slovenia, Latvia, and Poland) had the lowest values for this indicator, ranging from 23% to 36% of the EU average (Figure 5).

The subsidy rate was included in the FNVA calculation. It should be noted that, on average, subsidies represented 58% of FNVA/AWU in the countries analysed. In absolute terms, Finland, Sweden, and Denmark had the highest amounts of subsidies, while in relative terms, Finland (where the amount of subsidies exceeded the FNVA), Sweden, the Czech Republic, Lithuania, and Latvia had the highest amounts. The dairy farms in the Netherlands, Denmark, Belgium, and Ireland were the least dependent on subsidies.

Operating costs ranged from 166 Eur/t (Poland) to 357 Eur/t (Finland), with an average cost of 249 Eur/t. This indicator varied between countries depending on production intensity and productivity, it had the lowest coefficient of variation of all the indicators examined. The average gross margin (gross margin with coupled payment = milk price + coupled payment - total operating costs) in the countries analysed was 108 Eur/t. It is worth noting that the gross margin did not correlate with high national productivity indicators, e.g., Poland had the highest gross margin (152 Eur/t), while in this country, both herd size, cow productivity, and labour productivity were low among the countries analysed. It can be concluded that a large proportion of small extensive farms rely mainly on on-farm feed production and have relatively low outside costs, which allows them to achieve relatively high gross margins (Figure 6).

The amount of depreciation, which reflects the use of capital on the farm and the level of mechanisation, varied from 24 Eur/t to 99 Eur/t (142 Eur/cow to 848 Eur/cow, respectively). The need for capital was determined by the natural conditions (northern countries, mountainous areas), which determined the construction of the farm and the equipment, and by the technology used on the farm. Given that dairy farms in the EU countries were located in different natural conditions, the coefficient of variation of the annual depreciation amount per dairy cow was 41%. On small farms and farms with lower productivity, the depreciation amount per cow was relatively higher, as the farm did not achieve economies of scale. In Lithuania, the amount of depreciation per tonne of milk was 31.5% above the EU average.

For the multi-criteria evaluation, 11 indicators were selected to assess the main productivity and economic indicators of dairy farms in the context of the sustainability paradigm and the country’s economic situation.

Table 5. Ranking results of the countries according to selected indicators.

Country	Rank 1	TOPSIS		EDAS		SAW
		Criterion Estimate	Rank	Criterion Estimate	Rank	Criterion Estimate	Rank
Denmark	1	0.7758	1	0.9713	1	0.8268	1
Netherlands	2	0.6254	2	0.7942	2	0.7343	2
Ireland	3	0.5372	3	0.6897	3	0.6918	3
Luxemburg	4	0.5065	4	0.5600	5	0.6392	4
Belgium	5	0.4683	5	0.5922	4	0.6330	5
Germany	6	0.4146	7	0.5380	6	0.5945	6
Sweden	7	0.4291	6	0.4974	7	0.5900	7
Estonia	8–9	0.3375	10	0.4529	8	0.5529	8
France	8–9	0.3373	11	0.4296	9	0.5412	9
Czech Republic	10	0.3620	9	0.3919	10	0.5359	10
Slovakia	11	0.3838	8	0.3139	11	0.5161	11
Hungary	12	0.2503	12	0.2893	12	0.4999	12
Finland	13	0.2316	14	0.2129	13	0.4937	13
Poland	14	0.2399	13	0.1664	15	0.4822	14
Austria	15	0.2212	15	0.1812	14	0.4813	15
Latvia	16	0.1943	17	0.1045	16	0.4292	16
Lithuania	17	0.2076	16	0.0498	17	0.4270	17
Slovenia	18	0.1662	18	0.0430	18	0.4166	18

¹ Average ranking results of three multi-criteria methods used (TOPSIS, EDAS, and SAW).

presents the cumulative evaluation estimates of the dairy farms in the selected countries and the places determined by them. The results obtained using the MCDM suggest that the best performing dairy farms were those in Denmark, the Netherlands, and Ireland. The scale of production and the favourable relationship between output and inputs allowed the farms to perform best. It should be noted that the countries with the best scores were also the most export oriented in terms of dairy products. In these countries, it can be argued that the scale of dairy farms and the efficient operation of dairy farms were the key factors at the level of dairy exports. The relatively highest performing group of farms identified in this research corresponds to the fourth cluster acknowledged in the study accomplished by Poczta et al. in 2020 [47], defined as relatively large-scale highly intensive farms (Poczta et al. in 2020 [47] shaped five clusters of EU dairy farms which differed in the characteristics of production potential: group I—small-scale medium-extensive farms: Finland, Latvia, Lithuania, Slovenia, Poland, Austria, Romania, and Croatia, group II—small-scale extensive farms: Portugal, Spain, and Bulgaria, group III—large-scale extensive farms: Slovakia, Hungary, Estonia, and Czech Republic, group IV—relatively large-scale highly intensive farms: Sweden, Luxembourg, UK, the Netherlands, Ireland, and Denmark, and group V—medium-scale, medium-intensive farms: France, Italy, Germany, and Belgium.). The results of the study show that high productivity and intensive production have a positive impact on the performance of dairy farms, which is confirmed by Latruffe et al. [82], who in their research showed that low-intensity farming farms have better economic performance despite poorer environmental performance. The importance of productivity is demonstrated by Olagunju et al. [79], who found that productivity is the most important factor in improving and maintaining international competitiveness.

The worst results in the MCDM assessment were obtained by farms in Slovenia, Latvia, and Lithuania. These farms are characterised by extensive production and low resource utilisation. According to the Poczta et al. [47] clustering of EU dairy farms, this corresponds to cluster I: small-scale medium-extensive farms. The viability of the farms is supported by relatively high state support, and export competitiveness is ensured by low milk prices. Milk production is relatively insensitive to price changes in the short term, as farmers’ fixed costs are high compared to their variable costs [89], and the nature of dairy production (continuous milk production, raw milk cannot be stored as it is perishable) makes it necessary for farmers to continue production even in unfavourable economic conditions. Due to the fact that a relatively high proportion of total costs are fixed, farmers are generally better off even when they are losing money, as they would lose much more if they were to close down [90]. However, in the longer term, the economic performance of these farms possibly will limit their expansion, or these farms will not be attractive to inheritors and they will be forced to withdraw from dairy production, so a farm orientation towards productivity improvement would lead to better economic performance.

Farm size and labour productivity lead to better economic performance of dairy farms, which is also in line with Parzonko and Bórawski [33], who assessed the economic competitiveness of dairy farms in the six EU countries with the highest milk production. The results of this study show that the countries with the largest farms had the best farm competitiveness and the scale of production applies a considerable impact on economic efficiency, including labour productivity. This is likely to contribute to further consolidation of dairy farms and an overall reduction in the number of such farms. The Latruffe et al. [82] study found that the technology gap is the main driver of differences between higher and lower-productivity farms, although inefficiencies are higher among extensive farms.

The study on dairy farm incomes [45] showed that Ireland, Belgium, Denmark, Austria, Germany, and the Netherlands had the highest incomes among the EU countries included in our study. All these countries, except Austria, also scored highest in our study. Gołaś in 2017 [88] found that forage area, herd size, cow productivity, milk price, and energy and labour costs are the most influential factors on dairy farm profitability.

Syrůček et al. in 2022 [46] analysed the breaking points of dairy farms in EU countries in order to determine the minimum milk yield and milk price that would lead to a certain level of profitability. The best results, similar to the present study, were obtained for Irish dairy farms, while the lowest results were calculated for Lithuanian dairy farms. It was found that an increase in milk yield per cow of more than three times or a milk price level of 602.1 Eur/t would lead to zero profitability in Lithuania. It was concluded that low production volumes and low productivity indicators lead to poor performance of Lithuanian dairy farms [91].

It is important to note that the results of our study show that two parameters are very important for the performance of dairy farms: herd size and milk yield per cow. This is also supported by many studies on the technical efficiency of dairy farms [31].

Despite the low MCDM score, Lithuania, Latvia, and Poland remain exporting countries, but their competitive advantage is due to low production costs, less dependence on resource prices, and relatively big public support, but scaling up is a prerequisite for maintaining competitiveness, ensuring profitability, and maintaining sufficient resources for economic development.

The study confirmed the well-known rule that one of the main factors determining the competitiveness of dairy farms is the scale of production. Raw milk is a raw material for further processing, and with rising input prices, farmers are forced to continuously increase the scale of production, which in turn increases labour productivity. Increasing the scale of production enables technical progress, but the main constraints are land area and the exploitation of natural resources (pollution and environmental factors).

Our research reveals which farms in EU countries are maintaining their competitiveness in the light of the EU’s liberalisation of the milk market. The research shows that it is the large farms with economies of scale that form the backbone of EU dairy exports. The EU agricultural policy’s aim to reduce the environmental impact of farming and to produce milk in a more sustainable way leads to the promotion of an extensive farming model. In this context, the results of the study would be very useful for the selection of policy support measures to maintain the competitiveness of dairy farms on the domestic market and on world markets. The shift towards more extensive farming should not hinder the achievement of the other objectives of the CAP, i.e., to ensure the competitiveness of the sector, food security, and a dignified income from farming. Therefore, farmers should be compensated for losses due to more extensive production practices.

5. Conclusions

An economic analysis of dairy farms in selected EU countries where milk self-sufficiency exceeds national needs is important in order to assess how the sustainability paradigm (the shift towards more extensive production methods) will affect the development and competitiveness of the EU dairy sector in both domestic and foreign markets. Our results show that countries that are leaders in dairy exports obtained the best estimates of dairy farm performance. These countries have the highest farm size, production scale, productivity, and income indicators. Other countries achieve competitiveness in the dairy sector through lower milk prices, higher utilisation of their own resources, and higher levels of support.

However, it must be stressed that our results cannot be interpreted per se as unconditional support for intensive production systems. In fact, further research is needed on the factors influencing the performance of dairy farms in order to find a compromise between the economic performance of farms and environmental and animal welfare requirements.

Our paper aimed to contribute to the debate on the advantages and disadvantages of large farms and intensive and extensive production systems. The results of our analysis provide valuable information on the drivers that determine the level of milk production and the export orientation.

The EU is promoting a shift towards more extensive production technologies to contribute to the sustainability of agriculture by tackling biodiversity loss and GHG emissions. The results of the analysis show a high degree of heterogeneity on dairy farms, production technologies are linked to natural production conditions and future agricultural policies, and public payments may be decisive for the successful achievement of policy objectives. In order to maintain the development of the EU dairy sector and its competitiveness on world markets, the transition must ensure a sufficient level of income for dairy producers.