Efficiency Analysis of Sheep Farms in Cyprus

Abstract

In this study, an empirical analysis was applied to measure the efficiency level of dairy farms in Cyprus and estimate the capacity of sheep farmers to support the increasing demand for halloumi cheese. Data Envelopment Analysis was used on data from 50 dairy sheep farms in Cyprus, which operate under extensive, semi-intensive, and intensive systems. The main features of the most efficient farms are presented, and a comparative financial analysis is implemented between the efficient and less efficient farms. The results indicate room for improvement in extensive and semi-intensive dairy sheep farming and verify that the transition that takes place in sheep farming towards more intensive systems constitutes the optimal approach. The most efficient farms operate under semi-intensive and intensive dairy sheep farming and achieve higher milk yields than the farms operating under extensive systems. Feeding constitutes the main cost driver, exceeding 60% in both efficient and inefficient farms, while labor wages and fixed capital cost varies between 25% and 30% of the total production cost for both efficiency groups. The findings indicate that the farms should utilize economies of scale to reduce production costs and utilize fixed capital endowments at full capacity.

1. Introduction

The sheep population in the European Union (EU) stood at 73.9 million live animals in 2004 and has since followed a downward trend, reaching its lowest point in 2023 with 57.5 million live sheep. In contrast, in Cyprus, there were 278,980 sheep (ewes and lambs) in 2004, and—except for the years 2008 and 2014—the population has continued to grow, reaching its peak in 2023 with 353,530 [1]. This increase in sheep population, but also in milk production and ewes’ yields, is mainly attributed to the increasing demand in the international markets for halloumi cheese.

Halloumi cheese, traditionally made from sheep or goat milk, or a combination of the two, with or without added cow’s milk, has been produced in Cyprus since 1554. Following Cyprus’s accession to the EU in 2004, milk and cheese producer associations worked diligently to register halloumi as a product of Protected Designation of Origin (PDO). The application was submitted in 2014, published in 2015, and halloumi was finally registered as a PDO in 2021 [2]. Since 2004, halloumi has seen a remarkable increase in export volume, becoming the primary export commodity of Cyprus and accounting for 20% of the country’s total value exports. In this regard, dairy sheep farming constitutes an integral part of the Cyprus economy with a considerable contribution not only to national income but also to social cohesion and environmental sustainability [3,4].

To accommodate the increasing demand for halloumi cheese in international markets, high-yield milk rich in fats and protein is required. Consequently, since 2004, milk production has increased by 150%. The sheep sector has experienced a notable population surge of 27%, overshadowing the declining trend in the goat sector, which has seen a decrease of 43%. Between 2004 and 2023, the productivity increased by 196% for goats and 206% for sheep. Driven by this demand and in combination with the grazing restrictions in High Nature Value areas of the country (53.7% of Cyprus’s agricultural land is classified as HNV farmland), the small ruminant sector has been experiencing a rapid intensification process. Extensive pasture-based, low-input and low productivity systems are witnessing a constant decline in favour of semi-intensive systems, more productive and more reliant on capital investments, purchased feedstuff, and hired skilled labour. More lately, highly intensive farms, where animals are kept inside throughout the year with no or minimal grazing, are emerging, aiming to maximize milk productivity and minimize production costs through specialization. As a result, according to the Yearly review of the sheep and goat sector in Cyprus for 2023 [5], the number of animal holdings has decreased. In 2023, the sector was comprised of 2066 farms rearing 353,530 sheep (ewes and lambs) and 233,920 goats (does and kids) in total [1]. At the same time, the size of flocks has increased, and traditional local breeds have been replaced by high-productive crossbreeds [6].

By 2029, it is required that at least 50% of the milk used in halloumi production must be sourced from sheep and/or goats [2]. Ensuring a steady and sustainable milk supply from small ruminants will be vital for maintaining the growth and success of the halloumi market. Regarding the milk marketing model, sheep and goat farmers in Cyprus mainly prefer to deliver their milk to dairy industries and local small dairies [6], and very few produce halloumi on the farm. The dairy industry is growing and specializing to support the development of the halloumi value chain.

Currently, the primary challenge facing halloumi production is the insufficient supply of sheep and goat milk to meet the rising international demand at competitive prices. To address this issue, the sheep sector must enhance its efficiency and adapt to the increased demand for milk. Ongoing activities and initiatives include mainly genetic improvement in the national flock, both in terms of productivity and of feeding efficiency [6], also taking into consideration that Cyprus depends heavily on imported feedstuff. In the pursuit of this increased productivity, it is vital for the sector to properly acknowledge the particularities of the formidable sheep production systems of Cyprus. Multidimensionality must be an integral part of the sheep sector in Cyprus to safeguard the benefits of the traditional, more extensive systems that promote biodiversity, protect environment, provide employment in marginal areas and produce high quality, “greener” products complying this way to the objectives defined in the European Green Deal, Farm-to-Fork, and EU Biodiversity Strategy for 2030 strategies [7,8,9].

The main objective of this paper is to explore the efficiency of Cypriot dairy sheep farms from diverse production systems in the utilization of the existing resources and to detect the sources of (in) efficiency in their operation. More specifically, we want to answer the following questions. First, how much can dairy sheep farms increase their productivity without requiring more of the available inputs? Secondly, which are the optimal farms and under which farming system do they operate? Third, are sheep farms resilient and sustainable in the long run? In addition, farm profitability is evaluated, and financial results such as gross revenues and capital return for the efficient and the inefficient farms are presented in detail.

During our research, primary farm accounting data from 50 dairy sheep farms in 2023 in Cyprus, from the prevailing production systems in the island (intensive, semi-intensive, extensive), have been collected, using a structured questionnaire. Efficiency differences are analysed, and a technical and economic comparison between efficient and inefficient farms, revealing key structural and economic features of the top-performing farms, is performed. Many studies have been conducted to measure the technical efficiency (TE) level of small ruminant sectors in European countries that share common characteristics with the Cypriot sheep sector using Data Envelopment Analysis or Stochastic Frontier Analysis. Kyrgiakos et al. [10] and Toro-Mujica et al. [11] provide a detailed review of these studies and their findings.

This is the first study that estimates the efficiency level of sheep farms in Cyprus and one of the few analysing economic performances. The only recent study that relates to the economic performance of sheep farms in Cyprus is by Hadjipavlou et al. in 2021 [6], who investigated the impact of diverse technical and economic factors on dairy small ruminant farmer market choices. Older studies include the work of Papas and Papachristodoulou in 1975 [12], Panayiotou in 1989 [13], and Papachristoforou and Markou in 2006 [4]. We contribute to the literature by analysing primary farm data from sheep farms that cover the whole spectrum of farming types and all the different climatic and soil zones in the island. The findings of this study can contribute to the debate about the capacity of sheep farmers to provide the required amounts of milk to support the halloumi industry while also securing the position of semi-intensive and extensive farms in the value chain.

2. Materials and Methods

2.1. Data Envelopment Analysis—DEA

Data Envelopment Analysis (DEA) is a data-oriented approach for performance evaluation and improvement. DEA uses mathematical programming techniques and models to evaluate the performance of peer-production units in terms of multiple inputs used and multiple outputs produced [14]. It compares Decision Making Units (DMUs) by measuring how efficiently they convert inputs (i.e., labor, land, capital) into outputs (i.e., revenues, products). DEA constructs a piece-wise frontier over the data and estimates the Technical Efficiency (TE) score of the production units relative to that frontier [15,16]. Each unit uses different amounts of inputs to produce outputs, and efficiency is measured relative to the highest observed performance in the sample, formulating an efficiency frontier. The efficient score ranges from 0 to 1 based on the best-performing farms. The units that lie on the frontier are efficient (efficiency score = 1). The main advantages of DEA are that the method handles simultaneously multiple inputs and outputs, does not require assuming a specific functional form for the production process, and is useful for performance evaluation of various sectors (agriculture, finance, education, healthcare, etc). Researchers in a number of fields, including agricultural economics and animal husbandry, have quickly recognized that DEA is an excellent methodology for modeling operational processes, benchmarking, and revealing best-observed practices [10]. Assuming that there are n production units, DMUj (j = 1, …, n), each producing a single output y, and using multiple inputs (x₁, x₂, …, x_m)., the output-oriented DEA model that allows for varying efficiency scales (VRS) in the linear programming formulation is expressed algebraically for unit k among n units as follows:

Max φ

(1)

Subject to:

$$\sum _{j=1}^n{\lambda }_j{x}_{\mathit{ij}}=x_{i\kappa }\forall i\left(\mathit{input}\mathit{constraints}\right)$$

(2)

$$\sum _{j=1}^n{\lambda }_j{y}_j=\phi y_{\kappa }\left(\mathit{output}\mathit{constraint}\right)$$

(3)

$${\lambda }_j\ge 0 \forall j$$

(4)

$$\sum _{j=1}^n{\lambda }_j=1 \left(\mathit{convexity}\mathit{constraint}\right)$$

(5)

where x_ij = input i_i of unit j_j, y_j = output of unit j_j, x_ik, y_k = observed inputs and output of DMU k, where TE = 1/φ is the technical efficiency score (ϕ ≥ 1) and λj is the weight of production unit j in the reference set. If ϕ = 1, then the production unit k is efficient (the production unit is on the efficient frontier, no other unit operates better, and the unit cannot produce more output with the given inputs). If φ > 1, then the unit k is considered inefficient and can increase output by (ϕ − 1) × 100%. Hence, in the output-oriented models, the efficiency score reflects how much output can be scaled up without requiring additional inputs [17]. One of the most attractive features of DEA is that it assesses efficiency using sample real-life ‘best practice’ benchmarks instead of a theoretical maximum [18], and the coexistence of different farming types that do not apply a standardized production technology with homogeneous characteristics produces a high TE score with high variance. A detailed description of the DEA methodology, the various DEA models, and their extensions can be found in [14,19,20,21].

2.2. Empirical Model

Technical and economic data collected from 50 dairy sheep farms in 2023 were used for this empirical analysis. These farms were in the rural areas of the municipalities of Nicosia (8 farms), Limassol (9 farms), Larnaca (17 farms), Famagusta (10 farms), and Paphos (6 farms), representing the main soil and climatic conditions in Cyprus. The sampling framework and its geographic distribution were developed based on official livestock data [5] and expert knowledge. A stratified random sampling technique was applied to ensure the proportional representation of the prevailing dairy sheep production systems in Cyprus: extensive, semi-intensive, and intensive.

Within each stratum, farms were randomly selected to participate in the survey. The farms were discerned into the three farm types based on the level of grazing. Farms in which sheep are mainly grazing are extensive, whereas those relying primarily on the use of concentrated feed and animals have zero or limited access to pasture are considered intensive. The farms in which feeding is based on purchased or on-farm produced forage and concentrates supplemented by grazing are classified as semi-intensive

This typology accounts for the main production systems of the island and complies with the typology developed within the iSAGE (Innovation for Sustainable Sheep and Goat Production in Europe) HORIZON Project (www.isage.eu, accessed on 4 May 2025) and described in Theodoridis et al. [22]. Most of the sample farms were semi-intensive (22 farms, 44% of the sample), while 19 farms (38%) and 9 farms (18%) of the sample farms were intensive and extensive, respectively.

Data were collected using a questionnaire to record fixed capital endowments (facilities and machinery, land reclamation, flock composition and value), labor requirements and salaries for family and hired workers, land and inputs for the production of feedstuff (acreage and land rent, expenses for on-farm production of feed (seeds, fertilizers, pesticides, fuel, irrigation), purchased feed (quantities and prices), expenses related to animal production (fuel, detergents, water, drugs, veterinary services etc.), as well as milk yields and prices, meat yields and prices and various types of farm subsidies. Based on this accounting data, technical and economic indices were estimated at the farm level.

An output-oriented VRS DEA model was applied using the DEA Frontier software (version 2008) [23]. Following the model specification implemented in Theodoridis et al. [24], the following input variables were used in the DEA model: the flock size (number of ewes), the total human labor measured in annual hours, the variable cost including expenses for feeding stuff (procured and produced on the farm) expenses for fuel, water, drugs, veterinary services etc., measured in € and the fixed cost (annual expenses of fixed capital) measured in €. As an output variable, the gross revenue (products × prices, including budgetary payments that consist of all forms of subsidies (per hectare direct payments and per head support)) was used, measured in €, to consider also the effect of the milk price. With the outcomes of the DEA model as a starting point, a comparative technical and economic analysis revealed differences in the financial results and technical indicators of the sample based on their level of TE (efficient and inefficient farms) as well as of the three farm types under examination. This approach reveals the practices that the best farms utilize, implying re-adjustments that can be effectuated in inefficient farms to maximize their performance.

3. Results and Discussion

The output-oriented VRS DEA model (Variable Returns to Scale Data Envelopment Analysis) is a method used to measure the technical efficiency of farms by evaluating how much they can increase their outputs using the same level of inputs, compared to the best-performing peers. It assumes that farms may operate under different scales (not all at optimal size) and focuses on maximizing output. The efficiency analysis using the output-oriented VRS DEA model reveals notable performance disparities among the dairy sheep farms in the sample. The average technical efficiency (TE) score is 0.844. The standard deviation of 0.145 further underscores substantial variability, suggesting that some farms operate near peak efficiency, while others significantly underperform. These dairy sheep farms could increase their gross revenues by 16.6% using the same level of inputs, provided that the farmers apply the best observed practices in the utilization of existing resources.

The classification of farms by efficiency level in

Table 1. Frequency distribution of technical efficiency (TE) scores.

Efficiency Score	Data Envelopment Analysis
	No of Farms	% of Farms	Mean TE
<0.7	9	18%	0.608 (0.037) *
0.7–0.8	10	20%	0.747 (0.026)
0.8–0.9	6	12%	0.850 (0.037)
0.9–1.0	14	28%	0.941 (0.018)
1.0	11	22%	1.000 (0.000)
Total	50	100%	0.844 (0.145)

* Figures in the parentheses are standard deviations.

provides additional insight into the structure and dynamics of the sector. In the full efficiency class, the dominating farm type is the semi-intensive (8 of the 11 farms, 22% of the farms), suggesting this production system may be more conducive to efficient resource use. A proportion of 28% of the farms (14 of the 50 farms) operate near the efficient frontier, having an efficiency score between 0.9 and 1.0 (Mean TE score of 0.941), with the extensive farm type dominating this class, with 9 of the 14 farms. Six farms, accounting for 12% of the sample, exhibit efficiency scores between 0.8 and 0.9, with a mean efficiency score of 0.850, dominated by the semi-intensive farms (three of the six farms). Ten farms (20% of the sample) have TE scores between 0.7 and 0.8. In this efficiency class, domination is shared between semi-intensive and intensive types of farms (4 of the 10 farms for each type). Only nine farms (18% of the sample) lag considerably from fully efficient farms by achieving TE scores less than 0.70 (mean efficiency score of 0.608). The extensive type of farm dominates this efficiency class (five out of the nine farms). This stratification suggests that semi-intensive systems tend to be associated with higher efficiency, while extensive systems show more variability, with both high-performing and lagging farms.

To ensure the robustness of the results, alternative DEA specifications were applied. The Input-oriented VRS model resulted in a Mean TE = 0.873, while the Measure-Specific VRS model (with herd size and fixed capital treated as non-controllable inputs) resulted in a Mean TE = 0.828. In addition, the output-oriented VRS model with five input variables resulted in a Mean TE = 0.869. These values are consistent with the base model, and importantly, the identity of fully efficient farms remains unchanged across all DEA variants. This consistency reinforces the reliability and stability of the core findings.

In

Table 2. Technical Efficiency (TE) scores for Cypriot dairy sheep farms per farm type.

Farm Types	No of Farms	No of Inefficient Farms (TE < 1)	No of Efficient Farms (TE = 1)	Mean TE
Extensive	9	8	1	0.693
Semi-intensive	22	14	8	0.886
Intensive	19	17	2	0.868
Total sample	50	39	11	0.844

, the TE scores of the farms per farm type are presented. Most of the efficient farms were semi-intensive (72.7% of the most efficient farms), although they account for 44% of the sample farms, while only two efficient farms are intensive (18.2%), and one (9.1%) is extensive. The semi-intensive and intensive farms exhibit the higher Mean TE score (0.886 and 0.868, respectively), while the Mean TE score of extensive farms is considerably lower (0.693). These findings show that the farms that adopt a more intensive production pattern have a higher level of managerial ability and allocate their resources more rationally. On the other hand, the extensive farms can significantly increase their efficiency and improve their resilience if they are willing to apply modern management practices and differentiate their marketing strategies [24].

The Mean TE of the sample farms is similar to that found by Theodoridis et al. [25] for intensive dairy Chios sheep farms in Greece (TE = 0.8) and to that found by Theodoridis et al. [26] for dairy farms that reared either Manech or Basco-bearnaise and Lacaune breeds in the Western Pyrenees and Roquefort areas in France (TE score = 0.857). Moreover, Theodoridis et al. [26] reported in 2021 a lower Mean TE value for French semi-extensive meat sheep farms (0.71), which is, however, similar to the Mean TE score estimated for the extensive farms in this study (Table 2). This finding also conforms with the findings for organic dairy sheep farms in Spain; 0.66 in Toro-Mujica et al. [11]. Although no efficiency studies have been conducted on sheep farms in Cyprus, the diversity of farm types in Cyprus reported by Papachristofou and Markou [4] in 2006 and Hadjipavlou et al. [6] in 2021, has been acknowledged in our farm sampling and is reflected in the estimated efficiency scores, in the construction of the efficient frontier and of course in the selection of the best farms.

The DEA model applied in this study aims to attain efficiency targets by maximizing output levels using the same level of resources [23]. These feasible improvements in production value for each farm type are presented in Figure 1. The results, which reflect the potential revenue gains within the production possibility set (1 – θ) [17], indicate that the value of production on Cypriot sheep farms could increase substantially, particularly under extensive systems. Specifically, extensive farms could achieve a 57.7% increase in output, raising the average value of production from €170.6 thousand to €269.0 thousand per farm through optimized management practices. For semi-intensive farms, a potential increase of 11.1% is observed, corresponding to a rise in average output from €340.0 thousand to €378.0 thousand per farm. Intensive farms could experience a 14.3% increase, with average output improving from €470.3 thousand to €537.3 thousand, provided that best-observed production practices are implemented.

These findings indicate that there is high potential for increasing the productivity of dairy sheep within its farm type through enhanced management practices to meet the increasing demand for sheep milk from the dairy industry. The estimation of the improvement in farms’ productivity is based on all sample farms that belong to diverse production systems, which means that the character and the type of farms (especially the more extensive) can be significantly altered during the transition towards higher efficiency. Hence, one must be cautious and make suggestions about the appropriate production and management practices, always in the context of the specific farming type. Additionally, it should be noted that the DEA model applied in this study is output-oriented and no input expansion is required (e.g., more land and labor) to achieve the efficiency targets, but better allocation and utilization of existing resources.

Sample farms were categorized into two distinctive efficiency groups, i.e., inefficient and efficient farms, based on their TE score.

Table 3. Technical features of the efficient and inefficient farms.

Technical Data	Efficiency Groups		Average Farm TE = 0.844 (n = 50)
	Inefficient TE < 1 (n = 39)	Efficient TE = 1 (n = 11)
Herd size (ewes per farm)	624	776	658
Total production (×1000 lit/farm)	141	257	166
Milk yield (lt/ewe)	225	331	252
Land (ha/ewe) *	0.175	0.166	0.172
Non-irrigated (ha/ewe) *	0.152	0.136	0.147
Irrigated (ha/ewe)	0.023	0.030	0.025
Grazing land (ha/ewe)	0.42	0.13	0.34
Total labor (h/ewe/year)	13.58	14.88	13.91
Family	4.36	3.87	4.23
Hired	9.22	11.01	9.68

* expressed in irrigated equivalent hectares.

outlines the key technical indicators for the average farms in both efficiency groups, providing valuable insights into their structure, organization, and productivity level. The most efficient farms are larger in size, rearing on average 776 ewes, 24.4% more animals than the less efficient farms, and achieve higher milk yields, producing 331 L of sheep milk per ewe annually, 47.1% higher than the inefficient farms (225 L/ewe). The increased milk yield is mainly driven by economies of scale since the production cost of milk decreases with further intensification by increasing variable capital and by the increasing demand for halloumi. No significant difference is indicated in the use of land for home-grown feed and human labor among the efficient and inefficient farms, although it was expected that the most efficient farms, which are larger in size, utilize better available labor due to economies of scale, high-skilled labor, and the use of labor-saving modern technologies. On the contrary, efficient farms work 1.3 h more per ewe and resort mainly to hired labor (74% of total human labor used in the farm). The less efficient farms also use hired workers (9.22 h/ewe annually), but to a lesser extent (68% of total human labor). A similar finding was reported for transhumant farms in Greece [27], where only better-performing farms resorted to hired labor because only for them the productivity of additional labor could justify the additional labor costs. Moreover, the animals reared on the most efficient farms are mainly confined to modern facilities and do not graze. In general, the analysis shows that most of the efficient farms are large farms operating under more intensive patterns, relying on hired labor and the provision of concentrate feed, having limited or no access to pastures.

The results presented in

Table 4. Composition of gross revenue per ewe.

Composition of Gross Revenue	Inefficient TE < 1 (n = 39)		Efficient TE = 1 (n = 11)		Average Farm TE = 0.844 (n = 50)
	€/ewe	%	€/ewe	%	€/ewe	%
Milk (sold to dairies)	324	67.2	486	66.8	365	66.9
Lamb Meat	90	18.7	154	21.1	107	19.6
Cull animals’ meat	9	1.9	13	1.8	10	1.8
Cheese	3	0.6	9	1.2	5	0.9
Budgetary payments	56	11.6	66	9.1	59	10.8
Total revenues (€/ewe)	482	100.0	728	100.0	546	100.0
Total revenues (€/farm)	300,890		564,996		358,994

show that there is a big difference in the gross revenues between the two efficient groups; the efficient farms achieve 728 €/ewe, i.e., 51% higher revenues than the inefficient farms, which is attributed to the high milk yields of efficient farms. The gross output for the whole sample farm is 546 €/ewe. Unfortunately, there are no results regarding the economic performance reported from other studies in Cyprus. The composition of the revenues does not differentiate substantially between the efficiency groups. The revenues from milk deliveries to industries are the main source of income in all efficiency groups, contributing 67% on average to their gross output. The second-largest revenue stream is lamb meat production, contributing 19.6% on average. Notably, this share is higher in relatively efficient farms (21.2%) than in fully efficient ones (18.7%). The share of the value of culled animals contributes by 1.8% in total revenues, while the value of on-farm cheese production is very low, contributing by 0.9% to the total sample farm (0.6% and 1.2% for the inefficient and the efficient farms). The analytical results show that only 8 of the 50 farmers produce cheese in their facilities, usually for their own consumption.

In general, farms with diversified income sources are better able to adapt to change. In contrast, heavy reliance on a single product reduces resilience to economic shocks [28]. Since milk production represents a major share of gross revenues for Cypriot sheep farms, both groups, since the average milk price for the inefficient group of farms is €1.44/litter and €1.46/litter for the efficient group, are particularly vulnerable to fluctuations in milk prices. This vulnerability is also related to the role of budgetary payments in their overall income. The share of budgetary payments for the whole sample farm is not trivial, accounting for 10.8% of the gross revenue, a share which is higher in the inefficient (11.6%) than in the efficient farms (9.1%). The Cypriot sheep farms exhibit moderate to low resilience to policy shifts, and sudden policy changes can destabilize their farm incomes. Moreover, the farms have low adaptive capacity since the limited product diversification and farms’ reliance on a single commodity (milk) leads to intensified risks if policies shift production focus and alter market conditions (such as trade rules and price supports). These findings align with prior studies in the meat sheep sector (covering both extensive and intensive farming systems) [27], transhumance systems [29], and the broader agricultural sector [30,31], which similarly identified a negative relationship between public subsidies and farm efficiency.

The cost structure for both efficient groups and the average sample farm is presented in

Table 5. Cost structure of efficient and inefficient farms.

Expenses per ewe	Efficiency Groups		Average Farm (TE = 0.844) (n = 50)
	Inefficient TE < 1 (n = 39)	Efficient TE = 1 (n = 11)
Land rent (€/ewe)	15.35	12.38	14.58
Labor wages (€/ewe)	59.76	70.79	62.63
Feed cost (€/ewe)	286.15	363.83	306.32
-Home-grown feed (€/ewe)	38.23	37.91	38.15
-Purchased feed (€/ewe)	247.92	325.92	268.17
Other variable costs (€/ewe)	34.18	63.27	41.73
Fixed capital cost (€/ewe)	59.43	93.64	68.32
Total expenses (€/ewe)	454.87	603.91	493.58
Milk production cost (€/lit) *	1.37	1.24	1.33

* using the proportional costing method.

. In line with the use of human labour, labour cost on efficient farms—which encompass both the alternative cost of family labour and hired labour wages—was €11 higher than in the inefficient farms (18.4% increase), while feeding cost was increased by 27% (77 €/ewe higher on the efficient farms). Purchased and home-grown feed represent the primary cost factor in a farm, and their effective use is highly associated with milk yields and, hence, productivity and sustainability [32,33,34]. In Cyprus, the expenses for feed account for 62% of the total expenses, which is in line with the findings of Åžahinli and Özçelik [35], Milan et al. [36], and Theodoridis et al. [26]. Of course, the exact share of feeding cost in small ruminant farms varies by production system and region/country. The results show that the efficient farms depend more heavily on purchased feed (89.5% of total feeding cost), which indicates that feed procurement from the market enhances efficiency, although overreliance on purchased feed weakens resilience against market shocks. Market-dependent farms are more vulnerable to sudden price increases and supply chain disruptions. In general, as the livestock production systems adopt more intensive patterns, the farmers prefer to purchase concentrate feed from the market and to produce forage on the farm [37,38,39]. However, it should be stressed that due to the climate and soil conditions in Cyprus, higher rates of on-farm—or even of domestic—feedstuff production are not a realistic option.

Moreover, fixed cost, which is the second most critical cost driver in sheep farms after feed cost, and comparable to the cost of labor input, is much higher in the efficient farms than in the inefficient (93.64 €/ewe and 59.43 €/ewe, respectively), although fixed cost is allocated to a larger herd size in the efficient farms. On the whole sample, farm fixed cost accounts for 13.9% of the total production cost (15.5% of total cost in efficient farms compared to 13.06% of total cost in inefficient farms). In most of the efficiency studies, it is indicated that the fixed cost per productive animal is reduced as the level of efficiency increases because farms utilize economies of scale and less efficient farms are characterized by unreasonable, ill-advised investments and mismanagement of capital [26,40]. The findings in this study suggest that most of the efficient farms are herd-size expanding and have made high investments in buildings, machinery, and high-productive animals, on which they should capitalize in the short term through further intensification by using more variable capital.

On average, the total cost for the average farm was 493.58 €/ewe, 603.91 €/ewe for the efficient, and 454.87 €/ewe for the inefficient farm. However, although the total expenses per ewe are higher by 32.8% in efficient farms, the milk production cost is lower (1.24 €/lit) than in inefficient farms (1.37 €/lit), which illustrates the high productivity of the efficient farms per ewe and also per farm. We were not able to benchmark these production costs with previous studies, because milk production cost varies among countries but also among regions in the same country, depending on the key cost drivers, which are defined by the production system, the soil and climatic conditions, and the structural characteristics of the farms. Although there are available data on the farm-gate prices at the EU level (https://www.clal.it/en/?section=latte_europa, accessed on 13 May 2025), there are no recent studies on the cost of sheep milk production in Cyprus, and comparisons with other countries could lead to fallacious results.

The financial results (

Table 6. Financial results of the efficient and inefficient farms.

Financial Results (€/ewe)	Inefficient TE < 1 (n = 39)	Efficient TE = 1 (n = 11)	Average Farm TE = 0.844 (n = 50)
Gross revenues	481.97	728	545.84
Variable cost	320.33	427.09	348.05
Gross margin *	161. 64	300.91	197.79
Farm Income **	124.33	207.81	146
Capital return (%) ***	4.77%	10.67%	6.82%
Fixed cost	134.54	176.81	145.53
Profit or loss	27.09	124.09	52.27

* Gross Margin = Gross revenues − Variable cost. ** Farm Income = Land rent + Labor wages + Interest + Net Profit. *** Capital return (%) = [Net Profit (capital interest included)/Capital Employed (variable cost included)] × 100.

) indicate that a higher level of efficiency is associated with a higher value of production. This aligns with the findings of [11,24,25,39,40,41], who estimated the level of technical efficiency in dairy sheep farms. Consequently, gross margin (gross revenue minus the variable cost) was also higher in the efficient farms by 86.2% compared to inefficient ones, from 161.64 €/ewe to 300.91 €/ewe, revealing that effective management is essential for the operation of a modern dairy farm, particularly in an uncertain and volatile economic environment. Soaring inflation tightens farmers’ budgets, while rising costs of fuel, feed, and agrochemical inputs outstrip any increases in milk prices. Strong managerial skills—financial planning, income optimization, and strategic decision-making—become vital to sustaining operations, improving efficiency, and ensuring long-term viability in an increasingly challenging industry. The promising outlook for sheep farming in Cyprus is further supported by the high capital return, which ranges from 4.77% in inefficient farms to 10.67% in the efficient ones. The capital return gap between the two farm groups indicates that efficient farms make better use of their investments and demonstrates that intensification is both feasible and profitable for the island’s dairy sheep sector.

Another interesting result is that net economic margin (profit or loss), which is an important indicator for farms of entrepreneurial nature, is positive for both efficient and inefficient farms (124.09 €/ewe and 27.09 €/ewe, respectively). Nevertheless, it merits to be pointed out that, if budgetary payments are not included in gross revenues, inefficient farms operate with net losses (27.09 €/ewe − 56.00 €/ewe = −28.91 €/ewe), indicating that these farms are less resilient with negative prospects under unfavorable policy environments while efficient farms are economic viable in the long term and have strong potential (124.09 €/ewe − 56.00 €/ewe = 68.09 €/ewe).

4. Conclusions

The main conclusion of the analysis of the efficiency of a sample of 50 dairy sheep farms in Cyprus was that intensification has beneficial results on farm productivity and efficiency. The results indicated that higher capital expenses for feedstuff and fixed capital, along with larger flock sizes and more use of hired labor, increase productivity and reduce milk production costs. The fact that intensification is an ideal pathway for sheep farms in Cyprus is also reflected in the fact that feeding expenses are high—due to low feeding autonomy compromised by the soil and climate conditions of the island—as well as in the high capital investments undertaken, which require further specialization. Since milk production represents a major share of gross revenues for sheep farms, milk price is a core factor for the economic viability of the sector. The Common Agriculture Policy (CAP) budgetary payments play a significant role in the economic viability of the sector, especially for the inefficient group of farms. The measures of the policy in force are critical for balancing viability, financial stability, and sustainability in sheep farming, and future CAP reforms must address disparities for smallholders and rare breed custodians to ensure sector resilience. However, the inefficient and most vulnerable farms must not rely only on the CAP payments, but they should adopt the optimized management practices applied by successful operations and invest in cutting-edge technology, highly skilled workers, proper nutritional management, and modern marketing strategies. Key focus areas that are associated to the efficiency level of the farms also include production practices such as tailored, genomic-based breeding programs aligned with specific goals, routine data collection, use of a specific criteria for choosing the best replacement animals, use of elite flocks, advanced reproductive techniques such as assisted reproduction techniques (artificial insemination), frequently reviewed use of rams and reproduction plans, efficient feeding systems including increased forage and pasture quality, innovative grazing practices, use of by-products, and digital technology applications.