Sustainable Coexistence: Wind Energy Development and Beekeeping Prosperity—A Propensity Score Matching Approach

Abstract

Beneath the promise of clean energy, the rapid rise of wind energy farms has stirred mounting concern for pollinator-dependent livelihoods—particularly in beekeeping. This study investigates the effect of wind energy farms on honey-related income using data from six provinces in Turkiye’s Aegean region and the propensity score matching method. Results show that beekeepers operating near wind energy farms experience significantly higher incomes—an average treatment gain of 45,107 TL, with treated groups earning 56,515 TL more—backed by several robust statistical evidence such as placebo and bootstrap techniques. Certain groups—such as younger, nomadic, and family-trained beekeepers, and those receiving financial support—exhibit greater resilience. The findings highlight the need for land-use strategies that balance renewable energy development with ecological and economic concerns. Introducing bee-friendly vegetation around turbines is proposed as a practical solution. This approach can foster a mutually beneficial relationship between wind energy farms and beekeeping, supporting both rural livelihoods and the broader goals of sustainable development.

1. Introduction

Energy is an indispensable component of contemporary industrial processes, playing a pivotal role in elevating living standards, propelling socioeconomic advancement, and enhancing societal well-being [1,2]. Strategically pivotal in shaping the future of the world, energy is recognized as a critical resource. However, it is imperative to acknowledge the ongoing necessity of energy to support life and promote long-term economic expansion [3,4]. The prevailing concern of our era pertains to the sustainability of energy, as global energy demand continues to escalate at a rate of 4–5% annually. This escalation can be attributed to a multitude of factors, including the rapid expansion of industrialization, population growth, rising consumption levels, technological advancements, and rising living standards [5,6]. Non-renewable energy sources, particularly fossil fuels such as coal, oil, and natural gas, which are intimately intertwined with myriad aspects of daily life, have been the chief means of addressing this need [7]. The utilization of fossil fuels has been the predominant catalyst for the surge in primary energy consumption on a global scale over the past four decades [8]. Nevertheless, the overreliance on these energy sources has led to a considerable surge in atmospheric carbon dioxide (CO₂) levels, thereby expediting global climate change, amplifying the greenhouse effect, and disrupting ecological equilibrium [7]. In 2016 alone, approximately 50 billion tons of CO₂ equivalent were released into the atmosphere, constituting 73.2% of global greenhouse gas (GHG) emissions [9]. The imperative for transitioning to alternative, sustainable energy sources is underscored by the limited availability of fossil fuel reserves, which are estimated to deplete within centuries despite their geologic origins. These reserves have profound long-term ramifications for the environment, atmosphere, and human health [4,7,10]. Consequently, renewable energy has emerged as a pragmatic and indispensable substitute for fossil fuels [11]. Renewable energy plays a pivotal role in the clean energy transition by reducing reliance on imported energy, conserving natural resources, and enhancing energy security [12]. In light of these benefits, governments worldwide are intensifying their efforts to adopt ecologically beneficial, health-conscious, and sustainable energy solutions [13]. Among the various clean energy sources, wind energy farms (WEF) have emerged as the most popular, fastest-growing, and most promising option, attracting substantial investment [14,15,16,17]. In 2022, wind energy farms exhibited the most substantial growth among all energy sources, with an increase of 17% (+312 TWh) from 1848 TWh in 2021 to 2160 TWh in 2022, marking the most significant absolute increase in global power generation [18].

The inherent qualities of wind energy, including its clean, limitless, and eco-friendly characteristics, have been demonstrated to promote economic growth, reduce reliance on imported energy, and establish a viable pathway towards a safe and sustainable energy future [7,14,19]. Moreover, the reliability of wind energy has been identified as a secure resource for the foreseeable future [20]. The environmental impacts of wind energy are notable for their absence of deleterious effects on natural flora or human health and for their absence of CO₂ emissions, contributions to acid rain, or warming of the atmosphere, in contrast to the effects of fossil fuels [20,21]. Furthermore, wind energy avoids radioactive consequences and conserves fossil fuels [20]. The economic impact of wind energy is also beneficial, contributing to higher overall revenue and job growth due to its rapidly advancing technology [12,20,21]. A study conducted in Croatia between 2007 and 2016 revealed that WEF had the greatest economic impact in 2013, producing €91 million in gross domestic product (GDP) and 3766 full-time equivalent (FTE) jobs. By contrast, the operation of WEF through the intermediate consumption channel generated approximately 72% of the total gross value added (GVA) attributed to their functioning. Unlike the investment channel, this pathway revealed stronger indirect employment effects. In 2016 alone, wind energy farm operations contributed €114 million in GVA to the Croatian economy and supported 481 full-time equivalent (FTE) jobs [22]. Moreover, wind energy’s sustainability and economic feasibility position it as a leading renewable energy source [23]. Consequently, numerous countries worldwide are developing WEF to supplant fossil fuels and augment the share of clean energy in their national grids [24].

Despite the numerous advantages associated with WEF, inadequate installation practices can give rise to certain disadvantages. These include radio and television signal interference within a 2–3 km radius, noise pollution, adverse landscape alterations, and bird fatalities [25]. Moreover, WEF has the potential to adversely impact the ecosystem through mechanisms such as soil erosion, forest loss, and vegetation damage [20]. The adverse impacts on migrating birds, bats, and bee colonies disrupt the natural equilibrium, as evidenced by studies conducted by Nazir et al. [26], Gürbüz et al. [27], Guest et al. [28], and Leroux et al. [25]. Additionally, noise, stray voltage, air pressure fluctuations, turbulence, and electromagnetic fields have been identified as potential sources of disturbance to bees [29]. For instance, flying against the wind can result in bees expending energy, consuming a significant amount of the nectar they gather, and making it more difficult for them to visit flowers and carry pollen and nectar [30,31]. Humming noises and stray voltage can also cause disruptions in bee behavior [29]. Furthermore, it has been observed that wind can disperse pollen, desiccate nectar, and impede bees’ access to it [30]. Concerns have been raised regarding the potential for shade, light flicker, collisions, and turbine blade vibrations to confuse bees. Moreover, studies have also identified stray voltage and blade noise as factors that may exacerbate the negative impacts [25,28,32,33]. Consequently, recognizing and mitigating these hazards is imperative to ensure the continuity of beekeeping operations and the generation of revenue from honey production. A rigorous examination is necessary to determine the impact of beehives situated in proximity to WEF on honey production and the subsequent income derived from this activity. This investigation will contribute to a more comprehensive understanding of the issue.

This study investigates the impact of WEF on beekeepers’ honey production income, focusing particularly on the economic implications in Turkiye. By empirically estimating the influence of proximity to WEF on honey-related revenue compared to a counterfactual scenario in which beekeepers operated at a greater distance, the study provides a robust economic evaluation of the local effects of renewable energy. The current research makes several novel contributions to the literature. Firstly, although the biological and behavioral effects of wind turbines on bee colonies have been widely studied [25,28,32,33,34,35], this is the first study of its kind to adopt an economic perspective, linking these ecological dynamics directly to honey production income. Secondly, the study employs both conventional statistical methods and the Propensity Score Matching (PSM) technique to generate robust empirical evidence and deliver a rigorous analysis of income disparities from honey production—not only as a function of proximity to WEF, but also across beekeepers with varying operational characteristics, regardless of their spatial location. Thirdly, it contributes to the emerging discourse on the interdependence between renewable energy infrastructure and agricultural livelihoods, an area that remains under-explored in the context of sustainable development. By examining individual and operational characteristics alongside spatial factors, the study provides a nuanced understanding of how environmental infrastructure can shape rural economies. Conducted in Turkiye’s rapidly expanding wind energy farm region along the Aegean coast, the findings are locally relevant and have wider global significance. They highlight the potential for agricultural activity to coexist with renewable energy development, emphasizing the importance of strategic infrastructure planning to maximize ecosystem services and rural economic resilience. Furthermore, these findings have implications that extend beyond Turkiye, offering a transferable framework applicable to countries with similar geographical and topographical features. This study provides a solid basis for developing informed energy infrastructure and rural development strategies that are aligned with wider sustainable development goals. By presenting beekeeping as both an ecological indicator and a beneficiary of renewable energy infrastructure, the research encourages policymakers to recognize that agriculture is a complementary aspect of the green energy transition, rather than being in competition with it for land use. Building on the reviewed literature and theoretical considerations, the following hypotheses are formulated: Is there a statistically measurable difference in honey-related income among beekeepers based on their proximity to WEF, regardless of whether the effect is positive or negative? Should such a relationship be identified, the analysis will further assess whether this difference persists after controlling for beekeeper and business characteristics, including age, education, and insurance status, among others. Establishing and evaluating this association would provide valuable insights for the formulation of policies that effectively balance the expansion of clean energy with biodiversity conservation and the strengthening of rural resilience.

The sections that follow are arranged as follows: The “Materials and Methods” section will address the data collected and the methodology employed, while the “Results and Policy Implications” section will be thoroughly reviewed first, providing a concise and perceptive summary of the main conclusions of the study as well as their wider importance. Lastly, the “Conclusion” section will provide the study with a convincing wrap-up.

2. Data and Methods

2.1. Data

This study utilizes a comprehensive survey that was conducted in Turkiye’s Aegean Region, a pivotal area for wind energy farms and beekeeping. The Aegean Region accounts for 20% of the nation’s beekeeping operations and approximately 55% of its total wind energy potential. The study encompasses six pivotal provinces: Aydin, Balikesir, Manisa, Canakkale, Izmir, and Mugla. Initially, a stratified sample strategy was considered for the selection of beekeeping businesses; however, data restrictions posed a significant challenge, precluding the collection of specific business-level information. Consequently, proportional sampling was employed in lieu of stratified sampling, as the latter was deemed impractical. The availability of total hive counts by province from the Turkish Statistical Institute (TSI) and the Ministry of Agriculture and Forestry (MAF) was acknowledged; however, detailed data on the distribution of hives among individual businesses were not provided. A statistical power analysis determined that a minimum sample size of 135 surveys per group (WEF and non-WEF) was necessary to achieve the desired level of precision. To enhance data reliability and mitigate potential response errors, the sample size was increased by around 10%, resulting in a total of 300 surveys—150 from beekeepers operating near WEF and 150 from those in non-WEF areas. In line with established methodological frameworks and given the relatively uniform distribution of hives across provinces, the surveys were distributed evenly across six provinces. Specifically, 25 surveys were conducted in each province, with 25 in the WEF category and 25 in the non-WEF category. This ensured geographic representativeness and comparability between the treatment and control groups. The study population consisted exclusively of sedentary (non-nomadic) beekeepers, whose fixed apiary locations enabled a more precise evaluation of the spatial impact of WEF. The WEF exposure zone was defined as the area within a 10-km radius of wind energy farms, in line with previous studies indicating that turbine-related impacts are typically observed within a range of 10 to 20 km [36,37,38]. Beekeepers located beyond this radius constituted the control group. To capture data reflective of the most pertinent period for honey production and associated revenue, all surveys were administered during the honey harvest season in September and October. This timing was chosen to maximize the accuracy and consistency of responses relating to production output and income. Finally, all statistical analyses for this study were performed using the R programming environment.

2.2. Method

The economic impact of honey production in WEF areas necessitates the evaluation of the income levels of beekeepers operating within these contexts. The primary objective of this study is to estimate the income differential that beekeepers in WEF areas would have experienced if they had not been producing honey in these zones. To this end, beekeepers operating outside of designated WEF zones are utilized as a reference group, with these individuals being considered as though they represented the counterfactual scenario for WEF-based beekeepers. This is due to the inherent impossibility of being within and outside of a WEF area simultaneously. The employment of this methodology facilitates the utilization of the PSM method, thereby enabling a meaningful comparison to be made. This is achieved by implementing Nearest Neighbor Matching (NNM) with replacement, a reliable technique that ensures an objective assessment of the treatment effect. In order to quantify the influence of WEF areas on beekeepers’ profits, we estimate the Average Treatment Effect (ATE) as well as the Average Treatment Effect on the Treated (ATT).

PSM is a statistical matching method that takes into consideration discernible variations between the control group, consisting of beekeepers operating outside of WEF areas, and the treatment group, consisting of beekeepers working within WEF areas. It lessens selection bias by ensuring that comparisons are made between beekeepers with comparable traits (e.g., age, experience, production capacity, geographic location, etc.). Utilizing a logit regression model, the propensity score, which denotes the probability of a beekeeper operating within a WEF area, is calculated as follows [39,40,41,42,43]:

$$e\left(X_{\mathrm{i}}\right)=P\left({\mathrm{T}}_{\mathrm{i}}=1|X_{\mathrm{i}}\right)={\displaystyle \frac{1}{1+e^{-\left({\mathsf{\beta }}_0+{\mathsf{\beta }}_1{\mathrm{X}}_{\mathrm{i}1}+\dots +{\mathsf{\beta }}_{\mathrm{k}}{\mathrm{X}}_{\mathrm{ik}}\right)}}}$$

(1)

where the treatment variable is denoted by T_i, taking a value of 1 if beekeeper i is located within a WEF region (treatment group) and 0 if beekeeper i is located outside the area (control group). The vector of beekeeper i’s attributes is denoted by X_i. In order to ensure a more accurate comparison, beekeepers with comparable probabilities of beekeeping in WEF areas can be paired using the propensity score e(X_i). A beekeeper from the control group with a comparable propensity score is paired with each beekeeper in the treatment group using the NNM approach. The revenue of beekeeper i in the treatment group (e.g., WEF region) is shown here by Y_1i, and the income of beekeeper i in the control group (e.g., non-WEF region) is shown here by Y_0i. During the matching process, each beekeeper in the treatment group is paired with one or more control beekeepers. The ATE, which captures the overall impact of WEF area operations on honey production income, is estimated as follows:

$$\mathrm{A}\hat{\mathrm{T}}\mathrm{E}={\displaystyle \frac{1}{\mathrm{N}}}{\displaystyle \sum _{\mathrm{i}=1}^{\mathrm{N}}\left({\mathrm{Y}}_{\mathrm{i}}^{\mathrm{T}=1}-{\mathrm{Y}}_{\mathrm{i}}^{\mathrm{T}=0}\right)}$$

(2)

where ${\mathrm{Y}}_{\mathrm{i}}^{\mathrm{T}=1}$ and ${\mathrm{Y}}_{\mathrm{i}}^{\mathrm{T}=0}$ reflect the average amount of income (in Turkish Lira, TL) from honey production in matched groups. On the other hand, ATT focuses solely on the income effect for beekeepers already operating in WEF areas, defined as follows:

ATT=E[Y₁−Y₀∣T=1]

(3)

which measures the difference between the actual income of treated beekeepers and the estimated income they would have earned had they not operated in WEF areas. The empirical estimation of ATT is given by the following formula:

$$\mathrm{A}\hat{\mathrm{T}}\mathrm{T}={\displaystyle \frac{1}{{\mathrm{N}}_{\mathrm{D}}}}{\displaystyle \sum _{\mathrm{i}\in \mathrm{D}}\left({\mathrm{Y}}_{\mathrm{i}}^{\mathrm{T}=1}-{\mathrm{Y}}_{\mathrm{NN}\left(\mathrm{i}\right)}^{\mathrm{T}=0}\right)}$$

(4)

where N_D is the number of treated beekeepers (operating in WEF areas), ${\mathrm{Y}}_{\mathrm{i}}^{\mathrm{T}=1}$ is the observed income of beekeeper iii in the treatment group, while ${\mathrm{Y}}_{\mathrm{NN}\left(\mathrm{i}\right)}^{\mathrm{D}=0}$ is the income of the nearest neighbor matching (NNM) beekeeper in the control group. In this model, for each individual in the treatment group, the individual in the control group whose characteristics are closest is found, and the ATT value is calculated by taking the income difference.

Given the structure of the sample, the non-metric multidimensional scaling (NMDS) with replacement is utilized in this study as opposed to one-to-one matching, which is a more appropriate strategy. The potential discrepancy in the number of beekeepers in the treatment and control groups is a primary factor influencing this decision. Under one-to-one matching, there would be only one control beekeeper paired with each treatment beekeeper. However, it is ineffective to limit the employment of control beekeepers to a single match, as some of them may resemble several treated beekeepers. Furthermore, requiring unique matches could result in suboptimal pairings when the control group is small, as this would lead to matching beekeepers with widely disparate propensity scores, which would increase bias and lower statistical power. The replacement technique overcomes this challenge by enabling the most suitable control beekeepers to be reused as the best matches for treated beekeepers. This strategy enhances estimation accuracy and reduces selection bias, as demonstrated by empirical research [44,45,46]. The replacement technique has been shown to enhance the quality of matches in propensity score matching models. Furthermore, Abadie and Imbens [47] have highlighted that replacement ensures optimal comparisons between treatment and control units, a statistical benefit that is particularly relevant in small control groups. Consequently, in scenarios where the control group is limited in size, replacement-based NNM emerges as a reliable technique for estimating causal effects. The technique has been shown to reduce selection bias, enhance statistical precision, and reduce poor matches by permitting control beekeepers to be matched more than once [47]. This method guarantees a more accurate assessment of the income impact of conducting business in WEF regions, offering important information about the financial ramifications.

2.3. Justifications for Using Explanatory Variables

The present study analyzes the factors influencing site selection for beekeeping activities and examines how these factors impact honey revenue, with a particular focus on the effects of proximity to wind energy farms. The study draws on established theoretical frameworks, prior research, and empirical data in order to evaluate the determinants of location choice in beekeeping and the significance of their interactions with income outcomes. The methodological approach employed is designed to rigorously test the research hypotheses and ensure the validity and applicability of the selected variables.

The current research investigates how various factors influence beekeeping practices and explores their implications for honey income and business decision-making. Utilizing data collected from six provinces within Turkiye’s Aegean Region, the study highlights the presence of regional disparities in the interactions between wind energy farms and beekeeping operations. Furthermore, it explores how generational perspectives shape beekeeping decisions and production processes. The first group, designated the “Baby Boomer” generation, comprises individuals born prior to 1965; the second group, known as Generation X, includes those born between 1965 and 1980; and the third group, referred to as Generation Y, consists of those born after 1980. The findings of the study indicate that the perceived impact of wind turbines on beekeeping varies across different generations. The emphasis placed on the advantages of contemporary technologies and the prioritization of environmental sustainability by Generation Y are noteworthy. Conversely, Baby Boomers characteristically may adopt a circumspect stance towards emerging technologies, motivated by concerns regarding the potential erosion of conventional methods. In contrast, Generation X adopts a more balanced perspective, evaluating both the risks and benefits. The present study has sought to demonstrate the divergent attitudes of different generations towards technology and environmental change. As indicated by international literature, the age of beekeepers is a pivotal factor in shaping responses to climate change and technological advancements [48,49]. It is noteworthy that age differences have also been identified as a factor influencing preferences regarding different types of contractual arrangements [50,51]. As demonstrated in a substantial number of studies, age is a pivotal factor in agricultural decision-making processes that influence honey production strategies [52,53,54,55]. In this context, an examination of the intersection of generational attitudes, environmental perceptions, and approaches to technology provides valuable insights for developing sustainable and forward-looking strategies across both sectors—wind energy farms and beekeeping.

The marital status and educational attainment of beekeepers are considered to have a significant impact on beekeeping techniques. It is evident that those beekeepers with a strong educational background are better equipped to evaluate the advantages and disadvantages of sites located in proximity to wind energy farms. This ability empowers beekeepers to select optimal locations, thereby maximizing honey production and securing higher income. Moreover, research has demonstrated a correlation between marital status and site preferences, with married beekeepers often exhibiting a preference for more productive locations due to familial obligations. It is broadly recognized in the extant literature that both factors are critical in shaping key aspects of beekeeping operations, particularly in relation to honey yield, sensitivity to climate change, and adaptability to new technologies. As demonstrated in the literature [45,46,49,51,52,53,56,57,58,59], there is ample evidence to support the hypothesis. The positive impact of family support on enterprise growth and operational efficiency has been demonstrated. However, it is challenging to ascertain whether higher income among less educated beekeepers is indicative of economic necessity or greater involvement in the activity. Conversely, beekeepers with a higher level of education may engage in beekeeping as a leisure pursuit or a secondary occupation. Despite the unavailability of such detailed labor allocation data in the specified dataset, an attempt was made to address this limitation by incorporating the working proportion variable (i.e., the proportion of household members involved in beekeeping), which may serve as a proxy for labor intensity. Nevertheless, it must be recognized that such a variable does not fully resolve the issue, and the potential for confounding between education and income may persist. This study further explores how beekeeping knowledge and training influence hive placement decisions, financial returns from honey production, and practical variables such as vehicle ownership, which facilitates access to hives and enables regular maintenance and harvesting. This is of particular pertinence in remote areas in proximity to wind energy farms. Research has indicated that beekeepers who have access to quality training or extension services are more capable and willing to enhance productivity and environmental sustainability [45,46,49,52,56,59,60]. Such training and technical capacity building directly contribute to more efficient and sustainable apiculture practices. Additionally, beekeepers’ ability to cope with physical demands and make sound decisions during production may be linked to health indicators such as Body Mass Index (BMI). A lower BMI is generally associated with better health and greater physical endurance in demanding conditions, whereas a higher BMI may reflect potential health issues and reduced work capacity. Beekeepers with elevated BMI may delay essential operational tasks, ultimately limiting honey yield and overall beekeeping income.

The present study also investigates the impact of cooperative memberships and various agricultural registration systems on honey production income. Participation in cooperatives may demonstrate how to enhance operational processes for beekeepers, particularly those operating in proximity to wind energy farms. This objective is realized through the provision of training opportunities, shared resources, and technical support, thereby optimizing overall profitability. Research across various academic disciplines has consistently highlighted the pivotal role of such factors in strengthening the resilience, productivity, and innovation capacity of beekeeping systems [49,51,59,61]. The study goes on to examine financial variables such as government assistance, credit obligations, and agricultural insurance. Governmental support can play a pivotal role in mitigating risks in areas adjacent to wind energy farms, thereby fostering investment in advanced tools and inputs. The extant literature indicates that access to credit exerts a substantial influence on producers, manifesting both in a structural capacity and in terms of psychological impact. This impact fosters greater risk tolerance and openness to the adoption of new practices [49,50,51,59,62]. The present study employs a comprehensive approach by evaluating environmental factors, including the road distance between hives and major roads, as well as individual characteristics such as age and experience. The objective of this study is to achieve a comprehensive understanding of the intricate interrelationships among these factors and their impact on the profitability of honey production. As demonstrated by the findings of numerous studies, including those conducted by Sahle et al. [63] and Tadesse et al. [59], it has been established that these variables are significant in determining the performance of beekeeping operations. This analytical strategy elucidates the interplay between environmental conditions and beekeeping practices, thus offering critical insights into the development of sustainable beekeeping systems. Furthermore, the approach under discussion aligns with broader research in the field of agricultural innovation, which has repeatedly demonstrated a positive correlation between technology adoption, increased risk tolerance, and public support [49,50,57,61,64,65,66]. In conclusion, the capacity of a beekeeping enterprise to adapt to evolving conditions remains a fundamental determinant of its success, irrespective of geographic location. The Results section presents the findings of a meticulous analysis along with descriptive statistics, including the variance inflation factor (VIF).

Lastly, whilst the present study aims to capture the effects of wind energy farms through spatial and socio-economic indicators, it is important to acknowledge two key limitations related to the explanatory variables. Firstly, although environmental factors such as air pressure, vibrations, and noise are conceptually discussed as potential stressors for honeybee activity, objective environmental indicators (e.g., measured noise levels, vibration intensities, pressure variability, or exact turbine proximity) were not included as independent variables in the regression due to the unavailability of high-resolution georeferenced or instrumental data. It is recommended that future research endeavors integrate such measurements, whether through experimental or observational methodologies, with a view to more precisely isolating the biophysical mechanisms through which turbines may influence bee productivity. Secondly, since data collection occurred after the honey harvest period (November–December), the possibility of recall bias cannot be entirely ruled out. Furthermore, given the seasonality inherent in beekeeping practices, it is important to note that the outcomes observed may not fully generalize to other times of the year. While these constraints do not invalidate the current findings, they are recognized as significant considerations that emphasize the necessity for future longitudinal and multi-seasonal research to validate and extend the insights presented here.

3. Results

3.1. Descriptive Statistics and Group Mean Comparisons

As illustrated in

Table 1. Descriptive statistics of both dependent and explanatory variables.

Variable	Definition	Mean (Std. Dev.)	VIF
Dependent variables
Income	Annual income (Turkish Lira, TL) generated from hive production/1000	278.123 (257.228)	-
Explanatory variables:
Dummy variables:
Izmir	1 if the beekeeper has produced honey in Izmir; 0 otherwise	0.167 (0.373)	2.273
Aydin	1 if the beekeeper has produced honey in Aydin; 0 otherwise	0.167 (0.373)	2.149
Mugla	1 if the beekeeper has produced honey in Mugla; 0 otherwise	0.167 (0.373)	2.664
Manisa	1 if the beekeeper has produced honey in Manisa; 0 otherwise	0.167 (0.373)	2.291
Balikesir	1 if the beekeeper has produced honey in Balikesir; 0 otherwise	0.167 (0.373)	1.857
Canakkale	1 if the beekeeper has produced honey in Canakkale; 0 otherwise (reference group)	0.167 (0.373)	-
Age Cohort < 1965	1 if the beekeeper was born prior to 1965; 0 otherwise (reference group)	0.317 (0.466)	-
Age Cohort 1965–1980	1 if the beekeeper was born between 1965 and 1980; 0 otherwise	0.623 (0.485)	1.377
Age Cohort > 1980	1 if the beekeeper was born after 1980; 0 otherwise	0.060 (0.238)	1.596
Elementary school	1 if the beekeeper has no formal education or attained primary education; 0 otherwise (reference group)	0.513 (0.501)	-
Secondary school	1 if the beekeeper is a secondary school graduate; 0 otherwise	0.127 (0.333)	1.273
High school	1 if the beekeeper is a high school graduate; 0 otherwise	0.197 (0.398)	1.401
College	1 if the beekeeper is a university graduate, including master’s and doctoral degrees; 0 otherwise	0.163 (0.370)	1.589
Widow/divorced	1 if the beekeeper is widowed or divorced; 0 otherwise (reference group)		-
Married	1 if the beekeeper is married; 0 otherwise	0.913 (0.282)	2.296
Never Married	1 if the beekeeper has never been married; 0 otherwise	0.047 (0.211)	2.482
Learning by others	1 if beekeeping is learned by other means; 0 otherwise (reference group)	0.240 (0.428)	-
Learning by father	1 if beekeeping was inherited from the father; 0 otherwise	0.490 (0.501)	2.040
Learning by courses	1 if learned through beekeeping courses; 0 otherwise	0.270 (0.445)	1.729
Car	1 if the beekeeper has a vehicle; 0 otherwise	0.830 (0.376)	1.157
Beekeeping registration	1 if the business has a Beekeeping Registration System; 0 otherwise	0.223 (0.417)	1.202
Farmer registration	1 if the business has a Farmer Registration System; 0 otherwise	0.417 (0.494)	1.562
Cooperative member	1 if the operator is a member of a cooperative/union related to beekeeping; 0 otherwise	0.943 (0.232)	1.196
Out of beekeeping job	1 if the operator has an activity other than beekeeping; 0 otherwise	0.777 (0.417)	1.606
Crop and animal production	1 if the business has crop and animal production other than beekeeping; 0 otherwise	0.493 (0.501)	1.827
Nomad	1 if the beekeeper is nomadic; 0 otherwise	0.853 (0.354)	1.415
Insurance	1 if there is bee insurance; 0 otherwise	0.430 (0.496)	1.830
Credit-based debt	1 if agricultural loan; 0 otherwise	0.377 (0.485)	2.380
Other bee production	1 if other bee products are produced alongside honey production; 0 otherwise	0.857 (0.351)	1.208
Credit use	1 if the beekeeper has used agricultural credit during the year; 0 otherwise	0.430 (0.496)	2.775
Livelihood	1 if beekeeping is the main source of livelihood; 0 otherwise	0.480 (0.500)	2.176
Wind energy farm region	1 if the beekeeper is located in the wind energy farms area; 0 otherwise	0.500 (0.501)	-
Continuous variables:
Road distance	Distance of the hive location to the main road (km)	2.910 (2.132)	1.134
Experience proportion	Experience rate of beekeepers by age	0.393 (0.191)	1.546
Working proportion	The ratio of the number of working persons in the family to the total family size	0.452 (0.369)	1.139
Subsidy amount	Amount of state support for beekeeping activities throughout the year (TL/10,000)	6.971 (4.842)	2.192
Body mass index	The beekeeper’s body mass index (weight divided by height squared)	26.710 (3.436)	1.185
Number of observations		300

, the mean values and standard deviations (Std. Dev.) of the target variable (income derived from honey production) and the independent variables identified in the study area are presented. The table also includes the correlation coefficients between income and the continuous independent variables. Furthermore, the Variance Inflation Factor (VIF) scores calculated for the independent variables are displayed in the final column of the table.

Within the Aegean Region, the provinces of Izmir, Aydin, Mugla, Manisa, Balikesir, and Canakkale each contributed an equivalent proportion of participants to the study, accounting for 16.7% of the total sample. The following is a concise demographic profile of the participants: The “Baby Boomer” generation, constituting 31.7% of the beekeepers, was born prior to 1965; the “X” generation, comprising 62.3%, was born between 1965 and 1980; and the “Y” generation, constituting 6%, was born after 1980. According to this distribution, the majority of beekeepers are middle-aged. The data further reveal that 51.3% of the participants have only completed basic school or have no diploma, whereas 12.7% have completed middle school, 19.7% have completed high school, and 16.3% have earned a higher degree. Additionally, 83% of the participants own a car, and 91.3% are married. A significant proportion of the population, 94.3%, is affiliated with beekeeping cooperatives, and 41.7% are registered with the Farmer Registration System (FRS). Furthermore, 49.3% of the participants engage in dual activities, encompassing beekeeping and crop or animal production, while 77.7% of them are employed in diverse occupations beyond beekeeping. Moreover, 35.7% of beekeepers reported having agricultural loan debt, while 85.3% of beekeepers were identified as migratory. Additionally, 85.7% of participants engaged in activities beyond honey production. Despite 48% of participants considering beekeeping their primary source of income, 43% of them opted to insure their bees with agricultural insurance. In regard to the acquisition of beekeeping knowledge, 49% of respondents indicated that they had been taught by their fathers, 27% by classes, and 24% by other sources (see Table 1 for details). Furthermore, the VIF scores were employed to assess multicollinearity among the independent variables. All VIF scores were found to be below the commonly accepted cutoff of 10, indicating that multicollinearity among the covariates was not a substantial concern. The impact of the binary and multi-level categorical factors on the average revenue from honey production prior to propensity score matching is demonstrated in

Table 2. Mean income values for all predictor variables across the entire dataset, along with corresponding statistical tests including correlation coefficients.

Note: *** p < 0.01, ** p < 0.05, * p < 0.10 and ref. refers to the reference group.

,

Table 3. Mean income values for all predictor variables across the WEF region dataset, along with corresponding statistical tests including correlation coefficients.

Note: *** p < 0.01, ** p < 0.05, * p < 0.10 and ref. refers to the reference group.

and

Table 4. Mean income values for all predictor variables across the non-WEF region dataset, along with corresponding statistical tests, including correlation coefficients.

Note: *** p < 0.01, ** p < 0.05, * p < 0.10 and ref. refers to the reference group.

in a preliminary manner. In the current study, when the variable of interest is categorical (e.g., with two or more levels), one level is selected as the reference (e.g., ref.), and the other levels are compared pairwise against this reference using a t-test to assess whether the differences in average income distribution are statistically significant. Conversely, if the variable is continuous, its relationship with honey production income is evaluated by calculating the correlation coefficient along with the corresponding statistical measures. Table 2, Table 3 and Table 4 present a contrast between the average income levels of beekeepers regardless of WEF location, the average income of beekeepers operating in WEF regions, and the average income of beekeepers operating exclusively in non-WEF zones, respectively.

These tables offer fundamental information for the comparison of how various regional factors affect beekeepers’ earnings. All data were analyzed without distinguishing between areas with and without wind energy farms (Table 2) and separately for areas with and without wind energy farms (Table 3 and Table 4, respectively). Statistically significant variables were identified, and the results were assessed accordingly.

The findings indicate that wind energy farms and environmental conditions exert a notable and substantial effect on honey income in the Aegean Region. According to the results of the general data set (Table 2), Balikesir has the lowest yearly honey income, whereas Mugla has the greatest. The mean annual income of beekeepers in Mugla is 181,000 TL higher than that of beekeepers in Canakkale (t-value = 3.177, p < 0.01). Mugla remains the region with the highest annual honey income among wind energy farm locations (Table 3; t-value = 2.192, p < 0.05). The data indicate that beekeepers in Mugla generate an average of 217,000 TL more than their counterparts in Canakkale within these locations. Furthermore, Table 4 reveals that Mugla consistently exhibits the highest income in areas devoid of wind energy farms (t-value = 2.660, p < 0.05), with Izmir ranking second. Within the region, beekeepers in Izmir generate an average of 115,000 TL more per season than those in Canakkale, while those in Mugla produce an average of 145,000 TL more. A noteworthy finding is that beekeepers operating within wind energy farm locations in every province within the Aegean Region consistently earn higher incomes compared to those not situated in such facilities. For instance, a beekeeper in Mugla is estimated to generate an average of 138,000 TL less if their production is conducted within wind energy farm sites as opposed to external locations. This finding underscores the tangible economic impact of wind energy farms on local honey production.

In wind energy farm regions and the overall dataset, beekeepers born after 1980 and those born between 1965 and 1980 exhibit higher average honey incomes compared to those born before 1965. Additionally, young beekeepers from Generation Y demonstrate a marked propensity for generating significantly higher incomes. In the overall dataset, the mean income of younger beekeepers is 187,000 TL higher than that of their older Baby Boomer counterparts (t-value = 2.786, p < 0.01). This discrepancy is particularly pronounced in wind energy farm locations, where the mean income gap reaches up to 279,000 TL (t-value = 2.719, p < 0.01). Also, notably, across all data groupings, beekeepers with a high school diploma or less consistently generate lower incomes compared to those with more advanced education. Specifically, in the overall dataset, the mean income of beekeepers with a university degree was found to be 131,000 TL less than that of those with only a basic school diploma or no formal education (t-value = −3.172, p < 0.01). In locales devoid of wind energy farms, such income disparity amounts to approximately 120,000 TL (t-value = −3.007, p < 0.01). However, it undergoes a substantial increase in regions encompassing wind energy farms, reaching approximately 162,000 TL (t-value = −2.248, p < 0.05).

The complete dataset (t-value = 2.928, p < 0.01), wind energy farm regions (t-value = 2.221, p < 0.05), and non-wind energy farm regions (t-value = 2.040, p < 0.05) all underscore the significance of learning pathways in shaping revenue from honey production. The findings indicate that individuals who received beekeeping training from their fathers consistently exhibit higher earnings compared to those who learned the trade from friends, family, or mentors. Specifically, the analysis reveals that beekeepers who received training from their fathers exhibit an average income that is 74,000 TL higher than those who learned the skills from other sources in areas devoid of wind energy farms. In regions with wind energy farms, the observed income disparity escalates significantly, reaching 137,000 TL. The findings of the analysis suggest that employment in wind energy farm areas confers a significant financial benefit to beekeepers who received training from their fathers, with an average salary that is 108,000 TL higher compared to their counterparts working outside these zones. Also, the results indicated that car ownership significantly influenced income. The findings, based on both the total dataset (t-value = 2.132, p < 0.05) and wind energy farm regions (t-value = 1.772, p < 0.10), suggest that beekeepers who possess vehicles consistently generate higher incomes compared to their counterparts without vehicles. Specifically, the analysis suggests that the addition of a vehicle to a beekeeper’s assets could potentially lead to an increase in income by approximately 84,000 TL. In wind energy farm regions, the increase in income for beekeepers owning vehicles is approximately 84,000 TL, aligning with the findings in the overall dataset.

In non-WEF areas, beekeepers with registered farms exhibit an average income that is 48,000 TL higher than their unregistered counterparts (t-value = 1.717, p < 0.10). However, in the WEF region, the discrepancy in income escalates to 156,000 TL (t-value = 2.949, p < 0.01). Furthermore, beekeepers who possess a registered farm and operate within a wind energy farm settlement experience an additional income increase of 152,000 TL. Cooperative membership has also been demonstrated to have a positive impact on income levels. In non-WEF regions, beekeepers who are members of cooperatives earn an average of 113,000 TL more than their non-member counterparts (t-value = 1.916, p < 0.10). Furthermore, a statistically significant income disparity among beekeepers engaged in both crop and livestock farming was identified across the entire sample (t = 3.161, p < 0.01), with notable regional variation. In non-WEF areas, the difference is marginally significant (t = 1.848, p < 0.10), whereas in WEF regions, it becomes substantially more pronounced (t = 2.755, p < 0.01). The mean income disparity in non-WEF regions is estimated to be approximately 52,000 TL; however, this figure is approximately threefold in WEF zones, reaching around 139,000 TL. Moreover, beekeepers engaged in both agricultural and livestock activities within WEF regions receive, on average, 122,000 TL more than their counterparts in non-WEF areas. Furthermore, a marked income disparity between nomadic and non-nomadic beekeepers is evident in non-WEF areas (t = 3.989, p < 0.01), and this difference becomes even more significant when considering the entire dataset (t = 5.080, p < 0.01) as well as within WEF regions (t = 3.705, p < 0.01). On average, migratory beekeepers across the entire sample earn 205,000 TL more than their stationary counterparts, which represents almost twice the income of non-nomadic producers. Within the geographical confines of WEF regions, this disparity in income amplifies to approximately 255,000 TL, signifying that nomadic beekeepers in these regions accrue earnings that are nearly 2.5 times higher than those of stationary beekeepers. In zones not designated as WEF, the income advantage for nomadic beekeepers remains substantial, averaging around 156,000 TL. Furthermore, nomadic apiculture practitioners operating within wind energy farm areas accrue approximately 100,000 TL more than their migratory counterparts outside these zones.

The findings of the current study demonstrate that beekeepers who possess insurance policies specifically tailored to the agricultural sector earn a significantly higher income from honey sales when compared with those who do not possess such insurance (t = 3.030, p < 0.01). This elevated income is evident in both areas that are proximate to wind energy farms (t = 2.589, p < 0.05) and those that are not (t = 1.763, p < 0.10). On average, insured beekeepers generate approximately 90,000 TL more, with this disparity rising to 133,000 TL in regions with wind energy farms. This finding suggests that insurance enhances financial stability, empowering beekeepers to make strategic investments and augment production. In the overall dataset (t = 3.003, p < 0.01), those in debt earn approximately 91,000 TL more, while in wind energy farm areas (t = 2.729, p < 0.01), this amount rises to an average of 144,000 TL. Furthermore, farmers who have taken out loans specifically for beekeeping show significantly higher honey revenues compared to those who have not, both in the overall dataset (t = 2.664, p < 0.01) and in WEF regions (t = 2.257, p < 0.05). On average, farmers who received loans earn 79,000 TL more, with those in WEF areas gaining an additional 116,000 TL. However, the income difference in non-WEF regions (45,000 TL) was found to be statistically insignificant. Also, beekeepers who regard beekeeping as their primary source of livelihood generate considerably higher honey revenues in comparison to those who perceive it as a secondary income. This phenomenon is consistent across all data points (t-value = 8.173, p < 0.01), in WEF areas (t-value = 6.292, p < 0.01), and in non-WEF areas (t-value = 5.579, p < 0.01). The income disparity is particularly pronounced, with beekeepers in non-WEF regions earning approximately 145,000 TL more than part-time beekeepers, while those in WEF regions experience an even more substantial increase of around 289,000 TL. Beekeepers in WEF areas earn almost twice the income of their counterparts in non-WEF regions. Also, as demonstrated in Table 2, the mean income of a randomly selected beekeeper residing within a WEF region is 314,970 TL, in comparison to 241,280 TL for those located in non-WEF areas, thus demonstrating a substantial discrepancy in earnings. This income disparity is found to be statistically significant (t = 2.503, p < 0.05), indicating that beekeepers in WEF zones, on average, earn approximately 74,000 TL more than their counterparts outside these areas.

In analyzing the correlation between honey income and continuous variables, a positive and statistically significant correlation is identified between beekeepers’ years of experience and their earnings. This correlation is evident across all data points (correlation coefficient (ρ) = 0.314, p < 0.01), in WEF regions (ρ = 0.308, p < 0.01), and in non-WEF regions (ρ = 0.375, p < 0.01). A significant and positive correlation has been identified between the subsidies received and the income from honey production, with this relationship being consistent across all data points (ρ = 0.746, p < 0.01). This correlation is evident in WEF areas (ρ = 0.835, p < 0.01) and in non-WEF areas (ρ = 0.595, p < 0.01).

In

Table 5. Maximum likelihood estimates of logistic regression for propensity score matching.

Note: *** p < 0.01, ** p < 0.05, * p < 0.10.

, the logit model is employed as an intermediary instrument within the propensity score matching method, although this is not the primary objective.

The effects of various variables remain relevant; therefore, the focus will be on statistically significant variables, and the subsequent discussion will be based on their marginal effects. The results indicate that beekeepers’ marital status significantly influences their likelihood of operating in WEF regions. Compared to widowed or divorced individuals, married and never-married beekeepers exhibit a higher probability of choosing these areas. Specifically, married beekeepers are 34% more likely to prefer WEF regions, while the likelihood for never-married beekeepers stands at approximately 28%. In contrast, registered beekeepers, defined as those officially listed in the farmer registration system, exhibit an 18% lower likelihood of operating in wind energy farm regions compared to their unregistered counterparts. A 4.5% decline in the probability of selecting WEF regions is observed when hives are situated at a greater distance from primary roadways. It is also noteworthy that for every additional TL received in beekeeping subsidies, the likelihood of selecting a WEF region increases by 1.6%.

3.2. Some Sensitivity and Robustness Checks

Prior to presenting our results for ATE, ATT, and average treatment on untreated (ATC), we performed a series of sensitivity analyses to evaluate the robustness of our findings and the potential impact of unobserved confounders. Conducting a sensitivity analysis is essential for assessing the stability of our estimates and understanding how hidden variables might influence the outcome. In the context of Propensity Score Matching, various sensitivity tests can be utilized to analyze the effects of unobservable factors. A notable method involves adjusting the calibration value, which determines the maximum acceptable distance for matching. To assess how variations in the calibration affect matching quality and the reliability of our results, we systematically altered the calibration value from 0.01 to 0.50 in increments of 0.01. Through this iterative approach, we determined that 0.05 was the optimal calibration value. This finding indicates that the maximum permissible difference in propensity scores between matched individuals should not exceed 0.05 standard deviation units, thereby enhancing the quality of the matching process. With this refined calibration value, we analyzed enhancements in the distribution of covariates between the treatment and control groups. The findings confirm that the matching procedure effectively balanced the groups, with the initial difference in propensity scores (0.8090) being significantly reduced to 0.0008 after matching, indicating a marked improvement in the comparability of treated and control observations. Additionally, t-tests were conducted to check for any remaining systematic differences between the treatment and control groups after matching. The results indicated no statistically significant differences, further validating the reliability of the matching method. In conclusion, these results suggest that the matching procedure successfully mitigated selection bias, ensuring that the estimated treatment effects are both credible and robust.

Meanwhile, matching methods are commonly employed in observational studies to estimate causal effects and ensure the distribution of observed covariates is balanced between treatment and control groups. However, achieving this balance does not guarantee that the model accurately reflects the true causal relationship. This is where placebo tests become essential, as they provide a robust means of assessing the internal validity of the matching process. A placebo test evaluates the effect of a pseudo-treatment on an outcome variable that should theoretically remain unaffected by the treatment [67,68]. Examples include pre-treatment measures, fixed individual characteristics, and variables outside the causal pathway. In this study, we created a placebo income variable by randomly rearranging the original income data to ensure that any observed effect would indicate model misspecification rather than a genuine causal effect. The empirical findings suggest that the treatment variable (wind energy farm region) has a statistically significant positive impact on income. The ATE is estimated at 45,107 TL, which is significant at the 5% level (p = 0.040). The ATT is even higher, at 56,515 TL, and is marginally significant at the 10% level (p = 0.064). By contrast, the ATC, estimated at 33,700 TL, is not statistically significant (p = 0.285). In comparison, the placebo test results for the placebo income variable serve as a valuable benchmark. The estimated placebo effects for the ATE, ATT, and ATC are 21,000 TL, 32,750 TL, and 9250 TL, respectively. However, none of these are statistically significant (p = 0.306, 0.265, and 0.747, respectively). This lack of significant placebo effects contrasts sharply with the original income findings, indicating that the model is unlikely to produce false-positive results. In essence, the placebo test bolsters the credibility and robustness of our causal claims by confirming that the substantial observed effects of the treatment on actual income are not due to model misspecification or inadequate matching. Consequently, the internal validity of the matching approach is affirmed, thereby increasing confidence in the study’s causal conclusions. Using the untreated group as the sole reference, a fictitious treatment variable was also created as an alternative validation method through a placebo test. The results provide additional evidence supporting the reliability of the previous findings. Detailed results can be obtained from the authors upon request.

Meanwhile, the decision to select an equal number of beekeepers from each province was influenced by practical constraints, including data accessibility and the challenges of fieldwork. Chief among these constraints is the prevalence of migratory beekeeping and the absence of a comprehensive registration system that records migratory beekeepers by province. However, given the non-uniform distribution and varying density of beekeepers across regions, this equal sampling strategy may pose a logical limitation and potentially restrict the generalizability of the findings. To address these potential biases and assess the reliability of our treatment effect estimates, we implemented a bootstrap procedure based on a propensity score model. Within the specified framework, the sampling process was divided into two subgroups: 200 observations were drawn from wind energy farm areas and 400 from non-wind energy farm areas using a resampling technique, resulting in a total of 600 observations per iteration. Across 2000 bootstrap replications, ATE, ATT, and ATC were computed, and 95% confidence intervals were constructed based on the resulting distributions (27,626.67, 132179.33) for ATE, (20,268.75, 113,124.62) for ATT, and (22,386.38, 146,885.62) for ATC. The consistent inclusion of the original ATE, ATT, and ATC point estimates within these confidence intervals indicates that the equal sampling approach across provinces did not meaningfully distort the model outcomes, thereby supporting the validity of our inferences.

Finally, it is imperative to demonstrate that the covariate balance has been achieved post-matching to evaluate the reliability of causal inferences. In this context, the present study also presents balance estimates and distribution-related findings regarding the PSM method. This approach is employed to estimate the average treatment effect (ATE), as demonstrated in Figure 1 and outlined in

Table 6. Standardized mean differences (SMD) before and after propensity score matching.

. Figure 1 compares the propensity score distributions of the treatment group (regions with wind energy farms) and the control group (regions without wind energy farms). The treatment group shows a sharper density concentration, particularly between 0.4 and 0.6, while the control group shows a broader, flatter distribution with a low-density tail. Though these distributions differ in shape, substantial overlap between the two groups confirms the validity of the common support assumption, affirming that the matching procedure was appropriately applied to robustly estimate the ATE. Table 6 further supports these findings by reporting standardized mean differences (SMD) before and after matching. Prior to matching, notable imbalances were observed across several covariates. However, after matching, these differences were significantly reduced, with most falling below the commonly accepted threshold of 0.1. At the same time, the distance (e.g., PSM distance) indicates the SMD of the propensity score variable used in the PSM analysis, both before and after matching. The value exceeds the threshold of 0.10 before matching and decreases to 0.001 after matching. This reflects stronger alignment between the groups. Similarly, the SMD for the variable “Credit use” decreased from −0.0200 to 0.0072. The presence of negative and near-zero SMD values further reinforces that covariate balance has been successfully achieved between the treatment and control groups. In summary, the observed discrepancies in density should not be construed as problematic; the primary concern in propensity score matching is not the absolute shape of the distributions, but rather the degree of overlap ensuring comparability. The visual evidence from the figure clearly demonstrates this overlap, validating the use of ATE estimation on strong statistical grounds.

3.3. Propensity Score Matching Results

After a detailed discussion of sensitivity test analyses, ATE, ATT, and ATC were estimated at 45,107 TL, 56,515 TL, and 33,700 TL, respectively (e.g., Table 5). All treatment effect estimates—ATE and ATT—are statistically significant at the 5% and 10% levels, respectively, with the exception of ATC, which is not statistically significant. These results indicate that a randomly selected beekeeper operating within a wind energy farm area is expected to earn, on average, 45,107 TL more than they would have earned if operating outside such an area. Additionally, beekeepers already producing in wind energy farm zones earn about 56,515 TL more than they would otherwise. Conversely, if a beekeeper relocates from a wind energy farm area to a different location, their honey production income is expected to decrease by approximately 33,700 TL. These results imply that the treatment significantly increases the income of those receiving it, although evidence regarding its effects on the control group (e.g., ATC) remains inconclusive.

4. Discussion

A statistical analysis elicits substantial disparities in honey income among beekeepers of varying ages. The observed income disparity among beekeepers can be attributed to a combination of interrelated factors, where younger apiculture practitioners exhibit a higher degree of adaptability to market fluctuations, largely due to their proficiency with advanced production methods and modern technological tools. Moreover, the comparatively higher physical capacity of the subjects may be a contributing factor to the increased productivity observed. Conversely, advancing age is frequently accompanied by a decline in physical vigor, which can impede operational efficiency and result in reduced honey yields and income. This assertion is corroborated by the findings of previous studies conducted by Malkamäki et al. [54] and Aksoy et al. [69], both of which reported a negative correlation between age and honey production. The findings under discussion highlight the economic advantage associated with youth, particularly in terms of enhanced production capacity and a preference for modern techniques. Furthermore, empirical evidence demonstrates that young beekeepers operating within wind energy farm zones accrue approximately 305,000 TL more than their counterparts working outside these zones. This observation underscores the production-related benefits linked to agility and emphasizes the potential economic gains derived from involvement in renewable energy areas. This provides further empirical evidence to support the extant literature on the demographic and environmental influences on agricultural productivity.

The findings reveal statistically significant discrepancies in honey production income inequalities amongst beekeepers with varying educational backgrounds, both within the whole dataset and in areas with and without wind energy farms. These results indicate that educational attainment exerts a substantial influence on income from honey production. This discrepancy can be attributed to the fact that beekeepers with limited educational attainment often perceive beekeeping as their primary source of income, thereby allocating more time and effort to honey production. In contrast, highly educated individuals typically regard beekeeping as a recreational pursuit or ancillary activity, allocating less time due to competing professional and personal commitments. This phenomenon is further reinforced by the high opportunity cost of time for those with advanced education. Furthermore, beekeepers with limited education tend to rely on a substantial foundation of practical experience and traditional knowledge accumulated over time, which might help elucidate their financial advantage. Additionally, many members of this group perceive beekeeping as a way of life rather than merely a business, which could potentially enhance production and efficiency. The findings of the current study demonstrate the impact of educational attainment on not only individual production tactics but also the overall sustainability of the economy and the disparity in income within the honey industry. The significant disparities in income according to education underscore the necessity for policy pertaining to the beekeeping industry to prioritize education and capacity-building initiatives. Further research on the social and economic aspects of beekeeping may provide valuable insights for the future of the sector. At the 5% level of significance, Adeokun et al. [70] found that education had a significant impact on income, with 53% of beekeepers having attained a higher education degree. However, other research has found that education level has little bearing on beekeeping revenue [57,58]. However, a notable exception emerges when examining beekeepers with university degrees operating within wind energy farm zones, who exhibit a mean income of 68,000 TL, surpassing their counterparts in non-wind energy farm regions [57,58]. This observation suggests that highly educated producers may possess a distinct advantage in capitalizing on the enhanced environmental, technical, and production resources available in wind energy farm zones.

A statistically significant discrepancy was identified, favoring beekeepers who apprenticed under their fathers with regard to the financial implications of honey production. This variation was attributed to the different methods of training and knowledge transfer employed by each group. This observation suggests that the integration of traditional knowledge and expertise with modern production practices can enhance efficiency, further exacerbating the observed income disparity. The environmental advantages associated with wind energy farm areas appear to be a contributing factor in this discrepancy. The findings underscore the pivotal role of cumulative knowledge and experience, complemented by the ready availability of family labor and well-established market networks, in shaping the higher profits of father-trained beekeepers. A number of studies have highlighted the potential of formal beekeeping training and education to enhance income and production [60,69,71]. However, other studies have found no clear relationship between beekeeping education and income [57,58].

The association between honey production income and other binary variables, such as car ownership, farmer registration, cooperative membership, additional livestock or agricultural production, a nomadic lifestyle, agricultural insurance, agricultural credit debt, loan usage, primary source of livelihood, and presence in wind energy farm regions, was further examined. Aksoy et al. [69] observed that beekeepers face challenges in fully leveraging available incentives due to constraints in transportation resources, such as trucks. These limitations hinder the full realization of available incentives. The analysis indicates that vehicle ownership plays a pivotal role in augmenting revenue, particularly in rural or isolated regions with constrained transportation infrastructure. The study highlights a pronounced advantage for beekeepers owning vehicles, particularly in wind energy farm locations, where those who drive their own vehicles report a median income 85,000 TL higher than those without vehicles. This underscores the significance of flexible and mobile transportation in enhancing beekeeping revenue, particularly in areas with limited or inadequate transportation infrastructure.

The findings of this study demonstrate a statistically significant relationship between the farmer registration system and the income derived from honey production. This association has been observed in both WEF regions and non-WEF locations, as well as in the overall dataset. The advantages of the farm registration system and cooperative membership, including enhanced market access, eligibility for government assistance programs, access to advanced information and technology, branding opportunities, shared sales channels, and reduced financial risks, likely contribute to increased income. A study by Tarekegn et al. [72] pointed out that households that participated in cooperatives increased the volume of honey they marketed by 58.4% compared to those that were not members. In a similar vein, Abera et al. [73] found that cooperative participation increases access to technical assistance, inputs, market intelligence, and credit options, all of which improve honey supply to markets. These findings suggest a positive relationship between cooperative membership and honey income. However, the extant literature contains contradictory views. Alemu et al. [46] and Minja and Nkumilwa [74] contend that there is no statistically significant correlation between household income and honey cooperative involvement. These conflicting results highlight the need for more investigation into the precise circumstances under which the farmer registration system and cooperative participation result in financial gains for the beekeeping industry. Moreover, contracts provide financial predictability and secure market access [45,46,59], protecting producers from price fluctuations and demand changes, and cooperatives facilitate economies of scale, enhancing access to better resources, training, and distribution channels. The enhancement of beekeepers’ incomes as well as the strengthening of the beekeeping industry’s resilience to climate change and environmental challenges are two key benefits of these mechanisms.

Beekeepers who engage in crop and livestock farming generate a substantially higher income from honey production in comparison to those who concentrate solely on beekeeping. This indicates that engaging in agricultural and livestock activities not only provides an additional income stream but also improves resource efficiency, aligns with the flowering periods, and fulfills the nutritional needs of bees, thereby enhancing honey production. This assertion is further substantiated by the findings of Vaziritabar and Esmaeilzade [75], who observed that 77% of farmers regard honey production as a means of diversifying and optimizing their income. In a similar vein, Wagner et al. [58] discovered that access to alternative income sources from farming has a positive effect on honey yield, which in turn contributes to overall household income. Such a finding underscores the economic viability of beekeeping in areas endowed with renewable energy infrastructure, signifying its promise for enhanced profitability. Moreover, the findings of this study demonstrate that nomadic beekeepers consistently generate higher revenue from honey than stationary beekeepers. These results suggest that the capacity to migrate confers considerable benefits, including access to nectar-rich areas, flexibility in production scheduling, and rapid adaptation to ecological changes, thereby significantly enhancing honey revenues. Beekeepers are progressively adopting more extensive migratory practices to augment production and income [76]. This phenomenon is further evidenced by Uzundumlu et al. [77], who found that migratory beekeeping practices positively impact honey yield in Turkiye. Meanwhile, a national study revealed that 71% of Turkish beekeepers engage in migratory beekeeping, with 55% doing so to increase production volume, while the remaining 45% migrate due to short regional flowering periods and the need for greater production diversity [78]. Furthermore, a notable financial disparity is observed between nomadic and migratory beekeepers within wind energy farm zones, with the former earning substantially higher incomes. This observation underscores the considerable economic benefits of mobile apiculture, particularly in regions characterized by the presence of renewable energy infrastructure.

It is notable that the income increase associated with loans is more significant in WEF regions, emphasizing the economic potential of financial support in these areas. The findings underscore the pivotal role of financial investments in augmenting beekeeping income. However, some studies suggest that access to credit may have a negative impact on beekeeping revenues [71,79], primarily due to high interest rates increasing production costs and ultimately lowering output [71]. This underscores the imperative for a comprehensive evaluation of the accessibility and efficacy of financial support services for beekeepers [66]. Moreover, financial assistance should extend beyond mere loans, encompassing capacity-building initiatives to enhance beekeepers’ knowledge and skills. The promotion of sustainable beekeeping is contingent on the implementation of practical, results-oriented training programs incorporating mentorship, which are deemed essential for long-term development in the sector [66]. Consequently, the integration of financial and educational support constitutes a pivotal strategy for enhancing the productivity and sustainability of the beekeeping industry in the country.

Apiculture-focused beekeepers consistently generate significantly higher honey revenues compared to those for whom it constitutes a supplementary activity. This discrepancy is probably due to the increased dedication of full-time beekeepers, as reflected in greater time allocation and financial investment aimed at optimizing operational efficiency, enhancing product quality, and accessing more lucrative market opportunities. Consequently, engaging in beekeeping as a primary profession has been demonstrated to enhance income potential and contribute significantly to the improvement of overall living conditions [80]. The results suggest that greater beekeeping experience directly correlates with higher income levels. This finding is consistent with the conclusions of earlier studies that stressed the pivotal role of experience in achieving economic success in beekeeping [70]. As beekeepers accumulate knowledge and develop skills over time, they adopt advanced production methods, invest in superior equipment and technology, establish robust customer relationships, gain deeper market insights, make more strategic business decisions, and produce higher-quality honey. These practices ultimately result in increased financial returns. The findings of this study indicate that subsidies also play a pivotal role in augmenting beekeepers’ income. A notable correlation was observed in WEF regions, suggesting that beekeepers in these areas may derive greater benefit from financial support or that subsidies exert a more substantial influence on their earnings. While subsidies also exert a positive impact in non-WEF regions, the effect appears to be less pronounced, thereby underscoring the significance of financial assistance in enhancing beekeeper income. The provision of subsidies enables beekeepers to invest in essential equipment, access training and advisory services to enhance their skills, receive assistance in disease and pest management for healthier bee colonies, and obtain quality control and marketing support to sell their products more effectively and at higher prices. Consequently, these elements collectively contribute to increased honey revenue. Therefore, the expansion of subsidy programs and the enhancement of state support are imperative for the enhancement of beekeepers’ earnings [81].

Conversely, the findings derived from the logistic regression model—utilized as the primary instrument in the propensity score matching approach—underscore the substantial influence of demographic, institutional, and spatial variables on beekeepers’ locational preferences, particularly within wind energy farm regions. The findings have identified marital status as a significant factor, with married and never-married beekeepers demonstrating a stronger propensity to operate in WEF zones. This trend may be indicative of the benefits associated with shared household labor and the pursuit of greater economic stability, both of which can support more intensive honey production in biodiverse WEF landscapes in the Aegean region. On the other hand, the lower probability of registered beekeepers opting for WEF areas may be attributable to regulatory constraints, their predilection for more conventional farming zones, or the misalignment between wind energy farm geography and officially designated farmland. This observation underscores a potential discord between established agricultural frameworks and the evolving spatial dynamics characteristic of renewable energy zones. Moreover, research has demonstrated that accessibility constitutes a critical factor in this regard. The negative correlation between distance from major roads and the probability of operating in WEF regions indicates that logistical factors, including transport ease and proximity to markets or urban centers, play a pivotal role in site selection. This finding is at odds with those of Wagner et al. [58], who reported no significant relationship between apiary proximity to roads and beekeeping income. This finding suggests that context-specific factors may influence the perceived importance of accessibility in different regions. It is evident that the positive effect of subsidies on WEF area selection implies that financial incentives can help offset the perceived risks or costs associated with operating in these zones [69]. The hypothesis is that such support may enable beekeepers to invest in superior equipment and more advanced production methods.

When accounting for different treatment effect estimators, the analysis reveals that beekeepers operating within WEF regions tend to earn substantially higher incomes compared to the hypothetical scenario in which they had conducted their activities elsewhere. This income advantage is largely attributable to the distinctive ecological features commonly found in WEF zones. These areas are often situated at higher altitudes, where diverse floral communities and abundant nectar sources prevail. Such ecological richness not only supports extensive foraging opportunities but also enhances pollination, aided by consistent wind patterns that facilitate pollen dispersal and plant reproduction. In addition to exhibiting high levels of floral abundance, WEF regions characteristically demonstrate reduced levels of agricultural chemical utilization, thereby minimizing contamination risks and reducing stress on bee colonies. Favorable microclimatic conditions, such as moderate temperatures and humidity, further contribute to improved colony health and honey yield. Collectively, these environmental characteristics create optimal conditions for productive and sustainable apiculture. Contrary to the hypothesis that renewable energy infrastructure might disrupt beekeeping activities, the findings of this study suggest that WEF can, in fact, foster ecological and economic resilience. The hypothesis that wind energy farms have a positive impact on the sustainability of apiculture is one that merits further investigation. Rather than simply coexisting with apiculture, it is suggested that wind energy farms may have a beneficial effect on the sustainability of beekeeping. The synergy between clean energy generation and biodiversity-based agricultural practices demonstrates how environmental sustainability and economic development can be mutually reinforcing. In the context of strategic location, wind energy farms have the potential to function as ecological havens, thereby supporting pollinator populations, particularly in high-altitude regions. These findings underscore the potential for the development of renewable energy to be synchronized with the conservation of biodiversity and the promotion of rural economic growth. In the context of global challenges, including climate change and pollinator decline, this unanticipated collaboration presents a viable paradigm for integrated sustainable development. Consequently, spatial planning of wind energy infrastructure should consider regions with high ecological value, particularly those conducive to pollination services, in order to maximize both environmental and economic benefits. By aligning renewable energy initiatives with the preservation of biodiversity and the promotion of sustainable agriculture, policymakers can foster a model of rural development in which clean energy supports not only climate goals but also food security and community livelihoods. Through careful but multifaceted planning and adaptive land-use strategies [33,34], wind energy and apiculture can be made to coexist, thereby fostering a symbiotic relationship that contributes to a greener, more resilient future in which both nature and industry can thrive in harmony. This dynamic challenges the conventional belief that industrial development must come at the cost of agriculture and biodiversity, instead offering a model where the growth of green energy and rural livelihoods can thrive in parallel.

From a policy standpoint, the implementation and expansion of cooperative and contract-based frameworks within WEF regions should be regarded as a strategic measure to promote rural development and sustainable agricultural practices. The integration of small-, medium-, and large-scale apiculture into more structured and inclusive value chains has the potential to cultivate a robust and equitable agri-food economy. It is imperative that policymakers recognize the pivotal role they play in facilitating the collective marketing of honey and secondary bee products, improving market accessibility, and delivering technical support through local cooperatives or public–private partnerships. As the critical importance of pollination for global food security is increasingly recognized, there is an urgent need for targeted investments in institutional infrastructure. It is recommended that investments in this domain encompass the establishment of regional beekeeping centers, the introduction of mobile hive support units, and the implementation of certification programs that promote environmentally sustainable honey production. In view of the ecological diversity that is characteristic of WEF regions, it is possible that these endeavors may be further augmented by integrating complementary activities such as berry cultivation, thereby generating additional economic value. Furthermore, the provision of incentives for apiculture in these regions, including subsidies, land-use advantages, and knowledge-sharing programs, has the potential to unlock the dual potential of WEF zones to contribute simultaneously to clean energy production and sustainable food systems. It is imperative to reconceptualize beekeeping in these contexts. This should not be regarded as an isolated economic endeavor but rather as a fundamental component of a comprehensive, sustainable development framework. The implementation of such integrated systems has been demonstrated to have the capacity to effect transformation of rural economies, to provide support for the conservation of biodiversity, and to demonstrate the compatibility and mutual reinforcement of the systems in question.

5. Conclusions

Wind energy is quickly becoming one of the most popular renewable energy sources, attracting significant investment due to its sustainability and environmental benefits. However, when built without proper planning, wind energy farms can disrupt ecosystems, affecting both humans and wildlife. Pollinators such as bees are particularly vulnerable to turbulence, electromagnetic fields, stray voltage, and turbine noise, which can alter their behavior, reduce their foraging efficiency, and ultimately affect honey production. Further complicating the delicate balance of pollination and biodiversity are variations in wind patterns, which can affect the availability of nectar and pollen. Despite these challenges, this study highlights that wind energy farms and agriculture—specifically beekeeping—can coexist harmoniously within a sustainable development framework. With strategic land-use planning, wind energy farms can be integrated into agricultural landscapes in ways that support ecological health and rural livelihoods. Mapping suitable locations, enhancing floral diversity, and designating wind energy farms as pollinator-friendly zones are practical steps toward this goal. To maximize this synergy, beekeepers must be empowered through targeted training, technical assistance, and adaptation strategies suited to wind energy farm environments. Financial incentives, such as grants and subsidies, can strengthen their economic resilience. Furthermore, branding honey from these areas as eco-friendly and sustainable could open up new markets and increase profitability. Ultimately, advancing this model requires continued ecological research on pollinator–wind interactions and a commitment to inclusive, multidisciplinary policymaking. Aligning renewable energy development with agricultural sustainability can transform potential conflicts into opportunities, fostering a balanced approach to sustainable development that benefits both nature and society.

The decision to sample an equal number of beekeepers from each province in this study was driven primarily by pragmatic considerations, such as limited data availability and logistical challenges encountered during fieldwork. Although this uniform sampling strategy improved cross-provincial comparability, it may have introduced sampling bias, which could compromise the results’ external validity and generalizability. This is a particular concern given the substantial heterogeneity in beekeeper density and spatial distribution across regions. Although our treatment effect estimates are robust, as supported by extensive bootstrap analyses, the non-stratified sampling design may have led to the underrepresentation or over-representation of certain beekeeper populations. This limitation underscores the need for caution when extrapolating the findings to the broader population. Future studies could benefit from integrating detailed administrative or cadastral data, such as that maintained by the Ministry of Agriculture or regional land registries, to enable genuinely stratified or proportionate sampling frameworks. In the absence of such data, institutional efforts toward establishing a comprehensive data infrastructure are critical. One example would be requiring migratory beekeepers to report their seasonal locations. These improvements would facilitate more statistically representative sampling strategies and allow for the application of advanced causal inference methods. These improvements would strengthen the internal validity and generalizability of future research outcomes. To reduce the likelihood of selection bias in observational data, this study also employed PSM. PSM cannot address unseen confounders such as the quality of abandoned habitats or beekeeping motivation, nor can it completely avoid the possibility of reverse causality, despite successfully balancing observable factors. Therefore, caution should be exercised when interpreting the data in a causal manner. The inability to use more rigorous causal inference techniques, such as difference-in-differences (DiD) or instrumental variables (IV), is a significant methodological disadvantage. However, PSM was the most practical approach given the limitations of the available data. This study clearly acknowledges this limitation, underscoring the importance of future investigations using stronger empirical designs to strengthen causal findings. Although we were unable to conduct additional sensitivity analyses using different thresholds (such as 15 or 20 km) due to limitations in spatial resolution and fieldwork, we acknowledge this as a methodological limitation. Looking ahead, we emphasize the importance of future research adopting higher-resolution geospatial classification and treatment effects frameworks that focus on biologically relevant distance intervals, such as under 2 km, 4 km, and 7 km. In this context, using parametric treatment effect models in conjunction with ordered probit sample selection could significantly enhance the reliability of causal inference and represent a notable methodological advancement for future studies. Furthermore, subsequent research endeavors could benefit from the establishment of control groups composed of beekeepers situated beyond wind energy farm zones yet operating in ecologically and topographically analogous regions. Such an approach has the potential to minimize contextual confounding and facilitate a more precise estimation of the causal impact of WEF proximity. It is contended that these methodological advancements are not merely technical adjustments; rather, they are fundamental to generating spatially nuanced, policy-relevant insights in environmentally complex and socioeconomically diverse rural contexts.