Understanding Why Farmers Leave: Validating Key Indicators for Farm Exit in İzmir, Türkiye

Abstract

This study analyses the factors that affect farmers’ decisions to quit farming, using the İzmir region of Türkiye as a case study. A sample of 195 farmers was surveyed to identify the factors influencing this decision. Factors such as the farm type, farm size, agricultural subsidies, land ownership, and age of farmers were found to significantly affect the decisions of the farmers. Sensitivity analysis was performed to examine the changes in the selected variables. The sensitivity analysis showed that young farmers, under financial pressure and tenancy uncertainty, are likely to stay in farming when they receive agricultural subsidies. Thus, targeted financial subsidies can help sustain the livelihood of young farmers efficiently. This study reveals the important relationship between policy interventions and the long-term economic viability of farming as a livelihood. This relationship is crucial for the overall viability of agriculture and the well-being of rural communities. Overall, the findings from this study will inform the development of policies to strengthen the resilience and sustainability of rural farming communities in Türkiye.

1. Introduction

Sustainable agriculture is characterized by the integration of three primary objectives: a healthy environment, financial success, and social and economic equity [1]. However, sustainability is threatened by a variety of serious problems. These include pollution, climate change, biodiversity loss, land degradation, poverty, rising production costs, farm loss, and rural depopulation [2,3]. Rural depopulation and an aging farmer population pose a significant threat to the sustainability of agriculture [4,5]. The declining interest in agriculture among rural youth compared to their urban counterparts poses a significant challenge to securing the future of food production [6,7,8,9,10,11,12,13]. This raises concerns regarding the future of agriculture, potential implications for rural areas, and food security.

Due to various factors, some farmers choose to abandon agriculture, as they are discouraged from the lack of a promising future. The primary drivers of this problem are technological advancements and economic concerns [14,15]. The decline in demand for agricultural labor can be attributed to the increase in labor productivity resulting from technological advances, the availability of alternative employment opportunities, and the growing trend of rural-to-urban migration. Technological advances provide a competitive advantage to large farms due to their easy access to information, financial resources, and managerial expertise [16].

Farmers who are unable to sustain their agricultural activities due to loan payment difficulties are among those who abandon farming. This indebtedness may have arisen due to expansion, inefficiencies in operations, or other factors. Agricultural enterprises may become unable to sustain their operations due to rising costs and debt burdens [17]. The declining revenue is an important factor that contributes to the exit of farmers. The volatility of input and output prices coupled with fluctuations in yields has resulted in significant unpredictability in agricultural revenue in recent years [18,19]. Volatility holds significant importance for small-scale farms, as they usually generate limited income and even minor fluctuations can cause a noticeable impact [20]. These issues are not unique to any one region and pose significant challenges to agricultural sustainability worldwide.

Globally, the agricultural sector faces declining interest among younger generations, who are increasingly migrating to urban areas in search of better opportunities. This trend is exacerbated by technological advancements that reduce the demand for agricultural labor and favor large-scale operations with better access to resources. Studies from various countries, including the USA, the EU, and several Asian countries, have consistently shown that economic pressures, technological changes, and demographic shifts significantly influence farmers’ decisions to continue or exit farming [14,21,22,23,24,25].

Türkiye encounters similar challenges related to the dynamics of its rural population. The rural population has shown a significant downward trend in recent years. Furthermore, it is noteworthy to mention that the demographic composition of the rural population in Türkiye has undergone a significant transformation over the years. In the past, most of the population residing in rural areas comprised predominantly young individuals. However, this situation has changed. The aging of the population in rural areas in Türkiye is mostly caused by rural urban migration, limited work opportunities resulting in unemployment, and insufficient income. By understanding what has led to these structural changes in the past—particularly farmers’ decisions to leave—we can better predict future patterns [15,26,27,28,29]. These changes mirror global trends but are influenced by Türkiye’s unique socio-economic and cultural context.

Understanding the specific factors influencing Turkish farmers’ decisions to leave agriculture is crucial for developing targeted policies. By examining the experiences and decisions of farmers in Türkiye, this study aims to contribute to the broader understanding of agricultural sustainability and resilience.

Given this background, this study has two objectives. The first objective is to investigate what factors might drive farmers to leave using data from farms in Türkiye. The second objective is to conduct a sensitivity analysis to examine the possible effects of changes in selected factors that influence farmers’ decisions to leave. A binary logistic regression model was used to assess the factors causing farm exits and to calculate the exit probabilities for different farms. The findings will offer insights not only for Türkiye but also for other countries facing similar challenges in agricultural sustainability.

This study aims to address the following research question: How do socio-economic and political factors influence the likelihood of farmers exiting the agricultural sector in Türkiye? Based on the literature and preliminary observations, this study proposes the following hypotheses:

H1. 

Larger farm sizes and higher agricultural subsidies decrease the likelihood of farmers exiting the agricultural sector in Türkiye.

H2. 

Higher debt levels and older age increase the likelihood of farmers exiting the agricultural sector in Türkiye.

The findings from this study will contribute to the development of policies aimed at strengthening the resilience and sustainability of rural farming communities in Türkiye.

This paper consists of six parts. Section 2 reviews the existing literature. Section 3 describes a framework for analyzing the probability of farm exit. Section 4 presents the findings and Section 5 discusses the factors affecting farm exit. Section 6 presents the conclusions.

2. Literature Review

Sustainable agriculture faces numerous global challenges, including climate change, biodiversity loss, land degradation, and economic pressures. These challenges threaten the integration of a healthy environment, financial success, and social equity, which are the primary objectives of sustainable agriculture. Worldwide, there has been a decline in agricultural labor due to technological advancements, increased labor productivity, and the availability of alternative employment opportunities in urban areas [30]. This trend is causing significant shifts in agricultural practices and farmer demographics. For example, large farms gain a competitive advantage through better access to information, financial resources, and managerial expertise [16].

Research across various countries has identified key factors influencing farmers’ decisions to exit agriculture. Economic pressures are a primary determinant, as declining agricultural revenue due to volatile input and output prices makes farming less viable [31,32]. Technological advancements also contribute by reducing the need for labor and increasing efficiency, favoring larger operations. However, it also presents a range of challenges such as high initial investment costs, potential job displacement in rural areas, and the risk of widening existing inequalities. Limited internet access and digital literacy in rural communities further hinder technology adoption.

Demographic factors such as age, education, and household dynamics play significant roles. Studies have shown that older farmers are more likely to exit farming as they approach retirement, while younger farmers may leave due to better opportunities elsewhere [33,34]. Social factors include the probability of the farmer having a successor and the transfer of the farm to that successor [31,35,36]. For several reasons, retaining young farmers is desirable. Firstly, they are crucial for the future of food security and sustainable agriculture [37]. However, attracting and retaining them is a challenge due to factors such as the increasing age of farmers and rising farming costs [38]. The absence of young farmers can result in farms leaving the sector, affecting its survival and competitiveness [39]. Jansuwan [40] proposes multifunctional farming as a successful pathway for the next generation of farmers to address this issue. However, there is a need to transform the perception of the agricultural sector to make it more attractive to young workers [41]. Despite these challenges, Hamilton [42] has found that younger farmers are more innovative and entrepreneurial, demonstrating higher levels of performance.

In addition, farm size and land ownership significantly impact exit decisions. Larger farms tend to have more resources and stability, reducing the likelihood of exit [24]. Land ownership provides security against borrowing constraints, whereas rented land increases uncertainty and the likelihood of exit [16]. Subsidies and government support also play crucial roles; increased financial support can mitigate economic pressures and encourage farmers to continue their agricultural activities [43]. Mishra et al. [44] and Devadoss et al. [45] also show that a decrease in the intensity of government payments increases the probability of farm exit, suggesting that government subsidies play an important role in keeping farm households in agriculture.

Research into the effects of these factors on farming has been conducted across various countries and commodities. These products include milk [46,47,48,49], grain [50], tobacco [34], hog [51], and livestock [52]. In addition, empirical work has been conducted in the EU [24,53], the USA [44,54], Switzerland [55], Croatia [56], Estonia [57], Norway [58], Thailand [40], Pakistan [21], and Türkiye [59].

In regions with similar socio-economic contexts to Türkiye, such as Pakistan, Thailand, and several EU countries, specific factors have been identified that influence farm exit decisions. For instance, in Pakistan, studies have shown that younger farmers are more likely to continue farming if provided with adequate support and resources, such as subsidies and training programs [21,60]. In Thailand, demographic shifts and economic pressures similarly influence farm exit decisions, with older farmers more likely to leave farming [40].

Türkiye encounters similar challenges related to the dynamics of its rural population. The rural population has shown a significant downward trend in recent years, with a notable demographic shift from predominantly young individuals to an aging population due to rural urban migration and limited work opportunities. This has resulted in unemployment and insufficient income in rural areas, mirroring global trends but also influenced by Türkiye’s unique socio-economic and cultural context [59]. The authors of [61] reveal that farm income, family traditions, and a lack of other job opportunities are the main factors influencing young farmers to stay in Turkish agriculture. In contrast to the findings of this study, [15] found that agricultural subsidies, tenure status, education, and farm size were the factors that influenced farmers to quit farming.

Most studies investigating this problem have focused on estimating the exit rates or their impact, rather than the experiences of farmers or households who exit [48,58,62,63]. Far less attention has been paid to the potential impacts of farmer welfare, climatic events, community characteristics, and access to water. In recent years, few studies have investigated the relationship between farmer welfare and farm exit intentions [64,65]. In parallel with the increasing importance of climatic events, studies have also been conducted to investigate the relationship between climatic conditions, climate risks, and farm exit intentions [60,66,67,68].

Although there is rich literature on this topic, the findings are ambiguous and contradictory. This is because the relationship between quitting farming and various factors is complex and dependent on a specific context or region. In addition, the ambiguous results may indicate that there are other factors that are not captured in the analysis. More research is therefore needed to address these limitations and deepen our insights into the factors influencing exit decisions. There is also a lack of empirical work examining the impact of specific policy interventions on farm exit decisions in Türkiye.

This study aims to fill this gap by providing data-driven insights into the factors influencing farmers’ exit decisions in Türkiye, using a binary logistic regression model to analyze data from 195 farmers in the İzmir province. By conducting a sensitivity analysis, this study also explores the potential effects of changes in key variables, such as farm size, subsidies, and debt, on farmers’ decisions. These findings contribute to both regional and global discussions on agricultural sustainability and inform policy interventions aimed at enhancing the resilience and sustainability of rural farming communities in Türkiye.

3. Materials and Methods

This study used data gathered through a survey. The survey was conducted on individual Turkish farms located in ten districts within the province of İzmir. İzmir lies between the latitudes 37°45′ and 39°15′ north and the longitudes 26°15′ and 28°20′ east. This province is characterized by diverse agricultural practices, including a wide range of crop types, farm sizes, and socio-economic conditions. İzmir is one of Türkiye’s most economically developed regions, with a strong agricultural sector that benefits from favorable climate conditions, fertile soil, and access to markets. This economic advantage adds value to production. İzmir alone accounts for around 22.0% of the Aegean Region’s agricultural gross value added (GVA) and around 3.7% of Türkiye’s agricultural GVA, ranking second [69].

This study used 195 randomly selected farmers. The sample size was determined using the finite population proportional sample size [70]. Accepted statistical parameters are a 7% standard error and a 95% confidence interval. The questionnaire included open-ended and closed-ended questions about the characteristics of the farmer and their family, farm production, various farming activities, as well as their views on both farm and non-farm activities. The participants were also asked about policy support and their future farming plans.

We aimed to estimate the likelihood of farmers deciding to exit based on a set of important variables. A logistic regression model was formulated to perform the empirical analysis. Logistic regression was used if the structure of the dependent variable was categorical, discrete, or ordinal. There was no prerequisite of a normal distribution and continuity assumption [71]. Logistic regression analysis is a method that calculates the estimated values of the dependent variable as a probability and allows classification in accordance with probability rules. The dependent variable can be binominal, multinomial, or ordinal. Logistic regression analysis can be applied in all three cases. Data analysis was performed using Stata 15. The EXIT variable, the dependent variable in our study, refers to the response to the question “Do you intend to continue farming in the foreseeable future?”. The value assigned for the variable was set as “1” in cases where the respondent gave a negative response and “0” in all other cases. Using this, a binominal logistic regression analysis was conducted. Based on the present model, the log likelihood function for estimation can be expressed as follows:

$$\ln \mathrm{L}={\sum }_{\mathrm{t}=2}^{\mathrm{T}}\left\{{\mathrm{E}\mathrm{X}\mathrm{I}\mathrm{T}}_{\mathrm{i}\mathrm{t}-1,\mathrm{i}\mathrm{t}}\mathrm{l}\mathrm{n}\mathrm{F}\left({\mathrm{X}}_{\mathrm{i}\mathrm{t}-1}\mathsf{\beta }\right)+\left(1-{\mathrm{E}\mathrm{X}\mathrm{I}\mathrm{T}}_{\mathrm{i}\mathrm{t}-1,\mathrm{i}\mathrm{t}}\right)\mathrm{l}\mathrm{n}\left[1-\mathrm{F}\left({\mathrm{X}}_{\mathrm{i}\mathrm{t}-1}\mathsf{\beta }\right)\right]\right\}$$

(1)

EXIT_it−1,it is a binomially distributed variable denoting the quitting of farming. It indicates whether observation unit i realizes the event at time point t. F(X_it−1β) represents the independent variable values of the first observation unit Xit−1 at time point t−1 and β is the coefficient of the logistic regression model. F(X_it−1β) denotes the probability of the event occurring under the conditions specified by this model. If EXIT_it−1,it = 1, i.e., the event has occurred, this term is 0. Otherwise, EXIT_it−1,it = 0 and this term becomes 1. In this case, the expression lnL sums the log likelihood for all time points from t = 2 to T for time point T. For each time point, the log likelihood or the complement of the likelihood, the log minus 1 likelihood expression, is summed according to the actual event status of the observation unit.

Given the likelihood of farm exit in consecutive years, we can interpret that a rational farmer engages in a discrete decision-making process at the end of each year. This decision involves choosing whether to continue farming or to exit the agricultural sector, and this choice is shown in the subsequent year. After conducting an analysis to estimate the probability of farmers abandoning their agricultural activities, we examined the partial effects of variables that were found to be statistically significant on these probabilities. The marginal effect of a parameter on farm exit can be conceptualized and quantified as follows:

$$\frac{\mathrm{\delta }\mathrm{E}\left(\mathrm{y}|\mathrm{x}\right)}{{\mathrm{\delta }}_{{\mathrm{x}}_{\mathrm{k}}}}=\mathrm{F}(\acute{\mathrm{\beta }}\mathrm{x})\left[1-\mathrm{F}(\acute{\mathrm{\beta }\mathrm{x})}\right]{\mathsf{\beta }}_{\mathrm{k}}$$

(2)

Table 1. Description of variables.

Variable	Description and Units	Supposed Impact	Mean	Std. D	No. of Farmers	%
EXIT	1 = plan to exit from farming; 0 = otherwise
AGE	Age, years	Positive	49.2	10.3	195	--
EDU	Education years	Positive	6.2	2.3	195	--
MEMB	Household members count	Negative	3.7	1.3	195	--
EMP	Off-farm employment Having off-farm employment = 1, otherwise = 0	Positive	--	--	33	16.9
SIZE	Farm size groups	Negative	10.07	133.5	--	--
	≤2.5 ha = 1, otherwise= 0		--	--	45	23.1
	2.6–5.0 ha = 1, otherwise= 0		--	--	40	20.5
	5.1–10.0 ha = 1, otherwise= 0		--	--	57	29.2
	10.1–20.0 ha = 1, otherwise= 0		--	--	28	14.4
	20.1 + ha = 1, otherwise = 0		--	--	25	12.8
TYPE	Farm type	Positive/negative	--	--
	Livestock Share of livestock production in total farm production is more than 50% = 1. otherwise = 0		--	--	30	15.4
	Cotton Share of cotton production in total farm production is more than 50% = 1. otherwise = 0		--	--	15	7.7
	Tomatoes Share of tomatoes production in total farm production is more than 50% = 1. otherwise = 0		--	--	18	9.2
	Grains Share of grains production in total farm production is more than 50% = 1. otherwise = 0		--	--	43	22.1
	Mixed Share of many productions in total farm production is more than 10% = 1. otherwise = 0		--	--	19	9.7
	Olive Share of olive production in total farm production is more than 50% = 1. otherwise = 0		--	--	12	6.2
	Fruits Share of fruits production in total farm production is more than 50% = 1. otherwise = 0		--	--	19	9.7
	Vegetables Share of vegetables production in total farm production is more than 50% = 1. otherwise = 0		--	--	25	12.8
	Potatoes Share of potatoes production in total farm production is more than 50% = 1. otherwise = 0		--	--	14	7.2
SUB	Agricultural subsidy having agricultural subsidy = 1, otherwise = 0	Negative	--	--	335	90.8
LAND	Own land or rented land having rented land = 1, otherwise = 0	Positive	63.3	91.6	112	57.4
DEBT	Debt having loan = 1, otherwise = 0	Positive	--	--	138	70.7

Source: authors’ own calculations.

presents the variables used in the binary logistic regression analysis. The variable EXIT is the dependent variable. It is expected that the various independent variables, such as farm income, farmer’s age, education, family size, off-farm employment, agricultural support, type of farming activity, rented land, and debt, will affect the exit decision, denoted as Xit. Subsequently, a simulation was conducted to examine the implications of changes in these variables on the probability of farm exit.

Demographic factors, such as age (AGE), education (EDU), off-farm employment (EMP), and the number of family members engaged in agricultural activities (MEMB), were employed to determine the structure and dynamics of a farmer’s household. In accordance with current trends, it is common for farmers to consider exiting the agricultural sector as they approach retirement age. The variables EDU and EMP represent the probability of engaging in employment within sectors other than agriculture. Based on empirical evidence, farmers who have obtained educational qualifications and are involved in non-agricultural occupations are more likely to show higher levels of competitiveness in the non-agricultural labor market [21,72]. The variable in question is expected to positively correlate with the probability of farmers exiting the agricultural sector. The involvement of family members who work on the farm plays a significant role in determining the agricultural succession dynamics. Therefore, it has a significant influence on the decision-making process of continuing or withdrawing from the agricultural sector. It is expected that agricultural enterprises characterized by a higher proportion of family members involved in farm operations will likely show lower levels of disengagement.

The SIZE variable is used as an indicator to measure the size of the farm in the binary logistic model. This denotes the aggregate landholding of a producer, encompassing both owned and leased land. Given that a larger farm size positively correlates with net income, it is anticipated to have an inverse relationship with the likelihood of exiting the agricultural sector. Our model employs five distinct categories to represent varying size groups (≤2.5 ha, 2.6–5.0 ha, 5.1–10.0 ha, 10.1–20.0 ha, 20.1+ ha), with the initial group, i.e., small farms, being accepted as the reference group.

The agricultural sector covers various farm types, including livestock, cotton, tomatoes, grains, mixed, olives, fruits, vegetables, and potatoes. These nine categories represent the various types of production. The reason for including the production type in the model is that it depends on the investments made in buildings and machinery. According to the model, farms that produce grain serve as the reference group to estimate the exit probabilities of farms.

The variable subsidy (SUB) represents the political conditions. As per the findings in [43], it can be inferred that government subsidies have an adverse effect on the likelihood of farm failure. Increased income can enhance the profitability of the agricultural sector relative to other vocations, thereby mitigating the likelihood of farmers leaving their occupations.

The LAND variable denotes the land ownership status of the farmer, i.e., whether the land is owned or rented. The binary variable denoted by “1” represents the presence of land rental by the farmer, while “0” indicates the absence of such rental. The ownership of LAND is expected to have a positive impact on the likelihood of exiting the agricultural sector. This is because individuals who own larger amounts of land are better equipped to overcome borrowing constraints, thereby overcoming the limitations on future farm expansion [16].

The DEBT variable pertains to the presence or absence of loans among farmers. The binary variable denotes the presence or absence of credit utilization by the farmer, with a value of “1” indicating the former and “0” indicating the latter. The likelihood of farmers withdrawing from the agricultural sector is anticipated to be influenced by DEBT. This is because elevated levels of debt heighten the possibility of insolvency, which diminishes the appeal of farming as a viable option [16].

4. Results

4.1. Basic Characteristics of Farms and Farmers

The agricultural sector holds considerable importance in the Turkish economy. In Türkiye, agricultural activities are predominantly characterized by family ownership, with a tendency towards smaller and fragmented farm sizes [73]. Nationwide, the average farm size is 6.1 hectares in TURKSTAT’s 2001 General Agricultural Census and 7.6 hectares in the 2016 Agricultural Enterprise Structure Survey, which accounted for the 2014/2015 production season. According to the latest data from the Farmer Registration System, it was 7.0 hectares as of 2017 [74]. In the same way, 12.8% of the farms that were studied can be categorized as large farms. According to the data presented in Table 1, most farms (87.2 percent) are classified as small farms. The average size of the farms in the sample is slightly larger that of farms in Türkiye, as indicated in

Table 2. Socio-economic characteristics of farms.

	Total	Exiting from Farming
		Non Exit (0)	Exit (1)
	Mean	Mean	Mean
Age of the farm operator (years)	49.21	48.73	50.74
Education of the farm operator (years)	6.22	6.19	6.32
Farming experience of the farm operator (years)	25.06	24.27	27.55
Household size (person)	3.69	3.72	3.62
Average farm size (ha)	10.09	10.90	7.40

Source: authors’ own calculations.

.

The Credit Bureau’s study [74] confirms that Turkish farmers’ average land size has shifted from 2.0–9.99 hectares to 10.0–49.9 hectares. Thus, regardless of the official average size, in practice, farmers have come a long way towards complying with the definition of “agricultural land with sufficient income” announced by Law No. 6537. They also determined that the geometric mean was 9.9 hectares and the median was 10.0 hectares. Therefore, they claim that 10.0 hectares is an indicator of agricultural production in Türkiye. Our study’s average enterprise size supports the Credit Bureau’s average. We can also say that Law No. 6537 has had an impact in İzmir.

Table 2 displays the survey results according to the socio-economic characteristics of the farms and farmers. The average age of operators in agricultural households was about 49 years. The average years of farming experience among the agricultural producers was 25 years. The average level of education achieved by the farmers in the examined population was around six years. The average household size was about four individuals.

According to the findings presented in

Table 3. Farm exit rates according to farm size.

Farm Size	Exiting from Farming		Average Farm Size (ha)		Average Farm Size (ha)
	Non Exit	Exit	Non Exit	Exit	Mean
	Count	Count	Mean	Mean
>2.5 ha	33.00	12.00	1.70	1.20	1.80
2.6–5.0 ha	29.00	11.00	4.00	4.40	4.20
5.1–10.0 ha	41.00	16.00	7.70	7.40	8.00
10.1–20.0 ha	22.00	6.00	15.5	13.0	15.60
20.1+ ha	23.00	2.00	34.1	45.4	36.50

Source: authors’ own calculations.

, a significant negative correlation is evident between the farm size and the probability of farmers exiting the agricultural sector. As the farm scale expands, there is a corresponding decrease in the percentage of farmers who intend to discontinue farming activities.

4.2. Results of the Binary Logistic Model

The binary logistic model was utilized to calculate a farmer’s likelihood of discontinuing farming based on the explanatory variables. The optimization of Equation (1) was achieved through the maximization of the log likelihood function. The binary logistic regression’s parameter estimates are presented in

Table 4. The goodness of fit of the binary logistic model.

Goodness of fit
Number of cases correctly predicted	81.03%
Area under ROC curve	0.7651
Goodness of fit (Pearson chi2)	210.64 (0.0294)
Linear restriction	chi^2(20) = 48.062, with p-value = 0.000416955

Source: author’s own calculations.

. The binary logistic model coefficients represent the estimated partial derivatives of the propensity to exit in relation to each explanatory variable. A negative coefficient denotes a mitigating effect of the factor on the likelihood of exit. A positive value suggests a tendency towards exit, indicating a higher likelihood of abandonment.

Regarding the predictive effectiveness of the binary logistic model, the decision of 158 out of 195 sampled farmers were accurately predicted, which corresponds to an accuracy rating of 81.03 percent. The ROC curve was constructed by graphing the sensitivity, which represents the likelihood of correctly detecting a true positive as positive, against 1-specificity, which represents the likelihood of incorrectly predicting a true negative as positive. The greater the model’s predictive accuracy, the larger the area under the ROC curve. The area under the ROC curve for the model is 0.7651 (Figure 1). The chi-square test of linear restriction was utilized to examine the model’s overall suitability. The null hypothesis states that the estimated joint coefficients are equal to zero. The null hypothesis is rejected at a significance level of 1%. At least one of the regressors is significantly different from zero and contributes to the outcome prediction (Table 4).

Table 5. The results of binary logistic regression.

Dependent Variable: EXIT (1) Plan to exit from Farming (0) Otherwise
Independent Variable	Coefficient	Standard Error	z	p-Value
Constant	−2.12	1.69	−1.25	0.211
AGE	0.03	0.02	1.80	0.072 *
EDU		0.09	1.39	0.165
MEMB	−0.12	0.15	−0.83	0.408
EMP	0.15	0.49	0.31	0.757
SUB	−1.36	0.56	−2.42	0.016 **
LAND	0.75	0.43	1.73	0.084 *
DEBT	0.78	0.46	1.70	0.090 *
SIZE 2.6–5.0 ha	−0.16	0.55	−0.29	0.772
SIZE 5.1–10.0 ha	−0.002	0.55	0.01	0.996
SIZE 10.1–20.0 ha	−0.51	0.69	−0.74	0.459
SIZE 20.1+ ha	−1.70	0.95	−1.78	0.076 *
Livestock production	−0.96	0.61	−1.58	0.114
Cotton production	−2.22	1.18	−1.88	0.060 *
Tomato production	−0.71	0.68	−1.03	0.305
Mixed production	−1.85	0.85	−2.17	0.030 **
Olive production	−1.59	0.95	−1.68	0.093 *
Fruits production	−3.09	1.22	−2.52	0.012 **
Vegetable production	−0.42	0.63	−0.67	0.502
Potato production	−0.90	0.74	−1.21	0.227
Log-likelihood	−88.671144
LR	Chi-square(19) = 38.04 [0.0059]
McFadden R2	0.1766
AIC	217.3423
BIC	282.8023
df	20

* p < 0.1. ** p < 0.05. *** p < 0.01. Source: authors’ own calculations.

shows that, according to the results of the binary logistic model, nine of the nineteen variables considered in the study affect the probability of farmers leaving the profession. Age, rented land, and debt status have a positive effect and increase the likelihood of farmers leaving the profession. However, subsidies, farm size, and farm type reduce their likelihood of abandoning agriculture. The effect of changes in the explanatory variables should be interpreted while considering the marginal effects (

Table 6. Marginal effects of the binary logistic model.

Variable	dy/dx	Std. Err.	Z	p > |z|	[95% C.I.]		Mean of X
AGE	0.0054731	0.0029499	1.86	0.064	−0.0003086	0.0112547	49.21
SUB	−0.2024824	0.0793694	−2.55	0.011	−0.3580437	−0.0469212	0.82
LAND	0.1108211	0.0624186	1.78	0.076	−0.0115171	0.2331593	0.57
DEBT	0.1166281	0.0671775	1.74	0.083	−0.0150374	0.2482935	0.65
SIZE 20.1+ ha	−0.1995191	0.0922326	−2.16	0.031	−0.3802916	−0.0187466	36.50
Cotton production	−0.3188595	0.1107733	−2.88	0.004	−0.5359713	−0.1017478	0.07
Mixed production	−0.2872554	0.1041315	−2.76	0.006	−0.4913494	−0.0831615	0.09
Olive production	−0.2612877	0.1234842	−2.12	0.034	−0.5033122	−0.0192632	0.06
Fruits production	−0.3666663	0.0874049	−4.20	0.000	−0.5379767	−0.1953559	0.09

Note: dy/dx for factor levels is the discrete change from the base level. Source: authors’ own calculations.

).

The results indicate an inverse relationship between farm size and the likelihood of farm exit, meaning that as the size of the farm increases, the probability of farmers exiting farming decreases. In other words, owners of larger farms are less likely to leave farming than owners of smaller farms. According to Table 6, a 10% increase in farm size decreases the probability of farm exit by 19.95%.

The dummy variables used to represent the farm structure show that agricultural enterprises that grow more cotton, olives, and fruits than cereals are more likely to be involved in farming activities. The above result is supported even more when looking at production systems that include both crops and livestock. In addition, a unit increase in farm diversification (cotton, mixed, olive, and fruit production) decreases the probability of existing from farming by 31.9%, 28.7%, 26.1%, and 36.7%, respectively. This study shows that farms that grow cotton, olives, and fruit instead of tomatoes, potatoes, and other vegetables tend to have lower levels of agricultural work. This trend might be because of the seasonal need for workers in crop production. The Aegean and Mediterranean coastal regions exhibit good climatic and geographical conditions that are favourable to the cultivation of fruits and vegetables.

Debt is another important and relevant reason for leaving farming. In this study, higher debt levels were found to be significantly associated with an increased likelihood of planning to quit farming. Each additional unit of debt corresponds to a significant increase in the probability of planning to quit farming. The results show that, all things being equal, a unit increase in the level of debt owed by farmers defined by a change from (Xi = 0) to (Xi = 1) results in an increase of 11.7% in the probability of quitting farming. This finding highlights the importance of debt management and financial considerations in farmers’ decisions to continue or exit farming. High debt levels can pressure farmers and contribute to their intentions to exit farming, which should be considered in agricultural policies and support initiatives.

Land ownership is one of the economic factors that can impact farm abandonment. A unit increase in the leasing of a large share of farmland increases the probability of quitting farming by 11.1%.

The variable representing age exhibits statistical significance and indicates that younger individuals are often less inclined to discontinue their involvement in the agricultural sector compared to older individuals. The marginal effect of a unit increase in the farmer’s age on the conditional probability of exiting from farming is 0.5% (Table 6). This implies that a unit increase in the farmer’s age increases the probability of exiting farming by 0.5%.

Other socio-demographic factors affecting farming abandonment include the level of education of the farmer and non-farm employment (EMP). In our study, since the level of education and off-farm employment variable was not statistically significant, no statistical effect was found. This can be explained by the fact that there were no farmers working outside agriculture, and there were only a few farmers with higher levels of education in the sample. These individuals may hold supervisory or managerial positions.

Studies in the literature suggest that family size may indirectly affect exit rates through its effect on factors such as off-farm work and age. Indeed, in our study, the family size (MEMB) was not found to be statistically significant. In this respect, our finding suggests that this variable is not a direct determinant of farming exit.

As expected, the agricultural subsidy (SUB) negatively impacts the likelihood of farm exit rates, and this effect was statistically significant at a 5% level. For a farmer, a 10% increase in agricultural subsidy decreases the probability of existing from farming by 20.25%.

Overall, the results show that the following factors influence the decision to leave farming among the surveyed sample of Turkish farmers: the farm type, farm size, agricultural subsidies, land ownership, debt level, and farmer age. The impact of variables assessing farmer age and land ownership appears to be significant, whereas the farm type, farm size, and agricultural subsidies appear to be prominent determinants for Turkish farmers.

4.3. Simulating the Impact of Changes in the Variables on the Probability of Quitting Farming

A sensitivity analysis was carried out in this study. The goal was to replicate the effect of changes in statistically significant variables on the likelihood of farmers making exit decisions. The simulation was run against a base scenario that reflects farmers who plan to exit vs. farmers who do not plan to exit.

In the first scenario, we investigated the potential of a farmer aged 49.21 who has rented land, is in debt, and does not receive agricultural subsidies. As a result, the dichotomous variable “subsidy” was set to 0, implying that the farmer receives no agricultural subsidies. Assuming that farmers have debts and rent land, the variables “debt” and “land” were set to one. The initial condition for the continuous variable “age” on the other hand was set to the sample mean. The simulation results suggest that as the farm size increases, the likelihood of farm exit reduces. Farmers engaged in all forms of production, except cotton and fruit production, have a higher likelihood of farm exit (

Table 7. Simulated impact of variables on the probability of exiting farming.

	Age = 49.21		Age = 49.21		Age = 35.00		Age = 35.00
	Sub = 0		Sub = 1		Sub = 0		Sub = 1
	Land = 1		Land = 0		Land = 1		Land = 0
	Debt = 1		Debt = 0		Debt = 1		Debt = 0
Variable	dy/dx	p > |z|	dy/dx	p > |z|	dy/dx	p > |z|	dy/dx	p > |z|
FARM SIZE
SIZE <2.5 ha	0.6240127	0.000	0.084092	0.064	0.495749	0.003	0.0515819	0.120
SIZE 2.6–5.0 ha	0.5853367	0.000	0.0724332	0.090	0.455394	0.008	0.0442129	0.161
SIZE 5.1–10.0 ha	0.623357	0.000	0.0838771	0.073	0.4950505	0.003	0.0514454	0.131
SIZE 10.1–20.0 ha	0.4985862	0.010	0.05214	0.142	0.370686	0.050	0.031557	0.192
SIZE 20.1+ ha	0.2318318	0.188	0.0164213	0.307	0.1516631	0.255	0.0097931	0.342
FARM TYPE
Livestock production	0.5739962	0.003	0.0693675	0.116	0.4438762	0.034	0.0422871	0.186
Cotton production	0.2778818	0.231	0.0208442	0.415	0.185637	0.300	0.0124534	0.434
Tomato production	0.63559	0.001	0.0879966	0.158	0.5081628	0.017	0.0540661	0.216
Grain production	0.7797786	0.000	0.163797	0.042	0.6771612	0.000	0.1039707	0.095
Mixed production	0.3580698	0.094	0.0299339	0.257	0.2483612	0.167	0.0179511	0.292
Olive production	0.4172044	0.080	0.0380933	0.272	0.2977824	0.174	0.0229213	0.323
Fruit production	0.1377829	0.320	0.0087627	0.427	0.0864756	0.369	0.0052094	0.451
Vegetable production	0.6987138	0.000	0.1137053	0.083	0.5787296	0.001	0.0706295	0.134
Potato production	0.5902543	0.004	0.0738088	0.173	0.4604321	0.037	0.0450785	0.224

Source: authors’ own calculations.

).

In the second scenario, we replaced the variable “subsidy” with one that indicated that the farmers received agricultural subsidies. The variables “debt” and “land” were set to zero, assuming that farmers have no loans and possess land. The sample mean was still the “age” variable. The findings of this example illustrate that as the farm size increases, the likelihood of farm exit reduces. When the two outcomes are compared, it is clear that the probability of abandoning farming is smaller in the second situation. The quitting rates for farmers who only produce grain and vegetables are 16.38% and 11.37%, respectively (Table 7).

In order to assist young farmers, the “Young Farmer Projects Support (YFPS)” was included in the “National Agricultural Project” in 2016. This program attempts to encourage young farmers to stay in agriculture by providing them with financial assistance, as well as to discourage migration from rural areas to the city. The age variable was set to 35 in this case. The “subsidy” variable was set to zero, implying that farmers do not receive agricultural subsidies. Assuming that farmers have debts and rent land, the variables “debt” and “land” are set to one. In this situation, the likelihood of young farmers quitting farming decreases as the size of the farm grows. Similarly, farmers involved in all forms of production, except for cotton, mixed, olives, and fruit production, have a higher probability of farm exit (Table 7). The “subsidy” variable was set to one in the fourth scenario, implying that farmers receive agricultural subsidies. Assuming that farmers have no debts and own land, the variables “debt” and “land” were set to zero. In this situation, younger farmers have no plans to stop farming, even if their farm size grows. Furthermore, only wheat farmers are likely to stop farming (Table 7).

5. Discussion

Research on farm abandonment has revealed many factors influencing farmers’ decisions to abandon the agricultural sector. In general, these factors are categorized as economic, social, and political. In our study, the farm size and the farm structure as economic factors affected farmers’ decisions to leave farming. Our study’s findings on farm size indicate that larger farms have a reduced chance of exiting farms, complementing the findings of [59] in terms of the economic dimension of the farming operation itself. This may be due to several reasons deriving from economies of scale. Many studies have emphasized the critical role of economies of scale in the sustainability of agricultural farms. Indeed, larger farms can enjoy economies of scale that small farms often cannot. These can be achieved through better market and technological access, the ability to undertake more efficient production practices, and the spreading of fixed costs over a substantial output. In addition, large farms have more opportunities for diversified production, which makes them less sensitive to market and environmental risks [75]. By growing several crops or raising different animal species, these farms are less affected by diseases, attacks of pests [76,77], or commodity price fluctuations [78,79] that might destroy small, less diversified farms. Larger farms can also have a comparative advantage with regard to access to agricultural subsidies. Subsidies often significantly influence the economic viability of farms, acting as a safety net and reducing the risk of farm exit [10]. On the other hand, for smaller farms, the combination of higher per-unit costs, lower bargaining power, more significant challenges in accessing credit, and an inability to leverage the latest advancements in agricultural technology contribute to a higher likelihood of exiting farming.

The contrast between our findings and those in [80] leads to an interesting discussion with respect to the farm size’s influence on exit probabilities in different economic contexts. As our analysis revealed, in Türkiye, similar to Canada, owners of larger farms have a lower tendency to exit farming. This aligns with the finding regarding economies of scale, where a larger farm size reduces the likelihood of exit due to improved profitability and viability, thus reducing the incentive to leave the profession. The results of [80] with regard to Israel, where owners of larger farms are more likely to exit, therefore seem counterintuitive. The results of [80] suggest this might be because of fewer planned farm exits and institutional constraints on land transactions in Israel. These might result in situations where large farms are under different pressures or land transactions are more difficult so it is advantageous to exit from the larger farm and into activities that are more likely to be profitable.

Indeed, previous studies [22,24,48,57,62,63] confirm that farm size is generally associated with lower exit probabilities across different regions and subsectors of agriculture. Our finding on the importance of farm size as a determinant of exit patterns in the Turkish agricultural sector is consistent with previous studies and supports their findings. Policy interventions that promote farm consolidation and cooperative farming could enhance the sustainability of small farms by enabling them to achieve economies of scale. These approaches can help small farms improve their market competitiveness and enhance their overall sustainability.

The finding regardging farmers’ ages from our study coincides with the findings of the study conducted by [59] in Niğde province. In addition, Refs. [22,72] found that age affects decisions to leave farming, but in slightly different contexts. The authors of [22] state that age affects farmers’ decisions to retire in a non-linear way, with farmers at retirement age showing different tendencies to exit farming or not to invest depending on their age. The authors of [72] observed that the probability of farm growth is highest in the 40–49 age group and younger operators are less likely to exit the farm. Our study agrees with the view that younger operators are less likely to exit the farm, which is consistent with the findings of [55,72] wherein younger farm operators are less likely to exit the farm. The authors of [63] also observed a similar trend among 35-year-old farmers. Meanwhile, the results of [56] indicate that younger people are more mobile and susceptible to economic influences. Thus, they may leave agriculture for non-farm jobs or endure temporary unemployment due to frictional issues. Older farmers are also more likely to retire [22]. The authors of [80] showed that an extra year in age raises the risk of leaving farming by 0.6 and 0.5 percentage points for all farms and primary farmers, respectively. Further, Ref. [48] found that older farmers leave farming more often. As farmers age, transferring from farming to non-farm employment becomes less practical because specialized human-capital investments are necessary and older farmers have a shorter time to retirement to recoup these investments. However, as the farm operator ages, the probability of farm leave increases, especially on farms with unlikely succession. While all studies agree that age is an important factor in the decision to leave farming, they explore different aspects of how age affects these decisions. The results from these studies are complementary rather than contradictory and provide a broader understanding of the role of age in farming cessation. The findings highlight the need to attract younger generations to farming. Creating an attractive environment for young farmers can ensure the long-term sustainability of the agricultural sector.

Studies have shown that there is a complex relationship between the level of education of farmers and the decision to quit farming. Indeed, Refs. [21,81] found that the level of education has an effect on farm exit. The expected effect of the education dummy variable on the probability of farm exit is not statistically significant in our study. This may be explained by the lack of farmers with advanced education in the study sample. Farmers with higher levels of education tend to work in supervisory or managerial roles, which reduces their incentives to look for new jobs. However, people with lower levels of education are less likely to find jobs in other fields.

Many studies show that off-farm employment increases the probability of leaving agriculture [31,48,62,80]. In Pakistan, factors such as family labor force characteristics, size of the cultivated land, and government assistance play a role in leaving farming [21]. In Canada, off-farm work is associated with both entry into and exit from farming, with a large proportion of farmers working off-farm and a high exit rate [82,83]. However, in the US, the effect of off-farm employment on farm turnover is less clear, with some studies finding no statistical effect [84]. Likewise, we found no statistical effect on farmers’ exit decisions.

A study conducted in another region of Türkiye revealed that the number of children significantly influenced the likelihood of continuing farming [59]. Similarly, a highly significant and negative effect on farm exit was reported by [55] for farms with larger families. However, in our study, no significant effect of family size was found.

The type of farming activity also significantly impacts exit decisions. The results show that farms engaged in cotton, mixed, olive, and fruit production are less likely to exit than those focused on grains. This finding aligns with previous studies indicating that the profitability and stability of different types of farming vary significantly [24,63]. The authors of [24] also found that a larger share of crop production significantly increases exit rates, while it significantly decreases in regions specializing in livestock production. Similarly, [58] found that dairy farms had lower rates of exit than cereal farms. Meanwhile, the results of [63] also note that dairy, beef, sheep, other livestock, and fruit farms are less likely to exit than cereal farms. Policies that support diversification into these more stable and profitable farm types could reduce exit rates. This could involve specific subsidies, technical assistance, and market development support for these crops.

Furthermore, this study underscores the importance of subsidies in sustaining agricultural activities. Similar to findings from the EU and the USA, subsidies in Türkiye reduce the economic pressures that drive farm exits [43,44]. The role of subsidies in increasing profitability discourages farmers from leaving the sector [24]. The findings from [44] show that government program subsidies may have slowed farm exits. Farm program payments could keep farms open that would otherwise close. This highlights the need for well-targeted subsidy programs that support not only farm incomes but also investments in modernizing agricultural practices and infrastructure. However, this result differs from another study [54], even if it affirms the conclusions of the previously mentioned studies. Their finding is that government payments strengthen exit likelihood.

Farmers eventually acquire more land, reducing the risk of farm abandonment. The role of land ownership in this process is also important. The authors of [85,86] highlight the impact of land fragmentation, the lack of family members to cultivate the land, and the structure of land ownership. Similar to the findings of [57,62], our analysis shows a negative relationship between farmers’ land ownership and exit rates from farming. Land ownership may indicate a strong emotional bond between the family and its institutional interests. Therefore, they rarely abandon agriculture. Moreover, a large land holding can increase reliability and financial stability [24].

The positive association between higher debt levels and farm exit likelihood is a critical finding that resonates with global studies on farm financial health. The authors of [87,88] found that high debt levels can negatively affect farm income and survival, but [88] also noted that debt can help build equity capital in family farms. This suggests that policies providing financial literacy training, access to low-interest loans, and debt restructuring options could be crucial in reducing exit rates. These measures can improve the financial stability of farms and decrease the likelihood of exits due to economic distress.

In conclusion, while the factors influencing farmers’ decisions to exit agriculture vary across regions, there are also shared concerns and opportunities on a global scale. Notably, the growing demand for organic products presents a significant opportunity to create a more economically viable sector for farmers, potentially encouraging them to remain in agriculture. Future research should therefore investigate how transitioning to organic farming impacts farmers’ decisions and livelihoods, providing more comprehensive and globally relevant insights into its role in supporting agricultural sustainability.

To address the broader challenges of agricultural sustainability, policies should focus on long-term initiatives that integrate economic, environmental, and social objectives. This includes promoting sustainable farming practices, enhancing rural infrastructure, and supporting rural communities’ overall quality of life.

6. Conclusions

This research examines the empirical correlation between farm exits in the İzmir province of Türkiye, as a case study, and various farm and family characteristics, as well as policy intervention measures. The findings suggest that exit rates are significantly impacted by factors such as farm type, farm size, and agricultural subsidies. From a policy standpoint, a crucial finding emerges: the implementation of subsidy payments has a substantial impact in mitigating the decrease in the quantity of farms. Furthermore, exit rates are impacted by factors such as farmers’ ages, their longevity in the profession, and their level of indebtedness. There is a higher probability that farms with larger sizes will remain engaged in agricultural activities. Meanwhile, the rate at which farms quit the agricultural sector is subject to variation based on the type of farm. Farmers who specialize in cultivating cotton, olives, and fruit crops are less inclined to discontinue their farming activities overall.

Although our study revealed a lower likelihood of agricultural abandonment among large farms, small farmers are also crucial for sustainability. To ensure the retention of small-scale farmers in agriculture, it is essential to implement targeted policies that provide financial support, access to markets, and technical assistance. Subsidies and grants can help alleviate the financial pressures faced by smallholders, enabling them to invest in sustainable farming practices and improve their productivity. Access to cooperative networks can also enhance their bargaining power and provide economies of scale. Additionally, educational programs focused on sustainable farming techniques and business management can equip small farmers with the skills they need to thrive in a competitive market. By creating an enabling environment for small-scale farmers, policymakers can enhance the resilience and sustainability of the agricultural sector.

Another significant finding is that young individuals are less inclined to abandon agricultural activities when presented with good circumstances. Farmers, regardless of their age, encounter challenges such as inadequate support, outrageous expenses, and shortages of equipment. Moreover, this circumstance poses a significant challenge to small-scale farmers sustaining their livelihoods. Nevertheless, it is crucial to retain small-scale farmers who possess limited educational attainment and rely on traditional agricultural practices in the near future. It is imperative to consider the implementation of short-term policy interventions that can effectively provide suitable incentives and selective subsidies. These measures aim to assist the small farms that have suffered, encouraging them to adapt their resource allocations to improve their overall effectiveness.

The Young Farmer Projects Support, which was initiated in 2016, primarily aims at retaining young farmers in the agricultural sector while also attracting a greater number of educated youths to engage in agricultural activities. This is based on the understanding that a higher degree of education has been found to enhance the performance of farm production. Furthermore, it enhances individuals’ willingness and capacity to simply accept innovations that align with strengthening circumstances and adapt to emerging market situations. Consequently, this would increase the level of competitiveness within the agricultural sector and support an absence of trade-offs between the principles of sustainability and profitability in rural regions. The implementation of strategies designed to enhance the effectiveness and financial viability of agricultural operations will contribute to the achievement of enduring economic sustainability. Farms that optimize resource utilization are likely to contribute to environmental sustainability.

Furthermore, the support of local efforts is of crucial importance for maintaining agricultural sustainability. Hence, maintaining and improving the quality of life of the local population is of the utmost importance. For now, implementing specific subsidies aimed at job creation, particularly for younger individuals engaged in farming, as well as targeted initiatives that educate farmers regarding alternative on-farm food systems, would effectively support the shift from farming to alternate forms of employment. Furthermore, policy-makers should explore the possibility of implementing initiatives to assist farmers in their smooth transition into emerging markets. This might include facilitating vertical integration within alternative production markets. Additionally, it may be beneficial to design these programs in a manner that considers the unique needs and characteristics of different age groups, in order to maximize the desired outcomes.

The technological shift in rural societies presents both opportunities and challenges. While it can speed up development with positive outcomes, at the same time, it displaces jobs and leads to alterations in farm ownership [50,51]. Automation diminishes the need for human labor, resulting in potential joblessness in regions with few other job opportunities. This problem stems from the digital divide, which disadvantages rural areas due to their limited access to technology, internet connectivity, and digital literacy, which affects their capacity to adapt and compete. Additionally, rapid technological shifts can affect traditional livelihoods and social structures, making it difficult for small-scale farmers to compete and leading to the decline of traditional farming knowledge and cultural heritage. However, despite these difficulties, technology itself is not a challenge. Technology can serve as an effective tool for rural development through careful planning and equitable implementation. To enable rural communities to successfully adjust and advance, policymakers and stakeholders should allocate resources towards the development of rural infrastructure, education, and support services. Technology has the potential to bring about positive change in rural communities by closing the gap in digital access and ensuring equal possibilities.

To address the broader challenges of agricultural sustainability, future research should replicate this study in different regions and countries to validate and extend our findings, include qualitative in-depth analyses of individual cases, and explore these factors in various contexts. Such comparative studies can enhance our understanding of the universal and context-specific factors influencing farmers’ decisions to leave farming, ultimately helping to develop comprehensive strategies for sustaining global agricultural sectors.