The aim of the research is to operationalise the notion of women’s empowerment in agriculture and to assess the effect of women’s empowerment on the marketisation of small farms, comparing emerging markets of different development levels – less developed Serbia and Moldova with more advanced economies. The study will also conceptually bridge the framework of the self-efficacy concept with the theory of social embeddedness.
We combine qualitative identification of different models of women’s empowerment and social involvement with quasi-experimental, multi-value treatment effect analysis. Data were collected via semi-structured surveys conducted on 3,172 small farms in Lithuania, Romania, Poland, Serbia and Moldova.
We conclude that, when viewed solely in terms of decision-making power, women’s empowerment may not be sufficient for building a market orientation and that a high level of women’s participation in social networks is decisive.
We posit that there is a significant knowledge gap in how the definition of women’s empowerment is operationalised within the agricultural sector. We address this issue by applying Bandura’s self-efficacy framework. We then examine how self-efficacy, as reflected by various women’s empowerment conditions, influences the marketisation of small farms. In this way, we make a conceptual contribution, revealing that self-efficacy can be more effective when embedded in a social network context, as referred to in the theory of embeddedness. From a practical perspective, we contribute to discussions about achieving the gender equality Sustainable Development Goal.
1. Introduction
The marketisation of small farms is of great importance in the context of the problems faced by the agricultural sector, not only in Central and Eastern European Countries (CEECs), but also in other parts of the world. In particular, it is land abandonment (Živanović et al., 2022; Czyżewski and Kryszak, 2023), which not only has negative socio-economic impacts, but also leads to a decline in soil quality, the erosion of uncultivated land and the disappearance of biodiversity of mid-field ecosystems. The sustainability of small farms is crucial for local food security, soil and landscape care and the preservation of traditions and cultural heritage (Toma et al., 2021).
The marketisation of small and medium-sized farms has posed problems for CEECs since their economic transition. Hanf and Gagalyuk (2018) highlighted this issue, referencing case studies from countries such as Serbia, Ukraine and Turkey and exploring potential solutions. A study in Lithuania (Droždz et al., 2021) confirmed this, demonstrating that low farm integration can be improved by increasing social activity, such as participating in educational programmes and organisations, or by strengthening social networks through workshops, counselling and producer groups.
In this context, the role of women is important. Women’s empowerment typically involves accepting women’s viewpoints, raising their social status and increasing awareness by participating in various social networks, such as educational and training systems, professional organisations and cultural networks (Alkire et al., 2013). Therefore, when discussing women’s empowerment, there are two interacting dimensions: (1) decision-making power and (2) participation in social networks (Alkire et al., 2012). In agriculture, the focus is on intra-household decision-making and participation in producers’ groups, associations and local social initiatives.
Prior to the systemic transition from a socialist to a market economy in the 1990s, women were equally involved in CEECs. In some countries, such as Hungary and Romania, they were even more involved than men, with figures reaching around 60%. The ‘feminisation of agriculture’ occurred in these countries during the 1960s and 1970s as a result of policies, industrialisation and the migration of men to cities. This led to an increase in the number of women in agriculture. It was also revealed during this time that women had the potential to manage farms efficiently and integrate them into the market (Chyłek et al., 1996). Nowadays, on average, 31.6% of farms across the EU are managed by women (Eurostat, 2022a, b; European Commission, 2021). However, this statistic varies significantly between countries. The gender imbalance among farmers was particularly pronounced in the Netherlands, where only 5.6% of farmers were female in 2020. Female farmers were also relatively uncommon in Malta (10.8% of all farmers), Germany (10.8%), Denmark (10.9%) and Ireland (11.4%). A much closer gender balance was seen in Latvia and Lithuania, with 44.8% and 44.9% of farmers being female, respectively.
Within the market economy framework, the positive effects of women’s empowerment on farming participation and market orientation have rarely been examined empirically, despite numerous hypotheses regarding how such a relationship might function. According to recent research by Quisumbing et al. (2021), women play an important role in various agricultural and non-agricultural value chains. However, in many countries, their contributions are undervalued or limited by prevailing social norms and gender barriers. A survey of 613 direct-market farmers in Canada found that female farmers have a unique experience of and understanding of short supply chains based on a desire for caring relationships and a belief in the benefits of direct marketing arrangements (Azima and Mundler, 2022). Rosener (1995) demonstrates that women tend to adopt a leadership style that is more interactive and participatory, encouraging input and information sharing from others and establishing and maintaining open communication channels with their subordinates. Other authors confirm this observation, advocating that, in organisational settings, women tend to be less hierarchical and more collaborative than men (Ravasi and Schultz, 2006) and to manage businesses in a more democratic way (Tang et al., 2011). Several European studies confirm this, showing that women are more likely than men to be involved in direct sales, organic farming, on-farm tourism initiatives and all forms of value-added agriculture (Annes et al., 2021). Women play also a key role in promoting the ethical, agroecological and cultural aspects of food, as well as supporting the smallholder family economy through the establishment of direct, personal relationships (Castelló and Romano, 2021; Nigh and González Cabañas, 2015).
However, the studies cited above mainly refer to Western European countries and relatively large farms. In CEECs, however, the impact of women’s empowerment on market orientation in small-scale farming has yet to be studied. We argue that there is a knowledge gap regarding the operationalisation of the definition of women’s empowerment in agriculture, as well as the social context in which women’s capabilities that favour the marketisation of small business entities become apparent.
We therefore attempt to bridge the concept of self-efficacy, as reflected in different conditions of women’s empowerment, with the theory of social embeddedness (Granovetter and Action, 1985). The social embeddedness framework has already been used to explain various farmer behaviours (Schwabe et al., 2022; Zhang et al., 2023; Zheng et al., 2022; Czyżewski et al., 2024).
The aim of the research is to operationalize the notion of women’s empowerment in agriculture and to assess the effect of women’s empowerment on the marketisation of small farms bridging the framework of self-efficacy concept with the social embeddedness theory, in selected CEECs.
Our contribution is conceptual, methodological as well as practical. We address the gap in knowledge regarding the operationalisation of women’s empowerment within the agricultural sector. To tackle this issue, we apply Bandura’s self-efficacy framework. We then examine how self-efficacy, as reflected by various conditions of women’s empowerment, influences the marketisation of small farms. We combine the self-efficacy framework with the theory of embeddedness, testing whether self-efficacy is more effective when anchored in a particular social network context. In terms of methodology, to the best of our knowledge, no previous study has combined qualitative measures of women’s empowerment with marketisation proxies using the multi-valued treatment effect approach, which is a quasi-experimental method that allows for causal inference. Furthermore, we adopted a comprehensive measurement of market orientation, following the definition by Zhang et al. (2020). Previous studies have focused either on analysing farm labour allocation – examining whether it was in accordance with business criteria – or on the proportion of agricultural output sold by different market channels (market participation) (Arinloye et al., 2015). Our study covers both aspects using a composite index that focuses specifically on the short-channel aspect requiring direct social involvement of farmers in relationships with customers. From a practical perspective, our study contributes to discussions about achieving the fifth Sustainable Development Goal of gender equality.
The remainder of the article is organised as follows. First, we present a literature review to provide background to the analysis, emphasising changes in the gender structure of agriculture, typical gender-based roles in agriculture and women’s capabilities for creating added value on farms. Next, we explain the design of our research and the methodological issues. Finally, we present and discuss the results and formulate general conclusions.
2. Literature review
2.1 Women’s empowerment and the marketisation of farms in European agriculture
The number of female farmers in European countries has generally increased (Annes et al., 2021; Bjørkhaug and Blekesaune, 2008; Oedl-Wieser, 2011; Hernández-Nicolás et al., 2019; Vidickienė, 2017) and their positive impact on farm productivity has been demonstrated in several Western European countries (Tsiaousi and Partalidou, 2023; Annes et al., 2021; Hernández-Nicolás et al., 2019).
This review was conducted in accordance with the preferred reporting items for systematic reviews and meta-analyses (PRISMA) guidelines (Page et al., 2021). The spatial scope was limited to European countries and papers were referenced that used data collected not only at the farm level, but also at regional and country levels. The review was finally limited to papers examining women’s empowerment and assessing the effects of women’s empowerment and gender social involvement on the marketisation of small farms in Europe as a whole and in individual European countries. After analysing the abstracts and content, we found 17 articles that were directly or indirectly related to social embeddedness theory (Granovetter and Action, 1985), which provided the theoretical framework for the study. The reviewed papers were grouped by subject area.
The largest group comprises articles on the role of women in agriculture and their attitudes towards agricultural activities (Tsiaousi and Partalidou, 2023; Safiliou-Rothschild et al., 2007; Annes et al., 2021; Commandeur, 2005; Hernández-Nicolás et al., 2019; Knežević et al. (2017) and Gorlach et al. (2012)), as well as articles on the hierarchy of roles in agriculture (Kallioniemi and Kymäläinen, 2012; Kempster et al., 2023; Dunne et al., 2021; Shortall, 2017; Thorsen, 2017; Castelló and Romano, 2021; Nigh and González Cabañas, 2015; Haugen and Brandth, 1994). According to Tsiaousi and Partalidou (2023), women’s holdings are larger than average and female managers prioritise value-added production, seeking to increase it through informal networks. Research in France and Greece has shown that women are more likely than men to be involved in direct sales, organic farming and on-farm tourism initiatives, as well as all forms of value-added agriculture (Annes et al., 2021).
The proportion of female entrepreneurs in agriculture is higher in rural areas where farming is less mechanised and agricultural land is used more extensively (Kallioniemi and Kymäläinen, 2012). Evidence from the Netherlands and the United Kingdom shows that farm women are perceived as less professional entrepreneurs than men and are underrepresented in stereotypical ways (Kempster et al., 2023). While most farm women receive positive feedback from the wider community, some also receive negative remarks and behaviour, including actions that demonstrate a lack of faith in their capabilities and physicality. A substantial literature review by Dunne et al. (2021) reveals that hierarchical and predetermined gender roles within families and local societies persist. This conclusion is also present in the works of Shortall (2017) and Thorsen (2017). Castelló and Romano (2021) and Nigh and González Cabañas (2015) argue that female rural entrepreneurs contribute to the regional economy; however, the practical need for business support is often blurred with personal demands, particularly with respect to farm-based enterprises, such as caring roles. They are creative, implementing ethical, agroecological and cultural dimensions of food to support the smallholder family economy and seeking more direct, face-to-face relationships.
These studies emphasise the positive effect that women have on the economic conditions of farms in highly developed European countries such as France and Greece, despite their declining social importance in the agricultural sector. This impact is realised through female farmers’ involvement in direct sales, value-added production, organic farming, tourism and participation in informal organisations. However, none of these articles focused on Central and Eastern European (CEECs), raising the question of whether the same conclusions can be drawn for agriculture in these countries. This article attempts to answer that question. The differences between these groups of countries may stem from their history. In socialist countries, women formed an integral part of the agricultural workforce. They worked on state-owned farms, assisting with fieldwork, animal husbandry and processing. Socialism proclaimed the emancipation of women through work. However, women performed heavy physical labour and rarely held managerial positions on farms (Fitzpatrick, 1999). The transition from a planned to a market economy changed the model of female employment. Many state-owned enterprises that primarily employed women, such as those in the textile industry, were closed down, so women were more likely to lose their jobs than men (Kolin, 2010). Accompanying this change in many countries was a revival of traditional family values, with women returning to the roles of mother and carer (Penn and Massino, 2009). For more on this topic, see section 3.
2.2 Theoretical development and hypotheses
The review paints a picture of women farmers who, on the one hand, have a sense of self-efficacy and an innate ability to build effective relationships with customers (Annes et al., 2021; Balaine, 2019; Meyerding and Lehberger, 2018; Gidarakou et al., 2008), but who also feel that their role is undervalued compared to that of male farmers (Kempster et al., 2023; Dunne et al., 2021; Shortall, 2017; Thorsen, 2017). Thus, the review provides grounds for the hypotheses posed.
The empowerment of women has been shown to boost the development of marketisation on small farms, as evidenced by the success of women in establishing short distribution channels and cooperation networks, as well as in starting additional profitable activities (e.g. agritourism). This justifies the following hypothesis which highlights the role of women in developing value-added agriculture (see also Table A1 in Appendix):
Women empowerment triggers the capability to build social relationships, positively affecting small farm marketization
To be successful in the above activities, women need self-efficacy, which enables them to make groundbreaking decisions and manage the farm effectively. However, sole decision-making power may not be sufficient for effectively orienting a small farm towards the market.
The concept of self-efficacy and its determinants were first theorised by Bandura (1995), after which they were used interchangeably with “perceived behavioural control” (PBC) (Ajzen, 1991; Schwarzer, 1992). PBC has frequently been studied as a driver of farmers’ behavioural intentions and actions (e.g. Wang et al., 2019; Mingolla et al., 2019).
According to Bandura (1995), self-efficacy can be developed through social modelling and social persuasion. In agriculture, social modelling involves observing how other farmers have succeeded, while social persuasion encourages farmers to recognise their own importance and competence. If self-efficacy is required for women to utilise their skills, social network embeddedness is a crucial factor in enabling this.
Social network embeddedness involves interactive relationships with economic actors, such as professional associations and business networks, as well as non-economic actors, such as education systems and cultural networks. These relationships are sources of social modelling and social persuasion (Granovetter and Action, 1985; Schwabe et al., 2022; Zhang et al., 2023). This is why we propose a second hypothesis, which claims that the social network embeddedness of women is even more important for marketisation than decision-making power:
Social network embeddedness of women is more important for marketization than decision making power.
We therefore examine women’s empowerment and their participation in social networks in the context of other determinants of marketisation identified in previous studies (Barrett et al., 2012 – see Section 3.2).
3. Data and methods
3.1 Data and variables
The five countries selected for the study – Lithuania, Poland, Romania, Serbia and Moldova – were chosen for their similar agrarian structures. These countries are characterised by small family farms spanning a few hectares each. The proportion of farms spanning under 10 hectares ranges from approximately 50% in Lithuania to over 90% in Romania (Eurostat, 2022a, b). This historical arrangement is related to the operation of centrally planned economies after the Second World War and the subsequent transition from the early 1990s onwards. The marketisation of the agricultural sector has created a dual system comprising a small number of large, strong enterprises and a substantial number of small-scale family farms (Stępień and Maican, 2020). For the latter, market orientation is crucial for improving economic health and guaranteeing sustainable rural development.
At the same time, it is interesting to compare the results for the selected countries with those for the highly developed countries quoted earlier. The specific historical, social and cultural conditions in the states of Central and Eastern Europe lead us to assume that the context of women’s empowerment and their participation in social networks differ in certain areas. Firstly, the patriarchal family model has long determined the role of women in these countries (Cernea, 1978). This role involves giving birth and raising children, as well as running the household. Activity beyond these areas is not widely accepted as the norm. This perception is perpetuated by the influence of Orthodox Christianity in Serbia, Moldova and Romania and of Catholicism in Poland and Lithuania. Secondly, women farm owners in the analysed countries tend to have low levels of participation in political and business life (e.g. village boards, municipal boards, cooperative boards and local and national government bodies). This primarily occurs in Romania (Burtea, 1999), Serbia (FAO, 2021) and Moldova (Rocca et al., 2013). It can be concluded that the rural lifestyle of women has not changed significantly for decades, with traditional patterns of how free time is spent remaining in place. Most of this time is spent resting in order to continue the physical efforts related to work at home and on the farm.
The third distinguishing feature is women’s participation in education systems. Rocca et al. (2013) found that increasing women’s education levels empowers them. Meanwhile, Moldova, Serbia and Romania have a visible gender gap in education: women are less educated and have lower rates of access to educational institutions. This may contribute to lower management competence and knowledge in these countries. Romania is the worst-performing EU country in terms of women’s entrepreneurship, particularly among those in poorer households (Robayo-Abril et al., 2023). Conversely, in Poland and Lithuania, women on farms tend to be highly educated, which can have positive effects on economic resilience, propensity to innovation, environmental awareness and management skills and cooperation with the surrounding market (Baležentis et al., 2021).
Information from the farms was obtained through face-to-face interviews using a semi-structured questionnaire. As the aim was to assess women’s empowerment and the extent to which smallholder farms are integrated into the market, two sampling criteria were employed: an agricultural area of up to 20 hectares and an economic size of up to 25,000 euros, as defined by the farm accountancy data network (FADN) methodology. Standard output (SO) is the average monetary value of an agricultural product’s annual output at farm gate prices, expressed in euros per hectare or per head of livestock.
The surveys were conducted by agricultural consultants in 2019. The sample size was 1,000 farms in Lithuania, 900 in Romania, 710 in Poland and 550 in both Serbia and Moldova. After eliminating erroneous and incomplete observations, outliers and farms with no female respondents, 3,172 interviews were included in the analysis. The distribution for each country is presented in Table 1. Basic descriptive statistics for the surveyed households, including UAA in hectares, production type (mixed type 3 or horticulture and permanent crops type 4), work input in annual work units (AWU), distance to the nearest city in kilometres and farmer age, are also included.
Descriptive statistics
| Lithuania | Poland | Romania | Serbia | Moldova | |||||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Variable | N | Meana | SD | Min | Max | N | Mean | SD | Min | Max | N | Mean | SD | Min | Max | N | Mean | SD | Min | Max | N | Mean | SD | Min | Max |
| WoEm1 | 875 | 0.46 | – | 0.0 | 1.0 | 531 | 0.73 | – | 0.0 | 1.0 | 739 | 0.69 | – | 0 | 1 | 505 | 0.82 | – | 0.0 | 1.0 | 522 | 0.58 | – | 0.0 | 1.0 |
| WoEm2 | 875 | 0.05 | – | 0.0 | 1.0 | 531 | 0.09 | – | 0.0 | 1.0 | 739 | 0.04 | – | 0.0 | 1.0 | 505 | 0.02 | – | 0.0 | 1.0 | 522 | 0.17 | – | 0.0 | 1.0 |
| WoEm3 | 875 | 0.25 | – | 0.0 | 1.0 | 531 | 0.14 | – | 0.0 | 1.0 | 739 | 0.19 | – | 0.0 | 1.0 | 505 | 0.09 | – | 0.0 | 1.0 | 522 | 0.11 | – | 0.0 | 1.0 |
| WoEm4 | 875 | 0.24 | – | 0.0 | 1.0 | 531 | 0.05 | – | 0.0 | 1.0 | 739 | 0.08 | – | 0.0 | 1.0 | 505 | 0.07 | – | 0.0 | 1.0 | 522 | 0.14 | – | 0.0 | 1.0 |
| Market Orient | 875 | 0.45 | 0.21 | 0.0 | 1.0 | 531 | 0.49 | 0.14 | 0.2 | 1.0 | 739 | 0.62 | 0.24 | 0.0 | 1.0 | 505 | 0.54 | 0.22 | 0.0 | 1.0 | 522 | 0.47 | 0.19 | 0.0 | 1.0 |
| AWU farm | 875 | 1.10 | 0.64 | 0.1 | 3.8 | 531 | 1.52 | 0.66 | 0.1 | 4.0 | 739 | 1.44 | 0.70 | 0.2 | 7.5 | 505 | 1.67 | 0.86 | 0.3 | 5.0 | 522 | 1.53 | 0.98 | 0.1 | 12.6 |
| Farm area (ha)b | 875 | 10.78 | 5.93 | 1.0 | 20.0 | 531 | 13.10 | 6.31 | 1.5 | 39.0 | 739 | 8.62 | 11.12 | 0.0 | 127.1 | 505 | 4.30 | 2.79 | 0.0 | 20.0 | 521 | 5.26 | 3.03 | 0.0 | 15.9 |
| Distance to city (km) | 875 | 11.78 | 5.91 | 1.0 | 35.0 | 531 | 11.70 | 7.42 | 0.0 | 54.0 | 739 | 27.40 | 22.99 | 0.0 | 106.0 | 505 | 17.97 | 8.96 | 0.0 | 50.0 | 522 | 18.75 | 10.72 | 0.0 | 60.0 |
| Age of farmer | 875 | 47.60 | 13.72 | 19.0 | 77.0 | 531 | 48.42 | 11 | 23.0 | 67.0 | 739 | 46.74 | 12.54 | 20.0 | 81.0 | 505 | 54.24 | 13.14 | 21.0 | 95.0 | 522 | 46.52 | 13.62 | 20.0 | 77.0 |
| dType3 | 875 | 0.61 | – | 0.0 | 1.0 | 531 | 0.42 | – | 0.0 | 1.0 | 739 | 0.42 | – | 0.0 | 1.0 | 505 | 0.52 | – | 0.0 | 1.0 | 522 | 0.24 | – | 0.0 | 1.0 |
| dType4 | 875 | 0.14 | – | 0.0 | 1.0 | 531 | 0.09 | – | 0.0 | 1.0 | 739 | 0.17 | – | 0.0 | 1.0 | 505 | 0.19 | – | 0.0 | 1.0 | 522 | 0.53 | – | 0.0 | 1.0 |
| Female farm manager | |||||||||||||||||||||||||
| SocPart | 430 | 0.50 | – | 0.0 | 1.0 | 98 | 0.24 | – | 0.0 | 1.0 | 200 | 0.31 | – | 0.0 | 1.0 | 79 | 0.44 | – | 0.0 | 1.0 | 132 | 0.55 | - | 0.0 | 1.0 |
| Male farm manager | |||||||||||||||||||||||||
| SocPart | 445 | 0.47 | – | 0.0 | 1.0 | 433 | 0.206 | – | 0.0 | 1.0 | 539 | 0.33 | – | 0.0 | 1.0 | 426 | 0.41 | – | 0.0 | 1.0 | 390 | 0.70 | - | 0.0 | 1.0 |
Means for categorical variables indicate the share
Total farm area includes rented land or municipal area available for common use (mainly grassland)
The research design is presented in Figure 1. Four types of women’s empowerment (WoEm1–WoEm4) were identified based on decision-making power and social network embeddedness. The first type (WoEm1) includes women who are not managers and therefore do not have decision-making power when it comes to running a farm. At the same time, their level of participation in social networks in terms of lifelong education, training, social events and organisation membership is low (i.e. it was reported that women do not participate in such activities at all). Participation in lifelong education and training provides an opportunity to meet new people, establish contacts and exchange views and opinions; therefore, it is a form of socialisation for women.
The organizational chart has three main stages, each represented by a large rectangular box. At the top level, there is a large box with the title “First stage — women’s empowerment levels.” This box is divided into four smaller, numbered text boxes. The first text box, labeled “1,” reads “A woman does not manage the farm and is not socially involved in educational and training networks.” The second text box, labeled “2,” reads “A woman does not manage the farm but is socially embedded in educational and training networks.” The third text box, labeled “3,” reads “A woman manages the farm, but is not socially involved; she rarely participates in professional training or represents the farm in organizations or at local business or cultural events.” The fourth text box, labeled “4,” reads “A woman manages the farm and is highly socially embedded; she regularly participates in professional training, represents the farm in organizations and at local events.” A downward-pointing arrow labeled “treatment effect on” extends from this top-level box. This arrow points to the second level, which is a large box titled “farm market-orientation proxied by.” This box is divided into three text boxes that are positioned horizontally in a series. The first box reads “Share of agricultural income in household income.” The second box reads “Share of commercial production in total production value.” The third box reads “Share of commercial production sold by short channels (direct sales, local market and shops).” An upward-pointing arrow labeled “treatment effect on” connects the bottom stage to this middle stage. The third and final level has a large box titled “Second stage — gender participation in social networks of.” Below this title, two more text boxes are present side by side. The left text box reads “woman manager,” and the right text box reads “man manager.”Research design. Source: Authors’ own work
The organizational chart has three main stages, each represented by a large rectangular box. At the top level, there is a large box with the title “First stage — women’s empowerment levels.” This box is divided into four smaller, numbered text boxes. The first text box, labeled “1,” reads “A woman does not manage the farm and is not socially involved in educational and training networks.” The second text box, labeled “2,” reads “A woman does not manage the farm but is socially embedded in educational and training networks.” The third text box, labeled “3,” reads “A woman manages the farm, but is not socially involved; she rarely participates in professional training or represents the farm in organizations or at local business or cultural events.” The fourth text box, labeled “4,” reads “A woman manages the farm and is highly socially embedded; she regularly participates in professional training, represents the farm in organizations and at local events.” A downward-pointing arrow labeled “treatment effect on” extends from this top-level box. This arrow points to the second level, which is a large box titled “farm market-orientation proxied by.” This box is divided into three text boxes that are positioned horizontally in a series. The first box reads “Share of agricultural income in household income.” The second box reads “Share of commercial production in total production value.” The third box reads “Share of commercial production sold by short channels (direct sales, local market and shops).” An upward-pointing arrow labeled “treatment effect on” connects the bottom stage to this middle stage. The third and final level has a large box titled “Second stage — gender participation in social networks of.” Below this title, two more text boxes are present side by side. The left text box reads “woman manager,” and the right text box reads “man manager.”Research design. Source: Authors’ own work
The second category (WoEm2) includes farms managed by men, where women are engaged in social networks. It was reported that women regularly participate in continued education systems, training courses, universities for the third age, social events and/or organisations and associations.
The third (WoEm3) includes farms managed by women with decision-making power who are minimally socially involved. Despite having the formal opportunity to represent the farm in professional organisations, producer groups, cooperatives, at local business or cultural events, etc., it was reported that they rarely take it. We decided to identify this level of empowerment if no more than one of the following three types of social involvement was reported: (1) participation in training, courses and lifelong education; (2) participation in cultural events; and (3) participation in professional organisations.
Finally, the fourth category (WoEm4) includes farms managed by women who are highly socially involved. We identified this level of empowerment if at least two of the three aforementioned options were chosen.
Therefore, the different types of women’s empowerment are characterised by the extent of women’s decision-making power and the scale of their participation (embeddedness) in social networks. Consequently, women’s empowerment was coded as a four-level dummy variable. In the modelling, the lowest level of women’s empowerment (WoEm1), where women have neither power nor embeddedness in social networks, was set as the baseline. The first stage of the analysis thus enables us to test H1, which states that women’s empowerment triggers capabilities that positively affect the marketisation of small farms by enabling them to build social relationships.
Modelling results confirm H1 when WoEm2, WoEm3 or WoEm4 have a positive impact on marketisation, with WoEm1 as the reference point.
However, the first stage does not conclusively answer whether the positive impact on marketisation is due to decision-making power or participation in social networks, since both are components of women’s empowerment. Therefore, we tested H2 in the second stage.
Modelling results that confirm H2 would be as follows: WoEm2 or WomEm4 positively impact martketization index and the effect is stronger than in case of WoEm3 (or there is the negative effect of WoEm3); and SocPart of woman manager positively impacts the marketization index.
Table 1 contains information about the proportion of each group of farms considered part of each category of women’s empowerment, together with an indication of participation in social networks (SocPart).
To isolate the effect of gender participation in social networks (PSN) on decision-making power, we analysed the PSN treatment effect separately for female- and male-led farms. We assumed that a farm was embedded in social networks if at least two out of three of the following conditions were met: (1) participation in training, courses and lifelong education; (2) participation in cultural events; and (3) participation in professional organisations.
However, it was not possible to compare male and female participation in social networks on any given farm, as the farm manager is always more involved in social networks by virtue of their function, such as having a formal opportunity to represent the farm in professional organisations.
Data in Table 1 reveals that Lithuania has the highest proportion of women involved in farm management and a high level of female socialisation. However, it also has the lowest level of marketisation. This raises the question of what is causing this phenomenon. It is worth noting that, alongside Latvia, Lithuania is among the EU countries with the highest proportion of female farm managers (approximately 45% compared to an EU average of 32%; Eurostat, 2020). This may be due to the prioritisation of women-managed farms in national policy and legislation, as well as in initiatives and programmes undertaken by the Lithuanian government and NGOs. The European Union has also supported rural development in Lithuania in previous years (Šukyte and Vaičiūniene, 2009). This is reflected in our study, which shows a relatively high percentage of cases from WoEm3 and WoEm4. However, as Baležentis et al. (2024) argue, these figures reflect legal ownership rather than actual farming activities. In practice, farms managed by women tend to be significantly smaller in area and have a lower production value. These constraints make it more difficult to achieve economies of scale and are also an obstacle to marketisation, given that a significant proportion of food production is for subsistence (hence the lower share of income and output sold from agriculture, i.e. the variables that determine the marketisation rate).
3.2 Socio-economic proxies of market orientation in small-scale farming
Several studies in the literature have examined the determinants of market orientation and participation. In a recent article, Zhang et al. (2020) wrote: “Market orientation in agriculture can be defined as the degree to which resources (land, labour and capital) are allocated to the production, exchange and sale of agricultural products”. In market-oriented farming (MOF), farms are managed as business operations”. Thus, market orientation focuses on the resource perspective, whereas market participation is usually perceived from the perspective of product distribution and customer relations. Nevertheless, the two concepts are inseparable and successful market orientation requires resources to be allocated according to market criteria, as well as products to be distributed by market channels that consider consumer preferences (Stępień et al., 2022). We therefore argue that a market-orientation indicator should cover both perspectives (Figure 1). Furthermore, the indicator should reflect the distance from farmer to final consumer in distribution channels: the closer to the consumer, the better adjusted the product is to market preferences.
The market-orientation index was based on three elements: (1) the proportion of a farm’s total income derived from agriculture, reflecting resource allocation according to market criteria; (2) the proportion of a farm’s total agricultural production value derived from market sales (a general market participation measure); (3) the proportion of market sales distributed through short supply channels (a direct market participation measure). Therefore, a higher proportion of total household income from agriculture, greater market participation and shorter market distances all increase the market-orientation indicator (see Figure 1). To construct the indicator, farms were asked three questions: (1) What proportion of total household income comes from agriculture? (2) What proportion of total agricultural production is sold? (3) What proportion of sales are made through short distribution channels? This question was supplemented with an explanation related to the definition of short selling. Respondents answered each question by estimating the above shares as a percentage. These three specific numbers could then be transformed into one synthetic measure.
We therefore computed a composite index of three variables using the CRITIC-TOPSIS method (Deng et al., 2000; Czyżewski et al., 2018), which determines the weights of the variables included in the synthetic measure. This approach eliminated the accusation of subjective weight assignment to the three criteria. CRITIC-TOPSIS gives greater weighting to characteristics with higher variability and lower correlations with the other components. The components are standardised using the zero-unitarisation method. The market orientation synthetic measure (MO) for farm i is then calculated using the following formula:
where u + represents the maximum and u- represents the minimum of the weighted and zero-unitarised j-characteristics :
The socio-economic control variables that are included as determinants of market orientation are listed below (for descriptive statistics see Table 1.):
Farm area (ha);
Work engaged (in all work units AWU);
Specialisation in permanent crops – dType4, e.g. orchards, vineyards (share of standard output ≥2/3);
Specialisation in mixed farming – dType3 (standard output share of animal production ≥1/3 and ≤2/3; plant production share ≥1/3 and ≤2/3);
Distance to city (in km);
Age;
The effects of education and standard output on the market orientation index were also tested, but were found to be statistically insignificant.
3.3 Propensity score as a tool for analysing the causal effect of women’s empowerment and gender participation in social networks
Propensity-based methodologies have recently been used to estimate the effect of qualitative variables (e.g. Ma et al., 2020). Mathematical proof shows that controlling a probability P(V) can produce the same results as measuring V directly. Thus, a propensity score (PS) can be defined as a function that converts variables into a treatment probability (Rosenbaum, 2002). However, there are two key considerations for effective propensity-based analysis: control variables with significant effects on the outcome and overlapping distributions of treated and untreated objects to balance the groups.
In the current study, AIPW estimators with four levels for WoEm (multinomial logit model) and two levels for SocPart (logit model) were used (referring to Cattaneo, 2010) and the average treatment effects (ATEs) in the population were estimated. This approach enables casual inference with regard to multi-category variables.
As a basic inverse propensity weighting ATE estimator has been proven to have poor small-sample properties, we used the improved AIPW approach suggested by Glynn and Quinn (2010):
where is a dataset of observable controlling variables, represents the outcome variable (0 or 1), is a treatment variable and is the calculated PS (Stata, 2013).
3.4 Model specification, robustness and balance
We used the same initial set of variables summarised in Table 1 and Figure 1 to predict the outcome (the market-orientation index) and the treatment status (whether the outcome was women’s empowerment or gender social involvement). These variables were selected based on a review of the literature (see Appendix, Tables A1–A4) and their effect on the market orientation index was tested using fractional regression models.
Using a stepwise procedure, we eliminated the variables with insignificant effects from the respective outcome models. The data indicated that market participation is most influenced by labour input (AWU) and the type of horticulture and permanent crops. In both cases, the relationship was positive and statistically significant for all countries except Serbia, where it was positive but not statistically significant. Type 3 of land use (i.e. mixed farms) was positively correlated in all countries except Poland. In Poland, however, mixed farms are different from those in other countries. They are mainly used for crop production, such as cereals, corn, oilseeds and industrial plants. Direct sales of these crops are not typical. By contrast, in other countries, mixed farms are involved in both plant and animal production, including the cultivation of vegetables and fruit and the production of milk and meat, i.e. products suitable for sale through short supply chains. The relationship between market participation and distance to the city was statistically significant for all countries. For Lithuania, Romania and Serbia, the relationship was positive (i.e. the closer to the city, the stronger the marketisation), while for Poland and Moldova, the relationship was negative. The positive direction of influence is easy to understand: proximity to cities provides greater opportunities for selling products through short supply chains, hence the higher level of marketisation. In Poland, the inverse relationship is the result of farms located close to cities earning a relatively large part of their income from work in the city rather than agriculture. Here, we observe a process of disappearance of the traditional role of farms. In Moldova, the situation arises because a significant proportion of the farms observed are vineyards. These farms are located far from the city, but are relatively highly commercialised and sell their products directly to customers. For the PS models, only those variables that met the restrictive balance conditions were retained (see Appendix, Tables A5–A6).
For two-level modelling, we employed the balance test outlined by Imai and Ratkovic (2014). In the multi-level AIPW model, we examined the weighted differences between the control and treatment groups, as well as the variance ratio, in accordance with the methodology outlined by Linden and Samuels (2013). All the weighted differences were within the recommended threshold of less than 0.1 and the variance ratio was between 0.5 and 2.0. This shows that the AIPW model sufficiently balanced the covariates in each country’s sample.
As with all research, our approach may have limitations. There is a risk of omitted variables bias in the outcome and treatment models, which may be the result of the specific context of each country. To ensure comparability of results, we use a similar initial set of variables. However, this approach may neglect some country-specific factors of marketisation.
4. Estimates of treatment effects
Table 2 shows the effect of each category of women’s empowerment on market participation (ATE). The reference category is Category 1, where the woman is neither a manager nor involved in social networks (e.g. training, events or organisations). The other three categories are estimated against this reference category (as in Figure 1). The table also includes information on how women’s and men’s participation in social networks translates into marketisation.
Average treatment effects of women’s empowerment and gender social participation
| Lithuania | Poland | Romania | Serbia | Moldova | ||||||
|---|---|---|---|---|---|---|---|---|---|---|
| ATE | Coef | p > z | Coef | p > z | Coef | p > z | Coef | p > z | Coef | p > z |
| WoEm (2 vs 1) | 0.107 | 0.000*** | −0.014 | 0.468 | −0.030 | 0.300 | 0.106 | 0.026** | −0.040 | 0.060* |
| WoEm (3 vs 1) | −0.012 | 0.489 | −0.036 | 0.007* | 0.040 | 0.075* | 0.089 | 0.000*** | −0.095 | 0.001*** |
| WoEm (4 vs 1) | 0.031 | 0.074* | 0.054 | 0.051* | −0.025 | 0.394 | 0.094 | 0.011** | −0.020 | 0.453 |
| Female farm manager | ||||||||||
| SocPart | 0.043 | 0.039** | 0.093 | 0.001*** | −0.046 | 0.179 | 0.003 | 0.956 | 0.080 | 0.028** |
| Male farm manager | ||||||||||
| SocPart | 0.026 | 0.161 | 0.007 | 0.606 | −0.079 | 0.000* | −0.010 | 0.659 | 0.015 | 0.488 |
| H1 | accepted | accepted | accepted | accepted | rejected | |||||
| H2 | accepted | accepted | rejected | rejected | accepted | |||||
Note(s): *,**,*** respectively significant effects at the level α = 0.1, α = 0.05, α = 0.01
The estimate shows that:
in Lithuania and Serbia, WoEm Category 2 (where a man manages the farm, but women are socially involved) has a market participation rate that is approximately 10% higher than Category 1. However, in Moldova, the relationship is negative.
A comparison of WoEm Category 3 (where a woman manages the farm but is minimally socially involved) with Category 1 yields mixed results: for Poland and Moldova, the relationship is negative, while for Romania and Serbia, it is positive.
Comparing Category 4 (farms managed by women who are highly socially involved) with Category 1 shows a positive relationship in Lithuania, Poland and Serbia, but the relationship is statistically insignificant in Romania and Moldova. Farm management by a woman who is also highly socially involved translates into a market participation increase of almost 10% in Serbia, around 5% in Poland and around 3% in Lithuania.
Further observations show that higher female participation in social networks increases market participation in Lithuania, Poland and Moldova (in Romania and Serbia, the relationship is statistically insignificant), while a significant positive relationship was only found for men in Romania.
5. Discussion
The first estimation shows that in Lithuania and Serbia, on farms where men are the managers but women are socially involved, the market participation rate is higher (by approximately 10%) than on farms where women are neither managers nor involved in social networks. In our opinion, this may be due to women’s greater involvement in improving farming practices. Women also often initiate decisions about selling through short channels. This is consistent with the findings of Annes et al. (2021). They found that women were more likely than men to be involved in direct sales, organic farming and the production of traditional local agri-food products. The sale of organic, traditional and local products most often occurs directly, eliminating middlemen and therefore utilising short supply chains. This is consistent with the findings of Tsiaousi and Partalidou (2023) and Annes et al. (2021), who argue that women are more likely than men to engage in direct sales and all forms of value-added agriculture. They are also more open to vocational training, technological expertise and union membership (Haugen and Brandth, 1994). Women add new activities to their regular tasks, such as selling farm products and getting involved in tourism (Commandeur, 2005).
In contrast, the impact was negative in Moldova, which is surprising. Previous studies have suggested that women in this country face discrimination and inequality in social and economic life and have limited opportunities to participate in decision-making processes (Ungureanu, 2023). Women living in rural Moldova are at high risk of poverty (FAO, 2022) and their involvement in social life does not improve farm income; in fact, it may even worsen it. This is all the more concerning given that they are generally older women (over 30% are over 65 years of age and almost 60% are aged 55+), with a very low level of education (87% have no agricultural education, 6% have vocational education and even fewer have secondary or higher education). No other country surveyed has such a clear education gap or such a high percentage of older women in agriculture (National Bureau of Statistics, 2014). Furthermore, the survey results by Gorlach et al. (2012) indicate that economically disadvantaged farms are run by older women. While this research focused on Poland, it does provide a basis for the claim that Poland and Moldova converge in this regard. Surveys from France, Greece and Ireland (Balaine, 2019; Gidarakou et al., 2008; Meyerding and Lehberger, 2018; Annes et al., 2021), as well as many other researchers who have indicated that men occupy the hegemonic position of farmer and women are subordinated to the role of farmer’s wife (Dunne et al., 2021; Shortall, 2017; Bjørkhaug and Blekesaune, 2008; Thorsen, 2017).
A review of the literature reveals that, although more women now have the opportunity to work in agriculture, many of them still face public scepticism and discrimination. This often leads to self-doubt, insecurity and isolation (Annes and Wright, 2015; Kempster et al., 2023). Our findings show that gender-role hierarchisation is more intense in Moldova than in the other analysed countries. However, changes in this regard are a long-term process. Nevertheless, the results of the research suggest that in some European countries, such as Lithuania, Austria, Norway, France and Spain, the proportion of women among farm managers is gradually increasing (Bjørkhaug and Blekesaune, 2008; Oedl-Wieser, 2011; Hernández-Nicolás et al., 2019; Vidickienė, 2017). Next, we compared farms where women are managers but minimally socially involved, with farms where women are not managers and not involved in social networks. The results were ambivalent: for Poland and Moldova, the relationship was negative, while for Romania and Serbia it was positive. This proves that even if a farm is managed by a woman, this is not a sufficient factor to achieve higher market participation. Furthermore, as previously mentioned, women in Moldovan agriculture are poorly educated and primarily manage underdeveloped farms without access to technology or financial resources, resulting in low production potential. What is produced is largely consumed on the farm and there are limited contacts with the market. In Poland, some women manage farms because they inherited property from a deceased husband. If these women are neither ambitious nor enterprising nor socially engaged, the financial condition of such a farm may be poor and the level of commercialisation low.
Significant differences in conclusions emerge when comparing farms managed by highly socially involved women with farms where women are neither managers nor involved in social networks, including training sessions, events and organisations. The strong socialisation of women in relation to their management of farms markedly increases market participation. Therefore, greater involvement in community life and participation in organisations or training by women appears to have a greater impact on the marketisation of farms than the same activities by men. Literature confirms that greater representation of women in management positions in agricultural entities (such as farms and cooperatives) leads to higher profits, lower operational risk and less debt (Hernández-Nicolás et al., 2019). Similarly, the presence of women in the management of an agricultural cooperative results in better decision-making and a different approach to risk, as well as bringing more diverse skills to the table, although this relationship may be weaker (Knežević et al., 2017). It should also be noted that women’s greater involvement in society stems from their needs; women farm workers have the same needs as other women with regard to earnings, financial independence, social contact, status and self-realisation (Gasson, 1992). According to the results of our research, this also favours the marketisation of small farms in most CEECs.
These findings emphasise the importance of including active women farmers in ongoing development programmes that aim to boost farmers’ competitiveness. This is consistent with previous reports on agriculture in Catalonia (Safiliou-Rothschild et al., 2007) and Greece, where young female farmers are engaged in agriculture, optimistic about farming and have seized opportunities to develop and become professional farmers (Gidarakou et al., 2008). An additional benefit of increasing women’s involvement in agriculture is that female rural entrepreneurs contribute to the regional economy. Farm women create alternative food networks — a market-oriented activity — that implement the ethical, agroecological and cultural dimensions of food. This supports the smallholder family economy and seeks more direct, face-to-face market relationships (Nigh and González Cabañas, 2015).
6. Conclusions
Times change, as does the role of women in farm management. Women are increasingly playing an active role in the management process, forging strong social bonds that unite the rural population. Women’s empowerment requires a certain level of self-efficacy; in other words, women must perceive a certain degree of behavioural control. This perception is primarily influenced by social participation, as Bandura (1995) suggested. We demonstrate that women’s empowerment has a stronger impact on farm marketisation when combined with social participation. We isolate and confirm the effect of the latter, which allows us to accept H2 in the majority of countries under study. Therefore, we bridge the concept of self-efficacy with the theory of social embeddedness (Granovetter and Action, 1985), arguing that sole decision-making power is not sufficient for effective small-farm management unless it is supported by interactive relationships with social networks.
Studies in the literature have shown that female farm management has a positive impact on farming efficiency and added value. This is due to women’s greater willingness to participate in direct sales, have closer contact with customers, create alternative sales channels (e.g. online) and develop food processing. In other words, women are more likely to increase the marketisation of a farm.
Decision-making power and participation in social networks are two interrelated attributes of women’s empowerment. Our research shows that participation in social networks is crucial when investigating the effect of women’s empowerment on the commercialisation of small farms. Being “embedded” in a social network catalyses the ability to build social relationships and thus helps to commercialise smallholdings. From an academic perspective, further research is needed to explain how social embeddedness affects women’s empowerment. We argue that participation in social networks builds the self-efficacy of individuals through social persuasion and modelling. Self-efficacy can enhance the effectiveness of decision-making and thus the impact of women’s empowerment. Nevertheless, this is a preliminary step that opens up a new avenue for further research.
The research we conducted has yielded a message for practitioners, especially policymakers. The message is that agricultural policy should support processes of horizontal integration, particularly in the small-scale agriculture sector. This could be achieved by creating associations and producer groups, as well as various innovative forms of cooperatives. These include joint purchasing and use of agricultural equipment, collective sales of products and the sharing of buildings. Secondly, it should ensure full equality of access to and membership of, such organisations and their boards for women. Thirdly, it should support various types of cultural initiatives in rural areas to preserve cultural heritage. Fourthly, it should reform the education and vocational training system to encourage positive social modelling. Farmers, especially women, should be able to see and learn from successful, thriving small farms, which play an important and recognised role in modern society. This would replace the current dominant image in CEECs of small units as hopeless and having no future. This new image of small-scale agriculture should emphasise its growing role in building local food security, providing environmental public goods, caring for the soil and landscape and preserving traditions and cultural heritage.
However, gender-based role categorisation and differing perceptions of women’s and men’s positions remain common. In countries with a more patriarchal social structure, for example, female farm management does not offer the same opportunities as in countries with a more equal social structure. In this context, it is important to promote good practices and highlight the economic and social benefits of empowering women.
The authors are aware that the research presented in the article is not without its limitations, which relate to the use of cross-sectional data from surveys. Furthermore, although the research includes data from five countries (Lithuania, Romania, Poland, Serbia and Moldova), these countries vary widely in terms of culture, economy and rural structures. Therefore, as explained in the methods section, some omitted variable bias may occur. For this reason, our findings should be generalised to other CEECs or countries outside the region with caution.
CRediT author statement
Bazyli Czyżewski: conceptualization; data curation; methodology; investigation; formal analysis; supervision; funding acquisition Marta Guth: conceptualization, methodology; literature review; discussion; writing – original draft; writing – review and editing Anna Matuszczak: conceptualization; methodology; literature review; discussion; writing – original draft Katarzyna Smędzik-Ambroży: conceptualization; methodology; formal analysis; discussion; writing – original draft Sebastian Stępień: conceptualization; methodology; investigation; data acquisition; data curation; writing – original draft; writing – review and editing; funding acquisition.
The supplementary material for this article can be found online.
