The impact of survival-entrepreneurship on female domestic autonomy in Kisumu, Kenya

(1)

The Impact of Survival-Entrepreneurship on Female

Domestic Autonomy in Kisumu, Kenya

Emma Kaatje Sofie Lucas Student number: 11786493

Graduate School of Social Sciences (UvA) Research MSc International Development Studies

2017-2019

Supervisor: dr. Nicky Pouw Email: e.k.s.lucas@outlook.com

Date: 15th of August 2019 Word Count: 29 990 words 

(2)

(3)

Acknowledgments

This thesis would not have been possible without the excellent support of my supervisor: Nicky Pouw. With her many critical notes she pushed me further in my writing and to improve the quality of this thesis. I will never again forget to write conclusions at the end of chapters. Further thanks goes to Eunice Likoko and Prof. J.B. Okeyo-Owuor from VIRED International in Kisumu, who guided me through Kisumu and research there and without whom my time in Kisumu would not have been lovely as it was. Thank you to Dr. Hannington Odame and Elsie Kangai from the Centre for African Bio- entrepreneurship (CABE) in Nairobi who welcomed me in Kenya and helped me at the start of my research project.

I am also grateful to my friends and family who helped me to digitalize the surveys when I felt it was a task I would never finish: Liesbeth, Hannah, John, Lisa, Roel, Adrienne, Senna: thank you! Both Lisa and Hannah asked me to note that they were fastest in typing in the questionnaires, and I do not want to pick sides, so special thanks to both of them. Without the help of these people, it is very likely that I would still be looking at the huge stack of papers that I brought back from Kisumu. More thanks to Liesbeth, Hannah, Lisa, Marieke and Senna for proofreading some or all chapters and being sparring partners when I was totally stuck. Also thank you to Lisa and Devin for allowing me to try out some different methods to talk about their relationship when I was writing my methodology in my proposal: you were the best guinea-pigs!

A special shout-out to my mother, who is probably the reason that I am so interested in the many ways in which women are empowered or not and continues to be a role-model. She was also invaluable to overcome (some of) the difficulties I faced putting all of my thoughts on paper. Thank you to both of my parents who let me stay in their house and brought me coffee when I felt a little hopeless about the process.

Lastly and most importantly, my gratitude goes out to all the women who participated in my research. Whether it was casual conversation where people told me about Kenyan culture and my (many) misconceptions about the way my questions should be phrased, or all women who took the times out their very busy days to partake in the surveys and focus-groups discussions: thank you very much, my research would not have been possible without your help. 

(4)

(5)

Abstract

Economic empowerment has been deemed the ‘magic bullet’ to increase female autonomy within development policy for decades. Existing research has addressed the relationship between work, specifically the effects of increased material wealth, and the decision-making power of women in their households. However, little research has been done on the causal mechanism that explain this relationship and how this is effect by the type of work. This study seeks to remedy this by answering the following question: “How does work impact female decision-making power in the household in Kisumu, Kenya”. Using an explanatory mixed-method design, data was gathered using surveys (n=275), focus groups and stakeholder-interviews. The research concludes that female survival-entrepreneurs are more likely to make decisions and be involved in decision-making processes in their household. Income, and control over income, allows women to make decisions without interference of their partner. Respect and status of women does not come from actual financial contributions, but rather the perceived contribution and general respect for those who work. Moreover, the research finds that decision-making should be seen as a process. It identifies the strategies women use to influence the different phases of decision making, which create a better understanding of the causal mechanisms of the relationship between paid work and female domestic autonomy. Based on these findings, I recommend a reconceptualization of autonomy and inclusion of information sharing in future research into the relationship.

Keywords: Bargaining Power, Domestic Autonomy, Economic Empowerment, Gender,

(6)

(7)

Tables

Figures

Table 1. Frequency table who makes decisions (women with partner) (n = 200) 37 Table 2. Frequency table who makes decisions (women without partner) (n = 75) 38

Table 3. Regression Model Daily Income (n=275) 39

Table 4. Regression Interaction Model Daily Income (n=275) 42

Table 5. Regression Type of Work (n=275) 45

Table 6. Regression Complete: Combination of Daily Income and Type of Work (n=275) 47

Table 7. Regression Financial decisions (n=275) 51

Table 8. Operationalization Table Domestic Autonomy 88

Table 9. Categories type of work 106

Table 10. All Other Variables Used in the Regression 106

Table 11. Frequency-table of explanations household-spending 107 Table 12. Frequency table of explanations work of respondent 107 Table 13. Frequency table of explanations production on farm 107

Table 14. Frequency table of explanations eating 108

Table 15. Frequency table explanations buying and selling 109

Table 16. Frequency-table of explanations chores 109

Table 17. Frequency table number of children 110

Table 18. Frequency table living location 110

Table 19. Frequency table raising children 111

Figure 1. Conceptual model 17

Figure 2. Phases of Decision-making in the Household 56

Figure 3. Adapted Conceptual Model 73

Map 1. Obunga Sampling Map 89

Map 2. Nyalenda Sampling Map 89

(11)

Chapter 1. Introduction

“We can already see that the emancipation of women and their equality with men are impossible and remain so, as long as women are excluded from social

production and restricted to domestic labor. The emancipation of women becomes

feasible only then when women are enabled to take part extensively in social

production, and when domestic duties require their attention in a minor degree.” (Engels, 1884: p. 196)

1.1. Background and Problem Statement

In 1884, Engels already argued that women need access to paid work if they were to be free from the survived forced upon them by the ‘patriarchal family structure’ (Anderson & Eswaran, 2009; Engels, 1884). Today, paid work and economic empowerment are still seen as a vital strategy to increase female autonomy in high-income and low-income countries alike. Female economic empowerment is a focus point in the strategy of the World Bank, both in their own policies and their influenced on the agenda of the broader development agenda (Bebbington, Lewis, Batterbury, Olson, & Siddiqi, 2007; Klugman et al., 2014; Vaz, Pratly, & Alkire, 2016). The policy focus on economic empowerment also explains the popularity of micro-credit programs, which are believed to provide not only economic empowerment but also general female autonomy through financial literacy and their own economic initiatives (Amin, Becker, & Bates, 1998).

These policies are based on the hypothesis that paid work leads to an increase in female autonomy. The argument behind this hypothesis is that paid work improves the bargaining position of women in their household, their community and in their workplace. An improved bargaining positions means women can make more decisions, and therefore increases their autonomy (Anderson & Eswaran, 2009; Meurs & Ismaylov, 2019). However, empirical studies do not present an uniform conclusion and the relationship between paid work and female autonomy changes in different contexts (Kabeer, 2011). Moreover, it is simplistic to assume all forms of paid work to be the same and to fit within a single hypothesis. The circumstances and consequences of paid work can be very different and some forms of paid work are very disempowering, such as forced prostitution, exploitative labour or slavery.

(12)

This study contributes to three gaps in the literature are identified in Chapter 2. Firstly, there is a lack of understanding about the causal mechanisms that underly the relationship between paid work and female autonomy. This is also a consequence of the predominance of quantitative methods used to study this topic. Secondly, the relationship between paid work and female autonomy is often generalized, and the different types of work are not studied specifically. This means there is a lack of understanding of the specific interaction between type of paid work and female autonomy, even though evidence suggests type of paid work is relevant. Lastly, there is little integration of cultural beliefs within a context and how these can influence the relationship between paid work and female autonomy.

1.2. Research Question and Approach

This research seeks to contribute to closing these gaps within the literature by answering the following research question: “How does female survival-entrepreneurship impact domestic female autonomy in informal settlements in Kisumu?”. This question allows for a focus on the type of work, in this case survival-entrepreneurship, and for cultural beliefs, by its focus on Kisumu specifically. Moreover, the questions also includes the causal mechanisms that explain the relationship. The question is answered in this study by adopting elements of a bargaining theoretical approach. It uses a critical realist position and in connection to that a mixed methodology. This mixed methodology approach consists of three inter-related methods. Firstly, a survey amongst 275 households was conducted in three informal settlements in Kisumu, which contains both quantitative and qualitative questions. Secondly, to expand on the surveys, focus groups discussions were held, with women from the three informal settlements where the surveys were conducted Lastly, the local assistant chiefs of the three informal settlements were interviewed about their role in the decision-making process and conflict resolution in the household, and to give more information about the areas where they are assistant chief.

The use of mixed methods means the quantitative results from the survey can be put into the cultural context in which decisions in the household take place with the qualitative data. Furthermore, these methods allow for the inclusion of the perspectives of women on their own autonomy. This also means the perception of women on the causality of the relationship can be included, and therefore, the study can contribute to the debate on the causal mechanisms in the relationship between paid work and female autonomy. There is also a social importance. Given the focus on economic empowerment in development policy, survival-entrepreneurship is a particularly

(13)

important form of work to study. For many women in low-income countries , survival-1 entrepreneurship is the only form of work that is accessible and feasible for them, because they are excluded from other forms of work. Increased information about this relationship could feed into improved policy and interventions targeted at increasing the autonomy of women, who are survival-entrepreneurs.

1.3. Thesis Outline

Chapter 2 explains the concepts of female domestic autonomy and survival-entrepreneurship. Furthermore, it introduces the bargaining theory and the three resources in this theory: the material, relational and cognitive resources. Lastly, it also presents alternatives to the bargaining theory. In Chapter 3, the research questions and methodology of the study are presented. The chapter presents the sub-questions. Moreover, the data collection and analysis of the quantitative and qualitative research methods are explained. Chapter 4 discusses the research location of the study. It discusses both the context of the national, county-level and three informal settlements where the research was conducted. Chapter 5 discusses who makes decisions in the household. It starts with different regression-models, and expands on these with the qualitative explanations women give for the way decision making is divided in their household. In Chapter 6, the process of decision-making in the household is discussed and the consequences of paid work according to the respondents. Five different phases are identified and linked to the strategies that women use to exert influence in these phases. Lastly, in Chapter 7, the Conclusion, the main research question is answered. An adapter conceptual model is presented and the chapter reflects on methodology. It also discusses policy recommendations and future endeavors of research.

Whilst I realize the term ‘low-income countries’ is not ideal, this is the word used in a lot of the literature and by many

1

of the organizations that discuss female paid work. Therefore, I opted for using this over other descriptions such as developing nations or the Global South.

(14)

Chapter 2. Literature Review

This chapter provides an overview of existing literature about the relationship between female domestic autonomy and paid work. The chapter starts with a conceptualization of female autonomy in section 2.1 and a conceptualization of survival-entrepreneurship in section 2.2. Secondly, the chapter addresses the argument of increased material resources and the economic ‘bargaining theory’ in section 2.3. This section also introduces two other determinants of domestic autonomy: relational and cognitive resources. The main argument in this section is that an increase in the resources that women can use in the bargaining process in the household, leads to an increase in decision making. Moreover, the chapter discusses the advantages and disadvantages of alternative theories for household autonomy, such as non-conflict based household decision-making in section 2.4. Section 2.5 presents the conceptual model and combines the different theories presented. Lastly, section 2.6 concludes this chapter.

Thee main research gaps are identified in the chapter: (1) The literature about paid work does not differentiate between different forms of work. (2) The literature acknowledges culture matters but does not discuss the specific ways in which this happens. (3) Many studies focus on the relationship, but few examine the underlying causal mechanisms that could explain the relationship.

2.1. Conceptualization of Autonomy

It is important to establish what female domestic autonomy means before we study the relationship between paid work and female domestic autonomy. This section discusses the conceptualization of autonomy in general and domestic autonomy specifically. Lastly, the importance of studying domestic autonomy is explained.

While there seems to be quite some agreement on the general conceptualization of autonomy, finding an agreed upon operationalization and measurement is more complex (Meurs & Ismaylov, 2019). The concepts: agency, control, empowerment, and autonomy are often used interchangeably in the literature (Beegle, Frankenberg, & Thomas, 2001). This study also similarly uses autonomy and decision-making power. Following the literature, this research conceptualizes autonomy as “the ability to make decisions about one’s own life and act on them to achieve a desired outcome, free of violence, retribution, or fear” (Klugman et al., 2014, p. 13). The relational aspect of autonomy means that in this case, the autonomy of a woman is relative to other members of her household, such as her partner, children, parents, co-wives or in-laws (Vaz et al., 2016).

(15)

Empowerment refers to the process of gaining more autonomy, which is a development rather than a state (Kabeer, 2008).

The meaning of autonomy is culturally bound and differs between countries and over time (Hanmer & Klugman, 2015). Interpretations of autonomy differ because the decisions that are deemed important and relevant differ across cultures. Moreover, the areas that are seen as the responsibility of one household-member, which means this person is the sole decision-making in this area, can differ (Li & Xiaoyu, 2011; Mabsout & Staveren, 2010). However, the broader conceptualization of autonomy as the ability to make decisions about issues that are deemed important is universal and cuts across cultures. There are many different domains in which women can gain or lose autonomy, such as the domestic domain, the political domain, or the financial domain (Hanmer & Klugman, 2015). Autonomy or empowerment in one area of a woman’s life does not necessarily lead to empowerment in other areas of her life (Heaton, Huntsman, & Flake, 2005). For example, if a woman gains political autonomy, that does not necessarily mean she also gains financial autonomy. This study focuses on female autonomy on the household-level, which this study calls domestic autonomy.

2.1.1. Conceptualization of Domestic Autonomy

Malhotra (2003) distinguishes between autonomy on the level of the household, community and the broader arena of women. She identifies six different dimensions of domestic autonomy: economic, socio-cultural, interpersonal, legal, political and psychological. The legal and political dimensions are not taken into account in this research, because they refer to the knowledge women have about their legal rights and political reality around them. Given that one of the mechanisms that is studied in this research is based on increasing knowledge of women, incorporating this within the conceptualization of domestic autonomy would make the dependent and independent variable similar. Moreover, given that these two dimensions focus on information, rather than making actual decisions, it does not fit within the conceptualization of autonomy as described by Klugman (2014). The psychological dimension is not included for similar reasons, because the self-perception of women is also studied as a mechanism rather than the independent variable in the research. This leaves three dimensions in the female domestic autonomy: the economic, socio-cultural and interpersonal dimension. The operationalization and measurement of these three dimensions can be found in Appendix A.

(16)

2.1.2. Importance of Domestic Autonomy

Domestic autonomy is important. Firstly, in a general sense, I argue that having the ability to make decisions over ones own life is a good in and of itself regardless of other consequences. It means women can control their lives and this is already a benefit. Moreover, increased domestic autonomy makes it easier to increase other areas of autonomy (WorldBank, 2012). Even though, higher autonomy in one area does not necessarily lead to autonomy in other areas, without the ability to make decision is in the household, women are less able to create the circumstances that allow them to make decisions outside the household. For example, without the ability to space or limit the number of children born, it is difficult for women to increase their income and plan their career or paid work (Canning & Schultz, 2012; Miller & Xiao, 1999). Moreover, it would be difficult for women to make decisions outside their household if they are restricted in their decisions within the home. Often women specifically, spend a lot of time in their household. Their household is where their most basic needs, such as food, shelter, and hygiene, are met. Women are unable to make decisions if they do not have these basics met. This means a household where women can make decisions can be a necessary condition to make decisions on a community level.

There is also a broader social importance. In households where women make more decisions, defined as autonomous within this study, the spending and decisions made in the household are more likely to benefit society. Women are more likely to spend money on the healthcare and education of their children. For example, households where women have more domestic autonomy, there are higher levels of immunization of their children (Ebot, 2015), lower levels of infant mortality (Boehmer & Williamson, 1996), and higher levels of school enrollment (Alfano, Arulampalam, & Kambhampati, 2010).

2.2. Survival-Entrepreneurship

This section explains the term survival-entrepreneurship. Survival-entrepreneurs are people who start an informal businesses to add to the income of their household, without a growth-strategy because they do not have enough capital (Vossenberg, 2016), opportunity or freedom. This type of paid work has the interest of both policy-makers and academics for its possible impact on economic and social development. The conceptualization and importance of survival-entrepreneurship is expanded on in the following paragraphs.

2.2.1. Conceptualization of Survival-Entrepreneurship

This study focusses on women who have become ‘survival’-entrepreneurs. It is difficult to conceptualize survival-entrepreneurship, because the majority of the literature is based on the

(17)

Western economic reality, ignoring poverty conditions, and is gender-blind. The traditional image of the entrepreneur is that of the self-made man, motivated by “the dream and the will to found a private kingdom, usually, but not necessarily, also a dynasty” (Schumpeter in Ahl 2016). However, in low-income countries many people start their business for different reasons entirely.

People who start their own business in low-income countries, often do not fit the traditional Schumpeterian definition of entrepreneurship (Banerjee & Duflo, 2007). They start informal businesses to contribute to the income of their household. It varies per country, but many of these entrepreneurs have multiple forms of income, to limit the risk for the household when one form of income declines (Grimm, Knorringa, & Lay, 2012). There is limited specialization, which means business often does grow to a size where it can provide for all household needs. The lack of specialization can also partly be explained because there is no intent to grow the business and there is a considerable strain on resources that limit the ability to invest in the business (Ahl, 2006; Banerjee & Duflo, 2007).

The traditional literature on entrepreneurship is less relevant to study entrepreneurs in low-income countries, where a large part of the economy is informal. Historically, entrepreneurship is often seen and analyzed through the expectation of the male entrepreneur (Ahl, 2006). The literature sees the defining characteristics of entrepreneurs as being risk-taking and innovative (Carland, Hoy, Boulton, & Carland, 1984). The entrepreneur is seen as a visionary who wants to change the world. Therefore, there is a lack of focus on power relations or gender equality (Ahl, 2006). Moreover, the literature is largely based on situations in developed states with a strong, formalized economy, where there is no lack of resources or supportive institutions (Berner, Gomez, & Knorringa, 2012)

Some literature argues that the type of informal entrepreneurship that is common in these countries should not be seen as entrepreneurial (Carland et al., 1984). However, as outlined by Berner, Gomez, et al. (2012), the entrepreneurs who start their business out of necessity, still risk their assets and apply their knowledge, while making decisions about the best way to run their business. In this sense, they are entrepreneurs. Moreover, these entrepreneurs often use the same skills that are labeled as ‘entrepreneurial’ by traditional literature (Grimm et al., 2012). For example, survival-entrepreneurs also respond to market fluctuations and “independently allocate factors of production” (Berner et al., 2012, p. 384). Therefore, this study classifies this type of business as entrepreneurs. However, the variety between different forms of entrepreneurship shows the need for a distinction between different forms of entrepreneurship. This distinction is made by looking at the goal of entrepreneurs. The group is divided between those who have a

(18)

growth-strategy and entrepreneurs who became entrepreneurs as a means of survival (Berner et al., 2012). The growth-strategy is similar to the description of an entrepreneur as given by Schumpeter (1983). The other significant group of people starts as an entrepreneur as a ‘survival-strategy’ (Berner et al., 2012; Grimm et al., 2012; Vossenberg, 2016). This group can often not access other forms of paid work, and begin their own small business as an alternative to earn an income. Even with entrepreneurship they face significant social, legal and economic barriers (Amine & Staub, 2009). This type of entrepreneurship is particularly relevant for those living in poverty (Banerjee & Duflo, 2007). Moreover, people who become survival-entrepreneurs have the lowest opportunity costs, for example, because they do not have have access to paid employment. As a result, survival-entrepreneurs are often women or the elderly. In West Africa, men own only 13% of survival-type businesses, while 76% of ‘growth’ entrepreneurial businesses are owned by men (Grimm et al., 2012). Moreover, survival-entrepreneurs are less likely to be literate, to speak the language of the elites and to have electricity. Different from the “conquering man” who becomes an entrepreneur, those who become entrepreneurs as a survival strategy are often the most disenfranchised in poor communities. This study refers to this group of entrepreneurs as survival-entrepreneurs.

2.2.2. Importance of Studying Survival-Entrepreneurs

Studies about paid work and female domestic autonomy argue type of work matters (Alam, 2011). The income, security, network, and other characteristics of the paid work, all change the relationship (For example in Heaton et al., 2005). Therefore, it is important to focus on one form of paid work rather than trying to generalize between all different forms. There are three reasons why this study decided to focus on survival-entrepreneurship as the specific form of paid work. Firstly, the focus of development policy on entrepreneurship means there is a need to understand survival-entrepreneurship from a perspective different than the Western economic angle. Secondly, there is a social importance of the group of women who become survival-entrepreneurs, and lastly, an academic interest because survival-entrepreneurship in many ways is an atypical case of paid work. Firstly, economic empowerment is treated as a ‘magic-bullet’ for international development. With a lack of alternatives for many women, survival-entrepreneurship is the most likely form of paid work. This means that if we want to see what the consequences are for women when they start to work, we should look at paid work realistically available to the women that development policy seeks to empower.

(19)

Secondly, there is social importance in studying this form of paid work. The women that become survival-entrepreneurs are often the poorest in a country. They are most likely to be excluded from formal employment (Banerjee & Duflo, 2007). As a result, this group of women is often described as the most deprived of decision-making power. If these women face most constraints in their decision-making, it is especially important to study in which ways they are constrained. Moreover, because there is little focus on the decision-making that they do have, it is difficult to establish in which areas they lack most autonomy. Without such an understanding, solving the barriers these women face in their decision-making remains extremely difficult.

Lastly, the case of survival-entrepreneurship is a specifically interesting form of paid work to study, because it is an atypical case of paid work. Survival-entrepreneurship is a form of paid work for women who are particularly disenfranchised and lack alternative forms of income. Women who become survival-entrepreneurs do so from necessity or lack of alternative options. This means that a lack of autonomy, at least in some areas, is the reason behind this form of work, which makes it more interesting to study whether survival-entrepreneurship still increases the autonomy of women. Moreover, there is a limited increase in the income or wealth of women because the profit margins of this type of work are often small (Grimm et al., 2012). The informal nature of the work and the frequent dependency on agricultural production means the income is insecure and unstable. The focus on income in the literature as the mechanism for change, means that this is an interesting case to study. With a low income, survival-entrepreneurs might not have the economic benefits of work that other women have, but they could still be affected by the other mechanisms that increase female autonomy as a result of paid work. Survival-entrepreneurship, therefore, challenges the assumptions about paid work. Moreover, one of the questions about the relationship is the direction. The literature often assumes that women who become survival-entrepreneurs lack autonomy and are especially vulnerable (Banerjee & Duflo, 2007), while paid work is also seen as empowering. This paradox makes this group interesting to study.

2.3. Bargaining Theory

This section explains bargaining theory. It starts with the theory that women use material resources to strengthen their bargaining position in the household. Thereafter, two other resources are discussed, which were added after critique on oversimplification of the bargaining theory: relational resources and cognitive resources.

The leading theory behind the argument of increased female autonomy as a result of paid work is the ‘bargaining theory’. This body of literature is based on the bargaining models of John

(20)

Nash (1953). It argues that an increase in the bargaining power of women leads to an increase in women’s autonomy. The theory assumes that the ‘utility function’ of members of a family differ (Manser & Brown, 1980; McElroy & Horney, 1981), as opposed to the traditional view that a household or a family share a common ‘utility function’ (Becker, 1981). Put simply, a different ‘utility function’ means there are different, contradicting preferences between different members of the household, resulting in the need for a compromise (Sen, 1990). There is empirical evidence that many household members have preferences that contradict at least partly, such as preferences about reproductive choices or resource allocation (Beegle et al., 2001). In the case of contradicting preferences, the theory argues that a decision in the household is made through the process of negotiation and bargaining.

2.3.1. Material Resources

If women have more ‘threat options’ during the bargaining process, they would have more influence in the negotiations and resulting decisions. In this case, this refers to their fall-back options if cooperation in the household breaks down. Fall-back options can refer to divorce or separation, but also non-cooperative behavior within the household or marriage (Ghysel, 2004). Examples are the ability of a woman to withhold income necessary for the household or her ability to provide for herself if she left the household. Kabeer (2005) observed this in women who received micro-credits to start up a business, who report to use non-cooperation as a bargaining tool. The income and assets of a woman are called her material resources.

However, the argument is not as straight forward as it might seem. Firstly, it matters whether the income women earn is earned independently or in cooperation with the rest of the income. If other members of the household are needed in the process, for example, when they help on the land or take care of chores when a woman focusses on her paid work, there is an interdependence. Interdependence creates difficulties to use the income as a bargaining tool (Heaton et al., 2005; Kabeer, 1997), especially in countries were women are dependent on men to “mediate many social interactions” (Kantor, 2003, p. 426). Furthermore, women need to have control over material resources before they can use it during negotiations (Eyben, Kabeer, & Cornwall, 2008; Fafchamps & Quisumbing, 2002). Lastly, predictions of future income impact the ‘threat’. Expected future income from paid work influences the strength of the negotiation-position (Majlesi, 2015). Unstable or insecure income means that women might not be able to take care of themselves financially in the future, which would limit their ability to make a credible threat to leave the

(21)

household for an extended period of time. Therefore, not only current income but also security and predictability factor into the bargaining position of women.

The survival business entrepreneurs often generate a low-profit in absolute terms (Grimm et al., 2012). As a result, one hypothesis could be that the increase of bargaining power using material resources is limited as well. At the same time, even with limited profit in absolute terms, the relative income could be high. Survival-entrepreneurs come from poor household (Grimm et al., 2012). So, for these specific households, the profit could make a significant difference, meaning the bargaining power would increase. Especially because these households often do not have sufficient funds to take care of all their needs, a small amount of money can still have large effects. Therefore, the argumentation of the bargaining theory applies fully to survival entrepreneurs and women would be empowered through their paid work. The income of a survival-entrepreneur is often unstable, which suggests women can not use their income for long-term threats.

One issue with the focus on the income of the original bargaining theory is that “these economic models fail to take into account the “extra-household socio-economic and legal institutions within which households are embedded” (Agarwal, 1997, p. 2). It assumes that the actual size of material resources brought in during negotiations influences households, rather than the perceived contribution (Sen, 1990). This shows, for example, in a lack of analysis of the way perceived contribution to the household is constructed through gendered and social norms or expectations (Agarwal, 1997). Moreover, culture can influence the choices women have, for example, strong cultural norms can prevent separation even if financially feasible (Kantor, 2003). Other cultural norms could influence the composition and hierarchy of the household. Therefore, it is too simplistic to look solely at the influences of material resources as a factor in female bargaining power. Cultural and social norms might explain some of the conflicting results of empirical research. It is difficult to generalize across countries about the characteristics of bargaining within the household and one study in multiple nations can establish different relationships between paid work and autonomy (Heaton et al., 2005; Kabeer, 2011). Some cases conclude very limited to no increase in autonomy (Blattman, Green, Annan, & Jamison, 2013; Kantor, 2003; Vaz et al., 2016), whilst other conclude female autonomy actually does increase significantly (Antman, 2014; Beegle et al., 2001; Boateng et al., 2014; Chakrabarti & Biswas, 2012; Kabeer, 2011). Moreover, the influence of paid work can differ across different decisions (Anwar, Shoaib, & Javed, 2013; Kishor & Subaiya, 2008). The reason these different norms have such a substantial impact is that they can create or prevent other constraints on female decision-making.

(22)

Kabeer, Mahmud, et al. (2011) propose a solution to this issue by introducing two resources believed to have an impact on female autonomy: cognitive and relational resources. While these resources improve the bargaining theory and add to the complexity of the model, there is a lack of understanding of the way the causal links interact.

2.3.2. Relational Resources

Relational resources refer to the relations women have with other people in the broadest sense of the word, and their ability to capitalize on this. It refers to the personal relations and also to her status and respect in general. Paid work influences the place a woman holds both within society and her household, changing her relational resources. It matters in what way general paid work, and survival-entrepreneurship specifically is perceived in the household and society (Kabeer, 1997; Sen, 1990). Relations resources means the bargaining model can incorporate the context of the household composition. While, especially Western, literature assumes women all live in a nuclear family, this is not the same in every culture, and this impacts the bargaining position of women (Debnath, 2015). The three main factors of relational resources described in the literature, influenced by paid work, are the status of a woman in the household, the threat of violence, and the networking opportunities outside the household.

Firstly, one aspect is status as a result of the paid work. Many different factors influence the status of a woman in the household, such as age (Hanmer & Klugman, 2015), the number of children a woman has (Kabeer, 2002) and the perceived contribution of a woman (Kabeer, 2011). Women who are believed to contribute to the household get a bigger voice, yet, intentions matter. A woman who seeks to contribute to the household income when it is in need of money, is often regarded higher than a woman who tries to seek an income to become independent or to increase her bargaining position (Chakrabarti & Biswas, 2012; Heaton et al., 2005; Kabeer, 2005, 2011). This means that the perceived reason that leads a woman to seek paid work matter for the increase (or decrease) of her status. If the status of a woman increases, her voice and therefore, preferences are weighed more during the negotiations. The societal perception of her work also influences her status. If a woman is employed in a line of work which is deemed shameful, such as prostitution, the income that is brought in is mot likely not discussed as openly, meaning it is less effective as a bargaining tool. Therefore, it is important that research includes the type of work women have. This is currently understudied, since most research focusses on material resources, rather than the type of work women have.

(23)

A second important factor in the relational resources of a woman is the possible increase in violent behavior from a partner. The perceived boss or patriarch of the household might feel threatened as a consequence of the paid work of a woman. The increase in material resources might mean that the husband feels he needs to rely on his wife, or that the income could be used against him if she needed to negotiate with him, meaning his power in the household is decreased (Hanmer & Klugman, 2015). Therefore, a woman with income could be perceived as a threat to the power balance in the household, leading men in the house to feel that they must reinforce the power balance in the household. One way to rebalance the power is to use existing physical differences, which could manifest itself through intimate partner violence or violently controlling behavior (Slegh, Barker, Kimonyo, Ndolimana, & Bannerman, 2013; Vossenberg, 2016). Even if a husband does not use violence, the risk of violence could still constrain the negotiation tools a woman can use. However, it is unlikely that there is a linear increase in the risk of violence when women start to earn more money. For example, if women start to make money, they have more freedom to chose their partner or to leave their partner, which can decrease domestic violence (Muthengi, Gitau, & Austrian, 2016). When the income of a woman starts to threaten the power of balance, it might work the other way around. The interaction between paid work and intimate partner violence complications the bargaining model. The increase of bargaining power as a result from material resources can also lead to a decrease in bargaining power as a result from relational resources, which shows the importance of a better understanding of the causal mechanisms. The possibly contradicting effects need further exploration. Current literature identifies the different resources, but the interaction between the causal mechanisms is understudied.

Thirdly, paid work can change relational resources by increasing and extending the relations of a woman outside the household. Women who hold a job outside the house, or who discuss work-related issues with other women, build up a network. Within the framework of the economic bargaining theory, this means an increase in possible fall-back options and therefore an increased bargaining opportunity (Ghysel, 2004). It also means that women have better opportunities to discuss previously private matters, making men more aware of the way their behavior could come across (Kabeer, 2005). Afraid of how they are perceived if they cross social norms, men are more likely to change their behavior accordingly. It depends on the different cultural norms, whether this means they are more aware of their perceived loss of masculinity or if they feel a bigger need to listen and respect their wives. This form of relational resources seems to strengthen cultural norms around female decision-making, whether that means strengthening or weakening constraints that

(24)

already exist. Therefore, understanding cultural context is important when discussing the impact of relational resources.

Survival-entrepreneurship is often home-based (Kabeer, 2005), making the increase in network marginal (Grimm et al., 2012). Moreover, survival-entrepreneurs are far more likely to hire or work with family, rather than members of their broader community (Grimm et al., 2012). This means women built few new relations through their work. The increase in material resources could still lead to an increase of perceived threat towards the household balance. However, this is unlikely to have a significant impact for two reasons. The context of most survival-entrepreneurs is that they start their business as a way of diversification rather than to gain independence, which makes it likely it is perceived as a contribution to the household, not as a way to get power. This would actually lead to an increase in the status that a woman holds in the household (Chakrabarti & Biswas, 2012; Heaton et al., 2005; Kabeer, 2005, 2011). Moreover, the survival businesses are only part of the income and not enough to provide fully for all of their needs, meaning the balance of power is not under threat (Grimm et al., 2012; Kabeer, 2005). Therefore, women are still within the ‘female’ domain, and this way of paid work does not threaten expectations such as a woman staying at home or a husband knowing what his wife is doing during the day.

2.3.3. Cognitive Resources

Cognitive resources refer to mental processes and skills. While higher levels of education are often positively linked to increased autonomy in many domains, such as political or economic autonomy (Hanmer & Klugman, 2015; Kabeer, 2002), it is too simplistic to reduce cognitive empowerment solely to formal education. Firstly, skills can be learned through other mechanisms that formal education and these skills acquired through a different means than formal education should also be included in the concept of cognitive resources. Women can develop skills from their paid work that enable them to make decisions in their households, such as keeping a budget or negotiation techniques. However, it remains difficult to establish the size of the effect of skills and knowledge on female autonomy. Interventions to create more autonomy through education only seem to have modest effects on autonomy (Heaton et al., 2005).

Secondly, cognitive resources can also refer to the beliefs a woman holds about herself and her rights or duties in the household. The way women think about their role in the household influences her autonomy (Hanmer & Klugman, 2015). Paid work is relevant in this aspect because women who work can gain confidence and get a sense of ‘mental strength’ (Amin et al., 1998; Kabeer, 2011),which could potentially increase female autonomy. Moreover, if women have paid

(25)

work, they often have increased confidence in their ability to decide on relevant topics (Blumberg, 2005). People who believe they have decision-making power, regardless of whether they objectively have, are more likely to take action and control in their lives (Diener & Biswas-Diener, 2005). At the same time, developing a sense or understanding of autonomy does not mean that oppressive structures change and that autonomy can be exercised (Rowlands, 1997). However, it might make them better able to work within these structures.

Based on the literature, I hypothesize that cognitive resources are partly increased partly when women become survival entrepreneurs. The majority of women who become survival entrepreneurs have very little education. Most often, they do not use financial note keeping (Grimm et al., 2012). However, there are still some skills that they could develop, such as skills in negotiation-techniques or even basic math skills. Their feeling of pride and confidence is expected to improve (Kabeer, 2005). Especially if the income is vital to the survival of the household, which is the case of survival entrepreneurship, the feeling of contribution to the household can improve self-confidence.

2.4. Alternative Explanations

Bargaining theory is not uncontested in the literature. This section discusses some of the criticism and alternative theories. Firstly, it discusses the assumption of individual egoistic preferences of household members. Secondly, it explain the importance of culture in theories. Lastly, it discusses the interplay between agency and structure.

One issue with the argument of bargaining theory is that it assumes that all members of the household are ‘egoistic’ and are mainly concerned with their own preferences. A proposed solution is the inclusion of the ‘caring agent’ into the model (Chiappori, 1992). This means that a member of the household cares about other members in the household and therefore included their preferences in their own preferences. Yet, this is often not included in the economic models, because it is difficult to express in statistical analysis. Moreover, the model assumes that a model is conflictual, rather cooperative (Becker, 1981). Some literature argues the household can be seen as one, rather than as a group of individuals. The assumption of individuality within the household can be seen as a cultural assumption.

Another field of literature argues that rather than looking at individuals and individual households, the main focus should be on broader culture and group formation. Firstly, it is argued “that country and community of residence predict women’s reported domestic empowerment better than their personal socioeconomic and demographic traits do” (Oppenheim Mason, 2005, p. 91).

(26)

Moreover, even when individual traits, such as paid work or survival-entrepreneurship, can be statistically linked to female autonomy, at times this significance is lost when accounted for the beliefs of women and men about female autonomy. The bargaining theory assumes decisions are made in a way that cuts across cultures and countries: different individuals in a household bargain together. However, this assumes the structures that guide society are consistent, rather than fluid and created.

The last critique on the bargaining theory fits within this broader scientific debate on agency and structure. While many scholars argue that there is a dichotomy between structure and agency, one strand of theory combines these two stances in ‘practice theory’. This position assumes a “dialectical rather than oppositional relationship between the structural constraints of society and culture on the one hand and the “practices” of social actors on the other” (Ortner, 2006, p. 2). Certain structures constrain the decision-making of women, but at the same time, women have the agency to negotiate and change these structures. The bargaining theory at times assumes the structure of the bargaining in the household is stable, and women are only improve their position within the bargaining framework, rather than change 'the rules’ of the bargaining. Practice theory argues that “systems of control can never work perfectly because those being controlled have both agency and understanding and thus can always find ways to evade and resist” (Ortner, 2006, p. 5 ). This means agency needs to be included within the models of bargaining to show how the structure can be mitigated or otherwise contested and used to the advantage of women and men.

2.5 Conceptual Model

This section provides a short summary of the conceptual model presented in Figure 1. The six sub-questions presented in Chapter 3 are also visually depicted within the model. As shown in the figure, survival-entrepreneurship comes from a lack of autonomy in the form of oppressive gender-norms, poverty and a lack of alternative employment. Embedded within the context of culture, survival-entrepreneurship influences three main resources: relational resources, material resources and cognitive resources. This is a non-linear process, since these three resources enhance, contradict and influence each other in multiple directions. Moreover, they influence survival-entrepreneurship and are in turn influenced by domestic autonomy. These resources influence the household-bargaining position. A stronger household position means women can make more decision in the household and therefore have more female domestic autonomy.

(27)

Figure 1. Conceptual model

Fe

m

ini

st

S

ta

nd

po

int

Th

eo

ry

P

ra

ct

ic

e Th

eo

ry

So

ci

al

-

C

ul

tu

ra

l C

on

te

xt

Gender Inequ al ity La ck o f al te rn ati ve ty pe s of wo rk C on str ai ne d Re so ur ce s N ec es si ty Survival - En tr ep re ne ur shi p Increase F emale D om es ti c A uto no m y

Decision Making Process

Increase in

M ate ri al Re so ur ce s R el ati on al Re so ur ce s C og ni ti ve Re so ur ce s Str on ge r In tr a-H ou se ho ld Ba rg ai ni ng P os iti on SQ 6 SQ 2 _{SQ 3} _{SQ 4} SQ 5 SQ 1

(28)

2.6. Conclusion

This section has provided a first insight in the way decisions within the household are made. The household-members bargain in the household, within the context of culture and gendered norms, using three main different resources. Survival-entrepreneurship means that women increase household income and are more independent from their partners. However, the nature of survival-entrepreneurship also means that the benefits for women are likely limited. This specific type of work is born out of necessity and yields limited financial results. Therefore, survival-entrepreneurship is expected to be a form of work that only marginally increases the female domestic autonomy of women.

Three research gaps were identified. Firstly, more research is needed into the ways specific types of work impact female domestic autonomy. While research acknowledges the impact of different forms of work, there is little research that builds on this to study the specific different characteristics of paid work that are important. This study seeks to contribute by focussing on survival-entrepreneurship specifically and look at the way this form of paid work compared to other forms of work, such as employment and growth-entrepreneurship. Secondly, current research often lacks an understanding of culture or social norms. The discussion of culture stays at an abstract level which states that culture is relevant, but it does not discuss the specific ways culture interacts with the link between paid work and female autonomy. The majority of the studies are quantitative and on a national level, which means they cannot account for local culture, except by using pre-defined quantitative variables. This research, uses a mixed-method approach, which is elaborated on in chapter 3. This approach allows for a broader study of cultural beliefs and a discussion with women about the way they perceive cultural expectations. A third gap in the literature is a lack of comparison between the different mechanisms that influence female autonomy. The literature often identifies different causal links, but does not discuss the way the resources are linked to each other and the way conflicting mechanisms affect autonomy. The studies that discuss the relationship often use statistical regression analyses, which means they lack an explanation of causal mechanisms. 

(29)

Chapter 3. Research Questions and

Methodology

This chapter starts with the explanation of the ontological and epistemological stance of this thesis: critical realism. Section 3.2 introduces the main research question of this study, and the six sub-questions used to answer the main research question. In the next section the methodology of this thesis is explained. It starts with the exploration of the mixed method approach used in section 3.3. In section 3.4 the debate between structure and agency in the general context of social science is discussed, highlighting the relevance for this specific research. Thirdly, the chapter elaborates on the unit of response and analysis in section 3.5 using feminist standpoint theory. The chapter moves on to explain the used mixed-method approach and it provides further explanation about the specific quantitative and qualitative methods used for data collection and analysis in 3.6 and 3.7. Lastly, it discusses the ethics in section 3.8 and limitations of the research identified before the fieldwork in section 3.9.

3.1. Critical Realism

This study takes a critical realist approach. This means that it takes the ontological stance of a “relatively autonomous existence of social reality” (Archer et al., 2016). Reality exists, at least partly, independent of human beings and human perceptions of the world (Mingers, 2006). At the same time, reality changes, and it is not static. Social structures are always localized in time and space, meaning concepts, such as female autonomy, have different meanings in different contexts at different moments in time. However, critical realism is not a social constructivist approach. Critical realism argues it is possible to differentiate between forms of knowledge and that there should be clearly defined criteria to assess the quality of the research. Moreover, critical realism believes science can build and improve its knowledge about social reality and get a more accurate representation of reality.

Pawson and Tilly (1997) describe how critical realism specifically studies the way mechanism (or causal relations) play out in different contexts. This is refered to as the ‘real' within critical realism, which “includes the entities and structures of reality and the causal powers inherent to them as they independently exist” (Wynn & Williams, 2012, p. 790). This fits well within the goal of the study, which seeks to explore the causal mechanisms between paid work of women and

(30)

increased autonomy in the household. While focussing on a general relationship, this study acknowledges the local cultures and therefore does not draw broad, generalized conclusions. Lastly, according to the critical realist approach, it is crucial to triangulate data by using multiple methods in research. “[Different] methods can be seen to be necessary to reveal different aspects of the constituency of phenomena, that is their ontic character, as structural, that is cause and effect, relations more broadly” (Downward & Mearman, 2007, p. 91). In line with this principle, this study uses a mixed-method approach to gather the data via different methods and from different sources, relying on both a quantitative and qualitative methodology. These methods are used to triangulate the data. The triangulation happens both through the methodology and the collection of data.

In agreement with the feminist standpoint theory, I argue that women are better able to identify oppressive structures within their household, than their husbands, often the benefactor of these structures (Hartsock, 1997; Hekman, 1997). Since women have to live within oppressive structures, they have to understand these structures to survive and circumvent the structures. This is in line with Ortner (2006) and practice theory, which argues that people have the agency to decide how they behave within social structures. People who benefit from these structures are not required to understand them in order to survive and have less of an understanding, in this case, about the decision-making processes in the household (Harding, 1991).

3.2. Research Questions

Flowing from the gaps identified in the literature review, this research seeks to answer the following research question: “How does female survival-entrepreneurship impact female domestic autonomy in Kisumu?”. To do so, it answers six main sub-research questions.

Sub-question 1: “Who makes decisions in the household in Kisumu?"

This first sub-question is a descriptive question which is meant to shed a light on the context of decision-making within households in Kisumu. Without a perspective on who makes decisions in the household, the changes cannot be put into context. This question was included to get a more general perspective on the way decisions in the household. This puts the importance of survival-entrepreneurship in a broader context of decision making in the household.

(31)

Sub-question 2: “How do material resources from survival-entrepreneurship influence female domestic autonomy in Kisumu?”

Sub-question 3: “How do relational resources from survival-entrepreneurship influence female domestic autonomy in Kisumu?”

Sub-question 4: “How do cognitive resources from survival-entrepreneurship influence female domestic autonomy in Kisumu?”

These three sub-research questions focus specifically on the relationship between the three resources identified in the literature and female autonomy. This allows for a more specific exploration of the influence of these three resources, which are currently still understudied, as explained earlier. This creates a better understanding of which part of survival-entrepreneurship influences female autonomy. This focus on causal mechanism also fits with the critical realist approach.

Sub-question 5: “In what way can women influence decisions made in the household in Kisumu?” The actions women can take to influence decisions made in the household contributes to the understanding of the causal mechanisms between survival-entrepreneurship. It fits within the literature of practice theory, since it studies the actions women take within the constraints of the household decision-making and the ways they negotiate with these structures.

Sub-question 6: “How do the levels of autonomy of women compare to women who have other types of paid work?”

This last research questions allow for a closer comparison between different forms of paid work. This sub-question was included for two reasons. As explained in Chapter 1, the type of paid work is relevant for the status, (stability of) income and skills learned. Therefore, comparisons between different types of work should be made. It contributes to the a more specific understanding of the impact of different resources. Furthermore, the comparison between different forms of paid work allows for a more accurate identification of the mechanisms that underly the relationship between paid work and female domestic autonomy.

3.3. Mixed Method Approach

This research uses a mixed-method approach, more specifically a sequential explanatory approach (Creswell & Clark, 2017). Given the extensive literature on this specific relationship, an exploratory design was not necessary. A sequential explanatory approach is a design with two

(32)

separate phases of data collection, starting with a quantitative phase and building on that with a second qualitative phase. In the first phase, data is gathered in the form of questionnaires. These questionnaires contain both quantitative questions and qualitative questions. After the first phase, a preliminary analysis of the quantitative data is conducted and based on those findings, further questions of inquiry are identified. These questions are explored further using focus groups and stakeholder interviews in the second phase of data collection. In the second phase further qualitative data is collected to expand on the results of the first phase of research. More specifically, this type of mixed-method research is called the “follow-up explanations variant” (Creswell & Clark, 2017).

However, the research is also partly convergent design. This design means that data analysis of the data from the first and second phase of research is partly done simultaneously. Since the questionnaires contain both closed and open-ended questions, the questionnaires generate both quantitative and qualitative data, this is what is described by Creswell and Clark (2017) as the “questionnaire variant” of convergent mixed-method design. In the analysis of the first phase of data collection, this data is quantified. In the second phase, the qualitative data from the questionnaires is analyzed in the same way as the data from focus groups and interviews with government officials.

3.4. Units of Analysis and Response

In this research female autonomy is the unit of analysis. The unit of response is individual women, whether in a relationship or not. The study specifically focusses on women. Men, or other individuals in the household, were not included in the sample. Women were most relevant to study because the research is about female autonomy. Therefore it is most important to look at the perspectives of women, because they are the only ones who know whether they can make decisions about their own life. If the research included men as respondents to discuss female autonomy, men would be asked to judge whether women were able to include their preferences in the decision-making. Given that preferences are often covert, men are unable to judge whether preferences of women are included in household decision-making. This is also inline with the feminist standpoint theory I take in section 3.1. Moreover, the literature suggests that women have an incentive to hide their negotiation strategies, which makes it likely that men do not recognize or understand these strategies. Secondly, because the study focusses on women who have paid work, the study limits itself to women of employable age, which in Kenya means over 18 years old (Fieldnotes, 11/2018). Contrary to some studies, both women with and without a partner are included in the sample. In female-headed households, there can still be other (male) adults in the household that a

The impact of survival-entrepreneurship on female domestic autonomy in Kisumu, Kenya