Women’s economic status, male authority patterns and intimate partner violence: a qualitative study in rural North West Province, South Africa

Abstract There are conflicting views on the impact of microfinance-only interventions on women’s economic empowerment and intimate partner violence in low and middle-income countries. Evidence suggests however that when microfinance is combined with complementary programmes (microfinance plus) it may be effective for empowering women and addressing intimate partner violence. We conducted in-depth interviews with adult women in rural South Africa who had received microfinance loans for more than a year and had recently completed gender training. We explored women’s perceptions on income generation; the effects on their relationships, including intimate partner violence; their notions of power; and perspectives on men’s reactions to their empowerment. Findings reveal that the notion of ‘power within the self’ is supported by women’s income generation, alongside a sense of financial independence and improved social support. Women reported increased happiness and reduced financial stress, although social norms and gender expectations about women subservience and male headship remain salient, particularly among older women. Furthermore, younger women appeared to tolerate abuse due to financial and caring responsibilities. These findings underpin the importance of complementary gender training programmes and of including men as participants for enhancing the effectiveness of economic strengthening interventions.


Introduction
Intimate partner violence is a major global public health challenge with one in three women ever having experienced lifetime physical and/or sexual violence (Devries et al. 2013). In addition to causing injury or loss of life, intimate partner violence is associated with a range of adverse health outcomes, including depressive symptoms, suicide and increased risk of HIV (Devries et al. 2013;Jewkes 2002). In South Africa, among ever-partnered women aged 18 years and older, 21 percent report lifetime physical intimate partner violence and 6 percent lifetime sexual intimate partner violence (Demographic Health Survey, 2016). Risk factors for intimate partner violence include women's poverty and low education, acceptability of intimate partner violence and gender inequitable norms (Jewkes, Levin, and Penn-Kekana 2002).
Research on both gender (Heise 2011) and development (Buller et al. 2018) posits that one approach to addressing intimate partner violence is poverty alleviation. Governments or donors target poor women in low-middle income countries with savings groups, microfinance programmes or cash transfers (Hidrobo, Peterman, and Heise 2016). These programmes are based on the notion that women with access to earnings and enterprise are more economically empowered, with economic empowerment being defined in terms of women's access to resources through income-generating activities (either employment or credit programmes). This includes measures such as a woman's control over her resources, decision-making power and autonomy or her contribution to household expenses (Vyas and Watts 2009).
Microfinance uses a group-lending approach to increase women's ability to generate income and secure livelihoods and has been identified as an effective poverty reduction tool (Dalal, Dahlstr€ om, and Timpka 2013). These programmes give women control over financial resources in order to create or access an independent source of income (Duvendack and Mader 2019). Microfinance only programmes had shown great promise in the development field for its potential to impact on women's economic empowerment (Vyas and Watts 2009). However, recent trial evidence has shown a modest impact of microfinance only interventions on women's economic empowerment, with effects being context dependent and heterogenous (Banerjee, Karlan, and Zinman 2016;Duvendack and Mader 2019). Furthermore, evaluations have shown that microfinance only does not necessarily result in decreases in women's experience of intimate partner violence (Meinck et al. 2019); evidence from Southeast Asia indicates increases (Bulte and Lensink 2019), with mixed results from sub-Saharan Africa (Gibbs, Jacobson, and Kerr-Wilson 2017).
There is however, evidence to suggest that when microfinance is combined with additional components to address, for example, unequal gender norms and power relations (microfinance plus), it may be effective in empowering women and reducing intimate partner violence (Gibbs, Jacobson, and Kerr-Wilson 2017). For example, a 2006 cluster randomised trial in South Africa of the Intervention with Microfinance and Gender Equity (IMAGE) programme demonstrated that a poverty reduction intervention combined with participatory gender training, achieved a 55 percent reduction in past year physical and/ or sexual partner violence relative to no programme (Pronyk et al. 2006).

Women's economic empowerment and intimate partner violence
There is acknowledgment across disciplines that associations between women's economic circumstances and intimate partner violence risk vary by the economic indicator , type of violence (Ranganathan et al. 2019), and context, though empirical evidence to explain such variations and unpack mechanisms of association is limited (Buller et al. 2018;Vyas and Watts 2009). Furthermore, the evidence for whether being economically empowered leads to a decrease in intimate partner violence is mixed (Buller et al. 2018).
An increase in women's financial independence enhances their household position through financial contributions and control over resources, increases their household economic status, and widens their support networks (Vyas, Mbwambo, and Heise 2015). This is supported by recent evidence from a social and economic empowerment trial in Afghanistan ), a cluster randomised trial of women's groups in Nepal (Gram, Morrison, Saville, et al. 2019) and three intimate partner violence prevention trials that included economic strengthening components plus gender training in Tajikistan, Rwanda and South Africa (Stern, McGhee, et al. 2020), where women's experience of work outside the home indicates a broader transformation of their livelihood opportunities. However, the effect of income generation by itself on women's risk of violence is contingent on a number of other factors, particularly men's reactions (Agarwal 1997;Heise 2011).
Unequal gender norms -the assumption that men should financially provide for the household; that domestic tasks are a woman's responsibility; and that men have the right to 'discipline' womenlimit women's ability to act in their own self-interests (Stern, Heise and McLean 2017). This is the case despite women being under pressure to seek out an income or employment because of poverty, male unemployment and household economic stress (Vyas, Mbwambo, and Heise 2015). Furthermore, if men perceive women as challenging their traditional role as household providers, access to an independent income may increase women's vulnerability to intimate partner violence (Buller et al. 2018). The relationship between women's work and intimate partner violence is further complicated by the role intimate partner violence may play in women's decisions to pursue work in the first place. Men may use violence as a control tactic to prevent their partner from earning money (Grabe, Grose, and Dutt 2015). However, abused women may be more likely than non-abused women to seek their own financial resources in order to gain the financial independence needed to escape an abusive relationship (Vyas and Watts 2009). The risk of increased intimate partner violence could decline over time, however, as both men's individual attitudes and broader social attitudes become more accepting of women's increased economic activity and financial autonomy (Ahmed 2005). This was evidenced in three studies included in a Stern, McGhee, et al. (2020) paper where participants reported shifts from harmful gender norms and roles towards more equitable decision-making and the sharing of household tasks and provider roles. Therefore, the relationship between women's empowerment and intimate partner violence is not straightforward and warrants further exploration (Raj et al. 2018;Vyas and Watts 2009).
It is also important to consider how the neoliberal notion of women's agency as 'powerful women exercising control over their own lives and relationships' may not fully apply in some contexts. Rather, there needs to be recognition of the everyday complexities of women's lives, including the patriarchal social structures that constrain them. The fact is women may experience gradual increases in relationship power, with agency distributed over time, across contexts and across strong social networks that promote interdependence in a cyclical way (Campbell and Mannell 2016;Kurtiş, Adams, and Estrada-Villalta 2016). Finally, there are limits to economic and social empowerment programmes' ability to fundamentally address inequitable gender norms and power relations within programmatic timeframes. Even with positive changes at the individual and relationship level, for meaningful change and sustainable impacts on intimate partner violence, individual-level interventions need to be complemented by structural initiatives that work at the community level to challenge broader patriarchal norms over time (Stern, McGhee, et al. 2020).

South African context of male providership
The notion of hegemonic masculinities was initially described in terms of configurations of masculinity that influence how individual men judge their 'success' as men (Connell 1987). Conceptualised in this way, hegemonic masculinity is an aspect of patriarchy that legitimates the subordination and control of women (Jewkes and Morrell 2010). In South African society, attributes of masculinity, such as physical strength, courage and an acceptance of hierarchical authority are valued, but importantly, there is also a demand that men are able to exercise control (over women and other men) . Furthermore, certain cultural practices place an expectation on men to provide for women that are institutionalised through the practice of lobola (bride price) before marriage (Hunter 2005). Consequently, there is an expectation of men needing to fulfil the provider role that determines the power relations between men and women relationships and their intimate partner violence risk (Conroy et al. 2015).
While dominant ideals of femininity in South Africa often entail subordination to men, women do not all experience controlling behaviours by their male partners to the same extent (Jewkes and Morrell 2010). Across the social spectrum there are men who adopt masculinities that incorporate counter-hegemonic practices, such as engagement in childcare and other caring responsibilities, or support for gender equality and opposition to violence against women. There are also women who are single mothers and economically independent of men (Walker 1990). Equally, it is important to consider these behaviours through a historical and cultural lens. In South Africa, the gendered division of labour has constantly evolved and shifted. Women have long engaged in domestic work and caringbut have likewise led households that are economically independent of men. The historical trajectory shows the dynamism and fluidity of gender relations, but it does not show that these women and men necessarily resist the fundamental gender order that subordinates women to men (Jewkes and Morrell 2010). It is possible to occupy opposing gender positions without challenging the gender order.

Study setting and purpose
We conducted this qualitative study alongside the quantitative follow-up study of the IMAGE programme in rural Mahikeng district, North West Province, South Africa (Ranganathan et al. 2019;Knight et al. 2020). The Mahikeng peri-urban and rural area is poor, lacking in infrastructure and amenities. It has high population density (Statistics South Africa, n.d.) and there are limited employment opportunities, with unemployment rates of 35.7 percent (and 47.1 percent among youth aged 15-34) (Statistics South Africa, 2016). A sizeable number of people depend on a cash economy supplemented by state-sponsored, non-contributory social grants (Statistics South Africa 2014).
The study was developed from discussions with project staff who highlighted the need for an in-depth understanding of the conflicting pressures on women's lives when enrolled in the IMAGE programme and their experiences of intimate partner violence. Using in-depth interviews, we explored women's perceptions on income generation through the microfinance plus programme; the effects on their relationships with male partners, including intimate partner violence; their notions of power; and their perspectives on men's reactions to their empowerment.

Study design: Intervention with Microfinance and Gender Equity (IMAGE) follow-up study
The study was embedded in the Intervention with Microfinance and Gender Equity (IMAGE) follow-up study whose purpose was to examine the effect of the scaled-up programme on women's experience of intimate partner violence, ten years after the original IMAGE programme (Knight et al. 2020). Women were eligible for inclusion in the main study if they were 18 years or older and had been enrolled for a year or more in the Mahikeng branch of SEF loan centres where the microfinance plus programme was recently completed. The IMAGE programme combined a poverty-focused microfinance initiative implemented by the Small Enterprise Foundation (SEF), with a ten-session participatory curriculum of gender training and HIV education known as Sisters for Life (SFL).
Loans are administered for the development of income generating activities with a group-lending model. Loan centres of approximately 40 women meet fortnightly in groups of five to repay loans, apply for additional credit, and discuss business plans, including savings options. Individual women run businesses, but groups of five women guarantee each other's loans. There are numerous loan conditions that women must adhere to: attending bi-monthly centre meetings, with few exceptions to skip the meeting; setting aside 2 percent of their loans towards savings; paying back loan instalments by 4-6-12 months depending on terms chosen by the group and cancelling or suspending of loan instalments only in case of death or illness. The loan interest rates depend on the loan amount and repayment terms, but for a loan of R1000 (US$55) the interest rate ranges from 16-30 percent (SEF Guidelines 2019).
'Sisters for Life' is a compulsory plus component of the programme that is put into practice during routine loan centre meetings. It is run just after the microfinance session by a separate training team and is facilitated by local women. Sisters for Life has two phases: phase one consists of ten one-hour training sessions, and covers topics including gender roles, cultural beliefs, power relations, self-esteem, intimate partner violence and HIV. Participatory methods aim to increase confidence and communication skills and to encourage critical thinking about the links between intimate partner violence and HIV (for the curriculum, see online supplementary Appendix 1). Since group-based learning can foster solidarity and collective action, phase two pf the programme encourages wider community mobilisation. All women undergo training on community engagement and leadership in four sessions over five months.

Data collection
The main IMAGE follow-up study (n ¼ 860) involved two rounds of data collection: a baseline survey (November-December 2016) immediately after participants received Sisters for Life training. This was followed by a second survey and qualitative interviews (November-December 2017), a year after the participants had completed the training and were still enrolled in the microfinance intervention.
For this qualitative study, we conducted in-depth interviews (IDIs) with 26 women participants who completed the baseline survey. To ensure the sample consisted of women who had experienced any type of lifetime intimate partner violence, we randomly selected 18 women from those who responded positively to the baseline intimate partner violence question adapted from the World Health Organization (WHO) Multi-country Study on Violence Against Women survey. Likewise, we randomly selected eight participants from those who had not responded positively to the baseline intimate partner violence questions. To maintain confidentiality, interviewers were unaware of the sampling method and the identity of the women who had responded positively or negatively to the intimate partner violence question. In this study, intimate partner violence encompassed physical, sexual and emotional abuse and controlling behaviours by an intimate partner (Ranganathan et al. 2019).
The IDIs took approximately 60-90 min, were conducted in Setswana in a private location, and followed a semi-structured, thematic interview guide. The semistructured interviews explored four domains: (1) participants' experiences of incomegeneration activities; (2) financial management and decision-making; (3) conceptualisations of power within the self and relationships; and (4) experiences of intimate partner violence and the intersection with male authority patterns.
Two female researchers from Social Surveys Africa, a research partner based in Johannesburg, conducted the interviews. Both researchers were involved in data interpretation and one of the researchers (MM) co-authored this paper. The researchers received ten days of additional training in the ethics of collecting violence data following WHO guidelines, such as access to support services through referral pathways. All interviews were digitally recorded, transcribed verbatim and translated from Setswana to English.

Analysis
A combined inductive and deductive approach was used in the analysis, drawing from the available data to generate reasonable explanations (Saldana 2009) The first author, with support from the two researchers who conducted the interviews, recorded initial findings and impressions from the data and field work. Using the guide, notes and existing insights from the literature on women's economic empowerment and intimate partner violence, we then developed an initial coding framework identifying key codes.
Four steps led to an analysis of the transcripts (1) reading transcripts until the content became familiar; (2) assigning codes to specific sections of text, refining the framework through an iterative process and identifying new themes; (3) categorising and displaying data to explore further dimensions and concepts in the data; (4) interpreting core meanings along with the different themes from the IDIs.
We also searched for relationships between themes or concepts identified from the analysis (Creswell 2003). The first author also discussed the findings with IMAGE programme and field research staff who helped with data verification and interpretation.

Ethics and informed consent
Ethical approval was granted by the London School of Hygiene and Tropical Medicine's Research Ethics Committee, and the Human Research Ethics Committee (Medical) at the University of Witwatersrand, Johannesburg. As part of our informed consent procedures, we asked participants if they would agree for us to include anonymised quotes from their interviews in potential papers.

Findings
In total, 26 women participated in the interviews. Participants' ages ranged from 24 and 65 years with a mean age of 46 years. The majority were married or living as married at the time of the interview (n ¼ 21, 81%), two were widows, two were separated from their husbands, one was divorced. All the women had childcare responsibilities either as mothers, grandmothers or stepmothers. Using the business loans provided by the programme, most of the women (n ¼ 22) ran small retail businesses such as fruit and vegetable vending, or second-hand clothing and tailoring businesses. These businesses were fully owned by women. On average, their net income in a month ranged from 80 to 1450 South African Rand (US$5-$90) depending on the success of the business.

Principal themes
Income generation, gender training and women's economic situation As our study sample included women who currently received or had received microfinance loans from the Small Enterprise Foundation (SEF), we first explored their reasons for entering employment and the importance of income in their lives. The dominant reasons were financial independence, realising opportunities and aspirations for a better life.
My reason was that my home needed an income. I wanted my house to be beautiful and normal, just like other homes. I did not want to go begging for food in my neighbour's house. Sheila, age 60 I wanted to be independent; I hated asking for money every now and then. I had to work because I had two kids to take care of. Grace, age 55 Most women reported that earning an income through their small business enabled them to save, and to afford items they need and want without resorting to usurious money lenders.
My income is important because whenever I want something, I can buy it myself. In previous years I would go and lend money from the mashonisa (loan sharks). But ever since I got involved with SEF, I'm no longer struggling like before. Lerato,age 40 Women acknowledged that even with small earnings they were better able to feed their children, provide minor items for the household, and realise their ambition of securing an education for their children.
That money helps me to run the household, I could buy a radio, or microwave … most importantly, I would like to see my kids getting a better education so that they can have a brighter future. Nonzane, age 56 Participants described an increased sense of confidence and satisfaction from running their small businesses and spoke about a sense of happiness and reduction in stress that came with an increase in financial resources.
[Being employed] changed me; I was very happy because I could feel that I was strong. I no longer have a lot of stress because I know that I get something every month end. I am now able to buy what we need with that money. When I was not working, I was so stressed on how I am going to support my family but not anymore. Jeannette, age 32 Some women expressed feelings of security from being able to save money for their family's future and from the safety it brought, in case their partners left them.
After joining SEF, I realised that saving is important, running a business is very helpful because even if your partner decides to leave you, there shouldn't be a gap because you have your savings. Esther, age 21 Relatedly, there was acknowledgement that income generation offered women a safety net from abuse.
When you don't have your own income to sustain yourself your partner will end up abusing you … As women, we need to regain our independence and work for ourselves instead of our partners. Boipelo, age 53 Some women raised concern about the fact that their businesses did not provide them with an adequate income to repay large loans. Others also expressed their concerns about the stress of loan repayments without a stable revenue source that was exacerbated when clients purchased items on credit. Women also had to adhere to perceived onerous loan repayment conditions. Further, many women appeared to be in similar types of businesses and in competition for clients. I joined SEF and started selling but people were buying on credit and they didn't want to pay me back. Some women from SEF were also buying on credit but they were not paying me. I felt as if I was working for free. Phodiso, age 47 Almost all participants expressed deep appreciation for the Sisters for Life (SFL) gender training. They described a strong sense of social support from loan members through regular group discussions that helped to build group cohesion and enabled them to manage the loans. This helped alleviate some of the stress associated with loan repayments. I have friends whom I share my problems with. SFL taught us not to bottle things up, it is better to talk to someone you trust about your problems. That makes me feel better. Thandi, age 51 Income and intimate relationships Regardless of whether they had experienced intimate partner violence, many participants reported improved relationships with their male partner once they earned their own income, especially when men were previously the only earning members of their household. This was particularly the case when partners were currently unemployed or temporarily employed. Some women mentioned that their partners did not humiliate them anymore and appreciated how women's economic contribution alleviated household financial stress.
He is no longer crushing my ideas. He is now supportive in every decision that I make at home and in business. Sylvia, age 46 Women reported cooperative interactions around household economic decision-making after their partners realised the value of the additional income and had dropped their scepticism of women's ability to run viable businesses and make household economic decisions.
It has made us get along, almost like we are friends now because he even gives me advice, like if something is not selling … .He is also bragging to his friend that he thought this SEF was a joke, but now he sees it's very worthy in our lives. Belina, age 53 Overall, many women mentioned feeling encouraged to continue being productive, with only two reports of partners' feelings of jealousy and one report of violence when the partner extracted resources for alcohol.
My business affected us a lot. In a bad way because I did not see any profit since my partner used to drink all the monies he collected from our customers. When I asked him what happened to his money he would beat me up. Cecilia, age 63 In some instances, women reported that sceptical partners were willing to divert from traditional gender roles and assist them with household chores, so that they could continue focusing on their businesses and meet household and child-related expenditures.
He is now treating me well … he loves me and he now respects me. We help each other at home, when I am washing the dishes he is cooking, when I am washing our clothes he is cleaning the house. We work hand in hand.

Lindiwe, age 47
Women's perceived power within the self and within the relationship A common theme was the link between feelings of independence, the ability to make decisions and feeling power within the self. Several women acknowledged that the receipt of loans, access to business opportunities and having the means to generate an income made them feel powerful. Some participants linked power with having disposable income for the accumulation of assets, home improvements or having access to services such as Small Enterprise Foundation in the community.
A woman who can stand on her own is a woman who has her own money, a woman who is independent, who decides on everything that happens in her household. Miriam, age 30 Participants described how the critical reflections sessions in the Sisters for Life training helped build self-confidence and self-awareness and enabled them to challenge power imbalances in their relationships. Almost all participants mentioned communication and problem-solving skills acquired through the Sisters for Life training for their ability to handle difficult situations or resolve conflicts.
That when you find yourself having a grudge with anyone, you don't have to keep quiet. SFL taught us to speak out about it, by confronting the person with who you have a grudge. We shouldn't keep it to ourselves but talk about it. Lerato, age 40 Finally, some younger participants attributed their ability to seek legal recourse and challenge their partners because of the training programme. I went to the police because I had knowledge; I wouldn't have gone there if I wasn't a member of SFL. You have to report the matter, irrespective of whether he is your husband. A husband has to respect her, she is also a human being who doesn't deserve to be beaten up. Poppy, age 24 Overall, most participants mentioned the ways by which the Sisters for Life training provided them with a platform for support and a set of diverse skills to maintain boundaries and mitigate triggers of intimate partner violence.

Perceptions of male authority and decision-making
A few women (n ¼ 7), particularly those who reported having experienced violence, mentioned their continued adherence to prevailing gender norms regarding male provision, despite their partner's appreciation of the extra income.
He was not happy about [my working] but he got used to it because he saw my salary and got happy at the end of the month. Husbands are still used to the idea that a man should work while the wife stays at home taking care of household chores and the kids. Disebo, age 56 For older participants especially, prevailing cultural norms meant respecting the traditional role of men as heads of households: women were expected to be submissive and show respect towards male partners.
Culture tells them to respect their husbands, no matter what, even if he was wrong Makhumo, age 63 There was acknowledgement that the power in a relationship tilted towards men, despite women gaining financial independence.
As a woman you have your limits, you must know how to respect your husband at all times. A husband has the power to beat up his wife, but I do not have the power to beat him up. Mmaphefo, age 58 Family responsibilities kept some younger women in abusive relationships. I wish I did not have a boyfriend. I wish I was only staying with my children. I just wish to be alone. He always threatens me that he is going to kill me and the children. I think my children will suffer if I break up with him … Yes, it is better for me to suffer instead of my children. Eva, age 30 Some participants' responses indicated traditional male authority was derived from men's role as the head of the household, despite women earning a higher income than their partners.
When a man tells you something, you ought to listen. There is a saying that goes "rre ke tlhogo ya lelapa" meaning the man is the head of the household. You must listen to the husband however you do have a right not to obey everything that he says. Phomolo, age 40 Relatedly, there appeared to be an expectation of male entitlement to women's earned money.
[If you do not give money to your husband] he wants that money. He wants it by force … Even if I try to ignore him, he would keep on reminding him to give him the money. Lindiwe, age 47 Interestingly, women reported that their partners' support towards their financial independence was conditional on whether it was deemed necessary; men assumed a provider role once they secured a source of income. However, some younger women were not willing to submit to male authority, indicating a certain level of self-confidence.
He was happy in the beginning. But in the long run after he got a job, he showed signs of someone who preferred me to be a stay-at-home wife. He wanted me to depend on him. But, I didn't leave my job. We were always fighting, but I told myself that I was not going to leave my job. Charlotte, age 39 Some reports indicated that despite women's financial contributions men controlled certain types of decision-making in the household. Decisions were taken either by the woman or jointly on matters perceived as minor, such as domestic responsibilities and childcare. However, for larger household purchases and in the case of a disagreement, it was ultimately the man's decision.
It is my husband. He always says "you have made so much money, with this money you should stock your snacks, with this money you should pay SEF, with this money you should do this and that." Kagiso, age 63 The SFL training curriculum also covered communication around sexual matters. There was an indication that for sexual decision-making, it was primarily a man's choice, irrespective of the woman's preference: some women were expected not to refuse sex with their intimate partners or would be coerced into having sex.
[If I refuse sex with him] Sometimes he fights me and forces me to have sex with him. He would remove my panties by force. It is as if he is raping me. Eva, age 30 Although women were economically independent, older women felt bound by traditional norms justifying male household authority and female subservience. Younger women tolerated a certain level of abuse because of family responsibilities. Furthermore, most household decisions appeared to be jointly made, other than decisions about large household purchases and those that were of a sexual nature.

Discussion
Findings from this study reveal the benefits of women's empowerment from microfinance loans and gender-training and the complexities surrounding the intersection with male authority patterns. The Small Enterprise Foundation microfinance programme opened doors for women to realise existing economic opportunities that in turn boosted their self-confidence. Furthermore, financial independence contributed to economic empowerment as women were in a better position to fulfil their caring responsibilities. Several women expressed feelings of happiness with their earnings as it reduced financial stress and relationship tensions. Financial self-reliance also enhanced women's understandings of power within the self. There was some acknowledgement that generating an income provided women with a safety net from abuse. But there were also concerns over the stress of loan repayments, the strict loan conditions and competition for clients in a difficult business environment. An important implication of this result for programmes is that providing cash transfers as seed capital for businesses instead of microfinance loans that have high interest rates may help alleviate loan repayment stress.
Women were overwhelmingly positive about the Sisters for Life gender training. They appear to have benefited from the social support with personal and business issues and this aligns with the quantitative findings from our study (Knight et al. 2020). In addition, the gender training played a key role in enhancing greater cooperation in the household. This finding aligns with research from Southern and Eastern Africa showing that social capital and strong relationships contributes to increased household harmony and raises women's self-confidence (Brody et al. 2015;Ranganathan et al. 2019;Stern, McGhee, et al. 2020). A similar microfinance plus programme in Tanzania found that gender training that sought to develop political consciousness enabled women to become more effective agents of change with partners, neighbours and children . In this study, women reported a greater sense of confidence to challenge their husbands and improve their ability to negotiate decisions in the household. Further, the training appeared to provide women with skills to navigate relationships, provide social support through the interactive group format and address various limiting normative beliefs that helped them to manage triggers of violence. This aligns with assertions from feminist theory that enhanced female autonomy and power improves women's negotiating power, due to its association with financial independence, feelings of improved confidence and a lower likelihood of tolerating male domination (Buller et al. 2018). Furthermore, women in our study appear to be expressing a form of 'distributed agency' that manifests gradually across time and with various support networks (Campbell and Mannell 2016). For instance, managing household tensions or intimate partner violence can be a slow process as these are but two of many challenges women face in their daily struggles for economic survival. Women also appear to rely on support networks, such as friends, other family members and children that provide the strength to cope with stresses in their lives.
Our data, however, reveal that despite women's sense of empowerment, they were constrained to some extent by the context in which dominant norms of male authority still exist. The provider role continues to influence relationships with women, particularly when men's own economic position is precarious (Jewkes et al. 2012). In this study, some women were still influenced by men's controlling behaviours and by expectations to remain submissive towards their partners. Older women appeared to be more compliant regarding this norm. Similar research in Rwanda suggests tolerance of divergence from norms around male provision and authority only for certain reasons, such as if a woman's job was of lower relative status and/or to meet basic household needs (Stern, Heise and McLean 2017). It appears that, in a few cases, violence was justified because of expectations of male household headship. This was particularly the case among younger women (aged <40 years) who appeared to tolerate abuse because of family and financial responsibilities, even though they were more likely to speak out against gendered expectations. In some cases, when men were supportive of their partners as breadwinners and shared household and childcare duties with them, there were less relationship tensions.
Although women were active in household and childcare related decision-making, there was some evidence suggesting that younger women were less able to exercise sexual decision-making power. This was consistent with quantitative findings from this study (Ranganathan et al. 2019) and research in East Africa (Stern and Heise 2019;Nyanzi et al. 2005). In this regard, there is modest evidence that inequitable gender norms support the right of men to initiate and decide the terms of sex. To address rigidly entrenched gender norms, especially within short time-spans, it is therefore important that men be involved to harness their support of women's empowerment. A trial in Rwanda (Dunkle et al. 2020) and research in the Democratic Republic of Congo (Vaillant et al. 2020), alongside other evidence (Armenti and Babcock 2016;Stern, McGhee, et al. 2020) have shown that when men and women are involved as couples, there is a significant reduction of intimate partner violence, including sexual intimate partner violence, as well as economic and psychological empowerment of women. There needs to be recognition however that some men may not be interested in participating in such programmes or that there may be a 'backlash' towards them, if men do not view the intervention as helpful. Hence, there needs to be a balance between making the engagement of men culturally resonant and allowing women safe spaces in which to be independent from men's influence.

Limitations
A number of limitations are worth mentioning. Although every effort was made to explain to participants that the research team was independent of the IMAGE programme, some individuals may have reported favourably on an intervention they valued. Additionally, the study design limited our ability to tease out programme effects by microfinance only versus microfinance plus. Beyond this, we do not include men in this research, as they were not part of the microfinance programme. Gaining men's perspectives would have vastly enriched this data. Finally, despite the extensive training of the interviewers, there may have been some under-reporting and reluctance on the part of women to disclose violence.

Conclusion
Study findings support the notion that microfinance plus can play an important role in supporting women's empowerment, especially with respect to strengthening women's agency and mitigating intimate partner violence. Furthermore, the group-based model of delivering gender training appeared beneficial for supporting women and should be considered by other similar interventions. Men's attitudes towards women's economic empowerment were perceived as ambiguous by participants: men tended to hold onto traditional gender roles associated with male provision and household headship, but nevertheless recognised the value of additional income in periods of unemployment. Future programme efforts should therefore consider men's involvement as willing participants from the outset to address household power dynamics that stem from traditional gender roles and notions of providership.