The moral economy of microfinance in rural Bangladesh: ‘Dharma’, gender and social change

: Microfinance is widely portrayed as a tool that empowers poor women to challenge constraining social institutions and make better, rational choices that will enable them to escape poverty. At the same time, the reported impact of microfinance is underwhelming, and scandals and poor practices widely documented. This then poses a puzzle: why do people still borrow from microfinance institutions? In response to this question, this article ethnographically examines the moral economy of microfinance in rural Bangladesh, examining how morality shapes the remit of legitimate economic activity and social change locally. For men, fieldworkers use stories of moral ‘idols’ to embed narratives of microfinance within existing rural social institutions to legitimise and normalize microfinance. For women, borrowing is presented as a means to perform ‘dharma’, a moral imperative to protect, retain or compete for male guardianship. The moral economy of microfinance is one that motivates people living in poverty to imitate and conform to prevalent gendered and hierarchical norms and values. These arguments highlight the significance of conformity to social and moral obligations in economic behaviour, and bear important implications for understanding the remits of social change.


INTRODUCTION
and Joynuddin (55) are locally known as idols1 in a rural para (neighbourhood) of Tangail district, Bangladesh2 .When they and I first met, in the winter of 2011, the idols recounted the story of how they managed to diversify their income sources and manage successful small-scale businesses.Sitting in his bari (village home) Joynuddin explained that nine years after he married Monowara in 1972, microfinance credit officers 'came to our house to inform us about loans'.'They explained their usefulness' he said, and that 'if my young wife takes it, it would help us improve our life […] and if we managed to return the first loan of BDT10,0003 they would give us a BDT20,000 loan next time'.Joynuddin remembers following the credit officers' advice not to spend the loan but to invest it in the mobile paan (betel nut) shop he had been running.In 1983, he borrowed BDT5,000 and repaid it weekly, then he borrowed a further BDT10,000 in 1985, and BDT20,000 in 1987.The latter two loans were invested in a small grocery shop he had opened with his son.The business was very successful and in 2007, Joynuddin stopped taking microfinance and sold his shop.The profit from the sale financed his son's migration to Malaysia.Bangladesh is the second largest microfinance market in the world, with some 25.6 million Bangladeshi borrowers (registered in 2017), and part of a global market estimated to reach 139 million low-income households, with loans worth USD114 billion globally (Convergences, 2018).
Stories like that of Joynuddin and Monowara helped fuel the expansion of the sector in Bangladesh, and similar cases have become part of a widely accepted discourse about the poverty-alleviating impact of microfinance.The representations of heroic entrepreneurs such as Monowara and Joynuddin began to encroach on development discourse and practice in the late 1980s.Initially microfinance was promoted as a technical solution and tool to support the poor negotiate adverse and imperfect markets (De Haan and Zoomers, 2005) and viewed as countering structuralist analyses of poverty.These stories portray microfinance's success as unbridling the entrepreneurial poor, generating a welcome 'revolution' in the mission to eradicate poverty.Although primarily concerned with financial transactions, microfinance acquired an orthodoxic and symbolic status as a primary catalyst of development; termed by some as 'microfinance evangelism' (Rogaly, 1996;Roy, 2010: 93).
As microfinance expanded over time, so did a growing body of research that raised concerns about the sector's rapid commericalisation, suggesting the mission and practice of microfinance had and sometimes polarising economic impacts of loans and debt for the very poor (Banerjee and Jackson, 2017;Bateman, 2012;Dichter and Harper, 2007;Hulme and Maîtrot, 2014;Mader, 2015;Maîtrot, 2019;Taylor, 2011).This then poses a puzzle: why, if these analyses are valid, do poor people still choose to borrow from microfinance institutions (MFIs, henceforth)?Common explanations of this in the literature point to the poor's limited access to alternative reliable sources of finance, the adverse inclusion or coercion of women borrowers by male family members (husbands, fathers, sons or father-in-law), and to rising debt levels and borrowing traps.Resonant with wider recent analyses of microfinance (Goodman 2017;Guérin and Kumar 2017;Kar 2018), this article suggests a different answer to this question: microfinance's expansion was only possible because it was embedded within cultural norms, rendering it a socially acceptable and indeed a moral imperative.
Drawing from a year long ethnography in rural Bangladesh, this article explores the 'moral economy' of microfinance by examining how gendered rationalities shape the meanings and values of economic choices and behaviours (Goodman 2017;Palomera and Vetta 2016;Simoni, 2016).More specifically, I argue that microfinance's expansion relied on an ecology of norms, values and practices, that encouraged clients to conform to existing class and gender inequalities.As hinted at in the opening vignette, although microfinance targets women, it is common for men to control loans and use them as they see fit.The role of MFI fieldworkers is crucial in this process as they construct narratives around idols that present microfinance as a moral imperative for poor clients: a means to pursue both moral aspirations and economic security.In this way microfinance participation becomes, de facto, a new norm to be adhered to, a moral compass for the gendered navigation of relational obligations.For women, borrowing then becomes a means to sustain, redeem or compete for male guardianship.For men, the option to borrow becomes a means to demonstrate their leadership and guardianship of their households.Crucially therefore morality or its performance is a medium of exchange that legitimizes the pursuit of change inherent in microfinance participation, thereby rethinking the place of individual agency in relation to social change in development (Arce and Long, 1999;Scott, 1976).
The article is structured as follows: the first section outlines how microfinance has been conceptualised as empowering women, enabling 'rational' choices and poverty reduction.It then surveys critical readings of such discouse, highlighting the concept of moral economies as a way of examining how morality shapes the remit of legitimate economic activity and social change.In the second section, I embed the analysis within the context of Bangladesh, examining the concept of dharma and the gendered social institutions women in Bangladesh navigate.The third section explains the methodology and background to this study.This is followed by two sections that develop key empirical arguments around the significance of conformity to social and moral obligations in the pursuit of security and change.Finally, the conclusion brings together different lines of argument to highlight how current development ideals around individual agency and empowerment often occur incrementally and if embedded within local moral frameworks.

AGENCY AND GENDER: MORAL ECONOMIES IN DEVELOPMENT
Central to the expansion of the microfinance sector globally has been a set of claims about the potential of entrepreneurs to use their agency to pursue economic and social change.This imaginary has in particular promoted the idea of women as powerful agents of change.In the 1980s, microfinance was perceived as representing an opportunity to improve the legitimacy and visibility of women's social struggles as well as gender inequalities in development (Fraser, 2009;Isserles, 2003).
Women and girls came to prominence in the global development agenda as under-tapped resources for change (Shain, 2013) and their individual agency was considered a strategic locus in development (Cornwall, 2007;Kabeer, 2011).At a moment when feminism was often associated with radical socialist ideas, the practice of microfinance (focusing on group-lending and women empowerment) projected the image of entrepreneurial rural women, particularly mothers (Hossain, 2017: 90) leading the social transformation by emancipating themselves from and rupturing existing dysfunctional and oppressive institutions.Rankin (2001) calls an ideology of the rational economic woman.Rational choice theory is underpinned by long-standing ideas about how the maximisation of utility functions drives individual decisions and behaviour.Focusing on answering the 'how' rather than 'why' people make certain choices, this approach focuses on economic utility and, by implication, regards individual tastes and preferences as exogenous factors (Kabeer, 2010).This theory of choice takes the decisions of individuals as its sole unit of analysis, giving little or narrow consideration to the power dynamics that shape the decision-making of its members (Ibid).Microfinance then represents an incursion of a gendered rational choice theory into development thinking in that it aimed to enable women to make better choices for their household.

The expansion of microfinance coalesced historically with what
Here, the targeting of women was justified by the fact that women were found to better repay loans than men, and therefore conform to notions of rational economic behaviour.Furthermore, loans allocated to women were also found to have better wellbeing outcomes for household members (Ibid).
As such, microfinance 'elevated' the position of women, who 'empowered' with loans, represented a 'better developmental investment ' (Berkovitch and Kemp, 2011: 160).This utilitarian view of poor women entrepreneurs has however been widely critiqued on a number of fronts.Notably, it has been argued to depoliticize the nature of women's relationships within the households and with the market (Elyachar et al., 2005: 29-31), and neglect the multi-dimensionality of women's experience of poverty (Saith, 2007) by emphasizing residual and absolute deficits rather than its relational characteristic.By contrast with rational choice theory, more anthropological understandings of agency put social context at the centre of decision-making, and considers how norms, beliefs, routines and customs operate to shape preferences, attitudes and behaviours (Kabeer, 2010).This view considers that within any household, an individual member's sense of self-worth, role and identity are always shaped by others.One implication of this perspective is that the economic behaviour of a household is not then simply seen as the aggregation of decisions made by its individual members, but reflective of negotiations across a household's hierarchy.For example, the distribution of labour, free time and resources within the household will depend on a number of factors including the perceived contribution made by specific members to the household's welfare, and its valuation.Crucially, this approach helps explain how (in)equalities within the household can be legitimized.Situating agency within social institutions therefore helps illuminate how the nature and quality of institutions (i.e.'the rules of the game') entangle the political, economic and financial realms constitutive of the 'social order' (Mosse, 2010;Townsend et al., 1999) in ways that shape opportunities for social change in development (North et al., 2009: 251-71).
Contrary to the narrative that microfinance helps the 'deserving poor' caught up in, or held back by the 'social' (Gray, 2010), a growing body of literature has examined the complex ways in which microfinance in practice has interplayed with social hierarchies.As Kabeer (2001: 80) argued two decades ago, any attempt at conceptualizing empowerment needs to be grounded 'in an understanding of the relationships of dependence, interdependence and autonomy which characterize gender relations in different cultures, the structures of risks, incentives and opportunities which they generate'.Rather than empower women, it has been argued microfinance and wider povertyalleviation efforts can make women 'subjects' of development (Fernando, 2006;Kabeer, 2005).
Microfinance's effects on intrahousehold dynamics are often argued to exacerbate existing oppressive social institutions and further entrench the gender and class-based inequalities it is meant to tackle (Fernando, 1997, Karim, 2011, Bateman and Maclean, 2017).It can erode the webs of interlocking ties and reciprocal networks upon which women's status and survival strategies often rely (Maclean, 2010;Molyneux, 2002).As such, people's disposition towards and experience of risk can be argued to be gendered (Maclean, 2013).Where microfinance debts intertwine with and capitalise on the invisibility of women's reproductive work, and 'social capital', it can lead to 'dispossession' in rural communities (Paprocki, 2016).
One recent line of argument developed in the context of South Asia is that for microfinance to thrive, it must in fact serve or be adapted so that it can serve existing social institutions.In the context of India for example, Kar (2018: 109) argues that microfinance in Kolkata is 'domesticated' into women's lives, 'enfolded into women's existing schedules of domestic labor, [thereby] limiting its impact on women's empowerment'.Goodman (2017) similarly examines how for Kumaonis in Uttarakhand state microfinance has been altered to fit existing 'exchange relationships' that govern the life of the family and community.It 'became another node in the already existing networks of assistance' (Ibid: 362), and gained popularity on this basis.For Guérin and Kumar (2017: 742) microfinance in rural South India is 'both shaped by and constitutive of local structures of power', meaning that microfinance strengthens patronage driven and patriarchal institutions.
In examining how microfinance is adapted and legitimised locally, Kar (2018) and Goodman (2017) have drawn attention to the role of morality in this process, deploying the concept of 'moral economy'. 4The concept of a moral economy refers to how shared ethical frameworks legitimise economic activity, enabling us to examine processes and remits of change.Whereas it has been argued by classical sociologists that development is a process that disembeds individuals from local rules, social institutions and contexts (Polanyi, 1944), a moral economy framework highlights the entanglement of economic activity and social institutions the rural poor have to navigate.By contrast with a 'homo oeconomicus' characterisation of agency, the notion of moral economy combats 'crass economic reductionism' that obliterates 'the complexities of motive, behaviour, and function', constructing 'an abbreviated view of economic man' (Thompson 1971: 78).In early and prominent usage within academia, the term was used to highlight how moments of social and political rupture occur with the expansion of capitalist relations, and conflict with other notions of moral exchange and activity (Thompson 1971, Scott 1976).It hence has served as a way of examining how socio-political disjuncture can be an indicator of the 'conflicting co-presence' of competing moral frameworks (Simoni, 2016: 462).
Equally however the concept of moral economy can be used as a way of examining how ideas about moral behaviour intertwine with economic relationships in ways that maintain rather than rupture the social reproduction of hierarchy (Palomera and Vetta 2016).It can hence 'highlight the ambiguous logics and values that guide and sustain livelihood practices, by looking at the dynamic fields of struggle around the boundaries of what is good and acceptable, their power hierarchies and the political projects they might inform' (Palomera and Vetta 2016: 415). 5Here I hence adapt Thompson's framework to accommodate for a more relational perspective that replaces the examination of one particular population group (i.e., the poor) to the study of the values, meanings and practices exchanged and emerging from the interplay between the poor and microfinance fieldworkers.The exploration of the 4 These authors apply the concept of moral economy in different ways.For Kar (2017) it is applied as a means to explore the behaviour of credit officers, while Goodman (2017) uses it to refer to the relationships of obligation and exchange that govern networks of assistance. 5Early usages of the concept highlighted the moral economy 'of' particular groups (Thompson 1971, Scott 1976).One interpretation of this has been to say that moral economy as a concept is thus particular to pre-capitalist forms of hierarchy and obligation.By contrast, more recent usages of the concept have highlighted the plurality of moral economies, and their modes of operation in any given field as highlighted by Fassin (2009).particular ways in which the 'social' and 'economic' are entwined represents a framework for approaching questions of why social actors pursue or resist change, and how this is conceptualised.

DHARMA, WOMEN AND CHANGE IN BANGLADESH
To begin to examine the moral economy of microfinance in rural Bangladesh, we need to examine the prevailing social institutions, which are characterised by pervasive inter-personal hierarchical networks of patronage (Devine, 2006;Lewis, 2011).The concept of dharma, referring to 'the proper order of things', is useful to understand the way inter-personal hierarchical relations and networks of patronage are legitimized in Bangladesh (Devine and White, 2013: 134).Etymologically, the word dharma is derived from the Sanskrit root dhr, meaning to sustain (Inden 1976: 19) which is suggestive of the attributes of 'what makes the life of a community persist' (Kotalova, 1996: 46).In practice, it comes close to Bourdieu's notion of habitus in that it is embodied in people's habits, rituals and lifestyle (Bourdieu, 1977).Conventionally translated as religion, dharma is a broader ontological and moral framework that encompasses a set of social norms and values that produce individual dispositions and define the socially-accepted contours of individual agency (Devine and White, 2013).As such, social institutions and the hierarchies and obligations they entail in this context can be seen as having an inherent moral meaning.
The nature of agency is deeply gendered in this context.A prominent way in which this is felt is through the idea of purdah, the practice of sexual segregation, which has a long history in Bengali Muslim, Hindu and Christian communities.It implies that women's honour is preserved by limiting their movement and interactions outside the private sphere (Rozario, 2006).While honour (izzat) is often seen more as men's responsibility, shame (lojja) or decency is often seen as women's attribute (Rozario, 1992: 86).Purdah represents an ecology of norms that govern the respective roles of men and women in society in ways, often internalised, that influence women's capacity to choose for themselves both symbolically and materially (White, 1992).Despite the forces of globalization and urbanisation, purdah largely remains a prevalent socio-cultural and ideological institution all Bengali women (extreme poor, poor, lower and middle-class women) navigate and negotiate (within themselves and with others) with varying degrees of agency (Kabeer, 2010).
Within this order, it is important to examine how economic relationships and social networks are construed in moral terms and embedded in one other.Because women's wellbeing, access and control over resources is often mediated by men (Kabeer, 2011), women play an important role in sustaining the reproduction of purdah as a modus operandi.For this reason, women's participation in development initiatives has at times been contested and opposed by more traditional forces including Islamic groups who perceived women's involvement in development as a violation of the norms of purdah, thereby threatening the proper order of things (Naher, 2010).
Although it is true that women play a central role in maintaining dharma, it is also true that the performance of dharma is a legitimate source of power for women (Blanchet, 2008).Visibly demonstrating adherence to the proper order of things is crucial for women.Women often work hard at preserving and developing good relationships with male patrons, protectors and guardians (father, brothers, husbands, and sons later on in life), whose moral duty is to provide them with material support and relative, if provisional, security (Kabeer, 2011).The mythico-ritualistic power of marriage is foundational to people's sense of dharma and therefore constitutive of the reproduction of the moral order of things.For rural poor women therefore, marriage constitutes a pillar of the undisputed doxa, dharma, that carries symbolic and practical significance, as it places them within a new hierarchical system of obligations, which in theory guarantee their husband's guardianship (Rozario, 1992;White, 2017).As the distribution of property is biased towards male household members through a dominant patrilineal structure (Karim, 2008), women are likely to become poor in the event of marital breakdown, illness or death of a male income earner (Maîtrot, 2017).In return for guardianship therefore, married women maintain a practice of purdah in ways that protect their izzat and uphold their moral status as well as that of their guardians' (Kabeer, 2001;Rozario, 1998).6

BACKGROUND, METHODOLOGY AND DATA
This study is based on a year long ethnography of the interactions between microfinance institutions' agents and villagers across four villages in the district of Tangail conducted in 2010-2011.Located about four hours north of Dhaka in central Bangladesh, this area is now considered one of the most densely served districts in terms of MFI coverage and referred to as 'the home of microfinance entrepreneurship' (Chaudhury and Matin, 2002: 47).All families in the four villages of my study became acquainted with microfinance organizations in the 1980s.
A common joke among researchers and villagers goes that in this particular district of Bangladesh, households are so used to being surveyed by researchers, that they could fill in questionnaires themselves.From my experience of working in Bangladesh, I was aware that an NGO affiliation could exacerbate this problem and bias my interactions with local residents and MFIs employees.I therefore conducted the research as independently as possible, free from any local institutions, introducing myself as a researcher from a foreign university.With a research assistant 7 , I conducted mixed-methods research between 2010 and 2011.The primary research methods were participant observation and informal interviews with villagers and MFI staff.In addition to this, I conducted six focus group discussions with current clients, nine in depth formal interviews with former clients and key informant interviews with local leaders.
When I began the research it was common for particularly curious locals to follow me through the towns, villages and on occasions into respondents' houses.As the novelty of my presence wore off, people became less interested and aware of my whereabouts and I was able to interview people more automously and privately, one-to-one, in their homestead verandas, in their homes, in paddy fields, at tea stalls and village markets.As much as possible in this context, I maintained a high level of confidentiality, whilst being careful not to bother informants in the evening whilst they were tired and wanted to rest after a day of work.All respondents gave their informed consent to participate in the study.The information collected was triangulated by asking similar questions to a range of key informants including police officers, village leaders, other family members, religious leaders, shopkeepers and microfinance agents.Fieldwork was conducted in Bangla.The data was transcribed, translated into English, entered into NVivo and coded thematically.
Using qualitative methods to explore the 'seamless fabric' of lived experiences (Latour, 1993: 7) in this way, is important in challenging proliferating interpretations and assumptions about social change in development (Arce and Long, 1999).Researching and interweaving multiple narratives of the lived experiences of microfinance clients -both past and current 8 -was crucial to examine individual agency in the face of the change and authority microfinance agents embody (De Sardan, 2005).These narratives allowed me to examine the various meanings, including ambiguities, of individual agency.

DEMONSTRATION, OBLIGATIONS AND IMITATION
Contrary to the discourse that microfinance helps individuals overcome exclusive and imperfect institutions, this section shows how the moral economy of microfinance relies upon and reinforces existing social institutions.Through a process of demonstration, MFI staff construct moral ideals and figures of 'development success', locally known as idols, to legitimize microfinance interventions.
Idols are idealised figures that represent the 'good life' most people desire and are used by MFI agents as examples for other clients to follow 9 .They are seen as embodying both virtue and material prosperity, where the latter can be cast as a fruit of the former.The erection of simple figures to be particularly in relation to women's discussion of sensitive issues.In practice, I conducted many interviews partly or entirely alone to better facilitate such discussions. 8Considering the small number of former clients in the studied population (28 out of 490), in-depth focus groups seemed an apposite tool to understand experiences and choices. 9It was unclear from the research whether this term stemmed from the villagers or microfinance agents however was widely used by both.revered by their contemporaries for their economic and moral merit reveals a creative bricolage that intertwines powerful economic motivations, spiritual sentiments and religious norms surrounding duty and piety.Idols make 'better choices' and are able to fulfil their moral obligations to their families: for men this means successfully looking after their families, and for women it means performing and conforming to moral expectations.By linking microfinance and morality in this way, borrowing from MFIs is normalised and becomes a moral act in its own right.Not taking out loans or not engaging with MFIs is hence seen as a risky strategy, an immoral choice, and as the wrong thing to do.As a result, microfinance for the rural poor takes on moral meanings embedded within traditional social institutions.

Valuable idols
In all four villages, microfinance was first introduced by agents who were strangers to the local communities.In many ways, the idea of poor women taking loans from formal institutions, headed up by strangers, and often male strangers, belied the proper 'order of things' (Devine and White, 2013: 134).Several Regional Managers and Branch Managers reported that the major challenge they faced was to preserve a good relationship with the general public.A Regional Manager who has worked for a MFI since 1974 reported that at the start, microfinance organizations put greater emphasis on the need to communicate and socialize with the public 'because in the end', he explains, 'it is the responsibility of the loan officers to convince clients to participate'.Respondents described that when microfinance agents started recruiting clients in their village, people were wary of being visibly associated with something unknown and potentially controversial 10 .An elderly respondent and former microfinance client remembered back to when microfinance agents first came to his village: There is a proverb in Bangla: 'if you put a candle close to a fire it will melt.'It is the same between men and women.Credit officers come to our houses to collect loans, but you can never say who they really are inside.Over the course of a year, this man will develop a relationship with the family, be in regular contact with its women.You do not know what sort of glances he will give and to who and it is natural that a young man's behaviour can slip.This creates an opportunity to do something wrong-ethically and religiously.So if we have the opportunity to avoid it, why don't we?
Microfinance agents are aware of the perception that microfinance loans may challenge power dynamics within households, making male household members particularly wary of microfinance and 10 It can be noted that a further means by which microfinance was legitimised in this context was through the development of savings groups among clients.When explaining their motivation to borrow and livelihood strategies, respondents rarely made reference to the influence of group dynamics as such, and placed far greater emphasis on personal relations with specific fellow borrowers and with MFIs.therefore less likely to participate.Male household members, conforming to their guardian role, act as gatekeepers and are generally distrustful towards unfamiliar men entering their homestead.Indeed, men sometimes express frustration at the fact that loans are given in women's names, as a middleaged man Masud described: The loans should be in the man's name, because women stay under the men and they don't even earn.They don't even know how to earn.I don't know what they think, but to me it should be in the man's name then they will feel the pressure and will work hard.
The women live in the house and feel afraid when the people of these institutions come to their house.
Such attitudes constituted a considerable challenge for microfinance agents11 .Taking a loan with interest is often perceived or presented as conflicting with Sharia law and the Islamic banking system, which is predicated upon challenging the asymmetrical contractual relationship between borrowers and lenders.Risks borne by borrowers in microfinance loan configurations are generally considered disproportionate according to Islamic law.Such principles and moral value systems permeated religious boundaries, with other religious groups also reportedly feeling reluctant to participate.In this case, it meant that initially, the large majority of rural households were reluctant to borrow, casting microfinance borrowers as immoral.
We know that global development projects develop local legitimacy through adaptation and 'translations' to fit specific contexts (Ferguson, 1994;Mosse and Lewis, 2006;Radhakrishnan, 2015;Sharma, 2008).In this case, faced with the challenges explored above, microfinance agents over time deployed discursive tactics to normalise its practice by aligning it to traditional gender norms, thereby skilfully navigating the local social institutions.It is important then to recognise that the discretionary power of credit agents is often central to shaping the moral economy of microfinance in this context (and can be argued to create a 'practice drift' at the frontline of MFIs in Bangladesh (Maîtrot 2019)).
This normalization was achieved as part of the wider commercialisation of the sector, in a variety of ways.The first encounter between a microfinance agent and a client typically occurs at the potential client's homestead with credit officers addressing male household members and often deliberately excluding the future client (i.e., the woman) from initial exchanges.This is in part a tactic that shows respect to the head of the household, while allowing the agent to introduce him/herself.Speaking to the female members of a household without requesting the household head's permission would be considered an insult to him12 .
Microfinance agents make a rapid informal judgement about sources of potential tensions between their aim to formalise loans and a household's view of the 'proper order of things'.This judgement takes the level of vulnerability and poverty of a household into account judging by the size and condition of the homestead, the quality of clothes worn by the household members, and number of assets populating their homestead.Another element affecting the agents' judgement is the religiosity of the household members.This requires recognizing contextual religious and cultural codes, 'the material and ideological resources and practices of households and families' (Standing, 1991:12).As well as reading into people's verbal and body communication, agents can observe the degree to which purdah is enforced, religious clothing can be used as part of a wider set of symbols to signal piousness and religiosity (hijab or veil, tupi, Islamic beard and so on).From this information, agents infer the nature of kinship and interpersonal relationships and gender hierarchies characterising the household.
In order to reconcile microfinance with local institutions characterising a household, agents often portray microfinance borrowing as serving the existing social institutions and their associated norms and morals.A key means used by agents to recruit as well as retain microfinance clients is constructing and telling stories of idols.Idols embody development success and depict accounts of individuals (generally male) or households who have established prosperous businesses that enable them to provide for their families.In this context idols take on almost legendary status with tales of success being spread to nearby villages.Clients may not have met these idols, but they will have heard of them and of their success stories 13 .Nusrat Jahan, an elderly widow, described how microfinance agents 'inspired' poor households to take loans through song and stories about idols.
She remembers Grameen Bank coming to her village in the 1980s, and singing 'a song about how loans can change our lives.That persuaded us to join and I managed to make a group of five members'.During participant observation, as well as the nine focus groups, respondents told stories of households who became dhoni manush (wealthy people) by using loans.Nazma (35) a currentclient of four MFIs, explained: 'We see a few successful people here and there, if you go to this place in Uttarnagar you will see lots of cows over there.This one guy started with BDT30,000 and now he has grown so big.He has a huge farm.He is rich.These stories convince us to take money [...].' Accounts of idols generally focus on the successes of men.This rhetoric plays to the gendered image of the self-made entrepreneurial man (Gunnerud, 1997), calling on their family obligations to enthuse them to participate in microfinance.Men living in poverty often express the view that they are not 13 This has parallels to the sociology on knowledge diffusion in successful agricultural extension programmes or family planning in rural Bangladesh.A common practice is to use entrepreneurs as knowledge diffusers about the new technology or practice (Mita and Simmons, 1995).able to meet their moral obligation to attend to the needs of their mother, father, wife and children.
Narratives about idols portrayed as honourable and principled men able to fulfil their moral duties with regards to their dependents, make borrowing and lending from microfinance institutions an obvious or rational option.Microfinance participation therefore becomes the magic bullet for men anxious to fulfil their moral obligations towards their families.Refusal to participate is disapproved by credit officers, held as an immoral decision, a betrayal of moral responsibility.Most male respondents reported having been encouraged by credit officers to allow their wives to access loans so that they (husbands) could then use the money as they see fit.In this way, their wives' modesty (extra-familial interactions outside their para, neighbourhood) is not affected and men are seen as good guardians, doing all they can to look after their households.Indeed, many men considered it a point of honour that their wives did not directly use the loan themselves, maintaining the norm that important financial dealings are men's responsibility.As a result, taking on loans via wives gave men the opportunity to retrieve a 'moral status' as provider and guardian of the household.This was important for men themselves but also for the household's reputation in the immediate community as 'male guardianship' is considered socially an indicator of 'the order of things' (Maitrot, 2017).Such stories of idols are therefore powerful idioms and symbols that legitimize and diffuse the idea of borrowing in order to improve livelihoods as socially acceptable and morally compelling.
Microfinance agents play an active role in constructing and instrumentalizing idols as symbols of success.The case of Joynuddin and Monowara illustrates this well.In 1989, shortly after Joynuddin had decided to quit his MFI (on his wife's behalf) because -as he described it -his business was prosperous and he no longer needed it, another MFI visited him to recruit him as an idol.They had heard about him and wanted to convince him to take more loans.So, he agreed and borrowed BDT5,000 and then BDT10,000.He managed to repay the latter loan but with great difficulty.Once he had repaid, he again decided to quit.However in his view, microfinance agents did not want to let him go because he was still a 'good client' and the community looked up to him and considered him an example of success -someone who had met all his moral obligations.Holding on to the couple's savings 14 , the microfinance agents told him that by quitting he would make it more difficult for them to recruit and retain clients 15 .As this case suggests, idols are critical narratives that legitimise and normalise microfinance activities. 14In Bangladesh throughout the 1990s and 2000s, microfinance borrowers generally needed to make compulsory savings for a period between 40 and 60 weeks before being granted a loan.While the compulsory savings component was originally introduced to protect institutions against defaulting clients, many microfinance agents breached their operation manual by refusing to return savings to 'good clients' in an attempt to discourage them from leaving their MFI (Maitrot, 2019). 15After months long negotiation with the MFI, Joynuddin finally recovered the savings and quit the MFI.Joynuddin and Monowara's success rested in their capacity to make astute judgements about opportunities and risks.
The identification and reference to idols was effective in all four villages I studied, influencing villagers' aspirations and behaviour.While the decision to participate in microfinance is often seen as indicative of 'progressive' character traits in clients (risk taking, entrepreneurialism, challenging social norms and so on), in the villages I studied a large part of the decision making process for individuals was driven by the imitation and following of examples set by peers (Khandker et al., 1998: 108).Indeed, the majority of microfinance clients (or their husband/father/brother or other male guardian) perceived loans not as an opportunity to do something new or innovative, but to reproduce or imitate what an idol had previously done.These dynamics are not dissimilar to those identified by Kar and Schuster in India and Paraguay (2016).In the area where I conducted the fieldwork, it was for example common for most respondents' husbands to indicate that they wanted to use loans to start a chicken or cow farm 'like this man in Dhanbari [a nearby town]'.Or as one respondent, Kamal, put it, 'I simply followed the others.Whatever they were doing I trusted them and I also did the same.If I saw someone taking loans and doing well, I also followed him to get the same successes'.These examples hence indicate how MFIs relied on narratives of idols to minimize the perception of risk associated with becoming an MFI borrower 16 (Radhakrishnan, 2015;Weber et al., 2004).In so doing credit agents pursue their own interests and the interests of their organizations by presenting credit borrowing as following the example of morally good community members.

Borrowing as caring
Normally, when a male household male member agrees to participate in microfinance, a woman in his household is required to consent and 'give her name' 17 so he can access loans.Most of the clients interviewed were married and had moved from their natal village (her baper bari or 'father's village') to their husband's (patri bari or husband's village) as is local custom (Balasubramanian, 2013).In this process, they join a new network (Chowdhury, 2010;Rozario, 1992), assume a new identity as a notun bou (young bride) soon to become a mother (White, 2017).As discussed earlier, with marriage comes a set of obligations and care duties towards the new family.For women in the villages studied, a motivation for giving formal consent to MFIs was to perform to expectations and fulfil their moral obligations.This often took the form of safeguarding morality, improving their status within the household, and ensuring bad reputation (durnam) does not fall on the household. 16Some authors have argued that rather than comparing alternatives before making a choice, decision-makers often simply imitate the choices made by others Pingle, M. (1995) 'Imitation versus rationality: An experimental perspective on decision making', The Journal of Socio-Economics 24(2): 281-315.Imitation is often perceived as relatively less costly. 17The phrase 'give their name' is used locally to refer to becoming signatory to a loan.This expression is often used by clients themselves, reflecting the commodification of women's names in the process of accessing microfinance loans.As has been widely documented, although MFIs largely target women, men remain the primary loan users (Goetz and Gupta, 1996).
For women, accessing loans is often considered a way of demonstrating love and respect to their guardians.The 'order of things' within households is implicitly defined by reciprocity through which individuals prove the legitimacy of their membership.Within large extended households for example, the legitimacy of a couple (son and daughter-in-law plus dependents) within the family unit is often at risk if those are perceived to be insufficiently contributing to the economic life of the household.Like many women I met, Naseema said she had taken MFI loans 'out of love for my family, because my husband requested I do so [my emphasis added] it'.A particularly common motivation for taking loans was the health, medical needs and wellbeing of elderly shashur and shashuri (father and mother-in law, respectively).In this context, fulfilling and performing a 'moral duty' by adhering to patriarchal norms and agreeing to take on loans, is imperative for women seeking to maintain good relationships within the household and preserve or improve their access to security through male guardianship.As a result, failing to meet expectations (or failing to demonstrate a willingness to meet them) is often interpreted as a transgression of social norms and risking the withdrawal of male guardianship.
Over time microfinance has also become a means through which women compete and demonstrate their deservedness of guardianship.The fact that other young bou borrow, creates an environment where comparisons are drawn across households, and of course microfinance borrowing becomes a socially accepted practice and a norm.Female respondents often explained that they felt pressured to take loans, and that taking loans was a new means of fulfilling social expectations, that includes their expectations of themselves.The extent to which microfinance participation has become a norm has increased as the market has become saturated, putting pressure on credit agents to recruit and retain clients (Maîtrot, 2019).Between five and eight MFIs were operating in each of the villages studied because the appetite for microfinance is high and competitive.This is not uncommon in Bangladesh, with the poor often having better access to microfinance than basic state services (Rutherford, 2009).
Credit agents are therefore in fierce competition and seek advantage by reminding borrowers of their 'duty' to their households, and how refusing to participate will harm their personal reputation in and outside of the household.If this approach fails, then credit agents will approach husbands or male guardians and remind them that their wives may be falling short in their duties.With two-thirds of the entire populations of the villages actively borrowing or having borrowed from microfinance organizations, the decision to borrow therefore represents a means to adhere to local norms.
The effects of this approach can be strongly felt at the household level.For example, daughters-in-law commonly display and express love by sacrificing their wellbeing for that of their guardian(s) and children, and clients reported that by refusing to participate, their guardian would compare them unfavourably vis-a-vis other bou in the vicinity or daughters-in-law in the households who did agree to take or top-up loans.We thus see a competition between female members of the households, anxious not to be seen as 'less caring'.Credit agents and husbands reinforce these expectations.Some women reported that when credit agents heard of their reluctance to 'give their name' for their husband to access loans, agents would 'convince' them, either directly or using husbands as intermediaries, by emphasizing their duty of care towards their family and their moral duty to support their husband.
In the case of elderly widows, this power configuration might differ.In poor households, after the death of the father, the mother will often become dependent on the guardianship of her son(s).If the parents were living independently from the son's household, the mother will often try to recover her son's guardianship because he would traditionally have a moral duty to care for her.In this setting, daughters-in-laws can refuse to 'give their name' for a loan and encourage their husbands to 'request' their mothers to borrow for them.Since microfinance debts are often written-off after the death of the client, asking an elderly woman to take out a loan could point to a pragmatic shift of caring responsibility on the part of the daughter-in-law.Loans offer an infrastructure for sustaining or regaining male guardianship, which ultimately reproduces, and shapes gendered norms, institutions and hierarchies, including sometimes by shifting responsibilities for care.In any case, women understand that by refusing to access loans they put themselves at risk of being denied guardianship and protection.
Although traditionally, making major financial decisions is not considered a 'female' responsibility or obligation (Kabeer, 1999), women's borrowing is at times also motivated by an aspiration to lead or support the pursuit of small business ventures.Married women often do not want to be seen as managing their own financial affairs as that is considered a husband's duty.Often, they described that performing humility and risk aversion constituted important features of the 'good wife'.Yet, this did not mean that some harboured aspirations to pursue or support economic ventures.Women reported that they too could contribute financially to the fortune and wellbeing of their households.Loans were therefore used by women to purchase land, rickshaws, and livestock or rent and work land (borga, sharecropping).In practice however -and as has been widely documented -loan use is often diverted towards so-called 'non-productive' purposes, or 'consumption loans' such as medical costs, emergency home repairs, wedding costs and dowry (Ahmed et al., 2001;Goetz and Gupta, 1996;Kabeer, 2001;Roodman and Morduch, 2014).These uses of loans are often contentiously presented by researchers as 'irrational' (Carvalho et al., 2016), but also by other villagers as 'bad choices'.
Presenting these choices as such because they do not strictly conform to the idea of what being an entrepreneur means, however de-politicize the context within which poor women make 'choices' and ignore the moral trade-offs they often face.Many female respondents explained 'not being able to use the loan properly' because they have to take care of their children, husband and in-laws.It is for them a moral and pragmatic pursuit of the duty to care.Women expressed feeling caught between two guardianship.In this context, I find women's agency to be shaped by complex inter-personal gendered hierarchies and webs of inter-dependencies.These social relations influence the terms on which women perceive themselves and their actions.Their perception of who they are, or want to be, is tied to how they behave towards others.This invites us to consider agency in relational term, always embedded within and shaped by a process of social relationships.As Burkitt (2015: 332) suggests understanding agency as such acknowledges the 'transactional' dimension(s) in people's action.
Further, this article shows agency is both constrained and constituted by the performance of morality, which is related to disposition to risk and expectations of loyalty, two notions that bear specific gendered meanings.Relations of kinship and male patronage structure 'the order of things' for women and constitute the main means of wealth redistribution (Devine and White, 2013;Lewis, 2017;Van Schendel, 2009).The choice to borrow is not a proposition to challenge established norms and values but rather an exercise of social imitation and moral conformity necessary to secure guardianship.Making a choice, as Kabeer has argued (1999: 497), implies being able to choose from a number of limited viable alternatives.In contexts of poverty, exclusion and patriarchy, the option of viable alternatives cannot be presumed.Expectations around morality or the 'right order of things', mediates choice and as this article has argued, can increase women's responsibilities without necessarily transforming the norms or values that discriminate against them.The analysis thus calls for greater attention to be given to the significance of moral economies as sites that enforce social imitation and moral compliance.It also reinforces the idea that current development ideals around individual agency and empowerment can only be enacted incrementally and if 'morally' legitimized at local levels.This challenges thus key assumptions about the relationship between development and social change.

Reviewer 1:
The argument often takes a historical/diachronic angle, relying on oral history of the early days of microfinance in the 1970s/80s.However, the analysis tends to be quite synchronic, and does not attempt to explain the shifts that appear in the data (e.g.microfinance from immoral/treated with suspicion to performing 'guardian role' for men, this is all on page 10).The author covers this ground so quickly that the shift from guardian/keeping loans out to guardian/accept loans as a market of success and aspiration is not terribly clear.That agents attempt to normalise microloans comes through the story, but how and with what effects to the existing moral frameworks of guardianship (how normative ideals of senior masculinity have changed, for example) would be usefully discussed with more clarity.
I would like to thank the reviewer for this comment.While I agree that it be interesting to produce a more balanced gender analysis on the interactions of microfinance with local moral frameworks, it is beyond the scope of this article to cover this extensively.I am working on a project that aims to analyse how development interventions impact ideas about masculinity.The section mentioned by the reviewer has now been amended slightly to acknowledge how the strategies developed by the MFI agents accompanied the commercialisation and widespread of the practice of microfinance.
A further analytic question that may be outside the scope of this paper but which might be useful to pursue in other analysis would be a discussion of jealousy around *idols.*The success of some village members is portrayed as unequivocally a source of hope and aspiration.However, work by scholars such as Julia Elyachar suggest that individual success can bring ambivalence and even fears of evil eye.This might be addressed in the context of the 'inspiration' proffered by idols.This is a really important point.I am interested in understanding how conflictual the process of improving one's living standards can be.In particular, some of the data collected suggests that some of the strategies deployed to manage livelihood improvements while eschewing jealousy and envy are often be gendered.This topic deserves robust empirical analysis of the anthropology of consumption, which is unfortunately beyond the scope of this paper.
A final empirical question is about repayment.The essay offers an excellent and thorough analysis of the moral frameworks that incite people to borrow.Much less is said about repayment -the gendered expectations around shame of default, struggles to pay, asking for help, calling on kin-based interdependencies, &c.Perhaps this might be a topic to address in the essay's conclusion.This is clearly important, yet largely beyond the scope of this paper.I have incorporated footnote 19 to make reference to this.

Reviewer 3:
It would be possible to draw together all the elements in the title, but it would take a great deal of skill to do so.Agency also weaves in and out, not always in a consistent way, and seems to emerge as a primary concern at the end.
It is good to see more of the work on gender in Bangladesh.It still, however, feels a little uneven.The constructions of purdah, and of dharma, for example, do not seem to me to be understood in a consistent way.The literature review does not set up clear questions to be explored.Overall, there are some interesting observations, but the theorization needs to be considerably more developed.The sections referred to above have been edited and clarified.
P2 My memory is that microfinance was first promoted as a way to give poor people (especially women) access to loans as the funds they required were too small for banks to be bothered with….?
The earliest microfinance model used the group lending methodology to reduce some of the management costs and financial risks of lending money to poor women who had great financial needs and limited earning and credit history.

P3 Problematic to suggest critiques without giving detail and then assuming they are 'valid'
The sentence on page 3 refers to critiques and concerns mentioned in the previous sentence.If that wasn't clear the sentences have now been edited in ways that hopefully clarify this link.

P3 Strange to claim opening story indicates microfinance as moral imperative, when the story itself makes no mention of morality.
P3 is microfinance a moral compass?The argument is not that microfinance is a moral compass but that its discourse is locally constructed as such, in order to gain interest and make it appealing to all.P3 'a means to sustain, redeem or compete for male guardianship' -needs some explanation, as it suggests such a counter-intuitive combination of worlds, using market language in relation to guardianship.I feel these dynamics are sufficiently explained in the body of the text and empirical sections and do not feel that using 'market language' here is at odds with the pursuit of economic security by the poor.

P3 introduction of 'individual agency in relation to social change and development'
widens the scope in a way that is perhaps over ambitious.The sentence has been slightly edited to 'thereby rethinking the place of individual agency in relation to social change in development' making it slightly more narrow and less ambitious.P4 I am not sure what claim of connection is being made by 'coalesced historically' with Rankin's work, but it does not seem to me an entirely convincing fit.I strongly feel that discourses drawing on powerful 'gender ideologies' and 'rational choice theory' co-evolved in the development sector and influenced the development of microfinance and reinforced its attractiveness to the international community.
P5 -isn't it ok to be a subject of development, so long as you are not subject to it?This paragraph seems to me a little undigested.
I have not addressed this point as such discussions are beyond the scope of this article.
P6 are 'the poor' properly referred to as 'one particular actor'?This doesn't sound like a very relational approach to me!The sentence has been edited.
P9 not sure how much 'confidentiality' would be possible in such contexts.Something on informed consent procedures should be included in methodology.
Issue of positionality and confidentiality have been expanded on.
P9 'idols' is such an interesting term in the context of predominantly Muslim rural BD.So much more could be made of this!I would like to thank the reviewer for this comment.The relevant paragraph has been expanded on page 10-11 to elaborate on the interpretation of the term 'idol' in this context.
There also needs to be more discussion of 'guardian'.
I agree with the reviewer's comment, but unfortunately this isn't possible due to the limited word count.

P10 Did microfinance officers enter homesteads?
The following sentence included in the manuscript addresses this question: 'Male household members, conforming to their guardian role, act as gatekeepers and are generally distrustful towards unfamiliar men entering their homestead.'P11 I think we need some evidence of the moral (Islamic) resistance to microfinance.For example, I don't think people in general realised how high the interest rate was, as it was paid back in weekly instalments.
The point made by the reviewer here is unclear as taking any form of interest on a loan is forbidden in Islam.The rate of interest taken is irrelevant when such considerations are taken in account.

P11 needs to be clear what is the basis of the information about calculations made by MFI workers
The following paragraph was added to address the reviewer's comment and clarify some of the key elements influencing MFI agents' calculations and strategies: Microfinance agents make a rapid informal judgement about sources of potential tensions between their aim to formalise loans and a household's view of the 'proper order of things'.This judgement takes the level of vulnerability and poverty of a household into account judging by the size and condition of the homestead, the quality of clothes worn by the household members, and number of assets populating their homestead.Another element affecting the agents' judgement is the religiosity of the household members.This requires recognizing contextual religious and cultural codes, 'the material and ideological resources and practices of households and families' (Standing, 1991:12).As well as reading into people's verbal and body communication, agents can observe the degree to which purdah is enforced, religious clothing can be used as part of a wider set of symbols to signal piousness and religiosity (hijab or veil, tupi, Islamic beard and so on).From this information, agents infer the nature of kinship and interpersonal relationships and gender hierarchies characterising the household.
P12 idols and being rich.Not clear how this is about a moral economy, which is typically about keeping things even in some way.In general, claims about moral content of arguments need to be better substantiated.
The substantiation of the dynamics analysed in the paper, I think, is sufficient to substantiate the claims made about the gendered meanings of morality and economic behaviour.P 14 I have no doubt that in some households taking a loan is seen as women's duty, but is it so in all?Again, it would be good to know the depth of this data -based on one example or many?Counter examples?
Because loans have by and large been made available to women only, the position of women as entry points for others to get access to loans is extremely common (and normalised).
P15 Ds in law asking that their mothers in law should take loans rather than themselves is interesting.Again, would be good to know more about this.
The comment above is not clear.I could not address it.

P16 'borga' needs to be translated as sharecropping.
This has been clarified.P16 nice point re duty of care.Reference to 'irrationality' though seems to me a bit of a deviation.
Two of the three references to 'irrationality' have been delete in order to address this comment.
P16 'idols' not a traditional moral framework -the knitting together of different fragments of 'modern' and 'traditional' discourse needs to be much more critically examined I would like to thank the reviewer for this suggestion -and hope the additional edits reflecting on 'idol' addresses it sufficiently.I agree that more could be said about the formation of the idol discourse, yet it is beyond the scope of this paper to discussion 'modern' and 'traditional' ideas and values.

P17 the notion of 'relational agency' might prove helpful.
This suggestion was very useful.Although I cannot develop it fully here I have edited the middle paragraph in the conclusion, which I think helps better capture the analytical point about the inter-personal construction of agency.I would like to thank the reviewer for this comment!
1st sentence: meaning of 'idols' needs to be given a context.Who calls them this (and what else might they be called by others!).This has been clarified in footnote 9.The writing needs some attention, it is rather imprecise.E.g.P2 'combat over-zealous structuralist ideas about poverty' -not clear whose ideas, what 'overzealous' might mean, etc. P3 'the poor' being coerced by 'husbands'.The loose style is irritating (at least to this!) reader.