Market, Freedom and the Illusions of Microcredit. Patronage, Caste, Class and Patriarchy in Rural South India

Abstract As a market tool, microcredit is expected to promote individual freedom, for women in particular. By drawing on a southern Indian case, this paper argues that microcredit is in fact shaped by the power structures it is supposed to eradicate. Even if they are partly reshaped, local structures of power remain unavoidable to protect populations (something that microcredit fails to do) but also to build the microcredit market and ensure its legitimacy, for donors, local political arenas and local populations. Far beyond microcredit, our findings question the uneasy relationships between markets and individual freedoms.


Introduction
'There are many visions for microfinance, including this one: microfinance for redemption. The dictionary defines redemption as "restoring one's honor and worth; setting one free". Isn't that the highest vision for microfinance: assisting people in restoring their honor and worth, and setting them free from the bondage of poverty?' (Reed, 2011, p. 53). By claiming this, and echoing an ancient humanist (and Christian) tradition, Sam Daley-Harris, director of the 2011 microcredit summit campaign, and whose role in the mediatisation of microfinance has been crucial (Bédécarrats, 2013), takes up an old but persistent idea: the market as a vehicle for individual dignity, honour and freedom. One may think about various definitions and approaches of microfinance, he says, but his is the right one. In Muhammad Yunus' long-standing fight for microcredit, individual freedom and women's empowerment have been major arguments. In various writing Yunus has described how microcredit can successfully free poor women from various bonds of dependency and hierarchy. 1 His autobiography discusses how he found his vocation as a banker for the poor through observing the subordination of women by local patrons (Yunus, 2003, pp. 8-9). As he summates in his book Creating a World without Poverty: 'What if you could harness the power of the free market to solve the problems of poverty, hunger, and inequality? To some, it sounds impossible. But Nobel Peace Prizewinner Muhammad Yunus is doing exactly that' (2008, back cover).
A similar rationale has been applied to microcredit organisations. Subsidies have been presented as too limited and uncertain, and as sources of interference, dependency and market 'distortion'. Such critique has been widespread in the microcredit industry and defended by various key stakeholders, 2 which in great part explains the market shift that has occurred over the last decade (Brière & Szafarz, microcredit claims to fight). If the Stateor, we could add, any other form of democratic institutioncannot offer adequate protection, people may have no choice but to seek out relations of dependency and hierarchy that protect yet exploit them (Wood, 2003), in that the protection in question can be partial, insufficient, and permanently contested and circumvented. In return, those who offer such protection build and maintain their power by controlling the poor and the powerless .
Coercion is conversely a pre-requisite of markets: rules and compliance to them are necessary for people to engage in contractual arrangements. If the rule of law is weak or non-existent, pre-existing social institutions fill the gap, but often do so in incomplete, arbitrary ways that primarily serve their own interests. Furthermore, 'when such regulative institutions are firmly socially anchored, the state is delegitimised and/or captured in ways which make the establishment of the [state regulation] even more difficult' (Harriss-White, 2003, p. 38).
It is precisely this dual (and mutually reinforcing) process that will be addressed here. Microcredit, although meant to be a powerful vehicle for individual freedom, in fact desperately needs local structures of power and domination for the two aforementioned reasons. The local population's need for protection is the first reason: microcredit fails to provide what local structures of power precisely can. Constraints on how microcredit can be implemented are a second factor: microcredit organisations rely on local power structures to develop their legitimacy and clients' loyalty at a lower cost. These local structures are in constant flux. As Picherit (2012) in Andhra Pradesh has also observed, dependency ties may be diluted, but nonetheless multiply and persist. Far beyond microcredit, our findings question the contradictions of present-day pro-market development policies. The market as a relation between equals does not exist, and even less the market as a tool that can deliver freedom. Concrete markets (and not abstract and fictive markets) are necessarily embedded, intertwined, and articulated with social institutions (Hillenkamp & Servet, 2015). And it is this articulation which defines the nature of social relationships that one particular (and concrete) market will deliver.

Methodology
This paper's findings are part of a long-term research programme on labour and finance in various districts of north and coastal rural Tamil Nadu. It was begun in 2003 and still continues today. Diverse research questions are being addressed (microcredit, segmentation of labour markets, migration, debt bondage, over-indebtedness) around a common thread: understanding the evolution of social and political relationships through labour and financial arrangements. This research programme has drawn on a wide range of methods, including a wide range of qualitative methods (semi-directive interviews, case studies, informal discussions and observation) and a number of household surveys. One of the authors spent two years on site, with numerous stays and visits to various villages, and the other author lives in Chennai, spending part of his time in the villages of the districts under study. Some results of our household surveys will be used here to contextualise the analysis and quantify the prevalence of informal labour, lack of social protection, the fundamental role of social and dependency relations in accessing resources and the limited role of microcredit. The analysis of local power structures and the multiple ways they articulate with microcredit will draw on qualitative analysis, which has its own criteria of rigour, namely contextualisation (looking at the contexts in which discourses and practices take place), triangulation (confronting different points of views) and saturation (diversifying observations until no new information is obtained).
110 semi-directive interviews were conducted with microcredit organisations' managers, loan officers, elected bodies, local politicians, moneylenders. 3 More than 150 group discussions were done with villagers, who were women for the most part but also men. These group discussions were very useful to gain an understanding of official discourses, social norms and accepted practices. 17 case studies were done with loan officers and microcredit group leaders, with whom we had a series of interviews over time. We closely studied eight microcredit organisations.
From the locations, microcredit organisations and individuals, to the sampling for qualitative data collection, the aim was ensuring diversity. Villages differ along economic, social and political lines, Market, freedom and the illusions of microcredit 743 from agrarian villages still partly dominated by upper castes to peri-urban villages in which Dalits (ex untouchables) have acquired a certain form of independence. 4 Microcredit organisations include non and for profit structures and vary in terms of size (from a few thousand to a few hundred thousand clients), sources of funding (local philanthropy, state subsidies, international subsidies, bank funding) and the profile of their founder (in terms of caste, gender and religion). 5 As for field officers and selfhelp-group (SHG) leaders, here too, caste, gender and religion were the main determinants. Group discussions were carried out with people who were available, and spontaneously organised along caste and gender lines since men, women and people from different castes almost never mix on a daily basis. As we shall see throughout the paper, on all levels, this diversity has been key in understanding the everyday politics of microcredit.
Whether for contextualisation or triangulation purposes, confronting discourses and practices and observing what people do is crucial. Most of the information we use here came from participatory observation in three areas: the daily life of microcredit organisations, women and local politics. We spent around 12 months (spread over 10 years) in their offices or with their staff in the field. We attended many staff meetings, training sessions, witnessed varied interactions with women, and interactions between the staff of different organisations and media events, which as we shall see are instrumental in building the legitimacy of microcredit organisations. We also spent time in villagesfor around 36 months, here too spread between 2003 and 2012sharing the daily life of women and participating in daily political life. A strong limitation of our work relates to upper castes, with whom little fieldwork has been done in the villages, for various reasons that cannot be detailed for reasons of space. Some of these interviews and discussions were recorded and partly transcribed, others not, as the conditions were not right (informal discussions, refusal, noisy environment).

Context: Neoliberalism and Patronage
Microcredit is both shaped by and constitutive of broad social, economic and political dynamics that should be taken into account if we want to understand its true effects. Tamil Nadu is the fifth largest economic power in India in terms of GDP, tenth in terms of 'human development', and is often presented as a success model for combining economic growth and social progress. Despite successive liberalisation measures from the 1990s, state interventionism remains rather high. State policies were long focused on agriculture and the 'green revolution', but over the past 30 years have expanded into industrialisation and the private sector while maintaining a longstanding tradition of social policies. There is currently a wide range of social welfare schemes focusing on issues such as food security (subsidised food, school meals), housing, employment and credit. But while Tamil Nadu is one of the most welfarist states in India, it is also known for a populist rhetoric of the 'people' that leaves the privileges of the ruling classes intact (Harriss, 2001). Much of employers' competitiveness is down to public assistance (free access to water and electricity, public contracts, soft loans, very low or absent taxes) and a cheap, docile labour force, most of which falls outside of any form of labour law.
Our 2010 survey was conducted on a representative sample of the local population (407 households), and confirms the virtual absence of any form of social protection. Although agricultural employment is on the decline (87.9% of households depend on non-farm labour), employment mostly remains informal: out of 1350 labour 'contracts' (households and individuals comprise multiple occupations), a tiny minority of jobs include bonuses (6.2%), health insurance (1.9%) and pensions (4.4%). The vast majority of regular employees have no written contracts. Working time is determined by mutual agreement between the supervisor and the employee, and can be broken off at any time due to an unexpected fall in production or conflicts between supervisors and workers. A certain form of neoliberalism, of which informal labour is an illustration, thus coexists with a wide range of safety nets.
Social policies are moreover implemented along the lines of the 'patronage democracy' which is typical of contemporary India (Chandra, 2004, pp. 47-82). Accessing strategic resources remains subject to political patronage, defined here as the use of state resources to reward individuals for their electoral support, and clientelism, meaning relationships based on political subordination in exchange for material reward (Harriss-White, 2003;Pattenden, 2011). Patronage and clientelism 6 are subject to ongoing change, however. Traditional landlords are no longer the 'oppressors' they used to be (Harriss, 2013). The patrons are instead labour intermediaries, local politicians and development brokers (Pattenden, 2011;Picherit, 2012). The resources circulating within patronage networks have also evolved. Far beyond commonplace resources such as jobs, land, credit or water, now public contracts, social governmental schemes and NGO benefits are a key issue (Picherit, 2012).
Though measuring patronage is very challenging, some of our data related to debt arrangements give an idea of its scope. Debt relationships are key in building and maintaining patronage relationships. Our 2010 survey found that almost 90 per cent of household loans (of which only a small proportion are microcredit, as we shall see below) included additional services by the lender (and free services by the borrower in return): these include 'employment' (15% of loans); 'financial support' (51.1%, this means that lenders introduce their clients to other lenders); acting as a guarantor for other lenders (28%); access to information (43.7%), mostly for governmental schemes, administrative procedures but also hospitalsthough health care is supposed to be free, health expenses strain household budgets and advice and contacts to identify and access to reliable and cheap services are in high demandand finally 'political support' (30.7%). This means facilitating access to governmental schemes such as those described above, obtaining administrative certificates (such as certificates of poverty level to access subsidised food, certificate of residence, age, caste, property). Very few directly approach the administrative bodies, but instead pay an intermediary in cash or by offering a service in return. Political support also means negotiating with the police or the court in case of dispute: neighbourhood, kin or inter-caste conflicts, vendor without a license, commercial fraud, false charges, risk of expropriation, and so forth. Far from being anecdotal, disputes set the backdrop of daily life. Allegiance ties are closely intertwined with local structures of power, which in turn are articulated by class, caste, gender and religion relationships. As elsewhere in Tamil Nadu and India, caste remains a fundamental feature of social, economic, ritualistic and political life. Mostly thanks to social policies and industrialisation, discrimination against the lowest castes (the ex-untouchables, who are also called Scheduled Castes or Dalits) has significantly reduced over the past 50 years while leading to numerous conflicts with middle caste (here, Vanniyars). Nevertheless it remains high, varying considerably by region and sub-category of Dalits, and resulting in restricted access to resources (education, jobs) and social and symbolic stigmatisation. In the regions we looked at there are very few Muslims (less than 3%) 7 and they are usually ranked rather high in local social hierarchies. There are also few Christians (3-6%), but in some villages they make up a significant proportion of the population and are usually Dalits. In keeping with a long tradition of social services begun by European missionaries, many NGOs are of a Christian background, but mostly operate as secular institutions (Mosse, 2012, pp. 226-227). This paper's main argument is that microcredit draws on these relationships of clientelism. While it may reshape them, it does not challenge their role in the redistribution of wealth, as it relies on them.

Microcredit and Patronage Networks
There are two dominant models in the Indian microcredit landscape. Until recently at least, microcredit NGOs were major players. Under the SBPL (SHG Banking Linkage Program), microcredit NGOs work in partnership with SHGs, local banks and refinancing public banks. SHGs consist of 15 to 20 people who circulate money amongst one other, and who are then eligible for external loans. In March 2012 in Tamil Nadu alone, 8 an estimated 925,392 SHGs fell under the SBPL (with around 13 million SHG members), and there were around 10.72 million microcredit borrowers, two thirds of whom (67%) were with the SBPL, and 95 per cent were women. The rest were private microfinance institutions (MFIs), most with non-banking financial company status. The difference is not in lending method (many MFIs also provide SHG loans, even if there is a growing tendency to provide individual Market, freedom and the illusions of microcredit 745 loans) but in terms of funding: MFIs look for financial sustainability and thus have strong constraints in terms of cost reduction and income generation.
Both NGOs and MFIs seriously need local patronage networks: the former to establish legitimacy with their donors and in local political arenas, and the latter to develop their markets. In both cases, the mass mobilisation of women, and women Dalits in particular, is key.

Microcredit NGOs and Patronage: Establishing Legitimacy
Most field workers and microcredit organisation managers expect four key qualities from women: availability, motivation, gratitude and loyalty. Not only do women have to form groups, which they do not necessarily like, but they also have to adapt their schedules to their credit officers'. As the credit officers have genuinely limited time, women must be available to have their accounts checked and to make payments according to their credit officer's schedule of rounds. They also have to be available in the event of unexpected visits, for instance of a donor. They must show motivation and enthusiasm for any new projectstraining, information or outreach meeting, collective entrepreneurship, and so forth, and participate in the many events microcredit NGOs and their partners organise. For events of great importance (examples will be given later), women must not only attend, but latecomers are often scolded and absentees may struggle or wait to get their next loans.
Officially, SHGs are ranked by their capacity to administer their internal loans efficiently and to manage their accounts properly, which can prove complicated in practice. Most groups need support from staff, and those who get substantial help are also expected to comply the best with all these restrictions. Some women spend countless hours on all these activities, which are considered a natural counterpart to the services they access. Women are perfectly aware of the contradictions between microcredit organisations' official discourses on empowerment and their daily practices. 'How can we become empowered' one asked, 'when we are used like toys?' 9 (Interview, January 2009).
Understanding the ambiguity of these forms of domination requires an analysis of local political economies and the multiple ways through which NGOs establish their legitimacy, both with their donors and local political arenas.
Subsidies from the Indian government are a frequent source of funding (all NGOs here have been using them), especially during the decade 2000-2010 when SHGs were heavily supported by successive state governments. But being eligible without paying too many bribes or facing too many delays requires specific contacts and/or playing the game of the ruling party (Picherit, 2015). This means participating massively (through women) in the numerous events the government or party organises (visits from deputies, the inauguration of new schemes, celebrating the success of such or such scheme, and so forth). Some NGOs juggle between the two main alternating parties. 10 One of them posted some of its staff in the office of the district collector. Others are close to one specific party (as with one NGO which started out as a literacy campaign movement and is close to the Communist party, CPI). Others, due to a lack of critical mass and contacts, rely on other NGOs to play a brokering role: they help the smaller one to access subsidies and in return use 'their' women for the events they hold where having a large turnout of women is a key success factor.
International funding comes through Christian networks (especially Caritas), multilateral organisations (here, ILO, UNDP, IFAD) and private foundations. These donors are less demanding in terms of media events but instead ask for evaluation reports, concrete and quantifiable results and sometimes field visits, which are often organised at the last minute and for which the women are supposed to be available. Some organisations are partly supported by local philanthropy and business networks from various sectors (jewellery, garment, real estate, water management, clinics). Investing in 'development' is a way to limit taxes but also to boost the image of being a benefactor. This needs media events, which entrepreneurs may personally take part in, hoping to get their picture in the press. Microcredit NGOs also need local communities' acceptance on the village level. Local power structures, however, are very diverse. The dominant castes can vary by region or even by village. Such diversity requires context-specific strategies and multiple alliances between unequally powerful landlords, caste leaders, local elected bodies, and local politicians who include Dalits, as their alliances can sometimes be essential for influencing the Dalit population. These alliances can also be greatly affected by who the NGO founders are: an NGO created and managed by a single protestant Dalit woman does not have the same access to resources as an upper caste businessman, nor even of a Catholic man from a middle caste background. Alliance networks and dealings between NGOs are facilitated or blocked by personal relationships, but are often mediated through kinship, caste or religious membership, which people draw upon differently based on their circumstances and the opportunities open to them (see the examples in the Online Appendix).
Microcredit also needs to build its legitimacy in the NGO world. Here again, regardless of funding source, organising mass events is key. The following words from one NGO loan officer sum up the situation effectively: 'Mass is our strength. Your success, your visibility, all this depends on how many women you have. Not only those you reach, but those you can mobilise for a mass event. It shows your strength. If you want to deliver something to other NGOs, to the government, this is how it works. Same for the public. I can get 400 women, you get 4000. What do you think the common man will think?' (Interview M2, June 2010).
This means that mass events become unavoidable, both for boosting NGOs' reputation and for strategic alliances. They can include events for specific causes such as Women's Day, HIV-AIDS Day, Children's Day, Consumer Rights Day and so forth, presentations of SHG's products, health awareness campaigns, visits from politicians, the inauguration of public schemes, human rights campaigns (for instance for Dalits, women, children, the disabled). The smallest NGOs muster 300-500 women and the largest ones 3000-4000. Target numbers are fixed ahead of time and then worked hard towards. Events vary in number and frequency too. Some NGOs organise almost one event per month while others just three or four a year. Official speeches by guests of honour, cultural performances such as songs and dances by SHG members, and award presentations are usual features. While these kinds of mass events are one of the most important ways to bring members together, they also help to project and publically display the strength of microcredit organisations and their allies. SHG leaders go on stage to publicly testify the positive achievements of SHGs, whether in terms of entrepreneurship, women's emancipation from men or collective action, even if these are very far from real achievements.
But what is the key problem here? All kinds of organisations, after all, have to organise publicity events. The substantive issue is that women pay the price for it, with the brunt falling on Dalit women. Although, as we have seen, the actors involved have very diverse characteristics, they all adhere to the rules of patronage discussed above, with women's groups representing an unprecedented opportunity for organising the mass events so key to building and reinforcing the patronage system. NGO staff, villagers, men, and women all agree that women more readily agree to take part in the mass events. As we were once told by an SHG member, 'SHG women are the first goats to cut on the mutton stall' (Interview, July 2008). But the women in the villages do not all face the same expectations from NGO staff. The few able to start a business are mostly non-Dalits, as the barriers are too great for Dalits (Guérin, D'Espallier, & Venkatasubramanian, 2015). It is, by contrast, mostly Dalit women who are required to attend mass events organised by NGOs or their partners and allies. At the 24 such events we attended between 2005 and 2012, Dalit women members were systematically over-represented. Even for NGOs targeting both Dalit and non-Dalits, the former would make up at least two thirds of the audience. Field officers and NGO managers were very clear about why this was the case, explaining that non-Dalit women were far more strictly controlled, had much less freedom to leave their homes, were more demanding and asked more questions of the NGO. In some NGOs, participating at mass events was a very explicit condition for loan eligibility only for Dalit women.

Microfinance Institutions, Patronage and Market Building
The above points could be taken as confirmation of the weaknesses of the subsidy-based approach of NGOs: a market-based approach, by contrast, by which organisations balance their costs, could be championed as a way to get rid of all these dependencies and interference ties, while allowing massscale interventions. The two microfinance institutions (MFIs) we studied here have non-banking Market, freedom and the illusions of microcredit 747 financial status and are mostly funded by banks, some of which are foreign, but which are mostly local (see Table A.3 in the Online Appendix for details). High growth rates and low operational costs have been the main strategies to achieve financial self-sufficiency (Augsburg & Fouillet, 2010). But these organisations still have no real freedom from funders or local power structures of power.
First, a particular kind of local legitimacy may also be needed to attract domestic banks. The MFIs we looked at were both also involved in organising mass events, albeit to a much lesser degree than NGOs, and we did come across other MFIs which do not. But more fundamentally, all the MFIs we encountered need these local power structures to build their markets by recruiting clients and ensuring their loyalty in terms of loan renewals and repayments. Our two MFIs saw an annual growth in client numbers of over 40 per cent over several years between 2006 and 2010. This growth was possible only through the support of local networks of influence and the nodal role of specific intermediaries. Far beyond our case studies, a similar phenomenon has been observed in other parts of Tamil Nadu and Andhra Pradesh, as discussed in detail by Arunachalam (2011). These intermediaries, who may be men or women, have various profiles and trajectories (as summarised in Table A.5 in the Online Appendix). It is, however, possible to identify some common trends. They all know how to read and write. 11 In line with local social markers (people almost never borrow from people who are inferior to them, they handle clients from their own caste or from an inferior caste. 12 The difference between NGO and MFI loan officers is rather blurred: the former are under more pressure to mobilise women for media events, the latter to disburse loans and collect repayments, but many experience both. They build up their legitimacy through their ability to provide various services to local populations, which in turn requires them to be inserted in the local networks of patronage described above. Two typical scenarios for this process can be discerned. In the first scenario, loan officers are well integrated locally and can easily recruit female clients, possibly to mobilise them for mass events and ensure repayments. In examples of the cases we came across, these loan officers, men or women, were field workers who had been involved for a long time in social work such as literacy or health campaigns. The men were close to a political party, which in the cases we came across was their caste party. 13 They didn't have any specific position in their party, but regularly carried out favours for the party and received them in return. An example is getting help to set up in a new village through the approval of local village leaders, or resolving conflicts between clients. In return, the NGO might send women clients to attend political events. The women loan officers have less close links to parties. 14 Other strategic networks also include trade unions (which are often very close to political parties; see Table A.5 in the Online Appendix for more details on each case), Christian networks and local parish, 15 local employers (landowners, local enterprises), local administrations and finally local lenders. All these networks are strongly intertwined.
If the loan officer has no particular networks or comes from elsewhere, there is no choice but to build alliances with local resource people, which is the second scenario. Under pressure from ambitious portfolio objectives, they simply have no time to identify creditworthy clients for themselves. They turn to resource people, or their wives, to create and then lead SHGs. These resource people are selected both for their capacity to mobilise women and their support networks. Here one finds only women as SHGs are mostly female-based. They have well established relationships in their neighbourhood and play a role of intermediary by providing various favours, ensuring a certain loyalty from their group members. One summarised the situation as follows: 'They [MFIs] need us. How do you want them to find clients? They don't ask for any guarantee, they have no time to spend in the village. They need strong people. They use us for that purpose. We snap our fingers and women come. They need us' (Interview F7, July 2010). They play a key role in recruiting creditworthy clients (by drawing on women's morality and reputation rather than business projects, which are virtually nonexistent), in ensuring repayments but also in managing competition: as we shall see below, they build loyalty by providing a large range of additional services which go far beyond their official function.
They have diverse profiles that depend on their networks and class position. Some are part of a dominant class (within their own community) through their husband's positions, including a retired teacher (F7), a permanent railway company employee (F9), and a village government president who had been in power for many decades (F8). Some were already playing an intermediary role for arranging credit or labour for upper castes (who always use intermediaries as they don't want to deal directly with lower castes) (F7, F13, F15). Some have a job or a position which facilitates the contact with women, including a child care worker in a crèche (F10), midwife (F14), door-to-door saree vendor (F9), a volunteer in the local parish (F4), Rosca manager (F6, F9) or moneylender (a frequent case).
As we shall see below, both loan officers and SHG leaders can use their position as an entry point for accumulation patterns (albeit at very variable levels). Many cannot however be condemned as parasites on the system betraying the initial objective of microcredit. They are a product of microcredit itself, which can only grow so quickly at a low cost by using pre-existing networks of influence.

The (Relative) Feminisation of Patronage Networks
It would be a tempting conclusion to say that structures of power and authority have remained unchanged. It is of course clear that in the context we studied, microcredit does not openly challenge pre-existing hierarchies and may even strengthen them through Dalit women's instrumentalisation. What we observed was, however, more ambiguous. All things being equal, it is in women's best interest to play the game insofar as these patronage networks allow them access to other resources and a certain form of protection. They therefore partially submit to patronage rules, while challenging it and trying to get the most possible advantages from it. Ultimately, all women widen the spectrum of the relationships they juggle with, which is key in an environment where accessing resources is above all a matter of contacts. However, this widening process is of course highly uneven.

The Demand for Both Credit and Protection
The women themselves have a high demand for credit, as a result of a poor and irregular income not only owing to the agrarian crisis and the hazards of migration (Taylor, 2011), but also rising consumption needs. Of eight organisations we studied, seven were first involved in rural development issues and slowly shifted to microcredit, sometimes abandoning some of their previous activities. This shift was the result not only of pressure from their donors, but also from villagers, including women, who otherwise threatened to leave. Some of the NGOs were very sceptical about microcredit and were slow to respond, but eventually gave in as their own existence was at stake.
We shall note however that microcredit fails to eradicate poverty and vulnerability for a variety of reasons. According to four surveys we did, most microcredit is used to make ends meet (daily expenses, health, debt repayment) (see Table A.4 in the Online Appendix). It has virtually no effect on employment creation: people might be obliged to work more to repay their loans, but business creation, often thought to be an objective of microcredit, fails to take place: preconditions for success (mostly access to providers, clients and the administration in good terms) are absent and loan modalities inadequate (Guérin et al., 2015). Households still juggle with a wide variety of loan sources and microcredit represents only a small share of their total debt (according to our 2010 survey, 7.6% in number and 4.3% in volume, out of 1910 loans identified) while more than one third of the families (37.5%) have access to it. As such, microcredit fails to solve the issue of protection that people desperately need. As mentioned above, microcredit may be used to finance health or education, but needs in terms of protection are much higher. They include access to employment, which as we have seen above remains mostly informal and irregular; access to government benefits, which have the merit of existing, but eligibility is subject to having the right contacts; access to reliable and cheap health services; access to credit, which remains the most common form of making ends meet, investment in child education, improving housing and organising social events and rituals.
Market, freedom and the illusions of microcredit 749

Challenging the System
As we have seen, women have to submit to various criteria to be eligible for microcredit. The unequal power relations are undeniable, but often conceal the subtle games of performances and negotiations that frequently take place. NGOs' control is never a given, nor to be taken for granted. It depends on NGOs' ability (often through individuals such as loan officers and SHG leaders) to provide regular services and protection, in the context of ongoing bargaining processes. The women themselves are very much aware of the deal, frequently telling us: 'NGOs should heat the milk without letting it boil over'. Some refuse to submit to NGOs' constraints and drop out, complaining about the multiple conditions, peer pressure, and the considerable time spent at pointless meetings. Some are asked to drop out by husbands or in-laws grown exasperated by the time women have to spend, or by rumours about their mobility and broken promises, especially in terms of job creation.
Many women 'play the game' however, in the sense that they gradually learn how to make the most of it. Over time, women dare to refuse activities which are too time-consuming, ask for compensation (a day's wages, a better lunch, higher loan amounts, privileged access to governmental schemes the NGO is eligible for), question the added value of some mass events, and ask credit officers to come to their doorsteps. Some of their requests can lead to conflictover the years we felt that the staff complained more and more about the women's 'lack of motivation' or 'loss of gratitude'. For instance, by 2010, monetary compensation for attending any kind of meetings seemed the norm, which had not been the case a few years previously. As a woman who had been involved in SHGs for around 10 years stated: 'before we were only wagging our tails, now we've learnt how to bark' (Interview F8, July 2012).
While women's mobility beyond their village remains strongly controlled, albeit unevenly from one caste to the next, their presence in public space is probably better accepted now. SHG membership as a label, to some extent, allows them go beyond their previous boundaries.

Protection and Accumulation
In theory, loan officers and SHG leaders (the division of tasks between the two is often vague) are in charge of training SHGs, organising monthly SHG meetings where savings and repayments are collected and borrowers are selected. They are also in charge of bookkeeping, maintaining relationships with microcredit organisations and sometimes the bank or officials (in the case of NGOs). In practice, they do much more than what is expected on paper and gradually become local intermediaries (or strengthen a pre-existing intermediary function).
Enlisting women to projects, whatever their purpose, is a key part of their role (saving and credit groups, training, awareness sessions, health campaigns, mass events such as those described above). This demands charisma and persuasiveness, not only with respect to SHG members but also their husbands and in-laws. But far beyond questions of personal charisma, the ability to mobilise greatly depends on providing regular support and help of various kinds (see Table A.5 in the Online Appendix). Some of these intermediaries have privileged access to administrative and public institutions (dispensaries, hospitals, the police, schools, district collectors), which is acquired or strengthened through their affiliation to their microcredit organisation, and they give the women a hand with their administrative procedures and paperwork. Some, including men, intervene in family problems (such as domestic violence, conflicts related to pregnancy outside marriage or adultery). Some help with health issues: identifying reliable doctors, techniques for avoiding queues at the dispensary or hospital, helping limit bribes for doctors or to get medicines at a lower price.
Last but not least, many of these intermediaries lend money themselves (or act as intermediaries for moneylenders), and for some this is almost a profession. Some began upon becoming SHG leaders or loan officers. Others have been acting as financial intermediaries a long time, but have used their new position to expand their clientele and sources of funding (microcredit being one of them). Their role as financers is very useful for ensuring repayments. Even if in theory the group is expected to do it, in practice leaders or loan officers are the ones who enforce payments, either by exerting social pressure, or by lending money themselves to defaulters. In return, some do not hesitate to use SHG loans for themselves (often for on-lending), to create fake members or even fake groups, to charge additional interest (1 or 2% in the cases we came across) or additional fees. All this is often considered as normal payback for their commitment, which is either badly paid (loan officers) or free (SHG leaders).
Ultimately, these intermediaries take on a dual role typical of any intermediary function within a patronage system: they provide some sort of protection while using their gatekeeper position for accumulation, though only on a limited manner and unevenly.
We can identify a continuum of profiles. First we find ordinary SHG members. Through the group and above all through their leader or loan officers, they expect better access to a wide range of resources (especially credit and governmental schemes, and sometimes employment, as some loan officers and SHG leaders are labour brokers). In return they have to 'participate' in various mass events organised by NGOs or their allies, and may also offer free services to the SHG leader or loan officer.
Then there are women who gain little or no material benefit for their leadership role (F12, F13, F14), but nevertheless enjoy social and symbolic gratitude: they are recognised locally and will be thanked and applauded publicly at the time of mass events, where they will be called on stage. However, this type of engagement can only sustain over time with the support of a progressive husband or in-laws who appreciate social work, or in the case of widows, separated or old women.
At the other end of the continuum (but still at the very end of a long chain of patronage) are those who get material gain from their engagement. It is here that men can be found (given the importance of male breadwinner norms, one can hardly expect men to engage in social work for free) but women are not absent, albeit here too with various reasons for being there. For some it is extra cash. In 2012, the amounts ranged from 1000 to 2000 INR monthly (F9, F10, F11), while the average salary of loan officers ranged from 3000 to 5000 INR. Others manage to get 10,000-45,000 INR per month, including among some women (F7, F8). This is, however, only possible to those who already belong to a dominant class such as those mentioned above, through their husband, and who benefit from adequate capital and networks.

Conclusion
The main purpose of this paper has been to highlight one of the contradictions of microcredit: while it is intended to foster individual freedom and empowerment, especially for women, its implementation in fact needs embeddedness into local structures of power. Because it takes place within neoliberal policies which fail to offer adequate protection to local populations, and because it is done at low cost, with high growth objectives and in a climate of competition, microcredit needs these power structures in order to develop. Both processes are mutually reinforcing. Local networks of influence, through which a large part of resources and services circulate (jobs, credit, governmental schemes, conflict resolution), be these political, religious, caste-based or philanthropic, see microcredit groups as an excellent opportunity to mobilise large groups of women and thus strengthen themselves. In return, microcredit organisations, whatever their status, have to rely on these power structures to build their own legitimacy, recruit clients and ensure their loyalty. And finally, women facing ongoing insecurity also play the game. They are very much aware of the hierarchical dimension of patronage networks. But they feel more comfortable with social and political mechanisms they know the rules of, and which they believe they can navigate to their advantage, however slight. Few but the better off in their own community significantly benefit by becoming additional gatekeepers. The others submit to the rules of patronage, including to the higher caste gatekeepers in their community: they have to display commitment and gratitude, and in return expect better access to strategic resources. While the rules of the operation of these patterns of resource redistribution and the exercise of power go virtually unquestioned, microcredit fuels their complexification and (relative) feminisation.
Market, freedom and the illusions of microcredit 751 Far beyond microcredit, our findings question the ambiguous relationships between neoliberalism, the market and individual freedom. As argued in the introduction, enthusiasm for microcredit stems from its supposed contribution to freedom and dignity. These values have been a major underlying point in neoliberal policies, and explain why they achieve such wide consensus (Harvey, 2005). What is extremely problematic here is to conceive the market as a contract between free and equal individuals, which is very far from reality. In as much as neoliberal policies ignore structural inequalities and their underlying power relationships, they cannot promote the freedom they claim. The analysis we have put forward here confirms that gender, caste, class and religion, which have usually been seen as components of 'civil society' are in fact powerful determinants of accumulation (Harriss-White, 2003, p. 15), including at the very bottom of the pyramid. Criticising the unequal dimension of markets is nothing new, particularly as regards India. But this is often taken to be an appropriation by dominant groups (and as indicated in the introduction, the critical literature on microcredit follows the same line). While this is certainly also the case here, we have also highlighted the inherent contradictions of the market itself.
Nineteenth century thinkers, whether liberal or Marxist, viewed money and the market as a powerful dissolver of social relations (Zelizer, 1994, pp. 12-24), including hierarchical relations. 16 Despite strong divergences with regards to the consequences of the market (viewed either as a form of freedom or as a new form of exploitation), there was a common tendency, which is still valid today, to overestimate the capacity of the market to abolish dependency ties. Simmel's work on money, however, although strongly insisting on the values of equality and individual freedom conveyed by money, highlights the concomitant role of intellectuality and the law: both are strong components of the recognition of individuals as equals (Simmel, 2004, pp. 446-448). In the continuation of Polanyi and Common's approaches, more recent work shows that market-based and 'releasable' debtswhich can be ended through reimbursement by contrast with 'unreleasable' debts embedded into a wide set of rights and obligationsrequire the existence of a fiscal and social state ensuring the protection of its citizens (Théret, 2009). It is precisely this protection, as already argued by Polanyi, which allows market relations to be a source of freedom and empowerment.
Far beyond microcredit, the development world is characterised by the growing importance of market forces as a means to allocate resources, promote access to basic goods and services and eradicate poverty, as illustrated by the growing success of 'Bottom of the Pyramid' or 'Social Business' models (Berndt, 2015). There is thus a pressing need to recognise the uneasy relations between market and individual freedoms: in the absence of the rule of law and basic rights, the market is very likely to rely on social and hierarchical relationships to flourish and sustain itself.