Agricultural modernisation in Rwanda

After the 1994 genocide against the Tutsi, a ‘government of national unity’ was established to restore peace in the country, and eventually came to be controlled by the Rwandan Patriotic Front (RPF) and its party leader, Paul Kagame (Thomson 2018: 87–96). When Kagame became president in 2000, the RPF was determined not only to turn Rwanda into a secure nation free from ethnic division, but also a middle-income country by 2020 (RoR 2000). With all the odds against them—Rwanda is small and landlocked, poor in natural resources and has a violent history—president Kagame and the RPF impressed the international community with their ambitious plans and their ability to implement development policies rapidly and effectively. Rwanda has a steadily growing economy, and has witnessed a massive drop in infant mortality, a spectacular decrease in malaria mortality rates, a national health insurance that covers nearly all citizens and near-universal primary school enrolment (Reyntjens 2013: 164; UNESCO 2017). With all these accomplishments, Rwanda has often been referred to as a ‘success model’ for African countries in terms of human development (Straus and Waldorf 2011: 6). However, Rwanda remains one of the poorest countries in the world, where the vast majority live from subsistence farming with too little land and livestock to produce enough food for themselves and their families.

In order to reduce poverty and increase national food security, exports and growth, the government started to pay increased attention to modernising the agricultural sector in the mid-2000s (Huggins 2017: 61). This transformation was centred on the regionalisation of crops prioritised by the state and cultivated on consolidated land, formalisation of land tenure, and investments in fertiliser and fortified seeds. This strategy was meant to increase yields and help small-scale subsistence farmers, who typically were growing a variety of crops (such as sorghum, beans, sweet potatoes, cassava, cabbage and bananas) on individual scattered farm fields, to integrate into commercial commodity chains.

A number of programmes and projects were implemented with the general objective of developing the rural economy. The most important of these was the Crop Intensification Programme (CIP) (Huggins 2017: 69–77). Under CIP, farmers are encouraged by local leaders and agronomists to consolidate land use in the marshlands, on the valley floors and on the hillsides, and to grow one of the CIP-prioritised crops: rice, maize, potatoes or climbing beans. These crops are prioritised based on marketability, adaptability, high nutritive value and responsiveness to farm inputs. The implementation of CIP involved the entire state administration, and small-scale farmers all over the country were required to play an active role in the transformation of the country by leaving their traditional farming behind. In this process, there was no room for farmers’ local knowledge and concerns to be considered.

The state made sure that people aligned themselves with government policies through a decentralised system of surveillance, control and performance contracts (called imihigo) (Purdeková 2012: 193). The Rwandan state administration is organised into six levels: the central government, 4 provinces (plus the city of Kigali), 30 districts, 416 sectors, 2,146 cells and 14,876 villages. The central government is responsible mainly for policy formulation, while the lower tiers are in charge of policy implementation (RoR 2012). Sector agronomists, acting in coordination with the sector executive secretary and the district agronomist, oversee the implementation of national agricultural policies on the village level, and organise ‘sensitisation’ campaigns to explain public policies and their advantages (Ansoms and Cioffo 2016: 9–10). From the viewpoint of the government, Rwanda could be a developed nation only if people broke with traditional subsistence agriculture and their ‘backward mentalities’.

The government aims to change these ‘mentalities’ through ‘sensitisation’, a ‘top-down exercise connoting persuasion, if not simply instruction’ (Purdeková 2015: 109). Policies are diffused to the masses through repetitive messaging from local leaders, public meetings and mass media. It is believed that poor peasants lack the capacity to know what is best for them, and that they need to be convinced of the correctness of development policies, to change their attitude and to adapt (see, for example, RoR 2006: 16; RoR 2009: 20; RoR 2013: 31, 43). Here, power and knowledge are linked: those at the top are the ones who know, and those who are subordinate are the ones who do not know; what is indigenous, coming from within and from the bottom of the social hierarchy, is rejected by the leaders (Hasselskog 2016; Newbury 1989; Pottier and Nkundabashaka 1992). Interaction boils down to one-way communication or, more precisely, one-way mobilisation; the authorities determine how things should be done without respect for the local context or for people's wishes. In cases where the authorities do not see success in the sensitising campaigns, that is, if villagers do not adapt voluntarily to modernisation policies, coercive measures are used. These range from public shaming to fines, uprooting crops and even jail (Huggins 2009: 299; Ingelaere 2011: 74; World Bank 2015: 103).

In order to implement policies in an effective and controllable way, imihigo performance contracts are signed every year between the president of Rwanda and local government institutions at each level, all the way down to the lowest administrative unit. These contracts state what a particular administrative unit will achieve and how these outcomes fit into the broader development agenda of the state (RoR 2011). These contracts are thus result-oriented, and each unit is required to demonstrate the targets met at the end of the fiscal year to the next higher level (Ingelaere 2011: 70; Sommers 2012: 74). The imihigo performance contracts, for example, specify the crop to be cultivated and the productivity level to be achieved for each village, as well as the expected number of national health insurance plans (mutuelle de santé) to be obtained and the number of children enrolled in school. This pushes local governments to put pressure on farmers to fulfil their imihigo targets (Chemouni 2014: 250). Imihigo thus works as ‘a mechanism for measuring alignment with government policies’ (Purdeková 2011: 485).

The new land law and policy launched in 2005 was also an efficient tool for making sure people adhered to agricultural modernisation policies. With no codified land tenure law prior to 2005, the new land law gave landholders ownership through title deeds with a lease of up to ninety-nine years (GoR 2004). The aim of the reform was to convert customary rights to land into formal rights, reduce land disputes, and encourage ‘proper’ land use.³ Under the customary system, farmers considered themselves to be absolute owners of the land (Huggins 2013: 105; Pottier 2002: 181; Pottier 2006). Under the new system, the state was the ultimate owner of land and farmers became leasers mandated to use the land according to the state's wishes. Obligatory registration of farmland under the 2005 law was implemented in 2010–2012. Titleholders were obliged to protect the land from erosion, safeguard its fertility, and follow the CIP directions.⁴ Landowners found to have ‘misused’ the land were prevented from accessing it (Ansoms et al. 2017: 416; Huggins 2013: 2).

Critics highlight that the aggressive implementation of modernisation policies led to farmers losing control of how they used their land (see, for example, Ansoms et al. 2017; Chemouni 2014; Pritchards 2013). Monocropping of state-demanded crops has moreover been criticised for putting subsistence farmers in a vulnerable position, as they were prohibited from growing the variety of crops that could ensure produce throughout the year. Planning miscalculations and authorities’ mismanagement have, moreover, left farmers without seeds and fertilisers for monocropping, farmers have had their crops uprooted without compensation or have been fined for not using state-recommended agricultural techniques (Berglund 2019; Cioffo et al. 2016; Huggins 2009, 2014, 2017: 220). The prohibition on growing subsistence crops has furthermore caused hunger and malnutrition (Twizeyimana 2009a, 2009b; UNDP 2007). These accounts contradict the government's official statistics that indicate that the living standards of rural Rwandans have improved with the implementation of agricultural modernisation policies (Ansoms 2008, Ansoms et al. 2017).

In 2014, Rwanda's Ministry of Agriculture and Animal Resources (MINAGRI) reported that since the CIP implementation annual agricultural growth had averaged 6 per cent (RoR 2014: 2). Official statistics furthermore suggest that the proportion of people living in poverty under the national definition fell from 57 per cent in 2005 to 39 per cent in 2014 (NISR 2015). Filip Reyntjens (2015), however, suggests instead that poverty in Rwanda increased by 6 per cent between 2010 and 2014.⁵ An Ansoms and colleagues (2017: 60) furthermore question the way the survey for the statistics of agricultural growth was conducted, and point out that smallholder farmers were pressured into meeting the authorities’ targets for monocropping on consolidated land through threats, which might have influenced their responses and could have resulted in inflated production figures. It has also been questioned whether the impressive results in poverty reduction and increased agricultural productivity are in fact a result of local authorities’ data manipulation and desire to present good results upward in the hierarchy (Ansoms et al. 2017).

In an environment where the state has its eyes at every level of society and deploys a range of pressure tactics and intimidation in order to implement policies, Rwandans refrain from expressing opinions that do not follow the government line (Purdeková 2015; Sundberg 2014; Thomson 2013). As a result, James Scott's (1990) concepts of ‘hidden’ and ‘public transcripts’ have become a popular way to analyse the Rwandan context (see, for example, Ansoms 2013; Ingelaere 2010; Thomson 2013). According to this reading, ordinary Rwandans claimed to agree with the state's development and modernisation discourse in the presence of authority and in the public arena, whereas when they were far from those in power, and in trusted company, they would speak their minds and articulate hidden, oppositional accounts of their situation. Although I certainly observed such tendencies among my interlocutors, as I shall argue, fear alone does not explain why villagers tried to adapt to the policies rather than oppose or resist them.

Certainly, government authoritarianism and paternalism explain much of the lack of open resistance, a ‘frontstage’ acceptance of policies and the general habit of obedience amongst subaltern Rwandans (see, for example, Mamdani 2001; Prunier 1997; Straus 2006). But subaltern Rwandans also had their own ideas, values and practices which at times overlapped with state projects, which can partly explain why they supported the state projects that made their lives miserable. To begin with, we need to understand how the villagers viewed ‘poverty’.

Poverty and impression management

In Iwacu, everyone was officially categorised as poor. The main source of livelihood was farming, but none of the villagers had sufficient land to produce the amount of food they needed, leaving them constantly struggling to make ends meet. The villagers were repeatedly told by the sector agronomist and other local authorities that they could escape poverty if they dedicated themselves to the state's agricultural modernisation policies. Escaping poverty was presented as a moral obligation and a matter of will, of motivation to be incorporated into the market economy, of taking responsibility for one's own welfare. Remaining in poverty, conversely, is caused by not working hard enough, or by being ignorant of how to adapt to a market economy or by lacking the skills to make ends meet. This view overlooks the structural obstacles to increasing production faced by poor rural farmers. Low productivity was not due to laziness or a bad attitude but to more mundane factors: land scarcity, poor soil quality, expensive fertiliser and lack of cash resources outside agriculture. Yet, this view of poverty as a matter of poor people's ‘capability’ was not only constructed from above and communicated by people in power to their subordinates – it was sustained and reproduced among poor villagers themselves.

Among the 150-odd households in Iwacu, no family was significantly better off than any other. The main source of people's livelihoods was the land, and like most rural Rwandans, the people of Iwacu were not self-sufficient in agricultural production. To cope with food shortages, they were obliged to work as daily agricultural labourers on other, relatively wealthy people's land outside the village. As work on others’ plots was seasonal and work outside agriculture was hard to find, for many months of the year families lacked sufficient food and/or money.

With this harsh reality, the agricultural modernisation policies only made life more difficult. The state-mandated rice in the marshlands did not grow as expected, which resulted in poor yields season after season. Seed potatoes and fertilisers were not distributed for monocropping on the terraces, as promised, and villagers were left with barren land and hunger. At the same time as monocropping was introduced in Iwacu, pressure to fulfil other imihigo targets were also put on the villagers, foremost of which was to obtain the compulsory national health insurance (mutuelle de santé) and enrol their children in school.

The villagers fell deeper into poverty and were worse off than they had been before the state's introduction of monocropping; yet, they were still not considered ‘poor enough’ for their health insurance and school enrolment fees to be covered by the state.⁶ That kind of poverty relief was reserved for landless, elderly and physically handicapped people – the ones falling into the national poverty category of ‘very poor’ and ‘the destitute’ (RoR 2007: 13; see also Ansoms 2008: 19; Gaynor 2014: 53). As my interlocutors all had shelter, a small plot of farmland and the capacity to work, they, like most rural Rwandans, fell into the third lowest category – ‘the poor’. As such, they were expected to adapt to modernisation policies without financial help.

My interlocutor Donata agreed with the importance of health insurance and schooling for her four children, but her meagre resources made it impossible to follow these policies. She had another, more pressing, problem. ‘The most challenging thing in my life right now is to get enough food. We don't know how to get enough food for the day’, she admitted. Not being able to provide for such basic needs as feeding one's family was a source of stress, worry and shame, and a topic carefully avoided. Yet Donata's story was hardly unique; providing a daily meal was a struggle for all villagers, regardless of what poverty category they belonged to. With reference to their everyday life, my interlocutors therefore thought that they merited the same kind of financial help for school and health insurance as people in the poorer categories, especially since these were obligations, not voluntary expenses. After one weekly co-operative meeting, I asked the members if they could plead for a reduction in the health insurance cost given their economic situation. Petero, the co-operative accountant, explained the impossibility of my suggestion, while the others nodded in agreement:

Saying that you cannot cover these health insurance expenses will not change anything! They [the leaders] would laugh at us! They would tell us: ‘How come you are in physical health, and yet you cannot get anything to eat? How come?’

According to this logic, a person with the ability to work should be able to work their way out of poverty and pay for health insurance and school. This view transfers the blame for poverty from structural factors to the poor themselves (cf. Walker 2014).

While many wished that they could be freed from the expectation of being able to cover health insurance and school fees like those categorised as ‘very poor’ and ‘destitute’, no one wanted to be considered that poor. They underlined an important difference between themselves and the poorest; they were still able to keep up a façade, to hide their misfortunes, unlike the poorest, whose conditions prevented them from such impression management (Goffman 1963). As a villager, Agatha, explained:

Some people work so hard, but they do not have shelter or a farm field, and everyone can see that. In that case, it is that person's condition that makes his poverty apparent. But for me, who has a house and a small farmland, I have to hide my needs … If you can hide your poverty, you should; if you are not able to hide it, it means that you are too poor even to hide it. Until you are not able to hide your poverty people pretend that you are rich. But, realistically, you are poor.

Thus, if you had land and the capacity to work, you were expected to be able to fend for yourself, or else pretend you did. Inability to fend for yourself came with shame and the loss of dignity and pride, which people carefully avoided.

Keeping up appearances

One day during rice harvest, co-operative members gathered beside the field to take a break from the hard work. Their harvest was poorer than the neighbouring rice co-operative's, and they talked about how discouraging that was. I had seen how all co-operative members had devised strategies to ensure a rich harvest, so I wondered how they could possibly blame themselves for the disappointing yield, which was due to poor seed quality and insect attacks. Christian, the president of the co-operative, usually the first one to speak, explained:

It is not that we ignore the fact that we have made efforts; we know it is not our fault. But there is this pressure on someone who seems poor in the society. Because, in the community, no matter how hard you have been working, no one will be happy to have a respectful relationship with someone who cannot afford to eat. Sometimes they can speak about you ‘Ah, you see this one cannot do this and this’, and it looks very bad in our society.

Seasonal fluctuation in food availability caused fear of food shortage and vulnerability, but also a worry about being regarded as something less, as people who could not fend for themselves. Donata explained, and the others agreed, that no one would happily maintain a relationship with someone who could not afford to eat. On a separate occasion, at her home, Monique clarified:

Being hungry is a common experience for everybody. So how come your neighbour made it and you did not? You do not have an excuse for failing. It looks bad. You cannot tell [others] that.

Even in a situation where everyone faced similar hardships in everyday life, not managing under these conditions was experienced as shameful. It was, in fact, precisely because they were all in the same situation that there was felt to be no excuse to drown when others appeared to be keeping their heads above water. If others could cope with the situation, and you could not, it must be because you were not working hard enough; because you made the wrong choices; or because you were ignorant or weak. Villagers had, as such, generally internalised the belief that they were responsible for their own poverty and had failed to live up to the prevailing social norms of successful and respectable persons, the practice of kwihagararaho (cf. Appadurai 2013: 185; Walker 2014: 109).⁷ If one failed to keep up appearances, one was judged by fellow villagers, as Christian described:

We never expose our poverty to people so that others may laugh at us or spread rumours about us and say that we are not able to cover our own expenses, for example. That is how Rwandans are.

What one communicates to others, and presenting oneself in a favourable light, was thus considered important so that people would not ‘laugh at you’. Consequently, the poorest of the poor were considered the least capable of caring for themselves and were seen as unable to retain their pride.

My interlocutors distinguished themselves from such poor people by talking about how they were, unlike themselves, not able to plan their harvest, focussing on immediate instead of long-term profit, and how they were not hard-working. This type of discourse had similarities with the local authorities’ efforts to sensitise people to follow modernisation policies. But this type of ‘othering’ (Lister 2004) was in the Rwandan context not just rhetoric: the poorest people felt that fellow villagers did not want to get close to them, as they had nothing to offer others (see also Ansoms 2009: 239). The anthropologist Danielle de Lame, who conducted fieldwork in a rural Rwandan community between 1988 and 1996, explains that ‘Rwandan solidarity does not extend to the Judeo-Christian notion of assisting the poor’. Instead, ‘local solidarity takes the form of mutual gift-giving. People with nothing to exchange are excluded’ (2005: 281). People who felt they were unable to contribute or reciprocate would withdraw rather than seek help from others, and they ended up being utterly left on their own. Rwanda has furthermore been described as an image-conscious society, where one has to guard one's status (Thomson 2018: 156), and as a ‘culture of competition’, where assessing, valuing and comparing is stressed as essential to ‘development’ as a marker of status (Sundberg 2014: 248). Relationships with those who could help one climb the social ladder were therefore a priority for my interlocutors, whereas helping someone with lower social status entailed a risk of being further pulled down than they already were.

There was thus a competitive element to resources and social networks; those able to make others believe they had something to offer were more likely to keep their social position, and those who failed to live up to such norms risked exclusion and stigma (Goffman 1963). This led to practices such as lying, hiding and covering up, practices that have been described as cultural habits in Rwanda (De Lame 2005; see also Sundberg 2014: 215).

Even if one could barely feed oneself and one's family, it was important not to give others the opportunity to see that one has less food or work capacity. During wedding parties and other celebrations in the village, the guests left at least a quarter of the food they were served, to show that they did not need to satisfy their hunger with other people's food. Along the same lines, children were forbidden to eat in the homes of others, as that too was a sign that their parents could not feed them at home. Sociability and being respected had higher priority than satisfying one's hunger. People used every device at their disposal to maintain their status and respect from others.

Hence, even if someone was ill or feeling weak, they tried to keep it a secret from others. For example, one of my interlocutors, Hakizimana, did not show up for the co-operative work one day. I went to his house, where I found him lying outside on a straw mat in the shade, unable to get up and greet me. I asked how he was doing. ‘I am fine’, he said, even though all the signs suggested that he was not. He had a high fever and had not eaten since the previous day. I took him to a doctor, who judged his condition as serious and that he needed immediate treatment for malaria. Hakizimana asked me not to reveal this to the others, as they might then judge him too weak to participate in the rice harvest. This type of concealment was common amongst the villagers. On another occasion, Christian (the co-operative president) had badly injured his leg. Yet he insisted on working as hard as everybody else with the rice harvest, and claimed that the limping and the huge wound on his calf was ‘nothing’. Eventually, the infected wound caused a high fever and he had to stay home. He asked me not to tell the others about his condition: ‘You don't want many people to know that you are sick and weak, because soon people might start to say that you might die’ and become someone people do not count on anymore. How you looked in front of others and how physically strong you were mattered in the village. People's perception of your ability to work was an important factor in being able to make a livelihood.

Failing to show yourself as strong and to keep up appearances was, in this sense, a negation of solidarity. Although villagers in Iwacu did not always relate to others in a calculating way, it was nevertheless the case that those without the ability to reciprocate, or those who felt that they had nothing to offer others, would rather withdraw than seek help. In this environment of competition over limited work opportunities and access to land, people tended to look after their own households rather than come together and question the harsh conditions imposed upon them by the government. The potential for social solidarity gave way to interpersonal competition and status-shaming.

Escaping poverty was thus not simply tied to coping strategies and channelling resources; it was rather a moral obligation. This moral obligation was a state project enacted through the state's ‘changing mentalities’ agenda. Escaping poverty was also a local moral obligation promoted by the villagers themselves. This overlapping, self-reinforcing dynamic can help us understand the overlap between the villagers’ ambiguous striving for development and the modernisation project promoted by the authorities. The villagers who had trouble making ends meet were unwilling to call out the authorities or risk the criticism of fellow villagers. Instead, they tried to achieve the state's impossible goals and to be recognised as hard-working, progressive and intelligent enough to follow the vision of modernisation and to fend for themselves. This was an ideal that they strove for, and that they wanted to identify with, although it was an effort accompanied by frustrations and disappointments, amongst which keeping one's fears a secret from neighbours was paramount.

Future hopes

From my interlocutors’ point of view, they did not have any option but to follow the government's directives as communicated by the district authorities and village leaders. Questioning decisions taken by those at the top of the hierarchy was unthinkable. This was not simply because of fear of surveillance and reprisals for diverging from the government's agenda; it was also because they perceived the subordinate's position as that of a follower, and were quick to position themselves in the societal hierarchy (cf. Thomson 2018: 153).

Although my interlocutors’ wishes about how to cultivate their land often differed from the plans of the government, they refrained from expressing their ideas (see Pottier 1989: 469). They did not see it as their place or role to make demands of those in authority. This view of themselves as lacking the right to determine conditions in their own lives was expressed not only by how the local authorities looked on them, but also by their own self-image. As elaborated, being poor was seen as a sign of lacking the capacity to be ‘successful’. If one was poor, one had no moral right to raise one's voice to those in power. Joseph put it simply: ‘We are poor, so we keep quiet’.

Villagers’ acceptance of the fact that they could not influence policies did not mean that the decisions taken by the government always made sense to them. Many of my interlocutors feared the risks of monocropping, and judged intercropping as more reliable. Despite the costs of monocropping, however, the villagers did not reject the state's modernisation policies outright: these projects also generated hope as an avenue to achieve the dream of a more stable and affluent life. Land was scarce and unproductive, and the population was increasing. This left little for the villagers’ children to inherit, making it impossible for future generations to continue with small-scale farming. A change was needed, and my interlocutors hoped or wanted to believe that the state's modernisation agenda would lead to long-term positive change. For this, some suffering before gain might be inescapable, they reasoned. In this sense, compliance was not simply forced upon them. It was a strategy for achieving development.

When the sector authorities explained that the marshlands could be accessed only if the farmers organised themselves into co-operatives to grow rice, my interlocutors quickly organised themselves to work according to this model. This was a rational choice on their part, since it enabled them to access the land.⁸ But it was also a matter of pride. They explained that the authorities saw them as fit for the task. The shift to monocropping initially gave the co-operative members a hope of rich yields and access to money, as promoted by the sector agronomist. However, my interlocutors saw how the fertile marshlands that had once produced a variety of crops and were an important contributor to food security ended up producing unreliable harvests of rice.

The technical shift to rice cultivation meant that co-operative members lost the right to choose when, what and how they cultivated, and control over daily work routines, as they had to adapt to the co-operative's arrangements and work schedule. Working as part of a co-operative took up a large portion of the day. However, because the rice harvests were poor, members also had to attend to subsistence activities on their personal plots or work on other people's plots simply in order to put food on the table.

Despite the disappointment of repeated poor yields and the struggle to organise work, growing rice in the marshlands was the only option left for my interlocutors if they wished to access the land. As such, they had to play by the government's rules. In this sense, their compliance with policies was a consequence of state domination, but my interlocutors also rationalised it in terms of the state's idea of how land is best used. Although many co-operative members recounted how the fertile marshlands were more profitable to them when they could grow what they wanted, they still simultaneously supported the monocropping quest. Monique said:

It [the shift to monocropping on consolidated land] is good, because before people could grow what they wanted down there, and some people were not very efficient, and some people did not really work their land, so the land was just there, without anyone really taking care of it. Now rice grows there.

We might interpret this statement as expressing Monique's fear of being critical of the state, a fear often observed in Rwanda by other researchers (see, for example, Begley 2013; Thomson 2013; Waldorf 2011). But Monique was also doing something else. She argued that the policies were necessary for others who, unlike her, had misused the land. Petero agreed that, in spite of the challenges they were facing with the rice cultivation, land consolidation in the marshlands was necessary:

It is still good that we all put the crops at the same time and we harvest at the same time. It is more effective. Before you could find someone who was not growing anything on his plot. And some people did not cultivate well. So, it is more effective now, and I think we will have some positive result of this in the future.

Such sentiments were common amongst my interlocutors; even though the harvests had been disastrous, many thought that the agricultural modernisation policies were necessary to prevent others from misusing the land, and that seeing profit from this new way of working the land was a matter of patience. Some of my interlocutors constructed a view of other villagers whom they saw as somehow not as progressive and hard-working as themselves. In a strange way, therefore, some villagers came to support the state's authoritarian measures against vulnerable villagers whom they viewed as weak, lazy or ignorant, even though they did not consider this treatment necessary for themselves.

Thus, the co-operative members rationalised the government's direction and, to some extent, accepted their personal suffering for the higher good of effective land use, even if this compliance directly increased their own suffering (cf. Appadurai 2013). This type of rationalisation and acceptance was also a symptom of a situation that they could see little possibility of escaping. Exiting the co-operative and giving up the marshlands would be even more disadvantageous than remaining. They had no other land to turn to in its place, as even their hillside plots had been terraced and assigned to monocropping. Compliance, therefore, and showing that they were worthy of the marshland were in their interest. My material thus shows a compliance continuum and three dimensions to compliance in the Rwandan context: (1) compliance as sheer subordination; (2) compliance as an expression of village values of escaping poverty and hard work; (3) and compliance as an overlap of interest, where villagers saw a personal gain in being ‘exemplary farmers’.

The co-operative members thought that their reputation as a group was critical to avoid losing access to the land. During bad harvests, it was even more important to show unity and hard work and to show the authorities that they did everything in their power to succeed with the harvest. However, the hard work without any return affected morale, and they came to see their obligations towards the co-operative differently. My interlocutors described a division between those who questioned the amount of time spent working for the co-operative and prioritised working elsewhere, and those who argued that they needed to work harder towards a long-term goal. The president Christian, together with the secretary Petero, advocated most strongly for the long-term goal, supported by most members. This group of members had to some extent embraced the government's discourse of how to become ‘developed’ and achieve success – which was by trying harder and making sacrifices today for subsequent profit. They accused those who were less motivated to work of having difficulty in understanding long-term investments and of being unable to see what it took to grow as a co-operative.

As the leaders of the co-operative could not offer the members any immediate carrots for their work, they were left only with the stick: they threatened to exclude those who did not show up for work, which would deprive them from accessing the marshlands. In this way, the co-operative leaders and their supporters operated as an extension of the state, trying to impose directives from the sector onto their fellow villagers, with punishments for those who were unwilling to invest the effort to make the monocropping work. The values of the agricultural modernisation policies thus came to be reproduced and sustained among the very persons who were being harmed by these policies. The project only had room for those who could demonstrate their ability to work the land according to the state's wishes. Those who were perceived as lacking this ability, or ambition, were seen as obstacles to success and were driven off the project by other members, or left voluntarily when they saw that they could not contribute in the way expected. Thus, people competed with each other, individually or as groups, on monocropping and land consolidation based on rules that they themselves had not set, and showed at least ‘frontstage’ loyalty to the state and its policies (Hirschman 1970), even though compliance affirmed their own subordination (Appadurai 2013: 185).

Most members remained. In between the disappointments of poor harvests and conflicts, they hoped that their reality would be different one day. As Donata put it:

I hope to see some change. It [the co-operative] is like a baby. You give birth to the baby, then you wait patiently for it to be big enough that you can tell it ‘Go there and bring me something’.

There was a general belief that the suffering caused by agricultural transformation today would be fruitful for future generations. These two contradictory views – (1) being upset with and worried about the immediate effects of modernisation policies; (2) while seeing a need for change that presupposed suffering before gain and supporting the agricultural modernisation techniques – were difficult for villagers to reconcile.

Rationalisation and compliance

Villagers admitted that adhering to state policies was unattainable for poor people like themselves. However, they were not critical of these policies per se. Most people could argue for the importance of school enrolment and health insurance, that terraces prevent erosion and that monocropping was likely to be more productive than traditional intercropping. This was despite the fact that they had small farm plots on which they needed to grow a variety of crops to ensure production throughout the year. What stood between them and full participation in the development project was not their mindset, they argued, but their limited resources. The constant reminders of the importance of these policies were thus reminders of how they had failed to live up to the very ideals they themselves to a large extent supported. Compliance with state policy was a means of affirming these ideals, and articulating the importance of government policy was a way to be close to power and not identify as ignorant farmers.

When the Iwacu hillsides were going to be terraced and prepared for the monocropping of Irish potatoes, villagers did not contest it. Such a project had been implemented on the nearby ‘Eastern hill’ a year earlier. The farmers cultivating the ‘Eastern hill’, after terraces had been built, were provided with chemical fertilisers and seed potatoes by the local authorities. Their rich harvest was noted by the Iwacu villagers, and the villagers were eager to see the same result on their own land.

However, for Iwacu the potato-cropping played out poorly. When the local authorities showed up to make terraces, villagers had crops growing on the hillsides, ready for harvest in a few weeks’ time. The authorities claimed to have no time to wait for the harvest and directed villagers to uproot the nearly harvested crops. Villagers were not compensated for this loss and worried about how to ensure having enough food for the coming dry season. To make matters worse, the sector authorities never distributed the seed potatoes and the fertiliser for the terraces as promised, leaving villagers with barren land for one-and-a-half years.

Villagers who depended on small plots to feed their children could not afford agricultural experiments, losses and errors. Yet it was exactly this uncertain experimentation that was imposed upon them. Nevertheless, their criticism was directed not at the policies themselves, but at the sector authorities and the sector agronomist's flawed implementation of these policies.

The problem as my interlocutors saw it was not only that the authorities had mismanaged the project; it was also that they were denied the opportunity to be modern, high-value producers, as they felt they had the capacity to be, just like the people on the ‘Eastern hill’.

Whereas the local authorities were commonly regarded as the long arm of the state, the villagers did not see them as working towards government goals. In fact, the local authorities were perceived to be failing the government and its vision of modernising Rwanda by not implementing the government's policies correctly. Through criticism of implementation, the villagers rationalised and legitimised the agricultural transformation. They blamed the sector officials for not providing the resources, technology and expertise they needed to transform themselves into the modern subjects the state had imagined. To reproduce the state discourse was a means by which the villagers could ‘perform’ development and affirm that they were not backwards-thinking. Blaming local authorities made it possible to continue supporting the modernisation project in the abstract and to feel included in the overall drive to modernise, thus rejecting the label of ignorant villagers who were inherently resistant to change. This commonly produced and sustained story of how the local authorities had excluded them from the modernisation project also enabled the villagers to feel equal to their more successful neighbours on the ‘Eastern hill’, who had received fertiliser and seeds on time.

Even though villagers argued that they would have had a better return from traditional intercropping than performing monocropping without the right conditions to succeed, the villagers knew that a return to old methods of cultivation was unlikely to be authorised. For this reason, they ended up being dependent on the authorities’ help to succeed with the monocropping.

The villagers expressed that they would have liked the agronomist and sector leaders to be more present locally, to understand their lives, to teach them and provide the right conditions to follow the modernisation policies successfully, and to understand the hardships that broken promises had brought. The villagers were reaching out for a closer relationship with the state. At the same time, the villagers also wanted the sector leaders to keep a distance, because they only came to the village when they suspected that the villagers opposed policies or when the imihigo targets were not being met. Villagers were thus both constantly surveilled and controlled by the state, while at the same time they felt excluded from the state project. It was the fear of being left utterly on their own without the right kind guidance in trying to adopt state-approved farming that marked their lived experience. Fear of abandonment was more troubling than the urge to autonomy.

This contradiction, of aspirations for the future and adversity in the present, left them with a feeling of abandonment and frustration. The villagers aspired to be part of something that they were prevented from accessing. As a result, they oscillated between support for and rejection of the modernisation project. They supported state policy, but they were frustrated with its implementation. They understood the ideas of monocropping and terracing, but they knew they could profit more from traditional intercropping. They wanted to be modern, but they did not want to be victims of modernisation. It is this oscillation that explains some of my ethnographic observations. Their compliance with agricultural modernisation policies was thus closely connected to the values they held and to a prosperous future that they strove to fit into. In the meantime, their reluctance to embrace state policies was connected to the reality they faced when the modernisation policies did not bring them what they hoped for and what they wanted to be a part of.

Conclusion

In Rwanda, compliance can be analysed as a continuum of voluntary participation, acquiescence, fear and oppression. The Rwandan population was closely surveilled and controlled, and experienced reprisals if they did not follow directives. Fear of repression is certainly part of the explanation of why relatively powerless people do not oppose or resist those in power. However, it is not – as I have tried to show – the only explanation. The overlap between local and state values explains why voluntary participation and acquiescence is more common than outright resistance.

Villagers in Iwacu connected poverty to a lack of dignity and some level of individual failure. With this perception that poverty emerges from a personal inability to look after oneself or to work hard, most people did their best not to let others see their shortcomings in making ends meet. In turn, this created a competitive environment, in which one was always mindful of the impression one made in order to retain dignity. Instead of coming together to question decisions taken above, they endeavoured to be part of the state development project and to fit themselves into an image of being developed and modern, even when the conditions for reaching this level of progress were simply impossible. This passive support or acquiescence reinforced the state ideals and norms amongst the villagers themselves. Hence, the overlap between state modernisation and local norms partly explains their compliance.

Even while the villagers’ range of strategies to resist government policies was limited, my research shows that, in many cases, villagers and elites shared the same visions and desires. Adherence to state policy, therefore, was not just ‘frontstage’ behaviour, expressed solely in public transcripts. It was also expressed in backstage forums: what Appadurai refers to as compliance with dominant norms (2013: 185). This can help us understand how and why people comply with policies produced from above. The ethnographic task is to discern how these norms are manifested in local lives and local perspectives. People's views about policy begin with their views about themselves.

Pride and dignity, although immaterial elements, provide an important perspective for understanding how subordinate groups relate to development strategies and other interventions imposed upon them. Ideas of how to be a respectable person (kwihagararaho), of how to project oneself to others in one's own community and how to maintain pride and dignity create and reproduce the norms and ideals by which people live their lives. People's norms and ideals invariably clash with the possibility of actually living up to them, resulting in frustration and anger that may be directed at those in power (cf. Scott 1990). But these frustrations can also be projected onto others who are seen as not meeting societal norms, thus reinforcing the pressure to live up to those norms. Oscillation between desire for an ideal and the frustration of not being able to live up to that ideal is a form of social practice worth studying further.

Identifying and highlighting the oscillation between coercive and voluntary compliance, between subordination and acquiescence, can help us discover underlying factors that might explain compliance with dominant powers, how these powers impose decisions on subordinates, and why subalterns do not ‘resist’ even when their living conditions are threatened. In sum, we need a more sophisticated, ethnographic understanding of repression, compliance and resistance. And we need to accept the ethnographic fact that subalterns might share the projects of those who hold power over them.