Introduction
Tanzania has recently made headlines as a hotspot for global land grabbing – the rapid surge in sales, leases and concessions of land both for speculation and investment driven by the triple crisis in finance, food and fuel in 2007/8 (Borras and Franco 2010; McMichael 2010; Third World Quarterly 2013). Reliable data are notoriously difficult to obtain. The last agricultural census – undertaken before the land grab in 2003 – reported one million hectares covered by large-scale farms, of which about one-third were state farms (URT 2006). Only in 2013 did the Tanzanian Ministry of Lands, Housing and Human Settlement and Development (MLHUD) produce a comprehensive list of all farms over 50 acres (MLHUD 2013). The list does not provide any reliable information on the total acreage covered by large properties, while guesstimates indicate that about one million hectares have been leased to foreign companies (GRAIN 2012).
These figures do not sit well with the image of a country dominated by small-scale farming, where large-scale farms are relatively absent, land is an abundant resource and capital and labour are the major constraints on rural development.
Early scholarly analyses of land grabs in Tanzania confined themselves to descriptive labels of the people affected in different ways by land grabs – alternatively named ‘actors’ (Benjaminsen and Bryceson 2012), ‘indigenous people’ (Gardner 2012), ‘rural communities’ and ‘local communities’ (Nelson et al. 2012), or with essentialising cultural labels (e.g. ‘the Maasai’) or conflating productive roles with social identities (e.g. ‘pastoralists’ versus ‘farmers’). But did these ‘local people’ – the ‘people of the land’ (Bernstein 2005) – live in a political and social vacuum before the land grab set in? This question has been addressed relatively late in the land grab debate (Borras and Franco 2013), which has often celebrated rural civil society through problematic simplifications (Brass 2007).
This paper sets out from the proposition that scholarly research on rural societies in East Africa ‘must recognise both the varied local class forces and the existence of conflicting interests within each area’ (Cliffe 1977). It makes two contributions in the land grab debate. First, it provides a historically grounded analysis of longer processes of land dispossession in Tanzania by focusing on the role of state-led dispossession.
Second, it points to the contrast between accumulation from above (either through state-led evictions or by capital-led grabs) and accumulation from below, through the class dynamics emerging from increasing social differentiation and rural class formation, underlying the trends of land concentration and dispossession.
Revisiting the land question
Scholarly and policy debates on Tanzanian agriculture, while diverging on many points, all tend to assume the predominance of small-scale farming in the country. The idea that landless people, landlords and big farmers are not a common occurrence in the Tanzanian countryside is still very much entrenched. Just as in the past, neo-populist Chayanovian scholars assumed that a homogenous Tanzanian peasantry existed (Hyden 1980), present-day economists tend to generalise about Tanzanian smallholders (Binswanger and Gautam 2010; Skarstein 2010; Hella et al. 2011). While in the 1970s ujamaa policies had encouraged class-blind scholarly analyses, the discussions on the conflicting class interests and class formation in the Tanzanian countryside had been opened up (Rodney 1975; Cliffe 1977; Saul 1974), recognising the uneven pattern of commodification of rural production (Bernstein 1979). Yet the flawed assumption that all people somehow involved in farming share some common general interest (Mbilinyi 1988) has been perpetuated up to the present, not least because the conflicts between small and large farmers are not examined in scholarly research. This happens as sociological appraisals of differentiation in the countryside (Sender and Smith 1990; Mueller 2011) are disjointed from analyses of social and political conflicts, where class dynamics and struggles over land and labour become more visible.
From the early 1990s onward, struggles in the countryside steadily increased and coalesced around the land question. Land became increasingly politicised with liberalisation policies, which encouraged land hoarding by individuals with political clout. This added to the myriad of land disputes emerging from villagisation, and was exacerbated by an extremely centralised land administration, that was becoming more and more inefficient and corrupt.
As part of the package of liberalisation and privatisation policies, international financial institutions pushed for land law reforms to property rights (URT 1994; Kiondo 1991; Gibbon 1996; Sundet 1997). As shown by the work of the Presidential Commission of Enquiry into Land Matters, or Land Commission (URT 1994), this emphasis was at odds with the priorities of the vast majority of Tanzanians who, rather than pressing for strong private property rights, were requesting a radical democratisation of the land tenure system, with devolved authority to the villages and stronger local control over land access.
The impressive report of the Land Commission is, incidentally, one of the few sources on collective land claims on Tanzanian state farms and ranches – an ‘extremely under-researched’ topic (Bryceson 1990, 198). As with several other parastatal agencies, state farms tended to heavily rely on foreign capital and state debts (Loxley and Saul 1975). Although the number of state farms has steadily increased throughout the 1970s and 1980s, it figures marginally in the scholarship (Greco 2010; Chachage and Mbunda 2009; Izumi 1998; Itika and Makauki 2007).
This paper presents research on a privatised state farm in southwest Tanzania (Mbarali District, Mbeya Region). This case is interpreted as a land grab ante litteram – ahead of its time, as the farm was established in 1991 and privatised in 2005, just before the global land rush set in. Yet it shares three features commonly associated with recent land grabs. The first is that it involved an attempted conversion from food to fuel crops (Borras and Franco 2010). The second is that the land grab was coupled with a substantial water grab via the appropriation of a large-scale irrigation infrastructure (Woodhouse and Ganho 2011; Water Alternatives 2012; Allan et al. 2012). Third, this farmland became financialised as the acquiring company became involved with a private equity fund (Daniel 2012).
Primary research included participant observation, a socio-economic survey of 67 households, 60 life histories in Kapunga village, and around 200 semi-structured interviews with key informants throughout Chimala Ward, Mbarali District. Evidence was collected throughout 12 months of ethnographic research and 6 months of archival and secondary research, conducted by the author as a doctoral student in Mbeya and Tanga Region in 2007/8.
After sketching the history of the study area and of the state farm, the paper presents the results of a survey on social differentiation, then goes on to document the process of class formation in this rural locality and the relevance of class dynamics in local land politics. The concluding section reflects on class dynamics in the wider context of financialised agribusiness in East Africa and on the organisation of political responses to land grabs.
The study area
Mbarali District is a key national rice farming area that broadly coincides with the hydrogeological basin of the Usangu Plains – an extended irrigated lowland, best understood as a ‘wetland in drylands’ and an area of high commodification (Woodhouse et al. 2000), where rural wage labour is common and land is rented and sold on de facto land markets. Predominantly pastoral until the late 1950s (Hazelwood and Livingstone 1978; Walsh 1984), Mbarali is nowadays a highly specialised area of commercial rice farming, dry-season irrigated horticulture and commercial cattle breeding. Rice farming in Mbarali has a long history, dating back at least to the 1920s and intensified at independence with the construction of the first rice state farm in Rujewa. Since the 1960s, the agricultural frontier has steadily expanded at the expense of pastoral and fishing activities. Dry-season pastures and fishing ponds have been converted into rice fields. This process of agrarian change, combined with sustained immigration flows, has brought about heightened competition for land and water (Charnley 1991, 1997; SMUWC 2001). Since 2006, the enlargement of the Ruaha National Park has led to the eviction of wealthy families owning large herds, besides the forcible resettlement of nine villages (Tenga et al. 2008). Based on a questionable narrative of environmental crisis in the Ruaha river basin,the government's decision to enlarge the Ruaha National Park caused further land enclosure; nine villages were forcibly resettled and 200,000 pastoralists and their herds evicted. In particular, the scapegoating of pastoral groups covered up the fact that the enlargement of the park is profitable for private touristic companies operating in the area (Lankford et al. 2004; Walsh 2007, 1996, 2006). Besides having shifted the ‘issue' to neighbouring regions (Maganga et al. 2009), this process of environmental enclosures and the attached evictions for conservation (Brockington and Igoe 2006) have aggravated land and water conflicts in Mbarali. Against this background of increasing competition over irrigated farmland, the following section presents the politics of land of one of the two large-scale irrigated rice estates in Mbarali.
A brief history of the National Food Corporation in Kapunga
Kapunga Rice Farm has a contentious history, from its early establishment to its divestment. In the mid 1960s, the then Usangu Farmers’ Cooperative had tried to establish rice farms in that remote, flooded lowland, failing in some years because of the lack of water control infrastructure (TNA 1966). In order to respond to irrigation problems, Kapunga Rice Farm was planned as an irrigation scheme, and paid for by the Tanzanian government through an African Development Bank (AfDB) loan co-financed by the Nigerian government. The farm was built by expropriating village land in 1985 (NAFCO 1985). From 1987 to 1989, 73 families were dispossessed and resettled to make space for the estate.1 A small group of local residents – all belonging to the early settler families of the area, recognised as village founders – opposed the resettlement, but to no avail.2 The farm started operating in 1991 under the management of the National Food Corporation (NAFCO) – one of the largest parastatal corporations involved in large-scale food production – which managed it until privatisation in 2005.
The farm, whose land has been levelled for the construction of water canals, has two components: the large-scale estate (3015 ha) and a smallholder scheme (800 ha). These are interdependent, as they rely on a single, shared water canal. The estate is composed of six-hectare rice plots, pooled together for mechanised production, plus the estate headquarters, the workers compounds, a large rice mill and, a warehouse, and other undeveloped areas (AfDB 1995). The smallholder scheme is subdivided into one-hectare plots, which were allocated to farmers from throughout Chimala Ward.
At its outset in 1991, the estate employed very little agricultural labour, as rice farming was completely mechanised (AfDB 1995). Tractors tilled the land, mechanical dry-seeders sowed the rice, an aeroplane sprayed the fields with fertilisers and pesticides and combine harvesters harvested rice. About 60 parastatal staff and their families moved from all over the country to the estate headquarters, which included a dispensary, a primary school and a canteen, serviced by an all-season road, water pipes, electricity and a bus service to Chimala town, which rapidly grew into a prosperous trade centre. Around the farm, local residents benefited from subsidised fertilisers and pesticides, cheap rental of tractors, cheap rice husking and the fringe benefits in free social services. Managers would let them access estate land that laid idle free of charge – mainly non-irrigated plots suitable for maize farming. This paternalistic mode of management aimed to defuse the latent class conflict between local residents and the staff of the parastatal, which was sharpened by the lack of the most basic infrastructural services in the surrounding villages.
On the estate, the model of capital-intensive mechanised farming was gradually undermined by unforeseen pest control problems. Rice production shrank. In 1997, the farmed area had decreased from the initial 3015 ha to only 256 ha (NAFCO 1997). The capital-intensive design, coupled with poor management and widespread corruption (TAC 1997), led the estate to run at a loss relatively soon. In 1997, the estate was at risk of bankruptcy and failed to pay the workers’ salaries (NAFCO 1999a, 1999b). Estate managers decided to rent out the farmland and machinery to private farmers, to use the rent income to keep the infrastructure up and running and to pay the staff.
De facto privatisation of NAFCO-Kapunga
From 1997 to 2005, the NAFCO management stopped having any role in production and limited itself to repairing and managing the irrigation infrastructure, managing the farm as a paternalistic landlord. The decrease of public funds to the farm was thus offset through land rent and the lease of agricultural machinery to local private farmers. This de facto privatisation of rice production on the estate mirrors the changing political context of the mid 1990s when, with impending privatisation policies, parastatals started to be dismantled. Since 1997, tenant farmers started running the estate by renting six-hectare plots. The organisation of production changed dramatically, as rice farming on the estate was demechanised, by employing wage workers in place of agricultural machinery. This caused a rapid, significant increase of wage labour, and migrant workers flocked to Chimala Ward from the surrounding areas.
NAFCO tenants adopted a rather homogeneous organisation of production. They all used tractors for land preparation and applied chemical fertilisers and pesticides, which they bought from NAFCO at subsidised prices. This notwithstanding, as the following section will show, they were far from being a homogeneous social group. I was able to distinguish four categories: local large farmers, local rural capitalists, NAFCO staff and urban professionals. With the exception of the local large farmers – who personally worked on their fields with their family members – the rest of the tenant farmers on the NAFCO estate hired wage workers and limited themselves to supervising the farm.
In general, all NAFCO tenants displayed higher levels of agricultural investment than small and middle-scale local farmers, coupled with an exclusive use of wage labour rather than family labour. In parallel to this process, they also started to acquire land and build irrigation canals in the farming areas surrounding the estate, quickly converting the village commons – mostly pastures, flooded lands and fishing areas – to rice farming.
Village authorities allocated these village commons to large farmers and rural capitalists who built irrigation canals to farm rice. This parallel process of land enclosure and commodification accelerated after 1997, as with NAFCO plots on offer, a wider variety of people moved to Chimala to invest in commercial rice farming.
By investigating the process of local land use change through oral history (Greco 2010), one important element has emerged: in the decade from 1996 to 2006, village authorities were called upon by small and middle-scale farmers to intervene in five different episodes of collective land conflicts. All the five episodes revolved around newly established irrigated rice farms and saw small and middle-scale local farmers contesting land allocations to local large farmers, rural capitalists and NAFCO staff – most of whom were also NAFCO tenant farmers – who had expanded their rice farms on village land, until then used as commons for dry-season grazing.
These land conflicts expressed a class contradiction, which village authorities tried to handle through village land allocations, where once a year all residents would apply for the individual allocation of village land against the payment of a nominal fee. These allocations accommodated both the interests of these richer groups and those of middle-scale and small farmers and defused the class contradiction, avoiding open conflict between the richer and the poorest sections of rural society. But what is the extent of social differentiation and how does it conflate into class dynamics? The following sections will address the extent of social differentiation by discussing the results of a household survey administered to a stratified sample of 80 households in Kapunga village during the agricultural season 2007–08.3
Land and social differentiation: the survey results
The survey showed that rural society in the study area is starkly divided. From the data, it emerged that the most vulnerable groups in this rural society are poor farmers and rural workers, who are too poor to farm and regularly sell their labour to others to ensure survival. They represent the lower tercile in the survey (wealth ranks 1 and 2). Middle-scale rural producers also sell their labour, but only sporadically and share similar levels of on-farm investment and diversification strategies. They represent the middle tercile (ranks 3 and 4). Large farmers and rural capitalists are the upper tercile (ranks 5 and 6) (see Table 1).
| Wealth rank | Category | Description | Tercile |
|---|---|---|---|
| 1 | Rural workers | Too poor to farm. Reliance on their own capacity to work. Casual employment. Lack of property and access to land and means of production. | Lower tercile |
| 2 | Small farmers | Small farms (0.2–5 acres) both owned and accessed. Hoe farming, occasional rental of tractors. | |
| 3 | Lower middle farmers | Average sized farms (2–10 acres) both owned and accessed. Hoe farming, occasional rent of tractors. Occasional use of pesticides and fertilisers. High degree of indebtedness to moneylenders. | Middle tercile |
| 4 | Upper middle farmers | Average sized farms (2–10 acres) both owned and accessed. Reliance on family labour. Occasional hiring of wage workers. Regular use of pesticides and fertilisers. Petty trade. Credit from SACCOs. | |
| 5 | Large farmers | Larger than average commercial farms (50–100 acres). High working capital (over TSh1,000,000). Credit from savings and credit cooperatives (SACCOs). Ownership of power tiller, rental of tractors. Regular use of pesticides and fertilisers. Mix of family labour and wage labour. Personally working on the farm. Double residence (village and town). Retail trade. | Upper tercile |
| 6 | Rural capitalists | Large commercial farms (>100 acres). Higher working capital (over TSh2,000,000). Credit from banks. Ownership of tractors. No family labour, only wage labour (use of paid farm workers) for farming. Not personally working or supervising the farm. Double residence (village and town). Wholesale trade. Moneylending often at usurious rates. |
Given the presence of the NAFCO estate and smallholder scheme, Table 2 presents a picture of the survey results according to the typologies of rice farmland.
| Typologies of rice farmland | Size | Irrigation system | Typologies of farmers | Level of investment | Wealth ranking |
|---|---|---|---|---|---|
| NAFCO estate plots | 6 hectares | Intake from Ruaha river; concrete all-weather water canals | 1 local rural capitalist | High | 6 |
| 2 local large farmers | 6 | ||||
| 3 NAFCO staff | 6 | ||||
| 4 urban professionals and civil servants | 6 | ||||
| NAFCO smallholder scheme | 1 hectare | Secondary concrete all-weather water canals from NAFCO estate | Upper medium-sized farmers Lower medium-sized farmers | Medium | 5 and 4 |
| Ordinary irrigated plots | Variable | Low input, temporary mud and sand furrows | Small, medium-sized and big farmers | Variable, according to farmer and size of plots | All |
The survey demonstrated that the main difference between the lower and the middle tercile is the level of investment in farming. Most households in the middle tercile show investment levels high enough to farm on the smallholder scheme, where the use of fertilisers is mandatory because of declining fertility. In the agricultural season 2007/8, the average investment on plots belonging to the scheme was higher than that on ordinary irrigated plots of similar size, implying that the scheme is the domain of lower and upper middle farmers (middle tercile) – many of them being early participants to the scheme who had profited from high land productivity in the early years of the scheme – while it excludes the poorest rural producers (lower tercile). Scheme farmers are best described as petty commodity producers engaging in simple commodity production without substantial accumulation (Bernstein 2010) – a highly unstable social formation, far from the ideal of a self-reproducing middle peasantry. Another significant result is that households in the upper tercile considered the scheme too small for their investment. They did not farm on the scheme, but rented NAFCO plots in the previous five years. Remarkably, two of these households owned large farms – of about 200 acres – of ordinary irrigated farmland.
Labour
A significant gender element emerged in that all female rural workers were rice gleaners, while the rest of farming operations showed no clear-cut gendered social division of labour. Women tended to perform all tasks of the rice labour cycle. Non-wage labour groups have almost disappeared in the area: only 13% of the sampled household members had participated in a non-wage labour group in the last five years, while 69% had relied on both family and wage labour for farming. More specifically, 34.5% of the sample had hired wage labour and 34.5% had both hired wage labour and provided wage labour themselves, the latter belonging to wealth ranks 1–3. The incidence of wage labour is explained by the crucial role of specialised wage workers for key operations, such as rice transplanting and weeding, where unskilled workers can potentially undermine the harvest of a whole farming season. This explains why also small and middle-scale farmers hire wage labour at peak labour times, although the bulk of the labour process is carried out by family labour.
Land
Landlessness is significant among the lowest tercile: about 70% of the household members are absolute landless people, as they neither possess land nor have access to it. None of them had ever farmed in the smallholder scheme in the past. In wealth rank 3, the percentage goes down to 40%. Poorer middle-scale farmers farm an average of 1 to 3 acres; 35% farmed on the smallholder scheme and 15% had farmed there in the past by paying rent. In rank 4, this percentage rises to 60%. The middle tercile portrays the unstable, shifting basis of a middle peasantry who struggles for survival through simple commodity production. No household in the middle tercile had ever rented NAFCO plots, as these require much larger capital investment.
Renting land is the main way to get access to good quality irrigated plots in the area. About 46% of the sample households rented land, while 28% both rented out their farm and rented in higher quality farmland. Almost the half of them belong to the upper tercile (wealth ranks 5 and 6) and a few upper middle producers (rank 4).
The superior quality of farmland and irrigation infrastructure is crucial in rice farming. Both middle and large farmers minimise crop failure by renting out inherited plots with low productivity and renting in higher quality plots. This explains why 14% of the landed farmers in the sample rent land, both in and out. Significantly, all of them belong to the upper tercile.
Another significant result is the fact that non-monetary access to land-borrowing or sharecropping agreements and usufruct rights were reported only in 18% of the sampled households – most of them being landless households – and were restricted to non-irrigated land. Irrigated farmland is never accessed for free, even among close relatives. As rice is a cash crop and hardly ever farmed for home consumption in Mbarali, commodification of agricultural production went hand in hand with commodification of land.
None of the rural workers and small farmers had access to the highest quality irrigated farms. None had farmed in the irrigation schemes in the past 10 years.
In contrast, about 77% of rural capitalists and large farmers (upper tercile) farmed either in the small irrigation scheme or on the estate; the three households which did not were engaging in commercial cattle keeping.
Capital
The scale of working capital and the role in production mark the difference between local large farmers and rural capitalists. The two groups have some common elements: they enlarge their landholdings, buy off the best quality farmland, hire wage labour, invest working capital in pesticides, fertilisers and agricultural machinery, and reinvest working capital both in farming and in trade and services. In contrast to rural capitalists, large farmers do personally work on their farms. Rural capitalists operate at a larger scale than large farmers and rely completely on wage workers. Their credit needs are too large to be met by informal credit institutions such as savings and credit cooperatives (SACCOs). Only three of them had farmed on the smallholder scheme and they lamented the small size of plots; most rented up to 30 hectares on the scheme.
To conclude, in the highly commodified rural life in Mbarali, NAFCO tenants stood out as having interests antagonistic to the rest of rural society.
De jure privatisation of NAFCO-Kapunga and the politics of land
In 2004, the government announced the imminent privatisation of NAFCO-Kapunga. NAFCO tenant farmers established a cooperative society – the Chimala Agricultural and Marketing Cooperative, or AMCO (NAFCO 2004). They then demanded that the government sell Kapunga Rice Farm to their coop (AMCO 2004).
The coop would have been in charge of the estate headquarters and storage facilities, while the estate farmland would have been redistributed to its members, keeping the existing formula of six-hectare plots. This plan amounted to a formalisation of the de facto to a de jure privatisation.
But the cooperative failed to mobilise local support in the crucial political circumstances of a potential reorganisation of property. Middle and small farmers refused to join the cooperative and support its claims, as the material interests of the coop members – ownership of large farms, commercial production through wage labour, and credit through the banking system to guarantee higher levels of capital investment – emerged as antagonistic to those of the large majority of rural society, composed of rural workers, small farmers and (upper and lower) middle farmers.
This vast majority realised that land redistribution to the cooperative would have consolidated rural capitalism, and therefore would not have defended their interests, but would instead have gone against them. They thus proposed an alternative land redistribution plan: the estate would be fragmented into small, one-hectare plots to allow more people to access the highest quality land in the area in parcels of a size manageable without large capital investment. This plan of redistribution by fragmentation was backed by the Rufiji Basin Water Office (RBWO 2005, and personal interview4) on the grounds of the greater water efficiency of low-input, small-scale irrigated farming – where temporary water canals allow the water to flow back to the river after watering the rice fields – in comparison to large-scale irrigated farming (Kadigi 2006).
Middle farmers took the lead in this initiative and succeeded in mobilising small farmers and rural workers towards the common goal, in a somewhat fragile inter-class alliance against rural capitalists. In the end, the large majority of Chimala Ward supported the plan for land redistribution by fragmentation. The two alternative proposals for land redistribution express the class contradiction between middle and small farmers (joined by rural workers), and rural capitalists (joined by large farmers and urban professionals).
The decision to privatise the farm was discussed in simultaneously with – and because of – the political campaign for the 2005 general political elections. The plan for redistribution by fragmentation was championed by the then Mbarali parliamentary candidate, who refused to back up the cooperative. Her populist stance won her considerable support during the political campaign. This modality of land politics, with promises of land redistribution from parastatal estates, is not new in Tanzanian multiparty politics (Kelsall 2000). The candidate was elected, but her perseverance in her lobbying against privatisation was condemned by her own party, which attacked her via a campaign of defamation. She dismissed the scandal by attributing it to the internal split of Chama Cha Mapinduzi (CCM, the ruling political party) into two competing factions in Mbarali District (Habari Tanzania 2006). But the party stance on privatisation was contested through a protest vote in Chimala Ward. The district party cadres had rejected the nominations that came out of the primary party polls – two popular young rural capitalists who had led the campaign in favour of redistribution campaign – and nominated some low-profile candidates who had played no role in the anti-privatisation campaign. Party supporters voted for opposition party candidates to protest against this lack of internal democracy within the party, which was also interpreted as a sign of the fact that the sale to a big investor had already been decided at the echelons of the party hierarchy at the national level. While at the ward level the discontent led to a protest vote, the mobilisation did not scale up to district and regional level. When the farm was eventually auctioned in 2006, the local MP was still pleading to the National Assembly for land redistribution to small producers (PSRC 2007; Parliament of Tanzania 2006) and the cooperative was holding protest demonstrations at the Parliament's headquarters in Dodoma. The press conveyed the impression that the farm had been sold at a throwaway price to politically connected investors (Mtanzania 2005; Kasi Mpya 2006). These allegations are difficult to prove as the auction documents have not been made public and no legal action has been taken.5
While the economic rationale of this privatisation cannot be clarified for lack of hard data, the political dynamic is clear: privatisation took place over the heads of the interests – if conflicting and contradictory – of the different groups of farmers, rural capitalists and professionals investing in commercial farming in Chimala, who furthered collective land claims for redistribution, although on diverging terms and without forging an alliance.
It is crucially against competing claims for land redistribution that the government defended the superiority of large-scale farms over small-scale farms. This is what makes this case a land grab ahead of its time: the decision to preserve the size of the estate intact, instead of fragmenting it and redistributing the land to farmers, which in turn affected successive decisions. The decision to keep the large-scale design called for an investor with substantial working capital – or with the credibility and connections to borrow it from financial institutions.
The investor – or a landlord in the making
In this context of political contestation, the new management took over the farm in 2006, taking heavy-handed decisions to curb local dissent. These included prohibiting access to estate land for local residents; closing off the estate road; revoking the usufruct agreements conceded to local residents to farm maize on idle estate land; and later evicting 13 families considered to be residing within the estate boundaries by burning down their mud-and-thatch houses after they had refused to move, without compensation. A wooden bridge on the main water canal, which connected the village to private rice fields along the estate border, was demolished, leaving about a hundred farmers cut off from their fields for the whole farming season. This was a retaliatory measure against one large farmer who was particularly vocal in the campaign against privatisation.
It is against this background that an intractable land and boundary dispute between the estate and the adjacent village of Kapunga arose: the new management considered it a village of squatters who had encroached upon private property, and they alternated threats of eviction with attempts at settling the dispute through political mediation.6 In the early stages of the privatisation, a generalised climate of political intimidation throughout Chimala Ward was a consequence of the connivance of district authorities with the investors, which partly explains why none of the victims of the abuses ever resorted to the law, even though some had the money and the capability to do so. Tenants on the smallholder scheme felt that the threats extended to the smallholder scheme as well, since the estate and the scheme were registered under a single land title.
Local residents resented the lack of investment and the suspension of farming on the estate. No employment was created. The appointment of foreign, non-Swahili speaking staff further complicated relations with local residents. The segregation of life on the estate compound differed starkly from the shared territoriality of the previous paternalistic mode of management.7
In this case of land grab ahead of its time, what has changed is not the organisation of production – which was substantially modified in 1997 with de facto privatisation – but simply the identity of the landlord – from a parastatal to a private company – and the rent levels, which have increased under private landlords. This change could be hardly illuminated by pre- and post analysis via a baseline survey, as recently proposed (Oya 2013), because the only thing that has changed with privatisation pertains to rent, not to capital or labour. In 2006, the management raised the rent levels and tenant farmers refused to comply with the increased rates. As a consequence, the estate stayed idle and rice farming almost stopped for two years. In 2008, negotiations among farmers and the estate were ongoing, mediated by local politicians. The company acted as a landlord in the making, in a hostile setting, by keeping control of the land and renting it out as before to local residents, without investing in production. The company extracted rent exactly as under the previous parastatal's management, but without extending the benefits of a paternalistic landlord over farmers.
The village administration reacted to the eviction threats by appealing to regional authorities for political mediation. The authorities proposed that the village surrender a few acres (the contested primary school, lying on the estate premises) as a symbolic compensation to restore relations of trust with the investor and bring an end to the land dispute. But the village assembly refused, not least because of the recurring abuses on the part of the local company management. Shortly thereafter, four of the anti-privatisation campaign leaders were arrested and charged with throwing pesticides on the estate rice fields.
The government delayed any decision on the legal status of the registration of Kapunga as a village. Yet the village had a regularly appointed civil servant (the Village Executive Officer) and a lawfully elected leader (the Village Chairman) presiding over the Village Council and Assembly. The legal status of registration is a not-so-veiled political threat to curb local political opposition in an area where opposition parties enjoy considerable support. In fact, if a village is not registered it has no right to have an autonomous village government and administration and, given the land dispute between the village and the estate, it carried overtones of potential threats of eviction against the residents.
To avoid resettlement, village authorities applied for a proper Village Land Certificate, as provided for under the Village Land Act no. 5 (URT 1999). But the government had removed from their posts, or transferred to another location, all the village politicians and administrators who had mobilised against privatisation in 2005. The district CCM electoral committee turned down the internal candidacies of the local leaders of the anti-privatisation campaign. The ward constituency reacted with a protest vote, by voting for the opposition parties. Consequently, protest leaders suffered intimidation by the company staff, which was tolerated by local authorities.
After the threats of eviction following the land dispute, political resentment built up and intimidation against members of opposition parties and leaders of the anti-privatisation campaign continued, to the point that in 2008 village elections were patrolled by riot police to prevent political unrest – a rather extreme measure in Tanzanian politics. The newly elected village secretary was chosen from the low-profile CCM candidates.
In accordance with the official party line, district and regional politicians defended the existing privatisation arrangement. While the Regional Commissioner called upon district and ward officials to ensure the protection of the investor from offences by local residents, local large farmers pledged the government to supervise the company, widely perceived as a trader and a landlord rather than as an agricultural investor. While the anti-privatisation campaign was met by authoritarian politics, in 2009 a parliamentary commission of enquiry investigated the land dispute between the village and the farm. Its report has not been disclosed and the dispute is still ongoing. From 2006 to 2010, no investment had materialised on the farm, while the inherited tenant system was maintained without resuming farming operations, but rather focusing on rice trading and land rent. The company was perceived not only as a hostile neighbour, but as a landlord in the making.
Besides the principal concern over land and water, a secondary preoccupation regarded the control of the large infrastructures. Given its control of the largest rice mill and storage facility in Chimala Ward, the company operated as a quasi-monopsony on rice, as it became the largest single buyer from small producers in the area. The large grain storehouses built by rural capitalists are much smaller than the estate's facilities and the widespread fear of monopsony by the large company was an additional reason for the political contestation of the sale.8
Class dynamics
The different scale of operation became evident both in the issue of control over marketing and in the land question – two elements of the agrarian question which spurred class conflict in Chimala Ward. With regard to land, the class contradiction was expressed by the existence of two opposing plans for land redistribution. The plan of redistribution by fragmentation into small plots was set out by small and middle-scale farmers against the plan of redistribution without dismantling, put forward by the rural capitalists’ cooperative.
An analytical point here is that a first set of class dynamics between large farmers and rural capitalists and the wider rural society determined the defeat of the contestation of the privatisation of Kapunga Rice Farm. The picture which emerges is one of a rural society highly divided along class lines, where class contradictions are attached to processes of accumulation – both from below and from above.
This appraisal calls for two additional observations. The first is that it is no longer the case that the national bourgeoisie has no role to play in production as in the past, when it fundamentally depended on – or coincided with – the state and its bureaucratic apparatus (Shivji 1972, 1975, 1976). There is accumulation from below, in the guise of substantial investment in commercial farming both from rural capitalists and from professionals and civil servants who, reinvesting part of their wage, have taken up a meaningful role in production.
Second, discussions of the agrarian question in Tanzania should not gloss over the existing trend towards class formation and land concentration. Evidence from Mbarali testifies that the existing trend towards class formation is crucially played out via land concentration in the hands of small groups, as fewer and fewer poor people can access productive and fertile land without having to pay for it. Although this may vary geographically, in highly commodified areas rent is a consolidating reality, as large farmers stand alongside increasingly large numbers of absolute landless. Larger data sets collected in other rural areas, such as Morogoro and Mtwara (Ellis and Mdoe 2003; Magongo and Da Corta 2011), have pointed out that the incidence of landlessness among the rural poor is rising, together with increasing land concentration and spiralling land prices.
It is therefore important to pose the question: is there a landlord class in the making? While it is not clear whether these features are prominent only in highly commoditised areas, there is nonetheless a generalised consolidation of the social structures of rent, attached to accumulation from below. This is going in the same direction, on a parallel line, with the speculative trend emerging in some of the larger land grabs – the accumulation from above, driven either by capital or by the state.
Throughout the last two decades, in Mbarali rural capitalists and urban professionals have been renting and buying larger, high quality farms – typically, irrigated farms close to tarmac roads. This trend of investment in commercial farming, based on the employment of wage labour, is coupled with aspirations to private property in the best farmlands in the country. The material interests of rural capitalists are closer to those of urban professionals investing in commercial farming than to their poorer fellow villagers. Class divisions are recognised and expressed as such by rural workers and small and middle-scale producers, who realise that their lower command over labour and capital is setting them apart from large farmers and rural capitalists. The latter have also realised that their interests converge with those of the heterogeneous group of people farming estate plots by reinvesting their wages in commercial agriculture. In contrast to the past (Gibbon and Neocosmos 1985), social differentiation in the countryside has been articulating along class lines. For this reason, when the issue of land redistribution was put on the table, small and middle-scale farmers allied with rural workers, while rural capitalists allied with urban professionals. The former reacted against the attempts of the latter to appropriate the best farmland in the area.
But it is important to recognise that this first class contradiction was superseded by a second class contradiction: while these two social groups confronted each other on land redistribution, they both ultimately came out defeated, as they failed to confront the state and (trans)national capital. At the national level, political alliances inside the dominant party took control of the privatisation without including any of the local constituencies.
In Mbarali, the failure of an attempted inter-class alliance on the issue of land redistribution underlines an embryonic class consciousness, as material collective interests become clear cut in the politics of land. This case of land grab ahead of its time shows that localised struggles against land grabs have progressive potential but need to confront the arena of national politics if progressive outcomes are to be secured.
What the poorest groups in rural society are asking for are opportunities to emancipate themselves from the dominant classes, in order to gain more control of the fundamental conditions of production, on which they have limited power.
There are three contributions to the land grab debate. First, I have provided evidence of the fact that, in some localities, there is no such a thing as ‘the local people’, or ‘communities’ in the struggle against land grabs, but rather a composite scenario of shifting class alliances, whose dynamics need to confront the state and its own class contradictions. Second, land dispossession and class struggles in the African countryside have a much longer history than that sometimes portrayed by the land grab debate. As a third point, the historical trajectory of this case underlines that governments are acting as intermediaries in the dialectics between ‘local and global processes’ (Araghi 1995) and that scholarly research must not only straddle the two to ‘escape the reductionist trap’ (McMichael and Buttel 1990), but also uncover the transformative potential of localised struggles and class alliances in the global phenomenon of land grabbing. To explore this third point further, the concluding section will show how the situation on the ground is constantly, rapidly changing, as reorganisations occur rapidly in the world of financialised agribusiness.
Agribusiness and finance in Mbarali
The world biofuel rush in 2007/8 offered opportunities for capital accumulation to the company that acquired Kapunga Rice Farm. In those years, the Export Trading Group (ETG)9 scaled up from a regional trade house to a global agribusiness company, and also shifted from trade in agricultural produce to wider farmland acquisition and biofuel plans.
Until that time, the company was a large family-run Tanzanian Asian trade house specialising in trade of staple foods and fertilisers. It had made its fortunes by sourcing emergency food relief stocks, as the local contractor for the World Food Programme (WFP) in Tanzania, and had built up large storage facilities and a private transportation fleet (ETG undated).
In 2007/8, ETG tried to convert Kapunga Rice Farm from rice farming to jatropha, through a joint venture with a South African biofuel company (Vermak 2010).10 Forced to abandon this conversion plan, in 2009 ETG received a syndicated loan of US$120 million from the International Finance Corporation (IFC) of the World Bank, to finance the strengthening of trade infrastructures for primary agricultural commodities in 11 countries in Africa, India and the United Arab Emirates (IFC 2009, 2010; Africa-Asia Confidential 2010). The IFC, one of ‘the global market makers’ and great predators in development finance (Bracking 2009), justified the loan as ensuring a highly developmental impact on smallholders through enhanced marketing infrastructures.
At the same time, in the aftermaths of the 2007/8 global financial crisis, it became apparent that private equity funds were targeting African agriculture (Daniel 2012). In the region, South African agribusiness groups started to increasingly rely on internal finance, operating on financial open markets, rather than on banks for credit, for example by establishing private equities. According to Lapavitsas (2011), this is one of the features of financialisation in advanced capitalist economies. In 2012, ETG received the first loan ever granted to an African company by the Carlyle Group, one of the largest private equity firms in the world (Reuters 2012). Remgro, a giant in South African agribusiness, functioned as the regional mediator of transnational flows of financial capital by establishing a private equity fund, the Pembani Remgro Infrastructure Fund, which, together with Standard Chartered Bank, entered into a partnership with the Carlyle Group to buy a stake of ETG worth US$210 million. The Tanzanian and simultaneously transnational nature of this company, its political connections with Tanzanian centres of power, and its multinational sources of finance make us wary both of the limits of analysing the agrarian question from a purely national perspective only – and of seeing the global land grab from just a transnational perspective. The former misses out the world historical dimension of capital; the latter, its ever-recurring necessity of using the state as intermediary and the regional dynamics of its unfolding – in this case, the trend of South Africanisation (Hall 2011) of Eastern African agrarian structures.
While the case of Mbarali shows that, for African agribusiness capital, financialisation meant a more direct and open involvement with private equity funds, the key point here is that the financialised and transnational nature of the company is not immediately evident to the rural producers in Chimala Ward. At the local level, two conspicuous features are its political connections and its Asian origins. The lack of investment in farming and the segregated life conducted by the company staff revived some entrenched stereotypes and social perceptions about Tanzanians of Asian descent – seen as closed communities with little social interaction outside their circles, compounded by a limited command of the Swahili language (Bujra 1992) and as trade-centred business community without expertise or interest in agricultural activities. Remarkably, notwithstanding this heavy, unresolved colonial legacy (Heilman 1998), collective action in Mbarali did not become xenophobic nor racialised – perhaps a sign of the deeply rooted political legacy of ujamaa, with its strong rejection of the political mobilisation of ethnicity and race.
If the overall progressive potential of this political mobilisation has not been realised, this is not because of lacking or weak political mobilisation from below, but rather because the refusal by small and middle-scale farmers to build a cross-class alliance unwittingly facilitated state intervention in favour of accumulation from above. Tanzanian rural capitalists and professionals investing in commercial farming operate on a much smaller scale than that of merchant and financial capitals involved in agribusiness in the region. Their defeat – with accumulation from above prevailing over accumulation from below – was effectively sanctioned through the government's reluctance to redistribute a large farm in a locality where not only rural capitalists, but also medium and small producers, together with the landless, claimed it.
Conclusions
To conclude, stating that the agrarian question is world-historical has at least two implications. The first is that any analysis at the national level interplays with longer and deeper historical processes of land dispossession. The case of Mbarali showed how accumulation from below has been halted by primitive accumulation by merchant capital, supported by finance capital and by extra-economic force. These were key to pre-empt claims of repossession of a state farm which was built on public land through public debt.
A cognate reflection pertains to the role of African governments in post-grab situations, as whereas investors abandon the investment site, state agencies often retain control over land and oppose land restitution to the original occupiers. This trend must be closely scrutinised, as such a scenario can widen the scope for rent on the part of state agencies, themselves expressions of class interests. This moment of primitive accumulation is fundamental to class formation – in this case, the consolidation of an agribusiness conglomerate in East Africa, through South African mediation of transnational financial capital.
The second is that a strategic focus on the power of the global forces at play must be coupled with a tactical attention to the politics and the agency of resisting groups, where they exist, by avoiding blank accounts of ‘local’ resistance. These risk missing a key tactical point, namely the potential of political responses to land grabs. Our analysis of the class dynamics in Mbarali has shown the limits of attributing homogeneous interests to farmers living in the same area – ‘the local’ – by pointing to the class interests which emerge when property is reorganised.
The global land grab debate has reminded scholars of the agrarian question that national dynamics are always to be understood in their broader world-historical dimension. Cases like the one presented above invite greater attention to the class dynamics in the land grab debate. If it is to make a difference in supporting the existing resistance against dispossession, then engaged scholarship must confront the class dynamics in the countryside, in order to uncover the political potential for more effective organisation and support the progressive elements contained in these shifting, fragile cross-class alliances.