Dan Brockington is leading an ESRC/DFID project exploring long-term livelihood change in Tanzania to revisit villages  first surveyed in the 1990s. We try and talk to the same families that were visited in the original surveys, conducting re-surveys and then sharing results and findings with community members. One of the features of this work is that we also visit these sites with the researchers who conducted the first work. Here we have re-blogged a series of entries and reflections about these revisits in which researchers consider what changes they have seen and what they might mean.

The fourth blog in the series is from Stefano Ponte, Professor in the Department of Business and Politics at Copenhagen Business School. In April 2016, he joined Christine Noe, Olivia Howland, and Dan Brockington for a pilot visit to two of the three original villages in Morogoro region: Mlali and Langali.

Twenty years ago, I lived in Tanzania for 18 months to carry out fieldwork for my PhD on changing rural livelihoods under market liberalization. This included work in six villages – three in Morogoro region and three in Ruvuma region. It was a time of rapid change, when agricultural cooperatives were facing new competition from private traders in purchasing export crops from farmers, while food crops had already been liberalized for a number of years. Cooperatives were weakening, and with that the provision of services to farmers, including fertilizers and agro-chemicals on credit, was also waning. New small shops were popping up in rural areas, and depending on whom you talked to, there was a sense of new opportunities arising, but also of loss of public service and solidarity.


I have been in Tanzania several times in between, including a six-month stint researching coffee markets in 2000. I was also part of a collaborative research project on standards in agro-food export with Sokoine University in 2005-2010, but I had not had the chance to go back to these particular villages since 1996, apart from a brief visit to one of the Morogoro villages (Mlali) in the early 2000s. The ‘Long term livelihood change in Tanzania’ project, coordinated by Dan Brockington at the University of Sheffield, spurred my interest in doing so, along with other researchers who worked in the country in the 1990s. In April 2016, I joined Christine Noe, Olivia Howland, and Dan Brockington for a pilot visit to two of the three original villages in Morogoro region: Mlali and Langali (now in Mvomero district). This involved, in each village, interviews with key informants, discussions with small groups of elders, two focus groups with elders (women and men), and in-depth interviews with a sample of the original household I had surveyed in 1996. To my surprise, we were able to identify and locate 90% of the original households in one village and 60% in another. I was pleased and relieved that so many of the original people I talked to were still alive, and living in the village, but I was also anxious about what I was going to find.



Mlali is located about 45 Km from Morogoro town, on the plains at the foot of the Uluguru mountains. It is easily reachable via tarmac road until Mzumbe University, and then a well-graded dirt road that we found easy to transit even at the height of the rainy season. The village is served by many minibuses trawling back and forth from Morogoro town. The center of the village has boomed. There is a new market area and at least 50-60 shops of all kinds, including two dedicated shops for agricultural inputs and various food joints. In 1996, there were between 5 and 10 shops – and much smaller ones. The village is still chaotic and dirty, with trash and plastic bottles strewn everywhere, just like I remembered it, but now at a different order of magnitude. One of the most obvious changes is the ubiquitous presence of young men on their pikipiki (motorbikes), a swarm going back and forth along the main road and main paths, or hanging around and bantering, waiting for their next client. Many of the pikipikis in the village are owned by well-off farmers. They are operated by young men, who pay a daily rate to use the pikipiki until they can buy the asset from the owner (a form of lease). Twenty years ago, only the very rich, the resident doctor at the local clinic, and a few local government officials had a motorbike – totaling less than five in Mlali. There are now several private cars, parked in front of well-built houses with porches and electricity, and even at least one private truck. There were no private cars in the village in 1996.

Mlali was not a place where collective action was particularly strong even in 1996 – by then the local primary cooperative society had already gone belly up, and the village farm had already been divided up and distributed to residents. The local economy at that time was driven by tomato cultivation, which had been introduced already in the early 1980s. Twenty years later, tomato cultivation is an even stronger driver of rural livelihoods in Mlali, at least for those who have enough resources to buy farm inputs, own land or can rent it, and can draw family labor, hire farm workers or use tractor services for this intensively cultivated crop. This also includes labor to carry water for irrigating tomato plants (a few wealthy farmers have installed drip irrigation). While paddy and maize are still farmed in the village, sorghum (the traditional food crop in this area) is less so. Sorghum grows more slowly than maize, and is not ready to harvest early enough to free up plots for the main tomato cultivation season. Therefore, farmers prefer to plant maize and then tomatoes after the maize harvest. There is a vibrant land rental market in Mlali, with plots being rented twice a year, depending on the planting season and with prices linked to the type of crop planted (premium prices are charged for the main tomato season).

A ranking exercise we carried out in focus groups allowed participants to define the characteristics of various layers of ‘wealth’ as they saw fit. In Mlali, rather than the amount of owned land, a key indicator of wealth is the amount of land put under tomato cultivation – whether the land is owned or rented does not matter. The key issue is thus access to capital or credit to rent land, buy inputs and employ workers or use tractor services for tomato cultivation. The current use of tractors for land preparation really surprised me, as there was no sign of this service twenty years earlier.

Households in the top two wealth layers, as defined in focus groups, seem to be able to use tractor services for all their cultivated land, with at least five tractors available in the radius of a few kilometers. But even those in the middle wealth ranks, who prepare land by hand, seem to fare relatively well (as indicated by the quality of housing), as long as they have either accumulated funds, can sell an asset or borrow from relatives or from their social network to buy inputs and on occasion hire labor. Without fertilizer and agrochemicals, tomato yields are much lower, and those who cannot afford them tend to rent out their plots to others. These tend to inhabit the lower wealth ranks, followed by those who, for health or other reasons, cannot take care of themselves.

These are preliminary findings from qualitative research that may have been skewed towards ‘winners’ in the village, as key informants and focus group participants are usually unlikely to include the worse off. Yet, at least from the focus group discussions and ranking exercises, there seems to have been a general movement from a pyramidal wealth structure in 1996 to more of a pointed egg structure. The middle wealth ranks, according to focus group participants, include a sizeable majority of households. We do not yet have data to assess inequality between top and bottom categories, but this is likely to have increased in the past two decades. At the same time, the size of the ‘middle ranks’ (in local terms) seems to have increased substantially. The meaning of different wealth rankings has also changed. While twenty years ago, for example, only the relatively well-off had metal sheets as roofing material — while the poor had thatched grass roofs — virtually all houses have metal sheets now. Mud walls were quite common then, while now baked bricks are the norm. Cement floors, plastered walls and indoor sanitation were only for the very rich then; they now have spread to the upper-middle ranks as well.

Pilot interviews with the original households that I had interviewed in 1996 confirmed these preliminary observations. We asked some of the same questions on farming, housing characteristics, household composition as I had done twenty years earlier. But we also followed up with asking basic information on the children who left the original household to form their own. The original households in 1996 had been selected at random from the village list, thus provide a less potentially biased picture. Yet, given the small number of re-interviews carried out so far, we are still talking about very preliminary results and reflections.

What emerged from the first few household interviews in Mlali suggests two trajectories of livelihood change: (1) a trajectory of accumulation based on expansion and intensification of tomato cultivation, translating into much improved housing quality, more assets, availability of funds to solve health problems, and some of the children having gone through secondary education (or even tertiary) and formal employment in urban areas; and (2) a trajectory of relative stagnation in tomato cultivation (less land, lower use of inputs) that in any case sufficed to invest in some of the children’s education, leading to at least one landing a government job or setting up business, and thus to remittances allowing for slightly better living conditions than in the past. In future research, to be carried out in the rest of 2016, we expect to find out more about the trajectories of decline, which were not clearly captured this time.



We carried out exactly the same research activities in the second village. Langali is placed on the Uluguru mountains, in an area known as Mgeta – a 90-minute drive uphill from Mlali. Given that it is the main rainy season, I had feared we would struggle to reach the village. I remember taking at least three hours to reach it in 1996 and with significant struggle. However, the road is in much better conditions than it used to be. Stretches have been upgraded to cement by a USAid-funded project that also brought electricity to Mgeta, a stunning set of efforts given the rough mountain terrain.

My first impression of Langali was quite similar that of Mlali: many new shops, a newly-refurbished market area, a cement road in the village center, lots of pikipikis, some private cars, and – at least in the village center – a substantial number of houses and businesses hooked up to electricity. Yet, part of the economic hive of activity that was typical of this village even in 1996 seemed to have been drawn away to the neighboring village of Nyandira, situated five kilometers uphill from Langali, where a modern market place for fruit and vegetables has been recently built with French aid funding.

The steep hills surrounding the village make farming quite a demanding activity that includes terracing and reclaiming land between boulders and on steep inclines. Furrow irrigation is used in most plots for vegetable cultivation, with water diverged from main streams. Many kinds of vegetables and fruit have been grown in Mgeta as cash crops since the 1950s, supplying surrounding urban areas and Dar es Salaam. Food crops are cultivated in plots further away from homesteads, and where no irrigation is necessary. The beauty of the area, cool climate and the complexity of farming systems have attracted many researchers to this area throughout the past 30-40 years. As a result, there has been a long-standing set of academic debates on the impact of demographic pressure on land use, farming systems and out-migration in Mgeta, including in Langali.

Many of the contributions that appeared in the 1990s portrayed a pessimistic picture of the Uluguru mountains, either because they saw a process of agricultural ‘involution’,[1] or because of a vicious cycle of subordination to market forces that undermined the so-called ‘reproduction of the peasantry’.[2] In both views, population pressure, land scarcity, and the crisis of agriculture loomed dauntingly over the future of rural livelihoods in Mgeta. When I published research from my 1996 fieldwork, I argued that agriculture was indeed not going through an easy transition in the area—especially for poorer farmers.[3] But while some livelihood options were becoming more limited, others were being more skillfully utilized. On the Uluguru mountains, I argued, land scarcity (or at least scarcity of suitable land close to the villages) was the main feature of agriculture; deforestation and soil erosion were major problems; and inputs had become increasingly expensive.  Under these circumstances, the main ways households could improve their quality of life—short of leaving the area altogether and in addition to relying on remittances from outside—were to expand land cultivated in other locations, to experiment with alternative farming systems, and to increase non-farm incomes. The inhabitants of Langali village were doing all the above, with a certain measure of success.

Fast forward to 2016, some of these observations are still relevant. However, they need to be further qualified. While cabbage was the top cash crop then, it has now been replaced by Irish potatoes. A variety of other vegetables (from carrots to cucumber) were cultivated then, and this is still the case now – but with tomatoes becoming the third main cash crop (after cabbage), while twenty years ago they were only a minor crop. Due to different altitude and climatic conditions, tomatoes ripen later in Mgeta than in Mlali, so they do not compete in the same seasonal market, making them profitable even in the smaller plots and harder terrain of the Uluguru mountains. A root disease is now affecting cabbage cultivation, and fertilizer needs to be applied to any crop (even maize) to obtain decent yields – given the lack of fallow and generally declining land fertility due to long-term intensive cultivation. In other words, reliance on farm inputs, although not new, has increased. Relatives who have moved down to the plains or in urban areas are key sources of cash or loans to purchase inputs. The other source is small livestock (pigs, goats, chicken, farmed fish) that can be sold according to need. If all else fails, households will rent out their plots to those who have funds to prepare land and buy inputs, and work as farm laborers. Land preparation, given the terrain, is exclusively done by hand.

Elders and focus groups participants did not report a massive outmigration flow, as portrayed in the 1990s literature, although they highlighted the lack of inward migration. They also reported, surprisingly, that most households had access to land through inheritance and use-rights from their clan, although plots have become increasingly fragmented in time. Also surprising was their claim that land per se was actually available, but mostly in locations far away from the village and where transport costs may not justify the effort after all. This was confirmed by direct observation of the landscape, where carefully-tended terraces existed next to plots that had not been cultivated in years – especially in areas further away from the village center. Focus group participants and interviews with elders suggested that the apparent availability of land is explained as follows: (1) plots may lack enough natural fertility and need to be left fallow for a few years (this also makes it unattractive for potential short term rental); (2) they may be too far away to be cultivated; (3) the household may not have the resources to prepare and tend for the land, including farm input purchases; and (4) even in this case, they may not want to rent it out for fear of not being paid at the end of the season.

Wealth rankings emerging from the focus groups indicated a pointed egg-shape distribution, but with a larger base than in Mlali. Similar to Mlali, wealth rankings followed an asset’s functional utility rather than its sedentary value. In other words, there is a difference between owning a productive asset and making it work to generate income. Land that cannot be used for vegetable production because of lack of means is not considered an indicator of wealth. Land that is productive, even if rented, counts as ‘wealth’. Although there are of course differences in housing quality between different wealth ranks, they are quite subtle in the middle ranks, highlighting again a process-oriented approach to wealth assessment – with important indicators being where households send their children to school, and whether they have access to funds for medical expenses immediately, through loans, through sale of livestock, or whether they have no ability to mobilize funds. According to focus group participants, the top two and middle two wealth ranks are populated by households who, one way or another, can cultivate vegetable plots with different degrees of input intensity. Those who work as farm workers in other people’s land (who may actually own land but can not afford inputs), the sick and otherwise unable to work belong to the bottom three wealth ranks.


Pilot interviews with the original households that I had interviewed in 1996 suggest a complex picture of livelihood trajectories, and so far one that suggests a more sizeable bottom wealth layer than in the focus groups. While there are examples of upper mobility that emerged in the past twenty years, they are related to inheritance and a mixture of good management and good luck (lack of disease outbreaks in vegetable cultivation and favorable prices), at least in the first few years of cultivation after inheritance. These are accompanied by more numerous instances of relative stagnation or decline, which are related to disease outbreaks, lack of resources to buy inputs, old age and limited support from migrated family members. Housing conditions, even in these instances, have improved, and some of the children who have left and migrated elsewhere have found permanent employment or opened a business, thus sending remittances back to Langali.

Summing up the preliminary findings

In both villages, a clear process of further commercialization has taken place in the past twenty years, although this had already started during my fieldwork back in 1996. Many more non-farm activities, businesses and transport services are evident in the two locations. And infrastructure (roads, marketplaces and electricity provision) has improved markedly.

Four groups of general trajectories of long-term livelihood change emerged from preliminary fieldwork:

  1. trajectories of accumulation based on expansion and intensification of tomato/vegetable cultivation – translating into much improved housing quality, more assets (including means of transport), diversification into other economic activities, ready availability of funds to solve health problems, and some of the children having gone through secondary education (or even tertiary) and formal employment in urban areas;

  2. trajectories of marginal improvement in tomato/vegetable cultivation (less land, lower use of inputs) that sufficed to: improve living conditions and types of assets; sometimes diversify in small business; access health services when needed (even if it entailed the sale of livestock or other small assets); and invest in some of the children’s education (secondary school). In some cases at least one child landed a government job or set up a profitable business, sending remittances that further improved access to working capital for farming;

  3. trajectories of stagnation and decline, where households could not get access to enough capital to purchase inputs, and thus limit their farming activities to food for partial self-sufficiency; members of these households tend to work in other people’s fields as daily laborers; while their housing conditions may not have become worse, these households tend to send their children to primary school only, and to have little or no diversification into non-farm activities; they also tend to struggle to find resources to take care of health problems;

  4. trajectories of pauperism, where people are not able or willing to take care of themselves, are not able to work for others, and rely on their family’s help or other people’s handouts.

With all the provisos that preliminary research and analysis entail, what seems to be emerging is a process of rural transformation. In Mlali, this is more clearly driven by agricultural accumulation. While the mainstay cash crop has remained the same, the intensity of cultivation has increased, as have input application, the use of improved planting material and access to tractor services. While this is not an agricultural revolution per se (and it is based on a high-chemical input model of modernization with all the drawbacks it entails), it seems to have provided enough resources for further investment in retail activities and transport services, at least for a substantial proportion of the population.

In Langali, rural transformation is less clearly rooted in agricultural accumulation, although this is happening for a minority of households. Farming systems are fairly similar to those observed twenty years go, but with a different mix of top cash crops. Input use is increasing and now necessary to obtain sufficient yields even for food crops, thus increasing risk and vulnerability in view of recent crop disease outbreaks. On the Uluguru mountains, rural transformation seems to stemming from a more complex combination of economic diversification, remittances and marginal productivity gains for farmers who have enough access to capital to buy inputs.

These processes are of course not universal, even within villages, and we do not yet know enough about the lower wealth groups in both locations. However, for the time being, it seems fairly reasonable to suggest that wealth segmentation seems to have switched from a pyramid structure to a ‘pointed egg’, with Langali having a larger base than Mlali.


The research team will be carrying out more interviews and wealth ranking exercises in the next few months, thus we will expect to know more about the trajectories of long-term livelihood change. Some of these trajectories are of course determined by the specific demographic profile of the households at the time of the original fieldwork in 1996. We also expect to be able to assess the relative importance of each wealth segment and thus the precise shape of its distribution. We will also seek to deepen our understanding of the livelihoods of those who left the original households, and find information about the households who have either completely moved away or are no longer existing. Finally, we will compile an updated profile of access to education and health services, which we have not examined in detail so far.

Stay tuned for more!

[1] See, e.g., van Donge, J.K. 1992a. ‘Agricultural decline in Tanzania: the case of the Uluguru Mountains’, African Affairs, 91: 73-94.

[2] See, e.g., Hadjivayannis, G.G. 1993. ‘Stratégies paysannes face à la crise: adoption des innovations, diversification des activités et differenciations sociales dans un village tanzanien:  Mgeta, Morogoro’. Université de Paris I, thèse de doctorat.

[3] See Ponte, S. 2001. ‘Trapped in Decline? Reassessing Agrarian Change and Economic Diversification on the Uluguru Mountains, Tanzania,’ Journal of Modern African Studies, 39(1): 81-100; and Ponte, S. 2002. Farmers and Markets in Tanzania: How Market Reforms Affect Rural Livelihoods in Africa, James Currey, Heinemann and Mkuki na Nyota: Oxford, Portsmouth NH, and Dar es Salaam.

No Comments have been posted yet.

Post a comment