Alternatives to Slash and Burn in Madagascar - Why are we Still Looking for Them?

Alternatives to slash-and-burn in Madagascar: why are we still looking for them?


The search for alternatives to slash-and-burn cultivation is a hot topic in agronomic research, as shown by decades of work conducted by the CGIAR organizations and their partners since 1991, with heavy funding provided by the GEF. This state of affair is quite a surprise as alternatives to slash-and-burn have been developed by farmers all over the world since at least 2000 years. This article studies this paradoxical situation in the case of Madagascar. It will show that Malagasy farmers, as their counterparts in the rest of the world, have developed technical alternatives to slash-and-burn, and continue to develop and adapt them in areas where slash-and-burn cultivation is still practiced. It will explain why rural development projects do not pay attention to these techniques and struggle to develop their own alternatives, which farmers do not adopt. Three lines of arguments will be drawn. First, development projects overlook the specificities of peasant economics. I will refer to the work of Chayanov (1922), an agronomist who “discovered” these specifities in early 20th century Russia. Second, the endogenous dynamics of landscape transformation are ignored, whereas the general pattern of these transformations is well known since the work of Boserup (1964). Third, the ethics of peasant societies is also misunderstood and misregarded, whereas, here again, it has been described by a landmark work (Scott 1976). These three books are classics in the fields of peasant and agrarian studies but are ignored or rejected by most development institutions. If they were brought back within the mindset of development actors, they could constitute the base of a substantive agronomy that would help development actors to free themselves from the knowledge biases induced by their culture and by political agendas. Combined together, political and substantive agronomy could constitute a pragmatic and reflexive science where empiricism and hermeneutics would be productively articulated, and from which more efficient development approaches could emerge.

Alternatives to slash-and-burn in Madagascar: why are we still looking for them?


The search for alternatives to slash-and-burn is a prominent topic among organizations supporting agricultural development and environmental conservation in low income countries. The Global Environment Facility, launched in 1991 to finance the implementation of agenda 21 of the Rio Earth Summit, supported an ambitious international collaborative effort, the Alternative to Slash-and-Burn Program, coordinated by ICRAF and still ongoing. This program reflects the commitment of the international community to support a transition to more sustainable land uses on forest frontiers. It is characterized by the belief that achieving sustainability is mostly a matter of technological innovation (Pollini 2009), although emphasis shifted toward landscape planning during the last years. At the core of the ASB’s efforts are agroforestry technologies whose development is the core mission of ICRAF since its creation in 1978. These technologies, which include improved fallow, alley cropping, and hedgerows planted on contour lines, have been extensively “adopted” by NGOs and projects supporting smallholder farmers all over the world. This is the case also in Madagascar, where agroforestry technologies have long been envisioned as the most promising alternatives to slash-and-burn cultivation (Pollini 2007, 2009). This heat for agroforestry, however, is on the decline since about a decade (Ref XX). New alternatives like cover crops and no till farming are now gaining more fame and are promoted, together with agroforestry, by international organization, under the label “conservation agriculture” (Serpantie 2009).

This dominant paradigm of agriculture sustainability and intensification, which can be called “technological optimism,” to use the words of Norgaard (1994), constitutes an “apolitical agronomy.” It is now increasingly challenged, both by agronomists who investigate more deeply the rationale of smallholder farmers (Serpantie 20XX) and by scholars who study the political economy of conservation and development programs (XX ref XX). This article is a contribution to this critique. It is a political agronomy analysis of the alternative to slash and burn paradigm in the case of Madagascar. It provides a case study to the more general analysis drawn in Pollini (2009).

I will start by showing that alternatives to slash-and-burn already exist and have been adopted by farmers all over the world since millennia, Madagascar not making the exception. I will then review current efforts to develop alternatives to slash-and-burn in Beforona, a rural municipality located on the eastern edge of the Malagasy rain forest. Beforona received support from French, Swiss and US aid agencies since several decades and is often displayed as a model of successful rural development intervention. I will show that although farmers in Beforona are developing their own alternatives, they receive very little support to do so, as projects prefer to develop and propose their own alternatives, which I will call “alternatives to the alternatives”. In the discussion section, I will review the reasons of this gap and will make a few suggestions to remedy this situation.


Data collection occurred between 2001 and 2007. In 2001, I spent 6 months in Ambodilaingo, a village located 16 kilometers south from the national road and belonging to the municipality of Beforona. From 2001 to 2007, I made frequent visits in Ambodilaingo and other villages of the municipality, to collect more data and to monitor a rural development project implemented by a small Malagasy NGO (Zanaky ny Ala ).

Data collection was made through informal discussions while touring the area, or during meals and other social gatherings. Interviews were not registered but were conducted with a translator, which let ample time to take notes. Transcripts may not be totally accurate as the translation may have introduced a bias, but they usually reflect ideas and statements that are simple enough to be trusted. Only the information that was verified by several informants or by direct observation are presented in this article.

The questions asked were very open and did not include any suggestion of answer, in order to let the informant speak freely. Every informant was given the opportunity to talk about the subject that he is most familiar with or most interested in. The same question was rarely asked twice, but the same information could be formulated several times in various speeches. No pre-designed questionnaire was used. These methodological choices are aimed at minimizing, as far as this is possible, the biases introduced by the culture that the interviewer carries with him, and by the social relation between the interviewer and the interviewee.

The development of alternatives to slash-and-burn cultivation: a global overview

The transition of slash-and-burn farming to more intensive land uses has been widely documented, and I will rely on a few comprehensive syntheses here. Boserup (1965) discovered the general pattern of this transition. She describes an “historical sequence” that goes from long fallow systems to short fallow, annual cropping, and multi-cropping systems. The key variable that discriminate between these systems is the frequency of cropping, which goes from once every 20-25 years in long fallow systems, to several times per year in multi-cropping systems, where crops are associated and where several cultivation cycles occur in a single year. The key driver of change would be population density, simply because long fallow becomes materially impossible when the population density is high, which forces to develop or adopt new systems.

The types of land uses developed and/or adopted by farmers when this transition occurs have been reviewed extensively by Mazoyer and Roudart (20XX), Dufumier (20XX), Bahuchet (1993), XX search for ref in English XX. They can be classified into (1) systems based on crop rotation on plowed land with short herbaceous fallow or no fallow, in general associated with animal husbandry and using manure as a fertilizer; (2) hydro-agricultural systems based on flooding or irrigation, also often associated with animal husbandry; (3) systems more or less specialized into the cultivation of perennial cash crops that can be cultivated on degraded land; (4) pastoral systems; (5) combinations of these four types. The labor productivity of these systems is usually very low in their early stage, until a larger productive capital is acquired (more efficient tools like plows and carts; larger herds producing more manure; irrigation infrastructure enabling a better control of the water) and new institutions are created (in order to manage conflicts that could be provoked by these technological changes).

A controversy exists about whether technological innovation is the cause or the consequence of population growth . It is thus worth to look more closely at Boserup’s insights on this issue. For Boserup (1965), unless there is population pressure (lack of land), there is no reason to shift to more intensive systems, i.e., to systems with higher frequency of cropping. This is because such systems produce more by unit of land but less by unit of labor invested, at least during the early transition, before farmers increase their productive capital by acquiring more efficient tools (plows, carts). Changing to a new system also requires making investments, in order to acquire these tools, create irrigations schemes, build stables where manure can be collected; and taking risks, because of the uncertainties generated by any new undertaking. When population increases, farmers have no choice but making these investments and taking these risks which they may have delayed heretofore. It is in this sense that we can say that intensification is induced by population growth. Boserup (1965: 22), however, recognizes that innovations are not developed or adopted all the time, and that “some regions which previously supported a more or less dense population are barren today.” The outcome of population growth can thus be environmental degradation and society collapse as well as intensification. In this sense, we can say that technology is a prerequisite to population growth. But it should not be concluded from this statement that innovation automatically leads to intensification either. Boserup (1965) refers to the work of Bloch (1931: Les caractères originaux de l’histoire rurale française, XX) who showed that many techniques adopted during agricultural revolutions in Europe existed before these revolutions occurred. An excessive focus on innovation could thus lead to promoting intensification in conditions where it is not rationale (where population pressure is not felt) and to overlooking the cause of non adoption of techniques that are already available (lack of capacity to make investments and take risks). Saying that innovation is a prerequisite to intensification should not be assimilated either to saying that supporting innovation is a prerequisite to intensification. The lack of endogenous innovation can have social causes (lack of social capital, vulnerability which results in the reluctance to change farming practices) that will persist and that will constrain adoption if innovation is externalized, while solving these social causes instead of externalizing innovation could unleash endogenous innovation.

In Madagascar, Slash-and-burn cultivation persists only on forest frontiers and in areas with low population density dominated by secondary vegetation (bush fallows). It used to be practiced in the central highlands (Raison 1984) although irrigated rice cultivation was already common practice in the 17th century. The central highlands are now among the most populated and developed rural areas of Madagascar. Almost no primary forest remains and the dominant land use is a combination of paddy fields in the bottom lands, crop rotation on slopes, animal husbandry using stables and carts to collect and transport forage and manure, and plantation forestry using exotic fast growing species (mostly eucalyptus). As almost no information is available regarding the transition from slash-and-burn cultivation to more intensive systems in the central highlands, we will directly shift to studying Beforona, a forest frontier municipality where this transition is currently ongoing. A particular attention will be given to Ambodilaingo, where most enquiries have been conducted.

Beforona and Ambodilaingo

The rural municipality of Beforona is located 160 kilometers East from Antananarivo, in the Moramanga District, Mangoro-Alaotra Region, Toamasina Province (Map 1). Ambodilaingo is located 16 kilometers walking south from the National Road 2, which passes through the main center of the municipality, where the city hall (XX mairie XX), the health center, the junior high school, several stores and a weekly market are found.
The area experiences a tropical wet climate, with an average annual rainfall of 2757 mm at the Marolafo research station (Nambena 2004). Rains occur all the year but take the form of drizzle from June to November. Cyclones can strike the area from February to May, with winds of more than 100 kilometers per hour that destroy crops and provoke flooding and landslides. The hydrographic network is dense, with water running all year round in all rivers and most streams. The average temperature ranges from about 16C° in July to 26C° in January.

The topography shows a regular pattern of small valleys dominated by small hills with steep slopes (Map XX). The altitude ranges from 500 to 700 meters, except for a north-south escarpment that culminates at 1300 meters, on the western edge. This mountain chain is the only part of the landscape that is still forested (Vohidrazana forest). Many families live on its verge to find more fertile soils to cultivate but clearing almost stopped in 2002, as a consequence of a fire and slash-and-burn cultivation repression campaign implemented by the government and its international partners (Pollini 2007).

The soils are ferralitic with strong desaturation, except on forest land and recently cleared land where they still have an elevated organic matter content, and in the bottom lands and lower parts of the hills where they benefit from the transfer of particles from the upper part of the slopes. The main constraints for cultivation are a low phosphorus availability and aluminum toxicity. Sheet and gully erosion is not a major threat in slash-and-burn system and the most dramatic cause of erosion in the area is landslides, which occur during cyclones (Brand and Rakotovao 1997; Brand et al. 1997).

Farmers live scattered in the whole territory in order to remain close to their fields. This is particularly true in Ambodilaingo, where about ten families permanently stay in the village for a population of about 130 households (in 2003). Annual population growth is 2.7%, which is close to the national average. The population is young and mostly composed of Betsimisaraka, although many of them would be of Bezanozano origin, and some Antesaka, Betsileo, Merina and Chinese migrants settled close to the road. The ancestors from the people in Ambodilaingo seem to be Bezanozano people who lived in the Mangoro area and moved eastward to Ambodilaingo about XX years ago (Minohery 20XX). Bezanozano people have a way of life and culture that presents some similarities to that of the Merina people, with whom they are related, including experience in the practice of irrigated rice cultivation. But when they moved to the Beforona area, they adopted the Betsimisiraka way of life and became Betsimisaraka. This finding is consistent with anthropological work that shows the non fixity of ethnic identity in Madagascar (Bloch XX, Astuti XX Kottakk XX). It has implication on the debate about the causes of adoption or non adoption of new technologies as it shows that the biophysical environment, rather than culture, determines the type of land uses that are adopted, which in turns has consequence on culture.

The economy in Beforona has always been based on agriculture. Tavy, a local form of Slash-and-burn cultivation, is the main land use, although more intensive systems are also practiced, as we will see. Logging produces a significant complementary income to families living close to the forest, including most people in Ambodilaingo, who hire their labor force to transport boards into the road, where they are collected. This activity, however, is on the decline since about 2006. Mining (corundum in Ambodilaingo, gold in a few other villages) also plays a role in the economy, but a marginal one.

Slash-and-burn cultivation (tavy) in ambodilaingo

An abundant literature exists about slash-and-burn cultivation, globally but also concerning the municipality of Beforona, whose agricultural systems have been addressed in tens of PhD and master theses (Ref XX). I will simply report here a description by a farmer to whom I asked to describe the system, from the first clearing of a primary forest to the following cultivation cycles:

When we clear the primary forest, we first grow rice, and sweet potato on part of the plot when the rice is 30 cm high. Clearing the forest is the basis of our way of living. We cultivate the land a second time after 3-4 years, with rice and sweet potato, in the same way, with cassava in addition. We do not wait more because we lack area to cultivate, because of population increase. After that, we wait six years before cultivating again, because the soil is not so fertile. The fallow period after this third cultivation cycle will have to be still longer….

The harvest is 20 vat [400 kilograms of white rice] for one vat of seedlings [20 kilograms] at the first cultivation cycle, if the harvest is good, and five vat [100 kilograms] if it is bad. It is often bad because of rain excess. For the second cultivation cycle, we harvest 10 vat [200 kilograms] from five kopy [10 kilograms] if the harvest is good, or five vat if it is bad. We have to use more seedlings: five kopy instead of four for the same area, because the soil is less fertile. The worst is if the rain falls in November because the seeds do not germinate. For the third cultivation cycle, we have to sow seven to eight kopy where we sowed four or five, because there are no trunks on the soil and because the soil is less fertile. We get eight to 15 vat….

[…] We associate beans, maize and tsidimy with the rice. The rice is sowed first. Maize grows better in the first cycle. We do not stop cultivating until the land is invaded by Tenina [Imperata cylindrica] and Rangotra [Pteridium aquilinum ]. This usually occurs at the fifth cultivation cycle, but these plants can disappear if we wait about 20 years. But the Tenina can also invade from the third cultivation cycle. For the fourth cycle of cultivation, we usually wait 8-10 years because the soil is not fertile. We sow one vat [20 kilograms] and harvest eight to 20 vat [400 kilograms]….
We often move, when the soil is not fertile anymore. […] In regards to the aspect, it is better to grow rice on east and northeast slopes. We can also cultivate on other exposures if the season is good….

Forest guards do not come anymore from many years. But clearing of forest is not so fast because this depends on the families. It is a very hard work and people are also afraid of the law. Forest edges are also more favorable for cultivating. This description is consistent with the results of other interviews and with what is found in the literature, although there are some local variations in term of duration of the fallow period, crop rotation and association. Villages located closer to the road, where population density is higher, have shorter fallow period and more degraded soil, and develop new systems that I will describe later. Regarding the quantity of rice that is sown and harvested, it varies greatly from one family to the other. The amount sowed ranges from about two to 15 kopy and the harvests from five to 50 vats (100 to 1000 kilograms), harvests beyond one ton per household being exceptional. Considering that an average family (parents and 4 children) requires about 800 kilograms of rice per year, these results match with the fact that the rice shortage period lasts three to nine months for most households.

The variation in the amount sowed is explained by the variety of strategies in terms of combination of cropping systems and competitions between various tasks in the working calendar:

I clear the land for cassava in August but I only can do a small area. Then I leave the vegetation drying, and I clear another land for rice, and go back to the cassava land for burning and planting. Then I burn the tavy and plant the rice. Weeding starts after one month. We do it twice and also weed the cassava. Weeding is continuous. We do not stop until rice fructification. The worst weeds are Bemimbo [Ageratum conizoides] and Fatakana [Pennisetum sp?]. We have to throw the weeds far from the plot in order to eliminate them. (a farmer from Ambodilaingo, 2001)
We sow tavy first, in October-November and we sow rice in paddy fields later, in November-December-January. We can sow in paddy fields at different seasons because there is no problem of water.” “Plowing the paddy is in January, at the same time as weeding in tavy. We always lack time. We also lack time for watching over paddy fields and tavy in the same time. (a farmer from Ambodilaingo, 2001) These testimonies show that slash-and-burn cultivation is not the only land use in the area. Since several decades, and maybe since their establishment, people in Ambodilaingo already implemented “alternatives to tavy”, as we will now see.

Tavy and other land uses in the past

Tavy has actually never been the only land use in Ambodilaingo. It has always been combined with irrigated rice cultivation, perennial crops, and animal husbandry. It is unclear what was the respective weight of these systems in the past but alternatives to tavy clearly played a more prominent role in the economy than they do now. Cattle husbandry, for instance, was an important activity, even if manure was not used, as shown by the following citations:

About half of the families had big cattle in 1959. Those who had animals had many: 15 to 30, sometimes more. Those who did not have animals loaned them for free to the others. They did not refuse because people had good relations at that time. Those who had large herds hired people to watch over animals and paid them by giving calves. Usually the herds were all put together to avoid being lost…. There was much area available for grazing….The animals were put in corrals during the night, but only during the cultivation season, in order to avoid damage to the crops….Animals had good care and were visited by veterinarians….They were used for earning money and for sacrifices. There were more sacrificed animals than sold animals. (an elder from Ambodilaingo 2001) We did not use manure. There was no taboo about that, but the land was fertile. (an elder in Ambodilaingo 2001)

Today, only a handful of families in Ambodilaingo have cattle.

Paddy fields also played a greater role in the economy in the past. This fact was linked to the abundance of cattle because animals were used to thread the soil in paddy field, a labor saving technique that enables to sow directly without requiring plowing the land:

The first paddy fields were created in 1914-1918. This crop did not conflict with livestock husbandry. There was a time when paddy fields covered more areas than now, because people had many more cattle. The animals were used for treading of land before sowing. If we do not have cattle, we can only cultivate small areas of paddy fields. (an elder from Ambodilaingo 2001) Cattle was used to tread paddy fields before sowing. My father had 50 head. He had a paddy field because treading was easy. Paddy rice was a big activity in ancient times because people had many cattle. (an elder from Ambodilaingo 2001) Later on, cattle husbandry declined, and with it irrigated rice cultivation. The following statements relate this decline:
There has been a big decrease of cattle in the 1970s. The main reason is that the people started to be bad and asked penalties when cattle damaged their crops….These damages mainly concerned cassava, sweet potatoes and coffee, in the bottom land and in the lower part of the slopes. As a consequence, cattle started to lack access to pastures. People did not build fences because agriculture officers told the herders to care for their herds. In ancient times, there was not frequent damage because there were fewer crops in bottom lands and people were not so bad. (an elder from Ambodilaingo 2001)

Other reasons for the decrease in cattle are the sale because of poverty and the sacrifices….Cattle also died because of diseases: some people bought cattle from other regions, and this brought the disease. The animals get quickly thin and die [when they get this disease]. The liver is full of small animals and cannot be eaten. This disease developed around 1983 and still exists. (an elder from Ambodilaingo 2001)
The same explanation of the decline of cattle husbandry (conflict with agricultural activities, lack of pasture, diseases, poverty) were given in many interviews, although the ranking of these causes varies from one informant to the other. As all these causes are linked to each other, the explanation should indeed be seen as being systemic: the development of agricultural activities, probably as a consequence of population growth, led to increasing tensions between herders and cultivators, a reduction of land available for grazing, lack of pasture, difficulties to feed the animal, and diseases.

The decline of cattle husbandry, however, was not the only cause for the decline of paddy fields:

Another reason of the decrease of paddy fields is the inundations that destroyed the dams and improvements. The ancient paddy fields became land for tavy or have been planted into banana trees. (an elder from Ambodilaingo 2001) The government made more significant effort, during the 1960s, to develop irrigation schemes. Several dams have been built in the commune of Beforona at that time, including in Ambodilaingo. But no support was provided for their maintenance, at least since the mid 1970s, and they were progressively destroyed by cyclones and heavy rains.
The decline of paddy fields could also be explained by changes in policies regarding forest clearing and the practice of tavy. A ban on the practice of tavy existed during the 1950s, but was levered after independence, in the 1960s: The main period concerning interdiction of clearing was after the visit of Mr. Abraham and the forest guards in 1954…. Interdiction mainly concerned the upper part of the slopes. People clearing forest at that time were sent to jail, but this did not occur anymore after the independency. (an elder from Ambodilaingo 2001).

During the colonial period, it was forbidden to clear the forest. There were forest guards. It was strictly forbidden to clear the primary forest and the crests. This lasted until independence. (an elder from Ambodilaingo 2001. Policies may have played a significant role because although alternatives existed, tavy was of paramount importance in the economy. One could get rich by investing its resources in it: In ancient times, rich people did not make cash crops but extended their tavy cultivation. (a farmer from Ambodilaingo 2001)

Another possible explanation is the overall economic context. It changed significantly during this same period, from a relative economic prosperity in the sixties (coffee was sold a good price and first necessity products were cheap), to the degradation of the terms of exchange and the failure of the “socialist revolution” in the seventies, and the deep economic crisis of the eighties. People in Ambodilaingo were impacted by these changes:
Coffee cultivation developed around 1950, under the authorities of the contremaitre . Coffee generated higher income at that time. One kilogram was sold for 150 F, while one kilogram of rice cost 30 F. Today, one kilogram of coffee is sold for 2500 F, and one kilogram of rice costs about 2000 F. (an elder from Ambodilaingo 2001)

The Malagasy economy started to recover during the nineties, due to the opening to foreign investments and the financial support provided by the international community. But for the villagers of Ambodilaingo, like for many farmers living in remote areas of Madagascar, this recovery never occurred. The period from the seventies to present is one of continuous degradation of their livelihood. As they put it, “we are always becoming poorer.” The renovation of the road and the opening to foreign investments and international aid led to some changes in Beforona but did not reach their village, except during a few years when the ginger price was high. Poverty is now dramatic.

More investigation is required to rank the various causes of the economic crisis and of the decline of alternatives to slash-and-burn cultivation (animal husbandry, paddy fields, coffee cultivation). But the picture that emerges is that of a progressive erosion of a productive capital that could have enabled the shift to more intensive land uses if it could have been maintained. Paddy fields could have expanded and substituted to tavy if irrigation schemes had been maintained and multiplied. Manure could have been used to fertilize crops on hill sides if stables and fences had been built, as a way to reduce crop damage by cattle. All this could have been achieved more easily if the price of coffee did not decline and if the terms of exchange remained favorable. Maintaining a ban on forest clearing, by limiting the land available for cultivation, could also have favored this transition. It would have had the same effect as population growth on intensification, following Boserup’s model, while the absence of regulation may have favored the return to more extensive systems. Hence a systemic view where all variables are linked is required if realistic and equitable development and environmental policies are to be implemented. In the next section, we will see that implementing alternatives to slash-and-burn remains an important objective of most farmers in the area, in spite of the dire situation that prevails today.

Paddy fields in Ambodilaingo

Irrigated rice cultivation persisted in the area, in spite of the decline of cattle husbandry. Only one family continues to practice the old system based on treading, simply because it has one of the largest herd. For the others, paddy fields cover smaller areas and the system rather matches with the following description:

I have four zebus for treading but this is not enough. I have to work the land with the angady. If we do not tread, we do not have a good harvest. We prepare the land for the nursery using an angady . To prepare the land with an angady takes much time. After we sow, we have to watch over the field for 15 days because of the Fody birds …. We cultivate this paddy field from 13 years. We were doing tavy before, on a land that we now want to sell. All this bottom land is ours and we cultivate it every year, on the same plots. We usually sow one vat [20 kilograms]. There are variations in the yield because of the cyclones. The knowledge for doing this comes from our ancestors…. People do not use the plow because there are not many zebus in this area…. We do not have to weed if we sow in October, because there is much water. We do not extend our paddy field because part of the land lacks water and cannot be cultivated properly. (a farmer from Ambodilaingo 2001)

One could wonder, then, why irrigated rice cultivation declined. One reason is the lack of suitable land as only a minority of families own land in bottom land. Other key reasons are the lack of labor and financial resources. Without treading by cattle, the system requires more labour input which has to be either provided by the family to the detriment of other activities, or paid for. Labor is especially a constraint when a piece of land has to be cleared for the first time, or after several years of fallowing, because the vegetation forms a dense mats with many herbaceous perennials (Rubus mollucanus, Lantana camara and Aframomum angustifolium) whose roots are difficult to eliminate with a hoe or angady. But the real constraint to intensification, behind the lack of labor and finance, could be the lack of food. Hard work like making land improvement and plowing requires a significant increase of caloric needs.

The following testimonies summarize the situation, although they don’t enable to determine which cause is the most important:
Only about 20 families own land in the large bottom land. Some families own land in the small bottom land. Less than half of the families own bottom land. (an elder from Ambodilaingo 2001)

Some families have no bottom land because it was already cleared when their ancestors arrived. They have to request some land if they want to cultivate. This is usually accepted, but only for annual crops. It is impossible to make paddy fields or home garden in this way, except after buying the land. (a farmer from Ambodilaingo 2001) Not everybody has land but everybody can ask for land for cultivating rice in the bottom land. This is for free. We do not ask for this land because it is much work to eliminate the weeds. One needs to have money and food to invest in cultivation in bottom land. Rice in bottom land is cultivated like in tavy if we cannot irrigate, but there are many weeds and if we do not do a good weeding, then we lose. Nobody has worked this land with a plow. We just use the angady and the machete. (a farmer from Ambodilaingo 2001) People tried to dig a canal. But they lacked food and did not finish in time. There is a problem of lack of zebu, but irrigation is very important also. (a farmer from Ambodilaingo 2001)

Home gardens in Ambodilaingo

The second alternative, the home garden, is mostly dedicated to cash crops, first of all bananas and coffee. This makes it a great complement to paddy fields and tavy, which mostly provide rice and other crops for family consumption. Coffee, which has the advantage of having a low weight for a high value and of being transportable over long distances, was an essential cash crop during the 1950s and 1960s but has declined since the 1970s, due to lowering price as already mentioned. Bananas mostly developed after 1990 because the rehabilitation of the road in the late eighties increased market opportunities.

Home gardens establishment and management does not seem to require much labor except for transporting the product to the road, where it can be sold. This is the main constraint to the development of this activity (Nambena 2007): There is no special time to weed banana trees. We do that when we can, little by little. (a farmer from Ambodilaingo 2001) It is easy to plant 1000 banana trees in one year if we want and if we have many friends. We get the seedlings from friends. But we do not plant that much because we have many other activities. For instance, now is the time for weeding the home garden, but we also have to clear the tavy. Most people just plant 10 to 20 banana trees. But if we specialize in this crop, we can plant more. If there would be a road, sure, we would plant more banana trees because we could sell. (a young farmer from Ambodilaingo 2001)
Bananas, on the other hand, have the advantage of being harvested all year round, providing a limited but permanent income. Farmers often carry a regime of bananas when they go to the market in Beforona, which occurs about once a week. Unfortunately the sale price is low: about 0.5 USD per kilogram. Collectors know that farmers will not return to their village with a 30 kilogram load. They impose their price and farmers sometimes prefer to dispose of bananas rather than accept this price. Bananas are also consumed, mainly during the rice shortage period.

The other products found in home gardens (e.g., avocado, jackfruits, grapefruits, oranges, limes, and sugar cane) are locally consumed and often go to waste. Home gardens, in sum, shift from the status of being a significant source of income, at the time where coffee was sold a good price, to the statute of being a marginal complement to the economy, though the sale of bananas and the supply of seasonal fruits. Home gardens could probably be intensified, in order to produce more coffee and compensate for the low price for instance. But this would necessarily require investments and the taking of risks that is not compatible with the level of poverty that prevails in the area. The lack of suitable land to establish home gardens is also a constraint to their development: Some people do their home garden far from the village and have their house out of the village, for looking after the home garden. There is a risk of exodus out from the village, because people go far to find land for the home garden. (an elder from Ambodilaingo 2001)

We prefer to build our house in the village because there is more communication between the people. But if we want a big home garden, we have to find the land quite far from the village. If we accept to move, it is quite easy to find such a land. (a young farmer from Ambodilaingo 2001)
It is hard to find a piece of land for the home garden because the soil is hard. There is available land but it already belongs to someone. If we want to borrow, the owner refuses. Much land is in this case because people start to improve the bottom land and progressively cultivate the upper part of the slope. There is no case of land renting. (a young farmer from Ambodilaingo 2001)

Other alternatives in Ambodilaingo

Some families raise pigs and this activity has the potential to replace cattle husbandry as a source of manure and income. Pig raising suffered from a taboo in the past but this taboo is currently being levered, probably because of economic necessities. The main constraint to this activity is the lack of resources to feed the animals. When the available labor does not suffice to produce enough food for the family nor does the income enable to buy it (during the rice shortage period, many families suffer greatly from a diet that is mostly based on cassava, which has much lower nutritional value than rice), no surplus is available to feed animals. This constrains also fowl husbandry.

Ginger is another cash crop that plays a significant role in the economy in Beforona, as we will see. It used to be grown in Ambodilaingo, but was abandoned when price went down, because of the cost of transportation: The cultivation and market for ginger started to develop in 1987. Ginger was sold for 3000 FMg/kilograms five years ago [about 0.45 USD, in 1996]. Now, it is sold for 750 FMg [about 0.12 USD]. (a farmer from Ambodilaingo 2001) According to certain farmers, this price drop would have been provoked by a few farmers producing large quantities in Beforona.
People in Abodilaingo also establish small permanent plots where they grow mostly beans and Chinese cabbages, on land that is superficially tilled.
The main source of income in Ambodilaingo is logging, which provides a regular revenue to any male capable of transporting boards from the forest where they are harvested to the road where they are collected. The salary is about 1.5 USD for transporting a 30 kilograms board over about 25 kilometers, which usually last 1.5 to 2 days and can be done 2 or 3 times a week. This activity developed in the late 1990 and is on the decline since 2006, as a consequence of a national ban on logging.

The constraints to adoption in Ambodilaingo

In spite of this overall decline since several decades, developing alternative to slash-and-burn cultivation remains a main endeavor for many farmers in the area. People know what to do but lack the resources to do it, as shown by the following testimonies:
Rich people invest their money to pay hired people to extend their home garden, their paddy fields, or for coffee or cassava cultivation. (a farmer from Ambodilaingo 2001) If I had money, I would invest in agriculture and in raising pigs. A 15 kilograms female costs about 200,000 FMg [about 30 USD]. We cannot buy a younger one because there is a big risk that it will die, because there are many diseases. We can sell the pig for 400,000 in April, after it gives babies. (a farmer from Ambodilaingo 2001) If we had money, we would invest in ginger and rice. (a group of farmers from Ambodilaingo 2001)

But they lack investment capacity and market opportunities to adopt these alternatives: There is much land to cultivate but the problem is to find money. The area is favorable to ginger and pineapple but the problem is to sell the products. There are big price variations. (a farmer from Ambodilaingo 2001) We grow pineapples that are very easy to sell in Beforona because they grow big here. But the transport cost is 300 FMg per kilograms [0.45 USD]. We also have problems with rats but we can eat pineapple. It is good for the children. (a farmer from Ambodilaingo 2001)
The construction of a road is often evoked as a way to solve these problems: The main problem is the transport to Beforona. We could grow coffee, banana, rice, cassava, sugarcane but it is difficult to sell the products. People would be very happy if there was a road, because they could sell easily. Now we go to Beforona once a week to sell our products and buy something. (a farmer from Ambodilaingo 2001)

As a way to verify that investment capacity was the main constraint to adopting alternative land uses, but also to stop be extracting information without providing any reward, I decided to create a small NGO (Zanaky ny Ala) that would achieve these investments, in partnership with the Food Security Program of ADRA , financed by USAID. Zanaky ny Ala built 7 irrigation dams in the municipality of Beforona, only three being totally functional, unfortunately. The cost ranged from 1 to a few thousand dollars per dam, including the digging of the main irrigation canals, each dam having about 10 to 40 beneficiaries, meaning that the construction cost equals no more than a few years of harvest . The dams and their irrigation canals were delivered to the communities, who were not required any contribution other than their work as they were paid above the minimal wage through a food for work system. Once this was done, the NGO left the beneficiaries alone with the dam. In all cases where the dam was functional (Marofody dam in Ambodilaingo, XX dam in Haranana, and Ankorakabe dam in Beforona), farmers proceed to the land improvements and organized for the maintenance of the canals. In Beforona, the dam enabled many families to shift from rice shortage to rice surplus, by shifting from one harvest to two harvests a year, i.e., by passing the stage from annual cropping to multiple cropping. This confirms that land use changes occurs spontaneously as soon as new resources create opportunity. The municipality identified tens of sites that could benefit from similar infrastructure.

Crop rotation and short fallow in Beforona

The alternatives developed by farmers in Ambodilaingo are well known in other villages of the municipality. Paddy fields, cattle raising and coffee cultivation have been unanimously described as activities that had more importance in the past and are in the decline since several decades. A distinction has to be made, however, between the villages located close to the roads and those located farther away, the threshold distance appearing to be about 7-8 kilometers. The rehabilitation of the road in the late 1980s induced more significant change for the villages located close to it, with a possible recovery of the economy, through the development of cash crops, mostly ginger and bananas. Villages located close to the road also have a higher population density, and hence shorter fallow period and more degraded land on hillsides. The conjunction of these two phenomenons (better economic situation and higher density of population) results in more interest for intensive systems. Farmers often mention that they “move down to the bottom land”, where they establish paddy fields and home gardens. Messerli (19XX) and Nambena (19XX) provide many relevant information regarding this dynamic. But short fallow systems are also encountered on hillsides as not all families have bottom land. In this case, the land is often plowed and new rotations are introduced. Ginger plays a paramount role in the development of these new systems as it amortizes the plowing effort, by being planted the first year of the rotation. New forms of land degradation arise as a consequence of plowing on steep slopes, but farmers experiment ways to control or mitigate soil losses, by digging drainage canals and by plantings grasses or corn on their edge. Certainly these systems are not perfect, but they verify an historical sequence of land uses (the Boserupian model) that occurred in most regions of the world where slash-and-burn cultivation was abandoned. For this reason alone, they should be paid attention by projects that work in the area, together with the more ancient alternatives (paddy fields, home gardens and animal husbandry) that farmers practice since generations. We will see in the next section that this is unfortunately not the case.

Project interventions in the area.

The support provided by projects working in the area mostly focuses on alternatives to the alternatives that were developed by farmers. An examination of the package proposed by the CDIA-LADIA, an agricultural research and extension center located in Marolafa (about 1 kilometer from the center of Beforona) enables to review these “alternatives to alternatives”. It would be worth also to examine the research conducted in the area, as they also reveal the same logic of overlooking farmers’ logic (Pollini 2007 ). I will nevertheless limit my analysis to the CDIA package because it summarizes the finding of these researches, or, at least, it represents their practical outcome.

The CDIA-LADIA center is regarded as a model farm showing the various agricultural techniques that farmers are expected to adopt, and about which they are trained. I will critically examine the cropping and livestock system it displays on the small watershed that it occupies. Up the hill are trees that have been planted by the CTFT project in the early 1970s. They mostly belong to the genera Pinus, Araucaria and Agathis. They reflect the concern about the loss of forest cover that always guided project interventions in the area. But in a context where pressure on primary forests is still elevated, and where farmers even clear eucalyptus plantations for practicing slash-and-burn cultivation, the rationality of planting trees on hillsides can be questioned. Planting trees on fallow land means less land available for agriculture, which may contribute to increase pressure on remaining primary forest, as these are considered as territory in more or less open access while plantation forests have an identifiable owner.
Under this woody crown, banana trees have been planted in association with coffee and a few fruit trees, revealing that projects are aware that home gardens are an alternative to slash-and-burn cultivation. These home gardens show some differences as compared with those of local farmers: they are not invaded by weeds or invasive species such as Rubus mollucanus, whatever is the period of the year, as they are weeded several times a year. As a consequence, their yield is higher but they also require more labor input. This contrasts with traditional home gardens where weeding can be done any time of the year, which leaves time to the production of rice and other subsistence crops. The key question, now, is whether the higher yield obtained compensates for the increased labor input to the point that labor productivity, calculated in consideration of the varying opportunity cost over the year, is increased by the new system. This question has not been answered but farmers give their own answer: by not adopting the new technique.

Looking down the home garden, we see hedgerows planted on contour lines that delineate plots either cultivated or in a state of herbaceous fallow, on land that farmers in the area commonly dedicate to the practice of tavy. The vegetation is never burned on this land, and compost and/or mulching is often applied. These new techniques represent a significant investment in terms of labor. A permanent cultivation of ginger, beans, rice, or other crops, in the alleys between the hedgerows, is proposed to amortize these investments, following the same logic that farmers apply when they grow ginger on plowed land. The outcome of combining hedgerows and other conservation practices with a ginger cultivation is “ecological ginger”, which is supposed to obtain higher yields and be more profitable than traditional systems: “[Ecological ginger yields] from 7.5 [to] 14 T/ha (traditional ginger yields ~ 5 tons/hectares), and even though it requires increased labor and financial investment (the use of compost, limited chemical fertilizer and hillside stabilization techniques such as contour vetiver planting), it is very popular along the eastern side of the forest corridor.” (LDI 2003, 74)

In spite of this optimistic statement from a LDI project’s annual report, the adoption of “ecological ginger” remains low indeed. Farmers contest the yields obtained by the CDIA, and, more importantly, explain that higher yield does not necessarily result in more production, because the new techniques also demand more labor. As a consequence, they can only apply the technique on a small plot. As a farmer of Beforona directly involved in the program as extension agent explained: “Fire is doing much work for us. It clears the land and kills the insects and the weeds. So if we do not use fire, we have to apply many supplementary techniques to clear and fertilize the land and to control the pests. Each of these techniques takes much time and then we cannot cultivate a large area. So I apply the new techniques on a small piece of land for demonstration and I continue to practice traditional systems on the rest of my land.”
“Sometimes I also use 80% or 90% of the technique. But what really takes much time is the preparation of compost. This I really cannot do.”
This assertion is consistent with Messerli’s (2003) calculation of the economic return of various alternatives to tavy. The supposed high profitability of ecological ginger is also contradicted by the fact that this crop is grown only sporadically and on limited surface in the CDIA itself, although bureaucratic issues (the fact that the center depends on top-down decisions taken by the projects that support it) can also explain this limitation.
The fourth element of CDIA’s model farming system is the bottom land. Small paddy fields are encountered there, but cover quite narrow surfaces. A small cement dam and a canal divert water from a small stream for irrigation. Two types of rice cultivation systems are implemented: SRI and SRA.
SRI is an innovative rice cultivation system that can achieve high yields without requiring chemical inputs (Stoop et al. 2002). It is based on a simple principle: whereas conventional systems imply a permanent inundation, SRI necessitates alternately inundating and drying the field, with a weekly frequency or more. This provokes an oxygenation of the soil that strongly benefits the root system, allowing rice to express its total biological potential if other factors are not limiting. Many thallus can be obtained from a single stem and more spikelets are born by each stem. The rice can be sowed at larger spacing and with just one seed per seed hole, which allows saving many seedlings. Yields of 10 tons per hectare and more can be achieved. Soil oxygenation is higher, which stimulates biological activity and makes nutrients more available and better preserved in the long-term. SRI works with all varieties of rice and does not require “modern” inputs such as fertilizers or machines.

The diffusion of SRI in Beforona, however, faces several problems: farmers criticize it for the high labor input it requires. The reason is simple: the oxygenation of the soil also benefits the weeds. The main function of the water table in rice cultivation systems is actually not to feed rice in water, but to eliminate weeds. In the absence of inundation, SRI plots have to be weeded three or four times (some technicians recommend weekly weeding). The consequence is a lower return per unit of labor that causes concern to farmers, as in the case ecological ginger. I am not able to provide hard data to substantiate this conclusion, but this is precisely the point that disserves attention: SRI is proposed to farmers despite the fact that its profitability has not been calculated in terms of return per unit of labor, and despite the fact that farmers’ experience already led to the conclusion that SRI produces less, considering the limited surface on which it can be applied. As farmers put it, “we do SRI on a small plot and traditional rice on the rest of the land.” They actually try it rather than adopt it, as they usually abandon it as soon as projects leave, a tendency that is observed all around Madagascar (Moser and Barrett 2003a, 2003b).

Moreover, even if SRI provided higher return per unit of labor (which may be possible when practiced on fertile land by farmers who became experts in the technique and can hire daily workers to solve bottleneck labor constraints), its adoption may still not be economically rationale. The reason is that in a context of high poverty and frequent disasters (cyclones) an economically rational farmer (one who maximizes his chances to feed his family) is one who maximizes production the worst years, rather than production the best years. SRI is riskier because transplanting from the nursery is done at an earlier stage (8 to 15 days after sowing), which renders the plot more vulnerable to flooding. Other risks (pests, diseases) may exist that can annihilate the return of the high labor a farmer invests in SRI, and no comparative assessment of these risks has been done to date.
SRI’s rationale, in sum, is to express the complete yield potential of rice. But even if we agreed about the ways to achieve this potential (which could also be probably achieved using more conventional techniques), this does not prove that SRI is a rational way to grow rice. Rice as a plant with its physiology should not be confused with rice as a crop. Achieving the full yield potential of rice as a plant implies that there are no limiting factors. This may rarely be the case in many non laboratory situations, i.e., when rice grows on a farm, unless farmers invest all their available resources on it, which would transform the least accident into a disaster. Some climatic risks, pest, diseases and labor constraints always remain and farming practices are aimed at mitigating these risks; not at betting maximum gains through maximum care. Maybe SRI has a potential in deltas where rice is almost the only crop, the biophysical environment is perfectly mastered and people have very small surfaces, but we are very far away from such a situation in Beforona.

The LDI and PTE projects were aware of these limitations and proposed another pathway for irrigated rice intensification: the SRA. SRA is based on more conventional techniques such as transplanting in rows, using varieties with higher yield potential and doing an earlier transplantation. Its diffusion started during the first republic (and maybe earlier), through support by the CTFT project. Some ingredients of the SRA package are sometimes adopted by farmers. The main advantage they notice is that new techniques (both SRA and SRI) allow saving seedlings due to a larger spacing and a lower number of seedling holes. SRA extension thus certainly contributed to the progress of rice cultivation in the area. However, the labor costs of the new techniques are also criticized, although to a lesser extent than in the case of SRI. The main work constraint results from transplanting in rows (this also applied to SRI). Rows allow mechanical weeding, which saves some time, but on the other hand, more time is needed for transplanting. Furthermore, weeding machines are not always available, are expensive with regard to the investment capacity of most households, and the labor gain for weeding does not necessarily compensate for the labor cost of transplanting into a row. In order to compare these two tasks, labor input has to be measured in terms of opportunity cost, which has not been done yet. To summarize regarding paddy fields, this technique has been correctly identified as a promising alternative to slash-and-burn. But support concerns mostly technical aspects, and the techniques proposed are not adapted to the local situation.

The CDIA further developed livestock husbandry activities. It raised improved breeds of pigs and chickens but these activities were not profitable due to the high cost of food supplements. They were abandoned in 2002 although they restarted later. The CDIA also created fish ponds but their production was low or uncertain for reasons that are not yet clear, despite the sustained efforts of a Peace Corps volunteer. Moreover, they necessitated labor investments that farmers cannot easily afford. Here again, a calculation remains to be done to know if the profit compensates for the investments, and to compare the profitability of fish ponds with other activities. Farmers have done their empirical “calculation” already and are reluctant to create fish ponds on their land. Fish ponds are even the subject of jokes among staff hired as daily worker by the CDIA.
The situation I described is indeed common in agricultural research and extension centers. The limits of this type of approach have been already analyzed and criticized in depth XX ref XX. The CDIA, similar to other centers of the same kind, is isolated from the farming systems it is supposed to support. In spite of this, the CDIA is the subject of many newspapers articles that speak in praise of its achievements. It may be that its implicit function is the demonstration of the project’s achievements to journalists and government officials, rather than to farmers. Representatives of the Malagasy government, of the donors, and even of citizens from donor countries (American congressmen) come to visit the center and leave with an enthusiasm that contrasts strongly with farmer’s criticisms and complaints. Hard data and graphs about high yield and monetary return are heralded as proofs of success in power point presentations while in the real world, just outside the window of the CDIA seminar room, farmers who tell the whole story of their land use and of the CDIA alternatives draw quite different conclusions. In this context, it is not a surprise that most activities at CDIA stopped once the projects that supported it came to term. Beyond the development and demonstration of new techniques, the CDIA also proposes training programs about agroforestry (contour strips, alley cropping, and improved fallow), conservation agriculture (mulching, zero tillage), home garden intensification, “ecological” ginger” cultivation, horticulture (litchis and vegetables), SRA and SRI, compost production, and swine and chicken husbandry. When beneficiary farmers are asked about their interest for the training they receive, they almost unanimously answer that the topics are interesting but that they cannot apply the proposed techniques because they do not have sufficient time or financial resources. In short, the content of training, upon which I had opportunities to give a few glances, reflects the same overlooking of local realities as revealed by the CDIA farm package.

Economic support

In summary for the previous sections, there is a great contrast between the alternatives to slash-and-burn cultivation developed by local farmers and the “alternatives to alternatives” proposed by the projects that work in the area. The first are attempts to take opportunities of new resources (water, emerging markets), through the mobilization of the labor force that is available on the farm. This is done cautiously, meaning that both slash-and-burn cultivation and the alternatives are implemented during a transition period that can be quite long. The second consists of a complete remobilization of resources toward new technologies that brake completely with the ancient system, as exemplified by the rejection of burning. The first is constrained by the lack of economic resources (finance, labor force, tools) while the second assumes that the challenge is mostly technological, and proposes state of the art technologies that require still more economic resources (more labor, more credit, more tools). Projects, however, are aware that economic constraints explain the low adoption. So they also propose financial services that I will briefly describe here:
LDI helps farmers to obtain funding through the OTIV system and the PSDR project, mostly for developing ecological ginger cultivation, livestock husbandry and vegetable cultivation in a few cases. LDI also helps the Kolo Harena cooperatives to establish commercial contracts, mostly for the sale of ginger.

Specific data about Beforona concerning these grants are not available, but an overview at regional or national level enables to capture the strategy of the project. For the overall LDI project (mostly the Toamasina and Fianarantsoa provinces) 143 farmers have been helped during the year 1998-1999, mobilizing 16,500 USD at OTIV. This allowed the cultivation of irrigated rice using intensified and sustainable techniques on 242 hectares of land. A second request of about 2,700 USD “helped 68 farmers produce upland beans on 48.3 hectares using anti-erosive and soil fertility improvement technologies that will stabilize the need for continued slash burn agriculture on the hillsides” (LDI 2000, 4). About 90,000 USD have further been mobilized, from OTIV and from the Bank of Africa, to fund 45 community granaries. During the year 2002-2003, OTIV loans “totaling 158,313,350 FMg were disbursed to 218 Kolo Harena farmers beneficiaries [110 USD per beneficiary] with 60% of these loans funded by the revolving fund, and the remainder as new loans.” (LDI 2003, 73)

Concerning the PSDR, this project accepted 18 grant requests during the year 2001-2002 (LDI 2002, 27). In the year 2002-2003,
with assistance from LDI staff, the Kolo Harena Federations begun submitting requests to the World Bank-funded PSDR project for grants involving as much as 60 million FMG [about 9200 USD]. These grants were destined to help the farmers develop new, or expand existing production activities (e.g., ducks, peanuts, fish farming) which are implemented with LDI and farmer extension agent support. Grants approved thus far to Kolo Harena total 20 billion FMG [about 3 million USD]. (LDI 2003, 11)

Considering this general information and the few examples given in the reports (not specifically from Beforona), it seems that a limited number of beneficiaries benefited from most of the aid, and that the amount of PSDR grants represented a large sum in comparison with the economic level of slash-and-burn farmers. Moor (1998b) calculated that households invest about nine to 13% of their revenue, which represents 22 USD for an annual revenue of about 175 USD. The farmers supported by Messerli (2003) had an annual benefit ranging from 700,000 to 900,000 FMg (100 to 140 USD) and invested 13 to 41% of this benefit into agriculture, which represented about 10 to 40 USD per year. These amounts appear to be insignificants at first glance but actually allowed farmers to significantly improve their livelihood. This even sometimes resulted in jealousy from the rest of the community (Messerli 2003).

In comparison, the amount of OTIV grants may be adequate in the case of richer or less poor farmers. For the poorest, they may represent amounts they could not reimburse, especially if a cyclone, a flood or any other natural calamity were to destroy their harvest. This would be particularly true in the case of remote villages having no access to market, such as Ambodilaingo. The PSDR grant represented a still more elevated amount. Several thousand USD were usually granted to Kolo Harena associations which grouped less than ten households (in the case of Beforona). However, PSDR loans were nonreimbursable, to the contrary of OTIV loans, meaning that PSDR grants represented no risk. This made them highly attractive but also made impossible the generalization of the approach to the whole population. In such a context, farmers with more education and economically better off, often migrants that never practiced tavy or were never dependant on this land use, were the first to benefit from the loans. Even nonfarmers (teachers, medical doctors, employees from projects, etc…) received nonreimbursable grants from PSDR in Beforona. Certainly these people deserve support because not being among the poorest does not mean being rich in Beforona. But if the objective of the projects was to reduce tavy and poverty, PSDR loans may have missed their target.

Coming back to the smaller loans made by OTIV, they could have been appropriate if properly managed. But they remained inefficient because they were conditioned to the adoption of fireless techniques, mostly for ginger cultivation, which did not prove to be profitable as we saw in the previous section. Many farmers had difficulty repaying the loans and complained about the low profitability and high risk of these “supports.”
In sum, economic support existed but missed its target, probably because it was linked to the adoption of techniques that were ill designed, and that only less poor farmers in search of social prestige and networking could adopt.

Interventions in Ambodilaingo

The perception of development projects by farmers living in Ambodilaingo confirms this inadequacy of both their technical alternatives and economic supports. Farmers in Ambodilaingo had limited direct contact with project staff, because LDI, like most other projects, focused its intervention on village located less than 10 kilometers off the road. The only significant intervention that was related to me was an attempt by LDI to intensify crop cultivation on bottom land, in collaboration with the village teacher. The proposed technique is similar to those proposed in Beforona (use of compost) and raise similar issues (more labor input): The difference with the traditional system is that the land is plowed, instead of just taking out the upper part of land with all superficial roots. This asks four times more work but plowing gets easier every year. I had a training in Andasibe for making compost and I will start soon to do that. (the teacher of Ambodilaingo 2001)

But even if projected intervened, their visit would be highly disputed because of the emphasis they put on not using fire. Farmers are aware of these technical choices, and of the political agenda associated to it. This creates caution, doubt and even hostility, as shown by the following citations:
I know the people from here. People do not dare to tell you that but I want to be sincere. People in Ambodilaingo are afraid of LDI because they have little means to cultivate the bottom land and only do tavy. People are afraid of LDI because LDI impedes doing tavy and cutting the forest. People here also observe that LDI does not really have a solution to replace tavy. They have some solutions, but it is impossible to do what they propose. LDI also tries to convince the foreign people that they have some solutions. (a farmer from Ambodilaingo, among the most educated 2001)
My own presence also created fear in the village, at least at the beginning: Maybe you did not follow that in the meeting, but there was a big discussion when you first came. The elders did not want you to come but the young people insisted for the village to receive you. People are not confident with LDI because LDI does not support them well for the new techniques. (a farmer from Ambodilaingo, same as for the previous citation 2001)

When you came for the first time with [the representative of the commune], I proposed that we wait for the agreement of the elders from the five hamlets of the fokontany, before giving approval. I have some difficulties feeding my 11 kids and I resisted at the time of your arrival, but the person who was accompanying you insisted on getting my immediate acceptance, without agreement from the whole fokontany. (an elder of Ambodilaingo 2001) The main concern is, of course, the attitude of visitors with regard to tavy: When one forbids tavy, one kills people. (a farmer of Ambodilaingo 2001) If you do not come for forbidding tavy, then we are safe. (a farmer of Ambodilaingo 2001)

Why such a discrepancy?

In this section, I will try to explain the reasons of this discrepancy between the faremrs’ and projects’ alternatives to slash-and-burn cultivation.
Before to start, it must be noted that the situation in the center of Beforona, where the CDIA is located, is quite different from Ambodilaingo, where I collected most of my data. The center of Beforona is better connected to markets, has a population with higher density, and more degraded soils, than remote villages like Ambodilaingo. Most localities of the municipality, however, have in common that their economy depends mostly on subsistence crops grown on slash-and-burn cultivation systems, although this is more true in the case of remote villages, and that farmers develop or try to develop paddy fields, home gardens, livestock husbandry and cash crops (mostly ginger) as alternatives, although this is more true close to the road. Another key difference is that remote villages of the western side of the municipality still have forest land with logging operation providing significant income (board transportation) in the southern half (where Ambodilaingo is located). Nevertheless, the LDI project, like other projects that preceded it, was targeting the whole municipality, claimed that its alternative land uses could be adopted in remote villages as well, and tried to some extent to intervene in remote villages, as we saw in the case of Ambodilaingo. So the comparison of the two situations remains relevant.
What are, then, the causes of discrepancy? A complete treatment would take much time. I undertook it in Pollini (2007) and summarizing it here would prove difficult. But one way to see it is to refer to three key books that are still classic work in the field of peasant studies, and whose finding are ignored by projects and experts working in the area. These are Peasant Economic, by Alexander Chayanov (1922), The Condition of Agricultural Growth, by Esther Boserup (1965), and The Moral Economy of the Peasants, by James Scott (1976). These three books provide deep insights of the functioning of peasant societies at three complementary levels or dimensions: the peasant farm, the agrarian landscape, and the peasant’s mind, respectively. Put together, they constitute a theory of agrarian change and culture, each focusing on key aspects of agrarian systems that are precisely the aspects that are overlooked by rural development projects, in Beforona like in most regions of the world. They could constitute the building blocks of a new agronomy liberated from the dogmas and fashions of mainstream agronomy, so I will briefly point out their key contents relevant to our case study.

Chayanov showed that peasant economics does not work like business economics. Peasants attempt to maximize the output for any input unit of the rarest resources (land, labor, or capital), and to minimize risks, whereas a “modern” farmer adopting a business would attempt to maximize the profit rate of any unit of capital he invested. So Labor in peasant economics is not just a cost like any other input. It is not subtracted as a cost in economic calculation (which is what most experts, students and staff do at CDIA when they assess the performance of farming systems in Beforona). Instead, the total output is divided by the labor input, which can be expressed in number of labor days, or in term of active workers on the farm, the later having the advantage of not being biased by the fact that every day of labor has a different opportunity cost. This issue of opportunity cost of labor is indeed given paramount importance in Chayanov’s work. This notion explains, for example, why it can be economically rationale for a peasant to buy a processing equipment (like a rice hulling machine) that results in high higher cost of processing in book keeping terms (if one establishes a conventional business plan, like is required to apply for an OTIV or PSDR loan), while it can be irrational to buy one that reduces this “cost”. The explanation is simply that if the days of labor saved by the machine can be invested in an activity that generates great income, the benefit of this activity could overcome the additional cost of the processing machine, so overall income will be higher. On the other hand, if the machine saves labor at a time when there are no other tasks to achieve, then it will actually save almost nothing, because this saved labor has almost no opportunity cost. It could even have a negative “cost”, if for example it gives an opportunity to socialize with friends or relatives and to increase social capital. Of course, if labor could be provided or supplied at the same rate than the cost given to it in book keeping, Chayanov’s model would not work, but such a situation rarely occurs in peasant societies, where most people tend to have labor constraints occurring in the same time. The example of the processing machine is the most appealing one, but this logic applies to any operation that involves labor. Addressing it would imply that agronomist elaborate extremely complex models where not a single day of labor would have the same cost, with variation occurring also from one year to the other and from one farmer to the next. If the benefits of increasing social capital by doing collective manual work were considered, the modeling exercise could become still more challenging. Such a formal calculation is indeed basically impossible to achieve, but farmers do it, in their own empirical way, as they experience every day their labor constraints and those of their neighbors, and the variations of their social capital in relation with the kind of activities they have, and adjust their system accordingly.

This example shows that there is a form of knowledge that is collectively produced over long periods of time and that may play a very decisive role in enabling communities of farmers to finely tune their system to biophysical and social constraints. Complexity science is starting to investigate how complex decisions (like “what farming system to adopt ?”) are elaborated from the interaction of a large number of agents who follow simple rules (like “how to use better my time today ?”). One striking example was presented by Lansing (19XX), who showed how farmers develop the most efficient pest control and irrigation schedules in Bali, just by adjusting to the decisions after comparing their results to that of a few neighbors. There is a whole domain of knowledge here that goes much beyond the finding of Chayanov, but that may render its finding more meaningful, almost a century later. This kind of knowledge may provide a justification for trusting farmer’s decisions and logics more than we trust our own. The emerging field of complexity science could indeed provide a theoretical justification from a paradigm shift in the way development is implemented, from a logic that focuses on design and planning, to a logic that focuses on incentives applied to free agents who are trusted for the rationality of their individual decisions. Such a model is indeed the one that enables the economic success of free market models. It would be worth to investigate why the defenders (or supposedly defenders) of such free market models (most donors of international aid) prefer central modes of planning when they are dealing with small farmers all over the world. But that would be a subject for another paper.

Boserup, beyond addressing the question of whether agricultural intensification is the cause or the consequence of population growth, showed that the key concept to understand agricultural intensification is that of frequency of cropping. When a farmer wants to increase production, there are three things he can do: he can increase yield; he can cultivate more land, and he can increase the frequency of cropping. Boserup explains that western thought about agricultural intensification overlooks the third strategy, which can be explained by the fact that during the last centuries, European farmers mostly depended on yield increase (through the development of new varieties and the use of chemical fertilization) and the conquest of new land (through colonization). But the more universal pattern of agricultural intensification proposed by Boserup, and evoked in the introduction of this article, rather emphasizes frequency of cropping, which shifts from 1/25 (forest fallow, where 1 year of cultivation alternates with about 25 years of fallow) to 1/8 (bush fallow, which is the situation that prevails in Ambodilaingo), 1/3 (short fallow, which is a situation encountered in some places in Beforona), 1/1 (annual cropping, like in rain fed paddy fields), 2/1 (multiple cropping, which can be practiced on irrigated paddy fields) or 3/1 (common practice in areas with very high population density). Hence it is possible to multiply several folds the density of population, from a few tens inhabitants per square kilometers, like in Ambodilaingo, to several hundreds, like in Java and Burundi, and feed it without increasing yield (slash-and-burn cultivation in forest fallow have quite high yields) and without clearing more land; just by increasing the frequency of cropping on the same land, which is indeed what was accomplished on the paddy fields irrigated by the Ankorakabe dam in Beforona. This requires new techniques, but not necessarily the same techniques than those proposed by agronomists who focus on yield.

Another key finding of Boserup is that more or less intensive systems usually coexist during periods of transitions that can last a while. This is because landscapes are heterogenous, so different frequency of cropping are justified depending on the location. This obviously applies to Beforona, where annual and even multiple cropping can easily be applied in bottom land, but not so easily on hill sides. Eventually, Boserup showed that fertility is as much a consequence of land use than the other way around. New techniques can end in building up soils that had been degraded during previous phases. This is consistent with the finding of Tiffen and Mortimore, which showed that there are Kuznet curves for soils in Machakos, Kenya. Maintaining fertility on slopes at all cost may thus not necessarily be the most rationale thing to do if considering the long term. If farmers in Ambodilaingo and Beforona successfully developed more intensive systems in bottom land, they may later be able to invest the resulting additional incomes in restoring land on slopes, by developing livestock husbandry and applying manure, for example (like are actually doing many farmers of the Malagasy highlands), or by planting perennial crops.

The third building block is the concept of moral economy, investigated by Thompson (19XX) in the case of the English working class, applied to the case of peasants by Scott (1976), but also found, even if the word is not used, in the writings of Bourdieu (19XX), Polanyi (19XX), Sahlins (19XX), and in the literature in economic anthropology and substantive economics, to which these two authors belong respectively. The moral economy is a system of social interaction determined by economic imperatives but framed within an ethical system where having enough, rather than having more, is the main imperative (the subsistence ethics), and where collective interests prevail over individual ones. It integrates collective experiential knowledge acquired along the time and ritualizes behavior whose rationality cannot be stated explicitly by individual actors. Like for the optimal allocation of labor resources in consideration of opportunity costs, it is an emergent property of complex systems. Hence the prevalence of collective interests over individual ones may be an optical illusion: the adaptive feature of the moral economy may be precisely to maximize individual gains measured using an indicator that would account both of their average (shortsighted economic aspect) and their variation (ethical aspect that may render the system more sustainable). This would be achieved through “centuries of trial and errors” that produce “the most stable and reliable yield possible under the circumstances,” in order to “iron out the ripples that might drown a man” (Scott 1976). Note that the moral economy both involves economic institutions (systems of redistribution and reciprocity), in order to address ethical issues that market mechanisms fail to consider, and techniques, in order to spread risks and achieve the subsistence ethics. The moral economy may indeed be a key dimension that any economic systems needs if it is to achieve sustainability, because it mitigates imbalances and maintains strong social ties between individuals. It should not, however, be seen as the miracle recipe to sustainability issues. The moral economy is engraved within culture and depends on strong social ties. It could be a cause of cultural and social inertia that could be maladaptive in a context of rapid change.

Coming back to the situation in Beforona, the moral economy explains the jealousy of farmers who do not get support from projects vis a vis those who get these supports. Such jealousies, often accompanied with sabotages that Scott (19XX) described as being “everydays’ forms of peasant resistance”, are well known by development project staff who interpret them as showing that peasants are backward and do not want to develop. Once the moral economy is understood, the interpretation changes. It can be shown that project beneficiaries are often better off farmers who take advantage of the supports they receive (like PSDR loans) to fill Market niches (by growing large quantities of ginger) and appropriate the most fertile land (by practicing usury, or “buying” land titles from corrupted civil servants), the outcome being the marginalization of the poorest farmers (like farmers in Ambodilaingo who abandon ginger cultivation because of price drop provoked by big producers; farmers moving out to forest land when migrants accaparate the land). A way to escape this is either redistribution: beneficiary farmers make gifts to the left behind, which is their way to show that they respect the moral economy; or sabotage, done by the left behind in case this redistribution does not occur. These logics show a certain form of economic rationality rather than backwardess, as soon as collective gains in the long term matter more than individual gains in the short term. Peasants want development indeed, but want it to be a collective process that does not let them behind, which obviously has an implication on sustainability. Note that I did not investigate into details these dynamics in the case of Beforona, however, but it is well know that they are common practice in Madagascar and I had a several indications that they exist in Beforona too. The moral economy enables also to interpret the frequent prejudice that peasant are lazy. Peasant are not lazy indeed, as shown by the incredible efforts that do to transport heavy boards or logs during days to purchase food and first necessity products. But they also value social life and leisure, which is a source of social capital that does matter also for their individual economy. The level of efforts they are ready to put in farming is determined by a balance between satisfying direct subsistence needs where having enough, not having more, matters (the subsistence ethics), and satisfying social life as a way to increase social capital and simply enjoy life. This equilibrium was clearly identified by Chayanov, who called it the “drudgery-utility” equilibrium.

Conclusion: toward a substantive agronomy

Agronomy is about finding better ways to produce food and other plant and animal products. It thus relates to an economic system. Paralleling the concept of substantive economics and its contrast with formal economics (economics as taught in conventional textbooks), I suggest, hoping that this will not lead to a new academic fashion, to open a new paradigm for agronomic research that I would call substantive agronomy. A substantive agronomy would be an agronomy that escapes from formal models elaborated along decades based on a minimal set of assumptions. Among these general assumptions is the belief that individual agents make their decisions only or mostly based on the computation, by themselves, of signals regarding the cost and the utility of the goods they want to purchase. Economists would argue that this model is encompassing because the utility function can include such things as participating social events, having more or less leisure, conforming to social norms, etc… I argue that it is not encompassing because there is a whole domain of knowledge that is built up through collective experience over the time and that, even if it could theoretically be reduced to a series of variable in the utility function, cannot be so in practice, simply because the function in question is not part of the cognized model (Rappaport 19XX). It is part of the operational model (Rappaport 19XX), which is why it plays an adaptive role, but it is not part of the cognized model, because nobody ever figured out this function. If the function is part of the system, it is simply because its outcome has been experienced as positive, which led to retaining the behavior from which he function arise along successive trial and errors. There is no need to figure out the causal relation between the behavior and the operational function to achieve this, but economists need to formulate this causal link if they are to add the proper terms to the utility functions they test. Among the three main ways to produce knowledge (logical statements arising out of the rational deliberation of individuals; collective experience that produces common sense knowledge, and memory that expresses its outcomes in the form of myths), economists, and the agronomists that follow their paths, may only take the first one seriously. A substantive agronomy would be an agronomy that takes seriously the three ways, as a way to account for the full substance of the reality, and design its tools accordingly.
A substantive agronomy will have to free itself from the dominant ways to do science, which are positivism or classical empiricism on one side XX ref Kanbur XX (which focuses on processes that can be directly perceived in the real world) and hermeneutic on the other side (which focuses on the interpretation of meaning and tends to disinterest itself from the real world processes about which these meaning refer). The works of Boserup and Scott, but also of Salhins and Polanyi, have in common that they are profoundly empirical but do not focus on directly measurable processes. They are based on a deep or long experience in contact with the real world (although Scott’s moral economy was mostly written based on literature review; scott will correct this limitation later when he will publish everyday’s forms of peasant resistance) but tell a story out of this experience. There is no hard fact, measurement, and statistics, to support their case. These hard facts are not there because they would probably lead to focus on a few aspects of the reality and to loose touch from the big picture, leading to higher internal validity but lower external validity. The story, however, does not disconnect from the reality, as it would if it arose out of ideology, or prejudices and preconception, rather than out of direct contact with the real world. The relevance of scientific knowledge may indeed be a function of this balance between what comes out from ideology and what comes out from direct experience of the world, rather than a matter of whether knowledge is the outcome of measurement and formalization or a matter of observation and description. In both these modes, ideology can be present, by telling what to measure and put in the model in the first case, or by telling what to observe and describe in the second case. A substantive agronomy will thus do both measurement and description, but will maximize direct and fresh experience while minimizing the sets of preestablished concepts it uses; or at least pick these concepts in the broadest range of disciplines and fields, as it cannot escape from the practical necessity of using reading grids if it is to achieve sufficient efficiency to translate its findings into real world actions. This broad and open analytical frame will enable it to capture a background noise that may render agrarian systems adaptive, resilient, evolutive, and functional, as well as the blasts that choke this noise.
Note that some schools, like the school of comparative agriculture in Paris XX ref XX, already perpetuate the Boserup/Chayanov heritage and could be the location of a substantive agronomy. These finding of Scott, Salhins, Polanyi, are also very influential in several academic branches, from economic anthropology to development sociology, and of course in the Agrarian Studies Programm chaired by James Scott himself and the Institute of Development Studies. But these schools did not make it to the mainstream, either because they are shy or reluctant to be involved in such mainstream real world activities, or because they are perceived as a threat to “aid as usual” and rejected. But maybe the never ending critiques and failure of “aid as usual”, and increasingly vocal critique of “science as usual,” visible in the field of complexity science for instance, will give opportunities to engage more institutions and individuals in this new paradigm after all.

The method: questioned if just story, But each story (eachcitation) consolidates hundred statements and decades of experienve.

Left over
We are used to contrast inductive versus deductive methods by saying that one goes from the specific to the general and the other from the general to the specific. So some scholars have brilliant theories and look for the factthat are consistent with them in the realities(idealism). Other have no theory and start form observation, but can handle only a few observation so produce narrow and incomplete often misleading theories (classical empiricism). We ned to start from the reality without narrowing the scpe, like we do in every day's life. Only doing so can we have a chance to add collective and historical knowledge to individualk knowledge that is insufficient to address sustainability issues, because it overlooks the emergent properties of self organizing systems.

the reality is too complex to be planned. We must trust self organizing processes. A new agronomy needs to consider this. It needs to open new areas of practice but not elaborate precise designs. Design has an area of application, however,: when sufficient resources can be provided to control all variables of the biophysical and social environment. So would the old agronomy still have prevalence in rich countries? Not so sure, because of sustainability issues. We now experience the limits of design and need to trust more self organization as an emergent property of a multitude of free agents working together. History showed us the success of this: the free market model. The notion needs to be extended by finding ways to set up agreed rules.

“It is clear that this challenge [sustainability] cannot be met. The world is far too complex for us to perceive and establish the conditions for sustainability” (Norgaard 1994, 23). Social and natural systems coevolve (Norgaard 1994). They behave as complex systems that self-organize and acquire properties that cannot be explained by how their parts function. We cannot plan them from outside, because planning requires analytical tools and designs that overlook such emerging properties. Instead, we need to trust agents and let them take decisions from which relevant properties will emerge. But we can set the frame within which these decisions can be taken (incentives, taxes). See complexity science.


BAHUCHET (S.), coordinateur, 1993.- Situation des populations indigènes des forêts denses humides. LACITO-CNRS/Centre d'anthropologie culturelle Université Libre de Bruxelles, rapport pour la Commission des Communautés Européennes, 460 p. ronéo, 1 atlas de 15 cartes.