Livestock Research for Rural Development 14 (2) 2002

http://www.cipav.org.co/lrrd/lrrd14/2/sant142.htm

Agroforestry for resource-poor farmers: a deeper understanding from the Brazilian Savannah

Pilar Santacoloma

Agricultural Support Division –AGS,
Food and Agriculture Organization of the United Nations (FAO), Rome, Italy
Pilar.Santacoloma@fao.org

Abstract

This paper presents an analytical approach of a socio-economic process occurring in smallholder farming systems in the south-west of the Brazilian Savannah and proposes agroforestry as an alternative strategy to increase farmers’ livelihood in a sustainable manner. The theoretical framework of this approach is based on the assessment of linkages between ecosystem attributes and resource management strategies at the farm-household level.

The results suggest that intensification and extensification are no longer suitable for the system under the current technological and resource conditions. Agroforestry might be an appropriate alternative strategy for improving economic parameters while regenerating resource deterioration.

Key words: Ecosystem attributes, stability, productivity, resource management strategies, agroforestry, agricultural sustainability.

Introduction

Increasing demand for livestock foods offers countless opportunities for improving resource-poor farmers. Rapid urbanization, high income and population growth in developing countries are driving forces explaining this expansion (Delgado et al 1999). However, an increase of this demand is being supplied by large-scale industrialized and intensive farms. Resource-poor farmers are not able to meet properly this livestock market expansion unless sustainable-oriented technologies and policies are addressed.

In the Brazilian savannah, resource-poor farmers based on live stock face serious difficulties to sustain their family livelihoods. Significant poverty levels, emigration and resource degradation are critical (Cunha et al 1994). The most serious environmental constraint is pastureland degradation, which is calculated covers about 10.3 millions hectares, nearly 50 percent of total savannah pastures (Carvahlo 1998). There is strong evidence that agroforestry can be an alternative strategy to recover fundamental components and attributes of ecosystems while reducing poverty and food insecurity for thousands of smallholders in developing countries (Arnold 1999).

This paper attempts to provide a sound understanding of the socio-economic process occurring in livestock-based smallholder systems in the south-west Brazilian savannah in order to identify opportunities for proper technological and policy interventions. Firstly, aspects of sustainability of the system are analysed via assessing linkages between farmer’s management decisions and attributes of ecosystems. Secondly, agroforestry models are simulated to assess impacts on the family income level and farm resources allocation.

Methodology

The study area is located in the south-west Brazilian savannah whose dualistic development pattern is typical in the whole region: mixed intensive mechanized agriculture for export-oriented commodities and traditional low-input agriculture practised by resource-poor farmers.

The study is mainly based on farm-household data obtained from a survey in 1997. The random sample includes all resource-poor family farmers in a micro-watershed (n=35). Statistical, geographical and historical information complements the socio-economic data. A micro-economic approach is used to estimate interrelations between selected ecosystems attributes and resource management strategies. Multiple regression models and Pearson’s correlation are the tools used to assess the nature of these relationships. Comparative-static linear programming models are developed for assessing impact. A basic model describing the current situation is compared with possible future scenarios resulting from the adoption of agroforestry.

Indicators

Stability, efficiency and productivity are key properties for sustainable agro-ecosystems (Conway and Barbier 1990). Stability is an attribute of ecosystems defined as the capacity of an ecosystem to renew itself and to regenerate in front of disturbances in its natural evolution (Redclift and Sage 1994). In productive terms it refers, thus, to the ability of an ecosystem to maintain productivity through time in face of long-term ecological constraints and social pressures (Altieri 1995). Normally, these constraints are relatively continuous and expected and therefore farmers can face them by introducing measures to counter their effects [A broad discussion regarding the properties of ecological systems stability and resilience can be found in Holling (1973). While resilience determines the persistence of relationships within a system to absorb changes of state, stability refers to the ability of a system to return to an equilibrium state after a temporary disturbance]. The capability response of eco-systems will depend on their intrinsic characteristics as well as on the nature of the stress and shocks.

It is believed that, particularly in the developing word, the disappointing adoption of management technologies has been the major obstacle for achieving stable development (Laming and Ashby 1992). Especially in areas with low resource endowment, better management of farm production systems enable farmers to minimize risks and withstand stress and shocks (Altieri 1995). In this paper it is assumed that farm management decisions have been strongly influenced by prevalence of instability. The system stability could be measured indirectly through instability sources-management decision relationships. Pearson’s correlation is used to assess the nature of these associations.

Another ecosystem property explored here is resources productivity. A profitable agro-ecosystem is indispensable to increases physical and social well-being of the farmers (Barbier 1987). This indicator is assumed as the relationship of total revenues minus variable costs per unit of resources use. Influences of resources management decisions on resources productivity are tested by using econometric models.

Instability sources and resources management strategies

Farm-family decisions expressed as management strategies are taken in response to particular cultural, socio-economic and physical factors. In this system, pastures account for approximately 80 percent of land used (65ha) indicating the livestock specialization. Improved pastures take into account the area in introduced pastures and occupy less than a third of the total grassland. Cropping intensity is quite low and takes into account the total cultivated crops on the total cultivated area in the given year. The crops cultivated are maize, rice and cassava and occupy nearly 8 percent of the farmland. Mechanisation, defined as ploughing practices on small plots using heavy rented tractors, is practiced by most of the farmers. Feeding livestock area includes harvested grass and sugarcane production. This practice although practiced by most of the farmers occupies very little of the farmland. Native and forest vegetation remains only as some patch of gallery forest around the streams. Only a third of the farmers maintain some land as fallow. Manure is used to fertilise coffee or fruit trees on backyards (see table 1).

Table 1: Resources management strategies

Management strategies and land-use

Coverage

% of farmers

Mechanisation (ha)

Improved pastures (ha)

Cropping intensity (index)

Feeding livestock area (ha)

Food self-sufficiency (index)

Native pasture (ha)

Forest (ha)

Fallow (years)

Manure (ha)

2.98

16.0

0.13

3.12

0.59

36.8

6.20

4 -5

0.34

Almost all farm-families produce for food self-subsistence (food produced/total food consumed in monetary terms) covering nearly two third of their needs. The relevant sources of instability identified by the farmers were ranked as of high or moderate importance. In the former group, land and capital scarcity are socio-economic constraints while low soil fertility is identified as an environmental constraint. This fact gives a picture of the farmer’s perception on their own resources restrictions. Problems like inefficient extension services, high production costs and debts ranked in the moderate importance group. Pasture degradation although classified as moderate acquires relevant dimension when associated with problems like deficient herd nutrition and illness. All the constraints represent aspects of the system unsuitability under the current technological and resources conditions (table 2).

Table 2: Sources of instability grouped by importance and scale levels
Regional and sub-regional level	Climatic fluctuations
	Inefficient marketing Inefficient extension services
	High production costs Debts
Farm-household level	Land scarcity-Capital scarcity
Crops activities level	Low soil fertility
Livestock activities level	Pastures degradation-Livestock under-nutrition
*Elements in “bold” are highly important;* others are of moderate importance

Table 3 shows the significant correlation between management strategies and sources of instability. Land scarcity was associated positively to mechanisation, feeding area, crop intensity and food self-sufficiency. The first three associations could be explained as logical, given that intensification practices are world-wide response to land shortages. The positive association of cropping intensity to cash scarcity and inefficient marketing reinforces this trend. The feeding practices, which are land and labour intensive, are also positive correlated to market constraints.

The positive relationship of food self-sufficiency and land scarcity contrasts with processes happening in other savannah regions, where farming intensification is reducing food production for self-consumption (Dos Santos et al 1998; Muriel and Sebastian 1998). Food self-sufficiency is also positively related to low soil fertility, water scarcity and inefficient marketing. These relationships could be interpreted as a farmer insurance perception against expected food shortages. Forest area is related positively to market constraints. It might be interpreted that farmers are discouraged to cut forest and extend agricultural land in face of less market opportunities. The more general conclusion of this section is the farming intensification in response to resources shortages and deterioration.

Table 3: Relationships between resources management strategies and sources of instability
	Resources management strategy
Sources of instability ¯	Mechanisation	Food self-sufficiency	Cropping intensity	Feeding area	Manure	% Forest	% Native pastures
Land scarcity	0.35 *	0.33 *	0.29 *	0.40 *
Inefficient marketing		0.29 *		0.41 *		0.50 *	-0.25 **
Liquidity scarcity			0.44 *
Debts			0.50 *	0.39 *	0.40 **
Low soil fertility					0.43 *	-0.27 **	-0.22 **
, * Significance levels at 1% and 5%, respectively

Resource productivity and resources management strategies

Productivity is considered the working capital, whereas labour productivity involves family and hired labour. Resources productivity presents a high heterogeneity in the farmer sample, with the higher variation in labour use (table 4). Some explanations could be found in factors like education level, family size and composition and hired labour availability (Santacoloma 2000).

Table 4: Resources use productivity
Productivity of resources	Mean (SD)
Capital (Reais/capital use)	0.627 (0.405)
Labour (Reais/ME)	38,801 (37,532)
Land (Reais/ha)	1153 (921)
ME: man-equivalent

Multiple regression models were used to explain how resource management strategies affect the systems performance (table 5). The stepwise method was applied to select the variables that fully establish conditions of significance. From the set of resource management strategies only a few can significantly explain the resource productivity. From the results, mechanisation, which is a strategy oriented to intensify resources use, has a negative impact on capital and land productivity. Otherwise, forestland use, which favours fallow use, and consequently land extensification, has also negative impact on labour and land resources productivity. This apparent contradiction might indicate that the system is unsustainable given the resources and technologies available.

Some possible explanations are affordable when details are examined. Firstly, mechanisation is relying on heavy rented machinery. Most probably this technology is not only unsuitable for the ecosystem conditions, but it is also an inappropriate choice in economic terms. It is also possible to argue that due to the small proportion of forest land (table 1) this no longer represents an alternative for soil regeneration with economic implications.

Table 5. Multiple regression models
Performance of resources/selected variables	Standardised coefficient	Significance of “t”	F value	Sig. of F	Adjusted R²
Capital
Mechanisation	-0.223	0.015
Food self-sufficiency	-0.298	0.04	3.63	0.024	36%
Native pastures %	0.464	0.015
Labour
Forest %	-0.353	0.039	4.61	0.03	26 %
Native pastures %	0.141	0.099	4.61	0.03	26 %
Land
Mechanisation	-0.225	0.06	2.69	0.091	21%
Forest %	-0.271	0.07	2.69	0.091	21%

Positive impacts of native pastures on capital and labour productivity could be associated with less land deterioration being evident in the economic indicators. Improved pastures are normally adopted when native pastures become degraded. So, native pastures represent a positive indicator of sustainability. This indicator should be not only its proportion within the land used, but also its species composition (Fujisaka 1997). Negative impacts of food self-sufficiency on capital productivity suggest negative trade-offs between social and economic goals in the system. From the results the need is evident to promote measures in order to recover the resource productivity and by doing so to increase farmers’ livelihood opportunities. One such measure is agroforestry.

Modelling the impact of integrating agroforestry into the farming system

Several benefits of integrating trees into agricultural systems have been identified. Growing trees on the farm helps to maintain agricultural productivity, enhance the supply of market products, and contributes to risk reduction and management (Arnold 1999). Impacts on productivity can be direct or indirect. Indirect benefits are related to those affecting nutrient recycling, soil and water. They include shade, reduction of wind velocity, improved moisture retention, reduction of erosion, suppression of weeds, and provision of nutrients through adding organic matter to the soil (Sanchez et al 1999). Maintenance of supplementary farm outputs, savings in fertiliser purchases and additional biomass for livestock feeding enhance direct productivity. In many cases, tree species can play a role as an intermediate input improving the income within the system besides the recognised ecological services they create. Regarding risk factors, the farmers are aware that trees are more resistant to environmental fluctuations and farming investment losses than with crops alone. In places strongly eroded, sowing trees may enable poor farmers to increase the amount of land they are able to work.

Model assumptions

The study uses a linear programming model for static comparative analysis. A basic model describing the current situation is set up and validated. The basic model seeks to reflect the defined reality using average parameters of the relation between input-output, the level of activities and restrictions in the farm-families sample (Hazell and Norton 1986). The objective function is to maximize family income subject to a set of resources and activities. Once validated, the model is compared with possible scenarios resulting from the application of the agroforestry strategies. The static linear programming model in this analysis can be mathematically represented as follows:

Max. Z = åP_iX_i - C_iX_i

Subject to åX_ia_ij =b_j, all_j= 1 to m

and X_i = 0 all_i= 1 to n

Where:

Z = objective function; X_i = level of activities, household and external relations; P_i = price of the output or service activity; C_i= cost per unit of the_i input; a_ij = represents the functional constraints and are called input-output coefficient; and b_j = the amount of _j resource available.

Two scenarios are depicted as following:

Scenario A1: Integrating nitrogen-fixing species into the fodder crops sub-system. In order to simplify the model, and considering agronomic properties, information related to Cratylia argentea is used. The main assumptions are based on results of experiments in Brazil (Xavier et. al 1990 cited by Argel and Lascano 1998; Carvalho 1999). Cratylia argentea, a native legume shrub from South-America, is tolerant to poor and acidic soils and resistant to droughts for long periods. The biomass production, when sowing at high densities, may rise to 14.5 tonnes/ha/year. Important nutritional characteristics make it an excellent source of protein for livestock. This specie is relatively simple to establish under suitable conditions, requiring relative low demands in labour. The specific assumptions of these scenarios are as follows:

§ The production of Cratylia argentea per shrub would be 430 g DM/ha/year (leaves and soft stems), equivalent to 2.58 ton DM/ha/year at a plant density of 6,000 trees/ha.

§ Replacements of concentrate as supplementary feeding for livestock by the leguminous fodder, which is rich in crude protein and is highly digestible (Argel and Lascano 1998).

§ Increase in productivity of sugar cane and cutter grass by 15 percent due to improvements in the chemical traits of the soil (Isichei and Moughalu 1992 cited by Carvahlo 1998).

§ Partial replacement of maize as a rich-energy source of fodder for livestock, 40 percent by sugar cane and cutter grass.

§ Increase in milk production by 10 percent. It is a very conservative assumption considering that increases of more than 50 percent have been reported (Argel and Lascano 1998).

§ Cost of sowing and maintenance of the Cratylia argentea in the first year: 120 Reais/ha.

§ Increase of labour use is 35 person/days. This includes labour for sowing and cutting activities.

Scenario A2: Besides the assumptions for Scenario A1, a mixture of nitrogen-fixing trees within the pastures in one hectare is added. The specific assumptions are:

§ Increase of grass productivity of 20 percent due to impact of trees on soil features.

§ Increase in milk production of 20 percent in relation to the basic model due to synergetic effects of rich energy and rich protein sources of fodder.

§ Decrease in the use of maize for fodder up to 80 percent regarding the basic model and replacement with alternative sources of energy.

§ Additional costs of establishment of the pasture and sowing of trees.

Results of the impacts of agroforestry on the economic results and resources use

Both the scenarios of agroforestry and silvo-pastoralism have significant positive impacts on the family income level. The introduction of agroforestry, Scenario A1 alone, would increase the family income by 4 percent. If, in addition, silvo-pastoralism is implemented, the increase in family income would be 11 percent (Figure 1). The impact on the cash balance would be positive and stronger than on the family income: the increase of cash availability is 28.2 percent for Scenario A1 and 98.9 percent for Scenario A2. Increases in cash availability would result from increases in cash inflows derived by higher sales of milk and maize, as well as by the decrease in feeding costs due to the elimination of concentrates.

Figure 1: Impact of agroforestry strategies into the Smallholder System

The levels achieved are very significant, although they hardly overcome the negative cash balance in the system. That is to say, in an optimal allocation of resources, the system would have positive economic results with strategies that are also environmentally friendly. However, additional incentives would be required to support this technological change during the transition phase. Both the knowledge regarding the better management of own resources and the financial resources for investments in capital naturally involve costs and risks. Institutional support to cover learning costs and investment costs are needed to encourage farmers to adopt this technology, which favours their short, medium and long term interests.

The most important impact on resource use for the crop sub-system is the increase in maize production. It might be that liberating a part of the production of maize for fodder could stimulate its production for sales. This increase would be significantly higher in Scenario A1 than in Scenario A2 (Table 6). Decreases in the live stock herd (-10%) and cassava production (-23%) might be interpreted as a result of competition for labour resources among crop, feeding and live stock sub-systems. However, the no-feasible solution for the use of permanent labour, due to its low coefficient values in the model, does not allow us to give a clearer picture about this conclusion. More significant result on the farm labour force is the increase of demand for seasonal labour, slightly higher in Scenario A1 than in Scenario A2. Apparently the integration of agroforestry would not have any significant impact in the allocation of labour force for off-farm activities. Another significant and positive impact is on rented land with increases to the upper limit in both tested scenarios.

In a general sense, the stronger impact of the tested strategy on the resource allocation would be in Scenario A1, although better economic profit would be obtained with Scenario A2.

Table 6: Impacts of integrating agroforestry into the system on the use of resources
	Basic model	Scenario A1	Scenario A2
Cash crops, ha
Maize	1.16	1.85
Cassava	0.33	0.25
Feed crops, ha
Sugar cane	1.08	1.08	1.08
Cuttergrass	0.82	0.82	0.82
Forage trees	0	0	1
Live stock, units	20	18	18
Rented land, ha	3.06	5.0	5.0
Labour, units
Permanent	0.52	nf	nf
Seasonal	18.8	25.1	23.5
Off-farm	124	124	124
nf: no feasible solution

Conclusions

The systems practiced by resource-poor farmers in the south-west Brazilian savannah are no longer sustainable under the current technological and resource conditions. Alternative strategies need to be addressed. In scenarios of optimal allocation of resources, the integration of agroforestry into these systems would deliver positive socio-economic impact. Its implementation would require additional financial and labour investments from the farmers. Specific training, extension and research on this subject, as well as financial support, are complementary measures needed to successfully implement this strategy.

Acknowledgements

Comments from John Dixon (FAO-AGSP) to a previous draft are very much appreciated.

References

Altieri M 1995 Agroecology. The science of sustainable agriculture. 2nd edition. Westview Press

Argel P and Lascano C E 1998 Cratylia argentea: una nueva leguminosa arbustiva para suelos ácidos en zones subhúmedas tropicales. In: Tercera Conferencia Electrónica de la FAO sobre Agroforestería para la producción animal en Latinoamerica

Arnold J E M 1999 Economic factors in farmers adoption of forest product activities. http://www.fao.org/docrep/w3735e/w3735e19.htm

Barbier E 1987 The concept of sustainable economic development. In.: Environmental Conservation. Vol. 14 No 2. Pp 101-110

Botero R and Russo R 1998 Utilización de arboles y arbustos fijadores de nitrógeno in sistemas sostenibles de producción animal en suelos ácidos tropicales. In:Tercera Conferencia Electrónica de la FAO sobre Agroforestería para la producción animal en Latinoamérica

Carvahlo M M 1998 Recuperação de pastagens degradadas em areas de relevo acidentado. In: Recuperação de areas degradadas. Sociedade Brasileira de Recuperação de areas degradadas. Pp 149-161

Carvahlo M M, Freitas V and Xavier D 1999 Comportamento de cinco leguminosas arbóreas exóticas em pastagem formada em solo de baixa fertilidade. In: Revista Arvore Vol 23 No 2 Pp 187-192

Conway G R and Barbier E 1990 After the Green Revolution. Sustainable agriculture for development. Earthscan Publications Ltda

Cunha A, Mueller C, de Andrade E and da Silva J E 1994 Uma avaliação da sustentabilidade da agricultura nos Cerrados. In: Relatórios de Pesquisas. IPEA- Instituto de Pesquisa Económica Aplicada. Projeto PNUD/BRA/91/1014

Delgado C, Rosegrant M, Steinfeld H, Ehni S and Courbois C 1999 Livestock 2020: the next food revolution. 20020 Brief 61. http://www.ifpri.org/2020/briefs/number21.htm

Dos Santos N A, Lothoré A and Ayres M 1998 Renda na agricultura familiar. Um estudo de caso no municipio de Silv(nia-Go. (Manuscript) Projecto de P&D Embrapa Cerrados

Fujisaja S 1997 Personal communication, CIAT, Cali, Colombia

Hazell Peter B R and Norton Roger D 1986 Mathematical programming for economic analysis in agriculture. Macmillan, New York

Holling C S 1973 Resilience and stability of ecological systems. Annual Review of Ecology and Systematics. Vol 4 pp 1-23

Laming D L and Ashby J A 1992 Adoption of natural resource management practices by resource-poor farmers, in Zobish, Acceptance of soil and water conservation strategies and technologies. Witzenhausen, DITSL.

Muriel A-F and Sébastien B 1998 A intensificaçon leiteira no municipio de Silvânia-GO: as implicaçoes en termos de sistemas forrageiros e sistemas de produço. (Manuscript) CIRAD.EMBRAPA CPAC

Redclift M and Sage C 1994 Strategies for sustainable development. Local agendas for the southern hemisphere. John Wiley & Sons, New York

Sanchez P, Izac A-M and Scott B 1999 Delivering on the promise of agroforestry. http://www.worldbank.org/cgiar/newsletter/sept99/agforest.htm

Santacoloma P 2000 An Ecologically and Economically Balanced Development in the Brazilian Savannah : The Case of Irai de Minas. Farming Systems and Resources Economics in the Tropics No 36. Hohenheim University, Dissertation. Wissenchaftsverlag Vauk Kiel

Solarte A 1997 Sustainable Livestock Systems based on Local Resources: CIPAVs Experiences. In: Second FAO Electronic Conference on Livestock Feed Resources within Integrated Farming Systems. September 1996-February 1997

Received 6 June 2001
(the editors apologise for the delay in reviewing and editing this paper for which they are solely responsible)

Go to top