7

LATIN AMERICA AND CARIBBEAN

REGIONAL SETTING

Characteristics of the region

The Latin America and Caribbean region covers some 205 million ha and encompasses 42 countries with a total estimated population in 2000 of 505 million. The size of the region, its wide range of favourable ecologies, and its low average population density of 0.25 persons per ha, combined with an urbanisation rate of 75 percent, have led to an extremely high level of biodiversity. According to the United Nations Environmental Programme (UNEP)1, it contains five of the ten richest countries in the world in terms of biodiversity and 36 percent of the main cultivated food and industrial species. The world's largest unfragmented tropical forest is found in the Amazon basin, and the region possesses 28 percent of the world's forest area - nearly one billion ha in 1994.

This abundance stems in part from the favourable climatic conditions of the region. It possesses 40 percent of the developing world's humid areas, and almost half of its total renewable water resources, but only four percent of the arid and semiarid lands. Some 90 percent of the region's land area is humid and subhumid.

The region contained some 160 million ha of cultivated land2 in 1999, including 18 million ha equipped for irrigation. This still represents only 18 percent of the estimated potential of the region3 and it is estimated that no more than one percent of the available water is currently utilised4. A further 600 million ha is under grazing and pastureland. The region is globally important in a number of crops and often achieves yields significantly above the developing world average.

With an average GNP per capita of US$3940 in 1998, it is the wealthiest of the developing regions and also the least dependent on agriculture - only eight percent of GDP was derived from the sector in 1998. As growth in agricultural value added is lower than for industry or services, that share is likely to continue declining. FAO nutritional data indicate that the average diet contains 120 percent of the required minimum daily allowance5, while the Economic Commission for Latin America and Caribbean (CEPAL) per capita food index for the region rose by 15 percent over the period 1980-19976.

Nevertheless, serious problems of equity exist. Not only do the wealthy control one of the highest proportions of resources of any region in the world7, but there is also a strong urban bias. According to 1997 estimates by CEPAL8, 54 percent of rural households were classified as poor, against only 30 percent from urban areas. Extreme poverty9 affected 31 percent of rural households but only 10 percent of urban ones. In total, 47 million rural inhabitants were classified as being in extreme poverty, and a further 78 million are in poverty. Internationally comparable poverty data vary extensively - from fewer than two percent of the population with an income of under US$1 per day in Uruguay (1989 data), to 40 percent in Guatemala and Honduras10. Equity problems are particularly evident in respect of land distribution11.

Major farming systems in South Asia and Caribbean - MAP

Major farming systems in Latin America and Caribbean

Due to its enormous latitudinal range, varied topography and rich biodiversity, Latin America and Caribbean has one of the most diverse and complex range of farming systems of any region in the world. Sixteen major systems have been defined for the purposes of this book and are briefly described below. These farming systems are summarised in Table 7.1 and all but the Urban Based System are presented graphically in the accompanying Map.

Irrigated Farming System

This farming system encompasses enormous areas of arid lands across Northern and Central Mexico and coastal and inland valley areas of Peru, Chile and Western Argentina. The total land area of almost 200 million ha contains only 7.5 million ha of cultivated land, but almost all is irrigated. Irrigated production occurs in many of the other farming systems as well, but always in a minority role. Nevertheless, the Irrigated Farming System accounts for only 40 percent of total irrigated area in the region (18.5 million ha). The presence of irrigation infrastructure allows a relatively high degree of intensification of production - generally commercially oriented - and supports an agricultural population of nearly 11 million. Key products within this system include rice, cotton, fruit, horticulture and vines. Poverty is low to moderate.

Forest Based Farming System

Centred on the Amazon basin and covering over 600 million ha, or 30 percent of the total region, this system comprises scattered indigenous and low-input settler agricultural activity, interspersed with extensive beef and occasional plantation farming - especially towards the margin of the area. Cultivated area is little more than one percent of the total, with negligible irrigation. Population density is very low, with the agricultural population of around 11 million, representing less than 0.02 persons per ha. Poverty is generally low to moderate.

Coastal Plantation and Mixed Farming System

This system covers 186 million ha, and has an estimated agricultural population of 20 million. There are 20 million ha of cultivated land of which 13 percent is irrigated. The system occupies some of the richest agricultural lands in the region, but also includes mangrove swamps and isolated areas of tropical forest. There are two major sub-systems: (a) small-scale family farms with mixed agriculture, in-shore fishing and frequent off-farm employment (e.g. tourism); and (b) large-scale plantations, typically export-oriented and often internationally owned, with intensive production and significant poverty among labourers. Otherwise, poverty is not prevalent.

Intensive Mixed Farming System

Centred on Eastern and Central Brazil, this intensive mixed agricultural system represents the heartland of Brazilian agriculture, and occupies an estimated 81 million ha with an agricultural population of almost 10 million. There are approximately 13 million ha of cultivated land, of which about eight percent is irrigated. Coffee, horticulture and fruit are important products. Poverty levels are relatively low in this system.

Cereal-Livestock (Campos) Farming System

The Campos represent a gradation in moisture, and often soil quality, from the intensive system described above. Covering just over 100 million ha in Southern Brazil and Northern Uruguay, the system has an estimated rural population of about seven million, and is strongly oriented to livestock and rice production. There are an estimated 18 million ha of cultivated land, of which 10 percent is irrigated. Poverty is low to moderate.

Moist Temperate Mixed-Forest Farming System

This system comprises one of the few temperate farming systems within the region and is strongly reminiscent of New Zealand in its topography and climate. It is a small system, comprising only 13 million ha, restricted to the coastal zone of Central Chile. The system has an agricultural population of little more than one million and is characterised by extensive natural and plantation forest (over one million ha) interspersed with dairy, sheep and some crops, such as sugar beet, wheat and barley. Cultivated area is only 1.6 million ha, with negligible irrigation. The agricultural population of the system is estimated at slightly over one million. Poverty is generally low.

Maize-Beans (Mesoamerican) Farming System

Stretching from Central Mexico to the Panama Canal and with an estimated agricultural population of about 11 million - including a substantial indigenous population - this system covers 65 million ha and is historically and culturally based upon the production of maize and beans for subsistence. Although there are 2.4 million ha of irrigation within the system (40 percent; the highest concentration of irrigation outside the irrigated farming system), the historical loss of the better valley lands to non-indigenous settlers and commercial operations has led to extensive and severe poverty and serious land degradation in many areas.

Intensive Highlands Mixed (Northern Andes) Farming System

Covering 43 million ha and with an agricultural population of four million, this system contains two distinct sub-systems, generally differentiated by altitude: (i) the well-developed intermontane valleys and lower slopes - the heartland of Andean coffee and horticultural production; and (ii) the highlands and upper valleys where temperate crops, maize and pigs predominate and where the traditional indigenous culture is strongly established. Total cultivated area is estimated at 4.4 million ha and some 20 percent is irrigated. Poverty is generally moderate in the lower areas, but is extensive, and often severe, at higher altitudes.

Extensive Mixed (Cerrados and Llanos) Farming System

Covering the enormous wooded and open savannah areas of Central-Western Brazil and Eastern Colombia, Venezuela and Guyana, this system encompasses 23 million ha and has an agricultural population of around 10 million. Less than 15 percent is cultivated (31 million ha), and irrigation is almost entirely absent. Only recently starting to become intensively developed, this frontier system offers enormous potential for future agricultural growth in livestock, cereals and soya, among other crops. Poverty is relatively low, although higher among landless immigrants.

Temperate Mixed (Pampas) Farming System

Covering 100 million ha in Central and Eastern Argentina and Uruguay, this system was originally largely devoted to livestock but now contains nearly 20 million ha of cultivated land, although there is only negligible irrigation. The impetus for growth has come from demand for such crops as wheat, soybean and sunflower as well as horticultural production for Buenos Aires and Montevideo. The agricultural population is now estimated at almost seven million, and further intensification of production is expected within the system. Poverty is generally low.

Dryland Mixed Farming System

Due to its location near the coast of Northeast Brazil and in the Yucatan peninsula of Mexico, this large system of nearly 130 million ha has a well-established economic and productive structure and an agricultural population of about 10 million, but faces severe moisture and soil quality constraints. Despite frequent droughts, little more than two percent of the 18 million ha of cultivated land is irrigated. It is a system with extensive and severe poverty among small-scale producers, who exist alongside large-scale extensive ranches, and often depend on seasonal migration and wage labour for survival. Land degradation is a serious problem.

Extensive Dryland Mixed (Gran Chaco) Farming System:

Stretching from North-central Argentina, through Paraguay and into Eastern Bolivia, this system of 70 million ha has only recently been economically developed and still has a rural population of less than two million. Total cultivated area is estimated at under eight million ha, and irrigation is negligible. Unlike the Cerrados and Llanos areas, the growth potential of the Gran Chaco is severely limited by soils and moisture. Extensive poverty is found among the small colonists.

High Altitude Mixed (Central Andes) Farming System

Again divided into two distinct sub-systems, the Central Andean system covers 120 million ha and has a total agricultural population of over seven million. Through most of Peru the system occupies the steep valleys of the high Sierra, while from Southern Peru through Western Bolivia into Northern Chile and Argentina, the altiplano is the predominant landform. Throughout the zone the key characteristics are production at an altitude of more than 3200 m, a dependence on indigenous grains, potatoes, sheep and llamas, and a very strong indigenous culture. Where altitude and moisture permit, the same temperate crops are cultivated as in the Northern Andes. More than a third of the total cultivated area of 3.1 million ha is irrigated. Poverty is extensive and often very severe in this system.

Pastoral Farming System

As the Pampas extend southwards, they become drier and cooler, merging eventually into the very sparsely populated plains of Patagonia, covering some 67millionha, where sheep and cattle ranching is the only widespread agricultural activity. Cultivated area is negligible, and there is no reported irrigation in the system. Poverty is low to moderate among the agricultural population of less than one million.

Sparse (Forest) Farming System

At the southern end of the Andes, lower temperatures combined with continued high altitudes, render cultivation generally sub-marginal. The agricultural population which number no more than a quarter of a million (almost 150 ha per person) is largely dependent upon livestock grazing, forestry and tourism for income and cultivate less than 0.5 percent of the land area. Poverty is low to moderate, reflecting the low population densities.

Urban Based Farming System

In common with all other regions of the world, specific peri-urban and intra-urban agricultural systems have developed to serve major conurbations and population centers throughout the region. Focusing on perishable products with high levels of demand but only limited space requirements, these urban-based systems typically include horticulture, poultry and dairy, but off-farm income is usually also integral to the family unit, with many members engaged in agriculture on a part-time basis. An estimated three percent of the agricultural population lives within this system.

Region-wide trends in Latin America and Caribbean

The following section summarises regional trends with reference where appropriate to the position of the region vis a vis all developing countries12. After outlining selected projections with regard to population, hunger and poverty, the following section lists some key trends affecting farming systems in the areas of: (i) natural resources and climate; (ii) science and technology; (iii) trade liberalisation and markets; (iv) policies, institutions and public goods; and (v) information and human capital.

Population, hunger and poverty

During 2000-2030, population is expected to increase 40 percent to reach 725 million13. This is lower than the overall 47 percent rate projected for developing countries as a whole, but higher, for example, than East Asia. In fact, the rate of regional population growth has declined dramatically in the last 40 years; from 2.8 percent per annum in the 1960s to about 1.6 percent in the 1990s. The proportion of total population living in rural areas14 is projected to decline from 25 to 17 percent over the next 30 years, leaving rural populations marginally lower than at present (from 128 to 121 million); but significant sub-regional differences are anticipated. The poorer countries are expected to maintain high rates of overall population growth, resulting in an absolute increase in rural populations in such areas as Central America, Bolivia, Paraguay and Haiti. On the other hand, countries such as Argentina and Brazil are likely to experience declines in rural population of 20 percent or more. In general, those countries with projected overall population increases of 50 percent or more to 2030 will see increasing rural populations.

During 2000-2030, the average per capita daily nutrient intake within the region is expected to increase by 10 percent from 2791 to 3080 calories, which will maintain average intake in LAC above the developing world average to 2030. This increase in calorie intake is expected to derive principally from meat and vegetable oils (33 percent each) and dairy (18 percent). Roots and tuber consumption is expected to decline. The number of people suffering from undernourishment - currently 53 million - is projected to decline to 32 million by 2030. This represents a drop from 11percent of the population to 5 percent, but is only half the current international target.

Natural resources and climate

Cultivated land has expanded by 47 percent since 1961, but cropping intensity increased only one percent during this period. During 2000-2030, it is projected to expand a further 20 percent (depending on the evolution of farming systems); one third of the 40-year historical trend rate of 1.76 percent per annum. However, this rate may be underestimated, given the enormous potential for agricultural expansion in the Cerrados, Llanos, Chaco and Amazon basin15.

During 2000-2030, the irrigated area is expected to increase from 18 million to 22 million ha, but remain constant in relative terms at 14 percent16 of cultivated land. Irrigation efficiency is low, and only 8.5 million ha of the installed area is thought to be in use. Surface irrigation accounts for almost 90 percent of all irrigated areas. During the period 2000-2030, only minor increases in water use and efficiency are expected.

Over the ten year period 1982/1984-1992/1994, the area under pasture and grazing land increased by a total of three percent in South America and 6.2 percent in Central America to reach 600 million ha17. Pasture lands in Guatemala increased by an astonishing 65 percent (albeit from a small base) to 2.6 million ha. At the end of the 1980s, the deforestation rate was estimated at 7.4 million ha per annum, equivalent to 0.8 percent per annum18. This rate has since appeared to decline in South America, but accelerated in Central America - to 1.3 percent per annum.

Recent experience has shown how vulnerable considerable parts of the region are to climatic variations; including hurricanes in Central America, flooding and loss of fisheries in Pacific South America and drought in Northeast Brazil. Climatic changes discussed in Chapter 1 are likely to exacerbate these risks and even reduce potential yields. However, it is still impossible to predict specific impacts in different geographical regions with any degree of confidence, given that the influence of climatic change is by no means uniform.

Science and technology

In the last 20 years, the value of regional agricultural production has grown at 2.8 percent per annum. However, due to the slowdown in total population growth, and a relatively low income elasticity of demand, the growth in demand for food and raw materials has been declining in recent years and for the next 30 years it is estimated at no more than 2.4 percent per annum. Table 7.2 shows the historical rate of growth of the major crops of the region. The major cereals have all grown strongly in the last thirty years - almost entirely due to yield increases - and the region now accounts for more than one quarter of the developing world's output of maize. Cereal output is expected to continue to expand, albeit at a slower rate than in recent decades. Fruits and vegetables have also exhibited strong growth; the area dedicated to fruit has expanded faster than for any other crop category in this period.

The strongest performing crop category, however, has been the oilcrops, especially soya and sunflower. Production of oilcrops has increased from three million tons per annum in 1970 to almost 16 million tons in 2000 - a rate of growth of almost six percent per annum. Since 1961, more than three quarters of all developing country growth in soya production, and over 40 percent of increased world output, has originated in the region. Growth in output is expected to continue to 2030. Among the major crop categories, only fibres (principally cotton) have show an absolute decline - despite significant yield increases - but this trend is projected to reverse in the coming decades.

Overall crop production growth to 2030 is projected at 1.7 percent per annum Expected strong gains in cultivated area will be associated with a slow rise in cropping intensity - only an 11 percent increase to 2030. During 2000-2030, average crop yields are forecast to grow by less than 50 percent and reflects the greater expansion in cultivated area compared with other regions.

The 356 million cattle within the region constituted 26 percent of the developing world total and have increased by 1.6 percent per annum in the last three decades, faster than developing countries as a whole. In contrast, growth in other species has been slower. However, the rate of increase in all species has declined in the past decade. The cattle population is forecast to grow at 0.9 percent per annum to 2030, while the population of sheep and goats will expand by 0.7 percent per annum. Numbers of pigs and poultry are expected to grow by 0.9 and 1.6 percent per annum respectively.

From 1970 to 1990, agricultural labour productivity increased at about two percent per annum compared with developing country averages of between 3.5 and 4.5 percent19. The slow growth in labour productivity in the region only partly reflects the abundance of land. Growth rates have been particularly low in areas, such as the Andes and Central America, where a higher concentration of small farmers exists.

During the past decade, fertiliser consumption has expanded at the rate of 2.1 percent per annum, to reach 88 kg/ha of nutrient; close to the average rate for the developing world20. Fertiliser use to 2030 is projected to increase slowly - at about one percent per annum.

Trade liberalisation and market development

Average agricultural tariffs in 1995 (from 10 to 20 percent) were considerably lower than ten years previously (20 to 60 percent). Nevertheless, the position of the different countries and products varies widely. The region currently accounts for a significant portion of world trade in a number of commodities including: coffee (Brazil, Colombia, Central America); orange juice (Brazil); bananas (Ecuador, Honduras, Costa Rica); table grapes and contra-seasonal temperate fruits (Chile); vegetables (Mexico); cut flowers (Colombia, Ecuador); pineapples (Costa Rica, Guyana), and shrimp (Ecuador, Honduras). Strong growth is foreseen in products that are currently significant in industrialised countries; either due to rising land and labour costs (sugar, cotton, citrus juice, vegetables) or to perceived environmental costs (pork, mushrooms, and possibly chickens).

During 1995-1997, the region had an annual net trade deficit in cereals of 16 million tons and this is projected to double by 2030; a decline in self-sufficiency from 90 percent to 87 percent. However, it is the only developing region with a net positive livestock trade (874000 tons per annum) and livestock exports are expected to triple by 2030, in contrast to other developing regions. Current net imports of 6.3 million tons of dairy products are expected to grow by 2030, in line with population increases.

The broad shift to the free market in the last decades of the 20th century has freed many constraints on rural labour markets in Latin America (less so in the Caribbean). Consequently, the share of off-farm income in the household livelihoods of poor farmers has been increasing; a trend that is expected to continue to 2030. In more densely settled farming systems, household members work in the local area, whereas in remote or extremely poor areas farm people are often forced into seasonal migration in search of work.

The possibility of a reversal of trade liberalization exists, in which case countries would shift agricultural and food policies towards support for self-sufficiency, including high tariff barriers for basic grains and other staples. The opening up of agricultural frontier lands, which is in great measure a response to the liberalisation of markets, would slow down or even reverse. The shift away from production of basic staples - by small-scale producers in systems affected by poverty - would also be in doubt, since domestic prices for these products could be expected to rise. On the other hand, if trade liberalisation is accelerated, adjustment within farming systems would be even more difficult, poverty would likely increase rapidly and out-migration to urban areas would further accelerate. In fact, poverty figures suggest that this has already been the case since the mid-1990s. However, further trade liberalisation would boost the development of frontier lands and accelerate shifts in cropping patterns towards products still protected in industrial countries, such as sugar cane, cotton, orange juice and tobacco.

Policies, institutions and public goods

Nowhere in the developing world has the logic of structural adjustment and economic liberalisation been carried forward faster, and more profoundly, that in Latin America. Starting with Chile in the 1970s, the region has seen the privatization of large areas of national economies previously under government control; including banking, telecommunications, energy, transport and of particular importance here, agricultural marketing and finance. Ministries of agriculture in many countries - previously often major bureaucracies employing many thousands of staff - were ruthlessly cut back. Services to the sector, such as extension and research, were also significantly reduced. In theory, such reductions should have been met by an increase in the role of the private sector. However, only those products associated with major international commodities, such as bananas, coffee, and citrus, have found it relatively easy to attract private sector research and development. Crops of interest only to smaller producers, or without significant extra-regional demand, have found little private sector support, although several international agencies21 have maintained programmes related to basic crops - mainly cassava and potato.

It appears unlikely that the trend towards reduced public sector participation in agriculture will be reversed soon, as budgetary constraints limit the ability of governments to assume the enormous recurrent costs - and consequent fiscal deficits - that were once accepted. However, two factors may ease the situation in future. The first is the probable transition of many small producers towards export and market-oriented crops where more private sector interest exists. The second is the mounting evidence that scope exists for effective public/private partnerships in the provision of agricultural support services.

A second major trend that has emerged in the last decades of the 20th century is that of decentralization. Closely linked to the process of structural adjustment, decentralization has often been used primarily as a way of shedding fiscal responsibilities to local or regional levels of government. Nevertheless, the process of decentralization may yet have far reaching impacts on the pattern of rural development within the region, by transferring decision making to levels much closer to the rural poor and affecting the delivery and financing of services to producers. It will also require considerable strengthening of local government and community level capacity in planning, implementation and financial control.

Whether or not the trend towards decentralization will endure through the next three decades is difficult to predict. Real centralization implies a considerable reduction in power at central government level, and may be resisted by existing power structures when and if it is apparent that the process is more than cosmetic. Local elites may also capture a high proportion of the benefits, rendering decentralization ineffective.

Information and human capital

Latin America and Caribbean faces many of the same challenges with respect to information and human capital as other developing regions - perhaps more so, given the market-orientation of the region as a whole. Market requirements are becoming more exacting, technologies are changing more rapidly, and skills development is progressively more vital. Historically, a high proportion of resources within the region has been devoted to secondary and tertiary education, benefiting largely wealthier urban populations. By contrast, rural primary education, communications and information services have been weak or absent. However, in the last three decades there has been a significant improvement in educational attainment in many rural areas, and literacy has increased noticeably, although rural education budgets are still extremely limited.

The transition process being brought about by globalisation will penalise small farmers who are not prepared to respond, and this realization is beginning to influence educational and information priorities at national level. In fact, the privatisation of telecommunications has probably done more to expand information linkages into rural areas than any other policy decision taken in recent years. Privatized energy provision is also accelerating rural electrification, an essential precursor for many modern services. It is likely that wealthier farming systems, such as the Intensive Mixed Farming System in Brazil, will be among the first to benefit from new information technologies, but whether they will reach poor farming households in Guatemala, Haiti or Bolivia in the next few decades remains to be seen.

Selection of farming systems for analysis

Four of the systems delineated above have been selected for more detailed analysis - using selection criteria based largely upon the prevalence of poverty, population and growth potential - are described in more detail in the following sections of the Chapter.

Within the region there are three farming systems clearly associated with extensive and often severe poverty, namely:

• Dryland Mixed Farming System of Northeast Brazil and Yucatan;
• Maize-Beans Farming System of Mesoamerica; and
• High Altitude Mixed Farming System of the Central Andes.

An additional farming system was also selected, as it represents one of the major agricultural frontiers left in the developing world, namely the:

Extensive Mixed Farming System of the Cerrados and Llanos.

EXTENSIVE MIXED (CERRADOS AND LLANOS) FARMING SYSTEM

Characteristics of the system

The Extensive Mixed Farming System covers about 230 million ha, of which some 190 million ha, known as the Cerrados, are in Central Brazil, and a further 40 million ha - the Llanos - extend across parts of Guyana, Southern Venezuela and Eastern Colombia. The total system population of 24 million is approximately 40 per-cent agricultural22. The zone has a subhumid tropical climate (rainfall from 1000 to 2000 mm per annum) with a clearly defined dry season, although the Llanos tend to be more humid than the Cerrados. Natural vegetative cover ranges from open grassland through woody savannah to gallery woods along rivers.

Historically, the frontier savannah areas - with their isolation from the cities of the coast, acidic nutrient-poor soils, and lengthy dry seasons - were judged suitable only for extensive ranching. At the beginning of the 1970s only three percent was under cultivation. Since then, however, growth has been rapid, especially in the better soils of the Cerrados. Upland rice is often the first crop to be planted, and is still dominant in Llanos. In the Cerrados, however, soybeans and maize have now assumed greater importance. Total cultivated area within the system is currently estimated at over 30 million ha, including permanent crops. However, an estimated 40millionha of the Cerrados have been severely degraded by poor land management. These are predominantly areas established in the early years of Cerrado settlement, using upland rice and Brachiaria sown simultaneously.

Large farms (larger than 500 ha), often with absentee owners, have traditionally dominated the Cerrados and Llanos and still account for approximately ten percent of holdings. The vast majority - an estimated 70 percent of all production units in the Cerrados and probably higher in the Llanos - still have ranching as their primary activity. This accounts for 40 to 45 million ha of pasture and 60 million head of cattle across the farming system.

Despite the predominance of ranching, there are a growing number of large mixed and arable farms in the Cerrados zone, which comprise many of the 50 percent of holdings that fall within the 10 to 100 ha size range (see Box 7.2). These mid-size units tend to employ considerable seasonal labour. Already the Cerrados account for some 20 percent of the national output of beans - a major staple in Brazil - while coffee production is expanding in the East. Other significant contributions to national agricultural output from the Cerrados include 34 percent of soya production, 21 percent of maize and 21 percent of rice. The importance of agriculture is, however, much more restricted in the Llanos. Although still less than one percent (0.4 million ha), the use of irrigation is expanding.

The role of medium-sized farms appears to be growing in importance in both the Cerrados and the Llanos. Evidence from the Llanos indicates that in the period from 1961 to 1997, holdings over 1000 ha fell by more than a third, while the number of holdings under 20 ha also declined. Similarly, fewer than 10 percent of holdings in the Cerrados are under 10 ha, and there is evidence from some states that the number has declined since 1970, suggesting land consolidation is underway.

Studies of agricultural potential for the Cerrados estimate total land area suitable for cultivation in excess of 100 million ha. The Llanos, with more serious soil suitability problems, could provide perhaps an additional 10 to 15 million ha. This represents three times current cultivated land use. The rapid expansion of agriculture within the system has, however, caused considerable soil degradation and a reduction in native biodiversity.

The overall level of poverty in the system is much less severe than in the semiarid Dryland Mixed Farming System which borders the Cerrados. Two-thirds of the farms are owned, while only a small percentage are rented. The number of landless labourers emigrating into the Cerrados is increasing, however. Some degree of bankruptcy among farmers has also been recorded.

The historically low population density in the Extensive Mixed Farming System has meant that infrastructure development has been limited, and costs for transportation and storage have been relatively high in comparison to more settled agricultural areas. The high cost of inputs, low soil nutrient levels, and low land prices, is reflected in low land productivity. Yields of major crops tend to be lower than national averages (80 to 90 percent for soya and maize, but as low as 50 percent for rice in the Cerrados). Investment in education, training and other government services within the system has been poor.

Trends and issues in Extensive Mixed (Cerrados and Llanos) System

The current growth of the Extensive Mixed Farming System is expected to continue and even accelerate over the coming years, although growth rates will be closely linked to market access and demand for soya, beef, cereals and other crops. Total cultivated area can be expected to approach 70 million ha by 2030, or over 100 million ha including pasture. Livestock will no longer be the dominant activity. Land speculation may become a problem. Data from 1995 for the Llanos zone already indicate a 12 percent annual turnover of farm properties; considerably higher than the national average of 3.5 percent for Venezuela as a whole. The structure of farms within the system in the future will have a major impact on the nature of development. If ownership stays predominantly with the large haciendas, considerable poverty may develop within the system. If, however, land distribution becomes more equitable - as current trends suggest may be happening - this is unlikely. Intensification could also be expected to occur more quickly.

Considerable investment can be expected in transportation systems, storage facilities, social infrastructure, and processing - partly financed by the private sector. These changes are expected to reduce transaction costs, increase farm income, facilitate diversification and expand off-farm employment (attracting still more immigrants to the system). Given the current rate of population growth and migration, it is expected that the total population of the system will double in the next 30 years. However, agricultural populations will probably increase at a slightly slower rate.

Good potential exists for agricultural diversification into such activities as acid-soil tolerant fruits (pineapple and passion fruit), mango and avocado. Dry-season food crops such as pigeon pea and pearl millet are also expected to expand in the Cerrados. Both of these categories may be attractive, especially to smaller producers. However, as cropping intensities rise, increasing demands for irrigation water will require effective planning and if this is poorly managed, could lead to conflict.

The key trend for annual cropping in the Cerrados is likely to be the adoption of no-till cultivation methods. Already practised over several million ha, it is expected to grow rapidly in the future. The main advantages of this system include: (i) more timely sowing; (ii) improved yields; (iii) earlier harvesting that permits a second (cover) crop; (iv) conservation of soil organic matter; and (v) reduced production costs. However, not all experiences with no-till agriculture have been successful. Inappropriate herbicide application, lack of suitable technologies for smaller farmers, and poor training of extensionists have all caused problems.

Field level research undertaken by the Brazilian Agricultural Research Corporation (EMBRAPA), indicates that substantial gains in crop and livestock yields are technically feasible over the next thirty years if environmentally sustainable production approaches are widely adopted. Maize and field bean yields have the potential to increase by over 100 percent (to 5.5 t/ha and 3.2 t/ha respectively), rice by 75 percent (to 2.8 t/ha) and soybean by 40 percent (to 3.5 t/ha). Meat off-take from beef cattle would increase by a projected 300 percent, reaching 0.2 t/ha per annum. The projected increases would derive exclusively from improvements in production systems, not improved varieties, while the strong increases in annual livestock of-take per ha assume the recovery of large areas of degraded pasture land.

In summary, key issues facing the frontier savannah farming system over the next 30 years include:

• The distribution of farm sizes and ownership within the system, and the impact of that distribution on poverty and production intensification;
• The creation of a wide range of infrastructure to support the accelerating economic development of the zone;
• Achievement of further agricultural expansion without damaging the relatively fragile soils of the savannah zone; and
• The ability to reverse degradation of large areas of poorly managed pasture lands.

Priorities for Extensive Mixed (Cerrados and Llanos) System

The main strategies for poverty reduction in this system are intensification of production - coupled with improvements in infrastructure - and an increase in farm size among smaller producers. Secondary sources for poverty reduction are enterprise diversification and increased off-farm income. In order to respond to the challenges described above, a series of priorities are indicated. These include the development of mechanisms to promote the expansion of small to medium sized production units, together with strengthened technical assistance services to smaller producers, tax regimes favouring holdings under 100 ha, financing for land acquisition, and the promotion of co-operative and other forms of joint services for input and output marketing. New agro-industries and post-harvest operations (at family, co-operative and corporate level) should be facilitated in order to create new demand opportunities and off-farm employment.

It is important to validate and disseminate information on intensified integrated production systems and diversification options - especially for smaller farmers - including effective water use, no-tillage methods and integrated crop-livestock systems. To be effective, this requires improved communications with farmers and their participation in adaptive research and trials. Efforts should be focused on restoring the estimated 40 million ha of degraded pasture lands - larger than the current total cultivated area - in order to channel agricultural growth to existing areas and reduce pressure for clearance of new savannah lands.

Specific interventions can be expected to change significantly over the next 30 years. However, innovative technologies and approaches already exist for reversal of degraded pasture lands (see Box 7.3). While mainly tested on large holdings, these approaches are also relevant to smaller producers and have resulted in dramatic increases in sustainable stocking capacity.

As moisture availability in the system varies, different crop combinations prove optimal. These include many crops that are still little known in the savannah context, such as castor beans, pearl millet, grain amaranth, kenaf, pigeon pea and quinoa. The integrated production systems tested have reversed soil degradation and achieved significant yield increases. Field data indicate that beef production can increase 300 percent (on a per ha basis) across the system, while maize and beans outputs rise over 100 percent. Net average incomes per ha for participating farmers have increased from US$200 to US$350. While larger farmers may need little support for investment, the adoption of mixed cropping technologies among smaller farmers requires assistance. The agricultural potential of the system justifies increased field work in applied research, seed multiplication and integrated pest management systems. Potential may exist for utilisation of these technologies in the Llanos and also in the West African savannahs.

DRYLAND MIXED FARMING SYSTEM

Characteristics of the system

The Dryland Mixed Farming System includes two principal areas within Latin America: (i) the Northeast of Brazil, comprising about 20 percent of the country or 110 million ha; and(ii) the Yucatan peninsula of Mexico, extending into the Northern Peten zone of Guatemala, covering about 17 million ha. The agricultural popu-lation of the system is estimated at about 500000 in Yucatan and almost 11 million in Brazil (see Box 7.4). As much as half of the rural population of Brazil live within this system24. Annual precipitation in Northeast Brazil varies from 400 to 600 mm in the drier west of the Brazilian system, to 1000 mm in the East. In the Yucatan, the average range is 600 to 1500 mm. In both regions, soils are mainly shallow and stony with areas of low forest vegetation and elevations are low to moderate. Good quality agricultural land is scarce and there are few erosion control measures. About 18 million ha of the semiarid area are subject to severe desertification in Northeast Brazil.

The long dry season, frequent droughts and uncertain patterns of precipitation typical of the system, make farming a highly uncertain process for the vast majority of producers without access to supplemental irrigation. In both zones, more than 80 percent of farmers practise semi-subsistence production (see Box 7.5). Crop failure - especially in maize and rice - is common if rains are late. In Yucatan, shifting cultivation is a traditional subsistence practice of the Mayas that is still widely used today. This form of agroforestry is a succession of two components: the cultivation phase (milpa) where maize, beans and squash are grown together, and the resting phase (acahual) where wild brush and trees take over and replenish the soil. Cultivation traditionally lasts two or three years and fallow between five and twenty years, depending on soil, vegetation and land availability.

Land distribution is strongly bimodal. In Northeast Brazil, there are about two million farmers within the system, cultivating an estimated area of 15 million ha. However, more than half (59 percent) have holdings of less than five ha and account for only 6.1 percent of the total arable area (another 22 percent have holdings from 5-20 ha). At the other extreme, only 8.2 percent have holdings over 50 ha, but these account for 61 percent of all land25. Larger holdings are concentrated very heavily on maize (often for feed), sugar cane towards the coast, and livestock - with 49 percent of farmers registered as producing beef, 55 percent dairy and 40 percent poultry.

In the Yucatan, each ejidatario26 usually has between three and eight ha under cultivation with an average 4 to 4.5 ha. Total sown area in Yucatan in 1995 was 1.1 million ha. Just over half (58 percent) of Northeast Brazilian farmers are owners, although this is skewed towards the larger landholders. Sharecroppers and tenants account for a further 17 percent, and the remaining 25 percent are informal occupants.

Small producers within the farming system have become poorer in the last few decades. More than 50 percent of rural families in Northeast Brazil live in chronic and severe poverty, with an average family income of only US$366 per annum (compared with a national average of US$938, and US$1744 for the south)27. Farmers normally do not use improved seed, fertilisers, pest and disease control or mechanisation. Crop yields reflect this low level of technology and inputs. Average yields across the system are: maize (one t/ha); beans (0.45 t/ha); cassava (9.9 t/ha); and, rainfed rice (1.59 t/ha). Nevertheless, Northeast Brazil accounts for more than 30 percent of national production of beans and cassava28. In the 1996 Census, 20 million ha of native pasture and 12 million ha of improved pasture were recorded for the northeast, only about 15 percent of which was found on holdings of less than 50 ha. These pasture areas are thought to have increased significantly since the census.

Irrigated agriculture is poorly developed; partly as a result of shortage of water availability, partly due to poor soil conditions, and partly because of investment requirements beyond the reach of most small producers. In Yucatan, there were only an estimated 47000 ha of irrigated crop production in 1995 - less than 5 percent of the total cropped area - while in Brazil the total is under 400000 ha. By contrast, there are abundant water conservation measures in place, including reservoirs, retention barriers desalinisation, etc., but none work well during periods of severe drought. Frequently, the installed technologies are not understood by those they are supposed benefit.

Infrastructure throughout the marginal drylands system is poorly developed. Public services such as health and education are only available in some locations and are generally insufficiently funded to be operated effectively. Poorly maintained and unsurfaced roads and a low degree of market development add to the problems facing economic progress in the region.

In Yucatan a crisis has arisen, as increasing population pressure has forced a reduction in the fallow period of the shifting agricultural production pattern. As a result, insufficient time is given to the milpa land to recover its fertility after the cultivation period. Yields, and the ability to resist drought, have been reduced, lowering levels of food security in rural communities. The rapid growth of tourism in the coastal zones of Yucatan, and its consequent demand for labour, has also had a profound direct impact on the structure, makeup and economic status of the farming system. Many families have migrated permanently to the state's new coastal cities such as Cancun and Cozumel29. A similar pattern of migration is also occurring in the Brazilian northeast. Because of the marginal nature of the area and the recurrent droughts faced by producers, the rural population has become very dependent on periodic government assistance. Since no long-term solution has been developed, this population represents a migratory time-bomb, with some estimates placing the number of potential migrants at 8 to 13 million.

Trends and issues in Dryland Mixed System

The key sectors that offer potential for regional income growth are tourism, services and agro-industry. By year 2030, the productive structure of this farming system will probably be more concentrated than today, as many sub-marginal producers will have left. There is a low probability that the system will be able to finance its own development, so it will continue to depend on government resources. Total population is projected to grow at one percent per annum. Income distribution will continue to be highly skewed and any decrease in poverty will rely heavily on government action programs.

The operating capacity of public institutions, including the research and extension sector, is expected to increase. However, improved response to small farmer needs, and expanded co-operation with the private sector, is unlikely to occur unless significant measures are undertaken to restructure the organisation and management of research organisations. On the other hand, the organisation and democratisation of society will be stronger.

Drought effects are likely to be aggravated over the years, as climatic instability increases and degradation of natural resources reduces the capacity of the system to resist long dry periods. The impact will be high, further degrading soils and vegetation and increasing population outflow. These negative effects will be mitigated by growing irrigation activities - with an expected 0.5 million ha of new irrigated land. Much of this irrigation will be devoted to the production of tropical fruit for export, creating nearly an estimated 300000 jobs. A key constraint is the widespread perception that lack of water is purely a result of limited natural resources. In reality, it arises largely from inequitable land distribution, inappropriate technologies, and poor resource management. Land degradation in sub-marginal holdings further exacerbates the problem, although opportunities do exist for low water-demand production systems (e.g. cashew).

Yucatan faces a particular challenge with respect to the sustainability of shifting cultivation practices. Technical changes within the milpa system are limited to the introduction of external inputs such as herbicides, improved varieties and, to some extent, fertiliser. Crop diversification, or changing the cropping cycles on the same milpa, are possible adaptations. However, if fallow periods continue to decline, the only possible outcome will be serious soil and vegetation degradation and drastically lower yields. This may well happen, unless out-migration occurs at a faster rate than population growth. Other important issues to be considered are:

• reform of the existing system of agricultural land to permit consolidation of holdings and increased efficiency;
• control of desertification in the driest areas; and
• using ecotourism potential to create employment opportunities.

Priorities for Dryland Mixed System

In contrast to other regions of the world that have already made the transition to modern agricultural practices, the Dryland Mixed Farming System has failed to undertake the necessary changes to stimulate a broad and equitable process of agricultural and socio-economic development. Even major advances in technology will not permit a poor agricultural population of 10 million people to escape from poverty over the next 30 years, given existing resource constraints. If farm populations can be reduced, and a more equitable land distribution achieved, then the possibility exists for increased earnings for those remaining - together with a halt to the ever-increasing degradation of the natural resource base.

Two major strategic approaches to poverty reduction can thus be proposed: (i) alternative livelihoods, principally through exit from agriculture within the system; and (ii) increased off-farm incomes. In addition, agricultural growth through increased farm size, diversification, and some intensification of production will contribute to the reduction of poverty.

Provision of alternative livelihood opportunities to sub-marginal farm families within the region would include incentives for the creation of agro-industry and other rural occupations, such as training of workers and tax benefits. Assistance in relocating sub-marginal farmers to areas of agricultural expansion (i.e. the Cerrados) is also important; possibly including compensation for exit from currently occupied lands and access to finance to buy land in expansion areas. A further element in this strategic approach is the provision of skills training and infrastructural improvements to facilitate the movement of people (especially youth and women) into regional urban areas.

The second strategic approach consists of providing those remaining on-farm with the potential for increased earnings through land purchase and diversification into higher value crops (possibly directed at the tourist sector). This may exclude many renters who have no initial land base from which to expand. Experience in a number of countries has shown, however, that financing land purchase costs alone is not enough. Any programme of this type must also address: (i) on-farm investments and working capital requirements; (ii) dissemination of technologies that improve productivity in relation to moisture constraints - including development and dissemination of drought resistant varieties and species and encouragement for the replacement of maize; and (iii) zero-tillage technologies appropriate for small producers; small-scale irrigation where feasible.

Introduction of farming practices that reverse the degradation of the natural resource base should also receive a high priority. These practices include: (i) the expanded utilisation of legumes and fodder crops (e.g. Mucuna pruriens and Canavalia ensiformis); (ii) zero or limited grazing systems for small-stock; and (iii) greater attention to the potentials of native vegetation. Land use planning should also be given priority to improve the identification of areas under risk. Meteorological research leading to specific drought forecasting all over the region would also be valuable.

In addition to these major strategic approaches, public financing must be shifted away from emergency drought relief and similar programmes into activities that provide hope for future avoidance of these conditions.

MAIZE-BEANS (MESOAMERICAN) FARMING SYSTEM

Characteristics of the system

The Maize-Beans (Mesoamerican) Farming System extends over an area of approximately 65 million ha. It occupies mostly upland areas from the Panama Canal to the highland plateaux of Guatemala and Central Mexico. The system is distinguished by: (i) the significant proportion of indigenous population30; (ii) the central role, both agriculturally and culturally, of maize and beans; (iii) the small size of holdings - typically under five ha and in El Salvador, less than two ha31; (iv) the high degree of on-farm consumption of production (over 65 percent in Honduras)32; and (v) the importance of seasonal migration of wage labour to lowland agricultural and coffee estates. Coffee, and intensive small-scale irrigated vegetable production (in areas close to roads and urban centres), are important income sources and often critical in determining the degree of poverty within a community.

Cultivated land may extend as high as 3500 m above sea level in the Guatemalan highlands, but the majority is between 400 and 2000 m, with a precipitation from 1000 to 2000 mm per annum. There is extensive irrigation, much of it under the control of larger farmers, although small-scale systems play an important role in horticulture. Most soils are of volcanic origin, and relatively fertile, but on slope lands tend to be thin and subject to erosion33. FAO studies from the 1980s estimated severe erosion on as much as 45 percent of all land in El Salvador and 25 to 35 percent in Guatemala34. Forest cover was severely reduced during the 20th century, leaving large areas of contiguous forest only in inaccessible areas and in National Parks. A distinctive sub-system in Central Mexico35 shares the cultural and agronomic elements of the main system, but is agro-ecologically different, consisting of high altitude plateau (2000 to 3000m elevation) with lower temperatures and poorer soils.

The system has an estimated agricultural population of some 11 million people, half of them in Mexico36. This ranges from about 50 percent of national agricultural population in Guatemala to less than 20 percent in Honduras, while arable area within varies from 40 percent (El Salvador) to 10 percent (Panama) of national totals37. In 1989, it was estimated that there were 1.4 million producers of basic grains in Central America38, the vast majority of whom would be included in the system.

Large-scale estates, frequently in excess of 100 ha, are interspersed throughout the system. Often controlled by absentee landlords or corporations, these estates have traditionally been dedicated to the commercial production of coffee and beef although more recently rubber, cut flowers and foliage production have increased in importance. In addition, commercially-operated family farms created by European settlers are clustered in the more fertile valley areas of the system, and often produce vegetables and dairy products as well as coffee. Most indigenous producers control their family parcels under customary arrangements but lack legal title, rendering access to formal credit sources very difficult. (see Box 7.7).

Yields tend to be low; with maize averaging 1 to 2 t/ha39; by contrast average maize yields in Sinaloa State, where irrigation is widespread, reach 6 t/ha while system yields of beans (0.6 to 0.9 t/ha) are lower than the Mexican national averages of over 1 t/ha40. Nevertheless, small-scale indigenous producers are important contributors to national output of these crops. In 1999, over 50 percent of the area sown to maize in Mexico was within the system boundaries. In total, some 6 to 7 million ha of maize are harvested annually within the system41. The importance of the system is even more pronounced for coffee, which requires altitudes over 500 m for proper development and fruiting. Almost all regional coffee production derives from the mesoamerican system area. However, large coffee estates account for much of that output.

Extensive poverty is present throughout the system, reaching levels as high as 80 percent in the Guatemalan Departments of Huehuetenango and Quiché. A regional average of 60 percent was estimated by IICA in 199142. Malnutrition is also widespread, especially in the period before harvest. Returns per unit of land tend to be greater - but population densities much higher and holding sizes smaller - in areas where coffee production is practised. Food insecurity is exacerbated by significant variability in coffee prices from year to year.

Public infrastructure is sparse or completely absent away from local administrative centres, especially in those areas where there has been prolonged armed conflict in recent decades (Chiapas in Mexico, Guatemala, El Salvador and Nicaragua), and many indigenous communities are several hours from the nearest road. Similarly, the availability of education, health care and other services is minimal. To compensate, many indigenous communities have high levels of social cohesion, enabling community mobilisation for a wide range of tasks. Community controlled forest areas are also common, but tend to be very fragmentary.

Trends and issues in Maize-Beans (MesoamericaN) System

Fragmentation due to inheritance and other causes has increased the number of holdings and decreased the average size over recent decades. The number of holdings with maize in Guatemala more than doubled, from 321000 to 667000 between 1964 and 1996, (however not all may be within the system)43. This is faster than the rate of population increase. Although urbanisation rates are also increasing, projections indicate no significant decline in rural populations in Mexico and Central America over the next 30 years. This suggests that pressure on access to land - an underlying cause of so much of the civil conflict in the region in the last 20 years - will continue.

With so little unexploited land available, existing trends are likely to continue: (i) exploitation of ever steeper slopes; (ii) intensification of traditional production systems; and (iii) diversification of production. Expansion onto steep slopes will generate only short-term gains, as underlying soil structure is difficult to maintain and erosion will increase. In addition, increasing climatic variability resulting in torrential rainfall and flooding, is already believed to be responsible for major damage to many degraded areas. This trend will probably intensify.

The positive relationship between population density and yield levels in Central America44 suggests that yields of both maize and beans are likely to increase over time, perhaps by 50 to 100 percent. However, a downward trend in farmgate prices for basic grains is likely to increase pressure on semi-subsistence farmers. Falling international prices, which are projected to decline even further over at least the next ten years45 and trade liberalisation, have eroded the degree of protection previously enjoyed by national producers. Although major advances in maize yields may well occur in industrialised countries in the future, the reliance of farmers on self-produced seed, plus widespread soil limitations, would likely limit the benefits to system farmers.

Increased diversification will offer some compensation. National urban populations with increased incomes will demand more high value products, and improved technologies in post-harvest handling will continue to expand speciality markets in industrial countries. Many of these products are labour intensive and can be well suited to family labour. Diversification has already occurred in horticultural and fruit production in peri-urban and other favourable areas throughout the mesoamerican system and will likely accelerate over the next 30 years.

Four principal development household strategies are thus foreseen for small producers within the mesoamerican system. Firstly, those with favourable resource endowments (including the possibility of irrigation), location, and human capital will probably emerge from poverty as specialist suppliers of export-oriented products, despite their small holding sizes (see Box 7.8).

Other cases of diversification into intensive horticultural and fruit production include sweet onions in Nicaragua and chilli peppers in Belize. Maize and beans will continue to be grown for household consumption only, but perhaps with dramatically higher yields, as this group will have the resources to adopt new technologies. However, it is not believed that this strategy will ever represent a viable option for more than a relatively small percentage of mesoamerican system smallholders - perhaps 15 percent or less.

Secondly, a larger proportion of smallholders are likely to continue to rely on the farm for their basic sustenance, but turn increasingly to off-farm employment as a means of earning the income needed to finance basic household expenses (medicine, education, clothes, etc.). This income may also finance greater input use, raising yields. The growth of export-oriented production on the part of commercial farmers and pioneer smallholders will create employment opportunities, both in agribusiness, as well as in services. Expanded tourism may also offer a source of employment.

A third path will consist of families who abandon their holdings and move to urban centres. It is likely that, over the next three decades, this exit strategy will be followed by a sizeable proportion of the rural population, reaching perhaps as high as 20 to 25 percent for areas such as El Salvador where population densities are high.

The fourth path comprises those who neither benefit from the new developments nor migrate in search of work. This group will account disproportionately for the extension of cultivated area on steep slopes, risking severe environmental and human loss from flooding and landslides in search of increased production. Agricultural output will bring declining real income as relative returns diminish and with few resources to purchase inputs, yields will stagnate or decline. Poverty will worsen for this group, perhaps dramatically. Unfortunately, they are also likely to be the largest group, and may account for more as many as half the population in some zones.

Throughout the system, public infrastructure and services are likely to continue to improve, especially as a result of significantly expanded post-conflict rural investment programmes in countries such as Nicaragua, El Salvador and Guatemala. Private sector and civil society participation in rural areas is also expected to increase in importance.

Priorities for Maize-Beans (MesoamericaN) System

The system is characterised by extensive and often severe poverty. Yet focusing attention on increasing yields of traditional products can produce limited poverty alleviation at best. The problem facing producers is not one of insufficient foodstuffs, but rather the need for higher cash incomes to meet household needs. Where no alternative sources of cash exist they are forced to sell output that would otherwise be consumed within the household, hence creating secondary malnutrition.

For those producers occupying degraded or fragile slope land and unable or unwilling to return such land to forest cover, the priority must be to promote more sustainable patterns of production that can be readily adopted with few resources, and can provide rapidly realisable benefits in output or labour use. Specific interventions include: (i) permanent high value crops; (ii) reduced tillage; (iii) higher plant populations; (iv) contour cultivation; (v) improved varieties, (vi) live barriers; (vii) intercropping; (viii) dispersed tree cover; and (ix) mulching. The benefits of these technologies and the feasibility of their adoption have been clearly demonstrated by a number of innovative projects within the system. However, while such practices can both increase yields and lead to more sustainable use of natural resources, they will have only limited impact on increasing household incomes unless integrated into diversification and marketing programmes, and can only be seen as part of any solution.

Diversification offers the greatest potential rewards but, even among those who can make the transition, support will be needed. The private sector is best equipped to provide this, as entrepreneurs understand most clearly the needs and demands of the market. Nevertheless, a crucial role remains for Government in ensuring that the business environment is appropriate for small farmer-private sector interaction. Important measures include: (i) ensuring a competitive environment for enterprises, with low barriers to entry and controls over monopoly powers; (ii) promotion of farmer associations and small enterprise groupings, that can act as effective channels for financing, technical assistance and management training; and (iii) increasing the participation of producers and traders in the setting of policies and negotiation and monitoring of trade agreements. The development of land markets would accelerate the transition, permitting producers to buy-out less successful neighbours and form the nucleus of a commercial family farming sector.

Off-farm employment and other rural employment may initiate an upward spiral of employment, earnings expenditure, and increased demand for goods and services among those unable to diversify production. Investment by medium and large scale agro-industries, assembly plants and tourism can be promoted by offering non-financial incentives. These could include: (i) public infrastructure provision in zones where companies are establishing operations; (ii) extension and organisational assistance to contract suppliers; (iii) training of company staff; and (iv) assistance in dealing with municipalities in land acquisition, local taxation, etc.

The reduction of rural poverty among those unable to diversify production or obtain off-farm employment must focus on facilitating a successful exodus of people from these areas. Empirical evidence strongly suggests that the poorest segments of the population do not migrate as much as wealthier rural inhabitants; they lack the resources to do so47. An appropriate strategy would be to tackle this problem on two fronts - providing resources for migration and attempting to increase the probability of successful absorption at their destination. Start-up capital for migrants might be provided through: (i) payments for the transfer of customary land rights to groups interested in conservation and biodiversity, thus taking sub-marginal land out of circulation; and (ii) financing the purchase of land rights by more successful small producers. Literacy and vocational training is also important for would-be migrants and their families, to provide basic employment skills.

HIGH ALTITUDE Mixed (CENTRAL ANDEs) FARMING SYSTEM

CHARACTERISTICS OF THE SYSTEM

The High Altitude Mixed (Central Andes) Farming System extends over 121 million ha from Cajamarca in Northern Peru, through Bolivia into Northern Chile and Northeast Argentina (see Box 7.9). An estimated 40 percent of the territory of Peru and Bolivia falls within the system boundaries, as well as smaller portions of Chile and Argentina. In the north a series of interwoven ranges dissected by longitudinal valleys make up the Peruvian Sierra. Further south, at about latitude 14°South, the Andes divide into two major ranges, enclosing the plateau or altiplano of Peru, Bolivia and Northern Chile and Argentina. This vast treeless area of land is composed of undissected tableland above the level of 3500 m and is characterised by interior drainage. Some of the agricultural land reaches 4500 m48. Precipitation is concentrated within a single wet season of variable length and ranges from 150 mm in the western ranges to 1000 mm per annum in the eastern ranges. Although the soils, as well as their capability for agricultural production, are extremely diverse their fertility is typically low. All the lands in the system are affected by severe soil erosion.

Agro-ecologically the zone is extremely complex49. The great variation of soil types and the frequent sharp changes in altitude are accompanied by dramatic changes in temperature, humidity and rainfall. An aridity gradient exists from east to west across the Central Andes as well as from north to south along the length of the chain. Annual mean temperature varies greatly with altitude. Mean daily temperature is generally below 10°C and frost is common, especially during the dry season.

Apart from altitude, other key features that characterise the Central Andean High Altitude Farming System are the overwhelmingly indigenous population, the virtual absence of large holdings - unusual in Latin America - and the low levels of external input use. Most of the system was originally part of the Inca Empire and has retained strong cultural elements inherited from that period, which have changed only superficially. Poverty within the system is extensive. In 1997, 68 percent of the total rural population of Peru fell below the national poverty line and this proportion is believed to be higher for Bolivia, although recent data are not available.

According to the 1994 Census, almost 60 percent of holdings in the Peruvian Sierra were under 3 ha with 1.5 to 2.5 ha cultivated (often in scattered parcels), while in the altiplano they reach 15 to 20 ha, with 1.5 to 2 ha cultivated. These small cultivated areas, combined with the poor productivity of such high altitude arid lands, result in endemic poverty, widespread soil degradation and erosion, and out-migration has become an important element of the system. There is an extensive flow of permanent and seasonal migration to the recently opened land on the eastern side of the Andes and to the Amazonian lowlands beyond. Nevertheless, despite out-migration, it is estimated that the rural population within the system grew by 1.6 percent per annum during the period 1960-1990 - this growth rate is expected to decline in future. The most densely populated areas of this farming system are the Peruvian Sierra with about 0.3 inhabitants per ha and the Bolivian altiplano where in certain parts, such as the areas surrounding La Paz, Cochabamba and Potosi, it can reach 0.4 inhabitants per ha. The remaining arid and subarid parts of Chile and Argentine are very sparsely populated. The average rural population density for the whole farming system is 0.06 persons per ha (17 ha per person).

The cultivated area within the system is about 3.1 million ha and represents as much as a third of the national cultivated area of Peru and Bolivia. An estimated 1.2 million ha are irrigated, mostly in the arid western section in Peru. On the basis of statistical data from 1994, an estimated 1.2 million farms in Peru50 and 0.6 million in Bolivia51, fall within the system. The total number of farms for the entire system would be over two million. Legal rights over property and other resources are not commonly established. In the Peruvian Sierra, approximately one-third of holdings are legally owned, while two-thirds are held under customary arrangements. There is almost no formal land rental, although there may well be informal arrangements. The agricultural population of some 7 million people, three-quarters of them in Peru52, rely on a subsistence-based mixed production system including potatoes, pre-Colombian Andean crops (e.g. quinoa and chenopodium) and barley, maize, and lima bean. Sheep are important in the Peruvian Sierra, while the camelidae (llama, alpaca) dominate further South. Guinea pigs are also of major importance at higher altitudes. The potential for irrigation is reasonable. However, given the topography of the system, much of the water only becomes available on the lower flanks of the western ranges.

In the Peruvian Sierra, typical land use patterns distinguish three major production systems according to altitude: (i) the valley floor where corn, quinoa, chenopodium and potatoes are grown. Where water is available, small-scale traditional irrigation systems may make horticultural production possible; (ii) the intermediate slopes, where drier western terraces are dedicated to barley and grains and the better eastern slopes are dominated by tuber cultivation; and (iii) the high hills where the cultivation of more frost-resistant crops is combined with pastoral activities.

In the altiplano sub-zone further to the south, the land use pattern chiefly depends on the gradient of rainfall. In the most humid areas agriculture is widespread, while under drier conditions extensive livestock production predominates. In some of the arid and semiarid areas, such as the `Valles Altos' of Bolivia, extensive degradation of soils has caused major changes in prior production patterns, as traditional crops such as maize have become infeasible.

Average system yields are constrained not only by agro-ecological conditions but also by the limited use of external inputs and the low productivity of family labour. In the Peruvian Sierra, less than 10 percent of smallholders are estimated to use purchased seed, but nearly 70 percent use organic fertiliser of one form or another. Maize yields typically do not exceed 1 t/ha, with quinoa yielding 0.85t/ha and potatoes 10 t/ha. Other cereals (wheat, barley) average approximately 1 to 1.2t/ha, but these yields may reflect the participation of larger producers. Further south, in the altiplano sub-system, yields drop further: potatoes from 4 to 5 t/ha; quinoa 0.6 t/ha and wheat and barley about 0.6 to 0.7 t/ha.

During the mid-1980s and the beginning of 1990s, structural and sectoral adjustment programmes caused major changes in national economies. Food production grew at annual rates of three percent and five percent for Bolivia and Peru respectively53 while agricultural exports expanded at an even higher rate; eight percent and nine percent respectively for the same two countries. Food imports also rose, but at a lower rate, resulting in a net improvement within the domestic agricultural sector. Nevertheless, the evidence suggests that most of these gains have been captured by the modern agricultural sector, bringing little benefit to the producers of the high altitude system. The modern agricultural sector has also been the beneficiary of most public resources devoted to the rural areas, including the development of improved agricultural technologies. The system is characterised by a widespread lack of basic infrastructure in such areas as education, health, roads, and markets. In Bolivia, the proportion of overall public resources devoted to the agricultural sector has not been consistent with the contribution of this sector to the economy. Agricultural GDP was 16 percent of the national total in 1999 but the sector received only six percent of the total public budget. The allocation of resources to the agricultural sector is more equitable in Peru, where both expenditures and agricultural GDP are around 7 percent.

Trends and issues in High Altitude MIXED (Central Andes) System

Over the next 30 years, rural populations within the system are not expected to fall significantly. Pressure on access to land will persist as declining real prices for agricultural products drive producers to expand their cultivated areas and increase stocking levels. The low productivity explains the apparent contradiction between the low population/cropland ratio and the persistent high pressure on land. The widespread soil erosion occurring on the slopes of the Peruvian Sierra and the altiplano will likely worsen without substantial changes in cropping patterns and natural resource management practices. Strong economic growth nationally, and in the lower valleys and Amazonian lowlands in particular, may accelerate out-migration. Although historically a relatively abundant resource, labour may join humidity, soils and slope as a limiting factor of production if this occurs.

The evolution of land tenure and landholding distribution during this period is an extremely complex matter. It is expected to be mainly influenced by whether or not adequate land registration, financing, and market development for land will be achieved, requiring considerable legislative institutional reform54.

Although crop yields are low by regional standards, the adoption of available improved technologies requires levels of input use that are not feasible for most producers under current circumstances. In fact, natural resource degradation and possible labour shortages will increasingly challenge current production levels. Also, opportunities for crop diversification are more limited than in other zones, although expansion of irrigation and protected cultivation under plastic tunnels may provide limited opportunities.

The best opportunities for many producers may lie in improved community organisation to benefit from the potential for added value in existing products. Stronger marketing linkages to end-users (supermarkets, institutions, and restaurants) and the processing of raw materials (milk, potatoes and grains) could significantly increase family earnings, while speciality products such as quinoa and camelidae wool may provide profitable export opportunities.

Key issues that will have to be addressed if poverty reduction is to be achieved and economic growth promoted within the system, include:

• Unsustainable natural resource management practices, resulting in soil degradation and erosion;
• Low prices for agricultural outputs;
• Lack of off-farm employment and income generating opportunities;
• Low factor productivity, especially with respect to labour and land;
• Inadequate public sector investment in infrastructure and agricultural services.

Priorities for High Altitude MIXED (Central Andes) System

The extreme agro-ecological conditions, fragmented landholdings, poor soils and lack of off-farm employment opportunities have resulted in extremely high poverty levels within the Central Andean high altitude system and render the sustainable development of the system both a necessity and challenge. Soil degradation on the hillsides in the Peruvian Sierra and in the altiplano compromises the long-term potential for land productivity55. Although poverty is recognised as a dominant force affecting soil dynamics in the mountainous areas, the manner in which it influences production systems is complex. Erosion control projects implemented within the system have had only limited success, and must recognise that purely technical responses will fail to resolve the problem. Instead, they must be better integrated with the creation of wider economic opportunities for small famers, while recognising the limitations imposed by low labour productivity and fragmented holdings.

The concentration by small-scale producers on low value commodities for domestic markets has left them exposed to stagnant demand and with consumers who are unwilling to pay higher prices for such aspects as varietal improvements, post-harvest handling or packaging. Yet a number of the products grown within the system have significant potential for international specialist and niche markets. Anecdotal evidence suggests that major increases in prices can be achieved by tapping into these markets for llama and alpaca wool, quinoa, speciality potato varieties and similar products, especially if the potential exists for origin-labelling or control. Fair trade agreements may also offer income increasing opportunities, but are inherently limited in scale. Diversification for export markets will require improved producer organisation to co-ordinate and bulk volumes required for export shipments; as well as efficient input supply mechanisms (including finance). Technical assistance would be needed to ensure adequate quality control, develop post-harvest handling and packaging, and establish effective market linkages.

Off-farm income, particularly related to agro-industry, can contribute significantly to rural development in small-farm areas. Agro-industry often induces rapid technical change among participating small farmers, contributes to demand for local production56 and creates off-farm employment (see Box 7.9). However, the imperfect functioning of markets has traditionally inhibited a broad expansion of small-farm oriented agro-industry. This is particularly true in relation to rural credit57, land markets, insurance for agriculture activities, information, technology and specialised inputs. Agro-industries have developed a number of strategies to compensate for the effects of these imperfect market, including the provision of farmer credit, technology and inputs, as well as renting of land from small producers59.

Although it is difficult to develop recommendations for such programmes and policies outside the context of a specific activity, they would generally include the reduction of information costs and improved market linkages for producers in order to move toward assured market access and financing of production costs. The framework should include a strategy to strengthen technical assistance. The reduction of transactions costs arising in the first stage of the linkage between the industry and farmers60 is important, including rapid, simple and enforceable arbitrage mechanisms to resolve the conflicts that often arise between farmers and agro-industry in their transactions. Facilitating access to credit and removing legal and market constraints affecting small farmers with regard to enlarging their landholdings, are also of great importance. Similar priority should be accorded to reducing the costs of training - both to farmers and to agro-industry - as well as to supporting farmers' organisations.

Increased land and labour productivity is a further priority area, as the High Altitude Farming System of the Central Andes is characterised by very low factor productivity. Average smallholding size is relatively large by the standards of many farming systems, but the very poor returns to land and labour leave producers with little benefit for their efforts. While low productivity is in part a result of unfavourable climatic and soil fertility conditions, it can be argued that the current system fails to utilise effectively those resources that are available. Priority must be given to participatory research and extension on adapted crop varieties for this high altitude zone (short season, drought tolerant, etc.). It is worth noting that successful trials have been made in promoting indigenous technologies - such as the use of large earth planting beds dissected with deep trenches that reduce frost damage by avoiding settling of cold air on growing plants - and more work would be worthwhile in this area.

It is important to accelerate the regularisation and registration of farm ownership, the lack of which seriously hinders land market development and provide strategic support to land buyers - mainly information and notary services - including identifying of available land and the provision of more readily-accessible mortgage facilities.

The achievement of productivity increases is closely related to public investment in infrastructure and services, in improving human capital within the farm population through education and health care, provision of specific government programmes to promote the development of the agro-industrial sector, as well as the completion of the reforms in factor61 markets. The conditions to achieve these objectives do not yet appear to have been met. At the root of any strategy for rural development is the premise that macroeconomic stability and more appropriate relative prices, are not sufficient by themselves to bring about large-scale agricultural productivity growth. These must be complemented by government support, in particular in the form of more effective government investments in agriculture. Greater emphasis should therefore be placed on adequate public resource allocation to agriculture through suitable budgetary adjustments.

STRATEGIC PRIORITIES FOR LATIN AMERICA AND CARIBBEAN

The following strategic priorities derive from the analysis of the four selected farming systems supplemented by expert knowledge of the other systems. It is believed that the conclusions with regard to poverty reduction and agricultural growth presented below will be broadly relevant across the region as a whole62. There is considerable scope for reduction of agricultural poverty in Latin America and Caribbean, through production intensification, diversification (including value added), expansion of farm size, off-farm employment and exit from agriculture63. Table 7.4 indicates the relative importance of each of these strategies as factors in reducing poverty within individual farming systems in Latin America during the coming decades.

As can be seen, the importance of different household strategies varies significantly from system to system. In broad terms, lower potential systems derive a greater proportion of poverty reduction from farming system exit - migration to cities, abandoning farms for employment in the off-farm economy or moving to another farming system - than do higher potential systems, which tend instead to derive more benefits from productivity increases.

Across all systems except the most remote, it is expected that diversification will be a major source of agricultural poverty reduction. Diversification would include a shift into higher value non-traditional crops as well as added-value activities, such as grading, packaging, and on-farm processing. Increased off-farm employment and income generation is also believed to be generally important. However, strategies such as the expansion of farm size among small farmers is important as a source of poverty reduction in only a few cases.

Expected sources of poverty reduction have implications for the specific interventions in each of the five principal categories outlined in the introduction to this book. Each of these categories is briefly examined below:

Policies, institutions and public goods

Three key strategic areas are likely to dominate governmental and institutional roles within farming systems in Latin America over the next 30 years: (i) improving access to land and - to a lesser extent - water, in farming systems with extensive poverty and among poverty groups in more wealthy systems; (ii) promoting alternative occupations for the agricultural poor who do not have access to sufficient land and water to escape poverty; and (iii) strengthening public goods in rural areas. In all these areas government must inevitably play a major role, although preferably working in co-operation with civil society and private sector groups.

In many farming systems, a small minority of large farmers occupy large areas of land, which are often utilised only at relatively low intensities, while most producers are confined to smallholdings which are increasingly less viable. Civil conflict has often been a direct result of this situation. For the Maize-Beans (Mesoamerican) and Dryland Mixed Systems (arguably less so for the Central Andean System), poverty levels are directly related to problems of access to, and control of, natural resources - primarily land. Effective land policies is also important in frontier areas (e.g. the Extensive Mixed system) where in-migration from neighbouring poverty systems could lead to conflict. Key strategic priorities include: (i) improved functioning of land markets, through acceleration of cadastral and titling procedures, conflict resolution mechanisms, and changes in land tax structures; (ii) land banks to buy both marginal and large holdings and resell land with the objective of consolidating smaller commercial holdings; (iii) fiscal disincentives to under-utilised holdings and incentives for sale to land banks; and (iv) supported exit for sub-marginal producers including, where feasible, negotiating the purchase of traditional rights to land.

A vital aspect of policy reform is the enforcement of legislation prohibiting illegal seizure of lands and other resources. This applies not only to powerful interests capturing state land, but also to small producers illegally occupying private lands and colonists entering indigenous community lands. Where such risks are perceived to exist, land may not be put on the market or offered for rental due to the fear of attracting squatters. This has become a major issue in recent years in the coastal zone of Guatemala, for example.

Despite the opportunities that may exist for diversification and increasing output value among small-scale farmers, there will inevitably be many marginal and sub-marginal farmers who simply lack the human, financial, locational and natural resource assets to benefit from these opportunities. Two broad alternatives exist: local off-farm employment and outmigration.

Off-farm employment offers a major route for escape from poverty in severely constrained farming systems. Policies, institutional support and public goods can all play a major role in promoting such employment. Natural resource-based industries offer employment opportunities, but the disadvantages of locating facilities in rural, compared with urban, areas must be overcome. National and local governments can co-operate with the private sector in the design of integrated programmes in which larger potential employers (agro-industries etc.) would be offered incentives to offset the perceived advantages of urban operations. Support measures might include: (i) improved infrastructure in the area of the plant and its suppliers, perhaps together with a long-term commitment from the employer to part-finance maintenance; (ii) provision of training for future staff of the employer; (iii) supply organisation among raw material providers (crops, livestock, etc.); and (iv) the creation of rapid, transparent and binding dispute arbitration and settlement mechanisms for suppliers and the employer.

To promote the development of small-scale enterprises, key interventions would include: (i) simplification of small enterprise registration and approval procedures; (ii) assistance in preparing realistic business and investment proposals; (iii) training in simple accounting and administrative procedures; (iv) more rapid and flexible financing of investments; (v) priority in providing basic services (e.g. electricity, telephone, water, etc.); and (vi) encouragement and resource provision to associations representing small scale rural enterprises.

Out-migration has been the traditional response for those who are not able to participate in diversification or off-farm employment activities. In recent years, the focus has been exclusively on retaining the rural population in situ, and discouraging out-migration. Yet, if little potential exists for substantial increases in quality of life within the system, this retention policy must be questioned. Indeed, it is timely to identify and implement measures to ensure that any process of out-migration is a positive one, both for those migrating and for those remaining within the system. These would include innovative measures, such as targeted out-migration incentives, to increase the human and financial capital of migrants and thus ensure better economic possibilities in the future. Such incentives could involve providing capital to departees who hand over control of lands currently occupied to the financing agency. Where these lands are sub-marginal for agricultural purposes they would be passed to conservation authorities for reforestation and watershed protection measures. Otherwise, land banks would finance their acquisition by more successful neighbours, thus contributing to land consolidation.

Training - particularly of women and young people - in vocational skills of direct relevance to off-farm employment or migration, needs considerable strengthening throughout the region. This should be seen as second in importance only to basic primary education, as a social service in rural areas.

One key area of public goods is infrastructure. Although infrastructure is increasingly accepted as a largely private sector activity, such areas as rural roads, electrification and water capture and distribution for irrigation will have to remain public goods, and are critical for diversification and intensification activities in many areas. The withdrawal of the state from many rural institutional activities (banking, extension, marketing) renders strengthened support and oversight capacity for civic and private institutions essential.

Trade liberalisation and market development

The liberalisation of trade and markets is probably the dominant force governing the evolution of farming systems, and thus rural hunger and poverty, within the region. Imported products increasingly compete in national markets, often undercutting local producers - at least in easily-served urban markets. A rapid transition to free market conditions will increase the prevalence of poverty in the short term, as producers face declining prices for traditional products and difficulties in the transition to new higher-margin products. Those systems with severe poverty are most seriously affected, as they often lack the human, financial and technological resources needed to adjust. In the longer run, trade liberalisation should also create opportunities for market development and diversification, although not all system participants are likely to be able to benefit from such changes.

Governments have legitimate roles with respect to markets and trade. However, experience has shown that many direct state-supported interventions merely impede the transition process while providing little real benefit to poor farm households. Instead, the most effective strategy is often to establish the conditions for an active and competitive private and civil sector in rural areas as outlined in the previous sub-section (see also the Maize-Beans [Mesoamerican] Farming System case study). Apart from creating an environment that is supportive of small business establishment and growth (see previous sub-section), such a strategy would comprise interventions to assist farmers to organise themselves to respond to new opportunities, including: (i) training group leaders in commercial management and administration; (ii) providing information, genetic material and technical assistance to promote the adoption of new crops; (iii) promoting value-added activities (selection, packaging, processing) through technical and financial assistance for quality standards, brand creation and targeted marketing; and (iv) improving access to financing for required investments.

One traditional area of public intervention has declined in many farming systems as a result of restructuring and decentralization: setting and enforcing market-related norms and standards. Activities in areas such as weights, measures, quality standards, disease monitoring and tracking, and phyto-sanitary controls are all vital public functions if market and trading systems are to function effectively.

Information and human capital

Accelerated rates of change appear inevitable in traditional systems, and the process will require both improved information and human resources. Provision of information by the public sector involves substantial recurring costs and is frequently irrelevant to real market needs, so the service is best provided by private sector buyers and traders, or through trade or producer associations. Typically, the most effective support that Government can offer for information dissemination is the improvement of communications channels, whether they be cellular telephone networks, open licensing of local radio stations, or simply improved access roads to rural areas.

Human resource development should focus on vocational training, and numeracy. Literacy may not necessarily be the first priority. Not all rural inhabitants are farmers, and agricultural training should be only one possible training alternative, alongside skills such as sewing, mechanics, welding, cooking etc. These skills may be particularly important for future out-migrants (especially young people) in helping them to find employment at their destinations. Training should also be able to address the needs of specific employers in rural areas (whether agro-industry or non-agricultural) as a way of removing disincentives to rural enterprise operations.

Over a thirty year period, the use of new communications technologies, such as the Internet, may revolutionise the transfer of information and skills to poorer rural communities. However, considerable infrastructure development, as well as profound changes in government thinking about services and service provision, will be needed before the promised revolution can occur.

Science and technology

A number of pioneering projects have shown that a range of technologies already exist that can contribute to improved natural resource management and drought tolerance. These include (i) increasing soil organic matter content e.g. by legumes (Mucuna pruriens and Canavalia ensiformis); (ii) no-tillage cultivation, multi-cropping and small-scale irrigation linked to terracing on lower slopes and in semiarid areas; and (iii) vegetative barriers, contour protection, permanent crops and agroforestry on steeper slopes. The zero or controlled grazing of livestock, especially goats, when combined with fodder crops and trees, is another promising technology improvement, as is the integrated management of fragile savannah soils.

However, significant changes are needed in the field of applied agricultural research. Greater use must be made of participative approaches to areas such as improved varietal selection and field trials of new potential diversification crops. Achieving this will require considerable restructuring of national research organisations in many LAC countries, with particular emphasis on disbursement mechanisms for research grants and on staff training in participatory methods. Responsiveness of the research system to market demand must be increased. Although hybrid genetic material would be acceptable for diversification, it is likely that material capable of on-farm multiplication would be a prerequisite for traditional crops.

Increased research is required on short-season and drought-tolerant crops suitable for small producers, and in particular on technologies that increase labour productivity in systems with high levels of poverty. To increase the ability of smaller producers to compete effectively in growing international markets, research will also be needed in such areas as: (i) adapting existing and future post-harvest technologies to the needs of smaller producers; (ii) appropriate IPM and organic cultivation practices and tools (e.g. biological controls); and (iii) field testing of new varieties/species and determining optimal agronomic practices.

Natural resources and climate

Many regional farming systems are experiencing increasing levels of natural resource degradation. In the absence of primogeniture or active land markets, fragmentation of holdings is a natural consequence of population growth over time. Without appropriate policy incentives or technologies to increase yields and improve soil fertility, the result is soil mining and expansion of cultivation into sub-marginal areas. Predicted population increases in `poverty' systems - such as the Maize-Beans (Mesoamerican) System and the High Altitude Mixed (Central Andes) System - will only exacerbate these pressures. Frontier systems with lower population densities face a different set of constraints. Settlement in these systems has been historically sparse due, in part at least, to limitations in agricultural potential. Poor management practices can result in widespread damage and degradation to the natural resource base, as already recognised in the Cerrados (Extensive Mixed Farming System).

Although the solution to these problems may lie partly in interventions described in other sub-sections - including increased off-farm incomes - specific responses appropriate to densely populated systems do exist. Among these, the development and implementation of effective, community-level natural resource management plans is extremely important. Such interventions must include technical assistance and incentives for adoption, plus an emphasis on demonstrating rapidly realisable benefits from watershed improvements, forestry and other resource management activities. Other promising interventions centre upon moisture conserving technologies in dryer areas to combat the droughts and desertification (e.g. North-east Brazil and Central Andes), and effective watershed protection (Mesoamerica and Northern Andes). Both impacts are likely to become more common as a result of global climatic changes and risk reduction mechanisms will need to be strengthened.

For frontier systems, intervention priorities include the development of a detailed knowledge base on natural resources and their characteristics within the system, and the linkage of this knowledge base to planning tools and appropriate resource management approaches. This would be backed by research on the development of crop varieties adapted to limitations of frontier zones (e.g. aluminum tolerance, post-harvest characteristics), and dissemination of results. Of fundamental importance, however, are land settlement and development policies which encourage appropriate patterns of land use, through the use of instruments such as land taxes (regional and municipal); land grants; facilitated credit for investment or working capital; and, eligibility for support services (marketing, extension, veterinary services, etc.).

Larger farms are clearly better equipped in terms of resources to adjust to the new economic realities, but it should not be assumed that larger farms are always more competitive. The experience of Guatemala, for example, has shown that smallholders can be more efficient than commercial farms with respect to labour-intensive crops. All farming systems have potential for competitive production by small farmers, as well as for increased off-farm income generation, although the proportion achieving these goals will vary substantially from system to system. Where agro-ecological conditions are favourable and markets easily accessible, for example in the Coastal Plantation and Mixed Farming System, a relatively high proportion of smallholders may be able to follow these strategies. However, where agro-climatic conditions and tenure patterns impose severe constraints, fewer producers may be able to make the transition, and system exit assumes greater importance. This is likely to be the case in the High Altitude Mixed Farming System of the Central Andes and the Dryland Mixed Farming System of Northeast Brazil and the Yucatan.

Due to its abundant water and largely tropical and sub-tropical climate, the region is well endowed with high potential farming systems. However, many of the frontier areas - such as the Extensive Mixed Farming System of the Cerrados and Llanos, or the Forest Based System of the Amazon - possess fragile soils with nutritional limitations. And although development strategies will clearly be influenced by the agronomic potential of the system involved, the key role of land distribution in the region renders these differences less significant than in many other parts of the world. Even in high potential systems, improvements in yields or cropping intensity of traditional crops would have only a limited potential to reduce poverty among small farmers. By contrast, diversification and increased off-farm employment comprise essential strategic approaches for small producers in both high and low potential systems.

Conclusions

The Latin America and Caribbean region offers a sharp contrast between extensive frontier areas with low population densities and a high future growth potential, and established densely-populated systems - many with extensive poverty. Yet these two extremes share a number of common challenges that define a clear strategic focus for the region over the next thirty years, and although it is impossible based solely on the foregoing regional analysis to prescribe specific national actions, the overall situation demands a clear strategic focus for agricultural development and poverty reduction based on three, inter-linked, regional initiatives:

Sustainable resource management. Sustainable management of natural resources and the reversal of resource degradation are of prime importance, both in established high population systems and in rapidly growing frontier areas. However, a clear positive impact on producer incomes is essential if widespread adoption is to occur. Components include: dissemination of proven technologies for smallholders, notably green mulching, small-scale no-till, vegetative barriers, terracing and zero grazing; expanded attention to the selection, testing and dissemination of varieties appropriate for small producers, with an emphasis on permanent and tree crops; financing producer out-migration from unsustainable lands to permit reforestation; promotion of off-farm employment opportunities, so as to reduce pressure on densely populated areas; and, improving knowledge of frontier area lands and their fragile soils.

Improved resource access. A fundamental precondition for development in the region is improved access to - and control over - land by poorer rural populations. Components include: strengthened cadastral, titling and registration services; dispute resolution services; land banks financing consolidation of smallholdings, as well as purchase, breakup and resale of larger holdings; improved management measures for communally owned lands, including protection from invasion and colonisation; and, taxation policies that provide incentives for sustainable land use.

Increased small farm competitiveness. It is essential to increase the capacity of the smaller producers within farming systems, to respond adequately to trade liberalisation and market development. Components include: training of farmer groups in commercial management and administration techniques; improved technologies for high value production; reduced barriers to entry for small enterprises, as well as the strengthening of enterprise associations; the facilitation or part financing of the development of rural market-related infrastructure such as roads, communications and market facilities; and provision of incentives for relocation of agroprocessing and other enterprises to rural areas - including training of personnel, infrastructure provision and technical support to contracted producers.

1 United Nations Environmental Programme 2000.

2 Annual plus permanent crop areas.

3 FAO 2000a.

4 United Nations Environmental Programme 2000.

5 FAO estimates that an average food intake of 2000-2300 kcal per day is required for average activity.

6 FAO 1999b.

7 World Bank data show Brazil to have the highest income inequality in the world (47.9 percent controlled by the top 10 percent of the population), and that 11 of the 20 most inequitable countries in the world are in LAC.

8 Echevarria 2000.

9 See Annex 5 for an explanation of extreme or dollar poverty.

10 World Bank 2000a (Table 2).

11 Some of the highest GINI (inequity) coefficients for land ownership in the world are found in LAC, reaching over 0.9 in Peru, Paraguay and Venezuela and little less in Colombia and Brazil (Deininger 2000 pers. comm).

12 Unless otherwise stated, historical data is taken from the FAOSTAT statistical system, while future projections are largely extracted from FAO (2000a).

13 United Nations Population Division 1999.

14 Rural is defined to exclude cities with a population of more than 50000 inhabitants, and peri-urban areas with densities of more than 1000 persons/km2. Thus small towns would be included in the definition of rural.

15 This expansion may not be sustainable, however, unless careful resource management practices are followed.

16 FAO 2000c.

17 World Resources Institute 2000.

18 FAO 1995.

19 Many industrialized countries have shown growth in labour productivity as high as six percent per annum for periods of 20 years or more. Indeed, this is also the case for the Pampas, Southern Brazil, and the low valleys of Colombia.

20 World Bank 2000f.

21 These include the International Center for Tropical Agriculture (CIAT), the International Potato Center (CIP) and the International Maize and Wheat Improvement Center (CIMMYT).

22 The proportion of non-agricultural population within the system is higher than may be expected due to the presence of Brasilia, which lies within the system.

23 Spehar 2000.

24 This farming system boundaries has been defined to exclude the heavily urbanised coastal zone of Northeast Brazil.

25 Instituto Brasileiro de Geografia e Estatistica 1996.

26 An ejidario is a member of a community managed, state-owned ejido, or area of land, that bestows inheritable user rights, but not ownership, to its members. Since 1992, however, liberalisation of legal controls on ejidos has meant that more and more ejidarios have converted their use rights into private ownership.

27 Instituto Brasileiro de Geografia e Estatistica 1996.

28 Superintendency for the Development of the Northeast 1999.

29 Torres 1997.

30 Indigenous people account for 66 percent of total population in Guatemala and 29 percent in Mexico, but are lower in other Central American countries. Within the system itself, the proportion can be estimated at 60 to 80 percent.

31 Comité de Acción para el Desarrollo Social & Económico de Centroamerica/European Commision 1990.

32 Comité de Acción para el Desarrollo Social & Económico de Centroamerica/European Commision 1990.

33 FAO 1999a.

34 MAG/FAO/CIP/IICA 1996, p.63.

35 Comprising much of the states of Hidalgo, Tlaxcala, Guanajuato, Queretero, and Mexico.

36 Instituto Nacional de Estadistica Geografica e Informatica 1995.

37 MAG/FAO/CIP/IICA 1996.

38 Comité de Acción para el Desarrollo Social & Económico de Centroamerica/European Commision 1990.

39 Yields in the altiplano area of Central Mexico tend to be higher (from 2 to 2.5 t/ha) reflecting proximity to urban markets.

40 Secretaria de Agricultura, Ganaderia, Desarrollo Rural, Pesca y Alimentación 1998.

41 Secretaria de Agricultura, Ganaderia, Desarrollo Rural, Pesca y Alimentación 1998.

42 FAO 1999a.

43 Comité de Acción para el Desarrollo Social & Economico de Centroamerica/European Commisión (1990), p. 17 for 1964-1979 data; 1996 data from: Unidad Sectorial de Planificacion Agropecuaria y de Alimentación (1996).

44 In 1985 Panamá, with a population density of 28 persons/km2 recorded average maize yields of 0.93 t/ha while El Salvador, with a population density of 265 persons/km2 recorded 1.84 t/ha. Other Central American countries fell between these points.

45 World Bank 2000e.

46 Abstracted from Case Study 5, Annex 1.

47 Inter-American Development Bank 1999.

48 The latitude limit to potatoes cultivation is 4200 m at latitude 15° South.

49 More than 60 biotic zones of the 103 found in the world are represented in the Central Andes and Altiplano near Titicaca Lake.

50 Instituto Nacional de Estadistica e Informatica 1996.

51 Muñoz 1999.

52 The system rural population is estimated at 40 percent of total rural population in Peru and 45 percent in Bolivia.

53 The growth rate was calculated for the period 1985-1996 for Bolivia and 1990-1996 for Peru.

54 The present regulation of the land market was established in Peru by the land reform of 1953, which remains popular among small farmers, but restricts the consolidation of holdings.

55 Morales and Knapp 1987.

56 Well documented in CEPAL, FAO & GTZ (1998).

Export-oriented agroindustries frequently require products with higher input levels and costs than traditional crops.

58 de Grandi 2001.

59 A very innovative link between farmers and agro-industry has been established in the Ica Valley of Peru. Here a tomato paste processor rents land from local farmers, who work on their own land, but as employees of the processing plant.

60 Agro-industries generally prefer to be linked to the commercial sector, as transactions costs are lower than those incurred in the small farm sector.

61 Land market reform is unlikely in the medium term future as the necessary preparatory steps (as undertaken in Mexico as part of their land reform process) still have not been started in Bolivia and Peru.

62 There is also poverty reduction and growth potential within the Forest Based System, covering the majority of the Amazon basin and adjacent humid areas. Clearly major economic growth - including agriculture - can be expected to occur within this system over the next 30 years. However, serious natural resource and political constraints render its development highly controversial.

63 See Chapter 1 for a discussion of these household strategies for escape from poverty.