Chapter 9: Agriculture and Rural Poverty

Contents - Previous - Next

9.1 Introduction
9.2 Incidence of poverty
9.3 Rural poverty and agricultural growth
9.4 Interventions to improve access to land
9.5 Rural finance
9.6 Marketing
9.7 Selected direct anti-poverty interventions

9.1 Introduction

Over I billion people in the developing countries are poor, with a substantial majority of them living in rural areas. The development of agriculture can play a direct role in rural poverty alleviation, since the majority of rural poor depend on agricultural activity for providing the main source of their income and employment. It also contributes indirectly to alleviate rural poverty because the state of agriculture influences that of the non-farm rural economy. It can also play an important role in alleviating general poverty, since the agricultural sector makes a significant contribution to overall economic growth through its linkages with other sectors of the economy (Chapter 7). Furthermore, the increasing recognition of poverty as the root cause of problems of hunger and undernutrition assigns agricultural development a pivotal role in efforts to improve nutrition through increasing the quantity, quality and variety of food supplies, and through creating employment and income earning opportunities for the poor.

In this chapter, the focus is on a review of: (a) the empirical evidence concerning the effects of agricultural and more general economic growth on the incidence of poverty; (b) the role of interventions to improve access to land and develop the rural finance and marketing systems; and (c) the scope for direct interventions to create employment, or facilitate access to food. The related topic of the non-farm rural sector is covered in Chapter 7 and that of future developments in agricultural technology and research requirements in Chapters 4 and 12. This chapter was prepared in parallel with, and drew upon, a more comprehensive study of issues in rural poverty alleviation. This latter study (Gaiha, 1993) can be usefully consulted for more in-depth discussion of the topics in this chapter and in Chapter 10.

9.2 Incidence of poverty

The only available comprehensive estimates of poverty are those provided by the World Bank. The relevant table from the World Bank is reproduced here (Table 9.1). These estimates indicate that over 1.1 billion people in the developing countries were poor in 1990. This is approximately 30 percent of

Table 9.1 Estimates of the Magnitude and Depth of Poverty in the Developing World, 1985-90

Source: Reproduced by permission from World Bank (1933d).

Note. The poverty estimates are for 86 countries, representing about 90 percent of the population of developing countries. They have been updated from those used in The World Development Report 1990 and are based on national household sample surveys from 31 countries, representing roughly 80 percent of the population of developing countries, and on an econometric model to extrapolate poverty estimates to the remaining 55 countries. The estimates do not include the countries of Indochina or of the former Soviet Union. The poverty line is $31.23 per person per month at 1985 prices. It is derived from an international survey of poverty lines and represents the typical consumption standard of a number of low-income countries. The poverty line in local currency is chosen to have constant purchasing power parity across countries based on 1985 purchasing power parity exchange rates. The head count index is the percentage of the population below the poverty line. The poverty gap index is the distance of the average income of the poor below the poverty line (zero for the non-poor) expressed as a percentage of the poverty line. their total population. Apparently the incidence of poverty declined in the two decades to the mid-1980s, but no significant progress has been made since then and the absolute numbers of the poor increased (Lipton and Ravallion, 1993). The highest incidence of poverty is encountered in South Asia and sub-Saharan Africa, where nearly 50 percent of the population is estimated to be below the poverty line, with the former accounting for the bulk of the world's poor because of its large population. The data and methods used to estimate the incidence of poverty and to construct poverty profiles are subject to many limitations. Some methods are better suited than others as aids in the design of policies for alleviation of poverty or mitigation of its effects (for a review see Ravallion and Bidani, 1994).

Comprehensive estimates of rural poverty are not available. The available data for a number of developing countries indicate that the incidence of poverty is highest in the rural areas (Table 9.2). High concentrations of poverty also occur in urban areas. In Latin America and the Caribbean, much of the increase in total poverty during the crisis decade of the 1980s occurred in urban areas (World Bank, 1993d). Rural and urban poverty are linked through rural urban migration flows, as well as other factors.

Table 9.2 Rural poverty in the 1980s

Source: Reproduced by permission from World Bank (1990).

Evidence concerning access to education and health services, housing and sanitation also indicates that the incidence of poverty is comparatively more severe in rural areas. While much of the information comes from case studies, some comparative estimates on urban-rural differences in access to safe water and health services in the developing countries are given in Table 9.3.

From a policy perspective it is necessary to know the characteristics of the population groups making up the rural poor. A recent study on rural poverty identifies the small farmers, the landless, the women, the nomadic pastoralists, the artisanal fishermen, the indigenous ethnic groups and the displaced persons as the functionally vulnerable groups in the rural sector and indicates that alleviating poverty in each may require different policy instruments and approaches (Jazairy et al., 1992). Although few indicators are disaggregated by gender, the available data on literacy indicate wide gender disparities. Thus the female adult illiteracy rate in 1990 was 46 percent compared with a total rate of 36 percent. In 1987 there were only 81 females per 100 males in primary education and 75 females per 100 males in secondary education. Women's inferior educational endowments, combined with reduced access to assets, cultural constraints on engaging in certain types of employment and their continuing daily responsibilities for children, make it even harder for them to escape the poverty trap.

Table 9.3 Rural-urban disparities in access to selected services

Source: UNDP (1992).

The regional patterns of poverty presented here are fairly similar to those presented earlier for the incidence of undernutrition (Chapter 2). Data on related indicators such as the prevalence of underweight children or the incidence of micronutrient malnutrition further confirm the close association of deprivation in the nutrition/health nexus with that of poverty (FAD/WHO, 1992; WHO, 1992). It can be expected that progress in poverty alleviation would be largely reflected in amelioration of these poverty-related indicators of deprivation in the nutrition, health and other areas. However, public policy which targets directly the areas of nutrition, health, education and housing will also be essential. The fact that countries with similar per caput incomes have widely differing rates of undernutrition, morbidity and illiteracy, proves that there is considerable scope for such public policy. Moreover, the increasing recognition of the role of human resources in overall development strengthens the case for such public policies.

9.3 Rural poverty and agricultural growth

A general case can be made from several studies that when the economy grows the incidence of poverty (percent of population below the poverty line) tends to decline. Whether such a relationship holds in all cases, depends essentially on what happens to income distribution when the economy grows. It is possible that income distribution could become less equitable, thus offsetting partially or completely the potential benefits of such growth to the poor. Whether this happens is an empirical question. It is difficult to establish whether systematic relationships exist between overall economic growth and changes in the distribution of income. Comparative studies seem to indicate that economic growth is as likely to be associated with increases as with decreases in inequality. A recent study illustrates such diverging effects by comparing the experiences of rural India and Brazil (Daft and Ravallion, 1992). According to this study, in India the positive growth effects on poverty were reinforced by improvements in the distribution of income. In Brazil, on the other hand, a worsening of the income distribution led to one-half of the potential poverty reduction effects of growth being cancelled.

With regard to the relationship between agricultural growth and rural poverty, the empirical evidence seems to lend support to the common-sense proposition that the distribution of benefits from increased agricultural production will approximately reflect the initial distribution of productive assets and of access to inputs and services, as well as changes in the distribution of such assets brought about by the process of agricultural growth itself. It is, therefore, possible for agricultural growth to be associated with worsening of the distribution of income. If, moreover, the deterioration is sufficiently severe, it can even cause parts of the rural population to become poorer in absolute terms. This seems to have been the case in Latin America in the 1980s when the incomes of the rural poor declined despite considerable increases in aggregate agricultural production. In this region, the most severely affected category-landless labourers-suffered a 23 percent fall in real wages between 1980 and 1987 (UN, 1992).

The specific impacts of agricultural growth on different socioeconomic categories of rural producers and labourers, as well as the mechanisms through which these impacts are mediated, depend on the nature of the growth processes and the structural factors underlying the social organization in rural areas. Much of the empirical knowledge on these matters comes from studies (mostly from India) of situations in which rapid agricultural growth occurred as a result of the spread of the green revolution.

The evidence in these studies indicates that the advent of the new technology, in the form of biochemical innovations, was associated with reductions in rural poverty, e.g. the proportion of the poor among the cultivating households declined and so did the severity of the poverty of those households which remained poor. At the same time, however, some of the poor became poorer and some of the non-poor were pushed below the poverty line. Much of the adverse impact on selected groups of the rural population was the consequence of the initial inequality in access to land. Since smallholders were impeded by restricted credit markets, input supply problems, limited access to extension services, risk aversion and tenure insecurity, the benefits of the new technology tended to accrue mainly to large landholders. The latter expanded acreage through resumption of land for personal cultivation (by evicting tenants) and/ or through leasing or purchase of land from small landowners. As a result, the distribution of gross cropped area became more unequal and the percentage of landless households increased from 25 percent to 35 percent (data and analysis for rural India for 1968-70 from Gaiha, 1987).

On the other hand, the longer period (1974-84) evidence from North Arcot (a small region in South India) indicates that where small-scale, owner occupied farms dominated, the effects were more favourable to the poor, largely because there was a conducive institutional setting, in which state and local governments made credit and modern inputs available to small farmers, and invested heavily in infrastructure (Hazell and Ramasamy, 1991). Even then, the early adopters of the modern varieties (MVs) were typically large farmers. But eventually, over 90 percent of the paddy area was planted with the MVs, with no systematic differences by farm size, and similar yields were obtained on large- and small-scale farms alike. There was no evidence of a general increase in land concentration or of loss of land by smallholders and there were sizeable absolute gains in income of all household categories and a decline in the incidence of absolute poverty (for more general discussion of technical change and rural poverty see Lipton and Longhurst, 1989).

In addition to the fact that agricultural growth may be associated with at least some parts of the rural population becoming worse-off economically, concern with rural poverty must also encompass those sections which are chronically poor and too marginal to be affected by agricultural growth, one way or the other. These sections comprise people living in remote, resource poor regions, without any infrastructure; backward sections of society, debarred from owning assets, denied access to education and condemned to menial occupations; and the disabled and the aged, incapable of augmenting their incomes above a bare subsistence level.

Over a longer term perspective, the sustainability issues related to agricultural growth assume increasing importance for the rural poverty problem. Continued population growth in the context of rural poverty, and in particular when it occurs in conditions of inequality of access to land, tends to push the rural poor to expand agriculture into ecologically fragile areas. This causes deforestation and exploitation of the land in ways which damage its productive potential. This process sets the stage for continued poverty of the part of the population concerned. Examples of this process abound, from the Himalayan and Andean ecosystems, large parts of Africa and the colonization experiences (spontaneous or officially sponsored) of the tropical rainforest (e.g. Brazil, Indonesia). To the extent that this process takes place in parallel with production increases in the higher potential areas, it provides another example of agricultural growth bypassing, and often making worse-off, part of the rural poor, e.g. by lowering the prices they receive for their own production. The sustainability issue and its relation to rural poverty is also relevant for the better off agricultural lands, in so far as more intensive agriculture, if not carefully managed, can reduce the productive potential of the land and water resources and threaten the sustainability of the poverty-reduction effects obtained initially.

It can be concluded that, on balance, agricultural growth can be expected to bring about reductions in rural poverty (Lipton and Ravallion, 1993).

However, some parts of the rural population, including the rural poor, may become worse-off economically in the process of agricultural growth. The structural characteristics of the rural economy at the inception of agricultural growth play a predominant role in the distribution of benefits from higher production. Scale-neutral technical change can benefit large and small farmers alike, if institutional rigidities do not stand in the way. With regard to the latter, an activist public policy in the area of institutions, research, credit, etc., can be instrumental in ensuring the wider spread of benefits.

9.4 Interventions to improve access to land

It was noted above that the structural characteristics of the rural economy-particularly land ownership and land tenure systems-play a decisive role in determining the distribution of benefits and the rural poverty effects of agricultural growth. The pros and cons of interventions and the lessons of experience from efforts to enhance the access of the poor to land are discussed below with regard to: (a) redistribution of ownership rights; (b) regulation of tenancy contracts; and (c) the role of land titling.

Land redistribution

The most recent attempt to take stock of progress in redistributive land reform was undertaken in 1991 for the quadrennial FAO report on progress under the Programme of Action of the World Conference on Agrarian Reform and Rural Development-WCARRD (FAO, 1991b). The report concludes that progress has been limited, mainly because the implementation of land distribution programmes was strongly affected by political realities.

The equity case for land redistribution from large landowners to the landless and/or small owners of land rests on at least three considerations: (a) the landless/small owners are usually poorer than large landowners; (b) in general, but with important exceptions, total employment and production per hectare increases as farm size decreases; and (c) inequality in the distribution of land conditions the poverty effects of agricultural growth not only because of the resulting unequal distribution of the income attributed to land but also because it breeds social stratification patterns inimical to the poor in many other areas, e.g. the distribution of political power or access to credit.

The efficiency case requires that land redistribution increases, or at least does not reduce, farm output and the potential for growth. Since often an inverse relationship between farm size and output per hectare is observed, land redistribution has the potential of increasing output. In most cases, the inverse relationship is due to a higher cropping intensity, and a more labour-intensive and higher-value crop mix, on smaller farms. Land quality differentials may account for part of this inverse relationship, e.g. when large farms contain a higher proportion of inferior quality land compared with the small farms.

Controlling for these differences in land quality can attenuate the inverse relationship, but it does not cancel it.

The inverse relationship, however, can be modified with the onset of green revolution where land augmenting technology tends to equalize the yields attained by small and large farms. But differences in labour use per hectare tend to remain, and this is an argument in support of the continued relevance of the efficiency case for land reform even under the new technology. This is because in the context of rural market imperfections, particularly of labour markets, large farms tend to use factor proportions skewed in favour of capital (mechanization) and against labour beyond the proportions dictated by social allocative efficiency considerations, even in the absence of policies distorting prices in favour of capital, i.e. even when governments have implemented policies to "get prices right". This is a rather important consideration in the debate on what is the appropriate mix of policies for getting prices right and those for bringing about changes in basic structural characteristics of the rural economies (see Platteau, 1992).

The case for a more equal land distribution is further strengthened when linkages with non-agricultural activities in the rural sector are considered. Some South Asian evidence suggests that in villages with relatively equal land (and farm income) distributions, the share of locally produced labour-intensive non-farm goods in total consumption is higher compared with that in villages characterized by higher inequality. Thus, it can be expected that a more equal land distribution contributes to rural poverty alleviation indirectly via its effects on rural non-farm employment.

The extent to which the changes brought about by policy interventions to redistribute land prove durable and the persistence of the unavoidable upheavals in the production structures are important elements in the land reform process. Chile's experience with land reform is instructive. Initiated during 1964 to 1970, land redistribution was extended during 1970 to 1973 and reversed during 1973 to 1976. Expropriation of land was stopped. Land worked by the campesinos in the transition period (asentamientos) was partly confirmed to them, partly restored to the previous owners and the balance was sold at public auction. New farm enterprises with strong financial resources entered the sector. The modern sector became strongly export oriented. In parallel, the lack of technical and credit support induced many small farmers to retreat into more traditional patterns of subsistence production. Producer cooperatives were dismantled. But more than half of the land expropriated was ultimately confirmed to the beneficiaries (Jarvis, 1989; Gomez and Echenique, 1991). The crisis of 1982-83 resulted in a notable reversal of policy. From 1987 the Government started financing a considerable part of the costs of technical assistance to smallholder beneficiaries of agrarian reform (Meller, 1988).

Policy interventions also led to structural changes in production in the Philippines where land reform was limited to rice and maize lands which were managed predominantly through tenancy arrangements. This limitation induced landlords to divert their land to other crops, often at the expense of both efficiency and equity. For instance, rice lands with higher income-earning and labour absorptive capacity were converted to less labour-intensive crops such as coconuts.

Success is more likely to be achieved when distributing state-owned land, where resistance is less than when trying to redistribute land away from large landholders. In the Philippines, between 1987 and 1990, two-thirds of the total land distribution targets of state-owned land were fulfilled. However, only 2 percent of the target for private land redistribution was achieved, due to disputes with owners over appropriate compensation (information from country working papers for the preparation of FAO, 1991b).

Another important issue is the extent to which the state of modernization of agriculture is related to the process and chances of success of redistributive reforms. It was noted above that under the new technology, the inverse relationship between yields and farm size tends to weaken. This can weaken the efficiency case for reform. The experience also shows that the threat of reform can be instrumental in prompting larger farmers to promote modernization as a defensive action. For example, in some Latin American countries, the threat of expropriation and incentive policies (input subsidies, tax breaks) were successful in inducing large farms to modernize, hence increasing agricultural output. One outcome of this modernization, however, was to render expropriation with compensation very costly. Further, as was recently found in Colombia, larger farmers often successfully used their influence to extract promises from the government that their land would not be expropriated if they modernized. As a consequence, redistribution of land to the poor was negligible. Interestingly, modernization of agriculture had the opposite effect on land redistribution in the Philippines. There, compensation had been fixed at pre-green revolution land values and the economic gains associated with modern seed and fertilizer rice technology allowed the beneficiaries to capture significant economic surpluses.

Government support to the beneficiaries of land reform is an essential component of the whole operation. The case of Mexico illustrates the pitfalls of unsupported land reform. There, the reform was not accompanied by significant productivity gains as most of the small farmers of the agrarian reform sector were left with non-irrigated land, and government support policies were not always effective. Even in the case of small farmers who had some irrigated lands, the government-supported cooperatives did not provide the needed services. By contrast, most of the irrigated land was left in the medium and large holdings and government support policies were heavily skewed in their favour.

The preceding discussion is more relevant to cases where the reform aims at creating more equal distribution of landholdings to be owned and operated as individual units. But there have been experiences with alternative ownership and exploitation structures in the post-reform periods. The creation of producer cooperatives is a case in point. Here the general conclusion is that the experiments with producer cooperatives led to disappointing results particularly in some Latin American countries. In Peru, for example, earlier reforms had led to some two-thirds of agricultural land being controlled by producer cooperatives in 1979. However, these cooperatives suffered from serious diseconomies of scale and work incentive problems and many were broken up in the early 1980s and the land was distributed as individual holdings.

In Nicaragua, producer cooperatives were initially thought to be better suited for the large-scale production of export products such as coffee, cotton and beef. Subsequently, the emphasis on land distribution moved away from the establishment of producer cooperatives towards direct distribution to individuals. This followed the realization that dividing large farms into smaller holdings would not necessarily cause a reduction in output if adequate credit and other support were provided.

Land reform will continue to be a relevant issue in the future in the quest for poverty alleviation and more equity in the rural areas. However, it may cease to be the burning issue it once was, especially in those countries where the nonagricultural sector will be increasingly the main source of additional employment and income-earning opportunities and land will lose its primacy as the main form of wealth. As indicated in Chapter 3, a number of developing countries are expected to have over the next 20 years economic growth rates high enough to imply that the bulk of additional wealth will be generated in the non-agricultural sector. It can be expected that trends for increasing farm size will emerge in such situations, just as they did in the developed countries. This is because pressures build up for the incomes of people in farming to follow (though not necessarily become equal to) those that can be potentially earned in the non-agricultural sector. A combination of more land per person and higher income earned per unit of land is normally the result of these pressures, brought about by technological change and the flow of labour from agriculture to other activities, though not necessarily always urban-based ones.

Many developing countries may not, however, enter this phase of transition in the foreseeable future. In many low-income countries with unfavourable overall growth prospects, high incidence of rural poverty and continued high population growth rates, the number of people seeking to make a living in agriculture will continue to grow. In these circumstances, the distribution of land ownership and the potential role of interventions to change it towards patterns more conducive to poverty alleviation and enhanced equity will continue to be live issues.

It is to be noted, however, that a more equitable distribution of a growing agricultural income can only contribute in limited ways to make significant dents in rural poverty directly, so long as the population dependent on agriculture continues to grow. This is because even an optimistic agricultural growth assumption (e.g. around 3.5 percent p.a. in gross value terms) will probably mean a growth in average per caput incomes of the growing agricultural population of under 2.0 percent p.a. Welcome as such an outcome is for its potential for reducing rural poverty, it cannot be compared with the long-term benefits obtainable from a combination of vigorous non-agricultural growth and declining agricultural population.

Tenancy reforms

Tenancy is the term commonly used to refer to those land tenure arrangements (legal and customary) which regulate access to land in forms other than acquisition of ownership rights. It refers to all those situations where a person's access to land is through some arrangement with another person or entity who enjoys superior land rights. Policies to reform tenancy arrangements are often predicated on grounds of both efficiency and equity or poverty alleviation. It must, however, be noted that regulation of tenancy contracts could result in a contraction of the supply of land for tenancy and may thus lead to an increase in the number of the landless in rural areas as tenants are evicted (as happened in the Philippines, India and Sri Lanka in South Asia, and numerous Latin American countries; Osmani, 1988).

One of the major issues concerns the relative merits of alternative arrangements for renting land, e.g. fixed payment for a definite period of time (fixed rent), sharecropping, labour service or mixtures thereof. A major thrust of tenancy reform policies has been to restrict or prohibit sharecropping contracts. The concern with sharecropping was also associated with the feudal conditions existing in agrarian societies. In retrospect, however, it has been found that such policies can have unintended negative effects on the poor. There is now increasing realization that under special circumstances, sharecropping contracts can be an efficient vehicle for risk-sharing with positive impacts on both efficiency and equity. For example, evidence shows that there is a higher implicit rent on sharecropped rented land (which may reflect a risk premium) and a higher frequency of share tenancy in areas with variable weather. Cost-sharing arrangements in sharecropping can allow poor farmers to have access to certain inputs which they would not otherwise obtain because of their limited access to financial resources. For example, because fixed rents on leased lands generally must be paid in advance, poor farmers without access to credit may be prevented from leasing land. This constraint is overcome under share tenancy as payments are made only at harvest.

The experiences of tenancy reform in China, Laos and Vietnam indicate that changing from forms of socialized farming systems to ones based on the household economy, where allocative decisions, ownership of other productive means, and longer land-use rights are given to the individual households, can bring substantial efficiency and equity gains. In China, for example, the increases in agricultural production, and the development of the non-farm rural economy were spectacular and led to strong broad-based rural economic growth and significant reductions in the incidence of rural poverty. This process has slowed down in recent years. Vietnam, after the implementation of land tenure reforms, became self-sufficient in food grains for the first time and subsequently a net exporter of rice.

Concerning Africa, recent evidence suggests that most indigenous land tenure systems are adapting efficiently to changes in resource availability. As a policy option, it may therefore be better to concentrate on providing an appropriate legal and institutional environment for more efficient transactions than to restrict land sales and rental markets with tenancy legislation.

Land titling

Three arguments are usually put forward in support of land titling:

1. Titling is assumed to increase tenure security with a view to promoting investment in land and water conservation and capital inputs, and adoption, where appropriate, of permanent crops
2. By providing collateral, titles may increase access to institutional credit
3. Land titles are considered necessary for the development of land markets which are essential for promoting commercial development of agriculture.

Recent evidence from Africa suggests that the expected positive relationship between tenure security or the extent of land rights, particularly inheritance, and long-term investment in the form of land improvements hold in some areas, but not in others. Further, formal ownership titles are not necessary for tenure security, since under most communal tenure systems, a farmer has usage rights to specific plots which are often hereditary. In other places where such hereditary rights are not available, however, lack of title does bias production decisions in favour of short-cycle crops. Although some evidence from Africa shows that titling may not have a significant effect on access to credit, evidence from some countries in Asia and Latin America is rather more favourable, with significant increases in access to institutional credit after land titling.

Land titling often aggravates inequality since wealthier and more influential individuals can obtain greater rights than they formerly enjoyed. In these conditions the poor are exposed to increased risk of landlessness and loss of common property resources after implementation. The position of women in land titling requires a special focus. Land titling programmes, for example, tend to concentrate on the land parcel as the relevant target unit, paying little attention to distribution of land rights within the household. Granting land titles to male household heads tends to diminish women's control over land usage and transfers. Land titling may also cause the loss of secondary rights in land, such as the right to gather fuelwood, which are of particular importance to women. There is thus a strong case for designing land legislation and reforms to target women as direct beneficiaries.

Advances in electronic data processing have opened up new horizons for traditional land titling and cadastre systems. There is no inherent reason why land registration and cadastre cannot be made sensitive to special cultural conditions on the one hand, and to equity and efficiency criteria of sustainable rural development, on the other. With the evolution currently taking place in thinking about land tenure and the shift from social to private property models, the compilation of land records (land registration and cadastre), and local community involvement in land regulation (taxation, zoning, etc.) can be expected to be major activity areas in the coming decades.

Contents - Previous - Next