2.4 Issues of agricultural resources, environment and sustainability

Contents - Previous - Next

General considerations

Concern with the state of the environment and the dwindling quantity (per caput) of land and water resources as well as their degradation requires that the conclusions of the preceding section be amplified to address questions like the following: To what extent may the resource and environmental constraints impinge on the prospects for increasing food supplies and assuring access to food by all, the very essence of food security? Can such progress be achieved while ensuring that the gains made and the potential for further gains are maintained for future generations, the very essence of sustainability? The rest of this chapter endeavours to put the overall issue in a proper perspective. More specific discussion is to be found in Chapters 4, 11, 12 and 13. These chapters provide estimates of the pressures that are likely to be put on agricultural resources in the process of increasing production in the period to 2010. They also explore the options for policy responses to minimize the unavoidable trade-offs between increasing production and pressures on the environment.

The preceding section singled out the rate of increase in per caput food supplies of countries and population groups with inadequate access to food as a practical proxy for measuring progress towards solving problems of food security and undernutrition. It also highlighted a number of interdependent factors as being instrumental in increasing per caput food supplies: poverty reducing economic growth; the multiple role of agricultural growth in the majority of the developing countries (increasing food supplies and providing employment and income-earning opportunities to the poor, both directly and indirectly via the growth linkages of agriculture); enhanced capacity to import food; and public policy.

It follows, therefore, that important as the agricultural resource constraints are in conditioning the prospects for food production and generation of incomes in agriculture, the wider environmental constraints can also affect in important ways the prospects of eliminating undernutrition because of their possible effects in restraining the overall economic growth rate and the potential for reducing poverty. For example, reducing emissions of greenhouse gases and the existence of non-agricultural resource constraints may cause the world economic growth rate to be lower than what it would otherwise be. Those low-income countries which depend, actually or potentially, on a buoyant world economy for their development will find it more difficult to improve their economic growth rates and reduce poverty. Moreover, the adverse environmental impacts (local, but with global implications) which often accompany the accelerated growth in the use of energy in the transition from low- to middle-income levels in the low-income countries will tend to make such transition more difficult.' These are examples of how the more general environmental and resource constraints, and not only the agricultural ones, may impinge on the prospects for reducing undernutrition.

Given the above considerations, the extent to which the agricultural resources are adequate or otherwise to produce as much food as required to increase per caput food supplies for a growing population in sustainable ways must be examined in the context of these grander themes concerning the overall resource and environmental constraints. It may well be, for example, that at the global level the binding constraints would not be those impinging directly on the production of food but rather those standing in the way of achieving economic growth rates and patterns adequate to eliminate poverty in the not too-distant future.

There is another sense in which the agricultural resources may not be the binding constraint to making progress towards elimination of undernutrition, at least not in a global context and in the longer term. It was noted in the preceding section that: (a) progress achieved in the last few decades points to ever increasing per caput food consumption levels; (b) a considerable number of countries have made the transition from low and medium-low levels to medium-high ones; and (c) beyond these levels the growth of per caput consumption of food tends to slow down before stopping altogether when it reaches physiologically maximum levels.

It follows that a combination of continuation of these developments in per caput food consumption and the expected slowdown in population growth will eventually translate into a slowdown in the rate at which pressures are exerted on world agricultural resources for increasing food production. That is, the world may reach at some future date a stage when very little additional growth in global food production would be necessary to maintain adequate food supplies for all. The experience of many developed countries in which there is little scope for further expansion of aggregate agricultural output for domestic use, and in which land has often to be taken out of production, is telling. The gist of the matter is, therefore, whether this stage can be reached while maintaining sufficient agricultural resources for continuing production at a nearly steady-state level in sustainable ways and with enough natural habitat intact for it to continue providing its essential life-support functions. Some further discussion on these matters is given in Chapter 3 in an attempt at speculating on possible developments beyond the year 2010.

The preceding discussion provides the background for discussing in the remainder of this chapter the relative significance of agricultural resource constraints for making progress towards reducing undernutrition. This means essentially the prospects for increasing per caput food supplies for the population groups with inadequate access to food, taking into account both the supply side of the problem (e.g. can enough food be produced in sustainable ways?) as well as the demand side (e.g. in what ways may agricultural resource constraints play a role in the process of enhancing access to food by the poor?). The link between the two sides is provided by the fact that the great majority of the poor depend for employment and income on the exploitation of those very agricultural resources.

Land and water resources in the quest for sustainable responses to the food problem

In Section 2.3, the food problem was defined in terms of a few measurable variables (per caput food supplies by country, incidence of undernutrition) and analysed in terms of others (agricultural production, per caput incomes, distribution, food imports). In this analysis, the possible role of agricultural resources was not considered explicitly. It can be hypothesized that such role is subsumed in that of some of the variables considered, notably agricultural production and per caput incomes. The widespread concerns with agricultural resources, the environment and sustainability require that an attempt be made to consider such a role more directly. This issue is addressed in terms of: (a) what is known about the land and water resources; and (b) how the constraints relating to these resources may enter the determination of the rate of progress towards solving the food problem.

The state of knowledge on the extent and use of agricultural resources and the historical evolution of such use leaves much to be desired. For example, the data on cropping patterns by agroecological zones used in this study had to be compiled from fragmented sources and supplemented by expert judgements (see Chapter 4). Likewise, the data on the state of degradation of irrigated lands or erosion of rainfed lands are limited and very little is known about how such states have been evolving over time. Furthermore, there is a dearth of systematic information on yet unexploited irrigation potential. The data on water resources (river flows, acquifers), such as they exist, need to be interfaced with those on land (terrain, soil, etc., characteristics) and analysed in the context of a host of socioeconomic factors affecting their use, before comprehensive and credible estimates of the potential for irrigation expansion can be obtained.

Potential of land and water resources for rainied crop production in the developing countries

Given these data shortcomings, resort may be made to the second best option offered by the rather more systematic data in the soil and climatic inventories used in FAO's Agroecological Zones (AEZ) work for the developing countries. These data (recently re-elaborated for this study) permit the derivation of estimates of land stocks of varying quality with potential for growing crops under rainfed conditions. Examples of the kind of data thus obtained are given in Table 2.4 for South Asia and tropical South America (more comprehensive data for the developing countries are shown in Chapter 4).

It is noted that such evaluation indicates the potential for rainfed crop production of land in its natural state, i.e. not taking into account improvements or deteriorations brought about by human activity. The results should, therefore, be read with this caveat in mind, because it is well known that much of the land in agricultural use has been modified in the course of time, for better or for worse, by human intervention. Some account of such alterations is taken in the process of accounting for irrigated land, e.g. in the extreme, desert land with no agricultural potential in its natural state is added to the agricultural land if it has been irrigated.

The above data on land convey also significant information on water resources for agriculture. On the latter subject, a distinction is made between water supplies from rainfall which are directly utilizable in rainfed agriculture if falling or ending up in soils with appropriate qualities, in particular the capacity to retain humidity in the root zone for the length of time required by the growth cycles of the different crops; and the part of rainfall which feeds into water bodies like rivers and acquifers and which, together with stocks of fossil water, can be used for agriculture only through human intervention (irrigation). As noted, the data on this latter resource are not adequate for a full evaluation. But the data on supplies from directly utilizable rainfall are part and parcel of the evaluation of the above mentioned AEZ data set on land resources for agriculture. This is because when evaluating any piece of land for its suitability to produce one or more crops at "acceptable" yields under alternative technologies (see notes to Table 2.4), the rainfall regime and the soil's water holding capacity are key elements in the solution.

In the end, declaring that, for example, South Asia has some 50 million ha of land with terrain/soil characteristics "very suitable" or "suitable" for rainfed agriculture which receive rainfall and have water holding characteristics sufficient for a growing period of 180-269 days, is equivalent to making a statement on water availabilities for rainfed agriculture. This is counting not rainfall in abstract terms (e.g. in millimetres) but more precisely that part which ends up in soils with other desirable characteristics for farming. Barring changes in the rainfall regimes and the quality of soils, it can be assumed that this is a perennial resource. This estimate is perhaps more robust than those of water resources for irrigation. The latter are subject to greater uncertainty concerning their permanence over time because of: (a) possible reduction of the water supplies due to overexploitation; (b) the risk of deterioration of the irrigated lands (waterlogging, salinization) and of infrastructure (siltation, etc.); and (c) possible diversion of water supplies to competing non-agricultural uses.

In practice, therefore, the classification of land with agricultural potential according to the length of growing period (LOP) criterion goes some way towards defining the water constraints for agriculture. It has the added advantage that the estimates thus obtained are less subject to uncertainty compared with those referring to water resources for irrigation, as discussed in Chapter 4. The importance of improving the data and knowledge about these aspects is evident from the fact that at present some 37 percent of the gross value of crop output (and 50 percent of that of cereals) of the developing countries comes from irrigated lands. The estimates thus derived for the land and water availabilities for rainfed agriculture are supplemented with two pieces of additional relevant information, viz.: (a) the extent to which they are irrigated, including an estimation of irrigation of land not suitable for rainfed production in its natural state (rows B, C in Table 2.4); and (b) if they are used currently for crop production (not including land used for fodder, whether cultivated or natural grass for grazing).

This estimate of the current use status makes it possible to obtain as residual the land with crop production potential of varying quality which is not yet in agricultural use. However, its mere existence does not mean that it should be considered as available for expansion of crop production in the future (see below). The stark contrast between the situations in South Asia and tropical South America is evident (Table 2.4). It becomes even starker when expressed in terms of population densities, given that South Asia's population is 1.1 billion and that of tropical South America only 240 million. Moreover, South Asia has 65 percent of its economically active population in agriculture (265 million), while tropical South America has only 25 percent (22 million). There are, therefore, even starker differences in terms of agricultural land actually or potentially available per person in the agricultural labour force. This latter variable is the key one for understanding the forces that may shape the future in terms of the population-resources balance. As already noted, this balance has two main dimensions: (a) how much more food must be produced, which is directly linked to the growth of total population and the per caput consumption of food; and (b) how many people are, or will be, making a living out of the exploitation of agricultural resources. The relevant variable here is the size of the population economically active in agriculture.

Table 2.4 Examples of land balance sheet: South Asia and tropical South America


Moisture regime (LGP, days)*

Land quality

South Asia††

Tropical South America‡‡


In crop prod. use



In crop prod. use


Land with crop prod. potential by class                
1 . Dry semi-arid 75-119 VS,S,MS 29.2 22.1 7.1 9.9 3.5 6.3
2. Moist semi-arid 120-179 VS,S 82.4 61.0 21.4 32.2 10.8 21.5
3. Sub-humid 180-269 VS,S 50.6 45.4 5.2 121.5 47.4 74.2
4. Humid 270+ VS,S 6.0 25.1 3.3 329.5 44.0 516.2
5. Marginal, moist semi-arid, sub-humid, humid 120+ MS 22.4 230.7
6. Fluvisols/gleysols Nat. flooded VS,S 21.3 21.6 0.6 65.2 8.1 99.1
7. Marginal fluvisols/gleysols Nat. flooded MS 0.9 42.0
A. Total 1 to 7     212.8 175.2 37.6 831.0 113.7 717.3
B. --of which irrigated       48.1     4.5  
C. Additional irrig. from hyperarid land     15.3 15.3   0.9 0.9  
D. Total land with crop prod. potential (A+ C)     228.1 190.5 37.6 831.9 114.6 717.3
E. Land without crop prod. potential     204.9     532.7    
E.1 –Hyperarid     45.6     22.7    
E.2 –Other constraints     159.3     510.0    
F. Total forest area     61.1     802.9    
F.1 –Could be on land without crop prod. potential     55.0     310.2    
F.2 -Minimum forest area on land with crop prod. potential||         6.1     492.7
G. Total in human settlements and Infrastructure**     25.7 (0.023 ha/ person)     11.3 (0.046 ha/ person)    
G.1 -On land without crop prod. potential     8.9     4.7    
G.2 -Balance on land with crop prod. potential         16.8     6.6
H. Protected areas     15.6     143.6    
H.1 -On land without crop prod. potential     10.3     52.1    
H.2 -Balance on land with crop prod. potential         5.3     91.5

*LGP =length of growing period is the number of days during the year when temperature and rainfed soil moisture permit plant growth.
VS = very suitable, in the sense that obtainable yields can be 80% or higher of those obtainable in land without constraints; S - suitable, yields 40-80% of the constraint-free ones; MS =marginally suitable, yields 20-40%.
Could be partly in grazing use; additional grazing land is the fallow part of the land used in crop production.
Data from the FRA 90 assessment (see Chapter 5 on forestry).
Maximum amount of land on which trees, but not crops, could exist.
Balance of forest area (F2 = F - F1) which by necessity must be on land with crop prod. potential. As such, it is the minimum overlap between forest and agricultural land. It could be much larger in reality.
**For method of estimating land in human settlements and infrastructure see Chapter 4.
Bangladesh, India, Nepal, Pakistan, Sri Lanka.
Bolivia, Brazil, Colombia, Ecuador, Paraguay, Peru, Venezuela.

As noted, the existence of land with crop production potential does not necessarily mean that such land may be so used. In the first place, part of it is used for human settlements (habitation, industry, infrastructures) and more of it will be so used in the future following population growth. There are no reliable data on how much land is occupied by human settlements. Sporadic data for some countries have been used to derive the estimates shown in Table 2.4 (Row G, for methods of estimation see Chapter 4). These estimates are subject to an unknown, though probably very large, margin of error.

Secondly, part of the area with crop production potential overlaps with forest. The extent of this overlap is not precisely known. The table provides minimum estimates obtained by first deducting from the total forest area the part which, on agroecological criteria, could exist on the land without agricultural potential (Row F.1 in Table 2.4). The balance of the total forest area must be by definition on the land with agricultural potential (Row F.2). This is a minimum estimate of overlap and the actual one is probably much larger. Considerations of environment and sustainability dictate that a good part of the land with forest should not be considered for prospective agricultural expansion. Indeed some areas, not always under forest, are legally protected (Rows H.1, H.2 in Table 2.4). Moreover, forest lands do contribute to food security and any gains in food production from conversion to agriculture must be netted out for the food security losses incurred. This is because significant numbers of people, much beyond the communities of forest dwellers, depend on sustainably managed forests and trees either as a source of complementary food supplies, or even more importantly, as a source of off farm income.

It is evident from the preceding discussion that the question "How much more land can be drawn into crop production?" cannot be answered only, or even predominantly, on the basis of the data presented here. For one thing, the extent of, perhaps multiple, overlap between land with crop production potential, forest, human settlements and protected areas is not known with an acceptable margin of error. For another, more of the land with crop production potential (whether presently used or not) will be occupied by human settlements in the future following population growth. Further, a host of other factors (socioeconomic, technological, etc.) will determine the combination of area expansion and growth of yields that will underpin the future growth of production. In the event, and as explained in Chapter 4, the additional land to come under crop production by year 2010 may be about 4 million ha in South Asia and 20 million ha in tropical South America. In addition, continued population growth would probably require additional land for human settlements and infrastructures of some 9 million ha and 3 million ha in the two regions, respectively.

Declining land/person ratios

As noted, the continuous decline of agricultural resources per caput following population growth is one of the major reasons why concern is expressed in relation to the population-food supply balance. The other reason has to do with the deterioration of the quality and food production potential of the resources. The data discussed above may be used to shed some light on the nature and significance of the decline in the resources/ person ratio (hereafter referred to as land/person ratio). The values of this latter ratio in the different developing countries span a very wide range, from the very low to the very high. For example, at the very low end are countries like Egypt, Mauritius, Rwanda, etc., with ratios of land-in-use of under 0.1 ha per person in the total population and virtually zero reserves for further expansion. At the other extreme, countries like Argentina or the Central African Republic (CAR) have land-in-use ratios of close to I ha per person and considerable reserves."

With population growth, more and more countries will be shifting closer to values of the land/person ratios typical of those encountered currently in the land-scarce countries. Does this matter for their food and nutrition? An approach to obtaining a first partial answer is to examine if the currently land scarce countries are worse-off nutritionally (in terms of per caput food availabilities) compared with the more land-abundant ones. This is attempted in Table 2.5 with the land/person ratios adjusted as indicated in Note 11. The picture emerging from the table just confirms what is known, i.e. there is no apparent close relationship between the land/person ratios and per caput food supplies. If anything, many land-abundant countries have low per caput food supplies, while most of the nutritionally better-off countries seem to be precisely those with the highest land scarcities. At the same time, most of these latter countries have considerable cereal imports.

Should this evidence be interpreted to mean that the perceived threat of ever declining land/person ratios is misplaced? Not necessarily. In the first place, the national land/person ratio, even if adjusted for land quality differentials, is just one of the many factors that determine per caput food supplies. Its importance cannot be evidenced without an analysis accounting for the role of these other factors (essentially respecting the clause "other things being equal"). Secondly, the high dependence of the land-scarce, good-nutrition countries on imported cereals means that the perceived threat of the declining land/person ratios must be understood in a global context. That is, a decline in an individual country's land/person ratio may not threaten its own food welfare provided there is enough land elsewhere (in the actual or potential exporting countries) to keep the global land/person ratio from falling below (unknown) critical minimum values; and, of course, provided that the people in the land-scarce country do not depend in a major way on the local land and water resources for a living. Countries like Korea (Rep.) and Mauritius are in this class.

Table 2.5 Distribution of developing countries by per Caput land-in-use and food supplies, data for 1988/90

Land per caput* (ha)

Food supplies per Caput (cal/day)

Under 2000 2000-2100 2100-2300 2300-2500 2500-2700 Over 2700
Under 0.10 Rwanda (8)       Jamaica(140) T. Tobago(213)
Jordan (338)
Korea, Rep. (225)
Mauritius (190)
0.10-0.19 Burundi(4) Kenya(1) Yemen(134) Venezuela(126) Indonesia(10) Egypt(163)
Somalia(29) Bangladesh (20) Lesotho(117) Dominican Rep. (94) Lebanon(188)
Namibia(49) Haiti (36) Sri Lanka(60) El Salvador(36) S. Arabia, (265)
Vietnam(- 11) Philippines(35)
Liberia(47) Colombia(27)
Guatemala(36) Laos(14)
Honduras(33) Gabon(74)
0.20-0.29 Ethiopia(15) Peru(65) India(1) Myanmar(-4) Malaysia (140) Costa Rica (120)
Malawi(13) Panama(53) Ecuador (46) Algeria(251)
Nepal(2) Korea, PDR (27)
Nigeria(5) Libya(401)
0.30-0.39 Sierra Leone (37)   Tanzania (2) Chile(14) Swaziland (134) Turkey(14)
Mozambique (30) Gambia (97) Suriname(-60) Cuba(235)
Botswana (148) Mauritania(117) Tunisia(219)
Pakistan(6) Mexico(77)
Nicaragua(46) Iran(101)
Thailand (- 113) Syria(114)
0.40-0.50 Angola(49) Bolivia(19) Zimbabwe (-40)   Uruguay (- 158) Morocco(58)
Afghanistan (17) Sudan(16) Togo(21) C.Ivoire(50) Iraq(223)
Over 0.50 Chad(8) Zambia(15) Niger(26) Senegal(82) Paraguay (-70) Brazil(18)
CAR(15)   Cameroon (43) Benin(22) Argentina
(- 289)
Guinea(40) Guyana(10)
Burkina Faso(17)

*Land per caput adjusted to measure land of roughly comparable production potential (see note 11).
Numbers in parentheses are net cereal imports in kg per caput. A minus sip denotes net exports.
Calorie data for 1988/90 before the latest (1994) revision of the FAO food balance sheets.

It follows from the above that declining land/person ratios can threaten the food welfare of those land-scarce countries which depend on agriculture in a major way for a living. And this irrespective of the fact that their own population growth may not have a significant impact on the global land/person ratios. Most countries in this class are those in the upper left quadrant of Table 2.5. Only a combination of much more productive agriculture (in practice, resort to land-augmenting technologies that would halt or reverse the declines) and vigorous non-agricultural growth will free them from the bondage of the ever declining land/person ratios.

In conclusion, the declining land/person ratios do matter for per caput food supplies in two senses. In the global context and for countries with high actual or potential dependence on food imports they matter mainly if the declines threaten to push the global ratio below (unknown) critical values, even after allowing for the reprieve to be had from land-augmenting technologies. Should this happen, the effects would be manifested in terms of rising food prices which would affect mostly the poor. It has not happened so far despite continuous declines in the global land/person ratios. How close the world is to eventual critical values and whether such values are likely to be reached before the world achieves stationary population and acceptable per caput food supplies for all is a matter of conjecture.

In the local context declines in the land/person ratios do matter for food supplies, nutrition and incomes, mainly for the countries with limited access to imported food and high dependence on agriculture for the maintenance and improvement of living standards and, consequently, of food welfare. If and when such dependence is reduced, the pressures on the global land/person ratios will assume increasing importance also for them.

The possible role of land-augmenting (in practice, yield-increasing) technologies was referred to above for the reprieve such technologies can afford in relation to the consequences of the inexorable declines of the land/ person ratios. However, some of the perceived threats to progress towards solving the food problem have precisely to do with the risks to the productive potential of the agricultural resources stemming from the application of these very technologies, e.g. loss of irrigated land to salinization and waterlogging, loss of yield potential and increased risk of crop failures because of pesticide resistance, etc. In addition, efforts to bring new land into cultivation or to use existing agricultural land more intensively can often be associated with degradation (e.g. from reduced fallows, from exposure of fragile soils to erosion following deforestation) and may not add permanently to total productive potential. In the following subsection an attempt is made to address what are hypothesized to be the more fundamental processes driving human activity towards degradation of the productive potential of agricultural resources.

Agricultural activity and degradation of agricultural resources

As noted, there is sufficient (though not comprehensive, nor detailed) evidence establishing that the productive potential of at least part of the world's land and water resources is being degraded by agricultural activity (e.g. soil erosion, waterlogging and salinization of irrigated lands). In addition, agricultural activity generates other adverse environmental impacts (e.g. threat to biodiversity, pollution of surface and ground water sources). Chapter 11 presents some evidence of these processes. While recognizing that agricultural activity often contributes to maintaining or restoring the productive capacity of land and water resources, this concluding subsection attempts to provide a framework of thinking through why human activity may end up destroying rather than preserving or enhancing this capacity.

The most commonly held view is that these processes are somehow related to continuing demographic growth, in two senses: (a) more food must be produced and this tends to draw into agricultural use land and water resources not previously so used and/or causes such resources to be used more intensively-both processes may generate adverse impacts on the quality of the resources themselves as well as on the broader environment; and (b) in many developing countries population growth is accompanied by increases in the number of persons living off the exploitation of agricultural resources with the consequence that the amount of resources per person declines.

In the normal course of events the decline in per caput resources would tend to raise their value to the persons concerned (being often their main or only income-earning asset) and would lead to their more efficient use, including maintenance and improvement of their productive potential. The fact that much of the agricultural resource base has been improved for agricultural use by human activity in the historical period is testimony to this process. Yet it is often observed that under certain conditions this caring relationship tends to break down with the result that people destroy rather than conserve and improve the productive potential of the resources (see Harrison, 1992).

Understanding why this happens is the most important insight needed for policy responses to promote sustainable development. When this destructive relationship is observed in conditions of poverty, it is commonly taken for granted that poverty explains the behaviour of people vis-a-vis the resources. The hypothesized mechanism works (in economic parlance) via the shortening of the time horizon of the poor. In plain language, this means that in conditions of abject poverty the need for survival today takes high precedence over considerations for survival tomorrow. The poor simply do not have sufficient means to provide for today and also invest in resource conservation and improvement to provide for tomorrow.

However, this proposition is far from being a sufficiently complete explanation of processes at work useful for formulating policy responses. For one thing, there is plenty of empirical evidence that this process is not at work in many situations of poverty. For another, it is often observed that agricultural resource degradation occurs also when such resources are exploited by the non-poor (a matter discussed below). It also occurs, and often more so, in conditions where poverty is declining rather than increasing, e.g. when the opening up of income-earning opportunities outside agriculture leads to the abandonment (because they are no longer worth it) of elaborate resource conservation practices, such as the maintenance of terraces to conserve small poor quality land patches on hillsides, etc. (for examples from the Latin American sierras, see de Janvry and Garcia, 1988). Another example of degradation associated with alleviation rather than aggravation of poverty is given by the opening up of profitable opportunities to grow cassava in some Asian countries for export to the EC where it substituted high-priced cereals in the feed sector. It is thought that part of this expansion of cassava production had adverse effects on the land and water resources which were not "internalized" in the export price (see Chapter 13).

It follows from the above that more complex processes are at work and the simple correlation between poverty and environmental degradation can be an oversimplification. This is well recognized, and research work on understanding the role of other variables which mediate the relationship between poverty and environmental degradation can provide valuable insights. Such work emphasizes, for example, the vital importance of institutions governing access to resources (e.g. to common property or open access resources) and how such institutions come under pressure when population density increases; inequality of access to land and landlessness; policies which distort incentives against the use of technology that would contribute to resource conservation, e.g. by depressing the output/fertilizer price ratio and making fertilizer use uneconomic where increased use is vital for prevention of soil mining; and the knock-on effects of policies which facilitate interactions between the non-poor and the poor in ways which lead to degradation, e.g. when deforestation and expansion of agriculture are facilitated by incentives to logging operations which open access roads into, and make possible agricultural settlement of, previously inaccessible forest areas which may have soils that cannot easily sustain crop production.

Understanding the role of these and other mediating variables and getting away from the simple notion that degradation can be explained by poverty alone is important for formulating and implementing policies for sustainable agriculture and resource conservation. It is important because the policy environment in the future will continue to be characterized by pressures on agricultural resources related, in one way or another, to rural poverty. Indeed, the numbers of the rural poor depending on the exploitation of agricultural resources will probably increase further in some countries and decline in others.

It was noted earlier that both processes can be associated with resource degradation. Therefore, the key policy problem is how to minimize adverse environmental impacts of both processes. Chapter 12 presents what are essentially technological options for policy responses; and Chapter 13 deals with policies that would contribute to minimizing the unavoidable trade-offs between agricultural development and the environment.

Poverty-related degradation of agricultural resources is only part of the story. It is well known that part of the degradation process is related to the actions of people who are not in the poor category. This issue has two aspects: the first one has to do with consumption levels and patterns of the non-poor, in both the developed and the developing countries. For example, some 30 percent of world cereals output is used as animal feed and a good part of the production of soybeans and other oilseeds is also related to livestock production. Most of the livestock output produced in concentrate-feeding systems is consumed by the medium- and high-income people. To the extent that production of cereals and oilseeds causes degradation (as indeed it does in some places, though not in others) it can be said that part of the degradation is caused by actions of the rich, not the poor. It would perhaps be more correct to say that it is caused by interactions between rich and poor, e.g. when expansion of soybeans production in South America raised the price of land there. This induced small farmers to sell land to large soybean operations and move to other areas to colonize new lands. The example cited earlier concerning the expansion of cassava production for export to Europe is another case. Both cases are to some extent related to policies of other countries which maintained artificially high prices for cereals used in livestock production and increased incentives for production and export of these cereal substitutes (for more discussion, see Alexandratos et al., 1994).

Many other cases could be cited to illustrate the complex interactions between the behaviour of the poor and the non-poor resulting in the building up of pressures on agricultural resources. Without a thorough understanding of such complex processes at work leading to resource degradation, it would be difficult to design and implement appropriate policy responses. Accounting for the factors which determine the actions of the poor and the non-poor alike is required even if, in a poverty-focused strategy, the priority objective were to minimize the degradation of the resources operated by the poor.

The second aspect has to do with the fact that resource degradation is also associated with agriculture practiced by farmers who are not poor. Soil erosion believed to be associated with some of the grain production in North America is a case in point; excessive fertilizer and other agrochemical use in Europe is another; and effluents from intensive livestock operations are in the same category. These are all examples of actions by the non-poor with adverse environmental effects. It all goes to show that associating resource degradation with poverty addresses only part of the issue.

In the end, the focus of policy has to recognize that resource degradation has different consequences for different countries and population groups. For the poor countries, the consequences can be very serious because their welfare depends heavily on the productive potential of their agricultural resources (Schelling, 1992). Therefore, from a purely developmental and conventional welfare standpoint, it is right that preoccupation with resource degradation problems focuses primarily on the developing countries. At the same time, it must be recognized that resource degradation not only in the developing countries, but anywhere on the planet, particularly in the major food exporting developed countries, can make more difficult the solution of the food security problems of the poor if it reduces the global food production potential. Therefore, controlling resource degradation in the rich countries assumes priority even in strategies focused primarily on the food security of the poor; and this irrespective of the fact that the welfare of the rich countries, as conventionally measured by, for example, per caput incomes, may not be seriously threatened by moderate degradation of their own resources. There are, of course, other compelling reasons why the rich countries give high priority to controlling degradation of their own resources, as an objective in its own right.


1. The food available for direct human consumption (Food) is computed from the following food balance sheet equation for each country, food product and year:

Total food supply = Production + (imports-exports) + (beginning stocks-ending stocks), and Food = Total food supply-Animal feed-Industrial non-food uses-Seed-Waste (from harvest to retail). The resulting estimate of per caput food availabilities can differ from the amounts of food people actually ingest in a nutritional sense, e.g. because of further waste in the post-retail stage (e.g. at the household level). For discussion of this issue and comparisons see FAO (1983).

2. World production of cereals in per caput terms peaked in the historical period at 342 kg in the three-year period 1984/86 and remained below this level in all moving three-year averages since then. It had fallen to 317 kg in 1987/89 (1988 was the year of the North America drought) and recovered to 325 kg by 1990/92.

3. The developing countries span a very wide range of this indicator, from under 15 percent (Jordan, Argentina, Chile, Trinidad and Tobago) to over 80 percent (many countries in sub-Saharan Africa).

4. It is recognized that world market price movements do not always convey unbiased information for evaluating the issue at hand. This is because they have been influenced, generally downwards, by the agricultural support and protection policies of major developed countries. In addition, the non-incorporation into the product prices of alleged losses of natural capital generated by the production process (reduction of the production potential of land and water resources following their degradation) also tends to bias downwards the market prices. In short, the downward movement of world market prices has in part reflected the rich countries' consumer and taxpayer subsidies (see O'Brien, 1994) and perhaps, but to an unknown extent, also an environmental subsidy of all countries which ignore the costs of the alleged natural capital losses.

5. These indicative numbers of per caput food supplies are distinct from the related concept of national average requirements for dietary energy. The latter are derived as the weighted average of the requirements of the individuals in a population group. Individual requirements vary widely with (mainly) age, sex, body weight and level of physical activity. For example, they are 1900 calories/day for a woman in the 18-30 age bracket, with a body weight of 50 kg and light physical activity. They rise to 3700 calories/day for a man of 70 kg in the same age bracket but with heavy physical activity (FAO/WHO/UNU, 1985, Tables 15, 42, 45). These are still averages, since requirements vary among persons of the same age, weight, etc. National average requirements have been computed for individual countries. For example, they were 2130 calories/day for Haiti for moderate activity level (FAO, 1990c). To make this estimate conceptually comparable to that used here for per caput food supplies it must be raised by a margin to account for food losses at the post-retail level. If the margin were 10 percent, Haiti would have a requirement at the retail level of some 2350 calories/person/day. This would have been adequate for meeting the country's nutritional needs if food were distributed to individuals exactly according to their requirements. Since this is never the case in any country, it results that the national average must be still higher if enough food were to be available for people at the lower end of the distribution to have potentially access to food to meet their energy requirements. In the event, Haiti's per caput food supplies were 2000 calories in 1988/ 90. The preceding discussion in no way implies that the solution to the food problem is to be pursued by means of policies aimed at raising per caput food supplies to whatever level is suggested by the above considerations. Such policies must be seen as a necessary complement to, not a substitute for, policies which address directly the root cause of the problem, the inadequate food entitlements of the poor. However, in countries where most of the poor are in agriculture, the two policies go in tandem.

6. Physiological considerations (i.e. a person needs a minimum food intake for survival and there is an upper physiological limit to how much food a person may consume) dictate that the scope for distributional inequalities is more limited in the case of food intakes than in other "unbounded" variables, e.g. income. Consequently, it may be hypothesized that very low or relatively high levels of average per caput food supplies will be associated with more equal distribution of food intakes compared with those that could prevail when the overall average is the middle range. This is a useful hypothesis for looking at the distributional issue when we know nothing about prospective developments in the other determining variables of the distribution of food intakes. Thus, it can be expected that the distribution of food intakes will tend to become more equal in those developing countries in which the average per caput food supplies will continue to edge upwards towards the 3000 + calorie level. These considerations are taken into account in Chapter 3 when considering the issue of the possible evolution of the incidence of undernutrition in the future.

7. These countries are (in ascending order of their 1961/63 rank): Libya, China, Algeria, Saudi Arabia, Indonesia, Iraq, Korea (Rep.), Iran, Korea (DPR), Tunisia, Morocco, Egypt. Each of them had increased per caput food supplies by at least 40 percent over the historical period considered. Thus, countries which had in 1988/90 2600 calories are included only if they started with less than 1860 calories in 1961/63. Likewise, those with 3300 in 1988/90 are included only if they started with less than 2350 calories. The data used here are those available before the latest (1994) revision of the FAO Food Balance Sheets. Calorie data for all countries are given in the Appendix 3.

8. All developing countries (among the 93 of this study) with per caput food supplies under 2100 calories in 1988/90 (19 countries, listed in ascending order of calories in 1988/90): Ethiopia, Chad, Afghanistan, Mozambique, Central African Rep., Somalia, Angola, Sierra Leone, Rwanda, Burundi, Namibia, Haiti, Bolivia,

Zambia, Peru, Bangladesh, Sudan, Malawi, Kenya. All of them were in 1988/90 in the range 1700 calories (Ethiopia) to 2065 calories (Kenya). In 1961/63 the range was from 1720 calories (Somalia) to 2300 calories (Chad).

9. The role of the public policy can be appreciated from the following World Bank statement: "Egypt probably has the lowest level of malnutrition among countries with comparable levels of per capita income. Egypt in effect is a vast welfare state which provides the bulk of its people cheap food, cheap oil, and, for those living in rent-controlled housing, cheap though insufficient shelter" (cited in Yitzhaki, 1990).

10. See, for example, Kennedy (1993: 192). A case can be made that some developing countries, but not all of them, would gain from policies to reduce GHG emissions if such policies removed economically wasteful energy market distortions (see OECD, 1992).

11. A country may have plenty of land of poor agricultural quality, but not for this reason should it be classified as land-abundant. For example, it is estimated that Niger has land-in-use of 1.5ha per person in the total population, but with 95 percent of it in the dry semi-arid category. Other countries have much less land-inuse per person but of better quality, including quality improvements brought about by irrigation. For example, Pakistan has only 0.16 ha/person but 86 percent of it is irrigated. The land/person ratios must, therefore, be adjusted before a meaningful comparison among countries can be attempted. Adjustments are made using the following weights: 1.0 for sub-humid; 0.81 for fluvisols/gleysols and 0.35 for marginally productive fluvisols/gleysols; 0.31 for dry semi-arid, 0.88 for moist semiarid, 0.85 for humid; 0.35 for the marginal areas in the moist semi-arid, sub-humid and humid zones; and 2.2 for irrigated. These weights roughly reflect potential cereal yields. After such adjustments, Niger's land/person ratio falls to 0.50ha and that of Pakistan rises to 0.31 ha. This is land of fairly comparable production potential after the adjustments. The comparisons in Table 2.5 are on the basis of the thus adjusted land/person ratios.

12. Reference to land-augmenting technologies and the growth of other sectors just underlines the fact that aggregate resources for producing food and/or incomes cannot be considered as given. In the process of development scarce resources are substituted by less scarce ones and the total is augmented by additions of man-made capital, the most important component of which is human ingenuity. Whether there are ultimate limits to this process is another question (see, for example, Daly, 1992; Daly and Townsend, 1993; and for a less pessimistic view, Pearce and Warford, 1993, whose book goes under the suggestive title World Without End).

13. Yield-augmenting technologies, generally those associated with the introduction of modern varieties, have come occasionally under fierce criticism, both because of their adverse effects on resources and the environment as well as because they tend to disrupt traditional farming systems and the associated social structures (e.g. Shiva, 1991). Yet such effects need to be evaluated in relation to those, probably much more severe ones, that would have resulted from the increasing pressures of the growing population on resources, traditional farming systems and social structures if these technologies had not made it possible to achieve quantum jumps in food production. The issue is not one of having or not having yield-augmenting technologies. It is more one of minimizing their adverse effects. Agricultural research is increasingly oriented towards this goal, e.g. breeding resistance to pests into modern varieties so as to reduce dependence on chemical pesticides (see Chapter 4).

Contents - Previous - Next