Chapter 7 - Economic consequences of land degradation
Back to contents - Previous file - Next file
Introduction: economic and social consequences
Economic valuation of natural resources and degradation
Land degradation in South Asia: the orders of magnitude of the economic costs
Macroeconomic impact of land degradation
Introduction: economic and social consequences
Chapters 7 and 8 are concerned with the economic and social consequences of land degradation: its implications for the community as a whole, and thus for governments, and its effects upon the people, the rural population of the areas affected. One of the main social consequences is, in fact, also an economic one, namely reduced income for the farmers; whilst analysis at the macroeconomic lever is ultimately based upon aggregating the effects of degradation upon individual farmers.
This chapter covers the economic consequences of land degradation considered at the national and regional lever: its cost to the countries and their people. The effects on the rural population, including reduced incomes, are discussed in Chapter 8.
Economic valuation of natural resources and degradation
Concepts and approaches in natural resource accounting
Natural resource accounting is a relatively new concept. Its implications for land degradation are so great that an introductory outline of the concept and methods is called for.
The basis is simple: that natural resources, such as minerals, soils and forests, have an economic value. This is called natural capital, to be distinguished from manufactured capital such as roads, factories and machinery. Changes in manufactured capital - construction and depreciation - have always been considered in both financial and economic analysis. Until recently, changes in natural capital have not been given money values, nor included in cost-benefit and other forms of economic analysis. Changes in natural capital are not currently included in systems of national accounting, although there is pressure for their inclusion.
Natural resources have formerly been priced only in terms of their cost of use: minerals were priced only at the costs of extracting them, forests at the logging costs. In the case of soils, these were treated as the 'land' factor in classical economics, priced at the market value of farmland. In effect, the capital value of the resources themselves was priced as zero. It was assumed that they were so abundant as to have no scarcity value.
This led to some gross distortions in the apparent creation of wealth. Minerals are extracted, or forests cut clown, and the money received from their sale is treated as national income. The decrease in the reserves of minerals or area of forest does not appear in the accounts. Wealth appears to have been created, based on the 'free' natural resources.
The example of minerals refers to a non-renewable natural resource, that of forest to a renewable one. The situation for soil and water resources is more complex. The milder forms of land degradation, for example soil nutrient depletion, can be reversed by changes in management; the resource is renewable, and the degradation reversible. In the case of two severe forms of degradation, salinization and waterlogging, land productivity can be restored by reclamation, even where degradation has progressed to the point of total loss of production; the degradation is reversible, although at high costs, as shown by the SCARP projects of Pakistan. Soil degradation may be reversible or irreversible, as discussed below.
In the case of soil erosion, some of the effects may appear to be reversible, through checking further erosion by soil conservation programmes and restoring lost nutrients and organic master. Where land has been lost by gullying, or severe sheet erosion has removed the soil clown to a gravelly residue, degradation is clearly irreversible. It should be noted, however, that this applies also to any actual loss of soil material, or reduced profile depth, since the rate of natural soil formation is extremely slow on a human time scale.
Two recent case studies, in the Philippines and Indonesia, illustrate the orders of magnitude which may be involved. In the Philippines it has been estimated that there is an annual rate of natural resource depletion equivalent to 4% of the gross domestic product (World Bank, 1989; Cruz and Repetto, 1992). For Indonesia, inclusion of the loss of timber, oil and soil resources had the effect of reducing gross domestic produce by about 20%, whilst gross domestic investment was reduced to low, and in one year negative, values. The annual depletion of soil fertility was calculated as 4% of the value of crop production, or as large as the annual increase in production (Repetto et al., 1989; Magrath and Arens, 1989).
Discussions of methods of natural resource accounting, drawn upon as the basis for the present discussion, include Ahmad et al., (1989), Chisholm and Dumsday (1987), Lutz and El Serafy (1988), Pearce and Turner (1990), Peskin (1989), Pezzey (1992) and Southgate (1989). A report of a seminar specific to Asia is given in Sun (1989), and a consideration of natural resource accounting for India by Parikh et al. (1992).
Methods for the valuation of soil resources
Soil resources have been valued chiefly as the basis for analyzing the economics of soil conservation projects (Bojö, 1992; Magrath, 1989). There has also been extensive cost-benefit analysis of reclamation projects to counter salinization and waterlogging. Five methods are found:
For the case of soil erosion, restoration costs have been incompletely assessed in previous analyses. Suppose that land has lost 5 cm topsoil. It is not sufficient to value the cost of installing soil conservation measures, followed by improved land management, for one is still working with the depleted, shallower, soil. If the soil is to be restored to its former conditions then it is necessary to:
One way to replace lost soil would be to buy it, as can be clone from a horticultural supplier. This solution succeeds in putting a market money value on soil volume, but is environmentally unacceptable, since it is robbing one area to restore another. The only true way to restore lost soil is by taking land out of production until it is restored by the natural process of weathering. The rate of this process varies by orders of magnitude for different rock types, but for consolidated strata it has been estimated as a maximum of 500 mm per thousand years (Saunders and Young, 1983). In soil conservation circles a top rate of "one inch in 30-300 years", or about 0.1-1.0 mm per year, has been quoted (Schertz, 1983). Assuming natural erosion to be very slow, then even at the fastest of these rates, it would mean putting land under fallow for 50 years to restore the lost 5 cm. This again is unrealistic as a practical proposition, but it gives a measure of the true resource loss incurred.
Provisional nature of these estimates
The natural resource accounting studies noted above, of the Philippines and Java, required a large input of effort, which it is out of the question to attempt here for the whole of South Asia. Nevertheless, a major objective of this study would be missing if some attempt were not made to estimate the economic cost of land degradation.
It should be stated at the outset that such estimates are highly approximate. They are made with the intention of indicating the orders of magnitude of the costs involved, with the objective of stimulating more detailed studies on a national and local scale.
The best researched cases are the impact of salinization and waterlogging in India and Pakistan (for example Ahmad and Kutcher, 1992; ESCAP, 1990b). A small number of local case studies have been identified, covering impacts of salinization, lowering of the water table, erosion, and soil fertility decline (Vittal et al., 1990; Joshi and Tyagi, 1991; Joshi and Jha, 1992; Chaudhary and Aneja, 1991). It is likely that other such studies exist.
For comparative purposes, the main approaches used here will be those of lost production, nutrient replacement, and reclamation or restoration. The starting point is the estimates for the extent and degree of each type of land degradation obtained in Chapters 4 and 5. Some extremely broad assumptions as to typical yields, yield reductions and farm production economics are necessary. To reduce the problems of differences between prices, types of land use, etc. between countries, in some cases calculations are clone first for India (as having more than half the population and agricultural production of the region) and then extended, still more approximately, to the region.
Best estimates for extent of land degradation are given in Table 18. Data on land use, fertilizer use, agricultural production and prices are taken from FAO statistics.
The term production loss refers to the production lost as a result of land degradation, that is:
Production loss = production from non-degraded land - production from degraded land, with the same inputs and management
Relative production loss is production loss as a percentage of production from non-degraded land. It should be noted that in many cases, farmers do not accept reduced production but instead, counteract decreased soil productivity by increasing inputs.
The following assumptions are made with respect to the effect of degrees of degradation on agricultural production:
|Degree of degradation||Relative production loss|
These are relatively low, or modest, assumptions; that is to say, the true figures may well be higher. In particular, by definition, strong degradation should mean that the land has been abandoned, with 100% loss of production; however, it has been assumed that 25% has been saved in some way by the ingenuity, backed by need, of the local population.
The above assumptions are critical: that is, results for the economic effects of degradation show a high degree of sensitivity, in many cases pro rata, to the values assumed.
Generalized prices in the region (1992), taken as the basis for calculations, were taken as:
|Cereals||US$ 150 per tonne|
|Fertilizer||US$ 300 per tonne nutrients|
Land degradation in South Asia: the orders of magnitude of the economic costs
Production loss basis An estimate will first be made for India. Approximately 61 % of India's agricultural land is under cereals, with an average yield of 1.9 t/ha. It is assumed that erosion affects cereal-growing land in the same way as total land. On this basis, and with the above assumptions on proportional losses of production, the loss in cereal production is as follows:
|Light degradation||168 000 t|
|Moderate degradation||3 980 000 t|
|Strong degradation||10 935 000 t|
|Total cereal production loss||15 083 000 t|
A production loss of 15 Mt cereals is equivalent to 8% of India's total cereal production.
Assuming similar proportional losses to other forms of production, the 8% loss may be scaled up to 25 Mt cereal equivalent, representative of the loss to total agricultural production.
At an approximate price of US$150 per tonne for cereals or cereal equivalent, the value of lost production is US$2 260 M per year.
For the region as a whole, it would be possible to carry out country by country calculations for land use, production, prices, etc., such as were clone in the study of Java by Magrath and Arens (1989). Such detail, however, would not be justified in view of the uncertainty of, and sensitivity to, the production losses. An approximation in terms of reduced production as cereal equivalent is:
|Light degradation||3 107 000 t|
|Moderate degradation||13 528 000 t|
|Strong degradation||19 237 000 t|
|Total production loss||35 872 000 t|
A cereal equivalent production loss of 36 Mt is about 9% of the total agricultural production of the region. At a price of US$150 per tonne cereal equivalent, the value of lost production due to water erosion is US$5 400 M per year.
This approximate result may be expressed in another way. If all land in the region were non-degraded, that is, in the condition it was in prior to recent population pressure, then with today's lever of inputs and methods of management, an additional production of 36 Mt cereal equivalent could be expected in the region.
Nutrient replacement basis For the calculation of nutrient replacement it is necessary to estimate current annual rates of erosion associated with degrees of land degradation (the GLASOD survey includes estimates of the extent to which rates of degradation have recently accelerated). The following rates are assumed:
|Degree of degradation||Current rate of erosion|
|Light||10t/ha per year|
|Moderate||20t/ha per year|
|Strong||50t/ha per year|
It is assumed that eroded soil contains 0.2% nutrients. On this basis, for India, the annual loss of nutrients through water erosion is:
|Light degradation||58 000 t nutrients per year|
|Moderate degradation||688 00.0 t nutrients per year|
|Strong degradation||1 260 000 t nutrients per year|
|Total nutrient loss||2 006 000 t nutrients per year|
Total mineral fertilizer use in India is 12.5 Mt of nutrients per year. The loss through erosion of 2 Mt is 16% of this. Expressed in another way, India would have to increase its fertilizer use by this amount each year just to replace nutrients lost through erosion.
At a representative fertilizer cost of US$300 per tonne nutrients, the loss through erosion is US$600 M per year.
A similar approximate calculation for the region as a whole gives a loss through water erosion of 3.4 Mt nutrients per year. This is equivalent 20% of total fertilizer use in the region. Its value is approximately US$1 020 M per year.
The two estimates obtained for the effect of water erosion are not directly comparable. That obtained for production reflects the cumulative effect of past erosion, whereas the longer estimate based on nutrient replacement is an annual value. However, in replacing the lost nutrients, the farmer is counteracting only of the effects of erosion. These also include loss of soil organic master and reduction in soil profile depth, leading to degradation of soil physical properties and, in particular, water holding capacity. Nutrient loss, and thus nutrient replacement, is only one element in the effects of erosion.
Restoration or reclamation The reclamation of land subject to gully (ravine) erosion is often undertaken, but with the aim of preventing further extension of gullying. Such attempts meet with variable success, and it is rarely possible to restore productivity to anything like its former state.
For land with moderate degradation, a specimen calculation for restoring lost soil is as follows. As above, moderate degradation is assumed to correspond to a current rate of erosion (in excess of replaceable) of 20t/ha per year (equivalent to an horizon thickness of 1.33 mm).
Let it be assumed that replacement of soil by natural processes is at a rate ranging from 0.1-1.0 mm per year, equivalent to 1.5-15.0 t/ha.
To replace the soil lost in one year of erosion at 20t/ha would require following for between 1 and 13 years, and hence a loss of production between 50% and 93%. This is clearly unrealistic as a practical proposition. It indicates, however, that the full cost of erosion is substantially higher than the estimates obtained above, which are on a medium-term basis only. In practical terms, loss of soil material is largely irreversible. To achieve long-term sustainability, erosion must be limited to the rate of soil formation.
Off-site costs of erosion Deforestation and erosion lead to greatly increased sediment load in rivers, causing problems of poorer water quality, river bed sedimentation and reservoir sedimentation. A review of the effects of soil conservation measures upon sediment yield is given by Doolette and Magrath (1990, p.203ff.). Reductions in sediment yield through conservation measures are frequently as high as 95 %. The economic effect is most clearly seen in sedimentation of reservoirs. For eight Indian reservoirs, the presently assessed life as a percentage of that anticipated on design ranges from 23-79%, with four below 40% (FAO/RAPA, 1992, p.216). In developed countries, off-site costs of erosion are often assessed as substantially higher than on-site costs, although in less developed countries, the reverse may be the case (P. Faeth, D. Knowler, persona! communications).
Off-site costs have not been assessed in this study but their existence, and certainly appreciable magnitude, should be taken into account.
It is difficult to obtain even the most approximate estimate of the economic cost of wind erosion. The land affected is partly under arable use and partly used for livestock production, and there is virtually no basis for estimating the effects of erosion upon production. Yet if its seriousness is to be appreciated, some value must be given.
If the degree and extent of wind erosion is compared with that of water erosion, the total impact of the two is comparable. Areas affected by moderate and strong wind erosion are similar to those of water erosion, around 35 and 12 M ha respectively. The area for light wind erosion is only 40% of that for water erosion, but on the assumptions used above, this has a relatively small effect on production.
Being confined to dry climates, the average productivity of land affected by wind erosion will be less than that affected by water erosion. Suppose that on average it is one third as productive. The production loss from water erosion was assessed at US$5 400 M per year. For an impact of similar severity, the production loss from wind erosion is of the order of US$1800 M per year.
This impact is very unevenly distributed in the region, being entirely for the countries of the dry zone: Afghanistan, Iran, Pakistan and the dry region of India.
Soil fertility decline
Production loss basis There is no doubt that soil fertility decline is occurring over large parts of the region. Data for assessing its effects are, however, tentative in two respects: the area covered, and the magnitude of the depression of crop yields.
The estimates of area are (as a conservative assumption) dominated by the light degree of degradation, at 38.5 M ha compared with 3.9 M ha affected to moderate or strong degrees. As a further simplifying assumption, the total figure only will be taken, that of 42.4 M ha affected, to at least a light degree, by soil fertility decline.
Two alternative assumptions are then made, which can be justified by the available experimental data. These are that the average effect of fertility to decline is to reduce crop yields, in the absence of additional inputs, by 5% or 10%. Using the same basis as for water erosion, an average cereal equivalent yield of 1.9 t per hectare, gives a production loss of:
For a 5% yield loss: 4 028 000 t
For a 10% yield loss: 8 056 000 t
At a price of US$150 per tonne, the loss to the region from soil fertility decline is tentatively estimated at US$600 M - 1200 M per year.
Replacement cost basis As already noted, farmers with soils of declining fertility frequently attempt to maintain yields by additional inputs, primarily fertilizers. Some research results have shown that quite high rates of fertilizer application are necessary where the soil has been degraded by prolonged cropping. However, let it be assumed that on average, yields on the 42.4 Mha of degraded soils can be maintained by an average input of either 50 or 100 kg nutrients per hectare. The cost is taken as US$300 per tonne of nutrients. Additional inputs and their cost are then:
|At 50 kg ha¹:||2.12 Mt||US$636 M|
|At 100 kg ha¹:||4.24 Mt||US$1 272 M|
This is of the same order of magnitude, US$0.6 - 1.3 billion, as the estimate on a production loss basis.
This reasoning, however, applies only to the short term. The additional of unbalanced fertilizers, without other measures to improve the soil, is a cause of fertility decline. Fertilizer rates needed to maintain crop yields can therefore be expected to increase with time, raisin" the cost.
More fundamental measures are needed to restore soil fertility, particularly through the improvement of organic master statue. These management measures also have a cost, for example the opportunity cost as lost fodder or fuel of returning crop residues to the soil. The combination of such methods for soil improvement with continued, and more balanced, use of fertilizers is necessary for sustained land use in the medium and long term.
Percentage yields obtained under four crops at different water table depths are given in Ahmad and Kutcher (1992, p.42). Taking their own data for areas with shallow water tables, and yield reductions for wheat as representative, this gives a yield loss for Pakistan of 1.57 Mt, or about US$240 M per year. On the basis of comparative areas affected, the loss for India would be substantially higher. This gives a total loss from waterlogging in excess of US$500 m per year.
There have been more attempts to asses the impact of salinization than is the case for other forms of degradation. This is partly because its effects are substantial and visibly apparent, partly because the degree of degradation can be readily quantified, and also because it occurs on irrigated areas which have received large financial investments.
Estimates will first be compared for Pakistan. Experimental work on percentage yield losses for different values of salinity is summarized in Ahmad and Kutcher (1992, p.45). The impact differs between crops, with cotton tolerant, rice intolerant of salinity. Taking the data for wheat, and matching values of soil conductivity to degrees of degradation, the following production losses will be assumed:
|Degree of degradation||Relative production loss|
Estimating in terms of wheat equivalent, using the average Pakistan wheat yield of 1.84t/ha, production losses are as follows:
|Light degradation||524 000 t|
|Moderate degradation||1 196 000 t|
|Strong degradation||2 392 000 t|
|Total wheat equivalent loss||4 112 000 t|
Valued at US$1501t wheat this equals a loss of US$617 M per year. These values would be altered by taking the crop mix into account, but the order of magnitude would remain the same.
This may be compared with other estimates. ESCAP (1990b) state, "A 20 per cent reduction in yield of, say, wheat in Pakistan on about 3 M hectares of salt-affected land would result in a loss of about 1.2 M tonnes of grain on a very conservative estimate. This would amount to some US$150 M." Ahmad and Kutcher (1992) assess salinity levers, areas affected and yield decreases for Pakistan, concluding, "If these numbers are anywhere near correct, soil salinity is "robbing" Pakistan of about 25% of its potential production of cotton and rice, or about US$2.5 billion per year!".
It is not possible to obtain comparable estimates for the region as a whole. The problem is of the same order of magnitude in India, therefore the above estimates may first be doubled, to US$1234 M per year. In areal extent, salinization is dominated by Iran, and the salinized area exceeds that of total arable land. It is therefore difficult to make broad assumptions on which to base an estimate. Taking as a very minimal value a loss of some US$300 M, the total loss to the region from salinization is not less than US$1 500 M per year.
Cost of reclamation Salinization and waterlogging can be reversed, and the land productivity partly restored, by reclamation. This has been clone most notably in the case of the series of Salinity Control and Reclamation Projects (SCARPs) in Pakistan continuing from 1959 to the present. The main elements in the technology involved are:
From 1969-85, SCARP projects covered 3.5 M ha, and a further 2.8 M ha are at present being reclaimed. It is stated that as a result of SCARPs, soil salinity has been reduced from 40% to 28%, and 80 000 ha of land are being restored to production each year (ESCAP, 1989b; 1990b, p. 26).
The cost of such reclamation measures is huge. The opportunity cost of the water used for leaching is that of the production it could have given if used for irrigation. Reclamation costs are currently about US$500/ha (Ahmad and Kutcher, 1992). For Pakistan, the cost of reclaiming 3.3 M hectares of affected land has been estimated at US$9 billion (Ahmad and Kutcher, 1992). With an area affected of the same magnitude, the cost to India would be similar.
The cost of reclaiming salinized and waterlogged areas is considerably higher than that of prevention by good design and management of irrigation schemes.
Lowering of the water table
The consequence of lowering of the water table, where it has developed in areas of non-saline groundwater, could be expressed in economic terms as the added cost of tubewell pumping from greater depth. The true shadow price of electricity, and not its subsidized price, should be used. It is likely that large farmers can afford this added cost, and that the more serious effect is upon small farmers with holdings, and capital resources, too small to justify deep tubewells.
A production loss basis would underestimate the cost of lowering the water table, since this is a clear case of non-sustainable use of a resource. Restoration of water table levers would require reduction in water use to less than the rate of natural recharge, with consequent loss of production, for long periods. This is an unrealistic scenario, and the economic cost would be vast. Because of these complexities, coupled with inadequate data, the cost of lowering the water table has not been assessed in this study.
Macroeconomic impact of land degradation
The difficulties and uncertainties that arise in assessing the economic effects of land de gradation must again be emphasized. Sources of error arise at all stages of assessment:
These sources of error are cumulative: an over- or under-assessment at one stage is multiplied by errors at successive stages. For stages 1 and 2, the sensitivity of the total estimate is almost pro rata with errors of estimate. If the area affected by a given extent and severity of degradation is over- or underestimated by 50%, the deduced economic impact will be underestimated by the same amount. For the physical effects the relation is somewhat less simple, for example, total loss of land reduces inputs as well as production, but the sensitivity of the final economic result to the proportional loss of production is still high.
It would be possible to argue that the overall degree or uncertainty is such that no total figure should be quoted. To do this would be to miss the objective of this analysis, which is to signal that the problem of land degradation appears, on present evidence, to be of a magnitude that is significant in relation to the total wealth of the countries concerned.
Using the incomplete data above, and basing a summary on the method of lost production, the estimates are given in Table 21.
Summing these estimates for the direct, on-site, costs gives a total of US$9.8 - US$11 billion per year. Thus, in round figures, the cumulative effect of human-induced land degradation is estimated to cost countries of the region a sum of the order of US$10 billion per year.
TABLE 21 - Provisional estimates of the cost of land degradation in the region
|Type of degradation||Cost, billion US$ per year||Notes|
|Water erosion||5.4||On-site effects only|
|Wind erosion||1.8||Assessed relative to water erosion|
|Fertility decline||0.6-1.2||Tentative estimate|
|Lowering of water table||Not assessed|
The gross domestic product of the eight countries (1989) is US$488B, and their combined agricultural domestic product US$145 billion. The estimate obtained for the on-site effects of land degradation upon productivity is equivalent to 2 % of the gross domestic product of the region, or 7% of its agricultural gross domestic product. The inclusion of off-site effects of water erosion would increase this value substantially.
The value of resources
The above discussion has been conducted largely on the basis of the user value of land resources, their value for agricultural production. There are, in addition, values which are not directly quantifiable but which are known. For example, a soil cover is needed to stabilize runoff and provide base flow; where there is no such cover, runoff is immediate as floods. There are also user values which are not yet known; for example, a century ego, the resource potential of bauxitic soils was not appreciated.
However, natural resources also possess a primary value over and above the sum of their user, or secondary, values. In the case of soils, the primary value represents the outcome of processes of soil formation - rock weathering, pedogenesis, biological activity - which have taken of the order of 10 000 - 100 000 years. Within the human time span, soil cannot be created (other than in extremely small amounts). The primary value represents the difference between land without soil and land with soil. For plant resources, the primary value includes the processes of evolution.
The primary value of soils is not only to the present generation, nor to the 20-50 years commonly included in obtaining net present values by the procedure of cash flow discounting. Soils have been a resource for the past 2000 and more years. There is no reason to support that the population will not be dependent upon them for a least the same length of time into the future. Moreover, if there is continued population increase, land resources will certainly increase in relative value in the future.
Complex questions of economic analysis are involved in assessing primary values, which it is inappropriate to discuss here. One simple means of obtaining a minimum figure is to estimate the sum of today's user values and multiply this by, say, 2 000 without discounting. This represents the value to future generations of today's soil resources. Whilst not attempting such an assessment here, the essential point is that land resources have a value, for future generations, over and above the sum of either their current user values, or their discounted net present value.
The existence of a 'contrary' view has already been noted. Expressed in terms of investment appraisal, this stases that reports of land degradation may be greatly exaggerated; and that unless and until better data are obtained, the problem does not meet the criteria for development investment.
For reasons given in Chapter 5, Section Discussion, this view is rejected. The present study does not seek to magnify the seriousness of the problem. It is an attempt to obtain the best objective estimates on the basis of available data. The assumptions made in calculation of economic values err on the side of caution. On this basis, the best estimate that can be obtained is that land degradation is costing countries of the region an economic loss of the order no less than US$10 billion, equivalent to 7% of their combined agricultural gross domestic product.
Efforts should certainly be made to improve the quality of the data, not only on the degree of degradation but also its effects upon production. However, action to check degradation should not be withheld until such improved data are available. The loss of productive resources is already considerable and is becoming more serious year by year.