CONSULTATIVE GROUP ON INTERNATIONAL AGRICULTURAL RESEARCH
TECHNICAL ADVISORY COMMITTEE
A Framework for Prioritizing Land Types in Agricultural Research
FOOD AND AGRICULTURE ORGANIZATION OF THE UNITED NATIONS
A Framework for Prioritizing Land Types in Agricultural Research
The Report of the Study of CGIAR Research Priorities for Marginal Lands1 (the Marginal Lands Report), noted that many different terms are used to designate the productive capacity of land. With the aim of achieving greater uniformity of terminology, the Report presented a six-celled matrix for evaluating land-related issues. This paper takes the principles, which the Marginal Lands Study adopted for its examination of the relation between marginal lands and rural poverty, and develops them as a conceptual framework for more-general use by the CGIAR in its priority setting. Land is defined as an area of the earth's surface which comprises the major biophysical attributes that influence land use, including the atmosphere, soil, geology and hydrology, plant and animal populations, and the results of past and present human activity such as terracing, drainage and irrigation.
1 TAC(1997a). Report on the Study of CGIAR Research Priorities for Marginal Lands. Working Document SDR/TAC:IAR/96/18.1, TAC Secretariat, FAO, Rome.
The CGIAR has adopted the practice of setting its priorities and strategies according to the poverty weighted values of commodities and sectors, with allowance for important modifying factors such as those that influence the probability of success in research2. These principles could also be applied to prioritizing the System's work on different types of land. Alston, Norton and Pardey3 have listed four efficiency criteria for estimating the expected benefits from commodity research programs: current value of production, maximum yield increase (or decrease prevented) if the research were successful, probability of research success, and the likely rate and extent of adoption of its results. The framework outlined in this paper would assist in assigning values to these criteria, so that potential benefits could be estimated for research on different categories of land. Information about the likely costs of such research would be needed for a prospective cost/benefit analysis.
2 TAC (1997b). Medium Term Resource Allocation 1998-2000: Centre Proposals and TAC Recommendations. Document SDR/TAC:IAR/97/6, TAC Secretariat, FAO, Rome.
3 Alston, J M, Norton, G W and Pardey, P G (1995) Science Under Scarcity, pp.477-8. Cornell University Press: London.
In the Marginal Lands Report, the six-celled matrix was defined by the present value of land use on one axis, and on the other, the potential for enhancement of production through research. Three levels of present use value were identified:
· high present use value - favoured agricultural lands,
· low present use value - marginal agricultural lands,
· lands at low or zero intensity of use.
Two levels of potential to expand production based on research were identified, high and low. Figure 1 reproduces the table that was presented in the Marginal Lands Report.
2. Issues in the Development of a Generalized Framework
In developing these principles for more-general use by the CGIAR, TAC recognized the following issues. Firstly, in the real world there tends to be continuous variation from zero to high in both the present value of land use and the potential to enhance its productivity through research. While the six quadrants* are helpful in visualizing the characteristics of the different types of land, it is a matter of choice as to where the dividing lines are drawn.
* The terminology of the Marginal Lands Report has been adopted here.
Secondly, the Marginal Lands Report defined the potential to expand production in terms of research only. While research is the primary interest of the CGIAR, there are arguments for preceding the analysis of research priorities with a review of the potential benefits from investments of all kinds. It may be that for well-known socioeconomic reasons (not needing further research) existing technologies are not being used to enhance the value of production. Alternatively, further research may be required to find out why apparently suitable technologies are not being used.
Generally, two main sets of factors determine the scope for enhancement of value of land use through investment. Firstly, there are the biophysical attributes of soil, climate and terrain. Man-made physical capital, such as the provision of irrigation water and the terracing of steep slopes, also could be included here. Secondly, there are socioeconomic elements associated with (a) the creation, dissemination and utilization of knowledge, (b) access to inputs, markets and credit facilities, (c) output/input ratios, and (d) a group of motivational factors determined to a large extent by the incentives created by society and its economic, cultural and other institutions.
In other words, and as developed more fully in the TAC Marginal Lands study, the actual realization of the potential for improvements in production or productivity depends on people having control over land (use rights), and specifically on their knowledge of what can and needs to be done; their access to the resources required to implement what needs to be done; and their incentives to do what needs to be done.
Thirdly, the point needs to be made that the present productive value of any particular area of land depends greatly on what it is used for. Land that is too dry for cropping may be very profitable for grazing ruminant livestock. In more humid climates, terrain that is unsuitable for cultivation of annual crops may be well suited to perennial tree crops and forests. So there will be a different matrix for each type of land use in agriculture and forestry.
Fourthly, it must be recognized that in practice the productivity of agricultural land can also be reduced. The use of land for agriculture sets in train a set of potentially degrading processes (a downward movement in Figure 1), which have to be countered if value is not to be lost. While investment in research is usually thought of in the context of enhancing land value, it can also serve to prevent or reduce potential loss of value.
During the past 150 years, agricultural research has continually expanded the opportunities for productivity enhancement, often far beyond the intrinsic value of the land for agriculture when it was first used for that purpose. However, the possibility that these inherent degrading processes will reduce the productive value of agricultural land has been of increasing global concern recently, under the rubric of sustainability.
Finally, there does not appear to be any strict logical relationship between lands at low or zero present intensity of use (for any particular form of agriculture or forestry) in quadrants 5 and 6 of Figure 1 and those in the other four quadrants of the matrix. Land that is available for future development, say grazing land with good potential for grain cropping, could in theory translate into any of quadrants 1 to 4 when developed for that purpose.
3. The Proposed Framework
With these issues in mind, TAC proposes that the Group adopt a modified version of the scheme described in the Marginal Lands Report. The priority of any particular category of land would be rated by plotting its present productivity per unit area of land, against the potential for change through investment in research. This is equivalent to using quadrants 1 to 4 of Figure 1 without the dividing lines, and reversing the direction of the axis for potential to expand production. Another important change would be to define the X axis to include prevention of degradation, rather than enhancement of production only, as in Figure 1. This is consistent with sustainability reasoning, by allowing for investment in research that forestalls loss of productivity through land degradation. Land that is awaiting development would be classified into a single separate category, noting that in practice there is likely to be a spectrum of feasibility, from land that is available immediately using existing technologies and under existing investment conditions, through to land that would become attractive for development only after the discovery of new technologies or under the pressure of increased demand. Having a separate category of land awaiting development, instead of quadrants 5 and 6 set in a fixed relationship to the rest of Figure 1, allows newly-developed land to fall anywhere in the graph of productivity against potential returns from research.
The location of any particular land type within the framework of Figure 2 is very likely to change with time. Implementation of improved technologies will move land from the bottom right of the figure towards the top left of it. Thus, the adoption of green revolution technologies in Asia over the period 1961 to 1990 lifted average rice yields from 1800 to 3700 kilograms a hectare4. In the process, the potential benefit from further investment in conventional agricultural research was sharply reduced, at least for those lands using the new technologies at levels close to their economic optima, and with the existing stock of basic knowledge. Some significant scientific breakthrough, say in yield potential or in resistance to pests, diseases or soil constraints, was then needed to move the most productive green revolution rice lands back towards the right hand side of Figure 2. Moreover, research to prevent degradation of such high-yielding lands may then have offered a more attractive investment than yield enhancing research. Generally, unchecked degradation will move lands downwards in Figure 2, to the right for those kinds of degradation with good potential for reversal through research, to the left for those lacking that potential.
4 Plucknett, D L (1993) Science and Agricultural Transformation. IFPRI Lecture Series No. 1.
4. How the Framework Might Be Used
In order to use the efficiency criteria described by Alston, Norton and Pardey, it is necessary to know the total area of each category of land as well as its current productivity per unit area. This permits current values of production to be calculated. The other three criteria, the maximum increase (or decrease forestalled) if the research were successful, the probability of such success, and the likely rate and extent of adoption, can all be factored into the potential benefits achievable through investment in research.
One of the greatest difficulties in attempting to set agricultural research priorities objectively is the lack of reliable global data sets. In the case of agricultural land, reasonable data are likely to be obtainable on area and current productivity, but at present the potential change achievable through research can be estimated only on the basis of well informed judgement. Because of deficiencies in existing knowledge of values for the efficiency criteria in research, in most cases it may be possible to agree only on the relative rankings of different types of land on the X axis, without it being possible to assign any exact values to them.
Members of the CGIAR may find most value in assessments of the relative priorities of research on different classes of land within the mandated region of a single centre and the national systems with which it works. Agro-Ecological Zones (AEZs) would be one obvious basis of comparison. Another would be on political boundaries, because of the powerful influence of social and institutional factors on the implementation of new agricultural technologies. Also, the data permit poverty weightings to be made only for political boundaries at present.
The following paragraphs offer some very preliminary suggestions of questions that would be worth asking about the priorities of different land types for production of annual grain crops, ruminant livestock and forests.
In the case of croplands, one of the triggers for TAC's review of the marginal lands issue was the perception by some people that they warranted a higher priority in CGIAR research than they had hitherto received. It is true that the major impact of green revolution cropping technologies has been in irrigated areas, and the possibility that some categories of rainfed land may now offer a higher rate of return on research deserves serious consideration. The highest immediate priority may be for socioeconomic research to understand better why the biophysical components of the very successful green revolution technologies have not been used to a greater extent on rainfed cropping lands. Government policies may be part of the answer. In developing countries, irrigated agriculture has generally been subsidized much more heavily than rainfed agriculture.
The case for additional CGIAR crop and livestock research to enhance the productivity of marginal lands is questionable. The information available to the Marginal Lands Review indicated that they had certainly not been neglected by the centres. Indeed, the Report argues strongly that the primary problem of poor people in low potential croplands is not technology as such, but deficiencies in the prevailing institutional and policy arrangements.
Concern about land degradation seems to have motivated much of the external interest in CGIAR research priorities for grazing lands used for ruminant livestock production. Although the very large area of grazed rangeland that is too dry for cropping or for growing improved forages does not support a large proportion of developing-country livestock or people, its research priority versus that of higher rainfall regions needs to be assessed objectively. Despite the pathetic appearance of arid grazing lands during prolonged drought, they can recover rapidly in subsequent periods of normal rainfall5. In the current state of knowledge, most experts would assess them as offering relatively low potential rates of return from investment in agricultural research.
5 de Haan, C, Steinfeld, H and Blackburn H (1997). Livestock and the Environment. Finding a Balance. European Commission Directorate-General for Development, Development Policy, Sustainable Development and Natural Resources.
There is a significant area of land in sub-Saharan Africa awaiting development for ruminant livestock production, if the constraint of trypanosomiasis could be overcome. This category of land has, by implication, received a high priority in CGIAR investments. The chief question mark over current biotechnical research on trypanosomiasis has attached to its probability of success, not the priority of the associated land type.
There has been very limited interest in comparisons of land types for tropical forestry. A far more important question is that of research priorities amongst alternative forms of land use within the same land type, notably in the remaining forests of the humid tropical lowlands. The framework of Figure 2 could be adapted to this purpose. Much of the land that was deforested in South America during the last two decades went into ruminant livestock production, whereas cropping and logging were the major causes of deforestation in Asia and sub-Saharan Africa6. The critical issue of prioritization is to estimate relative rates of return from research on the use of lowland rainforest for conservation of biodiversity plus production of non-wood products, versus that from research on the use of the same land for logging, for forestry plantations, for tree crops (agroforestry), or for annual food crops. The challenge is to find common units of measurement that allow conservation of biodiversity to be plotted along with subsistence and dollars on the value axis.
6 Alexandratos, N (Ed.) World Agriculture: Towards 2010. An FAO Study. (New York: John Wiley and Sons, 1995)
Figure 1. Table 2.2 from TAC Working Document: Report of the Study on CGIAR Research Priorities for Marginal Lands. SDR/TAC:IAR/96/18.1
Present Use Value
High Potential to Expand Production
Low Potential to Expand Production
Favoured Agricultural Lands
Marginal Agricultural Lands
Lands at Low or Zero
Figure 2. A framework for prioritizing land types in agricultural research.