3.1. Introduction
3.2. Dimensions of CGIAR Priority Setting
3.3. The Process
3.4. Agricultural Research, Productivity and Income Growth
3.5. Economic Growth and Poverty Alleviation
3.6. Agricultural Research and the Conservation of Natural Resources
3.7. Methods for Setting Priorities for Agricultural Research
3.8. Poverty as an Element in Priority Setting
3.9. The Efficiency Criterion and Priority Setting
3.10. Other Elements Considered in Priority Setting
3.11. Concluding Remarks
TAC's framework for priority setting, the principal theme of this Chapter, rests squarely on the CGIAR's mission and goals and its focus on poor people, particularly women. These concerns were established at Lucerne and have been reaffirmed since. (See especially Milestones of Renewal.) 1 The Group aims to achieve its mission and goals efficiently and effectively. Priority setting, and subsequently resource allocation, rest on four points: a people-centred approach, goals, efficiency and current allocations. Discussions within the Group on the research agenda suggest that current resource allocations are roughly consistent with its goals and, for the most part, with TAC's earlier recommendations. Hence, TAC employed current allocations as its point-of-departure in applying priorities to resource allocations for 1998-2000. These elements have guided TAC's work on priority setting.
1 CGIAR. Milestones of Renewal. October, 1996
Conceptually, these primary considerations tie in well with the work of the centres and are the subject of several sections of this Chapter, especially 3.4, 3.5 and 3.6. Subsequent sections (3.7, 3.8 and 3.9) describe how such considerations influenced decision making on activities, and particularly, commodities and sectors.
TAC's framework for priority setting can be divided into activities, sectors, commodities and Systemwide activities. This represents a change from the framework for the 1992 undertaking in that TAC has not explicitly considered regions as a separate dimension of decision making (see below).
The activities considered number 19 and consist of those listed in the 1992 report. (See Annex I.) In the course of its deliberations, TAC has noted some potential improvements in the categories of activities; these will be discussed with the Group at ICW97.
TAC considered four sectors: crops, livestock, forestry and fisheries. The commodity portfolio reviewed in this round of priority setting is similar to that utilized in 1992; deliberations first cover sectors and then commodities within crops. As in 1992, Systemwide activities have been included in these deliberations. (See Chapter 7.)
With regard to regional allocations, TAC proposes that these emerge from the primary considerations in priority setting rather than be imposed a priori. This approach assures a more even-handed treatment of the overarching goals of the System and of any future opportunities which may occur. What is more, since the introduction of standardized project reporting, the CGIAR's internal information system can now more accurately estimate the System's investment by region.
In view of the commitments made by the System as a part of the renewal process, TAC presented its first round of recommendations to the Group at MTM96. Immediately after MTM96, TAC started work with the centres on the MTPs for 1998-2000, taking into consideration the outcome of the discussions that took place at the meeting. The centres MTPs for 1998-2000 were reviewed by TAC in March 1997 and will be reviewed by the Group at MTM97, prior to deliberations on financing the 1998 research agenda.
Work on the 1998-2000 priorities started at TAC 66 in March 1995, with emphasis on the major dimensions of priority setting. This was followed by TAC 67 in July 1995, which was attended by specialists in several areas identified as being of particular importance by TAC 66. At ICW95, TAC invited the Group to comment on its interpretation of the Group's important underlying concerns. Elements of that discussion are evident in many of the points that follow here. The findings of ICW95 were integrated into the report to TAC 68 in December 1995 and laid the foundations for MTM96. In addition, since mid-1994 TAC has been engaged in meetings with representatives of NARS aimed at encouraging greater interaction with TAC and the centres on mutual concerns, including priority setting.
At MTM96, the orientation of TAC's efforts was endorsed, certain elements within the analysis were sharpened, and the process towards developing the MTPs was discussed. From mid-1996 to February 1997, TAC has been interacting with centres on the development of MTPs. These exchanges, largely aimed at ensuring that MTPs contain the information necessary for the Group's decision making, have contributed to the orientation and clarity of the MTPs.
Comparison of the 1992 decision variables for priority setting, with those employed in 1996, show both similarities and differences. The similarities are concerns for poverty, the mix of quantitative and qualitative judgements and, for commodities, a point of departure based on value of production and current allocations. The differences lie in this round's more forward-looking perspective, its approach to quantitative work and the fact that fewer decision variables are employed.
TAC notes that in most developing countries, especially the poorest, a large portion of the population works in Agriculture and a high proportion of the average family budget is spent on food. It would seem to follow then that increased productivity 2 in Agriculture is a necessary condition to achieving income growth in poorer economies. However, as income levels increase, the role of Agriculture decreases (see Table 1), offering less scope for agriculturally driven growth.
2 Productivity is used in its broad sense, relating the value of product to the cost of inputs. Value is taken to include the relevant characteristics of product, rising as those characteristics (e.g., taste or storability) improve, and includes the value ascribed to use, e.g., home consumption. Cost, meanwhile, is taken to include the opportunity cost of the bundle of resources utilized in production. Productivity is seen to be broader than, for example, yield and includes, where relevant, non-market considerations. An increase in productivity implies the same product (appropriately valued) can be produced with fewer inputs.
Table 1: Agriculture's Share in National Income, Employment, and Export Earnings by Country Income Groups, 1989
|
Country Group* |
Average Per Capita Income |
Agriculture's Share in % |
||
|
GNP |
Employment |
Export Earnings |
||
|
Low Income |
||||
|
Less than US$ 250 per capita |
190 |
49.3 |
76.5 |
65.0 |
|
US$ 251-500 per capita |
380 |
33.1 |
68.2 |
44.3 |
|
Middle-Income |
||||
|
US$ 501-1,000 per capita |
800 |
22.8 |
52.7 |
40.5 |
|
US$ 1,001-2,400 per capita |
1,580 |
12.8 |
29.8 |
39.8 |
|
Upper-Middle Income |
||||
|
US$ 2,401-6,000 per capita |
3,530 |
8.7 |
21.8 |
16.5 |
|
High Income |
||||
|
US$ 6,001-16,000 per capita |
11,520 |
4.6 |
8.8 |
18.9 |
|
Greater than US$ 16,001 |
20,460 |
2.8 |
4.4 |
10.8 |
* Country income groupings arc taken from the 1991 WDR but with subdivisions within groups that give roughly equal numbers of countries in each.Source: World Development Report 1991, except agricultural employment data which were obtained from the 1991 FAO Production Yearbook; as quoted in "Agricultural Sector Review," World Bank, Agriculture and Natural Resources Department, July 1993.
Analysis supports the notion in the previous paragraph. IFPRI studies show that an increase of US$ 100 in agricultural income gives rise to more than US$ 50 of added income in the rest of the economy. Data from Africa suggest even higher multipliers in that region. 3 Moreover, current research demonstrates that, especially in the poorest countries, the multiplier effects of agriculture on the remainder of the economy are greater than those for investment in any other sector. (See Box 1.) The statement in the World Bank's Agricultural Sector Review 4 is even more emphatic: "broad-based agricultural growth, involving small and medium-size farms and driven by productivity-enhancing technological change, offers the only way to create productive employment and alleviate poverty on the scale required". One of the conclusions reached in an illuminating country study by Ravallion and Datt is that "Fostering the conditions for growth in the rural economy - in both primary and tertiary sectors - must thus be considered central to an effective strategy for poverty reduction in India". 5 They also observe that both urban and rural poor gain from rural sector growth, but that urban growth had no discernible impact on rural poverty and, indeed, had adverse effects within urban areas.
3 Delgado, C. Bringing Previously Disadvantaged Rural People into the Economic Mainstream: The Role of Smallholder Agricultural Production in sub-Saharan Africa. IFPRI, Washington, DC, 1996.4 World Bank, Agriculture and Natural Resources Department, July 1993.
5 Ravallion, M. and Datt, G. How Important to India's Poor is the Sectoral Composition of Growth? The World Bank Economic Review. 10:1-26. p. 19, 1996.
Historically, productivity increases in Agriculture come from improved technologies, education for producers, better infrastructure, improved institutions and more effective policy. Improved technologies have provided the most noteworthy and the most reliable source of increased productivity (see the World Bank observation in the previous paragraph). Improved policies and, it is beginning to be perceived, improved institutions, have also proven fruitful. Research in all these make up the portfolio of the CGIAR.
How does increased productivity lead to higher incomes? Agricultural producers gain directly via lower production costs, then increased returns stimulate increased spending on consumption and on investment which in turn, augments the incomes of workers in other sectors who then increase their own spending. The result is rising rounds of spending across the economy and the multiplier effects described by IFPRI. This scenario reflects Agriculture's role as an engine of growth.
Moreover, the lower food prices accompanying increased output serve as a lubricant to growth in the remainder of the economy, usually through their effect on wages and, via wages, on savings and further investment. That investment, in turn, contributes to an ever larger capital stock, again with further implications for growth, that ultimately draws workers into non-Agricultural pursuits through higher wages. This outcome will be of rising importance as urbanization proceeds in developing countries. In addition, lower food prices directly benefit rural consumers, especially poor rural consumers.
Clearly some of the consequences for the poor are direct and more immediate while others, especially those associated with non-Agricultural employment, are indirect and longer term in nature. (See also Box 1.)
These considerations throw light on the poverty/food insecurity nexus. Food security rests on the availability of food and on access to it. Globally, there is little question about the availability of food. However, as argued so cogently in Alexandratos, this is not the critical question. 6 For even when food is available, there are millions of people across the developing world, and especially in poorer countries, who do not have access to food because they have inadequate incomes. "It is now widely recognized that failure to alleviate poverty is the main reason why under-nutrition persists." 7 This leads Alexandratos to emphasize that the production of food serves two important functions which contribute to food security: one is to assure availability while the other and, for developing countries, the more important is that the production of food is a major source of income for the poor. In the short term, transfers, either internally or via food aid, may assure food access. In the long term, however, incomes must increase if access is to be gained and maintained, and increased income must rest ultimately on increased productivity. The alleviation of poverty through increased productivity, therefore, is the way to improve the nutritional status of the poor and to augment their food security.
6 Alexandratos, N. (ed). World Agriculture: Towards 2010, An FAO Study (New York: John Wiley & Sons, 1995).7 Ibid.
|
Box 1: Linking Agricultural Research to Growth This theme warrants treatment at three levels: that of the farm, of the region, and of the nation. At the farm level, growth permits an increase of spending in several dimensions. Clearly it can lead to higher levels of consumption. More important for the long term, it can lead to on-farm investments in future productivity growth including those in natural resources management, to investment into non-farm income diversification (including small business activities), and to investments in education. Beyond that, through increased savings, farm level growth augments the capacity and willingness to accept risk. At the regional level, agricultural growth leads to powerful demand linkages to the non-farm economy. These include consumption demands, demands for farm inputs, demands for marketing and agricultural processing services, and demands for investment goods. Empirical studies show that the first three lead to annual income multipliers of between US$ 0.5 and US$ 0.8 per dollar of additional value generated in agriculture, with most of that coming in rural towns. Of the three, consumption demands play the largest role. More recent work suggests that demand for investment goods are more powerful even than consumption spending when their impact on future productivity and income generation is considered, i.e., there are large dynamic multipliers. Agricultural growth in high potential areas often generates additional labour demands and higher wages which attract seasonal migrant farmer workers from poorer regions, leading to important income multipliers elsewhere. Examples include the migrations from eastern India to Punjab and Haryana, and the Burkina Faso labour migrations to the cocoa farms in Côte d'Ivoire. While these payments tend to reduce the host region's multipliers, they can increase multipliers in the home region, as income from non-local sources can be an important catalyst for investment, and subsequent growth in output, in the home region. Regional multipliers vary with farm size. Medium-sized family farms generate the largest regional multipliers. Growth via smaller farms usually has larger regional multipliers than large farms because the income generated per hectare is larger, more labour intensive practices are used, and small farms use a larger share of locally produced goods in production, consumption and investment. Very small farms generate small multipliers because they have few cash transactions. Regional multipliers also tend to reflect the state of infrastructure. With poorer roads, for example, much of an income increase is spent on diversifying the diet, a result starkly shown when comparing villages with good versus bad road connections. As well, the level of income tends to influence the multipliers which tend to become larger with regional economic growth, at least through middle-income status. Therefore, favoured environments well served by infrastructure would be expected to have larger regional multipliers than marginal environments with less infrastructure. For national multipliers and with well-functioning markets, the differences in demand linkages, for the reasons described above, would not operate. All investments would generate the same long-term income multipliers and these would be small or even zero because all resources remain fully employed. Markets and conditions are rarely so efficient, and labour and capital in particular are not notably mobile across regions and sectors. The targeting of agricultural growth by region and type of farm does matter. In the more real world, the regional multipliers are probably good guides to the relative differences in national multipliers. It should be noted that where labour is immobile, investments in poor regions may generate smaller regional and national income multipliers than investments in high potential areas, but may be necessary to reach the very poorest of the nation. In countries with large poor urban populations, however, the larger income multipliers from investments in high potential areas may have a bigger impact on poverty. These issues need further research. A number of other macro or national level benefits accompany agricultural growth. Among these are: capital flows out of agriculture through the banking system add to investment potential elsewhere in the economy; increased entrepreneurial activity in rural areas, whether in on-farm or non-farm pursuits, augments managerial skills which can be applied in other parts of the economy; lower food prices have implications for wage rates that make the economy more competitive in international markets; balance of payments are strengthened through increased exports or through reduced imports; and the higher incomes in agriculture provide a broader tax base for the public development of infrastructure and education. As for comparisons with growth linkages through other strategies, recent IFPRI work shows that development strategies based on agriculture are superior to industrial-led strategies for the poorer developing countries and, as in the earlier discussion, that targeting the small farm sector is better than targeting the large farm sector. These studies measure success in terms of growth rates but also show that the income distribution shifts towards the poor. As expected, these results are sensitive to assumptions about the mobility of capital and labour. Importantly for the CGIAR, the superiority of agricultural-led growth diminishes with economic development. These results suggest that efforts aimed at poverty alleviation should be via productivity growth in agriculture, focus on poorer countries and emphasize rural-labour intensive activities. |
Summary
The major impediment to appropriate nutrition and food security is access to food by the poor. Alleviating their poverty will require increased productivity, leading to both higher incomes for producers (via lower costs) and to consumers (through lower prices). Improved technologies, policies and institutions are proven vehicles for achieving increased productivity in Agriculture. Improvements in these areas ultimately rest on research of the kind emphasized by the CGIAR. In turn, higher productivity in Agriculture has been found to be among the most effective means for achieving broader economic growth. How growth affects poverty is the subject of the next section.
For a brief, lucid and penetrating treatment of over 200 years of research on poverty and economic growth, see the review by Lipton and Ravallion, with its nearly 600 references. 8 In this they refer to a paper by Ravallion 9 commenting on 16 studies chosen because their data enabled a comparison of economic growth with the reduction in the proportion of people below a commonly defined poverty line. (See 3.8.) The studies showed that the proportion of such people decreased with economic growth. Moreover, higher-order measures of poverty (See 3.8) also showed declines in poverty, implying that "benefits from growth are typically felt well below the poverty line". 10 The review further states that "...even when growth has been associated with rising inequality, it appears that poverty has typically fallen". 11 All of this is not to suggest that governments should concentrate attention solely on promoting growth. A number of factors influence the relationship between growth and the reduction of poverty, and governments, through targeting and policy, may have a role in promoting patterns of growth more congruent with poverty alleviation. 12 Among such strategies is that of "shifting investments towards rural labour-intensive activities". 13
8 Lipton, M. and Ravallion, M. "Poverty and Policy," in Behrman, J. and Srinivasan, T.N. (eds.). Handbook of Development Economics, Vol.3. Amsterdam: North Holland Press, 1995.9 Ravallion, M. "Growth and Poverty: Evidence from the Developing Countries in the 1980s". Economics Letters, Forthcoming.
10 Lipton and Ravallion, op. cit., p. 2607.
11 Ibid.
12 Ibid., p. 2608.
13 Ibid., p2609.
Recent studies reinforce the view that economic growth reduces poverty. Deininger and Squire 14 reviewed some ninety studies reflecting long-term economic growth and income distribution. In five out of six studies, the real income of the lower 20% of the population increased with growth. This is the strongest recent evidence in support of this critical association.
14 Deininger, K. and Squire, L. A New Data Set Measuring Income Inequality. The World Bank Economic Review. 10, 565-591, 1996.
It can be argued that strategies that slow the rate of economic growth will, sooner or later, impede the reduction of poverty and that growth itself is the best pro-poor strategy. This point of view is, to some extent, consistent with an observation by Lipton and Ravallion that a key question in assessing the effects on poverty of macro-economic policy is: "did adjustment raise or lower the rate of growth?". 15 However, another view is that barriers to the effective functioning of markets - whether because of policy, shortage of information or inertia in labour markets - argue for the implementation of pro-poor strategies These considerations make up one strand of the rationale underlying the emphasis on labour-intensive activities in rural areas; they also support TAC's decision to include a modifier for poverty in its analysis of priorities.
15 Ibid., p. 2613.
Summary
With regard to the evidence that economic growth leads to poverty alleviation TAC has concluded that there is strong empirical evidence that economic growth in poorer countries leads to poverty alleviation, even in cases where growth leads to less equal distribution of income. However, TAC recognizes that growth alone is not sufficient to achieve the Group's pro-poor goals and that a further emphasis on poverty in TAC's priority setting is required.
The link between CGIAR's activities and the health of the environment is more apparent than the link to poverty alleviation. Conservation requires technologies that directly serve the various purposes of producers, including those pertaining to natural resources, or which can be made to do so through policy (see below). Productivity increasing, resource-conserving technologies and policies are products of research. The CGIAR is closely involved in research relevant to conservation to which much of its work contributes, directly or indirectly.
Over the years, TAC has examined the subject of conservation on a number of occasions and has concerns in several areas. These concerns include questions relating to which resources should be conserved and the extent to which the System should look beyond producer gains to off-site effects. Other concerns relate to the nature of the System's role in resource conservation; options include developing more suitable technologies for sustainable production, developing effective research paradigms through the experience of developing more suitable technologies, leading the way in the science of resource conservation, and advocacy for conservation itself.
Sustainability has been a major theme for TAC as it is for all those concerned with natural resources. The meanings ascribed to sustainability vary from maintaining nature's endowment intact to ensuring "resiliency", so that further clarification is required to convey meaning. TAC's 1996 paper, 16 consistent with the System's concern for the poor, focuses on the conservation of natural resources of concern to present and future generations of poor producers and consumers and, to borrow a term from Alexandratos, 17 in the context of returns and "... timescales which meet their differing circumstances or risk perceptions". TAC also includes the benefits from involvement in global environmental concerns such as greenhouse gases and life systems, as well as ethical, aesthetic and amenity considerations. Much of CGIAR's work on natural resources, indeed much of the System's total endeavour, has implications for both the local and the global environment.
16 Technical Advisory Committee, Consultative Group on International Agricultural Research, Priorities and Strategies for Soil and Water Aspects of Natural Resources Management Research in the CGIAR (TAC Secretariat, FAO, Rome, 1996).17 Alexandratos, op. cit., p. 400.
Previous reference has been made to policy as an agent in promoting the use of environmentally friendly technologies. One example is the GEF's policy of offsetting the added costs of carbon-efficient power plants in order to limit greenhouse gases without raising the cost of power to users. This is an industrial example of reducing the trade-offs between development and the environment. Certainly a major challenge for policy in Agriculture is the need to influence the behaviour of many thousands of largely independent agents. Reducing trade-offs between development and the environment, but with limited institutional capital and while restricting rent-seeking behaviour, is a daunting task. TAC is aware of this challenge and is concerned for the orientation of CGIAR research on policy issues. (See Chapter 5.)
Increases in productivity have implications for the environment, some of which have been negative in the past. 18 However, increases in productivity have also had a positive, albeit indirect, effect on the environment. In particular, it is frequently observed that the lower product prices accompanying increases in productivity in favoured areas, reduce encroachment onto fragile lands and forest margins.
18 See Alexandratos, Ibid., and World Food Summit Technical Paper 6 for reasoned discussions of the issues.
The two primary concerns of the System are embedded in a nexus of three problems: poverty, environmental degradation and population growth. Poverty, the pivotal element, has immediate consequences for the environment, as the poor are driven to focus on the near term, with less attention to the long term when the benefits of conservation come into play. Poverty is also strongly associated with population growth, (see Table 2) either because of the direct benefits from large families, as seen by poor individuals, or because poverty limits what society can do to reduce growth rates through investment in education (especially for women) and better health. In turn, rapid population growth, combined with poverty, promises more damage to natural resources in the future as degradation of natural resources ultimately threatens productivity, further reinforcing poverty. The above leads to the conclusion that the way forward is through poverty alleviation. In this context, Alexandratos opines that "...much of the mismanagement of natural resources in developing countries relates to poverty and to the lack of economic growth to provide better and sustainable livelihoods outside of subsistence agriculture". 19
19 Alexandratos, op. cit., p. 400.
Table 2. Relationship between per capita income in 1990 and annual population growth, 1980-92, among 116 countries by income group. i
|
Countries |
Mean GDP ii |
Mean Growth Rate |
|
Lowest |
631 |
2.51 |
|
Low income |
1092 |
2.58 |
|
Low middle |
1822 |
2.36 |
|
Middle middle |
3008 |
2.29 |
|
Upper middle |
4573 |
1.51 |
|
Low upper |
9185 |
1.17 |
|
Upper upper |
14832 |
0.77 |
i Calculations by Lant Pritchett, Senior Economist, World Bank, via personal communication. Countries are evenly divided among the income groups.ii Data are from the Penn World Table 5.6 and are in terms of dollars of Purchasing Power Parity.
It is ironic that efforts to address today's natural resource concerns may well have the effect of heightening tomorrow's damage; to the extent that strong efforts to protect resources constrain income growth and thereby give rise to larger populations which may well add to the pressure on resources in the future. The interaction among the three elements: poverty, environmental degradation and population growth, adds layers of complication to effective decision making in this arena. An awareness of these interactions argues for what Alexandratos calls "strategies for reducing trade-offs between productivity and conservation". 20
20 Ibid., p. 20. (Not page 20, check)
TAC has not included in its priority setting a special modifier for degradation of soil and water. The priority accorded to research in this area is dependent upon the effects of soil and water degradation on present and future productivity, and the global environment. 21 Data and measures on these effects are not yet available.
21 Technical Advisory Committee, op. cit.
More generally, while much has been said about themes pertaining to NRM, a careful assessment gives evidence of three important weaknesses. Firstly, there is a lack of standardized methodologies for surveys and little attention has been given to the basic principles of sampling theory. This means that the soil and water data that is available for developing countries are not on a broad enough scale to allow a reliable assessment of current trends and the condition of these resources in order to judge future food security. For example, to date IFPRI's reviews of the literature on land degradation in the developing world found that there have been few global studies of the impacts of degradation on agricultural production and that estimates of the monetary value of those impacts vary widely. Consequently, IFPRI simply summarized the information in terms of hot spots where degradation is known to be serious and bright spots where improvement is known to be occurring.
Another common defect of the available literature is that little distinction has been made between forms of degradation that are comparatively easy to stop or reverse and those that are not. For example, plant nutrient depletion and overgrazing of vegetation receive about the same weight as the removal of topsoil by erosion and the contamination of water resources with salts, fertilizer nutrients or pesticide residues. Yet nutrient depletion has been readily reversed with fertilizers in industrialized countries and in a growing number of developing countries. Again, research carried out by CGIAR in tropical Africa, and by many industrialized country institutes, has shown that overgrazed grasslands can recover quite rapidly under reduced stocking. In contrast, the effects of soil loss and water contamination may be very difficult and costly to reverse.
No serious attempt has been made to reconcile contradictions between expert opinion and production statistics. The two global surveys of NRM experts quoted in the IFPRI review found that there has been very serious land degradation since the mid-century. For example, the GLASOD study rated 38% of 1.5 billion hectares of cropland as being degraded. Over the same period, world food production has grown dramatically and prices have fallen. Either the NRM "experts" have seriously overestimated the severity of degradation for agricultural production in the short term, or the growth in production would have been potentially much larger in the absence of the land degradation, or some combination of the two; on the published evidence it is impossible to decide which. Another very worrying possibility is that the recent growth in food production bears a hidden cost to the environment in the increased use of pesticides and fertilizers, loss of biodiversity, and expansion onto new land, and that this price will have to be paid in the future. However, available data provides little guidance about the magnitude of this cost.
The CGIAR alone cannot fill this knowledge gap, but it can work with others to do so and also ensure that its own NRM research has this as one of its objectives. (See Chapter 5.) The areas of research that the Group has endorsed for the strengthening of soil and water aspects of NRM in the System are listed in Chapter 4.
Summary
The struggle of the poor to survive underlies a good part of the threat to biodiversity, land and water in developing country Agriculture. In this context, TAC has concluded that emphasis be given first to the implications of conservation on present and future productivity, broadly defined to include on-farm and off-farm benefits and costs, and secondly to broader global environmental effects. 22 TAC notes the importance of positive indirect linkages between productivity and conservation, where productivity increases can lead to reduced pressure on more marginal lands, and the potentially negative direct linkages where intensification of production sometimes leads to burdens on soil and water quality. In the latter case, the importance of pursuing opportunities to reduce trade-offs between development and the environment has been emphasized. In addition, TAC notes the now broader view of conservation guiding the CGIAR System, i.e., the view resting on four classes of capital - physical, human, natural and institutional - each of them important to future generations. 23
22 Ibid.23 Monitoring Environment Progress: A Report on Work in Progress (??), Vice Presidency for Environmentally Sustainable Development. The World Bank, 1995, Washington D.C.
To set priorities is to order activities in ways that are most consistent with objectives. Agricultural research is an especially difficult arena for convincing priority setting. Apart from uncertainties, other issues intervene. In Science Under Scarcity: Principles and Practices for Agricultural Research, Evaluation and Priority Setting, Alston, Norton, and Pardey review the various elements that challenge the setting of priorities for agricultural research. 24 The book is the latest, and arguably the most complete, presentation of the issues influencing priority setting in research, and includes a presentation of the advantages and disadvantages of various approaches. A prime point is the problem of dealing with multiple objectives. The authors favour a single objective based on income growth while recognizing that in the absence of other distributive instruments or mechanisms, research may lead to preferred effects on income distribution.
24 Alston, J. Norton, G. and Pardey, P. Science Under Scarcity: Principles and Practices for Agricultural Research, Evaluation and Priority Setting (Ithaca: Cornell University Press, 1995).
There are several approaches for those who would set public research priorities and allocate public resources so as to promote economic efficiency. Alston et al advocate a model based on estimated economic surplus. As an alternative, applicable in limited circumstances only, they include congruence allowance. This approach simply posits that, in the absence of other evidence, research should be "funded in proportion to the value of production". 25 TAC has used congruence analysis as the point of departure in its work on priorities relating to commodities.
25 Ibid.
Alston et al also note that there are many situations in which insufficient information limits the use of quantitative methods, and that qualitative estimates of criteria should be obtained where available and combined with structured discussion where quantitative approaches are impossible. 26 TAC has followed this approach in dealing with activities and also notes its strong support for the idea that over time, and with appropriate investment, better informed measures can be developed.
26 Ibid., p. 479.
In the previous round of priority setting, a multi-dimensional model brought several modifiers into the analysis. While this approach is not without its shortcomings, it was felt that, on balance, it was the preferred approach, given the System's then desire for a more quantitative orientation to priority setting than in the past. With 1997 as the base year for the new round of priority setting, and to the extent that the 1993 TAC recommendations were adhered to in the 1997 allocations, the previous priority setting will influence the current round of planning because of its influence on the base. In addition, the residual affects of the 1992 a priori judgements about regional priorities will still be visible.
Summary
In setting priorities, TAC has relied most on structured discussion supported by quantitative and qualitative information, the so-called Delphi Approach. Its earlier recommendations and current budget allocations have influenced the point of departure. Other major factors were the locus and level of poverty, opportunities for women, science, alternative sources of supply, recent experience, possibility of success and prospects beyond research, such as the quality of extension services.
The commodities that make up the CGIAR portfolio were initially selected by the System because of their importance to poor producers and consumers. Therefore, the CGIAR portfolio itself reflects a concern for poverty. The following relates to the further integration of poverty into TAC's analysis, starting with a brief review of measurement themes, and concluding with the measures applied to modify congruence analysis and guide structured discussions about other elements that should influence priorities and resource allocations.
Measurement
The definition of poverty, and its amelioration, measurement and effect on behaviour, has motivated discussion and study for over 200 years. For a brief lucid and penetrating treatment of the evolution of ideas on the theme see Lipton and Ravallion. 27 Many of the elements that follow were drawn from their piece.
27 Lipton, M. and Ravallion, M. op. cit.
While poverty imposes severe limits on welfare, efforts to describe and measure it are complicated by the fact that perceptions of poverty and welfare vary from place to place. "However, it is not controversial that inadequate command over commodities is the most important dimension of poverty, and a key determinant of other aspects of welfare, such as health, longevity, and self-esteem." 28
28 Ibid., p. 2553.
Spurred on by the growing emphasis on greater efficiency in reaching the poor, development assistance agencies have sought a pragmatic measure of poverty, clearly related to well-being and with tolerable costs of implementation. The better known attempts include assessing basic needs (health, food, education, water, shelter and transport), capabilities (which sees consumption as permitting desired activities) and human development. However, these and other criteria have proved difficult to implement. What has emerged are measures based on control over commodities, gathered from consumption or income figures - co-related with the main indicators of basic needs, such as health, longevity and self-esteem - which are comparable over time and space.
Initially, poverty was measured by the number of people, or the proportion of people, below a poverty line, defined in terms of income adjusted for purchasing-power parity. This was followed by a measure that took account of the gap between the poverty line and those below it in order to give a measure of the depth of poverty. In turn, this was augmented by a measure reflecting the severity of poverty via the squared poverty gap measure. 29 Each of these measures is "exclusive", in the sense that all weight is given to those below a poverty line.
29 Foster, J. Greer, J. and Thorbecke, E. "A Class of Decomposable Poverty Measures", Econometrica 52 (3): 761-765.
It can be argued that all members of society, the poorest as well as the less poor, should be included when calculating the degree of poverty within a country. A country's mean income is a simple "inclusive" measure that does precisely that. However, in the view of many, an "inclusive" measure should reflect the severity of poverty, and a country's mean income does not do this.
Ravallion has demonstrated that the theoretical differences between "inclusive" and "exclusive" measures are blurred by the accommodation of measurement errors and uncertainties about what should be brought into a measure of well-being. 30 Using data from 40 developing countries, he goes on to show that there is little empirical difference between "exclusive" poverty measures, such as the three referred to above, and an "inclusive" measure which incorporates income inequality by giving increasing weight, even moderately increasing weight, to the poorest. Given this, Ravallion concludes that the choice of measure will make little difference to the perception of relative poverty. That conclusion facilitates the work of TAC. (See Box 1, Annex II.)
30 Ravallion, M. "Measuring Social Welfare With and Without Poverty Lines", American Economic Review, 84 (2), 1994: 359-364.
In TAC's poverty measure (see Annex II), the critical variables pertain to average income, income distribution and a critical level of average income above which the country no longer figures directly in TAC's priority setting. Differences in these variables have strong impacts on the level of the poverty indicator, hence on TAC's recommendations. Annex II contains tables showing the sensitivity of the poverty weighted shares to various assumptions about these and other variables, viz., the importance of the depth of poverty, gender considerations, rural/urban proportions and the rate of growth in income. (See 3.10.)
Implementation
As noted earlier, one strategy for allocating research resources uses congruence analysis where greater relative value leads to greater relative resources. Following decisions by the Group, TAC has employed poverty-weighted congruence analysis. This uses the poverty indicator of the preceding discussion (and Annex II) as a multiplier in determining value. Poorer countries have a higher poverty indicator than do less poor countries; hence the products of poorer countries have higher weights applied to their estimated value of production. For example, the weight for every dollar's worth of product from Ethiopia is 0.98; for China 0.23 and for Turkey 0.11. (See Annex II for calculations.)
Summary
The decision to feature poverty in priority setting led TAC to seek a measure that would permit comparisons over space and time, be sensitive to differences in the inequality of income distribution among countries, give increased weight to the poorest within a country, facilitate comparison of different ethical views (or value judgements) about the importance of the degree of poverty, assure consistency with good practice, and accommodate the limitations imposed by the data. The measure described in Annex II has these qualities. Finally, as applied in congruence analysis, the measure clearly shifts priorities towards the products of the poorest countries, favouring technology development for their commodities.
Earlier discussion noted the Group's concern for efficiency and effectiveness in priority setting. This concern led to the identification of three primary considerations.
The first of these is that the centres' products should be international public goods. The reason being that given the high opportunity costs of CGIAR resources, products limited to use in a single country that also meet the CGIAR's opportunity cost criteria would be so valuable in terms of local opportunity costs that the country itself would finance the effort, and at lower costs, thereby becoming its own alternative supplier. The reason why the CGIAR will have higher opportunity costs than a nationally focused programme is that the latter will treat benefits to other countries as spillovers, not counting them in calculating opportunity costs, while the CGIAR, given its international role, will count all benefits in calculating its opportunity costs. As the quality of being an international public good is either present or absent, this characteristic is viewed by TAC as a necessary condition for consideration in priority setting.
The second consideration is alternative sources of supply for the products of the CGIAR. If there are suppliers who can offer products at costs below those of the CGIAR, who are a reliable source of supply and who provide the products in an even-handed way, then CGIAR investment is questionable and will be limited.
The third standard emerging from concern for efficiency rests on the probability of creating an impact. The higher that probability, the higher the expected value of the benefit stream and the more likely is CGIAR participation. Of special importance in shaping the probability of success is new science. As well, new science lowers the cost of output dependent on science. Both considerations will, other things being equal, raise investment in related products.
Although other factors were examined in exploring the implications of the efficiency criterion (see the following Section), it is these three, together with the obvious requirement for critical mass, that were deemed most important.
Recognizing that the products of research initiated today will have an impact many years in the future, TAC's analysis favoured future income and poverty, together with future relative values of production. Reasonable measures of income growth provide the vehicle for projecting income. These projections, in turn, promise to have a strong influence on relative incomes across countries, hence the value of the poverty indicator and the weighting given to various commodities and activities. The influence of these factors suggested the application of sensitivity analysis and the results are reviewed in Annex II.
In 1992, TAC used a price series based on 1987-89 prepared by the Secretariat in close collaboration with the centres. Shortly after the conclusion of that work, a new set of international prices was provided by Pradado Rao of FAO. This series attempts to put all prices on the same footing and appears to represent an improvement on the earlier set used by TAC. No one, however, has built a similar data set for future prices. FAO, IFPRI and the World Bank predict prices for about half of the CGIAR commodities. TAC recognizes that, for example, urbanization will lead to higher relative prices for wheat (because of its convenience) and that income growth will boost relative prices of livestock and feedstocks, especially maize, and reduce the relative price of rice as people, particularly Asians, seek more varied diets. Even so, there is little analysis to give precision to these probable occurrences. Fortunately, because of substitutability among commodities, TAC requires only relative prices, seemingly more stable than absolute prices, even in the face of the forces mentioned above. Given these considerations, and the absence of alternatives, TAC has used the Rao construction to estimate relative future prices for CGIAR commodities.
Given the Group's concern for rural women, TAC examined further adjustments in the value of production by giving added weight when females are engaged in production. The argument made by the Group at ICW95 was that income gains to women would contribute more to the well-being of women and children than the same income gains to men. An added consideration was that, on average, women in developing countries are poorer than men. 31 Further, it is said that women had been given relatively less attention in the past and heavier weighting now was required to redress the balance. Beyond that, some argue for giving extra weighting to women in the expectation that this will hasten the reduction in population growth rates and its implications for the future conservation of natural resources as well as for incomes. Offsetting these considerations, somewhat, are studies that suggest that the production activities of poor female farmers are more similar to those of poor male farmers than better-off female farmers. Finally, it can be argued that CGIAR concerns can only be fully met if women's access to the fruits of their labour are known. While anecdotal evidence abounds, there are few systematic studies.
31 This point rests in some measure on the widely held view that some 60% of those below the poverty line are females. In a personal communication, Hilary Feldstein reports that further study suggests that the number is 54%, rather than 60%. Subsequent work suggests the proportions might be even closer to 50 percent. Quisumbing, A. Haddad, L. and Pena, C. "Gender and Poverty: New Evidence from 10 Developing Countries," Food Consumption and Nutrition Division Discussion Paper No 9 (Washington, DC: International Food Policy Research Institute).
Clearly it was not the intention of the Group that TAC set priorities exclusively in terms of women, simply that poor females, or more generally females, receive a higher weighting than males. For its purposes, TAC framed one analysis in which females and males, whatever their location and degree of poverty, were given equal weight and one set in which females had 25% more weighting than males. The implications for modified values of production are presented in Annex II, Table 7 while the implications for priorities are discussed in Chapter 6. TAC also notes that when setting priorities in terms of projects, centres will be better able to tailor activities to the needs of women than TAC which sets priorities at the global level. Also, it seems likely that national programmes and NGOs, working as they do on specific local problems, will be even more able to respond to gender-based issues. Through its review processes, TAC will assess the extent to which centres have taken the Group's admonition into account in setting their priorities and will monitor the extent to which NARS are incorporating such considerations into their own work.
As a result of discussions at ICW95, TAC set out two scenarios based on different weighting for rural and urban dwellers (see Annex II, Table 8). Group members developed several arguments for favouring rural dwellers. One was the belief that, on average, and within any given poor country, rural people are poorer than urban dwellers. A second was that urban dwellers have more access to socially provided services than do rural dwellers, hence the former's real incomes are understated vis-à-vis the latter. A third was that rural people are more numerous than urban people, a fourth was that the CGIAR, with its emphasis on agriculture, can do more for rural people than for urban people, and, finally, that growth in poorer countries has a stronger effect on overall growth than does growth elsewhere.
TAC has observed that increased productivity in agriculture affects poverty alleviation in two ways: firstly through the immediate direct returns to those holding the resources (especially labour) favoured by the productivity change and later as induced rounds of spending add further to the demand for factors (again especially labour); and, secondly through the positive consequences for real income of a decline in the price of food. Both rural and urban populations benefit from these effects. Furthermore, for most countries, the full realization of poverty alleviation will require a shift of labour from Agricultural to non-Agricultural pursuits. This suggests that a concern for rural poverty must recognize developments in urban areas as being critically important in the long run. One fact which does differentiate rural/Agricultural poor from urban poor is that the effects of productivity changes are more quickly felt in rural than in urban areas. Differences in lag time then, would suggest some added weighting to rural populations. Moreover, recent findings cited earlier favour an emphasis on the rural poor by offering evidence consistent with the premise that growth there has a stronger effect on overall poverty, rural and urban, than does growth elsewhere. 32
32 Ravallion and Dan, op cit.
Annex II has two versions of rural/urban weighting which reflect the implications of this judgement for modified congruence allowance: one with equal weighting and the second giving added weighting for rural populations. It is noteworthy that there is little difference between the implications of the two, at least at the level at which TAC recommends priorities. However, as the centres frame their priorities around projects, they may well have more options for favouring rural rather than urban populations. Careful note will be taken of centre reactions as a part of the review process.
Some Group members expressed concern for marginal environments and advocated that such environments be given added priority. Several considerations seemed to motivate this concern. One was that such areas contain either most of the poor or the poorest of the poor, However, recent evidence casts doubt on the first hypothesis (see, e.g.. Fan and Hazell 33) and the latter calls for further research. A second was that such areas suffer most from degradation and a further concentration of CGIAR resources would be congruent with the System's interest in conserving natural resources. Possibly a third was the view that, as relatively more research has been done on favoured environments than on marginal environments, the effects of diminishing returns in the former would suggest moving more resources to the latter.
33 Fan, S. and Hazell, P. Should the Indian Government Invest More in Less-Favored Areas? IFPRI, Washington, D.C., 1996.
TAC notes that more than one third of the System's resources are now destined to marginal environments and that any change in that level should emerge from the general considerations that are shaping priority setting, such as, value of production, probabilities of success and alternative sources of supply, together with perceptions on the locus, level and consequences of poverty. Centre priorities on such work will be followed via the newly emerging project framework.
Attention was given to the possibility of incorporating labour intensity in priority setting. Some commodities and some sectors utilize more labour than others per hectare or per unit value of production. The available data were reviewed but misgivings about its accuracy and current relevance were expressed. Moreover, the measures required to estimate the demand for labour are not available. TAC concluded that at its level of priority setting there is little opportunity, at this time, to bring labour intensity directly and systematically into the analysis of priorities, except as the degree of labour intensity is indirectly reflected by poverty.
Finally, TAC reflected on the possibility of employing the production function for knowledge, i.e., the description of the relationship between inputs to research and outputs from research in its assessments. Although conceptually of great importance, there is little information about the relationship, hence little opportunity to use it in priority setting at TAC's level. TAC will join those engaged in developing the information required in order to permit the future use of this relationship in priority setting. IFPRI has recently published a paper on this theme. 34
34 Huang, J., Rozelle, S. and Rosegrant, M. "China's Food Economy to the Twenty-First Century: Supply, Demand, and Trade. IFPRI, January 1997
Summary
In brief, it was agreed that: estimated benefit flows should be for 2010 and beyond; added weighting to female farmers and rural populations makes little apparent difference given the level of generality at which TAC shapes priorities but that centre priority setting should be carefully monitored to account for these considerations; the balance between favoured and marginal environments is best achieved through the major factors shaping priority setting rather than through a priori judgements; and, neither differences in labour intensity nor in research production functions can shape TAC's priority setting at present, although they could, perhaps, influence centres' priorities for projects.
This Chapter has listed and described the major elements on which TAC based its priority setting and commented on several other elements that were considered. Two other important considerations, the view of regional fora and national programmes and the perceptions of the centres, are the subjects of subsequent chapters. The recommendations in resource allocations emerging from the resulting priorities are discussed from Chapter 5 onwards.
