low high low high
Species diversity/genetic diversity (public value)
Figure 5-2. Sites with high benefit-cost ratios for on-farm conservation Source: Adapted from Smale and Bellon (1999, p. 395).
An example of an "I" location was found in the uplands of Nepal for rice.3 Farmers recognized many varieties but their genetic diversity when characterized was found to be relatively low. No modern varieties compete in that environment, so there were no opportunity costs associated with growing farmers' cultivars. An example of a "II" location was found in Mali.4 There, despite 26 years of climatic change and drought, analysis of seed samples demonstrated that farmers maintained the same overall level of sorghum diversity, though its spatial distribution had shifted. This is strong empirical evidence that farmers depended on sorghum diversity to manage risk ex ante (as compared to those in more favored, less isolated environments who manage risk through markets, ex post). Geneticists also consider the range of traits found among those varieties to be important.
3 Project entitled "In Situ Conservation of Agrobiodiversity On-Farm," implemented by the National Agricultural Research Council, Nepal, and the International Plant Genetic Resources Institute.
4 Project entitled "Development of strategies for in situ conservation and utilization of plant genetic resources in desert-prone areas of Africa," funded by International Fund for Agricultural Development and implemented by the Institut d'Économie Rural, Unité des Ressources Génétiques and International Plant Genetic Resources Institute.
Based on empirical findings, we know that in other places in the world, rural people depend on the diversity of their crops and varieties to cope with climatic risk, match them to specific soil and water regimes, and meet a range of consumption needs when markets are unreliable. These locations are often characterized as "less favored," or "marginalized"; the people who live in them are often considered to be poor on a global scale.
With economic analysis, genetic analyses, and other scientific information, we can ascertain which locations are "promising candidates" for managing rare or diverse crop genetic resources on farms. Metrics can then be used to rank areas according to their expected social and private value.
To rank locations along the horizontal axis, diversity indices developed by scientists can serve as proxies for the public value of a set of crop varieties or populations. Indices are scalars constructed from any one of several types of data (Meng et al., 1998). For example, data may measure the physical characteristics of crop plants grown in controlled experiments. Alternatively, data may summarize the patterns in DNA taken from plant tissue and observed under a microscope. Option values other than those subsumed in scientists' assessments of information value are not likely to be estimable but would generally be positive (Brown, 1990).
Adaptations of econometric models of variety choice, and nonmarket valuation methods combined with random utility models, enable us to rank candidate locations along the vertical axis using probabilistic statements. Some empirical evidence from related analyses is summarized next.
3.2. Predicting locations where landraces will continue to be grown
Broadly speaking, three generic factors are hypothesized to determine the likelihood that modern varieties are attractive to farmers and, hence, the opportunity cost of growing landraces: population density, agroecology, and development of commercial markets.
The pattern of diffusion of modern varieties of wheat, rice, and maize illustrate this point. Population density, or the ratio of the supply of labor to the supply of land, explains much about where the transition from low-yield, land-extensive cultivation to land-intensive, double, and triple crop systems has occurred (Boserup, 1965; Hayami and Ruttan, 1985; Pingali, 1997). The genetic changes embodied in seed constitute one type of intensification, which refers more broadly to the increase in output per unit of land used in production (or yield). Predictably, the adoption of modern rice varieties in the less industrialized world has been most complete in densely populated areas of their cultivation where traditional mechanisms for enhancing yields per unit area have been exhausted (Pingali, 1997).
Population densities interact with agroecological conditions in explaining the adoption of modern varieties. Since the initial adoption and rapid diffusion of the first semidwarf varieties of wheat in the irrigated areas of the Asian subcontinent during the 1970s, more widely adapted descendants of these varieties spread gradually into environments less favored for wheat production and rain-fed areas. Today, wheat landraces are cultivated extensively only in portions of the drier production zones of the West Asia-North Africa region and highlands of Ethiopia. Similarly, the adoption of modern varieties of rice is virtually complete in irrigated areas and uneven in rainfed zones, while they are largely absent in the uplands and deepwater areas.
In contrast with wheat and rice, maize is grown over a greater range of latitudes, altitudes, temperatures, and moisture regimes. Maize also has the greatest proportion of area in the less-industrialized world that is still planted to landraces. For many of the environments in which maize is grown, suitable improved materials have not been developed by centralized breeding programs (Byerlee, 1996; Perales, Brush, and Qualset, 1998). The maize germplasm that performs well in temperate climates of industrialized countries cannot be introduced directly into the nontemperate regions of less-industrialized countries without considerable additional breeding for adaptation (Morris, 1998), nor are there always economic incentives for a commercial seed industry. Thus, even if adaptation problems could be overcome through breeding, farmer demand for improved seed may be small (Morris, 1998). In maize, agroecological factors have interacted with the development of commercial seed systems in slowing the expansion of area in modern varieties.
As the orientation of crop production shifts from subsistence toward commercial objectives, the locus of crop improvement and seed distribution also moves from individual farmers toward an organized seed industry composed of specialized private and public organizations (Morris, 1998). Maize has moved substantially faster than rice and wheat in terms of an increased reliance on commercially produced seed. In a stylized depiction of the maize seed industries in developing countries, subsistence production is characterized by open-pollinated varieties improved through farmer selection and on-farm seed production with local seed markets governed by custom. In a fully commercial system, the predominant seed type is a hybrid that is purchased annually. Seed is a globally traded product of specialized research that is both privately and publicly funded. The exchange of seed and the genetic resources used to improve it are enabled and protected by strict forms of intellectual property rights (Morris, Rusike, and Smale, 1998).
In rice and wheat, which are self-pollinating crops, the incentives for privatization of research have not been as strong as for maize, although this depends on the institutional and economic context. In the industrialized countries, profound changes in science and in intellectual property protection over the past 20-30 years have been associated with a higher rate of investment in agriculture by the private sector than the public sector and a shift in the composition of private investment from agricultural machinery and processing into chemical research and plant breeding (Alston, Pardey, and Smith, 1998; Fuglie et al., 1996). Privatization is greatest in the maize seed industry in industrialized countries and is increasing in the less-industrialized agricultural economies, but has only occurred to a limited extent for wheat—in Europe. Almost all of the seed research for rice has been and continues to be conducted by the public sector, and most research has occurred in Asia (Pray, 1998).
Historical experience in the diffusion of modern varieties of wheat, rice, and maize lead us to predict that, for highly bred, staple food crops, the opportunity costs of landrace cultivation will be lowest, and farmers' economic incentives to grow landraces higher, in less densely populated, commercially isolated areas for which either public or private breeding systems are unlikely to develop well-adapted materials. Where conditions are otherwise, the costs of designing institutions and mechanisms to encourage conservation are likely to be quite high—unless consumers demand specialized traits that are both unique to landraces and difficult to transfer genetically into modern varieties. On the other hand, the costs of designing institutions to support on-farm conservation is likely to be considerably lower for crops that are not so highly bred and have received less research attention but play significant roles in the strategies of small-scale farmers. These crops are sometimes known as "neglected" or "underutilized" species from the perspective of professional plant breeding and modern agricultural systems.
3.3 Predicting locations where crop biodiversity will continue to be managed by farmers
An extensive microeconomics literature on variety adoption and a growing set of case studies about on-farm management of crop biodiversity provide us with empirical evidence about the types of social and economic contexts that lead to a higher probability that farmers will continue to manage biologically diverse crop genetic resources. The first factor, which is important on a local as well as a regional scale, is agroecological heterogeneity. Environmental heterogeneity has also been advanced as an explanation of farmers' continued use of landraces in Turkey (Brush and Meng, 1998). Across a series of villages with differing agroclimatic conditions, heterogeneity in agroecological conditions increased the number of different crops and the varieties of maize, beans, and squash varieties grown by farmers in the state of Puebla, Mexico (Van Dusen, 2000). In the highlands of Ethiopia, land fragmentation, soil erosions, and the numbers of plots on farms had effects on the variety diversity that varied among crops (Benin et al„ 2003).
The second factor that is important at both local and regional levels is the extent to which households trade on markets. The more removed a household is from a major market center, the higher the costs of buying and selling on the market and the more likely that the household relies primarily on its own production for subsistence. Van Dusen (2000) found that the more distant the market, the greater the number of maize, beans, and squash varieties grown by farmers. In the highlands of Ethiopia, findings regarding market distances depended on the cereal crop, how the variable was measured (distance from farm to household, distance from household to all weather road, or distance from community to market), and whether the diversity among crops or the diversity within crops was the dependent variable (Benin et al., 2003). In environmentally sensitive areas of Hungary, small-scale farmers of a more isolated site with poorer soils and fewer food markets valued ancestral varieties and higher levels of crop species richness more than farmers in an economically developed site with good road quality, fertile soils, and numerous food markets (Birol and Gyovai, 2003).
Small-scale farmers' choice to grow more than one variety simultaneously is likely to reflect their need to address numerous concerns that no single variety can satisfy (Bellon, 1996). Case studies demonstrate that in many of the regions of the developing world where landraces are still grown, either markets for commercially produced seed, markets for the crop output, or markets for the multiple attributes farmers demand from their varieties are incomplete (Brush and Meng, 1998; Brush, Taylor, and Bellon, 1992; Smale, Bellon, and Aguirre, 2001). This means that the traits demanded by farmers (grain quality, fodder, suitability for a certain soil type) cannot be obtained through the production of modern varieties or procured through impersonal market transactions, so that farmers must rely on their own or neighbors' production for their supply (Renkow and Traxler, 1994; Bellon and Taylor, 1993). Knowing the crop characteristics that most matter to farmers, how these are distributed across varieties and populations, and to what extent farmers can meet their needs through market transactions is, therefore, important. In Turkey, concern for bread quality in wheat, in addition to high household transaction costs such as transportation and uncertain prices, were associated with the choice to grow landraces rather than modern varieties (Brush and Meng, 1998).
"Promising candidates" may also be locations where both modern varieties and landraces are grown, if growing both types in that location represents an economic equilibrium. Though modern varieties have long been equated with a loss of diversity on farms (Frankel, 1970), like any new or exotic type that is introduced, a modern variety can add to the portfolio of distinct agro-morphological types grown in a community precisely because it has been bred with the ideal type of other farmer-breeders or professional breeders in mind. Modern varieties may possess a trait not found in the local varieties grown in a community (Louette, Charrier, and Berthaud, 1997), suiting particular production niches but not others (Gauchan and Smale, 2002). With cross-pollinating species, farmer seed management or deliberate introgression may mean that the introduction of modern varieties generates new types that exhibit traits from both (vom Brocke, 2001; Bellon and Risopoulos, 2001).
Farmers often choose to grow both landraces and modern varieties. Viewed in the conventional microeconomic literature as partial adoption, this observed pattern has been explained theoretically through attitudes toward risk and uncertainty, missing markets, and differential soil quality or nutrient response combined with fixity or rationing (reviewed in Meng, 1997; Smale, Just, and Leathers, 1994). Though treated as a transitional period to full adoption, the coexistence of modern varieties and landraces may represent an economic equilibrium if one or several of these aspects persist despite economic change. Meng, Taylor, and Brush (1998) concluded that multiple factors, including missing markets, yield risk, grain quality, and agroclimatic constraints, influence the probability that a Turkish household will grow a wheat landrace; a change in any single economic factor is unlikely to cause farmers to cease growing it.
Zimmerer (1996) found that the capacity of farmers to grow diverse food plants (including maize) in Peru and Bolivia depends on whether they can cultivate them in combination with commercially developed, high-yielding varieties that were easier to sell because they had more uniform grain quality. So far, introduction of modern varieties of wheat and maize has not meant that any single variety dominates or that modern varieties have replaced farmers' varieties in the Ethiopia highlands (Benin et al., 2003), most likely because they have limited adaptation and farmers face many economic constraints in this location. It is just as likely that small amounts of seed of modern varieties broadens the variety set of these farmers by meeting a particular purpose. Neither the physical terrain nor the market infrastructure network is favorable for specialized, commercial agriculture.
The relationship of household characteristics, such as wealth and income sources, to crop diversity depends on the measurement and empirical setting. Smale, Bellon, and Aguirre (2001) found that variety attributes such as suitability for food preparation (tortillas) far outweighed the importance of household characteristics in explaining the number of maize landraces grown by individual farmers and the average share of maize area planted to each. In three sites in Nepal, based on a composite variable for wealth rank, Rana et al. (2000) found that poor households cultivate more coarse-grained, drought-tolerant varieties of rice, while wealthier households grew high-quality varieties for premium market prices and special food preparations. In the state of Puebla, Mexico, Van Dusen (2000) found that the greater the wealth of the household, as measured by house construction and ownership of durable goods, the less likely the household is to plant a diversity set of maize, beans, and squash varieties. Benin et al. (2003) found that while larger farms were associated with more varieties of any major cereal crop grown in the highlands of Ethiopia, livestock assets had conflicting effects among cereals.
In many parts of the developing world, off-farm migration generates a growing proportion of the income of farm households. Brush, Taylor, and Bellon (1992) found that off-farm employment was negatively associated with maintenance of potato diversity in the Andes, indicating that the opportunity cost of cultivating many varieties—which requires laborintensive seed selection and procurement tasks—is significantly higher where other employment possibilities exist. Van Dusen found that overall diversity in the milpa system decreased as local labor markets intensify, or as more migration to the United States occurs, though these effects were not as pronounced when each crop was considered singly. Yet off-farm income can also release the cash income constraint faced by some farmers, enabling them to shift their focus from growing varieties for sale to growing the varieties they may prefer to consume. In Chiapas, Mexico, Bellon and Taylor (1993) found that off-farm employment was associated with higher levels of maize diversity. Meng (1997) found the existence of off-farm labor opportunities to have no statistically significant effect on the likelihood of growing wheat landraces in Turkey.
Was this article helpful?