Factors influencing the adoption of conservation agriculture
It might be assumed that CA is more profitable in steep-sloping, high rainfall tropical regions (e.g. Latin America) than in flatter temperate areas (e.g. Canada, the United States), since the former would be subject to a higher risk of erosion under conventional tillage. But such a generalisation would hide a number of the complexities that make the analysis of financial returns from CA difficult. For example, in 7 of the 12 recent cost studies reviewed for this study (Appendix 1), reduced or no-tillage showed higher net returns than conventional tillage, and most of these studies involved temperate regions.
One of the first comprehensive financial analyses of CA on large farms in developed countries (Crosson, 1981) compared the on-farm costs of conventional tillage with conservation tillage in the United States. More recent reviews have tended to reinforce its conclusion that CA has a small cost advantage over conventional tillage but that site-specific conditions could alter this result in various ways (Table 3). The following input cost aspects form the basis for these general conclusions.
This is the most important cost item for larger producers and so the impact of CA on these expenditure items is critical. Most analyses suggest that CA reduces machinery costs. Zero or minimum tillage means that farmers can use a smaller tractor and make fewer passes over the field. This also results in lower fuel and repair costs. However, this simple view masks some complexities in making a fair comparison. For example, farmers may see CA as a complement to rather than as a full substitute for their existing practices. If they only partially switch to CA (e.g. on some fields or in some years), then their machinery costs may rise as they must now provide for two cultivation systems, or they may simply use their existing machinery inefficiently on their CA fields.
To capture such complexity, economists distinguish between short-run and long-run costs, where the former assumes no adjustment to existing capital equipment and the latter assumes such an adjustment. A comparative study of CA and conventional tillage in Wisconsin (Mueller et al., 1985) found that short-run average costs under CA exceeded long-run average costs by about 7 percent. The short-run average costs per hectare for CA were greater than for conventional tillage. However, after adjustments to capital, CA costs fell below those of conventional tillage in the long run.
Similarly, the expectation is for fuel costs to be lower under CA, and this is generally the finding in most studies. Falling fuel prices should encourage greater adoption of CA. One study (Uri, 1998a) shows that the price of crude oil has a statistically significant but relatively minor effect on the intensity of CA (but not adoption by new farmers). It finds that a 10 percent increase in the United States in the price of crude oil is associated with an expansion in planted hectares under CA of 0.4 percent, with the expansion being concentrated primarily on existing CA farms.
Offsetting lower machinery costs are higher herbicide applications under CA, especially during the early adoption period and with no-till. Indeed, herbicides substitute for the use of machinery to keep weeds under control. Site-specific factors are important as perennial weeds can present problems for CA. Nonetheless, herbicide application rates and the ability to fully control weeds under CA in all situations remains a controversial and continuing area of CA research. Recent assessments have tended to argue that herbicide applications decline over time and may eventually fall to a level equal to that of conventional tillage (USDA, 1998). Insect control is less an issue in conventional and CA comparisons. As most pesticides are petroleum based, crude oil prices are liable to affect their cost to farmers. If so, then a higher crude price would mean higher herbicide costs, partially offsetting CA's relative cost advantage stemming from lower machinery fuel requirements (this may explain the small response found by Uri).
Much attention has focused on the apparent reduction in labour requirements under CA. This reduction follows from the decreased demand for labour for land preparation at the beginning of the growing season. Some estimates put this reduction at 50-60 percent during this time period. On large mechanized farms in the developed world the true impact of this saving is small as labour costs account for under 10 percent of total per acre costs (Table 3). However, on some farms in the developed world, the trend towards increased off-farm work has made even the relatively small labour savings under CA attractive. Indeed, some case studies have cited the time savings provided by CA as the primary motivation for the adoption of conservation tillage (Wandel and Smithers, 2000).
Most comparative analyses of the costs of conventional tillage versus conservation tillage assume that other production inputs remain unchanged following a switch to CA. A debate continues concerning fertilizer use under CA as there is evidence, that CA adoption affects nitrogen use by crops and leaching. Uri (1997) finds some increase in fertilizer use by maize farmers adopting conservation tillage in the United States. Additionally, if the application of fertilizers under CA requires greater management skill, then application costs could rise even if application rates do not. A more general finding is that CA requires greater management skills and it may be costly for farmers to acquire these. CA may also affect seed purchases as farmers may be able to avoid some pest problems by investing in more resistant seed varieties. However, this increases costs.
The comparative data in Table 3 reveal a consistent picture in recent decades concerning conservation tillage costs in the United States. More recent estimates tend to show a wide range for CA, recognizing the variation in site-specific conditions (e.g. drainage, rainfall). Perhaps more significantly, the cost items listed in Table 3 represent only a subset of total costs as other production inputs and land were assumed to remain constant under either cultivation system. Putting the cost savings attributable to CA in the context of these total costs, any cost advantage amounts to about 5-10 percent in 1979 and probably about the same in the 1990s.
Also missing from many cost comparisons of conventional and conservation tillage is an analysis of risk factors. One aspect of risk is a recognition that yields might vary under the different cultivation systems. Much debate has centred on whether switching to CA leads to higher or lower yields. As the results for temperate climates are often contradictory, and any differences are usually not statistically significant, most analysts simply assume no change in yield. Similarly, the impact of adopting CA on yield variability and risk is controversial. Some studies argue that CA increases yield variability in many situations, thereby worsening risk (Fox et al. 1991). By contrast, Australian research shows a reduced variability in crop yields with CA (Kirby et al., 1996), while work in Canada indicates that the net returns were higher under CA than conventional practices in bad years, but lower when averaged over time. Firm conclusions on whether risk is increased or reduced under CA remain elusive.
More certain are the impacts of CA on cropping intensity. With reduced field preparation time, the cropping cycle is shorter, allowing more crops in a given period and even double cropping where it was not possible previously. Where this benefit is available from CA, more efficient utilization of the fixed land resource results in higher annual net returns per hectare. Moreover, farmers may adjust their cropping strategy when switching to CA. Hence, yield trials comparing the same crop under either cultivation system may not represent reality. In fact, fully adopting CA involves switching to a suitable crop rotation that will probably differ from the conventional cropping strategy used previously. For this reason, some writers have called for a broader whole farm approach to comparative assessments in temperate agriculture (Diebel et al., 1993).
Overall, a comparison between conventional and conservation practices in temperate agro-ecological zones hinges on two offsetting effects. One involves CA's labour and possibly machinery cost savings, while the other involves higher herbicide costs, at least initially, under CA. Depending upon the magnitude of each of these effects, CA may appear either more or less costly. For example, in Saskatchewan, Canada, researchers found that the higher herbicide costs characterizing CA overwhelmed any cost savings associated with labour, fuel, machine repair and overheads (Zentner et al., 1991). Similarly, Stonehouse and Bohl (1993) used a linear programming model to argue that conservation tillage in a cash-crop farm system is not profitable. However, most developed-country studies reviewed find that CA demonstrates at least minor cost savings over conventional practices. However, these savings have not been sufficient to induce adoption by large numbers of farmers on large mechanized farms. These farmers may resist new practices unless there is a promise of much higher financial returns.
One of the success stories for CA has been in Latin America (Box 2). Large-scale mechanized farming is common in many parts of Latin America and farmers have adopted CA on large portions of this cultivated area. While most of the comparative cost analysis presented above for temperate northern regions would apply here, the advantage of CA in Latin America has been more pronounced. In part, this greater advantage reflects physical and climate factors, but also the differences in the nature of the technology adopted. While most studies in the United States document adoption of conservation tillage alone, in Latin America the technology is much closer to the concept of CA described in Chapter 1. That is, it is liable to include not just tillage adjustments but also changes in cover crops and mulching practices as well as the incorporation of crop rotations and other changes.
In Paraguay, yields under conventional tillage declined 5-15 percent over a period of ten years, while yields from zero tillage increased 5-20 percent (Sorrenson et al., 1997 and 1998). Savings in fertilizer and herbicide inputs dropped by an average of 30-50 percent over the same period. In Brazil, over a 17-year period, maize and soybean yields increased by 86 and 56 percent, respectively, while fertilizer inputs for these crops fell by 30 and 50 percent, respectively. In addition, soil erosion in Brazil fell from 3.4-8.0 t/ha under conventional tillage to 0.4 t/ha under no-till, and water loss fell from approximately 990 to 170 t/ha.
Latin America has the highest rate of adoption of no-till practices in the world. The first recorded attempt at mechanized zero tillage was in sub-tropical Brazil between 1969-1972 and in 1981/2 in tropical Brazil. The first field testing of no-till was in the state of Parana in 1972. By 1999, the percentage of the total cultivated area under no-tillage had reached 52 percent in Paraguay, 32 percent in Argentina and 21 percent in Brazil. No-tillage accounts for 95 percent of all conservation tillage in Latin America (44 percent in the United States). At first, the adoption of zero tillage in Latin America was only gradual, due to herbicide and planter limitations and the high incremental costs of adoption (Box 1). However, as farmers received support from farmer NGOs, the public sector and private interests, adoption increased significantly. For example, small, medium and large-scale farm operators in Paraguay have detailed considerable improvements in on-farm profitability and the reduction of risk. The studies also point to the crucial role of skilled personnel for training farmers in new management skills and the importance of credit availability for the purchase of new no-till machinery. By providing institutional and financial support, government has played a crucial role in creating incentives for adoption. Smallholders have been a special target as they lack the capacity to raise funds and retrain on their own. The World Bank reiterated these observations in its review of a project in Brazil promoting sustainable agriculture, modern forms of land management, and soil and water conservation. It considered rural extension to be a pivotal element in the project. In addition, monetary incentives were highly successful in motivating group formation among farmers, leading to an increase in cooperation and social capital. It recognized rapid paybacks and government financial incentives and support as key influences on adoption.
Sources: Sorrenson et al., 1997 & 1998; World Bank, 2000.
As a result, the financial benefits for farmers in Latin America who have adopted CA have been striking. However, these take time to fully materialize. Sorrenson (1997) compared the financial profitability of CA on 18 medium and large-sized farms with conventional practice in two regions of Paraguay over 10 years. He found that by the tenth year net farm income had risen on the CA farms from under US$10 000 to over US$30 000, while on conventional farms net farm income fell and even turned negative. Medium and large-scale farmers have experienced:
In Latin America and in other developing regions, CA is a technology with potential appeal for smallholders. However, adopting CA on a small, possibly non-mechanized, farm involves some different considerations when compared to a large mechanized farm. For example, as smallholders use few purchased inputs, discussions on large increases in herbicide costs may not be relevant. Even if smallholders accept the need for herbicides, they may be unable to finance their purchase. In addition, few smallholders use significant amounts of fertilizer so that a debate over the impact of CA on fertilizer use is largely irrelevant. Ultimately, the availability of credit to assist with CA's increased need for purchased inputs plays an important role. If smallholders hire land preparation equipment, then a switch to CA should be relatively simple as there are no machinery investment implications. Short-run costs would be close to long-run costs when switching to CA.
The majority of smallholders worldwide do land preparation and weeding manually, and adopting CA has its greatest impact on the labour used in these activities. In a comparative analysis of traditional bush fallow systems with no-till and alley cropping in Nigeria, labour savings under the no-till technology were substantial (Ehui et al., 1990). Whereas alley cropping required from 126 to 151 person-days/ha/year and the bush fallow system needed from 67 to 102 days, the no-till technology required 58 days (with an allowance for land clearing in each case). These labour inputs amounted to more than 50 percent of total production costs for each technology. However, higher herbicide and equipment costs penalized the no-till technology and it was only preferred under conditions of higher population pressure, which penalizes alternative fallow systems. In studies of smallholders in Latin America, net farm income and returns to labour were much higher under CA than conventional practice. Table 4 supports this observation for adopters of CA in Paraguay.
In judging the attractiveness of CA in smallholder systems in Africa, Latin America and elsewhere, labour savings are a key factor. A further point related to labour is that as the labour savings come at both the land preparation and weeding stages (assuming herbicide use), there are liable to be implications for the gender division of labour. In most smallholder systems in Africa, male household members are responsible for land preparation (with a contribution to sowing), while female household members are responsible for weeding. Herbicide use may require some adjustment in these responsibilities as male household members usually handle pesticides. Male household members may resist the additional labour demand during the weeding period, so creating a barrier to the adoption of CA.
Furthermore, certain conditions can enhance the relative financial attractiveness of CA. For example, rising land pressure tends to increase the attractiveness of CA relative to bush fallowing. An additional consideration is land quality. Studies of the net returns from mulching, an important component in smallholder CA, suggest that the benefits of this practice increase with the quality of cropland (Lamers et al., 1998). Successful instances of CA adoption in Latin America have demonstrated the importance of credit as an important enabling factor. This is because of the need to finance specialized planting equipment and herbicides.
Most financial analyses of CA concentrate on a comparison with conventional practice, whether this is conventional tillage or bush fallow. However, farmers can often select from a number of alternative conservation practices, in which case CA is just one option of perhaps several. This is especially true for smallholder systems as an absence of prior machinery investments and the small-scale adaptability of many soil and water conservation techniques makes adoption relatively easy in physical and financial terms.
To consider CA's attractiveness in relation to alternative conservation practices to a smallholder, a database of over 130 different analyses of individual soil and water conservation technologies was compiled. The analyses concentrated on Africa and Latin America with all technologies coded according to whether they constituted a CA-related technology (Group 1) or not (Group 2), as specified by the World Overview of Conservation Approaches and Technologies (WOCAT) technology classification system. Group 1 includes measures aimed primarily at enhancing soil cover and organic matter, while Group 2 technologies are generally linear, cross-slope approaches intended to reduce erosion from wind or runoff. Information about farm-level financial returns was entered in the database for each technology. The results for each of the two technology groups were sorted based on whether technology adoption provided a positive or negative net present value (NPV). Table 5 presents the results of this procedure.
The analysis presented in Table 5 is somewhat crude as many studies employ differing assumptions about project life, discount rates, land opportunity costs, etc. Moreover, the classification of technologies is not precisely consistent with the definition of CA presented earlier. Nonetheless, the results in Table 5 do indicate that CA and, more broadly, agronomic improvements tend to show higher net returns at the farm level than do other techniques (e.g. vegetative, structural and other improvements). Arguably, this relative attractiveness of CA is more pronounced than was the case from the comparison of only CA and conventional tillage. Thus, when faced with numerous alternatives to conventional practice, CA and related approaches may offer the best possible returns in many situations. Site-specific factors would determine which individual technology offered the best returns for individual farmers.
In summing up the financial evidence in support of CA, a few words of caution are in order. While it is true that CA often conforms to what Pampel and van Es (1977) term an `environmentally profitable practice' (i.e. good for environment and profitable), this is not always so. Particular location constraints might result in reduced yields, or institutional factors may favour alternative practices (Stonehouse, 1995).
Thus, it is necessary to consider site-specific conditions in determining the financial attractiveness of CA. Even where the financial incentives may appear attractive, a consideration of non-financial factors is required to understand the actual and potential adoption of CA.
A number of studies have sought to identify barriers to adoption beyond the obvious divergence between on-farm costs and wider social benefits under CA (Smit and Smithers, 1992; Pierce, 1996; Cary and Wilkenson, 1997). For example:
In part, the need to consider factors other than net returns reflects farmers' competing objectives in farm management, i.e. profitability versus low investment or minimum subsistence food requirements. Competing technologies may meet individual objectives to varying degrees. In terms of maximizing net financial returns, Tables 3-5 suggest that CA can provide better net returns than either conventional practice or other conservation technologies, subject to local site conditions. Table 6 compares various attributes of CA technologies and other soil conservation techniques at the farm level in West Africa. The qualitative analysis applies four criteria representing different smallholder objectives, of which one is financial profitability (Table 5). While consistent with the net returns analysis in Table 5, the results in Table 6 allow for a much broader evaluation, highlighting assorted shortcomings or advantages of individual technologies that may not be apparent in a financial analysis alone.
The influences other than net returns shown in Table 6 represent only a small subset of the many non-financial factors thought to influence conservation technology adoption. Table 7 lists those other factors found to influence the adoption of CA in a statistically significant sense (based on a review of statistical results contained in Appendix 2). A review of the many studies contributing to Table 7 suggests that results are often not conclusive. Conditions may be too site specific to allow much generalization based on statistical studies alone.
Farm-level factors vary from farm operation to farm operation and higher level factors are also at work, such as the transmission of information (via policy-related activities and social processes). Furthermore, the variables discussed below, and their broader categories, do not act independently, but rather interact to influence adoption.
Since Ryan and Gross (1943) first showed that the adoption of agricultural innovations is typically uneven from farmer to farmer, researchers have directed attention to certain characteristics and attributes of farmers in an effort to explain this unevenness. In the case of soil conservation technology adoption, Gould et al. (1989) emphasize awareness on the part of farm operators to soil erosion or other soil problems as an obvious prerequisite to adoption. Indeed, farmer awareness or perception of soil problems is frequently found to positively correlate with CA adoption (Stonehouse, 1991). Similarly, the central place of information and knowledge in CA adoption, in terms of being aware of soil problems and potential solutions, should lead the level of education of a farm operator to correlate positively with adoption. Education, be it specific or general, generally correlates positively with the adoption of CA practices, notwithstanding some findings of insignificance or even negative correlation (Rahm and Huffman, 1984; Marra and Ssali, 1990; Warriner and Moul, 1992).
Age and/or experience are difficult factors to link to CA adoption, given that studies have shown both a positive and negative correlation. Based on a study of conservation tillage adoption in Wisconsin, Gould et al. (1989) showed that older and more experienced farmers were more likely than their younger colleagues to recognize soil problems. However, they were less likely than their younger colleagues to address the problems once recognized. In contrast, several studies have found that income correlates positively with the adoption of soil-erosion control practices (Okoye, 1998; Wandel and Smithers, 2000).
Studies of the adoption of conservation tillage and other CA-type practices have often given significant attention to farm size (or sometimes planted area). Many studies have found that farm size correlates positively with adoption (Westra and Olson, 1997). However, other studies have shown no significant relationship (Agbamu, 1995; Uri, 1999b), or even a negative correlation (Shortle and Miranowski, 1986). Hence, the overall impact of farm size on adoption is inconclusive.
Some studies have found that the presence of soil erosion and other soil problems on the farm correlates positively with conservation tillage adoption (Stonehouse, 1991). However, farmer awareness of and concern for soil problems is probably the more critical factor affecting adoption. Another important farm characteristic is underlying land productivity. In the case of no-till and mulch tillage, Uri (1997) shows that in the United States adoption is more likely on farms with low rather than high levels of soil productivity. In addition, a good fit between CA and the farm's production goals encourages adoption.
A more complex factor liable to affect adoption is land tenure. In simple terms, privatizing land should lead to better incentives for the adoption of conservation technologies. However, studies of the privatization of land or titling have not shown that this is necessary to motivate sustainable practices and, in some instances, it has had the opposite effect. As a result, it appears that producers may accept titling because it guarantees land rights, but this does not necessarily bring about changes in their land management. In contrast, there are numerous studies indicating that traditional institutions governing access to land resources in developing regions are flexible in responding to internal and external pressures. Table 8 summarizes the empirical evidence provided by a number of African studies addressing both private title and customary tenure. It shows that the former institutional arrangement does not bestow any advantage over the latter, in terms of investment incentives. Thus, general claims that titling will lead to increased investment in land improvements should be viewed with caution.
Without knowledge of the practices associated with CA via some information or communication channel, adoption is improbable. Indeed, studies of innovation adoption and diffusion have long recognized information as a key variable, and its availability is typically found to correlate with adoption (de Harrera and Sain, 1999). Information becomes especially important as the degree of complexity of the conservation technology increases (Nowak, 1987).
Information sources that positively influence the adoption of CA-type practices can include: other farmers; media; meetings; and extension officers. However, with respect to this latter source, Agbamu (1995) shows that contact alone will not promote adoption if information dissemination is ineffective, inaccurate or inappropriate. Studies have not always shown that the ease of obtaining information correlates with adoption.
In technical terms, the characteristics and availability of CA technologies are crucial factors in adoption. However, de Harrera and Sain (1999) note that availability does not imply individual ownership of the necessary machinery as lease/hire arrangements proliferate. Furthermore, potential adopters must believe that the technology will work. Technical factors interact with biophysical factors, e.g. soil type, rainfall or topography can encourage/facilitate or discourage/limit CA adoption. While some studies have shown that farm operations located within regions of steep slopes and erodible soils have a greater tendency to use CA practices, other studies have found these variables to be insignificant.
CA adoption is seldom strictly a function of individual profit maximization alone, but also can reflect non-individual or societal interests. More specifically, Lynne (1995) argues that farmer decision making usually reflects a compromise between private economic utility and collective utility. Producers often identify this latter interest as `the right thing to do', at least in those places where stewardship is part of the cultural norm. The argument runs that for many producers the pride associated with stewardship makes up for limits in financial rewards (Campbell et al., 1999). Examples of such stewardship motives governing land management arrangements include the Landcare movement in Australia (Sobels et al., 2001). In contrast, Van Kooten et al. (1990) modelled the trade-offs between stewardship and net returns on wheat-fallow farms in Saskatchewan, Canada. Their study found that farmers make improvements in agronomic practices to benefit soil quality only under extreme degrees of concern (e.g. stewardship). This result holds despite such practices representing no more than a 5 percent sacrifice in net returns.
In addition to stewardship motives, collective action may be necessary to implement CA on a regional basis. Cooperative arrangements govern numerous activities within village agricultural systems. Although the discussion usually focuses on common property resources, even private land use may overlay with cooperative arrangements governing various aspects of farm management (Pretty, 1995). For example, contour ploughing, stone lines and other structural works require cooperation amongst several or many farmers in order to be effective conservation strategies. Many dimensions of CA fit the cooperative model, including the formation and operation of farmers' groups, dissemination of information, pest control and the purchase of agrochemical inputs. Box 3 provides a more general discussion of collective action in relation to sustainable agriculture.
If CA requires collective action or high levels of social organization to help it gather momentum, then widespread adoption may be related to a society's social capital. The role of social capital in fostering or retarding the collective action needed in promoting new conservation technology is of growing interest (Box 3). In the broadest sense, social capital refers to the interconnectedness among individuals in society and considers relationships as a type of asset. Several studies have examined the influence of social capital on technology adoption in either developed or developing countries. For example, kinship, or more exactly `connectedness to others', can influence the adoption of conservation technology. Some studies have shown that the expectation of farmland inheritance can have a bearing on conservation behaviour amongst farmers, although other studies testing for this have not shown a positive correlation. Similarly, higher levels of social capital help explain the adoption of fertilizer and soil conservation practices in Peru (Isham, 2000; Swinton, 2000), while one study has related the success of peasant committees in Paraguayan villages to the level of social capital in these communities (Molinas, 1998). Such institutions at the local level have been an important catalyst in the adoption and diffusion of CA.
In conclusion, the inconsistent and sometimes contradictory results obtained from studies of the adoption of CA-type practices tend to suggest that the decision-making process is highly variable, and that outcomes may be specific to particular people, places and situations. This makes the task of developing a policy framework to promote CA adoption particularly challenging.
Collective action can have benefits over individual decision making when the tasks at hand require coordinated group activity (e.g. various agricultural and conservation practices). For example, it may reduce the costs of repeated transactions amongst many individuals by establishing a single set of rules and avoiding individualized negotiation and transaction. However, collective action is not automatic in the diffusion of improved technologies such as CA, especially where information is lacking or the underlying physical processes of land degradation are slow and barely perceptible. Additionally, some individuals may benefit from collective action without contributing, and this may result in a lack of collective incentives. Using game theory to model behaviour in collective action situations, researchers have tried to understand what factors may foster collective behaviour. For example, if repetition and observability characterize group activities, the result may well be cooperation, but only if:
In general, the key variables influencing the potential success of collective action are: the number of decision-makers, especially the minimum number required to attain a collective benefit; discount rates, which influence the magnitude of future benefits from collective action; a similarity of interests among agents; and the presence of some individuals with leadership or other assets. In part, the behaviour needed to foster collective or socially responsible actions may hinge on the level of social capital in a community. The World Bank (1998) reviewed various definitions of this term and found they ranged from a fairly narrow view relating to the interconnectedness among individuals, via associations, societies, etc., to a much broader view encompassing the entire social and political environment. In simple terms, if conservation activity requires cooperation, then the degree of interconnectedness and the enabling social environment may be a critical determinant. The various indicators of a community's or nation's level of social capital include the number and type of associations, homogeneity within communities, levels of trust in others, reliance on networks of support, presence of natural leaders, etc.