Risk,learning, and technology adoption

This article explores how decision makers learn and use information, with an application to the adoption of biotechnology in agriculture. The empirical analysis relies on experimental and survey data measuring risk preferences, learning processes, and the adoption of genetically modified (GM) seeds among U.S. grain farmers. While controlling for risk aversion, we link individual learning rules with the cognitive abilities of each decision maker and their actual GM adoption decisions. We find evidence that very few individuals are Bayesian learners, and that the population of farmers is quite heterogeneous in terms of learning rules. This suggests that Bayesian learning (as commonly assumed in the analysis of agricultural technology adoption) is not an appropriate characterization. In addition, we do not find a strong relationship between observed learning styles and the timing of GM seed adoption. To the extent that learning is a key part of the process of technology adoption, this suggests the presence of much unobserved heterogeneity in learning among farmers.     

A dynamic adoption model with Bayesian learning: an application to U.S. soybean farmers

Adoption of agricultural technology is often sequential, with farmers first adopting a new technology on part of their lands and then adjusting their use of the new technology in later years based on what was learned from the initial partial adoption. Our article explains this experimental behavior by using a dynamic adoption model with Bayesian learning, in which forward‐looking farmers take account of future impacts of their learning from both their own and their neighbors’ experiences with the new technology. We apply the analysis to a panel of U.S. soybean farmers surveyed from 2000 to 2004 to examine their adoption of the genetically modified (GM) seed technology. We compare the results of the forward‐looking model to that of a myopic model, in which farmers maximize current benefits only. Results suggest that the forward‐looking model fits data better than the myopic model does. And potential estimation biases arise when fitting a myopic model to forward‐looking decision makers.     

Early adoption,experience, and farm performance of GM corn seeds

The current article explores the characteristics that distinguish early from late adopters of GM corn and measures the productivity impacts of early adoption, for a sample of farmers in Minnesota and Wisconsin. The results of the adoption analysis confirm that size, education as well as specialization are positively correlated with early adoption. In addition, these results also show that farms that are mostly worked by family labor but hire some off‐farm labor are more likely to adopt GM seeds earlier in the diffusion process. The productivity analysis demonstrates the superiority of stacked varieties. At the same time, we find no evidence of a direct impact of experience on yields. Given the previously documented impact of early adoption on the use of stacked varieties, we conclude that experience plays a role through the adoption of these new technologies but does not play a role in allowing the producer to use the technology more efficiently, once it has been adopted.     

The adoption of intensive monocrop horticulture in southern Cameroon

Results from a 1997 survey of 208 households in the humid forest zone of southern Cameroon indicate that African policy makers seeking to intensify agricultural production should focus attention on the horticultural sub‐sector. The survey, which gathered information on horticultural production practices, found that the average expenditure on agro‐chemical inputs by horticultural producers using monocrop production systems was 190US$/ha, which greatly exceeds the FAO reported national average expenditure of 6.50 US$/ha. A logit model of monocrop adoption indicated that the size of land holding per household had a negative effect on adoption, congruent with population‐driven technical change and that increases in unit transportation costs significantly decreased the probability of adoption. These findings suggest that policy makers should target horticultural intensification in areas of higher population density and promote investment in rural roads. The age of the household head had a significant negative and elastic effect on adoption, which in combination with an increase in the cohort of younger farmers in the rural population induced by macro‐economic events contributed to the spread of intensified horticulture. In the study area, roughly two‐thirds of rural households also produce cocoa and the quantity of cocoa produced was positively associated with adoption of intensive horticultural systems suggesting that export crop promotion indirectly facilitated diversification of agriculture. Women's participation in intensive monocrop production was limited and efforts to promote their greater involvement are recommended.     

The impact of social networks on hybrid seed adoption in India

This article adds to the literature about the impact of social networks on the adoption of modern seed technologies among smallholder farmers in developing countries. The analysis centers on the adoption of hybrid wheat and hybrid pearl millet in India. In the local context, both crops are cultivated mainly on a subsistence basis, and they provide examples of hybrid technologies at very different diffusion stages: while hybrid wheat was commercialized in India only in 2001, hybrid pearl millet was launched in 1965. The analysis is based on surveys of wheat and millet farmers in the state of Maharashtra. Comprehensive data on farmer characteristics and social interactions allow for identifying individual networks, thereby improving upon previous research approaches that employed village-level variables as proxies for network effects. Using econometric models, we find that individual social networks play an important role for technology adoption decisions. While village-level variables may be used as suitable proxies at later diffusion stages, they tend to underestimate the role of individual networks during early phases of adoption.     

Agricultural diversification and integrated pest management in Bangladesh

We study factors associated with a shift toward diversified, high‐valued vegetable crops and the incentives associated with the use of IPM methods for vegetable producers in Bangladesh. The primary objective is to measure how IPM technologies affect the crop and technology choices of low‐income rice farmers. A three‐season household optimisation model is used to study crop and technology choice under price and yield uncertainty. The model is parameterised using data from vegetable farms and experimental IPM trials conducted in Bangladesh. Simulation results show that access to IPM technology and IPM availability combined with access to credit increase household welfare and lead to higher rates of vegetable adoption. Off‐farm employment opportunities work against vegetable cultivation and IPM use by risk‐averse farmers. Implications for policy and extension efforts are highlighted.     

International research on vegetable improvement in East and Southern Africa: adoption,impact, and returns

There is a lack of evidence for impact at scale of vegetable research and development, although the importance of vegetables for human nutrition and smallholder incomes is generally understood. We therefore study adoption and impact of improved tomato and African eggplant varieties developed through international agricultural research, released by national agricultural research and extension systems, and supplied to farmers by private seed companies in East and Southern Africa from 1990 to 2014. The study finds that in 2014, varieties developed by the World Vegetable Center accounted for 50% of tomato and 98% of African eggplant commercial seed production in East and Southern Africa. For Tanzania alone, investment in crop improvement generated economic gains of US$ 255 million for tomato and US$ 5 million for African eggplant up to 2014. The internal rate of return is 26% for tomato and 12% for African eggplant, though we project the latter to increase to 26% by 2024 as the variety was released only in 2007. These findings support the view that agricultural policy and investment reoriented towards contemporary nutritional challenges will give high returns to investment.     

Tradeoffs and complementarities in the adoption of improved seeds,fertilizer, and natural resource management technologies in Kenya

There is widespread consensus that agricultural technology has an important role to play for poverty reduction and sustainable development. There is no consensus, however, about the types of technologies that are best suited for smallholder farmers in Africa. While some consider natural resource management (NRM) technologies as most appropriate, others propagate input intensification with a stronger role of the private sector. In the public debate, these two strategies are often perceived as incompatible. Environmental non‐governmental organizations in particular consider low‐external input strategies as the only sustainable form of agriculture, a view that has considerable influence on policymakers and the international donor community. Most existing research studies on smallholder innovation focus on the adoption of individual technologies, so that comparisons between different types of technologies in the same context are not easily possible. We use representative data from maize‐producing households in Kenya and a multivariate probit model to analyze the adoption of different types of technologies simultaneously. Results indicate that NRM technologies and strategies that build on external inputs are not incompatible. Interesting complementarities exist, which are not yet sufficiently exploited because many organizations promote either one type of technology or the other, but rarely a combination of both.     

Analyzing technology adoption using microstudies: limitations, challenges, and opportunities for improvement

Policy makers and interest groups have many questions about the use of improved technologies in developing country agriculture. These include the roles of policies, institutions, and infrastructure in the adoption of improved technologies and their impact on productivity and welfare. Most micro‐level adoption studies, however, cannot address these important policy issues. Drawing on an extensive review of the literature on the adoption of agricultural technologies, this article suggests alternative approaches for designing technology adoption studies to make them useful for policy makers. It explores the generic limitations of cross‐sectional adoption studies carried out in small numbers of communities and discusses some problems faced in conducting such studies. Recommendations include the use of sampling approaches that allow data from microstudies to be generalized to higher levels of aggregation, adherence to clearly defined terms that are standardized across studies, and careful examination of the assumptions that often underlie such studies.     

The economic impact of cowpea research in Senegal: an ex-post analysis with disadoption

This study evaluates the impact of a long‐term research collaboration between the Senegalese Institute for Agriculture Research (ISRA), and the Bean/Cowpea Collaborative Research Support Program (CRSP) funded by the U.S. Agency for International Development (USAID). Adoption of the primary outputs from this collaboration, a nonchemical cowpea storage method and two short‐season cowpea varieties, were tracked through a series of farmer interviews in 1996, 1997, and 2004 in the main cowpea growing area of Senegal. Combined with information from an impact assessment in the late 1980s, this study draws on ex‐post adoption data from most of the product life cycle for the storage technology. The 2004 survey found that 60.4% of cowpea production in the main cowpea growing area of Senegal is stored in metal drums and that 69% of households use this storage technology; ISRA/CRSP improved short season varieties were determined to account for 3.6% of cowpea production. Baseline economic analysis of the most recent survey data, which includes benefits and costs of the storage technology and improved varieties, reveal an IRR to donor funding of 13%. The benefits of these technologies appear to be evenly distributed in the rural population on the basis of gender and ethnic background. Comparison of the rates of technology adoption over time, however, reveals that use of the storage technology may be decreasing; inclusion of this technology disadoption trend reduces the IRR by only 0.4% from the baseline estimate. Overall, estimated returns reveal that this collaboration has been a good investment for the U.S. and Senegalese governments, and other donors.     

Modeling the length and shape of the R&D lag: an application to UK agricultural productivity

This article updates total factor productivity (TFP) growth in UK agriculture from 1953–2005 and shows that public and private research and returns to scale explain TFP. Cointegration and causality tests are used to investigate the validity of attempts to explain UK agricultural productivity with R&D and related technology variables. Then, the length and shape of the lag structures are modeled and compared with the structures that are commonly imposed on the data. The rates of return (ROR) to R&D using the data determined lags differ considerably from those obtained by imposing lag shapes. These comparisons show that the ROR to public R&D are sensitive to the lag shape as well as its length and that the omission of other technology variables, such as mechanical and chemical patents pertaining to agriculture and farm size can bias the ROR.     

Social learning and technology adoption: the case of coffee pruning in Peru

A unique natural experiment involving a coffee pruning technology is used to study social learning. The yield effects of pruning take two years to appear, a characteristic that aids in identifying social learning apart from correlated unobservable variables that are a concern in the social learning and technology adoption literature. Panel data are employed that start with a private initiative which introduced systematic pruning in central Peru and that contain the population of participating growers. Results show a jump of at least 0.15 in the probability of adoption two years after the first pruning in a grower's group.     

Diffusion and spillover of new technology: a heterogeneous-agent model for cassava in West Africa

Understanding what determines the geographic spread of innovations can help guide the funding and implementation of research and extension programs. Our approach uses household survey data as model parameters, to simulate behavior across the entire surveyed population and avoid the aggregation bias associated with representative‐farm models. Such a “heterogeneous agent” approach allows us to infer the distribution of a technology's impacts across one set of households, and predict the potential for spreading to another set that shares similar characteristics with respect to natural resource endowments and farming systems. We apply the technique to new cassava varieties in West Africa, finding a strongly poverty‐alleviating impact, with substantial spillover potential from Nigeria to neighboring countries.     

Taking a new look at empirical models of adoption: average treatment effect estimation of adoption rates and their determinants

This article shows that the observed sample adoption rate does not consistently estimate the population adoption rate even if the sample is random. It is proved that instead the sample adoption rate is a consistent estimate of the population joint exposure and adoption rate, which does not inform about adoption per se. Likewise, it is shown that a model of adoption with observed adoption outcome as a dependent variable and where exposure to the technology is not observed and controlled for cannot yield consistent estimates of the determinants of adoption. The article uses the counterfactual outcomes framework to show that the true population adoption rate corresponds to what is defined in the modern policy evaluation literature as the average treatment effect (ATE), which measures the effect or impact of a "treatment" on a person randomly selected in the population. In the adoption context, a "treatment" corresponds to exposure to the technology. The article uses the ATE estimation framework to derive consistent nonparametric and parametric estimators of population adoption rates and their determinants and applies the results to consistently estimate the population adoption rates and determinants of the NERICA (New Rice for Africa) rice varieties in Côte d'Ivoire. The ATE methodological approach developed in the article has significant policy implications with respect to judging the intrinsic merit of a new technology in terms of its potential demand by the target population independently of issues related to its accessibility and in terms of the decision to invest or not in its wide-scale dissemination.     

The impact of safety nets on technology adoption: a difference‐in‐differences analysis

This article contributes to a growing body of empirical literature relating credit constraints and incomplete insurance to investment decisions. We use panel data from rural Ethiopia to investigate whether participation in a safety net program enhances fertilizer adoption. Using a difference‐in‐differences estimator and inverse propensity score weighting, we find that participation in Ethiopia's food‐for‐work (FFW) program increased fertilizer adoption in the short run, but not in the long run. Results also indicate that the intensity of fertilizer usage increased with livestock holdings for FFW‐participant households, providing some evidence that the intervention helped asset‐rich farm households more than asset‐poor households. We find no significant effects of free distribution on fertilizer adoption or intensification. Our results are consistent with the hypothesis that safety nets can be viewed as mechanisms that allow households to take on more risk to pursue higher profits. The results highlight the importance of safety net programs, their effectiveness in ensuring farmers that they will be protected against uninsured shocks, and how that assurance can translate into productivity‐enhancing behavior.     

Knowledge of vitamin A deficiency and crop adoption: Evidence from a field experiment in Mozambique

Vitamin A deficiency is a widespread public health problem in Sub-Saharan Africa. This paper analyzes the impact of an intervention fighting vitamin A deficiency through the promotion of production and consumption of orange-fleshed sweet potato (OFSP). We conducted a randomized evaluation of OFSP-related training to female farmers in Mozambique, who were also the primary caretakers of preschool children. The treatment consisted of group and individual-level training where basic knowledge about nutrition was taught, and planting and cooking skills related specifically to OFSP were developed. We find considerable increases in nutrition-related knowledge, as well as knowledge about cooking and planting OFSP, which persist after more than a year. We also observe clear evidence of adoption of OFSP for production in the short- and medium-run, which spreads through social networks.     

Estimating returns to soil conservation adoption in the northern Ethiopian highlands

Land degradation in the form of soil erosion and nutrient depletion presents a threat to food security and sustainability of agricultural production in many developing countries. Governments and development agencies have invested substantial resources to promote soil conservation practices as part of an effort to improve environmental conditions and reduce poverty. However, limited rigorous empirical work has been done on the economics of soil conservation technology adoption. This article investigates the impact of stone bunds on value of crop production per hectare in low and high rainfall areas of the Ethiopian highlands using cross-sectional data from more than 900 households, with multiple plots per household. We have used modified random effects models, stochastic dominance analysis (SDA), and matching methods to ensure robustness. The parametric regression and SDA estimates are based on matched observations obtained from the nearest neighbor matching using propensity score estimates. This is important, because conventional regression and SDA estimates are obtained without ensuring that there actually exist comparable conserved and nonconserved plots on the distribution of covariates. We use matching methods, random effects, and Mundlak's approach to control selection and endogeneity bias that may arise due to correlation of unobserved heterogeneity and observed explanatory variables.
We find that the three methods tell a consistent story. Plots with stone bunds are more productive than those without such technologies in semi-arid areas but not in higher rainfall areas, apparently because the moisture conserving benefits of this technology are more beneficial in drier areas. This implies that the performance of stone bunds varies by agro-ecology type, suggesting the need for designing and implementing appropriate site-specific technologies.     

Why don't producers adopt best management practices? An analysis of the beef cattle industry

Cattle producers were surveyed to determine their adoption rates and reasons for non‐adoption of 16 best management practices. Potential reasons for non‐adoption included unfamiliarity, non‐applicability, high cost, still considering adoption, and preference not to adopt. The two most commonly cited reasons for non‐adoption were unfamiliarity and non‐applicability of the practice. Results highlight the importance of educational efforts in encouraging adoption, as well as farm type and financial situation of the farmer.     

Estimating overall returns to international agricultural research in Africa through benefit-cost analysis: a "best-evidence" approach

This study offers a "best evidence" approach to summarizing recent benefit-cost analyses of international agricultural research in Africa. First, from an extensive literature review and the resulting global inventory of impact studies, 23 studies are identified that calculate aggregate rates of return for Consultative Group on International Agricultural Research (CGIAR) and partner investments in Sub-Saharan Africa (SSA). These studies are then appraised against a review framework consisting of principles, criteria, and indicators for study rigor. Subsequently, the economic benefits reported by studies grouped on the basis of its analytical rigor are aggregated and set against total investment by the CGIAR and national agricultural research systems to determine if the total investment to date can be justified by documented benefits under a range of assumptions. As a result, the study finds that aggregate investment is justified under a fairly wide range of suppositions. Under all scenarios, the vast majority of documented benefits stem from a relatively limited array of activities with a majority of benefits stemming from biological control (80%). Close to 20% of total benefits result from crop genetic improvement, and less than 1% result from all other activities. The implications of these results for research investment strategies in SSA and impact assessment are discussed.     

The economic and poverty impacts of maize research in West and Central Africa

This article assembles the results of three multicountry surveys on variety performance and adoption patterns to measure the impacts of maize research in West and Central Africa from 1981 to 2005, and uses cost data since 1971 to compute social rates of return on public investments in maize research in the region. Adoption of modern varieties increased from less than 5% of the maize area in the 1970s to about 60% in 2005, yielding an aggregate rate of return on research and development (R&D) investment of 43%. The estimated number of people moved out of poverty through adoption of new maize varieties rose gradually in the 1980s to more than one million people per year since the mid 1990s. Over half of these impacts can be attributed to international maize research at IITA and CIMMYT. The article concludes with a discussion of strategic options to enhance the impacts of maize research in the region.