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.     

Modelling the adoption of organic horticultural technology in the UK using Duration Analysis

Duration Analysis, which allows the timing of an event to be explored in a dynamic framework, is used to model the adoption of organic horticultural technology in the UK. The influence of a range of economic and non-economic determinants is explored using discrete time models. The empirical results highlight the importance of gender, attitudes to the environment and information networks, as well as systematic effects that influence the adoption decision over the lifetime of the producer and over the survey period.     

The Great Chinese Famine (1959–1961) and farm households’ adoption of technology: evidence from China*

The diffusion of new technology is an important driver of agricultural development, especially in the developing world. In this research, we follow the persistence of major historical events, employing a difference-in-differences method to carefully examine the long-term effect of China’s 1959–1961 famine on farm households’ current decisions to adopt technology. Further, we combine a mediating regression procedure with a bootstrap method to explore the mechanism of impact in this relationship. Overall, this study provides strong empirical evidence that the Great Famine attenuated technology adoption; moreover, a 1% increase in exposure to famine in childhood and adolescence resulted in a 0.137% decrease in the probability of technology adoption when controlling for village dummies. An analysis of mediating effects reveals that risk preferences account for the channel of famine persistence.     

Welfare impacts of maize–pigeonpea intensification in Tanzania

This article examines the driving forces behind farmers’ decisions to adopt improved pigeonpea and maize and estimates the causal impact of technology adoption on household welfare using data obtained from a random cross‐section sample of 613 small‐scale farmers in Tanzania. We use seemingly unrelated and recursive bivariate probit regressions to test the endogeneity and joint decision making of pigeonpea–maize production. A double hurdle model is used to analyze the determinants of the intensity of technology adoption conditional on overcoming seed access constraints. To address the impact of adoption on welfare, the article employs both propensity score matching and switching regression techniques. Results from bivariate probit models show that unobservable factors cause both decisions to be correlated but the finding does not support the conjecture that both decisions are made jointly. Overall the analysis of the determinants of adoption identifies inadequate local supply of seed, access to information, human capital, and access to private productive asset as key constraints for pigeonpea technology adoption. The causal impact estimation from both the propensity score matching and switching regression suggests that maize/pigeonpea adoption has a positive and significant impact on income and consumption expenditure among sample households.     

The impact of strengthening agricultural extension services on women farmers: Evidence from Ethiopia

This paper evaluates the effect of the Rural Capacity Building Project (RCBP), which aimed at promoting growth by strengthening the agricultural service systems in Ethiopia, and by making them more responsive to smallholders’ needs, in particular women farmers. We examine the gender‐differentiated impact of the RCBP using panel data on 1,485 geographically dispersed households in project and control kebeles. We find that women farmers’ access and satisfaction with extension services increased significantly immediately after the start of the project, but that effect did not last into the medium term. The project led to an increase in the adoption of high‐value crop farming, area of land cultivated, and economic participation of household members, benefiting male‐ and female‐headed households equally. Results point to the positive impact of incorporating women's needs and constraints in the design of the agricultural extension system. However, the project was not able to reduce the preexisting gender gap in agricultural outcomes.     

Nontraditional crops,traditional constraints: The long‐term welfare impacts of export crop adoption among Guatemalan smallholders

This article documents the long‐term welfare effects of household nontraditional agricultural export (NTX) adoption. We use a panel dataset that spans the period 1985–2005, and employ difference‐in‐differences estimation to investigate the long‐term impact of nontraditional agricultural export adoption on changes in household consumption status and asset position in the Central Highlands of Guatemala. Given the heterogeneity in adoption patterns, the analysis differentiates the impact estimates based on a classification of households that takes into account the timing and duration of nontraditional agricultural export adoption. The results show that while, on average, welfare levels have improved for all households irrespective of adoption status and duration, the extent of improvement has varied across groups. Long‐term adopters exhibit the smallest increase in the lapse of two decades, in spite of some early gains. Conversely, early adopters who withdrew from nontraditional agricultural export production after reaping the benefits of the boom period of the 1980s are found to have fared better and shown greater improvements in durable asset position and housing conditions than any other category.     

资本禀赋、生态认知与农户水土保持技术采用行为研究——基于生态补偿政策的调节效应

本文基于黄土高原地区陕西、甘肃和宁夏3省份1152户农户调查数据,构建资本禀赋、生态认知、生态补偿政策和农户水土保持技术采用行为的理论模型,运用Heckman样本选择模型进行实证分析,揭示资本禀赋、生态认知、生态补偿政策对农户水土保持技术采用行为的影响机理。研究结果表明:(1)资本禀赋中,农用机械数量、林地面积对农户水土保持技术采用及采用程度都具有显著的正向影响,耕地面积对采用程度有显著的正向作用,总收入对采用行为有显著的负向作用,相互信任对采用行为有显著的正向作用;(2)农户生态认知对其水土保持技术采用行为具有显著的正向作用;(3)生态补偿政策对农户采用水土保持技术具有显著的正向作用,并且对资本禀赋、生态认知与水土保持技术采用及采用程度关系中具有正向调节效应。     

The impact of water price uncertainty on the adoption of precision irrigation systems

We examine whether a firm is more or less likely to adopt precision technology when input prices are stochastic. The results are important to determining whether programs and contracts that reduce input price uncertainty may deter the adoption of conservation practices. An economic model of the technology adoption decision shows that the net effect of input price risk is ambiguous and depends on several factors including the shutdown effect, the mean price effect, the precision expansion effect, and the risk aversion effect. An empirical implementation of the model relies on data on water price and irrigation technology adoption observed in a California irrigation district over the period 1999–2002. The results show that a stable input price increases the adoption of precision technology, but the impact depends on crop choice and land quality characteristics.     

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.     

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.     

Risk, uncertainty, and learning in adoption of a crop innovation

Risk and uncertainty have often been suggested as causes of poor adoption of rural innovations, but empirical evidence has been scarce. This study focuses on a new crop‐type, chickpeas, in Western Australia to gather such evidence. The empirical models developed are based on a theoretical framework that conceptualizes adoption as a dynamic decision process involving information acquisition and learning‐by‐doing by growers who vary in their managerial abilities, risk preferences, and their perceptions of the profitability and riskiness of the innovation. Learning encompasses improvements in skill as well as reductions in uncertainty. An annual face‐to‐face survey of over 100 farmers was conducted over 3 years, eliciting the farmers' risk attitudes and their subjective distributions of yields and prices. Two limited dependent variable models, Tobit and Probit, are used to estimate the empirical model. There is a high degree of goodness‐of‐fit for both models. The study provides strong empirical support for the primarily economic character of the adoption decision, and highlight the importance of economic risk in the process. The two risk‐related factors with greatest impact on the adoption decision were risk aversion and relative riskiness of the innovation. Risk aversion tended to reduce adoption, and to do so to a greater extent as relative riskiness and scale increased. Results also reveal the key role that trialing of the innovation plays in adoption.     

Impact of conservation agriculture technology on household welfare in Zambia

This article examines the determinants and impact of conservation agriculture (CA) technology adoption on farm household welfare in Zambia. To account for selection bias from both observable and unobservable factors, an endogenous switching regression model is employed to estimate the impact of the technology on continuous outcomes like farm output, throughput accounting ratio (TAR), poverty gap, and severity of poverty. A recursive bivariate probit model is however used for the estimation of impact of adoption on a binary outcome like poverty headcount. The empirical findings demonstrate that the adoption of CA technology increases maize output, and farm TAR and reduces household poverty. Moreover, the results reveal that farmers’ years of schooling, social networks, access to credit, extension services, and machinery as well as soil quality positively influence adoption of CA technology.     

Neighbors follow early adopters under stress: panel data analysis of submergence‐tolerant rice in northern Bangladesh

The benefits of a stress‐tolerant crop become visible under the stresses that the crop is tolerant against. We investigate the adoption of submergence‐tolerant rice in northern Bangladesh by using panel data of 461 households, interviewed in 2013 and 2015. The sample households were randomly selected in 31 villages where submergence‐tolerant rice seeds were given to a small number of farmers in each village. The results from the household fixed‐effects model indicate that submergence in the previous year increased the adoption of submergence‐tolerant rice. The adoption impact was larger among farmers who were neighbors of early seed recipients. Our results suggest that being neighbors of early seed recipients may have helped the neighboring farmers observe the benefits of the submergence‐tolerant rice under submergence. The results indicate the importance of the information flow from early recipients to neighboring farmers.     

Vulnerability,shocks and persistence of poverty: estimates for semi‐arid rural South India

This paper focuses on the vulnerability of rural households to poverty when a negative crop shock occurs. The analysis is based on the ICRISAT panel survey of households in a semi‐arid region in south India during 1975–84. Using a dynamic panel data model that takes into account effects of crop shocks, an assessment of vulnerability of different groups of households is carried out. What is somewhat surprising is that even sections of relatively affluent households are highly vulnerable to long spells of poverty when severe crop shocks occur. As such crop shocks are frequent in a harsh production environment, there must be a shift of emphasis in anti‐poverty measures from meeting income shortfalls among the poor to enabling the vulnerable to protect themselves better against these shocks.     

Economic impacts of smallholder farmers’ adoption of drought-tolerant maize varieties

Drought and high temperatures are major threats to sustainable food production and consequently the livelihoods of the majority of Africans who depend on fragile agricultural systems. As a response to these threats, climate-smart agricultural technologies, such as drought-tolerant maize (DTM) varieties, have been developed and promoted on the continent. It is well-known that the adoption of improved technologies generally impacts positively on the wellbeing of adopters. Nevertheless, the magnitude of the impact of any technology or intervention is always an empirical question. Therefore, this study sought to determine the factors that influence the adoption of DTM and subsequently estimate how yield, commercialization intensity, and farm income are affected by adoption. To establish causation, we relied on observations from 200 farm households in the Northern Region of Ghana and estimated an instrumental variable regression. Consistent with findings reported in the literature, we found that DTM adoption is primarily driven by access to seed, extension service, labor availability, and location of farm households. In addition, we found that DTM adoption positively impacts on yield and commercialization intensity. The magnitude of impact is not trivial. For example, the yield of farm households increased by more than 150% (936 kg/ha) following DTM adoption. These results imply that policy-makers and development practitioners must support research and promotion of climate-smart agriculture to improve adoption and welfare indicators, such as yield and commercialization.     

Household welfare implications of sustainable land management practices among smallholder farmers in Ghana

This study employs farm household data to examine the impact of adoption of sustainable land management practices on consumption and poverty outcomes using multivalued treatment effects approaches. The generalized propensity score approach was employed to account for selectivity bias due to observed characteristics among households. Using a doubly robust treatment effects estimator, we found that increasing intensity of adoption improved per capita consumption expenditure, reduced poverty headcount and poverty gap among farm households. The results of the multivalued treatment effects approach revealed significant differences between adopters of different adoption intensities, a finding that is not possible with binary treatment effect approaches. The findings also revealed that the average treatment effect of moving from low intensity to high intensity adoption levels differed across quantiles of per capita consumption. We also use a dose-response function to demonstrate that the treatment effect of intensity of adoption on per capita consumption and poverty outcomes is nonlinear, with optimal adoption level occurring between 60–70 % of adoption intensity dose. It is therefore recommended that initial government support in sustainable land management is necessary to sustain and upscale individual household efforts towards investment in sustainable land management.     

Ecosystem impacts of pesticide reductions through Bt cotton adoption

This article examines the ecosystem impacts of transgenic Bt cotton technology resulting from reduced chemical pesticide use. Employing unique panel data from smallholder farmers in India, negative environmental and health effects of pesticide use are quantified with the environmental impact quotient (EIQ), with and without Bt technology. An environmentally sensitive dynamic production function is estimated, treating the environmental risk of pesticide toxicity as an undesirable output in the production process. Negative externalities are significantly lower in Bt than in conventional cotton. The reduction in EIQ through Bt technology adoption has increased from 39 per cent during 2002–2004 to 68 per cent during 2006–2008. Bt adoption has also contributed to higher environmental efficiency. Environmental efficiency is influenced by the quality of Bt technology: high‐quality Bt seeds are associated with higher environmental efficiency than lower‐quality seeds.     

Gender differentials in farming efficiency and profits: The case of rice production in the Philippines

Women make essential contributions to the agricultural and rural economies in all developing countries. Rural women are involved in a variety of production and farm management activities. In the Philippines women engage more intensively in agricultural work than men. However, Filipino women’s actual contribution to food production and the rural economy remains undervalued, if not invisible. Using the average treatment effect and farm-level data from the Philippines, this study investigates the effect of gender on farming efficiency, profits, and costs of rice production. Results indicate that female-headed farm households, despite having limited access to land, have higher values of rice production than their male counterparts. However, female-headed households have higher fixed, seed and labor input costs, consequently earning lower profits. In addition, female-headed farm households have lower irrigation costs. Findings from this study also indicate that women are less efficient in farming, but are more likely to adopt improved seed varieties.     

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.     

Estimating Input Complementarities with Unobserved Heterogeneity: Evidence from Ethiopia

The level of technology adoption is often characterised as low in Africa. Recent evidence, however, points to the coexistence of substantial heterogeneity across farm households and the lack of a suitable mix of inputs for farmers to take advantage of input complementarities. We use a random coefficients multivariate probit model to quantify the complementarities between agricultural inputs and alternative forms of unobserved heterogeneity effects in modeling farmers' technology adoption decisions. The empirical analysis reveals that, conditional on various types of unobserved heterogeneity effects, farmers' technology adoption decisions exhibit strong complementarity for some inputs. The analysis also reveals substantial unobserved heterogeneity effects. We show that ignoring these behavioural features (unobserved heterogeneity and input complementarity) has important implications in quantifying the effect of some policy interventions that are meant to facilitate technology adoption. In particular, ignoring these features leads to significant overestimation of the effectiveness of extension services.