Although farmers’ trust and understanding surely influences demand for weather index insurance, resolving these concerns has not proven sufficient to solve the demand issues that beset this financial product. Demand for insurance does increase when farmers observe payouts over time ; receiving payouts in the previous year has a strong effect on increasing subsequent demand, increasing demand by almost 30% . However, not receiving payouts when a fair price has been paid has a strong negative effect on subsequent demand. Since the latter state is the ‘normal’ year for insurance consumers, this does not bode well for the adoption of insurance and its commercial viability. Cole et al. explicitly tests whether efforts to improve trust can improve take-up of insurance, finding that “in Andhra Pradesh, households were more likely to purchase insurance if an agent from a well-known microfinance institution endorsed the product, but in Gujarat, a similar endorsement in a marketing video had no effect” . In another study from Gujarat, Gaurav et al. found that receiving an invitation to a financial literacy training increased take-up by 5.3 percentage points, but the cost of the training was more than three times the full cost of premiums. Basis risk, the risk that the official weather observation will not accurately reflect a farmer’s loss,hydroponic net pots can also dampen take-up. For example, a farmer may experience poor weather conditions that damage their harvest, but if rainfall at the weather station is adequate, there would be no pay out.
Mobarak and Rosenzweig find in Uttar Pradesh that for every kilometer increase in perceived distance from the weather station, demand for weather index insurance decreases by 6.4% . This suggests that improvements in index design could help resolve basis risk, using improved data to more closely align the experienced conditions of smallholders’ plots with the measured conditions at data collection facilities used to set the index. Since basis risk is largely covariate for a geographic area, another promising approach appears to be the provision of insurance to groups that are already providing informal risk pooling of idiosyncratic risks among their membership . Index insurance can be a complement to informal risk mitigation where these informal risk pooling arrangements are smoothing idiosyncratic risks.Where insurance projects have been successful in achieving widespread uptake they tend to increase the appetite for activities vulnerable to risk . Magruder emphasizes, however, that these studies are “a handful of promising results that suggest the potential for risk reduction to spur technology adoption,” necessarily drawn from exceptional contexts where insurance take-up rates were sufficient to detect the impacts of insurance on productive technology adoption. This shift to higher risk and potentially more profitable production can have the somewhat counter intuitive effect of increasing the overall exposure of agricultural activity to rainfall volatility.
Insured households are better financially insulated , but landless laborers, whose income relies on harvesting crops, may become more exposed to risk as a result . This is concerning if the poorest rural households have limited alternatives should agricultural wage labor opportunities disappear.Scientists have developed stress-tolerant crops to protect farmers and help the broader agricultural system cope with extreme weather. These breeder-selected varieties of common seeds are agronomically designed to maintain high yields if a drought or flood occurs. Dar et al. conducted a two-year randomized evaluation with the International Rice Research Institute to study the effects of a flood-tolerant rice variety, Swarna-Sub1, on rice yields and farmer behavior in Odisha, India. Switching from Swarna to SwarnaSub1 cultivation does not require significant changes in farmer behavior. Flood-tolerant Swarna-Sub1 rice reduced risk for smallholder farmers and encouraged additional investment in their farms, resulting in substantially increased yields in both flood and non-flood years . Yields of this flood resistant rice variety were as good as regular varieties in normal conditions and superior during floods, and the yield gain went disproportionately to low-caste farmers because of the less-desirable risk-prone location of their lands due to generations of social marginalization . Higher yields increased farmers’ revenue by approximately US$47 per hectare relative to farmers in comparison villages, and 36% of this additional revenue was reinvested in their land. The results show how farmers respond to risk reduction by crowding in other investments and technological changes, which effectively double farmers’ expected gains: first from the agronomic benefits of the improved seed itself, and an equal benefit reaped from unlocking more productive practices when protected from risk.
These researchers are now evaluating the long-term effects of Swarna Sub-1, as well as the yield, welfare, and labor market impacts of other drought- and saline-tolerant crop varieties. Yet longer-term analysis shows that even improved varieties like SwarnaSub1, with demonstrated impacts and effective demand, have not been widely adopted. This shifts attention to other constraints, particularly the importance of seed supply and extension systems for technology diffusion, which will be discussed in the remaining sections in this chapter.Given insufficient demand for individual-level insurance, and the limited availability of agronomic technologies that could single handedly protect farmers across a wide range of agro-climatic conditions, ongoing research is focused on adjustments to, and combinations of, these risk mitigation approaches . Could institutions engage in risk-sharing to move risk away from particularly vulnerable smallholder farmers? For example, perhaps financial institutions or governments could serve as clients for meso-level insurance to see if that benefits smallholder farmers. “Under this arrangement, [e.g. the World Bank Group’s Global Index Insurance Facility and African Reinsurance Corporation plan], a government or institution would reimburse insurers above a set loss ratio. This decreases risk and costs for insurers and could lead to lower premiums for farmers” .Or, given willingness to pay for individual insurance has been a challenge, perhaps free or subsidized insurance could be offered as a form of social protection, achieving a multiplier effect by releasing farmers’ production decisions from risk constraints. Ongoing research is testing strategic combinations of financial products including index insurance, precautionary savings, and emergency credit , to understand if these bundled products can protect smallholders across a spectrum of risks of varying severity . Or perhaps financial products could be combined with risk-protective seeds in an attempt to better mitigate risk under a wide range of conditions. An evaluation is underway by Carter et al. that offers a combination of drought tolerant maize and index insurance in Tanzania and Mozambique, given index insurance could protect against the extremely adverse events that prove so taxing that even stress-resistant seeds fail .Farmers face a range of potential production technologies and practices to choose from, each of which may have different risk profiles and different suitability for a farmer’s own plots. Many technologies have heterogeneous returns that vary based on local plot characteristics and complementary input choices or agronomic practices . In addition,blueberry grow pot any single year farmers can only observe the performance of the technology under one weather realization, and understandably have trouble predicting outcomes under a range of different conditions they could experience in the future. A variety of specific information is therefore necessary for farmers to make good decisions as to which technologies to use at which specific points in time. Extension services have been a common approach used to inform farmers and encourage technology adoption, and have traditionally been one of national agriculture ministries’ main types of expenditure . The available literature has compiled summary statistics on the type of learning outcomes that we should hope to see as a result of large investments in extension systems. These statistics match anecdotal understanding that often status quo extension systems are characterized by limited supply of extension agents in communities, and even where available, low engagement with services, low adoption of recommendations, as well as low information diffusion beyond a select few contact farmers .For example, in Mozambique, countrywide extension coverage is as low as 1.3 agents per 10,000 rural people , while in Malawi approximately 50% of government extension positions remain unfilled .
Glendenning et al. estimated that fewer than 6% of the agricultural population in India reported having received information from the Government of India’s decades-long extension program. Extension workers may simply shirk responsibilities, or choose to focus their attention on villages or individuals based on their convenience to reach or perceived potential , and may neglect more marginalized farmers from poorer or female-headed households . Rural extension services are difficult and costly to monitor, limiting the potential to hold extension workers accountable and providing little direct incentive to show up for work or complete their duties . Duflo et al. shows another, perhaps intuitive, reason that extension can be ineffective – if it is promoting the use of a technology that is not profitable. They found that test plots using fertilizer recommendations from the Kenyan Ministry of Agriculture did not encourage fertilizer adoption by farmers. In this case, Duflo et al. calculate that these fertilizer recommendations would have increased farmers’ yields if applied, but would have actually reduced farmer profits. For food security and broader economic reasons, governments are often interested in maximizing yields, and thus extension programs may make regional-level fertilizer and other input use recommendations that target yield outcomes that farmers disregard as unlikely to be profitable. Gearing extension service recommendations based on what is profitable is important to drive adoption. Traditional extension models that directly train “contact” farmers typically do so given budget limitations, in the hopes that these contact farmers will share information and encourage new practices among other farmers. Analyses of panel data suggest that farmers learn from observing the decisions and experiences of people in their social networks . Conly and Udry in particular examine the importance of heterogeneity in observable characteristics of demonstrators for the transferability and diffusion rate of a particular technology, and Rogers discusses how “trialability,” the degree to which a potential adopter can try something out on a small scale first before adopting it completely, can also be an important factor for adoption. We present findings from randomized evaluations that show how agricultural extension can be more effective in both the initial design of channels and pedagogy, and the use of social networks to encourage the spread of technology adoption.Evidence presented in this section shows that well-designed information provision attuned to smallholders’ information needs and social networks can encourage poor farmers to invest in new technologies. There is a need to support farmers’ decision-making when introducing unfamiliar inputs, and early evidence demonstrates how to adapt information provision systems to support smallholder farmers’ in technology adoption. The inefficiencies that cripple extension systems and lead to information constraints can also impede the physical availability of technologies. In Uganda, Bandiera et al. worked alongside BRAC’s roll out of their female extension worker model in Uganda to understand how social networks, credit constraints, and expectations about the returns to technology affect adoption decisions of improved seed varieties and modern farming practices such as zero tillage, line sowing, and disease prevention. Increases in agricultural productivity were achieved by effectively targeting the key “accessibility” constraint in this context — farmers previously had little access to quality improved seeds. However, social connections mediated input market access: extension workers play an important role in selling seeds, and farmers in their networks had better access to technologies compared to less-connected farmers. Emerick et al. are extending their work on improved seed and extension in Odisha to understand how agents in input value chains can effectively provide both information and technology. Given that a lack of information is unlikely the primary constraint to technology adoption where inputs are not locally accessible, and improved quantities, qualities, or varieties of output may not be profitable in current market structures, we turn our attention to supply chains and markets.Output market dynamics can also affect smallholder decision-making. Information about market conditions and prevailing prices could influence farmers’ decisions of how, when, and where to sell their harvest, while search and transport costs, and relationships with traders, mediate producers’ access to potential points of sale. In Sub-Saharan Africa in particular, farmers are limited in their ability to access lucrative options for sale .