This is the fourteenth in this year's series of posts by students on the job market.
Adoption of profitable agricultural technologies is strikingly low in sub-Saharan Africa (World Development Report 2008). Economists and policy makers worry about the under-adoption of technology because it constrains agricultural productivity and, more recently, makes farmers more vulnerable to the adverse effects of climate change (Dell et al, 2014). When deciding whether to adopt an agricultural technology, farmers might not fully internalize the benefits that they generate for other farmers. In equilibrium this leads to lower technology adoption than what would be socially optimal (Pigou, 1920).
In my job market paper, I study the relationship between positive externalities and technology adoption. I investigate whether positive externalities are empirically important for farmers when making adoption decisions, and which policies would exploit these externalities to increase adoption and welfare. The type of agricultural technologies I study is pest control practices. Since 2016, African farmers have experienced one of the worst infestations in recent years, by a new pest called the Fall Armyworm (FAW). My study is set in Uganda, where the FAW first appeared in early 2017 causing maize harvest losses worth about 1% of Ugandan GDP (FAO). Because the pest is new to the sub-Saharan ecosystem, knowledge of the FAW and practices to avoid infestation is very low among farmers. In late 2018 I offered 799 farmers in 103 villages the possibility to purchase an agronomist-led individual training on pest control practices to handle the FAW. This is what I learned about technology demand and the role of spillovers.
Farmers value the technology
To establish whether farmers are interested in the technology, I measure farmers’ truthful valuations for the training using a variant of the incentive-compatible Becker-DeGroot-Marschak (1964) (BDM) mechanism. I elicit farmers’ maximum willingness-to-pay (WTP) for the training using a set of increasing take-it-or-leave-it offers. Farmers can formulate WTP ranging from 0 to 34 USD. I find the median farmer I interview is willing to pay 17 USD to receive a training on the FAW. This is equivalent to four days’ wage of agricultural labor, a substantial amount considering that governments and NGOs usually provide agricultural trainings for free.
Farmers value that others have the technology
I use the BDM mechanism and elicit how much a farmer values that another farmer in her village receives the training to study whether farmers attach value to the externalities generated by others’ adoption. This is a novel metric that allows researchers to measure the monetary value of others’ choices, and I use it to calculate the monetary value farmers attach to spillovers. I argue that farmers who are willing to pay some amount for another farmer to receive training, expect to benefit from it. The majority of farmers are willing to pay considerable amounts for others to receive the training. The median WTP for another farmer in the village to receive the training is 9 USD or two days’ wage of agricultural labor.
Externalities matter
I single out two types of externalities that I expect to be the most relevant. The first is a contagion externality; if a farmer adopts a pest-control technology, other farmers face a lower risk of being infected (Miguel and Kremer, 2004). The second is a knowledge externality; if a farmer adopts a pest-control technology, other farmers can learn about it (they can talk to or observe the farmer) (Foster and Rosenzweig, 1995). To separate these channels, I generate random variation in perceived contagion and knowledge externalities with a belief update intervention. To identify contagion externalities, in randomly selected cases I inform the farmer of the true distance between her plot and another farmer’s plot, and then elicit her WTP for that farmer. Because the FAW spreads by proximity and farmers are not perfectly informed about distances (95% have wrong beliefs in my sample), revealing the true plot distance causes the farmer to update her beliefs over the probability that the FAW could spread from the other farmer’s plot to hers. I find that when the other farmer’s plot is further away than expected, and therefore contagion externalities are lower than expected, a farmer reduces her WTP for the other farmer. At means, the effect is about 0.43 USD or 10% of the daily wage. This shows that farmers value contagion externalities.
To identify knowledge externalities, in randomly selected cases I inform the farmer that we are organizing a meeting with another farmer, and then elicit her WTP for the other farmer to receive the training. This causes the farmer to update her belief over the probability that she learns from the other farmer. I do not find that farmers respond to this treatment, which is surprising since village meetings are a common tool to spread information among similar populations. But I provide evidence that farmers are more willing to pay for farmers whom they know, are related to, and go to for advice. Taken together, these results are suggestive evidence that knowledge externalities matter, in line with the literature on social learning that finds farmers learn from each other (Foster and Rosenzweig, 1995; Bandiera and Rasul 2006; Conley and Udry 2010).
Social benefit of adoption is larger than the private benefit
An interesting feature of collecting data on WTP for others is that it allows me to calculate the social benefit generated by training a certain farmer, using the WTP of all the other farmers in her village. I calculate that the social benefit of training one farmer is at least 18 times higher than the training cost; the private benefit of receiving training, however, is lower than the cost on average. This is the bad equilibrium: no one invests because the net private benefit is negative, but investing would generate large social benefits that outweigh the costs.
Escaping the low technology equilibrium
My results have two important policy implications. First, because the social benefits depend on who receives the training, policy makers can increase the impact of agricultural extension policy by taking into account the social and geographic position of the targeted individuals. Second, my estimates suggest that subsidies are valuable not only for their direct effect (increasing take-up), but also for their indirect welfare effect through positive externalities.
Benedetta Lerva is a PhD student at the Institute for International Economic Studies (IIES), Stockholm University. Her website is linked here.
Join the Conversation