Technology adoption under uncertainty

Many technology adoption decisions require investment at two or more points in time. The first investment is typically associated with take-up and subsequent investments with the use or implementation of the technology. After take-up, new information about the cost of subsequent investments is acquired, and adopters may decide to abandon the technology if they learn that following-through will not be worth it after all. We study this dynamic adoption problem in the case of agroforestry in Zambia. A field experiment that varies the payoffs from taking up seedlings and following through to keep the trees alive allows us to estimate the new information that arrives after take-up. We observe that, while farmers are responsive to the incentives offered in the experiment, a large share of the payoffs from adoption are not known to them at the time they make their take-up decision. Importantly, this makes it difficult for farmers to self-select into the take-up decision, meaning that a higher initial cost for the technology does not make follow-through more likely. As a result, subsidies for take-up are less cost effective, but also less necessary since many farmers facing uncertainty about the costs of follow-through choose to take-up so that they have the option to follow-through should the new information that arrives after take-up contain good news about the profitability of the technology.