Read on Twitter
1/N 
New WP!!! (ungated http://documents.worldbank.org/curated/en/496531576687282363/pdf/Factor-Market-Failures-and-the-Adoption-of-Irrigation-in-Rwanda.pdf) w/ all-star coauthors @mariaruthjones @loeserjohn and Jeremy Magruder. We find: failures in land and labor markets generate misallocation that limits adoption of productive technologies. \\begin{thread} https://twitter.com/nberpubs/status/1222987851442245632
New WP!!! (ungated http://documents.worldbank.org/curated/en/496531576687282363/pdf/Factor-Market-Failures-and-the-Adoption-of-Irrigation-in-Rwanda.pdf) w/ all-star coauthors @mariaruthjones @loeserjohn and Jeremy Magruder. We find: failures in land and labor markets generate misallocation that limits adoption of productive technologies. \\begin{thread} https://twitter.com/nberpubs/status/1222987851442245632
2/N Central puzzle in development: agricultural productivity and adoption of proven technologies in SSA have remained low. This has important welfare implications: 82% of people living in extreme poverty in Africa earn most of their incomes from farming.
3/N In our study schemes, a main canal carries water from a reservoir to the plots, following a slow gradient along the hillside. We aerially sample 1,700 plots w/in 50m of the canal across 3 sites and interview the HHs cultivating these "sample plots" over 4 years.
4/N We establish a first stage—but we also find our main puzzle! Adoption of irrigation jumps discontinuously at the boundary of the command (irrigated) area—but only 30% of farmers in command area adopt … WHY? We use this jump to produce three results.
5/N First, we establish that productivity is not the problem: yields net of expenditures increase by 46-70% (TOT). However, profitability depends crucially on how HH labor is valued --> labor market frictions likely important for farmers’ adoption decision.
6/N Important to note: this result comes solely from an increase in dry season irrigated horticulture, with no increase in dry season share of cultivated land. Instead, irrigation enables crop rotations to shift from low productivity bananas to high productivity horticulture.
7/N Second, we ask: is this low level of adoption inefficient? It may not be: if returns are very heterogeneous, observed low adoption may actually be efficient. Alternatively, adoption is constrained by at least 2 market failures --> inefficiently low.
8/N We exploit our spatial discontinuity in access to irrigation as a random productive shock that increases demand for inputs on the "sample plot" to produce a test for separation failure.
9/N 
Intuition: the irrigation shock on a farmer’s “sample plot” does not affect the optimal allocation on a farmer’s other plot. However, if farmers are constrained (and land markets are incomplete), the shock may force them to substitute away from the other plot.
Intuition: the irrigation shock on a farmer’s “sample plot” does not affect the optimal allocation on a farmer’s other plot. However, if farmers are constrained (and land markets are incomplete), the shock may force them to substitute away from the other plot.
10/N We observe a separation failure: when a farmer’s "sample plot" receives the irrigation shock, they are less likely to be able to use irrigation on their other plot, with no change in cultivation on either plot. Farmers are "stretched" --> reduced irrigation on other plot.
11/N Market failures constrain irrigation adoption, and adoption is inefficient. Eliminating this substitution would increase adoption of irrigation by at least 21%. We’ve found land market are incomplete--which other market failure(s) produce this inefficiency? Two things.
12/N First, we adapt Benjamin (1992) for 2 plots, 3 constraints: credit/input, insurance, labor. As a result of the irrigation shock, we predict: if financial constraint --> wealthy HHs should not substitute; if labor constraint --> large HHs should not substitute.
13/N We examine heterogeneous impacts for wealthier farmers and farmers with larger households. We find that, if anything, wealthy households substitute more, large households substitute much less --> suggests that labor market failures constrain adoption.
14/N However, demonstrating that labor and land markets fail does not allow us to reject that other market failures matter here!
15/N Second, we fielded a RCT to test for credit/knowledge constraints: we randomized small seed and fertilizer "minikits" for horticulture. If credit/knowledge constrain adoption, HHs who haven't irrigated should take up the minikits and increase adoption of horticulture.
16/N Results from this test lead us reject that credit/information limit adoption in this context. Despite a strong first stage, we find no effects on adoption of horticulture: take up was limited to farmers who has used horticulture previously!
17/17 We’ve shown that land and labor misallocation leads to inefficient adoption of a profitable technology: small HHs have too much irrigated land, large HHs have too little. This outlines the need for more evidence on role of factor markets in technology adoption. \\end{thread}
tagging the correct @mariaruth83 now :)))
