Publications

This study evaluates the impacts of a pilot project that introduced weather and climate
information services (CIS) on technical efficiency (TE) and sorghum productivity (Y) using a total
of 210 plotlevel data comprising of 92 users and 118 nonusers of CIS in the Upper West Region
of Ghana. First, we estimate a Stochastic Frontier model to measure the level of TE using a
Cobb-Douglas functional form with the assumption of an exponential distribution for the
identification of TE scores. Secondly, we measure the impact of the adoption of CIS on TE and Y
separately by addressing the potential bias stemming from the existence of unobserved
characteristics using a Control Function estimator. Thirdly, we employ a Recursive Structural
Equation System to deal with the simultaneous problems of the endogenous treatment of CIS
into Y, the endogenous covariate of TE into Y, and the reverse causality between Y and TE.
Overall, our findings are robust to the different methodologies with strong evidence that the
pilot project through the adoption of CIS has a substantial positive effect on improving TE and Y
in the study area. Our empirical results consistently estimate approximately 6% increase in TE
and 35% sorghum yield improvement corresponding to 150 Kg/Ha increased productivity among
CIS users. Furthermore, when we simultaneously estimate the combined effects of the adoption
of CIS and the level of TE on sorghum productivity with and without the problem of reverse
causation between Y and TE, the median value of the average treatment effects (ATE) is 10%.
Also, improving the level of TE has a higher payoff among users than nonusers of CIS resulting in
increased sorghum productivity of 5% when we compare the average treatment effects on the
treated (ATET) and the average treatment effects on the untreated (ATEU). These results
underscore a valuable policy insight and the importance of privileging the wide adoption of CIS
and promoting the efficient use of inputs with best-recommended climate-smart agricultural
practices such as crop management and increasing trainings to raise awareness in future project
expansion. However, it appears that the magnitudes of the impacts of the adoption of CIS on Y
using the ATE, ATET, and ATEU are sensitive to whether we address the potential reverse
causality between Y and TE. These findings indicate that more caution should be considered in
the evaluation of the impacts of a project that promotes agricultural innovations including
information communication technologies on farm productivity and technical efficiency.