The Millennium Development Authority (MiDA) financed the construction of a new irrigation scheme in Kpong and the renovation of two irrigation schemes in Botanga and Golinga. This data contains baseline information for the impact evaluation of this activity.
Treatment groups involve households with farmers who belong to FBOs within the geographic perimeters of the new/renovated irrigation schemes; these treatment groups are provided contracted access to an anchor farm that will enable them to receive irrigation. Comparison groups are households outside the water supply perimeters provided by the new/renovated irrigation scheme with characteristics similar to those of households in the treatment groups.
The three main research questions this evaluation will try to answer (once endline data is collected), presented in form of hypotheses to be tested, are: i) new irrigation schemes will raise production from 2 to 3 crops per year; ii) irrigation will allow for diversification of crops and potentially higher yields; and iii) irrigation will increase labor requirements. From these three hypotheses, there are five indicators that can be used to measure the impact of these irrigation activities: (1) total household income; (2) total household income from crop production; (3) paid employment per household; (4) crop mix - annual production output (kilograms) for each of the five most imported crops per household (i.e. want to observe move from low to high-value crops); and (5) crop yield (i.e. monitor output per unit, kilogram/hectare per crop cycle).
To evaluate this program, once endline data is collected, NORC proposed to use a difference-in-difference approach and an IV approach based on a distance indicator (i.e. instrument treatment with “farmer's distance from anchor farm” if we can assume small farms closer to anchor farms are more likely to benefit from activity).
In terms of descriptive statistics there is no clear evidence that households in the treatment group are better or worse off than households in the comparison group at baseline. There are, however, important differences to consider. While the household head in the treatment group is more likely to be female, slightly less educated, and live in an informal dwelling, they were more likely to have children currently attending school. Households within the treatment group also had, on average, higher income, though the variance was high (as is the case with income in general), and the difference was not statistically significant. With respect to farming activities, there are two important differences to highlight across experimental groups. First, households in comparison groups were more likely to own their own plots. This is important because the impact of irrigation activities could be confounded if the households in the treatment group are less likely to make long-term investments than households in the comparison group. Second, households in the treatment group owned smaller plots of land in terms of area, on average.