The Shifting Effect of One Acre Fund on Agricultural Production Possibility Frontier: An Econometrics Analysis of Treatment Effect Estimation of Maize Yield in Busia County, Kenya

Abstract

Maize is a staple food and its adequacy is a measure of food security in Busia County and across Kenya. Despite myriad efforts to increase its yield, the national level has remained low at 1.6 tons/ha relative to world average of 5.5 tons/ha and Busia County experiences much lower yield of 1.2 tons/ha. This low yield in the county coupled by high population growth rate arising from a fertility rate of 6 percent in the county when national level is 4.6 percent and a large population of women within productive age puts the county’s food security at stake. In attempts to support maize farming, One Acre Fund, a non-governmental organization has intervened through provision of farm inputs such as fertilizer and seeds as well as extension services among others. Thus this study sought to investigate the effect of interventions by One Acre Fund on maize yield in Busia County in Kenya. The study used primary data collected from a sample of 264 maize farmers’ family heads using questionnaires. The county was first stratified into seven constituent sub-counties out of which two sub-counties, Nambale and Teso North were randomly selected. The study sample was then drawn randomly from these two sub-counties. Robust regression result showed that improved inputs (using certified seeds and organic fertilizers as proxies) and organic manure had the effect of increasing maize yield. Propensity score matching estimation results using three matching algorithms interchangeably which are: nearest neighbor, kernel matching and radius matching with a radius of 0.01 indicated that One Acre Fund farmers managed to achieve significantly higher yields than similar non-members. The study recommends expansion of One Acre Fund membership and increased use of organic and inorganic fertilizer.

Keywords: Maize yield, One Acre Fund, Busia County, Propensity Score Matching, Average Treatment Effect on the Treated, improved inputs