Spatial dependence and spillover effects in crop diversification practices: farm household level evidence from Ghana

Diversification has been practiced by several economies globally due to its potential to drive structural economic transformation as well as its prospects in reducing the risks associated with crop failure under monocrop production regimes. Despite its widespread promotion and adoption among farm households in developing countries, little is known about how farm households’ crop diversification decisions are spatially interdependent, leading to underexploited spatial spillovers. This study utilizes the geo-coded farm household-level information in the wave 3 cross-sectional data of the Ghana Socioeconomic Panel Survey to examine spatial dependence in crop diversification practices and to estimate the spatial spillover effects. Using inverse distance spatial weighting matrices for farm households that are 2, 10, and 50 km radii apart, the results indicate the existence of positive spatial autocorrelation in crop diversification practices in Ghana. The findings from the Spatial Durbin Model, which reports the average marginal effects, reveal that irrigation access, farm size, extension access, tractor and animal plow access, social networking, tenure security, communal labor arrangement and agroecological factors exhibit significant spatial spillover effects. The Spatial Durbin Model results for the 10 km radius and 50 km radius inverse distance spatial weighting matrices were also estimated as robustness checks. The results are quite robust when estimates are compared across the different inverse distance spatial weighting matrix specifications. Potential policy implications include encouraging farm households to join social networking groups, developing and promoting collective irrigation schemes and shared mechanization services due to the positive spillover effects and externalities they generate. Crop diversification programs should also be targeted at agroecological zones.