Returns to Agricultural Public Spending in Ghana : Cocoa versus Noncocoa Subsector

Using public expenditure and agricultural production data on Ghana from 1970 to 2012, this paper assesses the returns to public spending in the agricultural sector, taking into consideration expenditures on agriculture as a whole and then separately for expenditures in the cocoa versus the noncocoa subsectors. Production functions for the agricultural sector as a whole are estimated first, and then separately for the two subsectors, to obtain elasticities of land productivity with respect to total and sectorial agricultural expenditure. Different regression methods and related diagnostic tests are used to address potential endogeneity of agricultural expenditure, cross-subsector dependence of the production function error terms, and within-subsector serial correlation of the error terms. The estimated elasticities are then used to calculate the rate of return (ROR) to expenditures in the sector as a whole and within the two subsectors. The elasticities are estimated at 0.43 for total agricultural expenditure; 0.13 for aggregate expenditure in the noncocoa subsector; and 0.19-0.53 for expenditure in the cocoa sector, depending on aggregation or disaggregation of expenditure on the Ghana Cocoa Board and other industry costs. The ROR is estimated at 141-190 percent for total agricultural expenditure, 124 percent for expenditure in the noncocoa subsector, and 11-39 percent for expenditure in the cocoa subsector. The relatively higher ROR in the noncocoa subsector is mostly due to a much lower expenditure-to-productivity ratio. Implications are discussed for raising overall productivity of expenditure in the sector, as well as for further studies, such as obtaining actual time-series data on some of the production factors in the two subsectors and obtaining information on the quality of sectorial expenditures to model different time-lag effects of spending in the different subsectors