Impact of genetically modified maize on risk, output, and cost among smallholders in South Africa

Abstract

Previous research in low-income countries reveals that genetically modified (GM) maize has the potential to increase yield and reduce labor use; however, other issues, especially regarding Roundup Ready (RR) maize, remain mostly unexplored. This research examines the impact of GM maize on yield, cost, and risk among 184 smallholders during the 2009-10 maize production season in two regions in KwaZulu-Natal, South Africa; Hlabisa and Simdlangetsha. Two hybrid maize varieties; Pannar and Carnia, and three GM varieties; Bt, RR, and BR (stacked with Bt and RR) are produced. In both regions, producers of RR and BR maize pay 47% more per kilogram of seed and use 44% less labor per hectare compared to other varieties. Due to low labor costs, net returns from RR and BR varieties are 25% and 40% higher than other varieties in Hlabisa and Simdlangetsha, respectively. Stochastic dominance analysis is used to compare net returns of all five varieties in both regions. RR maize is second-degree stochastic dominant to all other varieties in Simdlangetsha, while no variety is stochastically dominant in Hlabisa. Stochastic efficiency with respect to a function (SERF) analysis indicates that RR maize is the preferred variety for producers over the entire range of risk preferences in both regions. While average maize gross returns are $713 per hectare, risk premiums between $18 and $221 must be paid to RR maize producers, depending on region and farmer risk preference, to persuade them to switch to the second-most preferred variety. Econometric analysis indicates significant yield gains of at least 8% from RR maize, although the yield gain varies greatly when input endogeneity is taken into account. Elasticities of output with respect to labor are 0.41 and 0.82 for RR and non-RR maize respectively, and 0.61 and 0.33 with respect to land. A cost function analysis indicates that RR maize has 19% lower costs per maize plot, which increases to at least a 35% advantage when controlling for selectivity bias. Nonparametric kernel density estimation also reveals consistently lower total and average costs of RR maize at most levels of output, suggesting technological benefits to smallholder farmers from RR maize not available through conventionally-bred hybrids.