Switching costs in transient technology use: Panel evidence from Kenyan maize production

Abstract

The idea of switching costs and how they affect agricultural technology adoption has received little attention in the literature. We employ a dynamic switching model to show that switching costs generally discourage the adoption of new technologies by farmers, but the extent of the effect depends on the level of switching costs, as well as farmer perceptions and foresight. We provide an empirical strategy to identify switching costs and estimate their effects on adoption under different farmers’ perceptions about adoption uncertainty. An application to hybrid seed adoption in Kenya confirms the existence of switching costs. The correlated-random-effects probit results suggest there is heterogeneity in dominant farmer types across agro-ecological zones in Kenya. Area-specific policies are suggested to promote technology adoption in different environments.