Technology adoption, capital deepening, and international productivity differences

I document that cross-country differences in capital intensity are much larger in the agricultural sector than in the nonagricultural sector. To explain this fact, I build a model featuring technology adoption with fixed costs among heterogeneous farmers. More productive farmers operating larger farms pay the fixed cost and adopt a modern capital-intensive technology, while less productive ones choose a traditional labor-intensive technology. The model is calibrated using historical data on farmer adoption of mechanized technology in postwar America. This calibrated model is then employed to perform cross-country comparisons. Incorporating a technology adoption channel not only allows the model to predict substantial differences in agricultural capital intensity between rich and poor countries that an otherwise identical model would fall short in generating, but also improve explanatory power for cross-country agricultural productivity differences by 1.5-fold. •Cross-country differences in capital intensity are much larger in agriculture than in the nonagricultural sector.•In the U.S., nominal capital-output ratio was stable in nonagriculture while it increased in agriculture during 1940–1980.•Technology adoption is key to explaining differences in agricultural capital intensity between rich and poor countries.•Incorporating technology adoption helps a model explain 1.5-fold more of agricultural productivity differences between rich and poor countries.