Heterogeneous technology and panel data : the case of the agricultural production function

The paper presents empirical analysis of a panel of countries to estimate an agricultural production function using a measure of capital in agriculture absent from most studies. The authors employ a heterogeneous technology framework where implemented technology is chosen jointly with inputs to interpret information obtained in the empirical analysis of panel data. The paper discusses the scope for replacing country and time effects by observed variables and the limitations of instrumental variables. The empirical results differ from those reported in the literature for cross-country studies, largely in augmenting the role of capital, in combination with productivity gains, as a driver of agricultural growth. The results indicate that total factor productivity increased at an average rate of 3.2 percent, accounting for 59 percent of overall growth. Most of the remaining gains stem from large inflows of fixed capital into agriculture. The results also suggest possible constraints to fertilizer use.