| Previous studies of productivity patterns in U.S. agriculture have mostly used national aggregate data and have not adjusted for changes in the composition of inputs and outputs over time. Aggregation bias, and unaccounted-for changes in the composition of input and output, cloud the interpretation of the resulting measures.State-level, time-series data from 1949 through 1991 on inputs and outputs used in U.S. agriculture were compiled by Craig and Pardey (1996). These data were developed with corrections for composition changes (especially of labor and capital inputs) over time and space, and using state-level prices. These data were used in this study to analyze patterns of productivity and technical change in U.S. agriculture over time and space in the post-World War II period. Results from parametric, nonparametric, and index number approaches to measure both the rate of productivity growth, and the input and output technical change biases are presented and compared.U.S. agricultural output grew at an annual rate of 1.71 percent, and the use of inputs declined by 0.19 percent over the period 1949 to 1991. The 48-state aggregate average annual rate of productivity growth from the econometric, nonparametric, and index number procedures were 2.00, 1.91, and 1.90 percent respectively. On average, U.S. agricultural production experienced land-using, labor-saving, capital-saving, and purchased input-using technical change over the period 1949 to 1991. Patterns of output, input, and productivity growth, and technical change biases among states were quite diverse, and substantially different from national growth rates. |
