Pro-cyclical productivity patterns in United States agriculture

This dissertation focuses on the measurement of capital and productivity in U.S. agriculture. Two data sets that measure inputs and outputs at the state level were compared, revealing that the measures of capital input are drastically different. This finding led to an investigation of the construction of the capital input measures, and of a potential measurement issue related to capital inputs that could result in bias in estimates of productivity growth.; Some of the production accounts data were revised for use in this dissertation; the largest revision being a revaluation of the stock of service structures on U.S. farms, which had the effect of reducing the estimated annual flow of capital services in U.S. agriculture by roughly 20 percent from the original estimates.; Typical measures of capital input are constructed by assuming that the annual flow of capital services is proportional to the stock of capital. The proportionality assumption implies that farmers will not vary the utilization of capital stocks in the short run in response to changes in economic incentives or environmental conditions. This assumption, as well as other general difficulties related to measuring durable inputs, is commonly cited as an explanation of pro-cyclical patterns in measures of productivity growth.; I relax the proportionality assumption by defining a utilization response that depends on environmental conditions and farmers' terms of trade. The utilization response is then embedded in production functions to test for the significance of utilization-based measurement errors in primal models of production and productivity. I find that the utilization response is statistically significant and can result in substantial measurement errors for capital inputs. Furthermore, the utilization response is shown to contribute to the pro-cyclical patterns observed in measures of productivity growth.