Economics of Technological Change and Induced Innovation

摘要

My dissertation examines different aspects of technological progress and its impact on input allocations and productivity growth in U.S. agriculture. In the first chapter, I test whether the Hicks' Induced Innovation Hypothesis (IIH) holds in the industry for the period 1960--2004 using state-level input price and quantity data with major focus of allowing for a non-neutral innovation function. With homothetic production and innovation function, I derive a structural relationship between the expected input price ratio and input allocation under cost minimization which permits the effect of the IIH to be distinguished from general factor substitutions. The contributions are threefold: (1) I demonstrate that input saving behavior consistent with the IIH would be observed in response to an increase in the relative input price for a wide range of elasticities of substitution, (2) I document that the both relationships between factor augmentation and expected relative price and between factor-saving behavior and marginal research cost are not a monotonic function of the elasticity of substitution when the innovation function is accounted for, and (3) I provide a test procedure that is robust to the assumption of a time-variant and non-neutral innovation possibilities frontier in the innovation creating industry. Considerable supporting evidence for the IIH is found in the data for all inputs except land. My second chapter uses the same dataset as the first chapter and tests whether the IIH holds while accounting for adjustment costs of quasi-fixed inputs. The findings suggest that none of the input categories instantaneously adjust to their desired level in one production period in U.S. agriculture. However, empirical test for the IIH using modified research lags that are based on the adjustment rate estimates result in less support for the IIH. The third chapter investigate spillover effects of public research on productivity growth in U.S. agriculture. A spatial econometric model is used to reduce bias in productivity equation estimation. The findings show that a large portion of productivity growth has been driven by public research investment and that a state's research outcome can provide benefits to other states that collectively exceed its own benefit.