This collection of essays investigates producer use of information in decisionmaking, with applications to optimal resource use and learning. In the first essay, "Optimal Soil Conservation on Farmland," the economic costs of future soil degradation on cropland are shown to drive privately-optimal protection of stocks of soil nutrients and soil depth. A dynamic, simultaneous model of reversible and irreversible degradation for a Marlette, Michigan soil is presented, along with a sensitivity analysis with respect to costs and efficiency of crop residue management, input and output prices, and discount rates. Simulations show forward-looking producers increasingly engage in crop residue management over time to maintain high income streams, although at levels less than advocated by many conservation programs. Privately-optimal residue management decisions greatly diminish the amount of new technology required to address food security concerns.
In the second essay, "On-farm Experimentation and the Value of Information," a dynamicmodel of optimal firm learning is presented that studies the incentives to experiment in order to reduce payoff uncertainty. An application to precision nitrogen fertilization is presented which assumes the yield function holding for an individual plot is known only probabilistically. Bayesian updating allows for probability beliefs to change over time. Two learning strategies are compared. The dynamic optimization problem characterizes the active learning strategy, combining control and estimation. The static optimization problem characterizes the passive learning strategy, separating control and estimation. Active learning is a superior strategy for the producer, although it results in applying nitrogen fertilizer at higher rates than if a passive learning strategy is followed. Firm learning may explain seemingly-excessive fertilization rates on cropland.
In the final essay, "The Benefits of Off-site Research Results in Site-Specific Agriculture", optimal firm learning is modified to allow for a one-time update of probability beliefs from off-site sources, such as land-grant colleges. The objective of university researchers, who are assumed to minimize Type 1 and Type 2 hypothesis-testing error, is contrast with the firm's optimal learning objective that maximizes ex-ante returns. Results show that off-site research is more valuable to producers following passive rather than active learning strategies.