To address this objective, a land use allocation model is constructed using slope to reflect terrain heterogeneity. The model is formulated as a mathematical programming problem, with the objective of maximizing producer welfare subject to an exogenous land endowment and a series of production constraints. The model developed in this thesis differs from previous empirical models in several substantive ways. First, crop and livestock production activities are explicitly modeled as either separable or nonseparable activities. The advantage to doing so is that it gives the model the flexibility to choose the optimal degree of integration between the two. The model also diverges from previous studies by incorporating a common set of variables that affect the economic and environmental aspects of commodity production. Specifically, the spatial allocation of land use practices impacts economic and environmental outcomes via a yield damage function and differentiated rates of soil erosion. These two aspects are expected to improve the model’s predictive ability.

One of the primary benefits of the model is that it can be used to identify the economic factors driving landscape-level production patterns. The analysis demonstrates that the land use allocation is relatively insensitive to changes in commodity prices. Therefore, altering the level of commodity-based income support payments is insufficient to attain environmental improvements. Several hypothetical “green” policy instruments are simulated to estimate the cost to producers of reducing environmental damages. The results indicate that limiting soil erosion to an environmentally acceptable level with either a regulatory standard or a tax reduces the average return to land by ten percent. A program of green subsidy payments for less erosive land management practices cannot attain the same standard with less cost to producers. Overall, the inelastic response of land use change to commodity prices indicates that targeting the use of productive inputs, as opposed to commodity outputs, may be a more efficient means of encouraging agricultural producers to provide environmental benefits.

The empirical application used throughout is the olive fruit fly in California. Chapter 2 develops and estimates a pest damage function, using an exceptional dataset collected by entomologists. The econometric model accounts for simultaneity that arises because common factors affect both pest population dynamics and host susceptibility. The estimated damage function aptly reflects biologically-driven heterogeneity in infestation rates across hosts. The implication for pest management is that growers may combat infestation by targeting the pest population or by manipulating host susceptibility.

Given randomness in common factors, mainly temperature, Chapter 3 incorporates both types of management tactics into a stochastic dynamic programming model of the incentives facing individual growers. The model allows growers to apply an insecticide at numerous times during the growing season, but growers may also curtail the growing season to reduce damage. Calibrated to various cultivars and growing regions, the results suggest that growers will change harvest timing to reduce their use of insecticide treatments. Some growers avoid insecticide use altogether by harvesting early.

Chapter 4 examines the individual benefits to membership in a PCD with heterogeneous commercial and non-commercial growers of a host. The analysis suggests that growers who eliminate insecticide treatments with early harvest lose 5.4-7.0 percent of seasonal returns by participating. In many cases, the estimated gains to other producers exceed these losses. However, the code governing PCDs in California mandates that all producers contribute a uniform levy to fund the institution. Redefining the structure of PCDs to reflect heterogeneity in returns to membership may enhance institutional efficiency and encourage cooperation among growers.