Carolyn R. Harper

Pest Externalities from Agricultural Inputs

Agricultural inputs such as water, pesticide, and even time may have the unintended effect of stimulating some pest populations, leading to crop losses. A conceptual model is developed to contrast optimal use of pesticide and nonpesticide inputs with myopic use patterns which ignore pest externalities. Under most conditions, optimal management is found to entail reduced input levels. These issues are illustrated for Imperial Valley cotton using biological simulation. Correct calculation of the relative profitability of conventional and integrated pest management techniques, such as a shortened growing season, are found to depend crucially on whether pest externalities are taken into account.

Pesticides and Worker Safety

A key problem in pesticide regulation is uncertainty about health risks. Trade-offs between economic benefits and worker health safety are examined using an empirical illustration. Alternative decision rules for regulation under uncertainty are considered: a safety fixed rule, which protects individuals from excessive health risks, and uncertainty-adjusted cost-benefit analysis, which evaluates aggregate trade-offs between health and economic welfare. These criteria may lead to opposite policy conclusions, suggesting that the most appropriate public policy is a safe minimum standard (SMS), which allows weighing of costs and benefits only after some minimum acceptable level of health safety has been assured.

Groundwater protection in mixed land-use aquifers

Compared with groundwater pollution episodes of a point-source nature, nonpoint-source contamination makes for particularly difficult policy design. This is especially true in the context of a mixed land-use aquifer, where the same pollutant may derive from various human land-use activities and in different concentrations. To data, the emprirical literature attempting to estimate the relative loadings of pollutants from alternative land uses is rather sparse. Yet this information is vital to a variety of numerical computer models used to predict likelihood of groundwater contamination, and the statistical results are useful in their own right for regional policy formation. Regression analysis is applied to estimate loadings of nitrate and sodium from various land uses. The model is then used to illustrate how well-intended local groundwater protection policies that fail to recognize land-use substitution and cross-pollutant effects may be misdirected.