Scrutiny of the risks that pesticides pose to human health and the environment is putting pressure on the U.S. Environmental Protection Agency to cancel registered uses of some pesticides. Some agricultural scientists fear that regulatory decisions will be made without sufficient knowledge of how they will affect crop production; some economists question whether analysis of the costs likely to result from cancellation of such registrations would adequately capture the complexity of crop production. Appropriate scientific experimentation and correction of shortcomings in economic analysis might affect decisions to cancel or not to cancel pesticide registrations.
The use of pesticides in agricultural production can increase crop yields, improve crop quality, and expand the range of farm management options—for example, the ability to use minimum or no tillage or to grow a crop earlier or later in the season. However, pesticide use can also pose risks to human health and the environment. In some cases, the risks may be great enough to warrant cancellation of a pesticide's registration for use on a particular crop or crops. But in other cases a pesticide's benefits, which are measured as the gain to growers and consumers from its use, may outweigh any risks use of the pesticide poses.
While extensive testing and research are conducted to provide data for estimating the risks posed by the use of a pesticide, scientific and economic data to support claims of the benefits of pesticide use are often not available. Current benefits analyses lack the sophistication and rigor required of risk assessments. Given the greater abundance of research regarding risks, many agricultural scientists fear that decisions to cancel registered uses of pesticides will be made relatively quickly over the next few years. Without the development of adequate benefits analyses, the consequences of such decisions would only become apparent as farmers sought ways to make do with fewer pesticides.
Of particular concern is the effect of cancelling the registration of pesticides for use on minor crops (such as snap beans, carrots, and strawberries), which often require complex pest control strategies, and crops grown in regions of the country that tend to have more severe pest problems (such as the Southeast). Losses in yield and quality could result if the available substitute pesticides prove less effective. Yet these losses are difficult to document in the absence of appropriate experimentation. Moreover, the substantial costs that some farmers would incur from these losses cannot be ascertained without more up-to-date economic data, which are necessary for more sophisticated economic analysis.
Benefits analysis
Under the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), the U.S. Environmental Protection Agency (EPA) is charged with the reregistration of all pesticides currently in use. It is during the reregistration process that the registration of a pesticide for use on a crop or crops may be cancelled. Amendments to FIFRA, which were approved in 1988, have expedited the process by requiring manufacturers of agricultural chemicals to supply data from scientific studies to demonstrate that currently registered pesticides do not pose an unreasonable risk to human health or the environment. Manufacturers have voluntarily withdrawn the registrations of many older pesticides whose patents have expired because those pesticides bring in profits too small to offset the cost of such testing. The effect that the unavailability of these pesticides would have on crop production may never be considered.
When studies identify no new evidence of risk, the reregistration process is completed. However, when a study identifies a hazard or risk, the EPA must determine whether or not the risk is unreasonable. This determination is based on risk-benefit analysis. Using such analysis, the EPA can estimate the costs of changes in agriculture that will accompany a decision to cancel the use of a pesticide and thus act to minimize any unintended consequences of pesticide regulation. Unfortunately, the benefits side of pesticide risk-benefit analyses suffers from a lack of scientific and economic data.
The benefits side of pesticide risk-benefit analysis suffers from a lack of scientific and economic data.
Benefits analyses at the EPA are conducted by the Biological and Economic Analysis Division within the Office of Pesticide Programs. Typically, the task is divided between a lead biologist in the Biological Analysis Branch and a lead economist in the Economic Analysis Branch. The EPA also utilizes benefits analyses prepared by the National Agricultural Pesticide Impact Assessment Program in the U.S. Department of Agriculture (USDA). In addition, the agency examines information submitted by chemical manufacturers as part of its benefits analyses.
Benefits are computed as the cost of switching to substitute pesticides, plus the value of any yield losses that occur when substitutes are used. The key pieces of agronomic information the EPA needs to compute pesticide benefits are pesticide usage patterns, including amounts applied and number of acres treated; the efficacy of a pesticide for controlling pest(s) that affect a particular crop as compared with other pesticides that are available; the costs of substitute pesticides or alternative pest control strategies; and the yield losses that would occur if a pesticide is banned for use on a crop or crops. To value yield losses, economists must derive the demand and supply schedules for an affected crop—that is, they must determine how such losses will affect the supply of the crop and how an increase in the crop's price will affect demand for the crop.
In one field-plot experiment conducted by the Rutgers Research and Development Center, the currently used pesticide pronamide was more effective as a weed killer in lettuce fields than the potential substitute pesticide bensulide
Weaknesses in scientific data
Detailed information on the extent of pesticide use in the United States has been spotty in past decades. With the National Agricultural Statistics Service, the USDA's Economic Research Service has conducted periodic surveys of pesticide usage on large-acreage commodities—such as field corn, wheat, soybeans, and potatoes—in the states where the majority of these commodities are produced. However, pesticide usage data for other crops grown in other states are rarely collected. (For example, surveys of pesticide use on fruits, which have a high economic value per acre, have not been conducted since the 1970s.) For pesticide usage data on these other crops, the EPA has had to rely on the best guesses of agricultural scientists, on estimates gleaned from private marketing reports, and on whatever state reports were available. However, the USDA has renewed efforts to obtain pesticide-use estimates for these crops. In 1991 it is expected to release the results of a survey of pesticide use on vegetables in the five major vegetable-producing states. A survey of pesticide use on fruits and nuts is expected to follow in 1992. Many state governments are also conducting their own pesticide-use surveys in response to information needs in formulating policy on issues such as water quality.
Much less progress has been made in obtaining data on the efficacy of individual pesticides for controlling pests. Currently, the EPA's strategy is to solicit the opinions of experts, usually crop protection scientists with states' cooperative extension services or land grant universities. Even when studies of differences in efficacy among pesticides are available, it is difficult to predict how these differences will affect crop yields.
Efforts to collect pesticide usage data are increasing, but less progress has been made in obtaining data on the efficacy of individual pesticides.
Furthermore, the use of expert opinion is often difficult to defend. A recent economic analysis of the proposed cancellation of the pesticide maneb submitted by the Pennwalt Corporation was not accepted by the EPA, partly because the comparative efficacy data, gathered by extensive communication with extension scientists, was judged to be inadequately documented. Problems with reliance on expert opinion extend to in-house analyses at the EPA. The agency is currently reevaluating how expert opinion is handled by its analysts after an administrative law judge gave little or no weight to expert predictions of the impact of cancelling the registrations of all pesticides containing arsenic. In an effort to make expert opinions defensible in administrative and judicial proceedings, the EPA policy now requires that such opinions be substantiated by experimental data or other credible information.
The need to solicit opinions regarding the comparative efficacies of various pesticides seems surprising when so much of the work of crop protection scientists involves the testing of pesticides. For example, hundreds of papers involving herbicide efficacy experiments are published annually in journals such as Weed Science and are presented at national and regional meetings. The Entomological Society of America annually publishes a volume entitled Insecticide and Acaricide Tests. In some cases, this research is used in the regulatory process. For example, experimental results from Insecticide and Acaricide Tests were used to determine comparative efficacy in the benefits analysis of diazinon use on turf.
However, much experimental work is not useful in benefits analyses during the reregistration process. Many tests are for newly registered or experimental products. Other tests compare pesticide-treated plots with untreated ones instead of plots treated with other pesticides. In some tests, pesticides are applied at higher or lower rates than the rate recommended by the manufacturer. Some experiments that do test the comparative efficacy of pesticides do not go on to measure the effects on yield of differences in pesticide performance. And in other cases a particular pest of concern in the experiment is not the pest of concern in the benefits analysis.
Experiments can be designed to provide information needed for the reregistration process, but the time required to plan and carry out the work can be a barrier. Reregistration is a risk-driven process. Once the EPA has compelling data regarding health hazards to pesticide applicators or consumers, it has a responsibility to act quickly in response to the risk. Generally, this leaves one year or less to prepare the benefits portion of the risk-benefit analysis. Thus, if experimental work is to be conducted for the benefits analysis, agricultural scientists must anticipate the pesticides and the crops for which information will be most needed. Given the possible combinations of pests, pesticides, crops, and regional growing conditions for which experiments could be designed, deciding which experiments will be most essential to the reregistration process is difficult at best.
Some field-plot experiments have been conducted in response to regulatory concerns. A few years ago there was evidence that the herbicide alachlor was contaminating groundwater in the state of Florida. In response to a probable ban on its use on peanuts in Florida, Daniel Colvin, an extension weed scientist with the University of Florida, obtained funding to conduct field-plot experiments to determine the differences in the efficacy and yield effects of alachlor and a number of alternative herbicides. His field experiments indicated that metolachlor performed as well as alachlor in controlling weeds. This was somewhat surprising since experience had indicated that metolachlor would not control certain weed species as well as alachlor, and expert opinion had predicted peanut yield losses of 13.2 percent in Alabama, Florida, and Georgia if metolachlor was substituted for alachlor. In 1990 the registrant of alachlor decided to withdraw the registration of that herbicide for use on peanuts in Florida on the basis of the state of Florida's concerns about groundwater contamination. Although Colvin's experiment did not play a role in deciding the fate of alachlor in Florida, it did provide extension weed scientists with information on how to advise farmers with regard to substitutes for alachlor. Some farmers have reported some problems since substituting metolachlor for alachlor. Colvin states that the reasons for these problems are not yet clear.
Does this mean that scientific field-plot data are not helpful in providing reliable estimates of yield losses needed in benefits analyses? There are some weaknesses in field-plot experimentation. Because of the small number of test plots and the natural variation in plot yields, large differences in yields must be observed between each pesticide treatment if these differences are to be considered statistically significant. In addition, field-plot tests do not reflect the full range of conditions under which a crop is grown. Yet experiments like Colvin's allow new combinations of herbicides to be tested that had not been tried before. Moreover, regulators at the EPA would agree that it is better to rely on experimental data than on expert judgment alone. However, plot experiments are not an adequate substitute for extensive knowledge of growers' experiences in the field.
Despite weaknesses in field-plot experimentation, it is better to rely on experimental data than on expert judgment alone.
Weaknesses in economic data
Economists working on benefits analyses are hampered by the inadequacy of available data on the revenues and costs of crop production. To value yield losses that may result when a substitute pesticide is used, economists must know what the demand for the affected crop is, as well as how the supply of that crop would shift if the pesticide currently used on it is banned. However, the number of USDA marketing studies on minor crops has decreased over the last several years. Moreover, economic models that can be used to examine the interaction between supply and demand are lacking for fruit and vegtable crops. In response to this situation, the EPA's Office of Policy Planning and Evaluation is funding the development of commodity models of the apple and processed tomato industries. These models will provide estimates of supply and demand functions for use in benefits analyses.
Some economists argue that pesticide usage should be evaluated with an entire crop production system in mind, rather than on a chemical-by-chemical basis, to more fully capture the complexities of crop production. Issues like managing or preventing resistance of pests to available pesticides have been mentioned in current analyses but not quantified as part of the benefits estimate. A commodity modeling approach would require substantial input from the agricultural scientific community. Complex relationships within the production system would have to be defined in such a way that the data input required for the analysis would not be an insurmountable barrier.
Is filling data gaps necessary?
The cry for more funding of research is heard in any situation in which information is lacking and important decisions must be made. But are the efforts needed to fill gaps in scientific and economic data really justified? Past benefits analyses have estimated that the costs to farmers of cancelling a pesticide registration are relatively small. Although these costs range from several million dollars to several hundred million dollars for more widely used pesticides, they are still small compared with the costs of other governmental regulations. For example, the total annual cost of complying with all federal air pollution laws in the United States is estimated to be $27 billion. Moreover, experience indicates that increased data input and more sophisticated analysis tend to drive the estimated benefits figure down, not up, for a given pesticide. Considering that much uncertainty exists in estimates of costs in terms of risk, is it necessary to be so precise in calculating costs to farmers?
Many think the effort would be worthwhile. They argue that if shortcomings in benefits analyses were corrected, the high cost to some farmers of some pesticide registration cancellations would be revealed. As a result, the EPA might not cancel some pesticide registrations that it normally would under current benefits analysis.
One shortcoming is the failure to separate the costs to users and nonusers of pesticides that would no longer be available. The net benefit from the use of a pesticide does not account for distributional effects in the costs of cancellation. A small subset of growers could conceivably incur the majority of the costs of cancellation since pest problems are not uniformly distributed through the entire producer population. A few pesticide users might suffer a substantial loss of income due to yield losses while others might benefit from the increased price of a commodity due to a decrease in the overall supply of that commodity. Thus cancellation of a pesticide registration may have a positive, if small, effect on nonusers, but a negative, often large, effect on some users. In some cases, regional inequities might result from cancellation of pesticides. Growers in certain regions of the country might not be able to compete with growers in other areas due to rising costs or increased yield losses from uncontrolled pests. Yet their losses could mean gains to growers in other regions as supply decreases and price increases.
Current benefits analyses do not separate the costs to users and nonusers of pesticides that would no longer be available; thus they cannot account for distributional effects in the costs of pesticide cancellation.
The net effect of pesticide cancellation on agricultural producers can be very small, as gains by some growers from an increase in price cancel out losses on the part of others. The impact on consumers also may be small—a price increase of a few cents for a certain commodity. But the removal of a pesticide could have drastic consequences for certain producers. In light of this outcome, efforts to supply scientific and economic data lacking in current benefits analyses appear warranted. It may be that the most appropriate funders for this type of analysis are regional commodity organizations that represent fruit and vegetable growers.
Leonard P. Gianessi is a fellow in the Quality of the Environment Division at RFF. Cynthia A. Puffer is a research assistant in the division. Since 1986, Gianessi and Puffer have been developing a national pesticide use database that is widely used by federal agencies in assessments of pesticide policies.
A version of this article appeared in print in the June 1991 issue of Resources magazine.
A version of this article appeared in print in the June 1991 issue of Resources magazine.
