Regulating Pesticides -- an Impossible Job

It's an old joke that one should never look too closely at how sausage is made or how legislation is passed. That's also true for how pesticides are regulated.

I began to look too closely when three of my colleagues were working on a pesticide data base. I watched them, asked questions, and learned how the Environmental Protection Agency (EPA) regulates the amount of pesticide in our food.

For those who haven't the stomach to delve into the details, here's the main conclusion. No one really knows how much of the 1.3 million tons of pesticides used annually by U.S. farmers shows up in our food. No one knows what a safe level of pesticide ingestion might be. The EPA is working in a haze of statistical and scientific uncertainty.

Now, for you brave ones who want to look at the innards, here's how the process works.

A company wanting to market a new pesticide must test it by feeding increasing amounts of it to laboratory animals. The toxicity of the pesticide is measured by how much it takes to kill 50% of the animals outright. Other tests of internal and external exposure determine whether the chemical causes cancers or birth defects.

The test data are submitted to the EPA. They are scientific observations, fairly incontestable as long as the test lab is honest. The EPA must then convert these animal results to permissible human exposures and farm application procedures. That requires a lot of assumptions.

First the EPA determines a No Observable Effects Level (N.O.E.L.) -- the daily dose that produces no visible ill effect on test animals. Assumptions: 1) no observable effect means no actual effect; 2) there is some safe level of exposure to this chemical (except for carcinogens, where the assumption is that no exposure is safe -- an assumption hotly contested by the chemical companies.)

Then comes the problem of converting data from animals to human beings. Assumption: humans are 10 to 100 times more sensitive to pesticides than are test animals. The EPA takes the N.O.E.L., divides it by some number between 10 and 100, and calls that the Acceptable Daily Intake (A.D.I.).

The EPA determines, for example, an A.D.I. of .05 milligrams of Captan per kilogram of body weight per day. Captan is a fungicide used on apples, tomatoes, peas, sweet corn, onions, beans, squash, carrots, and oranges. How can the EPA know how much of each of those foods you eat every day, so it can set permissible residue levels for each crop that won't add up to more Captan than you should be eating?

Assumption: we all eat the average American diet. The EPA has survey data on what a sampling of Americans eat. It averages us all together, you and me and your finicky two-year-old and my vegetarian grandfather.

Given the average diet it sets tolerances for each chemical and crop -- maximum permissible amounts of Captan on apples and beans. Then it lays down field procedures specifying how much Captan can be used on each crop, and what period must elapse between application and harvest, to assure that crops will meet tolerance.

A lot more assumptions creep in here. 1) The public's only exposure to a pesticide comes through food ingestion. 2) When several pesticides are present in the diet together, they do not do more damage than they would separately. 3) Farmers obey field guidelines. 4) Obeying field guidelines results in tolerances being met.

Crop samples are taken by the Department of Agriculture to test for compliance. Some of the samples do have pesticide levels over tolerance. Many crops have residues far below tolerance, and some farmers use no pesticides at all.

My colleagues did an experiment to test just one of this long string of assumptions. They calculated the pesticide exposures of people who do not happen to eat the average national diet. They used data on the actual diets of people of different ages, sexes, and ethnic groups. Their findings were especially striking for children.

According to their calculations, children 1-6 years old receive more than the acceptable daily intake of 10 out of 18 fungicides they investigated. Some of the calculated overexposure levels were enormous -- from 5 to 200 times the acceptable daily intake. (Calculations were made for only 18 fungicides out of 600 possible pesticides.)

Does this mean our children are being poisoned? Probably not, but the EPA can't be sure of that. All anyone can say is that, assuming that foods contain pesticides at the tolerance level, and assuming the EPA's definition of safe exposure, the average child eating an average child's diet, as revealed by one nutrition survey, is overexposed to 10 of the 18 pesticides investigated.

The EPA has no choice but to operate on assumptions. No one fully understands the effects of pesticides on the human body. The EPA Office of Pesticide Programs has a finite (and shrinking) budget. The task of regulating 600 chemicals on 376 crops in the hands of 3 million farmers who feed 240 million people is far beyond what any agency, no matter how well-funded, can handle.

My personal conclusion from this exercise in looking too closely is that, if we are to go on using pesticides, much more research and monitoring needs to be done. It should be paid for by the chemical companies who are profiting from this massive experiment in which we are all the white rats. And in the meantime, I'll get as much food as I can from my own garden and from farmers whose methods I know and trust.

Copyright Sustainability Institute
This article from The Donella Meadows Archive is available for use in research, teaching, and private study. For other uses, please contact Sustainability Institute, 3 Linden Road, Hartland, VT 05048, (802) 436-1277.

Read More

Who Sways the USDA on GMO Approvals?

 

Many corporations have gotten good at pulling the levers of government to tilt the odds in their favor, weakening regulations or securing perks, justified or not, to further their business interests. Economists use the term "regulatory capture" to describe the phenomenon whereby regulatory agencies serving the public instead end up advancing the interests of the companies they regulate. The main way companies accomplish this, economists theorize, is through lobbying and campaign contributions that convince legislators to pass laws in their favor.

Once those laws are passed, however, it's less clear how companies sway the regulatory agencies that enforce them, which are more isolated from the direct effects of money or persuasion.

"If a company can get enough farmers to support the product and they write letters, then the USDA is going to listen."

"Traditional theories of regulatory capture cannot be used the same on agencies," contends Shon R. Hiatt, an assistant professor at Harvard Business School. "There are a lot of checks and balances and firewalls in place."

So how are these agencies influenced?

Hiatt, who grew up on a dairy farm in Idaho, began asking that question through research on the controversial issue of genetically modified organisms (GMOs), agricultural products that are genetically altered to increase yield, incorporate pesticide properties, or exhibit other beneficial qualities. (Calgene's Flavr Savr tomato was the first genetically modified product to come to market, in 1992.) However, the organisms also potentially carry health and environmental risks. After reading about these dangers, Hiatt wondered how the US Department of Agriculture (USDA) decides which GMOs to approve—and how agribusiness influences the process.

Traditional theories break down

As Hiatt began investigating, he found that traditional theories of capture such as lobbying and campaign contributions had little effect on whether any particular GMO was approved. Even more direct means of influence such as scientific articles funded by industry or letters written by industry-friendly congresspeople were equally ineffective.

What did seem to affect the approval process, however, was the influence of third-party groups separate from Congress and industry, to which the department looked to justify its decisions.

We may think the primary goal of agencies such as the USDA is to protect public health and safety; based on previous economic theory, however, Hiatt started with a different assumption—the primary goal of an agency is really to protect its own legitimacy. After all, it's the perception of an agency's effectiveness by Congress and the White House that will determine its budget and the career trajectory of its top officials. Of course, there is an overlap between the appearance of doing a good job and actually doing one. "If the USDA weren't doing its job, it would have very little legitimacy," says Hiatt. But that subtle difference in perspective also has the potential to distort the agency's reliance on pure science in its approval of GMOs.

In his working paper "Lords of the Harvest: Third-Party Signaling and Regulatory Approval of Genetically Modified Organisms, written with Sangchan Park, an assistant professor at the National University of Singapore, Hiatt identifies two types of legitimacy important to the USDA. The first, "consequential" legitimacy, is the perception that the process produces effective results; the second, "procedural" legitimacy, is the perception that it is fairly following the rules of the process.

In both cases, the researchers found that the department looked to outside stakeholders in order to establish that legitimacy.

In the case of consequential legitimacy, Hiatt and Park found a strong influence of farmers associations, such as the Iowa Soybean Association or the Kansas Corn Growers Association, which have the power to judge whether the GMOs are performing as intended without side effects. While these groups might have some industry members, they are separate from the agribusiness companies that are introducing GMOs. In cases where they supported a particular organism, there was an 84 percent increase in the likelihood of approval.

"If a company can get enough farmers to support the product and they write letters, then the USDA is going to listen to that and say, 'We have to keep our stakeholders happy,' " says Hiatt.

Other agencies influence approvals

In the case of procedural legitimacy, the researchers found a strong influence from an unlikely source—the USDA's sister agency, the Food and Drug Administration. In the process of approving GMOs, companies have the option of consulting with the FDA to design nutritional labels for their products, earning a certificate of approval when they address FDA concerns. "They get these consultations and they are somewhat meaningless—they have little to do with the USDA approval process," says Hiatt.

Regardless of that fact, however, Hiatt and Park found that a positive endorsement by the FDA had a huge effect on USDA approval, increasing the likelihood by 157 percent.

Hiatt hypothesizes that in addition to receiving nutritional information on the GMOs, getting the green light from another agency might help insulate the department from criticism. "The USDA could be looking for a scapegoat," he speculates. "A positive signal from a fellow bureaucratic actor could diffuse the blame and provide political cover were the department to approve a faulty product."

These effects seemed to be even higher during instances where there was significant controversy or uncertainty. In cases where there were protests by activists over a particular GMO, the researchers found that the overall percentage of approvals went down, but the degree to which a positive endorsement by farmers associations increased the likelihood of approval by117 percent.

Results were even starker during a congressional election year, in which a heightened political environment presumably casts more scrutiny on agency decisions.

In those cases, farmers associations' influence increased the likelihood of approval by a whopping 400 percent. In addition to increasing the likelihood of approval, says Hiatt, third-party endorsements shorten the approval period. With farmers' approval, agricultural companies shaved about 162 days off the average approval time; with FDA consultation, they cut it down by about 257 days. That can translate to big bucks for companies.

"The average seed company earns about $2 million per day of revenue for high- selling GMO crops such as soybeans," notes Hiatt. "That's a substantial amount."

He stresses that these findings only concern the USDA, and only GMO approval; more research is necessary to determine whether the effect of third-party stakeholders on the USDA has an effect on other agencies or other policy issues. Conceivably, the same findings could hold true for other agencies: for example, the influence of doctors associations on the FDA drug approval process or consumer bureaus on rate increases by public utilities commissions.

Regulators must recognize influencers

To the degree these third-party stakeholders do have influence, it complicates the traditional models of regulatory capture. On the one hand, it is perhaps a relief for democracy if companies don't have such direct influence on the process. On the other, it opens up the possibility that firms could capture these third-party actors instead—for example, drug companies influencing doctors with incentives for prescribing drugs or sponsoring medical conferences.

Perhaps the larger lessons from Hiatt and Park's research, however, concern regulators themselves. It's important that agencies such as the USDA realize their susceptibility to these outside influences, less they short-circuit their reliance on scientific procedures. That is just what happened with the approval of GMO alfalfa and sugar beets: both received positive signals from farmers and the FDA, and were approved by the USDA. But environmental groups protested that these products were approved without a full environmental review, successfully suing to take them off the market. (Alfalfa was subsequently reapproved after a multi-year delay. Sugar beets are still pending approval.)

"Regulators need to be aware of the influence they are putting on these stakeholders and other regulatory agencies," says Hiatt. "In those cases, it's even more important they follow the same scientific procedures they usually do. If they find themselves cutting corners, they could run into problems."

About the author

Boston-based writer Michael Blanding is a fellow at the Edmond J. Safra Center for Ethics at Harvard University and author of The Coke Machine: The Dirty Truth Behind the World's Favorite Soft Drink.

Read More