Washington Legal Foundation Media Nosh
“Biotech Foods After the StarLink Corn Recall:
Is More Federal Regulation Needed?”
Remarks of Gregory Conko
Director of Food Safety Policy
I’d like to start by thanking Glenn Lammi and WLF for inviting me to speak here today about the future of biotechnology, and the appropriate framework for biotech regulation.
To judge whether or not the regulatory system for biotech foods is adequate, we first have to know both how the regulatory system is organized, and why it was set up that way. So let’s take a closer look at that structure.
Serious discussion about what sort of regulatory system was appropriate for biotech products began back in the early 1980s, when the science had progressed far enough for people to start thinking about bringing products to market. A general scientific consensus began to form, suggesting that the actual process of genetic engineering didn’t generate any unique risks.
A 1987 report published by the National Academy of Science’s National Research Council held that:
“The risks associated with the introduction of rDNA-engineered organisms are the same in kind as those associated with the introduction into the environment of unmodified organisms and organisms modified by other genetic techniques.”
The types of things that we’re concerned about with genetically engineered plants – potential weediness, biodiversity loss, introduction of endogenous toxins or allergens – had long been concerns for plants modified by more conventional means, and for plants transferred from one ecosystem to another.
The NRC report further found that “Assessment of the risks of introducing rDNA-engineered organisms into the environment should be based on the nature of the organism and the environment into which it will be introduced, not on the method by which it was modified.”
Both of these findings were repeated in two subsequent reports published by the National Research Council – one in 1989 and one in April of this year.
The 1989 report also found that “Recombinant DNA methodology makes it possible to introduce pieces of DNA … that can be defined in function and even in nucleotide sequence. With classical techniques of gene transfer … predicting the precise number or the traits that have been transferred is difficult, and we cannot always predict the phenotypic expression that will result. With organisms modified by molecular methods, we are in a better, IF NOT PERFECT, position to predict the phenotypic expression” (emphasis added).
What this means is that, in general, genetically engineered products will tend to be safer than conventional products, not more dangerous.
Consequently, when the Coordinated Framework for the Regulation of Biotechnology was finalized in 1986, it was based upon two principles: (1) that gene-spliced products didn’t need to be regulated differently than non-gene-spliced products, and (2) that the oversight ought to be concerned with the actual characteristics of the products, not with the mere fact that they were developed with recombinant DNA techniques.
Under the Coordinated Framework, USDA was charged with making sure that genetically engineered plants didn’t become environmental nuisances, because it was already in charge of making sure that non-genetically engineered plants didn’t become environmental nuisances.
EPA was charged with making sure that genetically engineered pest-protected plants weren’t harmful to the environment or to human health, because it was already in charge of making sure that other pest protection products weren’t harmful to the environment or to human health.
And the FDA was charged with making sure the genetically engineered food products were safe for human consumption, because it already had that role for non-genetically engineered foods.
But while the regulatory system is generally concerned with the characteristics of the products, only the FDA followed the general scientific thinking that genetically engineered and non-genetically engineered products should be regulated similarly.
For example, some varieties of canola, rapeseed, and soybean have been selectively bred to be herbicide tolerant, but only genetically engineered herbicide tolerant plants are regulated by the USDA. Other plants, like kidney beans and potatoes, are known to have high levels of endogenous pest resistant toxins that pose some small risk to human health, but only genetically-engineered pest resistant plants are regulated by the EPA.
Dozens of new plant varieties produced through hybridization and other, non-genetic engineering methods enter the market every year without any government regulation. But every genetically engineered product on the market has been tested and re-tested, going through several hundred – if not several thousand – different tests to ensure environmental and human health protection.
Which brings us to the EPA and StarLink corn.
As we heard earlier, StarLink corn has been approved by the USDA for commercial scale planting in the United States. And it has been approved by the EPA both for commercial scale planting and for use as animal feed and industrial uses.
Furthermore, StarLink passed all but one of EPA’s requisite tests for human health protection.
EPA was able to conclude that there was no toxicity concern with the Cry9C protein. But it was not convinced that it was not an allergen. In reality, what this means is that EPA decided that StarLink was guilty until proven innocent. And it set the evidentiary standard for innocence pretty darn high.
First, the Cry9C protein doesn’t share the same structure with known allergens. Second, commercial use of other Bt proteins in the Cry9 Class has shown no reason to suspect the Cry9C protein might be an allergen. Third, unlike with known allergenic proteins, where the allergen makes up a sizeable percentage of the food, there is a very low percentage of the Cry9C protein present in the StarLink corn grain. And fourth, in vitro tests on the blood serum from individuals who are allergic to other known food allergens showed no allergic response to the Cry9C protein.
The only test StarLink didn’t pass in full, was the gastric digestibility study. And even here, the results were fairly positive. In digestibility studies conducted under simulated gastric conditions, the Cry9C protein did digest within the typical gastric emptying time, within the range of normal gastric acidity. The only problem that arose was when the tested acidity was slightly lower than the normal range.
Given all this data, it would have been reasonable for the EPA to conclude that the Cry9C protein was highly unlikely to be a human allergen, and therefore, to approve StarLink corn for human consumption. And, at least until the recall began, it was widely assumed that EPA would approve it at some point in the not too distant future.
So, the question we ought to be asking here is not whether federal oversight of the biotech industry might be too lax, but rather, whether it might be too strict?
The EPA’s failure to approve StarLink corn for human consumption didn’t save anyone from any negative effects. There is no substantive evidence whatever that StarLink corn is now or has been a threat to human health.
On the other hand, EPA’s action may have caused substantial harm to its own reputation – and possibly to the reputation of the FDA. And it will almost definitely cause damage to the public confidence in genetically engineered foods, and to provide aid and comfort to the anti-biotech movement. All for no apparent reason other than EPA’s desire to “err on the side of caution.”
However, if it turns out that this fiasco does end up damaging public confidence in biotechnology, EPA may have caused more harm than good. Why? Because, on balance, genetically engineered products tend to be more safe, not less safe than conventional products.
Unfortunately, it’s problems like these that tend to make regulatory agencies over-react even more in the future. From the perspective of a standard public choice analysis, it’s easy to see why.
When it comes to health and safety regulation, regulatory agencies can make two kinds of mistakes. On the one hand, they can approve a product that later turns out to be dangerous – what statisticians call Type I error. This is the kind of problem we’re all familiar with, and consequently, the kind of mistake the public hopes regulatory agencies never make.
On the other hand, regulatory agencies can deny approval to a product that actually is safe and beneficial. This is what statisticians describe as a Type II error. When ordinary people think of regulatory mistakes, this isn’t usually what comes to mind. But it is important for us to recognize that Type II errors also cause harm to public health.
To dramatize this a little more clearly, let me put it in terms of another type of product, where the tradeoff is more apparent: a new medicine. When the FDA approves a new drug, it’s usually announced with great fanfare in lots of newspapers that the medicine will save the lives of so many people every year. But if the FDA fails to approve a drug that could save lives, real people stay sick longer, and many of them die.
Ideally, we’d like to have a regulatory system that minimizes both Type I and Type II errors – and one that lets individual consumers make their own choices about the appropriate tradeoffs where that’s possible.
In the real world, however, regulatory agencies are more prone to making Type II errors, because the political incentives they face drive them to avoid Type I mistakes. If a product they approve turns out to be even a little problematic, the agency is confronted by adverse media attention, lots of irate public interest groups, and usually are dragged in front of a congressional hearing, no matter how much benefit the product had in the meantime. Exactly what’s happened in the StarLink case.
However, if a regulatory agency fails to approve a safe and beneficial product, what happens? Usually nothing. There typically are no lengthy, brow-furrowing articles in The New York Times or The Washington Post. No letters of protest from Consumers Union or Public Citizen. And, most importantly, no congressional oversight hearings demanding to know why the product wasn’t approved. (Although the gay community’s protest movement regarding the slow approval of AIDS drugs provides one rare counter example.)
Certainly, it’s a lot less apparent how the public will be harmed by having one of half-a-dozen different varieties of Bt corn taken off the market. I can’t claim that the difference between having StarLink and not having StarLink is that great – though agriculture experts may be in a better position to argue that it is.
The real problem arises from the public and political reaction to EPA’s botched handling of the StarLink application. Having put it on the market only for limited use, and then having taken it off the market altogether in a cloud of uncertainty, could send a signal to the American public that (1) biotech products may be more dangerous than we originally thought, and (2) that government regulation of them is woefully inadequate. But neither of these perceptions is true.
If the StarLink fiasco turns out to be used as justification for ramping up regulation, that means that even fewer of the next generation of products will make it on to the market in a timely fashion. It also means that consumers will end up paying more for those that do.
Ultimately, the higher prices we’ll end up paying as a result of extraneous and unnecessary regulation will most hurt those marginal consumers in the US – and especially in developing countries – who are most in need of the safer, more nutritious, and more productive crop varieties that agricultural biotechnology will deliver in the next decade.