Genetically Modified Organisms, or GMOs, are a hot topic of controversy in the public arena, just like vaccines, or climate change. Public defenders of science, what are often referred to as “skeptics” (among whom I count myself), have taken sides on all these issues, trying to do their best to bring some sanity and evidence-based clarity to bear upon them. The problem is that while it is beyond doubt that vaccines do not, in fact, cause autism; and it is also pretty darn clear that human beings do, at the least in part, cause global warming; the thing about GMOs is slightly more complex. Make that a lot more complex.
The New York Times gets into the fray
The latest round got started with an in-depth article published by the New York Times and penned by Danny Hakim. It begins by acknowledging that to date there is no evidence at all that GMO technology is directly dangerous to human health, contra the standard — entirely unsubstantiated — position of many environmentalists. What, then, is the problem? Hakim suggests that there is another, hitherto largely ignored, issue: contra to industry hype, the use of genetically modified organisms has brought neither a substantial increase in yield, nor a significant reduction in the use of pesticides.
Well, that ought to be a matter pretty easy to settle empirically, yes? Not exactly. It is difficult to make sensible comparisons, because there haven’t been large scale controlled experiments set up to do so (other than the small trials conducted by the industry itself). So Hakim hit on an interesting idea: noting that the US and Canada opted for GMO technology while Europe rejected it, and noting also that the two blocks are both large, Western, advanced technological societies, why not compare them in terms of GMO and non-GMO yield and pesticide use? While far from perfect (again, it’s not a controlled experiment, more like a “natural” one, made possible by circumstances, not planning), such a comparison could prove informative.
The New York Times piece presents the results separately for the two issues: yield and pesticide use. Beginning then with the former, three graphs published in the article summarize the findings for rapeseed, corn, and sugar beet. In both the cases of rapeseed and corn, the lines plotting increased yield in North America vs Western Europe are exactly parallel, with no inflection upwards of the North American data after the introduction of GMOs, as it would be expected. For sugar beet, European (non-GMO) yield has actually grown faster than its North American counterpart.
Hakim also mentions independent reports bearing on the same issue: “[a] National Academy of Sciences report found that ‘there was little evidence’ that the introduction of genetically modified crops in the United States had led to yield gains.”
What about pesticides? Here the NYT article compares the US with France, the latter being Europe’s largest producer. France’s use of pesticides has fallen significantly in all categories, while the situation in the US is, again, more complicated: according to data from the US Geological Survey, the use of chemicals aimed at fungi and insects has fallen (like in France), but that of herbicides has increased by about 21% (unlike what happened in France).
Why is this a problem? Well, because pesticides enter our food chain, and are known to be highly deleterious to human health. The NYT piece reports a comment by David Bellinger, a professor at the Harvard University School of Public Health, “whose research has attributed the loss of nearly 17 million I.Q. points among American children 5 years old and under to one class of insecticides.” (I’m not sure it makes a lot of sense to cumulate I.Q. points within a population, but the figure would be troublesome even if more properly expressed as average loss per child.)
Hakim — who did get an entirely perfunctory and obviously unconvincing response from Monsanto — went on to point out a possible nefarious background: “The industry is winning on both ends — because the same companies make and sell both the genetically modified plants and the poisons. Driven by these sales, the combined market capitalizations of Monsanto, the largest seed company, and Syngenta, the Swiss pesticide giant, have grown more than sixfold in the last decade and a half. The two companies are separately involved in merger agreements that would lift their new combined values to more than $100 billion each.” (Bear this detail in mind, it will be challenged below.)
The NYT is careful to remind its readers that Monsanto did, in fact, promise a reduction in herbicides as a result of adopting its technology, back in 1994, as reported then by the Los Angeles Times. What happened instead was predictable, says the article: “The whole point of engineering bug-resistant plants ‘was to reduce insecticide use, and it did,’ said Joseph Kovach, a retired Ohio State University researcher who studied the environmental risks of pesticides. But the goal of herbicide-resistant seeds was to ‘sell more product,’ he said — more herbicide.”
Another predictable outcome — from a straightforward evolutionary biological perspective — is that weeds are becoming resistant to Roundup, the main herbicide liberally sprinkled on GMO crops that can withstand the poison. Monsanto is developing an alternative, which hasn’t been approved for commercial use yet, though that hasn’t stopped the company from beginning to sell to farmers seeds that are resistant to it.
Hakim continues his article by directly comparing the case stories of two farmers, Bo Stone in North Carolina, and Arnaud Rousseau in France. The first one uses GMOs, the second one doesn’t. The two plant identically looking corn seeds, but Stone spends $153 per 50,000-seed bag, while Rousseau spends only $85. Why does Stone continue to use GMO seeds, then? Because once you bought into the system it isn’t easy to do without, as genetic modifications cannot be purchased a la carte, but only in whatever combination the biotech companies sell them. Interestingly, Rousseau would like to have access to the same technology that Stone can use, because he has the feeling that his work would be easier (both farmers have seen increases in their yield over the years).
The issue of yield, however, does not concern just the profits of individuals farmers. As the NYT says: “With the world’s population expected to reach nearly 10 billion by 2050, Monsanto has long held out its products as a way ‘to help meet the food demands of these added billions,’ as it said in a 1995 statement. That remains an industry mantra.” (This is a second point to keep in mind before we turn to the critical analysis.) Indeed, the NYT comparison of US regional data with those from seven European nations (including the two largest producers, France and Germany) shows no broad yield advantage of GMO crops. Jack Heinemann, from the University of Canterbury in New Zealand, who has used UN data to make the comparison, is quoted as saying “Western Europe hasn’t been penalized in any way for not making genetic engineering one of its biotechnology choices.”
Steve Novella responds
My friend and fellow skeptic Steve Novella immediately published a response to the New York Time article, blasting it as “a hack piece with a biased narrative.” I knew that Steve had written before in defense of GMO technology, though I thought his defense was aimed at the sort of unfounded fears about food safety that the NYT begins its article dismissing. I also know Steve as a highly informed, seasoned skeptic, who is not prone to outbursts, so I was curious to dig more deeply into his unusually worded response. In this case, and with all due respect to Steve (I do respect him, a lot), I was disappointed.
The very first sentence of Steve’s piece is a handbook example of poisoning the well (a logical fallacy): “It is unfortunate that so many journalists begin with a narrative and then back fill the facts and points necessary to tell their narrative.” I wonder how he knows this, since he presumably does not have access to Hakim’s mind and intentions, nor does the NYT have a history of anti-GMO crusading.
Steve’s first actual target is Hakim’s claim that GMO technology has failed to deliver on two of its crucial promises: increased yield and lower use of herbicides. “Wrong. This is a very common anti-GMO trope. Genetic modification is a tool, and is not inherently tied in any way to the two currently most common applications, herbicide resistance and pest resistance.” This makes little sense in light of the fact — as clearly substantiated in the NYT article — that GMO companies made precisely those promises, very clearly, and very early on, presumably in order to gain both public support and regulatory approval. It is therefore perfectly legitimate to ask where we stand on both counts about twenty years after the introduction of such technologies.
Steve continues: “The promise of genetic modification, rather, is that it provides a tool for agricultural scientists to make more rapid and more specific changes to crop cultivars. The technology has completely fulfilled that promise. The technology works, it is safe with no demonstrable inherent risks.” Again, this is a rather strange way to put it. No, the promise of GMOs is not just to provide agricultural scientists with faster ways to change crops, it is to make life better for the farmers and the public at large. And yes, the technology works when “works” is narrowly defined, but whether it has delivered on its broader, and most important, promises, is very much at issue. (Keep in mind once again, that we are not talking about food safety here, since Hakim, Novella and I all agree that, so far, there is no issue there.)
Steve’s criticism gains more traction a bit later on, when he mentions that the NYT has used UN data (and other sources, actually, including the US Geological Survey) to make its own comparisons. The NYT article, Steve correctly says, is not a peer reviewed piece. Well, yes, but investigative journalism means just that: to gather evidence and analyze it critically. One can quibble with the specific methods used, but it is a questionable move to reject some conclusions on the only ground that a newspaper article is not a peer reviewed technical paper.
At any rate, Steve immediately provides his readers with two actual peer reviewed papers to counter the NYT claims. The first one, a meta-analysis of a large number of studies, published in PLOS One, found that “on average, GM technology adoption has reduced chemical pesticide use by 37%, increased crop yields by 22%, and increased farmer profits by 68%. Yield gains and pesticide reductions are larger for insect-resistant crops than for herbicide-tolerant crops. Yield and profit gains are higher in developing countries than in developed countries.” These conclusions do seem in direct contradiction of the NYT ones, but of course we need to remember that they are actually about different things, since the NYT comparison was only for the semi-controlled “experiment” provided by North America and Western Europe, while the PLOS One paper mixes developed and developing countries.
[As a side note, even the paper cited by Steve agrees that yield gains have been far lower for developed countries. This may mean, for instance, that it makes more sense for developing rather than for developed countries to switch to GMOs, given the other environmental and labor practice issues involved. But this sort of middle-of-the-road approach wouldn’t satisfy either supporters or critics of the technology, of course.]
Moreover, a superficial analysis of the PLOS One paper reveals a few problems of its own. (I told you: it’s complicated!) If you read the Materials and Methods section, you discover that the majority of the sources on which the meta-analysis was carried out are not peer reviewed. This is a big problem by Steve’s own standards, and indeed a rather unusual modus operandi on the part of the authors. Their justification for this unorthodox move isn’t particularly convincing: they claim that, first, it is actually difficult to separate peer reviewed papers from much of the “grey” literature, like non-peer reviewed contributions to conferences, governmental and non-governmental reports, and the like. Second, they say that “studies without peer review also influence the public and policy debate on GM crops; ignoring them completely would be short-sighted.”
The first reason is news to me. It is actually very clear what the difference is between peer and non-peer reviewed publications, it just takes a bit more effort to double check individual entries. But it is the second purported reason that is truly astonishing: sure, non-peer reviewed sources probably do have more sway on public opinion than technical papers, but the goal of the meta-review was not to assess influences on public opinion, it was to uncover empirical data concerning research questions. None of the above invalidates the PLOS One study, but it does mean that one ought to proceed with caution.
The second peer reviewed article cited by Steve appeared in GM Crops & Food, and concludes that there have been significant gains for farmers in the adoption of GMOs, mostly as a result of increased yield. Again, though, this isn’t really comparable to the NYT data, because of the issue mentioned above: the latter is a somewhat controlled comparison of similar economies that have and have not adopted GMOs, while the former is based on data from a mix of countries with very different economic and political situations.
Moreover, a bit of digging around reveals a few potential anomalies in this second paper as well. First off, the journal itself publishes a mix of research and policy papers, which is somewhat odd for a scientific publication. It looks more like an advocacy journal than a standard academic one. Second, the specific paper cited by Steve was funded by, you guessed it!, Monsanto. This in and of itself doesn’t disqualify it as a source, of course, but it is a pretty well established fact that industry-funded studies (for instance, in the pharmaceutical field) tend to find about one third more results positive for the industry that shelled the money, when compared to government-funded studies. Given the huge public controversy, and the large profits, involved in the GMO debate, one might, once more, want to be cautious.
Next, somewhat puzzlingly, Steve claims that “Hakim reproduces a common anti-GMO trope to combine Bt crops with herbicide resistant crops — two completely different applications.” But in fact Hakim does no such thing. The NYT article very clearly distinguishes the two, acknowledges that the use of insecticides has gone down in post-GMO USA, and then focuses on herbicides.
About the latter, Steve writes: “Hakim fails to point out that while glyphosate use has increased, it has replaced applications of much more toxic herbicides. If you measure only tons of herbicide you miss the point that overall herbicide toxicity has dramatically decreased, because glyphosate (despite claims of anti-GMO activists) is a very benign chemical.” Well, again, not exactly. The NYT article actually presents a very clear graph where the different kinds of herbicides are distinguished. Moreover, glyphosate may very well be “very benign,” but it is also by far the smallest contributor to the total in the graph.
Steve goes on accusing Hakim of a rather important imprecision: “Monsanto” is actually two companies, he says, one dealing with the chemicals, the other with the seeds, thus implying that the conflict of interest highlighted by Hakim is unfounded. But if you go on Monsanto’s own web site you will find a different, and more complicated (once more!) story, distinguishing between an “original” and a “new” Monsanto. The original Monsanto merged with Pharmacia back in 2000, and in the same year the new Monsanto was incorporated as a stand-alone company, which, however, is actually a subsidiary of the chemical one. So, it doesn’t look to me like the two Monsantos are really that distinct from each other, the maneuver probably having been carried out for tax benefits, shareholders short term profits, or both, as it is standard (and legal) for multinational corporations.
The last part of Steve’s critique of Hakim’s article begins thus: “I also reject the entire deeper premise of the article, that we should judge a technology by its current and early applications.” Well, much depends, of course, on what “early” means. Is twenty years not enough, especially when many billions of dollars and people’s livelihood are at stake?
Steve uses a couple of analogies to make his point, both flawed, I think. The first one concerns the human genome project, which he points out hasn’t cured any major human disease, contra to the hype that helped promote the original effort. Right, but the human genome project was largely an exercise in basic science, and it was funded as such, not because it promised to cure cancer (the National Science Foundation, for instance, clearly separates the actual immediate scientific merits that justify funding a given project from the possible “broader impact” of the same project). Similarly, Steve is right when he says that GMO technology per se has been successful as basic science, i.e., as a technique for experimental use. But that’s not what the controversy is about. The discussion concerns the use that the industry has made of that technology, therefore judgment very much needs to include a consideration of whether what the industry promised has been delivered or not.
The second example brought up by Steve is that of personal computers: “One might have argued (and many people did) even into the 1990s that computers did not improve productivity. They did not reduce use of paper (remember that one – the paperless office).” No, they didn’t, but I adopted personal computers at around the time Steve did, and they delivered immediately exactly what they promised: they helped me run some nice statistical analyses for my thesis, to produce graphics to accompany the same thesis, and to write the thesis to begin with. Oh, and video games, lots of video games. Even allowing for the failure of the paperless office, it is a stretch to argue that we had to wait more than two decades for personal computing to deliver. But that’s the relevant timeframe for GMOs.
The broader point made by Steve is that “How we implement [GMO] technology is an entirely separate question [from whether it allows us to develop new cultivars quickly and precisely], as separate as hardware and software.” No, it isn’t. Again, the promise of the technology lies in the broad scale human applications, not just in making scientists’ job easier. Let us not pretend otherwise.
So, what do I think of the entire controversy? I don’t know. I spent hours researching this single blog post, and the more one digs into the literature the more confusing things become. And I’m a professional biologist. Who worked on plants. For a couple of years I even worked on agricultural species, and for much of my career on invasive weeds, so I know first hand about herbicides, pesticides and all the rest. One thing, however, I am fairly confident of: the NYT article was not a “hack piece with a biased narrative.” If it needs to be criticized, by all means let’s do it, but we need to have good arguments and carefully laid out evidence, and Steve’s piece simply does not provide enough of either.
P.S.: There were several other rebuttals to Hakim’s piece, perhaps the most prominent of which was published by Andrew Kniss on his blog. I read that too, as well as a technical paper by the same author that is linked there. Kniss is a legitimate researcher in the field, and his arguments are complex and well worth considering. However, the technical paper he refers to is actually non-peer reviewed, and his own sources of funding (which are not reported on the paper itself, but are found under the “additional information” section of the web site) include the following: Bayer CropScience, Dow AgroSciences, DuPont, Monsanto, and Syngenta. That is, a who’s who of the GMO industry. Please realize that to note this isn’t an example of the genetic fallacy, it is a simple cautionary application of common sense. I noticed before that skeptics’ quick reach for the list of so called informal fallacies to throw at their critics is misguided and intellectually lazy; such “fallacies” often turn out to be reasonable heuristics to arrive at a first assessment of a complex situation. (See this paper that I co-authored with Maarten Boudry and Fabio Paglieri.)