Do GM crops really have higher yields? The answer is "No".
Brian John GM-Free Cymru
A strange article (1) has just been published in the pages of Nature Biotechnology (the very same journal which outraged the scientific community by pulling that "dummy proof" stunt on Irina Ermakova in 2007 and which can be counted upon only to publish positive news and research findings about GM crops and foods.) So we are justified, I think, in looking on any article contained within its pages with a degree of scepticism. Anyway, the article is a long-term data study (or meta-study) on GM and conventional crop yields, concentrating on commercialised lines planted in the state of Wisconsin.
Scepticism is justified when one examines the paper. It has probably been rewritten several times, and one gets the distinct impression that so dismal were the research findings about GM crop yields that the study team were forced to find something -- anything -- which they could use to give a positive spin. A silver lining had to be found, and they found it in the contorted argument that although GM crop yields are at worst very poor and at best nothing to celebrate, farmers growing GM varieties should be very grateful for a lower production risk (as measured by the variance, skewness and kurtosis of harvested yield). In their conclusions the authors state: "our results show how transgenic technology can improve farmers' ability to deal with a risky environment." Hmmm -- so how valuable is that going to be to farmers who are witnessing poor yields, yield drag, and increased herbicide weed resistance and the emergence of super-bugs in their GM maize fields? The authors claim that the benefit of reduced yield risk is "worth" a yield increase of 0.8 - 4.2 bushels per acre; but those figures are hypothetical and essentially meaningless.
The authors do not argue with the fact that GM crops are associated with increased agrichemical use, as shown by Charles Benbrook and others. They also show that over a 20-year period the yields of GM varieties as against their "conventional counterparts" varied between -12.22 bushels per acre and + 5.76 bushels per acre. Four of the groups of GM crops showed yield loss compared to conventional hybrids, and eight groups showed gains. But these gains were generally very small, at around 3 bushels per acre or less. Not worth writing home about. When the researchers looked at stacked GM events, they found some small yield gains and some losses. Quote: ".... with the exception of the ECB trait, we were surprised not to find strongly positive transgenic yield effects." More worryingly for the GM industry, they found significant gene interaction in stacked varieties, which had the overall effect of reducing yield. To quote: "...... the evidence of negative interaction effects among transgenes suggests that transgenic hybrids can perform more poorly than conventional hybrids." What the researchers are talking about here are unexpected occurrences of genetic variability and instability in stacked events, accompanied by a lack of uniformity -- effects which independent researchers have been warning about ever since stacked events were first mooted. This is a scientific and PR disaster for the GM corporations.
In addition to this strong empirical evidence of poor yields in GM crops, and a lack of stability in stacked events (also associated with poor yields) there are four further issues which the researchers fail to address:
1. In the studies reported in the paper, the non-GM crops used for comparative purposes are simply referred to as "conventional hybrid maize crops." Over 19,000 observations are reported on these crops, but we have no idea which varieties they were. The authors simply tell us that in all cases where GM varieties were measured, so were their "conventional benchmarks." This is simply not good enough. What were these "conventional benchmarks? Were they the non-GM isolines from which the GM crops were bred? Maybe, and maybe not. We need to know. For all we know, the researchers are simply reporting on GM trials conducted by or funded by the GM developers which were designed to demonstrate the safety and superiority of their new GM varieties -- as compared with carefully-chosen inferior conventional varieties. In other words, the comparisons might be fraudulent. And it is also possible that hidden within the 19,000 observations of 2,653 conventional hybrid varieties are the yield figures for many hybrids which were dismal failures and which were discontinued rather than being brought to the market. In other words, the mean yield figure of 186 bushels per acre for conventional hybrids of maize might be artificially depressed by the inclusion of many failed varieties, simply in order to demonstrate the superiority of the GM lines.
2. We have no idea how reliable the data may be which purportedly demonstrate a reduced risk of crop variability when GM crops are used. This is the key conclusion of the paper, and the one which the researchers have flagged up in their press release. According to Box 1 in the paper, 2653 conventional hybrids were tested, and 2096 GM hybrids. In Table 2, the estimated cost of risk for conventional hybrids is given as 6.36 bushels per acre, which is higher than the cost of risk for any of the GM varieties. The highest cost of risk for a GM hybrid is 5.58 bushels per acre, and the lowest is for one of the stacked hybrids, at 4.01 bushels per acre. So how reliable are these figures? Unfortunately, we have no idea, since we do not know whether the GM hybrids were tested in all cases side by side with their conventional isolines. Neither do we know whether there was a built-in bias in the tests which discriminated agains the conventional hybrids and in favour of the GM lines -- namely through the ongoing testing of failed conventional hybrids and the rejection of failed GM varieties prior to planting in field trials. As mentioned above, this is exactly the tactic used by the GM industry in order to make the data in assessment dossiers as convincing as possible to the regulators.
3. It is common knowledge that in GM breeding programmes, the GM multinationals and their research teams select the best conventional hybrids for their projects and then, if successful, remove those conventional hybrids from the seed catalogues. Then, if there is a demonstrable yield increase, this increase is put down - fraudulently - to the GM trait itself. This is another source of huge bias in this study. There is a distinct possibility that any yield increases that are observed in some -- and maybe many -- GM varieties have nothing whatsoever to do with the GM traits -- and this begs the following question: could these yield increases have been even greater if a conventional breeding programme had been followed on the most successful and adaptable conventional hybrids? (One might expect that, given that GM varieties are stressed by the disturbance of the plant genome and might fight that disruption by seeking to reject "intruder" genes rather than putting maximum energy into growth and yield.)
4. This study is a "meta study" incorporating research results from many different trials -- in different locations and in different environmental conditions. This brings some benefits -- with 19,000 observations -- in smoothing out variations in yield performance for both conventional hybrids and GM hybrids. But in no sense are we looking at a "controlled experiment", since there will have been many differences in management procedures between the conventional and GM crops studied. So there is no examination of a null hypothesis -- namely that there is no yield difference between a GM hybrid and a conventional hybrid. In most cases, we can assume that since plant breeders were seeking to develop (and promote and commercialise) GM varieties, management procedures will have been designed to maximise output in all cases -- while it may well be that management procedures for conventional or comparator crops might even have been designed to suppress yield increases! (This was exactly the criticism levelled -- with complete justification -- at the British "Farm Scale Evaluations" of GM crops more than a decade ago. The null hypothesis was not properly tested, and the trials were essentially corrupt.)
These are uncomfortable questions, and I hope that the research team responsible for this paper will be able to answer them. I cannot be sure, but I have a strong suspicion that there might have been deliberate bias in the paper, which should have been picked up in any responsible peer review process. But even with a built-in bias in favour of GM maize crops, the results are pretty devastating for the GM industry -- and they show that those who claim that GM crops have higher yields than conventional varieties are talking nonsense.
NOTE
1. Commercialized transgenic traits, maize productivity and yield risk by Guanming Shi, Jean-Paul Chavas & Joseph Lauer Nature Biotechnology, volume 31 number 2 FEBRUARY 2013, pp 111-114
Friday, 15 February 2013 17:32
NOTE: The new paper referred to: https://www.motherjones.com/files/maize_prod_nat-biotech_2013.pdf
The PDF can be downloaded.
The paper: Commercialized transgenic traits, maize productivity and yield risk by Guanming Shi, Jean-Paul Chavas & Joseph Lauer Nature Biotechnology, volume 31 number 2 FEBRUARY 2013, pp 111-114
Do GM crops really have higher yields? Tom Philpott Mother Jones, February 13 2013
http://www.motherjones.com/tom-philpott/2013/02/do-gmo-crops-have-lower-yields
According to the biotech industry, genetically modified (GM) crops are a boon to humanity because they allow farmers to "generate higher crop yields with fewer inputs," as the trade group Biotechnology Industry Organization (BIO) puts it on its web page.
Buoyed by such rhetoric, genetically modified seed giant Monsanto and its peers have managed to flood the corn, soybean, and cotton seed markets with two major traits: herbicide resistance and pesticide expression—giving plants the ability to, respectively, withstand regular lashings of particular herbicides and kill bugs with the toxic trait of Bacillus thuringiensis, or Bt.
Turns out, though, that both assertions in BIO's statement are highly questionable. Washington State University researcher Charles Benbrook has demonstrated that the net effect of GMOs in the United States has been an increase in use of toxic chemical inputs. Benbrook found that while the Bt trait has indeed allowed farmers to spray dramatically lower levels of insecticides, that effect has been more than outweighed the gusher of herbicides uncorked by Monsanto's Roundup Ready technology, as weeds have rapidly adapted resistance to regular doses of Monsanto's Rounup herbicide. http://www.motherjones.com/tom-philpott/2012/10/how-gmos-ramped-us-pesticide-use
And in a new paper (PDF) funded by the US Department of Agriculture, University of Wisconsin researchers have essentially negated the "more food" argument as well. The researchers looked at data from UW test plots that compared crop yields from various varieties of hybrid corn, some genetically modified and some not, between 1990 and 2010. While some GM varieties delivered small yield gains, others did not. Several even showed lower yields than non-GM counterparts. With the exception of one commonly used trait—a Bt type designed to kill the European corn borer—the authors conclude, "we were surprised not to find strongly positive transgenic yield effects." Both the glyphosate-tolerant (Roundup Ready) and the Bt trait for corn rootworm caused yields to drop.
Then there's the question of so-called "stacked-trait" crops—that is, say, corn engineered to contain multiple added genes—for example, Monsanto's "Smart Stax" product, which contains both herbicide-tolerant and pesticide-expressing genes. The authors detected what they call "gene interaction" in these crops—genes inserted into them interact with each other in ways that affect yield, often negatively. If multiple genes added to a variety didn't interact, "the [yield] effect of stacked genes would be equal to the sum of the corresponding single gene effects," the authors write. Instead, the stacked-trait crops were all over the map. "We found strong evidence of gene interactions among transgenic traits when they are stacked," they write. Most of those effects were negative—i.e., yield was reduced.
Overall, the report uncovers evidence of what is known as "yield drag"—the idea that manipulating the genome of a plant variety causes unintended changes in the way it grows, causing it to be less productive.
More encouragingly, the authors found that crop yields for GMO varieties are more stable year-to-year, that is, their yields fluctuate less than those of conventional varieties. As a result of this stabilizing effect, the authors conclude that "our results show how transgenic technology can improve farmers' ability to deal with a risky environment," especially given "current concerns about the effects of climate change on production uncertainty in agriculture."
Simply by planting Roundup Ready or Bt crops, they claim, farmers face less risk from yield fluctuations.
That may be true, but it's a long way from "generating higher crop yields with fewer inputs." And it's not clear at all that GMOs' marginal advantages over conventional seeds when it comes to risk mitigation trump the benefits offered by organic ag in that department. Here's how the authors of a major paper published in Nature last year put it:
"Soils managed with organic methods have shown better water-holding capacity and water infiltration rates and have produced higher yields than conventional systems under drought conditions and excessive rainfall." http://www.nature.com/nature/journal/v485/n7397/full/nature11069.html
PRESS RELEASE: VALUE OF MODIFIED CORN IS MORE IN REDUCING LOSSES THAN BOOSTING YIELDS
University of Wisconsin-Madison, USA Press Release, by Nicole Miller
http://www.news.wisc.edu/21505
14.02.2013
SUMMARY: "By analyzing two decades worth of corn yield data from Wisconsin, a team of UW-Madison researchers has quantified the impact that various popular transgenes have on grain yield and production risk compared to conventional corn. Their analysis, published online in a Nature Biotechnology correspondence article on Feb. 7, confirms the general understanding that the major benefit of genetically modified corn doesn‚t come from increasing yields in average or good years, but from reducing losses during bad ones. "For the first time we have an estimate of what genetically modified hybrids mean as far as value for the farmer," says UW-Madison and UW-Extension corn agronomist Joe Lauer, who led the study."
http://www.genet-info.org/information-services.html
VALUE OF MODIFIED CORN IS MORE IN REDUCING LOSSES THAN BOOSTING YIELDS
A team of UW-Madison researchers has found that yields of genetically modified corn vary widely compared to conventional crops, but production risk is reduced ˜ giving farmers more certainty about the yield levels they can expect.
While there‚s no end of robust and heated conversation about genetically modified foods, there are strikingly few comprehensive studies that put a numeric value on the costs and benefits.
Now, there‚s more to talk about.
By analyzing two decades worth of corn yield data from Wisconsin, a team of UW-Madison researchers has quantified the impact that various popular transgenes have on grain yield and production risk compared to conventional corn. Their analysis, published online in a Nature Biotechnology correspondence article on Feb. 7, confirms the general understanding that the major benefit of genetically modified (GM) corn doesn‚t come from increasing yields in average or good years, but from reducing losses during bad ones.
"For the first time we have an estimate of what genetically modified hybrids mean as far as value for the farmer," says UW-Madison and UW-Extension corn agronomist Joe Lauer, who led the study.
Lauer, who is also a UW-Extension corn agronomist, has been gathering corn yield and other data for the past 20 years as part of the Wisconsin Corn Hybrid Performance Trials, a project he directs. Each year, his team tests about 500 different hybrid corn varieties at more than a dozen sites around the state with the goal of providing unbiased performance comparisons of hybrid seed corn for the state‚s farmers. When GM hybrids became available in 1996, Lauer started including them in the trials.
"It‚s a long-term data set that documents one of the most dramatic revolutions in agriculture: the introduction of transgenic crops," says Lauer, who collaborated with UW-Madison agricultural economists Guanming Shi and Jean-Paul Chavas to conduct the statistical analysis, which considered grain yield and production risk separately.
Grain yield varied quite a bit among GM hybrids. While most transgenes boosted yields, a few significantly reduced production. At the positive end of the spectrum was the Bt for European Corn Borer (ECB) trait. Corn plants with this added transgene, which comes from the bacterium Bacillus thuringiensis, are protected from a damaging caterpillar. When the researchers combined the yield data from all of the ECB hybrids grown in the trials over the years, they found that the ECB plants out-yielded conventional hybrids by an average of more than six bushels per acre per year. GM hybrids with "stacked traits," or multiple transgenes, tended to have slightly improved yields ˜ an extra two or three bushels per acre. On the other hand, grain yields from hybrids with the Bt for Corn Rootworm (CRW) transgene trailed those of regular hybrids by a whopping 12 bushels per acre.
"A lot of farmers assume that if it‚s transgenic, it‚s great in terms of yield, but we know that putting a transgene into a corn hybrid isn‚t always successful," says Lauer. "You don‚t want to pay $75 dollars more per bag of seed to produce 12 bushels less per acre."
However, even among poor-performing groups of GM corn, there are individual varieties that perform quite well, Lauer notes. "It depends on how the transgene interacts with the underlying germplasm," he says. "My message to farmers is that every hybrid has to stand on its own."
Where transgenic corn clearly excels is in reducing production risk. The researchers found that every GM trait package ˜ whether single gene or stacked genes ˜ helped lower variability. For farmers, lower variability means lower risk, as it gives them more certainty about the yield levels they can expect.
This makes sense, explains Lauer. "The traits themselves don‚t add to yield. What they do is protect the yield, so any kind of yield advantage we can get from [a variety of hybrid corn] will be protected from pests," he says.
Lauer equates choosing GM to purchasing solid-performing, low-risk stocks. Just as safe stocks have relatively low volatility, yields from GM crops don‚t swing as wildly from year to year, and most important, their downswings aren‚t as deep.
GM crops help reduce downside risk by reducing losses in the event of disease, pests or drought. Economists Shi and Chavas estimated the risk reduction provided by modified corn to be equivalent to a yield increase ranging from 0.8 to 4.2 bushels per acre, depending on the variety.
Risk reduction associated with GM corn can add up to significant savings for farmers ˜ as much as $50,000 for 1,000 acres, calculates Lauer. "It depends on the price that farmers can receive for corn," he says.
The two factors quantified in this study, yield and production risk, are just a part of the overall picture about GM crops, says Lauer. He notes there are other quantifiable values, such as reduced pesticide use, as well as ongoing concerns about the safety and health of growing and eating genetically modified foods.
"There‚s a lot of concern about this biotechnology and how it‚s going to work down the road," says Lauer, "and yet farmers have embraced it and adopted it here in the U.S. because it reduces risk and the yield increases have been as good as, or some would argue a little better than, what we‚ve seen with regular hybrid corn."