Despite the fact that the global area of GM crops has grown rapidly over the past decade and now encompasses more than 170 million hectares it is still questioned whether current applications of genetic engineering in agriculture are beneficial to farmers or not. While several reviews and meta-analyses on the impact of GM crops exist, the evidence is still not regarded as conclusive by many. A new meta-analysis of the agronomic and economic impacts of GMOs has appeared in the journal PLOS ONE. It summarizes the findings of 147 original studies published until March 2014. Besides being the most up to date, this meta-analysis has two other features that make it an important addition to the existing body of meta-analyses and review articles: First, it includes not only studies from peer-reviewed journals but also those that are not (grey literature), thus providing a broader perspective. Second, the authors investigated the reasons why different studies found different results, providing several interesting insights.
The work was funded exclusively with public money, in part from the German Federal Ministry of Economic Cooperation and Development (BMZ) and the European Union’s Seventh Framework Programme (FP7/2007-2011). The authors are Wilhelm Klümper, a PhD student at the Department of Agricultural Economics and Rural Development at the University of Göttingen, Germany, and Matin Qaim*, a professor at the same institution and well-published researcher on the economics of GMOs.
The average findings reported in the literature are that GM technology in the main crops soybean, maize and cotton has:
- Reduced chemical pesticide use by 37%
- Increased crop yields by 22%
- Increased farmer profits by 68%
Furthermore, on average across studies, pesticide cost was significantly reduced (by 39%), while no significant impact was found on total production cost.
Some of the mechanisms driving these impacts are already well-known: The GMOs currently on the market are insect-resistant (IR) and/or herbicide tolerant (HT). IR traits substitute, to a certain extent, for chemical insecticides. HT traits, on the other hand, do not consistently reduce herbicide use, as they have led to the substitution of one herbicide for another. HT traits have often reduced pesticide costs because glyphosate, the most widely used herbicide compatible with HT crops, was cheaper than the more selective herbicides farmers previously employed. Yields are higher due to more effective pest control and thus lower crop losses. This effect is more pronounced in IR crops and in developing countries.
Apart from the type of GM trait (IR/HT) and the type of country (developing/developed) the paper also sheds light on several other reasons why some yield results are different from others. For example, it looks at whether funding from industry is associated with higher yield estimates. It is not.
Another interesting question is whether the type of publication matters. Indeed, studies published in peer-reviewed journals report 12 percentage points higher yield gains than studies published elsewhere (conferences, grey literature). This might seem as evidence of publication bias, whereby what is published in journals is artificially inflated and not representative of the data collected for studies. However, journal articles find a wide range of effects, some of which are even negative. Also, the literature not published in journals includes reports of NGO that are outspoken critics of GMOs. It is thus possible that the positive journal publication bias is the result of a negative bias in the grey literature.
A common criticism of the literature is the allegation that the evidence on impacts of GM crops stems only from a few selective datasets and is not representative. Indeed, there are datasets that are used for more than one publication. This can be the case when datasets are built over several years and an early publication reports impacts for the first period while later publication use the panel dataset for their analysis. Also, yield impacts may be the focus of a first paper and be reported again in a later paper which has another focus. Klümper and Qaim checked whether certain primary datasets have a big influence by weighting studies with the inverse of number of results derived from the same primary dataset. This approach yields almost the same results indicating that there is no reason to be concerned that the general conclusions are excessively influenced by a small number of datasets and researchers.
This new meta-analysis provides robust evidence on the positive impacts of the most widely used GM technologies and crops. The reason, however, that these crops are evaluated in a single meta-analysis in the first place is the fact that different GM crops are thrown into one and the same basket both in the public debate and in regulatory frameworks. It is, thus, useful to have a comprehensive paper on the effects of the most important GM crops and traits in order to inform the current public debate. Hopefully, at some point in the future, it will be possible to move the debate to a level where the question of whether a plant trait has been produced with genetic engineering of not will not be an issue anymore. The focus of the discussion can then move to whether a specific seed technology may be a good choice or not.
Reference: Klümper W, Qaim M (2014) A Meta-Analysis of the Impacts of Genetically Modified Crops. PLoS ONE 9(11): e111629. doi:10.1371/journal.pone.0111629