Category Archives: Policy related

The European Union’s hunger for soybean

In a recent paper published in Nature Food we show that halted soybean feed imports into the EU would favour ruminant animals such as cows and sheep over pigs and poultry. We also show that total production of animal-source foods in the EU would have to reduce if animal feed production is not to encroach on land currently used to produce food. Increased uptake of plant foods in EU diets is required maintain supply of nutrients.

The fires that raged in the Amazon during the 2019 fire season was unprecedented in number since records began in 2013. Although the drivers are numerous a paper in Global Change Biology has attributed the increase in fire count mainly to increased deforestation rates. When forests are cleared the vegetation is left to dry before it is burned off to make way for agriculture.

Agricultural expansion where natural vegetation is cleared and replaced with fields and pastures is one of the major threats to biodiversity and releases large quantities of greenhouse gases into the atmosphere. One of the recognised drivers of agricultural expansions into rainforests and other biomes in South America is demand for soybeans. The majority of soybeans produced in the region are destined for export markets. While domestic legal frameworks are important for forest protection, international trade has been shown to be an important driver of tropical deforestation and macroeconomic factors can often outweigh domestic regulations. Demand-side measures that reduce demand for soybean are therefore important to avoid further loss of forests.

Most of the soybean produced globally is “crushed” to separate the vegetable oil from the protein-rich meal. Each kg of soybean generates around 0.2 kg vegetable oil and 0.8 kg protein meal. Most of the revenue from the soybean crush is generated by selling the protein meal which is almost exclusively used to feed animals. This means that it is increased use of soybean meal, and ultimately demand for animal products, that primarily drives up soybean demand which leads to clearing of new areas in order to increase production.

Out of the total amount of soybean (beans, meal and oil) exported from South America around one-fifth is imported into the EU, where the vast majority is used to feed animals producing meat, milk and eggs. For example around 0.8 kg soybean is used to produce the average kg of EU pork, much more than needed for a kg of soy meat or tofu. If rough feeds (e.g. grass and silages) are excluded almost one-third of all protein fed to EU livestock comes from imported soybean. Demand for animal-source food within the EU thereby drives global soybean demand and tropical deforestation in South America.

Soybean is in itself an excellent crop with among the highest yields of protein per hectare, and being a legume, it fixates nitrogen from the air which reduces the need to apply nitrogen in the form of mineral fertilizers. Soybean is also one of few crops that is on par with animal-source foods when it comes to amino acid profile. The huge demand from the world’s livestock industry for cheap soybean does however make it hard to implement sustainable production.

When fed to animals, the majority of nutrients and energy present in the soybean are lost as heat or manure and only a small fraction is in the end retained in meat, milk and eggs that we eat. Redirecting the use of soybean from feed to eating it directly could therefore reduce the demand considerably. This would reduce pressure to deforest new areas and may also make it easier to implement more sustainable soybean production practices.

In a recent paper published in Nature Food we assess how dependant the EU livestock sector is on soybean imports by estimating how much meat, milk and eggs that would be possible to produce without soybean imports, given that use of EU cropland for animal feed production does not increase, which is important to avoid pushing production of food crops outside EU borders with potential negative environmental effects. We found that, depending on scenario for how to use EU cropland, 18-25% reduced animal-source food production in terms of edible fat and protein would be needed to completely eliminate soybean feed imports. Mainly it was the production of pork and chicken that was reduced, while beef, milk and eggs was less affected in the scenarios. To compensate this loss of nutrition to EU diets only between 17 and 22% of soybean that is currently imported to feed livestock would be needed to produce soy meat, tofu or other products for direct human consumption, thus considerably reducing the EU’s demand for land in deforestation-prone regions. Alternatively some land currently used to produce feed in the EU could be used to produce plant-source foods (resulting in the strongest reduction of animal-source foods presented above). In that case it would be possible to maintain supply of fat and protein to EU diets even without soybean imports.

The potentials for reduced cropland demand in South America were found to be large, but results also showed a risk of increasing cropland demand in South-East Asia, which is also a region where large scale deforestation occurs. Reduced use of soybean meal for animal feed results in less soybean oil being produced and if global demand for vegetable oils remain unchanged, this oil would need to be replaced by other vegetable oils. The most likely alternative today is palm oil. In absolute terms, the increased demand for cropland in South-East Asia was small compared to reduced demand in South America, but it is nevertheless an important potential trade-off that need to be considered. In one of the scenarios we explored the potential to increase rapeseed oil production within the EU to avoid this trade-off. Such a scenario would avoid increased palm oil demand, but instead rely on continued soybean imports to compensate for reduced consumption of meat. Which scenario is preferable will depend on how future global demand for different agricultural products develop. If for example transition from fossil fuels is achieved through the use of vegetable oil fuels a scenario with increased vegetable oil production in the EU is likely a good candidate.

In summary, the results from the new study show that there is a lot to gain by redirecting feeds of high food value, such as soybean, towards direct human consumption. To achieve these benefits without risking burden shifting it is however crucial that policies are developed with a holistic food systems approach and target soybean imports, dietary patterns, and livestock and crop production in conjunction.

Karlsson, J.O., Parodi, A., van Zanten, H.H.E., Hansson, P-A., Röös, E. (2020). Halting European Union soybean feed imports favours ruminants over pigs and poultry. Nature Food.

Soybean use for plant-based products shown in the figure was calculated based on data from the following sources:

Karlsson Potter, H., Lundmark, L. & Röös, E. (2020). Environmental impact of plant-based foods – data collection for the development of a consumer guide for plant based foods. Swedish University of Agricultural Sciences, NL Faculty/ Department of Energy and Technology.

Ercin, A. E., Aldaya, M. M., & Hoekstra, A. Y. (2012). The water footprint of soy milk and soy burger and equivalent animal products. Ecological Indicators, 18, 392-402.

Mejia, A., Harwatt, H., Jaceldo-Siegl, K., Sranacharoenpong, K., Soret, S., & Sabaté, J. (2018). Greenhouse Gas Emissions Generated by Tofu Production: A Case Study. Journal of Hunger & Environmental Nutrition, 13(1), 131-142.

Cai, T. D., Chang, K. C., Shih, M. C., Hou, H. J., & Ji, M. (1997). Comparison of bench and production scale methods for making soymilk and tofu from 13 soybean varieties. Food Research International, 30(9), 659-668.

Ecoinvent v3.6 – Products: Tofu production (GLO) and Soybean beverage production (GLO)

New paper on the methodology behind the WWF-vegoguide

How can the environmental impact of plant based foods be evaluated and communicated to consumers?

In a new paper published in the Journal of Cleaner Production, Hanna Potter Karlsson and Elin Röös describe the methodology behind the WWF-vegoguide presented in another blogpost. The guide was developed in cooperation between the researchers and WWF in a process described in the Fig. 1 below. WWF was the project owner and were responsible for the final design decisions regarding aspects such as which products to include, target audience for the guide, evaluation criteria and thresholds. The researchers were responsible for collecting footprint data, test the evaluation criteria, and provided feedback on the design to WWF. Views on the guide from external stakeholders like consumer and trade organizations were consulted in workshops.

Fig. 1
Fig.1. Process of developing the Vego-guide.
From Karlsson Potter and Röös (2020). J of Clean Prod.

The environmental impact categories to include in the evaluation of the foods were selected from the planetary boundaries framework (Steffen et al., 2015) and the mid-point categories of ReCiPe (Huijbregts et al., 2016) based on a set of criteria including their relevance for plant-based products, importance for guiding consumers, availability of scientifically accepted evaluation methods and data availability. Four indicators were finaly chosen: climate impact, biodiversity impact, water and pesticide use. Thresholds for rating the different product as green star, green, yellow and orange were designed to be aliged with the WWF Meat guide and to relate to the absolute food system boundaries as presented in the EAT-Lancet report (Willett et al. 2019). All products were compared on a per kg basis despite their different functions and nutrient content, which instead were considered by applying different thresholds for food groups, e.g. the protein group was allowed a larger share of emission space as these are more demanding to produce and more valuable in diets than carbohydrates.  

Read the full paper here:

100 000 ton per year instead of 30 000 – updated data about food waste from Swedish retail stores

The elefant in the room in this new report, was the new data on retail waste – 100000 tons instead of 30000 tons -, which now is based on figures reported directly by the retail chains via the voluntary agreement instead of statistics based on a few stores combined with the number of employees. Expressen, a Swedish tabloid newpaper, made an interview with me and choosed to put focus on how the stores try to hide the food waste data, by not including rejected fruit and vegetables and returned unsold bread and dairy products. From my perspective, this actual behaviour is true, but the incentive is probably not to hide data from the public, but to shuffle the waste to other actors for economical reasons. I also think the newpaper “scope” about this could have been that the Swedish environmental protection agency previously used a method that only catched a third of the actual waste, besides that the now published “actual” waste does not include everything arising at retail level. /Ingrid Strid, food waste researcher at SLU

Oranges in nets are often wasted when one fruit gets bad.