The EU renewables target of 10% of transport fuel by 2020 to be met mainly by biofuels has been heavily criticised for its potential impact on diverting land from food to fuel production and thus putting upward pressure on food prices. Another source of criticism is whether it does actually contribute to reducing overall greenhouse gas (GHG) emissions. The Commission sought to deflect this latter criticism by requiring that biofuels which count against the renewables target must show a direct GHG saving of 35% compared to the fossil fuel that they displace. This saving requirement is gradually increased to 50% and 60% for fuels produced by installations that start production after 2017 and 2018, respectively.
However, these GHG reduction requirements do not take into account the possible indirect effects of converting cropland to biofuel production, which will encourage indirect land use change (ILUC) by conversion of pasture or forest to produce the food crops displaced by biofuels. The Commission has beeen asked to examine the question of ILUC and measures to avoid it, and to report on this by the end of this year. Two recent studies illustrate the complexities involved.
In March this year, a team of researchers at the International Food Policy Research Institute produced a report for DG Trade arguing that net GHG emissions would fall as a result of the EU renewables mandate, even when ILUC was taken into account. Yesterday, this conclusion was flatly contradicted by a study prepared by the Institute for European Environmental Policy, which suggested that the mandate would lead to increased emissions equivalent to about 7% of total EU transport emissions in 2007.
The two studies are very different. The IFPRI study is a state of the art attempt using computable general equilibrium modelling to trace through the production and consumption responses on a global level to the EU mandate. The IEEP study uses secondary data to produce a more back-of-the-envelope assessment of the likely consequences of the policy. However, the strength of this study is that it takes the actual plans set out by Member States in their renewable energy action plans and not just hypotethical scenarios to evaluate the ILUC impact.
There are other obvious differences. The IEEP study projects a higher first generation biofuels target than the IFPRI study (8.8% as opposed to 5.6%) based on MS action plans and therefore projects higher total biofuel demand of 26.0 Mtoe in 2020 compared to the IFPRI figure of 17.8 Mtoe (in each case, the difference with the 10% target is made up by second generation biofuels and other sources of renewable energy).
The IEEP also show that MS plan a much higher share of biodiesel in EU consumption than assumed in the IFPRI study (72% compared to 45%). As a result they project a bigger ILUC effect – while the IFPRI study estimates an increase of 0.07-0.08% in global cropland area, the IEEP study calculates a figure of 4-7 million hectares which works out at around 0.3-0.5% of world arable land, between five and six times as large.
Apart from the more ambitious scenario, another important reason for the different indirect land use outcomes is that, in a CGE framework, higher prices encourage intensification of existing land use, so that higher yields could account for around half or more of the additional demand for biofuels (depending on the crop) and thus reduce the ILUC requirements correspondingly by half. This is an important feedback mechanism ignored in the IEEP study.
There are also big differences in the two estimates of the GHG savings from biofuel use. IEEP project savings of 17 Mt CO2 compared to the IFPRI figure 18 Mt CO2 in moving from the 2008 biofuels penetration of 3.3%, but their incremental increase (from 3.3% to 8.8% = 5.5% increase) is more than double the IFPRI one (from 3.3% to 5.6% = 2.3% increase). Their more pessimistic conclusions on savings seem to derive from their methodology of calculating these savings. They don’t examine savings directly from the use of particular feedstocks; instead, they use the minimum savings required to meet the sustainability criterion in the RED, i.e 35% in 2011 rising slowly to 50% by 2017. The IFPRI study assumes that sugarcane-based ethanol will play a bigger role in meeting EU demand, which allows them to factor in a larger direct savings than what the IEEP assume. However, the IEEP study reflects what Member States plan to do and these plans assume much greater reliance on biodiesel than the more carbon-efficient sugarcane.
For these reasons, the IEEP ILUC estimates of GHG emissions are much higher than the IFPRI ones. They estimate the additional GHG emissions associated with ILUC to be between 44 and 53 Mt CO2, compared to the IFPRI figure of 5.3 Mt CO2. This mainly reflects their much higher estimate of ILUC, which is five to six times higher. The remaining difference would probably be eliminated taking into account the recognition in the IFPRI study that marginal emissions go up if biofuels demand increases, so their emissions would increase more than proportionately if they modeled the more ambitious IEEP scenario.
The very different results of these two studies concerning the effect of the EU biofuels mandate on net GHG emissions illustrate the tightrope that the Commission is walking in preparing its own assessment later this year.
As the author of the IEEP report I just wanted to follow up on Alan’s very interesting comparative article. I think the article raises some important points and indeed our work differs significantly from that which has gone before, as it is based on the actual figures supplied by the Member States in the national renewable energy action plans. This limits our ability to use complex analysis as the only information supplied in these plans in the scale of biofuel usage up to 2020, the nature of the fuel ie biodiesel versus bioethanol, whether this is conventional ie produced from primarily food based feedstocks or advanced ie waste or ligno cellulosic based, and whether materials will be imported. Therefore, it was not possible to take into account the specific savings from different feedstocks – although we did try to make allowances for imported bioethanol for the more limited ILUC impacts anticipated for sugar cane.
I wanted to respond to Alan’s point regarding assumptions; the IEEP analysis is based on an assessment of modelling study outputs produced by the Joint Research Council. It is from here that the estimates of ILUC scale were identified and the selection of conversion factors verified with experts in the field. The JRC analysis compares these models outcomes to the IFPRI study and all deliver much higher assessments of ILUC than IFPRI and I would suggest anyone interested in this field take a look at the JRC’s comparative analysis that considers outputs from models generally considered more reliable such as AG-LINK and GTAP.
Secondly, the conversion factors used within the IEEP study, which are based on these models outcomes, do take into account assumptions (to a variable degree) regarding price increases, driving yield changes and also issues such as frontier yields. Therefore we do take account innovations in the agriculture sector due to higher prices etc. This is not explicitly discussed in our work but these issues were inherently integrated into the conversion factors used to assess ILUC extent in the analysis due to their basis on previous modelling efforts.
One final point is that we ran sensitivity analysis during our study based on much lower estimates of ILUC – more equivalent to the IFPRI values. This showed that even with much lower estimates of ILUC impact per unit fuel biofuels still failed to deliver the GHG emission reductions required under the renewable energy Directive.
The key messages emerging are, firstly that we need to innovate more in the biofuels sector to avoid negative consequences and better promote the biofuels that deliver the highest GHG savings. Secondly, ILUC is an issue that will undermine the achievement of GHG reductions under the renewable energy Directive and therefore should be addressed within a policy where the purpose ultimately is to help combat climate change.
I hope these clarifications are of use and I would be happy to discuss with others who want to explore the study’s assumptions and outputs in more detail.
Actually, the NGOs have based their analysis on the land predictions of existing economic modelling, and hence it actually already includes price induced yield change – the carbon consequences are large even though this is accounted for. Otherwise, good article.
@Catherine
Thanks for this response which will certainly help to clarify issues for anyone trying to come to grips with the apparently conflicting evidence in this area. I accept that the JRC coefficients used in the IEEP study in principle should incorporate the induced effect of price increases on yield intensification, that is a useful clarification. However, I would be cautious in concluding that AGLINK and GTAP are considered more reliable than the IFPRI-MIRAGE model. While there appears to be some debate about an outdated coefficient which IFPRI use for peatland carbon emissions, this is not central to their conclusions. The model itself addresses a number of weaknesses in previous studies. I suspect that the reason it was not fully evaluated in the JRC study was simply that it did not appear until March of this year, and thus was not around when the simulations run for the JRC study were planned.
In fact, as I argued the IFPRI study may be less different to the IEEP conclusions than the headline numbers suggest, once the differences in scenarios are taken into account. The main difference appears to be the biodiesel/bioethanol mix. Using the ratio embedded in the national action plans ratio discussed in the last action plans (more 25/75 instead of 45/55 assumed in the IFPRI study) multiplies by 3 the land use emissions by MJ in the IFPRI model. This underlines your conclusion that promoting the biofuels that deliver the highest GHG savings is hugely important.
The tabloid type title is a bit deceitful. This is a very constructive piece in the ILUC debate, in the middle of an ocean of propaganda and mandated science.
How US Congressmen from corn producing States have blocked attempts to include ILUC in official US Environmental Protection Agency’s life cycle analyses of biofuels (because they led to a less positive environmental balance and might put zillions of tax exemptions at risk) must be one of the worst political interference with science since this old earth/sun rotation story. The EU NGOs that have taken the Commission to court on biofuels have not necessarily helped independent research either. True, transparency is useful, and there had been too much unsubstantiated propaganda on the benefits of biofuels from some Member states ministries captured by vested interest (do I need to quote countries?). But the NGOs’ action has been damaging for well-intentioned researchers: some of us have been forced to release preliminary drafts of all kinds of unrelated research papers on biofuels that we were much too preliminary to be released. There has been a very unpleasant suspicion that we had been subjected to policy pressure just because results changed when better data came. Not very good for science either.
So, in this poor environment for research, thanks for this constructive post, Alan. Explaining why IEEP and IFPRI results differ makes more sense than saying one of them is wrong.
JC
PS: In a paper to be released soon (available upon request) we show that among possible causes of divergence in results among Life Cycle Analyses on biofuels (no ILUC there), the most statistically significant cause is whether the authors are affiliated to a farmers’ organisation or to an environmental organisation. Being in the middle does not mean you are unbiased, but still this raises question about possible mandated science from both sides. To my knowledge, IFPRI is not part of any of these groups. I think it is good that DG Trade asks for external expertise to an independent body in addition to the JRC. The JRC (a directorate of the Commission, even though it has long proven its strict independence of mind on biofuels) is a very legitimate source but not a more legitimate one by nature. In the same area, Ms Bowyer, are you actually saying that GTAP and LINKAGE are “more reliable” models than MIRAGE? I invite you to look at the results of simulations of a Doha agreement on developing countries with the different models cerca 2005-2007 and we can discuss which ones made more sense next time I go to the IEEP.