European Biodiesel Board launches pre-emptive critiques of new European Commission/IFPRI iLUC modelling

Over the last year, David Laborde of the International Food Policy Research Institute (IFPRI) has been working for the European Commission Directorate General for Trade on revisions to indirect land use change modelling published in 2010, using the Computable General Equilibrium MIRAGE economic model. The results of this revised modelling will be reflected in a forthcoming impact assessment by the European Commission of options to address indirect land use change caused by biofuel mandates.

In anticipation of the publication of the new MIRAGE results, the European Biodiesel Board has commissioned and released two separate critiques of the IFPRI MIRAGE model framework. The authors have been given access to the forthcoming IFPRI MIRAGE report in advance of its publication. There are three papers available:

There’s a lot of material, and until the IFPRI report itself has been released you might find it hard to assess all of it, but in the current policy context there are a few points we think are worth drawing attention to, in particular areas where the reports do not fully and/or consistently support the EBB position paper:

iLUC factors

Neither paper recommends the use of iLUC factors in European legislation, both arguing that there is substantial uncertainty in the modelling outcomes. Don O’Connor argues that the IFPRI iLUC numbers are likely to be overestimates. Neither paper, however, makes an argument that the idea of indirect land use change is ‘unscientific’ (as the EBB position paper represents it), with the Kiel Institute actively recommending iLUC should be addressed with increased direct GHG thresholds and a shift away from biodiesel and towards ethanol use to mitigate the risks.

Model quality

Don O’Connor’s paper highlights several areas in which CGE modelling with MIRAGE to date either does not or cannot model outcomes with complete accuracy. It focuses on data questions where iLUC may have been overestimated.

Kiel are actually very complimentary about the quality of the MIRAGE exercise. The Kiel report comments:

“The MIRAGE model by IFPRI used to address land use change caused by the European biofuel mandate represents a sophisticated modelling approach in the field of CGE modelling. It uses up-to-date data inputs and new methodological way to treat land and land use emissions on a global scale. The studies from 2010 and 2011 both transparently report the assumptions made and critical parameters chosen.”

And go on to say:

“For all biodiesel options, taking into account by-product allocation or not, the typical well-to-wheel values of the EU-RED plus land use change emission values from the Laborde’s Monte Carlo Simulation lead to higher emissions than the required 35% emission savings. These results are robust.”

On the basis of this robustness, Kiel also support a consideration of ways to reduce the contribution of biodiesel to the RED, given that biodiesel is modelled to have significantly higher impacts in terms of indirect land use change impacts than ethanol:

“We agree with Laborde (2011) that what could be learned from the model is that the distribution between ethanol and biodiesel play a major role and that thus this ratio should be under review for the National Renewable Energy Action Plans.”

Palm, vegetable oil substitutability and peatlands

The EBB position paper implicitly suggests that MIRAGE allows too much substitution of palm oil for rapeseed oil. The Kiel report, however, does not dispute this characteristic of the MIRAGE model, noting:

“The price competitiveness of palm oil leads to the substitution of non-energy uses of oils towards palm oil. However, since these demands cannot be met on current land areas devoted to palm oil production, there will be expansion, i.e. iLUC for palm oil plantations.”

O’Connor does argue that substitutability between oils for biodiesel may be too high, “Each feedstock gives the biodiesel some unique properties, including cloud point and stability.” This is reflected in the EBB position, “the study assumes important substitutions effects between vegetable oils, which does not to correspond to the reality of the European biodiesel market,” although Kiel find that assuming high palm oil substitutability for biodiesel use may be reasonable, “A recent study by Greenpeace Germany testing for biofuel admixtures in European filling stations found high shares of palm oil in the biodiesel shares (up to 80% in Italy), showing that this result not unrealistic.“

More important for the iLUC discussion is whether palm and rapeseed oil are mutually substitutable for food and other uses. O’Connor notes, “in the food sector the vegetable oils do not have full substitutability. Issues such as trans fat has favoured palm oil over soybean oil in many food applications.” While highlighting a potential model weakness and lower substitutability of soy oil for other food oils, this does seem to support the idea that palm can substitute strongly in the food market. Kiel notes that IFPRI have 78% of new rapeseed oil demand being met by rapeseed expansion, with the rest largely coming from palm. They discuss that the lower price of palm oil makes it an appealing substitute for more expensive oils. We do not believe that either paper supports an argument that the contribution of palm oil to meeting displaced demand for rapeseed oil has been overestimated by IFPRI.

O’Connor questions the peat emissions value used by MIRAGE, arguing the real value could be less. A recent ICCT report (which Don O’Connor would likely not yet have seen), however, reviewed the emissions literature and found that the value of 55 tCO2/ha/yr used by MIRAGE is in fact likely an underestimate, and that the emissions for the first 20 years of peat degradation could be nearly double this.


There is an inconsistency in opinions of co-product modelling by MIRAGE. The EBB position paper carries a clear message alleging that MIRAGE doesn’t model co-products properly, and Don O’Connor of (S&T)2 does indeed express concerns; but the Kiel Insitute actually says, “We think that by-products of bioethanol and biodiesel are well treated by the model.”

Amortisation for land use change

The paper’s differ over the choice of a 20 year amortisation period for land use emissions, with Don O’Connor characterising it as ‘arbitrary’ while Kiel state that they “consider the 20 year assumption as appropriate.”

EBB conclusions

At the end of the position paper informed by these studies, the EBB makes three recommendations, one of which is:

  • The Commission should remember the positive effect of biodiesel and biofuels to tackle Climate Change and meeting CO2 emissions reduction targets, advantages that ILUC debate wrongly diminishes.

This idea that, “The Commission should remember the positive effect of biodiesel and biofuels to tackle Climate Change”, is not well supported by these two papers. It is clear that biofuels mandates drive indirect land use change, and it is only by estimating the emissions this will cause that one can decide whether biodiesel has a positive effect in tackling climate change. The Kiel paper is supportive of the overall IFPRI conclusion on biodiesel – that it does not represent a good climate mitigation strategy. The O’Connor paper is less supportive, arguing that iLUC emissions are likely to be less than IFPRI finds, but still does not make a strong case that biodiesel iLUC emissions are insignificant. Any uncertainty in indirect land use change emissions is not around whether they exist, or whether they are in general positive, but around their magnitude – and the IFPRI modelling exercise suggests that biodiesel may well not deliver any GHG benefit compared to fossil diesel.



Alternative fuels Life-cycle analyses