Am iLUC-ing in the wrong place?

Posted Wednesday, 10 July 2013, 03:43

There is a modern parable that goes something like this:

An observer sees a man searching for something under a streetlight and asks what the man has lost. He says he lost his keys and they both look under the streetlight together. After a few minutes the observer asks if he is sure he lost them here, and the man replies: No, he lost them in the dark alley. The observer asks why he is searching here, and the man replies: "This is where the light is."

I was reminded of this parable this week while reading a proposal by Prof. Uwe Lahl to replace the economic modeling based indirect land use change (iLUC) reporting proposed for Europe’s Renewable Energy Directive and Fuel Quality Directive with a value based on a simple calculation that would spread global annual land use change emissions equally across all agricultural production. The calculation is detailed in an amendment to the Directives suggested by German MEP Christa Klass, and has been referred to elsewhere by Prof. Lahl as a “Potential Land Use Emissions Value”. In essence, the calculation involves working out global annual land use change emissions from deforestation, dividing that by total global agricultural production to give a value for average land use emissions per tonne of agricultural production, and then converts that number to a grams of carbon dioxide per megajoule (gCO2e/MJ) iLUC estimate that would be assigned to biofuel. Based on this calculation, Prof. Lahl reports a “Potential Land Use Emissions Value” of 16 gCO2e/MJ. Prof. Lahl proposes that this simple calculation would be preferable to economic modeling (see e.g. Laborde, 2011) as a method of estimating indirect land use change (iLUC) emissions, highlighting the uncertainty inherent in economic modeling exercises and arguing that this calculation is based on ‘actual data’.

It’s certainly true that this calculation is much simpler than running a model with thousands of parameters, and that it is based on using reported data from the previous year, but unfortunately there’s a lot that’s problematic about it. It provides a single average iLUC estimate for all biofuel feedstocks, ignoring the fact that some feedstocks will be more strongly associated with carbon loss than others. It considers only deforestation emissions, ignoring the fact that grassland conversion result in substantial carbon loss. It ignores things like co-products that reduce iLUC from some biofuel feedstocks. It ignores the fact that not all deforestation is driven by agriculture. But most importantly it is looking for an answer on iLUC in a place where there isn’t one, just like the drunk under the streetlight in the story above. The Lahl calculation provides an answer to the question, “What is the average annual carbon emission from deforestation for every tonne of agricultural production in the world?” But the question for policy makers considering biofuel mandates is different – the question is how much land use change will occur as a result of increasing biofuel demand. Current levels of agricultural production could (generally) continue without requiring deforestation – it’s when production expands that we expect to drive land use change. Using the same calculation but dividing by the year on year increase (from 2010 to 2011) in agricultural production, rather than the total global agricultural production gives a “Potential Land Use Emissions Value” of 300 gCO2e/MJ! This compares very unfavorably to the ~90 gCO2e/MJ of emissions caused by burning fossil fuels – if you felt this number gave a useful indication of iLUC, you would probably abandon all biofuel policy. 

The thing is that both of these emissions numbers, 16 or 300 gCO2e/MJ, can be thought of as the results of a ‘precise’ calculation based on actual data, but that doesn't make either of them a very useful estimate of iLUC emissions that are actually caused by biofuels. Prof. Lahl rejects the use of estimates from economic modeling because of the uncertainties in those assessments – and it is certainly true that there are uncertainties. Nevertheless, these types of economic study have already been used to apply regulatory iLUC factors in the United States, and as I argued in a peer reviewed paper last year, in the European Union context the modeling by IFPRI with the MIRAGE model provides the best available estimate of the iLUC emissions that will be caused by expanding European biofuel demand. This is because this economic modeling considers all of the key dynamics that affect actual land use outcomes (like yield increase, co-product use, switching between crops), rather than taking a semi-arbitrary pair of statistics from a previous year and dividing one number by the other. 

I took the streetlight parable from a conference paper I read this week called The need for consequential system modelling in Life Cycle Assessment for robust decision support, by Bo Weidema, Miguel Brandão and Jannick Schmidt. That paper also provides the following quote from Tribus and El-Sayed (1982):

It is much more important to be able to survey the set of possible systems approximately than to examine the wrong system exactly. It is better to be approximately right than precisely wrong.

Economic models of indirect land use change cannot give an exact, measured answer to how much iLUC will be caused by biofuel production, but they can look at possible scenarios for how the world will deal with increased biofuel demand (the Laborde modeling with MIRAGE looks at thousands of scenarios by doing Monte Carlo analysis to assess uncertainty). In contrast, calculations like the one proposed by Prof. Lahl generate the wrong answer precisely because they simply ask the wrong question.  

At best, the Lahl calculation provides a reminder that land use change emissions are significant and should be addressed*. But at worst, debating a calculation that misses the point is an unhelpful distraction as the political institutions of the European Union seek to find an appropriate compromise to reduce the risk of indirect land use change while providing some protection to investments made in good faith by the biofuel industry.  


* Prof. Lahl makes this point himself - you can read his ideas in full here. He has a proposal to limit imports from countries where deforestation occurs as a way to drive improvements in forest governance in the developing world, but my concern with this would be that it ignores the fundamentally connected nature of food markets - for instance increased rapeseed oil demand in Europe is clearly a driver of palm oil imports as I showed in a short paper earlier this year. Attempting to use these types of direct trade barriers against biodiesel imports does not recognise that indirect land use change occurs in this global way, and may be difficult given international trade law.