What is wrong with ‘real’ carbon offsets?

Abstract

Over the last several years, an extensive literature on greenhouse gas (GHG) emission offsets and offset programmes has emerged (CRS, 2007; OQI, 2008; Kotchen, 2009; Bushnell, 2010; Kollmuss et al., 2010; Sovacool, 2011; Gillenwater and Seres, 2012). Common to much of this literature and to all major GHG emission offset programmes is some elaboration of quality criteria for defining what constitutes an acceptable offset project. The purpose of this short essay, while supporting the use of offsets as a policy mechanism, is to argue that a ubiquitous element in these quality criteria is essentially meaningless and should be purged from the technical lexicon for this topic. Reports and studies on carbon offset programmes and markets make frequent reference to common quality principles for emission offset projects. If you have followed any of these programme design discourses you can probably repeat most of the principles by memory: real, additional, permanent, verifiable, no double counting, etc. Different programmes or protocols might add other points about GHG accounting concepts like leakage (i.e. changes in GHG emissions outside of the specified system boundary) or the quality of the baseline used (i.e. reference projection of emissions absent the specified intervention), or the principles might be further supplemented with good practice precepts that could apply to any programme, such as accuracy, completeness, or conservativeness. But, common in the literature as well as to almost all programmes, standards, and protocols is the explicit statement that offset projects must be ‘real’. However, there is a problem here. What does it mean for an offset project to be real? What would an unreal offset project be? How could we tell if it was not real, and is this something policy makers, offset programme administrators, and verification bodies should be concerned about? For years I have never really been clear what was meant by the term ‘real’ in the context of carbon offsets. But, apparently like almost everyone in this community of practice, I went along with using the lingua franca. All the literature and discussions I had with other experts confirmed that this term had a commonly understood and significant meaning. It must be worth using and repeating because everyone else was repeating it ... yes? Some things you do not question because they sound good, even if you are not sure what they mean. You assume everyone else understands and that maybe you are just missing some key rationale. An early, although possibly not the earliest, reference to the term ‘real’ in emissions trading comes from the US Clean Air Act. Specifically, the ‘New Source Review’ programme under the United States Clean Air Act of 1977, which required offsets under that programme to ‘real, creditable, quantifiable, permanent, and federally enforceable’ [emphasis added]. Given that this early form of an offset programme set ‘real’ as its leading quality criteria, it seems likely that the precedent and pattern started with this legislative language. But I am here to argue that ‘real’ is a meme that lives on only because it sounds good not because it is useful or meaningful. Or to be more blunt ... the term ‘real’ in the context of emission reduction offsets is little more than ambiguous gibberish and that there would be no loss in substance or meaning if we simply did away with the term entirely when speaking of offset quality criteria. Further, by relying on principles and language that lack useful and precise meaning, we risk doing long-term harm to the credibility of offset mechanisms as a class within our policy options toolkit. In part, the use of the continued use of the term is probably a historical accident; a carryover from US legislative language.