Getting real about “real” carbon offsets

August 3, 2012, by Michael Gillenwater

I am about to commit an act of minor heresy by telling you that something everyone repeats as gospel is flat bunk. The qualities of a good emission offset project are one of the most common refrains you hear in the carbon offsets community. You can probably repeat most of them by memory: real, additional, permanent, verifiable, etc. Different programs or protocols might add other points about leakage or accuracy, or conservativeness or some other offset quality principle. But common to almost all programs and standards and protocols is the criterion that offset projects or credits must be “real.”

Here is a question for you: 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 unreal, and is this something we should be concerned about?

For years I have never really been clear what we meant when we said offsets should be “real.” But I just went along with it; because everyone else was repeating it, it must be correct. Right? (I was not positive where the concept came from, but my colleague Derik Broekhoff has cited the U.S. Clean Air Act. See below for more historical background.1) Some things you don’t question because they sound good, even if you are not sure what they mean. You assume everyone else understands and just go with it.

I have even participated in repeating the use of the term “real” in my own publications, including our work through the Offset Quality Initiative. (This work was migrated and built upon in developing

But I am here to tell you that it is all a bunch of gibberish and we should do away with the term entirely and quit acting like it is unambiguously meaningful. We only do ourselves harm by employing vacuous language.

Now you may say that you have seen definitions of what “real” means in the context of offset projects, and that the definition seems reasonable to you. So, let me show you why you have been duped.

I will start with a previous effort of mine the Carbon Offset Research & Education (CORE) website (content updated and migrated to to show that I am not just picking on others. CORE refers to the meaning of “real” as:

Offsets should come from real projects that have actually been implemented or will be implemented

Here is a pretty clear case of a circular definition…using the term real to define real. Reading more closely, though, the intended meaning seems to be that offsets cannot come from imaginary projects, or even more specifically that forward crediting (i.e., issuing credits before emission reductions have been achieved) should not be allowed. Forward crediting is a legitimate issue. But it is not clear to me that the term “real” means a ban on forwarding crediting. Wouldn’t it be easier to say: “no forward crediting.”

The Regional Greenhouse Gas Initiative (RGGI) defined “real” as:

Offsets must represent actual emission reductions and not artifacts of incomplete or inaccurate accounting. The effects of a project on GHG emissions must be comprehensively accounted for, and “leakage” in emissions must be factored into the quantification of emission reductions.

RGGI throws in the concepts of completeness and accuracy in accounting as well as the analogous issue of leakage. So does “real” mean the same as accuracy and completeness? Or is real just a synonym for leakage?

In another RGGI document they went further:

For a greenhouse gas offset to be real an offset compliance unit must represent one ton of CO2-equivalent (CO2e) greenhouse gas emissions reduction or removal (carbon sequestration) that results from an identified emissions reduction activity (i.e., a clearly identified action or decision). Offset project emissions reductions or removals must not be an artifact of incomplete or inaccurate accounting. Therefore, a project emissions or carbon sequestration baseline and project emissions reductions or removals must be quantified using accurate quantification methodologies and conservative assumptions where appropriate to account for measurement uncertainty. Quantification methodologies must appropriately account for all relevant greenhouse gas emissions sources and sinks and identified project leakage.

Let’s unpack this one. It seems that “real” for RGGI means pretty much everything. There is a hint at additionality in the language on action or decision. They bring in completeness then credible baselines and next conservativeness followed by leakage. And it seems the principle of accuracy is repeated about half a dozen times. So maybe “real” is just some meta-level principle in line with the surfer dude use of the term. “Dude, that project is soooo real.”

In its presentations of how it interprets AB32, the California Air Resources Board has explained to stakeholders that it views “real” as having the following components:




Again, it seems “real” is all good things to all people, such that it is not clear to me what it doesn’t mean.

The organization formerly known as the Pew Center on Global Climate Change (now the Center for Climate and Energy Solutions or C2ES) defined “real” in its briefing on offsets as:

GHG emission reductions should represent actual emission reductions and not simply be artifacts of incomplete or inaccurate accounting.

Like with RGGI, it sounds like what they really meant to refer to is the principles of completeness and accuracy.

Well this is an interesting exercise, you may think. But what do the players that set the market terminology in the first place have to say? OK, let’s run through them.

The Clean Development Mechanism (CDM) rules states that offset projects must be “real” in various places, but I am not aware of the term actually being defined anywhere (if I’ve overlooked something, someone please point out to me in the comments section where they do).

The Verified Carbon Standard (VCS) goes with the imaginary friend test:

All the GHG emission reductions and removals and the projects that generate them must be proven to have genuinely taken place.

I’m not sure we need a new term and principle for something that is already pretty well covered under the concept of fraud. Lying and claiming that some nonexistent offset project activity is taking place would seem to be a pretty obvious case of fraud to me. If we needed to be explicit on it then instead of saying real, how about just “VCS does not allow fraud.” That would seem to address the issue without ambiguity.

From WRI and the GHG Protocol team we have the following:

An offset credit is real if it represents an actual net reduction or sequestration in emissions, and is not an artifact of incomplete or inaccurate emissions accounting, including leakage. Leakage is defined as an unintended increase in GHG emissions caused by a project. A frequently cited example of leakage is a forest sequestration project that simply shifts deforestation activities to other forest land, reducing or eliminating the net sequestration from the project.

Here, again, they seem to view real as being a synonym for both completeness and accuracy, which they also use as separate principles elsewhere. They state that real is also about leakage, and go on to try and explain leakage within their definition of “real.” Why not just talk about completeness and boundaries. How does referring to it as “real” help?

Our friends at the Climate Action Reserve (CAR) offer the following definition for “real”:

Estimated GHG reductions should not be an artifact of incomplete or inaccurate emissions accounting. Methods for quantifying emission reductions should be conservative to avoid overstating a project’s effects. The effects of a project on GHG emissions must be comprehensively accounted for, including unintended effects (often referred to as “leakage”).

Again, we have concepts related to completeness, accuracy, conservativeness (which is in conflict with accuracy), as well as leakage (which is a derivative of completeness).

ISO 14064 is silent and does not use term (it also completely avoids the topic of additionality, which is fascinating given that you can’t even conceive of an offset without the concept of additionality). The Gold Standard and Carbon Fix require that projects be “real” but do not appear to explain what they mean by using the term (again, someone please correct me if I missed something).

What does all this mean? Well, it would seem that we have all acted a bit like a bunch of used car salesman…spouting off terms that sound good to everyone but that don’t really mean anything, or at least not anything well-specified or widely agreed upon. (We might as well have required that offset projects be “beautiful” or “synergistic” or some other corporate buzzword that can be attached to any appealing concept.) It is amazing we have gotten so far in writing standards, laws, and protocols using such a vague catchall term.

So, here is what I am calling on all of us to do. Say what we mean and do away with the term “real” when referring to offset project quality criteria. If you mean completeness, then say “completeness.” If you mean “accuracy” then say that. If you want to forbid forward crediting, then say forward crediting is not allowed. If we want offsets and offset policies to be viewed as credible, then we need to be careful and speak with precision and a purpose instead of just saying what sounds good.

1 The “New Source Review” program under the United States Clean Air Act of 1977 required offsets to be “real, creditable, quantifiable, permanent, and federally enforceable.”

Image courtesy of

14 responses to “Getting real about “real” carbon offsets”

  1. Alex says:

    A very interesting observation. I appears to me as if the definitions and characteristics of “real” would be captured under validation/verification. If the project were not real, it could not be validated/verified.

  2. Mike Burnett says:

    Michael brings up a good point about offset nomenclature. The “real” concept is embedded everywhere, but nowhere is it tightly defined. It is a bit of a feel good chameleon term. And we could probably do without it, and do just fine.

    I worked with Michael, and representatives of many of the other organizations mentioned in his op-ed, on the Offset Quality Initiative (OQI). It is interesting to note that the OQI had an evolving treatment of “real” over time. The July 2008 “Ensuring Offset Quality” document contained the “Offsets Should Be Real – Project-based offset credits should represent actual emission reductions and not simply be artifacts of incomplete or inaccurate accounting” language. But the November 2009 “Assessing Offset Quality in the Clean Development Mechanism” did not include “real” as a key offset quality criterion. It was just dropped, while all of the other criteria were included.

    I do not recall why we made this decision to eliminate “real” a criteria. Perhaps for the very reasons that Michael brings up. It’s probably because of the first of my three concepts listed below.

    I have three ways of thinking about “real” as regards offsets.

    First, I think of real as the meta-test for offsets. All of the other criteria that we apply (additional, permanent, verifiable, completely and accurately quantified and monitored with a realistic baseline and adjusting for leakage) are there to ensure that the offset is real. So if it fails to meet one of these specific criteria, it is not “real.” This is the clearest way to think about it, I believe.

    A second way to think about “real” is that the reduction must come something that is “really” a “project.” This means that a discrete decision was made to install some technology, as opposed to some more nebulous reduction in emissions attributable to better operation of existing equipment. So under this perspective, it would have to be a “real project,” as distinct from a “real emission reduction,” which presumably could come from improving operations of existing equipment.

    Finally, “real” could mean that the emission reduction really happened, which of course you would test by applying the other specific sub-criteria under the meta-test for “real-ness.”

    We could probably do without the “real” word entirely, and merely apply the specific criteria, and nothing would be changed. It is likely a “relic” that lives on because it was there previously, and it is a comforting word.

    But that comfort might be important. Since an offset is, as I have defined it in my speeches, “something that is in the air, and is not there,” it is a nebulous commodity, with only the contract backed up be the specific criteria tests, to prove that reductions did occur. Specifically requiring it to be “real,” even if not defined (or definable) might not be a bad place to be.

    Hi to you all,

    Mike Burnett

  3. Matt Dil says:

    Great point, Michael! I think you are right – we don’t think critically about the term “real” and just accept that it must be legitimate because it sounds good. Having the term “real” in the criteria for offsets is redundant. Imagine if you saw a job advertisement that began with: ‘Must be a real person and actually possess the following qualifications…’

    I suspect the reason “real” was frequently included in the standards is because the authors were being defensive. They knew that offsets are a ‘tough sell’ – offsets are an abstract concept and people will argue that polluters need to do more to reduce their emissions on site. But hopefully this defensiveness will become less necessary for two reasons: (1) the public is becoming more educated about climate change, (2) offset providers are clearly communicating how their projects work and/or sourcing offsets from local projects that customers can identify with.

  4. Peter Martin says:

    Thanks for a very useful analysis that cuts like a knife through the woolly use of ‘real’. Next, I hope you will turn your attention to the huge difference between what gets conveniently counted as an offset in the milieu of carbon marketers and managers and what might actually make a difference to the atmosphere. Please see page 5 and Appendix I in ‘Beyond carbon neutral’ for some suggested starting points.

  5. Michael, a refreshing article, but I’m not still quite sure what your final conclusion is. If I said “offset credits, that are defined against a counterfactual baseline, are not real in the sense that they are physical or material” would that be a fair paraphrasing?

  6. Congratulations, Michael! And raising eyebrows at ‘real’ is to only scratch the surface of a long-overdue critique of the ‘Alice in wonderland’ world of concepts and terminology defining emissions offsetting and accounting policy and practice. I’ve been involved in this arcane world domestically in Australia for a while and internationally since the UNFCCC Bali Conference of the Parties and have had to deal with the vagaries of the debates and arguments over the development of both an international REDD mechanism and revision of accounting rules for land use, land use change and forestry (LULUCF) by Annex 1 countries.

    Some simplicity and ‘naturalness’ in terms used would help a lot but, it seems to me, that much of the confusion stems from failure to separate out responsibilities for project managers, national/sub-national regulators and for the international community. What is a ‘real’ issue at one level of responsibility is not always ‘real’ at another yet we persist in presuming that everyone should be responsible for everything.

    But then, as we’ve learnt watching the self-serving negotiation of LULUCF accounting rules by Annex 1 countries’ resource agencies, what’s the point of having ‘real’ projects if the accounting rules are perverse?

    Alistair Graham
    Tasmania, Australia
    7 August 2012

  7. Tom Baumann says:

    The term and concept of ‘Real’ GHG emission reductions/removal enhancements is certainly worthy of being revisited, if just for the sake to eliminate redundancy in order to improve understanding/uptake of offsets. I like Mike Burnett’s post on the issue (note to GHGMI – is it possible to turn on ‘like/dislike’ for comments?).

    In reference to Michael Gillenwater’s post and specifically on how ISO 14064-2 (project accounting) addresses ‘Additionality’ and ‘Real’ – allow me to clarify. ISO 14064-2 does not include ‘Real’ because during development of ISO 14064-2 ‘Real’ was regarded as a programmatic rule/criteria, which is outside the scope of ISO 14064-2.

    The SEI-GHGMI CORE website describes how ISO 14064-2 is a standard rather than a program, and how ISO 14064-2 addresses ‘additionality’ with a general requirement and reference-out to the program rules (link =

    ISO 14064-2 (Clause 5.4) specifies the following requirement in regards to additionality: “The project proponent shall select or establish, justify and apply criteria and procedures for demonstrating that the project results in GHG emissions reductions or removal enhancements that are additional to what would occur in the baseline scenario.”

    How additionality is incorporated into ISO 14064-2 is based on the core principles of ISO standards in general, i.e. that ISO standards not be a barrier to trade (WTO-TBT – anyone following development of ISO 14067 (product) will know this is a major issue). As such, ISO standards must be policy-neutral (extended to include program-neutrality). This neutrality has pros and cons. The main benefit intended (other than not being a trade barrier) is for the ISO standard to be internationally compatible among users – which is expected to yield a certain level of standardization. This is an important aspect of fungibility in the markets. On the other hand, the limitation of ISO standardization often relies on further specific requirements (from programs and sector specific methodologies) to ensure the appropriate degree of quality. This is of course very important for market confidence.

    As further explanation of ISO 14064-2 on the above issues, its Introduction states:
    “A standardized approach for quantification, monitoring and reporting is required for GHG projects and any resulting GHG emission reductions and/or removal enhancements, in order that they are comparable among
    intended users and GHG programmes. Accordingly, this part of ISO 14064 specifies a general, GHG programme-neutral framework and uses terms and concepts designed to be compatible with other requirements and guidance from relevant GHG policies and programmes, good practice, legislation and standards.
    This part of ISO 14064 deals with the concept of additionality by requiring that the GHG project has resulted in GHG emission reductions or removal enhancements in addition to what would have happened in the absence of that project. It does not use the term “additionality”…Thus the project proponent may apply additionality criteria and procedures, or define and use boundaries consistent with relevant legislation, policy, GHG programmes and good practice.”

    Although the concept/requirement of additionality is within the requirements of ISO 14064-2, the simple reason why the ‘term’ additionality is not present within the requirements of ISO 14064-2 is because of certain sensitivities/perceptions/politics of certain parties involved in the development of the standard.


  8. Thanks to everyone for the great comments. I’ll give some quick reactions.

    Alex, I think you are correct. If there is not a physical activity then there is nothing much to verify (assuming it is more than a desk review, which it should be). I think this falls under my point on fraud. Verification should, at a minimum, prevent acts of fraud.

    Mike, thanks for the lengthy commentary. I think you are correct in viewing “real” as a sort of meta term. But also then see it as not really being useful or needed outside the realm of prose. It is fine to use it there, but don’t pretend it is something appropriate for a technical specification regarding offset quality. As for the OQI reports, which you and I worked on together, we did drop it in the CDM report for the reasons I discuss here. I think it was probably my doing, having worked a lot on that draft.

    Matt, my point exactly. Where else do we need to clarify that something must be real? Or as I put it, do we really need to be explicit that offset projects should not be a product of fraudulent information? In a legal sense we should have project developers attest to the veracity of the information they provide offset programs and other stakeholders, but I don’t think “absence of fraudulent information” needs to be a core offset quality criteria.

    Peter, I think you are probably talking about additionality. Some other readers here will be familiar with my painfully long discussions of this topic, which you can find a entry point to here:

    John, you have perfectly foreshadowed my next blog, which is exactly about the concept of “counterfactual”. To give you a hint at what I will say, basically we have been mistaken on this concept as well, which I will explain in a future newsletter. As for my conclusion, apologies if i was not sufficiently unambiguous. It is that we should quit using the term “real” when specifying offset project quality criteria and instead be more precise in our specifications. Using a term that means something different and/or a ransom mix of a bunch of different things to different people does not help further the credibility of offsets as a policy mechanism. And if it was not clear in the blog, I do thing offsets are a credible policy mechanism.

    Alistair, thanks for sharing your experience. I am not sure, though, that simplification in terminology is what is needed. Of course we should not make rules any more complicated than they need to be. But the issue here is probably the need to be more precise and unambiguous, as well as move towards more uniform terminology to foster standardization and fungibility. As we move into REDD issues, things are going to avoidably get more complex, so best we get our house in order for the, relatively easier stuff in the non-LULUCF sectors.

    Tom, thanks as well. I should make it clear that Tom and I are good friends, colleagues, and co-founders of GHGMI. On the issue of additionality with respect to ISO 14064-2, I’ll just point to my discussion of it in the Part 1 paper I did on the topic showing that we all got it wrong back then on how to define additionality, and ISO was no exception. See Table 1.


  9. Judith Hull says:

    The concept of ‘real’ we had been working with in Environment Canada was very specific, simple to understand and essential – The project must be a specific action to reduce GHGs; reductions that arise simply by reducing production are not eligible.

  10. Judith,

    Thanks for sharing. I was not aware of what Environment Canada had done on this topic. My colleagues here at GHGMI, Tom Baumann and Patrick Hardy, will surely be interested.

    As for the approach you describe, I think I would still have some problem with it. It seems to be that you are really talking about issues of what the baseline is and issues of functional equivalence. If the emission reductions from my project does come from a reduction in production, but that reduction is due to other changes I make that allow me to use that output more efficiently, is this not eligible? The point is that what is the appropriate baseline can be a complex question, and boiling it down to a vague term like “real” seems to not really do proper service. But I am curious to hear more.

  11. Bob Washburn says:

    GHG offsets are the only ones I know of where the reductions are from the current base line emissions and not below BACT or some other defined limit. This is causing double counting of mediocre results and hiding the lack of progress toward what is necessary. If we take the target for reductions of GHG to achieve the 2C or whichever climate change target we chose, then the offsets need to be for reductions below what is required for that goal. Basically if the world reduction needs to be ~80% then offsets only count when the reduction is beyond that level. Taking credit for someone else’s 2% reduction, regardless of how “real” is greenwashing at its worst.

  12. Dave Newport says:

    Nice work. Indeed, language is important. While we are on about it, there are a few other terms bandied about incorrectly that I have been at war with for a few years.

    1) RECs are not offsets. However, folks routinely state they “have offset their carbon emissions with RECs.” That is wrong on several levels. Yes, it is inconvenient that the word offset can be a verb or a noun. But get it right. RECs can be said to “compensate for” or perhaps “balance” the GHG emissions from electricity consumption, but leave offsets out of the conversation.

    2) While we’re hammering RECs, if a facility/organization is buying RECs they are not “powered by wind.” There is no orange extension cord connected to a wind turbine. RECs connect load to renewable power by a contract, not a wire. To say “powered by wind” creates false images in lay peoples’ minds.

    3) If a facility/organization is buying offsets or RECs sufficient to equalize all their GHG emissions, they have not attained “climate neutrality.” They have attained “GHG neutrality.” The definition is “no net GHG emissions.” That definition does not define a state of the climate, it defines a state of emissions. Stop dragging the climate into it.

    Thanks for your well done rant–and tolerating mine.

  13. Bob and Dave, thanks for your comments.

    Bob, there are numerous ways to conceive of what a baseline should be. Your points are interesting because it gets at the issue of how burden is allocated for mitigating GHG emissions. If you say we all must take on responsibility for reducing our emissions by 80%, and define a baseline accordingly, then yes…only reductions beyond that would be additional. Such an approach does have practical implications in that you are then left with the question of how, in the first place, do you have policy that enforces everyone to reduce their emissions 80%.

    Dave, I share your skepticism on RECs. You should find these articles interesting:

    Gillenwater, Michael, “Redefining RECs (Part 1): Untangling attributes and offsets,” Energy Policy, Volume 36, Issue 6, June 2008, Pages 2109-2119.

    [Here for pre-publication discussion paper version]

    Gillenwater, Michael, “Redefining RECs (Part 2): Untangling certificates and emission markets,” Energy Policy, Volume 36, Issue 6, June 2008, Pages 2120-2129.
    [Here for pre-publication discussion paper version]

    Gillenwater, M., “Taking green power into account,” Environmental Finance, October 2008.

  14. Judith Hull says:

    Michael – In response to your August 9 note on my short comment, I would emphasize that isolating one criterion is always problematic. Clearly baseline setting with functional equivalence is key. We used the ‘real’ criterion to flag that just a cut in production would not be eligible. A project to make that production more efficient (even if total output were to decrease) may well be eligible. Judith

Leave a Reply

Your email address will not be published. Required fields are marked *