How do you explain additionality?

January 25, 2012, by Michael Gillenwater

If you have worked in the climate change space for very long, you have likely faced this question in one form or another. Try explaining carbon offsets to your sister-in-law and you have two choices. Either you give her a superficial response in an attempt to change the subject or you dive in and try and explain offsets. If you chose the latter, you will find it near impossible to avoid the concepts of a baseline and additionality.

Although I have done my share of dodging the question over the years, it has been the deep dive discussions on offsets (including the work we did as part of the Offset Quality Initiative) that led me to realize, not only did I not fully understand additionality, neither it seems did anyone else. (Or at a minimum any additionality experts out there seem to have serious trouble articulating their full comprehension of the topic.)

There are, of course, lots of opinions about how impossible or complicated additionality is to apply. But it was only when I started looking into the reports and other literature, with the sole purpose of studying how additionality and baselines were addressed, that I realized we had a real problem here. No wonder people are skeptical about offsets. If you look at the climate community’s own words on the subject, we don’t appear to have a handle on a concept we have championed as integral to the policies we have created. Language on additionality and baselines is vague, inconsistent, or both. No two authors seem to define these concepts in the same way without falling back on some platitude like “business as usual.”

Now, it is at this point where I expose a bit of my personality. I have a tendency to be a bit of a gadfly at times (some might go as far to say this time it is more of a Don Quixote complex). Would you spend several years researching and writing a paper on additionality? Well I did. And further, I wrote three papers.

(As an aside, some of the impetus for this research came out of my work comparing Renewable Energy Certificates and carbon offsets and thinking about what really justifies a claim that an activity actually reduces emissions.)

So before I go any further, here are the papers. We released an earlier version of them several months ago. Since then I have received comments from a number of you (thank you!). The versions we are releasing now incorporate the comments as well as some other improvements.

Each of the three papers does something different. Part 1 explores how we got here. Why is additionality important and why is it such a mess. It concludes with a definition for additionality and baseline that addresses some long standing problems.

Part 2 is a beast of a paper. So be prepared. It dives deep into how to be more rigorous in the application of additionality and baselines and do so in a way that enables the development of standardized approaches. Part 2 has an element of guidance to it, but more importantly, it walks you through the theoretical questions we need to answer as a community if we are to defend offsets as a legitimate policy mechanism.

Lastly, Part 3 takes a tangent into the world of credit stacking and ecosystem services. Once you have thought deeply about additionality, then you can reflect on how to deal with cases where you have multiple overlapping offset programs that are crediting multiple environmental benefits. As we know, many project activities produce benefits other than just greenhouse gas emission reductions.

All three papers are written in an academic style. In this blog, I am not going to try and give a complete summary of them in a non-academic style. But I will make a few points.

Defining the definitions

Hopefully, it is already clear to you that proper consideration of additionality and baselines is key to the environmental integrity of offsets. I would go even farther and say that the very concept of an “offset” requires the concept of additionality. You can’t say you have offset some harm unless you can show that you “caused” some equivalent extra good to occur elsewhere. Additionality is about this causal question.

To start, we need to clarify the precise “cause and effect” we are concerned with in the context of project-level accounting for emission offsets? For us, the “effect” is the implementation of the proposed project (the effect is not the reduction of emissions…see Part 2 paper for an explanation why).

Next, we need to specify our “cause”? You can’t try and predict an effect if you never bother to identify the cause with which you are concerned. Ignoring this issue is like saying:

Betsy: “Why should I trust that this offset credit is real?”

Tom: “Because, we caused the project to happen.”

Betsy: “OK, how?”

Tom: “Well we don’t know how, and we avoid thinking about what we did to cause it. But we know we did cause it to be implemented.”

In the offset community this line of thinking is epitomized by the vacuous, and unfortunately widely used, phrase: “the project would not have occurred otherwise.” This type of language is problematic because it is half a thought: Otherwise except for what?

The “cause” is the policy intervention recognized by an offset program. It might be limited to the economic incentive created by the GHG program (i.e., the risk-adjusted offset credit price signal), but it does not necessarily have to be limited as such (for why, see the Part 2 paper). The additionality question then becomes whether this intervention caused the proposed project to happen or whether there was no behavior change resulting from the intervention. How do we answer this question? By assessing whether the proposed project is the same as its baseline, which, if so, indicates that the policy intervention had no effect. Therefore, the definition of additionality is contingent upon the definition of a baseline. And what is a baseline? Well, it is what would occur in the absence of the policy intervention, holding all other factors constant. So, again, we are back to the importance of being precise about what we recognize as the policy intervention.

The problem here is that few GHG programs explicitly specify what they recognize as their policy intervention. Therefore, they leave it to validators, project developers, and media reporters to guess and play games with what is additional. Further, and probably more importantly, their lack of specificity makes it impossible to falsify a determination of additionality. We especially need to be precise about the recognized policy intervention before we can develop standardized approaches (i.e., we have to know what we are setting a standard for).

Avoiding linguistic traps

But my biggest annoyance when discussing additionality —and a very common trap that’s snared no shortage of climate policy wonks— is the circular definitions error. How many times have you read or heard someone say that additionality is about what “would occur without the project” (e.g., the Kyoto Protocol itself includes this type of language). The problem with this way of thinking is that we are trying to understand a cause and effect relationship, as discussed above. We are trying to decide if behavior is being changed. The project cannot cause itself to happen. That makes no sense. It is not the absence of the project that is the defining characteristic of a baseline. It is the absence of our recognized policy intervention. This is what is meant by circular definitions. Saying something causes itself puts you in a meaningless logical loop.

Critics will say that we can’t go trying to get inside the heads of every project developer and investor and predict why they are doing what they are doing. I agree. But it would be good enough to predict what a typically project developer would do. What would a reasonable project developer do under typical conditions in this industry or country under conditions where the recognized policy intervention is absent? This is our baseline.

If we are very precise about our recognized policy intervention we can then call on critics to use more rigorous testing and analysis in their arguments rather than vague challenges. Likewise, we can provide validators and methodology developers clear guidance on what their standardized approaches are to approximate.

Another reason there is so much confusion and frustration on the topic of additionality is that most GHG protocols and standards have provided little help. Both the ISO standard and the GHG Protocol for projects basically punted on the topic. We have gotten so used to sloppy and vague language on additionality and baselines that these non-definitions have started to sound like they actually mean something. Since no one seems to say anything that sounds carefully crafted, we assume it must be OK for us to do the same. It is a classic case of groupthink.

I know this is a challenging topic (and one that most really do not want to discuss), but it is not going away. Either, as a community, we deal head on with the conceptual challenges of additionality and baselines, or we should just walk away from offsets as a policy mechanism. I’m convinced that for us to make progress on offset policy, we first have to be far more precise in our thinking about additionality and baselines.

I look forward to hearing your thoughts and comments on my attempt to do just that.


Why is the GHG Management Institute publishing these discussion papers? To be clear we are not a traditional research institute or think tank. There are plenty of those already. And while we are highly cognizant of the value of research and academic inquiry (we even sponsor a peer-reviewed journal), it is not part of our mission to add to the ongoing 20+ year avalanche of policy white papers and reports. Indeed, we long ago determined that our greatest contribution comes in changing the way GHG management is taught, technical skills are developed, and the resulting practitioner class professionalizes. As we work to achieve that mission our research program selectively identifies neglected research questions we believe are key for the GHG community to grapple with for the benefit of all and to further develop ourselves professionally.

13 comments on “How do you explain additionality?

  1. Sallie on said:

    Hi Michael, your blog entry popped into my inbox just as I was pondering how one should best communicate to others (i.e. our projects and partners) how to set baselines for NAMAs. Given the not-necessarily-offset-nature of the NAMA, I was asking myself how precise we need to be in setting baselines for measuring GHG redux. On the one hand, the NAMA might not generate carbon credits but on the other hand in order for the NAMA-implementer to prove that reductions have happened, he/she needs to also show additionality and have a plausible baseline. Do you have any tips or good resources I should check out as we think about setting baselines for NAMAs? In the meantime, I will try diving into your three papers for further inspiration! Best, sallie

  2. If you want precise language in ISO standards, there is a process for that: join the ISO standards development teams, and work hard to fix the language in the standards by suggesting specific improvements. They would then be debated, commented upon, voted on, and then potentially incorporated into the standards. It is a continuous improvement process that works.

  3. Hector Ginzo on said:

    I mostly agree with you. (The reason for my ‘mostly’ is that I’m not sure how much I’ll agree with you after reading your parts 1 to 3. But I digress…) The problem with the political concept of additionality is that it is defined as a logical counterfactual: if not A, then B. Bringing A and B from the realm of Logics to the earthly scope of an international negotiation process that feeds on scientific facts and theories but is constrained by politicians’ agreement to mostly unsavoury policies for theirs constituencies is a very complex endeavour. It might be that replacing that counterfactual with a reasonable and sound proposal to reduce emissions with regard to some historical trend would be less problematic as to its implementation and comprehension, but I doubt it would be politically acceptable by a community of almost 200 countries differing in aims, idiosyncrasies, and political stripes. The path to an agreement among them is necessarily paved with loose concepts, like the one for additionality; otherwise, the path would lead to nowhere. (But I digress…) Cheers!

  4. Mark Purdon on said:

    Thanks for this posting and opportunity to comment. I agree with your statement that “What would a reasonable project developer do under typical conditions in this industry or country under conditions where the recognized policy intervention is absent? This is our baseline.”

    But how do you determine this? From my own research it is clear that demonstrating additionality is best done through comparison, rather than counterfactuals. Indeed, this is how the effectiveness of other policies is evaluated. For example, in a recent textbook on policy evaluation the authors clear state: “All causal analysis also requires comparison. Without comparison, there can be no counterfactual: what would have happened to outcome (Y) if there were no intervention (X) or if the intervention (X) had been different?” (Langbein & Felbinger, 2006: 59).

    It seems to me that the problem is that comparison is expensive and most offset systems don’t want to pay for it. It means gathering data not just on the offset project, but also for similar activities in the same sector but not claiming credits. Hard to get project developers to do this, though its possibly the role for government or another organization.


  5. Michael Gillenwater on said:


    Thanks for the note. Glad we have been timely. All three papers are intended to be scale neutral. Meaning, I attempted to think about additionality generically. So, I did not limit the framework to the traditional project scale or carbon offset context. So the general points should apply to NAMAs. Of course, as you increase in scale, then the possibility of finding representative case study decreases. As we saw yesterday, if the scale is the entire housing sector of Mexico, then it will be hard to find a control group on that scale. But within a country you could.


  6. Michael Gillenwater on said:


    Point well taken. In some ways I am critiquing ourselves. Our own Tom Baumann (Director, Strategic Solutions and Partnerships) was a leader in drafting the 14063 Part 2 standard. And the Institute is already on the USA Tag and an ISO liaison. (Although, I will admit that as a small non-profit, it is extremely difficult for us to be active in ISO, as the primary mode is flying to meetings in different parts of the world. Small organizations do not generally have that kind of funding.) It was the process of writing these papers that allowed me to reach the conclusions regarding how we need to improve the way we define additionality and baselines. So I fully agree, the next step is to work to do just that in the relevant standards.

    I’ll also note, that we include all the 14064 standards in our courses and teach them. But we also teach our learners to think critically about all programs and standards and not take what they read blindly.


  7. Michael Gillenwater on said:


    Looking forward to hearing your thoughts after you have a chance to look at the papers. One purpose of them is to provide a rigorous framework for how to remove politics from the process as much as possible. We could just have baselines that use historical data (that would not work for completely greenfield projects, though). But then we really can no longer claim that these are true offsets, which requires the concept of additionality. I’d be fine doing that, but then we just need to be honest and transparent about what we are going. If we are going to climate we are truly offsetting emissions then we have to consider a forward looking baseline.


  8. Michael Gillenwater on said:


    I could not agree more. And if you look at the Part 2 paper, this is exactly what I recommend. Although I think we can also use tools such as discrete choice or lab based behavioral econ studies as well. But the ideal is a case study with a control group. Yes this is expensive, but not impractical (for some scales and contexts). The expense can be justified if we go to a system of standardized approaches instead of project by project assessments.


  9. I think that you have defined the problem in exactly the right way. I have had several meetings, presentations and training deliveries where I have defined Additionality with great confidence and simplicity. I then look around and find at least 80% of the faces are registering confusion. Of the remaining 20% I suspect that most of them have an understanding but not necessarily the one I intended. The issue is that usually they are not specialists and lack the whole raft of context that specialists can apply. GHG accounting and carbon management are immature disciplines and as such lack a stable and reliable system of knowledge.

    It is inevitable that jargon and professional short forms of meaning will exist. Being able to communicate these without context from experience is possibly not best left to heavily engaged experts who live and breath it on a daily basis.

    I would add materiality and uncertainty to additionality for the list of confusing, but vital, concepts that are difficult to communicate outside of – respectively – accounting and engineering (though then with nuances to take into account).

    The question for me is – should we expect non practitioners to grasp these or is it merely incumbent upon the practitioners to come up with hueristics to guide a range of professional best practice. As Mike points out, they are not the same thing.

  10. I have written a few articles about ecosystem markets and this concept of additionality is easy to explain superficially but, as you say, extremely difficult to define – what’s more the definitions have to change depending on the type of offset you are talking about. An article entitled “Green Carbon” [] that I wrote for the North American Wholesale Lumber Association about forest carbon offsets under California’s new market addressed this question pretty extensively. It looked at the concept of additionality from the standpoint of someone implementing practices on the ground and the incentives they get. I found this very useful, and the way that the California program has handled this seems to be pretty reasonable. However, it also seems clear that unavoidably, additionality will remain an amorphous concept without a decent monitoring system and means of measuring baselines.

  11. Michael Gillenwater on said:


    Yes, agreed. Jargon will develop no matter what. Hopefully, though, the jargon helps to efficiently convey well understood issues, rather that obscure them.

    And agreed. Materiality and uncertainty are two more key concepts to understand. Uncertainty is, in particular, of interest to me. I helped develop the guidance for the UNFCCC and GHG Protocol on the topic. And of course, there is a connection between uncertainty and materiality.


  12. Michael Gillenwater on said:

    Thanks for sharing. I’ll take a look at the article.

  13. Completely agree that defining additionality is a critical task. For small organizations, it is beyond their scope. There may be political or social factors that affect the process of defining baseline. It all depends on what part you conduct your activities. Still, to standardize the whole system by bringing them into the same platform, the political factors must be reduced that will allow free flow of implementation process.

Leave a Reply

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


* Copy This Password *

* Type Or Paste Password Here *

15,958 Spam Comments Blocked so far by Spam Free Wordpress

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>