Why Did the Clean Development Mechanism Produce Low-Quality Carbon Offsets?

In an earlier post, I described several similarities between the implementation of carbon offsets under the 2015 Paris Agreement and the 1997 Kyoto Protocol’s Clean Development Mechanism (CDM). Based on the significant overlap between the two programs, I asked whether the Paris approach is likely to repeat the well-documented problems of the past. So what about the CDM didn’t work and why?

The main problem is that many CDM projects earned carbon credits for things that were going to happen anyway. This outcome violates a standard known as additionality, a common requirement across carbon markets (including the CDM).

Although technical definitions vary, the basic idea with additionality is that income from carbon credits must cause something new to happen beyond business-as-usual outcomes. Additionality is particularly difficult to establish because business-as-usual outcomes can’t be observed or verified—by definition, counterfactual scenarios don’t happen.

In the absence of direct evidence, carbon crediting methodologies look to other indicators, frequently based on subjective information and sometimes even qualitative narratives.

That’s not to say that we lack good evidence about the quality of project outcomes. For example, a landmark 2016 study commissioned by the European Commission reviewed the CDM’s major market segments and found that only 7% of projected CDM carbon credit supplies had a high likelihood of achieving additionality.

More recently, a group of researchers led by Georgetown University Professor Raphael Calel published a detailed study (here’s an open access version) of CDM-subsidized wind energy projects in India. Professor Calel and his colleagues identified “blatantly inframarginal” projects—those that would have happened anyway—when they matched a wind farm earning carbon credits with another, more difficult project built without carbon credits in the same state and year. The authors found that more than half of the CDM credits came from blatantly inframarginal projects, a lower-bound estimate of non-additionality.

So how did this happen? Part of the answer is that the CDM uses highly subjective methodologies. Prior to a major regulatory revision in 2009, for example, CDM projects could demonstrate additionality by claiming to face qualitative “barriers” associated with their financial feasibility, without providing quantitative evidence or analysis. This was true even for utility-scale power plants financed and built by large companies.

Even after the CDM required a formal quantification of financial barriers, it only took half-measures. The CDM investment analysis tool tells project developers how to calculate their expected financial returns, but clean energy projects need only show that their returns without income from offsets fall below national benchmarks. Notably, projects are not required to show that offset income makes them financially feasible on the same basis. That’s an odd omission: if the investment analysis tool provides a robust basis for determining additionality, then you would expect it to work both ways.

The more fundamental problem with offsets is information asymmetry. In the CDM’s bottom-up process, methodologies are developed by private parties for review and approval by rotating panels of experts. Projects that game these rules will be detected only if regulators know as much or more than project developers and want to act.

That’s a tall order in any sector or country, let alone in multiple sectors across the globe. It is especially difficult when the rules are written in an iterative, decentralized process—as they were in the CDM, and as they could be in the Paris Agreement carbon crediting mechanism again.

Danny Cullenward

Senior Fellow

Danny Cullenward is a Kleinman Center Senior Fellow. He is a climate economist and lawyer, a Research Fellow with the Institute for Carbon Removal Law and Policy, and the Vice Chair of California’s Independent Emissions Market Advisory Committee.