The central idea behind a carbon offset is that it can substitute for GHG emission reductions that an organization would have made on its own. For this to be true, the world must be at least as well off when you use a carbon offset credit as it would have been if you had reduced your own carbon footprint. When people talk about the “quality” of a carbon offset credit, they are referring to the level of confidence one can have that the use of the credit will fulfill this basic principle.
The concept sounds straightforward, but it is challenging to guarantee in practice. Quality has two main components. First and foremost, a quality offset credit must represent at least one metric tonne of additional, permanent, and otherwise unclaimed CO2 emission reductions or removals. Second, a quality offset credit should come from activities that do not significantly contribute to social or environmental harms.
A variety of terms are frequently used to define quality criteria for carbon offsets, including that associated GHG reductions must be “real,” “quantifiable,” and “verifiable.” Most of these terms have their origin in regulatory criteria established for air pollutant credits under the U.S. Clean Air Act (going back to 1977). However, these terms have distinct regulatory meanings under U.S. law that do not always translate meaningfully to carbon offsets. The term “real,” for example, has no commonly agreed definition across carbon offset programs and standards, and is often used as a vague catch-all.
For this guide, therefore, we have distilled the essential elements of carbon offset quality down to five criteria. In short, quality carbon offset credits must be associated with GHG reductions or removals that are:
- Not overestimated
- Not claimed by another entity
- Not associated with significant social or environmental harms
Carbon offset programs were created with the intention of ensuring the quality of carbon offset credits (see Carbon Offset Programs for more info). In the remainder of this section, we describe the approaches carbon offset programs use to address the quality criteria listed above. As indicated in High-Quality Offsets, however, many observers believe that carbon offset programs have a mixed track record. Part of the challenge is that offset quality is not black and white. The multiple criteria involved – plus the fact that critical criteria like “additionality” are a matter of confidence rather than absolute truth (see below) – means that quality exists along a continuum. Carbon offset programs, by contrast, are forced to make a binary decision: do they issue an offset credit or not? Most carbon offset programs will say that every credit they issue is equally valid, but buyers should feel justified in questioning this assertion. Think of scoring the quality of an offset on a 100-point scale. A carbon offset program may decide to issue credits for every GHG reduction that exceeds a score of 50. But as a buyer, is a score of 51 really “good enough”?
Astute buyers will understand this difficulty and actively seek out higher quality offset credits. For each offset quality criterion below, we highlight some questions that buyers can ask about specific offset projects to better ascertain their relative quality. Even for sophisticated buyers, however, getting detailed answers to these questions may be difficult. Thus, in Strategies for Avoiding Low-Quality Offset Credits, we identify a range of strategies buyers can use to steer clear of lower quality offset credits and improve the chances of acquiring higher-quality credits.
 This condition applies to GHG emissions, as well as to other social and environmental impacts. If global GHG emissions would be no greater as a result of using a carbon offset credit instead of reducing your own emissions, then the credit is said to preserve “environmental integrity” (Schneider and La Hoz Theuer 2019). However, it is also important that offset projects do not cause significant social or (non-climate) environmental harm. Both are important for offset quality.
 See Gillenwater (2012).
 For an in-depth discussion of these ideas, see Trexler (2019).