Additionality & Baselines: The Counterfactual That Decides Whether a Tonne Is Real
Every carbon dioxide removal (CDR) credit is, at bottom, a claim about a parallel universe. Additionality asks whether the tonne of CO₂ subtracted from the atmosphere would have been subtracted anyway, absent the credit revenue. Baselines are the quantitative expression of that parallel universe: the reference scenario against which “extra” removal is measured. Get the baseline wrong and everything downstream — the price, the corporate net-zero claim, the registry’s reputation — is built on a number that doesn’t correspond to any physical reality. This is why buyers doing serious diligence spend more time on counterfactual assumptions than on capture chemistry, and why what “net zero” actually means depends on the integrity of the removals it invokes.
The mechanics: three counterfactuals stacked on top of each other
A defensible baseline typically has to answer three questions at once. First, the financial counterfactual: would the project be built without credit revenue? A Direct Air Capture (DAC) plant costing hundreds of dollars per tonne to operate is trivially additional today; a woodlot that a landowner would have planted anyway is not. Second, the physical counterfactual: what would the carbon stock or flux have been on this land, in this reactor, in this ocean parcel, without intervention? Third, the policy counterfactual: is the activity already mandated, subsidised, or standard practice? Each of these can be gamed by adjusting assumptions a few percent, and because baselines are multiplied by project lifetime and hectares, small assumption changes swing issuance by orders of magnitude.
The voluntary forest-offset market is the cautionary tale. Crediting programmes let developers pick reference regions or growth trajectories that made avoided-deforestation projects look heroic against inflated business-as-usual loss rates. When independent researchers reconstructed the counterfactuals with matched controls, a large share of issued credits corresponded to no additional carbon on the ground. The collapse in buyer confidence that followed was not about the trees — it was about the baselines. The life-cycle-assessment literature on CDR makes a related point: without a defensible system boundary and reference case, headline removal numbers are essentially editorial choices.
How the tension bites, pathway by pathway
Engineered removal has the cleanest additionality story and the messiest baseline arithmetic. Nobody builds a Climeworks or Heirloom plant absent credit or tax-credit revenue, so financial additionality is a formality. The baseline fight moves to the LCA boundary: what grid emissions do you assign to the electricity? What counterfactual use did the limestone, sorbent, or waste heat have? A DAC tonne credited against a marginal fossil grid and one credited against curtailed renewables are physically different products. The same holds for bioenergy with carbon capture and storage (BECCS) developers like Svante/Carbon Alpha or Reverion: the biomass counterfactual (would it have decomposed? been burned? left standing?) drives the net number more than the capture rate does.
Biochar inverts the problem. Pyrolysis is cheap enough that some facilities — NetZero in tropical residues, Aymium in North America, Bio-Logical in Kenya, Alcom in Philippine rice husks — could plausibly operate on product revenue alone in some markets. Registries then have to decide whether the whole tonne is additional, some fraction is, or whether only facilities in specific geographies qualify. The baseline for the feedstock is equally contested: the meta-analytic literature on biochar shows agronomic benefits that themselves generate revenue, further eroding the “wouldn’t have happened anyway” claim. Meanwhile the counterfactual fate of crop residues — field burning, mulching, composting — swings the net credit by a factor of two or more.
Enhanced weathering and land-based pathways face the hardest baseline problem of all: natural background flux. When Metalplant spreads olivine in Albania, or Varaha works with smallholders across South Asia, or Grow Indigo and Klim credit regenerative practice changes, the “removal” is a delta on top of a noisy, weather-driven soil or weathering flux that would have existed anyway. Dynamic baselines — re-measured over time rather than fixed at project start — are more honest but they punish developers for good climate years and complicate financing.
The hard parts
Three questions remain genuinely unresolved.
Who sets the baseline, and how often? Static baselines are bankable but drift out of physical reality; dynamic baselines are accurate but hostile to project finance. No registry has solved this cleanly.
How conservative is conservative enough? Every tonne of buffer or discount protects credit integrity and erodes project economics. Push too hard and marginal projects — exactly the smallholder-integrated ones like Green Carbon, Carbon Remove, or Dark Earth Carbon — never get built. Push too little and you rebuild the forest-offset crisis in a new pathway.
What is the policy counterfactual in a world of tax credits? When a jurisdiction pays $180/tonne through a production credit, the private voluntary buyer has to ask what their additional dollar is buying. This will only sharpen as fossil-phaseout pathways tighten and CDR moves from voluntary novelty to compliance instrument.
Additionality is not a technical detail. It is the load-bearing assumption under every removal claim, and the pathway that treats it as a compliance checkbox rather than a live scientific question is the pathway that loses its market next.