To achieve the necessary scale for carbon dioxide removal (CDR) to meaningfully address climate change, we cannot rely on market signals or generalized climate policies alone; instead, we must embrace targeted industrial policies.
This means building real coalitions that will fight for these technologies because they see direct, tangible benefits, not just generic atmospheric benefits. It means designing policies that create durable political and technological lock-ins and deliberately lower barriers to growth, such as infrastructure, monitoring, and verification costs.
Given the central role governments will have to play, as CDR is one of the quintessential public goods, this approach carries the risk of government missteps and rent-seeking. But historical examples from renewable energy and successful industrial policies in other sectors clearly demonstrate that proactive and iterative policy support, grounded in concrete benefits, is not only effective but essential.
History tells us that big, complex industries don’t scale by accident, they succeed through focused and politically smart interventions. To realize the full potential of CDR, we must intentionally build it into a robust, politically resilient industry, using lessons learned from past successes and failures to guide policy that rewards learning by doing, weeds out unsuccessful approaches, and continuously adapts to real-world challenges and opportunities.
Creating a CDR industry
CDR is less a technology than a collection of approaches. It’s less like talking about solar and more like talking about renewables. In that space, there are debates over what should count as renewable, debates about how to measure and transfer ownership of renewable claims, permitting issues, mud slinging, the involvement of some of the world’s largest emitters, geopolitical tensions, but above all, there is progress.
If CDR is more like renewables than a single energy technology, we should look carefully at how that set of technologies reached maturity. Some progress came through technology‑agnostic markets, but much of it was driven by targeted, technology‑specific interventions such as Germany’s feed‑in tariffs for photovoltaics rather than relying solely on general carbon prices.
Lots of discussions about CDR focus on narrow questions, the impact of a given policy proposal or the metrics of a new technological approach. These are important questions but it’s useful to step back and look at some larger questions that focus on what it looks like for CDR to follow in the footsteps of other successful industries and rise from its current position to gigatons.
Why generic climate policies aren’t enough
There’s a general problem with climate policy: high‑ambition policies need coalitions to advocate for them, but potential winners are often diffuse while losers (e.g., incumbent emitters) are concentrated. This has often pushed policy toward incrementalism.
Carbon pricing illustrates this. Many jurisdictions that later adopted carbon prices first used targeted decarbonization tools like renewable portfolio standards. When prices were enacted without these more targeted policies coming first they were often weak, nudging marginal measures like fuel efficiency or fuel switching rather than catalyzing system change.
These weak price signals don’t create enough of an incentive to drive the investments in system changes required for deep decarbonisation. Without this style of investment we don’t see true changes in interest groups who could then advocate for the ratcheting up of these policies over time. A weak price could be followed by a stronger one but getting there requires the very coalitions that weren’t present to advocate for a stronger price mechanism in the first place.
One approach to durability is to target policies with some form of bipartisan support today. Bipartisan support is no bad thing but our focus is on policies that produce hard‑to‑reverse techno‑economic and socio‑political changes. In this sense you can view their durability less due to bipartisan support and more because they lead to outcomes that are robust to inevitable partisan forces.
Socio-political changes can be created through forming “winning coalitions” that benefit directly from, and therefore advocate for, certain climate policies. The hope is that over time these coalitions form the basis of support for ratcheting climate ambition and make advancing decarbonization in the future more politically feasible.
In the CDR space you could imagine farming groups becoming key supporters for enhanced weathering after they see more benefits from reduced soil acidification and crop yields, as well as direct payments for the carbon benefits. It is these groups who could then form a key constituency in pushing for things like farming subsidies for EW rather than this mainly being done by a set of more purely climate focused advocates.
Techno‑economic improvements aim to reduce the real cost of technologies directly or through building enabling infrastructure. These changes aim to make technologies more attractive in their own right because they are now cheaper but also reduce the political capital needed to push for a given level of policy ambition.
Building durable CDR support
For CDR to scale, it needs political and private coalitions that champion it for reasons beyond “the IPCC says we need this.” Co‑benefits are not decorative; they’re central to coalition formation. It will be far easier for enhanced weathering to scale if farmers are genuinely bought into its yield and cost advantages than if these are treated like a nice to have add‑on. Even if you care only about removing tonnes, co‑benefits support delivery: where projects are wanted and win support, they face less opposition and deliver more reliably over time.
Particularly in a world where many groups are facing the fact that they are going to have to adapt to a warming climate the more mitigation activities can be tied to adaptation outcomes the more naturally we are to see constituencies in support of these activities appear1.
Crucially, instruments that create stakes, regulations or direct payments that change firms’ operations often build more durable coalitions than generic carbon pricing. If you’re just slapped with a fee, you lobby to kill it; if your business model and capex have shifted under a program, you have a vested interest in sustaining it. 45Q is a good example of this; large‑scale infrastructure exposure among politically connected companies likely contributed to its resilience across administrations. Another emerging avenue on this theme is industrial integration of CDR that drives process changes firms will defend over time to prevent their investments from losing value.
Importantly, many of the above approaches will fail what today is often described as the north-star policy goal, being technology neutral. There are some intervention types that we want to be big tent, like ETS regimes, but if coalition building is done at the level of individual technologies, then we shouldn’t shy away from policies that also benefit a narrow set of technologies.
The complement to coalition‑building is driving changes in the technology itself to make coalition‑building easier. Any actions that reduce the real cost of CDR are largely no‑regrets moves.
For most providers, three cost buckets dominate:
- CapEx: building facilities
- OpEx: energy and other inputs
- Monitoring, Reporting, and Verification (MRV): measuring and verifying outcomes
The point here is not to artificially lower prices through subsidies (those can be useful for other reasons), but to push down underlying costs. There are lots of policies that we could discuss to reduce CapEx or OpEx costs but I wanted to spend some time talking about the third bucket: MRV.
This is a fairly unique aspect of carbon removal and gets at how it is both a physical activity but also, today, an informational product. With something like oil if the end product is substandard this can be directly measured. For carbon removal you might end up with 1,000t of carbon stored but if you don’t properly track the emissions, leakage, etc then this gives you a very incomplete picture of the impact of any activities, to ensure these are all accounted for requires robust MRV.
In CDR many pathways have MRV as a key part of their cost stack through areas like sampling and analysis or simply by using techniques that have uncertainty bands which can lead to lower quantities of credits, and therefore higher associated costs, if the crediting process takes into account uncertainty2. Bringing MRV costs down is therefore a powerful lever to shift pathways to durably cheaper levels.
Costs are likely to remain an ongoing challenge with CDR compared to other technologies like renewables. In the latter case once you could get the levelised cost of energy below that of other options on the market things really take off all on their own. For CDR given its nature as a waste disposal service there is less of a natural strike price to try to aim for. However, given most futures with large scale CDR envisage large scale public support, anything that can be done to reduce the amount these policies would cost is a true no regrets option.
Implementing industrial policy
When it comes to implementing these types of policies, we need to find mechanisms that are strong enough to impact real world incentives while also being aware that what the right technologies to back, and how we want to back them, may change over time. We can think of there as being three stages in the industrial policy playbook: support → evaluate → adjust.
Support. To create concentrated behavior‑changing benefits, you want direct, targeted policies over shallow, broad ones that fail to shift investment decisions. Reward behaviors that change an industry’s structure: long‑term offtake, storage network build‑out, siting with community upside, and integration into industrial processes.
Evaluate. Technologies that miss promises, or that have met their goals and no longer need the same level of support, should face ramp‑downs. The hope is that these early technology specific support mechanisms lead to the conditions where we have the conditions to create more robust generalisable policies like higher carbon prices, or put another way carrots today should buy the credibility to add sticks tomorrow.
Evaluation in carbon removal can be particularly challenging, given, as we mentioned previously, it is both a physical and informational product that is being produced. This evaluation process could include more work on benchmarking expected technology performance. This may not account for all the unexpected things that happen along a technology’s deployment journey but at least having the best ex-ante expectations of how a technology was going to perform is a useful signpost to go back to if the technology massively diverges from these expectations.
Aside from forecasts simply building up the expertise needed for evaluation should not be underestimated. Having enough capacity in whichever group is responsible for supporting a technology is crucial. You can see good examples of this with institutions like the California Air Resources Board or, in a different domain, with independent central banks who have built up the politically insulated human capital to evaluate monetary policy options to advance their chosen objectives.
Adjust. Any policies today should be thought about just as much as tools to help us in the process of discovery as much as a signal of support. We don’t know what technologies will truly have what it takes to scale so we want to try and support a broad basket of them but if things aren’t working cut our losses. It can help in this endeavour if the groups funding these technologies have explicit sequencing theses about what capabilities will be needed when, so today’s actions can be part of a roadmap to giving us the information we need for tomorrow’s options.
Avoiding the pitfalls
A classic objection to the types of interventions I’ve referenced above is that governments shouldn’t pick winners and will overpay for whatever climate impact occurs versus a first‑best carbon price if they do. Theoretically, yes, targeted policies transfer higher near‑term rents to specific groups and will often look inefficient in a static analysis. But the “economic efficiency of a policy is irrelevant if it is politically infeasible”. If extra rents accelerate learning and coalition formation that unlock more ambitious climate outcomes, and thereby lower long‑run costs, second‑best solutions can be best over time.
It’s also worth recalling the empirical record: most jurisdictions that enacted carbon prices had already leaned on targeted policies; where prices passed, many were weak and didn’t trigger the system‑level investments needed to build ratcheting constituencies.
The second concern is rent‑seeking lock‑in: policies that stick around after they’re no longer optimal, or worse, entrench dead ends. We’ve seen both in the past, renewable providers have lobbied against carbon pricing that might jeopardize generous subsidies and corn ethanol has maintained ongoing political and subsidy support despite questionable long run decarbonisation benefits.
CDR has two specific lock‑in risks:
- The classic technology/pathway lock‑in risk where pathways are backed that can’t scale in a cost effective manner
- A more CDR specific lock‑in risk around integrity where accounting regimes that overstate net removals could become entrenched, tilting the tables in favour of certain pathways and in the long run jeopardising both the credibility and climate benefit of the sector.
The debate over “what counts as CDR” is so lively in part because of the second lock-in risk. If there is a misclassification of what should count as CDR and that leads to a diversion of scarce resources to pathways that might not be able to deliver against the goals of CDR3, then we might lock ourselves out of having the full climate toolkit we need.
Part of the unique challenge in CDR is that we care about the net outcome: it’s not enough to put 100 tCO₂ underground; we must also count emissions used to achieve this as well as how these compare to the baseline emissions. Protocols and MRV choices can create de facto standards that skew more towards quantity or quality and it is key at this stage of the market that we try to build coalitions around quality.
If early coalitions form around volume at the expense of integrity, they can be hard to unwind and may ultimately erode trust and stall the market. This is not just a hypothetical concern but one that we have seen play out with the legacy VCM through practices like lax additionality or baseline rules that received justification in the name of supporting projects that had strong social or ecological benefits but ultimately they undermined the trust and quality of the core assets.
Ultimately, picking winners is not necessarily the goal we should be aiming for; rather, we should try to design systems that can identify winner and let go of the losers.
We need CDR, but we won’t get it at scale by accident. To build a new sector from the ground up will require building coalitions that advocate strongly for the underlying technologies across the political spectrum as well as deploying funds to help drive down the costs of the underlying approaches to make the lives of these coalitions even easier.
Achieving our goals might require departing from the technology neutral lens that is often aimed for but if we can put good systems in place to evaluate and adjust the approach to CDR funding over time we can ensure a healthy scale up of the entire sector. Ultimately, we want to design an ecosystem that supports ratcheting ambition rather than just hopes for it.
- This does raise some of the challenges around financial additionality. Approaches with the richest private co‑benefits may be the most challenging to fit into an offsetting based financial additionality framework. More so these same approaches are also the ones which may have the best political coalition building potential. ↩︎
- For example, at Isometric we would look at the expected CDR and the standard deviation based on the uncertainty associated with different datapoints, projects would then be credited one standard deviation below the mean to ensure crediting is conservative. This means that for two projects with the same expected removals the one with higher uncertainty will receive fewer credits. ↩︎
- Offsetting hard to abate emissions, helping us claw our way back from temperature overshoot and possibly being used to deal with legacy emissions ↩︎
Neil Hacker 