Nike executives made a startling discovery in the early 1990s.

The gas pumped into the soles of the Nike Air series of athletic shoes created an exceptionally resilient cushion. But the large, tightly-linked molecules of sulfur hexafluoride also make it the planet’s most powerful greenhouse gas — 24,300-times more potent than carbon dioxide. Air Jordans and Nike Inc.’s other popular shoes, it turns out, were overheating the planet as much as half the cars in its home state of Oregon.

The sneaker giant spent tens of millions of dollars over the next 14 years to find a replacement, eventually declaring in 2006 that it had eliminated all heat-trapping gases from its shoes. Instead of fading into the annals of climate history, however, Nike’s decades-old environmental problem has suddenly become relevant again — and it’s exposing a key weakness in today’s carbon market.

That’s because Nike, as it finished its gas substitution work two decades ago, decided to turn its sulfur hexafluoride reductions into a carbon-offset project. This allowed the company to sell credits for its eliminated emissions to other entities, who could then subtract that amount of pollution from their own environmental reports. Nike was awarded nearly 8 million carbon credits — roughly equivalent to the annual emissions of a large coal plant — by the predecessor to the American Carbon Registry (now called ACR), one of a handful of nonprofit registries that underpin the carbon market.

Nike’s project gained little traction at the time; and nobody has used any of these credits in nearly a decade. But the project re-emerged earlier this year, with more than 1.2 million Nike credits surfacing, just as the market for carbon offsets attempts to fix the chronic flaws that have sent it into a spiral.

The Nike credits showed up in the ACR buffer pool, which is a large cache of credits that acts as an insurance policy in case one of its carbon projects suffers a massive loss. If a wildfire rips through a protected forest, for instance, credits from the buffer pool replace the incinerated ones, thus preserving the market’s atmospheric claims.

All of the major registries in the carbon market use buffer pools to protect against such setbacks. But these insurance mechanisms have recently come under scrutiny from academics and think tanks for not stockpiling enough credits to guarantee against reversals, particularly as climate change exacerbates forest-damaging wildfires, droughts and pests.

ACR had long declined to say what credits are in its buffer pool. When it finally revealed the contents earlier this year, Nike credits accounted for 19% of the insurance pool, more than twice as much as any other project. ACR said they added the Nike credits to the buffer pool, but they declined to explain why. 

Nike’s project has long been controversial, even in the market’s infancy. “It was a somewhat notorious project for those of us in the North American carbon market 15 years ago,” says Derik Broekhoff, a senior scientist at the Stockholm Environment Institute.

Carbon credits are supposed to represent emissions reductions that happen because of the promise of payments — a concept known as “additionality.” This standard sets credit-worthy activity apart from climate-friendly activity that would have occurred anyway. Think of a nature preserve that refrains from clearcutting its own land. Since those trees were already safe, protecting them wouldn’t make for credible offsets. Additionality is what allows a purchaser of carbon credits to claim their payment caused a change in atmospheric emissions.

This wasn’t the case with Nike. The sneaker giant had first promised in 1997 to phase out planet-warming gases from its shoes amid blowback from green groups and European regulators. In a series of emails, Nike officials confirmed that their efforts to eliminate sulfur hexafluoride were driven by future regulations and weren’t influenced by carbon payments.

All of this should mean Nike’s emissions reductions are basically worthless as carbon credits. And yet Nike’s solution for its own Air problem now helps backstop the long-term integrity of scores of other carbon projects.

“If a project isn’t additional, it has no business being in a buffer pool,” says Mark Trexler, a consultant who helped Nike measure its carbon footprint for the first time in the 1990s.

Nike officials say the company ultimately decided against selling most of the credits. “Nike is proud of our emissions-reduction efforts and has continued to be at the forefront of industry action,” said a spokesperson in an email.

ACR officials, meanwhile, declined to be interviewed. But ACR’s representatives said in emails that Nike’s Air effort from two decades ago remains a credible carbon project. Instead of evaluating the motivations of each individual project on a case-by-case basis, ACR and other registries typically require projects to clear thresholds, such as proving their climate-friendly activity wasn’t required by law or undertaken to increase profits. Nike’s project, they said, met these requirements.

ACR officials have also emphasized that it has “robust, enforceable systems in place” to guard against reversals in its scores of carbon projects. No ACR project has relied on the buffer pool after suffering an unintentional setback, though that could soon change as a result of a July wildfire that burned through one of its forest projects in California.

This exposes a problem facing carbon markets, which have struggled against a wave of reports showing many projects deliver fewer climate benefits than advertised. Buyers have pulled back, with voluntary purchases of carbon credits plummeting from $2 billion in 2021 to $700 million last year, according to Ecosystem Marketplace. 

Some type of long-term insurance is critical for carbon credits to work. That’s because these credits are supposed to negate the impact of burning fossil fuels, which deposits new carbon dioxide into the atmosphere for hundreds or thousands of years. For the math to pencil out, carbon projects need to eliminate CO2 for a similarly long period of time.

But many offset projects store CO2 in fragile ecosystems prone to re-releasing heat-trapping gases. “The nasty thing about carbon dioxide is that when we pump it into the atmosphere, it stays there, effectively, forever,” says Grayson Badgley, a research scientist at CarbonPlan, a nonprofit that studies climate solutions. “That means, if you want to create an offset, you need to have something that resembles that permanence.”

Registries rely on buffer pools to address this. Any project that can suffer a reversal is required to put some credits, usually between 10% to 20% of their total, into a shared insurance account. If a wildfire then wipes out a forest project and sends thousands of carbon credits up in smoke, an equivalent number of credits would be canceled from the buffer pool, thus making the atmosphere whole.

Researchers have been warning about the dangers of under-capitalized buffer pools. Academics at the University of California at Berkeley found last year that a large group of forest projects on Verra, the world’s biggest carbon registry, underestimated risks for fires and other natural disturbances by a factor of ten. CarbonPlan likewise found that forest projects in California’s carbon market had already suffered more than 95% of the fire losses that were expected to occur over 100 years.

This summer’s fire season was particularly bad. CarbonPlan identified at least six forest carbon projects in the US that suffered substantial damage in fires, engulfing about 77,000 acres and likely causing the reversal of hundreds of thousands of carbon credits.

Human mistakes can also take a toll. One of the world’s biggest offset projects, a forest-protection effort in Zimbabwe known as Kariba, recently withdrew from Verra’s program after vastly overestimating its climate benefits. Nestle, McKinsey and scores of other companies had purchased over 20 million Kariba credits to meet their environmental commitments. Depending on what happens with the project, Verra may eventually need to use as much as 40% of its buffer pool to make up for these credits.

“When the buffer pool inevitably fails, which it will…nobody is standing behind it,” says Nandini Wilcke, co-founder and chief operating officer of CarbonPool, one of several companies working to develop new types of insurance for carbon projects. “What happens to all the other projects that should be covered by the pool?”

Verra officials say that their buffer pool remains strong and that they’ve updated risk measurement tools to better consider changing climate impacts on their projects. California officials, meanwhile, say their buffer account is robust and they are studying whether it needs to be adjusted as the wildfire seasons become more intense. Operators of the Kariba project didn’t respond to requests for comment.

Nike isn’t the only concerning project to show up in ACR’s buffer pool. Three Brazilian renewable energy projects account for over 1 million credits, or 16% of the total. Credits from renewable projects have long been criticized because clean energy has often become cheaper to build than fossil-fuel alternatives; and it’s unlikely small carbon payments tip the balance to get new power plants built.

“The wind farms would be built anyway,” said Haroldo Maia, the former financial director for Santos Energia, developer of one of the Brazilian projects in ACR’s buffer pool. Carbon credits, he added, represent “extra revenue that wasn’t foreseen in financial modeling of the company.” (The ACR spokesman said that credits from the Brazilian renewable-energy projects were “verified to meet additionality and other requirements.”)

It’s not clear exactly how the Brazilian projects arrived in ACR’s buffer account. The registry declined to comment on that. Projects tied to wind turbines aren’t required to put anything into the pool, since the climate benefits involved aren’t vulnerable to reversal.

ACR, like some other registries, allows project operators to purchase other credits to stick into the buffer pool in lieu of their own. 

For experts like Grayson Badgley of CarbonPlan, this lack of clarity isn’t satisfactory. “The buffer pool is supposed to be a replacement: ‘If something happens to your credit, I’m going to replace it with something that’s just as good,’” he says. “We have to be much more serious about permanence.”