I see carbon credits as doing two things: (1) financially incentivizing corporations to control their emissions. I see carbon credits as a tax where the money goes to carbon negative organizations instead of the government. Would we be better off if this "tax" were funneled elsewhere? Probably. But, organizations like the Mass Audubon are not a terrible place for it to go. I'd be curious what the research shows on whether the laws in California have incentivized corporations to reduce emissions below the prior trajectories they were on. (2) producing a lot of data on carbon sources and carbon sinks that we did not have before. What can the field do with the additional data? Are the corporate and forest survey data made sufficiently public to enable researchers and innovators to get additional value out of it?
The other problem with these carbon schemes is they in effect, prevent the restriction or prohibition of certain technologies. For example, why ban supersonic transport due to their carbon emissions when the aviation industry would argue that the whole carbon schema will effectively take society to the same place?
It's a ridiculous argument though, because if supersonic transport were banned, the entire pie would be smaller to begin with, and there would be fewer carbon credits created for the marketplace as a whole. Same argument for bitcoin mining, and NFT.
Credits should only apply to valuable economic sectors in which it may be cost-prohibitive to otherwise achieve net-zero emissions. And even then, credits should be rare and hard-to-get, scrutinized like a 100 carat diamond, and verified only by a government agency impervious to political influence.
Great post. I'd just say in particular for approaches that are not directly capturing carbon... it's actually a much harder problem than an A/B test on a website.
For example, the Brazilian study you reference uses synthetic controls, which are A COUNTERFACTUAL but not the only one. How well they work depends entirely on how well you believe those controls were constructed. And synthetic controls (as opposed to other techniques, like matching) are tough to investigate because they don't actually exist :-P .
Shameless plug: I wrote about some related issues surrounding evaluations (called SUTVA violations) in a recent post.
I see carbon credits as doing two things: (1) financially incentivizing corporations to control their emissions. I see carbon credits as a tax where the money goes to carbon negative organizations instead of the government. Would we be better off if this "tax" were funneled elsewhere? Probably. But, organizations like the Mass Audubon are not a terrible place for it to go. I'd be curious what the research shows on whether the laws in California have incentivized corporations to reduce emissions below the prior trajectories they were on. (2) producing a lot of data on carbon sources and carbon sinks that we did not have before. What can the field do with the additional data? Are the corporate and forest survey data made sufficiently public to enable researchers and innovators to get additional value out of it?
The other problem with these carbon schemes is they in effect, prevent the restriction or prohibition of certain technologies. For example, why ban supersonic transport due to their carbon emissions when the aviation industry would argue that the whole carbon schema will effectively take society to the same place?
It's a ridiculous argument though, because if supersonic transport were banned, the entire pie would be smaller to begin with, and there would be fewer carbon credits created for the marketplace as a whole. Same argument for bitcoin mining, and NFT.
Credits should only apply to valuable economic sectors in which it may be cost-prohibitive to otherwise achieve net-zero emissions. And even then, credits should be rare and hard-to-get, scrutinized like a 100 carat diamond, and verified only by a government agency impervious to political influence.
Great post. I'd just say in particular for approaches that are not directly capturing carbon... it's actually a much harder problem than an A/B test on a website.
For example, the Brazilian study you reference uses synthetic controls, which are A COUNTERFACTUAL but not the only one. How well they work depends entirely on how well you believe those controls were constructed. And synthetic controls (as opposed to other techniques, like matching) are tough to investigate because they don't actually exist :-P .
Shameless plug: I wrote about some related issues surrounding evaluations (called SUTVA violations) in a recent post.
https://databetter.substack.com/p/what-kermit-got-right-carbon-credits