thanks for your reply—a couple of reactions, hopefully quite nuanced (I agree with you a bunch, and disagree on others).
1. I did not mean to imply that you do not consider this possibility at all (you do!), but rather was reacting to the general rationale of the piece of using present-day evidence on behaviors as informative with regards to long-term prospects.
One could go away from your piece thinking there is a lot of evidence that should have one update against long-term PTC and alternative protein transition which seems not warranted by most of the kind of evidence you cite (I think).
2. I should note that my exposure to alt proteins is mostly from engaging with alt proteins arguments in the context of climate, I am not an animal welfare person so I am not so aware of how the issue has been framed in these contexts. From studying the modeling from groups like BCG and Blue Horizon it has always been clear to me that most of the benefits accrue more than a decade hence and that this is not a rapid transition issue and this is the frame from which I commented that I think current experimental evidence is not that informative. I agree it is misleading if the dominant narrative is one of rapid transition now.
3. I also agree with you that some of the arguments of the form “alternative proteins is definitely the most cost-effective animal advocacy opportunity” are likely overstated because it will take a combination of cheap substitutes (alt proteins), persuasion, norm change, etc. and there is no argument I am aware of that establishes the marginal dominance of further accelerating alt proteins.
4. That said, I do think we know more about how the transition away from animal proteins will go (conditional on it happening) than what you seem to suggest, if I understand you correctly. In particular, it seems clear that a substitute that is close to PTC conditions will be needed, where then regulation, norm change, etc. can do the rest to lead to widespread adoption.
At least that is the what the outside view from other recent transitions to socially more beneficial technologies would suggest, i.e. protecting the ozone layer, scaling renewables, electric cars, etc. and, conversely the failure to do achieve other transitions as of yet (say, industrial decarbonization, meat consumption, etc.). In none of those cases was it only the technological change towards PTC conditions (or their respective equivalent) that drove the change, but it was always necessary. And in all of those cases there were also norms and other non-PTC patterns against diffusion of alternatives that were more overcomeable (?) once we were closer to PTC-like conditions.
5. I think a crisper way to state 4 would be to say “from what we know from other transitions, we know that reaching a state close to PTC explains a lot of the variance in adoption so it seems reasonable as a best-guess prior that this will be the case for the animal protein transition as well.” That doesn’t mean other factors don’t matter, but I think it puts thing in perspective with regards to relative importance (again, this makes no statement about the marginal effectiveness of different actions).
Thank you, I appreciate the nuance! [Also, I realize it’s a long paper, so I quote some relevant passages, but apologies if you already read them. I figure it might help other folks following our thread as well.]
“One could go away from your piece thinking there is a lot of evidence that should have one update against long-term PTC” Reasonable, although I did try to avoid this and emphasize the results apply to current consumers. So I also agree it “seems not warranted by most of the kind of evidence you cite” and specifically didn’t cite evidence that focused on future prospects to maintain a narrow focus on current prospects. That said, all else equal, I do think a worse current situation is evidence in favor of a worse future situation.
That makes sense! I want to spend more time with those two reports if I have more time to focus on long-term plant-based meat prospects; thanks for pointing me towards them specifically.
Concur.
I think this is probably the crux of the our disagreement :) First, I think the PTC conditions you’re referring to are under-specified as I elaborate here, so it’s hard to refute without knowing specifically what you have in mind for PTC. (In the paper, I tried to focus on particular cases that limit the definitional issues or relied on working definitions that I specifically criticize, like passing blinded taste tests.) Second, I don’t think the sustainable innovations reference class is especially compelling: for example, most people had ~no understanding of CFC in refrigerants and didn’t think they were a basic necessity of a healthy life. Similarly, few people celebrate holidays that explicitly focus on the act of consuming fossil fuels. In contrast:
People feel a peculiar personal attachment to meat (Graça et al., 2015), believe that meat is necessary for health, feel that meat consumption is socially normative, and perceive meat as a nice and natural component of a healthy diet (Piazza et al., 2015). [And celebrate Thanksgiving where they ritualistically eat a turkey :)]
Some of my other work has touched on what I think is a more relevant reference class: plant-based analogs to animal products:
Our research has reviewed estimates of cross-price elasticities between margarine and butter (Mendez et al., 2023) and plant-based and dairy milk (Mendez & Peacock, 2021). The results suggest that behavior might be inconsistent across studies. Many estimates suggest that decreased margarine or plant-based milk prices result in increased consumption of the corresponding animal product (known as complementarity, the opposite of substitution).
“From what we know from other transitions, we know that reaching a state close to PTC explains a lot of the variance in adoption” indeed is the core claim that I don’t think there’s actually much clear evidence on. Attributing causation in these sorts of transitions is very difficult; I try to lay out some of the challenges for PTC in particular the last paragraph here. I’d be interested what evidence supports this claim.
thank you—strongly upvoted for quality of exchange!
In the interest of time (this has to be my last comment), I ignore the smaller disagreements and focus on what seem like the two cruxes we have here (opposite sequence in your comment, but I think answering in this order is easier here):
(a) Does PTC or PTC-likeness causally drive adoption?
(b) Are clean energy technologies a good comparator?
On (a), here is a visual from the latest IPCC report:
Of course, correlation does not equal causation, but we know from many richer accounts than those simplistic curve displays that the story usually goes (a) high-income country heavily subsidizes R&D and/or deployment of tech, (b) tech gets cheaper, (c) we get a self-amplifying dynamic that drives cost reductions and adoptions, (d) an increasing share of adoption is in countries without those high subsidies, i.e. in settings where the cost reductions (P) but also other improvements (e.g. C-like range for electric cars) drive increased adoption, i.e. it is not just “green subsidies everywhere, all at once” but rather “green subsidies drive cost reductions that enable global diffusion”.
For example, I don’t think anyone doubts that solar will expand massively and that this was causally enabled by cost reductions which were a function of early investments. It is clear that the world could have turned out different here for example if the conservatives had won the German elections in 1998.
On (b), if I understand you correctly you seem to be saying that there are a lot of food-specific considerations that make food special and clean energy comparators inappropriate. That may be and I am not a food system expert enough to weigh the Thanksgiving holiday vs. other forms of cultural lock-in for fossil fuels.
But I would be a bit cautious here as well because there are also many ways in which the food transition is easier than the energy transition so I think a list of ways in which the food transition feels less tractable feels incomplete. Here are a couple of ways how food could be easier: alternative proteins are a better meat replacement than renewables are replacements for coal (no equivalent to intermittency), the food industry is smaller and less powerful than the energy industry, changing protein sourcing is easier than changes in the energy system that require more infrastructure, etc. The point here is that I think focusing on particular considerations only becomes really convincing, I think, when you do a fairly complete accounting in all directions.
My personal take is that including evidence from technologies that have undergone those transitions feels important and that the lack of comparability because they are of a different domain is indeed a limitation but that the effects from undergone transitions provide important additional evidence to form expectations about the future.
Ditto, really appreciate your taking the time to so thoughtfully engage. :) A good day on the Forum! I’ll try to wrap up here as well.
(a) Thanks for this reference—I wasn’t aware of it! This definitely seems like useful evidence in the right direction and I agree with the XKCD’s comic sentiment. That said, it seems like there are still many possible contingencies where price might be a partial rather than full cause. This seems like a ripe area for further research.
(b) I agree, my list is incomplete, and these are good considerations. By the same token, I am no expert in clean power, so hadn’t thought about some of these challenges like intermittency. Along the same lines of this issue, focusing only on successful technologies also introduces a bias—for example, I imagine a similar graph as above which included nuclear would tell a different story.
I also agree there’s evidence to be gleaned from studying other sustainability technologies; we certainly shouldn’t ignore these other transitions. But I would like to see them studied more rigorously and systematically, including positive and negative instances; analogies and disanalogies; and contemporary techniques in causal inference.
I also think that the energy tech analogy might be useful, in particular the case of solar panels, which, unlike nuclear and other energy sources, are also consumer products that went from “rich persons vanity project” to “you’d be dumb not to buy one”.
Decades ago, solar cells were highly expensive, and mainly used for niche applications. There was environmentalists pressure towards clean energy, but the high cost meant only a few wealthy enthusiasts would undergo the switch, and the industry was small and non-influential.
The environmentalist movement was influential in changing this in several ways, first by getting funding for R&D into developing cheaper, more efficient solar panels, second by subsidizing solar energy so it was more cost competitive for consumers, third by encouraging governments to directly build solar farms, and fourth by increasing public support for clean energy so people felt good about buying solar.
The result of this was a feedback loop. Solar was cheaper to buy, and more fashionable, so more people bought it. And the more people bought it, the cheaper solar got, thanks to mass production techniques, so the more people bought it. And then the industry became large and influential. Eventually, solar became cost-competitive without subsidies in many places, so consumers felt it was a good investment, and then everyone in the neighborhood had it, so it became a normal thing that everyone did. (I’m thinking here of parts of australia where pretty much everyone has solar).
I wonder if you polled people about switching to solar in 1990, whether they might seem as reluctant as people are about substitute meat today.
The big difference between PTC meat-substitutes and solar might be in the effect of mass production: could a “feedback loop” of cheaper meat-substitutes causing more consumption causing cheaper meat-substitutes occur? I think there could also be a social feedback loop, where the more people go veggie, the more normalised it becomes, leading more people to go veggie, etc.
In my (non-expert) opinion, PTC in addition to public pressure and opinion changing movements could be much more effective than either of the two on their own. It seems obviously easier to persuade people to make easier changes, but you actually have to persuade them.
Hi Jacob,
thanks for your reply—a couple of reactions, hopefully quite nuanced (I agree with you a bunch, and disagree on others).
1. I did not mean to imply that you do not consider this possibility at all (you do!), but rather was reacting to the general rationale of the piece of using present-day evidence on behaviors as informative with regards to long-term prospects.
One could go away from your piece thinking there is a lot of evidence that should have one update against long-term PTC and alternative protein transition which seems not warranted by most of the kind of evidence you cite (I think).
2. I should note that my exposure to alt proteins is mostly from engaging with alt proteins arguments in the context of climate, I am not an animal welfare person so I am not so aware of how the issue has been framed in these contexts. From studying the modeling from groups like BCG and Blue Horizon it has always been clear to me that most of the benefits accrue more than a decade hence and that this is not a rapid transition issue and this is the frame from which I commented that I think current experimental evidence is not that informative. I agree it is misleading if the dominant narrative is one of rapid transition now.
3. I also agree with you that some of the arguments of the form “alternative proteins is definitely the most cost-effective animal advocacy opportunity” are likely overstated because it will take a combination of cheap substitutes (alt proteins), persuasion, norm change, etc. and there is no argument I am aware of that establishes the marginal dominance of further accelerating alt proteins.
4. That said, I do think we know more about how the transition away from animal proteins will go (conditional on it happening) than what you seem to suggest, if I understand you correctly. In particular, it seems clear that a substitute that is close to PTC conditions will be needed, where then regulation, norm change, etc. can do the rest to lead to widespread adoption.
At least that is the what the outside view from other recent transitions to socially more beneficial technologies would suggest, i.e. protecting the ozone layer, scaling renewables, electric cars, etc. and, conversely the failure to do achieve other transitions as of yet (say, industrial decarbonization, meat consumption, etc.). In none of those cases was it only the technological change towards PTC conditions (or their respective equivalent) that drove the change, but it was always necessary. And in all of those cases there were also norms and other non-PTC patterns against diffusion of alternatives that were more overcomeable (?) once we were closer to PTC-like conditions.
5. I think a crisper way to state 4 would be to say “from what we know from other transitions, we know that reaching a state close to PTC explains a lot of the variance in adoption so it seems reasonable as a best-guess prior that this will be the case for the animal protein transition as well.” That doesn’t mean other factors don’t matter, but I think it puts thing in perspective with regards to relative importance (again, this makes no statement about the marginal effectiveness of different actions).
Thank you, I appreciate the nuance! [Also, I realize it’s a long paper, so I quote some relevant passages, but apologies if you already read them. I figure it might help other folks following our thread as well.]
“One could go away from your piece thinking there is a lot of evidence that should have one update against long-term PTC” Reasonable, although I did try to avoid this and emphasize the results apply to current consumers. So I also agree it “seems not warranted by most of the kind of evidence you cite” and specifically didn’t cite evidence that focused on future prospects to maintain a narrow focus on current prospects. That said, all else equal, I do think a worse current situation is evidence in favor of a worse future situation.
That makes sense! I want to spend more time with those two reports if I have more time to focus on long-term plant-based meat prospects; thanks for pointing me towards them specifically.
Concur.
I think this is probably the crux of the our disagreement :) First, I think the PTC conditions you’re referring to are under-specified as I elaborate here, so it’s hard to refute without knowing specifically what you have in mind for PTC. (In the paper, I tried to focus on particular cases that limit the definitional issues or relied on working definitions that I specifically criticize, like passing blinded taste tests.) Second, I don’t think the sustainable innovations reference class is especially compelling: for example, most people had ~no understanding of CFC in refrigerants and didn’t think they were a basic necessity of a healthy life. Similarly, few people celebrate holidays that explicitly focus on the act of consuming fossil fuels. In contrast:
Some of my other work has touched on what I think is a more relevant reference class: plant-based analogs to animal products:
“From what we know from other transitions, we know that reaching a state close to PTC explains a lot of the variance in adoption” indeed is the core claim that I don’t think there’s actually much clear evidence on. Attributing causation in these sorts of transitions is very difficult; I try to lay out some of the challenges for PTC in particular the last paragraph here. I’d be interested what evidence supports this claim.
Hi Jacob,
thank you—strongly upvoted for quality of exchange!
In the interest of time (this has to be my last comment), I ignore the smaller disagreements and focus on what seem like the two cruxes we have here (opposite sequence in your comment, but I think answering in this order is easier here):
(a) Does PTC or PTC-likeness causally drive adoption?
(b) Are clean energy technologies a good comparator?
On (a), here is a visual from the latest IPCC report:
Of course, correlation does not equal causation, but we know from many richer accounts than those simplistic curve displays that the story usually goes (a) high-income country heavily subsidizes R&D and/or deployment of tech, (b) tech gets cheaper, (c) we get a self-amplifying dynamic that drives cost reductions and adoptions, (d) an increasing share of adoption is in countries without those high subsidies, i.e. in settings where the cost reductions (P) but also other improvements (e.g. C-like range for electric cars) drive increased adoption, i.e. it is not just “green subsidies everywhere, all at once” but rather “green subsidies drive cost reductions that enable global diffusion”.
For example, I don’t think anyone doubts that solar will expand massively and that this was causally enabled by cost reductions which were a function of early investments. It is clear that the world could have turned out different here for example if the conservatives had won the German elections in 1998.
On (b), if I understand you correctly you seem to be saying that there are a lot of food-specific considerations that make food special and clean energy comparators inappropriate.
That may be and I am not a food system expert enough to weigh the Thanksgiving holiday vs. other forms of cultural lock-in for fossil fuels.
But I would be a bit cautious here as well because there are also many ways in which the food transition is easier than the energy transition so I think a list of ways in which the food transition feels less tractable feels incomplete. Here are a couple of ways how food could be easier: alternative proteins are a better meat replacement than renewables are replacements for coal (no equivalent to intermittency), the food industry is smaller and less powerful than the energy industry, changing protein sourcing is easier than changes in the energy system that require more infrastructure, etc. The point here is that I think focusing on particular considerations only becomes really convincing, I think, when you do a fairly complete accounting in all directions.
My personal take is that including evidence from technologies that have undergone those transitions feels important and that the lack of comparability because they are of a different domain is indeed a limitation but that the effects from undergone transitions provide important additional evidence to form expectations about the future.
Ditto, really appreciate your taking the time to so thoughtfully engage. :) A good day on the Forum! I’ll try to wrap up here as well.
(a) Thanks for this reference—I wasn’t aware of it! This definitely seems like useful evidence in the right direction and I agree with the XKCD’s comic sentiment. That said, it seems like there are still many possible contingencies where price might be a partial rather than full cause. This seems like a ripe area for further research.
(b) I agree, my list is incomplete, and these are good considerations. By the same token, I am no expert in clean power, so hadn’t thought about some of these challenges like intermittency. Along the same lines of this issue, focusing only on successful technologies also introduces a bias—for example, I imagine a similar graph as above which included nuclear would tell a different story.
I also agree there’s evidence to be gleaned from studying other sustainability technologies; we certainly shouldn’t ignore these other transitions. But I would like to see them studied more rigorously and systematically, including positive and negative instances; analogies and disanalogies; and contemporary techniques in causal inference.
I also think that the energy tech analogy might be useful, in particular the case of solar panels, which, unlike nuclear and other energy sources, are also consumer products that went from “rich persons vanity project” to “you’d be dumb not to buy one”.
Decades ago, solar cells were highly expensive, and mainly used for niche applications. There was environmentalists pressure towards clean energy, but the high cost meant only a few wealthy enthusiasts would undergo the switch, and the industry was small and non-influential.
The environmentalist movement was influential in changing this in several ways, first by getting funding for R&D into developing cheaper, more efficient solar panels, second by subsidizing solar energy so it was more cost competitive for consumers, third by encouraging governments to directly build solar farms, and fourth by increasing public support for clean energy so people felt good about buying solar.
The result of this was a feedback loop. Solar was cheaper to buy, and more fashionable, so more people bought it. And the more people bought it, the cheaper solar got, thanks to mass production techniques, so the more people bought it. And then the industry became large and influential. Eventually, solar became cost-competitive without subsidies in many places, so consumers felt it was a good investment, and then everyone in the neighborhood had it, so it became a normal thing that everyone did. (I’m thinking here of parts of australia where pretty much everyone has solar).
I wonder if you polled people about switching to solar in 1990, whether they might seem as reluctant as people are about substitute meat today.
The big difference between PTC meat-substitutes and solar might be in the effect of mass production: could a “feedback loop” of cheaper meat-substitutes causing more consumption causing cheaper meat-substitutes occur? I think there could also be a social feedback loop, where the more people go veggie, the more normalised it becomes, leading more people to go veggie, etc.
In my (non-expert) opinion, PTC in addition to public pressure and opinion changing movements could be much more effective than either of the two on their own. It seems obviously easier to persuade people to make easier changes, but you actually have to persuade them.