The World Bank "WITS" report lists 2023 US exports of code 300290 "Human and animal blood; microbial cultures; tox" at $37.9 billion.
This is confusing, because the US government "DataWeb" portal you linked only provides four codes under 3002.90: 3002901000, 3002905210, 3002905220, 3002905250, which add up to a measly $1.3 billion in 2023.
Is the World Bank using a very slightly different classification which uses similar names to the USITC but has completely different values? Am I the fool for assuming that 300290 and 3002.90 would refer to the same kind of blood-related products?
I think I found the answer at the bottom of that link:
> HS Nomenclature used HS 1988/92 (H0)
That sounds... old. There's no version provided on the DataWeb link at first glance, but it includes codes such as 3002.14 which seem to have been created in 2022:
The WITS number of $37.9 B and the DataWeb number of $41.98 B are close enough for me to be happy assuming that this funkiness is down to the mapping between two different versions of the same classification system. That would also mean the analysis in your article still stands up: might as well use the most detailed and recent system when we can.
Excellent post. The Economist is reliably inaccurate, though it purports to cover economics, a field awash in data. This is because 1) many of the journalists and most of the editors appear innumerate, 2) the current format of the magazine is “inch deep” coverage of subject matter better suited to “mile deep”, and 3) as you note, there are no (few) bylines and little accountability.
Making the world a better place one calculation at a time. Thanks!
Also, if say 99.9% of the value of some of these products is the steps after getting the blood, then that would mean the value of blood included is much lower. I think that would be the number that people want to compare with their intuition about "how much blood does the US export?".
It's definitely true that most of the value in many of these products comes from things other than the blood. But I think that some of the bigger categories (e.g. blood plasma) get a large fraction of their value from blood. So I wouldn't think the number for that would be *that* much smaller. Maybe 0.2% at least?
A "do as I say, not as I do" kinda situation. I've been there. :)
I do agree that letting people reverse-engineer numbers is helpful, but I disagree with your digits article that it makes more sense to push for "show lots of digits" than to push for other reproducibility information. I think both will be about equally hard to get writers to adopt, and the full information is much more useful than a few digits. The digits approach also forces the average reader (who does not care about the source) to read useless information, so I think it's worse overall.
I noticed you've now corrected the last article to say 0.7%, despite your updated calculation here returning 0.6867%. Very reasonable IMO, and I would point out that perhaps whatever instinct leads you to consistently not follow your own advice will provide some insight into why others don't follow your advice either. :)
Though admittedly, your reasoning in the digits article only applies to cases where the reader can't easily find the source calculation some other way, so I think the best defense for your rounding would be to point out that, by including direct links to your sources in these cases, you obviate the need for the extra digits.
That is the defense I'd make. But I still I tend to agree with you that the fact that I can't follow my own advice even in the context of chastising others for not showing enough digits for _exactly the same number_ has got to signify something! 🤦
This was my thought. At the outset of this article, I was expecting the 2% figure to be off by orders of magnitude. Being off by a factor of 4 isn't that big a difference from what I was expecting
Regarding this post and the previous one, surely the low hanging fruit for addressing organ shortages is to tweak our current incentives for donors rather than have the poor selling kidneys to make rent, right?
E.g., instead of opting into being an organ donor when you obtain/renew your drivers license (I believe this is the most common way organ donation is addressed, at least in the US) you make donor status the default and charge a minimal fee to opt out; something like $10 should be sufficient. I like the moral symmetry of using DLs as well. If you are licensed to drive your 2,000+ pound metal box at lethal speed on the public thoroughfares, it’s reasonable that donation of your organs on death should be the default.
An opt-out system that requires familial consent is not an opt-out system at all. I.e., the problem is not that opt-out doesn’t work (or at least seriously help), but that true opt-out has not yet been tried.
IMHO, the issue here is that death generally, and cutting into the dead body of a loved one specifically, makes people squeamish. The idea of default organ harvesting gives people a feeling of moral disgust and they feel justified in it because the family of the deceased doesn’t have to feel the suffering of those not getting the organs. Our prosperity has given western society a warped view of suffering and death, but I’m definitely digressing…
Back on point: This issue is closely analogous to probate: you can opt-out of the laws of intestacy by creating a will. Absent that, the State has laws about who gets your property upon your death. Given the stakes, organs should be no different. A society where that was the norm could also put into place processes to maximize the number of viable cadaverous organs. (And I would counter any objections about contentious probate matters with exigent circumstances—an organ has a shelf life much shorter than, say, a bank account—and magnitude of harm)
I don't necessarily have a strong view here, other than that people dying from lack of kidneys is bad and we should do *something*. That said some countries have tried opt-out systems and it didn't seem to work that well in practice. The issue is that in practice the ultimate decision is typically left to the family—and it seems tyrannical to not leave it to the family, since that would mean taking organs against the wishes of loved ones. When places have moved to opt-out the families know that and are less likely to consent, meaning it doesn't seem to move the needle too much:
> Unfortunately, the opt-out system does not seem to have worked. A 2019 analysis of opt-in versus opt-out systems in OECD countries concluded that opt-out systems do not increase the supply of organs available from cadaveric donors in practice. This is because, in opt-out systems, families recognize that their loved ones have never given an explicit preference in favor of donation, and are therefore less willing to approve donation than in an opt-in system, where their deceased relative’s preferences have been made explicit. Perhaps removing familial consent could get around this, but doing so has proven very hard in practice where tried.
Excellent work. Proof again that we should resist the Gell-Mann Amnesia Effect as hard as we can, that journalists can be wrong about anything and everything, and that if a published claim is important to us, we should take the effort to verify it.
Whether it is or isn't, who cares? It's an empirical fact that journalists get facts wrong all the time. The problem is especially acute when the topic is political, but they even get it wrong on issues as bland as the percentage of American exports derived from blood. And they almost never cite sources which would allow independent verification.
In my view, the only expertise you can really trust is your own. For claims outside your domain, you need to root out the raw data yourself if you want to verify them. Obviously, this is a lot of work, which is why I say you should only bother if the claim matters to your life (or perhaps if curiosity is eating you up, as was evidently the driver for dynomight's article). Otherwise, the proper attitude to any non-obvious claim is, "Maybe. Maybe not."
I didn't check it too carefully, since it was just a small aside in the original post. That said, don't let me Amnesia you either—it wouldn't shock me at all if someone shows my calculation is still wrong. (At least my calculation is legible!)
Good point, and if the issue were important enough to me, I would do the legwork to verify the figures for myself. But it isn't, so I'm content to sit back and admire your efforts.
The World Bank "WITS" report lists 2023 US exports of code 300290 "Human and animal blood; microbial cultures; tox" at $37.9 billion.
This is confusing, because the US government "DataWeb" portal you linked only provides four codes under 3002.90: 3002901000, 3002905210, 3002905220, 3002905250, which add up to a measly $1.3 billion in 2023.
Is the World Bank using a very slightly different classification which uses similar names to the USITC but has completely different values? Am I the fool for assuming that 300290 and 3002.90 would refer to the same kind of blood-related products?
https://wits.worldbank.org/trade/comtrade/en/country/USA/year/2023/tradeflow/Exports/partner/ALL/product/300290
hmmmmm!
In short, I can't find how this is defined! I'd guess it's a different definition, but I'm puzzled for the same reasons you are.
I think I found the answer at the bottom of that link:
> HS Nomenclature used HS 1988/92 (H0)
That sounds... old. There's no version provided on the DataWeb link at first glance, but it includes codes such as 3002.14 which seem to have been created in 2022:
https://www.wcoomd.org/-/media/wco/public/global/pdf/topics/nomenclature/instruments-and-tools/hs-nomenclature-2022/table-i_en.pdf?la=en
The WITS number of $37.9 B and the DataWeb number of $41.98 B are close enough for me to be happy assuming that this funkiness is down to the mapping between two different versions of the same classification system. That would also mean the analysis in your article still stands up: might as well use the most detailed and recent system when we can.
Soundtrack: https://scruffamudda.bandcamp.com/track/how-much-blood
Excellent post. The Economist is reliably inaccurate, though it purports to cover economics, a field awash in data. This is because 1) many of the journalists and most of the editors appear innumerate, 2) the current format of the magazine is “inch deep” coverage of subject matter better suited to “mile deep”, and 3) as you note, there are no (few) bylines and little accountability.
Bloody good read.
Making the world a better place one calculation at a time. Thanks!
Also, if say 99.9% of the value of some of these products is the steps after getting the blood, then that would mean the value of blood included is much lower. I think that would be the number that people want to compare with their intuition about "how much blood does the US export?".
It's definitely true that most of the value in many of these products comes from things other than the blood. But I think that some of the bigger categories (e.g. blood plasma) get a large fraction of their value from blood. So I wouldn't think the number for that would be *that* much smaller. Maybe 0.2% at least?
Thanks! The 2% number seemed off but I was too lazy to check it, this is great
If people should always show lots of digits, why did your last article say 2% rather than 1.8%?
Fair point… I guess you can take your pick between the fact that I’d already verified that number was non-reproducible and rank hypocrisy.
A "do as I say, not as I do" kinda situation. I've been there. :)
I do agree that letting people reverse-engineer numbers is helpful, but I disagree with your digits article that it makes more sense to push for "show lots of digits" than to push for other reproducibility information. I think both will be about equally hard to get writers to adopt, and the full information is much more useful than a few digits. The digits approach also forces the average reader (who does not care about the source) to read useless information, so I think it's worse overall.
I noticed you've now corrected the last article to say 0.7%, despite your updated calculation here returning 0.6867%. Very reasonable IMO, and I would point out that perhaps whatever instinct leads you to consistently not follow your own advice will provide some insight into why others don't follow your advice either. :)
Though admittedly, your reasoning in the digits article only applies to cases where the reader can't easily find the source calculation some other way, so I think the best defense for your rounding would be to point out that, by including direct links to your sources in these cases, you obviate the need for the extra digits.
That is the defense I'd make. But I still I tend to agree with you that the fact that I can't follow my own advice even in the context of chastising others for not showing enough digits for _exactly the same number_ has got to signify something! 🤦
Even half a percent of goods exports by dollar amount seems very high.
This was my thought. At the outset of this article, I was expecting the 2% figure to be off by orders of magnitude. Being off by a factor of 4 isn't that big a difference from what I was expecting
factor of 3 error in random aside unacceptable for dynomight
If Dracula was an economist…
Regarding this post and the previous one, surely the low hanging fruit for addressing organ shortages is to tweak our current incentives for donors rather than have the poor selling kidneys to make rent, right?
E.g., instead of opting into being an organ donor when you obtain/renew your drivers license (I believe this is the most common way organ donation is addressed, at least in the US) you make donor status the default and charge a minimal fee to opt out; something like $10 should be sufficient. I like the moral symmetry of using DLs as well. If you are licensed to drive your 2,000+ pound metal box at lethal speed on the public thoroughfares, it’s reasonable that donation of your organs on death should be the default.
An opt-out system that requires familial consent is not an opt-out system at all. I.e., the problem is not that opt-out doesn’t work (or at least seriously help), but that true opt-out has not yet been tried.
IMHO, the issue here is that death generally, and cutting into the dead body of a loved one specifically, makes people squeamish. The idea of default organ harvesting gives people a feeling of moral disgust and they feel justified in it because the family of the deceased doesn’t have to feel the suffering of those not getting the organs. Our prosperity has given western society a warped view of suffering and death, but I’m definitely digressing…
Back on point: This issue is closely analogous to probate: you can opt-out of the laws of intestacy by creating a will. Absent that, the State has laws about who gets your property upon your death. Given the stakes, organs should be no different. A society where that was the norm could also put into place processes to maximize the number of viable cadaverous organs. (And I would counter any objections about contentious probate matters with exigent circumstances—an organ has a shelf life much shorter than, say, a bank account—and magnitude of harm)
I don't necessarily have a strong view here, other than that people dying from lack of kidneys is bad and we should do *something*. That said some countries have tried opt-out systems and it didn't seem to work that well in practice. The issue is that in practice the ultimate decision is typically left to the family—and it seems tyrannical to not leave it to the family, since that would mean taking organs against the wishes of loved ones. When places have moved to opt-out the families know that and are less likely to consent, meaning it doesn't seem to move the needle too much:
> Unfortunately, the opt-out system does not seem to have worked. A 2019 analysis of opt-in versus opt-out systems in OECD countries concluded that opt-out systems do not increase the supply of organs available from cadaveric donors in practice. This is because, in opt-out systems, families recognize that their loved ones have never given an explicit preference in favor of donation, and are therefore less willing to approve donation than in an opt-in system, where their deceased relative’s preferences have been made explicit. Perhaps removing familial consent could get around this, but doing so has proven very hard in practice where tried.
https://worksinprogress.co/issue/compensating-compassion/
Excellent work. Proof again that we should resist the Gell-Mann Amnesia Effect as hard as we can, that journalists can be wrong about anything and everything, and that if a published claim is important to us, we should take the effort to verify it.
I agree, but isn't this also the same sentiment that drives the anti expert world view of the MAGA crowd?
Whether it is or isn't, who cares? It's an empirical fact that journalists get facts wrong all the time. The problem is especially acute when the topic is political, but they even get it wrong on issues as bland as the percentage of American exports derived from blood. And they almost never cite sources which would allow independent verification.
In my view, the only expertise you can really trust is your own. For claims outside your domain, you need to root out the raw data yourself if you want to verify them. Obviously, this is a lot of work, which is why I say you should only bother if the claim matters to your life (or perhaps if curiosity is eating you up, as was evidently the driver for dynomight's article). Otherwise, the proper attitude to any non-obvious claim is, "Maybe. Maybe not."
I didn't check it too carefully, since it was just a small aside in the original post. That said, don't let me Amnesia you either—it wouldn't shock me at all if someone shows my calculation is still wrong. (At least my calculation is legible!)
"don't let me Amnesia you either"
Good point, and if the issue were important enough to me, I would do the legwork to verify the figures for myself. But it isn't, so I'm content to sit back and admire your efforts.