Read on Twitter
I have been digging into this & #39;research& #39; (thanks isolation!!) & I think people on here have got the reasons it& #39;s bad a little wrong. It& #39;s actually much worse: here is how the guy& #39;s calculations work...
(oh, image nicked from / brief obsession sparked by @JoMicheII)
This is the crucial bit of the paper & (12) is the crucial equation. It& #39;s not introduced on the basis of evidence from the UK post-2008 - that comes on the next page. So... where does it come from?
Turns out it& #39;s just a different form of an equation introduced earlier, as a causal relationship between life expectancy and GDP that "the J-value approach reveals"
It is actually just a mathematical implication of the "J-value" alternative method of estimating the value of life & effectiveness of safety procedures. Doesn& #39;t sound like a causal relationship... but they claim to have validated it in this study https://www.sciencedirect.com/science/article/pii/S0957582017302896">https://www.sciencedirect.com/science/a...
That paper& #39;s argument is that higher income countries have higher life expectancies because countries across the world implement public health measures if and only if they& #39;re recommended by the J-value approach
There& #39;s no very clear statement of the causal claim, but as best as I understand it: people live longer in rich countries because there are more & #39;worthwhile& #39; health interventions there
And that... is because the J-value calculation assumes a year of life saved is more valuable if the person is richer.
So the argument stepped through seems to be:
- lockdown makes the UK poorer
- because the UK is poorer, it& #39;ll spend less saving lives (partly because it has less money, partly because the lives are worth less now)
- once the UK spends less saving lives, life expectancy will fall
- lockdown makes the UK poorer
- because the UK is poorer, it& #39;ll spend less saving lives (partly because it has less money, partly because the lives are worth less now)
- once the UK spends less saving lives, life expectancy will fall
I kind of doubt the author actually believes or would argue this - the papers are full of lots of equations, lots of references to Karl Popper, and very little straightforward statement of what the causal chain is meant to be
I may well have made some mistakes, the papers are strange and confusing
But at minimum - the "lockdown will be worse than COVID" conclusion is & #39;validated& #39; by a causal study with these assumptions, every one of which seems ridiculous
But at minimum - the "lockdown will be worse than COVID" conclusion is & #39;validated& #39; by a causal study with these assumptions, every one of which seems ridiculous
And the paper makes no effort to test other explanations of the data, so the extent to which J-value& #39;s "causal explanation of GDP-life expectancy links" has been validated strikes me as basically zero
The J-value thing seems to be an alternative approach to cost-benefit analysis originally designed for eg. nuclear plant safety mechanisms. It may work well for that but its extension to macroeconomics and public health seems not much better than crankery.
Tagging in some people who were talking about this @Frances_Coppola @jdportes @crookedfootball @thefishareloose @JoMicheII
By the way, even with no methodological flaws - the paper implies it wouldn& #39;t be worth everyone in the UK giving up 4 months of life to prevent 393,000 deaths
A neat illustration of the aggregation problem that convinced me I& #39;m not a consequentialist
https://philpapers.org/rec/CARAH ">https://philpapers.org/rec/CARAH...
A neat illustration of the aggregation problem that convinced me I& #39;m not a consequentialist
https://philpapers.org/rec/CARAH ">https://philpapers.org/rec/CARAH...
