Yet there is no evidence to suggest the factors you describe have ever been part of the public discussion in those countries. And the US is the country with the most draconian drug laws in the Anglosphere.
https://uwaterloo.ca/tobacco-use-canada/adult-tobacco-use/sm...
https://www.cdc.gov/tobacco/data_statistics/fact_sheets/adul...
The government is not forbidding anybody from smoking.
The Canada numbers include teenagers (persons aged 12+) while the US ones do not (persons aged 18+).
Also, the Canada numbers went from 13% (lower than the reported US rate) in 2015 to 15.1% (higher than the US rate) in 2017. That 13 to 15.1% increase over two years seems to be a statistical anomaly as it goes sharply against the overall trend (according to your source):
"Despite the overall prevalence increase in the most recent survey year, from 1999 to 2017, the overall trend was an average annual decrease in prevalence of 3.2% of the previous year’s value"
TLDR; the average over the last several years is more or less the same.
This doesn't sound right at all. Why would the number of interacting parts increase exponentially? And why would that result in decreased economic efficiency rather than increased?
Doesn't this very idea run counter to one of the basic tenets of capitalism that "bigger is better" (due to economies of scale)?
>what's clear is that you can't linearly extrapolate one nation's experiences across nations of different sizes
Maybe true, but IMHO not at all for reasons of economic size. That is, all other things being equal, increasing the size of an economic system should make it more efficient, not less.