The estimate is, CO emissions are reduce 40X to pass the test. Meaning, the cars are 40 times too dirty to be on our roads.
I think you misunderstand both the testing and the defeat device. We do tailpipe testing here in the US, and the defeat device results in far cleaner tailpipe emissions during said tests. It's not just an OBD data-hiding thing, it actually changes the emissions.
edit - clarification, we may not do tailpipe tests on diesels here, didn't occur to me that gas tests would be different than diesel.
NOx, not CO. And it's up to, not always.
> Meaning, the cars are 40 times too dirty to be on our roads.
No, it means the cars are up to 40 times too dirty in this one respect to be on our roads.
I will address one thing, though:
> Anything over the mandated limit is a health risk, and equivocating over how much is like asking whether you'd want 3 years, or 10 years shaved off your life expectancy.
That is not equivocating. Also, pollution exists on a continuous scale. Saying everything over the limit is equivalently "bad" and everything under the limit is equivalently "good" is ignoring that. There is a big difference in effect between, say, twice the limit and 40 times the limit.