Automobile ownership was rare in Sabresia as recently as 30 years ago

"More than 10 years ago" is pretty vague isn't it? If the changes took place 20 years ago, well maybe you can argue that even though the number of accidents is up, the accidents were somehow less severe because of the safety changes. (Though even this feels like a bit of a stretch, since we'd have to assume there's such a thing as a safety improvement that doesn't reduce the total number of accidents, but magically causes the accidents that do occur to be minor ones.)

But what if the changes took place 35 years ago? Or 50 years ago? Then this tells us absolutely nothing about the discrepancy between now and 30 years ago.

And while you make a valid point in your analysis of (B) -- you could, in theory, tweak the numbers in such a way that serious injuries should go up even with fewer people in each car -- we have to make sure we're taking into account the language the author uses. As you noted, we're told there's a "significant" increase in the number of accidents and a "dramatic" decrease in the number of people per car. In your example, you've got a 400% increase in the number of accidents and a 50% decrease in the number of people per car, meaning that "significant" in this case represents a much more extreme change than "dramatic."

That seems unlikely to me. I'd think the "dramatic" change would be the larger one, and if that's the case, well, it could totally explain why accidents are up but injuries are down. 100% ironclad? No -- you're right about that. But it can still explain the paradox without having to make any questionable assumptions.

Ultimately, the language in (A) is vaguer, and the logical leap you have to make is more implausible, so that makes (B) the better answer.

I hope that clears things up!