Yes per 100,000 inhabitants, but to determine driver safety the correct metric is fatalities per x unit of distance driven, which is what the GP was using.

People in the US drive far more often than those in France, you'd expect for driving fatalities.