Interesting.

It looks like you’re right on traffic deaths. https://en.wikipedia.org/wiki/List_of_countries_by_traffic-related_death_rate The US (12.9) is actually a little below average on traffic related deaths compared to the global average per 100,000 (16.7).

It looks like the discrepancy comes from how different statistics interpret alcohol related deaths. The Wikipedia article I linked to initially, uses numbers from the Disease Control Priorities report https://www.dcp-3.org/ that counts underlying risk factors. That may be a high estimate and there’s some variation in how people talk about alcohol related deaths (eg fully vs partially attributable to alcohol https://www.cdc.gov/mmwr/volumes/73/wr/mm7308a1.htm)

So if you look at just deaths that are fully attributable to alcohol are 51,665 but deaths that are at least partially attributable to alcohol are at 178,000). Some of those traffic deaths are included. Around 11,000 traffic fatalities a year are attributed to drunk driving (https://www.bankrate.com/insurance/car/car-crash-statistics/#leading-causes-of-fatal-car-accidents).

It makes sense now that I think about it. It’s easy to tell when someone’s been in a vehicle crash and whether or not it was fatal. There’s a whole continuum of how much alcohol someone can consume and how much of a problem it is for them. It’s pretty obvious if they die from acute alcohol poisoning but if their alcohol use weakened their immune system and they die of COVID, how do we categorize it?