For men, [1] shows the median lifetime earnings for high school graduates is $882,300. The median lifetime earnings for men with a bachelor's degree is 1,517,200. The lowest field is Fine and Applied Arts at $843,900, which is lower than the median lifetime earnings for those with a high school education, up to the highest median field of engineering with earnings of $1,845,000.
For women, [1] shows median high school lifetime earnings of $458,900, median bachelor's earnings for all fields of $972,500, and lowest median bachelor's earnings of $652,100 (Fine and Applied Arts), and a highest of $1,169,100 (Business Administration). The interesting thing to note is for women there is only 1 of 18 field degree combinations (earnings are provided for bachelor's and associate's) where the woman's lifetime earnings decreases relative to a high school education alone (Fine and Applied Arts associates median earnings are $437,300). For men, 4 of 18 field degree combinations will result in lower lifetime earnings relative to high school education alone.
There is further analysis in [1] of the earnings percentiles by each field showing that men with a business degree earn the most at the 90 percentile. This holds for women with a business degree.
I don't see if this has been restricted to those whose highest education is a bachelor's degree or not, but there is also work that has shown increases in earnings by field for graduate or professional degrees [3]. It's possible that those earnings premiums have been captured in the reporting of [1], but [3] seems to come to the conclusion that most business, arts, and social sciences fields gain a large increase in salary for graduate or professional degrees.
Hopefully those familiar with similar analysis from the US could help to provide that information.
[1] https://www150.statcan.gc.ca/n1/pub/11-626-x/11-626-x2014040...
[2] https://www150.statcan.gc.ca/n1/daily-quotidien/150911/dq150...
[3] https://www150.statcan.gc.ca/n1/pub/75-006-x/2018001/article...
I think this is absolutely the key. If you just look at some of the figures given in [0], and do some simple math, you find out that someone with a bachelor's degree makes just a hair more than $26k per year more than someone with only a high school diploma. Over the course of, say, 40 years working, that comes out to a hair over $1M difference in favor of the college grad. So, it would seem that even if a 4 year degree cost $200k (which, sadly, is not an insane estimate these days [1]), it still seems like a no-brainer.
But, that's as far as statistics can take us, I think, because, as you mentioned, we don't see what happens to people who choose not to go to college, but who could have successfully gotten a bachelor's degree. There's no literal way for us to ethically conduct that experiment.
If I were to speculate a bit, I might wonder if the answer lies in looking at children of college graduates and comparing those children who went to college vs those who didn't. I have no idea if such a study is available or not, unfortunately.
The other idea I have is that, although it pays off in the long run, maybe acquiring a significant amount of debt at a young age makes it harder to get by in the initial, post-college years harder. Considering the cost of housing and health care, this seems like a reasonable hypothesis, but I'd also like to see if there's research on that.
---
[0]: https://www.northeastern.edu/bachelors-completion/news/avera...
[1]: https://www.collegeavestudentloans.com/blog/estimating-the-t...