A couple other possible confounds come to mind, such as highly-educated stay-at-home parents.
Something something protectionism, something something basic income.
It all seems like junk to me.
The downside to protectionism is it makes it harder for firms to find workers. The downside to basic income is that it's a massive expense to government, would require higher taxes, and may reduce incentives for people to work (which is why it's mostly being trotted out as a way for people to subsist when automation potentially wipes out labour).
I more mean the bleak pictures painted to gain these ideas favor are pretty much junk in my opinion.
I don't see "has a college degree" in the data. I think you are projecting it on the data.
It's more aptly described as "any college courses" and it's likely self-reported. Remember that the vast majority of "college" is not postgraduate degrees at elite Ivy League universities, but rather started (but didn't finish) at a community college or for-profit diploma mill.
I can easily imagine {teachers, college students, temporarily unemployed/underemployed taking care of dependents, retired persons} falling in the "college" + "bottom 20%" category.
But yes, you're right. It's not clear exactly what data is being graphed.
1) Going into debt for college makes you poor, and
2) A lot of college degrees are just not worth the cost, if you're expecting them to lead to a good income.
But these data don't have anything to do with "being poor" -- it's entirely about income, which is still positively correlated with more education.