I feel I am repeating myself, but these remain neglected:
1. The good outcomes for African immigrants to the United States mean we could and should take in more such immigrants, to mutual benefit.
2. In part these gains arise from selection, namely that it is not easy to get from Africa to the United States, at least not generally. So we should not make it too easy, even though we should take in more migrants. “Take in more, keep hard” sounds contradictory but it is not. And if African outcomes decline in quality at the margin, that is a sign that policy is working (more entrants), not that policy is failing.
3. We cannot let everyone in, and so at the margin there will always be cruelties when it comes to those who are denied entry, sent back, and so on. Right now even Canada may be sending back some Haitians (NYT). Those cruelties are relevant for assessing an immigration decision, but they are not decisive. If you cite the cruelties without also outlining a limiting principle for the appropriate margin where immigration ought to stop, you are arguing poorly and most of all fooling yourself. It is a good recipe for never thinking clearly again about any policy issue.
4. Adopting a cosmopolitan ethic will increase the margin at which immigration should be allowed. But still we cannot let everyone in, if only because of backlash effects. And if backlash effects are the binding constraint, the degree of cosmopolitanism in your ethic may not matter much for finding the appropriate rate of immigration.
5. Just to repeat, we really should take in more immigrants. Not only from Africa, but from many countries that are not major successes on the gdp or education front, India and Iran being two other obvious examples.
6. Will Wilkinson has an excellent NYT piece on making immigration deals with Trump.