Matt Yglesias has a post on the politics of immigration that is well worth a look:
I like to make the broad case for higher levels of immigration at all skill levels rather than the narrow case for skills. The basic issue with immigration is that it’s all about complementarities. When new immigrants arrive who are very similar to you personally the tendency is for your real wages to fall, but the real wages of everyone else (including the immigrant) to rise. And the important fact about this is that the gains are much larger than the losses. Because it’s a positive sum interaction, in principle you could compensate the losers with redistribution. But in the particular case of immigration, there’s no need to do that because very few people are very similar to you personally. The important thing is just to not restrict immigration to one particular kind of person. Only allowing computer programmers in is bad for computer programmers and only allowing plumbers in is bad for plumbers. But a broad-based expansion of legal immigration is good for everybody as long as the immigrants are genuinely here to get jobs. You wouldn’t want to throw our doors open to criminals or let every foreign 83 year-old start getting Medicare coverage, but more foreign workers in general is good for American workers in general while if you scrutinize any particular potential migrant you’re going to find plenty of people with a rational basis for wanting to keep him out.
This is a strong argument, in my view, yet I persist in favoring a skill-biased immigration system. The premise is that high-skilled immigrants are particularly attractive because, among other things, (a) they strengthen existing domestic talent agglomerations, (b) they are relatively unlikely to represent a net fiscal burden, (c) they are somewhat more likely to assimilate to the cultural norms that prevail among the self-reliant and upwardly mobile, and (d) they complement less-skilled and mid-skilled domestic workers in in-person services. The case for HSI is closely related to the case for increasing the density and connectivity of high-productivity regions: creating thick economic interrelationships between high-skilled and less-skilled workers can greatly contribute to collective wealth and well-being.
So why not also find complements for high-skilled domestic workers by sharply increasing the intake of less-skilled foreign workers? The main arguments against flow from (b) and (c), i.e., less-skilled foreign workers are somewhat more likely to represent a net fiscal burden and they might have a somewhat more difficult time assimilating to the cultural norms that prevail among the self-reliant and upwardly mobile, but the case on both counts is not extremely strong. A closely related reason is that an less-skilled workers potentially contributes to wage and wealth dispersion, as low-cost labor in in-person services increases the earning capacity of high-skilled workers while restraining wage growth for similarly-skilled domestic workers. The trouble with this latter argument, i.e., the argument from restraining wage growth, is, as immigration advocates often correctly observe, is that less-skilled foreign workers and less-skilled domestic workers tend to have quite different skill sets, e.g., different levels of English-language proficiency, highly specialized skills rooted in the culture of the source country, etc. I still think that an outcome in which a significant rise in HSI leads to tighter labor markets and higher wages for less-skilled and mid-skilled domestic service workers can reasonably be preferred to one in which a significant rise in HSI is balanced by a significant rise in LSI, and the society as a whole is wealthier but (i) wage and wealth dispersion increase considerably, thus prompting a politco-economic backlash and (ii) the assimilation process is somewhat slower, and cultural isolation cuts against upward mobility for a significant share of the population.
But I should stress that this is definitely not a no-brainer.