I want to follow up quickly to a couple of notes on my posting where I referred to the "Die at the Border" policy. I was not arguing for open borders. I don't think that anyone who has given the issue serious thought advocates open borders, since a literal open border policy would almost certainly imply an inflow of hundreds of millions of people in the next couple of decades.
My point is that we don't have open borders; instead we have very serious limitations on immigration. Immigration is restricted both by the danger of the border crossing and the prospect of deportation due to a random encounter with law enforcement (e.g. a traffic ticket). These threats ensure that most immigrants will not be well-educated, since well-educated people in the developing world will not take these risks to work in the United States.
This means that less-skilled workers in the United States have to worry about competition from undocumented workers, while the people who design and debate immigration policy (economists, lawyers, reporters) don't have to worry about professionals from developing countries slipping over the borders and undercutting their wages. The implication of the current immigration policy is that the people who design and debate it are largely its beneficiaries, since they can get low cost home repairs, bargain restaurant prices, and cheap nannies.
We can debate whether this is good immigration policy, but we first have to acknowledge the policy in place. The reason that most immigrants are less educated is not because of any shortage of more educated workers willing to immigrate to the United States, it's because our policy acts to exclude them.
As far as the evidence of the link between immigration and the declining wages of less-skilled workers, I can point to the papers by Borjas, Katz, and Freeman, but I confess to being more influenced by what I take as the stylized facts. If immigration was being driven by a shortage of people willing to take certain jobs, then we should expect to be seeing sharp increases in the relative wages of the occupations that have large shares of immigrant workers. In other words, we should have expected to see sharply rising wages in construction, restaurant and hotel work, as employers raised wages in a desperate effort to attract more workers. In fact, we see the opposite. Wages in these industries/occupations have fallen sharply relative to the wages of more skilled workers.
In a similar vein, accepting the view that the minimum wage has little impact on employment implies a belief that the demand for labor is relatively unresponsive to changes in wages. (In other words, a 15 percent hike in the minimum wage has little effect on the number of workers that firms are willing to hire.) Those of us who believe that the minimum wage has relatively little impact on employment, have a difficult time explaining how a large increase in labor supply will have little impact on wages. (If the demand for workers does not respond much to changes in wages, then it would take a large decline in wages to keep fully employed a workforce that has grown substantially due to immigration.)
Finally, a note on the immigration/wage studies, none of the studies seem to have examined the possible effect of immigrants on local housing markets. (I am prepared to stand corrected, if someone knows of such a study.) There are substantial differences in housing costs and the rate of rate of growth of housing costs across cities. For example, between 1980 and 2005, shelter costs rose by 181.3 percent in Los Angeles (4.2 percent annually), while they rose by just 133.4 percent (3.4 percent annually) in Detroit. (This is taken from the CPI shelter index.) This is a difference of 0.8 percentage points a year. Of course, Los Angeles has seen a large inflow of immigrants, while the Detroit area has seen a much smaller inflow.
Shelter accounts for 30 percent of family expenditures on average, and a considerably larger share for low-income families. If shelter costs rise more rapidly in cities with large immigration flows, then this would imply lower real wage growth, given the same nominal wage growth. This could have a substantial impact: in the LA-Detroit comparison, the difference in the growth of shelter costs would translate into a difference in real wage growth of 0.2-0.3 percentage points annually, for the same nominal wage growth.
A serious study would have to look at a large set of cities, and ideally at the housing actually occupied by lower income households, since there are likely different rates of housing inflation for different types of housing. However, given the importance of housing in consumption baskets, and the differences in price trends across cities, this could be an important part of the picture.