The American economy is geared to having a large quasi-legal or illegal immigrant presence; many Americans like the economic results but don't like the economic, social, legal, etc consequences. Since the US is not going to deport millions upon millions of immigrants, any conceivable deal requires BOTH a promise to restrict future illegal flow and something short of mass roundups and eviction for those already here.
This is worth digging into a little deeper. According to The Economic Logic of Illegal Immigration (pdf), a report from economist Gordon Hanson written in 2007, policies focused on halting and reversing the flow of illegal immigration would be actively harmful to the economy and would cause serious disruptions to the labor market in the regions where illegal immigrants are more concentrated.
To understand why, it's first important to know that illegal immigrants are highly responsive to market forces—much more so than legal immigrants, who have to contend with legal hassles that make it challenging to move from one job to another. Writes Hanson:
Illegal migrants tend to arrive in larger numbers when the U.S. economy is booming (relative to Mexico and the Central American countries that are the source of most illegal immigration to the United States) and move to regions where job growth is strong. Legal immigration, in contrast, is subject to arbitrary selection criteria and bureaucratic delays, which tend to disassociate legal inflows from U.S. labor-market conditions.
Not only can illegal immigrants move quickly to where the jobs are, but the types of jobs they tend to fill are those that Americans gradually have become, over the last few decades, either over-qualified for or less willing to do. It's true that we're just now coming out of a recession and unemployment is high, but at some point that will change and the previous trend is likely to continue. Hanson writes:
Due to steady increases in high school completion rates, native-born U.S. workers with low schooling levels are increasingly hard to find. Yet these workers are an important part of the U.S. economy—they build homes, prepare food, clean offices, harvest crops, and take unfilled factory jobs. Between 1960 and 2000, the share of working-age native-born U.S. residents with less than twelve years of schooling fell from 50 percent to 12 percent. Abroad, low-skilled workers are more abundant. In Mexico, as of 2000, 74 percent of working-age residents had less than twelve years of education. Migration from Mexico to the United States moves individuals from a country where their relative abundance leaves them with low productivity and low wages to a country where their relative scarcity allows them to command much higher earnings.
You can draw a couple of conclusions from this beyond the repudiation of the normal complaint that immigrants "steal" American jobs. The first is that even if you believe that immigration's impact on the economy is negligible at best (more on this in a minute), the beneficial impact on the immigrants themselves is huge. Maybe that shouldn't be a primary consideration in formulating national policy, but it should count for more than nothing.
Second, any policies that address immigration shouldn't just carve an easier path to legitimacy for illegal immigrants; they should also reduce the bureaucratic hassles experienced by legal immigrants and give them more flexibility to change jobs. At least from the narrow perspective of economics, this kind of solution would be vastly preferable to one that spends more of the country's resources on tightening up the borders, deporting illegals, and cracking down on employers.
I should point out that Hanson is no squishy pro-immigration idealist. This paper cites work by George Borjas in claiming that immigration does lower the wages of low-skilled Americans. That decline is at least partially offset by the disinflation impact of low-skilled immigrants on goods and services—and even the conclusion about wages has been challenged by other economists, but I'll leave that for another day.
Hanson also writes that the "net economic impact of immigration on the US economy appears to be modest," and accepts that lower-skilled immigrants are probably a (very small) drain on the country's public finances. Again, there are economists who disagree, but Hanson's point is that policies focusing on enforcing the border and reversing illegal immigration would have specific effects, and those effects are, on balance, damaging to national well-being. This is especially the case when you take into account the costs associated with beefing up the border.
Anyway, there are some things in the paper that can be quibbled with, as I just mentioned. But it's an excellent read, and I'll close by quoting from parts of his conclusion at length:
Sending all illegal immigrants home would reduce the U.S. labor force by 5 percent and the low-skilled U.S. labor force (workers with less than a high school education) by 10 percent or more. In 2005, illegal immigrants accounted for 24 percent of workers employed in farming, 17 percent in cleaning, 14 percent in construction, and 12 percent in food preparation. Losing this labor would likely increase prices for many types of non-traded goods and services, increase wages for low-skilled resident labor, decrease incomes of employers that hire these workers, and increase the incomes of taxpayers that pay for the public services these individuals use. The net impact of these changes would be small, although in some regions and industries the dislocation caused by the labor outflow would be considerable. ...
Because the aggregate gains or losses are small, any new policy that requires a major outlay of funds would be likely to lower U.S. economic well-being. In a rush to secure U.S. borders, some policymakers insist that major efforts are needed to prevent continued illegal inflows from abroad. While the goals of reducing illegality and establishing greater border control are laudable, it would be difficult to justify massive new spending in terms of its economic return. ...
Illegal immigration is a persistent phenomenon in part because it has a strong economic rationale. Low-skilled workers are increasingly scarce in the United States, while still abundant in Mexico, Central America, and elsewhere. Impeding illegal immigration, without creating other avenues for legal entry, would conflict with market forces that push for moving labor from low-productivity, low-wage countries to the high-productivity, high-wage U.S. labor market.
I'm in the process of aggregating some of the recent academic literature on immigration and the US economy for a series of blog posts. It's taking me longer than I expected, so in the meantime I'll sporadically share some of the more interesting bits I come across.