Second-generation Latinos have much higher levels of English language proficiency than Latino immigrants, as one might expect. This is often touted as evidence that concerns about assimilation are misplaced. The larger problem, however, is that while incomes for second-generation Latinos also tend to be much higher than for Latino immigrants, which makes sense given that second-generation Latinos have access to U.S educational opportunities, household incomes among second-generation Latinos lag behind household incomes for the population as a whole.
A recent Pew report on second-generation Americans offered a strikingly optimistic take regarding the economic progress of the children of immigrants, noting that the median household income, adjusted for a family of three, for second-generation Latinos was $48,400 in 2012. This is substantially higher than the median household income (again, adjusted for a family of three — assume this is the number I’m using henceforth), for Latino immigrants ($34,600) and all Latino adults ($39,200). This last number is of particular interest, as it captures third-generation and fourth-generation Latino adults, etc., as well as immigrants and second-generation Latinos. So what exactly is going on with third-generation-plus Latino households? The Pew report includes a chart which notes that median household income for these households is $43,600, i.e., that median household income for the third-generation-plus cohort is actually lower than it is for the second-generation cohort.
One needn’t overinterpret this finding. It could be that the children of today’s second-generation Latinos will greatly surpass their parents in educational attainment and household income. But consider the contrast between median household income for Latino adults and for all adults, including Latino adults, by immigrant generation, which Pew also provides. For immigrant adults, the median household income is $45,800, below the number for second-generation Latinos. Note that this cohort is 47 percent Hispanic, i.e., were we to exclude the Hispanic share of this population, the median household income number would be much higher. For second-generation adults, median household income is $58,100, considerably higher than the $48,400 number for Latino second-generation adults. And 35 percent of this second-generation cohort is Hispanic. Among third-generation-plus U.S. adults, a group that is much larger (177.7 million) than the previous cohorts (37.4 million immigrants, 19.7 million second-generation) and also less Latino (6 percent as opposed to 47 percent and 35 percent), median household income is $60,600. Suffice it to say, the gap between $60,600 and $43,600 is substantial, though of course this comparison is hardly apples-to-apples — it is entirely possible that the Latino third-generation-plus cohort is far more skewed towards third- than, say, fifth-generation adults.
In an ideal world, we wouldn’t be reduced to contrasting median household income across ethnic groups. Rather, we would track household incomes across immigrant generations by educational attainment. Many Latino immigrants are skilled, and these individuals have higher household incomes than their less-skilled counterparts, and there is good reason to believe that the same is true of their children and grandchildren.
What I find puzzling is this: immigration policy represents an effort to shape the future workforce. The U.S. invest considerable resources, both public and private, in increasing high school graduation and college completion rates. The goal of this investment in human capital is, among other things, to see to it that U.S. adults are more robust against economic shocks, e.g., an adult with a college degree is generally considered better able to navigate an economic transition than an adult with a high school diploma, and an adult with a high school diploma is generally considered better able to navigate an economic transition than an adult without a high school diploma. As the labor market position of less-skilled individuals has deteriorated, many have been forced to take jobs that lack security of tenure, that don’t offer health and pension benefits, and that don’t yield an income high enough to surpass the poverty level. In some cases, less-skilled individuals have found that exiting the formal labor force is preferable to remaining in it — this is particularly true of those facing various physical maladies, which are in some cases exacerbated by the stresses associated with low-wage work.
So why exactly would be would want to deliberately engineer an increase in the size of the less-skilled workforce? Why would we do this if we also believe that while the children of less-skilled immigrants will likely surpass their parents in levels of educational attainment, they will continue to lag behind other native-born adults? There are a number of arguments that are raised, e.g.:
1. One of the central arguments for increasing less-skilled immigration is that less-skilled immigrants will themselves be better off if they have access to the U.S. labor market. There are weaker and stronger versions of how seriously we ought to take this claim, ranging from the notion that restrictions on transnational labor mobility represent an egregious human rights violation — the contemporary equivalent of enslavement — to the notion that humanitarian considerations ought to play at least some role in U.S. immigration policy. I’m not unsympathetic to the view that humanitarian considerations ought to play some role, yet this would tend to move us towards guest worker policies that are limited to individuals from highly-indebted poor countries and that would discourage guest workers from forming strong attachments in the U.S. (as in Canadian guest-worker policy), so as to facilitate the replacement of one cohort of guest workers who, having been given an opportunity to earn U.S. wages for a limited period of time, can more easily be replaced by another, thus spreading the benefits of temporary access to the U.S. labor market as widely as possible. The problem, of course, is that such an arrangement might prove impracticable in the U.S. context, as it is likely to be regarded as inhumane.
2. There is considerable demand for less-skilled labor, in food preparation, eldercare, tourism, agriculture, and other labor-intensive services. One assumes, however, that demand is high at a given (low) wage. While it is easy to see why one would want to pay less for eldercare and not more, it is worth thinking through the implications of this line of thinking. To secure less-expensive eldercare, we will allow an increase in the influx of less-skilled immigrants willing to work at a low wage. Yet in order for the less-skilled immigrant in question to have a standard of living that we in the U.S. consider acceptable, the public sector would eventually have to provide a suite of means-tested benefits. To be sure, recent immigrants are not eligible for a wide range of benefits, yet eligibility increases over time and when the immigrant in question reaches various milestones, such as securing citizenship. That is, taxpayers at large are obligated to subsidize the consumers of low-cost eldercare services. We might instead choose to subsidize eldercare via some more direct route. Our decision regarding whether or not to take a direct or indirect route ought to rest in part on intergenerational considerations, e.g., how likely is it that children of less-skilled immigrants might also depend on means-tested benefits?
You’ll note that the argument above bears some resemblance to arguments for a statutory minimum wage. That is, because low-wage employees are in many cases forced to rely on means-tested benefits, low-wage employers can be understood as free riders. This view doesn’t strike me as quite right, as there is a real danger of pricing less-skilled individuals out of the labor force. It is important to strike a balance that reflects the needs of firms, the social benefits of labor force participation, and (to some limited extent) the dangers of free-riding, which leads me to favor a combination of a relatively low statutory minimum wage and relatively generous wage subsidies and work supports. But note how such a policy mix interacts with a large influx of less-skilled immigrants. Resources devoted to wage subsidies and work supports would have to be provided to individuals who arrived in the U.S. voluntarily, in part because doing so would benefit the native-born children of less-skilled immigrant parents. Is it necessarily wrong to believe that we ought to husband these resources carefully?
3. A number of advocates of increased less-skilled immigration have pointed me to the work of economist Jennifer Hunt, who observes, among other things, that:
(a) assuming the arrival of less-skilled immigrants increases wage inequality, it will increase the returns to graduating from high school, thus encouraging higher graduation rates;
(b) and that evidence drawn from U.S. Census data bears out this proposition. Specifically Hunt finds the following:
The increase in immigration in the 1990s caused the 2010 native high school completion rate, which stood at 87.8%, to be 1.3 percentage points higher than it would otherwise have been. I find native-born black people to be more responsive to immigration than natives as a whole. However, as the rate of immigration slowed in regions with more black people, the overall effect is very similar (1.4 percentage points). This is comparable with the completion gap between black people and white people, which was 7.3 percentage points in 2010. I also estimate a small negative net effect for native-born Hispanic people (-1.2 percentage points), but this is not statistically significantly different from zero.”
If we further distinguish immigrants by age and education, you also find support for immigration improving education through the labour market channel. Again, magnitudes are small. The 1990s increase in adult immigrant dropouts caused a 0.8 percentage point increase in 2010 native completion rates, with larger effects for native-born minorities. That said, the effects of more educated adult immigrants are not precisely estimated.
I find that immigrants have a small negative effect on natives as a whole; the 1990s increase in child immigrants reduced the native completion rate by 0.2 percentage points. The effect on native-born black people is 0.4 percentage points. The results for native-born Hispanic-descended people are more subtle because the effect of immigrant children on education depends strongly on the education of their parents. Native people of have a moderately sized positive response to the inclusion in the education system of child immigrants of more educated parents and only a moderately sized negative response to child immigrants of poorly educated parents. [Emphasis added]
What I find odd is that Hunt’s arguments have been characterized as a slam-dunk case of the unambiguous benefits of less-skilled immigration on educational attainment for the native-born. It seems no less plausible that in the absence of an increase in adult immigrant dropouts, the U.S. public sector might have been in a position to shift time and attention away from financing various social expenditures associated with this increase and towards increasing high school graduation rates for the native-born population. That is, is increasing the returns to high school graduation by exacerbating low-end wage inequality actually the lowest-cost strategy for increasing high school graduation rates?
I recommend reading Dylan Matthews interview with Michael Clemens of the Center on Global Development, a staunch advocate of increasing less-skilled immigration. The discussion is revealing, particularly as it pertains to the “W visas” proposed in the new Gang of Eight immigration reform bill. One of the more interesting aspects of Clemens’ argument is that we ought to look to earlier eras of mass low-skill immigration, e.g., the 1900s, a period during which the gap in educational attainment between native-born U.S. adults and low-skill immigrants was much smaller than it is today and when means-tested benefits were not widely available.
In the end, my differences with Clemens et al. are relatively modest. I favor a substantial amount of immigration. But I think the costs associated with an influx of less-skilled immigrants are higher than Clemens acknowledges, particularly if one believes that the labor market position of less-skilled workers might continue to deteriorate over time. Many labor-intensive tasks have become capital-intensive tasks over time, and it’s not obvious to me that this trend won’t continue. There are, to be sure, high-touch, nontradeable services that resist automation and offshoring, like eldercare. Yet it remains unclear to me why efforts to reduce the cost of eldercare ought to trump efforts to focus social resources on attacking intergenerational domestic poverty rather than the poverty of newcomers.
Some of those who disagree with me will presumably argue that as a conservative, I am indifferent to intergenerational domestic poverty. This betrays a lack of familiarity with my work and my interests. Others, including many libertarians, will argue that it is inappropriate to devote public resources to attacking integenerational domestic poverty; rather, we ought to eliminate licensing requirements and statutory minimum wages to do the job. I disagree with this view as well.
P.S. In a recent Reuters Opinion column, I offer an alternative framework that might be of interest.