The Agenda

NRO’s domestic-policy blog, by Reihan Salam.

The Undereducated American and U.S. Immigration Policy


Ruy Teixeira of the Center for American Progress argues that Americans are “not as educated as we should be, given the demands of today’s increasingly skill-driven economy.” And he makes a convincing case.

As a recent study by Georgetown University’s Center on Education and the Workforce suggests, today’s 74 percent premium for a college education over high school (that is, how much more an average college graduate earns than an average high school graduate) is too high, both because it signals that today’s economy is demanding more college graduates and because of it contributes to wage inequality. They recommend instead a Bachelor’s degree wage premium level of around 46 percent.

This reduction in the college wage premium can be obtained if we increase the supply of college educated workers by around 20 million over the next 15 years (right now, we’re on track to add only about 8 million). Such an increase in the college educated workforce would take a big bite out of inequality, rolling back about a quarter of the increase in inequality we’ve seen since 1979.

This can be done, but it will necessitate a much more aggressive approach than we currently have to bringing down college costs and making sure everyone who wants to has adequate resources to pursue a college degree. This is particularly the case for black and Latino youth who still lag substantially behind whites and Asians in their college completion rates.

The study Teixeira cites — “The Undereducated American,” by Anthony Carnevale and Stephen Rose — is eye-opening. One of their findings is that it is important that the share of workers with post-secondary education increase, as a higher college-educated share of the workforce will lead to higher wages for less-skilled workers. Under current trends, the U.S. will have 8 million more postsecondary-educated workers by 2025. But if we sharply increase the number of workers with post-secondary education to 20 million, the U.S. economy will look very different:

[I]f this alternative path is adopted, GDP will be $500 billion higher than current projections and there will be earnings increases across the board even though the earnings gaps between Bachelor’s degree holders and high school educated workers will decline. Figure 11 projects the average increase in earnings by education level. Earnings of workers with a Bachelor’s degree would rise by 6 percent (from $56,138 to $59,329 [2010 dollars]) and the earnings of workers with only a high school diploma would rise by 24 percent (from $32,179 to $40,003), which would bring the wage premium down to 46 percent. For those with an Associate’s degree, earnings would rise 15 percent (from $36,160 to $41,584).

Given the declining earnings of male high school graduates over the past 30 years, our predicted 24 percent rise in real high school earnings may seem improbable. But the supply and demand conditions for less-skilled workers will be very different under our 2025 scenario: Output per worker will be 25 percent higher than it is today, and the share of the labor force with only a high school diploma or less will decline from 40 percent to 25 percent, making high school workers more scarce. High school workers’ scarcity will result in higher wages. Further, as part of meeting the goal of producing more college graduates, we specified that the quality of the K-12 system would have to increase; therefore, the quality of the noncollege-educated group in 2025 is likely to be higher than their counterparts today. Therefore, our projected higher wages for high school graduates reflect both higher skills and a large decline in supply. [Emphasis added]

But what will happen if we increase the supply of workers with post-secondary education while also dramatically increasing the supply of less-skilled workers with high school diplomas are less? Under this scenario, less-skilled workers will be far less scarce, and so wage gains for these workers will presumably be much smaller. To those of us who believe that  robust wage gains for less-skilled workers would be a welcome development, the next question is this: how we can grow not just the number but also the share of the workforce with post-secondary education, as the share of the workforce with post-secondary education will ultimately determine the size of the college wage premium? 

One relevant factor is immigration policy. An immigration policy that is explicitly geared towards welcoming workers with post-secondary education while restricting the numbers of workers with less education would contribute, over time, to increasing the share of the workforce with post-secondary education. 

Recently, Shikha Dalmia replied to my argument that as long as the U.S. intends to welcome a finite number of immigrants, it makes sense to prioritize workers with post-secondary education over less-skilled workers. There are a few small issues with her characterization of my posts, e.g., while I do observe that limited English proficiency is one reason low-income immigrants are less likely to access social welfare services, it is not the only reason — linguistic isolation is not the only form of social isolation; complicated eligibility rules mean that recent arrivals are much less likely to collect benefits than earlier arrivals, and so less-skilled immigrants will tend to access benefits more over time, etc. So her argument, as I understand it, is unconvincing:

Salam’s first point that one reason low-skilled immigrants use less welfare is because their poor language skills render them less able to navigate the bureaucracy might be true — but so what? It doesn’t negate my central thesis that so long as America needs a labor class, a foreign-born one (that needs welfare but can’t access it) is cheaper than a homegrown one (that needs welfare and can get it). This might be unfair to foreign workers — although they are obviously better off with a job in America even with differential access to welfare than in their home country or they wouldn’t be here — but not to America. (I am merely describing the current state of affairs, not endorsing it.)

My point is that this is not a stable equilibrium — low-skilled immigrants are “cheaper” only as new arrivals; as they assimilate, they become more like native-born less-skilled workers. Moreover, the U.S. is affirmatively choosing to welcome less-skilled foreign workers. The presence of less-skilled native-born workers reflects a failure of U.S. K-12 policies and much else. U.S. policymakers are not choosing to rely on a homegrown less-skilled workforce rather than a foreign-born less-skilled workforce. Rather, the existence of a homegrown less-skilled workforce is a product of complex historical circumstances. If we could legislate away the legacy of enslavement, segregation, and entrenched poverty and in doing so dramatically increase the share of the workforce with a post-secondary education, I would recommend that we do so. But we can’t. We can, however, choose not to sharply increase the share of less-skilled workers by revising our immigration policies. And in doing so, we can preserve our limited resources to tackle intergenerational poverty to deal with intergenerational poverty that has domestic origins. Other countries face profound intergenerational poverty challenges, including middle-income countries like Mexico and low-income countries like Bangladesh. While it is undoubtedly true that allowing these countries to export less-skilled workers to the U.S. will help alleviate their poverty problems, it is not obvious that this is the best policy from the perspective of U.S. taxpayers or from the perspective of the native-born American poor. Some balancing of interests is appropriate, e.g., some humanitarian component could be a part of a sensible immigration policy. But the notion that we are substituting a cheap foreign-born less-skilled workforce for an expensive native-born less-skilled workforce doesn’t make much sense — the native-born less-skilled workforce isn’t going anywhere, where or not we sharply increase the overall number of less-skilled workers.

Dalmia acknowledges that regularizing the status of less-skilled workers might make them more expensive in terms of public social expenditures. Yet she states that it is impossible to predict whether this cost will exceed the economic benefit created by less-skilled migration. That may well be true, yet it is a much safer bet that immigrants with post-secondary education will deliver net fiscal benefits.

Dalmia insists that we face a knowledge problem: those of us who believe that U.S. immigration policy ought to favor skilled over less-skilled immigrants are engaging in a form of central planning. As evidence, she points to the fact that some Canadian immigrants with doctoral degrees in physics drive taxi cabs. My conjecture is that a Ph.D. physicist driving a cab in Year 0 will have a substantially higher household income, regardless of the profession she is in, in Year 2 or Year 10 than an individual with less than a high school diploma. (I’ll also note that some people choose to drive cabs because they prize autonomy over income — a minority, perhaps, but the point is that more educated workers have more options than less educated workers.) That is, Dalmia is taking a snapshot of skills mismatch to imply that having Ph.D. physicists driving cabs is necessarily a waste, and that we’d be better off if the immigrants driving cabs had the lowest possible skill level to serve in that particular occupational niche. This ignores the fact that economic life is dynamic, and that it makes sense to select immigrants who are likely to prove resilient over time. The lowest possible skill level in Year 0 might prove too low a skill level for Year 10. An immigration policy that undershoots in terms of skill level is likely to yield a larger immigrant underclass over time than an immigration policy that overshoots in terms of skill level. Dalmia’s critique of central planning in immigration policy reminds me of the notion that NGDP targeting represents central planning in monetary policy. Any immigration policy, like any monetary policy under a fiat money regime, implies a degree of central planning. It happens that some central plans are geared towards the near-term interests of agricultural firms and low-wage employers who fear the prospect of a scarce supply of less-skilled workers while others are geared towards other priorities. 

Finally, Dalmia writes the following:

As for Salam and Bigg’s objection that it makes little sense for America to import foreigners with an 8th grade education when it is trying to raise general educational attainment levels, it sounds perfectly reasonable, but here’s the thing: The goal of an education policy and an immigration policy are not identical. The primary purpose of an education is to produce well-honed, well-rounded, self-actualized individuals — not necessarily generate workers for an economy. The economy certainly shapes individual career choices — just as it is shaped by the availability of workers — but that doesn’t mean that at any given time individual aspirations are going to be perfectly in sync with the needs of an economy. A well-crafted immigration policy can help alleviate workforce imbalances and go a long way toward enhancing economic productivity. The fact that between 1995 and 2005, about 700,000 low-skilled foreigners every year managed to be gainfully employed (legally and illegally) in construction, agricultural, landscaping, restaurant, house-cleaning industries speaks for itself.

This is one interpretation of the primary purpose of an education. Another view is that one of the purposes of education policy is, for example, to reduce the share of the population that is likely to prove dependent on means-tested benefits. This is one of the justifications for a public sector role in education — to subsidize human capital investment for households with limited resources and credit constraints, as the spillover benefits of a more educated population are considerable. Dalmia treats workforce imbalances as static, yet a dynamic economy will change over time by substituting labor-intensive business models with capital-intensive business models if the supply of less-skilled workers shrinks over time. We have seen this happen in many economies and in many historical periods. We don’t lament the fact that our economy is severely imbalanced between illiterate horse-and-buggy drivers and cab drivers with post-secondary education, nor should we be.


Subscribe to National Review