December 09, 2005

Immigration Control Through The Ages

The New Republic makes the case for somewhat-open borders (they don't have the stomach to fling the doors open all the way, but okay), in part by noting that the demand for unskilled jobs will outpace the rate immigration over the coming decade, so we need more workers. Matthew Yglesias responds:
Supply and demand doesn't really dictate anything. If you had an airtight immigration control system, there wouldn't just be oodles of undone jobs. There's be some combination of higher wages, higher labor costs, less overall GDP growth, etc.
He's right that there wouldn't ever be "oodles of undone jobs." As for the rest, though, it brings up the question: what happened during the 1920s when the United States, in a bout of xenophobia, cracked down on immigration? Proponents of immigration controls often cite this period as proof that restrictions "work." After all: during the "roaring twenties" productivity was high, unemployment was low, nifty new consumer goods were practically spilling off the shelves, and, on the progressive side, labor unions were able to expand, supposedly unburdened by ethnic and racial rifts. (Michael Lind, among others, has made this argument.) Well, let's take a look then.

Certainly one effect of immigration controls was that businesses in the 1920s, finding unskilled workers in short supply, did what you'd expect: they substituted capital for labor. As an example, Bryan Lew and Bruce Cater argued that this explains why American farmers adopted the tractor in the 1920s more quickly than their counterparts in pro-immigration Canada. This economic shift was certainly good for some people—relatively skilled white-collar and blue-collar workers captured the bulk of the income gains here. But it's less certain whether relatively poor and unskilled blue-collar workers saw any rise in real wages between 1920 and 1929. And overall, inequality increased dramatically.

It's difficult to say how much of all of this was due to immigration, and how much might have happened anyway, but at any rate it's worth noting that restrictions weren't quite as good for unskilled workers in the 1920s as one might expect. (Conversely, immigration during the 1990s wasn't nearly as bad for unskilled native workers as one would expect—perhaps in part because, as Ethan Lewis has argued, firms adopted technological innovations to take advantage of the increased supply in unskilled labor, though it may have slowed firms from adopting changes to boost productivity among skilled workers.)

Now obviously during the Depression, World War II, and afterwards, wage and income inequality decreased dramatically, and eventually in the 1940s the floor started lifting for unskilled workers. Brad DeLong takes a crack at an explanation:
[Jeffrey] Williamson and [Peter] Lindert guess that about half of the reduction in wage inequality was due to a shift in the character of technological progress after 1929. Before 1929, productivity growth had been concentrated in manufacturing industry and other sectors that required a relatively skilled workforce. As the economy's structure shifted toward these sectors where productivity was growing most rapidly, demand for and returns to skills and capital rose and demand for unskilled labor fell. After 1929, productivity growth was much more balanced: productivity in agriculture and in service sector jobs that relied on unskilled labor more than skilled labor also grew rapidly.

The other half of the levelling comes from demographic factors: fewer children and more education per child both shifts the distribution of skills within the population—making more skilled and fewer unskilled workers—and diminishes the supply of unskilled workers. The levelling of the wage distribution was also encouraged by the growth in the number of jobs that were relatively low-skilled yet also paid high wages.
Perhaps continued immigration restrictions helped here. Now much of this "levelling" started to deteriorate in the 1970s, incidentally after Lyndon Johnson had loosened immigration controls in 1965 and women continued to enter the labor market (although it's worth noting that "unskilled" and lower-income women have been working since time immemorial). How did this affect things? One could note that most of the influx of unskilled workers into the labor market in the 1970s came to the service industry, which, according to Robert Gordon, did experience a productivity slowdown during that time. On the other hand, William Nordhaus argues that the bulk of the productivity slowdown in the 1970s was concentrated in energy-intensive industries, so the oil shock, and not immigration, seems to be the greatest factor here.

At any rate, this is a very crude sketch, and hardly the complete story. But one possible lesson, it seems, is that a restricted supply of unskilled labor may well lead to quality productivity gains. (And a surplus of unskilled workers may hurt those gains.) One way to tighten the supply, then, is to restrict immigration, as was done from the 1920s to 1965. But another way is for the government to loosen immigration controls and simply pursue full employment—through monetary policy, public work programs (preferably not involving military expansion) and other government policies, along with better education to get whatever proportion of skilled to unskilled workers is needed. If that's possible to achieve, then the effect should be essentially the same. I haven't seen anyone make this argument, so very likely it's wrong, but it doesn't seem wrong.

Immigration Control Through the Ages
-- Brad Plumer 3:05 PM || ||