Saturday, April 04, 2015

Debunking the Myth of the Job-Stealing Immigrant

"Debunking the Myth of the Job-Stealing Immigrant" is the title of a New York Times article by Adam Davidson.  Davidson is right on some key points, but ignores or is wrong on others.  Here are some excerpts from Davidson's article (in italics), along with my comments.

So why don’t we open up? The chief logical mistake we make is something called the Lump of Labor Fallacy: the erroneous notion that there is only so much work to be done and that no one can get a job without taking one from someone else. It’s an understandable assumption. After all, with other types of market transactions, when the supply goes up, the price falls. If there were suddenly a whole lot more oranges, we’d expect the price of oranges to fall or the number of oranges that went uneaten to surge.

The Lump of Labor Fallacy is real.  The idea that adding to a nation's population means more joblessness makes about as much sense as the notion, popular not so long ago, that computers would eliminate jobs and create joblessness. Computers did eliminate some kinds of jobs, but created enough others to avoid joblessness.   Here is another way to think about immigration's impact on jobs.  Suppose Britain  was physically moved to and attached to the US as another state.  Would that imply more unemployment?



However, even though the Lump of Labor Fallacy is real, there are likely to be short term effects, whereby at least some domestic workers are likely to be hurt.  Here is an example that provides useful perspective.

First, suppose immigration is not allowed.  The US is better off than Mexico, in part because its capital to labor ratio is higher - resulting in a higher marginal productivity of labor in the US than in Mexico.  Presumably, labor's wage reflects its marginal productivity, hence is higher in the US than in Mexico.  Mexican labor then has an incentive to move to the US to earn more.  Now allow unrestricted immigration.  Given static capital, this results in immigration of Mexicans to the US until the marginal product of labor is equal in both countries.  The new marginal product of labor is below the previous US marginal product of labor, hence labor's wage is lower.  Assuming that the marginal product of labor is monotonically declining, which is reasonable, this implies more jobs, not less, but at a lower wage.  The economy expands, but the output per capita, hence standard of living declines.

The above short term analysis does not apply to the intermediate and long term, because the stock of capital is not fixed.  The same economic forces that led to the previous US equilibrium with a particular capital to labor ratio will lead to a new equilibrium, with more capital and the same marginal product of labor as before the immigration.  Over time, labor's wage will return to its previous level.  The net impact will be an expanded economy with the original output per capita and the original standard of living.

The short term adverse effect described above neglects the increased need for infrastructure, food, clothing, etc. that immigrants require.  So the adverse effects are overstated because capital and infrastructure creation are likely to begin immediately, leading to considerable new demand for labor, e.g., construction, relative to pre-immigration levels.  Nothing assures that the net short term effect will be lower wages on average if the pre-immigration economy has excess capacity.  On the other hand, assuming excess capacity in the pre-immigration economy might be viewed as a debating "cheat".

But immigrants aren’t oranges. It might seem intuitive that when there is an increase in the supply of workers, the ones who were here already will make less money or lose their jobs. Immigrants don’t just increase the supply of labor, though; they simultaneously increase demand for it, using the wages they earn to rent apartments, eat food, get haircuts, buy cellphones. That means there are more jobs building apartments, selling food, giving haircuts and dispatching the trucks that move those phones. Immigrants increase the size of the overall population, which means they increase the size of the economy. Logically, if immigrants were “stealing” jobs, so would every young person leaving school and entering the job market; countries should become poorer as they get larger. In reality, of course, the opposite happens.

Davidson gets it partly right and partly wrong, here.  Immigrants are not the same as young people entering the workforce, because young people can be viewed as replacing retiring people.  The rest of his comment is correct in the long term, in that immigrants lead to increased demand for goods, hence motivate increased production, i.e., an expansion of the economy.  However, as noted above, without excess capacity, there will be a short term decrease in the average wage and reduction in the average standard of living even though the economy expands.

The single greatest bit of evidence disproving the Lump of Labor idea comes from research about the Mariel boatlift, a mass migration in 1980 that brought more than 125,000 Cubans to the United States. According to David Card, an economist at the University of California, Berkeley, roughly 45,000 of them were of working age and moved to Miami; in four months, the city’s labor supply increased by 7 percent. Card found that for people already working in Miami, this sudden influx had no measurable impact on wages or employment. His paper was the most important of a series of revolutionary studies that transformed how economists think about immigration. Before, standard economic models held that immigrants cause long-term benefits, but at the cost of short-term pain in the form of lower wages and greater unemployment for natives. But most economists now believe that Card’s findings were correct: Immigrants bring long-term benefits at no measurable short-term cost. (Borjas, that lone dissenting voice, agrees about the long-term benefits, but he argues that other economists fail to see painful short-term costs, especially for the poor.)

"no measurable impact" can cover a host of sins.  Some issues are measurement accuracy, statistical significance, data accuracy, and survey response accuracy.  In the face of reasonable theory, it is sometimes appropriate to question data analysis.  For example, how did Card allow for the fact that Miami's economy is "underground" to a significant extent, particularly with respect to jobs that are most likely to be influenced by immigrants?  It is unlikely that those with skilled jobs suffered, probably they gained.  But the the lower income jobs, many in the underground economy, may well have expanded while wages declined without Card ever knowing it.

To me, immigration is the greatest example of our faulty thinking, a shortsightedness that hurts others while simultaneously hurting ourselves. The State Department issues fewer than half a million immigrant visas each year. Using the 7 percent figure from the Mariel boatlift research, it’s possible that we could absorb as many as 11 million immigrants annually. But if that’s politically untenable, what about doubling the visas we issue each year? It would still be fewer than a million, or less than 0.7 percent of the work force. If that didn’t go too badly, we could double it again the next year. The data are clear. We would be better off. In fact, the world would be better off.

Davidson fails to appreciate that the US could be worse off even under circumstances that the world is better off.  Here is a dramatic example.  Assume that all countries suddenly live like the US, in that education, capital per capita, etc., are identical to US levels.  Then the world standard of living would be the same as the US standard of living. What could be better?  In fact, the world standard of living, and the new US standard of living, probably would be well below the original US standard of living.  Raw materials, e.g., metals, oil, probably would be considerably more expensive, as less accessible raw material deposits would have to be used, raising extraction cost, and non-renewable raw materials, e.g., oil, would, at some point, tend to become relatively scarce.

No comments: