New work on undocumented immigrants and the American labor force

working

Source: Flikr Creative Commons

Donald Trump calls for deporting all 11 million undocumented immigrants.  What effects would this have on the US economy?

To answer this question, we first need to have a sense of the labor supply behavior of undocumented immigrants.  That is, we need to know information such as their employment rate and their sensitivity to changes in US wages (i.e. their labor supply elasticity.)

George J. Borjas, one of the leading economists who study the economic effects of immigration, has a new working paper that is chock full of data, including estimates of employment rates and employment responsiveness to wages.  I’ll summarize his two key findings:

  1. Employment rates: “Undocumented immigrant men are far more likely to work than other groups, while undocumented immigrant women are far less likely to work (p. 27).” Further, “the employment gap that distinguishes undocumented men from the other groups widened dramatically over the past twenty years. By 2014, the probability that an undocumented man was employed in the CPS reference week was around 12 percentage points larger than that of native men. The probability that undocumented women are employed also grew at a relatively faster rate, but the increase was far less dramatic.

Here are Borjas’s data for men:

 Fig 6

In 1994 native-born and undocumented immigrant men had roughly the same employment rates.  By 2014, 85 percent of undocumented immigrant men had jobs compared to 81 percent of legal immigrants and 74 percent of native-born men.

One hypothesis that leapt to my mind was that lower-skilled undocumented men were substituting for similarly qualified legal and native-born men. Borjas investigates this along with a variety of other factors and concludes, “the key lesson resulting from this exercise is that the ‘usual suspects’ do not explain why the three nativity groups experienced such differential trends in labor supply over the 1994-2014 period (p. 19).”  He later notes, “The factor driving the relative increase in employment of legal and undocumented immigrants remains unexplained (p. 20).”

Here are the data for women:

Fig 7

Two things stand out in this picture.  First, the pattern of employment is reversed for women: native-born women have the highest employment rate followed by legal immigrants and then undocumented immigrants.  Second, the employment rates for immigrant women (both legal and undocumented) are much lower than their male counterparts, which isn’t true for native-born women.

  1. Labor supply elasticity: “The labor supply of undocumented workers is not as responsive to wage changes as the labor supply of the other groups in the population. In fact, the data clearly suggest that the labor supply of undocumented men is almost perfectly inelastic (p. 27).” That is, changes in American wages don’t seem to have much, if any, effect on the hours worked by undocumented workers.  In particular, the data indicate that male undocumented immigrants work an average of 36 hours per week regardless of their wage.  This compares to 34.5 hours per week on average for the population as a whole.

Borjas notes that his analysis is only a first step in analyzing the economic effects of various immigration reform proposals ranging from deportation to a path to citizen ship for undocumented immigrants.

There is certainly much more work to be done.  However, this first step is important as it tells us that the majority of undocumented immigrants have paid jobs and that they continue to work regardless of what happens to their wage.  If we’re going to have a sensible conversation about immigration policy we need to start with these estimates.