Immigrants, welfare reform, and the U.S. safety net

That is the title of an intriguing new economics paper by Marianne Bitler and Hilary W. Hoynes, official NBER version here.  Remember when they cut some benefits for immigrants, circa 1996?  That can form the basis for a natural experiment, because non-immigrant poor families did not experience a similar cut in benefits.

I urge extreme caution in the interpretation, but here is one result:

The difference-in-difference estimates show that poverty rates declined for children in immigrant-headed households compared to natives post-welfare reform (2008-2009) relative to pre-reform (1994-1995).

But why?  There is more:

This result is unexpected but may be explained by a change in the composition of immigrant children (see Figure 3).  That is…the difference-in-difference reflects the decrease in immigrant poverty in the 1994-1999 period.

You can take this as a mix of optimism about immigrants and skepticism about some welfare programs, or perhaps optimism about how a health job market helps immigrants more than non-immigrants.  I don’t in Figure 3 see any actual measurement of the composition of immigrants, although immigrant households do show rising income levels over the critical years.  Stick by the caution mentioned above.  In any case, following the decrease in welfare benefits immigrant households rely more heavily on earned income, which should be taken as good news.  I would rather offer fewer benefits to immigrants and take more people in, to the extent that is the choice.

I don’t think this paper gets to the bottom of the puzzle it is studying, but it is an important piece of work.