Heritage's Updated Study on the Welfare Costs of Immigrants: Garbage In, Garbage Out

The problem with many environmental studies, conservatives are fond of pointing out, is that they are too often "garbage in, garbage out."  This means that the results of the studies are so dependent on their modeling assumptions that they will pretty much prove whatever the researcher wants them to prove. This same criticism can now be directed at an updated Heritage Foundation study released yesterday examining the welfare costs of low-skilled immigrants.

Conducted by Robert Rector and Jason Richwine, the study found that amnesty for illegals would impose a net fiscal cost of $6.3 trillion over the course of their (immigrants', not Rector/Richwine's) lifetime.  "Given the U.S. debt of $17 trillion, the fiscal effects detailed in our study should be at the forefront of legislators' minds as they consider immigration reform," Rector and Heritage President Jim DeMint wrote in the Washington Post yesterday.  Actually, as I noted here and here, what should be at the "forefront of legislators' minds" is that a foreign-born underclass is a bargain for taxpayers compared to a native-born one and hence might help strengthen the safety net.

Be that as it may, these results are completely in line with Rector's controversial 2007 study, which is credited with having derailed President George W. Bush's immigration reform. That study found that low skilled immigrants imposed nearly $1 trillion in annual fiscal costs on American taxpayers.

However, it was riddled with methodological flaws most of which, unfortunately, the new study does not even acknowledge, let alone correct.

One big flaw that Alex Nowrasteh of the Cato Institute points to is that in this study Heritage abandons its usual methodology of dynamic scoring. Heritage criticizes CBO and other government agencies for applying a static methodology when it comes to examining the fiscal effects of tax cuts. This approach counts the revenues the government looses from tax cuts, but does not factor in the revenues the government gains from the economic growth and productivity gains that these cuts engender, thereby greatly exaggerating the impact of tax cuts on government coffers.

But Rector does the exact same thing in his study. He counts the costs that welfare for low-skilled immigrants would impose on government coffers, but neglects the fiscal impact of the economic growth that these immigrants would spur. Studies that have done a fuller accounting suggest that the GDP-growth spurred is greater than the fiscal costs of low-skilled immigrants.

A 2006 analysis by the Texas comptroller estimated that low-skilled unauthorized workers cost the state treasury $504 million more than they paid in taxes in 2005. Without them, however, the state's economy would have shrunk by 2.1 percent, or $17.7 billion, as the competitive edge of Texas businesses diminished.

Likewise, a 2006 study by the Kenan Institute at the University of North Carolina found that although Hispanic immigrants imposed a net $61 million cost on the state budget, they contributed $9 billion to the gross state product.

Furthermore, a 2012 paper for Cato Institute written by UCLA's Raul Hinojosa-Ojeda that deployed a dynamic model found that immigration reform would increase U.S. GDP by $1.5 trillion in the 10 years after enactment. Hinojosa-Ojeda also ran a simulation examining the economic impact of the policy favored by Heritage: the removal or exit of all unauthorized immigrants. The economic result would be a $2.6 trillion decrease in estimated GDP growth over the next decade.

The other big problem with this version of Rector's study, as with the last one, is that it compares the welfare costs of average American households with low-skilled immigrant ones to show how much more the latter consume in welfare. But if you compare the latter with lower-income American families, the advantage is on the side of immigrants.

Rector's original study was also criticized for engaging in single-entry bookkeeping. This means he counted what the (American-born) children of low-skilled immigrants cost in welfare but not the taxes they'd pay when they grew up. Rector responds to this criticism this time by noting: "Even if all the children of unlawful immigrants graduated from college, they would be hard-pressed to pay back $6.3 trillion in costs over their lifetimes." Perhaps, but that misses the point. If he doesn't want to count their contributions then he should remove them from the equation altogether.  Most of these kids are Americans anyway. Simply look at the welfare consumption of low-skilled immigrant adults. That number would be a lot, lot less than the eye-popping $6.3 trillion because these folks don't qualify for the full panoply of means-tested benefits that natives do.

But Rector completely gives his game away with his claims about the wage and employment effects of immigration that have nothing to do with the core question of his study:

[U]nlawful immigration appears to depress the wages of low-skill U.S.-born and lawful immigrant workers by 10 percent, or $2,300, per year. Unlawful immigration also probably drives many of our most vulnerable U.S.-born workers out of the labor force entirely. Unlawful immigration thus makes it harder for the least advantaged U.S. citizens to share in the American dream. This is wrong; public policy should support the interests of those who have a right to be here, not those who have broken our laws.

There is a vast and rich economics literature on this subject that Rector seems to be quite innocent of. Plenty of studies have found that immigrants don't depress but stimulate the labor market for natives because they allow more businesses to form. As for the wage effects, even restrictionists' favorite economist George Borjas' 2003 paper failed to find any. Borjas disaggregated the impact of low skilled immigration on different native groups and found that, over the long run, the overall impact on their wages was zero. Only one group, high school dropouts, felt a noticeable negative impact. However, a subsequent study by Giovanni Peri and Gianmarco Ottaviano failed to corroborate Borjas' findings even for native high-school dropouts. They found a positive long-run effect of 0.3 percent. In other words, no one — not even high school dropouts lose in the long run due to low-skilled immigration.

 Rector managed to derail immigration reform last time because responsible conservative academics did not have time to take a close look at his claims. Not so now. Many respectable figures, including folks at the American Enterprise Institute and former CBO director Douglas Holtz-Eakin, are questioning Heritage and its methodology.  For a great round up of conservative reaction to the study, read this excellent piece by WaPo's Jennifer Rubin.

(For more on this and other immigration-related issues buy Reason's latest ebook, Humane and Pro-Growth: A Reason Guide to Immigration Reform, a compendium of Reason's best work by its best writers over the last seven years for the bargain price of $2.99.)