Immigrant Welfare Consumption: A Response to Richwine

Jason Richwine recently published a short criticism of a new brief that Robert Orr and I wrote about immigrant and native benefit levels and use rates for means-tested welfare and entitlement programs.  This is another in a long series of blog post responses between those who support different methods for measuring native and immigrant welfare consumption so the response is wonky and does not revolve around a central question.  The title of Richwine’s criticism is “Obfuscating the Immigrant-Welfare Debate.”  Below, Richwine’s comments will be in quotes and my responses will follow.

“A few years ago I noted that ‘the amnesty movement has turned the political numbers game into an art form, systematically obscuring the trade-offs inherent in immigration policy.’ The movement has reached new heights of obfuscation with Alex Nowrasteh and Robert Orr’s Cato Institute study, ‘Immigration and the Welfare State.’”

Richwine hid half of our title: “Immigration and the Welfare State: Immigrant and Native Use Rates and Benefit Levels for Means-Tested Welfare and Entitlement Programs.”  Our entire title is important to defusing many of Richwine’s other complaints later in his piece.  The charge of obfuscation is serious but cutting off three-quarters of the words in our title does not enhance clarity.

“The Nowrasteh-Orr study says that’s all wrong. In fact, immigrants receive 39 percent less in welfare benefits than natives on a per capita basis. How is this possible? By including Social Security and Medicare as ‘welfare,’ for starters.”

As the title of our brief states, we included entitlement programs as part of the welfare state.  As we further explained in the first two sentences in our brief, we included them because they accounted for about 65 percent of all federal benefits outlays in 2016.  It is impossible to discuss the welfare state or the impact that immigrants have on it without including entitlement programs because they comprise its largest share.

It is normal to count entitlement programs like Social Security or Medicare as part of the welfare state.  Proponents of a larger welfare state like Jonathan Chait and Matt Bruenig consider Social Security to be part of the welfare state.  Even welfare-state skeptics like Robert Samuelson and others include it under the umbrella of the welfare state.

Many people use the term “welfare” to describe programs that they dislike while they use other terms to describe programs that they approve of.  Means-tested welfare programs fall into the former category, especially for conservatives, while entitlement programs fall into the latter.  Some may find that rhetorical distinction interesting, but it is not relevant in an analysis of how immigrants impact the welfare state and the myriad benefits it pays out.

“Lumping contributory entitlements with means-tested benefits is incredibly misleading in a group comparison. If I told you that Group A receives more food stamps than Group B, you could infer that a larger proportion of Group A is struggling economically and has turned to the taxpayer for support. By contrast, if I told you that Group X receives bigger Social Security checks than Group Y, then Group X probably earned higher wages, worked longer, contributed more to Social Security, and received a lower return on those contributions. Notice the difference there?”

There are several reasons to write the type of brief that Orr and I did.  One is to provide some evidence for one variable in a larger estimate of how immigrants affect the budget deficit.  The entire welfare state, including entitlement programs, is an important component of the cost side of immigration.  However, the consumption of means-tested welfare and entitlement programs is necessary but not sufficient to understand their fiscal impact as the taxes paid by them and by others because of a larger economy must also be counted and compared to alternative policies.

Richwine doesn’t seem concerned with the fiscal effects of immigration in his criticism but that question does concern him elsewhere.  Richwine’s flawed 2013 report for the Heritage Foundation on the estimated net-fiscal cost of amnesty finds that retirement programs (overwhelmingly Social Security and Medicare) account for 61 percent of the total increase in benefit outlays for amnestied illegal immigrants over a 50 year period.  Legitimate estimates of the fiscal impact of immigrants also reveal how important Social Security and Medicare are over any length of time when estimating the net fiscal impact of immigration.  If we care about how immigrants affect the national debt and immigrant consumption of retirement programs are a major source of outlays, then our analysis should include programs like Social Security and Medicare. That is just one good reason to include entitlement programs in analyses like these.

“Under the Nowrasteh-Orr model, an immigrant could be cashing a TANF check, shopping at the grocery store with food stamps, paying for doctors’ visits through Medicaid, living in a subsidized rental unit, heating it with energy assistance — and all the while be counted as receiving ‘less welfare’ than a native retiree who contributed to Social Security and Medicare his whole career and never once used a means-tested program. This is sophistry.”

Richwine writes as if Social Security and Medicare are like 401(k) programs where workers contribute and then draw down their own accounts in retirement. In reality, Social Security taxes the currently-working and then redistributes that revenue toward the currently-retired, which is why Social Security is running a deficit that is set to balloon over the next decade and continue to do so for the foreseeable future.  As a net present value, Social Security currently has about $34.2 trillion in unfunded liabilities over the infinite horizon and $12.5 trillion over the next 75 years.  Individual beneficiaries have to work a certain number of years to get benefits but the taxes they individually pay come nowhere close to covering the benefits they receive – just like other welfare state programs.

Orr and I also think that the dollar value of benefits received matters.  If immigrants use a wide number of cheap welfare programs then the cost can still be below that of natives who use only the more expensive entitlement programs.  Since entitlement programs are such an overwhelming portion of the American welfare state and they are facing a catastrophic funding shortfall, excluding them produces an incomplete welfare-state cost estimate.

“Social Security favors immigrants, as they have somewhat shorter careers, lower average incomes, longer lifespans, and appear to benefit more from spousal coverage. One study from the Social Security Administration calculated that late-Baby Boomers born in the U.S. will lose money to the system. Their Social Security contributions are greater than their benefits, such that participation ultimately results in a 0.4 percent tax on their lifetime earnings. Foreign-born participants, by contrast, receive a 1.2 percent subsidy from the system. (Illegal immigrants are actually pushing down the overall subsidy by paying into the system without being eligible to collect benefits. For legal immigrants, the subsidy is greater than 1.2 percent.) And that’s just Social Security. Medicare is even more progressive in its funding and payout structure, so it confers even greater net benefits on people with shorter careers and lower wages.”

Richwine omitted half the major findings in that paper.  According to that paper, immigrants are less likely to qualify for Social Security than natives are but, when they do, they receive a greater subsidy from the program.  The paper is limited entirely to those immigrants who receive benefits and excludes the tax contributions of those who do not receive benefits.  Richwine’s own wording here obfuscates far more than it clarifies. His statement could easily be interpreted as claiming that immigrants as a group harm the fiscal solvency of entitlement programs.  Let us be clear: Taxes paid by immigrants subsidize natives’ retirement benefits on net.

Furthermore, the paper that Richwine cites explains that Social Security is a welfare program that transfers tax revenue to poorer retirees.  As the authors write, “because of Social Security’s progressive benefit structure, more recent immigrants would be expected to have higher relative returns.”  Richwine cannot have it both ways.  He cannot claim that Social Security is a retirement system unlike welfare and that it is a progressive wealth redistribution program that favors poor immigrants.

“Entitlements aside, why do immigrants and natives appear to use about the same amount of means-tested benefits in the chart above? Because U.S.-born children of immigrants are counted in the native category. So if an impoverished immigrant signs up her U.S.-born child for Medicaid (or any other means-tested benefit), this would be considered native use of the welfare system. Treating parents as economic units separate from their dependent children is simply incorrect, as any program for children benefits the parents who are otherwise responsible for them.”

Eligibility for Medicaid benefits is determined by the individual characteristics of the beneficiary just as they are for most means-tested welfare and entitlement programs.  But including U.S.-born children of immigrants in the same household headed by an immigrant is not good enough to achieve Richwine’s goal.  All of the U.S.-born children of immigrants, including multiple generations of adult citizens living in independent households, must also be considered for his preferred accounting of benefit consumption to be accurate.  That is a considerably more difficult statistical task, but one that Orr and I are currently working on.  We do not know how that will turn out, but we will report the findings when (if) we arrive at a result derived by defensible methods.

Regardless, if you want to count the U.S.-born children of immigrants as a welfare cost of immigration then there is no good reason not to also count their grandchildren, great-grandchildren, and successive generations.  After all, very few Americans would be here today without any immigrants since the 1500s.  Any method that includes nearly all Americans as immigrants is transparently absurd.  We make the easiest and most defensible choice here: counting consumption by the immigrants.

“More broadly, it illustrates the unwillingness of advocates to give a single inch on any talking point.”

A criticism of previous Cato research on immigrant welfare use is that we limited it to the poor by comparing poor immigrant use rates and benefit levels to those of poor native-born Americans.  We did this to compare the actual populations that are eligible for means-tested welfare programs.  Because of that criticism, our primary model in this brief was to compare all immigrants to all natives without adjusting for age or poverty.  Under that adjusted method, we found that immigrants individually consume 39 percent less in welfare and entitlement benefits than natives.  Our second model adjusts for age and poverty.  It produced a smaller estimated difference in benefit-use between natives and immigrants.  However you feel about the conclusions reached in our brief, it is unfair to say that we have not given a single inch as we took previous criticisms to heart and adjusted our research accordingly. 

Our brief is transparent, we clearly explain our methods, and the entitlement programs fit neatly under the welfare state umbrella.  It is up to the readers to make their own judgment as to the quality of our research.  All we ask is that they read the entire piece with an open mind – including all of the title.