We at Open Borders: The Case have not blogged much about the empirics of the welfare state/fiscal burden objection. We do have some thoughts on the matter which we hope to blog in the future. But one reason why this is a low priority, at least for me, is that I think that whatever the specific truth regarding welfare use by immigrants (under the status quo, modest variations thereof, or open borders), the keyhole solution of building a stronger wall around the welfare state to prevent immigrants and non-citizens from accessing it is politically feasible. Other keyhole solutions, such as immigration tariffs and DRITI, often run against opposition from voters across the political spectrum, but denying immigrants benefits seems to be quite politically popular. At any rate, (open borders + strong wall around the welfare state) seems no less feasible than (open borders without strong wall around the welfare state).
Nonetheless, the empirics of welfare usage, both under the status quo and under open borders, are important in so far as these allow us to better understand and prepare for the impact of open borders. With this in mind, I link to Economic Development Bulletin No. 17 put out by the Cato Institute. The bulletin is Poor Immigrants Use Public Benefits at a Lower Rate than Poor Native-Born Citizens and it is authored by Leighton Ku and Brian Bruen. It is a shorter version of a Cato working paper by the same name. Here is the conclusion of the bulletin:
Low-income non-citizen adults and children generally have lower rates of public benefit use than native-born adults or citizen children whose parents are also citizens. Moreover, when low-income non-citizens receive public benefits, the average value of benefits per recipient is almost always lower than for the native-born. For Medicaid, if there are 100 native-born adults, the annual cost of benefits would be about $98,400, while for the same number of non-citizen adults the annual cost would be approximately $57,200. The benefits cost of non-citizens is 42 percent below the cost of the native-born adults. For children, a comparable calculation for 100 non-citizens yields $22,700 in costs, while 100 citizen children of citizen parents cost $67,000 in benefits. The benefits cost of non-citizen children is 66 percent below the cost of benefits for citizen children of citizen parents. The combined effect of lower utilization rates and lower average benefits means that the overall financial cost of providing public benefits to non-citizen immigrants and most naturalized immigrants is lower than for native-born people. Non-citizen immigrants receive fewer government benefits than similarly poor natives.
These results seem to be at odds with research and findings published by the Center for Immigration Studies, one of the relatively more respected think tanks on the restrictionist side. The authors of the Cato bulletin explain the discrepancy as follows (I’ve removed the internal footnotes in the quoted text):
A study by the Center for Immigration Studies (CIS)found that immigrant-headed households with children used more Medicaid than native-headed households with children and had higher use of food assistance, but lower use of cash assistance. The CIS study did not examine the average value of benefits received per recipient.
There are several reasons why our study differs from CIS’s study. First, CIS did not adjust for income, so the percent of immigrants receiving benefits is higher in their study in part because a greater percent of immigrants are low-income and, all else remaining equal, more eligible for benefits. Non-citizens are almost twice as likely to have low incomes compared with natives. We focus on low-income adults and children because public benefit programs are means-tested and intended for use by low income people. It is conventional in analyses like these to focus on the low income because it reduces misinterpretations about benefit utilization.
Second, CIS focused on households headed by immigrants while we focus on individuals by immigration status. Our study focuses on individuals because immigrant headed households often include both immigrants and citizens. Since citizen children constitute the bulk of children in immigrant-headed households and are eligible for benefits, CIS’s method of using the immigrant-headed household as the unit of analysis systematically inflates immigrants’ benefit usage. For example, 30 percent of U.S children receiving Medicaid or CHIP benefits are children in immigrant-headed families and 90 percent of those children are citizens.
Third, CIS focused on immigrants in general, including naturalized citizens, while we also included non-citizen immigrants. Naturalized citizens are accorded the same access to public benefits as native-born citizens and are more assimilated, meaning their opinions of benefit use are more similar to those of native born Americans. Separating non-citizens from naturalized Americans gives a clearer picture of which immigrant groups are actually receiving benefits.
I haven’t had time to study the data carefully, but the most obvious counter-response seems to be that even if immigrants use benefits at a lower rate than natives, the fact is also that on average they pay less in taxes, so that they are still bigger net fiscal drains than natives. A related argument is that even if they do better than low-income natives, this is too weak an argument, because low-income natives are even bigger fiscal drains. But low-income natives are here to stay, while immigrants can be denied entry, so it makes sense to admit immigrants only if they are net fiscal pluses. In this view, immigrants performing better than low-income natives, even if true, is not a good enough argument to support more immigration. The “net fiscal burden” argument is one that we will take up on this blog some other time for more detailed discussion.
Another related point that is highlighted by this paper is what relevant groups one should look at when studying the effects of immigration. The position taken by the CIS is that the relevant groups to look at are all foreign-born people, including citizens and non-citizens, as well as the minor children of the foreign-born. Others have taken the position that we should look only at the proportion of the population that comprises non-citizen immigrants, and that it would be cheating to include their citizen children in the calculation. I tend to be agnostic on this question framed generally, since a lot depends on what specific aspect is being studied. For this reason, I like the fact that the Cato paper explicitly separates naturalized citizens and non-citizen immigrants, as well as separating children based on both their own and their parents’ immigration status, and dutifully reports all numbers.