The welfare magnet: for more and more immigrants, America is becoming the land of welfare opportunities

National Review, March 11, 1996 by George Borjas

THE evidence has become overwhelming: immigrant participation in welfare programs is on the rise. In 1970, immigrant households were slightly less likely than native households to receive cash benefits like AFDC (Aid to Families with Dependent Children) or SSI (Supplementary Security Income). By 1990, immigrant households were more likely to receive such cash benefits (9.1 per cent v. 7.4 per cent). Pro-immigration lobbyists are increasingly falling back on the excuse that this immigrant - native 'welfare gap' is attributable solely to refugees and/or elderly immigrants; or that the gap is not numerically large. (Proportionately, it's 'only' 23 per cent).

But the Census does not provide any information about the use of noncash transfers. These are programs like Food Stamps, Medicaid, housing subsidies, and the myriad of other subsidies that make up the modern welfare state. And noncash transfers comprise over three quarters of the cost of all means-tested entitlement programs. In 1991, the value of these noncash transfers totaled about $140 billion.

Recently available data help provide a more complete picture. The Survey of Income and Program Participation (SIPP) samples randomly selected households about their involvement in virtually all means-tested programs. From this, the proportion of immigrant households that receive benefits from any particular program can be calculated.

The results are striking. The 'welfare gap' between immigrants and natives is much larger when noncash transfers are included [see table]. Taking all types of welfare together, immigrant participation is 20.7 per cent. For native-born households, it's only 14.1 per cent -- a gap of 6.6 percentage points (proportionately, 47 per cent).

And the SIPP data also indicate that immigrants spend a relatively large fraction of their time participating in some means-tested program. In other words, the 'welfare gap' does not occur because many immigrant households receive assistance for a short time, but because a significant proportion -- more than the native-born -- receive assistance for the long haul.

Finally, the SIPP data show that the types of welfare benefits received by particular immigrant groups influence the type of welfare benefits received by later immigrants from the same group. Implication: there appear to be networks operating within ethnic communities which transmit information about the availability of particular types of welfare to new arrivals.

The results are even more striking in detail. Immigrants are more likely to participate in practically every one of the major means-tested programs. In the early 1990s, the typical immigrant family household had a 4.4 per cent probability of receiving AFDC, v. 2.9 per cent of native-born families. [Further details in Table 1].

And that overall 'welfare gap' becomes even wider if immigrant families are compared to non-Hispanic white native-born households. Immigrants are almost twice as likely to receive some type of assistance -- 20.7 per cent v. 10.5 per cent.

The SIPP data also allow us to calculate the dollar value of the benefits disbursed to immigrant households, as compared to the native-born. In the early 1990s, 8 per cent of households were foreign-born. These immigrant households accounted for 13.8 per cent of the cost of the programs. They cost almost 75 per cent more than their representation in the population.

The disproportionate disbursement of benefits to immigrant households is particularly acute in California, a state which has both a lot of immigrants and very generous welfare programs. Immigrants make up only 21 per cent of the households in California. But these households consume 39.5 per cent of all the benefit dollars distributed in the state. It is not too much of an exaggeration to say that the welfare problem in California is on the verge of becoming an immigrant problem.

The pattern holds for other states. In Texas, where 8.9 per cent of households are immigrant but which has less generous welfare, immigrants receive 22 per cent of benefits distributed. In New York State, 16 per cent of the households are immigrants. They receive 22.2 per cent of benefits.

The SIPP data track households over a 32-month period. This allows us to determine if immigrant welfare participation is temporary -- perhaps the result of dislocation and adjustment -- or long-term and possibly permanent.

The evidence is disturbing. During the early 1990s, nearly a third (31.3 per cent) of immigrant households participated in welfare programs at some point in the tracking period. Only just over a fifth (22.7 per cent) of native-born households did so. And 10.3 per cent of immigrant households received benefits through the entire period, v. 7.3 per cent of native-born households.

Because the Bureau of the Census began to collect the SIPP data in 1984, we can use it to assess if there have been any noticeable changes in immigrant welfare use. It turns out there has been a very rapid rise.

During the mid-1980s, the probability that an immigrant household received some type of assistance was 17.7 per cent v. 14.6 per cent for natives, a gap of 3.1 percentage points. By the early 1990s, recipient immigrant households had risen to 20.7 per cent, v. 14.1 per cent for natives. The immigrant-native 'welfare gap,' therefore, more than doubled in less than a decade.