similar concerns about assuming that current immigrants will eventually use government benefits in the same way as immigrants who have been in the United States for more than 10 years in 1994-95, controlling for age and education. However, the case of benefits is quite different from that of taxes.

The Age Profiles of Taxes and Benefits

We begin by discussing the estimated age profiles of taxes and benefits for 1993-94, which are important building blocks for the later longitudinal calculations. We have used the merged 1994 and 1995 CPS, and occasionally other data sources, to estimate age profiles for 25 different state, local, and federal programs, most of which are disaggregated by educational level, immigrant generation, and, for immigrants, time since arrival. 9 It should be borne in mind that the benefits and taxes estimated from these data will reflect the fact that the U.S. economy ended a recession in the first quarter of 1991. Real gross domestic product growth averaged 2.5 percent in 1992-93; unemployment began to fall in the third quarter of 1992 and averaged 0.6 points lower in 1993 than in 1992. Incarceration costs were estimated using the Public Use Micro Sample (PUMS).10 The estimated age profiles were adjusted to match national program totals (using the National Income and Product Accounts, NIPA), when weighted by the distribution of the population by age and immigrant status and summed.


Specifically, these are Old-age, Survivors, and Disability Insurance (OASDI), Medicare, Medicaid (noninstitutional), Medicaid (institutional), Supplemental Security Income (SSI), AFDC, school lunch, food stamps, Special Supplemental Nutrition Program for Women, Infants, and Children (WIC), energy assistance, rent subsidy, public housing, earned income tax credit, unemployment insurance, elementary and high school, bilingual education, public college, federal student aid, incarceration costs, federal retirement, military retirement, railroad retirement, workers' compensation, state and local retirement, and refugee assistance. Time since arrival is categorized into 0 to 4, 5 to 9, and 10+ years. Immigrant generation is first, second, and third and later, the last representing the majority of the population. Education is categorized into less than high school, high school, and more than high school.

For some items, there is a question of how to allocate benefits to individuals. For example, should AFDC receipts be allocated to the mother or to the children, or proportionately to each? We allocate such costs to the mother. In the longitudinal analysis, we track an immigrant and all the immigrant's descendants, so the AFDC receipts will in any case be attributed to the woman at the age at which the relevant children are in the household. For an immigrant arriving as a young child, and perhaps eventually qualifying for AFDC as a child, it will matter how we handle the attribution, but this is not the most important case. For a measure of the static (cross-sectional) fiscal impact of immigrants, the attribution would matter only if we did not count the costs of U.S.-born children of immigrants as a part of the cost of immigrants; however, it seems clear that such costs should be included. If they are, then attribution does not matter. Most other program items are allocated on an individual basis, with the exception of housing and food stamps. The AFDC discussion applies to these items as well.


Using PUMS, we estimated a logistic regression for probability of institutionalization of the population under 60, by nativity, age, education, and time since arrival. Generally, these incarceration rates were considerably lower for immigrants than for others.

The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement