Local Fiscal Effects of Illegal Immigration: Report of a Workshop (1996)

Workshop participants noted that in order to produce accurate and reliable estimates of the fiscal costs of illegal immigration, information on the number and characteristics of illegal immigrants (and their families), the actual use of services, and the actual public cost of providing those services are needed. The workshop included extensive discussion about the use and costs of selected public services, including education, health care, and incarceration and criminal justice. Limited workshop discussion centered on the problem of ascertaining and measuring what revenues illegal immigrants actually contribute and the amount of their contributions. This section summarizes the workshop discussions of the fiscal effects of illegal immigration on education, health care, and criminal justice, and on revenues.

Each of the studies estimated the cost of providing elementary and secondary education to undocumented immigrants, including the cost of educating the U.S.-born children of undocumented immigrants. Although some of the case studies distinguished the country of birth of the children of illegal immigrants, claims to the federal government for the educational costs of illegal immigrants typically include costs based on the assumption that all children of undocumented aliens are themselves undocumented. However, an unknown proportion of children of illegal immigrants are native born and hence are U.S. citizens. This raises a question as to whether the costs of educating the citizen children of illegal immigrants should be included in the cost estimates (as a cost of illegal immigration)

or not included because they do not directly benefit the illegal immigrants. If the underlying question is what the costs would be if there were no illegal immigration, then it is appropriate to include costs for the legal children because the children would not be in the area if their parents had not immigrated.

All of the studies used per pupil expenditures to estimate the cost of educating the additional immigrant children entering the school system. Workshop participants identified six unresolved issues affecting the reliability of cost estimates for kindergarten to grade 12 (K-12) education: geographic variation in expenditures, program-specific variations in costs, capital costs, consideration of dynamic issues, mandated goods and locally chosen goods, and use of average versus marginal costs.

Average pupil expenditures vary widely within metropolitan areas and states. Because illegal alien school-age children are geographically concentrated, their schooling costs are likely to differ from state and city averages—even within school districts.

Participants noted that education costs vary across individuals within school districts, as well as across cities. Available studies of schooling costs suggest that it is more expensive to educate poor children. If children of illegal immigrants are poor, then they may be more expensive to educate, other factors being equal. This is an issue for further study.

The substantial variation in per pupil expenditures across school districts and among states is a common characteristic of U.S. elementary and secondary public education. Despite some intergovernmental financial support, school district budgets are still largely a function of local property tax revenues. Districts with wide variations in tax capacity but equal tax rates therefore collect significantly different amounts of revenue. At the same time, undocumented immigrant populations, who typically have very low household incomes, tend to live in areas with low-priced housing and low property values. This geographic variation in fiscal resources leads to variation in educational spending. This fact, along with the uneven geographical distribution of undocumented immigrants, strongly suggests that per pupil expenditures for undocumented children may diverge substantially from state or national average per pupil expenditures.

Per pupil expenditures differ across school districts not only by dollar amount but also by the number and type of programs provided. Even when programs are federally mandated, there is still significant variation across districts that reflect different choices made by the state and the districts about how to meet federal and state requirements. Because local districts often have programs that are related to

or are enhancements of mandated programs, it is difficult to separate the cost of mandated programs from the cost of program components initiated at the local level. As workshop participants observed, it is often a local decision to set taxation levels, to set priorities for funding for education and other local public service programs, and to define the quality and allocation of resources to programs.

Although there are few program cost estimates for illegal alien children, it is assumed that these children are more likely to be enrolled in programs such as bilingual education. In New York City, the school system estimates that bilingual education costs $2,000 per child annually (Dillon, 1994, citing an earlier report by Herman Badillo). It was not clear in the case studies the workshop examined if these costs were factored in, and, if so, how it was done.

Most of the case studies examined the costs for operating budgets for education. As with other cost estimates in the case studies, it is unclear if per pupil costs or average program costs take into account capital costs, such as providing new infrastructure, a pressing concern in many school districts with an aging capital stock. And it is unclear what proportion of capital costs (the costs of building a new school, for example) should be attributed to illegal immigrants. It is possible that annual school budgets in some areas include capital expenditures; in such cases, the education budget divided by the number of students would effectively take capital costs into account. After reviewing the six case studies, however, workshop participants were still uncertain about whether school cost estimates included capital costs and, if so, whether the costs were periodic or were project specific. One workshop participant suggested including depreciation in cost estimates as one way to factor in capital costs.

Costs of public services change over time, depending on social and economic changes separate from the effects of illegal immigration. A related issue, and one difficult to evaluate in monetary terms, is the social cost of adding children to a classroom or school if the addition of illegal immigrant children results in increases in classroom size. At what point, if any, does increasing class size affect the quality of education? A workshop participant pointed out that the concept of education quality has proven elusive and difficult to measure. The other side of the issue is the difficulty of measuring the positive externalities of public education—the value of socializing immigrant (and other) children into civic life, the cultural and personal value of a literate and numerate population, or the greater diversity that immigrants bring to the classroom. Educational investments have the long-term benefit of producing a more highly skilled workforce,

workshop participants noted. Debate about the impact of illegal immigrant children, according to workshop participants, might be more appropriate to discuss within a broader context of the value of education as a long-term social investment. Without providing answers, workshop participants suggested that a question is whether analysis should be limited to annual effects on government budgets or whether the effects should be considered over the lifetime of current residents.

Some public services are mandated by the federal or state government. For services that are offered by local jurisdictions (i.e., not federally mandated), it is less clear how and how much of the cost of their use by illegal immigrants to attribute to a federal responsibility. In education, the question of how best to balance responsibility across levels of government is of growing importance. The current federal role in primary and secondary education is a limited one, but as federal mandates increase, state and local governments are increasingly arguing that the federal government should subsidize local costs.

Using per pupil costs with cross-sectional data sidesteps cost and revenue issues if time is a critical consideration: the applicability of average cost or marginal cost may be time sensitive. A workshop participant pointed out that 10 years ago, when the New York City school system had significant excess student capacity, the marginal cost of adding some children to the school system was negligible. But today, with overcrowded schools, the city has requested $20 billion in direct impact aid from the federal government to cover the additional cost of educating undocumented children. Over time, costs are a function of the capacity utilization level, not simply the number of additional children. Cost estimation is especially difficult when the demand creates the need for substantial additional capital expenditures for new schools.

Estimates of the cost of providing health care to undocumented immigrants differed among the case studies based primarily on what services were included in the cost estimates.

The Urban Institute study of seven states (Clark et al., 1994) limited its analysis to emergency medical care for undocumented immigrants mandated under Medicaid. Because of a lack of central, uniform record keeping, administrative records were judged too unreliable to use. As a result, the researchers constructed "benchmark" cost estimates for each state. The benchmarks were

based on per capita costs of emergency care mandated under Medicaid for eligible legalized aliens. After legalization, in the 1980s, legalized aliens became eligible for emergency medical services for 5 years, under Medicaid. In addition, the federal government funds the nonfederal share of emergency Medicaid costs through special grants to states—the State Legalization Impact Assistance Grants (SLIAG). Using SLIAG data, which are available by state, the Urban Institute analysts calculated the actual cost of emergency medical services for a population (i.e., legalized aliens) presumed to be similar to undocumented immigrants. The benchmark was calculated as the per capita legalized alien Medicaid emergency service cost (by state and year), multiplied by the number of undocumented immigrants (by state and year). Because of factors assumed to encourage utilization by legalized aliens and discourage utilization by undocumented immigrants, the benchmarks were used to approximate the upper bound of emergency Medicaid cost for undocumented immigrants. Given the dearth of reliable administrative data, the benchmarks are an innovative approach for establishing a marker against which emergency Medicaid cost estimates can be evaluated.

Nevertheless, for the Urban Institute and other case studies, emergency medical service costs under Medicaid remain difficult to estimate. First, a medical emergency is determined by the physician, not Medicaid or the state, and definitions are variable. Second, emergency Medicaid spending is an open-ended entitlement. As a result, projecting emergency Medicaid costs on the basis of past medical practice may lag behind changes in medical judgment and practice and changes in the health conditions of illegal immigrants, observed one workshop participant.

Estimating total medical costs for state and local government is also difficult. In many states, localities are responsible for the medical expenses of indigent people, which are typically not included in cost estimates because data are not available by immigration status. On the revenue side, the estimates do not include the federal money that goes to underserved areas under Medicaid and Medicare disproportionate share reimbursement practices. A variety of other federal program funds directed at needy areas are also excluded from the cost estimates. Although these programs do not specifically target undocumented immigrants, the undocumented alien population is likely to live in medically underserved areas, and, presumably, benefit from such spending.

When studies attempt to attribute health care costs to undocumented immigrants, the estimates may lack overall credibility. The San Diego study, for example, attributed about 5 percent of in-patient admissions to undocumented immigrants, based on an estimate that the illegal immigrant population comprises approximately 5 percent of the total resident population (Parker and Rea, 1993:107). Because the characteristics of the undocumented population (typically young, healthy, and male) differ from the characteristics of the in-patient population (usually older and less healthy), however, the San Diego estimate appears inflated. In the Huddle studies of Texas and the United States (1993,

1994a), health care costs are extremely high because spending on long-term care is included.

An additional challenge of monitoring and measuring health care costs for illegal immigrants is the fact that the nature of emergency medical care often precludes establishment of identity before treatment. So, for the important area of emergency care, it would be difficult to restrict care to only legal residents. And, if emergency care is given, it is difficult to define the treatment duration for the care, observed one workshop participant.

Cost estimates for incarceration and criminal justice were discussed briefly during the workshop. Participants noted that the Urban Institute study (Clark et al., 1994) of seven states devised several innovations in calculating incarceration cost estimates (discussed below), while relying on new INS analysis of the number of illegal immigrants in prison. Overall, the various case studies are difficult to compare because two were limited to estimating the cost of incarceration in state prisons (the California and Urban Institute studies), while the other four attempted to estimate costs for the entire criminal justice system—that is, police, courts, prosecution and publicly paid defense attorneys, prisons and jails, and probation or parole. Participants noted that there are incidental and federal criminal justice costs that are omitted from local and state case studies, although they did not discuss them in detail. Incidental costs include the costs of monitoring or policing immigration policies, such as checking legal immigration status before hiring an employee or providing access to health services. Federal costs include judicial, incarceration, and deportation costs for illegal immigrants.

Table 2 displays the incarceration and criminal justice costs of undocumented immigrants in the six studies. Several differences were mentioned by workshop participants. Both the San Diego (Parker and Rea, 1993) and Los Angeles County (1992) studies included the same elements in their cost estimates, yet the estimated cost for San Diego, the much smaller and less populated county, was 36 percent higher than that for Los Angeles County. More broadly, the Urban Institute estimates of incarceration costs for seven states (Clark et al., 1994), with 86 percent of the undocumented population residing in those states, were $471 million, while Huddle (1993) estimated national costs at $1,031 million: clearly, the estimates cannot both be accurate.

Some of the differences across the studies stem from different cost components and from the misidentification of the undocumented criminal population. For the last point, workshop participants noted that the Urban Institute report (Clark et al., 1994) adds useful new methods and estimates to current analysis. Working with INS data analysts, the Urban Institute researchers have developed a way to identify illegal immigrants in the prison population. Previous studies have assumed that the relevant population was prisoners who could be deported

or excluded under INS rules. However, deportable or excludable prisoners include both legal criminal immigrants and illegal criminal aliens. The Urban Institute included only illegal criminal aliens (at the time the crime was committed) in their analysis.

A two-step process was used to identify illegal aliens in the prison systems. The first step was to identify all foreign-born prisoners in the states' prisons, based on a mandatory reporting requirement of the U.S. Department of Justice. The second step was to determine the immigration status of those prisoners. Initially, the data on the foreign born were matched to an INS automated database that identified immigration status. A random sample of the foreign-born prisoner population that was not matched was selected from the database. Then, the immigration status of the sample population was determined either through a manual match of INS data, a search of paper files in INS district offices, or by personal interview. The result was an estimate of the number of foreign-born prisoners who were both illegal immigrants and subject to deportation or exclusion under U.S. immigration laws.

Accurate and comprehensive measurement of revenues is crucial for a complete assessment of the fiscal effects of illegal immigration. None of the case studies discussed at the workshop carried out a thorough estimate of revenues. Such federal revenue sources as grants-in-aid and disproportionate share payments for health care services in underserved areas were included in some studies yet not in others. Property tax revenues, the most important revenue source for local school districts, were treated in radically different ways in the case studies: one study assumed property owners bore the entire tax burden, other studies assumed a 50-50 split between landlords and tenants, and the remaining studies did not state an assumption about the economic incidence. The result, quite predictably, was different estimates of immigrants' tax payments. When taxes are a function of income, it is clear that improved estimates of total income, including earnings not reported for state or federal tax purposes, are needed.¹ Estimates of remittances to one's home country are important for estimating the amount of money undocumented immigrants spend in the United States, which is useful for estimating sales taxes paid.

Beyond these empirical issues, workshop participants were also concerned that cross-sectional estimates for illegal immigrants tended to understate revenues and overstate costs. Income and, by inference, tax payments tend to increase with age and labor market experience, while costs, particularly education

costs, occur early in life. Undocumented immigrants are generally younger than legal immigrants and the native-born population. A cross-sectional analysis of this population pinpoints a specific time when service costs are high and revenues low, a balance that could change as the population ages. A related comment made by workshop participants is that the case studies often relied on point estimates developed from cross-sectional data. Too few studies, participants noted in general, develop an overall analysis taking into account the statistical uncertainty in the underlying estimates.

Workshop participants noted that there may be indirect economic benefits of illegal immigration, such as cheaper labor costs, but that there are few available empirical studies on this topic. Although such benefits may be offset by other factors, such as labor displacement, it would be useful to see attention to such indirect effects of immigration.

Some economists at the workshop suggested the possible uses of computable general equilibrium (CGE) models to better relate variant conditions in immigration and the economy to revenue and cost outcome conditions. Several workshop participants suggested that the value of CGE models was that they offer a modeling strategy that recognizes the complexity of population and economic factors. For use in the context of assessing costs and revenues associated with illegal immigration, however, further investigation is needed to determine their usefulness.