Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.
OCR for page 183
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration 5 The Current Fiscal Impact of Immigrants and their Descendants: Beyond the Immigrant Household Ronald D. Lee and Timothy W. Miller INTRODUCTION Does immigration lead to higher taxes for state, local, and federal taxpayers? The question appears to be straightforward, but on closer consideration ambiguities emerge. Any answer must address a conceptual experiment in which the fiscal situations under two different immigration scenarios are compared. The result will depend on the particular experiment that is envisaged, or on the choice of scenarios to be compared. The possible experiments and scenarios differ along many dimensions, but in this chapter we focus on the demographic dimension: How is the immigrant study population defined? One type of experiment takes the form: Suppose that the number of immigrants arriving during some time period were different. What would be the fiscal implications for the balance of the population? Because the fiscal impacts unfold over time, the question is intrinsically longitudinal. Fiscal impacts are distributed over the lifetime of the immigrants and the lifetimes of their descendants and can be summarized by taking a present value. All the consequences of the difference in the number of immigrants, including the change in the descendants of the immigrants, would have to be taken into account. Because the impacts would depend on changing historical and future economic contexts, these would have to be taken into account as well. We call this type of experiment, and the resulting calculations, longitudinal. The experiment could be historical: Suppose 10 percent fewer immigrants had arrived between 1905 and 1909, for example. Or it could hypothesize a different number of immigrants arriving today and trace the consequences into the future. We believe that these longitudinal conceptual
OCR for page 184
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration experiments are the most appropriate guides for policy formation. Nonetheless, they are very seldom carried out. Lee and Miller (1997) and The New Americans (National Research Council, 1997:Chap. 7) are the most comprehensive longitudinal calculations available. Such calculations are complex and necessarily involve many assumptions and projections. Cross-sectional calculations are a simpler alternative to the longitudinal ones, but they are rarely justified by a clearly formulated conceptual experiment (see National Research Council, 1997, Chapter 6 for a discussion of these issues). One such experiment might be: How would the fiscal situation change if all the immigrants in the United States in some year were to vanish? We could attempt to answer this question by calculating the difference between all the taxes paid by immigrants and the marginal cost of all the benefits received by them. We would ignore any cumulative effect of the presence of the immigrants in the past, for example on government debt. Such a cumulative effect would be relevant in the longitudinal calculation, but not for this conceptual experiment.1 We call this the "immigrant only" approach, because no U.S.-born descendants are included in the calculation. This approach is quite common in cross-sectional analyses of fiscal impacts. Note, however, that this calculation takes no account of the costs and taxes for the U.S.-born children of immigrants, many of whom impose heavy public-sector costs by attending public schools and colleges. Although these children are U.S. citizens, and not immigrants, their presence is a direct consequence of their parents' immigration. For this reason, it appears appropriate to include their fiscal impact, which is typically done by analyzing the taxes and costs of benefits for all immigrant households or families. These are defined as households or families with an immigrant head or adult member. How would the fiscal situation differ in some year if all immigrant households (including U.S.-born members) were to vanish in that year? We call this the "immigrant household" formulation. This approach is also quite common in cross-sectional analyses of fiscal impacts. However, there is a serious bias in this type of study, because it counts the U.S.-born children of immigrants only while they are young and living with their parents and while society is investing heavily in their education. As soon as they leave home and get jobs, they no longer live in immigrant households, and their contributions are lost from sight. But if the children are to be included while young and costly, they should also be included when they are older and paying taxes (as argued, for example, in MaCurdy et al., in this volume). Furthermore, these children may themselves have children, who also should be included. Thus, we propose a conceptual experiment in which all the immigrants vanish in some year, along with all the descendants of those immigrants, 1 Taking into account the current effect of past fiscal impacts would be a complicated task because we do not know what the past fiscal impacts have been.
OCR for page 185
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration and the fiscal implications are assessed. Note that only the descendants of living immigrants are assumed to vanish along with the immigrants themselves. Descendants of deceased immigrants, comprising the vast majority of the U.S. population, would remain. We call this the "concurrent descendants" formulation. We believe that this is the first study to calculate fiscal impacts on this basis. The primary reason for doing fiscal impact calculations is to inform current policy decisions about the number and characteristics of immigrants to admit to the United States. Cross-sectional approaches such as the three outlined above have serious problems in relation to this goal. The cross section includes immigrants who have arrived at different periods and have different characteristics (such as educational attainment or English language ability). In addition, the volume of immigration has changed dramatically over time. This means that the age distribution of immigrants in the population is very uneven. If there are relatively few immigrants of retirement age, then the cross-sectional analysis will reflect low immigrant costs of Social Security and Medicare benefits, even though we know that most immigrants will survive to collect these benefits in future years. Likewise, if there are relatively many children of immigrants, then the analysis will reflect high costs of education and low tax payments, even though we know that most of these children will grow up to pay taxes over a lifetime of work. Furthermore, because taxes and benefits may vary over time as policy changes, their current levels may be misleading. For example, when the baby boom generation retires between 2010 and 2035, payroll taxes will have to be raised or Social Security benefits reduced. The cross-sectional calculation cannot take this into account. Finally, the cross-sectional results will depend on whether governments are running surpluses or deficits in the study year. In the longitudinal analysis, realistic constraints can be placed on the ability of governments to maintain unbalanced budgets in the long run; in the cross-sectional analysis, this cannot be done. For these reasons, the type of cross-sectional analysis reported here is inherently limited in its relevance for policy choices, and is certainly less useful for this purpose than the type of longitudinal analysis reported in Lee and Miller (1997) and National Research Council (1997:Chap. 7). Almost all studies of which we are aware have taken the cross-sectional approach, other than Lee and Miller (1997) and National Research Council (1997:Chap. 7). Rothman and Espenshade (1992:411) survey the fiscal impact literature for the United States. Of the 17 studies they review, 9 take the immigrants-only approach, and the remaining 8 take the immigrant household (or family) approach. Vernez and McCarthy (1996:7–8) update this survey with a look at nine studies since 1992. Most of these are based on the immigrants-only approach, whereas a few are based on immigrant households or families. Two very recent studies reported in this volume (Garvey and Espenshade and Clune) are based on immigrant households. We are not aware of any previous study that
OCR for page 186
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration takes the concurrent descendant approach. It is worthwhile, therefore, to consider how the calculated impact depends on the definition of the study population.2 In this chapter we carry out calculations pertaining to all three of the cross-sectional experiments, at the federal, state, and local levels, using the same data and many of the same assumptions. In this way, we can isolate the effects of the setup of each conceptual experiment on the outcome. The actual results depend very sensitively on the formulation chosen, in both sign and magnitude. Although we believe that the longitudinal formulation is most informative, the cross-sectional formulations have dominated the policy debate. Within this class, we argue that the concurrent descendant formulation is strongly preferable to the immigrant-only or the immigrant household formulations. In this chapter we focus on the concurrent descendant analysis, which we report in some detail. We also report comparative results for the other approaches. DEFINING THE POPULATION OF IMMIGRANTS AND THEIR CONCURRENT DESCENDANTS In our cross-sectional analysis, we assess the current impact of past immigration policies that have led to the presence of 22.8 million immigrants and their additional descendants in the United States in 1994–1995. We restrict our attention to the impact of these immigrants in that year. Our group of interest is all immigrants resident in the United States in 1994–1995 and the surviving U.S. resident children and grandchildren of these still-living immigrants. Once an immigrant dies, we no longer assess the impact of that immigrant's descendants. We also report results for the immigrant-only and the immigrant household formulations. Our data sample is drawn from the March Current Population Survey (CPS) conducted in 1994 and 1995. The CPS is designed to represent the non-institutionalized population of the United States.3 We define immigrants to be those reporting themselves as foreign born. The CPS estimate of this population in 1994–1995 is 22.8 million. This includes legal immigrants, undocumented immigrants, and the so-called nonimmigrant groups (students, businessmen, and temporary workers). About 31 percent of this group reported themselves as naturalized citizens. 2 There are many variants on these basic methods. One can distinguish by country of origin and by time of arrival of the immigrant head of household. One can construct a quasi-longitudinal estimate by interpreting the fiscal impacts of householders at differing durations within the country as if they referred to the same immigrant householder over his or her life cycle (see Akbari, 1991). One can calculate fiscal impacts separately for first- and second-generation immigrants Clune (in this volume). 3 We have simply combined years to form an aggregate sample. The CPS is a rotating sample such that one-half of the households interviewed in March of 1994 are reinterviewed in March 1995. One-fourth of our observations come from households interviewed only in 1994, one-fourth from those interviewed only in 1995, and one-half from households interviewed twice, once in each year.
OCR for page 187
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration Second-generation immigrants (U.S.-born children of immigrants) are defined primarily as those U.S.-born respondents who report that both parents were foreign-born. There remains a group of children with one parent foreign-born and one parent U.S.-born, which we count as 50 percent second-generation immigrant and 50 percent third-plus-generation immigrant. We estimate that there are 21.1 million second-generation immigrants in the United States in 1994–1995. However, we count only the second-generation immigrants with surviving immigrant parents. To estimate the children of still-living immigrants, we assume that immigrants experienced the same mortality as the general population and that each parent (immigrant) is 30 years older than each child. 4 Of the 21.1 million U.S.-born children of immigrants, 65.5 percent have at least one parent alive by this calculation, which leads to our estimate of 13.8 million U.S.-born children of still-living immigrants. Of these, 5.6 million are the adult children of immigrants, who presumably are not members of immigrant households. We estimate the grandchildren of still-living immigrants by estimating the number of children of the 5.6 million adult children of still-living immigrants. The June 1994 CPS provides the number of children ever born by age for the children of immigrants, which is very similar to the figures for the third and higher generations. We estimate that there are 3.9 million grandchildren of still-living immigrants. Our group of interest is therefore composed of 40.4 million individuals: 22.8 million immigrants alive in 1994, their 13.8 million surviving U.S.-born children, and their 3.9 million grandchildren (born to U.S.-born children). How does the presence of these individuals affect federal, state, and local tax burdens? We turn now to estimation of the costs generated and the taxes paid by these individuals. ESTIMATION OF THE COSTS GENERATED AND TAXES PAID Immigrants, and indeed all people, affect the budget through four broad categories of government expenditures: public goods, servicing of public debt, congestible goods, and transfer programs (or assignable items). (See Table 5-1 for a breakdown of government budgets into these categories of expenditure.) We discuss each of these briefly here. A more detailed discussion can be found in Lee and Miller (1997) or National Research Council (1997:Chap. 7). MaCurdy et al. (in this volume) discuss the rationale for a categorization of this type. 4 This assumed gap of 30 years between first and second generations is somewhat greater than the average age at which immigrants give birth to allow for the childbearing occurring at younger ages before arrival in the United States. We used cohort survival values taken from the Social Security Administration (1992) actuarial report to estimate immigrant survival. This probably underestimates the number of second-generation immigrants with a surviving immigrant parent because the survival rates for immigrants are somewhat higher than for nonimmigrants (Swallen, 1996).
OCR for page 188
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration TABLE 5-1 Government Expenditures by Category Budget Category Federal Budget State and Local Budget Public Goods 24% 0% Congestible Goods 7 32 Transfers 55 53 Debt Servicing 14 15 Total 100 100 NOTE: Federal expenditures are for FY 1994; state and local are for FY 1994 and represent the total for all states and localities. SOURCE: The detailed list of expenditures is taken from 1996 US Statistical Abstract, the Annual Survey of Government Finances,and Analytical Perspectives, Budget of the US Government, FY1996. The table shows aggregates based on the principles discussed inthe text and appendix. Public Goods By definition, the cost of providing a given level of public goods does not increase with population size. Public goods are not congestible. The presence of immigrants and their descendants, therefore, does not increase the cost of providing public goods.5 In our analysis, we treat the following as public goods: national defense; expenditures on veterans; and research on health, science, space, and technology. All of these are programs of the federal government. We do not treat any of the expenditures at the state and local levels as public goods. Immigrants certainly share in the benefits from public goods; and in an accounting of the value—rather than the cost —of the benefits they receive, services from public goods would be counted. Servicing the Public Debt Tax-paying immigrants help shoulder the costs of servicing the federal debt, which in 1994 came to about $4.6 trillion. In 1994 the cost of servicing the debt represent 14 percent of the federal budget. In a longitudinal historical formulation it would matter whether immigrants themselves were partially responsible for causing this debt to be larger or smaller through their fiscal impacts in the past. However, in the strictly current cost versions it does not matter whether immigrants contributed to the debt in the past. If the immigrants were to vanish today, all other taxpayers would have to make larger tax payments to service the debt, regardless of its origin. Somewhat similarly, we note that there is a vast 5 However, the amount of public goods demanded may be influenced by population size. The larger the population, the cheaper the price per capita of providing the good and hence the greater the demand. The fact that the population of the United States was 260 million in 1994 means that public goods such as national defense, space exploration, and cancer research are relatively inexpensive.
OCR for page 189
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration federal debt in the form of unfunded obligations to pay pension and health benefits to future retirees through Social Security and Medicare. When the baby boom generation retires, it will not be possible for the federal government to meet these obligations without raising payroll tax rates or reducing benefits. Doing so will make the fiscal impact of immigrants more positive, which is taken into account in a longitudinal analysis. However, the current cross-sectional calculation is a snapshot for which neither the past nor the future matters. In calculating the fiscal impacts of immigrants and their descendants, we always treat them as incremental population members with zero marginal costs for public goods. Many of the tables in this chapter show fiscal impacts for the balance of the population. These calculations also treat the balance of the population as incremental, and therefore with marginal costs of zero for public goods. We treat the debt servicing payments of immigrants and the balance of the population in a similar, symmetric way. These procedures make sense so long as we add or delete a modest proportional increment to the population. For these reasons, the presence of public goods and public debt tends to make the fiscal impact of any incremental member of the population less negative or more positive. Congestible Goods Unlike public goods, congestible goods have nonzero marginal costs. Included in this category are expenditures on roads, fire and police protection, libraries, airports, sewers, and so on. In assigning costs to immigrants and to the balance of the population, we assume that marginal costs equal average costs for these items. Capital expenditures are included in these outlays. We also experimented with a more direct method of imputing the capital costs of replicating the social infrastructure (school buildings, roads, sewers, etc.) for immigrants by calculating the annual user cost of capital from an estimate of the net capital stock per capita in 1994. We found that it yielded very similar cost estimates. Transfer Programs Here we include all government expenditures that can be assigned to specific individuals. Many cash and in-kind transfers such as Social Security, Medicare, Medicaid, AFDC, earned income tax credits, energy assistance, food stamps, public housing, school lunches, and unemployment benefits are included. Also included are public education at all levels and the costs of incarceration. We draw on three types of information from the CPS for use in calculating transfers to and from immigrants and their families. The first type is responses to questions about amounts of benefits received. An example would be the amount of Social Security benefits received last year. In general, summing these responses over all individuals (appropriately weighted to reflect the nonin-
OCR for page 190
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration stitutionalized population) will not equal the total costs of the program as reported in the budget. For example, summing all the Social Security benefits reported to have been received in our CPS sample yields an estimate of $265 billion in Social Security costs. Actual costs were $321 billion. The difference in these sums is due to the inclusion of administrative costs in the budget number, the exclusion of nonresident and institutionalized populations in the CPS, sampling variation in the CPS, and misreporting of benefits received. To account for these discrepancies, we simply adjust all responses upward by the same percentage so as to yield the aggregate budget numbers. For example, all interviewer responses on the amount of Social Security benefits received are multiplied by 1.2 (=321/265). This will create a bias in our estimates of aggregate benefits received by immigrants, their immediate descendants, and the general population to the extent that misreporting differs among these groups.6 The second type of information is response to interview questions about program participation. For example, the CPS tells us whether individuals reported that they participated in the Medicare program. We assume that all program participants generate the same amount in costs, so we equally divide the total program costs among all program participants. This will distort our estimates to the extent that misreporting and/or program usage per participant differ between the general population and immigrants and their descendants. The third type of information is for items such as federal income tax and property taxes, which are constructed by CPS based on reported income and other household characteristics. Although the majority of variables were drawn directly from the CPS, we needed to construct several ourselves. The federal tax incidence of corporate and business taxes is assumed to be shifted to individuals who own the business capital. Therefore, we assign these taxes to individuals based on their income from dividends and interest as reported in the CPS. At the state and local levels, we assume that the mobility of businesses means that income, sales, and property taxes paid by businesses are payments for business services provided by the state and local governments. If states or localities attempted to collect more in taxes than the service they provided, businesses would simply move to a new locality. The tax incidences of payroll taxes and sales taxes are assumed to be shifted to the individuals who receive the wage or purchase the goods. We allocated all the payroll taxes (FICA) paid by both employee and employer to the employee. Sales taxes were estimated based on household income and immigrant status (immigrant households are assumed to remit $1,000 in income and hence pay lower sales taxes than nonimmigrant households) using an algorithm estimated 6 In the case of Social Security benefits, we observed a higher frequency of ''zero" responses among immigrants and lower reported benefits among those who did respond. However, these responses are consistent with the shorter work histories of recently arrived immigrants and lower incomes of immigrants. We do not interpret this as a sign of higher misreporting among immigrants.
OCR for page 191
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration by Sheffrin and Dresch (1995). Owners of rental units are assumed to shift the majority of the property tax burden to their renters. Property taxes paid by renters were estimated by assuming that 70 percent of the tax on rental properties was borne by renters and this aggregate amount was equally divided among all renters. (For a full discussion of the issue of tax incidence, see National Research Council, 1997:Chap. 6.) Medicaid costs and incarceration costs were assigned based on institutionalization rates by age and immigrant status, taken from the 1990 Census. Medicaid institutionalized costs (for nursing home care) were divided equally among the institutional population over age 65 (a smaller portion of the elderly immigrant population than of natives was found to be institutionalized, however). Incarceration costs were divided equally among the institutional population below age 65. Refugee costs were divided equally among all immigrants. In assessing the educational costs of immigrant and second-generation children, we took into account the state-to-state variations in per student costs and the distribution of immigrant children across states. We also took into account the special costs of bilingual (limited English proficiency; LEP) education. These special costs were based on a study of Florida schools that showed that LEP students were 1.44 times more costly than other students. Immigrant children (51%) and, to a lesser extent, second-generation children (37%) are likely to be classified as LEP and hence are significantly more costly than other students.7 We estimate bilingual education costs to all states and localities to be $6.559 billion. These estimates suffer from several limitations. First, they are based on a Florida estimate of bilingual costs that may not be representative of the experience of immigrants in other states. The Florida courts have intervened in school financing on behalf of bilingual students. Second, this method ignores the possibility that immigrant children live in poorer (or richer) school districts and thus generate higher (or lower) than average costs to the federal and state governments and lower (or higher) than average costs to local school districts. Third, the method assumes that the same proportions of first-generation school children are LEP across states, and similarly for second-generation children. 7 Clark (1994:Table A.5 "State Costs of Bilingual Education") reports the extra cost of bilingual education in Florida in 1990–1991 as $2,334 per bilingual student. This is 44 percent of average student costs in Florida ($5,276). Furthermore, she estimates 51 percent of foreign-born children are LEP. Weighting by the probability of having LEP foreign-born children of immigrants is estimated to be 1.22 times as expensive as other children in their state (0.51*1.44 + 0.49*1.00). Clark estimates 3.68 percent of U.S.-born children are LEP. We assume that all U.S.-born children defined as LEP are children of immigrants. Our CPS sample shows one in ten U.S.-born children are children of immigrants. Therefore, the proportion of LEP students among the U.S.-born children of immigrants is estimated to be 36.8 percent. U.S.-born children of immigrants are estimated as being 1.16 times as expensive as other students in their state (0.368*1.44 + 0.632*1.00). We apply these ratios to each state, allowing state-to-state differences in educational spending to influence immigrant costs. That is, we assume that the relative costs of educating an immigrant child do not vary between the states, but the absolute costs do.
OCR for page 192
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration Other Considerations This is at best a partial equilibrium analysis. We do not try to calculate all the indirect fiscal consequences arising from the ways in which immigrants affect the economy in general. For example, immigration might affect housing prices, wages, unemployment rates, prices of other goods and services, profits on capital, rents on land, the geographic location of the nonimmigrant population, international trade flows, and so on (National Research Council, 1997). It is our view that it would be hazardous to single out one or two of these indirect effects for attention while ignoring others. In particular, some studies have estimated an indirect fiscal impact arising from the displacement of other workers, who then may claim unemployment insurance, swell the welfare rolls, and pay lower taxes than otherwise. The National Research Council report (1997) finds that immigration probably raised or had little effect on the wages of most workers, but that it may have considerably depressed the earnings of the least educated workers such as high school dropouts. Clearly these are important issues. It is important to keep in mind, however, that immigration also probably raised the earnings of nonlabor factors. On net, the National Research Council (1997) report concluded that the aggregate income of nonimmigrants was probably raised by $1 billion to $10 billion. Such a net increase in income would have engendered a corresponding increase in taxes at all levels of government, in addition to the tax payments by immigrants that we have explicitly included. Whether immigration may also have raised various types of welfare payments significantly by worsening the economic situation of less educated nonimmigrant workers, and whether any such effect would outweigh the positive effect on tax payments, are questions for further research. Because of the presence of public goods, public debt, and current deficit spending, there can be substantial differences between total tax revenues and total costs of nonpublic goods. Public goods and public debt make fiscal impacts more positive or less negative, whereas current deficit spending tends to make fiscal impacts less positive or more negative. Calculations for FY 1994 show that federal tax revenues exceeded costs for nonpublic goods by $345 billion, whereas state and local taxes exceeded total costs by $17 billion. Therefore, our calculations of net fiscal impacts presented in the next section sum to these positive values, not zero. Further details on the treatment of federal, state, and local budgets are described in the Appendix. RESULTS Examining the Fiscal Impacts Among Governments In Tables 5-2, 5-3, and 5-4 we examine fiscal impacts under the "immigrant and concurrent descendants" approach. Results for all state and local
OCR for page 193
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration governments combined are presented in Table 5-2. The presence of immigrants and their concurrent descendants generated $89 billion in costs to states and localities across the United States. This group paid an estimated TABLE 5-2 Fiscal Impact on State and Local Governments (1994 in 1994 dollars) Aggregate (in $ billions) Per capita ($) Immigrants & Concurrent Descendants All Others Immigrants & Concurrent Descendants All Others Ratio Diff Taxes Income Tax 15.7 113.1 387 513 0.75 -126 Property Tax, Homeowners 10.4 76.1 257 345 0.74 -88 Property Tax, Renters 5.2 23.6 130 107 1.21 23 Sales Tax 17.6 127.8 435 579 0.75 -144 Unemployment & Workers Comp. Contributions 4.3 27.1 106 123 0.86 -17 Other Taxes 8.3 60.1 206 273 0.76 -67 Total Taxes 61.5 427.9 1,520 1,941 0.78 -421 Costs of Benefits Medicaid, Institutional 1.1 14.3 27 65 0.42 -38 Medicaid, Non-institutional + Other Medical Welfare 12.3 49.8 305 226 1.35 79 SSI 0.7 3.1 17 14 1.22 3 AFDC + Other Welfare 7.7 28.1 190 127 1.49 63 Food Stamps 0.3 1.5 8 7 1.19 1 Unemployment Compensation 3.3 18.3 81 83 0.97 -2 Worker's Compensation 1.2 9.3 30 42 0.71 -12 Bilingual Education 6.2 0.0 154 0 n.a. 154 Elementary and High School 42.2 191.6 1,043 869 1.20 174 Public College 7.5 29.1 186 132 1.41 54 Incarceration Costs 4.2 28.1 103 127 0.81 -24 Congestible Goods 2.1 11.4 52 52 1.01 0 Total Costs 88.8 384.5 2,197 1,744 1.26 453 Total: (Taxes - Costs) -27.4 43.4 -677 197 -3.44 -874
OCR for page 195
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration TABLE 5-3 Fiscal Impact on the Federal Government (1994 in 1994 dollars) Aggregate (in $ billions) Per capita (in dollars) Immigrants Others Immigrants Others Ratio Diff Taxes Income Tax 15.7 113.1 387 513 0.75 -126 Income Tax 65.5 477.6 1619 2166 0.75 -547 Corporate Tax 13.3 126.7 329 575 0.57 -246 Excise Tax 5.2 49.8 129 226 0.57 -97 FICA 62.7 398.8 1551 1809 0.86 -258 SMI contribution 1.0 10.0 24 45 0.53 -21 Other Taxes 5.7 41.4 140 188 0.75 -48 Total Taxes 153.3 1,104.3 3793 5008 0.76 -1215 Costs of Benefits OASDI 24.7 296.4 612 1344 0.46 -732 HI 9.8 102.6 243 465 0.52 -222 SMI 5.4 55.9 132 254 0.52 -122 Medicaid, Institutional 1.5 19.0 36 86 0.42 -50 Medicaid, Non-institutional 12.2 49.1 301 223 1.35 78 SSI 5.0 22.5 123 102 1.20 21 AFDC + Other Welfare 6.6 24.1 163 109 1.49 54 EITC Refund 2.4 8.5 60 39 1.56 21 School Lunch 1.7 5.8 41 27 1.55 14 Food Stamps 5.4 25.9 134 117 1.14 17 Energy Assistance 0.2 1.6 4 7 0.56 -3 Rent Subsidy 2.2 12.2 55 55 0.99 0 Public Housing 0.8 5.4 19 25 0.77 -6 Unemployment Compensation 1.1 6.1 27 28 0.98 -1 Refugee Aid 0.4 0.0 9 0 n.a. 9 Bilingual Education 0.2 0.0 5 0 n.a. 5 Elementary and High School 2.4 10.7 58 49 1.19 9 Public College 0.3 1.2 8 5 1.49 3 Federal Student Aid 1.6 8.0 39 36 1.07 3 Incarceration costs 0.3 2.0 7 9 0.77 -2 Federal Retirement 1.3 35.6 32 162 0.20 -130 Military Retirement 1.2 25.6 29 116 0.25 -87 Railroad Retirement 0.2 4.4 5 20 0.25 -15 Congestible Goods 15.8 86.1 391 391 1.00 0 Total Costs 102.5 808.7 2535 3668 0.69 -1133 OASDHI: Taxes - Costs 28.2 -0.2 696 -1 n.a. 697 All Other: Taxes - Costs 22.7 295.7 562 1341 n.a. -779 Total: Taxes - Costs 50.9 295.5 1258 1340 n.a. -82
OCR for page 196
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration TABLE 5-4 Per Capita Fiscal Impact in High- and Low-Immigration States (Immigrants and Concurrent Descendants) A. Change in per capita tax burden All taxes Federal State and Local The Nation $107 $231 -$124 High immigration states -$49 $231 -$280 All other states $182 $231 -$49 B. Population (millions) Total Immigrants and Concurrent Descendants All Others The Nation 260.9 40.4 220.5 High immigration states 101.6 29.7 71.9 All other states 159.3 10.7 148.6 NOTE: Per capita is calculated by dividing aggregate totals from Tables 5-2 and 5-3 by the residual population figures given in Panel B of the table. (This table gives fiscal impacts in 1994 measured in 1994 dollars.) fornia, Texas, New York, Florida, Illinois, and New Jersey) and those in the rest of the nation. The calculations assume that immigrants and their concurrent descendants live in the same states. If immigrants and their concurrent descendants were to have vanished in 1994, the average member of the remaining population would have had to pay $107 more in taxes, or suffer a comparable reduction in benefits. This figure can be translated into the tax increase per residual household by multiplying by the average household size of 2.6 members, to arrive at $278. The population figures in Table 5-4, Panel B, can be used to re-express these figures relative to the total national population instead of the residual population, if so desired. However, this national average masks important interstate differences. Per capita taxes in high-immigration states would have been $49 lower had there been no immigrants or descendants, whereas per capita taxes in low-immigration states would have been $182 higher. Within the group of high-immigrant states, presumably some states had higher burdens and others had lower burdens because of interstate differences in the characteristics of the immigrants and state fiscal policies. The same methods employed here could be used to make estimates of the net costs of immigration for individual states. But these calculations would be correct only under the strong assumption that immigrants resided in the same state as their concurrent descendants. The high rate of geographic mobility in the United States undermines any state-specific methods that include descendants of immigrants who have left their parental home. Taxpayers in high-immigration states bear the full increased burden of providing state and local government services to immigrants who reside in their
OCR for page 197
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration state, whereas they share the federal fiscal benefits of immigration with taxpayers throughout the nation. This substantial discrepancy arises in large part because states and localities fund education and other youth services, whereas the federal government funds Social Security, Medicare, and other services for the elderly. (See Goldstein, 1995, for a discussion of interstate transfers arising from differences in age distribution and the funding of education and social security.) It is sometimes suggested that the federal government should address these interstate discrepancies through a national policy of compensation to state and local governments. According to these calculations, the federal government realized a net fiscal gain of about $51 billion in 1994 from the presence of immigrants and their concurrent descendants. At the same time, there were net fiscal costs of $27 billion per year incurred in total by a subset of state and local governments. The federal government could compensate states and localities for the $27 billion net cost, leaving the remaining $24 billion of federal net gain as a benefit to be shared equally among all U.S. taxpayers. 9 There are many complicated issues regarding such a policy of federal compensation. The presence of immigrants and their descendants may confer nonfiscal economic net gains to the residents of states and local areas, for example through cleaper goods and services or through higher wages of skilled workers and higher returns to nonlabor factors. The federal government does not generally compensate states for other discrepancies; what is special about the case of immigration? We do not consider these and other issues here. Comparison of Results for Immigrants Only, Immigrant Households, and Concurrent Descendants Table 5-5 shows comparable measures of fiscal impact for the different demographic formulations of the cross-sectional conceptual experiment. In the most restrictive definition, only immigrants themselves are counted. Here we find that immigrants pay about $32 billion more in taxes overall than they generate in costs. This positive balance reflects the age distribution of the immigrant population: There are relatively many working-age people and relatively few children and elderly. 9 Two other levels of compensation are possible, but neither is immigrant neutral. In one alternative, taxpayers in high-immigration states could be compensated so that they are not bearing any net increased fiscal burden of immigration. That is, state and local costs would just offset their reduction in national taxes ($4 billion annually to high-immigration states). In this scenario, states would compete to discourage immigration. Under another alternative, taxpayers would receive a share of the national benefits of immigration ($51 billion) in direct proportion to the number of immigrants in their state. Under this scenario, states would compete for immigrants. States are indifferent to immigrants only in the basic compensation scenario in which states are compensated for the full state and local costs of immigrants, and the remaining national benefits are shared equally among all states.
OCR for page 198
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration TABLE 5-5 How the Aggregate Fiscal Impact Depends on the Definition of the Study Population (1994 in 1994 $ billions) A. Aggregate Fiscal Impact Study Population: Overall Federal State and Local 1. Immigrants Only 32.4 28.2 4.2 2. Immigrant Households -13.3 16.0 -29.3 3. Immigrants and Concurrent Children 29.5 48.9 -19.3 4. Immigrants and Concurrent Descendants (Children and Grandchildren 23.5 v v B. Population Subtotals Study Population: Number Cumulative Total 1. First Generation 22,766,711 22,766,711 2. Second Generation under age 20 8,201,368 30,968,079 3. Concurrent Second Generation age 20 and over 5,597,759 36,565,838 4. Concurrent Third Generation 3,862,610 40,428,448 B. Population Subtotals Study Population: Number Cumulative Total Number Cumulative Total 1. First Generation 22,766,711 22,766,711 2. Second Generation under age 20 8,201,368 30,968,079 3. Concurrent Second Generation age 20 and over 5,597,759 36,565,838 4. Concurrent Third Generation 3,862,610 40,428,448 Immigrant households include nearly all immigrants, plus nearly all U.S.-born children of immigrants up to the age of 20 or so, because such children will co-reside with their immigrant parents. Expanding the definition to include immigrants plus their U.S.-born children under the age of 20, the estimated fiscal impact flips from $32 billion to -$13 billion. The biggest change is at the state and local levels, where the impact shifts from $4 billion to -$29 billion. Evidently it makes a decisive difference which of the two most common demographic formulations of the problem is used. Measured impacts are strongly positive for immigrants only and strongly negative overall for immigrant households. The bias from excluding the adult U.S.-born children of immigrants, together with their children, becomes apparent when we recompute and take them into account. When we count all the relevant descendants of still-living immigrants we find a net positive fiscal impact of about $24 billion. There is a very large positive fiscal impact of $51 billion at the federal level, partially offset by a large negative fiscal impact of -$27 billion at the state and local levels. Looking separately at the federal, state, and local components, we find that at the state and
OCR for page 199
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration local levels, the results including all descendants are little different from the immigrant household results: -$27 billion versus -$29 billion. The impacts of the two missing groups (adult children of immigrants and grandchildren of immigrants) offset each other. This is not true at the federal level, where the discrepancy is severe. Including all concurrent descendants of immigrants triples the federal surplus! Comparison with a Longitudinal Study In a separate study, we analyzed the fiscal impact of immigrants in a forward-looking longitudinal model. For each age at arrival and education level, we calculated the impact of the immigrant and all future descendants by projecting their survival, fertility, return migration, educational attainment, taxes, and benefits (see Lee and Miller, 1997; National Research Council, 1997:Chap. 7). In that analysis, we also calculated the average impact of immigrants, weighting by the age and educational composition of recently arrived immigrants. The average present value (over a 300-year horizon and expressed in 1996 dollars) of the stream of net fiscal impacts of an immigrant arriving in 1994 to state and local governments combined is -$25,000, to the federal government $105,000, and overall is $80,000 per immigrant admitted. For the purpose of comparison, we can multiply these present values by the interest rate (3%) to convert to annual flows: -$750 at the state and local levels, $3,150 at the federal level, and a combined annualized fiscal impact of $2,400 per immigrant. The concurrent descendant cross-sectional approach yielded estimates of -$1,200 at the state and local levels, $2,230 at the federal level, and a combined annual fiscal impact of $1,030 per immigrant (in 1994 dollars). There is no reason to expect a close agreement for these two very different methods, except perhaps under certain hypothetical steady-state conditions. The cross-sectional method is affected by the historical trends in the volume and composition of immigration streams and in vital rates, because these shape the current age and educational distribution of immigrants and descendants. These do not enter into the longitudinal calculation. At the same time, the cross-sectional method does not take into account projected trends in taxes, benefits, educational attainment, and vital rates, unlike the longitudinal method. The longitudinal calculation refers to an immigrant arriving in 1994, whereas the cross-sectional calculation is based on immigrants arriving at various times in the twentieth century. CONCLUSIONS We have argued that the longitudinal approach is the preferred method of assessing the fiscal impact of immigrants for most purposes. Yet because this method is forward looking, it must inevitably rely on projections far into the
OCR for page 200
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration future and make many assumptions. Some people, therefore, may prefer the down-to-earth cross-sectional estimates of the fiscal impact of immigrants in a particular year, which appear to be more straightforward and to involve fewer assumptions. Indeed, almost all published estimates of fiscal impact are of this cross-sectional type. Many of these are limited in scope because they include only a subset of tax payments and benefits received. The estimates in this chapter reflect a broad coverage, including 28 types of transfers (assignable benefits), both in cash and in-kind; public goods; servicing the government debt; and the costs of congestible goods and services. All benefits and taxes included in government budgets are counted in one way or another in our calculations. Our main point in this chapter, however, is to investigate the effects of formulating the conceptual experiment in terms of differing subpopulations. All the analyses reviewed by Rothman and Espenshade (1992) or Vernez and McCarthy (1996), and the more recent studies of which we are aware, take either the immigrant-only approach or the immigrant household approach. The first is flawed for most purposes because it ignores all U.S.-born descendants of the immigrants and thus underestimates the costs caused by the immigrants. The second approach is flawed because it includes the U.S.-born descendants only while they are young, costly, and reside in their immigrant parents' homes. In this way, the U.S.-born children of immigrants are counted only during the ages in which society is investing heavily in their education, not when they in turn grow up to become taxpayers themselves. This tends to make the estimated fiscal impact be more negative, or less positive. Our calculations indicate that definition of the study population is critical to the outcome. If limited to immigrants themselves, the overall fiscal impact is $1,400 (taxes paid less costs generated) per immigrant. If limited to immigrants plus their U.S.-born children under the age of 20, corresponding to the immigrant household formulation, the average fiscal impact is about -$600 per immigrant (or -$400 per immigrant and young child). If extended to all descendants of living immigrants, the average fiscal impact is $1,000 expressed per immigrant, or $600 expressed per immigrant and descendants.10 Therefore, the most widely used method based on the immigrant household is the only one that returns a negative value. We argue that if the U.S.-born children are to be included in the calculation (and we believe they should be), then they should be included at all ages and not just while they represent heavy costs to society. Therefore, we believe that the calculation inclusive of all concurrent descendants is most appropriate within the category of cross-sectional calculations. However, all cross-sectional calculations give the wrong answer. The longitudinal calculation remains the method of choice. In the longitudinal formulation, the appropriate demographic specifica- 10 These per capita numbers are calculated from the numbers given in Table 5-5.
OCR for page 201
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration tion follows closely from the question posed. An incremental immigrant arrives, and a long chain of consequences follows, continuing after the immigrant's death through all of the descendants. In the cross-sectional approach, the appropriate demographic formulation has been less clear. We hope to have shown, however, that whatever the general weaknesses of the cross-sectional approach, the failure to include the effects of all concurrent descendants biases the outcome. In practice, the resulting distortions are very large and cannot be ignored. APPENDIX: DETAILED ASSUMPTIONS FOR FEDERAL, STATE, AND LOCAL CALCULATIONS Why Do Taxes Exceed Costs? By definition, federal, state, and local budgets must meet the basic accounting identity: total revenue equals total expenditures. Therefore, one might expect that taxes could not exceed costs in our calculations. However, for purposes of examining the fiscal impact of an incremental member of the population, we exclude from consideration (a) expenditures on public goods, payments for debt interest or principal, and purchase of financial assets; and (b) revenues from interest on assets, the sale of assets, and new debt issued. We assume that adding or subtracting an incremental member of the population will not have any impact on these types of expenditures and revenues in the current period. It is certainly true that immigration will affect public assets and public debt in the long run. This impact is considered in longitudinal studies, but cannot be incorporated into cross-sectional studies. At the federal level, we excluded (a) $346 billion in expenditures on public goods, (b) $203 billion in expenditures for debt repayment, and (c) $203 billion in revenue from net new borrowing. Therefore, federal tax revenues exceed federal expenditures by $345 billion. The fact that debt repayment equaled net new borrowing in FY 1994 is purely coincidental. For example, in FY 1996 federal debt repayment exceeded net new borrowing by $96 billion. The larger debt payments and smaller deficit in FY 1996 would make the federal fiscal impact of immigrants significantly more positive in 1996 relative to 1994. At the state and local levels, we excluded (a) $65 billion in expenditures for debt interest, (b) $167 billion in expenditures for debt repaid, (c) an estimated $46 billion in expenditures for financial assets or funds otherwise not accounted for, (d) $49 billion in revenue from interest earnings and the sale of assets, and (e) $211 billion in new debt issued. Therefore, state and local tax revenues exceed expenditures by $17 billion. This represents 1.2 percent of total revenues and
OCR for page 202
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration could be considered as a measurement error due to difficulties in estimation as a result of the statistical nature of the data and misreporting. Notes on the Federal Budget Federal budget expenditure program totals for FY 1994 are taken from Budget of the United States Government, FY 1996, Analytical Perspectives (Executive Office of the President, 1996). The following lists the programs and budget item codes taken from this source: OASDI , HI , SMI , SSI , AFDC + other welfare [609 and 506], earned income tax credit [item code 609], school lunch , food stamps , energy assistance , rent subsidy and public housing , unemployment compensation , refugee aid , K-12 expenditures , public college , direct student aid , federal retirement , military retirement , and railroad retirement . Federal public assistance is broadly defined to include (a) $4.7 billion for child and family services programs [code 506]; (b) $3.0 billion for payments to states for foster care and adoption assistance [code 506]; (c) $3.8 billion for social services block grants [code 506]; (d) $2.3 billion for rehabilitation services [code 506]; and (e) $16.8 billion for family support payments [code 609]. Federal public goods totaling $346 billion are defined to include (a) $282 billion for national defense [codes 051–054]; (b) $16 billion for general science, space, and technology [codes 251-232]; (c) $11 billion for health research and training [code 552]; and (d) $38 billion for veterans benefits and services [codes 701–705]. The reported outlay of $20.6 billion for public housing and rental subsidies [code 604] is allocated to public housing (30%) and rental subsidies (70%) according to the proportions observed in Table 577, 1996 US Statistical Abstract of the United States: 1996 (Bureau of the Census, 1996). Program expenditures for federal incarceration costs were taken from Table 517, Statistical Abstract of the United States: 1996 (Bureau of the Census, 1996). Program expenditures for the federal and state Medicaid programs were taken from Medicaid National Summary Statistics (Table 1). About 25 percent of Medicaid expenditures are for the institutionalized population (Table 170, Statistical Abstract of the United States: 1996. Bureau of the Census, 1996). Federal congestible goods were defined as a residual. Starting with total federal outlays of $1,461 billion, we subtracted (a) $203 billion in interest payments, (b) $346 billion in public goods, and (c) $809 billion in transfers (assignable costs). [Table 515, Statistical Abstract of the United States: 1996 (Bureau of the Census, 1996)]. Federal tax receipts consisted of (a) FICA payroll tax, (b) income taxes, (c) corporate taxes, (d) excise taxes, (e) supplementary medical insurance (Medi-
OCR for page 203
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration care, Part B) contributions, and (f) other taxes. The CPS includes estimates of FICA payroll tax and federal income tax. Corporate taxes and excise taxes were assigned to individuals according to their earnings from dividends and interest as reported in the CPS. Supplementary Medical Insurance (SMI) contributions were assigned to individuals based on their participation in Medicare. Other taxes were assigned to individuals according to their federal income tax payments. Notes on State and Local Budgets We assume that revenues from corporations (income tax and estimated property and sales taxes) and revenues from charges/fees/special assessments are exactly matched by state and local government expenditures on congestibles and in-kind transfers. These revenues and the congestible goods and in-kind transfers they purchase are removed from consideration. Nationally, corporations were assumed to pay 41.5 percent of total property taxes based on estimates from California (California Board of Equalization, 1995). Corporations were assumed to pay 35 percent of sales taxes based on estimates from California (Sheffrin and Dresch, 1995). Transfers and congestibles purchased with federal tax dollars are also removed from the analysis of state and local budgets to avoid double counting, because we counted these expenditures as part of the federal budget. We do not consider revenue nor expenditures for public employee retirement trusts. State and local budget expenditure totals for FY 1993–1994 are taken from the Annual Survey of Government Finances, 1993–94 collected by the Bureau of the Census (1997). The following lists the programs and their budget item codes taken from this source: other medical welfare [EFG32–36, 74 less Medicaid expenditures]; other welfare [EFG67–79 less Medicaid expenditures]; K-12 education [EFG9–13]; public college [EFG16–21]; incarceration costs [EFG4–5]; unemployment compensation [Y05–06]; worker's compensation [Y14–15]; income taxes [T40]; property taxes [T01]; general and selective sales taxes [T09–19]; unemployment compensation contributions [Y01]; worker's compensation contributions [Y14–15]; and other taxes that include fines and forfeits [U30], rents [U40], royalties [U41], donations [U50], lottery [U95], miscellaneous general revenues [U99], death and gift taxes [T50], documentary and stock transfer taxes [T51], severance [T53], and miscellaneous taxes [T99]. Total expenditure data for Medicaid, SSI, AFDC, general assistance, and food stamps were taken from Table 577, Statistical Abstract of the United States: 1996 (Bureau of the Census, 1996). Of the total property taxes collected from individuals, we assumed that 64 percent was collected from homes and 36 percent from rental properties based on California data (California State Board of Equalization, 1995). Furthermore, we
OCR for page 204
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration assumed that 70 percent of the tax on rental properties was borne by renters, so that renters paid an estimated 25 percent of total property taxes collected from individuals. The CPS had estimated taxes for income tax, property taxes, and payroll taxes. We used these estimates to assign the aggregate totals for state income tax, property tax paid by homeowners, and contributions for worker's compensation and unemployment insurance. We needed to estimate taxes for property tax paid by renters, sales tax, and other taxes. We assigned property tax paid by renters by equal shares among all renters. We assigned sales tax based on household income and immigrant status using an algorithm derived by Sheffrin and Dresch (1995). We assigned other taxes in proportion to state income tax paid. ACKNOWLEDGMENTS We are grateful to the members and staff of the Panel on Demographic and Economic Impacts of Immigration, Michael Clune, James Smith, and an anonymous referee for helpful comments. Research for this chapter was funded by a grant from the National Institute on Aging, AG11761. REFERENCES Akbari, Ather H. 1991. ''The Public Finance Impact of Immigrant Population on Host Nations: Some Canadian Evidence." Social Science Quarterly 72(2):334–346. Bureau of the Census 1997. Annual Survey of Government Finances, 1993–94. The 1994 Finance Estimate Detail Data. Ascii file: fin94est.txt. Bureau of the Census 1996. Statistical Abstract of the United States: 1996. Washington, D.C.: U.S. Government Printing Office. California State Board of Equalization 1995. Annual Report. Sacramento. Clark, Rebecca 1994. "The Costs of Providing Public Assistance and Education to Immigrants." PRIP-UI-34 Program for Research on Immigration Policy, The Urban Institute, Washington, D.C. Executive Office of the President 1996. Budget of the United States Government. FY1996. Analytical Perspectives. Washington, D.C.: U.S. Government Printing Office. Goldstein, Joshua 1995. "From California to Florida: Interstate Wealth Flows in a Pay-as-you-go Social Security System." Paper presented at the 1995 annual meeting of the Population Association of America, San Francisco. Health Care Financing Administration 1997. "Medicaid National Summary Statistics." Internet site, http://www.hcfa.gov/medicaid/195.htm (accessed 11/17/97).
OCR for page 205
The Immigration Debate: Studies on the Economic, Demographic, and Fiscal Effects of Immigration Lee, Ronald, and Timothy Miller 1997. "The Future Fiscal Impacts of Current Immigrants." Working paper of the project on Intergenerational Transfers, Department of Demography, University of California, Berkeley. National Research Council 1997. The New Americans: Economic, Demographic, and Fiscal Effects of Immigration. Panel on the Demographic and Economic Impacts of Immigration. James P. Smith and Barry Edmonston, eds. Washington, D.C.: National Academy Press. Rothman, Eric S., and Thomas J. Espenshade 1992. "Fiscal Impacts of Immigration to the United States." Population Index 58(3):381–415. Sheffrin, Steven M., and Marla Dresch 1995. Estimating the Tax Burden in California. Berkeley: California Policy Seminar. Social Security Administration, Office of the Actuary 1992. Life Tables for the United States Social Security Area, 1900–2080. Actuarial Study No. 107. (August). Swallen, Karen Caperton 1996. "Morbidity, Mortality, and Nativity: The Influence of Early-age Effects and Health Selection on Health at Old Age." Ph.D. dissertation, Department of Demography, University of California, Berkeley. Vernez, Georges, and Kevin F. McCarthy 1996. The Costs of Immigration to Taxpayers: Analytical and Policy Issues. Santa Monica, Calif.: The RAND Corporation.
Representative terms from entire chapter: