EXPENDITURES ON IMMIGRANTS, PAST AND PRESENTJulian L. Simon In countries receiving large numbers of international immi- grants, and most notably in the United States, government expend- itures on immigrants, the taxes paid by immigrants, and the net balance between them, constitute a subject of continuing academic and public interest. This subject is difficult conceptually; it is also a subject for which appropriate data are scarce. This note will a) review the tax-and-transfer situation in the U. S. as of the mid-1970s, and clarify a matter that has led to some misunderstandings; b) adduce data from similar recent studies for Canada, showing that the same pattern found in the U. S. also holds there despite the observations being related to a different country and to later times; c) mention recent studies for Europe that show a similar pattern; and d) present a crude, but nevertheless indubitable, calculation from recent aggregate data for the U. S. showing that the pattern of expenditures observed in 1975 is even more pronounced now. THE U. S. EXPERIENCE AS OF 1975 Table 1 shows the situation as of 1975 based on the Census Bureau's Survey of Income and Education (SIE), still the only satisfactory source of such data for the United States. The data for various entry cohorts make possible the appropriate present- value calculation on the assumption that the cohorts are suffi- ciently similar to each other to consider them as sequential observations relating to the same life cycle. [Table 1 about here] Because the SIE does not include an accounting of government expenditures on public goods, both immigrants and natives pay more in taxes than the amount of direct government expenditures on them as individuals or families. One cannot, therefore, simply compare the immigrants' taxes to their benefits; instead, the appropriate procedure is to compare the results relating to immigrants to those relating to natives. Because one can assume that for natives as a group the cost of total benefits received equals total taxes, the extent to which the net balance of such inflows and outgos for immigrants exceeds that for natives meas- ures the contribution of natives to the public coffers<1>. The table shows (see columns 15, 16, and 6) that this balance is positive for all cohorts except for the groups that arrived before 1949. Table 1 summarizes the data presented in my report to the Select Commission on Immigration and Refugee Policy in 1981 and in my 1989 book, leaving out detailed data on particular services and taxes (Simon 1981 and 1989). But in my 1984 article discuss- ing this subject (Simon 1994) the data for the oldest immigrant group were not shown, partly to save space and partly because those data are not strictly relevant to the issue at hand. This caused some misinterpretation of my argument<2>, and hence requires further explanation. There are two reasons why the data for the older immigrant groups are not relevant. First, applying any reasonable discount factor, economic events occurring 25 years and more in the future have little effect on current decisions based on present-value calculations. But the decisive factor is that, on average, the oldest persons by then have children who are paying amounts into the system as taxes that balance what their parents withdraw as transfer payments; the immigrants' excess payments represent a one-time windfall to the equilibrium system. The only conceptual analytic alternative would be to include in the calculation the costs and benefits of the children when grown (and their chil- dren, and so on) on the ground that these streams of people are part of the consequences of the immigrants. But as a practical matter this is not feasible. (For an extended explanation, see Simon, 1989: 109-110). So I conclude that my interpretation pertaining to the balance of expenditures on U. S. immigrants in 1975 (Simon, 1984) is valid, rather than flawed due to omission of data on earlier immigrant groups. Immigrants put considerably more money into the public coffers than is taken from the coffers to cover ex- penses on them. THE CANADIAN EXPERIENCE Akbari (1989; 1995) conducted two similar studies for Cana- da, one using 1981 Census data and the other using 1991 data from the Survey of Consumer Finances for 1991. With respect to gov- ernment expenditures on immigrants and natives, the results of his analysis of 1981 data are entirely consistent with the re- sults of my study for the U. S. using 1970s data; in turn, his study of 1991 data shows no major differences with the study of the 1981 data. Thus the U.S. and Canadian studies, for two different countries covering different periods, corroborate each other's conclusions. DATA FROM EUROPE In Europe, most countries provide a variety of benefits to families and individuals not provided in North America. As a typical example, Germany gives parents a monthly allowance of $50 or more for the first child, and $100 each for the second and subsequent children. There are also grants for maternity leaves, income tax rebates for families with children, and various pay- ments for health care. Akbari and I analyze the effects of such benefits on the economic impact of immigration using data from the 1989 German microcensus (Simon and Akbari, 1995). We find that each cohort of immigrants pays more in taxes than the cost of the direct services that they receive, just as do natives; the amounts of those positive balances compare favorably with those of natives. We also find, however, that due to having lesser amounts of education, the most recent immigrant cohorts earn considerably less than do older cohorts, and that in these cohorts the excess of taxes paid over services received is less than with earlier cohorts. On balance, the data do not support the conclusion that immigrants pay more taxes than the costs of the services they receive, but neither do they support the opposite conclusion. Other, admittedly rather perfunctory, studies cited in that publication also do not find clear-cut conclusions on the econom- ic effects of immigration in European countries. RECENT U. S. DATA ON EXPENDITURES ON IMMIGRANTS The appropriate mode of analysis of the effect of immigrants on the public purse is a present-value assessment based on the lifetime experience of the relevant cohorts of immigrants. Such analysis examines each of the past cohorts as a separate observa- tion. If the present-value analysis is to be meaningful, the immigrant cohorts must be similar from period to period relative to natives; this requirement was met reasonably well in the United States data through the 1970s. For the 1980s and 1990s, however, there are no data suitable for life-cycle analysis, not even household data on immigrants' incomes by cohorts, on which tax estimates could be based. Researchers have, however, attempted to make do with the existing statistical information, mostly considering data on the various entry cohorts lumped together. This section discusses what may be learned from this body of work. Welfare Expenditures From administrative records, Rebecca Clark (1994) calculated 1990s government welfare expenditures for immigrants' and na- tives' use, narrowly defined as consisting, of a) Aid to Families with Dependent Children (AFDC); b) food stamps; c) Supplemental Security Income (SSI); and d) General Assistance. She finds that foreign-born persons taken as a group have perhaps a 10-20 per- cent higher probability of being recipients of these goods and services than do natives. From Clark's data I estimate that the combined total of these federal expenditures is $404 per year per immigrant, while the average native receives $260. These data on welfare expenditures in the early 1990s are shown at the bottom portion of the bars comparing government expenditures on immi- grants and natives in Figure 1. [Figure 1 about here] Government Expenditures Other Than Welfare The four federal "welfare" programs discussed above do not include most government payments to the elderly, or local expend- itures for public schooling. To assess total government expendi- tures on various cohorts of immigrants and natives such outlays need to be included. Social Security and Medicare payments, by far the largest government transfer programs, accrue mainly to natives. This is because immigrants typically arrive when they are young and healthy, and also because older recent immigrants do not qualify for Social Security. Payments to immigrants under the Social Security and Medi- care programs are particularly difficult to estimate because such payments differ greatly by age groups. And the sizes of the various age groups of foreign-born residents differ greatly because of the attrition of the numbers of older immigrants due to mortality and the increasing rates of immigration in recent years (a difficulty avoided in the cohort analysis of the 1975 SIE data). Nevertheless, I shall make rough estimates for immi- grants who entered the United States since about 1970. Total federal expenditures of $305 billion in 1993 (estimat- ed) for Social Security and $133 billion for Medicare (arbitrari- ly reduced by 2 percent for payments to immigrants), divided by the total population, indicate expenditures per native of $1305 and $566 respectively (Statistical Abstract of the United States, 1993: Table 512, p. 330). The 1975 SIE data (discussed above) suggest that the average receipt per immigrant who arrived within the past 25 years was (and probably still is) less than a fifth of the average expenditure per native for these two programs - say $261 and $113 for 1992 respectively. (Some allowance for the public support of the aged immigrants is embodied in the rela- tively heavy SSI payments, accounted for in the preceding sec- tion, that substitute for Social Security.) As Figure 1 shows, government expenditures for these old-age programs dwarf govern- ment expenditures on welfare programs as narrowly defined. Government expenditures on public schools estimated by Clark are $522 per immigrant, and $922 per native. The expenditures are lower for the immigrant population because the proportion of children in that population is smaller than in the native popula- tion. (Separate data for recent cohorts of immigrants probably would show higher per capita expenditures than the average of $552 estimated for all immigrants.) These estimates suggest that the gap in dollar expenditure on schooling between natives and immigrants is roughly eight times the gap in welfare expendi- tures, but with a different sign: favoring natives rather than immigrants. *****note changes in next par. Maybe change fig. For unemployment compensation we can safely assume expendi- tures of roughly $138 per capita (Statistical Abstract, 1993: Table 585); based on experience with unemployment compensation in the 1970s, this amount is likely to be similar for immigrants and natives. The possibility of important error is small here. For Medicaid it is reasonable to assume [somewhat] higher expenditures for immigrants than for natives [[[[[[[xxxx in the same proportion as for the welfare programs discussed above; higher expenditures from these programs ]]]]]]]]] to reflect the fact that immigrants are somewhat poorer on average than natives. Federal and state Medicaid expenditures were about $90 billion and $70 billion respectively in the early 1990s, so expenditures per person may be estimated as about $627 for natives and $752 for immigrants. Total Government Expenditures The sum of all transfer payments, including schooling costs, provides the appropriate measure of government expenditures that should be used in any comprehensive assessment of the economic costs and benefits of immigration. Figure 1 shows that the expenditures on natives per capita are much higher - indeed, nearly twice as high - than the expenditures on immigrants per capita: $3800 versus $2200. For immigrants who arrived between 1970 and 1990, the total might be $100 higher than the estimate of $2200 for all immigrants taken together, owing to higher schooling costs. Expenditures on the elderly cause the overall estimates for natives to be almost double those for immigrants. Of course the estimates made here are rather unsatisfactory, in part because of the differing age composition of the immigrant population - more of whom came in recent years - but also because of weaknesses in the available data. Yet one can confidently draw two conclu- sions: 1) The slightly greater expenditures per immigrant in the narrowly-defined welfare programs are more than offset by other categories of government spending - indeed, are dwarfed by them; therefore the welfare programs, taken in isolation, do not de- serve attention in assessing the economic costs of immigration to the government. 2) Overall average expenditures for immigrants are not greater than for natives; rather, the opposite is the case. Additional Notes on the Estimates for the 1990s 1. Compared to 1975, the 1990s gap in expenditures between natives and immigrants is considerably wider in proportional as well as absolute dollar terms. 2. Though the reckonings for expenditures on the Social Security, Medicare, and Medicaid programs and on public schooling are very crude, deriving from aggregate rather than from micro level data, no conceivable inaccuracy in these reckonings could alter the overall conclusion. Assessing government expenditures on immigrants requires that some estimates be made for all the categories mentioned, because in absolute terms of spending these categories of expenditures are by far the most important, and hence the crudity cannot justify omitting them. 3. In considering the issues discussed here, it would be more appropriate to make assessments on a per-family rather than a per-person basis. But calculations based on family level data would surely parallel the individual level calculations. 4. It also would be better to make assessments for immi- grant cohorts defined by date of arrival. However, the main conclusions presented above surely hold especially strongly for the most recent cohorts of immigrants, because these conclusions flow from the fact that recent immigrants tend to be young and in the labor force, and hence do not receive the most expensive government transfers - Medicare and Social Security payments. Furthermore, it is recent immigrants that are the main subject of most public discussion; they also are the most relevant for policy decisions concerning immigration, in large part because the value of economic flows in the far future is greatly dimin- ished by proper discounting. Taxes Paid by Immigrants and Natives, and the Overall Balance An overall judgment on the effect of immigration on future expenditures should take account of the tax contributions of immigrants and natives as well as the expenditures on them. However, no recent study provides data on household incomes of natives and immigrants, and hence there is no solid basis for estimating the taxes paid by the two groups. As a proportion of the taxes paid by natives, the taxes paid by immigrants probably have fallen since the 1970s because, according to analyses by Borjas (1994, p. 1678), the relative earnings of immigrants have declined - the wage differential increasing by about 10.5 percent from the 1960s immigrants to the 1970s immigrants, and by about 7 percent from the 1970s immi- grants to the 1980s immigrants (arrived at by averaging over 5- year cohorts and pairs of decadal observations). But even allow- ing for progressivity in taxes, it seems doubtful that the im- plied total decline of 15 or 20 percent is great enough to bal- ance or outweigh the increase in the gap on the expenditures side that occurred since the 1976 SIE survey observations. A solid assessment must await further research; for now, there is no ground for revising the conclusion reached on the basis of ana- lyzing the data from the 1970s. RECENT TRENDS IN WELFARE USE BY IMMIGRANTS It was noted above that narrowly-defined welfare expendi- tures are a small part of total government expenditures on immi- grants. But because there has been public concern that welfare expenditures have been increasing, it is worth noting the evi- dence on recent trends that does exist. The direction of the effect depends upon which groups and programs one examines. Table 2, drawn from Passel and Fix (1994), shows that both native and foreign-born groups had bare- ly-perceptible declines from 1979 to 1989 in the proportion of the respective populations receiving welfare. Those immigrants who entered between 1980 and 1990 had a slightly lower rate of receipt than those who entered between 1970 and 1979, certainly giving no evidence of an increase (though it sometimes takes some time before immigrants learn to use the welfare system). [Table 2 about here] Among foreign-born persons aged 65 years and older, a great- er and growing proportion receive welfare than among natives. This is due to many immigrants having arrived too late to accumu- late enough work time to earn Social Security benefits; the SSI welfare payments, as was noted above, are a partial substitute for Social Security. CONCLUSIONS Based on good quality statistics relating to the 1970s, immigrants contributed more to the public coffers than they received in public services. When the data are displayed in full, as is done in summary form in Table 1, an issue that prompted criticism of this finding is clarified, and the conclu- sion set forth in Simon (1984) holds. This conclusion is corrob- orated by Canadian studies drawing on data from the 1980s and 1990s. Data for Europe on this topic present a mixed picture. Data for the recent period for the U. S. are crude. To the extent that one can draw meaningful conclusions from them, they support the conclusion drawn from the earlier, more substantive, data. They suggest that the immigrants' relative contribution may even have increased, though the ratio of taxes paid by the average immigrant relative to those paid by natives may have declined. If there has been any increase in welfare expenditures on immigrants relative to natives, it is probably confined to the narrowly-defined category of welfare payments, and probably only among older immigrants. page 1 article5 pdr1 February 14, 1996 REFERENCES Akbari, Ather H., 1995. "The Impact of Immigrants on Cana- da's Treasury, circa 1990," in Diminishing Returns, edited by Don J. DeVoretz, Policy Study 24, C. D. Howe Institute, The Laurier Institution. Akbari, Ather H., 1989. "The Benefits of Immigrants to Canada: Evidence on Tax and Public Services," Canadian Public Policy, 15(4): 424-435. Borjas, George , Review of The Economic Consequences of Immigration by Julian L. Simon, Journal of Economic Literature, Vol. XXIX, March, 1991, pp. 115-116. Borjas, George J., 1994. "The Economics of Immigration", Journal of Economic Literature, 32(4): 1167-1717. Clark, Rebecca L., 1994. "The Costs of Providing Public Assistance and Education To Immigrants," Washington, D.C.: The Urban Institute, (Revised August 1994). Fix, Michael, and Jeffrey S. Passel, 1994. Immigration and Immigrants: Setting the Record Straight, draft April 8, Washing- ton: Urban Institute, May. (Mimeographed) Passel, Jeffrey S. and Fix, Michael, 1994. "Statement," March 14, 1994. Rothman, Eric S., and Thomas J. Espenshade, "Fiscal Impacts of Immigration to the United States", Population Index, Vol. 58 #3, Fall, 1992, pp. 381-415. Simon, Julian L., 1981. "What Immigrants Take From and Give To the Public Coffers," in U.S. Immigration Policy As The Nation- al Interest, Staff Report of the Select Commission on Immigration and Refugee Policy, Appendix D, (Washington: GPO, pp. 223-261). ___, 1984. "Immigrants, Taxes, and Welfare in the United States." Population and Development Review, 10 (1): 55-69. ___, 1989. The Economic Consequences of Immigration to the U. S. Oxford: Basil Blackwell. ___, and Ather Akbari, 1995. "Welfare State Effects of Immigration Into Europe Through Transfers and Taxes." (Mimeo- graphed) Simon, Julian L. "A Correction of a Review by Borjas", Journal of Economic Literature, forthcoming, 1996. page 2 article5 pdr1 February 14, 1996 ENDNOTES **ENDNOTES** <1>: A misunderstanding of this point led George Borjas to mistakenly conclude that my calculations were in error. See Simon (forthcoming) which discusses Borjas's criticism. <2>: Consider, for example, the complaint in Rothman and Espenshade (1991, 386) that there was an "omission of older immigrant groups" and that this "gives a sense of incompleteness to Simon's findings. page 15 article5 pdr1 February 14, 1996