Findings (continued)

Socio-Demographic Characteristic

Insurance Coverage by Educational Attainment. One of the best predictors of the propensity to have insurance is a person’s education level, regardless of nativity or income. This is true for several reasons. First, because of the limited value of their labor in an economy that increasingly demands more skilled workers, those with little education are the most likely to hold jobs that do not offer health insurance as a fringe benefit or, if insurance is offered, the workers must pay a large share of the costs. Second, those with little education have the lowest incomes and may find it very difficult or simply impossible to afford to insurance on their own. Third, persons with few years of education have the most unstable employment histories—suffering higher unemployment rates and longer periods of unemployment than educated workers whose skills are more in demand. Since loss of insurance often accompanies loss of employment, higher rates of unemployment should have an effect on rates of insurance coverage. Table 2 reports the health insurance coverage of persons between 25 and 64 years of age in immigrant and native households by educational attainment. The table indicates that insurance coverage varies enormously based on education level, regardless of nativity. In 1998, 36.9 percent of all dropouts were without insurance; for persons with only a high school education, 19.8 percent were uninsured; for those with some college, 14.9 percent had no insurance; and 8.9 percent of individuals with four years of college were without health insurance. Since persons aged 25 to 64 in immigrant households are much more likely than those in native households to lack a high school education (31.8 percent compared to 10.4 percent), the high percentage of persons in immigrant households who have no insurance is at least partly explained by the higher proportion with few years of schooling.

Table 2 also shows, however, that education alone does not account for the lower rate of insurance coverage associated with immigrants. At every education level, adults in immigrant households are twice as likely as adults in native households to lack insurance. Thus, a lower level of education is by no means the only factor accounting for the high percentage of persons in immigrant households who lack insurance.

One way of examining the importance of education in explaining the lack of insurance associated with immigrants is to calculate the rate of coverage for persons 25 to 64 in immigrant households, assuming that their educational endowment was the same as persons in native households. In other words, what percentage of working age adults in immigrant households would lack insurance if their distribution across educational categories was the same as persons in native households and their rate of insurance coverage by education was unaltered. In 1998, if persons in immigrants households had the same educational endowment as natives, but retained the same rates of coverage by educational category, 30.9 percent of persons in immigrant households would have been without insurance. While this is less than the actual 35.3 percent who lacked insurance in 1998, it is still dramatically higher than the 15.2 percent for working-age adults in native households. This suggests that 4.4 percentage points or 22 percent of the 20.1 percentage point gap in the insurance coverage for persons 25 to 64 in immigrant and native households is accounted for by the lower levels of education associated with immigrants. While these results indicate that education is clearly an important factor, the causes of the problem are more complex. Other factors must also contribute significantly to the lack of insurance for individuals in immigrant households.

Insurance Coverage by Age. In addition to education, age also has an impact on one’s likelihood of having health insurance. Table 3 reports health insurance statistics by age for persons in immigrant and native households. As the table indicates, insurance coverage is less common among those 18 to 64 than for both children and older people. Most persons over 65 have insurance because all persons who receive Social Security, or who have a spouse who receives Social Security, automatically qualify for Medicare. While a larger share of persons in native households are over age 65 (12.5 percent compared to 7.9 percent for immigrants), this does not explain the difference in rates of insurance coverage between immigrant and native households. Even when retired persons are excluded, the gap between immigrants and natives remains.

Turning to children (17 and under) first, we see that overall they have a somewhat higher rate of insurance coverage than adults because children in low-income families are often eligible for government-provided means-tested insurance even when their parent are not. Since 1997, the federal government, along with the states, has had some success in increasing insurance coverage among children in low-income families under the State Children Health Insurance Program or SCHIP (Morse, 2000). Even so, about one in seven children in the United States was without health insurance in 1998.

Table 3 shows that children in immigrant households have much lower rates of insurance coverage than do their counterparts in native households. In 1998, 28.5 percent of children in immigrant households were without health insurance, compared to 13 percent in native households. The 3.1 million children in immigrant households who lack insurance now represent 28 percent of all uninsured children. There can be little doubt that recent immigration is making it more difficult and costly to reduce the number of children without insurance under SCHIP. Congress passed SCHIP in 1997 in an effort to increase insurance coverage among low-income children and it is expected to cost the federal government $20 billion over five years. By April of 2000, an estimated one million eligible children have been provided insurance as part of SCHIP (Morse, 2000). But immigration certainly has the potential to undo much of that progress. In 1998 there were 588,000 immigrants under the age of 19 (SCHIP’s target population) without insurance who had arrived in just the last four years. In addition, there were 114,000 children born to adult immigrants who arrived during the same period who did not have insurance in 1998. Thus, because of recent immigration the number of children without health insurance was more than 700,000 larger than it would otherwise have been, erasing most of the gains made so far under SCHIP. In the years to come, it is very likely that immigration may offset much of the progress made as a result of SCHIP

While the gap between children in immigrant and native households is very large, it is also extremely large for persons between the ages of 18 and 64 and for all subdivisions of that age group. Table 3 indicates that 37.9 percent of individuals in immigrant households between 18 and 64 lacked insurance in 1998, compared 16.9 percent of individuals in native households. This difference can only be described as huge, with those in immigrant households being more than twice as likely to lack insurance as those in native households. Table 3 also shows that differences in the age distribution between persons in immigrant and native households do not account for the dramatically lower rates of insurance coverage associated with immigrants. In some of the younger age groups, the gap between persons in immigrant and native households is actually larger than the gap among all persons in immigrant households and all persons in immigrant households. For example, there is a 23.1 percentage point gap between persons aged 18 to 24 in immigrant and native households. This difference is slightly larger than the 18.5 percentage point gap that exists when comparing all persons in immigrant households to all persons in native households. Therefore, the age distribution of persons in immigrant households would seem to account for little if any of the large difference between the two groups in insurance coverage. We must look elsewhere if we are to understand why so many immigrants and their children lack health insurance.

Insurance Coverage by Country and Region of Birth. Table 4 reports insurance coverage by regions of the world and country of birth based on the household head. Turning to region of birth first, the table reveals that while persons in immigrant households are much more likely to be without insurance than those in native households, there are substantial differences among immigrants. Persons in Mexican- and Central and South American-headed households are the most likely to lack health insurance. In contrast, households headed by Canadian and European immigrants are the most likely to have health insurance. The sizes of the differences between regions are very large in some cases. Persons in Mexican and Central American households, for example, are more than three times as likely to be uninsured as persons in households headed by immigrants from Europe and Canada.
While there are some immigrant groups who are about as likely as natives to have insurance, lack of health insurance coverage is widespread among immigrants and is not simply confined to a few groups. In addition to the very large share of persons in Mexican and Cental American households without insurance, persons in households headed by immigrants from Sub-Saharan Africa, the Middle East, South Asia, South America and the Caribbean are either twice as likely to be without insurance as are persons in native households or very nearly so. While it is clear that lack of health insurance is a much greater problem for some groups than others, low rates of insurance coverage is common among many immigrant groups.

In addition to insurance coverage by region of origin, Table 4 also reports the percentage uninsured for the 15 countries with the largest number of post-1970 immigrants living in the United States in 1998. Given the sample size of immigrants from some countries, however, the differences between countries should be interpreted with caution. Smaller sample size means that there is greater variability for individual country estimates than for the estimates based on region of origin. The estimates for individual countries should be used to make determinations of the relative differences between countries, and should not be seen as quantified absolute differences. Table 4 shows that like the data by region, rates of insurance coverage differ significantly by country. In fact, the variation between immigrant households from different countries is much larger than is the difference between persons in immigrant and native households. In 1998 for example, the percentage of persons without insurance in Salvadoran households was more than five times that of persons living in households headed by Canadian immigrants. Even among immigrants from the same region of the world, rates of insurance coverage vary significantly. For example, persons in Korean households are much more likely to be uninsured than immigrants from other East Asian countries. Among Caribbean immigrants, persons in Cuban household are much more likely to be uninsured than are persons in households headed by immigrants from that region.

What Accounts for the Large Differences Between Immigrant Groups? While Table 4 is not a fully developed explanatory model of insurance coverage, it does provide important clues to why there are such large differences between immigrant groups. The right side of the table reports different characteristics for persons in immigrant households by region and country of origin. The first column on the right-hand side reports the percentage of persons who are in or near poverty using the official federal threshold.¹⁰ Groups, such as Mexicans, Central Americans, and those from the Caribbean, who have the highest poverty rates, also tend to have the lowest rates of insurance coverage. The same is true for education levels, which is reported in the second column on the right. A large share of adults in households headed by immigrants from these regions lacks even a high school education. Of course, high poverty rates and low levels of education do not explain all of the differences. Persons living in Dominican-headed households, for instance, are much more likely to be insured, despite very high poverty rates, than persons living in Mexican or Central American households. The main reason Dominicans are more likely to have insurance is that they make use of Medicaid at much higher rates than do other groups. It would appear that high rates of Medicaid use can offset, at least in part, the very low incomes of many Dominican households. Another example of Medicaid offsetting higher poverty rates can be found among households headed by immigrants from the former Soviet Union. While their poverty rates are high, their very heavy use of the Medicaid system means that the percentage without insurance is relatively low. The last column in the table shows the percentage of persons living in households headed by pre-1990 immigrants. One would expect that those groups that have been in the country longer on average would be the most likely to have insurance. While there is some evidence for this in the table, the relationship between insurance coverage and duration of stay seems weaker than for the other causal factors reported in the table. Of course, it should be recalled from Table 1 that lack of insurance remains a severe problem even after immigrants and their children have been in the country for many years. Therefore, it is not surprising that the percentage of persons living in established households seems to account for little of the difference between immigrant groups.

Insurance Coverage for Refugees and Asylees. The bottom portion of Table 4 gives estimates of insurance coverage for immigrants from the major refugee-sending countries. Not all individuals from these countries were admitted as refugees, but by reporting estimates separately for persons in households headed by immigrants from these countries, it is possible to gain some insight into the extent to which the lack of insurance coverage for persons in immigrant households reflects immigrants admitted based on humanitarian criteria. The table shows that immigrants from these countries are actually less likely to be uninsured than immigrants from non-refugee countries. In 1998, 23.7 percent of people living in households headed by an immigrant from one of the major refugee-sending countries were uninsured, compared to 33.3 for persons in household headed by immigrants from non-refuge-sending countries. The table also shows that, not surprisingly, persons in households headed by immigrants from refugee countries have much higher rates of Medicaid use than do other immigrants. In fact, the rules governing use of the entire welfare system (including Medicaid) by immigrants admitted for humanitarian reasons are more generous than for non-humanitarian immigrants and even for natives. As a result, the percentage without insurance in refugee households is somewhat lower than for other immigrants. While Table 4 does not definitively identify refugees, one thing is clear from these estimates: The high percentage of persons without health insurance in immigrant households is not a function of humanitarian immigration. Persons in immigrant-headed households headed by immigrants from non-refugee countries are much more likely to be uninsured than are those in households headed by immigrants admitted for humanitarian reasons.

Insurance Coverage by Race. Race is also an important predictor of insurance coverage. Table 5 reports health insurance statistics by race for persons in immigrant and native households, with Hispanics treated as a separate and discrete category. Since region of the world is closely correlated with race, the pattern in Table 4 follows the same general pattern found in Table 5. White and Asian immigrant households are the most likely to have insurance, while Hispanic and black immigrant households are the most likely to be without insurance coverage. Table 5 shows that immigrant households have a very large impact on the overall rate of insurance coverage for all Hispanics and Asians. In 1998, 20.7 percent of all Asians were uninsured—6.8 percentage points higher than natives. However, when only Asians living in households headed by a native are examined, the rate is 14.4, roughly the same as the 13.9 percent for all natives. Persons in immigrant households also have a significant effect on the rate of insurance coverage for Hispanics. In 1998, 35.3 percent of all Hispanics were uninsured. For Hispanics in native households the rate was 24.8, 10 percentage points lower than for all Hispanics, and 20 points lower than Hispanics living in a households headed by an immigrant. These data indicates that immigrants and their children have a very large impact on rates of insurance coverage for some racial minorities. This strongly suggests that what could be interpreted only as reflective of racial disparities in the United States may in part actually reflect immigration policy.

Income, Poverty Status, Workforce Status, and Employer Size

This section examines health insurance coverage for persons in immigrant and native households based on their economic characteristics. As we will see, having health insurance is correlated with other measures of economic welling-being. In the analysis that follows, we further develop the relationship between insurance and such factors as income and work force status.

Insurance Coverage by Income. Table 6 shows health insurance coverage for persons in immigrant and native households by annual household income.¹¹ It shows that there is a strong positive relationship between income and health insurance coverage. Persons with low incomes are the most likely to lack insurance, while those with the highest are the most likely to have insurance. In 1998, 27.0 percent of all persons in households with incomes of less than $25,000 were without insurance compared to 7.3 percent of all persons in households with incomes of $75,000 a year or more. Since 36.3 percent of persons in immigrant households (compared to 27.5 percent of natives) live in households with incomes below $25,000, it is not surprising that immigrants and their dependents are much more likely to be uninsured. Moreover, only 18.5 percent of persons in immigrant households had incomes of more than $75,000 a year, compared to 23.8 percent of persons in native households. Their higher concentration at the bottom of the income distribution as well as their under-representation at the top of the income distribution accounts for some of the large difference in insurance coverage between immigrant and native households.

However, Table 6 also shows that persons in immigrant households are still much more likely than those in native households to lack insurance even after controlling for income. In every income category in Table 6, the gap in insurance coverage between the two groups is dramatic. At the bottom of the income distribution (<$25,000) for example, persons in immigrant households are nearly twice as likely to be uninsured as persons in native households with the same income. And for the other three income categories, the percentage uninsured in immigrant households is more than double that of persons in native households. Even among those with incomes of at least $75,000, the difference between persons in immigrant and native households is very large. In fact, the percentage of persons in immigrant households with incomes of over $75,000 who are uninsured is actually slightly higher than for all persons in native households— even though only about one in four people in native households lives in a household with an income that high. In 1998, 14.9 percent of persons in immigrant households with incomes over $75,000 were without insurance, compared to 13.9 percent of all persons in native households. The large percentage of high income immigrants without insurance is striking because even if an employer does not provide any assistance in paying for insurance, households with incomes of more than $75,000 should have the resources to purchase it on their own. Moreover, using the definition of households in this study, immigrant households with high incomes are not much larger than those of natives — both have fewer than four people in them on average. The fact that so many persons in high income immigrant households do not have insurance strongly suggests that this may be a conscious choice on the part of many immigrants, perhaps reflecting cultural attitudes about the value of insurance, rather than an inability to afford coverage.

Insurance Coverage by Poverty Status. It is well established that persons in poverty (also referred to as poor) are at the highest risk for being uninsured (Campbell, 1999). They are often uninsured because many do not work or, if they are employed, work at low-wage jobs that do not provide insurance coverage. Moreover, buying insurance on their own is likely to be very difficult for those in poverty because of their limited incomes. Of course, in many cases persons in poverty are eligible for Medicaid and this tends to increase their rate of insurance coverage. Those with income only just above the poverty threshold (referred to here as the near-poor) also tend to be without insurance for many of the same reasons as those below the poverty threshold.

Table 7 reports health insurance coverage by poverty status using the official government poverty threshold. (For a family of four in 1998, poverty was defined as having an income below $16,660.) The table shows that nearly one in three of all persons in poverty lacked health insurance in 1998. However, among the poor there are significant differences between immigrant and native households. In 1998, 28.8 percent of poor persons in native households were uninsured compared to 45.6 percent of persons in immigrants households. The same general pattern exists for the near-poor. Persons are considered to be near-poor in Table 7 if their income is between 100 percent and 200 percent of the poverty threshold—$33,320 for a family of four. In 1998, 42.9 percent of the near-poor in immigrant households were uninsured compared to 21.6 percent in native households. One relatively minor reason persons in immigrant households living in or near poverty are more likely to be uninsured is that Medicaid use among the poor and near poor in immigrant households is somewhat lower than for persons in native households. In 1998, 26.8 percent of persons in native households in or near poverty used Medicaid. In contrast, 24.3 percent of the poor and near poor in immigrant households used Medicaid. However, this small difference only accounts for a small part of the gap between the two groups. If people in immigrant households living in or near poverty used Medicaid at the same rate as their native-born counterparts, 42.2 percent would still be uninsured. While this is lower than the 44.1 percent who were actually uninsured in 1998, it is still much higher than the 24.5 percent of the poor and the near-poor in native household without insurance.

More important than Medicaid use in explaining the lack of insurance coverage associated with immigrants is the fact that persons in immigrant households are much more likely to be in or near poverty than natives. In 1998, more than 20 percent of persons in immigrant households were in poverty in contrast to less than 12 percent of persons in native households. Persons in immigrant households are also more likely to be near-poor than persons in native households—26.6 percent compared to 16.7 percent. Overall, the poor or near-poor accounted for 63.8 percent of the uninsured population living in immigrant households. Therefore, the higher rate of poverty and near poverty is one of the most important reasons why so many immigrants and their children lack insurance. If so many persons in immigrant households were not in or near poverty, the difference in rates of health insurance coverage between individuals in immigrant and native households would not be as large. However, as is the case with the factors examined so far in this report such as education and income, even after accounting for poverty status, immigrants and their children are significantly more likely to be without insurance than similarly situated natives and their children.

Insurance Coverage by Workforce Status. Since most Americans receive insurance through their employers, persons under the age of 65 who do not work tend to have a very low rate of insurance coverage compared to workers. Table 8 reports insurance coverage for persons aged 25 to 64 based on whether they worked in 1998 and, if they did work, whether they were employed full or part time. (Persons are considered full-time if they usually work at least 35 hours per week.) The table shows that 44.4 percent of persons in immigrant households who did not work were uninsured in 1998, compared to 21.7 percent of non-workers in native households. Thus, adults in immigrant households who did not work were about twice as likely to be uninsured as non-workers in native households. This huge difference is striking because by examining only persons in their primary working years, Table 8 does not include the large number of elderly natives who do not work but are covered by Medicare. Moreover, by excluding persons younger than 25 and older than 64, the table compares persons of very similar ages. Persons in this age group in immigrant households are on average 41 years of age while the average for those in native households is 43. This small difference in age cannot account for the huge difference in insurance coverage between the two groups.

Large differences also exist in insurance coverage for immigrant and native households when we examine full- and part-time workers separately. In 1998, 39.5 percent of persons in immigrant households employed part time were uninsured, compared to 20.7 percent of part-time workers in native households. Turning to full-time workers, the gap is also very large, with 31.9 percent of persons in immigrant households being uninsured compared to only 12.8 percent of persons in native households. This means that persons in native households working part time are much more likely to be covered by health insurance than are full-time workers in immigrant households. This is surprising because full-time workers are more likely to be offered insurance by their employers.

Part of the explanation for the huge difference in insurance coverage is accounted for by the large share of immigrants who hold low-skilled jobs. As shown in Table 2, a much larger share of working-age immigrants have few years of schooling. Because of the limited value of their labor, less-skilled workers often are not offered health insurance as a fringe benefit, or if insurance is offered the employee must pay a large percentage of the costs. However, even when we account for education and confine our analysis to only full-time workers aged 25 to 64, very large differences remain. In 1998 for example, 14.5 percent of full-time workers aged 25 to 64 with at least a four-year college degree in immigrant households were uninsured. In contrast, only 5.9 percent of such persons in native households were uninsured. As we will see later in this paper (Table 10), similar differences exist for other educational categories.

The large differences in rates of insurance coverage between workers in immigrant and native households have important implications not only for individual workers, but also for their dependents. Most children are covered by insurance provided by an employer to their parents and many adults are covered by a spouse’s insurance. Table 9 reports insurance coverage for all persons based on whether there is at least one full- or part-time worker in the household. It shows that persons in households where no one works are the most likely to be insured. At first glance this may seem surprising, until one considers that the vast majority of these individuals are either elderly and covered by Medicare or are welfare recipients and covered by Medicaid. Thus, most of the people in households where no one works do have health insurance.

Table 9 also shows that the vast majority of the uninsured (immigrant and native) live in households with at least one worker.¹² In 1998, 91.1 percent or 10.3 million of the 11.6 million of the uninsured in immigrant households lived in a households where at least one person worked. In comparison, a slightly lower percentage (88.2 percent) of the uninsured in native households lived in a household with a least one worker. While the majority of the uninsured live in households with at least one worker, there is stark difference in rates of insurance coverage between immigrant and native households with workers.

One third of persons in immigrant households with at least one worker were uninsured in 1998, compared to 14.3 percent of persons in native households with at least one full-time or part-time worker. The gap in insurance coverage for all persons in immigrant and native households is just as large when we turn to households where at least one of the workers is employed full time. In 1998, 33 percent of persons in immigrant households where there was at least one full-time worker present were uninsured, compared to 13.7 percent of persons in native households with at least one full-time worker. This 19.3 point gap is basically the same size as the 18.5 point difference that exists for all persons in immigrant and native households. The gap is also very large for households with only part-time worker(s), with 37.4 percent of persons in immigrant households lacking insurance in contrast to 21.9 percent of persons in native households. Overall the findings in Tables 8 and 9 indicate that work force experience in 1998 accounts for little of the difference in insurance coverage between the two groups. While the vast majority of the uninsured live in households with at least one worker, the differences in insurance coverage between persons in immigrant and native households can only be described as huge even after controlling for the presence of a worker(s) and the full/part time status of the worker(s).

Insurance Coverage by Employer Size and Education. Table 10 accounts for several factors at once. The table looks at only full-time workers between the ages of 25 and 64, and controls for both education and employer size. This last factor is especially important because smaller firms are less likely to offer insurance to their workers than larger firms. In 1998, only 29.3 percent of all persons working for companies with fewer than 25 employees were covered by insurance provided by their employer. In contrast, two-thirds of those working for employers with more than 500 employees had employer-based health insurance¹³ (Campbell, 1999). By confining our analysis to full-time workers between the ages of 25 and 64, Table 10 also accounts for both hours worked and age. Because only full-time workers are examined and the age range is limited, the average age for workers in immigrant households is 40 years compared to 41.6 years for those in native households. Thus, Table 10 compares workers of very similar characteristics.

The top portion of each cell in Table 10 reports the percentage of workers in each category who do not have health insurance provided by their employer. Many individuals who do not have insurance through their employer are instead covered by health insurance as part of a family member’s plan or have insurance they purchased on their own. Additionally, in some cases low-wage workers are eligible for insurance provided by the government though Medicaid and some individuals receive insurance coverage because of prior military service. Therefore, the second number in each cell reports the percentage of persons who have no insurance of any kind. Looking at the first cell in the upper left hand corner, Table 10 reads as follows: in 1998, 80.8 percent of full-time workers in immigrant households without a high school degree were not covered by insurance provided by an employer, and 66.9 percent of these same individuals where uninsured. This means that the overwhelming majority of dropouts in immigrant households working full time for small firms were not covered by employer-provided health insurance and were also unable to secure insurance by other means.

Turning next to the totals at the bottom of the table, we find that workers employed at smaller firms are the most likely not to be covered by employer-provided insurance. This is true for both persons in immigrant and native households. For all workers in immigrant households employed at firms with fewer than 25 employees, 73.5 percent did not have employer-provided health insurance and 59.9 percent of all workers in native households working for small firms did not receive insurance from their employers. In contrast, 29.1 percent of all workers in immigrant households employed by a firm with at least 500 employees lacked employer-provided health insurance and for those in native households working for large firms, 19.3 percent did not have health insurance provided by their employers. Since immigrants are more likely to work for small firms, this partly explains why so many immigrants and their children are uninsured. In 1998, 33 percent of persons in immigrant households worked for an employer with fewer than 25 employees compared to 25 percent of workers in native households. However, the larger share of workers from immigrant households employed by small firms does not account for the large difference in the rates of insurance coverage for persons in immigrant and native households. Table 10 shows that even after accounting for employer size, full-time workers in immigrant households are much less likely to be insured by their employer and they are also less likely to have insurance from some other source. As already pointed out, 73.5 percent of workers in immigrant households employed by a small firm (fewer than 25 workers) did not have employer-provide health insurance compared to 59.9 percent of workers in native households employed by firms of the same size. Thus, a very sizable gap (13.6 percentage points) remains between full-time workers in immigrant and native households even after employer size is taken into account.

Table 10 also shows that about one-third of the workers in immigrant households employed by a small company who do not have employer-provided health insurance are able to get insurance elsewhere. Of workers in native households working for small employers and who are not insured through their job, more than half were able to obtain health insurance from some other source. This shows that not only are workers in native households more likely to be covered by insurance from their employer, they are also more likely to get insurance from another source when not covered by employer-provided health insurance. This same pattern exists throughout every cell in Table 10.

As already discussed, a greater percentage of adults in immigrant households have few years of schooling. This contributes significantly to the high percentage without insurance (see Table 2). But Table 10 also shows that even controlling for education and employer size, full-time workers in immigrant households are much more likely to be uninsured than workers in native households. This can be seen clearly by looking at college educated workers, who are the most likely to have insurance. In 1998, 31.8 percent of full-time workers in immigrant households with four or more years of college did not have insurance from their employer, compared to 21.9 percent in native households. This nearly 10-percentage-point gap in the rate of employer-provided insurance between college graduates in immigrant and native households narrows somewhat among workers employed by larger companies. Of workers in immigrant households with at least a college education and employed by a firm with between 100 and 499 employees, 23.2 percent did not have insurance through their job. For those in companies with more than 500 employees, 17.4 percent did not have insurance from their employer. For natives the corresponding figures are 16.4 percent and 13.7 percent. Although the differences associated with educated workers in immigrant and native households are somewhat smaller for those employed by larger companies, college-educated workers in native households are still more likely to have employer-provided health insurance than college-educated workers in immigrant households. The same general patterns exist in the rest of the table. Workers in immigrant households are much more likely to be without employer-provided health insurance or health insurance of any kind than their counterparts in native households, even after controlling for education level and employer size.

While the CPS does not ask respondents why they do not have health insurance, it seems plausible that the difference in rates of employer-provided coverage may partly reflect attitudinal differences between immigrants and natives. Since there is often some cost to the employee, a larger percentage of workers in immigrant households may be choosing to keep the employee’s contribution and simply not enrolling in their employers insurance plan. Given the lower average income and education level of immigrant households, this is not surprising. However, immigrants with a college education should have incomes that allow them to purchase insurance, even if their employer makes them pay a large share of the costs. Even if we confine our analysis to only college-educated workers 25 to 64 years of age, employed full-time, in households with income of more than $75,000, we still find that those in immigrant households are about twice as likely as those in native households with the same characteristics to be without insurance.

In addition to not taking advantage of insurance when it is available, lower rates of employer-provided health insurance among immigrant workers may also reflect the fact that immigrants, in contrast to natives, do not think insurance is as important a fringe benefit when choosing where to work. Having come from societies where health insurance is rare, some immigrants may be more willing to work for employers who do not offer health insurance. If this is the case, it would be roughly analogous to the often-made argument that immigrants are willing to work for lower wages than natives. A difference in how they view the value of health insurance might explain why a much higher percentage of college-educated full-time workers in immigrant households do not have employer-provided health insurance.

Further support for this possibility can be found by looking at the percentage of workers by education who have no insurance of any kind. Looking at the most educated workers indicates that 14.5 percent of all full-time workers in immigrant households with at least a college-education did not have health insurance in 1998 (bottom of Table 10). In contrast, only 5.9 percent of college-educated full-time workers with at least a college education in native households were uninsured. Part of the reason for these differences in overall rates of insurance coverage is that a much larger percentage of educated natives who did not get insurance from their employer acquired it from another source. Of college-educated workers in immigrant households who did not have employer-provided health insurance, 56 percent were able to get insurance from another source. In contrast, of college-educated workers in native households who did not have employer-provided health care, 73 percent obtained insurance from another source. This would seem to support the argument that a smaller share of educated persons in immigrant households desire to have insurance than educated persons in native households. Since educated workers generally have the highest incomes and should therefore be able to afford coverage if it is not provided by their employers, this suggests that many immigrants may simply be choosing not to have health insurance, even if they have the financial resources to purchase it on their own.

If a larger percentage of immigrants than natives do not see health insurance as a necessity, there will be a number of negative implications. For the immigrant and his family it may mean a reduction in the amount and quality of the health care they receive. It can also be seen as irresponsible in a society that will provide emergency and some basic health care to all. By not getting insurance when it can be obtained, immigrants are in effect shifting the costs of their health care from themselves to the rest of society, which will be forced to provide more funding for services to the uninsured.

Of course, the decision is about whether to purchase health insurance, where to work, and what fringe benefits are desirable are complex and can be affected by many factors not considered here. While the evidence suggests that immigrants are less likely to want health insurance than natives, other explanations are certainly possible. It should also be recalled that other evidence presented in this study indicates that the primary reason persons in immigrant households are much more likely to lack insurance is their lower level of education and lower incomes. Still, the larger proportion of educated and affluent immigrants without insurance is clearly an important finding. Exploring this question in greater detail should be an area of future research.

Geographic Area

So far this report has only examined health insurance coverage based on the individual characteristics of persons in immigrant and native households at the national level. While certainly important, national figures may obscure important differences that exist across the country. For this reason, we now turn to an examination of insurance coverage by state and metropolitan area. The question of immigration’s impact on the size of the uninsured population at the sub-national level is especially important because state and local governments pay many of the costs associated with providing care to those without insurance.

Insurance Coverage by State. Table 11 provides data on insurance coverage for persons in immigrant and native households for the eight states with the largest number of persons living in immigrant households. In 1998, 76.6 percent of all persons living in immigrant households resided in one of these states. The first and second columns in the table report the percentage of persons in immigrant and native households without health insurance. The last two columns of the table provide the percentage uninsured in each state using a combined sample of the March 1997 and 1998 CPS. The fact that the figures change so little when 1997 data are added demonstrates the strength of the findings in the first two columns. The table indicates that in all eight states, persons in immigrant households are much more likely to be uninsured than are individuals in native households. The differences in rates of insurance coverage for persons in immigrant and native households in 1998 are statistically significant using a 90 percent confidence interval. In light of non-sample errors that exist in any sample, including the CPS, the smaller difference between the two groups in Massachusetts should be interpreted with caution, especially since the state has the smallest population living in immigrant households of those listed in Table 11.

The largest difference between immigrants and natives can be found in Arizona, where persons in immigrant households are almost three times as likely to have no insurance as persons in native households. Arizona is by no means the only state where the difference in insurance coverage is enormous. In Texas, California, New York, and Illinois, persons in immigrant households are more than twice as likely as those in native households to be uninsured. And while persons in immigrant households in Florida and New Jersey households are not twice as likely to be uninsured as those in native households, the difference in both states is very close to being so. Table 11 makes clear that with the possible exception of Massachusetts, lack of health insurance is a serious problem among persons in immigrant households in every major immigrant-receiving state in the country. Low rates of health insurance coverage among immigrants and their children is not simply confined to one state or even one part of the country. While Table 11 shows that the situation is worse in western states, lack of insurance among persons in immigrant households also is a serious problem in New York, Florida, Illinois, and New Jersey. Thus, despite the fact that immigrants differ significantly across states by country of birth and other demographic characteristics, health insurance coverage is a serious problem in immigrant communities throughout the country.

The fourth and fifth columns of Table 11 report the percentage of persons in immigrant households receiving Medicaid. In five of the eight major-immigrant receiving states, persons in immigrant households are much more likely to be on Medicaid than are persons in native households. In the other three states, Medicaid use for persons in immigrant households is roughly equal to that of natives. Since immigrant use of Medicaid is as high as or higher than that of natives in every state, it is clear that lack of access to Medicaid by itself cannot be seen as the reason why individuals living in immigrants households are so much more likely to be without insurance.

Combining the Medicaid use rates with the percentage uninsured indicates that an enormous proportion of immigrants and their children either have no insurance or must have it provided to them at taxpayer expense. For example, 63.9 percent and 62 percent, respectively, of persons in immigrant households in Arizona and Texas are on Medicaid or have no insurance, and more than half of the persons in immigrant households in California and New York also have no insurance or have it provided to them by the government. In the other four states, more than one-third of the population in immigrant households is uninsured or on Medicaid. Massachusetts in particular does not seem to be the "success" story it appears to be when only the percentage of persons in immigrant households without coverage is considered. While immigrant households in Massachusetts do have the highest rate of insurance coverage of any major immigrant receiving state, this fact partly reflects the very heavy use of Medicaid by immigrant households in that state. In 1998, nearly one out of five persons in Massachusetts living in an immigrant households used Medicaid — only New York has a higher rate of Medicaid associated with immigrants.

By itself, the high percentage of persons in immigrant households without insurance would not necessarily create a large burden for the health care system in these states were it not for the fact that so many immigrants have entered the country in the last three decades. If there were relatively few immigrants then a large percentage without health insurance would not necessarily be of concern to the nation as a whole.¹⁴ And dealing with the problem would likely be both easier and less costly. In order to measure the impact immigration has on the size of the uninsured population in each state, the sixth column of Table 11 shows the proportion of the entire uninsured population who live in immigrant households. In California, Arizona, and New York, the impact of immigration on the size of the uninsured population is difficult to overstate. In all three states, more than 40 percent of the uninsured live in immigrant households. In the other five states the impact of immigration on the size of the uninsured population is somewhat smaller, but still very large. In Texas, Florida, and New Jersey, immigrant households account for nearly a third of the uninsured and in Illinois, one fourth of the uninsured live in an immigrant households. Even in Massachusetts, one out of five uninsured residents lives in an immigrant household. There can be little doubt that in all of these states, immigration has had a significant effect on the size of the uninsured population. For these states, any discussion of the uninsured problem must include a discussion of immigration policy.

The seventh column in Table 11 shows the percentage of the total population in each state who live in immigrant households. It shows that the immigrant percentage of the uninsured (column six) is not proportional to their representation in the total population. In fact, in some states immigrants represent nearly twice the proportion of the uninsured as they do of the overall population. Further evidence of immigration’s impact on the problem of insurance coverage can be found in the third column of Table 11, which shows the overall percentage of the population in each state who are uninsured. By comparing the percentage of persons in native households who are uninsured (column two) with column three one can see that the presence of immigrants in most states dramatically increases the total share of the states’ populations who are without health insurance. Taken together, the results in Table 11 indicate that in the major immigrant-receiving states, immigration has had a substantial impact on the size of the uninsured population. Health insurance coverage is a significant problem for persons in immigrant households throughout the country. The problem is not confined to one state or even one region. Lack of health insurance among immigrants and their children is a problem national in scope.

Insurance Coverage by Metropolitan Area. In addition to examining insurance coverage by state, it also is possible to examine rates of insurance coverage in the nation’s largest metropolitan areas. Because of the sample size in the CPS, however, immigrant insurance coverage can only be analyzed in a few very large Metropolitan Statistical Areas (MSA). However, it is possible to use combined metropolitan areas, referred to as Consolidated Metropolitan Statistical Areas (CMSA) by the Census Bureau, to examine the impact of immigration at the local level using the March 1998 CPS.

In 1998, there were 18 officially designated CMSAs. Of these, nine have a sufficiently large number of persons living in immigrant households to examine them using the CPS. Table 12 is structured in the same manner as Table 11 except that it reports figures for CMSAs and not states. The table shows that in every CMSA with a large immigrant population, lack of health insurance coverage is dramatically higher among persons in immigrant households than it is for individuals in native households. While the CPS is not designed to measure differences in sub-populations within CMSAs, the difference between the two is statistically significant in every CMSA. The fact that the 1997-98 two-year average in the last column of the table shows very similar differences in rates of insurance coverage lends strong support to the findings in the first two columns of the table. However, the differences between immigrant and native households in Miami and Boston are close enough that non-sampling errors may create statistical insignificance. Therefore, the difference in those two cities should be interpreted with some caution.

Putting aside Miami and Boston, which account for only about 5 percent of the nation’s total population living in immigrant households, the difference between immigrant and native insurance coverage in the nation’s largest cities is enormous. In six of the nine metropolitan areas, persons in immigrant households are more than twice as likely to be without health insurance as persons in native households. While higher than that of natives in every metropolitan area, the percentage of persons in immigrant households without insurance varies significantly between cities. In 1998, It ranged from a high of 58.7 percent for persons in immigrant households living in the Houston CMSA to a low of 17.4 percent for persons in immigrant households in the Boston area. Table 12 also shows that even in metropolitan areas where the immigrant population tends to be more skilled and have higher incomes, such as San Francisco and Washington, a very large percentage of immigrants lack health insurance. The findings in Table 12 confirm the results reported in Table 11 — lack of health insurance among persons in immigrant households is a problem throughout the country.

As a share of the total uninsured population, the highest percentage is found in Los Angeles, where 64.1 percent of all uninsured persons live in immigrant households. Miami is next with 58 percent, and New York is third with 50.2 percent of the uninsured living in immigrant households. The fact that immigrants and their children account for more than half of the uninsured in these cities may not be too surprising given the very high level of immigration to these metropolitan areas in recent decades, although in each city immigrant households are still disproportionately represented among the uninsured. But even in Boston and Washington, cities with much smaller immigrant populations, persons in immigrant households account for almost one-fourth of the uninsured. The figures reported in Table 12 indicate that while the debate over the uninsured may not yet reflect it, in many parts of the country immigration has become a determinate factor in the size of the uninsured population. Without a change in immigration policy, immigrant households will likely account for a growing share of the uninsured in cities like Boston and Washington as well as other parts of the country in the near future.

Tables 11 and 12 show that in many parts of the country, any discussion of the uninsured must include a discussion of immigration policy. In the major immigrant-receiving cities and states, the size of the uninsured population is one-third larger to twice the size it would otherwise be were it not for immigrants and their children. Clearly, to exclude immigration policy from the debate over how to deal with the health care insurance issue would be to ignore a central aspect of the problem.