Underlying the increase of 700,000 nonelderly uninsured was an increase of 1.5 million uninsured adults which was offset by a reduction of 800,000 in the number of uninsured children. Both adults and children experienced a decline in employer-sponsored insurance (ESI) and an increase in public coverage. However, the increase in public coverage for children was substantially larger than that of adults reflecting higher income eligibility levels for children in Medicaid and the Children’s Health Insurance Program (CHIP), and increased enrollment in these programs in the face of the economic downturn. This increase was well in excess of the decline of ESI. The result was that the number of uninsured adults increased sharply while the number of uninsured children fell.
Health Insurance Trends, 2000-2008. The general trends in health insurance coverage that began as early as 2000 continued in 2008 with the share of the nonelderly population covered by employer-sponsored insurance (ESI) declining, the share covered by public program coverage increasing, and yet the share that is uninsured growing. Figures 1, 2, and 3 respectively depict these changes between 2000 and 2008. While these trends are somewhat disrupted by a revision in survey methodology that occurred in 2004, the direction of change is clear.3
There has been a continuing decline (with the exception of 2007 for adults) in the rate of employer-sponsored insurance throughout the decade (Figure 1). The decline in ESI has been on average slightly less than one percentage point each year. However, much of the decline in the percent of the nonelderly population with ESI has been offset by population growth.
Note: Each percent corresponds to the percentage of the age group with employer-sponsored insurance (ESI) in the given year. Each italicized number in ( ) represents the total number of people in millions in the age group with ESI coverage in that year.
Medicaid and CHIP coverage has continually increased over the period (Figure 2). This is particularly true for children. The percent of children with public coverage increased from 16.7% to 25.2% between 2000 and 2008. Public coverage for adults also increased but at a slower rate. The increased rates of Medicaid/CHIP coverage resulted in fairly large increases in the number of people with public coverage. We estimate, using the Current Population Survey, that 19.9 million children were covered by Medicaid and CHIP in 2008, up from 12.7 million in 2000; among adults the increase has been from 8.9 million in 2000 to 13.3 million in 2008.
The percent of the population that is uninsured has generally increased for adults while it has declined among children (Figure 3). Ignoring changes in survey methodology, the number of uninsured adults has increased from 30.2 million in 2000 to 37.6 million in 2008. The number of uninsured children has declined from 9.4 million in 2000 to 8.1 million in 2008. 4 The number of uninsured children fell slowly between 2000 and 2004 despite poor economic conditions. Between 2004 and 2006, the number of uninsured children rose. However, in the last two years the number has again declined, largely because of expansions in eligibility for children in the Medicaid and CHIP programs. The number of uninsured adults has increased throughout the period with the brief exception of a drop in 2007. In this issue paper we examine the single year of health insurance change between 2007 and 2008 in more detail.
Main Findings
Underlying the health insurance coverage changes between 2007 and 2008 is a striking change in the income distribution of the nonelderly population. The nonelderly population grew by
1.3 million; however the population below 200% of the federal poverty level (FPL) increased by 5.1 million, while the middle income group (those with income between 200% and 399% FPL) declined by 900,000 and the high income group (those with incomes at or above 400% FPL) declined by 2.9 million. Thus, the nonelderly population clearly shifted towards the poor and near poor. Since low-income people are less likely to have ESI and more likely to be uninsured, this shift in itself affected coverage.
Among all nonelderly, regardless of income, there was a decline in ESI from 62.9% to 62.0% of the population and an increase in Medicaid/CHIP coverage from 11.8% to 12.6% (Figure 4 shows the percentage point changes and the gains and losses in millions of people). The net result was that the uninsured rate increased from 17.2% to 17.4% (not statistically significant). The number of uninsured increased from 45.0 to 45.7 million. The number with ESI declined by 1.7 million, while the number with Medicaid/CHIP coverage increased by 2.3 million (Table 1).
* Statistically significant change (p<.05). Medicaid also includes CHIP, other state programs. Figure excludes changes in Medicare/Tricare/Other federal and private non group insurance for ease of presentation because changes in coverage are generally small. Source: Urban Institute.
Among adults, the rate of ESI fell from 64.5% to 63.3%, resulting in a decline of 1.3 million adults with ESI. The percent with public coverage increased from 6.8% to 7.2% or by 1.0 million people (Table 1). The increase in public coverage was not enough to offset the decline in ESI; as a result the percentage of uninsured adults increased from 19.7% to 20.4%, and the number of uninsured adults increased from 36.1 million to 37.6 million.
Among children the results were quite different. The rate of ESI fell from 59.4% to 58.8%. The increase in Medicaid/CHIP coverage was substantially greater however, growing from 23.5% to 25.2%. Consequently, the number of uninsured children declined by 800,000. Their uninsured rate declined by a full percentage point from 11.3% to 10.3%.
Changes in Health Insurance Coverage by Income, 2007-2008
Table 1 also shows changes in coverage for low, middle, and high income individuals. Most of the change in coverage was among those with low incomes. As noted above, the share of the nonelderly population below 200% FPL increased dramatically. Among this income group, there was a decline in the rate of ESI and an increase in public coverage. As a result there was no significant change in the uninsured rate among low-income people. But because of the increase in the size of the low-income population, the number of uninsured, low-income Americans increased by 1.4 million.
Among low-income adults there was a decline in the rate of ESI from 28.7% to 27.7%. Changes in both the Medicaid rate and uninsured rate were not statistically significant. But because of the large increase in the number of low-income adults (3.8 million), more adults gained coverage through Medicaid. However, this gain was not sufficient to offset the drop in ESI and the number of uninsured low-income adults increased by 1.8 million. There were slight declines in the number of uninsured among middle and higher income adults (not statistically significant).
Among children, all of the increase in public coverage was among those with incomes below 200% poverty. The percent of low-income children with public coverage increased from 47.7% to 50.0%, with the number increasing by 1.4 million. The decline in ESI among low-income children was not statistically significant. The large increase in public coverage resulted in a sharp decline in the uninsured rate among this group. The result was a decline in low-income uninsured children of 400,000. There were small declines in the number of uninsured children in the middle and higher income groups as well.
Changes in Health Insurance Coverage among Adults by Age
The increase in the number of uninsured occurred among adults in all age groups (Table 2). Young adults (19 to 34 year olds) had a large decline in the share covered by ESI, from 55.8% to 54.4%. Their Medicaid coverage increased from 8.1% to 8.9% but was not sufficient to offset the decline in employer coverage. As a result their uninsured rate increased, resulting in an increase of 600,000 young adults without coverage.
The same pattern was true for adults, age 35-54. While the changes in coverage rates were smaller than those among young adults, there was a decline in ESI and an increase in public coverage. There was also a significant decline in private non-group coverage. The result was an increase in the uninsured rate from 16.9% to 17.7% and an increase of 600,000 uninsured adults in this age range.
Among older adults (55 to 64 years old) there were no significant changes in the uninsured rate or any kind of coverage. But the number of adults in this age group increased by 1.0 million reflecting the continuing impact of the aging of the baby boom population. Because of the increase in population, the number of uninsured 55 to 64 year olds increased by 300,000.
Changes in Health Insurance Coverage by Race/Ethnicity
All of the increases in the number of uninsured occurred among the majority white population (Table 3). White Americans saw a large decline in ESI which was partially offset by increases in Medicaid and CHIP. Their uninsured rate increased from 12.2% to 12.7%. The number of white uninsured Americans increased by 800,000. All of this increase was among those with incomes under 200% FPL. The number of low-income white Americans grew by 2.6 million and the rate of ESI declined by 1.2 percentage points (data not shown). As a result, the number of uninsured in this group increased by 1.1 million.
Black Americans saw a decline in ESI but this was more than offset by increases in public coverage. This is in large part because blacks are twice as likely as whites to have low incomes and so are far more likely to be eligible for public insurance programs (data not shown). More Hispanics gained public coverage in 2008 as well, though their ESI rate did not change. Public coverage increased from 20.4% to 22.0% and as a result, the uninsured rate among Hispanics declined by 1.4 percentage points. However, because of growth in the Hispanic population, the number of uninsured Hispanics declined by only 200,000. Among other racial and ethnic groups there were no significant changes in health insurance coverage, but because of an increase in the size of this mixed subpopulation, the number of uninsured grew by 200,000.
Changes in Health Insurance Coverage by Citizenship Status
All of the increase in the number of uninsured occurred among native-born U.S. citizens (Table 4). These Americans experienced a decline in their ESI rate of 1.1 percentage points that was offset by an increased rate of public coverage, resulting in no change in their uninsured rate. However, the native-born population grew by 1.9 million. Thus, the number of uninsured among native-born citizens increased by 800,000. All of this increase in the uninsured occurred among those with incomes below 200% poverty (data not shown). As with the nonelderly population as a whole, the number of uninsured native-born adults increased and the number of uninsured native-born children fell (data not shown). There were no significant changes in coverage among naturalized citizens. In contrast, the size of the non-citizen population actually declined by 900,000 between 2007 and 2008, with an accompanying decrease in the number of non-citizens with ESI (600,000 fewer) and a slight decrease in the number of uninsured non-citizens (200,000 fewer, but not statistically different).
Changes in Health Insurance Coverage by Region
All regions, except the Northeast, saw declines in ESI that were at least partially offset by increases in public coverage (Table 5).5 The Midwest had the largest drop in ESI, from 68.7% to 67.1%. Much of this was offset by an increase in public coverage and there was no significant change in the Midwest’s uninsured rate. In the South, the rate of ESI declined from 59.2% to 58.4%. This was also offset by an increase in public coverage. A similar picture emerged in the West. The exception was a significant decline in private, non-group coverage resulting in a higher uninsured rate. As a result, the West saw the largest increase in the number of people without insurance (500,000).
The picture, not surprisingly, was different between adults and children (Table 5). Among children, there was an increase in public coverage in every region that was substantially greater than the change in ESI. There was a significant decline in the number of uninsured children in the Northeast and South. Among adults there was a statistically significant increase in the uninsured rate in the Northeast and West. There were declines in ESI in the Midwest and South and a significant decline in non-group coverage in the West. These were not fully offset by increases in public coverage. The result was that the number of uninsured adults increased in all regions but the Midwest.
.
Changes in Health Insurance Coverage among Workers
Tables 6 and 7 provide data on changes in health coverage among workers, age 18-64. Between 2007 and 2008, ESI declined among workers from 71.1% to 70.3%. There was a small increase in public coverage that was not sufficient to offset the decline in employer-coverage. As a result the uninsured rate among workers increased from 18.2% to 18.8% and 900,000 workers lost coverage. Some of this decline among workers was due to changes in employment, with people moving from higher to lower wage jobs that were much less likely to provide health insurance. Thus, not only did the rate of ESI for workers with family incomes below 200% FPL decline but the number of people affected by this decline increased (data not shown).
Among workers in low-ESI industries (industries where ESI covered less than 80% of employees in 2007), such as agriculture, construction, and services, there was a decline of ESI from 64.3% to 63.3% and a small increase in public coverage (Table 6). Their uninsured rate grew from 22.8% to 23.6% resulting in an increase of 900,000 workers without insurance in low-ESI industries. There were no changes in ESI or public coverage among workers in high-ESI industries (industries where 80% or more of employees are covered by ESI). There was, however, a decline in the number of workers in these industries. The shift from high to low ESI industries partially helps explain the increase in the number of uninsured workers.6
We also saw changes in coverage rates by firm size (Table 7). Workers in small firms or the self-employed were likely to see a decline in ESI. Their ESI rate fell from 52.0% to 51.1%. As a result, the uninsured rate increased from 30.1% to 31.1%, creating an increase of 600,000 uninsured workers. There were no coverage changes among those in medium-sized firms aside from a small increase in public coverage; there was however a decline in the number of workers in such firms. Surprisingly, given that large firms typically have higher and more stable rates of ESI compared to smaller firms, we found a decline in the rate of ESI from 82.2% to 81.0% among workers in large firms. This was not offset by an increase in public coverage. Thus, the uninsured rate increased from 10.7% to 11.5%, resulting in an increase of 500,000 workers in large firms without coverage.
Changes in Health Insurance Coverage Likely to Persist with the Deepening Recession
These results show that between 2007 and 2008, adults experienced a sharp decline in ESI and an increase in the number of uninsured. While children also saw a decline in ESI, albeit smaller than the decline among adults, they experienced a large increase in public coverage reflecting recent coverage expansions, and an actual decline in the number of uninsured. Modeling the changes in coverage that are likely to occur between 2008 and 2009 because of higher unemployment rates suggests that ESI rates will continue to decline.
Research done by Garrett together with other Urban Institute colleagues for the Kaiser Commission on Medicaid and the Uninsured showed that a 1 percentage point increase in the unemployment rate resulted in the number of children covered by ESI declining by almost one percentage point (0.95 percentage points) and enrollment in Medicaid and CHIP increasing by 0.79 percentage points.7 They found no statistically significant increase in the number of uninsured children. These results reflected the fact that broader coverage through Medicaid and CHIP can help to offset most of the decline in ESI among children.
For adults, the results were different. A one percentage point rise in the unemployment rate reduced the share of adults with ESI by 0.9 percentage points, increased the share with Medicaid by 0.2 percentage points and increased the share who were uninsured by 0.6 percentage points. Because there is much less Medicaid coverage for adults, the number of uninsured was predicted to increase in response to a decline in ESI. The Garrett model predicted the changes in coverage for adults in response to increasing unemployment reasonably well (at 6.2 percent unemployment in August 2008)8. This was less true for children; their coverage changes, including the reduction in the number of uninsured, were largely affected by Medicaid and CHIP expansions which more than offset the impact of unemployment on employer-sponsored coverage.
Between July 2008 and July 2009 the unemployment rate increased from 5.8% to 9.4%. This suggests that this year we will see ESI decline by about three percentage points. Some of this will be offset by increases in Medicaid and CHIP enrollment. However whether states can afford to increase enrollment to the extent seen in 2008, particularly for children, is unclear. States revenues have been sharply affected by the economic downturn; on the other hand, the American Recovery and Reinvestment Act provided substantial funding to states and required maintenance of eligibility standards. Thus, we would expect some increase in enrollment and perhaps more than is predicted by the model.
A three percentage point increase in the unemployment rate could likely result in an increase of about two percentage points in the uninsured rate among adults, according to the model. Thus we would expect to see a rather sharp increase in the number of uninsured adults in the coming year, somewhere between three and four million. How the number of uninsured children will respond to the decline in employer-sponsored insurance and the availability of public coverage is more difficult to predict.
Conclusion
As the incomes of Americans declined in the first year of what is proving to be a deep recession, the number of uninsured in the U.S. increased by 700,000 between 2007 and 2008. There were substantial declines in employer-sponsored insurance, particularly among adults. Because children are substantially more likely to be eligible for public coverage than adults, there was a sharp increase in Medicaid and CHIP enrollment among children that more than offset the decline in employer-sponsored insurance. In contrast, while there was some increase in public coverage of adults, it was much less than the decline in employer-sponsored insurance, and the number of uninsured adults increased by 1.5 million.
Almost the entire drop in health insurance coverage was among native-born Americans and among white Americans. The increased uninsurance rate among adults was spread across all age groups. The number of uninsured adults increased in all regions of the country except the Midwest. Every region saw increases in public coverage for children and the Northeast and South saw declines in the number of uninsured children. There were declines in coverage of workers in both small and large firms as well as in industries such as agriculture, construction, and services that already have low rates of employer coverage. This is further exacerbated by the continuing trend towards greater employment in industries where ESI covers fewer of its workers.
Monthly unemployment rates have continued to rise in 2009. Thus, there is every reason to expect sharp declines in ESI in 2009. How well states are able to compensate by increasing public coverage is, at this point, unknown. But it is highly likely that we will see a sharp increase in the number of uninsured Americans when 2009 data are released.
