The Economic Downturn and Changes in Health Insurance Coverage, John Holahan & Arunabh Ghosh The Urban Institute September 2004

Transcription

1 The Economic Downturn and Changes in Health Insurance Coverage, John Holahan & Arunabh Ghosh The Urban Institute September 2004 Introduction On August 26, 2004 the Census released data on changes in coverage between 2002 and 2003 showing that the number of uninsured Americans increased to 45 million. This was an increase of 1.4 million non-elderly uninsured between these two years and 5.1 million since In this paper we begin by reviewing the data in more detail than provided in the Census report, indicating some major trends that underlie the growth in the number of uninsured Americans. Most of the paper is devoted to exploring the changes that occurred over a longer period years in which the U.S. economy has been in a serious economic slowdown. The results show how insurance coverage has changed in response to the decline in employer sponsored insurance rates as well as changes in both employment and the distribution of income. The principal findings of the paper are that the number of uninsured grew in both the last year and over the last three years because of continuing declines in employer sponsored insurance. For children this decline was more than offset by increases in enrollment in Medicaid and the State Children s Health Insurance Program (SCHIP), and as a result the number of children without coverage declined. The same growth in public coverage did not occur for adults and as a result all of the increase in the uninsured was among adults. In addition, we show that changes in coverage were affected by both job losses and changes in the distribution of income. The number of people below 200% of

2 poverty, particularly below the poverty line, increased significantly and the number of middle income Americans declined. Thus, individuals moved into economic circumstances in which the likelihood of employer sponsored insurance was lower, and consequently the likelihood of being uninsured increased. While incomes and coverage declined between , the declines were even more pronounced over the longer period. In addition to the different outcomes for children and adults we find that three quarters of the growth in the uninsured between 2000 and 2003 was among low income Americans, i.e., those below 200% of the federal poverty line. 1 This means of course that about 25% of this growth was among those with incomes above 200% of poverty. Further, about 60% of the growth in the uninsured occurred among young adults, those aged 19-34, about 60% occurred among whites, and 70% among native citizens. Over half of the growth in the uninsured was in southern states. Finally, we document that while part of the reason for the decline in employer sponsored insurance was that a smaller share of the population was working, there was also a shift in the locus of employment - - from large and medium sized firms to small firms, and from industries that have historically had high rates of employer coverage to those with much lower rates of coverage. 2 1 The federal poverty level for a family of four in 2003was $18,810; an income of < $37,620 would place them below 200% of the federal poverty line. 2 The March CPS does not allow analysis of how much of the change in employer sponsored coverage was due to changes in offers or take-up rates. Analysis of change in employer coverage using National Survey of America s Families data for 1999 and 2002 showed that much of the decline was due to lower take up rates. Linda Blumberg and John Holahan, Decomposing the Recent Declines in Employer Sponsored Coverage, Health Policy Online, The Urban Institute, May 17,

3 Data and Methods In this paper we use data from the 2001, 2003, and 2004 March supplements to the Current Population Survey (now called the Annual Social and Economic Supplement). The Current Population Survey (CPS) is the most frequently cited national survey on health insurance of Americans. Its strengths and weaknesses have been widely documented. 3 There is debate over whether the CPS is measuring the number of uninsured for an entire year (as intended) or whether responses more closely reflect the number of uninsured at a point-in-time. In this paper we assume that the CPS is essentially a measurement of point-in-time coverage, primarily because the number of uninsured in the CPS has historically been closer to point-in-time estimates and well above the full year estimates of other surveys. 4 There is also an issue of whether the CPS understates Medicaid enrollment and thus, possibly overstates the number of uninsured. We believe that there is no solid evidence on how great a problem this is and further while a Medicaid undercount may be increasing the CPS uninsurance rate, it may also be overstating private coverage. 5 But despite those issues, the CPS provides the best available measure of changes in coverage over time. Trends in the CPS are relatively reliable because modifications to the survey have been infrequent. This paper is focused almost entirely on the trends in coverage and much less on the actual numbers of uninsured. 3 John Holahan, Genevieve Kenney and Len Nichols, Towards a Federal Survey of Health Insurance Coverage and Access, Paper prepared for Academy Health Conference on Federal Data Colleciton, May 2004; Kimball Lewis, Marilyn Ellwood and John L. Czajka, Counting the Uninsured: A Review of the Literature, Washington, D.C.: The Urban Institute. Assessing the New Federalism Occasional Paper No U.S. Congressional Budget Office, How Many People Lack Health Insurance and for How Long. 5 Kathleen Thiede Call et al., Uncovering the Missing Medicaid Cases and Assessing Their Bias for Estimates of the Uninsured, Inquiry, Vol. 38, No. 4 (Winter 2001/2002), pp

4 The CPS allows respondents to report multiple types of health insurance; in this analysis their responses are classified in a hierarchy with each respondent assigned only one type of coverage the type of coverage that is highest in the hierarchy. The hierarchy, in declining order, is as follows: employer coverage; Medicaid, SCHIP or state coverage (it is not possible to distinguish SCHIP from Medicaid coverage in the CPS); military, veterans or Medicare; private non group coverage, and uninsured. The annual Census reports that provide data on health insurance use household income. In this paper we use the income of the health insurance unit. The health insurance unit includes members of the nuclear family who can be covered under one health insurance policy (i.e., policyholder, spouse, children under age 19, and full time students under age 23). Counting the income of all members of the household can overstate income because it includes the income of all relatives and unrelated people living together. The income of the health insurance unit more accurately reflects the income available to individuals when purchasing private insurance or determining eligibility for public programs. In this paper we present data for three income groups using poverty thresholds. The intent is to focus analysis of the impacts on low income, middle income, and higher income Americans. The three income groups are those with incomes below 200% of the federal poverty level (FPL), 200% FPL-399% FPL, and 400% FPL and higher, each of which represented about one-third of the U.S. non-elderly population in the year An additional advantage of using poverty thresholds is that they adjust for family size and 4

5 inflation. The household income categories used by the Census Bureau do not adjust for either and have led to some misinterpretation of data in the Census Reports. 6 Changes in Coverage, As noted, the number of uninsured increased between 2002 and 2003 by 1.4 million. The number of uninsured adults actually increased by more than 1.5 million, while the number of children declined by about 160,000. The number of uninsured children declined because of the strong growth in public coverage that more than offset the decline in employer-sponsored insurance. There was a decline in employersponsored insurance for adults but little change in public coverage. As a result, the share of adults without insurance increased and, combined with the growth in the adult population, led to an increase in the number of uninsured adults of 1.5 million. Figure 1 shows the changes in employment and incomes that occurred in the past year. The number of individuals in families with Figure 1 Changes in Population, Employment and Incomes, Millions Population Employment Incomes 1.9* 0.3* 0.7* 1.9* 2.2* 0.4* 0.6* two full time workers * -1.5* declined by 1.1 million. There were gains in the Population Two Worker Families One Worker Families Families with Part Time Worker Families with No Worker Income <100% FPL % FPL % FPL 400%+ FPL number of Americans living in households with one full time worker (by 300,000) and in households with only part time workers (700,000). These gains just about offset the decline in the number of 6 John Holahan, Catherine Hoffman and Marie Wang, The New Middle Class of Uninsured Americans Is It Real? Kaiser Commission on Medicaid and the Uninsured, March

6 individuals living in households with two full time workers. The number of people living in households with no workers increased by 1.9 million, about the same as the population increased. In other words, there was little change in the number of Americans who were working (though there was a decline in working Americans as a share of the population), but there was shift from two worker households to part time workers and non working households. Figure 1 (right side) also shows that there was a decline in the number of people we regard as middle income Americans. The number of people between 200% and 400% of poverty fell by 1.5 million people. At the same time, there was a modest increase in the number of people above 400% of poverty (600,000), and a sharp increase in the number below 200% of poverty (2.7 million). Of this 2.7 million person increase in the low income population, 2.2 million were below 100% of the federal poverty line. Why did children hold their own? Table A1 shows that there was a small decline in the number of uninsured children because of the strong growth in public coverage. Children fared poorly in terms of economic circumstances -- there was a large decline in the number of children (.8 million) between 200%-399% of the federal poverty line and a large increase (1.1 million) in the number below 200% of poverty (all of the increase was below the poverty line, Table A1). Despite the growth in the number of low income children, there was no increase in the number who were uninsured. While there was a sharp drop in the number with employer coverage, there was an even larger increase in the number with public coverage. The bulk of the growth in public coverage seemed to occur below the federal poverty line (1.1 million of the total 1.6 million increase in children with public 6

7 coverage Table A1), suggesting that most of the growth was through Medicaid (through additions to the rolls not because of expansion of eligibility standards) rather than the State Children s Health Insurance Program (SCHIP). But the availability of coverage in both programs meant that the decline in employer-sponsored insurance that occurred for children did not increase the number of children without coverage. Why did adults do so poorly? Table A1 also shows that the picture was quite different for adults. Adults also experienced declines in incomes with the number below 200% of poverty increasing by 1.6 million between 2002 and 2003, more than the overall increase in the adult population (1.5 million). Low income adults experienced a large drop in employer coverage which was not offset by an increase in public coverage. As a result, the uninsured rate for low income adults increased by 1.5 percentage points. The increase in the uninsured rate, coupled with the increase in the number of low income adults, resulted in an increase of 1.4 million uninsured adults. Thus, practically all of the increase in the number of uninsured adults occurred among those with incomes below 200% of poverty. This does not mean that middle income adults were unaffected; rather they were affected as more middle income adults became low income adults due to increased unemployment and declining incomes. More than half (900,000) of the growth in uninsured adults occurred among year olds all of whom had incomes below 200% of the federal poverty line (Table A2). For these young adults there was a sharp decline in employer coverage that was only partially offset by an increase in public coverage. The result again was a large increase (2.3 percentage points) in the uninsured rate which led to the growth of 900,000 7

8 uninsured low income young adults. Of the overall increase in the number of uninsured adults, about two-thirds occurred among childless adults, and one-third among parents (Table A3) and about 60% was among women and 40% among men (Table A4). Three other results for the period are of interest. First, the majority of the increase in the uninsured occurred among whites (Table A5). Whites, blacks and Hispanics all experienced significant reductions in rates of employer sponsored coverage. Each also had increases in public coverage. But increases in public coverage for whites were smaller, and as a result the uninsurance rate for whites increased; in all, 800,000 white Americans lost coverage. Second, slightly more than half (800,000 persons) of the increase in uninsured Americans occurred in the south (Table A6). The U.S. population continued to increase in the south and west and remain stagnant or fell in the northeast and midwest. One implication of these population shifts is that people moved to areas of the nation with lower rates of employer sponsored insurance and higher uninsured rates. Finally, there was a continuation of recent trends in the shift of work from large firms to self-employment or work in small firms (Table A7). The latter have lower rates of employer sponsored insurance and higher uninsured rates. The longer period 2000 to 2003 In the remainder of this paper we turn to the data showing changes between This is a period in which the Bureau of Labor Statistics establishment survey showed a decline of 3.1 million jobs, though technically the nation was only in a recession from March 2001 until November This period is instructive because it shows how susceptible the U.S. health insurance system is to economic declines. 7 National Bureau of Economic Research, Report of the Business Cycle Dating Committee, Cambridge, MA, July 17,

9 Figure 2 presents data on the changes in economic circumstances in two different ways. The figure shows that, first, the population grew by 7.6 million people. Second, the number of people in families with two full time workers declined by 3.7 million. The number in families with one full time worker or a part time worker increased by 3.8 million and 2.2 million Figure 2 Changes in Population, Employment and Incomes, Population Employment Incomes Millions Population 7.6* -3.7* 3.8* Two Worker Families One Worker Families Families with Part Time Worker Families with No Worker 2.2* Income 5.4* 5.6* <100% FPL % FPL % FPL 400%+ FPL 2.1* -1.9* 1.8* respectively. Thus, while there was an increase of 2.3 million individuals in working households over this period there was a shift toward families with one or only part-time workers. But there was also an increase of 5.4 million individuals in households with no worker. These shifts are important because rates of employer sponsored insurance decline and uninsurance rates increase with decreasing numbers of workers in the households. Third, Figure 2 shows changes in the distribution of income since The number of people below the federal poverty line increased by 5.6 million while the number between 100%-199% of the federal poverty line increased by 2.1 million. In contrast, the number of people in the middle income group (200%-399% of the federal poverty line) fell by 1.9 million. The number of people in the highest income category (above 400% of poverty) increased by 1.8 million. Thus, there was a decline in the number of individuals in what one might regard as middle income and a substantial 9

10 increase in the number of people below 200% of poverty, particularly below the federal poverty line. These economic trends, as we will show below, had important implications for changes in the pattern of health insurance coverage. Table 1 provides detailed data on the changes in coverage between 2000 and 2003 for the non elderly as a whole, and separately for adults, and children. For the non elderly as a whole, employer-sponsored coverage fell by 3.9 percentage points, from 67.8% to 63.9%. Medicaid and state coverage increased by 2.0 percentage points, not enough to offset the decline in employer coverage. As a result the uninsured rate increased from 16.1% to 17.7% and the number of uninsured increased by 5.1 million. All of this was due to increases in the number of uninsured adults; the number of uninsured children actually fell slightly (Figure 3). Both children and adults had substantial losses in employer sponsored insurance. But children had large increases in public coverage while adults did not. Thus, the uninsured 6% 5% 4% 3% 2% 1% 0% -1% -2% -3% -4% -5% -6% Change in Population Change in Uninsured Figure 3 Changes in Health Insurance Coverage, Children vs. Adults, Percentage Point Changes Non-Elderly Children Adults 2.0%* 1.5%* -3.9%* -4.3%* 4.8%* -0.5%* -3.8%* 0.8%* 7.6 Million 1.3 Million 6.3 Million 5.1 Million -0.3 Million 5.4 Million 2.4%* Employer Medicaid Uninsured * Statistically significant change between 2000 and 2003 (p<.10). Medicaid also includes S-CHIP, other state programs. Figure excludes changes in Champus /Medicare and private non group insurance for ease of presentation because changes in coverage are generally small. Source: Urban Institute rate for adults increased by 2.4 percentage points but declined by.05 percentage points for children. 10

11 Table 1 Health Insurance Coverage, Nonelderly by Age and Health Insurance Unit Income NONELDERLY ADULTS CHILDREN Coverage Change Coverage Change Coverage Change Distribution within in Millions of people Distribution within in Millions of people Distribution within in Millions of people Income Category and Percentage Income Category and Percentage Income Category and Percentage All Incomes (millions of people) a a Employer 67.8% 63.9% -3.9% * a 68.9% 65.1% -3.8% * a 65.4% 61.1% -4.3% * a Medicaid and State 8.8% 10.9% 2.0% * 5.77 a 5.3% 6.1% 0.8% * 1.79 a 16.7% 21.5% 4.8% * 3.98 a CHAMPUS/Medicare 2.1% 2.3% 0.1% * 0.50 a 2.3% 2.7% 0.4% * 0.83 a 1.7% 1.3% -0.4% * a Private Nongroup 5.1% 5.3% 0.2% * 1.02 a 5.6% 5.8% 0.2% 0.63 a 3.9% 4.3% 0.4% * 0.39 a Uninsured 16.1% 17.7% 1.5% * 5.12 a 17.9% 20.3% 2.4% * 5.38 a 12.3% 11.8% -0.5% * Less than 200% of FPL a a a Employer 34.7% 30.3% -4.4% * a 33.9% 29.9% -4.0% * % 30.9% -5.2% * a Medicaid and State 24.0% 27.3% 3.3% * 4.79 a 16.1% 17.3% 1.2% * 1.58 a 36.5% 44.0% 7.5% * 3.22 a CHAMPUS/Medicare 3.5% 3.5% 0.0% 0.24 b 4.4% 4.7% 0.3% 0.40 a 2.1% 1.5% -0.6% * a Private Nongroup 6.3% 5.8% -0.4% * % 7.3% -0.7% * % 3.4% -0.1% 0.03 Uninsured 31.5% 33.1% 1.6% * 3.85 a 37.6% 40.8% 3.2% * 3.94 a 21.9% 20.3% -1.6% * to 399% of FPL a a Employer 77.5% 74.5% -3.0% * a 75.9% 72.7% -3.1% * a 81.1% 78.5% -2.6% * a Medicaid and State 2.5% 3.7% 1.2% * 0.85 a 1.5% 1.9% 0.4% * 0.21 a 4.7% 7.8% 3.1% * 0.64 a CHAMPUS/Medicare 1.9% 2.1% 0.2% % 2.4% 0.5% * 0.24 a 1.9% 1.4% -0.5% * a Private Nongroup 5.1% 5.4% 0.3% % 5.6% 0.2% % 4.8% 0.4% 0.04 Uninsured 12.9% 14.3% 1.3% * 0.72 a 15.3% 17.3% 2.1% * 0.95 a 7.9% 7.4% -0.5% a 400% of FPL and above a a b Employer 89.5% 87.9% -1.6% * % 87.4% -1.6% * % 89.3% -1.9% * 0.13 Medicaid and State 0.5% 0.6% 0.1% * 0.13 a 0.4% 0.4% 0.0% % 1.3% 0.5% * 0.12 a CHAMPUS/Medicare 1.0% 1.2% 0.2% # 0.16 a 1.1% 1.3% 0.3% * 0.19 a 1.0% 0.9% -0.1% Private Nongroup 4.0% 4.8% 0.8% * 0.84 a 4.0% 4.7% 0.7% * 0.52 a 3.9% 5.3% 1.3% * 0.31 a Uninsured 5.0% 5.5% 0.5% * 0.55 a 5.6% 6.2% 0.6% * 0.50 a 3.1% 3.2% 0.2% 0.06 Source: Urban Institute, Based on data from March Current Population Surveys, 2001, Note: Excludes persons aged 65 and older and those in the Armed Forces. * Indicates change in percent of people is statistically significant (at the 95% confidence level). # Indicates change in percent of people is statistically significant (at the 90% confidence level). a Indicates change in numbers of people is statistically significant (at the 95% confidence level). b Indicates change in numbers of people is statistically significant (at the 90% confidence level).

12 Medicaid and SCHIP increase coverage of children Children fared much better than adults (Table 2) because the increase in Medicaid and SCHIP coverage (4.8 percentage points) more than offset the decline in employersponsored insurance (4.3 percentage points). 8 Much of the increase in public coverage occurred among low income children (3.2 million), but there was also an increase in public coverage among middle income children (860,000). 9 Table 2 and Figure 4 show the data on low and middle income children in with more detailed income breaks. There was a large increase in the number of low income children (2.0 million), particularly the Change in Population Change in Uninsured Figure 4 Health Insurance Coverage of Children, , By Income Percentage Point Changes <100% FPL % FPL % FPL 400% FPL+ 7.5* -3.5* -3.5* -6.1* 6.7* * 3.1* 0.4* -1.9* 1.3* 0.5* Million 0.5 Million -1.3 Million 0.6 Million -0.2 Million 0.1 Million -0.2 Million 0.1 Million Employer Sponsored Medicaid Private/ Non Group Uninsured * Statistically significant change between 2000 and 2002 (p<.10). Medicaid also includes S- CHIP, other state programs. Figure excludes changes in Champus/Medicare and private non group insurance for ease of presentation because changes in coverage are generally small. Source: Urban Institute number of children in poverty (1.5 million). At the same time there was a decline of 1.3 million in the number of middle income children. Children in families below the poverty line saw a decline in the rate of employer-sponsored insurance, from 21.2% to 17.7%. This was more than offset by a 7.5 percentage point increase in the rate of public coverage. As a result, the share of poor children without insurance declined by 3.5 percentage points or by 200,000 children. The decline in the number of uninsured 8 In total (without the hierarchy), the number of children with Medicaid or SCHIP coverage increased from 15.6 million to 20.0 million, or by 28.1%. 9 Without the hierarchy, these increases were 3.4 million and 1.0 million respectively.

13 Table 2 Health Insurance Coverage, Children by Health Insurance Unit Income (upto 400% FPL) CHILDREN Coverage Distribution within Change in Millions of people Income Category and Percentage Less than 100% of FPL (millions of people) a Employer 21.2% 17.7% -3.5% * -0.3 b Medicaid and State 47.6% 55.1% 7.5% * 2.1 a CHAMPUS/Medicare 1.7% 1.4% -0.3% 0.0 Private Nongroup 3.2% 3.1% -0.2% 0.0 Uninsured 26.3% 22.8% -3.5% * % of FPL b Employer 52.1% 46.0% -6.1% * -0.7 a Medicaid and State 24.7% 31.4% 6.7% * 1.2 a CHAMPUS/Medicare 2.4% 1.5% -0.9% * -0.1 a Private Nongroup 3.7% 3.7% -0.1% 0.0 Uninsured 17.1% 17.4% 0.3% to 299% of FPL b Employer 76.7% 74.0% -2.7% * -0.7 a Medicaid and State 6.4% 10.9% 4.5% * 0.5 a CHAMPUS/Medicare 2.3% 1.4% -0.9% * -0.1 a Private Nongroup 4.9% 4.9% 0.0% 0.0 Uninsured 9.7% 8.8% -0.9% -0.2 a 300 to 399% of FPL a Employer 86.6% 84.6% -2.0% * -0.9 a Medicaid and State 2.5% 3.7% 1.2% * 0.1 a CHAMPUS/Medicare 1.4% 1.3% -0.1% 0.0 Private Nongroup 3.7% 4.7% 1.0% * 0.1 Uninsured 5.7% 5.6% -0.1% -0.1 Source: Urban Institute, Based on data from March Current Population Surveys, 2003, Note: Excludes persons aged 65 and older and those in the Armed Forces. * Indicates change in percent of people is statistically significant (at the 95% confidence level). # Indicates change in percent of people is statistically significant (at the 90% confidence level). a Indicates change in numbers of people is statistically significant (at the 95% level). b Indicates change in numbers of people is statistically significant (at the 90% level). 12

14 children would have been significantly greater had there not been the large increase in the number of children below the federal poverty line. The decline would also have been greater if all eligible children were enrolled, i.e., virtually all of the uninsured children (4.1 million) were eligible. Among near poor children (100%-199% FPL) there was an even larger decline in employer-sponsored insurance, from 52.1% to 46.0%. Again, there was an increase in public coverage of 6.7 percentage points and no statistically significant change in the percentage of children without coverage. In the income bracket between 200%-399% FPL there was also a decline in employer-sponsored insurance, but even here there was an increase in public coverage of children from 4.7% to 7.8%, most likely because of the State Children s Health Insurance Program. As a result children in this income bracket also did not experience a significant decline in coverage. Among higher income children (those in families with incomes above 400% of FPL) there was a sharp increase in the rate of non group coverage which helped to offset the decline in the rate of employer sponsored insurance. As a result there was only a slight increase in the number of high income children without coverage. Loss of employer coverage increases number of uninsured adults The rate of employer sponsored insurance among adults fell between 2000 and 2003 from 68.9% to 65.1% (Table 1). There was a small increase in public coverage but nowhere near enough to offset the decline in employer coverage and the increase in the population. As a result the uninsured rate increased by 2.4 percentage points, or by 5.4 million adult Americans. About three quarters of the growth in the uninsured adults occurred among those below 200% of poverty (Figure 5). The rate of employer- 13

15 sponsored insurance of low income adults declined by 4.0 percentage points while public coverage increased by only 1.2 percentage points. Low income adults also saw a drop in private nongroup coverage. The Figure 5 Health Insurance Coverage Of Adults, By Income , Percentage Point Changes 1.2* <200% FPL % FPL 400% FPL+ -0.7* 3.2* 0.4* * * -5 Change in 5.7 Million -0.6 Million Population Change in Uninsured 3.9 Million 1.0 Million 2.1* -1.6* 0.7* 0.6* Million 0.5 Million * Statistically significant change between 2000 and 2003 (p<.10). Medicaid also includes S- CHIP, other state programs. Figure excludes changes in Champus/Medicare and private non group insurance for ease of presentation because changes in coverage are generally small. Employer Sponsored Medicaid Private/ Non Group Uninsured Source: Urban Institute result was an increase in the uninsurance rate from 37.6% to 40.8% or 3.9 million uninsured low income adults. There were also increases in the uninsured among higher income adults. Here again, the main reason was the decline in employer-sponsored insurance which was only partially offset by small increases in public or private non group coverage. The uninsured rate increased among middle income adults from 15.3% to 17.3% and among higher income adults from 5.6% to 6.2%. Thus, there was an increase of 1.5 million uninsured adults with incomes above 200% of poverty. A question which often arises is whether the drop in employer sponsored insurance for children was the result of the expansion of public programs or whether the availability of the latter meant that an exogenous decline in employer coverage was less likely to mean a loss of coverage. The sizable drop in employer coverage for both low income and middle income adults, and for both parents and childless adults (see below), 14

16 suggest that the availability of Medicaid and SCHIP was not likely to have been the predominant cause of the decline in employer coverage for children over this period. 10 Declines in adult coverage related to age and income To better understand the decline in coverage among adults, we examined changes in incomes as well as changes in insurance coverage among adults across different age groups. Figure 6 shows that the number of adults below 200% of poverty increased by 5.7 million, while the number of middle income adults fell and the number of higher income adults increased by 1.2 million. These shifts varied among Millions Change in Population Figure 6 Changes in Adults, By Age and Income, All Adults Ages Ages Ages * -0.6* 1.2* <200% FPL % FPL 400%+ FPL 2.7* -0.6* -0.9* 2.5* -0.4* -0.7* 0.5* 0.7* 2.5* 6.3 Million 1.3 Million 1.4 Million 3.7 Million Source: Urban Institute different age groups. Among those between the ages of 19 and 34 there was a large increase (2.7 million) in the number below 200% of poverty. The numbers of young adults in the two higher income groups both declined. Those in the age group had a similar a 2.5 million increase in the number with incomes below 200% of poverty with reductions in the numbers in the two higher income groups. For the oldest age group (ages 55-64), there was a small increase in the low income group, a slightly larger increase in the number of middle income near elderly and a sizable increase in the 10 There is clear evidence that expansions of Medicaid and SCHIP have resulted in some displacement of private coverage. Genevieve Kenney and Debbie I. Chang. "The State Children's Health Insurance Program: Successes, Shortcomings, and Challenges." Health Affairs. September/October Volume 23, Number 5. 15

17 Table 3 Health Insurance Coverage, Nonelderly Adults by Age and Health Insurance Unit Income ADULTS AGE ADULTS AGE ADULTS AGE Coverage Change Coverage Change Coverage Change Distribution within in Millions of people Distribution within in Millions of people Distribution within in Millions of people Income Category and Percentage Income Category and Percentage Income Category and Percentage All Incomes (millions of people) b b a Employer 61.7% 55.7% -6.0% * -2.9 a 74.9% 71.1% -3.9% * -2.2 a 66.6% 68.1% 1.5% * 2.9 a Medicaid and State 6.0% 7.6% 1.5% * 1.0 a 4.5% 5.3% 0.8% * 0.8 a 6.1% 5.3% -0.8% * 0.0 CHAMPUS/Medicare 1.2% 1.4% 0.1% % 2.3% 0.4% * 0.3 a 6.3% 6.9% 0.6% # 0.4 a Private Nongroup 6.3% 6.5% 0.3% 0.2 b 4.6% 4.9% 0.3% # 0.3 a 7.4% 6.6% -0.7% * 0.1 Uninsured 24.7% 28.8% 4.1% * 2.9 a 14.1% 16.4% 2.4% * 2.2 a 13.6% 13.0% -0.6% 0.3 a Less than 200% of FPL a a a Employer 34.0% 28.2% -5.8% * -0.7 a 34.7% 31.7% -2.9% * % 31.2% -0.1% 0.2 Medicaid and State 13.5% 15.6% 2.1% * 1.0 a 18.3% 19.1% 0.8% 0.6 a 19.4% 18.1% -1.3% 0.0 CHAMPUS/Medicare 1.8% 1.8% -0.1% % 5.1% 0.1% 0.1 a 12.7% 14.9% 2.2% * 0.2 a Private Nongroup 8.9% 8.1% -0.8% # % 5.7% -0.8% * % 8.9% -0.2% 0.0 Uninsured 41.8% 46.3% 4.5% * 2.4 a 35.5% 38.4% 2.8% * 1.5 a 27.4% 26.8% -0.6% to 399% of FPL a a Employer 73.5% 69.0% -4.5% * -1.5 a 79.2% 76.2% -3.0% * -1.1 a 70.5% 70.9% 0.5% 0.5 a Medicaid and State 1.5% 1.9% 0.4% # % 1.9% 0.5% * 0.1 a 2.1% 2.2% 0.2% 0.0 CHAMPUS/Medicare 1.0% 1.4% 0.4% * 0.1 b 1.5% 1.9% 0.4% # 0.1 b 6.2% 6.9% 0.8% 0.1 a Private Nongroup 5.4% 6.2% 0.8% * % 4.9% 0.1% % 6.5% -1.7% * -0.1 Uninsured 18.6% 21.6% 3.0% * 0.4 a 13.2% 15.2% 2.0% * 0.4 a 13.1% 13.4% 0.3% % of FPL and above a b a Employer 88.4% 86.8% -1.6% * -0.8 a 90.4% 88.4% -2.0% * -1.4 a 84.7% 85.3% 0.7% 2.2 a Medicaid and State 0.5% 0.5% 0.1% % 0.3% 0.1% % 0.4% -0.3% * 0.0 CHAMPUS/Medicare 0.6% 0.6% 0.0% % 1.0% 0.3% * 0.1 a 2.7% 2.8% 0.1% 0.1 a Private Nongroup 3.5% 4.3% 0.8% * 0.1 b 3.6% 4.5% 0.9% * 0.3 a 5.9% 5.5% -0.3% 0.1 Uninsured 7.0% 7.7% 0.7% % 5.8% 0.8% * 0.3 a 6.0% 5.9% -0.1% 0.1 a Source: Urban Institute, Based on data from March Current Population Surveys, 2001, Note: Excludes persons aged 65 and older and those in the Armed Forces. * Indicates change in percent of people is statistically significant (at the 95% confidence level). # Indicates change in percent of people is statistically significant (at the 90% confidence level). a Indicates change in numbers of people is statistically significant (at the 95% confidence level). b Indicates change in numbers of people is statistically significant (at the 90% confidence level).

18 number of higher income near elderly. Overall, because of the baby boom, the age group increased by 3.7 million. Thus, the adult population shifted towards the age group more likely to have some form of insurance coverage, either private or public, and less likely to be uninsured. In the absence of such a shift in composition, the coverage picture would have been somewhat worse. More than half of the increase in the number of uninsured adults occurred among those between ages (Table 3, Figure 7). In this age group over 80% of the decline in coverage occurred among those below 200% of poverty. Low income young adults experienced a sharp decline in the rate of employer-sponsored insurance, some increase in public coverage but a large increase in the Change in Population Change in Uninsured Figure 7 Health Insurance Coverage Of Non-Elderly Adults, By Age, , Percentage Point Changes -6.0* Age Age Age * 4.1* -3.9* 0.8* 1.3 Million 1.4 Million 2.9 Million 2.2 Million 2.4* 1.5* -0.8* Million 0.3 Million * Statistically significant change between 2000 and 2002 (p<.10). Medicaid also includes S-CHIP, other state programs. Figure excludes changes in Champus/Medicare and private non group insurance for ease of presentation because changes in coverage are generally small. Employer Sponsored Medicaid Uninsured Source: Urban Institute uninsured rate (from 41.8% to 46.3%). The result was an increase of 2.4 million low income young adults without coverage. Young adults in the two higher income categories also experienced declines in employer sponsored insurance. The middle income group saw an increase in their uninsured rate of 3.0 percentage points which translated into an increase of 400,000 uninsured. Among middle aged Americans there was also a decline in employer-sponsored insurance and a sharp increase, from 14.1% to 16.4% in the uninsured rate. About two- 16

19 thirds of the increase in the uninsurance among middle aged Americans occurred among those below 200% of poverty, again largely because of the decline in rates of employersponsored insurance. But, higher income middle aged Americans also saw declines in the rate of employer-sponsored insurance as well as increases in their uninsured rates. Thus, there was an increase of 700,000 middle aged Americans above 200% of poverty without insurance. As noted earlier, the largest growth in population was among adults aged 55-64; this group also experienced substantial gains in incomes despite the slow economy. The near elderly also saw an increase in the likelihood of having employer-sponsored insurance of 1.5 percentage points. However there were declines in the rates of public coverage and private non group coverage. While there was no significant change in the uninsurance rate for this age group, there was an increase of 300,000 near elderly uninsured due to population growth. The increase in the number of uninsured in this age group was split among the three income categories. Childless adults affected differently than parents; males more than females Figure 8 shows the changes in coverage among adults in a different way - - focusing on the changes affecting parents and childless adults and males and females. Both parents and childless adults experienced sharp declines in rates of employersponsored insurance with parents experiencing a somewhat larger decline. But parents also experienced a larger increase in public coverage (1.3 million versus 600,000 for childless adults). The net result is that both experienced significant increases in their uninsurance rates. Because there are more childless adults than parents and because there was a larger increase in the number of childless adults over this period, the number of 17

20 childless adults who were uninsured grew by 3.8 million and uninsured parents by 1.6 million. Both males and females experienced sharp declines in the rates of employer-sponsored insurance; 4.1% percentage points and 3.5 percentage points respectively. The increase in public coverage was Change in Population Change in Uninsured Figure 8 Changes in Health Insurance Coverage Among Adults, Percentage Point Changes Childless Adults Parents Males Females 3 2.9* 2.5* 2.2* 1.9* 2 1.3* 0.6* 0.7* 1.0* 1-3.4* -4.2* -4.1* -3.5* 5.5 Million 0.8 Million 3.4 Million 2.9 Million 3.8 Million 1.6 Million 3.2 Million 2.2 Million Employer Sponsored Medicaid Uninsured * Statistically significant change between 2000 and 2002 (p<.10). Medicaid also includes S- CHIP, other state programs. Figure excludes changes in Champus/Medicare and private non group insurance for ease of presentation because changes in coverage are generally small. Source: Urban Institute somewhat less for men than women and as a result the former experienced a higher increase in the uninsurance rate (2.9 percentage points vs. 1.9 percentage points). Thus, the number of uninsured men increased by 3.2 million and women by 2.2 million. Change in patterns of work contributed to decline in employer sponsored coverage To further understand the changes in coverage of adults, we looked at changes affecting workers. Figure 9 shows changes in coverage by size of firm and Figure 10 depicts changes in types of industries. Industries are grouped by those with high and low rates of employer-sponsored insurance in Over this three-year period there was an increase of 200,000 Americans who were working, though there was a decline in the employment rate. This is in conflict with data from the Bureau of Labor Statistics 18

21 establishment survey that shows a large decline in the American workforce. 11 The establishment survey is generally regarded as more reliable for measuring employment trends because of its larger sample size. 12 The CPS, while having smaller samples of workers, does capture self employment and employment in new firms which is important for understanding changes in insurance coverage. Figure 9 Changes in Health Insurance Coverage of Workers, By Firm Size, , Percentage Point Changes Change in Workers Change in Uninsured All Workers 0.4* -2.8* 2.1* Self Employed -3.4* * 0.6* -4.2* 3.2* 0.2 Million 1.0 Million 2.2 Million 3.1 Million 0.6 Million Small Firms Less Than 25 Workers 1.6 Million -1.9* 0.5* 1.4* * * Statistically significant change between 2000 and 2002 (p<.10). Medicaid also includes S- CHIP, other state programs. Figure excludes changes in Champus/Medicare and private non group insurance for ease of presentation because changes in coverage are generally small. Large Firms Medium Firms 1000 or More Workers Workers 1.1* -0.9 Million -2.1 Million 0.5 Million 0.4 Million Employer Sponsored Medicaid Uninsured Source: Urban Institute According to the household survey, there was an increase of 200,000 new workers, as shown in Figure 9. But among workers there was an increase of 3.1 million without insurance. Among those working there was a decline of 2.8 percentage points in the rate of employer sponsored insurance which led to an increase in the uninsurance rate of 2.1 percentage points, and 3.1 million uninsured workers. This shows clearly that the increase in the uninsured was not simply among those who were no longer working, but that the decline in coverage also occurred among workers. Figure 9 shows that there was a decline of 2.1 million among workers in large firms and of.9 million in medium sized firms, but an increase of 1.0 million in self- 11 U.S. Department of Labor, Bureau of Labor Statistics, National Employment Hours and Earnings, January 1994 July 2004; U.S. Department of Labor, Bureau of Labor Statistics, Labor Force Statistics from the Current Population Survey: Employment Level, January 1994 July Kathleen Utgoff, Bureau of Labor Statistics, Statement before the Joint Economic Committee, September 5,

22 employment and 3.2 million in the number of in workers in small firms (less than 25 workers). Thus, there was a shift from medium and large sized firms that have relatively high rates in employer-sponsored insurance to self-employment and small firms where rates of employer-sponsored insurance are substantially lower. At the same time there was a sharp decline in the rate of employer-sponsored insurance in small firms and among the self-employed of 4.2 and 3.4 percentage points respectively, significantly larger than the drop in medium and large firms of 1.9 and 1.3 percentage points respectively. This led to an increase of 3.2 percentage points in the uninsurance rate of those working in small firms and 2.5 percentage points among the self-employed. Thus there was not simply a shift to self-employment and work in small firms but also a reduction in the likelihood of having employer-sponsored insurance and an increase in the likelihood of being uninsured. The net result is that 2.2 million of the 3.1 million increase in the number of uninsured workers occurred among the self-employed and those in small firms. Because of declines in employer sponsored in medium and large firms, there was also an increase in the number of uninsured in these firms despite the fact that each had fewer workers. Figure 10 provides the same data for industries with historically high and low rates of ESI coverage. 13 High ESI industries include finance, manufacturing, public administration and utilities. Low ESI industries include agriculture, construction, wholesale and retail trade, and most service industries. Over the period, there 13 All industries are included. The cutoff between high and low was an ESI rate of 80% in High ESI industries are: Mining, Manufacturing, Utilities, Finance/Insurance/Real Estate, Education, and Public Administration. Low ESI industries are: Agriculture, Construction, Transportation, Wholesale/Retail Trade, Information/Communication, Professional, Health and Social Services, Arts/Entertainment, Other Services, and Former Military. 20

23 was a decline of 2.3 million workers in high ESI industries and an increase of 2.5 million in low ESI industries. There were declines in the rates of employer sponsored insurance in both types of industries but a much larger decline in low ESI industries; 3.2 vs. 1.2 percentage points. Because of the sharp drop in the rate of employer sponsored insurance in the low ESI industries coupled with a shift in work towards those industries, the number of workers in Figure 10 Changes in Health Insurance Coverage of Workers, By Industry Type, , Percentage Point Changes Change in Workers Change in Uninsured All Workers -2.8* 0.4* 2.1* High ESI Industries -1.2* * Low ESI Industries -3.2* 0.5* 0.2 Million -2.3 Million 2.5 Million 3.1 Million 0.2 Million 2.9 Million * Statistically significant change between 2000 and 2002 (p<.10). Medicaid also includes S- CHIP, other state programs. Figure excludes changes in Champus/Medicare and private non group insurance for ease of presentation because changes in coverage are generally small. 2.5* Employer Sponsored Medicaid Uninsured Source: Urban Institute low ESI industries without coverage increased by 2.9 million. Majority of coverage declines were among white Americans We next examined changes in coverage by racial and ethnic group, citizenship and region of the country. The data in the next three figures include both adults and children. The most striking finding in Figure 11 is that about 60% of the increase in the uninsured over this period occurred among whites. 14 Whites had the largest shift in income distribution a decline in the middle income and an increase in the low income 14 A problem in analyzing data on race over this period was that there were changes in the wording of the race question on the CPS. Through the March 2002 CPS individuals were instructed to choose one race white, black, and so on. In the March supplements individuals were allowed to respond by indicating multiple races. Relatively few chose to report multiple races; thus we expect any effect on the results to be small. We characterized those individuals as white if they reported being white alone. Similarly blacks are those who reported being black alone. By counting only those reporting being white alone or black alone in 2003, we will understate the growth in the white or black population (and with any type of or lack of coverage) because those who report a combination in 2003 will be counted as other. They did not have the option of doing so in Hispanics are identified in response to separate questions; in this analysis they are reported as Hispanic regardless of race. 21

24 groups (Table A8). In part because of the decline in incomes whites had the greatest growth in the number of uninsured. Blacks, Hispanics, and others experienced increases in the low income population without any significant decline in middle and higher income groups. All groups experienced a decline in employer-sponsored insurance. Whites, blacks, and Hispanics had increases in public coverage, though the increases for blacks and Hispanics were, in Figure 11 Changes in Coverage In Race/Ethnicity, Percentage Point Changes Change in Population Change in Uninsured White Black Hispanic Other -3.7* 1.7* 1.4* -3.7* 2.9* * 3.1* * 2.5* 0.5 Million 0.5 Million 4.0 Million 2.6 Million 2.9 Million 0.4 Million 1.3 Million 0.5 Million * Statistically significant change between 2000 and 2002 (p<.10). Medicaid also includes S- CHIP, other state programs. Figure excludes changes in Champus/Medicare and private non group insurance for ease of presentation because changes in coverage are generally small Employer Sponsored Medicaid Uninsured Source: Urban Institute percentage point terms, greater than those of whites. Because the increase in public coverage among whites was smaller, whites had the only statistically significant increase in their uninsured rate. Thus, because of the declines in incomes coupled with an increase in the uninsured rate, the number of uninsured white Americans increased by 2.9 million. Despite the increase in the number of uninsured, the white uninsured rate remained well below that of other groups. The uninsured rate for whites in 2003 was 12.9% compared to 21.0% for blacks, 34.3% for Hispanics, and 20.2% for others (Table A8). Blacks experienced a sharp decline in employer coverage but a large increase in public coverage. There was no significant change in the uninsured rate and the number 22

25 of black uninsured increased by 400,000 due to population growth. Similarly, for Hispanics the rate of in employer coverage fell while the rate of public coverage increased. The number of uninsured Hispanics increased by 1.3 million, largely because of growth in their numbers. Majority of coverage declines were among native citizens We next look at changes in coverage by citizenship. Figure 12 shows about 70% of the growth in the uninsured occurred among native citizens with almost all of the balance among non citizens. The growth in uninsured among non citizens was disproportionate to their overall numbers but nonetheless most of the increases in uninsurance Figure 12 Changes in Coverage, By Citizenship, Percentage Point Changes Native Citizens Naturalized Citizens Non Citizens 2.1* 1.3* * 3.9* occurred predominately among native citizens * -2.1* -5.7* Employer Sponsored Medicaid Uninsured Native citizens saw Change in Population Change in Uninsured 5.6 Million 1.0 Million 3.6 Million 0.3 Million 1.0 Million 1.2 Million * Statistically significant change between 2000 and 2002 (p<.10). Medicaid also includes S- CHIP, other state programs. Figure excludes changes in Champus/Medicare and private non group insurance for ease of presentation because changes in coverage are generally small. Source: Urban Institute a decline of employersponsored coverage of 3.8 percentage points. As a result, 4.5 million lost employer coverage. Some of this decline in employer-sponsored insurance was offset by increases in public coverage, but nonetheless there was a 1.3 percentage point increase in the uninsured rate and an increase of 3.6 million uninsured. As with the population in general, the bulk of the growth in the number of uninsured native citizens was among those below 200% of poverty (Table A9). In addition, almost all of the growth in the number of the uninsured 23