Health Insurance for Children: Many Remain Uninsured Despite Medicaid
Expansion (Letter Report, 07/19/95, GAO/HEHS-95-175).
Pursuant to a congressional request, GAO reviewed the status of health
insurance for children, focusing on: (1) the impact of the Medicaid
expansion on children's health insurance coverage since 1989; (2)
changes in the demographic profile of children enrolled in the Medicaid
program and uninsured children since the Medicaid expansion; and (3) the
number of uninsured children who might be eligible for Medicaid.
GAO found that: (1) policy changes helped increase the number of
children enrolled in Medicaid by 4.8 million between 1989 and 1993, but
the overall number of uninsured children did not decline because
employment-based coverage for adults and children declined during the
same period; (2) children were not as affected by the loss of
employment-based insurance as adults because of expanded Medicaid
coverage; (3) the percentage of poor children who were uninsured
declined from 25 percent in 1989 to 20 percent in 1993, while the
percentage of near-poor children who were uninsured increased during
that time; (4) the Medicaid expansion increased the enrollment of
children less likely to be on Medicaid and by 1993, more than half of
Medicaid children had a working parent and almost half were not
receiving Aid to Families with Dependent Children benefits; (5) the
greatest increase in coverage was among children with at least one
full-time working parent; (6) the South region had the greatest increase
in the number of children enrolled in Medicaid, although it still has
the greatest number of uninsured children; and (7) at least 2.3 million
uninsured children were eligible but not enrolled in Medicaid because
their parents were unaware of their eligibility or had difficulty in
applying for Medicaid coverage.
--------------------------- Indexing Terms -----------------------------
REPORTNUM: HEHS-95-175
TITLE: Health Insurance for Children: Many Remain Uninsured
Despite Medicaid Expansion
DATE: 07/19/95
SUBJECT: Medicaid programs
Disadvantaged persons
Children
Eligibility criteria
Demographic data
Health insurance
Public assistance programs
Aid to families with dependent children
Employee medical benefits
IDENTIFIER: AFDC
**************************************************************************
* This file contains an ASCII representation of the text of a GAO *
* report. Delineations within the text indicating chapter titles, *
* headings, and bullets are preserved. Major divisions and subdivisions *
* of the text, such as Chapters, Sections, and Appendixes, are *
* identified by double and single lines. The numbers on the right end *
* of these lines indicate the position of each of the subsections in the *
* document outline. These numbers do NOT correspond with the page *
* numbers of the printed product. *
* *
* No attempt has been made to display graphic images, although figure *
* captions are reproduced. Tables are included, but may not resemble *
* those in the printed version. *
* *
* A printed copy of this report may be obtained from the GAO Document *
* Distribution Facility by calling (202) 512-6000, by faxing your *
* request to (301) 258-4066, or by writing to P.O. Box 6015, *
* Gaithersburg, MD 20884-6015. We are unable to accept electronic orders *
* for printed documents at this time. *
**************************************************************************
Cover
================================================================ COVER
Report to the Ranking Minority Member, Subcommittee on Children and
Families, Committee on Labor and Human Resources, U.S. Senate
July 1995
HEALTH INSURANCE FOR CHILDREN -
MANY REMAIN UNINSURED DESPITE
MEDICAID EXPANSION
GAO/HEHS-95-175
Medicaid and Uninsured Children
Abbreviations
=============================================================== ABBREV
AFDC - Aid to Families With Dependent Children
CHAMPUS - Civilian Health and Medical Program of the Uniformed
Services
CPS - Current Population Survey
EBRI - Employment Benefits Research Institute
MCCA - Medicare Catastrophic Care Amendments
OBRA - Omnibus Budget Reconciliation Act
Letter
=============================================================== LETTER
B-260450
July 19, 1995
The Honorable Christopher J. Dodd
Ranking Minority Member
Subcommittee on Children and Families
Committee on Labor and Human Resources
United States Senate
Dear Senator Dodd:
Access to needed medical care can detect and prevent problems
affecting the future health and lives of children. Many U.S.
children, however, receive less health care than others because they
lack health insurance. In 1993, 9.3 million U.S. children did not
have health insurance coverage at any time during the year. These
9.3 million children account for almost one-quarter of all uninsured
people.
Since the 1980s, the Congress has expanded Medicaid eligibility so
that more children could have health care coverage. Currently, the
104th session of the Congress is considering making the Medicaid
program a block grant to states, limiting the growth of program
expenditures, and removing eligibility and other requirements. Such
changes could greatly impact children's health care coverage if
states choose to remove guaranteed eligibility for children or reduce
current state eligibility levels.
Concerned about these issues, you asked us to
determine the impact of the Medicaid expansion on children's health
insurance coverage since 1989,
identify changes in the demographic profile of children enrolled in
the Medicaid program and uninsured children since the Medicaid
expansion, and
estimate the number of uninsured children who might be eligible for
Medicaid.
To answer these questions, we analyzed data from the Bureau of the
Census' March 1990 and March 1994 Current Population Survey (CPS),
linking information on children's health insurance status and their
parents' demographic characteristics.\1 \2 (See app. III.) To
estimate the number of Medicaid-eligible children that were uninsured
and not enrolled in the Medicaid program, we defined a group of
children from the March 1994 CPS data file that were Medicaid
eligible according to federal rules and analyzed their insurance
status.\3 We also analyzed relevant laws to understand changes in
Medicaid eligibility criteria affecting children. We released some
of the findings from our analysis in a correspondence to you.\4 Our
work was conducted between October 1994 and April 1995 in accordance
with generally accepted government auditing standards.
--------------------
\1 Each March, the Census Bureau asks questions about health
insurance coverage during the previous year. Thus, the 1990 and 1994
surveys report information for 1989 and 1993.
\2 The CPS data we are reporting include only those children we were
able to pair with an adult--98.4 percent of the sample in 1993. We
could not pair 1.6 percent of the children. Thus, in 1993, 9.6
million children (13.7 percent) were uninsured and 14 million (20.2
percent) received Medicaid, while 39.7 million (57 percent) had
employment-based insurance.
\3 We define the group of Medicaid eligible as children age 0 to 5
with family income at or below 133 percent of the federal poverty
level and children 6 to 10 with family incomes at or below the
federal poverty level. The Congress mandated Medicaid coverage for
all of these children born after September 30, 1983.
\4 Uninsured and Children on Medicaid (GAO/HEHS-95-83R, Feb. 14,
1995).
RESULTS IN BRIEF
------------------------------------------------------------ Letter :1
Policy changes to expand Medicaid eligibility for children helped to
increase the number of children enrolled in Medicaid by 4.8 million
between 1989 and 1993. However, the overall number of children who
are uninsured did not decline because employment-based coverage for
adults and children also declined during the same period.\5 Because
of expanded Medicaid coverage, children were not as affected by the
loss of employment-based insurance as adults. Without expanded
Medicaid coverage, many more children would have been uninsured in
1993.
Medicaid primarily serves children who are poor or near poor, so
expanding Medicaid eligibility primarily affected the insurance
status of these children.\6 The percentage of poor children who were
uninsured declined from 25 percent in 1989 to 20 percent in 1993. In
contrast, the percentage of children in families with above near-poor
income who were uninsured increased.\7
The Medicaid expansion has increased the enrollment of children less
likely to be on Medicaid previously--children in working families and
children not receiving Aid to Families With Dependent Children
(AFDC). By 1993, more than half of Medicaid children had a working
parent and almost half were not on AFDC.\8 The greatest increase in
coverage has been among children with at least one full-time\9
working parent, which increased from 13.2 percent in 1989 to 20.1
percent in 1993. The proportion of children in two-parent families
on Medicaid has also increased. Of all regions, the South\10 had the
greatest increase in the number of children enrolled in Medicaid,
though it still has the greatest number of uninsured children.
Although the Medicaid expansion has increased Medicaid coverage for
children in working families, these children remain the largest
segment of uninsured children. In fact, more uninsured children--61
percent--had a full-time working parent in 1993 than in 1989. Like
most U.S. children, most uninsured children live in two-parent
families with at least one working parent--but they have lower than
average income. Compared with more affluent workers, lower income
workers are more likely to work for businesses that do not offer
employee coverage, and, if they do, do not offer dependent coverage.
While the Medicaid expansion allowed many children to be covered who
might otherwise be uninsured, a substantial number of uninsured
children in 1993 were eligible but not enrolled in the program. At
least one-quarter of currently uninsured children--2.3 million--met
federal Medicaid age and income eligibility standards. Reasons that
eligible children are not covered may include their parents' lack of
knowledge about their potential eligibility and difficulties in
applying for Medicaid.
--------------------
\5 Some children, particularly children on Medicaid, have multiple
sources of insurance. We define "children on Medicaid" as children
on Medicaid who did not have any employment-based insurance. (See
app. III.)
\6 Poor children are children living in families with income at or
below the federal poverty level. For July 1, 1992, through June 30,
1993, the federal poverty income for a family of three was an annual
income of $11,570 or less. For this report, we define "near-poor"
children as children living in families with income between 101 and
150 percent of the federal poverty level. For a family of three,
this means an annual income between $11,571 and $17,357.
\7 We define "above near-poor" children as children living in
families with income above 150 percent of the federal poverty level.
For a family of three, this means an annual income above $17,357.
\8 Nor were they on other assistance.
\9 In this report, "full-time" work refers to working full time for
the entire year with no lapses in full-time employment. Parents who
worked full time for only part of the year are classified with
part-time workers as working less than full time.
\10 The South includes Alabama, Arkansas, Delaware, the District of
Columbia, Florida, Georgia, Kentucky, Louisiana, Maryland,
Mississippi, North Carolina, Oklahoma, South Carolina, Tennessee,
Texas, Virginia, and West Virginia.
BACKGROUND
------------------------------------------------------------ Letter :2
Most U.S. families access the health care system through health
insurance coverage. Without health insurance, many families face
difficulties getting preventive and basic care for their children.
Uninsured women, for example, are less likely to get early and
adequate prenatal care. Late and inadequate prenatal care is
associated with higher rates of low birth weight and prematurity,
serious illness, and handicap for children.\11 Children without
health insurance are less likely to have routine doctor visits, get
care for injuries, or have a regular source of medical care. When
they do seek care, they are more likely to get it through a clinic
rather than a private physician or health maintenance organization
(HMO).\12 \13 \14 Uninsured children are also less likely to be
appropriately immunized--important in preventing childhood
illness.\15 \16
Insured children in the United States either have privately or
publicly funded health insurance. In 1993, 89 percent of children
with private insurance got coverage through their parents'
employment. A small percentage of children have private,
individually purchased policies. An even smaller percentage are
children of military personnel who get publicly funded insurance
through their parents' employment. Most children with publicly
funded insurance get coverage through Medicaid.
The Medicaid program is a jointly funded federal-state entitlement
program that provides health insurance for both children and adults.
It is implemented through 56 separate programs (including the 50
states, the District of Columbia, Puerto Rico, and the U.S.
territories). States are required to cover some groups of children
and adults and may extend coverage to others. Children and their
parents must be covered if they receive benefits under the AFDC
program. In the past, most children received Medicaid because they
were on AFDC. Children and adults may also be eligible for the
program if they are disabled and have low incomes or if their medical
expenses are extremely high relative to family income.\17
Beginning in 1986, the Congress passed a series of laws that expanded
Medicaid eligibility for pregnant women, infants, and children. (See
table I.1 in app. I.) Before 1989, coverage expansions were
optional, although many states had expanded coverage.\18 Starting in
July 1989, states had to begin covering pregnant women and infants
with family incomes at or below 75 percent of the federal poverty
level. The Omnibus Budget Reconciliation Acts of 1989 and 1990 added
additional requirements that states had to implement in 1990 and
1991.
By 1993, states were required to cover (1) pregnant women, infants,
and children up to age 6 with family income at or below 133 percent
of the federal poverty level and (2) children aged 6 to 10 (born
after September 30, 1983) with family income at or below 100 percent
of the federal poverty level. Current law also requires that the
group of poor children over age 6 eligible for Medicaid continue to
expand year by year until all poor children up to age 19 are eligible
in the year 2002. In addition, states may expand Medicaid
eligibility for infants and children beyond these requirements by
either phasing in coverage of children up to age 19 more quickly than
required or by increasing eligibility income levels or both. As of
April 1995, 37 states and the District of Columbia had expanded
coverage for children beyond federal requirements. (See app. I.)
Children represent a large proportion of Medicaid recipients but a
small proportion of Medicaid expenditures. In 1993, 49 percent of
Medicaid recipients were children under age 21, but only 16 percent
of Medicaid medical vendor payments were for their care.\19
Nonetheless, Medicaid's overall cost and the rate of cost increases
have raised concerns about the program's impact on the federal
budget. Medicaid costs are projected to increase from about $131
billion to $260 billion by the year 2000, according to the
Congressional Budget Office. The Congress is currently considering
different options to lower the cost of the program, including
removing guaranteed eligibility and giving capped funding to the
states as block grants.
--------------------
\11 Prenatal Care: Medicaid Recipients and Uninsured Women Obtain
Insufficient Care (GAO/HRD-87-137, Sept. 30, 1987).
\12 Barbara Bloom, Health Insurance and Medical Care: Health of Our
Nation's Children, United States, Advance Data from Vital and Health
Statistics; No. 188, National Center for Health Statistics
(Hyattsville, Md.: 1990).
\13 David L. Wood and others, "Access to Medical Care for Children
and Adolescents in the U.S.," Pediatrics, Vol. 86, No. 5 (1990),
pp. 666-673.
\14 Mary D. Overpeck and Jonathan B. Kotch, "The Effect of U.S.
Children's Access to Care on Medical Attention for Injuries,"
American Journal of Public Health, Vol. 85, No. 3 (1995) pp.
402-404.
\15 Charles N. Oberg, "Medically Uninsured Children in the United
States: A Challenge to Public Policy," Pediatrics, Vol. 85, No. 5
(1990), pp. 824-833.
\16 David U. Himmelstein and Steffie Woolhandler, "Care Denied:
U.S. Residents Who Are Unable to Obtain Needed Medical Services,"
American Journal of Public Health, Vol. 85, No. 3 (1995), pp.
341-344.
\17 This applies in states with Medically Needy programs.
\18 Thirty-two states and the District of Columbia had expanded
coverage for pregnant women and infants, and 26 states and the
District of Columbia had expanded coverage for older children as of
December 1988.
\19 If disabled children under 21 are included, 23 percent of
Medicaid medical vendor payments in 1993 were for children under age
21.
THE INCREASED NUMBER OF
CHILDREN ON MEDICAID HELPED
OFFSET THE DECLINE IN
EMPLOYMENT-BASED INSURANCE FOR
CHILDREN
------------------------------------------------------------ Letter :3
Medicaid has become an increasingly important source of health
insurance for low-income children as employment-based insurance has
declined for both children and adults. Between 1989 and 1993, the
number of children covered by Medicaid increased 54 percent--from
13.6 percent of U.S. children in 1989 (8.9 million children) to 19.9
percent in 1993 (13.7 million children). This could have led to a
major decrease in the percentage of children uninsured. It did not,
however, because the decrease in children covered by employment-based
insurance offset the increase in U.S. children insured through
Medicaid. (See fig. 1.)
Figure 1: Medicaid Insured
One-Fifth of U.S. Children in
1993
(See figure in printed
edition.)
Note: CHAMPUS stands for the Civilian Health and Medical Program of
the Uniformed Services.
Comparing trends between children and adults clarifies Medicaid's
role. Between 1989 and 1993, the percentage of children with
employment-based health insurance decreased 9 percent. During the
same period, the percentage of adults aged 18 to 64 with such
insurance decreased 7 percent. Both children and adults lost
employment-based coverage.
Because of the Medicaid expansion, however, the decline in
employment-based insurance for children did not lead to an increase
in the proportion of uninsured children. Adults had a different
experience. (See fig. 2.) Between 1989 to 1993, the proportion of
adults who were uninsured rose 16 percent. In contrast, the
proportion of children who were uninsured was similar in 1989--13.3
percent--and 1993--13.5 percent.\20
Figure 2: Percentage of
Uninsured Adults Increased More
Than the Percentage of
Uninsured Children Between 1987
and 1993
(See figure in printed
edition.)
Notes: Data through 1991 are weighted to the 1980 census. Data from
1992 and 1993 are weighted to the 1990 census.
Data collection method changed in 1993 to computer-assisted
interviewing.
Source: Bureau of the Census, Income Statistics Branch.
Comparing the experience of adults and children suggests that
expanding Medicaid for children did not displace privately purchased
individual insurance. The proportions of children and adults with
privately purchased insurance were similar in 1989 and changed little
in 1993. An increased proportion of adults did not purchase
individual policies as more adults became uninsured. If Medicaid had
displaced privately purchased insurance for children, the proportion
of children with privately purchased insurance would have decreased,
but it did not.
The question of whether parents who could have employment-based
insurance for their families chose to drop or refuse coverage to get
Medicaid coverage for their children is more complicated. The
longitudinal data that would be needed to directly answer this
question are not available. Two researchers attempted to overcome
this limitation by developing an economic model using CPS data from
1987 through 1992. They estimated that expanding Medicaid coverage
for pregnant women and children did partially displace
employment-based coverage, being responsible for about 17 percent of
the decline in private insurance coverage between 1987 and 1992. The
rest of the decline in coverage was due to macroeconomic factors,
changes in the demographic mix of population, or changes in
employers' offering or generosity in covering health insurance for
workers and their families. For children, their analysis leads to an
estimate that 37 to 47 percent of children's increase in Medicaid
coverage was linked to a reduction in employment-based insurance
coverage.\21
Although the Medicaid expansion offset the decrease in
employment-based insurance, an increasing number of children either
have no health insurance or depend on publicly funded health
insurance. In 1993, 9.3 million children were uninsured, and 13.7
million were on Medicaid. These totals represent over one-third of
U.S. children.\22
--------------------
\20 The proportion declined between 1989 and 1992 and then rose 8
percent from 1992 to 1993.
\21 David M. Cutler and Jonathan Gruber, Does Public Insurance Crowd
Out Private Insurance? National Bureau of Economic Research, Working
Paper No. 5082 (Cambridge, Mass.:1995).
\22 In fact, this underrepresents the importance of Medicaid, since
it does not include children who had both employment-based and
Medicaid coverage in these years. If those children are included,
23.9 percent of U.S. children were on Medicaid in 1993.
MEDICAID EXPANSION HAD
GREATER IMPACT ON INSURANCE
STATUS OF POOR AND NEAR-POOR
CHILDREN
---------------------------------------------------------- Letter :3.1
The effect of the Medicaid expansion is clear when considering how
poor children, near-poor children, and higher income children fared
in the health insurance marketplace between 1989 and 1993. During
this period, the percentage of children of all income levels with
employment-based insurance declined, and Medicaid coverage for all
income levels expanded but to different degrees.
Employment-based insurance declined most for near-poor children.
(See table 1.) Meanwhile, an additional 11 percent of these children
obtained coverage under Medicaid. The proportion of near-poor
children who were uninsured did not change significantly. These
children were still most likely to be uninsured in 1993 and most
likely to be on Medicaid.
Poor children had a smaller decline in employment-based insurance
than near-poor children and a large increase in Medicaid coverage.
As a result, the percentage of poor children who were uninsured
actually declined from 25 percent in 1989 to 20 percent in 1993.
Unlike poor and near-poor children, children with family incomes
above 150 percent of the federal poverty level were more likely to be
uninsured in 1993 than 1989. This group of children had the highest
rates of employment-based coverage in 1989 (80.2 percent) and the
smallest decrease in such coverage. The Medicaid expansion cushioned
this group less since few of these children depend on Medicaid for
their coverage. However, the percentage of these children covered by
Medicaid had a small and statistically significant increase--from 7.5
percent to 9.1 percent.
Table 1
Change in Percentage of Poor, Near-
Poor, and Above Near-Poor Children Who
Had Different Types of Insurance or Were
Uninsured, 1989 and 1993
Percen
tage
point Percen
Health insurance of children Percen Percen differ t
by poverty level t 1989 t 1993 ence change
---------------------------- ------ ------ ------ ------
Poor\a
------------------------------------------------------------
Employment-based 17.8 14.0 - -22
3.9\d
Medicaid 50.7 61.3 +10.6\ +21
d
CHAMPUS 1.0 0.9 -0.1 -8
Private/individually 5.4 3.8 - -30
purchased 1.6\d
Uninsured 25.0 20.1 - -20
5.0\d
============================================================
Total 100 100
Near-poor\b
------------------------------------------------------------
Employment-based 46.9 40.6 - -13
6.2\d
Medicaid 13.7 24.9 +11.2\ +82
d
CHAMPUS 3.2 2.4 -0.8 -25
Private/individually 9.7 7.6 - -22
purchased 2.1\d
Uninsured 26.5 24.5 -2.0 -8
============================================================
Total 100 100
Above near-poor\c
------------------------------------------------------------
Employment-based 80.2 77.4 - -4
2.8\d
Medicaid 1.7 3.1 +1.5\d +88
CHAMPUS 2.5 2.0 - -21
0.5\d
Private/individually 8.1 8.4 +0.3 +4
purchased
Uninsured 7.5 9.1 +1.6\d +21
============================================================
Total 100 100
All U.S. children
------------------------------------------------------------
Employment-based 63.2 57.6 - -9
5.6\d
Medicaid 13.6 19.9 +6.3\d +46
CHAMPUS 2.2 1.8 - -21
0.5\d
Private/individually 7.7 7.2 -0.5 -7
purchased
Uninsured 13.3 13.5 +0.1 +1
============================================================
Total 100 100
------------------------------------------------------------
Note: Column totals may not add to exactly 100 percent due to
rounding. Percentages may not exactly compute due to rounding.
\a Poor families have income at or below 100 percent of the federal
poverty level.
\b Near-poor families have income between 101 to 150 percent of the
federal poverty level.
\c Above near-poor families have income above 150 percent of the
federal poverty level.
\d Statistically significant at the .05 level.
THE MEDICAID EXPANSION
INCREASED THE NUMBER OF
MEDICAID CHILDREN IN WORKING
AND NON-AFDC FAMILIES, AMONG
OTHER CHANGES
------------------------------------------------------------ Letter :4
Because of policy changes to expand children's eligibility, Medicaid
has become a more important source of insurance for children who were
less likely to get Medicaid in the past, such as children in working
families. The proportion of Medicaid children who had a working
parent increased between 1989 and 1993. By 1993, more than half the
children on Medicaid had a working parent, and almost half did not
depend on AFDC. Medicaid coverage also increased more during this
period among other children less likely to receive Medicaid in the
past--children in two-parent families, children of more educated
parents, white non-Hispanic and Hispanic children, and children
living in the South.
MEDICAID NOW INSURES MORE
CHILDREN IN WORKING,
NON-AFDC, AND TWO-PARENT
FAMILIES
---------------------------------------------------------- Letter :4.1
Changing Medicaid eligibility policies so that children are eligible
on the basis of income and age, even if they are not on AFDC,\23
allowed uninsured children of low-income working families to get
health insurance through Medicaid. A significant increase in the
number of children with a working parent on Medicaid has resulted.
Over half of Medicaid children had a working parent in 1993. (See
fig. 3.) The proportion of children in working families on Medicaid
grew, and the number of Medicaid children with a working parent
increased from 4 million in 1989 to 7.3 million in 1993--an 83
percent increase. Most working parents with children on Medicaid
worked less than full time for the entire previous year. However,
the percentage of Medicaid families with a full-time worker
increased. By 1993, 1 child in 5 on Medicaid had a parent who worked
full time all year.
Figure 3: By 1993, More Than
Half of Medicaid Children Had a
Working Parent
(See figure in printed
edition.)
Expanding Medicaid eligibility on the basis of income and age was a
major reason for the increase in Medicaid enrollment, but not the
only reason. AFDC enrollment also increased between 1989 and 1993.
The number of children in Medicaid on AFDC or AFDC combined with
other assistance increased by 1.3 million children--a 25 percent
increase.\24 But the expansion in non-AFDC children on Medicaid was
much greater.
Between 1989 and 1993, the number of non-AFDC children on Medicaid
doubled--from 3.2 million to 6.4 million children. This greater
increase in non-AFDC children increased the proportion of non-AFDC
children on Medicaid. (See fig. 4.) The percentage of children not
receiving AFDC or other assistance in 1993 increased from 36 percent
of all children on Medicaid in 1989 to 47 percent--almost half.
Figure 4: Children Not on AFDC
Became a Larger Proportion of
Children on Medicaid
(See figure in printed
edition.)
The Medicaid program now serves more children in two-parent families,
another group of children less likely to receive Medicaid in the
past. The percentage of Medicaid children in two-parent families
grew from 29.7 percent in 1989 to 35.6 percent in 1993. Children on
Medicaid in two-parent families were more likely to have a working
parent in 1993 than in 1989 (78.5 percent compared with 72 percent)
and also more likely to have at least one parent who worked full time
(40.6 percent compared with 29.2 percent).
--------------------
\23 AFDC used to be one of the main ways children could enter the
Medicaid program and receive health insurance. Other children whom
Medicaid covered were the disabled or those with extremely high
medical bills.
\24 The growth in AFDC children probably relates to the increase in
the number and proportion of children in poverty and thus eligible
for poverty-related support programs. The number of U.S. children
in poverty increased from 14 million (22 percent of U.S. children)
to 17 million (25 percent) in 1993.
MEDICAID COVERAGE INCREASED
MORE AMONG SOME GROUPS OF
CHILDREN
---------------------------------------------------------- Letter :4.2
In addition to an increase in the proportion of working and
two-parent families on Medicaid, Medicaid enrollment also increased
more for other groups of children less likely to receive Medicaid in
the past. Children whose parents lack a high school diploma are most
likely to receive Medicaid, since lower education and lower income
are related. However, enrollment increased for some children less
likely to be on Medicaid--those whose most educated parent had a high
school diploma up through a bachelor's degree. Meanwhile,
employment-based insurance decreased for children whose parents'
education varied from less than high school to those whose parents
had some college education. (See app. II.)
Also, although a higher proportion of African American children
receive Medicaid coverage than children in other racial/ethnic
groups, Medicaid coverage expanded more for white, non-Hispanic
children, and Hispanic children than for African American children.
In 1989, a child in the South was less likely to receive Medicaid
than a child in any other region, even though the South had the
highest percentage of poor, uninsured children (47 percent). With
the Medicaid expansion, enrollment increased most in the South, so by
1993 the percentage of poor, uninsured children in the South had
declined. Despite this decline, the South remains the region with
the highest percentage of uninsured children and the largest number
of poor, uninsured children. (For more detail on these changes, see
app. II.)
MOST UNINSURED CHILDREN HAVE
WORKING PARENTS WITH LOWER THAN
AVERAGE INCOME
------------------------------------------------------------ Letter :5
Despite the Medicaid expansion, children of working parents with
lower than average income still predominate among uninsured children.
Living with a full-time working parent was even less of a guarantee
that children would have health insurance coverage in 1993 than it
was in 1989. Most uninsured children live with a full-time working
parent, generally in a two-parent family. They differ from most
children and especially from insured children because many more of
them are poor or near poor.
MOST UNINSURED CHILDREN HAVE
AT LEAST ONE WORKING PARENT
---------------------------------------------------------- Letter :5.1
In 1989 and 1993, 89 percent of uninsured children had at least one
working parent. The expansion in Medicaid coverage for children in
working families did not decrease the percentage of uninsured
children with a working parent. In some respects, the status of
children in working families worsened. This is because the
percentage of uninsured children with a full-time working parent grew
between 1989 and 1993--from 57.2 percent (5 million) to 61.4 percent
(5.7 million), respectively. (See fig. 5.)
Figure 5: Most Uninsured
Children Had at Least One
Working Parent in 1993
(See figure in printed
edition.)
Unlike children on Medicaid, uninsured children generally resemble
most U.S. children because they live in two-parent families,
generally with a working parent. (See fig. 6.) In comparison with
all U.S. children, uninsured children in two-parent families are
less likely to live with a parent who worked full time the entire
previous year. Uninsured children are also slightly more likely to
live in one-parent families than are U.S. children in general.
Figure 6: Most U.S. Children
and Most Uninsured Children
Lived in Two-Parent Working
Families in 1993
(See figure in printed
edition.)
UNINSURED CHILDREN'S
FAMILIES HAVE LOWER INCOMES
---------------------------------------------------------- Letter :5.2
The families of uninsured children have less income than families of
insured children. In 1993, 24.7 percent of all U.S. children lived
in poor families. A larger proportion of uninsured children--36.7
percent--lived in poor families while only 6 percent of children with
employment-based insurance did (see fig. 7). About 57 percent of
uninsured children have family income at or below 150 percent of the
federal poverty level, compared with 14 percent of children with
employment-based insurance.
Figure 7: More Than Half of
Uninsured Children Were Poor or
Near Poor in 1993
(See figure in printed
edition.)
In previous work, we found that poor workers may not have
employment-based insurance for several reasons:\25
Health insurance is a more substantial share of total employee
compensation for low-wage workers than for higher wage workers,
so firms with predominantly low-wage workers are less likely to
offer health insurance--or, if they do, may not offer dependent
coverage since doing so is more expensive than covering the
worker alone.
A large share of small firms that employ low-wage workers are less
likely to offer health insurance. Small firms pay higher health
insurance premiums because insurers incur higher administrative
costs to serve small firms.
Even if coverage is offered, health insurance cost sharing
represents more of the household budgets of poor and near-poor
families so that families may decide they cannot afford health
insurance even if available.
--------------------
\25 Health Insurance: An Overview of the Working Uninsured
(GAO/HRD-89-45, Feb. 24, 1989); Access to Health Insurance: State
Efforts to Assist Small Businesses (GAO/HRD-92-90, May 14, 1992); and
Employer-Based Health Insurance: High Costs, Wide Variations
Threaten System (GAO/HRD-92-125, Sept. 22, 1992).
MANY UNINSURED CHILDREN ARE
MEDICAID ELIGIBLE BUT NOT ON
MEDICAID
------------------------------------------------------------ Letter :6
Although many uninsured children could be covered by Medicaid, they
are not. At least one-quarter of uninsured children in 1993--2.3
million--had family incomes that should have made them eligible for
Medicaid.\26 That year, at least 13.8 million children met the
federal age and poverty income eligibility requirements for the
program. Of these children, 2.4 million (17.3 percent) had
employment-based insurance. Almost 8.4 million (60.9 percent) were
on Medicaid, and about 675,000 (5 percent) either had individually
purchased private insurance or Civilian Health and Medical Program of
the Uniformed Services (CHAMPUS) insurance. Another 16.9
percent--2.3 million children--were uninsured.\27 More than half of
these uninsured Medicaid-eligible children--over 1.4 million--were
under age 6.
Compared with children on Medicaid, these children are more likely to
live in working, two-parent families. In 1993, 79.5 percent had at
least one working parent, and 42.5 percent had at least one parent
working full time. A total of 1.3 million (57.6 percent) of the
Medicaid-eligible uninsured children are in two-parent families; 1.2
million of these children have a working parent. The South has more
uninsured Medicaid-eligible children than any other region--1.4
million. In comparison to all uninsured children, uninsured
Medicaid-eligible children are more likely to be African American or
Hispanic.
--------------------
\26 These are children for whom the Congress mandated Medicaid
coverage--either under age 6 with family incomes at or below 133
percent of the federal poverty level or from age 6 to 10 (born after
September 30, 1983) with family incomes at or below the federal
poverty level.
\27 This underestimates the number of uninsured children possibly
eligible for Medicaid, since many states have expanded either the age
or the income eligibility criteria above the federal minimum. (See
app. I.)
POSSIBLE REASONS WHY MORE
MEDICAID-ELIGIBLE UNINSURED
CHILDREN ARE NOT ENROLLED
---------------------------------------------------------- Letter :6.1
Several possible reasons may explain why these families might not
have enrolled their children in Medicaid. Low-income families may
not know that their children could be eligible for Medicaid even if a
parent works full time or if the family has two parents. In a study
that interviewed AFDC recipients and former recipients who had begun
working but were still receiving Medicaid (so-called Transitional
Medicaid) in Charlotte, North Carolina, and Nashville, Tennessee,
researchers found that 41 percent of AFDC recipients and 23 percent
of former recipients did not understand that a parent could work full
time and receive Medicaid for his or her children. Sixty-two percent
of the AFDC recipients and 37 percent of the Transitional Medicaid
recipients did not know that children could be eligible for Medicaid
if they lived in an intact, two-parent family.\28
Another reason so many uninsured children are not on Medicaid may be
that getting enrolled in Medicaid is difficult for low-income
families. Many people who are potentially eligible for Medicaid
never complete the application process, and about half the denials
are for procedural reasons--that is, applicants did not or could not
provide the basic documentation needed to verify their eligibility or
did not appear for all the eligibility interviews.\29
Finally, some families may not seek Medicaid until they face a
medical crisis because they are not used to regular or preventive
medical care. In addition, medical and social service providers
report that some families do not want to enroll in Medicaid because
they consider it a welfare program and consider it stigmatizing.
--------------------
\28 Sarah C. Shuptrine, Vicki C. Grant, and Genny G. McKenzie, A
Study of the Relationship of Health Coverage to Welfare Dependency,
Southern Institute on Children and Families (Columbia, S.C.: 1994),
pp. 21-25.
\29 Health Care Reform: Potential Difficulties in Determining
Eligibility for Low-Income People (GAO/HEHS-94-176, July 11, 1994).
CONCLUSION
------------------------------------------------------------ Letter :7
Expanding children's Medicaid eligibility has significantly increased
the number of children with Medicaid as their health insurance. It
has also helped cushion the effect of declining employment-based
health insurance coverage for children. Because of expanded
eligibility, the proportion of children on Medicaid in working and in
two-parent families has grown. The Congress is currently considering
legislation to reform AFDC to encourage low-income mothers to work.
However, work for many lower income families does not include the
benefit of health insurance that it more often does for higher income
families. Clearly, having a full-time working parent and being in a
two-parent family does not ensure that a child will have health
insurance. Although Medicaid has begun to help close that gap for
some families, many more uninsured children are eligible for Medicaid
than have been enrolled.
Changes to the Medicaid program that remove guaranteed eligibility
and change the financing and responsibilities of the federal and
state governments may strongly affect health insurance coverage for
children in the future. Children account for only a small portion of
Medicaid costs. Because they represent almost half the participants,
however, any changes to Medicaid disproportionately affect children.
Changes to Medicaid that result in reducing the number of children
covered, without any accompanying changes in the health insurance
marketplace either to encourage employers to provide dependent health
insurance coverage, or to encourage families to purchase insurance,
or to provide other coverage options for children, could lead to a
significantly increased number of uninsured children in the future.
We did not seek written agency comments because this report does not
focus on agency activities. We discussed a draft of this report with
responsible Department of Health and Human Services officials in the
Health Care Financing Administration and included their comments
where appropriate.
---------------------------------------------------------- Letter :7.1
As arranged with your office, unless you publicly announce its
contents earlier, we plan no further distribution of this report
until 30 days after its issue date. At that time, we will send
copies to interested parties and make copies available to others on
request.
Please contact me at (202) 512-7125 if you or your staff have any
further questions. This report was prepared by Rose Marie Martinez,
Sheila Avruch, Paula Bonin, and Frank Ullman.
Sincerely yours,
Mark V. Nadel
Associate Director
National and Public Health Issues
STATES HAVE EXPANDED MEDICAID
ELIGIBILITY AND OTHER HEALTH
INSURANCE OPTIONS FOR CHILDREN
=========================================================== Appendix I
The Congress passed a series of laws beginning in 1986 that
substantially expanded Medicaid eligibility for pregnant women and
children. Some of these laws required eligibility expansions, and
others allowed states options to expand eligibility. (See table
I.1.)
Table I.1
A Series of Laws Expanded Medicaid
Eligibility for Children
Act Eligibility expansion
----------------------------- -----------------------------
The Omnibus Budget OBRA-86 (P.L. 99-509) gave
Reconciliation Act of 1986 states the option to expand
(OBRA-86) Medicaid income eligibility
thresholds above AFDC levels
up to the federal poverty
level for pregnant women and
infants, effective April 1,
1987. It also gave states the
option of phasing in coverage
for poor children up to age
5, effective October 1, 1990.
The Omnibus Budget OBRA-87 (P.L. 100-203)
Reconciliation Act of 1987 allowed states to raise
(OBRA-87) Medicaid income thresholds
for pregnant women and
infants as high as 185
percent of the federal
poverty level, effective July
1, 1988. It also amended the
statute to give states the
option of phasing in coverage
of poor children up to age 8,
effective October 1, 1988.
The Medicare Catastrophic MCCA (P.L. 100-360) mandated
Care Amendments of 1988 minimum coverage of pregnant
(MCCA) women and infants at the
federal poverty level, with a
2-year phase-in period,
effective for calendar
quarters beginning on or
after July 1, 1989. Affected
states were to raise income
limits to 75 percent of
poverty by July 1, 1989, and
to poverty level by July 1,
1990. MCCA also added Section
1902 (r) (2) to the Social
Security Act, which allows
states to use more liberal
criteria for Medicaid than is
used for the AFDC program to
determine Medicaid financial
eligibility, effective July
1, 1988. States can disregard
specific amounts of income
and other resources and allow
certain categories of
eligible populations to
qualify for Medicaid.
The Omnibus Budget OBRA-89 (P.L. 101-239)
Reconciliation Act of 1989 superseded MCCA's mandate
(OBRA-89) schedule by requiring states
to cover, at a minimum,
pregnant women and children
up to age 6 at 133 percent of
the federal poverty level,
effective for calendar
quarters beginning on or
after April 1, 1990.
The Omnibus Budget OBRA-90 (P.L. 101-508)
Reconciliation Act of 1990 required states to begin
(OBRA-90) (effective on or after July
1, 1991) to phase in coverage
of children born after
September 30, 1983, until all
children living below poverty
up to age 19 are covered; the
upper age limit will be
reached by October 2002.
------------------------------------------------------------
Many states have taken advantage of the options to expand Medicaid
eligibility for infants and children beyond federally required
minimum eligibility levels--either by increasing the ages of children
covered more quickly than the phase-in requires or increasing
eligibility income levels or both. As of April 1995, 33 states and
the District of Columbia had increased coverage for infants beyond
federal requirements by generally expanding coverage up to 185
percent of the federal poverty level. Eight states expanded coverage
for children aged 1 through 5, and 20 states expanded coverage for
children aged 6 and older beyond federal requirements. In all, 37
states and the District of Columbia have expanded eligibility for
either infants or children or both. (See table I.2.)
Table I.2
States That Have Expanded Medicaid
Eligibility Beyond Federal Requirements,
as of April 1995
Childr Children Upper
Infant en (1- (6 and age
State s 5) older) limit
------------------------ ------ ------ -------- --------
Arizona 140 133 100 under 14
California\a 200 133 100 under
12\b
Connecticut 185 185 185 under
12\b
Delaware 185 133 100 under 19
District of Columbia 185 133 100 under
12\b
Florida 185 133 100 under
12\b
Georgia 185 133 100 under 19
Hawaii\c 185 133 133 under 19
Indiana 150 133 100 under
12\b
Iowa 185 133 100 under
12\b
Kansas 150 133 100 under 15
Kentucky 185 133 100 under 19
Maine 185 133 125 under 19
Maryland\a 185 133 100 under
12\b
Massachusetts\a 185 133 100 under
12\b
Michigan 185 150 150 under 15
(born
after 6/
30/79)
Minnesota\a\,d 275 133 100 under
12\b
Mississippi 185 133 100 under
12\b
Missouri 185 133 100 under 19
New Hampshire 185 185 185 under
19\
New Jersey\a 185 133 100 under
12\b
New Mexico 185 185 185 under 19
New York\a 185 133 100 under
12\b
North Carolina 185 133 100 under
12\b
Oklahoma 150 133 100 under
12\b
Oregon\c 133 133 100 under 19
Pennsylvania\a 185 133 100 under
12\b
Rhode Island\c 250 250 100 under 12
(born
after 6/
30/83)
South Carolina 185 133 100 under
12\b
South Dakota 133 133 100 under 12
(born
after 6/
30/83)
Tennessee\c 185 133 100 under
12\b
Texas 185 133 100 under
12\b
Utah 133 133 100 under 18
Vermont 225 225 225 under 18
Virginia 133 133 100 under 19
Washington 200 200 200 under 19
West Virginia 150 133 100 under 19
Wisconsin 155 155 100 under
12\b
============================================================
States with expansions 34 8 20
------------------------------------------------------------
Source: Center on Budget and Policy Priorities.
Note: Percentages in bold type show expansions beyond federal
minimum requirements, either for age, family income, or both.
\a These states also use state funds to further expand health
insurance coverage for children through statewide programs--see
discussion in following text.
\b States are required to provide Medicaid coverage to children aged
6 or over born after September 30, 1983, with income below the
federal poverty level.
\c These states are implementing 1115 Medicaid waivers, which may
expand health coverage for low-income children through Medicaid--see
the following discussion.
\d Minnesota also provides Medicaid coverage for children under age
21 with family income up to 133 percent of the AFDC standard.
In addition, several states have recently received special waivers
allowing them to undertake statewide Medicaid demonstration projects,
several of which extend health insurance coverage to portions of the
uninsured, including children. Authorized by section 1115(a) of the
Social Security Act (42 U.S.C. 1315), these waivers typically enable
states to place all or some of their Medicaid population in managed
care arrangements. The waivers commonly require higher income
families to pay premiums or copayments, often on a sliding scale. In
addition to the states with approved waivers, other states have
waivers pending. Since 1991 and as of June 15, 1995, Delaware,
Florida, Hawaii, Kentucky, Massachusetts, Minnesota, Ohio, Oregon,
Rhode Island, and Tennessee have had their section 1115 demonstration
waivers approved. To date, only Oregon, Hawaii, Rhode Island, and
Tennessee have implemented their 1115 demonstration waiver
programs.\30 Waiver applications for seven other states are pending:
Illinois, Missouri, Nevada, New Hampshire, New York, Oklahoma, and
Vermont.
The states that operate 1115 waiver programs have generally expanded
eligibility:
Hawaii expanded Medicaid eligibility to all persons with income up
to 300 percent of the federal poverty level, with cost sharing
for most residents with incomes above the federal poverty level.
Oregon expanded Medicaid eligibility to all persons with income up
to the federal poverty level while limiting health coverage to a
ranked list of services.
Rhode Island expanded coverage to pregnant women and children up to
age 6 with family incomes at or below 250 percent of the federal
poverty level.
Tennessee expanded coverage to uninsured people without regard to
income level, but cost sharing is required for people who are
not Medicaid eligible or have family income above the federal
poverty level. To manage the program within its planned
enrollment levels, Tennessee is now only enrolling people who
are Medicaid eligible or considered uninsurable.
Several states have developed other types of programs to insure
children not eligible for Medicaid. Seven states have statewide
programs using state and other funds to expand coverage for children
beyond Medicaid eligibility levels. Some of these programs provide
only limited benefits compared with the Medicaid program. For
example, they may not cover inpatient care. We will issue a report
on some of these programs and other nonstatewide programs later this
year.
California covers children under age 2 with family income up to 250
percent of the federal poverty level.
Maryland covers children under 11 who have family income up to 185
percent of the federal poverty level with a limited benefit
package.
Massachusetts has an insurance buy-in program with fees based on a
sliding scale for children under 13 with no family income limit.
Minnesota has an insurance buy-in program with fees based on a
sliding scale for children and adults with family income up to
275 percent of the federal poverty limit.
New Jersey covers children up to age 1 with family income up to 300
percent of the federal poverty limit.
New York covers children under 15 (born on or after June 1, 1980)
with a limited benefit package. The program is open to all
income levels, but only children with family income below 160
percent of the federal poverty limit are fully subsidized.
Children with family income between 160 percent and 222 percent
of the federal poverty level are partially subsidized.
Pennsylvania covers children under 14 with family income up to 185
percent of the federal poverty limit for fully subsidized
insurance; families with income between 185 and 235 percent of
the federal poverty level can buy partially subsidized insurance
for their children; in some parts of the state, these children
get their partial premiums paid by their insurers.
--------------------
\30 Despite their initial interest, some states have postponed or
reconsidered their waiver implementation due to concerns about
potential costs and other issues.
MEDICAID COVERAGE EXPANDED MOST
FOR CERTAIN GROUPS OF CHILDREN
========================================================== Appendix II
The Medicaid expansion has increased Medicaid coverage more for some
groups of children--those whose parents had more than minimal
education, whites and Hispanics, and children in the South. Although
the Medicaid expansion had a greater impact on the South, children
living in the South are still most likely to be uninsured in 1993.
MORE CHILDREN OF HIGHER
EDUCATED PARENTS ARE NOW ON
MEDICAID
-------------------------------------------------------- Appendix II:1
Medicaid coverage expanded for children whose parents' education
ranged from less than high school to college degrees. (See fig.
II.1.) The largest percentage increase in Medicaid coverage--92
percent--was among children whose most highly educated parent had
some college education but not a 4-year college degree. Children who
had a parent with a graduate education were the only group with
little change in percentage covered by Medicaid.
Figure II.1: Percentage of
Children on Medicaid Expanded
for Children With Parents of
Almost All Educational
Attainments
(See figure in printed
edition.)
The Medicaid expansion for children of all but the most educated
parents coincided with a decline in employment-based coverage for
most of the same groups. Since 1989, the percentage of children with
employment-based insurance declined for children whose parents'
education ranged from less than high school to some college. (See
table II.1.) The decline was greatest for children of parents with
less than a high school diploma. Children with a parent who had a
bachelor's degree or more education had no decrease in
employment-based coverage.
Medicaid continued to insure higher proportions of children with the
least educated parents. (See fig. II.2 and table II.1.) Higher
education is strongly correlated with higher income, and Medicaid
predominantly serves poor children. Children with less educated
parents are also more likely to be uninsured. Almost 80 percent of
children whose most educated parent lacks a high school diploma were
either uninsured or on Medicaid in 1993.
Figure II.2: Children of the
Least Educated Parents Were
Most Likely to Be Uninsured or
on Medicaid in 1993
(See figure in printed
edition.)
Table II.1
Change in Percentage of Children of
Parents With Differing Education Levels
Covered by Different Types of Insurance
or Uninsured, 1989 and 1993
Percen
tage
point Percen
differ t
Health insurance of child 1989 1993 ence change
---------------------------- ------ ------ ------ ------
Less than a high school diploma
------------------------------------------------------------
Employment-based 24.5 17.9 - -27
6.6\a
Medicaid 40.9 53.0 +12.1\ +29
a
CHAMPUS 0.3 0.3 0.0 -5
Private/individual 5.2 2.8 - -46
2.4\a
Uninsured 29.1 26.0 - -11
3.1\a
============================================================
Total 100 100
High school graduate
------------------------------------------------------------
Employment-based 60.7 49.5 - -18
11.2\a
Medicaid 14.2 25.0 +10.8\ +76
a
CHAMPUS 2.3 1.5 - -33
0.8\a
Private/individual 7.7 7.2 -0.5 -6
Uninsured 15.0 16.7 +1.7\a +12
============================================================
Total 100 100
Some college
------------------------------------------------------------
Employment-based 71.8 64.0 - -11
7.8\a
Medicaid 7.4 14.2 +6.8\a +92
CHAMPUS 3.2 2.7 -0.4 -13
Private/individual 8.4 7.9 -0.5 -6
Uninsured 9.3 11.1 +1.8\a +20
============================================================
Total 100 100
Bachelor's degree\b
------------------------------------------------------------
Employment-based 81.8 79.9 -1.9 -2
Medicaid 1.8 3.3 +1.5\a +80
CHAMPUS 2.1 1.7 -0.4 -18
Private/individual 8.9 8.4 -0.4 -5
Uninsured 5.4 6.6 +1.2 +22
============================================================
Total 100 100
Graduate education\c
------------------------------------------------------------
Employment-based 83.8 83.9 +0.1 0
Medicaid 0.8 1.0 +0.2 +24
CHAMPUS 3.0 1.7 - -42
1.3\a
Private/individual 8.3 9.3 +1.1 +13
Uninsured 4.1 4.0 -0.1 -2
============================================================
Total 100 100
------------------------------------------------------------
Notes: Educational levels were defined differently in 1989 and 1993
so that they are not exactly comparable. Numbers may not add,
subtract, or compute exactly due to rounding.
\a Statistically significant at the .05 level.
\b In 1989, 4 years of college and, in 1993, a 4-year college degree.
\c In 1989, more than 4 years of college and, in 1993, a graduate
degree.
MEDICAID EXPANSION PRODUCED
GREATER COVERAGE FOR WHITES AND
HISPANICS
-------------------------------------------------------- Appendix II:2
The Medicaid expansion increased the number of children on Medicaid
more for whites and Hispanics between 1989 and 1993 than for African
Americans. (See fig. II.3.) The number of white children on
Medicaid increased 75 percent--from 3.1 million to 5.5 million--and
the number of Hispanic children increased by 79 percent--from 1.8
million to 3.2 million--since 1989. In contrast, the number of
African American children on Medicaid increased 30 percent--from 3.4
million to 4.4 million. In 1993, 41 percent of African American
children, 35 percent of Hispanic children, and 12 percent of white
children were on Medicaid. White children are less likely to be on
Medicaid, but, of all children on Medicaid, they represent the
largest segment (40.2 percent).
Figure II.3: Number of
Children on Medicaid Increased
Most for Whites and Hispanics
Between 1989 and 1993
(See figure in printed
edition.)
Nevertheless, children of racial and ethnic minorities are still more
likely to be uninsured. (See fig. II.4.) While only 10.5 percent of
white children were uninsured in 1993, 25.6 percent of Hispanic
children and 15.2 percent of African American children were
uninsured. Minority children have higher rates of being uninsured,
but white children make up about half (51.6 percent) of all uninsured
children.
Figure II.4: African American
and Hispanic Children Were More
Likely to Be Uninsured or on
Medicaid in 1993
(See figure in printed
edition.)
GROWTH IN MEDICAID ENROLLMENT
FOR CHILDREN GREATEST IN THE
SOUTH
-------------------------------------------------------- Appendix II:3
Between 1989 and 1993, the number of children on Medicaid increased
in all regions, but the greatest increase occurred in the South.
(See fig. II.5.) Compared to the Northeast and Midwest, the South
had higher percentages of uninsured children in poverty in both 1989
and 1993--38.7 percent in 1993. Despite this, the South had the
lowest percentage of its children on Medicaid in 1989--12.7 percent
in the South compared with 15.3 percent in the West. Southern states
historically have had stricter AFDC eligibility requirements relative
to the federal poverty level than other regions. Thus, Southern poor
children were less likely to be on AFDC and covered by Medicaid
through AFDC. When Medicaid coverage became mandated by age and
poverty, the greatest number of children benefiting were in the
South. Medicaid coverage increased to 20.5 percent of southern
children in 1993. Not surprisingly, the four states with the highest
Medicaid coverage of children are all in the South--the District of
Columbia (45.4 percent), Louisiana (29.9), Mississippi (27.9), and
Tennessee (26.8); 8 of the top 15 states for Medicaid coverage of
children are southern states. (See table II.1 for percentages of
Medicaid, uninsured, and employment-based insured children by state.)
Overall, the percentage of children on Medicaid increased for all
regions, and the disparities between regions decreased between 1989
and 1993.
Figure II.5: The South Had the
Largest Increase in Number of
Children on Medicaid Between
1989 and 1993
(See figure in printed
edition.)
Nevertheless, despite the Medicaid expansion, more uninsured children
live in the South than in any other region of the country. The South
has 43 percent of uninsured children--almost 4 million children.
Businesses in the South are less likely to offer health insurance
than businesses in other regions. Regional differences in health
insurance coverage among the employed may also reflect the greater
degree of industrialization and unionization in other parts of the
country and higher incidence of small and service-sector businesses
in the South.
Table II.2
Children Uninsured, on Medicaid, or With
Employment-Based Insurance, by State,
1993
Percen 95% S. Percen 95% Percen 95%
State Number t E.\a Number t S.E.\a Number t S.E.\a
-------- ------ ------ ------ ------ ------ ------ ------ ------ ------
Alabama 156,22 15.1 5.1 168,53 16.2 5.3 627,16 60.4 7.0
3 0 1
Alaska 13,063 8.3 3.4 31,180 19.7 4.9 84,929 53.8 6.1
Arizona 207,03 19.5 5.4 160,81 15.1 4.8 615,41 57.9 6.7
9 9 7
Arkansas 127,86 20.2 5.6 127,13 20.0 5.5 322,26 50.8 6.9
9 9 0
Californ 1,611, 17.5 1.9 2,361, 25.7 2.2 4,539, 49.3 2.5
ia 965 248 114
Colorado 87,008 9.8 4.4 124,75 14.0 5.1 555,31 62.3 7.2
9 9
Connecti 72,225 9.1 4.8 107,17 13.5 5.7 533,58 67.3 7.8
cut 3 4
Delaware 20,398 11.6 5.0 27,906 15.9 5.7 115,25 65.6 7.4
7
District 23,850 16.7 6.4 64,962 45.4 8.6 49,740 34.8 8.2
of
Columbi
a
Florida 622,31 17.5 2.6 803,81 22.6 2.9 1,766, 49.7 3.4
2 4 901
Georgia 270,98 15.6 5.0 304,00 17.5 5.2 924,66 53.3 6.9
9 0 8
Hawaii 25,549 9.2 4.4 40,909 14.8 5.4 159,97 57.7 7.5
8
Idaho 51,836 14.4 4.0 51,257 14.2 4.0 219,76 61.1 5.6
5
Illinois 330,05 10.6 2.3 637,29 20.4 3.0 1,848, 59.2 3.6
4 9 756
Indiana 139,01 8.9 4.0 253,09 16.1 5.2 1,021, 65.1 6.7
9 9 912
Iowa 61,839 8.0 3.7 65,398 8.5 3.8 534,20 69.3 6.2
0
Kansas 87,787 12.5 4.3 104,96 15.0 4.7 443,27 63.2 6.3
5 2
Kentucky 107,52 11.5 4.6 241,62 25.7 6.3 534,60 57.0 7.1
6 1 2
Louisian 275,28 22.2 5.7 369,69 29.8 6.3 510,24 41.2 6.8
a 4 9 8
Maine 30,728 8.7 3.9 62,208 17.5 5.3 218,18 61.4 6.7
2
Maryland 106,62 9.4 4.6 155,80 13.8 5.4 721,22 63.9 7.6
8 2 2
Massachu 122,23 8.6 2.2 258,12 18.2 3.1 919,49 65.0 3.8
setts 8 3 7
Michigan 207,66 7.9 1.9 603,88 23.1 3.0 1,646, 62.9 3.5
7 0 967
Minnesot 76,517 6.9 3.7 182,12 16.5 5.5 726,11 65.7 7.0
a 1 8
Mississi 123,22 16.8 4.8 204,05 27.8 5.8 344,81 47.0 6.4
ppi 3 0 2
Missouri 125,94 9.3 4.2 288,75 21.4 6.0 850,45 63.0 7.1
2 4 7
Montana 33,356 14.2 4.6 34,836 14.8 4.7 140,21 59.7 6.4
7
Nebraska 45,361 9.2 3.6 69,565 14.1 4.3 295,75 60.1 6.1
9
Nevada 63,168 17.4 5.1 33,158 9.1 3.9 230,98 63.7 6.5
1
New 32,073 11.0 5.0 33,129 11.3 5.1 208,65 71.3 7.3
Hampshi 8
re
New 227,40 11.4 2.4 338,48 17.0 2.8 1,244, 62.5 3.6
Jersey 1 8 667
New 112,26 24.2 5.5 79,222 17.1 4.8 232,44 50.2 6.4
Mexico 1 5
New York 490,46 10.9 1.8 1,111, 24.8 2.5 2,581, 57.6 2.8
6 452 922
North 214,18 13.7 2.6 286,51 18.4 2.9 890,64 57.2 3.8
Carolina 4 9 6
North 18,571 10.9 4.1 20,793 12.2 4.3 101,83 59.7 6.4
Dakota 1
Ohio 274,44 8.8 2.0 630,54 20.2 2.8 1,976, 63.5 3.4
5 7 099
Oklahoma 252,55 25.5 5.6 145,10 14.7 4.5 519,57 52.5 6.4
5 7 2
Oregon 93,859 11.8 4.8 107,33 13.5 5.1 527,17 66.4 7.0
5 5
Pennsylv 295,07 9.8 2.2 525,99 17.4 2.8 1,918, 63.6 3.6
ania 7 9 598
Rhode 26,058 11.3 5.3 53,741 23.3 7.1 139,56 60.4 8.2
Island 9
South 126,04 13.5 4.3 204,24 21.8 5.2 506,03 54.1 6.3
Carolina 6 0 0
South 26,192 12.2 3.8 32,810 15.2 4.2 122,15 56.8 5.8
Dakota 6
Tennesse 131,67 10.3 4.1 344,25 26.8 6.0 715,54 55.8 6.7
e 8 1 5
Texas 1,110, 21.1 2.7 1,031, 19.6 2.6 2,675, 50.9 3.3
747 685 381
Utah 69,240 10.1 3.4 49,813 7.3 2.9 489,70 71.6 5.0
1
Vermont 11,387 7.1 3.9 28,204 17.6 5.8 102,27 63.9 7.3
6
Virginia 215,87 12.9 4.1 226,28 13.5 4.2 1,038, 62.0 6.0
3 1 898
Washingt 119,22 9.1 4.0 151,19 11.6 4.4 854,26 65.4 6.6
on 5 8 4
West 76,217 18.6 5.9 92,902 22.7 6.4 221,32 54.0 7.6
Virginia 0
Wisconsi 96,832 7.2 3.4 204,57 15.2 4.7 962,46 71.7 5.8
n 3 9
Wyoming 19,841 13.8 5.2 19,444 13.5 5.2 89,421 62.1 7.4
================================================================================
United 9,266, 13.5 0.6 13,656 19.9 0.7 39,619 57.7 0.8
States 892 ,027 ,937
--------------------------------------------------------------------------------
\a S.E. represents sampling error. Each reported percent and number
estimate from the Current Population Survey has an associated
sampling error, the size of which reflects the precision of the
estimate. Sampling errors for percentage estimates were calculated
at the 95-percent confidence level, which means that the chances are
about 19 out of 20 that the actual percentage being estimated falls
within the range defined by our estimate, plus or minus the sampling
error. For example, we estimate that 13.5 percent of U.S. children
are uninsured; a 95-percent chance exists that the actual percentage
is between 12.9 percent and 14.1 percent.
METHODOLOGY
========================================================= Appendix III
To examine the impact of the Medicaid expansion on children, we
analyzed the Current Population Survey (CPS). The method we used to
define insurance status and to match children to parents resulted in
a conservative estimate of the number of children uninsured and on
Medicaid. In addition, two other aspects of our analysis affected
the results in different ways. First, we counted parental work
effort on the basis of whichever parent had the highest level of work
(such as working full time as opposed to part time), and we counted
parental education on the basis of whichever parent had attained the
highest educational level. Second, we used recently released sample
weights for the March 1990 CPS, which makes the data more equivalent
to the March 1994 CPS, although it would differ slightly from
previously published analyses of the March 1990 CPS.
ABOUT THE SURVEY
------------------------------------------------------- Appendix III:1
The CPS is the source of official government statistics on employment
and unemployment. Although the main purpose of the survey is to
collect information on employment, an important secondary purpose is
to collect information on the demographic status of the population,
such as age, sex, race, marital status, educational attainment, and
family structure. The CPS survey conducted every March also collects
additional data on work experience, income, noncash benefits, and
health insurance coverage of each household member at any time during
the previous year.
The CPS sample is based on the civilian, noninstitutionalized
population of the United States. About 57,000 households with
approximately 112,000 persons 15 years old and older and
approximately 33,000 children aged 0 to 14 years old are included.
It also includes Armed Forces members living in households with
civilians either on or off base. The households sampled by the CPS
are scientifically selected on the basis of area of residence to
represent the United States as a whole, individual states, and other
specified areas.\31
--------------------
\31 In January 1994, the survey was changed to improve data quality
by introducing computer-assisted data collection with a revised
questionnaire. The Census used both the old and new questionnaire
and methods to test differences between the two methods and found
some differences, such as in employment categorization. However, the
differences that they noted should not affect the variables reported
here.
DEFINITION OF INSURANCE STATUS
------------------------------------------------------- Appendix III:2
We defined insurance status using a hierarchy. If a child had
multiple coverage during the year, we counted that child under only
one type of coverage. (See table III.1.) We counted employment-based
insurance, which is the most common type for children to have, as the
primary insurance. If a child had both employment-based insurance
and Medicaid, that child was counted as having employment-based
insurance. Since most Medicaid children with multiple coverage had
Medicaid and employment-based coverage, our count of Medicaid
children better represents children who depended entirely on Medicaid
for any insurance coverage.\32 \33
Table III.1
Definitions of Insurance Status
Insurance status Definition
----------------------------- -----------------------------
Uninsured Did not have any health
insurance during the entire
year.
Employment-based Had health insurance
purchased through a parent's
employer or union for at
least part of the year.
Medicaid Did not have employment-
based health insurance at
all, but had Medicaid or
Medicare\a coverage for at
least part of the year.
CHAMPUS Did not have employment-
based insurance or Medicaid
or Medicare coverage at all,
but had CHAMPUS coverage for
at least part of the year.
Private/individually Did not have employment-
purchased based insurance, Medicaid or
Medicare coverage, or CHAMPUS
at all, but had private/
individually purchased health
insurance coverage for at
least part of the year.
------------------------------------------------------------
\a Few children have Medicare coverage--37,000 in 1993.
The Census has published a table with information from the CPS that
reports multiple coverage. It has different numbers and percentages
for insurance status of children with Medicaid, CHAMPUS, and
private/individually purchased health insurance than we reported
because it reports multiple coverage on the unmatched data set. (See
table III.2.)
--------------------
\32 The Census has published a table that reports multiple coverage,
which can be used for comparison. (See table III.2.)
\33 The Employment Benefit Research Institute (EBRI) reports multiple
coverage for children but defines some of its variables differently
than the Census does. Its estimates of uninsured children and
children with employment-based insurance coverage differ from those
of the Census, whereas ours do not.
MATCHING CHILDREN WITH
PARENTS
----------------------------------------------------- Appendix III:2.1
We matched children with parents to analyze family characteristics.
The Census considers a family to be two or more persons residing
together and related by birth, marriage, or adoption. The Census
develops family records for the householder (a person in whose name
the housing unit is owned, leased, or rented, or if no such person,
an adult in the household), other relatives of the householder with
their own subfamilies, and unrelated subfamilies. If the house is
owned, leased, or rented jointly by a married couple, the householder
may be either the husband or wife. We paired children to an adult
(aged 18 through 64) in their immediate family whom we call a parent.
After this pairing, we matched the adult family member to a spouse,
if any, to get "parents" in our file.
We were not able to match all children with parents. Because data in
this report are based only on the matched files, the number of
children reported in every insurance category is conservative. The
estimates of Medicaid and uninsured children are more conservative
than the estimate of children with employment-based insurance because
we were able to pair fewer Medicaid and uninsured children with an
adult than children with employment-based insurance. (See table
III.2.)
Table III.2
Number of Children by Type of Insurance
in Whole Data Set and After Matching
Parents With Children
(Numbers in thousands)
Number Percen Percen Percen Percen
Insurance type \a t Number t Number t t
------------------------ ------ ------ ------ ------ ------ ------ ------
Employment-based 39,745 57.0 39,745 57.0 39,620 57.7 0.3
Medicaid\a 16,693 23.9 14,128 20.3 13,656 19.9 3.3
CHAMPUS 2,307 3.3 1,215 1.7 1,209 1.8 0.5
Private/individual 7,272 10.4 5,103 7.3 4,927 7.2 3.5
Uninsured 9,574 13.7 9,574 13.7 9,267 13.5 3.2
================================================================================
Total 69,766 100 69,766 100 68,679 100 1.6
--------------------------------------------------------------------------------
Notes: Numbers or percents may not add or compute exactly due to
rounding.
\a Our analysis combined Medicaid and Medicare for children.
DETERMINING PARENTS' WORK
AND EDUCATIONAL STATUS
----------------------------------------------------- Appendix III:2.2
The way we matched parents with children to analyze the association
of work effort, education, and insurance for children helped develop
a more accurate picture of uninsured and Medicaid children with
working and more highly educated parents. We analyzed parent work
status on the basis of information about the parent who worked the
most. (See table III.3.) We also reported educational status on the
basis of whichever parent had the highest educational
status--graduate education, bachelor's degree or 4 years of college,
some college, high school diploma, or less than a high school
diploma. This allowed us to more accurately portray the work status
or education of parents in two-parent families.
Table III.3
Definition of Work Status of Parent or
Parents
Work status Reported as Definition
------------------------- ------------------------- --------------------------
Full time/full year Full time/full year Either parent worked full
time/full year.
Full time/part year Less than full time/full No parent worked full
year time/full year, but at
least one worked full time
part of the year.
Part time/full year Less than full time/full No parent worked full
year time, but at least one
parent worked part time
for the entire year.
Part time/part year Less than full time/full No parent worked either
year full time or full year,
but at least one parent
worked part time for part
of the year.
Not working Not working No parent worked at all
during the entire year.
--------------------------------------------------------------------------------
Conducting the analysis in this way allowed us to search for a parent
more likely to have insurance--either because they worked more or
were more educated. We found some interesting results from this
analysis. For example, we found more uninsured and Medicaid children
living with at least one parent who worked full time than if we had
not searched for employment status of both parents in two-parent
families. We also found fewer uninsured and Medicaid children who
lived with a parent who had less than a high school diploma.\34
--------------------
\34 EBRI reports parental work status, and other statuses in
two-parent families, on the basis of the parent (in two-parent
families) who earns the larger income.
USING WEIGHTS BASED ON THE
1990 CENSUS
----------------------------------------------------- Appendix III:2.3
The CPS is based on a sample of the U.S. population, and weights are
used to compute the estimates for the total population. The basic
weight represents the probability that individuals will be included
in the survey. The weights are computed on the basis of information
from the decennial censuses.
We used weights based on information from the 1990 census for both
1989 and 1993 to make them more equivalent. Information from the
1990 census was not available when the March 1990 CPS public use
survey tapes were first released so those tapes were originally
released with weights established through information from the 1980
decennial census. Since then, the Census Bureau released adjusted
weights for the March 1990 CPS that can be used in analyzing that CPS
file. We used weights adjusted to the 1990 census provided by the
Census Bureau for both the 1989 and 1993 data to make the data more
comparable and to make the 1989 data more accurate. We also did a
sample run on the 1989 data using the earlier census weights to
compare the differences. Using the more recent weights yields
slightly different results, such as a small increase in the number
and percentage of children uninsured or on Medicaid.
RELATED GAO PRODUCTS
============================================================ Chapter 0
Medicaid: Spending Pressures Drive States Toward Program Reinvention
(GAO/HEHS-95-122, Apr. 4, 1995).
Medicaid: Restructuring Approaches Leave Many Questions
(GAO/HEHS-95-103, Apr. 4, 1995).
Medicaid: Experience With State Waivers to Promote Cost Control and
Access Care (GAO/HEHS-95-115, Mar. 23, 1995).
Uninsured and Children on Medicaid (GAO/HEHS-95-83R, Feb. 14, 1995).
Health Care Reform: Potential Difficulties in Determining
Eligibility for Low-Income People (GAO/HEHS-94-176, July 11, 1994).
Medicaid Prenatal Care: States Improve Access and Enhance Services,
but Face New Challenges (GAO/HEHS-94-152BR, May 10, 1994).
Employer-Based Health Insurance: High Costs, Wide Variation Threaten
System (GAO/HRD-92-125, Sept. 22, 1992).
Access to Health Insurance: State Efforts to Assist Small Businesses
(GAO/HRD-92-90, May 14, 1992).
Mother-Only Families: Low Earnings Will Keep Many Children in
Poverty (GAO/HRD-91-62, Apr. 2, 1991).
Health Insurance Coverage: A Profile of the Uninsured in Selected
States (GAO/HRD-91-31FS, Feb. 8, 1991).
Health Insurance: An Overview of the Working Uninsured
(GAO/HRD-89-45, Feb. 24, 1989).