Revised October 21, 1999
Declining Medicaid Enrollment Undermines the Nation's Progress
in Insuring Low-Income Children
by Jocelyn Guyer, Matthew Broaddus, and Michelle Cochran
The latest data from the Census Bureau on the health insurance status of Americans show that the country has failed to make progress in recent years in reducing the number of low-income children who lack health insurance coverage. In 1998, nearly one of every four low-income children 7 million low-income children went without coverage. ("Low-income" is defined here as having family income below 200 percent of the poverty line, $26,006 a year for a family of three in 1998.)
The number of uninsured low-income children has not declined despite strong economic growth, the lowest unemployment rate in a quarter century, and continued expansions in low-income children's eligibility for publicly-funded coverage. To the contrary, over the past two years the number of uninsured children has grown, although not by an amount that is considered statistically significant.
This lack of progress cannot be attributed to a decline in the portion of low-income children with employer-based coverage.(1) The rate at which low-income children are covered through employer-based coverage has been holding steady in recent years. Instead, the failure to reduce the number of uninsured low-income children results largely from sharp declines in the number of poor children enrolled in Medicaid. These declines have occurred even though virtually all children with family incomes below the federal poverty line ($13,003 for a family of three in 1998) are eligible for Medicaid.(2)
What Does the Latest Census Data Say About the Effectiveness of CHIP?
The lack of progress in reducing the number of uninsured poor children cannot be attributed to shortcomings in the Children's Health Insurance Program. CHIP was just getting started in 1998; only seven states were enrolling children in their CHIP programs throughout all of 1998, moreover, the CHIP program is generally targeted at providing health insurance coverage to children in families with incomes above the poverty line.* As a result, the new Census data tell little about the effect of CHIP on poor children's enrollment trends.
While the lack of progress in reducing the number of uninsured poor children cannot be attributed to CHIP, the Census data do raise the troubling prospect that increases in children's coverage brought about by CHIP may be more than fully offset by declines in Medicaid enrollment in the years ahead.
* Testimony given by Nancy Ann Min DeParle, Administrator of the Health Care Financing Administration, before the Senate Finance Committee in April 1999.
Although there has been some concern that the Census Bureau's health insurance data might overstate the declines in Medicaid enrollment (see the attached technical note for a detailed discussion), the trends shown by these data are the latest contribution to a growing body of evidence indicating that changes in welfare policies are inadvertently and inappropriately spilling over to Medicaid and causing children to lose out on coverage for which they are eligible. As a result, the nation is losing a significant opportunity to reduce the number of low-income children without health insurance coverage. Without focused attention to the need to evaluate Medicaid eligibility properly as families move in and out of the welfare system, initiatives to expand coverage and reduce the number of uninsured children are likely to continue to fall short of expectations.
Medicaid Enrollment of Poor Children Falling Sharply
Since 1996, the total number of poor children in the United States has declined significantly, dropping from 14.5 million in 1996 to 13.5 million in 1998. The Census data show that the number of poor children enrolled in Medicaid has fallen even more sharply.
- Between 1996 and 1998, the number of poor children covered by Medicaid declined by 1.3 million, from 9.1 million in 1996 to 7.8 million in 1998.
- About half of this decline can be attributed to the drop in the number of children living in poverty.
- The other half of the decline is due to a lower Medicaid participation rate among poor children. In 1998, 57.8 percent of all poor children participated in Medicaid compared to 62.6 percent in 1996. This 4.8 percentage point decline in the Medicaid participation rate caused 643,000 poor children to lose Medicaid coverage.(3)
The decline in the Medicaid participation rate among poor children would not be troubling if the children not enrolled in Medicaid were gaining coverage through other sources and the overall percentage of poor children lacking any form of insurance was dropping. Unfortunately, this is not the case. Despite the robust economy and the decline in the number of children living in poverty, there has been no decline either in the number of poor children without health insurance or in the rate at which poor children are uninsured. Most poor children do not have access to coverage through sources other than Medicaid; in 1998, only about one in six poor children 17 percent were enrolled in employer-based coverage or additional sources of coverage other than Medicaid. As a result, lower Medicaid participation rates generally translate into lost opportunities to cover children who are eligible for Medicaid and who are not likely to have access to coverage from any other source.
If Medicaid Enrollment Had Not Fallen, Nearly Two-Thirds of a Million More Poor Children Could Have Been Covered
One way to measure the impact of declining Medicaid enrollment is to consider how many more children could have had health insurance coverage if the Medicaid participation rate among poor children had not fallen between 1996 and 1998:
- If Medicaid had covered the same proportion of poor children in 1998 that it covered in 1996, some 643,000 additional poor children could have had health insurance coverage last year.
- Moreover, the uninsured rate among poor children would have dropped from 23.3 percent in 1996 to 20.4 percent in 1998. Instead, the uninsured rate among poor children rose to 25.2 percent (although this increase was not statistically significant).
- If Medicaid participation rates had improved even modestly a goal sought by various initiatives to increase enrollment among eligible children significantly more progress could have been made. For example, if the participation rate had increased by two percentage points rather than dropped by five percentage points, nearly one million more poor children could have been enrolled. Assuming all of the newly enrolled children had been uninsured, the uninsured rate among poor children would have dropped to 18.4%.
Census Data Confirm Other Evidence That Medicaid-Eligible Children Increasingly Are Losing Out on Coverage
The latest Census data are consistent with a growing body of evidence showing that children frequently lose their Medicaid coverage and end up uninsured when their families are diverted from welfare or when they leave welfare for work, or for other reasons. Between 1996 and 1998, the average monthly number of families on cash welfare fell by 1.4 million, dropping from 4.4 million families in 1996 to 3 million in 1998.(4) Although federal law includes a number of protections designed to assure that children do not lose Medicaid when their families leave welfare, many states have not taken the steps necessary to assure that welfare policies do not lead to children losing out on health care coverage (see box on page 6).
In recent months, the Urban Institute, General Accounting Office, and Families USA have issued studies suggesting that the drop in the number of families on cash assistance has contributed to a decline in Medicaid enrollment among eligible children.
- An Urban Institute study conducted by Pamela Loprest, based on interviews with 1,300 families that stopped receiving welfare at some point between 1995 and 1997, found that more than half of the children in these families 53 percent lost their Medicaid coverage. The study also found that these children had only limited access to coverage through their parents' employers. As a result, a quarter of the children in families that left welfare ended up uninsured.(5)
- A different Urban Institute analysis of enrollment patterns in California and Florida during 1995 found that half of the children leaving cash assistance in both states lost Medicaid. During the period evaluated by the study, both states used waivers to alter their welfare programs and experienced declines in their cash assistance caseloads.(6)
- A recent General Accounting Office report found that the number of adults and children on Medicaid dropped by 7.4 percent between 1995 and 1997, from 23.6 million in 1995 to 21.8 million in 1997. The report concluded that "National declines in Medicaid enrollment raise questions about whether states have been able to 'de-link' welfare and Medicaid policies in a manner that consistently ensures Medicaid coverage for eligible individuals."(7)
- Finally, a study conducted for Families USA by Lewin Associates found that changes in welfare policy between 1995 and 1997 caused 1.25 million people to lose Medicaid. More than half of these 1.25 million people were uninsured in 1997.(8)
A number of states have conducted their own studies of how families that have left welfare are faring. Although these studies often look at somewhat different populations and do not always gather the same information, those that examine health insurance status consistently show that children leaving welfare frequently lose Medicaid coverage and are at risk of becoming uninsured. These state studies also generally show that the children in families leaving welfare remain eligible for Medicaid, either because their family incomes fall below the Medicaid eligibility limits for children or their families qualify for Transitional Medical Assistance. For example, a General Accounting Office review of selected state studies on how families are faring after leaving welfare found that average quarterly earnings for former recipients range from $2,378 to $3,786, earnings that fall below the federal poverty line for a family of four.(9)
Among the state-specific studies that address this issue, the portion of children found to be enrolled in Medicaid in the months after their families leave welfare has ranged from a low of 14 percent in Cuyahoga County, Ohio to a high of 74 percent in South Carolina. (The state-specific studies often look at Medicaid enrollment at different times relative to when a family has left welfare; the Ohio and South Carolina studies, however, both examined Medicaid enrollment of children 12 months after their families left welfare).
- The majority of the state studies released over the past year suggest that half or more of the children in families leaving welfare lose their Medicaid coverage.
- Studies in some states, such as New York and Ohio, found as many as 65 percent to 70 percent of children have lost their Medicaid coverage in the months after their families leave welfare. The Ohio study showed that after 12 months, more than 85 percent of the children in families that had left welfare in Cuyahoga County had no Medicaid coverage.
Why Are Changes in Welfare Policy and Procedures Causing Children to Lose Out on the Medicaid Coverage for Which They Are Eligible?
In the past, families with children generally were eligible for Medicaid only if they were receiving welfare. Under those rules, state eligibility systems linked the two programs; when a family was denied or terminated from welfare, the computer system would often automatically deny the Medicaid application or close the Medicaid case at the same time. The law has changed, however, and now Medicaid eligibility is "delinked" from eligibility for welfare. This change was adopted in the 1996 federal welfare law to prevent changes in federal and state welfare policies from affecting Medicaid.
Many states, however, have not revamped their systems to assure that welfare policies and procedures do not spill over to Medicaid. In many cases, the two programs remain linked through state computer systems, and Medicaid-eligible families continue to be denied or terminated from Medicaid based on welfare, rather than Medicaid, rules. In addition, families that may be diverted from receiving ongoing cash assistance or that do not end up receiving cash aid because of new welfare rules that promote quick attachment to the job market, are not always provided the opportunity to separately apply for Medicaid.
Moreover, families leaving welfare for work are not always receiving the Transitional Medicaid coverage ("TMA") to which they are entitled under federal law. TMA provides up to 12 months of coverage for families that would otherwise become ineligible for Medicaid due to earnings, and some states have federal waivers to provide TMA for longer periods. Although most states do not report separate data on TMA enrollment, the GAO recently examined TMA data from six states and found that Transitional Medicaid participation rates varied greatly, from only about four percent in Idaho to 94 percent in Connecticut.
In recent months, the federal government and a number of states have become increasingly concerned about the adverse impact of changes in welfare policy aid practices on children's enrollment in Medicaid. A number of states have begun to take steps to assure that welfare policy changes do not continue to cause families or children inappropriately to lose out on the Medicaid coverage for which they are eligible. For example, Pennsylvania, Washington, and Maryland are updating their computer systems, conducting staff training, and, in some cases, restoring Medicaid coverage to families that have been cut off Medicaid improperly in recent years. In addition, in August 1999, the Health Care Financing Administration launched a state-by-state review of states' policies for assuring that families receive the Medicaid coverage for which they are eligible when they are diverted from welfare or leave the welfare rolls.(10)
Many of the state-specific studies do not include data on the extent to which the children who miss out on Medicaid coverage are able to secure coverage through other sources. Those studies that examined this question have generally found that the children in families leaving welfare are unlikely to have coverage through other sources. Without Medicaid, these children are at high risk of becoming uninsured. For example, in Oklahoma, fewer than one in five children in families leaving welfare had private coverage. In Mississippi and South Carolina, only one in ten children in families leaving welfare had private coverage.
The new Census data on declining Medicaid enrollment underscore the need for the federal government and the states to take additional steps to assure that poor children are enrolled in Medicaid. States and the federal government need, in particular, to take steps to assure that changes in welfare policies do not result in children missing out on the coverage for which they are eligible. In the absence of such steps, the country is likely to continue to fail to make progress in reducing the rate at which low-income children lack health insurance coverage despite the strong economy, the new Children's Health Insurance Program, and the strong, bi-partisan commitment to enrolling more children in coverage.
Technical Note on the Medicaid Enrollment Data Available
from the Current Population Survey
There are two major sources of data on Medicaid enrollment trends the Census data from the Current Population Survey, which are discussed in this paper, and administrative data on Medicaid enrollment that states report to the Health Care Financing Administration. In addition, states typically keep administrative data on Medicaid enrollment trends that are not reported to HCFA, and, thus, are not consistently available for all 50 states. In recent years, all of these data sources have shown significant declines in the Medicaid enrollment nationwide of adults and children affected by welfare policy changes. Moreover, studies states have conducted on how families leaving welfare are faring, some of which are discussed in the text, also consistently show that the children in families that leave welfare frequently lose their Medicaid coverage. Although all of the available data sources suggest that welfare policy changes are leading children (and their parents) to lose out on Medicaid coverage for which they are eligible, the data sources do not all show Medicaid enrollment declines of the same magnitude. In general, the Current Population Survey data show the largest declines in Medicaid enrollment.
Current Population Survey Data on Medicaid Enrollment
The Current Population Survey data on Medicaid enrollment are gathered as part of a survey that the Census Bureau conducts each March. The Medicaid enrollment data from the CPS consistently show fewer people enrolled in Medicaid than the administrative data on program enrollment that the states provide to the Health Care Financing Administration. The disparity between the Census data and the HCFA administrative data has increased in recent years, creating the possibility that the CPS overstates the decline in Medicaid enrollment. HCFA has not yet released administrative data on Medicaid enrollment during 1998, making it impossible to determine whether the gap between CPS and HCFA administrative data on Medicaid enrollment continued to increase in 1998.
In particular, some analysts have questioned whether the increase in the number of poor children with "private" coverage (defined in this context as coverage from any source other than Medicaid, including employer-sponsored coverage) and the drop in Medicaid enrollment among poor children might actually reflect an increase in families mis-reporting their children's source of coverage. States' efforts to re-name their Medicaid programs, issue enrollment cards from managed care organizations, and otherwise make Medicaid look more like private coverage may have led an increasing number of families to report that their children are enrolled in private coverage when they are in fact enrolled in Medicaid.
If such misreporting has increased over time, however, it would explain away only a portion of the reported decline in the number of poor children on Medicaid. For example, even if the unrealistic assumption were made that all of the increase in the rate of private coverage among poor children that occurred between 1996 and 1998 reflected misreporting of children actually enrolled in Medicaid, the data still would indicate that there were nearly 900,000 fewer poor children on Medicaid in 1998 than in 1996. More than 40 percent of the decline in the number of poor children on Medicaid that is attributable to the decline in the Medicaid participation rate would remain.
HCFA's Administrative Data on Medicaid Enrollment
In recent years, the administrative data that HCFA gathers from the states on Medicaid enrollment also has shown drops in the number of adults and children enrolled in Medicaid, albeit less dramatic declines than those the CPS data show. For example, HCFA's administrative data, as compiled by the Urban Institute, show that Medicaid enrollment among non-disabled children and adults (the population most likely to be affected by changes in welfare policies) declined 1.6 million between fiscal year 1995 and fiscal year 1997. Among children, Medicaid enrollment declined by 600,000 with the number of children enrolled in cash-related Medicaid declining 2.2 million, and the number of children enrolled in non-cash related Medicaid increasing by 1.7 million.
In the past, it has been common to treat the HCFA administrative data as the benchmark against which the quality of the Medicaid enrollment data available from the CPS are evaluated. An increasing number of concerns about the quality of the HCFA administrative data are being raised, however, that call this strategy into question. For example, in a recent study on the effect of the welfare law on Medicaid enrollment, the General Accounting Office found numerous problems with the reliability of the HCFA administrative data. Among other things, GAO noted that the HCFA administrative data for 1997 mis-classified most of the adults and children enrolled in Medicaid in Louisiana as "elderly." The problems with the HCFA administrative data were sufficiently severe that GAO decided not to use those data in its study and instead conducted a 50-state survey to evaluate Medicaid enrollment trends. The administrative data GAO gathered from the states indicate that the Medicaid enrollment of adults and children fell by 7.4 percent between 1995 and 1997. In comparison, the administrative data gathered by HCFA show a drop of 5.1 percent and the CPS data show a drop of 12.3 percent. (Note that the data from the three sources are not directly comparable because they provide information on somewhat different populations and somewhat different time periods).
In sum, there is no single reliable source of data on Medicaid enrollment trends in all 50 states. Taken together, the data present a consistent picture that Medicaid enrollment among children is dropping, even though states are expanding Medicaid eligibility to more low-income children. Particularly given the growing awareness of the shortcomings of the HCFA administrative data, it may be appropriate to view the data available from the various sources as providing a range within which the actual decline in Medicaid enrollment falls, with the CPS data providing a high-end estimate of the decline.
1. Indeed, the portion of children, including poor children, with employer-based coverage appears to be increasing. Between 1996 and 1998, the portion of poor children with coverage from sources other than Medicaid, including employer-based coverage, increased from 14.1% 17.0%. At the same time, the participation rate of all low-income children in coverage sources other than Medicaid also increased from 34.0% in 1996 to 34.6% in 1998 (although this change was not statistically significant). The increase in employer-based coverage among children is a trend that is attracting increasing attention. See Paul Fronstin, "Employment-Based Health Insurance for Children: Why Did Coverage Increase in the Mid-1990's," Health Affairs, September/October 1999.
2. Under federal law, all states must provide Medicaid to children under the age of six with family income below 133 percent of the federal poverty line. States also are required to provide Medicaid to children born after September 30, 1983 with family income below 100 percent of the federal poverty line. Many states have expanded coverage beyond these federal minimum standards. As a result, all poor children under the age of 14 were eligible for Medicaid in 1998, and older children with family income below the poverty line were eligible for Medicaid in a majority of states. In addition, all members of families that might otherwise become ineligible for Medicaid due to earnings are eligible for six months of transitional Medicaid coverage, and they can qualify for an additional six months if their family income less child care expenses remains below 185 percent of the federal poverty line.
3. It is possible to break the decline in Medicaid enrollment among poor children down between the portion attributable to the decline in the total number of poor children and the portion attributable to the decline in the Medicaid participation rate among poor children. The portion of the decline in Medicaid enrollment among poor children attributable to the decline in the total number of poor children can be estimated by considering how many fewer children would have been enrolled in Medicaid if the Medicaid participation rate among poor children had remained unchanged. Since the Medicaid participation rate among poor children in 1996 was 62.6 percent, a decline of one million (997,000) in the total number of poor children would have resulted in 624,000 fewer poor children being enrolled in Medicaid. The remaining drop of 643,000 in the number of poor children on Medicaid is attributable to the drop in the Medicaid participation rate between 1996 and 1998 from 62.6 percent to 57.8 percent.
4. Based on data reported by the states to the Department of Health and Human Services.
5. Pamela Loprest, Families Who Left Welfare: Who Are They and How Are They Doing?, Urban Institute, 1999.
6. Marilyn R. Ellwood and Kimball Lewis, "On and Off Medicaid Enrollment: Patterns for California and Florida in 1995," Urban Institute, Occasional Paper Number 27, July 1999.
7. General Accounting Office, Medicaid Enrollment: Amid Declines, State Efforts to Ensure Coverage After Welfare Reform Vary, September 1999. Note that these figures are based on data collected by GAO directly from all states except Rhode Island and West Virginia.
8. Families USA, Losing Health Insurance: The Unintended Consequences of Welfare Reform, May 1999. The study estimated the impact of the welfare law after controlling for several factors, including changes in the economy and demographic changes. As noted in the methodology section of the study, some of the efforts made to control for the effect of economic changes on Medicaid enrollment may have led the study to understate the extent to which welfare policy changes caused families to enter the job market whereupon they became ineligible for coverage.
9. General Accounting Office, Welfare Reform: Information on Former Recipients' Status, April 1999. Similar findings emerged from the Center on Budget and Policy Priorities' review of welfare leavers studies; see Sharon Parrott, Welfare Recipients Who Find Jobs: What Do We Know About Employment and Earnings, Center on Budget and Policy Priorities, November 16, 1998.
10. For further discussion on the ways in which welfare changes may be affecting Medicaid enrollment, see Schott and Mann, Center on Budget and Policy Priorities, Assuring that Eligible Families Receive Medicaid When TANF Assistance is Denied or Terminated, November 1988; Mann, Center on Budget and Policy Priorities, The Ins and Outs of Delinking: Promoting Medicaid Enrollment of Children Who Are Moving In and Out of the TANF System, prepared for Covering Kids, March 1999.